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
246 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @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 -fno-zero-initialized-in-bss @gol
277 -fomit-frame-pointer -foptimize-register-move @gol
278 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
279 -freduce-all-givs -fregmove -frename-registers @gol
280 -frerun-cse-after-loop -frerun-loop-opt @gol
281 -fschedule-insns -fschedule-insns2 @gol
282 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
283 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol
284 -funroll-all-loops -funroll-loops @gol
285 --param @var{name}=@var{value}
286 -O -O0 -O1 -O2 -O3 -Os}
288 @item Preprocessor Options
289 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
291 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
292 -C -dD -dI -dM -dN @gol
293 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
294 -idirafter @var{dir} @gol
295 -include @var{file} -imacros @var{file} @gol
296 -iprefix @var{file} -iwithprefix @var{dir} @gol
297 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
298 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
299 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
301 @item Assembler Option
302 @xref{Assembler Options,,Passing Options to the Assembler}.
307 @xref{Link Options,,Options for Linking}.
309 @var{object-file-name} -l@var{library} @gol
310 -nostartfiles -nodefaultlibs -nostdlib @gol
311 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
312 -Wl,@var{option} -Xlinker @var{option} @gol
315 @item Directory Options
316 @xref{Directory Options,,Options for Directory Search}.
318 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
321 @c I wrote this xref this way to avoid overfull hbox. -- rms
322 @xref{Target Options}.
324 -b @var{machine} -V @var{version}}
326 @item Machine Dependent Options
327 @xref{Submodel Options,,Hardware Models and Configurations}.
329 @emph{M680x0 Options}
331 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
332 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
333 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
334 -malign-int -mstrict-align}
336 @emph{M68hc1x Options}
338 -m6811 -m6812 -m68hc11 -m68hc12 @gol
339 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
347 -mcpu=@var{cpu-type} @gol
348 -mtune=@var{cpu-type} @gol
349 -mcmodel=@var{code-model} @gol
351 -mapp-regs -mbroken-saverestore -mcypress @gol
352 -mepilogue -mfaster-structs -mflat @gol
353 -mfpu -mhard-float -mhard-quad-float @gol
354 -mimpure-text -mlive-g0 -mno-app-regs @gol
355 -mno-epilogue -mno-faster-structs -mno-flat -mno-fpu @gol
356 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
357 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
358 -msupersparc -munaligned-doubles -mv8}
360 @emph{Convex Options}
362 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
363 -margcount -mnoargcount @gol
364 -mlong32 -mlong64 @gol
365 -mvolatile-cache -mvolatile-nocache}
367 @emph{AMD29K Options}
369 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
370 -mlarge -mnormal -msmall @gol
371 -mkernel-registers -mno-reuse-arg-regs @gol
372 -mno-stack-check -mno-storem-bug @gol
373 -mreuse-arg-regs -msoft-float -mstack-check @gol
374 -mstorem-bug -muser-registers}
378 -mapcs-frame -mno-apcs-frame @gol
379 -mapcs-26 -mapcs-32 @gol
380 -mapcs-stack-check -mno-apcs-stack-check @gol
381 -mapcs-float -mno-apcs-float @gol
382 -mapcs-reentrant -mno-apcs-reentrant @gol
383 -msched-prolog -mno-sched-prolog @gol
384 -mlittle-endian -mbig-endian -mwords-little-endian @gol
385 -malignment-traps -mno-alignment-traps @gol
386 -msoft-float -mhard-float -mfpe @gol
387 -mthumb-interwork -mno-thumb-interwork @gol
388 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
389 -mstructure-size-boundary=@var{n} @gol
390 -mbsd -mxopen -mno-symrename @gol
391 -mabort-on-noreturn @gol
392 -mlong-calls -mno-long-calls @gol
393 -msingle-pic-base -mno-single-pic-base @gol
394 -mpic-register=@var{reg} @gol
395 -mnop-fun-dllimport @gol
396 -mpoke-function-name @gol
398 -mtpcs-frame -mtpcs-leaf-frame @gol
399 -mcaller-super-interworking -mcallee-super-interworking }
401 @emph{MN10200 Options}
405 @emph{MN10300 Options}
407 -mmult-bug -mno-mult-bug @gol
408 -mam33 -mno-am33 @gol
411 @emph{M32R/D Options}
413 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
418 -m88000 -m88100 -m88110 -mbig-pic @gol
419 -mcheck-zero-division -mhandle-large-shift @gol
420 -midentify-revision -mno-check-zero-division @gol
421 -mno-ocs-debug-info -mno-ocs-frame-position @gol
422 -mno-optimize-arg-area -mno-serialize-volatile @gol
423 -mno-underscores -mocs-debug-info @gol
424 -mocs-frame-position -moptimize-arg-area @gol
425 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
426 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
427 -mversion-03.00 -mwarn-passed-structs}
429 @emph{RS/6000 and PowerPC Options}
431 -mcpu=@var{cpu-type} @gol
432 -mtune=@var{cpu-type} @gol
433 -mpower -mno-power -mpower2 -mno-power2 @gol
434 -mpowerpc -mpowerpc64 -mno-powerpc @gol
435 -maltivec -mno-altivec @gol
436 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
437 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
438 -mnew-mnemonics -mold-mnemonics @gol
439 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
440 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
441 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
442 -mstring -mno-string -mupdate -mno-update @gol
443 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
444 -mstrict-align -mno-strict-align -mrelocatable @gol
445 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
446 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
447 -mcall-aix -mcall-sysv -mcall-netbsd @gol
448 -maix-struct-return -msvr4-struct-return
450 -mprototype -mno-prototype @gol
451 -msim -mmvme -mads -myellowknife -memb -msdata @gol
452 -msdata=@var{opt} -mvxworks -G @var{num} -pthread}
456 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
457 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
458 -mminimum-fp-blocks -mnohc-struct-return}
462 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
463 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
464 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
465 -mgas -mgp32 -mgp64 @gol
466 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
467 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
468 -mmips-as -mmips-tfile -mno-abicalls @gol
469 -mno-embedded-data -mno-uninit-const-in-rodata @gol
470 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
471 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
472 -mrnames -msoft-float @gol
473 -m4650 -msingle-float -mmad @gol
474 -mstats -EL -EB -G @var{num} -nocpp @gol
475 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
476 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
478 @emph{i386 and x86-64 Options}
480 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
481 -masm=@var{dialect} -mno-fancy-math-387 @gol
482 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
483 -mno-wide-multiply -mrtd -malign-double @gol
484 -mpreferred-stack-boundary=@var{num} @gol
485 -mmmx -msse -msse2 -msse-math -m3dnow @gol
486 -mthreads -mno-align-stringops -minline-all-stringops @gol
487 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
488 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
494 -march=@var{architecture-type} @gol
495 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
496 -mfast-indirect-calls -mgas -mjump-in-delay @gol
497 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
498 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
499 -mno-jump-in-delay -mno-long-load-store @gol
500 -mno-portable-runtime -mno-soft-float @gol
501 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
502 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
503 -mschedule=@var{cpu-type} -mspace-regs}
505 @emph{Intel 960 Options}
507 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
508 -mcode-align -mcomplex-addr -mleaf-procedures @gol
509 -mic-compat -mic2.0-compat -mic3.0-compat @gol
510 -mintel-asm -mno-clean-linkage -mno-code-align @gol
511 -mno-complex-addr -mno-leaf-procedures @gol
512 -mno-old-align -mno-strict-align -mno-tail-call @gol
513 -mnumerics -mold-align -msoft-float -mstrict-align @gol
516 @emph{DEC Alpha Options}
518 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
519 -mieee -mieee-with-inexact -mieee-conformant @gol
520 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
521 -mtrap-precision=@var{mode} -mbuild-constants @gol
522 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
523 -mbwx -mmax -mfix -mcix @gol
524 -mfloat-vax -mfloat-ieee @gol
525 -mexplicit-relocs -msmall-data -mlarge-data @gol
526 -mmemory-latency=@var{time}}
528 @emph{DEC Alpha/VMS Options}
532 @emph{Clipper Options}
536 @emph{H8/300 Options}
538 -mrelax -mh -ms -mint32 -malign-300}
542 -m1 -m2 -m3 -m3e @gol
543 -m4-nofpu -m4-single-only -m4-single -m4 @gol
544 -m5-64media -m5-64media-nofpu @gol
545 -m5-32media -m5-32media-nofpu @gol
546 -m5-compact -m5-compact-nofpu @gol
547 -mb -ml -mdalign -mrelax @gol
548 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
549 -mieee -misize -mpadstruct -mspace @gol
550 -mprefergot -musermode}
552 @emph{System V Options}
554 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
559 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
560 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
562 @emph{TMS320C3x/C4x Options}
564 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
565 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
566 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
567 -mparallel-insns -mparallel-mpy -mpreserve-float}
571 -mlong-calls -mno-long-calls -mep -mno-ep @gol
572 -mprolog-function -mno-prolog-function -mspace @gol
573 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
578 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
579 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
580 -mregparam -mnoregparam -msb -mnosb @gol
581 -mbitfield -mnobitfield -mhimem -mnohimem}
585 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
586 -mcall-prologues -mno-tablejump -mtiny-stack}
590 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
591 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
592 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
593 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
594 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
598 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
599 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
600 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
605 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
606 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
607 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
608 -minline-divide-max-throughput -mno-dwarf2-asm @gol
609 -mfixed-range=@var{register-range}}
613 -mextmem -mextmemory -monchip -mno-asm-optimize -masm-optimize @gol
614 -mbranch-cost=@var{n} -mcond-exec=@var{n}}
616 @emph{S/390 and zSeries Options}
618 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
619 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
620 -m64 -m31 -mdebug -mno-debug}
624 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
625 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
626 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
627 -mstack-align -mdata-align -mconst-align @gol
628 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
629 -melf -maout -melinux -mlinux -sim -sim2}
631 @emph{PDP-11 Options}
633 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
634 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
635 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
636 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
637 -mbranch-expensive -mbranch-cheap @gol
638 -msplit -mno-split -munix-asm -mdec-asm}
640 @emph{Xstormy16 Options}
644 @emph{Xtensa Options}
646 -mbig-endian -mlittle-endian @gol
647 -mdensity -mno-density @gol
648 -mmac16 -mno-mac16 @gol
649 -mmul16 -mno-mul16 @gol
650 -mmul32 -mno-mul32 @gol
652 -mminmax -mno-minmax @gol
653 -msext -mno-sext @gol
654 -mbooleans -mno-booleans @gol
655 -mhard-float -msoft-float @gol
656 -mfused-madd -mno-fused-madd @gol
657 -mserialize-volatile -mno-serialize-volatile @gol
658 -mtext-section-literals -mno-text-section-literals @gol
659 -mtarget-align -mno-target-align @gol
660 -mlongcalls -mno-longcalls}
662 @item Code Generation Options
663 @xref{Code Gen Options,,Options for Code Generation Conventions}.
665 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
666 -ffixed-@var{reg} -fexceptions @gol
667 -fnon-call-exceptions -funwind-tables @gol
668 -fasynchronous-unwind-tables @gol
669 -finhibit-size-directive -finstrument-functions @gol
670 -fno-common -fno-ident -fno-gnu-linker @gol
671 -fpcc-struct-return -fpic -fPIC @gol
672 -freg-struct-return -fshared-data -fshort-enums @gol
673 -fshort-double -fvolatile @gol
674 -fvolatile-global -fvolatile-static @gol
675 -fverbose-asm -fpack-struct -fstack-check @gol
676 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
677 -fargument-alias -fargument-noalias @gol
678 -fargument-noalias-global -fleading-underscore}
682 * Overall Options:: Controlling the kind of output:
683 an executable, object files, assembler files,
684 or preprocessed source.
685 * C Dialect Options:: Controlling the variant of C language compiled.
686 * C++ Dialect Options:: Variations on C++.
687 * Objective-C Dialect Options:: Variations on Objective-C.
688 * Language Independent Options:: Controlling how diagnostics should be
690 * Warning Options:: How picky should the compiler be?
691 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
692 * Optimize Options:: How much optimization?
693 * Preprocessor Options:: Controlling header files and macro definitions.
694 Also, getting dependency information for Make.
695 * Assembler Options:: Passing options to the assembler.
696 * Link Options:: Specifying libraries and so on.
697 * Directory Options:: Where to find header files and libraries.
698 Where to find the compiler executable files.
699 * Spec Files:: How to pass switches to sub-processes.
700 * Target Options:: Running a cross-compiler, or an old version of GCC.
703 @node Overall Options
704 @section Options Controlling the Kind of Output
706 Compilation can involve up to four stages: preprocessing, compilation
707 proper, assembly and linking, always in that order. The first three
708 stages apply to an individual source file, and end by producing an
709 object file; linking combines all the object files (those newly
710 compiled, and those specified as input) into an executable file.
712 @cindex file name suffix
713 For any given input file, the file name suffix determines what kind of
718 C source code which must be preprocessed.
721 C source code which should not be preprocessed.
724 C++ source code which should not be preprocessed.
727 Objective-C source code. Note that you must link with the library
728 @file{libobjc.a} to make an Objective-C program work.
731 Objective-C source code which should not be preprocessed.
734 C header file (not to be compiled or linked).
738 @itemx @var{file}.cxx
739 @itemx @var{file}.cpp
740 @itemx @var{file}.c++
742 C++ source code which must be preprocessed. Note that in @samp{.cxx},
743 the last two letters must both be literally @samp{x}. Likewise,
744 @samp{.C} refers to a literal capital C@.
747 @itemx @var{file}.for
748 @itemx @var{file}.FOR
749 Fortran source code which should not be preprocessed.
752 @itemx @var{file}.fpp
753 @itemx @var{file}.FPP
754 Fortran source code which must be preprocessed (with the traditional
758 Fortran source code which must be preprocessed with a RATFOR
759 preprocessor (not included with GCC)@.
761 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
762 Using and Porting GNU Fortran}, for more details of the handling of
765 @c FIXME: Descriptions of Java file types.
772 Ada source code file which contains a library unit declaration (a
773 declaration of a package, subprogram, or generic, or a generic
774 instantiation), or a library unit renaming declaration (a package,
775 generic, or subprogram renaming declaration). Such files are also
778 @itemx @var{file}.adb
779 Ada source code file containing a library unit body (a subprogram or
780 package body). Such files are also called @dfn{bodies}.
782 @c GCC also knows about some suffixes for languages not yet included:
788 @itemx @var{file}.chi
789 CHILL source code (preprocessed with the traditional preprocessor).
795 Assembler code which must be preprocessed.
798 An object file to be fed straight into linking.
799 Any file name with no recognized suffix is treated this way.
803 You can specify the input language explicitly with the @option{-x} option:
806 @item -x @var{language}
807 Specify explicitly the @var{language} for the following input files
808 (rather than letting the compiler choose a default based on the file
809 name suffix). This option applies to all following input files until
810 the next @option{-x} option. Possible values for @var{language} are:
812 c c-header cpp-output
814 objective-c objc-cpp-output
815 assembler assembler-with-cpp
818 f77 f77-cpp-input ratfor
823 Turn off any specification of a language, so that subsequent files are
824 handled according to their file name suffixes (as they are if @option{-x}
825 has not been used at all).
827 @item -pass-exit-codes
828 @opindex pass-exit-codes
829 Normally the @command{gcc} program will exit with the code of 1 if any
830 phase of the compiler returns a non-success return code. If you specify
831 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
832 numerically highest error produced by any phase that returned an error
836 If you only want some of the stages of compilation, you can use
837 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
838 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
839 @command{gcc} is to stop. Note that some combinations (for example,
840 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
845 Compile or assemble the source files, but do not link. The linking
846 stage simply is not done. The ultimate output is in the form of an
847 object file for each source file.
849 By default, the object file name for a source file is made by replacing
850 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
852 Unrecognized input files, not requiring compilation or assembly, are
857 Stop after the stage of compilation proper; do not assemble. The output
858 is in the form of an assembler code file for each non-assembler input
861 By default, the assembler file name for a source file is made by
862 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
864 Input files that don't require compilation are ignored.
868 Stop after the preprocessing stage; do not run the compiler proper. The
869 output is in the form of preprocessed source code, which is sent to the
872 Input files which don't require preprocessing are ignored.
874 @cindex output file option
877 Place output in file @var{file}. This applies regardless to whatever
878 sort of output is being produced, whether it be an executable file,
879 an object file, an assembler file or preprocessed C code.
881 Since only one output file can be specified, it does not make sense to
882 use @option{-o} when compiling more than one input file, unless you are
883 producing an executable file as output.
885 If @option{-o} is not specified, the default is to put an executable file
886 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
887 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
888 all preprocessed C source on standard output.
892 Print (on standard error output) the commands executed to run the stages
893 of compilation. Also print the version number of the compiler driver
894 program and of the preprocessor and the compiler proper.
898 Use pipes rather than temporary files for communication between the
899 various stages of compilation. This fails to work on some systems where
900 the assembler is unable to read from a pipe; but the GNU assembler has
905 Print (on the standard output) a description of the command line options
906 understood by @command{gcc}. If the @option{-v} option is also specified
907 then @option{--help} will also be passed on to the various processes
908 invoked by @command{gcc}, so that they can display the command line options
909 they accept. If the @option{-W} option is also specified then command
910 line options which have no documentation associated with them will also
915 Print (on the standard output) a description of target specific command
916 line options for each tool.
920 @section Compiling C++ Programs
922 @cindex suffixes for C++ source
923 @cindex C++ source file suffixes
924 C++ source files conventionally use one of the suffixes @samp{.C},
925 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
926 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
927 files with these names and compiles them as C++ programs even if you
928 call the compiler the same way as for compiling C programs (usually with
929 the name @command{gcc}).
933 However, C++ programs often require class libraries as well as a
934 compiler that understands the C++ language---and under some
935 circumstances, you might want to compile programs from standard input,
936 or otherwise without a suffix that flags them as C++ programs.
937 @command{g++} is a program that calls GCC with the default language
938 set to C++, and automatically specifies linking against the C++
939 library. On many systems, @command{g++} is also
940 installed with the name @command{c++}.
942 @cindex invoking @command{g++}
943 When you compile C++ programs, you may specify many of the same
944 command-line options that you use for compiling programs in any
945 language; or command-line options meaningful for C and related
946 languages; or options that are meaningful only for C++ programs.
947 @xref{C Dialect Options,,Options Controlling C Dialect}, for
948 explanations of options for languages related to C@.
949 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
950 explanations of options that are meaningful only for C++ programs.
952 @node C Dialect Options
953 @section Options Controlling C Dialect
954 @cindex dialect options
955 @cindex language dialect options
956 @cindex options, dialect
958 The following options control the dialect of C (or languages derived
959 from C, such as C++ and Objective-C) that the compiler accepts:
966 In C mode, support all ISO C89 programs. In C++ mode,
967 remove GNU extensions that conflict with ISO C++.
969 This turns off certain features of GCC that are incompatible with ISO
970 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
971 such as the @code{asm} and @code{typeof} keywords, and
972 predefined macros such as @code{unix} and @code{vax} that identify the
973 type of system you are using. It also enables the undesirable and
974 rarely used ISO trigraph feature. For the C compiler,
975 it disables recognition of C++ style @samp{//} comments as well as
976 the @code{inline} keyword.
978 The alternate keywords @code{__asm__}, @code{__extension__},
979 @code{__inline__} and @code{__typeof__} continue to work despite
980 @option{-ansi}. You would not want to use them in an ISO C program, of
981 course, but it is useful to put them in header files that might be included
982 in compilations done with @option{-ansi}. Alternate predefined macros
983 such as @code{__unix__} and @code{__vax__} are also available, with or
984 without @option{-ansi}.
986 The @option{-ansi} option does not cause non-ISO programs to be
987 rejected gratuitously. For that, @option{-pedantic} is required in
988 addition to @option{-ansi}. @xref{Warning Options}.
990 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
991 option is used. Some header files may notice this macro and refrain
992 from declaring certain functions or defining certain macros that the
993 ISO standard doesn't call for; this is to avoid interfering with any
994 programs that might use these names for other things.
996 Functions which would normally be built in but do not have semantics
997 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
998 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
999 built-in functions provided by GCC}, for details of the functions
1004 Determine the language standard. This option is currently only
1005 supported when compiling C@. A value for this option must be provided;
1011 ISO C89 (same as @option{-ansi}).
1013 @item iso9899:199409
1014 ISO C89 as modified in amendment 1.
1020 ISO C99. Note that this standard is not yet fully supported; see
1021 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1022 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1025 Default, ISO C89 plus GNU extensions (including some C99 features).
1029 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1030 this will become the default. The name @samp{gnu9x} is deprecated.
1034 Even when this option is not specified, you can still use some of the
1035 features of newer standards in so far as they do not conflict with
1036 previous C standards. For example, you may use @code{__restrict__} even
1037 when @option{-std=c99} is not specified.
1039 The @option{-std} options specifying some version of ISO C have the same
1040 effects as @option{-ansi}, except that features that were not in ISO C89
1041 but are in the specified version (for example, @samp{//} comments and
1042 the @code{inline} keyword in ISO C99) are not disabled.
1044 @xref{Standards,,Language Standards Supported by GCC}, for details of
1045 these standard versions.
1047 @item -aux-info @var{filename}
1049 Output to the given filename prototyped declarations for all functions
1050 declared and/or defined in a translation unit, including those in header
1051 files. This option is silently ignored in any language other than C@.
1053 Besides declarations, the file indicates, in comments, the origin of
1054 each declaration (source file and line), whether the declaration was
1055 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1056 @samp{O} for old, respectively, in the first character after the line
1057 number and the colon), and whether it came from a declaration or a
1058 definition (@samp{C} or @samp{F}, respectively, in the following
1059 character). In the case of function definitions, a K&R-style list of
1060 arguments followed by their declarations is also provided, inside
1061 comments, after the declaration.
1065 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1066 keyword, so that code can use these words as identifiers. You can use
1067 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1068 instead. @option{-ansi} implies @option{-fno-asm}.
1070 In C++, this switch only affects the @code{typeof} keyword, since
1071 @code{asm} and @code{inline} are standard keywords. You may want to
1072 use the @option{-fno-gnu-keywords} flag instead, which has the same
1073 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1074 switch only affects the @code{asm} and @code{typeof} keywords, since
1075 @code{inline} is a standard keyword in ISO C99.
1078 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1079 @opindex fno-builtin
1080 @cindex built-in functions
1081 Don't recognize built-in functions that do not begin with
1082 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1083 functions provided by GCC}, for details of the functions affected,
1084 including those which are not built-in functions when @option{-ansi} or
1085 @option{-std} options for strict ISO C conformance are used because they
1086 do not have an ISO standard meaning.
1088 GCC normally generates special code to handle certain built-in functions
1089 more efficiently; for instance, calls to @code{alloca} may become single
1090 instructions that adjust the stack directly, and calls to @code{memcpy}
1091 may become inline copy loops. The resulting code is often both smaller
1092 and faster, but since the function calls no longer appear as such, you
1093 cannot set a breakpoint on those calls, nor can you change the behavior
1094 of the functions by linking with a different library.
1096 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1097 option has no effect. Therefore, in C++, the only way to get the
1098 optimization benefits of built-in functions is to call the function
1099 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1100 built-in functions to implement many functions (like
1101 @code{std::strchr}), so that you automatically get efficient code.
1103 With the @option{-fno-builtin-@var{function}} option, not available
1104 when compiling C++, only the built-in function @var{function} is
1105 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1106 function is named this is not built-in in this version of GCC, this
1107 option is ignored. There is no corresponding
1108 @option{-fbuiltin-@var{function}} option; if you wish to enable
1109 built-in functions selectively when using @option{-fno-builtin} or
1110 @option{-ffreestanding}, you may define macros such as:
1113 #define abs(n) __builtin_abs ((n))
1114 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1119 @cindex hosted environment
1121 Assert that compilation takes place in a hosted environment. This implies
1122 @option{-fbuiltin}. A hosted environment is one in which the
1123 entire standard library is available, and in which @code{main} has a return
1124 type of @code{int}. Examples are nearly everything except a kernel.
1125 This is equivalent to @option{-fno-freestanding}.
1127 @item -ffreestanding
1128 @opindex ffreestanding
1129 @cindex hosted environment
1131 Assert that compilation takes place in a freestanding environment. This
1132 implies @option{-fno-builtin}. A freestanding environment
1133 is one in which the standard library may not exist, and program startup may
1134 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1135 This is equivalent to @option{-fno-hosted}.
1137 @xref{Standards,,Language Standards Supported by GCC}, for details of
1138 freestanding and hosted environments.
1142 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1143 options for strict ISO C conformance) implies @option{-trigraphs}.
1145 @cindex traditional C language
1146 @cindex C language, traditional
1148 @itemx -traditional-cpp
1149 @opindex traditional-cpp
1150 @opindex traditional
1151 Formerly, these options caused GCC to attempt to emulate a pre-standard
1152 C compiler. They are now only supported with the @option{-E} switch.
1153 The preprocessor continues to support a pre-standard mode. See the GNU
1154 CPP manual for details.
1156 @item -fcond-mismatch
1157 @opindex fcond-mismatch
1158 Allow conditional expressions with mismatched types in the second and
1159 third arguments. The value of such an expression is void. This option
1160 is not supported for C++.
1162 @item -funsigned-char
1163 @opindex funsigned-char
1164 Let the type @code{char} be unsigned, like @code{unsigned char}.
1166 Each kind of machine has a default for what @code{char} should
1167 be. It is either like @code{unsigned char} by default or like
1168 @code{signed char} by default.
1170 Ideally, a portable program should always use @code{signed char} or
1171 @code{unsigned char} when it depends on the signedness of an object.
1172 But many programs have been written to use plain @code{char} and
1173 expect it to be signed, or expect it to be unsigned, depending on the
1174 machines they were written for. This option, and its inverse, let you
1175 make such a program work with the opposite default.
1177 The type @code{char} is always a distinct type from each of
1178 @code{signed char} or @code{unsigned char}, even though its behavior
1179 is always just like one of those two.
1182 @opindex fsigned-char
1183 Let the type @code{char} be signed, like @code{signed char}.
1185 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1186 the negative form of @option{-funsigned-char}. Likewise, the option
1187 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1189 @item -fsigned-bitfields
1190 @itemx -funsigned-bitfields
1191 @itemx -fno-signed-bitfields
1192 @itemx -fno-unsigned-bitfields
1193 @opindex fsigned-bitfields
1194 @opindex funsigned-bitfields
1195 @opindex fno-signed-bitfields
1196 @opindex fno-unsigned-bitfields
1197 These options control whether a bit-field is signed or unsigned, when the
1198 declaration does not use either @code{signed} or @code{unsigned}. By
1199 default, such a bit-field is signed, because this is consistent: the
1200 basic integer types such as @code{int} are signed types.
1202 @item -fwritable-strings
1203 @opindex fwritable-strings
1204 Store string constants in the writable data segment and don't uniquize
1205 them. This is for compatibility with old programs which assume they can
1206 write into string constants.
1208 Writing into string constants is a very bad idea; ``constants'' should
1212 @opindex fshort-wchar
1213 Override the underlying type for @samp{wchar_t} to be @samp{short
1214 unsigned int} instead of the default for the target. This option is
1215 useful for building programs to run under WINE@.
1218 @node C++ Dialect Options
1219 @section Options Controlling C++ Dialect
1221 @cindex compiler options, C++
1222 @cindex C++ options, command line
1223 @cindex options, C++
1224 This section describes the command-line options that are only meaningful
1225 for C++ programs; but you can also use most of the GNU compiler options
1226 regardless of what language your program is in. For example, you
1227 might compile a file @code{firstClass.C} like this:
1230 g++ -g -frepo -O -c firstClass.C
1234 In this example, only @option{-frepo} is an option meant
1235 only for C++ programs; you can use the other options with any
1236 language supported by GCC@.
1238 Here is a list of options that are @emph{only} for compiling C++ programs:
1241 @item -fno-access-control
1242 @opindex fno-access-control
1243 Turn off all access checking. This switch is mainly useful for working
1244 around bugs in the access control code.
1248 Check that the pointer returned by @code{operator new} is non-null
1249 before attempting to modify the storage allocated. The current Working
1250 Paper requires that @code{operator new} never return a null pointer, so
1251 this check is normally unnecessary.
1253 An alternative to using this option is to specify that your
1254 @code{operator new} does not throw any exceptions; if you declare it
1255 @samp{throw()}, G++ will check the return value. See also @samp{new
1258 @item -fconserve-space
1259 @opindex fconserve-space
1260 Put uninitialized or runtime-initialized global variables into the
1261 common segment, as C does. This saves space in the executable at the
1262 cost of not diagnosing duplicate definitions. If you compile with this
1263 flag and your program mysteriously crashes after @code{main()} has
1264 completed, you may have an object that is being destroyed twice because
1265 two definitions were merged.
1267 This option is no longer useful on most targets, now that support has
1268 been added for putting variables into BSS without making them common.
1270 @item -fno-const-strings
1271 @opindex fno-const-strings
1272 Give string constants type @code{char *} instead of type @code{const
1273 char *}. By default, G++ uses type @code{const char *} as required by
1274 the standard. Even if you use @option{-fno-const-strings}, you cannot
1275 actually modify the value of a string constant, unless you also use
1276 @option{-fwritable-strings}.
1278 This option might be removed in a future release of G++. For maximum
1279 portability, you should structure your code so that it works with
1280 string constants that have type @code{const char *}.
1282 @item -fdollars-in-identifiers
1283 @opindex fdollars-in-identifiers
1284 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1285 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1286 @samp{$} by default on most target systems, but there are a few exceptions.)
1287 Traditional C allowed the character @samp{$} to form part of
1288 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1290 @item -fno-elide-constructors
1291 @opindex fno-elide-constructors
1292 The C++ standard allows an implementation to omit creating a temporary
1293 which is only used to initialize another object of the same type.
1294 Specifying this option disables that optimization, and forces G++ to
1295 call the copy constructor in all cases.
1297 @item -fno-enforce-eh-specs
1298 @opindex fno-enforce-eh-specs
1299 Don't check for violation of exception specifications at runtime. This
1300 option violates the C++ standard, but may be useful for reducing code
1301 size in production builds, much like defining @samp{NDEBUG}. The compiler
1302 will still optimize based on the exception specifications.
1304 @item -fexternal-templates
1305 @opindex fexternal-templates
1307 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1308 template instantiation; template instances are emitted or not according
1309 to the location of the template definition. @xref{Template
1310 Instantiation}, for more information.
1312 This option is deprecated.
1314 @item -falt-external-templates
1315 @opindex falt-external-templates
1316 Similar to @option{-fexternal-templates}, but template instances are
1317 emitted or not according to the place where they are first instantiated.
1318 @xref{Template Instantiation}, for more information.
1320 This option is deprecated.
1323 @itemx -fno-for-scope
1325 @opindex fno-for-scope
1326 If @option{-ffor-scope} is specified, the scope of variables declared in
1327 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1328 as specified by the C++ standard.
1329 If @option{-fno-for-scope} is specified, the scope of variables declared in
1330 a @i{for-init-statement} extends to the end of the enclosing scope,
1331 as was the case in old versions of G++, and other (traditional)
1332 implementations of C++.
1334 The default if neither flag is given to follow the standard,
1335 but to allow and give a warning for old-style code that would
1336 otherwise be invalid, or have different behavior.
1338 @item -fno-gnu-keywords
1339 @opindex fno-gnu-keywords
1340 Do not recognize @code{typeof} as a keyword, so that code can use this
1341 word as an identifier. You can use the keyword @code{__typeof__} instead.
1342 @option{-ansi} implies @option{-fno-gnu-keywords}.
1344 @item -fno-implicit-templates
1345 @opindex fno-implicit-templates
1346 Never emit code for non-inline templates which are instantiated
1347 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1348 @xref{Template Instantiation}, for more information.
1350 @item -fno-implicit-inline-templates
1351 @opindex fno-implicit-inline-templates
1352 Don't emit code for implicit instantiations of inline templates, either.
1353 The default is to handle inlines differently so that compiles with and
1354 without optimization will need the same set of explicit instantiations.
1356 @item -fno-implement-inlines
1357 @opindex fno-implement-inlines
1358 To save space, do not emit out-of-line copies of inline functions
1359 controlled by @samp{#pragma implementation}. This will cause linker
1360 errors if these functions are not inlined everywhere they are called.
1362 @item -fms-extensions
1363 @opindex fms-extensions
1364 Disable pedantic warnings about constructs used in MFC, such as implicit
1365 int and getting a pointer to member function via non-standard syntax.
1367 @item -fno-nonansi-builtins
1368 @opindex fno-nonansi-builtins
1369 Disable built-in declarations of functions that are not mandated by
1370 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1371 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1373 @item -fno-operator-names
1374 @opindex fno-operator-names
1375 Do not treat the operator name keywords @code{and}, @code{bitand},
1376 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1377 synonyms as keywords.
1379 @item -fno-optional-diags
1380 @opindex fno-optional-diags
1381 Disable diagnostics that the standard says a compiler does not need to
1382 issue. Currently, the only such diagnostic issued by G++ is the one for
1383 a name having multiple meanings within a class.
1386 @opindex fpermissive
1387 Downgrade messages about nonconformant code from errors to warnings. By
1388 default, G++ effectively sets @option{-pedantic-errors} without
1389 @option{-pedantic}; this option reverses that. This behavior and this
1390 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1394 Enable automatic template instantiation at link time. This option also
1395 implies @option{-fno-implicit-templates}. @xref{Template
1396 Instantiation}, for more information.
1400 Disable generation of information about every class with virtual
1401 functions for use by the C++ runtime type identification features
1402 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1403 of the language, you can save some space by using this flag. Note that
1404 exception handling uses the same information, but it will generate it as
1409 Emit statistics about front-end processing at the end of the compilation.
1410 This information is generally only useful to the G++ development team.
1412 @item -ftemplate-depth-@var{n}
1413 @opindex ftemplate-depth
1414 Set the maximum instantiation depth for template classes to @var{n}.
1415 A limit on the template instantiation depth is needed to detect
1416 endless recursions during template class instantiation. ANSI/ISO C++
1417 conforming programs must not rely on a maximum depth greater than 17.
1419 @item -fuse-cxa-atexit
1420 @opindex fuse-cxa-atexit
1421 Register destructors for objects with static storage duration with the
1422 @code{__cxa_atexit} function rather than the @code{atexit} function.
1423 This option is required for fully standards-compliant handling of static
1424 destructors, but will only work if your C library supports
1425 @code{__cxa_atexit}.
1429 Emit special relocations for vtables and virtual function references
1430 so that the linker can identify unused virtual functions and zero out
1431 vtable slots that refer to them. This is most useful with
1432 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1433 also discard the functions themselves.
1435 This optimization requires GNU as and GNU ld. Not all systems support
1436 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1440 Do not use weak symbol support, even if it is provided by the linker.
1441 By default, G++ will use weak symbols if they are available. This
1442 option exists only for testing, and should not be used by end-users;
1443 it will result in inferior code and has no benefits. This option may
1444 be removed in a future release of G++.
1448 Do not search for header files in the standard directories specific to
1449 C++, but do still search the other standard directories. (This option
1450 is used when building the C++ library.)
1453 In addition, these optimization, warning, and code generation options
1454 have meanings only for C++ programs:
1457 @item -fno-default-inline
1458 @opindex fno-default-inline
1459 Do not assume @samp{inline} for functions defined inside a class scope.
1460 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1461 functions will have linkage like inline functions; they just won't be
1464 @item -Wctor-dtor-privacy @r{(C++ only)}
1465 @opindex Wctor-dtor-privacy
1466 Warn when a class seems unusable, because all the constructors or
1467 destructors in a class are private and the class has no friends or
1468 public static member functions.
1470 @item -Wnon-virtual-dtor @r{(C++ only)}
1471 @opindex Wnon-virtual-dtor
1472 Warn when a class declares a non-virtual destructor that should probably
1473 be virtual, because it looks like the class will be used polymorphically.
1475 @item -Wreorder @r{(C++ only)}
1477 @cindex reordering, warning
1478 @cindex warning for reordering of member initializers
1479 Warn when the order of member initializers given in the code does not
1480 match the order in which they must be executed. For instance:
1486 A(): j (0), i (1) @{ @}
1490 Here the compiler will warn that the member initializers for @samp{i}
1491 and @samp{j} will be rearranged to match the declaration order of the
1495 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1498 @item -Weffc++ @r{(C++ only)}
1500 Warn about violations of the following style guidelines from Scott Meyers'
1501 @cite{Effective C++} book:
1505 Item 11: Define a copy constructor and an assignment operator for classes
1506 with dynamically allocated memory.
1509 Item 12: Prefer initialization to assignment in constructors.
1512 Item 14: Make destructors virtual in base classes.
1515 Item 15: Have @code{operator=} return a reference to @code{*this}.
1518 Item 23: Don't try to return a reference when you must return an object.
1522 and about violations of the following style guidelines from Scott Meyers'
1523 @cite{More Effective C++} book:
1527 Item 6: Distinguish between prefix and postfix forms of increment and
1528 decrement operators.
1531 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1535 If you use this option, you should be aware that the standard library
1536 headers do not obey all of these guidelines; you can use @samp{grep -v}
1537 to filter out those warnings.
1539 @item -Wno-deprecated @r{(C++ only)}
1540 @opindex Wno-deprecated
1541 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1543 @item -Wno-non-template-friend @r{(C++ only)}
1544 @opindex Wno-non-template-friend
1545 Disable warnings when non-templatized friend functions are declared
1546 within a template. With the advent of explicit template specification
1547 support in G++, if the name of the friend is an unqualified-id (i.e.,
1548 @samp{friend foo(int)}), the C++ language specification demands that the
1549 friend declare or define an ordinary, nontemplate function. (Section
1550 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1551 could be interpreted as a particular specialization of a templatized
1552 function. Because this non-conforming behavior is no longer the default
1553 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1554 check existing code for potential trouble spots, and is on by default.
1555 This new compiler behavior can be turned off with
1556 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1557 but disables the helpful warning.
1559 @item -Wold-style-cast @r{(C++ only)}
1560 @opindex Wold-style-cast
1561 Warn if an old-style (C-style) cast to a non-void type is used within
1562 a C++ program. The new-style casts (@samp{static_cast},
1563 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1564 unintended effects, and much easier to grep for.
1566 @item -Woverloaded-virtual @r{(C++ only)}
1567 @opindex Woverloaded-virtual
1568 @cindex overloaded virtual fn, warning
1569 @cindex warning for overloaded virtual fn
1570 Warn when a function declaration hides virtual functions from a
1571 base class. For example, in:
1578 struct B: public A @{
1583 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1591 will fail to compile.
1593 @item -Wno-pmf-conversions @r{(C++ only)}
1594 @opindex Wno-pmf-conversions
1595 Disable the diagnostic for converting a bound pointer to member function
1598 @item -Wsign-promo @r{(C++ only)}
1599 @opindex Wsign-promo
1600 Warn when overload resolution chooses a promotion from unsigned or
1601 enumeral type to a signed type over a conversion to an unsigned type of
1602 the same size. Previous versions of G++ would try to preserve
1603 unsignedness, but the standard mandates the current behavior.
1605 @item -Wsynth @r{(C++ only)}
1607 @cindex warning for synthesized methods
1608 @cindex synthesized methods, warning
1609 Warn when G++'s synthesis behavior does not match that of cfront. For
1615 A& operator = (int);
1625 In this example, G++ will synthesize a default @samp{A& operator =
1626 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1629 @node Objective-C Dialect Options
1630 @section Options Controlling Objective-C Dialect
1632 @cindex compiler options, Objective-C
1633 @cindex Objective-C options, command line
1634 @cindex options, Objective-C
1635 This section describes the command-line options that are only meaningful
1636 for Objective-C programs; but you can also use most of the GNU compiler
1637 options regardless of what language your program is in. For example,
1638 you might compile a file @code{some_class.m} like this:
1641 gcc -g -fgnu-runtime -O -c some_class.m
1645 In this example, only @option{-fgnu-runtime} is an option meant only for
1646 Objective-C programs; you can use the other options with any language
1649 Here is a list of options that are @emph{only} for compiling Objective-C
1653 @item -fconstant-string-class=@var{class-name}
1654 @opindex fconstant-string-class
1655 Use @var{class-name} as the name of the class to instantiate for each
1656 literal string specified with the syntax @code{@@"@dots{}"}. The default
1657 class name is @code{NXConstantString}.
1660 @opindex fgnu-runtime
1661 Generate object code compatible with the standard GNU Objective-C
1662 runtime. This is the default for most types of systems.
1664 @item -fnext-runtime
1665 @opindex fnext-runtime
1666 Generate output compatible with the NeXT runtime. This is the default
1667 for NeXT-based systems, including Darwin and Mac OS X@.
1671 Dump interface declarations for all classes seen in the source file to a
1672 file named @file{@var{sourcename}.decl}.
1675 @opindex Wno-protocol
1676 Do not warn if methods required by a protocol are not implemented
1677 in the class adopting it.
1681 Warn if a selector has multiple methods of different types defined.
1683 @c not documented because only avail via -Wp
1684 @c @item -print-objc-runtime-info
1688 @node Language Independent Options
1689 @section Options to Control Diagnostic Messages Formatting
1690 @cindex options to control diagnostics formatting
1691 @cindex diagnostic messages
1692 @cindex message formatting
1694 Traditionally, diagnostic messages have been formatted irrespective of
1695 the output device's aspect (e.g.@: its width, @dots{}). The options described
1696 below can be used to control the diagnostic messages formatting
1697 algorithm, e.g.@: how many characters per line, how often source location
1698 information should be reported. Right now, only the C++ front end can
1699 honor these options. However it is expected, in the near future, that
1700 the remaining front ends would be able to digest them correctly.
1703 @item -fmessage-length=@var{n}
1704 @opindex fmessage-length
1705 Try to format error messages so that they fit on lines of about @var{n}
1706 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1707 the front ends supported by GCC@. If @var{n} is zero, then no
1708 line-wrapping will be done; each error message will appear on a single
1711 @opindex fdiagnostics-show-location
1712 @item -fdiagnostics-show-location=once
1713 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1714 reporter to emit @emph{once} source location information; that is, in
1715 case the message is too long to fit on a single physical line and has to
1716 be wrapped, the source location won't be emitted (as prefix) again,
1717 over and over, in subsequent continuation lines. This is the default
1720 @item -fdiagnostics-show-location=every-line
1721 Only meaningful in line-wrapping mode. Instructs the diagnostic
1722 messages reporter to emit the same source location information (as
1723 prefix) for physical lines that result from the process of breaking
1724 a message which is too long to fit on a single line.
1728 @node Warning Options
1729 @section Options to Request or Suppress Warnings
1730 @cindex options to control warnings
1731 @cindex warning messages
1732 @cindex messages, warning
1733 @cindex suppressing warnings
1735 Warnings are diagnostic messages that report constructions which
1736 are not inherently erroneous but which are risky or suggest there
1737 may have been an error.
1739 You can request many specific warnings with options beginning @samp{-W},
1740 for example @option{-Wimplicit} to request warnings on implicit
1741 declarations. Each of these specific warning options also has a
1742 negative form beginning @samp{-Wno-} to turn off warnings;
1743 for example, @option{-Wno-implicit}. This manual lists only one of the
1744 two forms, whichever is not the default.
1746 These options control the amount and kinds of warnings produced by GCC:
1749 @cindex syntax checking
1751 @opindex fsyntax-only
1752 Check the code for syntax errors, but don't do anything beyond that.
1756 Issue all the warnings demanded by strict ISO C and ISO C++;
1757 reject all programs that use forbidden extensions, and some other
1758 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1759 version of the ISO C standard specified by any @option{-std} option used.
1761 Valid ISO C and ISO C++ programs should compile properly with or without
1762 this option (though a rare few will require @option{-ansi} or a
1763 @option{-std} option specifying the required version of ISO C)@. However,
1764 without this option, certain GNU extensions and traditional C and C++
1765 features are supported as well. With this option, they are rejected.
1767 @option{-pedantic} does not cause warning messages for use of the
1768 alternate keywords whose names begin and end with @samp{__}. Pedantic
1769 warnings are also disabled in the expression that follows
1770 @code{__extension__}. However, only system header files should use
1771 these escape routes; application programs should avoid them.
1772 @xref{Alternate Keywords}.
1774 Some users try to use @option{-pedantic} to check programs for strict ISO
1775 C conformance. They soon find that it does not do quite what they want:
1776 it finds some non-ISO practices, but not all---only those for which
1777 ISO C @emph{requires} a diagnostic, and some others for which
1778 diagnostics have been added.
1780 A feature to report any failure to conform to ISO C might be useful in
1781 some instances, but would require considerable additional work and would
1782 be quite different from @option{-pedantic}. We don't have plans to
1783 support such a feature in the near future.
1785 Where the standard specified with @option{-std} represents a GNU
1786 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1787 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1788 extended dialect is based. Warnings from @option{-pedantic} are given
1789 where they are required by the base standard. (It would not make sense
1790 for such warnings to be given only for features not in the specified GNU
1791 C dialect, since by definition the GNU dialects of C include all
1792 features the compiler supports with the given option, and there would be
1793 nothing to warn about.)
1795 @item -pedantic-errors
1796 @opindex pedantic-errors
1797 Like @option{-pedantic}, except that errors are produced rather than
1802 Inhibit all warning messages.
1806 Inhibit warning messages about the use of @samp{#import}.
1808 @item -Wchar-subscripts
1809 @opindex Wchar-subscripts
1810 Warn if an array subscript has type @code{char}. This is a common cause
1811 of error, as programmers often forget that this type is signed on some
1816 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1817 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1821 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1822 the arguments supplied have types appropriate to the format string
1823 specified, and that the conversions specified in the format string make
1824 sense. This includes standard functions, and others specified by format
1825 attributes (@pxref{Function Attributes}), in the @code{printf},
1826 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1827 not in the C standard) families.
1829 The formats are checked against the format features supported by GNU
1830 libc version 2.2. These include all ISO C89 and C99 features, as well
1831 as features from the Single Unix Specification and some BSD and GNU
1832 extensions. Other library implementations may not support all these
1833 features; GCC does not support warning about features that go beyond a
1834 particular library's limitations. However, if @option{-pedantic} is used
1835 with @option{-Wformat}, warnings will be given about format features not
1836 in the selected standard version (but not for @code{strfmon} formats,
1837 since those are not in any version of the C standard). @xref{C Dialect
1838 Options,,Options Controlling C Dialect}.
1840 @option{-Wformat} is included in @option{-Wall}. For more control over some
1841 aspects of format checking, the options @option{-Wno-format-y2k},
1842 @option{-Wno-format-extra-args}, @option{-Wformat-nonliteral},
1843 @option{-Wformat-security} and @option{-Wformat=2} are available, but are
1844 not included in @option{-Wall}.
1846 @item -Wno-format-y2k
1847 @opindex Wno-format-y2k
1848 If @option{-Wformat} is specified, do not warn about @code{strftime}
1849 formats which may yield only a two-digit year.
1851 @item -Wno-format-extra-args
1852 @opindex Wno-format-extra-args
1853 If @option{-Wformat} is specified, do not warn about excess arguments to a
1854 @code{printf} or @code{scanf} format function. The C standard specifies
1855 that such arguments are ignored.
1857 Where the unused arguments lie between used arguments that are
1858 specified with @samp{$} operand number specifications, normally
1859 warnings are still given, since the implementation could not know what
1860 type to pass to @code{va_arg} to skip the unused arguments. However,
1861 in the case of @code{scanf} formats, this option will suppress the
1862 warning if the unused arguments are all pointers, since the Single
1863 Unix Specification says that such unused arguments are allowed.
1865 @item -Wformat-nonliteral
1866 @opindex Wformat-nonliteral
1867 If @option{-Wformat} is specified, also warn if the format string is not a
1868 string literal and so cannot be checked, unless the format function
1869 takes its format arguments as a @code{va_list}.
1871 @item -Wformat-security
1872 @opindex Wformat-security
1873 If @option{-Wformat} is specified, also warn about uses of format
1874 functions that represent possible security problems. At present, this
1875 warns about calls to @code{printf} and @code{scanf} functions where the
1876 format string is not a string literal and there are no format arguments,
1877 as in @code{printf (foo);}. This may be a security hole if the format
1878 string came from untrusted input and contains @samp{%n}. (This is
1879 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1880 in future warnings may be added to @option{-Wformat-security} that are not
1881 included in @option{-Wformat-nonliteral}.)
1885 Enable @option{-Wformat} plus format checks not included in
1886 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1887 -Wformat-nonliteral -Wformat-security}.
1889 @item -Wimplicit-int
1890 @opindex Wimplicit-int
1891 Warn when a declaration does not specify a type.
1893 @item -Wimplicit-function-declaration
1894 @itemx -Werror-implicit-function-declaration
1895 @opindex Wimplicit-function-declaration
1896 @opindex Werror-implicit-function-declaration
1897 Give a warning (or error) whenever a function is used before being
1902 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1906 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1907 function with external linkage, returning int, taking either zero
1908 arguments, two, or three arguments of appropriate types.
1910 @item -Wmissing-braces
1911 @opindex Wmissing-braces
1912 Warn if an aggregate or union initializer is not fully bracketed. In
1913 the following example, the initializer for @samp{a} is not fully
1914 bracketed, but that for @samp{b} is fully bracketed.
1917 int a[2][2] = @{ 0, 1, 2, 3 @};
1918 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1922 @opindex Wparentheses
1923 Warn if parentheses are omitted in certain contexts, such
1924 as when there is an assignment in a context where a truth value
1925 is expected, or when operators are nested whose precedence people
1926 often get confused about.
1928 Also warn about constructions where there may be confusion to which
1929 @code{if} statement an @code{else} branch belongs. Here is an example of
1944 In C, every @code{else} branch belongs to the innermost possible @code{if}
1945 statement, which in this example is @code{if (b)}. This is often not
1946 what the programmer expected, as illustrated in the above example by
1947 indentation the programmer chose. When there is the potential for this
1948 confusion, GCC will issue a warning when this flag is specified.
1949 To eliminate the warning, add explicit braces around the innermost
1950 @code{if} statement so there is no way the @code{else} could belong to
1951 the enclosing @code{if}. The resulting code would look like this:
1967 @item -Wsequence-point
1968 @opindex Wsequence-point
1969 Warn about code that may have undefined semantics because of violations
1970 of sequence point rules in the C standard.
1972 The C standard defines the order in which expressions in a C program are
1973 evaluated in terms of @dfn{sequence points}, which represent a partial
1974 ordering between the execution of parts of the program: those executed
1975 before the sequence point, and those executed after it. These occur
1976 after the evaluation of a full expression (one which is not part of a
1977 larger expression), after the evaluation of the first operand of a
1978 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
1979 function is called (but after the evaluation of its arguments and the
1980 expression denoting the called function), and in certain other places.
1981 Other than as expressed by the sequence point rules, the order of
1982 evaluation of subexpressions of an expression is not specified. All
1983 these rules describe only a partial order rather than a total order,
1984 since, for example, if two functions are called within one expression
1985 with no sequence point between them, the order in which the functions
1986 are called is not specified. However, the standards committee have
1987 ruled that function calls do not overlap.
1989 It is not specified when between sequence points modifications to the
1990 values of objects take effect. Programs whose behavior depends on this
1991 have undefined behavior; the C standard specifies that ``Between the
1992 previous and next sequence point an object shall have its stored value
1993 modified at most once by the evaluation of an expression. Furthermore,
1994 the prior value shall be read only to determine the value to be
1995 stored.''. If a program breaks these rules, the results on any
1996 particular implementation are entirely unpredictable.
1998 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
1999 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2000 diagnosed by this option, and it may give an occasional false positive
2001 result, but in general it has been found fairly effective at detecting
2002 this sort of problem in programs.
2004 The present implementation of this option only works for C programs. A
2005 future implementation may also work for C++ programs.
2007 The C standard is worded confusingly, therefore there is some debate
2008 over the precise meaning of the sequence point rules in subtle cases.
2009 Links to discussions of the problem, including proposed formal
2010 definitions, may be found on our readings page, at
2011 @w{@uref{http://gcc.gnu.org/readings.html}}.
2014 @opindex Wreturn-type
2015 Warn whenever a function is defined with a return-type that defaults to
2016 @code{int}. Also warn about any @code{return} statement with no
2017 return-value in a function whose return-type is not @code{void}.
2019 For C++, a function without return type always produces a diagnostic
2020 message, even when @option{-Wno-return-type} is specified. The only
2021 exceptions are @samp{main} and functions defined in system headers.
2025 Warn whenever a @code{switch} statement has an index of enumeral type
2026 and lacks a @code{case} for one or more of the named codes of that
2027 enumeration. (The presence of a @code{default} label prevents this
2028 warning.) @code{case} labels outside the enumeration range also
2029 provoke warnings when this option is used.
2033 Warn if any trigraphs are encountered that might change the meaning of
2034 the program (trigraphs within comments are not warned about).
2036 @item -Wunused-function
2037 @opindex Wunused-function
2038 Warn whenever a static function is declared but not defined or a
2039 non\-inline static function is unused.
2041 @item -Wunused-label
2042 @opindex Wunused-label
2043 Warn whenever a label is declared but not used.
2045 To suppress this warning use the @samp{unused} attribute
2046 (@pxref{Variable Attributes}).
2048 @item -Wunused-parameter
2049 @opindex Wunused-parameter
2050 Warn whenever a function parameter is unused aside from its declaration.
2052 To suppress this warning use the @samp{unused} attribute
2053 (@pxref{Variable Attributes}).
2055 @item -Wunused-variable
2056 @opindex Wunused-variable
2057 Warn whenever a local variable or non-constant static variable is unused
2058 aside from its declaration
2060 To suppress this warning use the @samp{unused} attribute
2061 (@pxref{Variable Attributes}).
2063 @item -Wunused-value
2064 @opindex Wunused-value
2065 Warn whenever a statement computes a result that is explicitly not used.
2067 To suppress this warning cast the expression to @samp{void}.
2071 All all the above @option{-Wunused} options combined.
2073 In order to get a warning about an unused function parameter, you must
2074 either specify @samp{-W -Wunused} or separately specify
2075 @option{-Wunused-parameter}.
2077 @item -Wuninitialized
2078 @opindex Wuninitialized
2079 Warn if an automatic variable is used without first being initialized or
2080 if a variable may be clobbered by a @code{setjmp} call.
2082 These warnings are possible only in optimizing compilation,
2083 because they require data flow information that is computed only
2084 when optimizing. If you don't specify @option{-O}, you simply won't
2087 These warnings occur only for variables that are candidates for
2088 register allocation. Therefore, they do not occur for a variable that
2089 is declared @code{volatile}, or whose address is taken, or whose size
2090 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2091 structures, unions or arrays, even when they are in registers.
2093 Note that there may be no warning about a variable that is used only
2094 to compute a value that itself is never used, because such
2095 computations may be deleted by data flow analysis before the warnings
2098 These warnings are made optional because GCC is not smart
2099 enough to see all the reasons why the code might be correct
2100 despite appearing to have an error. Here is one example of how
2121 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2122 always initialized, but GCC doesn't know this. Here is
2123 another common case:
2128 if (change_y) save_y = y, y = new_y;
2130 if (change_y) y = save_y;
2135 This has no bug because @code{save_y} is used only if it is set.
2137 @cindex @code{longjmp} warnings
2138 This option also warns when a non-volatile automatic variable might be
2139 changed by a call to @code{longjmp}. These warnings as well are possible
2140 only in optimizing compilation.
2142 The compiler sees only the calls to @code{setjmp}. It cannot know
2143 where @code{longjmp} will be called; in fact, a signal handler could
2144 call it at any point in the code. As a result, you may get a warning
2145 even when there is in fact no problem because @code{longjmp} cannot
2146 in fact be called at the place which would cause a problem.
2148 Some spurious warnings can be avoided if you declare all the functions
2149 you use that never return as @code{noreturn}. @xref{Function
2152 @item -Wreorder @r{(C++ only)}
2154 @cindex reordering, warning
2155 @cindex warning for reordering of member initializers
2156 Warn when the order of member initializers given in the code does not
2157 match the order in which they must be executed. For instance:
2159 @item -Wunknown-pragmas
2160 @opindex Wunknown-pragmas
2161 @cindex warning for unknown pragmas
2162 @cindex unknown pragmas, warning
2163 @cindex pragmas, warning of unknown
2164 Warn when a #pragma directive is encountered which is not understood by
2165 GCC@. If this command line option is used, warnings will even be issued
2166 for unknown pragmas in system header files. This is not the case if
2167 the warnings were only enabled by the @option{-Wall} command line option.
2171 All of the above @samp{-W} options combined. This enables all the
2172 warnings about constructions that some users consider questionable, and
2173 that are easy to avoid (or modify to prevent the warning), even in
2174 conjunction with macros.
2177 @opindex Wno-div-by-zero
2178 @opindex Wdiv-by-zero
2179 Warn about compile-time integer division by zero. This is default. To
2180 inhibit the warning messages, use @option{-Wno-div-by-zero}. Floating
2181 point division by zero is not warned about, as it can be a legitimate
2182 way of obtaining infinities and NaNs.
2185 @opindex Wno-multichar
2187 Warn if a multicharacter constant (@samp{'FOOF'}) is used. This is
2188 default. To inhibit the warning messages, use @option{-Wno-multichar}.
2189 Usually they indicate a typo in the user's code, as they have
2190 implementation-defined values, and should not be used in portable code.
2192 @item -Wsystem-headers
2193 @opindex Wsystem-headers
2194 @cindex warnings from system headers
2195 @cindex system headers, warnings from
2196 Print warning messages for constructs found in system header files.
2197 Warnings from system headers are normally suppressed, on the assumption
2198 that they usually do not indicate real problems and would only make the
2199 compiler output harder to read. Using this command line option tells
2200 GCC to emit warnings from system headers as if they occurred in user
2201 code. However, note that using @option{-Wall} in conjunction with this
2202 option will @emph{not} warn about unknown pragmas in system
2203 headers---for that, @option{-Wunknown-pragmas} must also be used.
2206 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2207 Some of them warn about constructions that users generally do not
2208 consider questionable, but which occasionally you might wish to check
2209 for; others warn about constructions that are necessary or hard to avoid
2210 in some cases, and there is no simple way to modify the code to suppress
2216 Print extra warning messages for these events:
2220 A function can return either with or without a value. (Falling
2221 off the end of the function body is considered returning without
2222 a value.) For example, this function would evoke such a
2236 An expression-statement or the left-hand side of a comma expression
2237 contains no side effects.
2238 To suppress the warning, cast the unused expression to void.
2239 For example, an expression such as @samp{x[i,j]} will cause a warning,
2240 but @samp{x[(void)i,j]} will not.
2243 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2246 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2247 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2248 that of ordinary mathematical notation.
2251 Storage-class specifiers like @code{static} are not the first things in
2252 a declaration. According to the C Standard, this usage is obsolescent.
2255 The return type of a function has a type qualifier such as @code{const}.
2256 Such a type qualifier has no effect, since the value returned by a
2257 function is not an lvalue. (But don't warn about the GNU extension of
2258 @code{volatile void} return types. That extension will be warned about
2259 if @option{-pedantic} is specified.)
2262 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2266 A comparison between signed and unsigned values could produce an
2267 incorrect result when the signed value is converted to unsigned.
2268 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2271 An aggregate has a partly bracketed initializer.
2272 For example, the following code would evoke such a warning,
2273 because braces are missing around the initializer for @code{x.h}:
2276 struct s @{ int f, g; @};
2277 struct t @{ struct s h; int i; @};
2278 struct t x = @{ 1, 2, 3 @};
2282 An aggregate has an initializer which does not initialize all members.
2283 For example, the following code would cause such a warning, because
2284 @code{x.h} would be implicitly initialized to zero:
2287 struct s @{ int f, g, h; @};
2288 struct s x = @{ 3, 4 @};
2293 @opindex Wfloat-equal
2294 Warn if floating point values are used in equality comparisons.
2296 The idea behind this is that sometimes it is convenient (for the
2297 programmer) to consider floating-point values as approximations to
2298 infinitely precise real numbers. If you are doing this, then you need
2299 to compute (by analysing the code, or in some other way) the maximum or
2300 likely maximum error that the computation introduces, and allow for it
2301 when performing comparisons (and when producing output, but that's a
2302 different problem). In particular, instead of testing for equality, you
2303 would check to see whether the two values have ranges that overlap; and
2304 this is done with the relational operators, so equality comparisons are
2307 @item -Wtraditional @r{(C only)}
2308 @opindex Wtraditional
2309 Warn about certain constructs that behave differently in traditional and
2310 ISO C@. Also warn about ISO C constructs that have no traditional C
2311 equivalent, and/or problematic constructs which should be avoided.
2315 Macro parameters that appear within string literals in the macro body.
2316 In traditional C macro replacement takes place within string literals,
2317 but does not in ISO C@.
2320 In traditional C, some preprocessor directives did not exist.
2321 Traditional preprocessors would only consider a line to be a directive
2322 if the @samp{#} appeared in column 1 on the line. Therefore
2323 @option{-Wtraditional} warns about directives that traditional C
2324 understands but would ignore because the @samp{#} does not appear as the
2325 first character on the line. It also suggests you hide directives like
2326 @samp{#pragma} not understood by traditional C by indenting them. Some
2327 traditional implementations would not recognize @samp{#elif}, so it
2328 suggests avoiding it altogether.
2331 A function-like macro that appears without arguments.
2334 The unary plus operator.
2337 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2338 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2339 constants.) Note, these suffixes appear in macros defined in the system
2340 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2341 Use of these macros in user code might normally lead to spurious
2342 warnings, however gcc's integrated preprocessor has enough context to
2343 avoid warning in these cases.
2346 A function declared external in one block and then used after the end of
2350 A @code{switch} statement has an operand of type @code{long}.
2353 A non-@code{static} function declaration follows a @code{static} one.
2354 This construct is not accepted by some traditional C compilers.
2357 The ISO type of an integer constant has a different width or
2358 signedness from its traditional type. This warning is only issued if
2359 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2360 typically represent bit patterns, are not warned about.
2363 Usage of ISO string concatenation is detected.
2366 Initialization of automatic aggregates.
2369 Identifier conflicts with labels. Traditional C lacks a separate
2370 namespace for labels.
2373 Initialization of unions. If the initializer is zero, the warning is
2374 omitted. This is done under the assumption that the zero initializer in
2375 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2376 initializer warnings and relies on default initialization to zero in the
2380 Conversions by prototypes between fixed/floating point values and vice
2381 versa. The absence of these prototypes when compiling with traditional
2382 C would cause serious problems. This is a subset of the possible
2383 conversion warnings, for the full set use @option{-Wconversion}.
2388 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2392 Warn whenever a local variable shadows another local variable, parameter or
2393 global variable or whenever a built-in function is shadowed.
2395 @item -Wlarger-than-@var{len}
2396 @opindex Wlarger-than
2397 Warn whenever an object of larger than @var{len} bytes is defined.
2399 @item -Wpointer-arith
2400 @opindex Wpointer-arith
2401 Warn about anything that depends on the ``size of'' a function type or
2402 of @code{void}. GNU C assigns these types a size of 1, for
2403 convenience in calculations with @code{void *} pointers and pointers
2406 @item -Wbad-function-cast @r{(C only)}
2407 @opindex Wbad-function-cast
2408 Warn whenever a function call is cast to a non-matching type.
2409 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2413 Warn whenever a pointer is cast so as to remove a type qualifier from
2414 the target type. For example, warn if a @code{const char *} is cast
2415 to an ordinary @code{char *}.
2418 @opindex Wcast-align
2419 Warn whenever a pointer is cast such that the required alignment of the
2420 target is increased. For example, warn if a @code{char *} is cast to
2421 an @code{int *} on machines where integers can only be accessed at
2422 two- or four-byte boundaries.
2424 @item -Wwrite-strings
2425 @opindex Wwrite-strings
2426 When compiling C, give string constants the type @code{const
2427 char[@var{length}]} so that
2428 copying the address of one into a non-@code{const} @code{char *}
2429 pointer will get a warning; when compiling C++, warn about the
2430 deprecated conversion from string constants to @code{char *}.
2431 These warnings will help you find at
2432 compile time code that can try to write into a string constant, but
2433 only if you have been very careful about using @code{const} in
2434 declarations and prototypes. Otherwise, it will just be a nuisance;
2435 this is why we did not make @option{-Wall} request these warnings.
2438 @opindex Wconversion
2439 Warn if a prototype causes a type conversion that is different from what
2440 would happen to the same argument in the absence of a prototype. This
2441 includes conversions of fixed point to floating and vice versa, and
2442 conversions changing the width or signedness of a fixed point argument
2443 except when the same as the default promotion.
2445 Also, warn if a negative integer constant expression is implicitly
2446 converted to an unsigned type. For example, warn about the assignment
2447 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2448 casts like @code{(unsigned) -1}.
2450 @item -Wsign-compare
2451 @opindex Wsign-compare
2452 @cindex warning for comparison of signed and unsigned values
2453 @cindex comparison of signed and unsigned values, warning
2454 @cindex signed and unsigned values, comparison warning
2455 Warn when a comparison between signed and unsigned values could produce
2456 an incorrect result when the signed value is converted to unsigned.
2457 This warning is also enabled by @option{-W}; to get the other warnings
2458 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2460 @item -Waggregate-return
2461 @opindex Waggregate-return
2462 Warn if any functions that return structures or unions are defined or
2463 called. (In languages where you can return an array, this also elicits
2466 @item -Wstrict-prototypes @r{(C only)}
2467 @opindex Wstrict-prototypes
2468 Warn if a function is declared or defined without specifying the
2469 argument types. (An old-style function definition is permitted without
2470 a warning if preceded by a declaration which specifies the argument
2473 @item -Wmissing-prototypes @r{(C only)}
2474 @opindex Wmissing-prototypes
2475 Warn if a global function is defined without a previous prototype
2476 declaration. This warning is issued even if the definition itself
2477 provides a prototype. The aim is to detect global functions that fail
2478 to be declared in header files.
2480 @item -Wmissing-declarations
2481 @opindex Wmissing-declarations
2482 Warn if a global function is defined without a previous declaration.
2483 Do so even if the definition itself provides a prototype.
2484 Use this option to detect global functions that are not declared in
2487 @item -Wmissing-noreturn
2488 @opindex Wmissing-noreturn
2489 Warn about functions which might be candidates for attribute @code{noreturn}.
2490 Note these are only possible candidates, not absolute ones. Care should
2491 be taken to manually verify functions actually do not ever return before
2492 adding the @code{noreturn} attribute, otherwise subtle code generation
2493 bugs could be introduced. You will not get a warning for @code{main} in
2494 hosted C environments.
2496 @item -Wmissing-format-attribute
2497 @opindex Wmissing-format-attribute
2499 If @option{-Wformat} is enabled, also warn about functions which might be
2500 candidates for @code{format} attributes. Note these are only possible
2501 candidates, not absolute ones. GCC will guess that @code{format}
2502 attributes might be appropriate for any function that calls a function
2503 like @code{vprintf} or @code{vscanf}, but this might not always be the
2504 case, and some functions for which @code{format} attributes are
2505 appropriate may not be detected. This option has no effect unless
2506 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2508 @item -Wno-deprecated-declarations
2509 @opindex Wno-deprecated-declarations
2510 Do not warn about uses of functions, variables, and types marked as
2511 deprecated by using the @code{deprecated} attribute.
2512 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2513 @pxref{Type Attributes}.)
2517 Warn if a structure is given the packed attribute, but the packed
2518 attribute has no effect on the layout or size of the structure.
2519 Such structures may be mis-aligned for little benefit. For
2520 instance, in this code, the variable @code{f.x} in @code{struct bar}
2521 will be misaligned even though @code{struct bar} does not itself
2522 have the packed attribute:
2529 @} __attribute__((packed));
2539 Warn if padding is included in a structure, either to align an element
2540 of the structure or to align the whole structure. Sometimes when this
2541 happens it is possible to rearrange the fields of the structure to
2542 reduce the padding and so make the structure smaller.
2544 @item -Wredundant-decls
2545 @opindex Wredundant-decls
2546 Warn if anything is declared more than once in the same scope, even in
2547 cases where multiple declaration is valid and changes nothing.
2549 @item -Wnested-externs @r{(C only)}
2550 @opindex Wnested-externs
2551 Warn if an @code{extern} declaration is encountered within a function.
2553 @item -Wunreachable-code
2554 @opindex Wunreachable-code
2555 Warn if the compiler detects that code will never be executed.
2557 This option is intended to warn when the compiler detects that at
2558 least a whole line of source code will never be executed, because
2559 some condition is never satisfied or because it is after a
2560 procedure that never returns.
2562 It is possible for this option to produce a warning even though there
2563 are circumstances under which part of the affected line can be executed,
2564 so care should be taken when removing apparently-unreachable code.
2566 For instance, when a function is inlined, a warning may mean that the
2567 line is unreachable in only one inlined copy of the function.
2569 This option is not made part of @option{-Wall} because in a debugging
2570 version of a program there is often substantial code which checks
2571 correct functioning of the program and is, hopefully, unreachable
2572 because the program does work. Another common use of unreachable
2573 code is to provide behavior which is selectable at compile-time.
2577 Warn if a function can not be inlined and it was declared as inline.
2581 @opindex Wno-long-long
2582 Warn if @samp{long long} type is used. This is default. To inhibit
2583 the warning messages, use @option{-Wno-long-long}. Flags
2584 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2585 only when @option{-pedantic} flag is used.
2587 @item -Wdisabled-optimization
2588 @opindex Wdisabled-optimization
2589 Warn if a requested optimization pass is disabled. This warning does
2590 not generally indicate that there is anything wrong with your code; it
2591 merely indicates that GCC's optimizers were unable to handle the code
2592 effectively. Often, the problem is that your code is too big or too
2593 complex; GCC will refuse to optimize programs when the optimization
2594 itself is likely to take inordinate amounts of time.
2598 Make all warnings into errors.
2601 @node Debugging Options
2602 @section Options for Debugging Your Program or GCC
2603 @cindex options, debugging
2604 @cindex debugging information options
2606 GCC has various special options that are used for debugging
2607 either your program or GCC:
2612 Produce debugging information in the operating system's native format
2613 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2616 On most systems that use stabs format, @option{-g} enables use of extra
2617 debugging information that only GDB can use; this extra information
2618 makes debugging work better in GDB but will probably make other debuggers
2620 refuse to read the program. If you want to control for certain whether
2621 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2622 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2623 or @option{-gvms} (see below).
2625 Unlike most other C compilers, GCC allows you to use @option{-g} with
2626 @option{-O}. The shortcuts taken by optimized code may occasionally
2627 produce surprising results: some variables you declared may not exist
2628 at all; flow of control may briefly move where you did not expect it;
2629 some statements may not be executed because they compute constant
2630 results or their values were already at hand; some statements may
2631 execute in different places because they were moved out of loops.
2633 Nevertheless it proves possible to debug optimized output. This makes
2634 it reasonable to use the optimizer for programs that might have bugs.
2636 The following options are useful when GCC is generated with the
2637 capability for more than one debugging format.
2641 Produce debugging information for use by GDB@. This means to use the
2642 most expressive format available (DWARF 2, stabs, or the native format
2643 if neither of those are supported), including GDB extensions if at all
2648 Produce debugging information in stabs format (if that is supported),
2649 without GDB extensions. This is the format used by DBX on most BSD
2650 systems. On MIPS, Alpha and System V Release 4 systems this option
2651 produces stabs debugging output which is not understood by DBX or SDB@.
2652 On System V Release 4 systems this option requires the GNU assembler.
2656 Produce debugging information in stabs format (if that is supported),
2657 using GNU extensions understood only by the GNU debugger (GDB)@. The
2658 use of these extensions is likely to make other debuggers crash or
2659 refuse to read the program.
2663 Produce debugging information in COFF format (if that is supported).
2664 This is the format used by SDB on most System V systems prior to
2669 Produce debugging information in XCOFF format (if that is supported).
2670 This is the format used by the DBX debugger on IBM RS/6000 systems.
2674 Produce debugging information in XCOFF format (if that is supported),
2675 using GNU extensions understood only by the GNU debugger (GDB)@. The
2676 use of these extensions is likely to make other debuggers crash or
2677 refuse to read the program, and may cause assemblers other than the GNU
2678 assembler (GAS) to fail with an error.
2682 Produce debugging information in DWARF version 1 format (if that is
2683 supported). This is the format used by SDB on most System V Release 4
2688 Produce debugging information in DWARF version 1 format (if that is
2689 supported), using GNU extensions understood only by the GNU debugger
2690 (GDB)@. The use of these extensions is likely to make other debuggers
2691 crash or refuse to read the program.
2695 Produce debugging information in DWARF version 2 format (if that is
2696 supported). This is the format used by DBX on IRIX 6.
2700 Produce debugging information in VMS debug format (if that is
2701 supported). This is the format used by DEBUG on VMS systems.
2704 @itemx -ggdb@var{level}
2705 @itemx -gstabs@var{level}
2706 @itemx -gcoff@var{level}
2707 @itemx -gxcoff@var{level}
2708 @itemx -gdwarf@var{level}
2709 @itemx -gdwarf-2@var{level}
2710 @itemx -gvms@var{level}
2711 Request debugging information and also use @var{level} to specify how
2712 much information. The default level is 2.
2714 Level 1 produces minimal information, enough for making backtraces in
2715 parts of the program that you don't plan to debug. This includes
2716 descriptions of functions and external variables, but no information
2717 about local variables and no line numbers.
2719 Level 3 includes extra information, such as all the macro definitions
2720 present in the program. Some debuggers support macro expansion when
2721 you use @option{-g3}.
2726 Generate extra code to write profile information suitable for the
2727 analysis program @code{prof}. You must use this option when compiling
2728 the source files you want data about, and you must also use it when
2731 @cindex @code{gprof}
2734 Generate extra code to write profile information suitable for the
2735 analysis program @code{gprof}. You must use this option when compiling
2736 the source files you want data about, and you must also use it when
2742 Generate extra code to write profile information for basic blocks, which will
2743 record the number of times each basic block is executed, the basic block start
2744 address, and the function name containing the basic block. If @option{-g} is
2745 used, the line number and filename of the start of the basic block will also be
2746 recorded. If not overridden by the machine description, the default action is
2747 to append to the text file @file{bb.out}.
2749 This data could be analyzed by a program like @code{tcov}. Note,
2750 however, that the format of the data is not what @code{tcov} expects.
2751 Eventually GNU @code{gprof} should be extended to process this data.
2755 Makes the compiler print out each function name as it is compiled, and
2756 print some statistics about each pass when it finishes.
2759 @opindex ftime-report
2760 Makes the compiler print some statistics about the time consumed by each
2761 pass when it finishes.
2764 @opindex fmem-report
2765 Makes the compiler print some statistics about permanent memory
2766 allocation when it finishes.
2768 @item -fprofile-arcs
2769 @opindex fprofile-arcs
2770 Instrument @dfn{arcs} during compilation to generate coverage data
2771 or for profile-directed block ordering. During execution the program
2772 records how many times each branch is executed and how many times it is
2773 taken. When the compiled program exits it saves this data to a file
2774 called @file{@var{sourcename}.da} for each source file.
2776 For profile-directed block ordering, compile the program with
2777 @option{-fprofile-arcs} plus optimization and code generation options,
2778 generate the arc profile information by running the program on a
2779 selected workload, and then compile the program again with the same
2780 optimization and code generation options plus
2781 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2782 Control Optimization}).
2784 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2785 when it is used with the @option{-ftest-coverage} option. GCC
2786 supports two methods of determining code coverage: the options that
2787 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2788 write information to text files. The options that support @code{gcov}
2789 do not need to instrument every arc in the program, so a program compiled
2790 with them runs faster than a program compiled with @option{-a}, which
2791 adds instrumentation code to every basic block in the program. The
2792 tradeoff: since @code{gcov} does not have execution counts for all
2793 branches, it must start with the execution counts for the instrumented
2794 branches, and then iterate over the program flow graph until the entire
2795 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2796 a program which uses information from @option{-a} and @option{-ax}.
2798 With @option{-fprofile-arcs}, for each function of your program GCC
2799 creates a program flow graph, then finds a spanning tree for the graph.
2800 Only arcs that are not on the spanning tree have to be instrumented: the
2801 compiler adds code to count the number of times that these arcs are
2802 executed. When an arc is the only exit or only entrance to a block, the
2803 instrumentation code can be added to the block; otherwise, a new basic
2804 block must be created to hold the instrumentation code.
2806 This option makes it possible to estimate branch probabilities and to
2807 calculate basic block execution counts. In general, basic block
2808 execution counts as provided by @option{-a} do not give enough
2809 information to estimate all branch probabilities.
2812 @item -ftest-coverage
2813 @opindex ftest-coverage
2814 Create data files for the @code{gcov} code-coverage utility
2815 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2816 The data file names begin with the name of your source file:
2819 @item @var{sourcename}.bb
2820 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2821 associate basic block execution counts with line numbers.
2823 @item @var{sourcename}.bbg
2824 A list of all arcs in the program flow graph. This allows @code{gcov}
2825 to reconstruct the program flow graph, so that it can compute all basic
2826 block and arc execution counts from the information in the
2827 @code{@var{sourcename}.da} file.
2830 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2831 option adds instrumentation to the program, which then writes
2832 execution counts to another data file:
2835 @item @var{sourcename}.da
2836 Runtime arc execution counts, used in conjunction with the arc
2837 information in the file @code{@var{sourcename}.bbg}.
2840 Coverage data will map better to the source files if
2841 @option{-ftest-coverage} is used without optimization.
2843 @item -d@var{letters}
2845 Says to make debugging dumps during compilation at times specified by
2846 @var{letters}. This is used for debugging the compiler. The file names
2847 for most of the dumps are made by appending a pass number and a word to
2848 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2849 Here are the possible letters for use in @var{letters}, and their meanings:
2854 Annotate the assembler output with miscellaneous debugging information.
2857 Dump after computing branch probabilities, to @file{@var{file}.15.bp}.
2860 Dump after block reordering, to @file{@var{file}.30.bbro}.
2863 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
2866 Dump after the first if conversion, to the file @file{@var{file}.18.ce}.
2869 Dump after delayed branch scheduling, to @file{@var{file}.32.dbr}.
2872 Dump all macro definitions, at the end of preprocessing, in addition to
2876 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2877 @file{@var{file}.07.ussa}.
2880 Dump after the second if conversion, to @file{@var{file}.27.ce2}.
2883 Dump after life analysis, to @file{@var{file}.16.life}.
2886 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.10.addressof}.
2889 Dump after global register allocation, to @file{@var{file}.22.greg}.
2892 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2895 Dump after reg-to-stack conversion, to @file{@var{file}.29.stack}.
2898 Dump after post-reload optimizations, to @file{@var{file}.23.postreload}.
2901 Dump after GCSE, to @file{@var{file}.11.gcse}.
2904 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2907 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2910 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
2913 Dump after local register allocation, to @file{@var{file}.21.lreg}.
2916 Dump after loop optimization, to @file{@var{file}.12.loop}.
2919 Dump after performing the machine dependent reorganisation pass, to
2920 @file{@var{file}.31.mach}.
2923 Dump after register renumbering, to @file{@var{file}.26.rnreg}.
2926 Dump after the register move pass, to @file{@var{file}.19.regmove}.
2929 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2932 Dump after the second scheduling pass, to @file{@var{file}.28.sched2}.
2935 Dump after CSE (including the jump optimization that sometimes follows
2936 CSE), to @file{@var{file}.09.cse}.
2939 Dump after the first scheduling pass, to @file{@var{file}.20.sched}.
2942 Dump after the second CSE pass (including the jump optimization that
2943 sometimes follows CSE), to @file{@var{file}.13.cse2}.
2946 Dump after null pointer ellimination pass ti @file{@var{file}.08.null}.
2949 Dump after the second flow pass, to @file{@var{file}.24.flow2}.
2952 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2955 Dump after the peephole pass, to @file{@var{file}.25.peephole2}.
2958 Produce all the dumps listed above.
2961 Print statistics on memory usage, at the end of the run, to
2965 Annotate the assembler output with a comment indicating which
2966 pattern and alternative was used. The length of each instruction is
2970 Dump the RTL in the assembler output as a comment before each instruction.
2971 Also turns on @option{-dp} annotation.
2974 For each of the other indicated dump files (except for
2975 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
2976 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
2979 Just generate RTL for a function instead of compiling it. Usually used
2983 Dump debugging information during parsing, to standard error.
2986 @item -fdump-unnumbered
2987 @opindex fdump-unnumbered
2988 When doing debugging dumps (see @option{-d} option above), suppress instruction
2989 numbers and line number note output. This makes it more feasible to
2990 use diff on debugging dumps for compiler invocations with different
2991 options, in particular with and without @option{-g}.
2993 @item -fdump-translation-unit @r{(C and C++ only)}
2994 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
2995 @opindex fdump-translation-unit
2996 Dump a representation of the tree structure for the entire translation
2997 unit to a file. The file name is made by appending @file{.tu} to the
2998 source file name. If the @samp{-@var{options}} form is used, @var{options}
2999 controls the details of the dump as described for the
3000 @option{-fdump-tree} options.
3002 @item -fdump-class-hierarchy @r{(C++ only)}
3003 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3004 @opindex fdump-class-hierarchy
3005 Dump a representation of each class's hierarchy and virtual function
3006 table layout to a file. The file name is made by appending @file{.class}
3007 to the source file name. If the @samp{-@var{options}} form is used,
3008 @var{options} controls the details of the dump as described for the
3009 @option{-fdump-tree} options.
3011 @item -fdump-tree-@var{switch} @r{(C++ only)}
3012 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3014 Control the dumping at various stages of processing the intermediate
3015 language tree to a file. The file name is generated by appending a switch
3016 specific suffix to the source file name. If the @samp{-@var{options}}
3017 form is used, @var{options} is a list of @samp{-} separated options that
3018 control the details of the dump. Not all options are applicable to all
3019 dumps, those which are not meaningful will be ignored. The following
3020 options are available
3024 Print the address of each node. Usually this is not meaningful as it
3025 changes according to the environment and source file. Its primary use
3026 is for tying up a dump file with a debug environment.
3028 Inhibit dumping of members of a scope or body of a function merely
3029 because that scope has been reached. Only dump such items when they
3030 are directly reachable by some other path.
3032 Turn on all options.
3035 The following tree dumps are possible:
3038 Dump before any tree based optimization, to @file{@var{file}.original}.
3040 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3042 Dump after function inlining, to @file{@var{file}.inlined}.
3045 @item -fpretend-float
3046 @opindex fpretend-float
3047 When running a cross-compiler, pretend that the target machine uses the
3048 same floating point format as the host machine. This causes incorrect
3049 output of the actual floating constants, but the actual instruction
3050 sequence will probably be the same as GCC would make when running on
3055 Store the usual ``temporary'' intermediate files permanently; place them
3056 in the current directory and name them based on the source file. Thus,
3057 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3058 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3059 preprocessed @file{foo.i} output file even though the compiler now
3060 normally uses an integrated preprocessor.
3064 Report the CPU time taken by each subprocess in the compilation
3065 sequence. For C source files, this is the compiler proper and assembler
3066 (plus the linker if linking is done). The output looks like this:
3073 The first number on each line is the ``user time,'' that is time spent
3074 executing the program itself. The second number is ``system time,''
3075 time spent executing operating system routines on behalf of the program.
3076 Both numbers are in seconds.
3078 @item -print-file-name=@var{library}
3079 @opindex print-file-name
3080 Print the full absolute name of the library file @var{library} that
3081 would be used when linking---and don't do anything else. With this
3082 option, GCC does not compile or link anything; it just prints the
3085 @item -print-multi-directory
3086 @opindex print-multi-directory
3087 Print the directory name corresponding to the multilib selected by any
3088 other switches present in the command line. This directory is supposed
3089 to exist in @env{GCC_EXEC_PREFIX}.
3091 @item -print-multi-lib
3092 @opindex print-multi-lib
3093 Print the mapping from multilib directory names to compiler switches
3094 that enable them. The directory name is separated from the switches by
3095 @samp{;}, and each switch starts with an @samp{@@} instead of the
3096 @samp{-}, without spaces between multiple switches. This is supposed to
3097 ease shell-processing.
3099 @item -print-prog-name=@var{program}
3100 @opindex print-prog-name
3101 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3103 @item -print-libgcc-file-name
3104 @opindex print-libgcc-file-name
3105 Same as @option{-print-file-name=libgcc.a}.
3107 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3108 but you do want to link with @file{libgcc.a}. You can do
3111 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3114 @item -print-search-dirs
3115 @opindex print-search-dirs
3116 Print the name of the configured installation directory and a list of
3117 program and library directories gcc will search---and don't do anything else.
3119 This is useful when gcc prints the error message
3120 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3121 To resolve this you either need to put @file{cpp0} and the other compiler
3122 components where gcc expects to find them, or you can set the environment
3123 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3124 Don't forget the trailing '/'.
3125 @xref{Environment Variables}.
3128 @opindex dumpmachine
3129 Print the compiler's target machine (for example,
3130 @samp{i686-pc-linux-gnu})---and don't do anything else.
3133 @opindex dumpversion
3134 Print the compiler version (for example, @samp{3.0})---and don't do
3139 Print the compiler's built-in specs---and don't do anything else. (This
3140 is used when GCC itself is being built.) @xref{Spec Files}.
3143 @node Optimize Options
3144 @section Options That Control Optimization
3145 @cindex optimize options
3146 @cindex options, optimization
3148 These options control various sorts of optimizations:
3155 Optimize. Optimizing compilation takes somewhat more time, and a lot
3156 more memory for a large function.
3158 Without @option{-O}, the compiler's goal is to reduce the cost of
3159 compilation and to make debugging produce the expected results.
3160 Statements are independent: if you stop the program with a breakpoint
3161 between statements, you can then assign a new value to any variable or
3162 change the program counter to any other statement in the function and
3163 get exactly the results you would expect from the source code.
3165 With @option{-O}, the compiler tries to reduce code size and execution
3166 time, without performing any optimizations that take a great deal of
3171 Optimize even more. GCC performs nearly all supported optimizations
3172 that do not involve a space-speed tradeoff. The compiler does not
3173 perform loop unrolling or function inlining when you specify @option{-O2}.
3174 As compared to @option{-O}, this option increases both compilation time
3175 and the performance of the generated code.
3177 @option{-O2} turns on all optional optimizations except for loop unrolling,
3178 function inlining, and register renaming. It also turns on the
3179 @option{-fforce-mem} option on all machines and frame pointer elimination
3180 on machines where doing so does not interfere with debugging.
3182 Please note the warning under @option{-fgcse} about
3183 invoking @option{-O2} on programs that use computed gotos.
3187 Optimize yet more. @option{-O3} turns on all optimizations specified by
3188 @option{-O2} and also turns on the @option{-finline-functions} and
3189 @option{-frename-registers} options.
3197 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3198 do not typically increase code size. It also performs further
3199 optimizations designed to reduce code size.
3201 If you use multiple @option{-O} options, with or without level numbers,
3202 the last such option is the one that is effective.
3205 Options of the form @option{-f@var{flag}} specify machine-independent
3206 flags. Most flags have both positive and negative forms; the negative
3207 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3208 only one of the forms is listed---the one which is not the default.
3209 You can figure out the other form by either removing @samp{no-} or
3214 @opindex ffloat-store
3215 Do not store floating point variables in registers, and inhibit other
3216 options that might change whether a floating point value is taken from a
3219 @cindex floating point precision
3220 This option prevents undesirable excess precision on machines such as
3221 the 68000 where the floating registers (of the 68881) keep more
3222 precision than a @code{double} is supposed to have. Similarly for the
3223 x86 architecture. For most programs, the excess precision does only
3224 good, but a few programs rely on the precise definition of IEEE floating
3225 point. Use @option{-ffloat-store} for such programs, after modifying
3226 them to store all pertinent intermediate computations into variables.
3228 @item -fno-default-inline
3229 @opindex fno-default-inline
3230 Do not make member functions inline by default merely because they are
3231 defined inside the class scope (C++ only). Otherwise, when you specify
3232 @w{@option{-O}}, member functions defined inside class scope are compiled
3233 inline by default; i.e., you don't need to add @samp{inline} in front of
3234 the member function name.
3236 @item -fno-defer-pop
3237 @opindex fno-defer-pop
3238 Always pop the arguments to each function call as soon as that function
3239 returns. For machines which must pop arguments after a function call,
3240 the compiler normally lets arguments accumulate on the stack for several
3241 function calls and pops them all at once.
3245 Force memory operands to be copied into registers before doing
3246 arithmetic on them. This produces better code by making all memory
3247 references potential common subexpressions. When they are not common
3248 subexpressions, instruction combination should eliminate the separate
3249 register-load. The @option{-O2} option turns on this option.
3252 @opindex fforce-addr
3253 Force memory address constants to be copied into registers before
3254 doing arithmetic on them. This may produce better code just as
3255 @option{-fforce-mem} may.
3257 @item -fomit-frame-pointer
3258 @opindex fomit-frame-pointer
3259 Don't keep the frame pointer in a register for functions that
3260 don't need one. This avoids the instructions to save, set up and
3261 restore frame pointers; it also makes an extra register available
3262 in many functions. @strong{It also makes debugging impossible on
3265 On some machines, such as the VAX, this flag has no effect, because
3266 the standard calling sequence automatically handles the frame pointer
3267 and nothing is saved by pretending it doesn't exist. The
3268 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3269 whether a target machine supports this flag. @xref{Registers,,Register
3270 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3272 @item -foptimize-sibling-calls
3273 @opindex foptimize-sibling-calls
3274 Optimize sibling and tail recursive calls.
3278 This option generates traps for signed overflow on addition, subtraction,
3279 multiplication operations.
3283 Don't pay attention to the @code{inline} keyword. Normally this option
3284 is used to keep the compiler from expanding any functions inline.
3285 Note that if you are not optimizing, no functions can be expanded inline.
3287 @item -finline-functions
3288 @opindex finline-functions
3289 Integrate all simple functions into their callers. The compiler
3290 heuristically decides which functions are simple enough to be worth
3291 integrating in this way.
3293 If all calls to a given function are integrated, and the function is
3294 declared @code{static}, then the function is normally not output as
3295 assembler code in its own right.
3297 @item -finline-limit=@var{n}
3298 @opindex finline-limit
3299 By default, gcc limits the size of functions that can be inlined. This flag
3300 allows the control of this limit for functions that are explicitly marked as
3301 inline (ie marked with the inline keyword or defined within the class
3302 definition in c++). @var{n} is the size of functions that can be inlined in
3303 number of pseudo instructions (not counting parameter handling). The default
3304 value of @var{n} is 600.
3305 Increasing this value can result in more inlined code at
3306 the cost of compilation time and memory consumption. Decreasing usually makes
3307 the compilation faster and less code will be inlined (which presumably
3308 means slower programs). This option is particularly useful for programs that
3309 use inlining heavily such as those based on recursive templates with C++.
3311 @emph{Note:} pseudo instruction represents, in this particular context, an
3312 abstract measurement of function's size. In no way, it represents a count
3313 of assembly instructions and as such its exact meaning might change from one
3314 release to an another.
3316 @item -fkeep-inline-functions
3317 @opindex fkeep-inline-functions
3318 Even if all calls to a given function are integrated, and the function
3319 is declared @code{static}, nevertheless output a separate run-time
3320 callable version of the function. This switch does not affect
3321 @code{extern inline} functions.
3323 @item -fkeep-static-consts
3324 @opindex fkeep-static-consts
3325 Emit variables declared @code{static const} when optimization isn't turned
3326 on, even if the variables aren't referenced.
3328 GCC enables this option by default. If you want to force the compiler to
3329 check if the variable was referenced, regardless of whether or not
3330 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3332 @item -fmerge-constants
3333 Attempt to merge identical constants (string constants and floating point
3334 constants) accross compilation units.
3336 This option is default for optimized compilation if assembler and linker
3337 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3339 @item -fmerge-all-constants
3340 Attempt to merge identical constants and identical variables.
3342 This option implies @option{-fmerge-constants}. In addition to
3343 @option{-fmerge-constants} this considers e.g. even constant initialized
3344 arrays or initialized constant variables with integral or floating point
3345 types. Languages like C or C++ require each non-automatic variable to
3346 have distinct location, so using this option will result in non-conforming
3349 @item -fno-function-cse
3350 @opindex fno-function-cse
3351 Do not put function addresses in registers; make each instruction that
3352 calls a constant function contain the function's address explicitly.
3354 This option results in less efficient code, but some strange hacks
3355 that alter the assembler output may be confused by the optimizations
3356 performed when this option is not used.
3360 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3361 @option{-fno-trapping-math}.
3363 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3365 This option should never be turned on by any @option{-O} option since
3366 it can result in incorrect output for programs which depend on
3367 an exact implementation of IEEE or ISO rules/specifications for
3370 @item -fno-math-errno
3371 @opindex fno-math-errno
3372 Do not set ERRNO after calling math functions that are executed
3373 with a single instruction, e.g., sqrt. A program that relies on
3374 IEEE exceptions for math error handling may want to use this flag
3375 for speed while maintaining IEEE arithmetic compatibility.
3377 This option should never be turned on by any @option{-O} option since
3378 it can result in incorrect output for programs which depend on
3379 an exact implementation of IEEE or ISO rules/specifications for
3382 The default is @option{-fmath-errno}.
3384 @item -funsafe-math-optimizations
3385 @opindex funsafe-math-optimizations
3386 Allow optimizations for floating-point arithmetic that (a) assume
3387 that arguments and results are valid and (b) may violate IEEE or
3388 ANSI standards. When used at link-time, it may include libraries
3389 or startup files that change the default FPU control word or other
3390 similar optimizations.
3392 This option should never be turned on by any @option{-O} option since
3393 it can result in incorrect output for programs which depend on
3394 an exact implementation of IEEE or ISO rules/specifications for
3397 The default is @option{-fno-unsafe-math-optimizations}.
3399 @item -fno-trapping-math
3400 @opindex fno-trapping-math
3401 Compile code assuming that floating-point operations cannot generate
3402 user-visible traps. Setting this option may allow faster code
3403 if one relies on ``non-stop'' IEEE arithmetic, for example.
3405 This option should never be turned on by any @option{-O} option since
3406 it can result in incorrect output for programs which depend on
3407 an exact implementation of IEEE or ISO rules/specifications for
3410 The default is @option{-ftrapping-math}.
3412 @item -fno-zero-initialized-in-bss
3413 @opindex fno-zero-initialized-in-bss
3414 If the target supports a BSS section, GCC by default puts variables that
3415 are initialized to zero into BSS@. This can save space in the resulting
3418 This option turns off this behavior because some programs explicitly
3419 rely on variables going to the data section. E.g., so that the
3420 resulting executable can find the beginning of that section and/or make
3421 assumptions based on that.
3423 The default is @option{-fzero-initialized-in-bss}.
3426 The following options control specific optimizations. The @option{-O2}
3427 option turns on all of these optimizations except @option{-funroll-loops}
3428 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3429 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3430 but specific machines may handle it differently.
3432 You can use the following flags in the rare cases when ``fine-tuning''
3433 of optimizations to be performed is desired.
3435 Not all of the optimizations performed by GCC have @option{-f} options
3439 @item -fstrength-reduce
3440 @opindex fstrength-reduce
3441 Perform the optimizations of loop strength reduction and
3442 elimination of iteration variables.
3444 @item -fthread-jumps
3445 @opindex fthread-jumps
3446 Perform optimizations where we check to see if a jump branches to a
3447 location where another comparison subsumed by the first is found. If
3448 so, the first branch is redirected to either the destination of the
3449 second branch or a point immediately following it, depending on whether
3450 the condition is known to be true or false.
3452 @item -fcse-follow-jumps
3453 @opindex fcse-follow-jumps
3454 In common subexpression elimination, scan through jump instructions
3455 when the target of the jump is not reached by any other path. For
3456 example, when CSE encounters an @code{if} statement with an
3457 @code{else} clause, CSE will follow the jump when the condition
3460 @item -fcse-skip-blocks
3461 @opindex fcse-skip-blocks
3462 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3463 follow jumps which conditionally skip over blocks. When CSE
3464 encounters a simple @code{if} statement with no else clause,
3465 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3466 body of the @code{if}.
3468 @item -frerun-cse-after-loop
3469 @opindex frerun-cse-after-loop
3470 Re-run common subexpression elimination after loop optimizations has been
3473 @item -frerun-loop-opt
3474 @opindex frerun-loop-opt
3475 Run the loop optimizer twice.
3479 Perform a global common subexpression elimination pass.
3480 This pass also performs global constant and copy propagation.
3482 @emph{Note:} When compiling a program using computed gotos, a GCC
3483 extension, you may get better runtime performance if you disable
3484 the global common subexpression elmination pass by adding
3485 @option{-fno-gcse} to the command line.
3489 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3490 attempt to move loads which are only killed by stores into themselves. This
3491 allows a loop containing a load/store sequence to be changed to a load outside
3492 the loop, and a copy/store within the loop.
3496 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3497 subexpression elimination. This pass will attempt to move stores out of loops.
3498 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3499 can be changed to a load before the loop and a store after the loop.
3501 @item -fdelete-null-pointer-checks
3502 @opindex fdelete-null-pointer-checks
3503 Use global dataflow analysis to identify and eliminate useless checks
3504 for null pointers. The compiler assumes that dereferencing a null
3505 pointer would have halted the program. If a pointer is checked after
3506 it has already been dereferenced, it cannot be null.
3508 In some environments, this assumption is not true, and programs can
3509 safely dereference null pointers. Use
3510 @option{-fno-delete-null-pointer-checks} to disable this optimization
3511 for programs which depend on that behavior.
3513 @item -fexpensive-optimizations
3514 @opindex fexpensive-optimizations
3515 Perform a number of minor optimizations that are relatively expensive.
3517 @item -foptimize-register-move
3519 @opindex foptimize-register-move
3521 Attempt to reassign register numbers in move instructions and as
3522 operands of other simple instructions in order to maximize the amount of
3523 register tying. This is especially helpful on machines with two-operand
3524 instructions. GCC enables this optimization by default with @option{-O2}
3527 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3530 @item -fdelayed-branch
3531 @opindex fdelayed-branch
3532 If supported for the target machine, attempt to reorder instructions
3533 to exploit instruction slots available after delayed branch
3536 @item -fschedule-insns
3537 @opindex fschedule-insns
3538 If supported for the target machine, attempt to reorder instructions to
3539 eliminate execution stalls due to required data being unavailable. This
3540 helps machines that have slow floating point or memory load instructions
3541 by allowing other instructions to be issued until the result of the load
3542 or floating point instruction is required.
3544 @item -fschedule-insns2
3545 @opindex fschedule-insns2
3546 Similar to @option{-fschedule-insns}, but requests an additional pass of
3547 instruction scheduling after register allocation has been done. This is
3548 especially useful on machines with a relatively small number of
3549 registers and where memory load instructions take more than one cycle.
3551 @item -ffunction-sections
3552 @itemx -fdata-sections
3553 @opindex ffunction-sections
3554 @opindex fdata-sections
3555 Place each function or data item into its own section in the output
3556 file if the target supports arbitrary sections. The name of the
3557 function or the name of the data item determines the section's name
3560 Use these options on systems where the linker can perform optimizations
3561 to improve locality of reference in the instruction space. HPPA
3562 processors running HP-UX and Sparc processors running Solaris 2 have
3563 linkers with such optimizations. Other systems using the ELF object format
3564 as well as AIX may have these optimizations in the future.
3566 Only use these options when there are significant benefits from doing
3567 so. When you specify these options, the assembler and linker will
3568 create larger object and executable files and will also be slower.
3569 You will not be able to use @code{gprof} on all systems if you
3570 specify this option and you may have problems with debugging if
3571 you specify both this option and @option{-g}.
3573 @item -fcaller-saves
3574 @opindex fcaller-saves
3575 Enable values to be allocated in registers that will be clobbered by
3576 function calls, by emitting extra instructions to save and restore the
3577 registers around such calls. Such allocation is done only when it
3578 seems to result in better code than would otherwise be produced.
3580 This option is always enabled by default on certain machines, usually
3581 those which have no call-preserved registers to use instead.
3583 For all machines, optimization level 2 and higher enables this flag by
3586 @item -funroll-loops
3587 @opindex funroll-loops
3588 Unroll loops whose number of iterations can be determined at compile
3589 time or upon entry to the loop. @option{-funroll-loops} implies both
3590 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3591 option makes code larger, and may or may not make it run faster.
3593 @item -funroll-all-loops
3594 @opindex funroll-all-loops
3595 Unroll all loops, even if their number of iterations is uncertain when
3596 the loop is entered. This usually makes programs run more slowly.
3597 @option{-funroll-all-loops} implies the same options as
3598 @option{-funroll-loops},
3600 @item -fprefetch-loop-arrays
3601 @opindex fprefetch-loop-arrays
3602 If supported by the target machine, generate instructions to prefetch
3603 memory to improve the performance of loops that access large arrays.
3605 @item -fmove-all-movables
3606 @opindex fmove-all-movables
3607 Forces all invariant computations in loops to be moved
3610 @item -freduce-all-givs
3611 @opindex freduce-all-givs
3612 Forces all general-induction variables in loops to be
3615 @emph{Note:} When compiling programs written in Fortran,
3616 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3617 by default when you use the optimizer.
3619 These options may generate better or worse code; results are highly
3620 dependent on the structure of loops within the source code.
3622 These two options are intended to be removed someday, once
3623 they have helped determine the efficacy of various
3624 approaches to improving loop optimizations.
3626 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3627 know how use of these options affects
3628 the performance of your production code.
3629 We're very interested in code that runs @emph{slower}
3630 when these options are @emph{enabled}.
3633 @itemx -fno-peephole2
3634 @opindex fno-peephole
3635 @opindex fno-peephole2
3636 Disable any machine-specific peephole optimizations. The difference
3637 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3638 are implemented in the compiler; some targets use one, some use the
3639 other, a few use both.
3641 @item -fbranch-probabilities
3642 @opindex fbranch-probabilities
3643 After running a program compiled with @option{-fprofile-arcs}
3644 (@pxref{Debugging Options,, Options for Debugging Your Program or
3645 @command{gcc}}), you can compile it a second time using
3646 @option{-fbranch-probabilities}, to improve optimizations based on
3647 the number of times each branch was taken. When the program
3648 compiled with @option{-fprofile-arcs} exits it saves arc execution
3649 counts to a file called @file{@var{sourcename}.da} for each source
3650 file The information in this data file is very dependent on the
3651 structure of the generated code, so you must use the same source code
3652 and the same optimization options for both compilations.
3654 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3655 note on the first instruction of each basic block, and a
3656 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3657 These can be used to improve optimization. Currently, they are only
3658 used in one place: in @file{reorg.c}, instead of guessing which path a
3659 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3660 exactly determine which path is taken more often.
3662 @item -fno-guess-branch-probability
3663 @opindex fno-guess-branch-probability
3664 Do not guess branch probabilities using a randomized model.
3666 Sometimes gcc will opt to use a randomized model to guess branch
3667 probabilities, when none are available from either profiling feedback
3668 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3669 different runs of the compiler on the same program may produce different
3672 In a hard real-time system, people don't want different runs of the
3673 compiler to produce code that has different behavior; minimizing
3674 non-determinism is of paramount import. This switch allows users to
3675 reduce non-determinism, possibly at the expense of inferior
3678 @item -fstrict-aliasing
3679 @opindex fstrict-aliasing
3680 Allows the compiler to assume the strictest aliasing rules applicable to
3681 the language being compiled. For C (and C++), this activates
3682 optimizations based on the type of expressions. In particular, an
3683 object of one type is assumed never to reside at the same address as an
3684 object of a different type, unless the types are almost the same. For
3685 example, an @code{unsigned int} can alias an @code{int}, but not a
3686 @code{void*} or a @code{double}. A character type may alias any other
3689 Pay special attention to code like this:
3702 The practice of reading from a different union member than the one most
3703 recently written to (called ``type-punning'') is common. Even with
3704 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3705 is accessed through the union type. So, the code above will work as
3706 expected. However, this code might not:
3717 Every language that wishes to perform language-specific alias analysis
3718 should define a function that computes, given an @code{tree}
3719 node, an alias set for the node. Nodes in different alias sets are not
3720 allowed to alias. For an example, see the C front-end function
3721 @code{c_get_alias_set}.
3723 @item -falign-functions
3724 @itemx -falign-functions=@var{n}
3725 @opindex falign-functions
3726 Align the start of functions to the next power-of-two greater than
3727 @var{n}, skipping up to @var{n} bytes. For instance,
3728 @option{-falign-functions=32} aligns functions to the next 32-byte
3729 boundary, but @option{-falign-functions=24} would align to the next
3730 32-byte boundary only if this can be done by skipping 23 bytes or less.
3732 @option{-fno-align-functions} and @option{-falign-functions=1} are
3733 equivalent and mean that functions will not be aligned.
3735 Some assemblers only support this flag when @var{n} is a power of two;
3736 in that case, it is rounded up.
3738 If @var{n} is not specified, use a machine-dependent default.
3740 @item -falign-labels
3741 @itemx -falign-labels=@var{n}
3742 @opindex falign-labels
3743 Align all branch targets to a power-of-two boundary, skipping up to
3744 @var{n} bytes like @option{-falign-functions}. This option can easily
3745 make code slower, because it must insert dummy operations for when the
3746 branch target is reached in the usual flow of the code.
3748 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3749 are greater than this value, then their values are used instead.
3751 If @var{n} is not specified, use a machine-dependent default which is
3752 very likely to be @samp{1}, meaning no alignment.
3755 @itemx -falign-loops=@var{n}
3756 @opindex falign-loops
3757 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3758 like @option{-falign-functions}. The hope is that the loop will be
3759 executed many times, which will make up for any execution of the dummy
3762 If @var{n} is not specified, use a machine-dependent default.
3765 @itemx -falign-jumps=@var{n}
3766 @opindex falign-jumps
3767 Align branch targets to a power-of-two boundary, for branch targets
3768 where the targets can only be reached by jumping, skipping up to @var{n}
3769 bytes like @option{-falign-functions}. In this case, no dummy operations
3772 If @var{n} is not specified, use a machine-dependent default.
3776 Perform optimizations in static single assignment form. Each function's
3777 flow graph is translated into SSA form, optimizations are performed, and
3778 the flow graph is translated back from SSA form. Users should not
3779 specify this option, since it is not yet ready for production use.
3783 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3784 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3788 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3789 Like @option{-fssa}, this is an experimental feature.
3791 @item -fsingle-precision-constant
3792 @opindex fsingle-precision-constant
3793 Treat floating point constant as single precision constant instead of
3794 implicitly converting it to double precision constant.
3796 @item -frename-registers
3797 @opindex frename-registers
3798 Attempt to avoid false dependencies in scheduled code by making use
3799 of registers left over after register allocation. This optimization
3800 will most benefit processors with lots of registers. It can, however,
3801 make debugging impossible, since variables will no longer stay in
3802 a ``home register''.
3804 @item -fno-cprop-registers
3805 @opindex fno-cprop-registers
3806 After register allocation and post-register allocation instruction splitting,
3807 we perform a copy-propagation pass to try to reduce scheduling dependencies
3808 and occasionally eliminate the copy.
3810 @item --param @var{name}=@var{value}
3812 In some places, GCC uses various constants to control the amount of
3813 optimization that is done. For example, GCC will not inline functions
3814 that contain more that a certain number of instructions. You can
3815 control some of these constants on the command-line using the
3816 @option{--param} option.
3818 In each case, the @var{value} is an integer. The allowable choices for
3819 @var{name} are given in the following table:
3822 @item max-delay-slot-insn-search
3823 The maximum number of instructions to consider when looking for an
3824 instruction to fill a delay slot. If more than this arbitrary number of
3825 instructions is searched, the time savings from filling the delay slot
3826 will be minimal so stop searching. Increasing values mean more
3827 aggressive optimization, making the compile time increase with probably
3828 small improvement in executable run time.
3830 @item max-delay-slot-live-search
3831 When trying to fill delay slots, the maximum number of instructions to
3832 consider when searching for a block with valid live register
3833 information. Increasing this arbitrarily chosen value means more
3834 aggressive optimization, increasing the compile time. This parameter
3835 should be removed when the delay slot code is rewritten to maintain the
3838 @item max-gcse-memory
3839 The approximate maximum amount of memory that will be allocated in
3840 order to perform the global common subexpression elimination
3841 optimization. If more memory than specified is required, the
3842 optimization will not be done.
3844 @item max-gcse-passes
3845 The maximum number of passes of GCSE to run.
3847 @item max-pending-list-length
3848 The maximum number of pending dependencies scheduling will allow
3849 before flushing the current state and starting over. Large functions
3850 with few branches or calls can create excessively large lists which
3851 needlessly consume memory and resources.
3853 @item max-inline-insns
3854 If an function contains more than this many instructions, it
3855 will not be inlined. This option is precisely equivalent to
3856 @option{-finline-limit}.
3861 @node Preprocessor Options
3862 @section Options Controlling the Preprocessor
3863 @cindex preprocessor options
3864 @cindex options, preprocessor
3866 These options control the C preprocessor, which is run on each C source
3867 file before actual compilation.
3869 If you use the @option{-E} option, nothing is done except preprocessing.
3870 Some of these options make sense only together with @option{-E} because
3871 they cause the preprocessor output to be unsuitable for actual
3875 @item -include @var{file}
3877 Process @var{file} as input before processing the regular input file.
3878 In effect, the contents of @var{file} are compiled first. Any @option{-D}
3879 and @option{-U} options on the command line are always processed before
3880 @option{-include @var{file}}, regardless of the order in which they are
3881 written. All the @option{-include} and @option{-imacros} options are
3882 processed in the order in which they are written.
3884 @item -imacros @var{file}
3886 Process @var{file} as input, discarding the resulting output, before
3887 processing the regular input file. Because the output generated from
3888 @var{file} is discarded, the only effect of @option{-imacros @var{file}}
3889 is to make the macros defined in @var{file} available for use in the
3890 main input. All the @option{-include} and @option{-imacros} options are
3891 processed in the order in which they are written.
3893 @item -idirafter @var{dir}
3895 @cindex second include path
3896 Add the directory @var{dir} to the second include path. The directories
3897 on the second include path are searched when a header file is not found
3898 in any of the directories in the main include path (the one that
3899 @option{-I} adds to).
3901 @item -iprefix @var{prefix}
3903 Specify @var{prefix} as the prefix for subsequent @option{-iwithprefix}
3906 @item -iwithprefix @var{dir}
3907 @opindex iwithprefix
3908 Add a directory to the second include path. The directory's name is
3909 made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was
3910 specified previously with @option{-iprefix}. If you have not specified a
3911 prefix yet, the directory containing the installed passes of the
3912 compiler is used as the default.
3914 @item -iwithprefixbefore @var{dir}
3915 @opindex iwithprefixbefore
3916 Add a directory to the main include path. The directory's name is made
3917 by concatenating @var{prefix} and @var{dir}, as in the case of
3918 @option{-iwithprefix}.
3920 @item -isystem @var{dir}
3922 Add a directory to the beginning of the second include path, marking it
3923 as a system directory, so that it gets the same special treatment as
3924 is applied to the standard system directories.
3928 Do not search the standard system directories for header files. Only
3929 the directories you have specified with @option{-I} options (and the
3930 current directory, if appropriate) are searched. @xref{Directory
3931 Options}, for information on @option{-I}.
3933 By using both @option{-nostdinc} and @option{-I-}, you can limit the include-file
3934 search path to only those directories you specify explicitly.
3938 When searching for a header file in a directory, remap file names if a
3939 file named @file{header.gcc} exists in that directory. This can be used
3940 to work around limitations of file systems with file name restrictions.
3941 The @file{header.gcc} file should contain a series of lines with two
3942 tokens on each line: the first token is the name to map, and the second
3943 token is the actual name to use.
3947 Do not predefine any nonstandard macros. (Including architecture flags).
3951 Run only the C preprocessor. Preprocess all the C source files
3952 specified and output the results to standard output or to the
3953 specified output file.
3957 Tell the preprocessor not to discard comments. Used with the
3962 Tell the preprocessor not to generate @samp{#line} directives.
3963 Used with the @option{-E} option.
3966 @cindex dependencies, make
3969 Instead of outputting the result of preprocessing, output a rule
3970 suitable for @command{make} describing the dependencies of the main
3971 source file. The preprocessor outputs one @command{make} rule containing
3972 the object file name for that source file, a colon, and the names of all
3973 the included files, including those coming from @option{-include} or
3974 @option{-imacros} command line options.
3976 Unless specified explicitly (with @option{-MT} or @option{-MQ}), the
3977 object file name consists of the basename of the source file with any
3978 suffix replaced with object file suffix. If there are many included
3979 files then the rule is split into several lines using @samp{\}-newline.
3980 The rule has no commands.
3982 Passing @option{-M} to the driver implies @option{-E}.
3986 Like @option{-M} but do not mention header files that are found in
3987 system header directories, nor header files that are included,
3988 directly or indirectly, from such a header.
3990 This implies that the choice of angle brackets or double quotes in an
3991 @samp{#include} directive does not in itself determine whether that
3992 header will appear in @option{-MM} dependency output. This is a
3993 slight change in semantics from GCC versions 3.0 and earlier.
3997 @option{-MD} is equivalent to @option{-M -MF @var{file}}, except that
3998 @option{-E} is not implied. The driver determines @var{file} based on
3999 whether an @option{-o} option is given. If it is, the driver uses its
4000 argument but with a suffix of @file{.d}, otherwise it take the
4001 basename of the input file and applies a @file{.d} suffix.
4003 If @option{-MD} is used in conjunction with @option{-E}, any
4004 @option{-o} switch is understood to specify the dependency output file
4005 (but @pxref{-MF}), but if used without @option{-E}, each @option{-o}
4006 is understood to specify a target object file.
4008 Since @option{-E} is not implied, @option{-MD} can be used to generate
4009 a dependency output file as a side-effect of the compilation process.
4011 With Mach, you can use the utility @code{md} to merge multiple
4012 dependency files into a single dependency file suitable for using with
4013 the @samp{make} command.
4017 Like @option{-MD} except mention only user header files, not system
4020 @item -MF @var{file}
4023 When used with @option{-M} or @option{-MM}, specifies a
4024 file to write the dependencies to. If no @option{-MF} switch is given
4025 the preprocessor sends the rules to the same place it would have sent
4026 preprocessed output.
4028 When used with the driver options @option{-MD} or @option{-MMD},
4029 @option{-MF} overrides the default dependency output file.
4031 Another way to specify output of a @code{make} rule is by setting
4032 the environment variable @env{DEPENDENCIES_OUTPUT} (@pxref{Environment
4037 When used with @option{-M} or @option{-MM}, @option{-MG} says to treat missing
4038 header files as generated files and assume they live in the same
4039 directory as the source file. It suppresses preprocessed output, as a
4040 missing header file is ordinarily an error.
4042 This feature is used in automatic updating of makefiles.
4046 This option instructs CPP to add a phony target for each dependency
4047 other than the main file, causing each to depend on nothing. These
4048 dummy rules work around errors @code{make} gives if you remove header
4049 files without updating the @code{Makefile} to match.
4051 This is typical output:-
4054 /tmp/test.o: /tmp/test.c /tmp/test.h
4059 @item -MQ @var{target}
4060 @item -MT @var{target}
4063 By default CPP uses the main file name, including any path, and appends
4064 the object suffix, normally ``.o'', to it to obtain the name of the
4065 target for dependency generation. With @option{-MT} you can specify a
4066 target yourself, overriding the default one.
4068 If you want multiple targets, you can specify them as a single argument
4069 to @option{-MT}, or use multiple @option{-MT} options.
4071 The targets you specify are output in the order they appear on the
4072 command line. @option{-MQ} is identical to @option{-MT}, except that the
4073 target name is quoted for Make, but with @option{-MT} it isn't. For
4074 example, @option{-MT '$(objpfx)foo.o'} gives
4077 $(objpfx)foo.o: /tmp/foo.c
4080 but @option{-MQ '$(objpfx)foo.o'} gives
4083 $$(objpfx)foo.o: /tmp/foo.c
4086 The default target is automatically quoted, as if it were given with
4091 Print the name of each header file used, in addition to other normal
4094 @item -A@var{question}(@var{answer})
4096 Assert the answer @var{answer} for @var{question}, in case it is tested
4097 with a preprocessing conditional such as @samp{#if
4098 #@var{question}(@var{answer})}. @option{-A-} disables the standard
4099 assertions that normally describe the target machine.
4103 Define macro @var{macro} with the string @samp{1} as its definition.
4105 @item -D@var{macro}=@var{defn}
4106 Define macro @var{macro} as @var{defn}. All instances of @option{-D} on
4107 the command line are processed before any @option{-U} options.
4109 Any @option{-D} and @option{-U} options on the command line are processed in
4110 order, and always before @option{-imacros @var{file}}, regardless of the
4111 order in which they are written.
4115 Undefine macro @var{macro}. @option{-U} options are evaluated after all
4116 @option{-D} options, but before any @option{-include} and @option{-imacros}
4119 Any @option{-D} and @option{-U} options on the command line are processed in
4120 order, and always before @option{-imacros @var{file}}, regardless of the
4121 order in which they are written.
4125 Tell the preprocessor to output only a list of the macro definitions
4126 that are in effect at the end of preprocessing. Used with the @option{-E}
4131 Tell the preprocessing to pass all macro definitions into the output, in
4132 their proper sequence in the rest of the output.
4136 Like @option{-dD} except that the macro arguments and contents are omitted.
4137 Only @samp{#define @var{name}} is included in the output.
4141 Output @samp{#include} directives in addition to the result of
4144 @item -fpreprocessed
4145 @opindex fpreprocessed
4146 Indicate to the preprocessor that the input file has already been
4147 preprocessed. This suppresses things like macro expansion, trigraph
4148 conversion, escaped newline splicing, and processing of most directives.
4149 The preprocessor still recognizes and removes comments, so that you can
4150 pass a file preprocessed with @option{-C} to the compiler without
4151 problems. In this mode the integrated preprocessor is little more than
4152 a tokenizer for the front ends.
4154 @option{-fpreprocessed} is implicit if the input file has one of the
4155 extensions @samp{i}, @samp{ii} or @samp{mi}. These are the extensions
4156 that GCC uses for preprocessed files created by @option{-save-temps}.
4160 Process ISO standard trigraph sequences. These are three-character
4161 sequences, all starting with @samp{??}, that are defined by ISO C to
4162 stand for single characters. For example, @samp{??/} stands for
4163 @samp{\}, so @samp{'??/n'} is a character constant for a newline. By
4164 default, GCC ignores trigraphs, but in standard-conforming modes it
4165 converts them. See the @option{-std} and @option{-ansi} options.
4167 The nine trigraph sequences are
4170 @expansion{} @samp{[}
4173 @expansion{} @samp{]}
4176 @expansion{} @samp{@{}
4179 @expansion{} @samp{@}}
4182 @expansion{} @samp{#}
4185 @expansion{} @samp{\}
4188 @expansion{} @samp{^}
4191 @expansion{} @samp{|}
4194 @expansion{} @samp{~}
4198 Trigraph support is not popular, so many compilers do not implement it
4199 properly. Portable code should not rely on trigraphs being either
4200 converted or ignored.
4202 @item -Wp,@var{option}
4204 Pass @var{option} as an option to the preprocessor. If @var{option}
4205 contains commas, it is split into multiple options at the commas.
4208 @node Assembler Options
4209 @section Passing Options to the Assembler
4211 @c prevent bad page break with this line
4212 You can pass options to the assembler.
4215 @item -Wa,@var{option}
4217 Pass @var{option} as an option to the assembler. If @var{option}
4218 contains commas, it is split into multiple options at the commas.
4222 @section Options for Linking
4223 @cindex link options
4224 @cindex options, linking
4226 These options come into play when the compiler links object files into
4227 an executable output file. They are meaningless if the compiler is
4228 not doing a link step.
4232 @item @var{object-file-name}
4233 A file name that does not end in a special recognized suffix is
4234 considered to name an object file or library. (Object files are
4235 distinguished from libraries by the linker according to the file
4236 contents.) If linking is done, these object files are used as input
4245 If any of these options is used, then the linker is not run, and
4246 object file names should not be used as arguments. @xref{Overall
4250 @item -l@var{library}
4251 @itemx -l @var{library}
4253 Search the library named @var{library} when linking. (The second
4254 alternative with the library as a separate argument is only for
4255 POSIX compliance and is not recommended.)
4257 It makes a difference where in the command you write this option; the
4258 linker searches and processes libraries and object files in the order they
4259 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4260 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4261 to functions in @samp{z}, those functions may not be loaded.
4263 The linker searches a standard list of directories for the library,
4264 which is actually a file named @file{lib@var{library}.a}. The linker
4265 then uses this file as if it had been specified precisely by name.
4267 The directories searched include several standard system directories
4268 plus any that you specify with @option{-L}.
4270 Normally the files found this way are library files---archive files
4271 whose members are object files. The linker handles an archive file by
4272 scanning through it for members which define symbols that have so far
4273 been referenced but not defined. But if the file that is found is an
4274 ordinary object file, it is linked in the usual fashion. The only
4275 difference between using an @option{-l} option and specifying a file name
4276 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4277 and searches several directories.
4281 You need this special case of the @option{-l} option in order to
4282 link an Objective-C program.
4285 @opindex nostartfiles
4286 Do not use the standard system startup files when linking.
4287 The standard system libraries are used normally, unless @option{-nostdlib}
4288 or @option{-nodefaultlibs} is used.
4290 @item -nodefaultlibs
4291 @opindex nodefaultlibs
4292 Do not use the standard system libraries when linking.
4293 Only the libraries you specify will be passed to the linker.
4294 The standard startup files are used normally, unless @option{-nostartfiles}
4295 is used. The compiler may generate calls to memcmp, memset, and memcpy
4296 for System V (and ISO C) environments or to bcopy and bzero for
4297 BSD environments. These entries are usually resolved by entries in
4298 libc. These entry points should be supplied through some other
4299 mechanism when this option is specified.
4303 Do not use the standard system startup files or libraries when linking.
4304 No startup files and only the libraries you specify will be passed to
4305 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4306 for System V (and ISO C) environments or to bcopy and bzero for
4307 BSD environments. These entries are usually resolved by entries in
4308 libc. These entry points should be supplied through some other
4309 mechanism when this option is specified.
4311 @cindex @option{-lgcc}, use with @option{-nostdlib}
4312 @cindex @option{-nostdlib} and unresolved references
4313 @cindex unresolved references and @option{-nostdlib}
4314 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4315 @cindex @option{-nodefaultlibs} and unresolved references
4316 @cindex unresolved references and @option{-nodefaultlibs}
4317 One of the standard libraries bypassed by @option{-nostdlib} and
4318 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4319 that GCC uses to overcome shortcomings of particular machines, or special
4320 needs for some languages.
4321 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4322 Collection (GCC) Internals},
4323 for more discussion of @file{libgcc.a}.)
4324 In most cases, you need @file{libgcc.a} even when you want to avoid
4325 other standard libraries. In other words, when you specify @option{-nostdlib}
4326 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4327 This ensures that you have no unresolved references to internal GCC
4328 library subroutines. (For example, @samp{__main}, used to ensure C++
4329 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4330 GNU Compiler Collection (GCC) Internals}.)
4334 Remove all symbol table and relocation information from the executable.
4338 On systems that support dynamic linking, this prevents linking with the shared
4339 libraries. On other systems, this option has no effect.
4343 Produce a shared object which can then be linked with other objects to
4344 form an executable. Not all systems support this option. For predictable
4345 results, you must also specify the same set of options that were used to
4346 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4347 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4348 needs to build supplementary stub code for constructors to work. On
4349 multi-libbed systems, @samp{gcc -shared} must select the correct support
4350 libraries to link against. Failing to supply the correct flags may lead
4351 to subtle defects. Supplying them in cases where they are not necessary
4354 @item -shared-libgcc
4355 @itemx -static-libgcc
4356 @opindex shared-libgcc
4357 @opindex static-libgcc
4358 On systems that provide @file{libgcc} as a shared library, these options
4359 force the use of either the shared or static version respectively.
4360 If no shared version of @file{libgcc} was built when the compiler was
4361 configured, these options have no effect.
4363 There are several situations in which an application should use the
4364 shared @file{libgcc} instead of the static version. The most common
4365 of these is when the application wishes to throw and catch exceptions
4366 across different shared libraries. In that case, each of the libraries
4367 as well as the application itself should use the shared @file{libgcc}.
4369 Therefore, the G++ and GCJ drivers automatically add
4370 @option{-shared-libgcc} whenever you build a shared library or a main
4371 executable, because C++ and Java programs typically use exceptions, so
4372 this is the right thing to do.
4374 If, instead, you use the GCC driver to create shared libraries, you may
4375 find that they will not always be linked with the shared @file{libgcc}.
4376 If GCC finds, at its configuration time, that you have a GNU linker that
4377 does not support option @option{--eh-frame-hdr}, it will link the shared
4378 version of @file{libgcc} into shared libraries by default. Otherwise,
4379 it will take advantage of the linker and optimize away the linking with
4380 the shared version of @file{libgcc}, linking with the static version of
4381 libgcc by default. This allows exceptions to propagate through such
4382 shared libraries, without incurring relocation costs at library load
4385 However, if a library or main executable is supposed to throw or catch
4386 exceptions, you must link it using the G++ or GCJ driver, as appropriate
4387 for the languages used in the program, or using the option
4388 @option{-shared-libgcc}, such that it is linked with the shared
4393 Bind references to global symbols when building a shared object. Warn
4394 about any unresolved references (unless overridden by the link editor
4395 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4398 @item -Xlinker @var{option}
4400 Pass @var{option} as an option to the linker. You can use this to
4401 supply system-specific linker options which GCC does not know how to
4404 If you want to pass an option that takes an argument, you must use
4405 @option{-Xlinker} twice, once for the option and once for the argument.
4406 For example, to pass @option{-assert definitions}, you must write
4407 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4408 @option{-Xlinker "-assert definitions"}, because this passes the entire
4409 string as a single argument, which is not what the linker expects.
4411 @item -Wl,@var{option}
4413 Pass @var{option} as an option to the linker. If @var{option} contains
4414 commas, it is split into multiple options at the commas.
4416 @item -u @var{symbol}
4418 Pretend the symbol @var{symbol} is undefined, to force linking of
4419 library modules to define it. You can use @option{-u} multiple times with
4420 different symbols to force loading of additional library modules.
4423 @node Directory Options
4424 @section Options for Directory Search
4425 @cindex directory options
4426 @cindex options, directory search
4429 These options specify directories to search for header files, for
4430 libraries and for parts of the compiler:
4435 Add the directory @var{dir} to the head of the list of directories to be
4436 searched for header files. This can be used to override a system header
4437 file, substituting your own version, since these directories are
4438 searched before the system header file directories. However, you should
4439 not use this option to add directories that contain vendor-supplied
4440 system header files (use @option{-isystem} for that). If you use more than
4441 one @option{-I} option, the directories are scanned in left-to-right
4442 order; the standard system directories come after.
4444 If a standard system include directory, or a directory specified with
4445 @option{-isystem}, is also specified with @option{-I}, it will be
4446 searched only in the position requested by @option{-I}. Also, it will
4447 not be considered a system include directory. If that directory really
4448 does contain system headers, there is a good chance that they will
4449 break. For instance, if GCC's installation procedure edited the headers
4450 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4451 original, buggy headers to be found instead of the corrected ones. GCC
4452 will issue a warning when a system include directory is hidden in this
4457 Any directories you specify with @option{-I} options before the @option{-I-}
4458 option are searched only for the case of @samp{#include "@var{file}"};
4459 they are not searched for @samp{#include <@var{file}>}.
4461 If additional directories are specified with @option{-I} options after
4462 the @option{-I-}, these directories are searched for all @samp{#include}
4463 directives. (Ordinarily @emph{all} @option{-I} directories are used
4466 In addition, the @option{-I-} option inhibits the use of the current
4467 directory (where the current input file came from) as the first search
4468 directory for @samp{#include "@var{file}"}. There is no way to
4469 override this effect of @option{-I-}. With @option{-I.} you can specify
4470 searching the directory which was current when the compiler was
4471 invoked. That is not exactly the same as what the preprocessor does
4472 by default, but it is often satisfactory.
4474 @option{-I-} does not inhibit the use of the standard system directories
4475 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4480 Add directory @var{dir} to the list of directories to be searched
4483 @item -B@var{prefix}
4485 This option specifies where to find the executables, libraries,
4486 include files, and data files of the compiler itself.
4488 The compiler driver program runs one or more of the subprograms
4489 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4490 @var{prefix} as a prefix for each program it tries to run, both with and
4491 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4493 For each subprogram to be run, the compiler driver first tries the
4494 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4495 was not specified, the driver tries two standard prefixes, which are
4496 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4497 those results in a file name that is found, the unmodified program
4498 name is searched for using the directories specified in your
4499 @env{PATH} environment variable.
4501 The compiler will check to see if the path provided by the @option{-B}
4502 refers to a directory, and if necessary it will add a directory
4503 separator character at the end of the path.
4505 @option{-B} prefixes that effectively specify directory names also apply
4506 to libraries in the linker, because the compiler translates these
4507 options into @option{-L} options for the linker. They also apply to
4508 includes files in the preprocessor, because the compiler translates these
4509 options into @option{-isystem} options for the preprocessor. In this case,
4510 the compiler appends @samp{include} to the prefix.
4512 The run-time support file @file{libgcc.a} can also be searched for using
4513 the @option{-B} prefix, if needed. If it is not found there, the two
4514 standard prefixes above are tried, and that is all. The file is left
4515 out of the link if it is not found by those means.
4517 Another way to specify a prefix much like the @option{-B} prefix is to use
4518 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4521 As a special kludge, if the path provided by @option{-B} is
4522 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4523 9, then it will be replaced by @file{[dir/]include}. This is to help
4524 with boot-strapping the compiler.
4526 @item -specs=@var{file}
4528 Process @var{file} after the compiler reads in the standard @file{specs}
4529 file, in order to override the defaults that the @file{gcc} driver
4530 program uses when determining what switches to pass to @file{cc1},
4531 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4532 @option{-specs=@var{file}} can be specified on the command line, and they
4533 are processed in order, from left to right.
4539 @section Specifying subprocesses and the switches to pass to them
4541 @command{gcc} is a driver program. It performs its job by invoking a
4542 sequence of other programs to do the work of compiling, assembling and
4543 linking. GCC interprets its command-line parameters and uses these to
4544 deduce which programs it should invoke, and which command-line options
4545 it ought to place on their command lines. This behavior is controlled
4546 by @dfn{spec strings}. In most cases there is one spec string for each
4547 program that GCC can invoke, but a few programs have multiple spec
4548 strings to control their behavior. The spec strings built into GCC can
4549 be overridden by using the @option{-specs=} command-line switch to specify
4552 @dfn{Spec files} are plaintext files that are used to construct spec
4553 strings. They consist of a sequence of directives separated by blank
4554 lines. The type of directive is determined by the first non-whitespace
4555 character on the line and it can be one of the following:
4558 @item %@var{command}
4559 Issues a @var{command} to the spec file processor. The commands that can
4563 @item %include <@var{file}>
4565 Search for @var{file} and insert its text at the current point in the
4568 @item %include_noerr <@var{file}>
4569 @cindex %include_noerr
4570 Just like @samp{%include}, but do not generate an error message if the include
4571 file cannot be found.
4573 @item %rename @var{old_name} @var{new_name}
4575 Rename the spec string @var{old_name} to @var{new_name}.
4579 @item *[@var{spec_name}]:
4580 This tells the compiler to create, override or delete the named spec
4581 string. All lines after this directive up to the next directive or
4582 blank line are considered to be the text for the spec string. If this
4583 results in an empty string then the spec will be deleted. (Or, if the
4584 spec did not exist, then nothing will happened.) Otherwise, if the spec
4585 does not currently exist a new spec will be created. If the spec does
4586 exist then its contents will be overridden by the text of this
4587 directive, unless the first character of that text is the @samp{+}
4588 character, in which case the text will be appended to the spec.
4590 @item [@var{suffix}]:
4591 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4592 and up to the next directive or blank line are considered to make up the
4593 spec string for the indicated suffix. When the compiler encounters an
4594 input file with the named suffix, it will processes the spec string in
4595 order to work out how to compile that file. For example:
4602 This says that any input file whose name ends in @samp{.ZZ} should be
4603 passed to the program @samp{z-compile}, which should be invoked with the
4604 command-line switch @option{-input} and with the result of performing the
4605 @samp{%i} substitution. (See below.)
4607 As an alternative to providing a spec string, the text that follows a
4608 suffix directive can be one of the following:
4611 @item @@@var{language}
4612 This says that the suffix is an alias for a known @var{language}. This is
4613 similar to using the @option{-x} command-line switch to GCC to specify a
4614 language explicitly. For example:
4621 Says that .ZZ files are, in fact, C++ source files.
4624 This causes an error messages saying:
4627 @var{name} compiler not installed on this system.
4631 GCC already has an extensive list of suffixes built into it.
4632 This directive will add an entry to the end of the list of suffixes, but
4633 since the list is searched from the end backwards, it is effectively
4634 possible to override earlier entries using this technique.
4638 GCC has the following spec strings built into it. Spec files can
4639 override these strings or create their own. Note that individual
4640 targets can also add their own spec strings to this list.
4643 asm Options to pass to the assembler
4644 asm_final Options to pass to the assembler post-processor
4645 cpp Options to pass to the C preprocessor
4646 cc1 Options to pass to the C compiler
4647 cc1plus Options to pass to the C++ compiler
4648 endfile Object files to include at the end of the link
4649 link Options to pass to the linker
4650 lib Libraries to include on the command line to the linker
4651 libgcc Decides which GCC support library to pass to the linker
4652 linker Sets the name of the linker
4653 predefines Defines to be passed to the C preprocessor
4654 signed_char Defines to pass to CPP to say whether @code{char} is signed
4656 startfile Object files to include at the start of the link
4659 Here is a small example of a spec file:
4665 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4668 This example renames the spec called @samp{lib} to @samp{old_lib} and
4669 then overrides the previous definition of @samp{lib} with a new one.
4670 The new definition adds in some extra command-line options before
4671 including the text of the old definition.
4673 @dfn{Spec strings} are a list of command-line options to be passed to their
4674 corresponding program. In addition, the spec strings can contain
4675 @samp{%}-prefixed sequences to substitute variable text or to
4676 conditionally insert text into the command line. Using these constructs
4677 it is possible to generate quite complex command lines.
4679 Here is a table of all defined @samp{%}-sequences for spec
4680 strings. Note that spaces are not generated automatically around the
4681 results of expanding these sequences. Therefore you can concatenate them
4682 together or combine them with constant text in a single argument.
4686 Substitute one @samp{%} into the program name or argument.
4689 Substitute the name of the input file being processed.
4692 Substitute the basename of the input file being processed.
4693 This is the substring up to (and not including) the last period
4694 and not including the directory.
4697 This is the same as @samp{%b}, but include the file suffix (text after
4701 Marks the argument containing or following the @samp{%d} as a
4702 temporary file name, so that that file will be deleted if GCC exits
4703 successfully. Unlike @samp{%g}, this contributes no text to the
4706 @item %g@var{suffix}
4707 Substitute a file name that has suffix @var{suffix} and is chosen
4708 once per compilation, and mark the argument in the same way as
4709 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4710 name is now chosen in a way that is hard to predict even when previously
4711 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4712 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4713 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4714 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4715 was simply substituted with a file name chosen once per compilation,
4716 without regard to any appended suffix (which was therefore treated
4717 just like ordinary text), making such attacks more likely to succeed.
4719 @item %u@var{suffix}
4720 Like @samp{%g}, but generates a new temporary file name even if
4721 @samp{%u@var{suffix}} was already seen.
4723 @item %U@var{suffix}
4724 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4725 new one if there is no such last file name. In the absence of any
4726 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4727 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4728 would involve the generation of two distinct file names, one
4729 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4730 simply substituted with a file name chosen for the previous @samp{%u},
4731 without regard to any appended suffix.
4733 @item %j@var{SUFFIX}
4734 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4735 writable, and if save-temps is off; otherwise, substitute the name
4736 of a temporary file, just like @samp{%u}. This temporary file is not
4737 meant for communication between processes, but rather as a junk
4740 @item %.@var{SUFFIX}
4741 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4742 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4743 terminated by the next space or %.
4746 Marks the argument containing or following the @samp{%w} as the
4747 designated output file of this compilation. This puts the argument
4748 into the sequence of arguments that @samp{%o} will substitute later.
4751 Substitutes the names of all the output files, with spaces
4752 automatically placed around them. You should write spaces
4753 around the @samp{%o} as well or the results are undefined.
4754 @samp{%o} is for use in the specs for running the linker.
4755 Input files whose names have no recognized suffix are not compiled
4756 at all, but they are included among the output files, so they will
4760 Substitutes the suffix for object files. Note that this is
4761 handled specially when it immediately follows @samp{%g, %u, or %U},
4762 because of the need for those to form complete file names. The
4763 handling is such that @samp{%O} is treated exactly as if it had already
4764 been substituted, except that @samp{%g, %u, and %U} do not currently
4765 support additional @var{suffix} characters following @samp{%O} as they would
4766 following, for example, @samp{.o}.
4769 Substitutes the standard macro predefinitions for the
4770 current target machine. Use this when running @code{cpp}.
4773 Like @samp{%p}, but puts @samp{__} before and after the name of each
4774 predefined macro, except for macros that start with @samp{__} or with
4775 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4779 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4782 Current argument is the name of a library or startup file of some sort.
4783 Search for that file in a standard list of directories and substitute
4784 the full name found.
4787 Print @var{str} as an error message. @var{str} is terminated by a newline.
4788 Use this when inconsistent options are detected.
4791 Output @samp{-} if the input for the current command is coming from a pipe.
4794 Substitute the contents of spec string @var{name} at this point.
4797 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4799 @item %x@{@var{option}@}
4800 Accumulate an option for @samp{%X}.
4803 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4807 Output the accumulated assembler options specified by @option{-Wa}.
4810 Output the accumulated preprocessor options specified by @option{-Wp}.
4813 Substitute the major version number of GCC@.
4814 (For version 2.9.5, this is 2.)
4817 Substitute the minor version number of GCC@.
4818 (For version 2.9.5, this is 9.)
4821 Substitute the patch level number of GCC@.
4822 (For version 2.9.5, this is 5.)
4825 Process the @code{asm} spec. This is used to compute the
4826 switches to be passed to the assembler.
4829 Process the @code{asm_final} spec. This is a spec string for
4830 passing switches to an assembler post-processor, if such a program is
4834 Process the @code{link} spec. This is the spec for computing the
4835 command line passed to the linker. Typically it will make use of the
4836 @samp{%L %G %S %D and %E} sequences.
4839 Dump out a @option{-L} option for each directory that GCC believes might
4840 contain startup files. If the target supports multilibs then the
4841 current multilib directory will be prepended to each of these paths.
4844 Output the multilib directory with directory separators replaced with
4845 @samp{_}. If multilib directories are not set, or the multilib directory is
4846 @file{.} then this option emits nothing.
4849 Process the @code{lib} spec. This is a spec string for deciding which
4850 libraries should be included on the command line to the linker.
4853 Process the @code{libgcc} spec. This is a spec string for deciding
4854 which GCC support library should be included on the command line to the linker.
4857 Process the @code{startfile} spec. This is a spec for deciding which
4858 object files should be the first ones passed to the linker. Typically
4859 this might be a file named @file{crt0.o}.
4862 Process the @code{endfile} spec. This is a spec string that specifies
4863 the last object files that will be passed to the linker.
4866 Process the @code{cpp} spec. This is used to construct the arguments
4867 to be passed to the C preprocessor.
4870 Process the @code{signed_char} spec. This is intended to be used
4871 to tell cpp whether a char is signed. It typically has the definition:
4873 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4877 Process the @code{cc1} spec. This is used to construct the options to be
4878 passed to the actual C compiler (@samp{cc1}).
4881 Process the @code{cc1plus} spec. This is used to construct the options to be
4882 passed to the actual C++ compiler (@samp{cc1plus}).
4885 Substitute the variable part of a matched option. See below.
4886 Note that each comma in the substituted string is replaced by
4890 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4891 If that switch was not specified, this substitutes nothing. Note that
4892 the leading dash is omitted when specifying this option, and it is
4893 automatically inserted if the substitution is performed. Thus the spec
4894 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4895 and would output the command line option @option{-foo}.
4897 @item %W@{@code{S}@}
4898 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4901 @item %@{@code{S}*@}
4902 Substitutes all the switches specified to GCC whose names start
4903 with @code{-S}, but which also take an argument. This is used for
4904 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4905 GCC considers @option{-o foo} as being
4906 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4907 text, including the space. Thus two arguments would be generated.
4909 @item %@{^@code{S}*@}
4910 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4911 argument. Thus %@{^o*@} would only generate one argument, not two.
4913 @item %@{@code{S}*&@code{T}*@}
4914 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4915 (the order of @code{S} and @code{T} in the spec is not significant).
4916 There can be any number of ampersand-separated variables; for each the
4917 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4919 @item %@{<@code{S}@}
4920 Remove all occurrences of @code{-S} from the command line. Note---this
4921 command is position dependent. @samp{%} commands in the spec string
4922 before this option will see @code{-S}, @samp{%} commands in the spec
4923 string after this option will not.
4925 @item %@{@code{S}*:@code{X}@}
4926 Substitutes @code{X} if one or more switches whose names start with
4927 @code{-S} are specified to GCC@. Note that the tail part of the
4928 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4929 for each occurrence of @samp{%*} within @code{X}.
4931 @item %@{@code{S}:@code{X}@}
4932 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4934 @item %@{!@code{S}:@code{X}@}
4935 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4937 @item %@{|@code{S}:@code{X}@}
4938 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4940 @item %@{|!@code{S}:@code{X}@}
4941 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4943 @item %@{.@code{S}:@code{X}@}
4944 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4946 @item %@{!.@code{S}:@code{X}@}
4947 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4949 @item %@{@code{S}|@code{P}:@code{X}@}
4950 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4951 combined with @samp{!} and @samp{.} sequences as well, although they
4952 have a stronger binding than the @samp{|}. For example a spec string
4956 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4959 will output the following command-line options from the following input
4960 command-line options:
4965 -d fred.c -foo -baz -boggle
4966 -d jim.d -bar -baz -boggle
4971 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4972 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4973 or spaces, or even newlines. They are processed as usual, as described
4976 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4977 switches are handled specifically in these
4978 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4979 @option{-W} switch is found later in the command line, the earlier switch
4980 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4981 letter, which passes all matching options.
4983 The character @samp{|} at the beginning of the predicate text is used to indicate
4984 that a command should be piped to the following command, but only if @option{-pipe}
4987 It is built into GCC which switches take arguments and which do not.
4988 (You might think it would be useful to generalize this to allow each
4989 compiler's spec to say which switches take arguments. But this cannot
4990 be done in a consistent fashion. GCC cannot even decide which input
4991 files have been specified without knowing which switches take arguments,
4992 and it must know which input files to compile in order to tell which
4995 GCC also knows implicitly that arguments starting in @option{-l} are to be
4996 treated as compiler output files, and passed to the linker in their
4997 proper position among the other output files.
4999 @c man begin OPTIONS
5001 @node Target Options
5002 @section Specifying Target Machine and Compiler Version
5003 @cindex target options
5004 @cindex cross compiling
5005 @cindex specifying machine version
5006 @cindex specifying compiler version and target machine
5007 @cindex compiler version, specifying
5008 @cindex target machine, specifying
5010 By default, GCC compiles code for the same type of machine that you
5011 are using. However, it can also be installed as a cross-compiler, to
5012 compile for some other type of machine. In fact, several different
5013 configurations of GCC, for different target machines, can be
5014 installed side by side. Then you specify which one to use with the
5017 In addition, older and newer versions of GCC can be installed side
5018 by side. One of them (probably the newest) will be the default, but
5019 you may sometimes wish to use another.
5022 @item -b @var{machine}
5024 The argument @var{machine} specifies the target machine for compilation.
5025 This is useful when you have installed GCC as a cross-compiler.
5027 The value to use for @var{machine} is the same as was specified as the
5028 machine type when configuring GCC as a cross-compiler. For
5029 example, if a cross-compiler was configured with @samp{configure
5030 i386v}, meaning to compile for an 80386 running System V, then you
5031 would specify @option{-b i386v} to run that cross compiler.
5033 When you do not specify @option{-b}, it normally means to compile for
5034 the same type of machine that you are using.
5036 @item -V @var{version}
5038 The argument @var{version} specifies which version of GCC to run.
5039 This is useful when multiple versions are installed. For example,
5040 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5042 The default version, when you do not specify @option{-V}, is the last
5043 version of GCC that you installed.
5046 The @option{-b} and @option{-V} options actually work by controlling part of
5047 the file name used for the executable files and libraries used for
5048 compilation. A given version of GCC, for a given target machine, is
5049 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.
5051 Thus, sites can customize the effect of @option{-b} or @option{-V} either by
5052 changing the names of these directories or adding alternate names (or
5053 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
5054 file @file{80386} is a link to the file @file{i386v}, then @option{-b
5055 80386} becomes an alias for @option{-b i386v}.
5057 In one respect, the @option{-b} or @option{-V} do not completely change
5058 to a different compiler: the top-level driver program @command{gcc}
5059 that you originally invoked continues to run and invoke the other
5060 executables (preprocessor, compiler per se, assembler and linker)
5061 that do the real work. However, since no real work is done in the
5062 driver program, it usually does not matter that the driver program
5063 in use is not the one for the specified target. It is common for the
5064 interface to the other executables to change incompatibly between
5065 compiler versions, so unless the version specified is very close to that
5066 of the driver (for example, @option{-V 3.0} with a driver program from GCC
5067 version 3.0.1), use of @option{-V} may not work; for example, using
5068 @option{-V 2.95.2} will not work with a driver program from GCC 3.0.
5070 The only way that the driver program depends on the target machine is
5071 in the parsing and handling of special machine-specific options.
5072 However, this is controlled by a file which is found, along with the
5073 other executables, in the directory for the specified version and
5074 target machine. As a result, a single installed driver program adapts
5075 to any specified target machine, and sufficiently similar compiler
5078 The driver program executable does control one significant thing,
5079 however: the default version and target machine. Therefore, you can
5080 install different instances of the driver program, compiled for
5081 different targets or versions, under different names.
5083 For example, if the driver for version 2.0 is installed as @command{ogcc}
5084 and that for version 2.1 is installed as @command{gcc}, then the command
5085 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
5086 2.0 by default. However, you can choose either version with either
5087 command with the @option{-V} option.
5089 @node Submodel Options
5090 @section Hardware Models and Configurations
5091 @cindex submodel options
5092 @cindex specifying hardware config
5093 @cindex hardware models and configurations, specifying
5094 @cindex machine dependent options
5096 Earlier we discussed the standard option @option{-b} which chooses among
5097 different installed compilers for completely different target
5098 machines, such as VAX vs.@: 68000 vs.@: 80386.
5100 In addition, each of these target machine types can have its own
5101 special options, starting with @samp{-m}, to choose among various
5102 hardware models or configurations---for example, 68010 vs 68020,
5103 floating coprocessor or none. A single installed version of the
5104 compiler can compile for any model or configuration, according to the
5107 Some configurations of the compiler also support additional special
5108 options, usually for compatibility with other compilers on the same
5111 These options are defined by the macro @code{TARGET_SWITCHES} in the
5112 machine description. The default for the options is also defined by
5113 that macro, which enables you to change the defaults.
5127 * RS/6000 and PowerPC Options::
5130 * i386 and x86-64 Options::
5132 * Intel 960 Options::
5133 * DEC Alpha Options::
5134 * DEC Alpha/VMS Options::
5138 * System V Options::
5139 * TMS320C3x/C4x Options::
5147 * S/390 and zSeries Options::
5151 * Xstormy16 Options::
5155 @node M680x0 Options
5156 @subsection M680x0 Options
5157 @cindex M680x0 options
5159 These are the @samp{-m} options defined for the 68000 series. The default
5160 values for these options depends on which style of 68000 was selected when
5161 the compiler was configured; the defaults for the most common choices are
5169 Generate output for a 68000. This is the default
5170 when the compiler is configured for 68000-based systems.
5172 Use this option for microcontrollers with a 68000 or EC000 core,
5173 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5179 Generate output for a 68020. This is the default
5180 when the compiler is configured for 68020-based systems.
5184 Generate output containing 68881 instructions for floating point.
5185 This is the default for most 68020 systems unless @option{--nfp} was
5186 specified when the compiler was configured.
5190 Generate output for a 68030. This is the default when the compiler is
5191 configured for 68030-based systems.
5195 Generate output for a 68040. This is the default when the compiler is
5196 configured for 68040-based systems.
5198 This option inhibits the use of 68881/68882 instructions that have to be
5199 emulated by software on the 68040. Use this option if your 68040 does not
5200 have code to emulate those instructions.
5204 Generate output for a 68060. This is the default when the compiler is
5205 configured for 68060-based systems.
5207 This option inhibits the use of 68020 and 68881/68882 instructions that
5208 have to be emulated by software on the 68060. Use this option if your 68060
5209 does not have code to emulate those instructions.
5213 Generate output for a CPU32. This is the default
5214 when the compiler is configured for CPU32-based systems.
5216 Use this option for microcontrollers with a
5217 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5218 68336, 68340, 68341, 68349 and 68360.
5222 Generate output for a 520X ``coldfire'' family cpu. This is the default
5223 when the compiler is configured for 520X-based systems.
5225 Use this option for microcontroller with a 5200 core, including
5226 the MCF5202, MCF5203, MCF5204 and MCF5202.
5231 Generate output for a 68040, without using any of the new instructions.
5232 This results in code which can run relatively efficiently on either a
5233 68020/68881 or a 68030 or a 68040. The generated code does use the
5234 68881 instructions that are emulated on the 68040.
5238 Generate output for a 68060, without using any of the new instructions.
5239 This results in code which can run relatively efficiently on either a
5240 68020/68881 or a 68030 or a 68040. The generated code does use the
5241 68881 instructions that are emulated on the 68060.
5245 Generate output containing Sun FPA instructions for floating point.
5248 @opindex msoft-float
5249 Generate output containing library calls for floating point.
5250 @strong{Warning:} the requisite libraries are not available for all m68k
5251 targets. Normally the facilities of the machine's usual C compiler are
5252 used, but this can't be done directly in cross-compilation. You must
5253 make your own arrangements to provide suitable library functions for
5254 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5255 @samp{m68k-*-coff} do provide software floating point support.
5259 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5262 @opindex mnobitfield
5263 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5264 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5268 Do use the bit-field instructions. The @option{-m68020} option implies
5269 @option{-mbitfield}. This is the default if you use a configuration
5270 designed for a 68020.
5274 Use a different function-calling convention, in which functions
5275 that take a fixed number of arguments return with the @code{rtd}
5276 instruction, which pops their arguments while returning. This
5277 saves one instruction in the caller since there is no need to pop
5278 the arguments there.
5280 This calling convention is incompatible with the one normally
5281 used on Unix, so you cannot use it if you need to call libraries
5282 compiled with the Unix compiler.
5284 Also, you must provide function prototypes for all functions that
5285 take variable numbers of arguments (including @code{printf});
5286 otherwise incorrect code will be generated for calls to those
5289 In addition, seriously incorrect code will result if you call a
5290 function with too many arguments. (Normally, extra arguments are
5291 harmlessly ignored.)
5293 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5294 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5297 @itemx -mno-align-int
5299 @opindex mno-align-int
5300 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5301 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5302 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5303 Aligning variables on 32-bit boundaries produces code that runs somewhat
5304 faster on processors with 32-bit busses at the expense of more memory.
5306 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5307 align structures containing the above types differently than
5308 most published application binary interface specifications for the m68k.
5312 Use the pc-relative addressing mode of the 68000 directly, instead of
5313 using a global offset table. At present, this option implies @option{-fpic},
5314 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5315 not presently supported with @option{-mpcrel}, though this could be supported for
5316 68020 and higher processors.
5318 @item -mno-strict-align
5319 @itemx -mstrict-align
5320 @opindex mno-strict-align
5321 @opindex mstrict-align
5322 Do not (do) assume that unaligned memory references will be handled by
5327 @node M68hc1x Options
5328 @subsection M68hc1x Options
5329 @cindex M68hc1x options
5331 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5332 microcontrollers. The default values for these options depends on
5333 which style of microcontroller was selected when the compiler was configured;
5334 the defaults for the most common choices are given below.
5341 Generate output for a 68HC11. This is the default
5342 when the compiler is configured for 68HC11-based systems.
5348 Generate output for a 68HC12. This is the default
5349 when the compiler is configured for 68HC12-based systems.
5352 @opindex mauto-incdec
5353 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5358 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5360 @item -msoft-reg-count=@var{count}
5361 @opindex msoft-reg-count
5362 Specify the number of pseudo-soft registers which are used for the
5363 code generation. The maximum number is 32. Using more pseudo-soft
5364 register may or may not result in better code depending on the program.
5365 The default is 4 for 68HC11 and 2 for 68HC12.
5370 @subsection VAX Options
5373 These @samp{-m} options are defined for the VAX:
5378 Do not output certain jump instructions (@code{aobleq} and so on)
5379 that the Unix assembler for the VAX cannot handle across long
5384 Do output those jump instructions, on the assumption that you
5385 will assemble with the GNU assembler.
5389 Output code for g-format floating point numbers instead of d-format.
5393 @subsection SPARC Options
5394 @cindex SPARC options
5396 These @samp{-m} switches are supported on the SPARC:
5401 @opindex mno-app-regs
5403 Specify @option{-mapp-regs} to generate output using the global registers
5404 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5407 To be fully SVR4 ABI compliant at the cost of some performance loss,
5408 specify @option{-mno-app-regs}. You should compile libraries and system
5409 software with this option.
5414 @opindex mhard-float
5415 Generate output containing floating point instructions. This is the
5421 @opindex msoft-float
5422 Generate output containing library calls for floating point.
5423 @strong{Warning:} the requisite libraries are not available for all SPARC
5424 targets. Normally the facilities of the machine's usual C compiler are
5425 used, but this cannot be done directly in cross-compilation. You must make
5426 your own arrangements to provide suitable library functions for
5427 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5428 @samp{sparclite-*-*} do provide software floating point support.
5430 @option{-msoft-float} changes the calling convention in the output file;
5431 therefore, it is only useful if you compile @emph{all} of a program with
5432 this option. In particular, you need to compile @file{libgcc.a}, the
5433 library that comes with GCC, with @option{-msoft-float} in order for
5436 @item -mhard-quad-float
5437 @opindex mhard-quad-float
5438 Generate output containing quad-word (long double) floating point
5441 @item -msoft-quad-float
5442 @opindex msoft-quad-float
5443 Generate output containing library calls for quad-word (long double)
5444 floating point instructions. The functions called are those specified
5445 in the SPARC ABI@. This is the default.
5447 As of this writing, there are no sparc implementations that have hardware
5448 support for the quad-word floating point instructions. They all invoke
5449 a trap handler for one of these instructions, and then the trap handler
5450 emulates the effect of the instruction. Because of the trap handler overhead,
5451 this is much slower than calling the ABI library routines. Thus the
5452 @option{-msoft-quad-float} option is the default.
5456 @opindex mno-epilogue
5458 With @option{-mepilogue} (the default), the compiler always emits code for
5459 function exit at the end of each function. Any function exit in
5460 the middle of the function (such as a return statement in C) will
5461 generate a jump to the exit code at the end of the function.
5463 With @option{-mno-epilogue}, the compiler tries to emit exit code inline
5464 at every function exit.
5470 With @option{-mflat}, the compiler does not generate save/restore instructions
5471 and will use a ``flat'' or single register window calling convention.
5472 This model uses %i7 as the frame pointer and is compatible with the normal
5473 register window model. Code from either may be intermixed.
5474 The local registers and the input registers (0--5) are still treated as
5475 ``call saved'' registers and will be saved on the stack as necessary.
5477 With @option{-mno-flat} (the default), the compiler emits save/restore
5478 instructions (except for leaf functions) and is the normal mode of operation.
5480 @item -mno-unaligned-doubles
5481 @itemx -munaligned-doubles
5482 @opindex mno-unaligned-doubles
5483 @opindex munaligned-doubles
5484 Assume that doubles have 8 byte alignment. This is the default.
5486 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5487 alignment only if they are contained in another type, or if they have an
5488 absolute address. Otherwise, it assumes they have 4 byte alignment.
5489 Specifying this option avoids some rare compatibility problems with code
5490 generated by other compilers. It is not the default because it results
5491 in a performance loss, especially for floating point code.
5493 @item -mno-faster-structs
5494 @itemx -mfaster-structs
5495 @opindex mno-faster-structs
5496 @opindex mfaster-structs
5497 With @option{-mfaster-structs}, the compiler assumes that structures
5498 should have 8 byte alignment. This enables the use of pairs of
5499 @code{ldd} and @code{std} instructions for copies in structure
5500 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5501 However, the use of this changed alignment directly violates the Sparc
5502 ABI@. Thus, it's intended only for use on targets where the developer
5503 acknowledges that their resulting code will not be directly in line with
5504 the rules of the ABI@.
5510 These two options select variations on the SPARC architecture.
5512 By default (unless specifically configured for the Fujitsu SPARClite),
5513 GCC generates code for the v7 variant of the SPARC architecture.
5515 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5516 code is that the compiler emits the integer multiply and integer
5517 divide instructions which exist in SPARC v8 but not in SPARC v7.
5519 @option{-msparclite} will give you SPARClite code. This adds the integer
5520 multiply, integer divide step and scan (@code{ffs}) instructions which
5521 exist in SPARClite but not in SPARC v7.
5523 These options are deprecated and will be deleted in a future GCC release.
5524 They have been replaced with @option{-mcpu=xxx}.
5529 @opindex msupersparc
5530 These two options select the processor for which the code is optimized.
5532 With @option{-mcypress} (the default), the compiler optimizes code for the
5533 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5534 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5536 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5537 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5538 of the full SPARC v8 instruction set.
5540 These options are deprecated and will be deleted in a future GCC release.
5541 They have been replaced with @option{-mcpu=xxx}.
5543 @item -mcpu=@var{cpu_type}
5545 Set the instruction set, register set, and instruction scheduling parameters
5546 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5547 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5548 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5549 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5551 Default instruction scheduling parameters are used for values that select
5552 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5553 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5555 Here is a list of each supported architecture and their supported
5560 v8: supersparc, hypersparc
5561 sparclite: f930, f934, sparclite86x
5566 @item -mtune=@var{cpu_type}
5568 Set the instruction scheduling parameters for machine type
5569 @var{cpu_type}, but do not set the instruction set or register set that the
5570 option @option{-mcpu=@var{cpu_type}} would.
5572 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5573 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5574 that select a particular cpu implementation. Those are @samp{cypress},
5575 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5576 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5580 These @samp{-m} switches are supported in addition to the above
5581 on the SPARCLET processor.
5584 @item -mlittle-endian
5585 @opindex mlittle-endian
5586 Generate code for a processor running in little-endian mode.
5590 Treat register @code{%g0} as a normal register.
5591 GCC will continue to clobber it as necessary but will not assume
5592 it always reads as 0.
5594 @item -mbroken-saverestore
5595 @opindex mbroken-saverestore
5596 Generate code that does not use non-trivial forms of the @code{save} and
5597 @code{restore} instructions. Early versions of the SPARCLET processor do
5598 not correctly handle @code{save} and @code{restore} instructions used with
5599 arguments. They correctly handle them used without arguments. A @code{save}
5600 instruction used without arguments increments the current window pointer
5601 but does not allocate a new stack frame. It is assumed that the window
5602 overflow trap handler will properly handle this case as will interrupt
5606 These @samp{-m} switches are supported in addition to the above
5607 on SPARC V9 processors in 64-bit environments.
5610 @item -mlittle-endian
5611 @opindex mlittle-endian
5612 Generate code for a processor running in little-endian mode.
5618 Generate code for a 32-bit or 64-bit environment.
5619 The 32-bit environment sets int, long and pointer to 32 bits.
5620 The 64-bit environment sets int to 32 bits and long and pointer
5623 @item -mcmodel=medlow
5624 @opindex mcmodel=medlow
5625 Generate code for the Medium/Low code model: the program must be linked
5626 in the low 32 bits of the address space. Pointers are 64 bits.
5627 Programs can be statically or dynamically linked.
5629 @item -mcmodel=medmid
5630 @opindex mcmodel=medmid
5631 Generate code for the Medium/Middle code model: the program must be linked
5632 in the low 44 bits of the address space, the text segment must be less than
5633 2G bytes, and data segment must be within 2G of the text segment.
5634 Pointers are 64 bits.
5636 @item -mcmodel=medany
5637 @opindex mcmodel=medany
5638 Generate code for the Medium/Anywhere code model: the program may be linked
5639 anywhere in the address space, the text segment must be less than
5640 2G bytes, and data segment must be within 2G of the text segment.
5641 Pointers are 64 bits.
5643 @item -mcmodel=embmedany
5644 @opindex mcmodel=embmedany
5645 Generate code for the Medium/Anywhere code model for embedded systems:
5646 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5647 (determined at link time). Register %g4 points to the base of the
5648 data segment. Pointers are still 64 bits.
5649 Programs are statically linked, PIC is not supported.
5652 @itemx -mno-stack-bias
5653 @opindex mstack-bias
5654 @opindex mno-stack-bias
5655 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5656 frame pointer if present, are offset by @minus{}2047 which must be added back
5657 when making stack frame references.
5658 Otherwise, assume no such offset is present.
5661 @node Convex Options
5662 @subsection Convex Options
5663 @cindex Convex options
5665 These @samp{-m} options are defined for Convex:
5670 Generate output for C1. The code will run on any Convex machine.
5671 The preprocessor symbol @code{__convex__c1__} is defined.
5675 Generate output for C2. Uses instructions not available on C1.
5676 Scheduling and other optimizations are chosen for max performance on C2.
5677 The preprocessor symbol @code{__convex_c2__} is defined.
5681 Generate output for C32xx. Uses instructions not available on C1.
5682 Scheduling and other optimizations are chosen for max performance on C32.
5683 The preprocessor symbol @code{__convex_c32__} is defined.
5687 Generate output for C34xx. Uses instructions not available on C1.
5688 Scheduling and other optimizations are chosen for max performance on C34.
5689 The preprocessor symbol @code{__convex_c34__} is defined.
5693 Generate output for C38xx. Uses instructions not available on C1.
5694 Scheduling and other optimizations are chosen for max performance on C38.
5695 The preprocessor symbol @code{__convex_c38__} is defined.
5699 Generate code which puts an argument count in the word preceding each
5700 argument list. This is compatible with regular CC, and a few programs
5701 may need the argument count word. GDB and other source-level debuggers
5702 do not need it; this info is in the symbol table.
5705 @opindex mnoargcount
5706 Omit the argument count word. This is the default.
5708 @item -mvolatile-cache
5709 @opindex mvolatile-cache
5710 Allow volatile references to be cached. This is the default.
5712 @item -mvolatile-nocache
5713 @opindex mvolatile-nocache
5714 Volatile references bypass the data cache, going all the way to memory.
5715 This is only needed for multi-processor code that does not use standard
5716 synchronization instructions. Making non-volatile references to volatile
5717 locations will not necessarily work.
5721 Type long is 32 bits, the same as type int. This is the default.
5725 Type long is 64 bits, the same as type long long. This option is useless,
5726 because no library support exists for it.
5729 @node AMD29K Options
5730 @subsection AMD29K Options
5731 @cindex AMD29K options
5733 These @samp{-m} options are defined for the AMD Am29000:
5738 @cindex DW bit (29k)
5739 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5740 halfword operations are directly supported by the hardware. This is the
5745 Generate code that assumes the @code{DW} bit is not set.
5749 @cindex byte writes (29k)
5750 Generate code that assumes the system supports byte and halfword write
5751 operations. This is the default.
5755 Generate code that assumes the systems does not support byte and
5756 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5760 @cindex memory model (29k)
5761 Use a small memory model that assumes that all function addresses are
5762 either within a single 256 KB segment or at an absolute address of less
5763 than 256k. This allows the @code{call} instruction to be used instead
5764 of a @code{const}, @code{consth}, @code{calli} sequence.
5768 Use the normal memory model: Generate @code{call} instructions only when
5769 calling functions in the same file and @code{calli} instructions
5770 otherwise. This works if each file occupies less than 256 KB but allows
5771 the entire executable to be larger than 256 KB@. This is the default.
5775 Always use @code{calli} instructions. Specify this option if you expect
5776 a single file to compile into more than 256 KB of code.
5780 @cindex processor selection (29k)
5781 Generate code for the Am29050.
5785 Generate code for the Am29000. This is the default.
5787 @item -mkernel-registers
5788 @opindex mkernel-registers
5789 @cindex kernel and user registers (29k)
5790 Generate references to registers @code{gr64-gr95} instead of to
5791 registers @code{gr96-gr127}. This option can be used when compiling
5792 kernel code that wants a set of global registers disjoint from that used
5795 Note that when this option is used, register names in @samp{-f} flags
5796 must use the normal, user-mode, names.
5798 @item -muser-registers
5799 @opindex muser-registers
5800 Use the normal set of global registers, @code{gr96-gr127}. This is the
5804 @itemx -mno-stack-check
5805 @opindex mstack-check
5806 @opindex mno-stack-check
5807 @cindex stack checks (29k)
5808 Insert (or do not insert) a call to @code{__msp_check} after each stack
5809 adjustment. This is often used for kernel code.
5812 @itemx -mno-storem-bug
5813 @opindex mstorem-bug
5814 @opindex mno-storem-bug
5815 @cindex storem bug (29k)
5816 @option{-mstorem-bug} handles 29k processors which cannot handle the
5817 separation of a mtsrim insn and a storem instruction (most 29000 chips
5818 to date, but not the 29050).
5820 @item -mno-reuse-arg-regs
5821 @itemx -mreuse-arg-regs
5822 @opindex mno-reuse-arg-regs
5823 @opindex mreuse-arg-regs
5824 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5825 registers for copying out arguments. This helps detect calling a function
5826 with fewer arguments than it was declared with.
5828 @item -mno-impure-text
5829 @itemx -mimpure-text
5830 @opindex mno-impure-text
5831 @opindex mimpure-text
5832 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5833 not pass @option{-assert pure-text} to the linker when linking a shared object.
5836 @opindex msoft-float
5837 Generate output containing library calls for floating point.
5838 @strong{Warning:} the requisite libraries are not part of GCC@.
5839 Normally the facilities of the machine's usual C compiler are used, but
5840 this can't be done directly in cross-compilation. You must make your
5841 own arrangements to provide suitable library functions for
5846 Do not generate multm or multmu instructions. This is useful for some embedded
5847 systems which do not have trap handlers for these instructions.
5851 @subsection ARM Options
5854 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5859 @opindex mapcs-frame
5860 Generate a stack frame that is compliant with the ARM Procedure Call
5861 Standard for all functions, even if this is not strictly necessary for
5862 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5863 with this option will cause the stack frames not to be generated for
5864 leaf functions. The default is @option{-mno-apcs-frame}.
5868 This is a synonym for @option{-mapcs-frame}.
5872 Generate code for a processor running with a 26-bit program counter,
5873 and conforming to the function calling standards for the APCS 26-bit
5874 option. This option replaces the @option{-m2} and @option{-m3} options
5875 of previous releases of the compiler.
5879 Generate code for a processor running with a 32-bit program counter,
5880 and conforming to the function calling standards for the APCS 32-bit
5881 option. This option replaces the @option{-m6} option of previous releases
5885 @c not currently implemented
5886 @item -mapcs-stack-check
5887 @opindex mapcs-stack-check
5888 Generate code to check the amount of stack space available upon entry to
5889 every function (that actually uses some stack space). If there is
5890 insufficient space available then either the function
5891 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5892 called, depending upon the amount of stack space required. The run time
5893 system is required to provide these functions. The default is
5894 @option{-mno-apcs-stack-check}, since this produces smaller code.
5896 @c not currently implemented
5898 @opindex mapcs-float
5899 Pass floating point arguments using the float point registers. This is
5900 one of the variants of the APCS@. This option is recommended if the
5901 target hardware has a floating point unit or if a lot of floating point
5902 arithmetic is going to be performed by the code. The default is
5903 @option{-mno-apcs-float}, since integer only code is slightly increased in
5904 size if @option{-mapcs-float} is used.
5906 @c not currently implemented
5907 @item -mapcs-reentrant
5908 @opindex mapcs-reentrant
5909 Generate reentrant, position independent code. The default is
5910 @option{-mno-apcs-reentrant}.
5913 @item -mthumb-interwork
5914 @opindex mthumb-interwork
5915 Generate code which supports calling between the ARM and Thumb
5916 instruction sets. Without this option the two instruction sets cannot
5917 be reliably used inside one program. The default is
5918 @option{-mno-thumb-interwork}, since slightly larger code is generated
5919 when @option{-mthumb-interwork} is specified.
5921 @item -mno-sched-prolog
5922 @opindex mno-sched-prolog
5923 Prevent the reordering of instructions in the function prolog, or the
5924 merging of those instruction with the instructions in the function's
5925 body. This means that all functions will start with a recognizable set
5926 of instructions (or in fact one of a choice from a small set of
5927 different function prologues), and this information can be used to
5928 locate the start if functions inside an executable piece of code. The
5929 default is @option{-msched-prolog}.
5932 @opindex mhard-float
5933 Generate output containing floating point instructions. This is the
5937 @opindex msoft-float
5938 Generate output containing library calls for floating point.
5939 @strong{Warning:} the requisite libraries are not available for all ARM
5940 targets. Normally the facilities of the machine's usual C compiler are
5941 used, but this cannot be done directly in cross-compilation. You must make
5942 your own arrangements to provide suitable library functions for
5945 @option{-msoft-float} changes the calling convention in the output file;
5946 therefore, it is only useful if you compile @emph{all} of a program with
5947 this option. In particular, you need to compile @file{libgcc.a}, the
5948 library that comes with GCC, with @option{-msoft-float} in order for
5951 @item -mlittle-endian
5952 @opindex mlittle-endian
5953 Generate code for a processor running in little-endian mode. This is
5954 the default for all standard configurations.
5957 @opindex mbig-endian
5958 Generate code for a processor running in big-endian mode; the default is
5959 to compile code for a little-endian processor.
5961 @item -mwords-little-endian
5962 @opindex mwords-little-endian
5963 This option only applies when generating code for big-endian processors.
5964 Generate code for a little-endian word order but a big-endian byte
5965 order. That is, a byte order of the form @samp{32107654}. Note: this
5966 option should only be used if you require compatibility with code for
5967 big-endian ARM processors generated by versions of the compiler prior to
5970 @item -malignment-traps
5971 @opindex malignment-traps
5972 Generate code that will not trap if the MMU has alignment traps enabled.
5973 On ARM architectures prior to ARMv4, there were no instructions to
5974 access half-word objects stored in memory. However, when reading from
5975 memory a feature of the ARM architecture allows a word load to be used,
5976 even if the address is unaligned, and the processor core will rotate the
5977 data as it is being loaded. This option tells the compiler that such
5978 misaligned accesses will cause a MMU trap and that it should instead
5979 synthesise the access as a series of byte accesses. The compiler can
5980 still use word accesses to load half-word data if it knows that the
5981 address is aligned to a word boundary.
5983 This option is ignored when compiling for ARM architecture 4 or later,
5984 since these processors have instructions to directly access half-word
5987 @item -mno-alignment-traps
5988 @opindex mno-alignment-traps
5989 Generate code that assumes that the MMU will not trap unaligned
5990 accesses. This produces better code when the target instruction set
5991 does not have half-word memory operations (i.e.@: implementations prior to
5994 Note that you cannot use this option to access unaligned word objects,
5995 since the processor will only fetch one 32-bit aligned object from
5998 The default setting for most targets is @option{-mno-alignment-traps}, since
5999 this produces better code when there are no half-word memory
6000 instructions available.
6002 @item -mshort-load-bytes
6003 @itemx -mno-short-load-words
6004 @opindex mshort-load-bytes
6005 @opindex mno-short-load-words
6006 These are deprecated aliases for @option{-malignment-traps}.
6008 @item -mno-short-load-bytes
6009 @itemx -mshort-load-words
6010 @opindex mno-short-load-bytes
6011 @opindex mshort-load-words
6012 This are deprecated aliases for @option{-mno-alignment-traps}.
6016 This option only applies to RISC iX@. Emulate the native BSD-mode
6017 compiler. This is the default if @option{-ansi} is not specified.
6021 This option only applies to RISC iX@. Emulate the native X/Open-mode
6024 @item -mno-symrename
6025 @opindex mno-symrename
6026 This option only applies to RISC iX@. Do not run the assembler
6027 post-processor, @samp{symrename}, after code has been assembled.
6028 Normally it is necessary to modify some of the standard symbols in
6029 preparation for linking with the RISC iX C library; this option
6030 suppresses this pass. The post-processor is never run when the
6031 compiler is built for cross-compilation.
6033 @item -mcpu=@var{name}
6035 This specifies the name of the target ARM processor. GCC uses this name
6036 to determine what kind of instructions it can emit when generating
6037 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6038 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6039 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6040 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6041 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6042 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6043 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6044 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6045 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
6046 @samp{arm1020t}, @samp{xscale}.
6048 @itemx -mtune=@var{name}
6050 This option is very similar to the @option{-mcpu=} option, except that
6051 instead of specifying the actual target processor type, and hence
6052 restricting which instructions can be used, it specifies that GCC should
6053 tune the performance of the code as if the target were of the type
6054 specified in this option, but still choosing the instructions that it
6055 will generate based on the cpu specified by a @option{-mcpu=} option.
6056 For some ARM implementations better performance can be obtained by using
6059 @item -march=@var{name}
6061 This specifies the name of the target ARM architecture. GCC uses this
6062 name to determine what kind of instructions it can emit when generating
6063 assembly code. This option can be used in conjunction with or instead
6064 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6065 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6066 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
6068 @item -mfpe=@var{number}
6069 @itemx -mfp=@var{number}
6072 This specifies the version of the floating point emulation available on
6073 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6074 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6076 @item -mstructure-size-boundary=@var{n}
6077 @opindex mstructure-size-boundary
6078 The size of all structures and unions will be rounded up to a multiple
6079 of the number of bits set by this option. Permissible values are 8 and
6080 32. The default value varies for different toolchains. For the COFF
6081 targeted toolchain the default value is 8. Specifying the larger number
6082 can produce faster, more efficient code, but can also increase the size
6083 of the program. The two values are potentially incompatible. Code
6084 compiled with one value cannot necessarily expect to work with code or
6085 libraries compiled with the other value, if they exchange information
6086 using structures or unions.
6088 @item -mabort-on-noreturn
6089 @opindex mabort-on-noreturn
6090 Generate a call to the function @code{abort} at the end of a
6091 @code{noreturn} function. It will be executed if the function tries to
6095 @itemx -mno-long-calls
6096 @opindex mlong-calls
6097 @opindex mno-long-calls
6098 Tells the compiler to perform function calls by first loading the
6099 address of the function into a register and then performing a subroutine
6100 call on this register. This switch is needed if the target function
6101 will lie outside of the 64 megabyte addressing range of the offset based
6102 version of subroutine call instruction.
6104 Even if this switch is enabled, not all function calls will be turned
6105 into long calls. The heuristic is that static functions, functions
6106 which have the @samp{short-call} attribute, functions that are inside
6107 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6108 definitions have already been compiled within the current compilation
6109 unit, will not be turned into long calls. The exception to this rule is
6110 that weak function definitions, functions with the @samp{long-call}
6111 attribute or the @samp{section} attribute, and functions that are within
6112 the scope of a @samp{#pragma long_calls} directive, will always be
6113 turned into long calls.
6115 This feature is not enabled by default. Specifying
6116 @option{-mno-long-calls} will restore the default behavior, as will
6117 placing the function calls within the scope of a @samp{#pragma
6118 long_calls_off} directive. Note these switches have no effect on how
6119 the compiler generates code to handle function calls via function
6122 @item -mnop-fun-dllimport
6123 @opindex mnop-fun-dllimport
6124 Disable support for the @code{dllimport} attribute.
6126 @item -msingle-pic-base
6127 @opindex msingle-pic-base
6128 Treat the register used for PIC addressing as read-only, rather than
6129 loading it in the prologue for each function. The run-time system is
6130 responsible for initializing this register with an appropriate value
6131 before execution begins.
6133 @item -mpic-register=@var{reg}
6134 @opindex mpic-register
6135 Specify the register to be used for PIC addressing. The default is R10
6136 unless stack-checking is enabled, when R9 is used.
6138 @item -mpoke-function-name
6139 @opindex mpoke-function-name
6140 Write the name of each function into the text section, directly
6141 preceding the function prologue. The generated code is similar to this:
6145 .ascii "arm_poke_function_name", 0
6148 .word 0xff000000 + (t1 - t0)
6149 arm_poke_function_name
6151 stmfd sp!, @{fp, ip, lr, pc@}
6155 When performing a stack backtrace, code can inspect the value of
6156 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6157 location @code{pc - 12} and the top 8 bits are set, then we know that
6158 there is a function name embedded immediately preceding this location
6159 and has length @code{((pc[-3]) & 0xff000000)}.
6163 Generate code for the 16-bit Thumb instruction set. The default is to
6164 use the 32-bit ARM instruction set.
6167 @opindex mtpcs-frame
6168 Generate a stack frame that is compliant with the Thumb Procedure Call
6169 Standard for all non-leaf functions. (A leaf function is one that does
6170 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6172 @item -mtpcs-leaf-frame
6173 @opindex mtpcs-leaf-frame
6174 Generate a stack frame that is compliant with the Thumb Procedure Call
6175 Standard for all leaf functions. (A leaf function is one that does
6176 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6178 @item -mcallee-super-interworking
6179 @opindex mcallee-super-interworking
6180 Gives all externally visible functions in the file being compiled an ARM
6181 instruction set header which switches to Thumb mode before executing the
6182 rest of the function. This allows these functions to be called from
6183 non-interworking code.
6185 @item -mcaller-super-interworking
6186 @opindex mcaller-super-interworking
6187 Allows calls via function pointers (including virtual functions) to
6188 execute correctly regardless of whether the target code has been
6189 compiled for interworking or not. There is a small overhead in the cost
6190 of executing a function pointer if this option is enabled.
6194 @node MN10200 Options
6195 @subsection MN10200 Options
6196 @cindex MN10200 options
6197 These @option{-m} options are defined for Matsushita MN10200 architectures:
6202 Indicate to the linker that it should perform a relaxation optimization pass
6203 to shorten branches, calls and absolute memory addresses. This option only
6204 has an effect when used on the command line for the final link step.
6206 This option makes symbolic debugging impossible.
6209 @node MN10300 Options
6210 @subsection MN10300 Options
6211 @cindex MN10300 options
6212 These @option{-m} options are defined for Matsushita MN10300 architectures:
6217 Generate code to avoid bugs in the multiply instructions for the MN10300
6218 processors. This is the default.
6221 @opindex mno-mult-bug
6222 Do not generate code to avoid bugs in the multiply instructions for the
6227 Generate code which uses features specific to the AM33 processor.
6231 Do not generate code which uses features specific to the AM33 processor. This
6236 Do not link in the C run-time initialization object file.
6240 Indicate to the linker that it should perform a relaxation optimization pass
6241 to shorten branches, calls and absolute memory addresses. This option only
6242 has an effect when used on the command line for the final link step.
6244 This option makes symbolic debugging impossible.
6248 @node M32R/D Options
6249 @subsection M32R/D Options
6250 @cindex M32R/D options
6252 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6257 Generate code for the M32R/X@.
6261 Generate code for the M32R@. This is the default.
6263 @item -mcode-model=small
6264 @opindex mcode-model=small
6265 Assume all objects live in the lower 16MB of memory (so that their addresses
6266 can be loaded with the @code{ld24} instruction), and assume all subroutines
6267 are reachable with the @code{bl} instruction.
6268 This is the default.
6270 The addressability of a particular object can be set with the
6271 @code{model} attribute.
6273 @item -mcode-model=medium
6274 @opindex mcode-model=medium
6275 Assume objects may be anywhere in the 32-bit address space (the compiler
6276 will generate @code{seth/add3} instructions to load their addresses), and
6277 assume all subroutines are reachable with the @code{bl} instruction.
6279 @item -mcode-model=large
6280 @opindex mcode-model=large
6281 Assume objects may be anywhere in the 32-bit address space (the compiler
6282 will generate @code{seth/add3} instructions to load their addresses), and
6283 assume subroutines may not be reachable with the @code{bl} instruction
6284 (the compiler will generate the much slower @code{seth/add3/jl}
6285 instruction sequence).
6288 @opindex msdata=none
6289 Disable use of the small data area. Variables will be put into
6290 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6291 @code{section} attribute has been specified).
6292 This is the default.
6294 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6295 Objects may be explicitly put in the small data area with the
6296 @code{section} attribute using one of these sections.
6299 @opindex msdata=sdata
6300 Put small global and static data in the small data area, but do not
6301 generate special code to reference them.
6305 Put small global and static data in the small data area, and generate
6306 special instructions to reference them.
6310 @cindex smaller data references
6311 Put global and static objects less than or equal to @var{num} bytes
6312 into the small data or bss sections instead of the normal data or bss
6313 sections. The default value of @var{num} is 8.
6314 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6315 for this option to have any effect.
6317 All modules should be compiled with the same @option{-G @var{num}} value.
6318 Compiling with different values of @var{num} may or may not work; if it
6319 doesn't the linker will give an error message---incorrect code will not be
6325 @subsection M88K Options
6326 @cindex M88k options
6328 These @samp{-m} options are defined for Motorola 88k architectures:
6333 Generate code that works well on both the m88100 and the
6338 Generate code that works best for the m88100, but that also
6343 Generate code that works best for the m88110, and may not run
6348 Obsolete option to be removed from the next revision.
6351 @item -midentify-revision
6352 @opindex midentify-revision
6353 @cindex identifying source, compiler (88k)
6354 Include an @code{ident} directive in the assembler output recording the
6355 source file name, compiler name and version, timestamp, and compilation
6358 @item -mno-underscores
6359 @opindex mno-underscores
6360 @cindex underscores, avoiding (88k)
6361 In assembler output, emit symbol names without adding an underscore
6362 character at the beginning of each name. The default is to use an
6363 underscore as prefix on each name.
6365 @item -mocs-debug-info
6366 @itemx -mno-ocs-debug-info
6367 @opindex mocs-debug-info
6368 @opindex mno-ocs-debug-info
6370 @cindex debugging, 88k OCS
6371 Include (or omit) additional debugging information (about registers used
6372 in each stack frame) as specified in the 88open Object Compatibility
6373 Standard, ``OCS''@. This extra information allows debugging of code that
6374 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6375 Delta 88 SVr3.2 is to include this information; other 88k configurations
6376 omit this information by default.
6378 @item -mocs-frame-position
6379 @opindex mocs-frame-position
6380 @cindex register positions in frame (88k)
6381 When emitting COFF debugging information for automatic variables and
6382 parameters stored on the stack, use the offset from the canonical frame
6383 address, which is the stack pointer (register 31) on entry to the
6384 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6385 @option{-mocs-frame-position}; other 88k configurations have the default
6386 @option{-mno-ocs-frame-position}.
6388 @item -mno-ocs-frame-position
6389 @opindex mno-ocs-frame-position
6390 @cindex register positions in frame (88k)
6391 When emitting COFF debugging information for automatic variables and
6392 parameters stored on the stack, use the offset from the frame pointer
6393 register (register 30). When this option is in effect, the frame
6394 pointer is not eliminated when debugging information is selected by the
6397 @item -moptimize-arg-area
6398 @opindex moptimize-arg-area
6399 @cindex arguments in frame (88k)
6400 Save space by reorganizing the stack frame. This option generates code
6401 that does not agree with the 88open specifications, but uses less
6404 @itemx -mno-optimize-arg-area
6405 @opindex mno-optimize-arg-area
6406 Do not reorganize the stack frame to save space. This is the default.
6407 The generated conforms to the specification, but uses more memory.
6409 @item -mshort-data-@var{num}
6410 @opindex mshort-data
6411 @cindex smaller data references (88k)
6412 @cindex r0-relative references (88k)
6413 Generate smaller data references by making them relative to @code{r0},
6414 which allows loading a value using a single instruction (rather than the
6415 usual two). You control which data references are affected by
6416 specifying @var{num} with this option. For example, if you specify
6417 @option{-mshort-data-512}, then the data references affected are those
6418 involving displacements of less than 512 bytes.
6419 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6422 @item -mserialize-volatile
6423 @opindex mserialize-volatile
6424 @itemx -mno-serialize-volatile
6425 @opindex mno-serialize-volatile
6426 @cindex sequential consistency on 88k
6427 Do, or don't, generate code to guarantee sequential consistency
6428 of volatile memory references. By default, consistency is
6431 The order of memory references made by the MC88110 processor does
6432 not always match the order of the instructions requesting those
6433 references. In particular, a load instruction may execute before
6434 a preceding store instruction. Such reordering violates
6435 sequential consistency of volatile memory references, when there
6436 are multiple processors. When consistency must be guaranteed,
6437 GCC generates special instructions, as needed, to force
6438 execution in the proper order.
6440 The MC88100 processor does not reorder memory references and so
6441 always provides sequential consistency. However, by default, GCC
6442 generates the special instructions to guarantee consistency
6443 even when you use @option{-m88100}, so that the code may be run on an
6444 MC88110 processor. If you intend to run your code only on the
6445 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6447 The extra code generated to guarantee consistency may affect the
6448 performance of your application. If you know that you can safely
6449 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6455 @cindex assembler syntax, 88k
6457 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6458 related to System V release 4 (SVr4). This controls the following:
6462 Which variant of the assembler syntax to emit.
6464 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6465 that is used on System V release 4.
6467 @option{-msvr4} makes GCC issue additional declaration directives used in
6471 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6472 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6473 other m88k configurations.
6475 @item -mversion-03.00
6476 @opindex mversion-03.00
6477 This option is obsolete, and is ignored.
6478 @c ??? which asm syntax better for GAS? option there too?
6480 @item -mno-check-zero-division
6481 @itemx -mcheck-zero-division
6482 @opindex mno-check-zero-division
6483 @opindex mcheck-zero-division
6484 @cindex zero division on 88k
6485 Do, or don't, generate code to guarantee that integer division by
6486 zero will be detected. By default, detection is guaranteed.
6488 Some models of the MC88100 processor fail to trap upon integer
6489 division by zero under certain conditions. By default, when
6490 compiling code that might be run on such a processor, GCC
6491 generates code that explicitly checks for zero-valued divisors
6492 and traps with exception number 503 when one is detected. Use of
6493 @option{-mno-check-zero-division} suppresses such checking for code
6494 generated to run on an MC88100 processor.
6496 GCC assumes that the MC88110 processor correctly detects all instances
6497 of integer division by zero. When @option{-m88110} is specified, no
6498 explicit checks for zero-valued divisors are generated, and both
6499 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6502 @item -muse-div-instruction
6503 @opindex muse-div-instruction
6504 @cindex divide instruction, 88k
6505 Use the div instruction for signed integer division on the
6506 MC88100 processor. By default, the div instruction is not used.
6508 On the MC88100 processor the signed integer division instruction
6509 div) traps to the operating system on a negative operand. The
6510 operating system transparently completes the operation, but at a
6511 large cost in execution time. By default, when compiling code
6512 that might be run on an MC88100 processor, GCC emulates signed
6513 integer division using the unsigned integer division instruction
6514 divu), thereby avoiding the large penalty of a trap to the
6515 operating system. Such emulation has its own, smaller, execution
6516 cost in both time and space. To the extent that your code's
6517 important signed integer division operations are performed on two
6518 nonnegative operands, it may be desirable to use the div
6519 instruction directly.
6521 On the MC88110 processor the div instruction (also known as the
6522 divs instruction) processes negative operands without trapping to
6523 the operating system. When @option{-m88110} is specified,
6524 @option{-muse-div-instruction} is ignored, and the div instruction is used
6525 for signed integer division.
6527 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6528 particular, the behavior of such a division with and without
6529 @option{-muse-div-instruction} may differ.
6531 @item -mtrap-large-shift
6532 @itemx -mhandle-large-shift
6533 @opindex mtrap-large-shift
6534 @opindex mhandle-large-shift
6535 @cindex bit shift overflow (88k)
6536 @cindex large bit shifts (88k)
6537 Include code to detect bit-shifts of more than 31 bits; respectively,
6538 trap such shifts or emit code to handle them properly. By default GCC
6539 makes no special provision for large bit shifts.
6541 @item -mwarn-passed-structs
6542 @opindex mwarn-passed-structs
6543 @cindex structure passing (88k)
6544 Warn when a function passes a struct as an argument or result.
6545 Structure-passing conventions have changed during the evolution of the C
6546 language, and are often the source of portability problems. By default,
6547 GCC issues no such warning.
6550 @c break page here to avoid unsightly interparagraph stretch.
6554 @node RS/6000 and PowerPC Options
6555 @subsection IBM RS/6000 and PowerPC Options
6556 @cindex RS/6000 and PowerPC Options
6557 @cindex IBM RS/6000 and PowerPC Options
6559 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6567 @itemx -mpowerpc-gpopt
6568 @itemx -mno-powerpc-gpopt
6569 @itemx -mpowerpc-gfxopt
6570 @itemx -mno-powerpc-gfxopt
6572 @itemx -mno-powerpc64
6578 @opindex mno-powerpc
6579 @opindex mpowerpc-gpopt
6580 @opindex mno-powerpc-gpopt
6581 @opindex mpowerpc-gfxopt
6582 @opindex mno-powerpc-gfxopt
6584 @opindex mno-powerpc64
6585 GCC supports two related instruction set architectures for the
6586 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6587 instructions supported by the @samp{rios} chip set used in the original
6588 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6589 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6590 the IBM 4xx microprocessors.
6592 Neither architecture is a subset of the other. However there is a
6593 large common subset of instructions supported by both. An MQ
6594 register is included in processors supporting the POWER architecture.
6596 You use these options to specify which instructions are available on the
6597 processor you are using. The default value of these options is
6598 determined when configuring GCC@. Specifying the
6599 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6600 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6601 rather than the options listed above.
6603 The @option{-mpower} option allows GCC to generate instructions that
6604 are found only in the POWER architecture and to use the MQ register.
6605 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6606 to generate instructions that are present in the POWER2 architecture but
6607 not the original POWER architecture.
6609 The @option{-mpowerpc} option allows GCC to generate instructions that
6610 are found only in the 32-bit subset of the PowerPC architecture.
6611 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6612 GCC to use the optional PowerPC architecture instructions in the
6613 General Purpose group, including floating-point square root. Specifying
6614 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6615 use the optional PowerPC architecture instructions in the Graphics
6616 group, including floating-point select.
6618 The @option{-mpowerpc64} option allows GCC to generate the additional
6619 64-bit instructions that are found in the full PowerPC64 architecture
6620 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6621 @option{-mno-powerpc64}.
6623 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6624 will use only the instructions in the common subset of both
6625 architectures plus some special AIX common-mode calls, and will not use
6626 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6627 permits GCC to use any instruction from either architecture and to
6628 allow use of the MQ register; specify this for the Motorola MPC601.
6630 @item -mnew-mnemonics
6631 @itemx -mold-mnemonics
6632 @opindex mnew-mnemonics
6633 @opindex mold-mnemonics
6634 Select which mnemonics to use in the generated assembler code. With
6635 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6636 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6637 assembler mnemonics defined for the POWER architecture. Instructions
6638 defined in only one architecture have only one mnemonic; GCC uses that
6639 mnemonic irrespective of which of these options is specified.
6641 GCC defaults to the mnemonics appropriate for the architecture in
6642 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6643 value of these option. Unless you are building a cross-compiler, you
6644 should normally not specify either @option{-mnew-mnemonics} or
6645 @option{-mold-mnemonics}, but should instead accept the default.
6647 @item -mcpu=@var{cpu_type}
6649 Set architecture type, register usage, choice of mnemonics, and
6650 instruction scheduling parameters for machine type @var{cpu_type}.
6651 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6652 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6653 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6654 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6655 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6656 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6658 @option{-mcpu=common} selects a completely generic processor. Code
6659 generated under this option will run on any POWER or PowerPC processor.
6660 GCC will use only the instructions in the common subset of both
6661 architectures, and will not use the MQ register. GCC assumes a generic
6662 processor model for scheduling purposes.
6664 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6665 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6666 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6667 types, with an appropriate, generic processor model assumed for
6668 scheduling purposes.
6670 The other options specify a specific processor. Code generated under
6671 those options will run best on that processor, and may not run at all on
6674 The @option{-mcpu} options automatically enable or disable other
6675 @option{-m} options as follows:
6679 @option{-mno-power}, @option{-mno-powerc}
6686 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6701 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6704 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6709 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6712 @item -mtune=@var{cpu_type}
6714 Set the instruction scheduling parameters for machine type
6715 @var{cpu_type}, but do not set the architecture type, register usage, or
6716 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6717 values for @var{cpu_type} are used for @option{-mtune} as for
6718 @option{-mcpu}. If both are specified, the code generated will use the
6719 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6720 scheduling parameters set by @option{-mtune}.
6725 @opindex mno-altivec
6726 These switches enable or disable the use of built-in functions that
6727 allow access to the AltiVec instruction set. You may also need to set
6728 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6732 @itemx -mno-fp-in-toc
6733 @itemx -mno-sum-in-toc
6734 @itemx -mminimal-toc
6736 @opindex mno-fp-in-toc
6737 @opindex mno-sum-in-toc
6738 @opindex mminimal-toc
6739 Modify generation of the TOC (Table Of Contents), which is created for
6740 every executable file. The @option{-mfull-toc} option is selected by
6741 default. In that case, GCC will allocate at least one TOC entry for
6742 each unique non-automatic variable reference in your program. GCC
6743 will also place floating-point constants in the TOC@. However, only
6744 16,384 entries are available in the TOC@.
6746 If you receive a linker error message that saying you have overflowed
6747 the available TOC space, you can reduce the amount of TOC space used
6748 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6749 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6750 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6751 generate code to calculate the sum of an address and a constant at
6752 run-time instead of putting that sum into the TOC@. You may specify one
6753 or both of these options. Each causes GCC to produce very slightly
6754 slower and larger code at the expense of conserving TOC space.
6756 If you still run out of space in the TOC even when you specify both of
6757 these options, specify @option{-mminimal-toc} instead. This option causes
6758 GCC to make only one TOC entry for every file. When you specify this
6759 option, GCC will produce code that is slower and larger but which
6760 uses extremely little TOC space. You may wish to use this option
6761 only on files that contain less frequently executed code.
6767 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6768 @code{long} type, and the infrastructure needed to support them.
6769 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6770 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6771 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6776 @opindex mno-xl-call
6777 On AIX, pass floating-point arguments to prototyped functions beyond the
6778 register save area (RSA) on the stack in addition to argument FPRs. The
6779 AIX calling convention was extended but not initially documented to
6780 handle an obscure K&R C case of calling a function that takes the
6781 address of its arguments with fewer arguments than declared. AIX XL
6782 compilers access floating point arguments which do not fit in the
6783 RSA from the stack when a subroutine is compiled without
6784 optimization. Because always storing floating-point arguments on the
6785 stack is inefficient and rarely needed, this option is not enabled by
6786 default and only is necessary when calling subroutines compiled by AIX
6787 XL compilers without optimization.
6791 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6792 application written to use message passing with special startup code to
6793 enable the application to run. The system must have PE installed in the
6794 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6795 must be overridden with the @option{-specs=} option to specify the
6796 appropriate directory location. The Parallel Environment does not
6797 support threads, so the @option{-mpe} option and the @option{-pthread}
6798 option are incompatible.
6802 @opindex msoft-float
6803 @opindex mhard-float
6804 Generate code that does not use (uses) the floating-point register set.
6805 Software floating point emulation is provided if you use the
6806 @option{-msoft-float} option, and pass the option to GCC when linking.
6809 @itemx -mno-multiple
6811 @opindex mno-multiple
6812 Generate code that uses (does not use) the load multiple word
6813 instructions and the store multiple word instructions. These
6814 instructions are generated by default on POWER systems, and not
6815 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6816 endian PowerPC systems, since those instructions do not work when the
6817 processor is in little endian mode. The exceptions are PPC740 and
6818 PPC750 which permit the instructions usage in little endian mode.
6824 Generate code that uses (does not use) the load string instructions
6825 and the store string word instructions to save multiple registers and
6826 do small block moves. These instructions are generated by default on
6827 POWER systems, and not generated on PowerPC systems. Do not use
6828 @option{-mstring} on little endian PowerPC systems, since those
6829 instructions do not work when the processor is in little endian mode.
6830 The exceptions are PPC740 and PPC750 which permit the instructions
6831 usage in little endian mode.
6837 Generate code that uses (does not use) the load or store instructions
6838 that update the base register to the address of the calculated memory
6839 location. These instructions are generated by default. If you use
6840 @option{-mno-update}, there is a small window between the time that the
6841 stack pointer is updated and the address of the previous frame is
6842 stored, which means code that walks the stack frame across interrupts or
6843 signals may get corrupted data.
6846 @itemx -mno-fused-madd
6847 @opindex mfused-madd
6848 @opindex mno-fused-madd
6849 Generate code that uses (does not use) the floating point multiply and
6850 accumulate instructions. These instructions are generated by default if
6851 hardware floating is used.
6853 @item -mno-bit-align
6855 @opindex mno-bit-align
6857 On System V.4 and embedded PowerPC systems do not (do) force structures
6858 and unions that contain bit-fields to be aligned to the base type of the
6861 For example, by default a structure containing nothing but 8
6862 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6863 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6864 the structure would be aligned to a 1 byte boundary and be one byte in
6867 @item -mno-strict-align
6868 @itemx -mstrict-align
6869 @opindex mno-strict-align
6870 @opindex mstrict-align
6871 On System V.4 and embedded PowerPC systems do not (do) assume that
6872 unaligned memory references will be handled by the system.
6875 @itemx -mno-relocatable
6876 @opindex mrelocatable
6877 @opindex mno-relocatable
6878 On embedded PowerPC systems generate code that allows (does not allow)
6879 the program to be relocated to a different address at runtime. If you
6880 use @option{-mrelocatable} on any module, all objects linked together must
6881 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6883 @item -mrelocatable-lib
6884 @itemx -mno-relocatable-lib
6885 @opindex mrelocatable-lib
6886 @opindex mno-relocatable-lib
6887 On embedded PowerPC systems generate code that allows (does not allow)
6888 the program to be relocated to a different address at runtime. Modules
6889 compiled with @option{-mrelocatable-lib} can be linked with either modules
6890 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6891 with modules compiled with the @option{-mrelocatable} options.
6897 On System V.4 and embedded PowerPC systems do not (do) assume that
6898 register 2 contains a pointer to a global area pointing to the addresses
6899 used in the program.
6902 @itemx -mlittle-endian
6904 @opindex mlittle-endian
6905 On System V.4 and embedded PowerPC systems compile code for the
6906 processor in little endian mode. The @option{-mlittle-endian} option is
6907 the same as @option{-mlittle}.
6912 @opindex mbig-endian
6913 On System V.4 and embedded PowerPC systems compile code for the
6914 processor in big endian mode. The @option{-mbig-endian} option is
6915 the same as @option{-mbig}.
6919 On System V.4 and embedded PowerPC systems compile code using calling
6920 conventions that adheres to the March 1995 draft of the System V
6921 Application Binary Interface, PowerPC processor supplement. This is the
6922 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6924 @item -mcall-sysv-eabi
6925 @opindex mcall-sysv-eabi
6926 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6928 @item -mcall-sysv-noeabi
6929 @opindex mcall-sysv-noeabi
6930 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6934 On System V.4 and embedded PowerPC systems compile code using calling
6935 conventions that are similar to those used on AIX@. This is the
6936 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6938 @item -mcall-solaris
6939 @opindex mcall-solaris
6940 On System V.4 and embedded PowerPC systems compile code for the Solaris
6944 @opindex mcall-linux
6945 On System V.4 and embedded PowerPC systems compile code for the
6946 Linux-based GNU system.
6950 On System V.4 and embedded PowerPC systems compile code for the
6951 Hurd-based GNU system.
6954 @opindex mcall-netbsd
6955 On System V.4 and embedded PowerPC systems compile code for the
6956 NetBSD operating system.
6958 @item -maix-struct-return
6959 @opindex maix-struct-return
6960 Return all structures in memory (as specified by the AIX ABI)@.
6962 @item -msvr4-struct-return
6963 @opindex msvr4-struct-return
6964 Return structures smaller than 8 bytes in registers (as specified by the
6968 @opindex mabi=altivec
6969 Extend the current ABI with AltiVec ABI extensions. This does not
6970 change the default ABI, instead it adds the AltiVec ABI extensions to
6974 @itemx -mno-prototype
6976 @opindex mno-prototype
6977 On System V.4 and embedded PowerPC systems assume that all calls to
6978 variable argument functions are properly prototyped. Otherwise, the
6979 compiler must insert an instruction before every non prototyped call to
6980 set or clear bit 6 of the condition code register (@var{CR}) to
6981 indicate whether floating point values were passed in the floating point
6982 registers in case the function takes a variable arguments. With
6983 @option{-mprototype}, only calls to prototyped variable argument functions
6984 will set or clear the bit.
6988 On embedded PowerPC systems, assume that the startup module is called
6989 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6990 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6995 On embedded PowerPC systems, assume that the startup module is called
6996 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
7001 On embedded PowerPC systems, assume that the startup module is called
7002 @file{crt0.o} and the standard C libraries are @file{libads.a} and
7006 @opindex myellowknife
7007 On embedded PowerPC systems, assume that the startup module is called
7008 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
7013 On System V.4 and embedded PowerPC systems, specify that you are
7014 compiling for a VxWorks system.
7018 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7019 header to indicate that @samp{eabi} extended relocations are used.
7025 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7026 Embedded Applications Binary Interface (eabi) which is a set of
7027 modifications to the System V.4 specifications. Selecting @option{-meabi}
7028 means that the stack is aligned to an 8 byte boundary, a function
7029 @code{__eabi} is called to from @code{main} to set up the eabi
7030 environment, and the @option{-msdata} option can use both @code{r2} and
7031 @code{r13} to point to two separate small data areas. Selecting
7032 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7033 do not call an initialization function from @code{main}, and the
7034 @option{-msdata} option will only use @code{r13} to point to a single
7035 small data area. The @option{-meabi} option is on by default if you
7036 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7039 @opindex msdata=eabi
7040 On System V.4 and embedded PowerPC systems, put small initialized
7041 @code{const} global and static data in the @samp{.sdata2} section, which
7042 is pointed to by register @code{r2}. Put small initialized
7043 non-@code{const} global and static data in the @samp{.sdata} section,
7044 which is pointed to by register @code{r13}. Put small uninitialized
7045 global and static data in the @samp{.sbss} section, which is adjacent to
7046 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7047 incompatible with the @option{-mrelocatable} option. The
7048 @option{-msdata=eabi} option also sets the @option{-memb} option.
7051 @opindex msdata=sysv
7052 On System V.4 and embedded PowerPC systems, put small global and static
7053 data in the @samp{.sdata} section, which is pointed to by register
7054 @code{r13}. Put small uninitialized global and static data in the
7055 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7056 The @option{-msdata=sysv} option is incompatible with the
7057 @option{-mrelocatable} option.
7059 @item -msdata=default
7061 @opindex msdata=default
7063 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7064 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7065 same as @option{-msdata=sysv}.
7068 @opindex msdata-data
7069 On System V.4 and embedded PowerPC systems, put small global and static
7070 data in the @samp{.sdata} section. Put small uninitialized global and
7071 static data in the @samp{.sbss} section. Do not use register @code{r13}
7072 to address small data however. This is the default behavior unless
7073 other @option{-msdata} options are used.
7077 @opindex msdata=none
7079 On embedded PowerPC systems, put all initialized global and static data
7080 in the @samp{.data} section, and all uninitialized data in the
7081 @samp{.bss} section.
7085 @cindex smaller data references (PowerPC)
7086 @cindex .sdata/.sdata2 references (PowerPC)
7087 On embedded PowerPC systems, put global and static items less than or
7088 equal to @var{num} bytes into the small data or bss sections instead of
7089 the normal data or bss section. By default, @var{num} is 8. The
7090 @option{-G @var{num}} switch is also passed to the linker.
7091 All modules should be compiled with the same @option{-G @var{num}} value.
7094 @itemx -mno-regnames
7096 @opindex mno-regnames
7097 On System V.4 and embedded PowerPC systems do (do not) emit register
7098 names in the assembly language output using symbolic forms.
7102 Adds support for multithreading with the @dfn{pthreads} library.
7103 This option sets flags for both the preprocessor and linker.
7108 @subsection IBM RT Options
7110 @cindex IBM RT options
7112 These @samp{-m} options are defined for the IBM RT PC:
7116 @opindex min-line-mul
7117 Use an in-line code sequence for integer multiplies. This is the
7120 @item -mcall-lib-mul
7121 @opindex mcall-lib-mul
7122 Call @code{lmul$$} for integer multiples.
7124 @item -mfull-fp-blocks
7125 @opindex mfull-fp-blocks
7126 Generate full-size floating point data blocks, including the minimum
7127 amount of scratch space recommended by IBM@. This is the default.
7129 @item -mminimum-fp-blocks
7130 @opindex mminimum-fp-blocks
7131 Do not include extra scratch space in floating point data blocks. This
7132 results in smaller code, but slower execution, since scratch space must
7133 be allocated dynamically.
7135 @cindex @file{varargs.h} and RT PC
7136 @cindex @file{stdarg.h} and RT PC
7137 @item -mfp-arg-in-fpregs
7138 @opindex mfp-arg-in-fpregs
7139 Use a calling sequence incompatible with the IBM calling convention in
7140 which floating point arguments are passed in floating point registers.
7141 Note that @code{varargs.h} and @code{stdarg.h} will not work with
7142 floating point operands if this option is specified.
7144 @item -mfp-arg-in-gregs
7145 @opindex mfp-arg-in-gregs
7146 Use the normal calling convention for floating point arguments. This is
7149 @item -mhc-struct-return
7150 @opindex mhc-struct-return
7151 Return structures of more than one word in memory, rather than in a
7152 register. This provides compatibility with the MetaWare HighC (hc)
7153 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7154 with the Portable C Compiler (pcc).
7156 @item -mnohc-struct-return
7157 @opindex mnohc-struct-return
7158 Return some structures of more than one word in registers, when
7159 convenient. This is the default. For compatibility with the
7160 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7161 option @option{-mhc-struct-return}.
7165 @subsection MIPS Options
7166 @cindex MIPS options
7168 These @samp{-m} options are defined for the MIPS family of computers:
7172 @item -march=@var{cpu-type}
7174 Assume the defaults for the machine type @var{cpu-type} when generating
7175 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7176 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7177 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7178 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7179 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7180 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
7182 @item -mtune=@var{cpu-type}
7184 Assume the defaults for the machine type @var{cpu-type} when scheduling
7185 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7186 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7187 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7188 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7189 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7190 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
7191 @var{cpu-type} will schedule things appropriately for that particular
7192 chip, the compiler will not generate any code that does not meet level 1
7193 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
7194 or @option{-mabi} switch being used.
7196 @item -mcpu=@var{cpu-type}
7198 This is identical to specifying both @option{-march} and @option{-mtune}.
7202 Issue instructions from level 1 of the MIPS ISA@. This is the default.
7203 @samp{r3000} is the default @var{cpu-type} at this ISA level.
7207 Issue instructions from level 2 of the MIPS ISA (branch likely, square
7208 root instructions). @samp{r6000} is the default @var{cpu-type} at this
7213 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
7214 @samp{r4000} is the default @var{cpu-type} at this ISA level.
7218 Issue instructions from level 4 of the MIPS ISA (conditional move,
7219 prefetch, enhanced FPU instructions). @samp{r8000} is the default
7220 @var{cpu-type} at this ISA level.
7224 Assume that 32 32-bit floating point registers are available. This is
7229 Assume that 32 64-bit floating point registers are available. This is
7230 the default when the @option{-mips3} option is used.
7233 @itemx -mno-fused-madd
7234 @opindex mfused-madd
7235 @opindex mno-fused-madd
7236 Generate code that uses (does not use) the floating point multiply and
7237 accumulate instructions, when they are available. These instructions
7238 are generated by default if they are available, but this may be
7239 undesirable if the extra precision causes problems or on certain chips
7240 in the mode where denormals are rounded to zero where denormals
7241 generated by multiply and accumulate instructions cause exceptions
7246 Assume that 32 32-bit general purpose registers are available. This is
7251 Assume that 32 64-bit general purpose registers are available. This is
7252 the default when the @option{-mips3} option is used.
7256 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7257 explanation of the default, and the width of pointers.
7261 Force long types to be 64 bits wide. See @option{-mlong32} for an
7262 explanation of the default, and the width of pointers.
7266 Force long, int, and pointer types to be 32 bits wide.
7268 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
7269 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
7270 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
7271 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
7272 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
7273 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
7274 are 32 bits, and longs are 64 bits wide. The width of pointer types is
7275 the smaller of the width of longs or the width of general purpose
7276 registers (which in turn depends on the ISA)@.
7288 Generate code for the indicated ABI@. The default instruction level is
7289 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
7290 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
7291 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
7296 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7297 add normal debug information. This is the default for all
7298 platforms except for the OSF/1 reference platform, using the OSF/rose
7299 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7300 switches are used, the @file{mips-tfile} program will encapsulate the
7301 stabs within MIPS ECOFF@.
7305 Generate code for the GNU assembler. This is the default on the OSF/1
7306 reference platform, using the OSF/rose object format. Also, this is
7307 the default if the configure option @option{--with-gnu-as} is used.
7309 @item -msplit-addresses
7310 @itemx -mno-split-addresses
7311 @opindex msplit-addresses
7312 @opindex mno-split-addresses
7313 Generate code to load the high and low parts of address constants separately.
7314 This allows GCC to optimize away redundant loads of the high order
7315 bits of addresses. This optimization requires GNU as and GNU ld.
7316 This optimization is enabled by default for some embedded targets where
7317 GNU as and GNU ld are standard.
7323 The @option{-mrnames} switch says to output code using the MIPS software
7324 names for the registers, instead of the hardware names (ie, @var{a0}
7325 instead of @var{$4}). The only known assembler that supports this option
7326 is the Algorithmics assembler.
7332 The @option{-mgpopt} switch says to write all of the data declarations
7333 before the instructions in the text section, this allows the MIPS
7334 assembler to generate one word memory references instead of using two
7335 words for short global or static data items. This is on by default if
7336 optimization is selected.
7342 For each non-inline function processed, the @option{-mstats} switch
7343 causes the compiler to emit one line to the standard error file to
7344 print statistics about the program (number of registers saved, stack
7351 The @option{-mmemcpy} switch makes all block moves call the appropriate
7352 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7353 generating inline code.
7356 @itemx -mno-mips-tfile
7357 @opindex mmips-tfile
7358 @opindex mno-mips-tfile
7359 The @option{-mno-mips-tfile} switch causes the compiler not
7360 postprocess the object file with the @file{mips-tfile} program,
7361 after the MIPS assembler has generated it to add debug support. If
7362 @file{mips-tfile} is not run, then no local variables will be
7363 available to the debugger. In addition, @file{stage2} and
7364 @file{stage3} objects will have the temporary file names passed to the
7365 assembler embedded in the object file, which means the objects will
7366 not compare the same. The @option{-mno-mips-tfile} switch should only
7367 be used when there are bugs in the @file{mips-tfile} program that
7368 prevents compilation.
7371 @opindex msoft-float
7372 Generate output containing library calls for floating point.
7373 @strong{Warning:} the requisite libraries are not part of GCC@.
7374 Normally the facilities of the machine's usual C compiler are used, but
7375 this can't be done directly in cross-compilation. You must make your
7376 own arrangements to provide suitable library functions for
7380 @opindex mhard-float
7381 Generate output containing floating point instructions. This is the
7382 default if you use the unmodified sources.
7385 @itemx -mno-abicalls
7387 @opindex mno-abicalls
7388 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7389 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7390 position independent code.
7393 @itemx -mno-long-calls
7394 @opindex mlong-calls
7395 @opindex mno-long-calls
7396 Do all calls with the @samp{JALR} instruction, which requires
7397 loading up a function's address into a register before the call.
7398 You need to use this switch, if you call outside of the current
7399 512 megabyte segment to functions that are not through pointers.
7402 @itemx -mno-half-pic
7404 @opindex mno-half-pic
7405 Put pointers to extern references into the data section and load them
7406 up, rather than put the references in the text section.
7408 @item -membedded-pic
7409 @itemx -mno-embedded-pic
7410 @opindex membedded-pic
7411 @opindex mno-embedded-pic
7412 Generate PIC code suitable for some embedded systems. All calls are
7413 made using PC relative address, and all data is addressed using the $gp
7414 register. No more than 65536 bytes of global data may be used. This
7415 requires GNU as and GNU ld which do most of the work. This currently
7416 only works on targets which use ECOFF; it does not work with ELF@.
7418 @item -membedded-data
7419 @itemx -mno-embedded-data
7420 @opindex membedded-data
7421 @opindex mno-embedded-data
7422 Allocate variables to the read-only data section first if possible, then
7423 next in the small data section if possible, otherwise in data. This gives
7424 slightly slower code than the default, but reduces the amount of RAM required
7425 when executing, and thus may be preferred for some embedded systems.
7427 @item -muninit-const-in-rodata
7428 @itemx -mno-uninit-const-in-rodata
7429 @opindex muninit-const-in-rodata
7430 @opindex mno-uninit-const-in-rodata
7431 When used together with @option{-membedded-data}, it will always store uninitialized
7432 const variables in the read-only data section.
7434 @item -msingle-float
7435 @itemx -mdouble-float
7436 @opindex msingle-float
7437 @opindex mdouble-float
7438 The @option{-msingle-float} switch tells gcc to assume that the floating
7439 point coprocessor only supports single precision operations, as on the
7440 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7441 double precision operations. This is the default.
7447 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7448 as on the @samp{r4650} chip.
7452 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7453 @option{-mcpu=r4650}.
7459 Enable 16-bit instructions.
7463 Use the entry and exit pseudo ops. This option can only be used with
7468 Compile code for the processor in little endian mode.
7469 The requisite libraries are assumed to exist.
7473 Compile code for the processor in big endian mode.
7474 The requisite libraries are assumed to exist.
7478 @cindex smaller data references (MIPS)
7479 @cindex gp-relative references (MIPS)
7480 Put global and static items less than or equal to @var{num} bytes into
7481 the small data or bss sections instead of the normal data or bss
7482 section. This allows the assembler to emit one word memory reference
7483 instructions based on the global pointer (@var{gp} or @var{$28}),
7484 instead of the normal two words used. By default, @var{num} is 8 when
7485 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7486 @option{-G @var{num}} switch is also passed to the assembler and linker.
7487 All modules should be compiled with the same @option{-G @var{num}}
7492 Tell the MIPS assembler to not run its preprocessor over user
7493 assembler files (with a @samp{.s} suffix) when assembling them.
7497 Pass an option to gas which will cause nops to be inserted if
7498 the read of the destination register of an mfhi or mflo instruction
7499 occurs in the following two instructions.
7503 Do not include the default crt0.
7505 @item -mflush-func=@var{func}
7506 @itemx -mno-flush-func
7507 @opindex mflush-func
7508 Specifies the function to call to flush the I and D caches, or to not
7509 call any such function. If called, the function must take the same
7510 arguments as the common @code{_flush_func()}, that is, the address of the
7511 memory range for which the cache is being flushed, the size of the
7512 memory range, and the number 3 (to flush both caches). The default
7513 depends on the target gcc was configured for, but commonly is either
7514 @samp{_flush_func} or @samp{__cpu_flush}.
7517 These options are defined by the macro
7518 @code{TARGET_SWITCHES} in the machine description. The default for the
7519 options is also defined by that macro, which enables you to change the
7522 @node i386 and x86-64 Options
7523 @subsection Intel 386 and AMD x86-64 Options
7524 @cindex i386 Options
7525 @cindex x86-64 Options
7526 @cindex Intel 386 Options
7527 @cindex AMD x86-64 Options
7529 These @samp{-m} options are defined for the i386 and x86-64 family of
7533 @item -mcpu=@var{cpu-type}
7535 Tune to @var{cpu-type} everything applicable about the generated code, except
7536 for the ABI and the set of available instructions. The choices for
7537 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7538 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7539 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7540 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp}
7541 and @samp{athlon-mp}.
7543 While picking a specific @var{cpu-type} will schedule things appropriately
7544 for that particular chip, the compiler will not generate any code that
7545 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7546 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7547 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7548 AMD chips as opposed to the Intel ones.
7550 @item -march=@var{cpu-type}
7552 Generate instructions for the machine type @var{cpu-type}. The choices
7553 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7554 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7563 @opindex mpentiumpro
7564 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7565 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7566 These synonyms are deprecated.
7568 @item -mfpmath=@var{unit}
7570 generate floating point arithmetics for selected unit @var{unit}. the choices
7575 Use the standard 387 floating point coprocessor present majority of chips and
7576 emulated otherwise. Code compiled with this option will run almost everywhere.
7577 The temporary results are computed in 80bit precesion instead of precision
7578 specified by the type resulting in slightly different results compared to most
7579 of other chips. See @option{-ffloat-store} for more detailed description.
7581 This is the default choice for i386 compiler.
7584 Use scalar floating point instructions present in the SSE instruction set.
7585 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7586 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7587 instruction set supports only single precision arithmetics, thus the double and
7588 extended precision arithmetics is still done using 387. Later version, present
7589 only in Pentium4 and the future AMD x86-64 chips supports double precision
7592 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7593 @option{-msse2} switches to enable SSE extensions and make this option
7594 effective. For x86-64 compiler, these extensions are enabled by default.
7596 The resulting code should be considerably faster in majority of cases and avoid
7597 the numerical instability problems of 387 code, but may break some existing
7598 code that expects temporaries to be 80bit.
7600 This is the default choice for x86-64 compiler.
7603 Attempt to utilize both instruction sets at once. This effectivly double the
7604 amount of available registers and on chips with separate execution units for
7605 387 and SSE the execution resources too. Use this option with care, as it is
7606 still experimental, because gcc register allocator does not model separate
7607 functional units well resulting in instable performance.
7610 @item -masm=@var{dialect}
7611 @opindex masm=@var{dialect}
7612 Output asm instructions using selected @var{dialect}. Supported choices are
7613 @samp{intel} or @samp{att} (the default one).
7618 @opindex mno-ieee-fp
7619 Control whether or not the compiler uses IEEE floating point
7620 comparisons. These handle correctly the case where the result of a
7621 comparison is unordered.
7624 @opindex msoft-float
7625 Generate output containing library calls for floating point.
7626 @strong{Warning:} the requisite libraries are not part of GCC@.
7627 Normally the facilities of the machine's usual C compiler are used, but
7628 this can't be done directly in cross-compilation. You must make your
7629 own arrangements to provide suitable library functions for
7632 On machines where a function returns floating point results in the 80387
7633 register stack, some floating point opcodes may be emitted even if
7634 @option{-msoft-float} is used.
7636 @item -mno-fp-ret-in-387
7637 @opindex mno-fp-ret-in-387
7638 Do not use the FPU registers for return values of functions.
7640 The usual calling convention has functions return values of types
7641 @code{float} and @code{double} in an FPU register, even if there
7642 is no FPU@. The idea is that the operating system should emulate
7645 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7646 in ordinary CPU registers instead.
7648 @item -mno-fancy-math-387
7649 @opindex mno-fancy-math-387
7650 Some 387 emulators do not support the @code{sin}, @code{cos} and
7651 @code{sqrt} instructions for the 387. Specify this option to avoid
7652 generating those instructions. This option is the default on FreeBSD@.
7653 As of revision 2.6.1, these instructions are not generated unless you
7654 also use the @option{-funsafe-math-optimizations} switch.
7656 @item -malign-double
7657 @itemx -mno-align-double
7658 @opindex malign-double
7659 @opindex mno-align-double
7660 Control whether GCC aligns @code{double}, @code{long double}, and
7661 @code{long long} variables on a two word boundary or a one word
7662 boundary. Aligning @code{double} variables on a two word boundary will
7663 produce code that runs somewhat faster on a @samp{Pentium} at the
7664 expense of more memory.
7666 @item -m128bit-long-double
7667 @opindex m128bit-long-double
7668 Control the size of @code{long double} type. i386 application binary interface
7669 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7670 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7671 impossible to reach with 12 byte long doubles in the array accesses.
7673 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7674 structures and arrays containing @code{long double} will change their size as
7675 well as function calling convention for function taking @code{long double}
7678 @item -m96bit-long-double
7679 @opindex m96bit-long-double
7680 Set the size of @code{long double} to 96 bits as required by the i386
7681 application binary interface. This is the default.
7684 @itemx -mno-svr3-shlib
7685 @opindex msvr3-shlib
7686 @opindex mno-svr3-shlib
7687 Control whether GCC places uninitialized local variables into the
7688 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7689 into @code{bss}. These options are meaningful only on System V Release 3.
7693 Use a different function-calling convention, in which functions that
7694 take a fixed number of arguments return with the @code{ret} @var{num}
7695 instruction, which pops their arguments while returning. This saves one
7696 instruction in the caller since there is no need to pop the arguments
7699 You can specify that an individual function is called with this calling
7700 sequence with the function attribute @samp{stdcall}. You can also
7701 override the @option{-mrtd} option by using the function attribute
7702 @samp{cdecl}. @xref{Function Attributes}.
7704 @strong{Warning:} this calling convention is incompatible with the one
7705 normally used on Unix, so you cannot use it if you need to call
7706 libraries compiled with the Unix compiler.
7708 Also, you must provide function prototypes for all functions that
7709 take variable numbers of arguments (including @code{printf});
7710 otherwise incorrect code will be generated for calls to those
7713 In addition, seriously incorrect code will result if you call a
7714 function with too many arguments. (Normally, extra arguments are
7715 harmlessly ignored.)
7717 @item -mregparm=@var{num}
7719 Control how many registers are used to pass integer arguments. By
7720 default, no registers are used to pass arguments, and at most 3
7721 registers can be used. You can control this behavior for a specific
7722 function by using the function attribute @samp{regparm}.
7723 @xref{Function Attributes}.
7725 @strong{Warning:} if you use this switch, and
7726 @var{num} is nonzero, then you must build all modules with the same
7727 value, including any libraries. This includes the system libraries and
7730 @item -mpreferred-stack-boundary=@var{num}
7731 @opindex mpreferred-stack-boundary
7732 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7733 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7734 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7735 size (@option{-Os}), in which case the default is the minimum correct
7736 alignment (4 bytes for x86, and 8 bytes for x86-64).
7738 On Pentium and PentiumPro, @code{double} and @code{long double} values
7739 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7740 suffer significant run time performance penalties. On Pentium III, the
7741 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7742 penalties if it is not 16 byte aligned.
7744 To ensure proper alignment of this values on the stack, the stack boundary
7745 must be as aligned as that required by any value stored on the stack.
7746 Further, every function must be generated such that it keeps the stack
7747 aligned. Thus calling a function compiled with a higher preferred
7748 stack boundary from a function compiled with a lower preferred stack
7749 boundary will most likely misalign the stack. It is recommended that
7750 libraries that use callbacks always use the default setting.
7752 This extra alignment does consume extra stack space, and generally
7753 increases code size. Code that is sensitive to stack space usage, such
7754 as embedded systems and operating system kernels, may want to reduce the
7755 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7771 These switches enable or disable the use of built-in functions that allow
7772 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7774 @xref{X86 Built-in Functions}, for details of the functions enabled
7775 and disabled by these switches.
7778 @itemx -mno-push-args
7780 @opindex mno-push-args
7781 Use PUSH operations to store outgoing parameters. This method is shorter
7782 and usually equally fast as method using SUB/MOV operations and is enabled
7783 by default. In some cases disabling it may improve performance because of
7784 improved scheduling and reduced dependencies.
7786 @item -maccumulate-outgoing-args
7787 @opindex maccumulate-outgoing-args
7788 If enabled, the maximum amount of space required for outgoing arguments will be
7789 computed in the function prologue. This is faster on most modern CPUs
7790 because of reduced dependencies, improved scheduling and reduced stack usage
7791 when preferred stack boundary is not equal to 2. The drawback is a notable
7792 increase in code size. This switch implies @option{-mno-push-args}.
7796 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7797 on thread-safe exception handling must compile and link all code with the
7798 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7799 @option{-D_MT}; when linking, it links in a special thread helper library
7800 @option{-lmingwthrd} which cleans up per thread exception handling data.
7802 @item -mno-align-stringops
7803 @opindex mno-align-stringops
7804 Do not align destination of inlined string operations. This switch reduces
7805 code size and improves performance in case the destination is already aligned,
7806 but gcc don't know about it.
7808 @item -minline-all-stringops
7809 @opindex minline-all-stringops
7810 By default GCC inlines string operations only when destination is known to be
7811 aligned at least to 4 byte boundary. This enables more inlining, increase code
7812 size, but may improve performance of code that depends on fast memcpy, strlen
7813 and memset for short lengths.
7815 @item -momit-leaf-frame-pointer
7816 @opindex momit-leaf-frame-pointer
7817 Don't keep the frame pointer in a register for leaf functions. This
7818 avoids the instructions to save, set up and restore frame pointers and
7819 makes an extra register available in leaf functions. The option
7820 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7821 which might make debugging harder.
7824 These @samp{-m} switches are supported in addition to the above
7825 on AMD x86-64 processors in 64-bit environments.
7832 Generate code for a 32-bit or 64-bit environment.
7833 The 32-bit environment sets int, long and pointer to 32 bits and
7834 generates code that runs on any i386 system.
7835 The 64-bit environment sets int to 32 bits and long and pointer
7836 to 64 bits and generates code for AMD's x86-64 architecture.
7839 @opindex no-red-zone
7840 Do not use a so called red zone for x86-64 code. The red zone is mandated
7841 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7842 stack pointer that will not be modified by signal or interrupt handlers
7843 and therefore can be used for temporary data without adjusting the stack
7844 pointer. The flag @option{-mno-red-zone} disables this red zone.
7848 @subsection HPPA Options
7849 @cindex HPPA Options
7851 These @samp{-m} options are defined for the HPPA family of computers:
7854 @item -march=@var{architecture-type}
7856 Generate code for the specified architecture. The choices for
7857 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7858 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7859 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7860 architecture option for your machine. Code compiled for lower numbered
7861 architectures will run on higher numbered architectures, but not the
7864 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7865 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7869 @itemx -mpa-risc-1-1
7870 @itemx -mpa-risc-2-0
7871 @opindex mpa-risc-1-0
7872 @opindex mpa-risc-1-1
7873 @opindex mpa-risc-2-0
7874 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7877 @opindex mbig-switch
7878 Generate code suitable for big switch tables. Use this option only if
7879 the assembler/linker complain about out of range branches within a switch
7882 @item -mjump-in-delay
7883 @opindex mjump-in-delay
7884 Fill delay slots of function calls with unconditional jump instructions
7885 by modifying the return pointer for the function call to be the target
7886 of the conditional jump.
7888 @item -mdisable-fpregs
7889 @opindex mdisable-fpregs
7890 Prevent floating point registers from being used in any manner. This is
7891 necessary for compiling kernels which perform lazy context switching of
7892 floating point registers. If you use this option and attempt to perform
7893 floating point operations, the compiler will abort.
7895 @item -mdisable-indexing
7896 @opindex mdisable-indexing
7897 Prevent the compiler from using indexing address modes. This avoids some
7898 rather obscure problems when compiling MIG generated code under MACH@.
7900 @item -mno-space-regs
7901 @opindex mno-space-regs
7902 Generate code that assumes the target has no space registers. This allows
7903 GCC to generate faster indirect calls and use unscaled index address modes.
7905 Such code is suitable for level 0 PA systems and kernels.
7907 @item -mfast-indirect-calls
7908 @opindex mfast-indirect-calls
7909 Generate code that assumes calls never cross space boundaries. This
7910 allows GCC to emit code which performs faster indirect calls.
7912 This option will not work in the presence of shared libraries or nested
7915 @item -mlong-load-store
7916 @opindex mlong-load-store
7917 Generate 3-instruction load and store sequences as sometimes required by
7918 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7921 @item -mportable-runtime
7922 @opindex mportable-runtime
7923 Use the portable calling conventions proposed by HP for ELF systems.
7927 Enable the use of assembler directives only GAS understands.
7929 @item -mschedule=@var{cpu-type}
7931 Schedule code according to the constraints for the machine type
7932 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7933 @samp{7100}, @samp{7100LC}, @samp{7200}, and @samp{8000}. Refer to
7934 @file{/usr/lib/sched.models} on an HP-UX system to determine the
7935 proper scheduling option for your machine.
7938 @opindex mlinker-opt
7939 Enable the optimization pass in the HPUX linker. Note this makes symbolic
7940 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
7941 in which they give bogus error messages when linking some programs.
7944 @opindex msoft-float
7945 Generate output containing library calls for floating point.
7946 @strong{Warning:} the requisite libraries are not available for all HPPA
7947 targets. Normally the facilities of the machine's usual C compiler are
7948 used, but this cannot be done directly in cross-compilation. You must make
7949 your own arrangements to provide suitable library functions for
7950 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7951 does provide software floating point support.
7953 @option{-msoft-float} changes the calling convention in the output file;
7954 therefore, it is only useful if you compile @emph{all} of a program with
7955 this option. In particular, you need to compile @file{libgcc.a}, the
7956 library that comes with GCC, with @option{-msoft-float} in order for
7960 @node Intel 960 Options
7961 @subsection Intel 960 Options
7963 These @samp{-m} options are defined for the Intel 960 implementations:
7966 @item -m@var{cpu-type}
7974 Assume the defaults for the machine type @var{cpu-type} for some of
7975 the other options, including instruction scheduling, floating point
7976 support, and addressing modes. The choices for @var{cpu-type} are
7977 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
7978 @samp{sa}, and @samp{sb}.
7985 @opindex msoft-float
7986 The @option{-mnumerics} option indicates that the processor does support
7987 floating-point instructions. The @option{-msoft-float} option indicates
7988 that floating-point support should not be assumed.
7990 @item -mleaf-procedures
7991 @itemx -mno-leaf-procedures
7992 @opindex mleaf-procedures
7993 @opindex mno-leaf-procedures
7994 Do (or do not) attempt to alter leaf procedures to be callable with the
7995 @code{bal} instruction as well as @code{call}. This will result in more
7996 efficient code for explicit calls when the @code{bal} instruction can be
7997 substituted by the assembler or linker, but less efficient code in other
7998 cases, such as calls via function pointers, or using a linker that doesn't
7999 support this optimization.
8002 @itemx -mno-tail-call
8004 @opindex mno-tail-call
8005 Do (or do not) make additional attempts (beyond those of the
8006 machine-independent portions of the compiler) to optimize tail-recursive
8007 calls into branches. You may not want to do this because the detection of
8008 cases where this is not valid is not totally complete. The default is
8009 @option{-mno-tail-call}.
8011 @item -mcomplex-addr
8012 @itemx -mno-complex-addr
8013 @opindex mcomplex-addr
8014 @opindex mno-complex-addr
8015 Assume (or do not assume) that the use of a complex addressing mode is a
8016 win on this implementation of the i960. Complex addressing modes may not
8017 be worthwhile on the K-series, but they definitely are on the C-series.
8018 The default is currently @option{-mcomplex-addr} for all processors except
8022 @itemx -mno-code-align
8023 @opindex mcode-align
8024 @opindex mno-code-align
8025 Align code to 8-byte boundaries for faster fetching (or don't bother).
8026 Currently turned on by default for C-series implementations only.
8029 @item -mclean-linkage
8030 @itemx -mno-clean-linkage
8031 @opindex mclean-linkage
8032 @opindex mno-clean-linkage
8033 These options are not fully implemented.
8037 @itemx -mic2.0-compat
8038 @itemx -mic3.0-compat
8040 @opindex mic2.0-compat
8041 @opindex mic3.0-compat
8042 Enable compatibility with iC960 v2.0 or v3.0.
8046 @opindex masm-compat
8048 Enable compatibility with the iC960 assembler.
8050 @item -mstrict-align
8051 @itemx -mno-strict-align
8052 @opindex mstrict-align
8053 @opindex mno-strict-align
8054 Do not permit (do permit) unaligned accesses.
8058 Enable structure-alignment compatibility with Intel's gcc release version
8059 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8061 @item -mlong-double-64
8062 @opindex mlong-double-64
8063 Implement type @samp{long double} as 64-bit floating point numbers.
8064 Without the option @samp{long double} is implemented by 80-bit
8065 floating point numbers. The only reason we have it because there is
8066 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8067 is only useful for people using soft-float targets. Otherwise, we
8068 should recommend against use of it.
8072 @node DEC Alpha Options
8073 @subsection DEC Alpha Options
8075 These @samp{-m} options are defined for the DEC Alpha implementations:
8078 @item -mno-soft-float
8080 @opindex mno-soft-float
8081 @opindex msoft-float
8082 Use (do not use) the hardware floating-point instructions for
8083 floating-point operations. When @option{-msoft-float} is specified,
8084 functions in @file{libgcc.a} will be used to perform floating-point
8085 operations. Unless they are replaced by routines that emulate the
8086 floating-point operations, or compiled in such a way as to call such
8087 emulations routines, these routines will issue floating-point
8088 operations. If you are compiling for an Alpha without floating-point
8089 operations, you must ensure that the library is built so as not to call
8092 Note that Alpha implementations without floating-point operations are
8093 required to have floating-point registers.
8098 @opindex mno-fp-regs
8099 Generate code that uses (does not use) the floating-point register set.
8100 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8101 register set is not used, floating point operands are passed in integer
8102 registers as if they were integers and floating-point results are passed
8103 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8104 so any function with a floating-point argument or return value called by code
8105 compiled with @option{-mno-fp-regs} must also be compiled with that
8108 A typical use of this option is building a kernel that does not use,
8109 and hence need not save and restore, any floating-point registers.
8113 The Alpha architecture implements floating-point hardware optimized for
8114 maximum performance. It is mostly compliant with the IEEE floating
8115 point standard. However, for full compliance, software assistance is
8116 required. This option generates code fully IEEE compliant code
8117 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8118 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8119 defined during compilation. The resulting code is less efficient but is
8120 able to correctly support denormalized numbers and exceptional IEEE
8121 values such as not-a-number and plus/minus infinity. Other Alpha
8122 compilers call this option @option{-ieee_with_no_inexact}.
8124 @item -mieee-with-inexact
8125 @opindex mieee-with-inexact
8126 This is like @option{-mieee} except the generated code also maintains
8127 the IEEE @var{inexact-flag}. Turning on this option causes the
8128 generated code to implement fully-compliant IEEE math. In addition to
8129 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8130 macro. On some Alpha implementations the resulting code may execute
8131 significantly slower than the code generated by default. Since there is
8132 very little code that depends on the @var{inexact-flag}, you should
8133 normally not specify this option. Other Alpha compilers call this
8134 option @option{-ieee_with_inexact}.
8136 @item -mfp-trap-mode=@var{trap-mode}
8137 @opindex mfp-trap-mode
8138 This option controls what floating-point related traps are enabled.
8139 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8140 The trap mode can be set to one of four values:
8144 This is the default (normal) setting. The only traps that are enabled
8145 are the ones that cannot be disabled in software (e.g., division by zero
8149 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8153 Like @samp{su}, but the instructions are marked to be safe for software
8154 completion (see Alpha architecture manual for details).
8157 Like @samp{su}, but inexact traps are enabled as well.
8160 @item -mfp-rounding-mode=@var{rounding-mode}
8161 @opindex mfp-rounding-mode
8162 Selects the IEEE rounding mode. Other Alpha compilers call this option
8163 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8168 Normal IEEE rounding mode. Floating point numbers are rounded towards
8169 the nearest machine number or towards the even machine number in case
8173 Round towards minus infinity.
8176 Chopped rounding mode. Floating point numbers are rounded towards zero.
8179 Dynamic rounding mode. A field in the floating point control register
8180 (@var{fpcr}, see Alpha architecture reference manual) controls the
8181 rounding mode in effect. The C library initializes this register for
8182 rounding towards plus infinity. Thus, unless your program modifies the
8183 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8186 @item -mtrap-precision=@var{trap-precision}
8187 @opindex mtrap-precision
8188 In the Alpha architecture, floating point traps are imprecise. This
8189 means without software assistance it is impossible to recover from a
8190 floating trap and program execution normally needs to be terminated.
8191 GCC can generate code that can assist operating system trap handlers
8192 in determining the exact location that caused a floating point trap.
8193 Depending on the requirements of an application, different levels of
8194 precisions can be selected:
8198 Program precision. This option is the default and means a trap handler
8199 can only identify which program caused a floating point exception.
8202 Function precision. The trap handler can determine the function that
8203 caused a floating point exception.
8206 Instruction precision. The trap handler can determine the exact
8207 instruction that caused a floating point exception.
8210 Other Alpha compilers provide the equivalent options called
8211 @option{-scope_safe} and @option{-resumption_safe}.
8213 @item -mieee-conformant
8214 @opindex mieee-conformant
8215 This option marks the generated code as IEEE conformant. You must not
8216 use this option unless you also specify @option{-mtrap-precision=i} and either
8217 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8218 is to emit the line @samp{.eflag 48} in the function prologue of the
8219 generated assembly file. Under DEC Unix, this has the effect that
8220 IEEE-conformant math library routines will be linked in.
8222 @item -mbuild-constants
8223 @opindex mbuild-constants
8224 Normally GCC examines a 32- or 64-bit integer constant to
8225 see if it can construct it from smaller constants in two or three
8226 instructions. If it cannot, it will output the constant as a literal and
8227 generate code to load it from the data segment at runtime.
8229 Use this option to require GCC to construct @emph{all} integer constants
8230 using code, even if it takes more instructions (the maximum is six).
8232 You would typically use this option to build a shared library dynamic
8233 loader. Itself a shared library, it must relocate itself in memory
8234 before it can find the variables and constants in its own data segment.
8240 Select whether to generate code to be assembled by the vendor-supplied
8241 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8259 Indicate whether GCC should generate code to use the optional BWX,
8260 CIX, FIX and MAX instruction sets. The default is to use the instruction
8261 sets supported by the CPU type specified via @option{-mcpu=} option or that
8262 of the CPU on which GCC was built if none was specified.
8267 @opindex mfloat-ieee
8268 Generate code that uses (does not use) VAX F and G floating point
8269 arithmetic instead of IEEE single and double precision.
8271 @item -mexplicit-relocs
8272 @itemx -mno-explicit-relocs
8273 @opindex mexplicit-relocs
8274 @opindex mno-explicit-relocs
8275 Older Alpha assemblers provided no way to generate symbol relocations
8276 except via assembler macros. Use of these macros does not allow
8277 optimial instruction scheduling. GNU binutils as of version 2.12
8278 supports a new syntax that allows the compiler to explicitly mark
8279 which relocations should apply to which instructions. This option
8280 is mostly useful for debugging, as GCC detects the capabilities of
8281 the assembler when it is built and sets the default accordingly.
8285 @opindex msmall-data
8286 @opindex mlarge-data
8287 When @option{-mexplicit-relocs} is in effect, static data is
8288 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8289 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8290 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8291 16-bit relocations off of the @code{$gp} register. This limits the
8292 size of the small data area to 64KB, but allows the variables to be
8293 directly accessed via a single instruction.
8295 The default is @option{-mlarge-data}. With this option the data area
8296 is limited to just below 2GB. Programs that require more than 2GB of
8297 data must use @code{malloc} or @code{mmap} to allocate the data in the
8298 heap instead of in the program's data segment.
8300 When generating code for shared libraries, @option{-fpic} implies
8301 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8303 @item -mcpu=@var{cpu_type}
8305 Set the instruction set and instruction scheduling parameters for
8306 machine type @var{cpu_type}. You can specify either the @samp{EV}
8307 style name or the corresponding chip number. GCC supports scheduling
8308 parameters for the EV4, EV5 and EV6 family of processors and will
8309 choose the default values for the instruction set from the processor
8310 you specify. If you do not specify a processor type, GCC will default
8311 to the processor on which the compiler was built.
8313 Supported values for @var{cpu_type} are
8319 Schedules as an EV4 and has no instruction set extensions.
8323 Schedules as an EV5 and has no instruction set extensions.
8327 Schedules as an EV5 and supports the BWX extension.
8332 Schedules as an EV5 and supports the BWX and MAX extensions.
8336 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8340 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8343 @item -mtune=@var{cpu_type}
8345 Set only the instruction scheduling parameters for machine type
8346 @var{cpu_type}. The instruction set is not changed.
8348 @item -mmemory-latency=@var{time}
8349 @opindex mmemory-latency
8350 Sets the latency the scheduler should assume for typical memory
8351 references as seen by the application. This number is highly
8352 dependent on the memory access patterns used by the application
8353 and the size of the external cache on the machine.
8355 Valid options for @var{time} are
8359 A decimal number representing clock cycles.
8365 The compiler contains estimates of the number of clock cycles for
8366 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8367 (also called Dcache, Scache, and Bcache), as well as to main memory.
8368 Note that L3 is only valid for EV5.
8373 @node DEC Alpha/VMS Options
8374 @subsection DEC Alpha/VMS Options
8376 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8379 @item -mvms-return-codes
8380 @opindex mvms-return-codes
8381 Return VMS condition codes from main. The default is to return POSIX
8382 style condition (e.g.@ error) codes.
8385 @node Clipper Options
8386 @subsection Clipper Options
8388 These @samp{-m} options are defined for the Clipper implementations:
8393 Produce code for a C300 Clipper processor. This is the default.
8397 Produce code for a C400 Clipper processor, i.e.@: use floating point
8401 @node H8/300 Options
8402 @subsection H8/300 Options
8404 These @samp{-m} options are defined for the H8/300 implementations:
8409 Shorten some address references at link time, when possible; uses the
8410 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8411 ld.info, Using ld}, for a fuller description.
8415 Generate code for the H8/300H@.
8419 Generate code for the H8/S@.
8423 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8427 Make @code{int} data 32 bits by default.
8431 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8432 The default for the H8/300H and H8/S is to align longs and floats on 4
8434 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8435 This option has no effect on the H8/300.
8439 @subsection SH Options
8441 These @samp{-m} options are defined for the SH implementations:
8446 Generate code for the SH1.
8450 Generate code for the SH2.
8454 Generate code for the SH3.
8458 Generate code for the SH3e.
8462 Generate code for the SH4 without a floating-point unit.
8464 @item -m4-single-only
8465 @opindex m4-single-only
8466 Generate code for the SH4 with a floating-point unit that only
8467 supports single-precision arithmetic.
8471 Generate code for the SH4 assuming the floating-point unit is in
8472 single-precision mode by default.
8476 Generate code for the SH4.
8480 Compile code for the processor in big endian mode.
8484 Compile code for the processor in little endian mode.
8488 Align doubles at 64-bit boundaries. Note that this changes the calling
8489 conventions, and thus some functions from the standard C library will
8490 not work unless you recompile it first with @option{-mdalign}.
8494 Shorten some address references at link time, when possible; uses the
8495 linker option @option{-relax}.
8499 Use 32-bit offsets in @code{switch} tables. The default is to use
8504 Enable the use of the instruction @code{fmovd}.
8508 Comply with the calling conventions defined by Hitachi.
8512 Mark the @code{MAC} register as call-clobbered, even if
8513 @option{-mhitachi} is given.
8517 Increase IEEE-compliance of floating-point code.
8521 Dump instruction size and location in the assembly code.
8525 This option is deprecated. It pads structures to multiple of 4 bytes,
8526 which is incompatible with the SH ABI@.
8530 Optimize for space instead of speed. Implied by @option{-Os}.
8534 When generating position-independent code, emit function calls using
8535 the Global Offset Table instead of the Procedure Linkage Table.
8539 Generate a library function call to invalidate instruction cache
8540 entries, after fixing up a trampoline. This library function call
8541 doesn't assume it can write to the whole memory address space. This
8542 is the default when the target is @code{sh-*-linux*}.
8545 @node System V Options
8546 @subsection Options for System V
8548 These additional options are available on System V Release 4 for
8549 compatibility with other compilers on those systems:
8554 Create a shared object.
8555 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8559 Identify the versions of each tool used by the compiler, in a
8560 @code{.ident} assembler directive in the output.
8564 Refrain from adding @code{.ident} directives to the output file (this is
8567 @item -YP,@var{dirs}
8569 Search the directories @var{dirs}, and no others, for libraries
8570 specified with @option{-l}.
8574 Look in the directory @var{dir} to find the M4 preprocessor.
8575 The assembler uses this option.
8576 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8577 @c the generic assembler that comes with Solaris takes just -Ym.
8580 @node TMS320C3x/C4x Options
8581 @subsection TMS320C3x/C4x Options
8582 @cindex TMS320C3x/C4x Options
8584 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8588 @item -mcpu=@var{cpu_type}
8590 Set the instruction set, register set, and instruction scheduling
8591 parameters for machine type @var{cpu_type}. Supported values for
8592 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8593 @samp{c44}. The default is @samp{c40} to generate code for the
8598 @itemx -msmall-memory
8600 @opindex mbig-memory
8602 @opindex msmall-memory
8604 Generates code for the big or small memory model. The small memory
8605 model assumed that all data fits into one 64K word page. At run-time
8606 the data page (DP) register must be set to point to the 64K page
8607 containing the .bss and .data program sections. The big memory model is
8608 the default and requires reloading of the DP register for every direct
8615 Allow (disallow) allocation of general integer operands into the block
8622 Enable (disable) generation of code using decrement and branch,
8623 DBcond(D), instructions. This is enabled by default for the C4x. To be
8624 on the safe side, this is disabled for the C3x, since the maximum
8625 iteration count on the C3x is @math{2^23 + 1} (but who iterates loops more than
8626 @math{2^23} times on the C3x?). Note that GCC will try to reverse a loop so
8627 that it can utilise the decrement and branch instruction, but will give
8628 up if there is more than one memory reference in the loop. Thus a loop
8629 where the loop counter is decremented can generate slightly more
8630 efficient code, in cases where the RPTB instruction cannot be utilised.
8632 @item -mdp-isr-reload
8634 @opindex mdp-isr-reload
8636 Force the DP register to be saved on entry to an interrupt service
8637 routine (ISR), reloaded to point to the data section, and restored on
8638 exit from the ISR@. This should not be required unless someone has
8639 violated the small memory model by modifying the DP register, say within
8646 For the C3x use the 24-bit MPYI instruction for integer multiplies
8647 instead of a library call to guarantee 32-bit results. Note that if one
8648 of the operands is a constant, then the multiplication will be performed
8649 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8650 then squaring operations are performed inline instead of a library call.
8653 @itemx -mno-fast-fix
8655 @opindex mno-fast-fix
8656 The C3x/C4x FIX instruction to convert a floating point value to an
8657 integer value chooses the nearest integer less than or equal to the
8658 floating point value rather than to the nearest integer. Thus if the
8659 floating point number is negative, the result will be incorrectly
8660 truncated an additional code is necessary to detect and correct this
8661 case. This option can be used to disable generation of the additional
8662 code required to correct the result.
8668 Enable (disable) generation of repeat block sequences using the RPTB
8669 instruction for zero overhead looping. The RPTB construct is only used
8670 for innermost loops that do not call functions or jump across the loop
8671 boundaries. There is no advantage having nested RPTB loops due to the
8672 overhead required to save and restore the RC, RS, and RE registers.
8673 This is enabled by default with @option{-O2}.
8675 @item -mrpts=@var{count}
8679 Enable (disable) the use of the single instruction repeat instruction
8680 RPTS@. If a repeat block contains a single instruction, and the loop
8681 count can be guaranteed to be less than the value @var{count}, GCC will
8682 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8683 then a RPTS will be emitted even if the loop count cannot be determined
8684 at compile time. Note that the repeated instruction following RPTS does
8685 not have to be reloaded from memory each iteration, thus freeing up the
8686 CPU buses for operands. However, since interrupts are blocked by this
8687 instruction, it is disabled by default.
8689 @item -mloop-unsigned
8690 @itemx -mno-loop-unsigned
8691 @opindex mloop-unsigned
8692 @opindex mno-loop-unsigned
8693 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8694 is @math{2^31 + 1} since these instructions test if the iteration count is
8695 negative to terminate the loop. If the iteration count is unsigned
8696 there is a possibility than the @math{2^31 + 1} maximum iteration count may be
8697 exceeded. This switch allows an unsigned iteration count.
8701 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8702 with. This also enforces compatibility with the API employed by the TI
8703 C3x C compiler. For example, long doubles are passed as structures
8704 rather than in floating point registers.
8710 Generate code that uses registers (stack) for passing arguments to functions.
8711 By default, arguments are passed in registers where possible rather
8712 than by pushing arguments on to the stack.
8714 @item -mparallel-insns
8715 @itemx -mno-parallel-insns
8716 @opindex mparallel-insns
8717 @opindex mno-parallel-insns
8718 Allow the generation of parallel instructions. This is enabled by
8719 default with @option{-O2}.
8721 @item -mparallel-mpy
8722 @itemx -mno-parallel-mpy
8723 @opindex mparallel-mpy
8724 @opindex mno-parallel-mpy
8725 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8726 provided @option{-mparallel-insns} is also specified. These instructions have
8727 tight register constraints which can pessimize the code generation
8733 @subsection V850 Options
8734 @cindex V850 Options
8736 These @samp{-m} options are defined for V850 implementations:
8740 @itemx -mno-long-calls
8741 @opindex mlong-calls
8742 @opindex mno-long-calls
8743 Treat all calls as being far away (near). If calls are assumed to be
8744 far away, the compiler will always load the functions address up into a
8745 register, and call indirect through the pointer.
8751 Do not optimize (do optimize) basic blocks that use the same index
8752 pointer 4 or more times to copy pointer into the @code{ep} register, and
8753 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8754 option is on by default if you optimize.
8756 @item -mno-prolog-function
8757 @itemx -mprolog-function
8758 @opindex mno-prolog-function
8759 @opindex mprolog-function
8760 Do not use (do use) external functions to save and restore registers at
8761 the prolog and epilog of a function. The external functions are slower,
8762 but use less code space if more than one function saves the same number
8763 of registers. The @option{-mprolog-function} option is on by default if
8768 Try to make the code as small as possible. At present, this just turns
8769 on the @option{-mep} and @option{-mprolog-function} options.
8773 Put static or global variables whose size is @var{n} bytes or less into
8774 the tiny data area that register @code{ep} points to. The tiny data
8775 area can hold up to 256 bytes in total (128 bytes for byte references).
8779 Put static or global variables whose size is @var{n} bytes or less into
8780 the small data area that register @code{gp} points to. The small data
8781 area can hold up to 64 kilobytes.
8785 Put static or global variables whose size is @var{n} bytes or less into
8786 the first 32 kilobytes of memory.
8790 Specify that the target processor is the V850.
8793 @opindex mbig-switch
8794 Generate code suitable for big switch tables. Use this option only if
8795 the assembler/linker complain about out of range branches within a switch
8800 @subsection ARC Options
8803 These options are defined for ARC implementations:
8808 Compile code for little endian mode. This is the default.
8812 Compile code for big endian mode.
8815 @opindex mmangle-cpu
8816 Prepend the name of the cpu to all public symbol names.
8817 In multiple-processor systems, there are many ARC variants with different
8818 instruction and register set characteristics. This flag prevents code
8819 compiled for one cpu to be linked with code compiled for another.
8820 No facility exists for handling variants that are ``almost identical''.
8821 This is an all or nothing option.
8823 @item -mcpu=@var{cpu}
8825 Compile code for ARC variant @var{cpu}.
8826 Which variants are supported depend on the configuration.
8827 All variants support @option{-mcpu=base}, this is the default.
8829 @item -mtext=@var{text-section}
8830 @itemx -mdata=@var{data-section}
8831 @itemx -mrodata=@var{readonly-data-section}
8835 Put functions, data, and readonly data in @var{text-section},
8836 @var{data-section}, and @var{readonly-data-section} respectively
8837 by default. This can be overridden with the @code{section} attribute.
8838 @xref{Variable Attributes}.
8843 @subsection NS32K Options
8844 @cindex NS32K options
8846 These are the @samp{-m} options defined for the 32000 series. The default
8847 values for these options depends on which style of 32000 was selected when
8848 the compiler was configured; the defaults for the most common choices are
8856 Generate output for a 32032. This is the default
8857 when the compiler is configured for 32032 and 32016 based systems.
8863 Generate output for a 32332. This is the default
8864 when the compiler is configured for 32332-based systems.
8870 Generate output for a 32532. This is the default
8871 when the compiler is configured for 32532-based systems.
8875 Generate output containing 32081 instructions for floating point.
8876 This is the default for all systems.
8880 Generate output containing 32381 instructions for floating point. This
8881 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8882 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8886 Try and generate multiply-add floating point instructions @code{polyF}
8887 and @code{dotF}. This option is only available if the @option{-m32381}
8888 option is in effect. Using these instructions requires changes to
8889 register allocation which generally has a negative impact on
8890 performance. This option should only be enabled when compiling code
8891 particularly likely to make heavy use of multiply-add instructions.
8894 @opindex mnomulti-add
8895 Do not try and generate multiply-add floating point instructions
8896 @code{polyF} and @code{dotF}. This is the default on all platforms.
8899 @opindex msoft-float
8900 Generate output containing library calls for floating point.
8901 @strong{Warning:} the requisite libraries may not be available.
8904 @opindex mnobitfield
8905 Do not use the bit-field instructions. On some machines it is faster to
8906 use shifting and masking operations. This is the default for the pc532.
8910 Do use the bit-field instructions. This is the default for all platforms
8915 Use a different function-calling convention, in which functions
8916 that take a fixed number of arguments return pop their
8917 arguments on return with the @code{ret} instruction.
8919 This calling convention is incompatible with the one normally
8920 used on Unix, so you cannot use it if you need to call libraries
8921 compiled with the Unix compiler.
8923 Also, you must provide function prototypes for all functions that
8924 take variable numbers of arguments (including @code{printf});
8925 otherwise incorrect code will be generated for calls to those
8928 In addition, seriously incorrect code will result if you call a
8929 function with too many arguments. (Normally, extra arguments are
8930 harmlessly ignored.)
8932 This option takes its name from the 680x0 @code{rtd} instruction.
8937 Use a different function-calling convention where the first two arguments
8938 are passed in registers.
8940 This calling convention is incompatible with the one normally
8941 used on Unix, so you cannot use it if you need to call libraries
8942 compiled with the Unix compiler.
8945 @opindex mnoregparam
8946 Do not pass any arguments in registers. This is the default for all
8951 It is OK to use the sb as an index register which is always loaded with
8952 zero. This is the default for the pc532-netbsd target.
8956 The sb register is not available for use or has not been initialized to
8957 zero by the run time system. This is the default for all targets except
8958 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
8959 @option{-fpic} is set.
8963 Many ns32000 series addressing modes use displacements of up to 512MB@.
8964 If an address is above 512MB then displacements from zero can not be used.
8965 This option causes code to be generated which can be loaded above 512MB@.
8966 This may be useful for operating systems or ROM code.
8970 Assume code will be loaded in the first 512MB of virtual address space.
8971 This is the default for all platforms.
8977 @subsection AVR Options
8980 These options are defined for AVR implementations:
8983 @item -mmcu=@var{mcu}
8985 Specify ATMEL AVR instruction set or MCU type.
8987 Instruction set avr1 is for the minimal AVR core, not supported by the C
8988 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8989 attiny11, attiny12, attiny15, attiny28).
8991 Instruction set avr2 (default) is for the classic AVR core with up to
8992 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8993 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8994 at90c8534, at90s8535).
8996 Instruction set avr3 is for the classic AVR core with up to 128K program
8997 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8999 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9000 memory space (MCU types: atmega8, atmega83, atmega85).
9002 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9003 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9004 atmega64, atmega128, at43usb355, at94k).
9008 Output instruction sizes to the asm file.
9010 @item -minit-stack=@var{N}
9011 @opindex minit-stack
9012 Specify the initial stack address, which may be a symbol or numeric value,
9013 @samp{__stack} is the default.
9015 @item -mno-interrupts
9016 @opindex mno-interrupts
9017 Generated code is not compatible with hardware interrupts.
9018 Code size will be smaller.
9020 @item -mcall-prologues
9021 @opindex mcall-prologues
9022 Functions prologues/epilogues expanded as call to appropriate
9023 subroutines. Code size will be smaller.
9025 @item -mno-tablejump
9026 @opindex mno-tablejump
9027 Do not generate tablejump insns which sometimes increase code size.
9030 @opindex mtiny-stack
9031 Change only the low 8 bits of the stack pointer.
9035 @subsection MCore Options
9036 @cindex MCore options
9038 These are the @samp{-m} options defined for the Motorola M*Core
9048 @opindex mno-hardlit
9049 Inline constants into the code stream if it can be done in two
9050 instructions or less.
9058 Use the divide instruction. (Enabled by default).
9060 @item -mrelax-immediate
9061 @itemx -mrelax-immediate
9062 @itemx -mno-relax-immediate
9063 @opindex mrelax-immediate
9064 @opindex mrelax-immediate
9065 @opindex mno-relax-immediate
9066 Allow arbitrary sized immediates in bit operations.
9068 @item -mwide-bitfields
9069 @itemx -mwide-bitfields
9070 @itemx -mno-wide-bitfields
9071 @opindex mwide-bitfields
9072 @opindex mwide-bitfields
9073 @opindex mno-wide-bitfields
9074 Always treat bit-fields as int-sized.
9076 @item -m4byte-functions
9077 @itemx -m4byte-functions
9078 @itemx -mno-4byte-functions
9079 @opindex m4byte-functions
9080 @opindex m4byte-functions
9081 @opindex mno-4byte-functions
9082 Force all functions to be aligned to a four byte boundary.
9084 @item -mcallgraph-data
9085 @itemx -mcallgraph-data
9086 @itemx -mno-callgraph-data
9087 @opindex mcallgraph-data
9088 @opindex mcallgraph-data
9089 @opindex mno-callgraph-data
9090 Emit callgraph information.
9094 @itemx -mno-slow-bytes
9095 @opindex mslow-bytes
9096 @opindex mslow-bytes
9097 @opindex mno-slow-bytes
9098 Prefer word access when reading byte quantities.
9100 @item -mlittle-endian
9101 @itemx -mlittle-endian
9103 @opindex mlittle-endian
9104 @opindex mlittle-endian
9105 @opindex mbig-endian
9106 Generate code for a little endian target.
9114 Generate code for the 210 processor.
9118 @subsection IA-64 Options
9119 @cindex IA-64 Options
9121 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9125 @opindex mbig-endian
9126 Generate code for a big endian target. This is the default for HPUX@.
9128 @item -mlittle-endian
9129 @opindex mlittle-endian
9130 Generate code for a little endian target. This is the default for AIX5
9137 Generate (or don't) code for the GNU assembler. This is the default.
9138 @c Also, this is the default if the configure option @option{--with-gnu-as}
9145 Generate (or don't) code for the GNU linker. This is the default.
9146 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9151 Generate code that does not use a global pointer register. The result
9152 is not position independent code, and violates the IA-64 ABI@.
9154 @item -mvolatile-asm-stop
9155 @itemx -mno-volatile-asm-stop
9156 @opindex mvolatile-asm-stop
9157 @opindex mno-volatile-asm-stop
9158 Generate (or don't) a stop bit immediately before and after volatile asm
9163 Generate code that works around Itanium B step errata.
9165 @item -mregister-names
9166 @itemx -mno-register-names
9167 @opindex mregister-names
9168 @opindex mno-register-names
9169 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9170 the stacked registers. This may make assembler output more readable.
9176 Disable (or enable) optimizations that use the small data section. This may
9177 be useful for working around optimizer bugs.
9180 @opindex mconstant-gp
9181 Generate code that uses a single constant global pointer value. This is
9182 useful when compiling kernel code.
9186 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9187 This is useful when compiling firmware code.
9189 @item -minline-divide-min-latency
9190 @opindex minline-divide-min-latency
9191 Generate code for inline divides using the minimum latency algorithm.
9193 @item -minline-divide-max-throughput
9194 @opindex minline-divide-max-throughput
9195 Generate code for inline divides using the maximum throughput algorithm.
9197 @item -mno-dwarf2-asm
9199 @opindex mno-dwarf2-asm
9200 @opindex mdwarf2-asm
9201 Don't (or do) generate assembler code for the DWARF2 line number debugging
9202 info. This may be useful when not using the GNU assembler.
9204 @item -mfixed-range=@var{register-range}
9205 @opindex mfixed-range
9206 Generate code treating the given register range as fixed registers.
9207 A fixed register is one that the register allocator can not use. This is
9208 useful when compiling kernel code. A register range is specified as
9209 two registers separated by a dash. Multiple register ranges can be
9210 specified separated by a comma.
9214 @subsection D30V Options
9215 @cindex D30V Options
9217 These @samp{-m} options are defined for D30V implementations:
9222 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
9223 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9224 memory, which starts at location @code{0x80000000}.
9228 Same as the @option{-mextmem} switch.
9232 Link the @samp{.text} section into onchip text memory, which starts at
9233 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9234 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9235 into onchip data memory, which starts at location @code{0x20000000}.
9237 @item -mno-asm-optimize
9238 @itemx -masm-optimize
9239 @opindex mno-asm-optimize
9240 @opindex masm-optimize
9241 Disable (enable) passing @option{-O} to the assembler when optimizing.
9242 The assembler uses the @option{-O} option to automatically parallelize
9243 adjacent short instructions where possible.
9245 @item -mbranch-cost=@var{n}
9246 @opindex mbranch-cost
9247 Increase the internal costs of branches to @var{n}. Higher costs means
9248 that the compiler will issue more instructions to avoid doing a branch.
9251 @item -mcond-exec=@var{n}
9253 Specify the maximum number of conditionally executed instructions that
9254 replace a branch. The default is 4.
9257 @node S/390 and zSeries Options
9258 @subsection S/390 and zSeries Options
9259 @cindex S/390 and zSeries Options
9261 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9266 @opindex mhard-float
9267 @opindex msoft-float
9268 Use (do not use) the hardware floating-point instructions and registers
9269 for floating-point operations. When @option{-msoft-float} is specified,
9270 functions in @file{libgcc.a} will be used to perform floating-point
9271 operations. When @option{-mhard-float} is specified, the compiler
9272 generates IEEE floating-point instructions. This is the default.
9275 @itemx -mno-backchain
9277 @opindex mno-backchain
9278 Generate (or do not generate) code which maintains an explicit
9279 backchain within the stack frame that points to the caller's frame.
9280 This is currently needed to allow debugging. The default is to
9281 generate the backchain.
9284 @itemx -mno-small-exec
9285 @opindex msmall-exec
9286 @opindex mno-small-exec
9287 Generate (or do not generate) code using the @code{bras} instruction
9288 to do subroutine calls.
9289 This only works reliably if the total executable size does not
9290 exceed 64k. The default is to use the @code{basr} instruction instead,
9291 which does not have this limitation.
9297 When @option{-m31} is specified, generate code compliant to the
9298 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9299 code compliant to the Linux for zSeries ABI@. This allows GCC in
9300 particular to generate 64-bit instructions. For the @samp{s390}
9301 targets, the default is @option{-m31}, while the @samp{s390x}
9302 targets default to @option{-m64}.
9308 Generate (or do not generate) code using the @code{mvcle} instruction
9309 to perform block moves. When @option{-mno-mvcle} is specifed,
9310 use a @code{mvc} loop instead. This is the default.
9316 Print (or do not print) additional debug information when compiling.
9317 The default is to not print debug information.
9322 @subsection CRIS Options
9323 @cindex CRIS Options
9325 These options are defined specifically for the CRIS ports.
9328 @item -march=@var{architecture-type}
9329 @itemx -mcpu=@var{architecture-type}
9332 Generate code for the specified architecture. The choices for
9333 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9334 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9335 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9338 @item -mtune=@var{architecture-type}
9340 Tune to @var{architecture-type} everything applicable about the generated
9341 code, except for the ABI and the set of available instructions. The
9342 choices for @var{architecture-type} are the same as for
9343 @option{-march=@var{architecture-type}}.
9345 @item -mmax-stack-frame=@var{n}
9346 @opindex mmax-stack-frame
9347 Warn when the stack frame of a function exceeds @var{n} bytes.
9349 @item -melinux-stacksize=@var{n}
9350 @opindex melinux-stacksize
9351 Only available with the @samp{cris-axis-aout} target. Arranges for
9352 indications in the program to the kernel loader that the stack of the
9353 program should be set to @var{n} bytes.
9359 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9360 @option{-march=v3} and @option{-march=v8} respectively.
9364 Enable CRIS-specific verbose debug-related information in the assembly
9365 code. This option also has the effect to turn off the @samp{#NO_APP}
9366 formatted-code indicator to the assembler at the beginning of the
9371 Do not use condition-code results from previous instruction; always emit
9372 compare and test instructions before use of condition codes.
9374 @item -mno-side-effects
9375 @opindex mno-side-effects
9376 Do not emit instructions with side-effects in addressing modes other than
9380 @itemx -mno-stack-align
9382 @itemx -mno-data-align
9383 @itemx -mconst-align
9384 @itemx -mno-const-align
9385 @opindex mstack-align
9386 @opindex mno-stack-align
9387 @opindex mdata-align
9388 @opindex mno-data-align
9389 @opindex mconst-align
9390 @opindex mno-const-align
9391 These options (no-options) arranges (eliminate arrangements) for the
9392 stack-frame, individual data and constants to be aligned for the maximum
9393 single data access size for the chosen CPU model. The default is to
9394 arrange for 32-bit alignment. ABI details such as structure layout are
9395 not affected by these options.
9403 Similar to the stack- data- and const-align options above, these options
9404 arrange for stack-frame, writable data and constants to all be 32-bit,
9405 16-bit or 8-bit aligned. The default is 32-bit alignment.
9407 @item -mno-prologue-epilogue
9408 @itemx -mprologue-epilogue
9409 @opindex mno-prologue-epilogue
9410 @opindex mprologue-epilogue
9411 With @option{-mno-prologue-epilogue}, the normal function prologue and
9412 epilogue that sets up the stack-frame are omitted and no return
9413 instructions or return sequences are generated in the code. Use this
9414 option only together with visual inspection of the compiled code: no
9415 warnings or errors are generated when call-saved registers must be saved,
9416 or storage for local variable needs to be allocated.
9422 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9423 instruction sequences that load addresses for functions from the PLT part
9424 of the GOT rather than (traditional on other architectures) calls to the
9425 PLT. The default is @option{-mgotplt}.
9429 Legacy no-op option only recognized with the cris-axis-aout target.
9433 Legacy no-op option only recognized with the cris-axis-elf and
9434 cris-axis-linux-gnu targets.
9438 Only recognized with the cris-axis-aout target, where it selects a
9439 GNU/linux-like multilib, include files and instruction set for
9444 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9448 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9449 to link with input-output functions from a simulator library. Code,
9450 initialized data and zero-initialized data are allocated consecutively.
9454 Like @option{-sim}, but pass linker options to locate initialized data at
9455 0x40000000 and zero-initialized data at 0x80000000.
9459 @subsection MMIX Options
9460 @cindex MMIX Options
9462 These options are defined for the MMIX:
9466 @itemx -mno-libfuncs
9468 @opindex mno-libfuncs
9469 Specify that intrinsic library functions are being compiled, passing all
9470 values in registers, no matter the size.
9475 @opindex mno-epsilon
9476 Generate floating-point comparison instructions that compare with respect
9477 to the @code{rE} epsilon register.
9479 @item -mabi=mmixware
9481 @opindex mabi-mmixware
9483 Generate code that passes function parameters and return values that (in
9484 the called function) are seen as registers @code{$0} and up, as opposed to
9485 the GNU ABI which uses global registers @code{$231} and up.
9488 @itemx -mno-zero-extend
9489 @opindex mzero-extend
9490 @opindex mno-zero-extend
9491 When reading data from memory in sizes shorter than 64 bits, use (do not
9492 use) zero-extending load instructions by default, rather than
9493 sign-extending ones.
9496 @itemx -mno-knuthdiv
9498 @opindex mno-knuthdiv
9499 Make the result of a division yielding a remainder have the same sign as
9500 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9501 remainder follows the sign of the dividend. Both methods are
9502 arithmetically valid, the latter being almost exclusively used.
9504 @item -mtoplevel-symbols
9505 @itemx -mno-toplevel-symbols
9506 @opindex mtoplevel-symbols
9507 @opindex mno-toplevel-symbols
9508 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9509 code can be used with the @code{PREFIX} assembly directive.
9513 Generate an executable in the ELF format, rather than the default
9514 @samp{mmo} format used by the @command{mmix} simulator.
9516 @item -mbranch-predict
9517 @itemx -mno-branch-predict
9518 @opindex mbranch-predict
9519 @opindex mno-branch-predict
9520 Use (do not use) the probable-branch instructions, when static branch
9521 prediction indicates a probable branch.
9523 @item -mbase-addresses
9524 @itemx -mno-base-addresses
9525 @opindex mbase-addresses
9526 @opindex mno-base-addresses
9527 Generate (do not generate) code that uses @emph{base addresses}. Using a
9528 base address automatically generates a request (handled by the assembler
9529 and the linker) for a constant to be set up in a global register. The
9530 register is used for one or more base address requests within the range 0
9531 to 255 from the value held in the register. The generally leads to short
9532 and fast code, but the number of different data items that can be
9533 addressed is limited. This means that a program that uses lots of static
9534 data may require @option{-mno-base-addresses}.
9537 @node PDP-11 Options
9538 @subsection PDP-11 Options
9539 @cindex PDP-11 Options
9541 These options are defined for the PDP-11:
9546 Use hardware FPP floating point. This is the default. (FIS floating
9547 point on the PDP-11/40 is not supported.)
9550 @opindex msoft-float
9551 Do not use hardware floating point.
9555 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9559 Return floating-point results in memory. This is the default.
9563 Generate code for a PDP-11/40.
9567 Generate code for a PDP-11/45. This is the default.
9571 Generate code for a PDP-11/10.
9573 @item -mbcopy-builtin
9574 @opindex bcopy-builtin
9575 Use inline @code{movstrhi} patterns for copying memory. This is the
9580 Do not use inline @code{movstrhi} patterns for copying memory.
9586 Use 16-bit @code{int}. This is the default.
9592 Use 32-bit @code{int}.
9597 @opindex mno-float32
9598 Use 64-bit @code{float}. This is the default.
9603 @opindex mno-float64
9604 Use 32-bit @code{float}.
9608 Use @code{abshi2} pattern. This is the default.
9612 Do not use @code{abshi2} pattern.
9614 @item -mbranch-expensive
9615 @opindex mbranch-expensive
9616 Pretend that branches are expensive. This is for experimenting with
9617 code generation only.
9619 @item -mbranch-cheap
9620 @opindex mbranch-cheap
9621 Do not pretend that branches are expensive. This is the default.
9625 Generate code for a system with split I&D.
9629 Generate code for a system without split I&D. This is the default.
9633 Use Unix assembler syntax. This is the default when configured for
9638 Use DEC assembler syntax. This is the default when configured for any
9639 PDP-11 target other than @samp{pdp11-*-bsd}.
9642 @node Xstormy16 Options
9643 @subsection Xstormy16 Options
9644 @cindex Xstormy16 Options
9646 These options are defined for Xstormy16:
9651 Choose startup files and linker script suitable for the simulator.
9654 @node Xtensa Options
9655 @subsection Xtensa Options
9656 @cindex Xtensa Options
9658 The Xtensa architecture is designed to support many different
9659 configurations. The compiler's default options can be set to match a
9660 particular Xtensa configuration by copying a configuration file into the
9661 GCC sources when building GCC@. The options below may be used to
9662 override the default options.
9666 @itemx -mlittle-endian
9667 @opindex mbig-endian
9668 @opindex mlittle-endian
9669 Specify big-endian or little-endian byte ordering for the target Xtensa
9675 @opindex mno-density
9676 Enable or disable use of the optional Xtensa code density instructions.
9682 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
9683 will generate MAC16 instructions from standard C code, with the
9684 limitation that it will use neither the MR register file nor any
9685 instruction that operates on the MR registers. When this option is
9686 disabled, GCC will translate 16-bit multiply/accumulate operations to a
9687 combination of core instructions and library calls, depending on whether
9688 any other multiplier options are enabled.
9694 Enable or disable use of the 16-bit integer multiplier option. When
9695 enabled, the compiler will generate 16-bit multiply instructions for
9696 multiplications of 16 bits or smaller in standard C code. When this
9697 option is disabled, the compiler will either use 32-bit multiply or
9698 MAC16 instructions if they are available or generate library calls to
9699 perform the multiply operations using shifts and adds.
9705 Enable or disable use of the 32-bit integer multiplier option. When
9706 enabled, the compiler will generate 32-bit multiply instructions for
9707 multiplications of 32 bits or smaller in standard C code. When this
9708 option is disabled, the compiler will generate library calls to perform
9709 the multiply operations using either shifts and adds or 16-bit multiply
9710 instructions if they are available.
9716 Enable or disable use of the optional normalization shift amount
9717 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
9723 Enable or disable use of the optional minimum and maximum value
9730 Enable or disable use of the optional sign extend (@code{SEXT})
9734 @itemx -mno-booleans
9736 @opindex mno-booleans
9737 Enable or disable support for the boolean register file used by Xtensa
9738 coprocessors. This is not typically useful by itself but may be
9739 required for other options that make use of the boolean registers (e.g.,
9740 the floating-point option).
9744 @opindex mhard-float
9745 @opindex msoft-float
9746 Enable or disable use of the floating-point option. When enabled, GCC
9747 generates floating-point instructions for 32-bit @code{float}
9748 operations. When this option is disabled, GCC generates library calls
9749 to emulate 32-bit floating-point operations using integer instructions.
9750 Regardless of this option, 64-bit @code{double} operations are always
9751 emulated with calls to library functions.
9754 @itemx -mno-fused-madd
9755 @opindex mfused-madd
9756 @opindex mno-fused-madd
9757 Enable or disable use of fused multiply/add and multiply/subtract
9758 instructions in the floating-point option. This has no effect if the
9759 floating-point option is not also enabled. Disabling fused multiply/add
9760 and multiply/subtract instructions forces the compiler to use separate
9761 instructions for the multiply and add/subtract operations. This may be
9762 desirable in some cases where strict IEEE 754-compliant results are
9763 required: the fused multiply add/subtract instructions do not round the
9764 intermediate result, thereby producing results with @emph{more} bits of
9765 precision than specified by the IEEE standard. Disabling fused multiply
9766 add/subtract instructions also ensures that the program output is not
9767 sensitive to the compiler's ability to combine multiply and add/subtract
9770 @item -mserialize-volatile
9771 @itemx -mno-serialize-volatile
9772 @opindex mserialize-volatile
9773 @opindex mno-serialize-volatile
9774 When this option is enabled, GCC inserts @code{MEMW} instructions before
9775 @code{volatile} memory references to guarantee sequential consistency.
9776 The default is @option{-mserialize-volatile}. Use
9777 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
9779 @item -mtext-section-literals
9780 @itemx -mno-text-section-literals
9781 @opindex mtext-section-literals
9782 @opindex mno-text-section-literals
9783 Control the treatment of literal pools. The default is
9784 @option{-mno-text-section-literals}, which places literals in a separate
9785 section in the output file. This allows the literal pool to be placed
9786 in a data RAM/ROM, and it also allows the linker to combine literal
9787 pools from separate object files to remove redundant literals and
9788 improve code size. With @option{-mtext-section-literals}, the literals
9789 are interspersed in the text section in order to keep them as close as
9790 possible to their references. This may be necessary for large assembly
9793 @item -mtarget-align
9794 @itemx -mno-target-align
9795 @opindex mtarget-align
9796 @opindex mno-target-align
9797 When this option is enabled, GCC instructs the assembler to
9798 automatically align instructions to reduce branch penalties at the
9799 expense of some code density. The assembler attempts to widen density
9800 instructions to align branch targets and the instructions following call
9801 instructions. If there are not enough preceding safe density
9802 instructions to align a target, no widening will be performed. The
9803 default is @option{-mtarget-align}. These options do not affect the
9804 treatment of auto-aligned instructions like @code{LOOP}, which the
9805 assembler will always align, either by widening density instructions or
9806 by inserting no-op instructions.
9809 @itemx -mno-longcalls
9811 @opindex mno-longcalls
9812 When this option is enabled, GCC instructs the assembler to translate
9813 direct calls to indirect calls unless it can determine that the target
9814 of a direct call is in the range allowed by the call instruction. This
9815 translation typically occurs for calls to functions in other source
9816 files. Specifically, the assembler translates a direct @code{CALL}
9817 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
9818 The default is @option{-mno-longcalls}. This option should be used in
9819 programs where the call target can potentially be out of range. This
9820 option is implemented in the assembler, not the compiler, so the
9821 assembly code generated by GCC will still show direct call
9822 instructions---look at the disassembled object code to see the actual
9823 instructions. Note that the assembler will use an indirect call for
9824 every cross-file call, not just those that really will be out of range.
9827 @node Code Gen Options
9828 @section Options for Code Generation Conventions
9829 @cindex code generation conventions
9830 @cindex options, code generation
9831 @cindex run-time options
9833 These machine-independent options control the interface conventions
9834 used in code generation.
9836 Most of them have both positive and negative forms; the negative form
9837 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9838 one of the forms is listed---the one which is not the default. You
9839 can figure out the other form by either removing @samp{no-} or adding
9844 @opindex fexceptions
9845 Enable exception handling. Generates extra code needed to propagate
9846 exceptions. For some targets, this implies GCC will generate frame
9847 unwind information for all functions, which can produce significant data
9848 size overhead, although it does not affect execution. If you do not
9849 specify this option, GCC will enable it by default for languages like
9850 C++ which normally require exception handling, and disable it for
9851 languages like C that do not normally require it. However, you may need
9852 to enable this option when compiling C code that needs to interoperate
9853 properly with exception handlers written in C++. You may also wish to
9854 disable this option if you are compiling older C++ programs that don't
9855 use exception handling.
9857 @item -fnon-call-exceptions
9858 @opindex fnon-call-exceptions
9859 Generate code that allows trapping instructions to throw exceptions.
9860 Note that this requires platform-specific runtime support that does
9861 not exist everywhere. Moreover, it only allows @emph{trapping}
9862 instructions to throw exceptions, i.e.@: memory references or floating
9863 point instructions. It does not allow exceptions to be thrown from
9864 arbitrary signal handlers such as @code{SIGALRM}.
9866 @item -funwind-tables
9867 @opindex funwind-tables
9868 Similar to @option{-fexceptions}, except that it will just generate any needed
9869 static data, but will not affect the generated code in any other way.
9870 You will normally not enable this option; instead, a language processor
9871 that needs this handling would enable it on your behalf.
9873 @item -fasynchronous-unwind-tables
9874 @opindex funwind-tables
9875 Generate unwind table in dwarf2 format, if supported by target machine. The
9876 table is exact at each instruction boundary, so it can be used for stack
9877 unwinding from asynchronous events (such as debugger or garbage collector).
9879 @item -fpcc-struct-return
9880 @opindex fpcc-struct-return
9881 Return ``short'' @code{struct} and @code{union} values in memory like
9882 longer ones, rather than in registers. This convention is less
9883 efficient, but it has the advantage of allowing intercallability between
9884 GCC-compiled files and files compiled with other compilers.
9886 The precise convention for returning structures in memory depends
9887 on the target configuration macros.
9889 Short structures and unions are those whose size and alignment match
9890 that of some integer type.
9892 @item -freg-struct-return
9893 @opindex freg-struct-return
9894 Return @code{struct} and @code{union} values in registers when possible.
9895 This is more efficient for small structures than
9896 @option{-fpcc-struct-return}.
9898 If you specify neither @option{-fpcc-struct-return} nor
9899 @option{-freg-struct-return}, GCC defaults to whichever convention is
9900 standard for the target. If there is no standard convention, GCC
9901 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9902 the principal compiler. In those cases, we can choose the standard, and
9903 we chose the more efficient register return alternative.
9906 @opindex fshort-enums
9907 Allocate to an @code{enum} type only as many bytes as it needs for the
9908 declared range of possible values. Specifically, the @code{enum} type
9909 will be equivalent to the smallest integer type which has enough room.
9911 @item -fshort-double
9912 @opindex fshort-double
9913 Use the same size for @code{double} as for @code{float}.
9916 @opindex fshared-data
9917 Requests that the data and non-@code{const} variables of this
9918 compilation be shared data rather than private data. The distinction
9919 makes sense only on certain operating systems, where shared data is
9920 shared between processes running the same program, while private data
9921 exists in one copy per process.
9925 In C, allocate even uninitialized global variables in the data section of the
9926 object file, rather than generating them as common blocks. This has the
9927 effect that if the same variable is declared (without @code{extern}) in
9928 two different compilations, you will get an error when you link them.
9929 The only reason this might be useful is if you wish to verify that the
9930 program will work on other systems which always work this way.
9934 Ignore the @samp{#ident} directive.
9936 @item -fno-gnu-linker
9937 @opindex fno-gnu-linker
9938 Do not output global initializations (such as C++ constructors and
9939 destructors) in the form used by the GNU linker (on systems where the GNU
9940 linker is the standard method of handling them). Use this option when
9941 you want to use a non-GNU linker, which also requires using the
9942 @command{collect2} program to make sure the system linker includes
9943 constructors and destructors. (@command{collect2} is included in the GCC
9944 distribution.) For systems which @emph{must} use @command{collect2}, the
9945 compiler driver @command{gcc} is configured to do this automatically.
9947 @item -finhibit-size-directive
9948 @opindex finhibit-size-directive
9949 Don't output a @code{.size} assembler directive, or anything else that
9950 would cause trouble if the function is split in the middle, and the
9951 two halves are placed at locations far apart in memory. This option is
9952 used when compiling @file{crtstuff.c}; you should not need to use it
9956 @opindex fverbose-asm
9957 Put extra commentary information in the generated assembly code to
9958 make it more readable. This option is generally only of use to those
9959 who actually need to read the generated assembly code (perhaps while
9960 debugging the compiler itself).
9962 @option{-fno-verbose-asm}, the default, causes the
9963 extra information to be omitted and is useful when comparing two assembler
9968 Consider all memory references through pointers to be volatile.
9970 @item -fvolatile-global
9971 @opindex fvolatile-global
9972 Consider all memory references to extern and global data items to
9973 be volatile. GCC does not consider static data items to be volatile
9974 because of this switch.
9976 @item -fvolatile-static
9977 @opindex fvolatile-static
9978 Consider all memory references to static data to be volatile.
9982 @cindex global offset table
9984 Generate position-independent code (PIC) suitable for use in a shared
9985 library, if supported for the target machine. Such code accesses all
9986 constant addresses through a global offset table (GOT)@. The dynamic
9987 loader resolves the GOT entries when the program starts (the dynamic
9988 loader is not part of GCC; it is part of the operating system). If
9989 the GOT size for the linked executable exceeds a machine-specific
9990 maximum size, you get an error message from the linker indicating that
9991 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9992 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9993 on the m68k and RS/6000. The 386 has no such limit.)
9995 Position-independent code requires special support, and therefore works
9996 only on certain machines. For the 386, GCC supports PIC for System V
9997 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9998 position-independent.
10002 If supported for the target machine, emit position-independent code,
10003 suitable for dynamic linking and avoiding any limit on the size of the
10004 global offset table. This option makes a difference on the m68k, m88k,
10007 Position-independent code requires special support, and therefore works
10008 only on certain machines.
10010 @item -ffixed-@var{reg}
10012 Treat the register named @var{reg} as a fixed register; generated code
10013 should never refer to it (except perhaps as a stack pointer, frame
10014 pointer or in some other fixed role).
10016 @var{reg} must be the name of a register. The register names accepted
10017 are machine-specific and are defined in the @code{REGISTER_NAMES}
10018 macro in the machine description macro file.
10020 This flag does not have a negative form, because it specifies a
10023 @item -fcall-used-@var{reg}
10024 @opindex fcall-used
10025 Treat the register named @var{reg} as an allocable register that is
10026 clobbered by function calls. It may be allocated for temporaries or
10027 variables that do not live across a call. Functions compiled this way
10028 will not save and restore the register @var{reg}.
10030 It is an error to used this flag with the frame pointer or stack pointer.
10031 Use of this flag for other registers that have fixed pervasive roles in
10032 the machine's execution model will produce disastrous results.
10034 This flag does not have a negative form, because it specifies a
10037 @item -fcall-saved-@var{reg}
10038 @opindex fcall-saved
10039 Treat the register named @var{reg} as an allocable register saved by
10040 functions. It may be allocated even for temporaries or variables that
10041 live across a call. Functions compiled this way will save and restore
10042 the register @var{reg} if they use it.
10044 It is an error to used this flag with the frame pointer or stack pointer.
10045 Use of this flag for other registers that have fixed pervasive roles in
10046 the machine's execution model will produce disastrous results.
10048 A different sort of disaster will result from the use of this flag for
10049 a register in which function values may be returned.
10051 This flag does not have a negative form, because it specifies a
10054 @item -fpack-struct
10055 @opindex fpack-struct
10056 Pack all structure members together without holes. Usually you would
10057 not want to use this option, since it makes the code suboptimal, and
10058 the offsets of structure members won't agree with system libraries.
10060 @item -finstrument-functions
10061 @opindex finstrument-functions
10062 Generate instrumentation calls for entry and exit to functions. Just
10063 after function entry and just before function exit, the following
10064 profiling functions will be called with the address of the current
10065 function and its call site. (On some platforms,
10066 @code{__builtin_return_address} does not work beyond the current
10067 function, so the call site information may not be available to the
10068 profiling functions otherwise.)
10071 void __cyg_profile_func_enter (void *this_fn,
10073 void __cyg_profile_func_exit (void *this_fn,
10077 The first argument is the address of the start of the current function,
10078 which may be looked up exactly in the symbol table.
10080 This instrumentation is also done for functions expanded inline in other
10081 functions. The profiling calls will indicate where, conceptually, the
10082 inline function is entered and exited. This means that addressable
10083 versions of such functions must be available. If all your uses of a
10084 function are expanded inline, this may mean an additional expansion of
10085 code size. If you use @samp{extern inline} in your C code, an
10086 addressable version of such functions must be provided. (This is
10087 normally the case anyways, but if you get lucky and the optimizer always
10088 expands the functions inline, you might have gotten away without
10089 providing static copies.)
10091 A function may be given the attribute @code{no_instrument_function}, in
10092 which case this instrumentation will not be done. This can be used, for
10093 example, for the profiling functions listed above, high-priority
10094 interrupt routines, and any functions from which the profiling functions
10095 cannot safely be called (perhaps signal handlers, if the profiling
10096 routines generate output or allocate memory).
10098 @item -fstack-check
10099 @opindex fstack-check
10100 Generate code to verify that you do not go beyond the boundary of the
10101 stack. You should specify this flag if you are running in an
10102 environment with multiple threads, but only rarely need to specify it in
10103 a single-threaded environment since stack overflow is automatically
10104 detected on nearly all systems if there is only one stack.
10106 Note that this switch does not actually cause checking to be done; the
10107 operating system must do that. The switch causes generation of code
10108 to ensure that the operating system sees the stack being extended.
10110 @item -fstack-limit-register=@var{reg}
10111 @itemx -fstack-limit-symbol=@var{sym}
10112 @itemx -fno-stack-limit
10113 @opindex fstack-limit-register
10114 @opindex fstack-limit-symbol
10115 @opindex fno-stack-limit
10116 Generate code to ensure that the stack does not grow beyond a certain value,
10117 either the value of a register or the address of a symbol. If the stack
10118 would grow beyond the value, a signal is raised. For most targets,
10119 the signal is raised before the stack overruns the boundary, so
10120 it is possible to catch the signal without taking special precautions.
10122 For instance, if the stack starts at absolute address @samp{0x80000000}
10123 and grows downwards, you can use the flags
10124 @option{-fstack-limit-symbol=__stack_limit} and
10125 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
10126 of 128KB@. Note that this may only work with the GNU linker.
10128 @cindex aliasing of parameters
10129 @cindex parameters, aliased
10130 @item -fargument-alias
10131 @itemx -fargument-noalias
10132 @itemx -fargument-noalias-global
10133 @opindex fargument-alias
10134 @opindex fargument-noalias
10135 @opindex fargument-noalias-global
10136 Specify the possible relationships among parameters and between
10137 parameters and global data.
10139 @option{-fargument-alias} specifies that arguments (parameters) may
10140 alias each other and may alias global storage.@*
10141 @option{-fargument-noalias} specifies that arguments do not alias
10142 each other, but may alias global storage.@*
10143 @option{-fargument-noalias-global} specifies that arguments do not
10144 alias each other and do not alias global storage.
10146 Each language will automatically use whatever option is required by
10147 the language standard. You should not need to use these options yourself.
10149 @item -fleading-underscore
10150 @opindex fleading-underscore
10151 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
10152 change the way C symbols are represented in the object file. One use
10153 is to help link with legacy assembly code.
10155 Be warned that you should know what you are doing when invoking this
10156 option, and that not all targets provide complete support for it.
10161 @node Environment Variables
10162 @section Environment Variables Affecting GCC
10163 @cindex environment variables
10165 @c man begin ENVIRONMENT
10167 This section describes several environment variables that affect how GCC
10168 operates. Some of them work by specifying directories or prefixes to use
10169 when searching for various kinds of files. Some are used to specify other
10170 aspects of the compilation environment.
10172 Note that you can also specify places to search using options such as
10173 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
10174 take precedence over places specified using environment variables, which
10175 in turn take precedence over those specified by the configuration of GCC@.
10176 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
10177 GNU Compiler Collection (GCC) Internals}.
10182 @c @itemx LC_COLLATE
10184 @c @itemx LC_MONETARY
10185 @c @itemx LC_NUMERIC
10190 @c @findex LC_COLLATE
10191 @findex LC_MESSAGES
10192 @c @findex LC_MONETARY
10193 @c @findex LC_NUMERIC
10197 These environment variables control the way that GCC uses
10198 localization information that allow GCC to work with different
10199 national conventions. GCC inspects the locale categories
10200 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
10201 so. These locale categories can be set to any value supported by your
10202 installation. A typical value is @samp{en_UK} for English in the United
10205 The @env{LC_CTYPE} environment variable specifies character
10206 classification. GCC uses it to determine the character boundaries in
10207 a string; this is needed for some multibyte encodings that contain quote
10208 and escape characters that would otherwise be interpreted as a string
10211 The @env{LC_MESSAGES} environment variable specifies the language to
10212 use in diagnostic messages.
10214 If the @env{LC_ALL} environment variable is set, it overrides the value
10215 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10216 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10217 environment variable. If none of these variables are set, GCC
10218 defaults to traditional C English behavior.
10222 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10223 files. GCC uses temporary files to hold the output of one stage of
10224 compilation which is to be used as input to the next stage: for example,
10225 the output of the preprocessor, which is the input to the compiler
10228 @item GCC_EXEC_PREFIX
10229 @findex GCC_EXEC_PREFIX
10230 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10231 names of the subprograms executed by the compiler. No slash is added
10232 when this prefix is combined with the name of a subprogram, but you can
10233 specify a prefix that ends with a slash if you wish.
10235 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10236 an appropriate prefix to use based on the pathname it was invoked with.
10238 If GCC cannot find the subprogram using the specified prefix, it
10239 tries looking in the usual places for the subprogram.
10241 The default value of @env{GCC_EXEC_PREFIX} is
10242 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10243 of @code{prefix} when you ran the @file{configure} script.
10245 Other prefixes specified with @option{-B} take precedence over this prefix.
10247 This prefix is also used for finding files such as @file{crt0.o} that are
10250 In addition, the prefix is used in an unusual way in finding the
10251 directories to search for header files. For each of the standard
10252 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10253 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10254 replacing that beginning with the specified prefix to produce an
10255 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10256 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10257 These alternate directories are searched first; the standard directories
10260 @item COMPILER_PATH
10261 @findex COMPILER_PATH
10262 The value of @env{COMPILER_PATH} is a colon-separated list of
10263 directories, much like @env{PATH}. GCC tries the directories thus
10264 specified when searching for subprograms, if it can't find the
10265 subprograms using @env{GCC_EXEC_PREFIX}.
10268 @findex LIBRARY_PATH
10269 The value of @env{LIBRARY_PATH} is a colon-separated list of
10270 directories, much like @env{PATH}. When configured as a native compiler,
10271 GCC tries the directories thus specified when searching for special
10272 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10273 using GCC also uses these directories when searching for ordinary
10274 libraries for the @option{-l} option (but directories specified with
10275 @option{-L} come first).
10277 @item C_INCLUDE_PATH
10278 @itemx CPLUS_INCLUDE_PATH
10279 @itemx OBJC_INCLUDE_PATH
10280 @findex C_INCLUDE_PATH
10281 @findex CPLUS_INCLUDE_PATH
10282 @findex OBJC_INCLUDE_PATH
10283 @c @itemx OBJCPLUS_INCLUDE_PATH
10284 These environment variables pertain to particular languages. Each
10285 variable's value is a colon-separated list of directories, much like
10286 @env{PATH}. When GCC searches for header files, it tries the
10287 directories listed in the variable for the language you are using, after
10288 the directories specified with @option{-I} but before the standard header
10291 @item DEPENDENCIES_OUTPUT
10292 @findex DEPENDENCIES_OUTPUT
10293 @cindex dependencies for make as output
10294 If this variable is set, its value specifies how to output dependencies
10295 for Make based on the header files processed by the compiler. This
10296 output looks much like the output from the @option{-M} option
10297 (@pxref{Preprocessor Options}), but it goes to a separate file, and is
10298 in addition to the usual results of compilation.
10300 The value of @env{DEPENDENCIES_OUTPUT} can be just a file name, in
10301 which case the Make rules are written to that file, guessing the target
10302 name from the source file name. Or the value can have the form
10303 @samp{@var{file} @var{target}}, in which case the rules are written to
10304 file @var{file} using @var{target} as the target name.
10308 @cindex locale definition
10309 This variable is used to pass locale information to the compiler. One way in
10310 which this information is used is to determine the character set to be used
10311 when character literals, string literals and comments are parsed in C and C++.
10312 When the compiler is configured to allow multibyte characters,
10313 the following values for @env{LANG} are recognized:
10317 Recognize JIS characters.
10319 Recognize SJIS characters.
10321 Recognize EUCJP characters.
10324 If @env{LANG} is not defined, or if it has some other value, then the
10325 compiler will use mblen and mbtowc as defined by the default locale to
10326 recognize and translate multibyte characters.
10331 @node Running Protoize
10332 @section Running Protoize
10334 The program @code{protoize} is an optional part of GCC@. You can use
10335 it to add prototypes to a program, thus converting the program to ISO
10336 C in one respect. The companion program @code{unprotoize} does the
10337 reverse: it removes argument types from any prototypes that are found.
10339 When you run these programs, you must specify a set of source files as
10340 command line arguments. The conversion programs start out by compiling
10341 these files to see what functions they define. The information gathered
10342 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10344 After scanning comes actual conversion. The specified files are all
10345 eligible to be converted; any files they include (whether sources or
10346 just headers) are eligible as well.
10348 But not all the eligible files are converted. By default,
10349 @code{protoize} and @code{unprotoize} convert only source and header
10350 files in the current directory. You can specify additional directories
10351 whose files should be converted with the @option{-d @var{directory}}
10352 option. You can also specify particular files to exclude with the
10353 @option{-x @var{file}} option. A file is converted if it is eligible, its
10354 directory name matches one of the specified directory names, and its
10355 name within the directory has not been excluded.
10357 Basic conversion with @code{protoize} consists of rewriting most
10358 function definitions and function declarations to specify the types of
10359 the arguments. The only ones not rewritten are those for varargs
10362 @code{protoize} optionally inserts prototype declarations at the
10363 beginning of the source file, to make them available for any calls that
10364 precede the function's definition. Or it can insert prototype
10365 declarations with block scope in the blocks where undeclared functions
10368 Basic conversion with @code{unprotoize} consists of rewriting most
10369 function declarations to remove any argument types, and rewriting
10370 function definitions to the old-style pre-ISO form.
10372 Both conversion programs print a warning for any function declaration or
10373 definition that they can't convert. You can suppress these warnings
10376 The output from @code{protoize} or @code{unprotoize} replaces the
10377 original source file. The original file is renamed to a name ending
10378 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10379 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10380 for DOS) file already exists, then the source file is simply discarded.
10382 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10383 scan the program and collect information about the functions it uses.
10384 So neither of these programs will work until GCC is installed.
10386 Here is a table of the options you can use with @code{protoize} and
10387 @code{unprotoize}. Each option works with both programs unless
10391 @item -B @var{directory}
10392 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10393 usual directory (normally @file{/usr/local/lib}). This file contains
10394 prototype information about standard system functions. This option
10395 applies only to @code{protoize}.
10397 @item -c @var{compilation-options}
10398 Use @var{compilation-options} as the options when running @code{gcc} to
10399 produce the @samp{.X} files. The special option @option{-aux-info} is
10400 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10402 Note that the compilation options must be given as a single argument to
10403 @code{protoize} or @code{unprotoize}. If you want to specify several
10404 @code{gcc} options, you must quote the entire set of compilation options
10405 to make them a single word in the shell.
10407 There are certain @code{gcc} arguments that you cannot use, because they
10408 would produce the wrong kind of output. These include @option{-g},
10409 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10410 the @var{compilation-options}, they are ignored.
10413 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10414 systems) instead of @samp{.c}. This is convenient if you are converting
10415 a C program to C++. This option applies only to @code{protoize}.
10418 Add explicit global declarations. This means inserting explicit
10419 declarations at the beginning of each source file for each function
10420 that is called in the file and was not declared. These declarations
10421 precede the first function definition that contains a call to an
10422 undeclared function. This option applies only to @code{protoize}.
10424 @item -i @var{string}
10425 Indent old-style parameter declarations with the string @var{string}.
10426 This option applies only to @code{protoize}.
10428 @code{unprotoize} converts prototyped function definitions to old-style
10429 function definitions, where the arguments are declared between the
10430 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10431 uses five spaces as the indentation. If you want to indent with just
10432 one space instead, use @option{-i " "}.
10435 Keep the @samp{.X} files. Normally, they are deleted after conversion
10439 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10440 a prototype declaration for each function in each block which calls the
10441 function without any declaration. This option applies only to
10445 Make no real changes. This mode just prints information about the conversions
10446 that would have been done without @option{-n}.
10449 Make no @samp{.save} files. The original files are simply deleted.
10450 Use this option with caution.
10452 @item -p @var{program}
10453 Use the program @var{program} as the compiler. Normally, the name
10454 @file{gcc} is used.
10457 Work quietly. Most warnings are suppressed.
10460 Print the version number, just like @option{-v} for @code{gcc}.
10463 If you need special compiler options to compile one of your program's
10464 source files, then you should generate that file's @samp{.X} file
10465 specially, by running @code{gcc} on that source file with the
10466 appropriate options and the option @option{-aux-info}. Then run
10467 @code{protoize} on the entire set of files. @code{protoize} will use
10468 the existing @samp{.X} file because it is newer than the source file.
10472 gcc -Dfoo=bar file1.c -aux-info file1.X
10477 You need to include the special files along with the rest in the
10478 @code{protoize} command, even though their @samp{.X} files already
10479 exist, because otherwise they won't get converted.
10481 @xref{Protoize Caveats}, for more information on how to use
10482 @code{protoize} successfully.