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 -fomit-frame-pointer -foptimize-register-move @gol
277 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
278 -freduce-all-givs -fregmove -frename-registers @gol
279 -frerun-cse-after-loop -frerun-loop-opt @gol
280 -fschedule-insns -fschedule-insns2 @gol
281 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
282 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol
283 -funroll-all-loops -funroll-loops @gol
284 --param @var{name}=@var{value}
285 -O -O0 -O1 -O2 -O3 -Os}
287 @item Preprocessor Options
288 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
290 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
291 -C -dD -dI -dM -dN @gol
292 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
293 -idirafter @var{dir} @gol
294 -include @var{file} -imacros @var{file} @gol
295 -iprefix @var{file} -iwithprefix @var{dir} @gol
296 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
297 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
298 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
300 @item Assembler Option
301 @xref{Assembler Options,,Passing Options to the Assembler}.
306 @xref{Link Options,,Options for Linking}.
308 @var{object-file-name} -l@var{library} @gol
309 -nostartfiles -nodefaultlibs -nostdlib @gol
310 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
311 -Wl,@var{option} -Xlinker @var{option} @gol
314 @item Directory Options
315 @xref{Directory Options,,Options for Directory Search}.
317 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
320 @c I wrote this xref this way to avoid overfull hbox. -- rms
321 @xref{Target Options}.
323 -b @var{machine} -V @var{version}}
325 @item Machine Dependent Options
326 @xref{Submodel Options,,Hardware Models and Configurations}.
328 @emph{M680x0 Options}
330 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
331 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
332 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
333 -malign-int -mstrict-align}
335 @emph{M68hc1x Options}
337 -m6811 -m6812 -m68hc11 -m68hc12 @gol
338 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
346 -mcpu=@var{cpu-type} @gol
347 -mtune=@var{cpu-type} @gol
348 -mcmodel=@var{code-model} @gol
350 -mapp-regs -mbroken-saverestore -mcypress @gol
351 -mepilogue -mfaster-structs -mflat @gol
352 -mfpu -mhard-float -mhard-quad-float @gol
353 -mimpure-text -mlive-g0 -mno-app-regs @gol
354 -mno-epilogue -mno-faster-structs -mno-flat -mno-fpu @gol
355 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
356 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
357 -msupersparc -munaligned-doubles -mv8}
359 @emph{Convex Options}
361 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
362 -margcount -mnoargcount @gol
363 -mlong32 -mlong64 @gol
364 -mvolatile-cache -mvolatile-nocache}
366 @emph{AMD29K Options}
368 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
369 -mlarge -mnormal -msmall @gol
370 -mkernel-registers -mno-reuse-arg-regs @gol
371 -mno-stack-check -mno-storem-bug @gol
372 -mreuse-arg-regs -msoft-float -mstack-check @gol
373 -mstorem-bug -muser-registers}
377 -mapcs-frame -mno-apcs-frame @gol
378 -mapcs-26 -mapcs-32 @gol
379 -mapcs-stack-check -mno-apcs-stack-check @gol
380 -mapcs-float -mno-apcs-float @gol
381 -mapcs-reentrant -mno-apcs-reentrant @gol
382 -msched-prolog -mno-sched-prolog @gol
383 -mlittle-endian -mbig-endian -mwords-little-endian @gol
384 -malignment-traps -mno-alignment-traps @gol
385 -msoft-float -mhard-float -mfpe @gol
386 -mthumb-interwork -mno-thumb-interwork @gol
387 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
388 -mstructure-size-boundary=@var{n} @gol
389 -mbsd -mxopen -mno-symrename @gol
390 -mabort-on-noreturn @gol
391 -mlong-calls -mno-long-calls @gol
392 -msingle-pic-base -mno-single-pic-base @gol
393 -mpic-register=@var{reg} @gol
394 -mnop-fun-dllimport @gol
395 -mpoke-function-name @gol
397 -mtpcs-frame -mtpcs-leaf-frame @gol
398 -mcaller-super-interworking -mcallee-super-interworking }
400 @emph{MN10200 Options}
404 @emph{MN10300 Options}
406 -mmult-bug -mno-mult-bug @gol
407 -mam33 -mno-am33 @gol
410 @emph{M32R/D Options}
412 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
417 -m88000 -m88100 -m88110 -mbig-pic @gol
418 -mcheck-zero-division -mhandle-large-shift @gol
419 -midentify-revision -mno-check-zero-division @gol
420 -mno-ocs-debug-info -mno-ocs-frame-position @gol
421 -mno-optimize-arg-area -mno-serialize-volatile @gol
422 -mno-underscores -mocs-debug-info @gol
423 -mocs-frame-position -moptimize-arg-area @gol
424 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
425 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
426 -mversion-03.00 -mwarn-passed-structs}
428 @emph{RS/6000 and PowerPC Options}
430 -mcpu=@var{cpu-type} @gol
431 -mtune=@var{cpu-type} @gol
432 -mpower -mno-power -mpower2 -mno-power2 @gol
433 -mpowerpc -mpowerpc64 -mno-powerpc @gol
434 -maltivec -mno-altivec @gol
435 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
436 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
437 -mnew-mnemonics -mold-mnemonics @gol
438 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
439 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
440 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
441 -mstring -mno-string -mupdate -mno-update @gol
442 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
443 -mstrict-align -mno-strict-align -mrelocatable @gol
444 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
445 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
446 -mcall-aix -mcall-sysv -mcall-netbsd @gol
447 -maix-struct-return -msvr4-struct-return
449 -mprototype -mno-prototype @gol
450 -msim -mmvme -mads -myellowknife -memb -msdata @gol
451 -msdata=@var{opt} -mvxworks -G @var{num} -pthread}
455 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
456 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
457 -mminimum-fp-blocks -mnohc-struct-return}
461 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
462 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
463 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
464 -mgas -mgp32 -mgp64 @gol
465 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
466 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
467 -mmips-as -mmips-tfile -mno-abicalls @gol
468 -mno-embedded-data -mno-uninit-const-in-rodata @gol
469 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
470 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
471 -mrnames -msoft-float @gol
472 -m4650 -msingle-float -mmad @gol
473 -mstats -EL -EB -G @var{num} -nocpp @gol
474 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
475 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
477 @emph{i386 and x86-64 Options}
479 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
480 -masm=@var{dialect} -mno-fancy-math-387 @gol
481 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
482 -mno-wide-multiply -mrtd -malign-double @gol
483 -mpreferred-stack-boundary=@var{num} @gol
484 -mmmx -msse -msse2 -msse-math -m3dnow @gol
485 -mthreads -mno-align-stringops -minline-all-stringops @gol
486 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
487 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
493 -march=@var{architecture-type} @gol
494 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
495 -mfast-indirect-calls -mgas -mjump-in-delay @gol
496 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
497 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
498 -mno-jump-in-delay -mno-long-load-store @gol
499 -mno-portable-runtime -mno-soft-float @gol
500 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
501 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
502 -mschedule=@var{cpu-type} -mspace-regs}
504 @emph{Intel 960 Options}
506 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
507 -mcode-align -mcomplex-addr -mleaf-procedures @gol
508 -mic-compat -mic2.0-compat -mic3.0-compat @gol
509 -mintel-asm -mno-clean-linkage -mno-code-align @gol
510 -mno-complex-addr -mno-leaf-procedures @gol
511 -mno-old-align -mno-strict-align -mno-tail-call @gol
512 -mnumerics -mold-align -msoft-float -mstrict-align @gol
515 @emph{DEC Alpha Options}
517 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
518 -mieee -mieee-with-inexact -mieee-conformant @gol
519 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
520 -mtrap-precision=@var{mode} -mbuild-constants @gol
521 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
522 -mbwx -mmax -mfix -mcix @gol
523 -mfloat-vax -mfloat-ieee @gol
524 -mexplicit-relocs -msmall-data -mlarge-data @gol
525 -mmemory-latency=@var{time}}
527 @emph{DEC Alpha/VMS Options}
531 @emph{Clipper Options}
535 @emph{H8/300 Options}
537 -mrelax -mh -ms -mint32 -malign-300}
541 -m1 -m2 -m3 -m3e @gol
542 -m4-nofpu -m4-single-only -m4-single -m4 @gol
543 -m5-64media -m5-64media-nofpu @gol
544 -m5-32media -m5-32media-nofpu @gol
545 -m5-compact -m5-compact-nofpu @gol
546 -mb -ml -mdalign -mrelax @gol
547 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
548 -mieee -misize -mpadstruct -mspace @gol
549 -mprefergot -musermode}
551 @emph{System V Options}
553 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
558 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
559 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
561 @emph{TMS320C3x/C4x Options}
563 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
564 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
565 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
566 -mparallel-insns -mparallel-mpy -mpreserve-float}
570 -mlong-calls -mno-long-calls -mep -mno-ep @gol
571 -mprolog-function -mno-prolog-function -mspace @gol
572 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
577 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
578 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
579 -mregparam -mnoregparam -msb -mnosb @gol
580 -mbitfield -mnobitfield -mhimem -mnohimem}
584 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
585 -mcall-prologues -mno-tablejump -mtiny-stack}
589 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
590 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
591 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
592 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
593 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
597 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
598 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
599 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
604 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
605 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
606 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
607 -minline-divide-max-throughput -mno-dwarf2-asm @gol
608 -mfixed-range=@var{register-range}}
612 -mextmem -mextmemory -monchip -mno-asm-optimize -masm-optimize @gol
613 -mbranch-cost=@var{n} -mcond-exec=@var{n}}
615 @emph{S/390 and zSeries Options}
617 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
618 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
619 -m64 -m31 -mdebug -mno-debug}
623 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
624 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
625 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
626 -mstack-align -mdata-align -mconst-align @gol
627 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
628 -melf -maout -melinux -mlinux -sim -sim2}
630 @emph{PDP-11 Options}
632 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
633 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
634 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
635 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
636 -mbranch-expensive -mbranch-cheap @gol
637 -msplit -mno-split -munix-asm -mdec-asm}
639 @emph{Xstormy16 Options}
643 @emph{Xtensa Options}
645 -mbig-endian -mlittle-endian @gol
646 -mdensity -mno-density @gol
647 -mmac16 -mno-mac16 @gol
648 -mmul16 -mno-mul16 @gol
649 -mmul32 -mno-mul32 @gol
651 -mminmax -mno-minmax @gol
652 -msext -mno-sext @gol
653 -mbooleans -mno-booleans @gol
654 -mhard-float -msoft-float @gol
655 -mfused-madd -mno-fused-madd @gol
656 -mserialize-volatile -mno-serialize-volatile @gol
657 -mtext-section-literals -mno-text-section-literals @gol
658 -mtarget-align -mno-target-align @gol
659 -mlongcalls -mno-longcalls}
661 @item Code Generation Options
662 @xref{Code Gen Options,,Options for Code Generation Conventions}.
664 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
665 -ffixed-@var{reg} -fexceptions @gol
666 -fnon-call-exceptions -funwind-tables @gol
667 -fasynchronous-unwind-tables @gol
668 -finhibit-size-directive -finstrument-functions @gol
669 -fno-common -fno-ident -fno-gnu-linker @gol
670 -fpcc-struct-return -fpic -fPIC @gol
671 -freg-struct-return -fshared-data -fshort-enums @gol
672 -fshort-double -fvolatile @gol
673 -fvolatile-global -fvolatile-static @gol
674 -fverbose-asm -fpack-struct -fstack-check @gol
675 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
676 -fargument-alias -fargument-noalias @gol
677 -fargument-noalias-global -fleading-underscore}
681 * Overall Options:: Controlling the kind of output:
682 an executable, object files, assembler files,
683 or preprocessed source.
684 * C Dialect Options:: Controlling the variant of C language compiled.
685 * C++ Dialect Options:: Variations on C++.
686 * Objective-C Dialect Options:: Variations on Objective-C.
687 * Language Independent Options:: Controlling how diagnostics should be
689 * Warning Options:: How picky should the compiler be?
690 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
691 * Optimize Options:: How much optimization?
692 * Preprocessor Options:: Controlling header files and macro definitions.
693 Also, getting dependency information for Make.
694 * Assembler Options:: Passing options to the assembler.
695 * Link Options:: Specifying libraries and so on.
696 * Directory Options:: Where to find header files and libraries.
697 Where to find the compiler executable files.
698 * Spec Files:: How to pass switches to sub-processes.
699 * Target Options:: Running a cross-compiler, or an old version of GCC.
702 @node Overall Options
703 @section Options Controlling the Kind of Output
705 Compilation can involve up to four stages: preprocessing, compilation
706 proper, assembly and linking, always in that order. The first three
707 stages apply to an individual source file, and end by producing an
708 object file; linking combines all the object files (those newly
709 compiled, and those specified as input) into an executable file.
711 @cindex file name suffix
712 For any given input file, the file name suffix determines what kind of
717 C source code which must be preprocessed.
720 C source code which should not be preprocessed.
723 C++ source code which should not be preprocessed.
726 Objective-C source code. Note that you must link with the library
727 @file{libobjc.a} to make an Objective-C program work.
730 Objective-C source code which should not be preprocessed.
733 C header file (not to be compiled or linked).
737 @itemx @var{file}.cxx
738 @itemx @var{file}.cpp
739 @itemx @var{file}.c++
741 C++ source code which must be preprocessed. Note that in @samp{.cxx},
742 the last two letters must both be literally @samp{x}. Likewise,
743 @samp{.C} refers to a literal capital C@.
746 @itemx @var{file}.for
747 @itemx @var{file}.FOR
748 Fortran source code which should not be preprocessed.
751 @itemx @var{file}.fpp
752 @itemx @var{file}.FPP
753 Fortran source code which must be preprocessed (with the traditional
757 Fortran source code which must be preprocessed with a RATFOR
758 preprocessor (not included with GCC)@.
760 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
761 Using and Porting GNU Fortran}, for more details of the handling of
764 @c FIXME: Descriptions of Java file types.
771 Ada source code file which contains a library unit declaration (a
772 declaration of a package, subprogram, or generic, or a generic
773 instantiation), or a library unit renaming declaration (a package,
774 generic, or subprogram renaming declaration). Such files are also
777 @itemx @var{file}.adb
778 Ada source code file containing a library unit body (a subprogram or
779 package body). Such files are also called @dfn{bodies}.
781 @c GCC also knows about some suffixes for languages not yet included:
787 @itemx @var{file}.chi
788 CHILL source code (preprocessed with the traditional preprocessor).
794 Assembler code which must be preprocessed.
797 An object file to be fed straight into linking.
798 Any file name with no recognized suffix is treated this way.
802 You can specify the input language explicitly with the @option{-x} option:
805 @item -x @var{language}
806 Specify explicitly the @var{language} for the following input files
807 (rather than letting the compiler choose a default based on the file
808 name suffix). This option applies to all following input files until
809 the next @option{-x} option. Possible values for @var{language} are:
811 c c-header cpp-output
813 objective-c objc-cpp-output
814 assembler assembler-with-cpp
817 f77 f77-cpp-input ratfor
822 Turn off any specification of a language, so that subsequent files are
823 handled according to their file name suffixes (as they are if @option{-x}
824 has not been used at all).
826 @item -pass-exit-codes
827 @opindex pass-exit-codes
828 Normally the @command{gcc} program will exit with the code of 1 if any
829 phase of the compiler returns a non-success return code. If you specify
830 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
831 numerically highest error produced by any phase that returned an error
835 If you only want some of the stages of compilation, you can use
836 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
837 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
838 @command{gcc} is to stop. Note that some combinations (for example,
839 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
844 Compile or assemble the source files, but do not link. The linking
845 stage simply is not done. The ultimate output is in the form of an
846 object file for each source file.
848 By default, the object file name for a source file is made by replacing
849 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
851 Unrecognized input files, not requiring compilation or assembly, are
856 Stop after the stage of compilation proper; do not assemble. The output
857 is in the form of an assembler code file for each non-assembler input
860 By default, the assembler file name for a source file is made by
861 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
863 Input files that don't require compilation are ignored.
867 Stop after the preprocessing stage; do not run the compiler proper. The
868 output is in the form of preprocessed source code, which is sent to the
871 Input files which don't require preprocessing are ignored.
873 @cindex output file option
876 Place output in file @var{file}. This applies regardless to whatever
877 sort of output is being produced, whether it be an executable file,
878 an object file, an assembler file or preprocessed C code.
880 Since only one output file can be specified, it does not make sense to
881 use @option{-o} when compiling more than one input file, unless you are
882 producing an executable file as output.
884 If @option{-o} is not specified, the default is to put an executable file
885 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
886 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
887 all preprocessed C source on standard output.
891 Print (on standard error output) the commands executed to run the stages
892 of compilation. Also print the version number of the compiler driver
893 program and of the preprocessor and the compiler proper.
897 Use pipes rather than temporary files for communication between the
898 various stages of compilation. This fails to work on some systems where
899 the assembler is unable to read from a pipe; but the GNU assembler has
904 Print (on the standard output) a description of the command line options
905 understood by @command{gcc}. If the @option{-v} option is also specified
906 then @option{--help} will also be passed on to the various processes
907 invoked by @command{gcc}, so that they can display the command line options
908 they accept. If the @option{-W} option is also specified then command
909 line options which have no documentation associated with them will also
914 Print (on the standard output) a description of target specific command
915 line options for each tool.
919 @section Compiling C++ Programs
921 @cindex suffixes for C++ source
922 @cindex C++ source file suffixes
923 C++ source files conventionally use one of the suffixes @samp{.C},
924 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
925 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
926 files with these names and compiles them as C++ programs even if you
927 call the compiler the same way as for compiling C programs (usually with
928 the name @command{gcc}).
932 However, C++ programs often require class libraries as well as a
933 compiler that understands the C++ language---and under some
934 circumstances, you might want to compile programs from standard input,
935 or otherwise without a suffix that flags them as C++ programs.
936 @command{g++} is a program that calls GCC with the default language
937 set to C++, and automatically specifies linking against the C++
938 library. On many systems, @command{g++} is also
939 installed with the name @command{c++}.
941 @cindex invoking @command{g++}
942 When you compile C++ programs, you may specify many of the same
943 command-line options that you use for compiling programs in any
944 language; or command-line options meaningful for C and related
945 languages; or options that are meaningful only for C++ programs.
946 @xref{C Dialect Options,,Options Controlling C Dialect}, for
947 explanations of options for languages related to C@.
948 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
949 explanations of options that are meaningful only for C++ programs.
951 @node C Dialect Options
952 @section Options Controlling C Dialect
953 @cindex dialect options
954 @cindex language dialect options
955 @cindex options, dialect
957 The following options control the dialect of C (or languages derived
958 from C, such as C++ and Objective-C) that the compiler accepts:
965 In C mode, support all ISO C89 programs. In C++ mode,
966 remove GNU extensions that conflict with ISO C++.
968 This turns off certain features of GCC that are incompatible with ISO
969 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
970 such as the @code{asm} and @code{typeof} keywords, and
971 predefined macros such as @code{unix} and @code{vax} that identify the
972 type of system you are using. It also enables the undesirable and
973 rarely used ISO trigraph feature. For the C compiler,
974 it disables recognition of C++ style @samp{//} comments as well as
975 the @code{inline} keyword.
977 The alternate keywords @code{__asm__}, @code{__extension__},
978 @code{__inline__} and @code{__typeof__} continue to work despite
979 @option{-ansi}. You would not want to use them in an ISO C program, of
980 course, but it is useful to put them in header files that might be included
981 in compilations done with @option{-ansi}. Alternate predefined macros
982 such as @code{__unix__} and @code{__vax__} are also available, with or
983 without @option{-ansi}.
985 The @option{-ansi} option does not cause non-ISO programs to be
986 rejected gratuitously. For that, @option{-pedantic} is required in
987 addition to @option{-ansi}. @xref{Warning Options}.
989 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
990 option is used. Some header files may notice this macro and refrain
991 from declaring certain functions or defining certain macros that the
992 ISO standard doesn't call for; this is to avoid interfering with any
993 programs that might use these names for other things.
995 Functions which would normally be built in but do not have semantics
996 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
997 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
998 built-in functions provided by GCC}, for details of the functions
1003 Determine the language standard. This option is currently only
1004 supported when compiling C@. A value for this option must be provided;
1010 ISO C89 (same as @option{-ansi}).
1012 @item iso9899:199409
1013 ISO C89 as modified in amendment 1.
1019 ISO C99. Note that this standard is not yet fully supported; see
1020 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1021 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1024 Default, ISO C89 plus GNU extensions (including some C99 features).
1028 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1029 this will become the default. The name @samp{gnu9x} is deprecated.
1033 Even when this option is not specified, you can still use some of the
1034 features of newer standards in so far as they do not conflict with
1035 previous C standards. For example, you may use @code{__restrict__} even
1036 when @option{-std=c99} is not specified.
1038 The @option{-std} options specifying some version of ISO C have the same
1039 effects as @option{-ansi}, except that features that were not in ISO C89
1040 but are in the specified version (for example, @samp{//} comments and
1041 the @code{inline} keyword in ISO C99) are not disabled.
1043 @xref{Standards,,Language Standards Supported by GCC}, for details of
1044 these standard versions.
1046 @item -aux-info @var{filename}
1048 Output to the given filename prototyped declarations for all functions
1049 declared and/or defined in a translation unit, including those in header
1050 files. This option is silently ignored in any language other than C@.
1052 Besides declarations, the file indicates, in comments, the origin of
1053 each declaration (source file and line), whether the declaration was
1054 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1055 @samp{O} for old, respectively, in the first character after the line
1056 number and the colon), and whether it came from a declaration or a
1057 definition (@samp{C} or @samp{F}, respectively, in the following
1058 character). In the case of function definitions, a K&R-style list of
1059 arguments followed by their declarations is also provided, inside
1060 comments, after the declaration.
1064 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1065 keyword, so that code can use these words as identifiers. You can use
1066 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1067 instead. @option{-ansi} implies @option{-fno-asm}.
1069 In C++, this switch only affects the @code{typeof} keyword, since
1070 @code{asm} and @code{inline} are standard keywords. You may want to
1071 use the @option{-fno-gnu-keywords} flag instead, which has the same
1072 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1073 switch only affects the @code{asm} and @code{typeof} keywords, since
1074 @code{inline} is a standard keyword in ISO C99.
1077 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1078 @opindex fno-builtin
1079 @cindex built-in functions
1080 Don't recognize built-in functions that do not begin with
1081 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1082 functions provided by GCC}, for details of the functions affected,
1083 including those which are not built-in functions when @option{-ansi} or
1084 @option{-std} options for strict ISO C conformance are used because they
1085 do not have an ISO standard meaning.
1087 GCC normally generates special code to handle certain built-in functions
1088 more efficiently; for instance, calls to @code{alloca} may become single
1089 instructions that adjust the stack directly, and calls to @code{memcpy}
1090 may become inline copy loops. The resulting code is often both smaller
1091 and faster, but since the function calls no longer appear as such, you
1092 cannot set a breakpoint on those calls, nor can you change the behavior
1093 of the functions by linking with a different library.
1095 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1096 option has no effect. Therefore, in C++, the only way to get the
1097 optimization benefits of built-in functions is to call the function
1098 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1099 built-in functions to implement many functions (like
1100 @code{std::strchr}), so that you automatically get efficient code.
1102 With the @option{-fno-builtin-@var{function}} option, not available
1103 when compiling C++, only the built-in function @var{function} is
1104 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1105 function is named this is not built-in in this version of GCC, this
1106 option is ignored. There is no corresponding
1107 @option{-fbuiltin-@var{function}} option; if you wish to enable
1108 built-in functions selectively when using @option{-fno-builtin} or
1109 @option{-ffreestanding}, you may define macros such as:
1112 #define abs(n) __builtin_abs ((n))
1113 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1118 @cindex hosted environment
1120 Assert that compilation takes place in a hosted environment. This implies
1121 @option{-fbuiltin}. A hosted environment is one in which the
1122 entire standard library is available, and in which @code{main} has a return
1123 type of @code{int}. Examples are nearly everything except a kernel.
1124 This is equivalent to @option{-fno-freestanding}.
1126 @item -ffreestanding
1127 @opindex ffreestanding
1128 @cindex hosted environment
1130 Assert that compilation takes place in a freestanding environment. This
1131 implies @option{-fno-builtin}. A freestanding environment
1132 is one in which the standard library may not exist, and program startup may
1133 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1134 This is equivalent to @option{-fno-hosted}.
1136 @xref{Standards,,Language Standards Supported by GCC}, for details of
1137 freestanding and hosted environments.
1141 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1142 options for strict ISO C conformance) implies @option{-trigraphs}.
1144 @cindex traditional C language
1145 @cindex C language, traditional
1147 @opindex traditional
1148 Attempt to support some aspects of traditional C compilers.
1153 All @code{extern} declarations take effect globally even if they
1154 are written inside of a function definition. This includes implicit
1155 declarations of functions.
1158 The newer keywords @code{typeof}, @code{inline}, @code{signed}, @code{const}
1159 and @code{volatile} are not recognized. (You can still use the
1160 alternative keywords such as @code{__typeof__}, @code{__inline__}, and
1164 Comparisons between pointers and integers are always allowed.
1167 Integer types @code{unsigned short} and @code{unsigned char} promote
1168 to @code{unsigned int}.
1171 Out-of-range floating point literals are not an error.
1174 Certain constructs which ISO regards as a single invalid preprocessing
1175 number, such as @samp{0xe-0xd}, are treated as expressions instead.
1178 String ``constants'' are not necessarily constant; they are stored in
1179 writable space, and identical looking constants are allocated
1180 separately. (This is the same as the effect of
1181 @option{-fwritable-strings}.)
1183 @cindex @code{longjmp} and automatic variables
1185 All automatic variables not declared @code{register} are preserved by
1186 @code{longjmp}. Ordinarily, GNU C follows ISO C: automatic variables
1187 not declared @code{volatile} may be clobbered.
1192 @cindex escape sequences, traditional
1193 The character escape sequences @samp{\x} and @samp{\a} evaluate as the
1194 literal characters @samp{x} and @samp{a} respectively. Without
1195 @w{@option{-traditional}}, @samp{\x} is a prefix for the hexadecimal
1196 representation of a character, and @samp{\a} produces a bell.
1199 This option is deprecated and may be removed.
1201 You may wish to use @option{-fno-builtin} as well as @option{-traditional}
1202 if your program uses names that are normally GNU C built-in functions for
1203 other purposes of its own.
1205 You cannot use @option{-traditional} if you include any header files that
1206 rely on ISO C features. Some vendors are starting to ship systems with
1207 ISO C header files and you cannot use @option{-traditional} on such
1208 systems to compile files that include any system headers.
1210 The @option{-traditional} option also enables @option{-traditional-cpp}.
1212 @item -traditional-cpp
1213 @opindex traditional-cpp
1214 Attempt to support some aspects of traditional C preprocessors.
1215 See the GNU CPP manual for details.
1217 @item -fcond-mismatch
1218 @opindex fcond-mismatch
1219 Allow conditional expressions with mismatched types in the second and
1220 third arguments. The value of such an expression is void. This option
1221 is not supported for C++.
1223 @item -funsigned-char
1224 @opindex funsigned-char
1225 Let the type @code{char} be unsigned, like @code{unsigned char}.
1227 Each kind of machine has a default for what @code{char} should
1228 be. It is either like @code{unsigned char} by default or like
1229 @code{signed char} by default.
1231 Ideally, a portable program should always use @code{signed char} or
1232 @code{unsigned char} when it depends on the signedness of an object.
1233 But many programs have been written to use plain @code{char} and
1234 expect it to be signed, or expect it to be unsigned, depending on the
1235 machines they were written for. This option, and its inverse, let you
1236 make such a program work with the opposite default.
1238 The type @code{char} is always a distinct type from each of
1239 @code{signed char} or @code{unsigned char}, even though its behavior
1240 is always just like one of those two.
1243 @opindex fsigned-char
1244 Let the type @code{char} be signed, like @code{signed char}.
1246 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1247 the negative form of @option{-funsigned-char}. Likewise, the option
1248 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1250 @item -fsigned-bitfields
1251 @itemx -funsigned-bitfields
1252 @itemx -fno-signed-bitfields
1253 @itemx -fno-unsigned-bitfields
1254 @opindex fsigned-bitfields
1255 @opindex funsigned-bitfields
1256 @opindex fno-signed-bitfields
1257 @opindex fno-unsigned-bitfields
1258 These options control whether a bit-field is signed or unsigned, when the
1259 declaration does not use either @code{signed} or @code{unsigned}. By
1260 default, such a bit-field is signed, because this is consistent: the
1261 basic integer types such as @code{int} are signed types.
1263 However, when @option{-traditional} is used, bit-fields are all unsigned
1266 @item -fwritable-strings
1267 @opindex fwritable-strings
1268 Store string constants in the writable data segment and don't uniquize
1269 them. This is for compatibility with old programs which assume they can
1270 write into string constants. The option @option{-traditional} also has
1273 Writing into string constants is a very bad idea; ``constants'' should
1276 @item -fallow-single-precision
1277 @opindex fallow-single-precision
1278 Do not promote single precision math operations to double precision,
1279 even when compiling with @option{-traditional}.
1281 Traditional K&R C promotes all floating point operations to double
1282 precision, regardless of the sizes of the operands. On the
1283 architecture for which you are compiling, single precision may be faster
1284 than double precision. If you must use @option{-traditional}, but want
1285 to use single precision operations when the operands are single
1286 precision, use this option. This option has no effect when compiling
1287 with ISO or GNU C conventions (the default).
1290 @opindex fshort-wchar
1291 Override the underlying type for @samp{wchar_t} to be @samp{short
1292 unsigned int} instead of the default for the target. This option is
1293 useful for building programs to run under WINE@.
1296 @node C++ Dialect Options
1297 @section Options Controlling C++ Dialect
1299 @cindex compiler options, C++
1300 @cindex C++ options, command line
1301 @cindex options, C++
1302 This section describes the command-line options that are only meaningful
1303 for C++ programs; but you can also use most of the GNU compiler options
1304 regardless of what language your program is in. For example, you
1305 might compile a file @code{firstClass.C} like this:
1308 g++ -g -frepo -O -c firstClass.C
1312 In this example, only @option{-frepo} is an option meant
1313 only for C++ programs; you can use the other options with any
1314 language supported by GCC@.
1316 Here is a list of options that are @emph{only} for compiling C++ programs:
1319 @item -fno-access-control
1320 @opindex fno-access-control
1321 Turn off all access checking. This switch is mainly useful for working
1322 around bugs in the access control code.
1326 Check that the pointer returned by @code{operator new} is non-null
1327 before attempting to modify the storage allocated. The current Working
1328 Paper requires that @code{operator new} never return a null pointer, so
1329 this check is normally unnecessary.
1331 An alternative to using this option is to specify that your
1332 @code{operator new} does not throw any exceptions; if you declare it
1333 @samp{throw()}, G++ will check the return value. See also @samp{new
1336 @item -fconserve-space
1337 @opindex fconserve-space
1338 Put uninitialized or runtime-initialized global variables into the
1339 common segment, as C does. This saves space in the executable at the
1340 cost of not diagnosing duplicate definitions. If you compile with this
1341 flag and your program mysteriously crashes after @code{main()} has
1342 completed, you may have an object that is being destroyed twice because
1343 two definitions were merged.
1345 This option is no longer useful on most targets, now that support has
1346 been added for putting variables into BSS without making them common.
1348 @item -fno-const-strings
1349 @opindex fno-const-strings
1350 Give string constants type @code{char *} instead of type @code{const
1351 char *}. By default, G++ uses type @code{const char *} as required by
1352 the standard. Even if you use @option{-fno-const-strings}, you cannot
1353 actually modify the value of a string constant, unless you also use
1354 @option{-fwritable-strings}.
1356 This option might be removed in a future release of G++. For maximum
1357 portability, you should structure your code so that it works with
1358 string constants that have type @code{const char *}.
1360 @item -fdollars-in-identifiers
1361 @opindex fdollars-in-identifiers
1362 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1363 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1364 @samp{$} by default on most target systems, but there are a few exceptions.)
1365 Traditional C allowed the character @samp{$} to form part of
1366 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1368 @item -fno-elide-constructors
1369 @opindex fno-elide-constructors
1370 The C++ standard allows an implementation to omit creating a temporary
1371 which is only used to initialize another object of the same type.
1372 Specifying this option disables that optimization, and forces G++ to
1373 call the copy constructor in all cases.
1375 @item -fno-enforce-eh-specs
1376 @opindex fno-enforce-eh-specs
1377 Don't check for violation of exception specifications at runtime. This
1378 option violates the C++ standard, but may be useful for reducing code
1379 size in production builds, much like defining @samp{NDEBUG}. The compiler
1380 will still optimize based on the exception specifications.
1382 @item -fexternal-templates
1383 @opindex fexternal-templates
1385 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1386 template instantiation; template instances are emitted or not according
1387 to the location of the template definition. @xref{Template
1388 Instantiation}, for more information.
1390 This option is deprecated.
1392 @item -falt-external-templates
1393 @opindex falt-external-templates
1394 Similar to @option{-fexternal-templates}, but template instances are
1395 emitted or not according to the place where they are first instantiated.
1396 @xref{Template Instantiation}, for more information.
1398 This option is deprecated.
1401 @itemx -fno-for-scope
1403 @opindex fno-for-scope
1404 If @option{-ffor-scope} is specified, the scope of variables declared in
1405 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1406 as specified by the C++ standard.
1407 If @option{-fno-for-scope} is specified, the scope of variables declared in
1408 a @i{for-init-statement} extends to the end of the enclosing scope,
1409 as was the case in old versions of G++, and other (traditional)
1410 implementations of C++.
1412 The default if neither flag is given to follow the standard,
1413 but to allow and give a warning for old-style code that would
1414 otherwise be invalid, or have different behavior.
1416 @item -fno-gnu-keywords
1417 @opindex fno-gnu-keywords
1418 Do not recognize @code{typeof} as a keyword, so that code can use this
1419 word as an identifier. You can use the keyword @code{__typeof__} instead.
1420 @option{-ansi} implies @option{-fno-gnu-keywords}.
1422 @item -fno-implicit-templates
1423 @opindex fno-implicit-templates
1424 Never emit code for non-inline templates which are instantiated
1425 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1426 @xref{Template Instantiation}, for more information.
1428 @item -fno-implicit-inline-templates
1429 @opindex fno-implicit-inline-templates
1430 Don't emit code for implicit instantiations of inline templates, either.
1431 The default is to handle inlines differently so that compiles with and
1432 without optimization will need the same set of explicit instantiations.
1434 @item -fno-implement-inlines
1435 @opindex fno-implement-inlines
1436 To save space, do not emit out-of-line copies of inline functions
1437 controlled by @samp{#pragma implementation}. This will cause linker
1438 errors if these functions are not inlined everywhere they are called.
1440 @item -fms-extensions
1441 @opindex fms-extensions
1442 Disable pedantic warnings about constructs used in MFC, such as implicit
1443 int and getting a pointer to member function via non-standard syntax.
1445 @item -fno-nonansi-builtins
1446 @opindex fno-nonansi-builtins
1447 Disable built-in declarations of functions that are not mandated by
1448 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1449 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1451 @item -fno-operator-names
1452 @opindex fno-operator-names
1453 Do not treat the operator name keywords @code{and}, @code{bitand},
1454 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1455 synonyms as keywords.
1457 @item -fno-optional-diags
1458 @opindex fno-optional-diags
1459 Disable diagnostics that the standard says a compiler does not need to
1460 issue. Currently, the only such diagnostic issued by G++ is the one for
1461 a name having multiple meanings within a class.
1464 @opindex fpermissive
1465 Downgrade messages about nonconformant code from errors to warnings. By
1466 default, G++ effectively sets @option{-pedantic-errors} without
1467 @option{-pedantic}; this option reverses that. This behavior and this
1468 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1472 Enable automatic template instantiation at link time. This option also
1473 implies @option{-fno-implicit-templates}. @xref{Template
1474 Instantiation}, for more information.
1478 Disable generation of information about every class with virtual
1479 functions for use by the C++ runtime type identification features
1480 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1481 of the language, you can save some space by using this flag. Note that
1482 exception handling uses the same information, but it will generate it as
1487 Emit statistics about front-end processing at the end of the compilation.
1488 This information is generally only useful to the G++ development team.
1490 @item -ftemplate-depth-@var{n}
1491 @opindex ftemplate-depth
1492 Set the maximum instantiation depth for template classes to @var{n}.
1493 A limit on the template instantiation depth is needed to detect
1494 endless recursions during template class instantiation. ANSI/ISO C++
1495 conforming programs must not rely on a maximum depth greater than 17.
1497 @item -fuse-cxa-atexit
1498 @opindex fuse-cxa-atexit
1499 Register destructors for objects with static storage duration with the
1500 @code{__cxa_atexit} function rather than the @code{atexit} function.
1501 This option is required for fully standards-compliant handling of static
1502 destructors, but will only work if your C library supports
1503 @code{__cxa_atexit}.
1507 Emit special relocations for vtables and virtual function references
1508 so that the linker can identify unused virtual functions and zero out
1509 vtable slots that refer to them. This is most useful with
1510 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1511 also discard the functions themselves.
1513 This optimization requires GNU as and GNU ld. Not all systems support
1514 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1518 Do not use weak symbol support, even if it is provided by the linker.
1519 By default, G++ will use weak symbols if they are available. This
1520 option exists only for testing, and should not be used by end-users;
1521 it will result in inferior code and has no benefits. This option may
1522 be removed in a future release of G++.
1526 Do not search for header files in the standard directories specific to
1527 C++, but do still search the other standard directories. (This option
1528 is used when building the C++ library.)
1531 In addition, these optimization, warning, and code generation options
1532 have meanings only for C++ programs:
1535 @item -fno-default-inline
1536 @opindex fno-default-inline
1537 Do not assume @samp{inline} for functions defined inside a class scope.
1538 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1539 functions will have linkage like inline functions; they just won't be
1542 @item -Wctor-dtor-privacy @r{(C++ only)}
1543 @opindex Wctor-dtor-privacy
1544 Warn when a class seems unusable, because all the constructors or
1545 destructors in a class are private and the class has no friends or
1546 public static member functions.
1548 @item -Wnon-virtual-dtor @r{(C++ only)}
1549 @opindex Wnon-virtual-dtor
1550 Warn when a class declares a non-virtual destructor that should probably
1551 be virtual, because it looks like the class will be used polymorphically.
1553 @item -Wreorder @r{(C++ only)}
1555 @cindex reordering, warning
1556 @cindex warning for reordering of member initializers
1557 Warn when the order of member initializers given in the code does not
1558 match the order in which they must be executed. For instance:
1564 A(): j (0), i (1) @{ @}
1568 Here the compiler will warn that the member initializers for @samp{i}
1569 and @samp{j} will be rearranged to match the declaration order of the
1573 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1576 @item -Weffc++ @r{(C++ only)}
1578 Warn about violations of various style guidelines from Scott Meyers'
1579 @cite{Effective C++} books. If you use this option, you should be aware
1580 that the standard library headers do not obey all of these guidelines;
1581 you can use @samp{grep -v} to filter out those warnings.
1583 @item -Wno-deprecated @r{(C++ only)}
1584 @opindex Wno-deprecated
1585 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1587 @item -Wno-non-template-friend @r{(C++ only)}
1588 @opindex Wno-non-template-friend
1589 Disable warnings when non-templatized friend functions are declared
1590 within a template. With the advent of explicit template specification
1591 support in G++, if the name of the friend is an unqualified-id (i.e.,
1592 @samp{friend foo(int)}), the C++ language specification demands that the
1593 friend declare or define an ordinary, nontemplate function. (Section
1594 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1595 could be interpreted as a particular specialization of a templatized
1596 function. Because this non-conforming behavior is no longer the default
1597 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1598 check existing code for potential trouble spots, and is on by default.
1599 This new compiler behavior can be turned off with
1600 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1601 but disables the helpful warning.
1603 @item -Wold-style-cast @r{(C++ only)}
1604 @opindex Wold-style-cast
1605 Warn if an old-style (C-style) cast to a non-void type is used within
1606 a C++ program. The new-style casts (@samp{static_cast},
1607 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1608 unintended effects, and much easier to grep for.
1610 @item -Woverloaded-virtual @r{(C++ only)}
1611 @opindex Woverloaded-virtual
1612 @cindex overloaded virtual fn, warning
1613 @cindex warning for overloaded virtual fn
1614 Warn when a function declaration hides virtual functions from a
1615 base class. For example, in:
1622 struct B: public A @{
1627 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1635 will fail to compile.
1637 @item -Wno-pmf-conversions @r{(C++ only)}
1638 @opindex Wno-pmf-conversions
1639 Disable the diagnostic for converting a bound pointer to member function
1642 @item -Wsign-promo @r{(C++ only)}
1643 @opindex Wsign-promo
1644 Warn when overload resolution chooses a promotion from unsigned or
1645 enumeral type to a signed type over a conversion to an unsigned type of
1646 the same size. Previous versions of G++ would try to preserve
1647 unsignedness, but the standard mandates the current behavior.
1649 @item -Wsynth @r{(C++ only)}
1651 @cindex warning for synthesized methods
1652 @cindex synthesized methods, warning
1653 Warn when G++'s synthesis behavior does not match that of cfront. For
1659 A& operator = (int);
1669 In this example, G++ will synthesize a default @samp{A& operator =
1670 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1673 @node Objective-C Dialect Options
1674 @section Options Controlling Objective-C Dialect
1676 @cindex compiler options, Objective-C
1677 @cindex Objective-C options, command line
1678 @cindex options, Objective-C
1679 This section describes the command-line options that are only meaningful
1680 for Objective-C programs; but you can also use most of the GNU compiler
1681 options regardless of what language your program is in. For example,
1682 you might compile a file @code{some_class.m} like this:
1685 gcc -g -fgnu-runtime -O -c some_class.m
1689 In this example, only @option{-fgnu-runtime} is an option meant only for
1690 Objective-C programs; you can use the other options with any language
1693 Here is a list of options that are @emph{only} for compiling Objective-C
1697 @item -fconstant-string-class=@var{class-name}
1698 @opindex fconstant-string-class
1699 Use @var{class-name} as the name of the class to instantiate for each
1700 literal string specified with the syntax @code{@@"@dots{}"}. The default
1701 class name is @code{NXConstantString}.
1704 @opindex fgnu-runtime
1705 Generate object code compatible with the standard GNU Objective-C
1706 runtime. This is the default for most types of systems.
1708 @item -fnext-runtime
1709 @opindex fnext-runtime
1710 Generate output compatible with the NeXT runtime. This is the default
1711 for NeXT-based systems, including Darwin and Mac OS X@.
1715 Dump interface declarations for all classes seen in the source file to a
1716 file named @file{@var{sourcename}.decl}.
1719 @opindex Wno-protocol
1720 Do not warn if methods required by a protocol are not implemented
1721 in the class adopting it.
1725 Warn if a selector has multiple methods of different types defined.
1727 @c not documented because only avail via -Wp
1728 @c @item -print-objc-runtime-info
1732 @node Language Independent Options
1733 @section Options to Control Diagnostic Messages Formatting
1734 @cindex options to control diagnostics formatting
1735 @cindex diagnostic messages
1736 @cindex message formatting
1738 Traditionally, diagnostic messages have been formatted irrespective of
1739 the output device's aspect (e.g.@: its width, @dots{}). The options described
1740 below can be used to control the diagnostic messages formatting
1741 algorithm, e.g.@: how many characters per line, how often source location
1742 information should be reported. Right now, only the C++ front end can
1743 honor these options. However it is expected, in the near future, that
1744 the remaining front ends would be able to digest them correctly.
1747 @item -fmessage-length=@var{n}
1748 @opindex fmessage-length
1749 Try to format error messages so that they fit on lines of about @var{n}
1750 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1751 the front ends supported by GCC@. If @var{n} is zero, then no
1752 line-wrapping will be done; each error message will appear on a single
1755 @opindex fdiagnostics-show-location
1756 @item -fdiagnostics-show-location=once
1757 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1758 reporter to emit @emph{once} source location information; that is, in
1759 case the message is too long to fit on a single physical line and has to
1760 be wrapped, the source location won't be emitted (as prefix) again,
1761 over and over, in subsequent continuation lines. This is the default
1764 @item -fdiagnostics-show-location=every-line
1765 Only meaningful in line-wrapping mode. Instructs the diagnostic
1766 messages reporter to emit the same source location information (as
1767 prefix) for physical lines that result from the process of breaking
1768 a message which is too long to fit on a single line.
1772 @node Warning Options
1773 @section Options to Request or Suppress Warnings
1774 @cindex options to control warnings
1775 @cindex warning messages
1776 @cindex messages, warning
1777 @cindex suppressing warnings
1779 Warnings are diagnostic messages that report constructions which
1780 are not inherently erroneous but which are risky or suggest there
1781 may have been an error.
1783 You can request many specific warnings with options beginning @samp{-W},
1784 for example @option{-Wimplicit} to request warnings on implicit
1785 declarations. Each of these specific warning options also has a
1786 negative form beginning @samp{-Wno-} to turn off warnings;
1787 for example, @option{-Wno-implicit}. This manual lists only one of the
1788 two forms, whichever is not the default.
1790 These options control the amount and kinds of warnings produced by GCC:
1793 @cindex syntax checking
1795 @opindex fsyntax-only
1796 Check the code for syntax errors, but don't do anything beyond that.
1800 Issue all the warnings demanded by strict ISO C and ISO C++;
1801 reject all programs that use forbidden extensions, and some other
1802 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1803 version of the ISO C standard specified by any @option{-std} option used.
1805 Valid ISO C and ISO C++ programs should compile properly with or without
1806 this option (though a rare few will require @option{-ansi} or a
1807 @option{-std} option specifying the required version of ISO C)@. However,
1808 without this option, certain GNU extensions and traditional C and C++
1809 features are supported as well. With this option, they are rejected.
1811 @option{-pedantic} does not cause warning messages for use of the
1812 alternate keywords whose names begin and end with @samp{__}. Pedantic
1813 warnings are also disabled in the expression that follows
1814 @code{__extension__}. However, only system header files should use
1815 these escape routes; application programs should avoid them.
1816 @xref{Alternate Keywords}.
1818 Some users try to use @option{-pedantic} to check programs for strict ISO
1819 C conformance. They soon find that it does not do quite what they want:
1820 it finds some non-ISO practices, but not all---only those for which
1821 ISO C @emph{requires} a diagnostic, and some others for which
1822 diagnostics have been added.
1824 A feature to report any failure to conform to ISO C might be useful in
1825 some instances, but would require considerable additional work and would
1826 be quite different from @option{-pedantic}. We don't have plans to
1827 support such a feature in the near future.
1829 Where the standard specified with @option{-std} represents a GNU
1830 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1831 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1832 extended dialect is based. Warnings from @option{-pedantic} are given
1833 where they are required by the base standard. (It would not make sense
1834 for such warnings to be given only for features not in the specified GNU
1835 C dialect, since by definition the GNU dialects of C include all
1836 features the compiler supports with the given option, and there would be
1837 nothing to warn about.)
1839 @item -pedantic-errors
1840 @opindex pedantic-errors
1841 Like @option{-pedantic}, except that errors are produced rather than
1846 Inhibit all warning messages.
1850 Inhibit warning messages about the use of @samp{#import}.
1852 @item -Wchar-subscripts
1853 @opindex Wchar-subscripts
1854 Warn if an array subscript has type @code{char}. This is a common cause
1855 of error, as programmers often forget that this type is signed on some
1860 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1861 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1865 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1866 the arguments supplied have types appropriate to the format string
1867 specified, and that the conversions specified in the format string make
1868 sense. This includes standard functions, and others specified by format
1869 attributes (@pxref{Function Attributes}), in the @code{printf},
1870 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1871 not in the C standard) families.
1873 The formats are checked against the format features supported by GNU
1874 libc version 2.2. These include all ISO C89 and C99 features, as well
1875 as features from the Single Unix Specification and some BSD and GNU
1876 extensions. Other library implementations may not support all these
1877 features; GCC does not support warning about features that go beyond a
1878 particular library's limitations. However, if @option{-pedantic} is used
1879 with @option{-Wformat}, warnings will be given about format features not
1880 in the selected standard version (but not for @code{strfmon} formats,
1881 since those are not in any version of the C standard). @xref{C Dialect
1882 Options,,Options Controlling C Dialect}.
1884 @option{-Wformat} is included in @option{-Wall}. For more control over some
1885 aspects of format checking, the options @option{-Wno-format-y2k},
1886 @option{-Wno-format-extra-args}, @option{-Wformat-nonliteral},
1887 @option{-Wformat-security} and @option{-Wformat=2} are available, but are
1888 not included in @option{-Wall}.
1890 @item -Wno-format-y2k
1891 @opindex Wno-format-y2k
1892 If @option{-Wformat} is specified, do not warn about @code{strftime}
1893 formats which may yield only a two-digit year.
1895 @item -Wno-format-extra-args
1896 @opindex Wno-format-extra-args
1897 If @option{-Wformat} is specified, do not warn about excess arguments to a
1898 @code{printf} or @code{scanf} format function. The C standard specifies
1899 that such arguments are ignored.
1901 Where the unused arguments lie between used arguments that are
1902 specified with @samp{$} operand number specifications, normally
1903 warnings are still given, since the implementation could not know what
1904 type to pass to @code{va_arg} to skip the unused arguments. However,
1905 in the case of @code{scanf} formats, this option will suppress the
1906 warning if the unused arguments are all pointers, since the Single
1907 Unix Specification says that such unused arguments are allowed.
1909 @item -Wformat-nonliteral
1910 @opindex Wformat-nonliteral
1911 If @option{-Wformat} is specified, also warn if the format string is not a
1912 string literal and so cannot be checked, unless the format function
1913 takes its format arguments as a @code{va_list}.
1915 @item -Wformat-security
1916 @opindex Wformat-security
1917 If @option{-Wformat} is specified, also warn about uses of format
1918 functions that represent possible security problems. At present, this
1919 warns about calls to @code{printf} and @code{scanf} functions where the
1920 format string is not a string literal and there are no format arguments,
1921 as in @code{printf (foo);}. This may be a security hole if the format
1922 string came from untrusted input and contains @samp{%n}. (This is
1923 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1924 in future warnings may be added to @option{-Wformat-security} that are not
1925 included in @option{-Wformat-nonliteral}.)
1929 Enable @option{-Wformat} plus format checks not included in
1930 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1931 -Wformat-nonliteral -Wformat-security}.
1933 @item -Wimplicit-int
1934 @opindex Wimplicit-int
1935 Warn when a declaration does not specify a type.
1937 @item -Wimplicit-function-declaration
1938 @itemx -Werror-implicit-function-declaration
1939 @opindex Wimplicit-function-declaration
1940 @opindex Werror-implicit-function-declaration
1941 Give a warning (or error) whenever a function is used before being
1946 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1950 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1951 function with external linkage, returning int, taking either zero
1952 arguments, two, or three arguments of appropriate types.
1954 @item -Wmissing-braces
1955 @opindex Wmissing-braces
1956 Warn if an aggregate or union initializer is not fully bracketed. In
1957 the following example, the initializer for @samp{a} is not fully
1958 bracketed, but that for @samp{b} is fully bracketed.
1961 int a[2][2] = @{ 0, 1, 2, 3 @};
1962 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1966 @opindex Wparentheses
1967 Warn if parentheses are omitted in certain contexts, such
1968 as when there is an assignment in a context where a truth value
1969 is expected, or when operators are nested whose precedence people
1970 often get confused about.
1972 Also warn about constructions where there may be confusion to which
1973 @code{if} statement an @code{else} branch belongs. Here is an example of
1988 In C, every @code{else} branch belongs to the innermost possible @code{if}
1989 statement, which in this example is @code{if (b)}. This is often not
1990 what the programmer expected, as illustrated in the above example by
1991 indentation the programmer chose. When there is the potential for this
1992 confusion, GCC will issue a warning when this flag is specified.
1993 To eliminate the warning, add explicit braces around the innermost
1994 @code{if} statement so there is no way the @code{else} could belong to
1995 the enclosing @code{if}. The resulting code would look like this:
2011 @item -Wsequence-point
2012 @opindex Wsequence-point
2013 Warn about code that may have undefined semantics because of violations
2014 of sequence point rules in the C standard.
2016 The C standard defines the order in which expressions in a C program are
2017 evaluated in terms of @dfn{sequence points}, which represent a partial
2018 ordering between the execution of parts of the program: those executed
2019 before the sequence point, and those executed after it. These occur
2020 after the evaluation of a full expression (one which is not part of a
2021 larger expression), after the evaluation of the first operand of a
2022 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2023 function is called (but after the evaluation of its arguments and the
2024 expression denoting the called function), and in certain other places.
2025 Other than as expressed by the sequence point rules, the order of
2026 evaluation of subexpressions of an expression is not specified. All
2027 these rules describe only a partial order rather than a total order,
2028 since, for example, if two functions are called within one expression
2029 with no sequence point between them, the order in which the functions
2030 are called is not specified. However, the standards committee have
2031 ruled that function calls do not overlap.
2033 It is not specified when between sequence points modifications to the
2034 values of objects take effect. Programs whose behavior depends on this
2035 have undefined behavior; the C standard specifies that ``Between the
2036 previous and next sequence point an object shall have its stored value
2037 modified at most once by the evaluation of an expression. Furthermore,
2038 the prior value shall be read only to determine the value to be
2039 stored.''. If a program breaks these rules, the results on any
2040 particular implementation are entirely unpredictable.
2042 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2043 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2044 diagnosed by this option, and it may give an occasional false positive
2045 result, but in general it has been found fairly effective at detecting
2046 this sort of problem in programs.
2048 The present implementation of this option only works for C programs. A
2049 future implementation may also work for C++ programs.
2051 The C standard is worded confusingly, therefore there is some debate
2052 over the precise meaning of the sequence point rules in subtle cases.
2053 Links to discussions of the problem, including proposed formal
2054 definitions, may be found on our readings page, at
2055 @w{@uref{http://gcc.gnu.org/readings.html}}.
2058 @opindex Wreturn-type
2059 Warn whenever a function is defined with a return-type that defaults to
2060 @code{int}. Also warn about any @code{return} statement with no
2061 return-value in a function whose return-type is not @code{void}.
2063 For C++, a function without return type always produces a diagnostic
2064 message, even when @option{-Wno-return-type} is specified. The only
2065 exceptions are @samp{main} and functions defined in system headers.
2069 Warn whenever a @code{switch} statement has an index of enumeral type
2070 and lacks a @code{case} for one or more of the named codes of that
2071 enumeration. (The presence of a @code{default} label prevents this
2072 warning.) @code{case} labels outside the enumeration range also
2073 provoke warnings when this option is used.
2077 Warn if any trigraphs are encountered that might change the meaning of
2078 the program (trigraphs within comments are not warned about).
2080 @item -Wunused-function
2081 @opindex Wunused-function
2082 Warn whenever a static function is declared but not defined or a
2083 non\-inline static function is unused.
2085 @item -Wunused-label
2086 @opindex Wunused-label
2087 Warn whenever a label is declared but not used.
2089 To suppress this warning use the @samp{unused} attribute
2090 (@pxref{Variable Attributes}).
2092 @item -Wunused-parameter
2093 @opindex Wunused-parameter
2094 Warn whenever a function parameter is unused aside from its declaration.
2096 To suppress this warning use the @samp{unused} attribute
2097 (@pxref{Variable Attributes}).
2099 @item -Wunused-variable
2100 @opindex Wunused-variable
2101 Warn whenever a local variable or non-constant static variable is unused
2102 aside from its declaration
2104 To suppress this warning use the @samp{unused} attribute
2105 (@pxref{Variable Attributes}).
2107 @item -Wunused-value
2108 @opindex Wunused-value
2109 Warn whenever a statement computes a result that is explicitly not used.
2111 To suppress this warning cast the expression to @samp{void}.
2115 All all the above @option{-Wunused} options combined.
2117 In order to get a warning about an unused function parameter, you must
2118 either specify @samp{-W -Wunused} or separately specify
2119 @option{-Wunused-parameter}.
2121 @item -Wuninitialized
2122 @opindex Wuninitialized
2123 Warn if an automatic variable is used without first being initialized or
2124 if a variable may be clobbered by a @code{setjmp} call.
2126 These warnings are possible only in optimizing compilation,
2127 because they require data flow information that is computed only
2128 when optimizing. If you don't specify @option{-O}, you simply won't
2131 These warnings occur only for variables that are candidates for
2132 register allocation. Therefore, they do not occur for a variable that
2133 is declared @code{volatile}, or whose address is taken, or whose size
2134 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2135 structures, unions or arrays, even when they are in registers.
2137 Note that there may be no warning about a variable that is used only
2138 to compute a value that itself is never used, because such
2139 computations may be deleted by data flow analysis before the warnings
2142 These warnings are made optional because GCC is not smart
2143 enough to see all the reasons why the code might be correct
2144 despite appearing to have an error. Here is one example of how
2165 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2166 always initialized, but GCC doesn't know this. Here is
2167 another common case:
2172 if (change_y) save_y = y, y = new_y;
2174 if (change_y) y = save_y;
2179 This has no bug because @code{save_y} is used only if it is set.
2181 @cindex @code{longjmp} warnings
2182 This option also warns when a non-volatile automatic variable might be
2183 changed by a call to @code{longjmp}. These warnings as well are possible
2184 only in optimizing compilation.
2186 The compiler sees only the calls to @code{setjmp}. It cannot know
2187 where @code{longjmp} will be called; in fact, a signal handler could
2188 call it at any point in the code. As a result, you may get a warning
2189 even when there is in fact no problem because @code{longjmp} cannot
2190 in fact be called at the place which would cause a problem.
2192 Some spurious warnings can be avoided if you declare all the functions
2193 you use that never return as @code{noreturn}. @xref{Function
2196 @item -Wreorder @r{(C++ only)}
2198 @cindex reordering, warning
2199 @cindex warning for reordering of member initializers
2200 Warn when the order of member initializers given in the code does not
2201 match the order in which they must be executed. For instance:
2203 @item -Wunknown-pragmas
2204 @opindex Wunknown-pragmas
2205 @cindex warning for unknown pragmas
2206 @cindex unknown pragmas, warning
2207 @cindex pragmas, warning of unknown
2208 Warn when a #pragma directive is encountered which is not understood by
2209 GCC@. If this command line option is used, warnings will even be issued
2210 for unknown pragmas in system header files. This is not the case if
2211 the warnings were only enabled by the @option{-Wall} command line option.
2215 All of the above @samp{-W} options combined. This enables all the
2216 warnings about constructions that some users consider questionable, and
2217 that are easy to avoid (or modify to prevent the warning), even in
2218 conjunction with macros.
2221 @opindex Wno-div-by-zero
2222 @opindex Wdiv-by-zero
2223 Warn about compile-time integer division by zero. This is default. To
2224 inhibit the warning messages, use @option{-Wno-div-by-zero}. Floating
2225 point division by zero is not warned about, as it can be a legitimate
2226 way of obtaining infinities and NaNs.
2229 @opindex Wno-multichar
2231 Warn if a multicharacter constant (@samp{'FOOF'}) is used. This is
2232 default. To inhibit the warning messages, use @option{-Wno-multichar}.
2233 Usually they indicate a typo in the user's code, as they have
2234 implementation-defined values, and should not be used in portable code.
2236 @item -Wsystem-headers
2237 @opindex Wsystem-headers
2238 @cindex warnings from system headers
2239 @cindex system headers, warnings from
2240 Print warning messages for constructs found in system header files.
2241 Warnings from system headers are normally suppressed, on the assumption
2242 that they usually do not indicate real problems and would only make the
2243 compiler output harder to read. Using this command line option tells
2244 GCC to emit warnings from system headers as if they occurred in user
2245 code. However, note that using @option{-Wall} in conjunction with this
2246 option will @emph{not} warn about unknown pragmas in system
2247 headers---for that, @option{-Wunknown-pragmas} must also be used.
2250 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2251 Some of them warn about constructions that users generally do not
2252 consider questionable, but which occasionally you might wish to check
2253 for; others warn about constructions that are necessary or hard to avoid
2254 in some cases, and there is no simple way to modify the code to suppress
2260 Print extra warning messages for these events:
2264 A function can return either with or without a value. (Falling
2265 off the end of the function body is considered returning without
2266 a value.) For example, this function would evoke such a
2280 An expression-statement or the left-hand side of a comma expression
2281 contains no side effects.
2282 To suppress the warning, cast the unused expression to void.
2283 For example, an expression such as @samp{x[i,j]} will cause a warning,
2284 but @samp{x[(void)i,j]} will not.
2287 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2290 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2291 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2292 that of ordinary mathematical notation.
2295 Storage-class specifiers like @code{static} are not the first things in
2296 a declaration. According to the C Standard, this usage is obsolescent.
2299 The return type of a function has a type qualifier such as @code{const}.
2300 Such a type qualifier has no effect, since the value returned by a
2301 function is not an lvalue. (But don't warn about the GNU extension of
2302 @code{volatile void} return types. That extension will be warned about
2303 if @option{-pedantic} is specified.)
2306 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2310 A comparison between signed and unsigned values could produce an
2311 incorrect result when the signed value is converted to unsigned.
2312 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2315 An aggregate has a partly bracketed initializer.
2316 For example, the following code would evoke such a warning,
2317 because braces are missing around the initializer for @code{x.h}:
2320 struct s @{ int f, g; @};
2321 struct t @{ struct s h; int i; @};
2322 struct t x = @{ 1, 2, 3 @};
2326 An aggregate has an initializer which does not initialize all members.
2327 For example, the following code would cause such a warning, because
2328 @code{x.h} would be implicitly initialized to zero:
2331 struct s @{ int f, g, h; @};
2332 struct s x = @{ 3, 4 @};
2337 @opindex Wfloat-equal
2338 Warn if floating point values are used in equality comparisons.
2340 The idea behind this is that sometimes it is convenient (for the
2341 programmer) to consider floating-point values as approximations to
2342 infinitely precise real numbers. If you are doing this, then you need
2343 to compute (by analysing the code, or in some other way) the maximum or
2344 likely maximum error that the computation introduces, and allow for it
2345 when performing comparisons (and when producing output, but that's a
2346 different problem). In particular, instead of testing for equality, you
2347 would check to see whether the two values have ranges that overlap; and
2348 this is done with the relational operators, so equality comparisons are
2351 @item -Wtraditional @r{(C only)}
2352 @opindex Wtraditional
2353 Warn about certain constructs that behave differently in traditional and
2354 ISO C@. Also warn about ISO C constructs that have no traditional C
2355 equivalent, and/or problematic constructs which should be avoided.
2359 Macro parameters that appear within string literals in the macro body.
2360 In traditional C macro replacement takes place within string literals,
2361 but does not in ISO C@.
2364 In traditional C, some preprocessor directives did not exist.
2365 Traditional preprocessors would only consider a line to be a directive
2366 if the @samp{#} appeared in column 1 on the line. Therefore
2367 @option{-Wtraditional} warns about directives that traditional C
2368 understands but would ignore because the @samp{#} does not appear as the
2369 first character on the line. It also suggests you hide directives like
2370 @samp{#pragma} not understood by traditional C by indenting them. Some
2371 traditional implementations would not recognize @samp{#elif}, so it
2372 suggests avoiding it altogether.
2375 A function-like macro that appears without arguments.
2378 The unary plus operator.
2381 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2382 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2383 constants.) Note, these suffixes appear in macros defined in the system
2384 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2385 Use of these macros in user code might normally lead to spurious
2386 warnings, however gcc's integrated preprocessor has enough context to
2387 avoid warning in these cases.
2390 A function declared external in one block and then used after the end of
2394 A @code{switch} statement has an operand of type @code{long}.
2397 A non-@code{static} function declaration follows a @code{static} one.
2398 This construct is not accepted by some traditional C compilers.
2401 The ISO type of an integer constant has a different width or
2402 signedness from its traditional type. This warning is only issued if
2403 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2404 typically represent bit patterns, are not warned about.
2407 Usage of ISO string concatenation is detected.
2410 Initialization of automatic aggregates.
2413 Identifier conflicts with labels. Traditional C lacks a separate
2414 namespace for labels.
2417 Initialization of unions. If the initializer is zero, the warning is
2418 omitted. This is done under the assumption that the zero initializer in
2419 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2420 initializer warnings and relies on default initialization to zero in the
2424 Conversions by prototypes between fixed/floating point values and vice
2425 versa. The absence of these prototypes when compiling with traditional
2426 C would cause serious problems. This is a subset of the possible
2427 conversion warnings, for the full set use @option{-Wconversion}.
2432 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2436 Warn whenever a local variable shadows another local variable, parameter or
2437 global variable or whenever a built-in function is shadowed.
2439 @item -Wlarger-than-@var{len}
2440 @opindex Wlarger-than
2441 Warn whenever an object of larger than @var{len} bytes is defined.
2443 @item -Wpointer-arith
2444 @opindex Wpointer-arith
2445 Warn about anything that depends on the ``size of'' a function type or
2446 of @code{void}. GNU C assigns these types a size of 1, for
2447 convenience in calculations with @code{void *} pointers and pointers
2450 @item -Wbad-function-cast @r{(C only)}
2451 @opindex Wbad-function-cast
2452 Warn whenever a function call is cast to a non-matching type.
2453 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2457 Warn whenever a pointer is cast so as to remove a type qualifier from
2458 the target type. For example, warn if a @code{const char *} is cast
2459 to an ordinary @code{char *}.
2462 @opindex Wcast-align
2463 Warn whenever a pointer is cast such that the required alignment of the
2464 target is increased. For example, warn if a @code{char *} is cast to
2465 an @code{int *} on machines where integers can only be accessed at
2466 two- or four-byte boundaries.
2468 @item -Wwrite-strings
2469 @opindex Wwrite-strings
2470 When compiling C, give string constants the type @code{const
2471 char[@var{length}]} so that
2472 copying the address of one into a non-@code{const} @code{char *}
2473 pointer will get a warning; when compiling C++, warn about the
2474 deprecated conversion from string constants to @code{char *}.
2475 These warnings will help you find at
2476 compile time code that can try to write into a string constant, but
2477 only if you have been very careful about using @code{const} in
2478 declarations and prototypes. Otherwise, it will just be a nuisance;
2479 this is why we did not make @option{-Wall} request these warnings.
2482 @opindex Wconversion
2483 Warn if a prototype causes a type conversion that is different from what
2484 would happen to the same argument in the absence of a prototype. This
2485 includes conversions of fixed point to floating and vice versa, and
2486 conversions changing the width or signedness of a fixed point argument
2487 except when the same as the default promotion.
2489 Also, warn if a negative integer constant expression is implicitly
2490 converted to an unsigned type. For example, warn about the assignment
2491 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2492 casts like @code{(unsigned) -1}.
2494 @item -Wsign-compare
2495 @opindex Wsign-compare
2496 @cindex warning for comparison of signed and unsigned values
2497 @cindex comparison of signed and unsigned values, warning
2498 @cindex signed and unsigned values, comparison warning
2499 Warn when a comparison between signed and unsigned values could produce
2500 an incorrect result when the signed value is converted to unsigned.
2501 This warning is also enabled by @option{-W}; to get the other warnings
2502 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2504 @item -Waggregate-return
2505 @opindex Waggregate-return
2506 Warn if any functions that return structures or unions are defined or
2507 called. (In languages where you can return an array, this also elicits
2510 @item -Wstrict-prototypes @r{(C only)}
2511 @opindex Wstrict-prototypes
2512 Warn if a function is declared or defined without specifying the
2513 argument types. (An old-style function definition is permitted without
2514 a warning if preceded by a declaration which specifies the argument
2517 @item -Wmissing-prototypes @r{(C only)}
2518 @opindex Wmissing-prototypes
2519 Warn if a global function is defined without a previous prototype
2520 declaration. This warning is issued even if the definition itself
2521 provides a prototype. The aim is to detect global functions that fail
2522 to be declared in header files.
2524 @item -Wmissing-declarations
2525 @opindex Wmissing-declarations
2526 Warn if a global function is defined without a previous declaration.
2527 Do so even if the definition itself provides a prototype.
2528 Use this option to detect global functions that are not declared in
2531 @item -Wmissing-noreturn
2532 @opindex Wmissing-noreturn
2533 Warn about functions which might be candidates for attribute @code{noreturn}.
2534 Note these are only possible candidates, not absolute ones. Care should
2535 be taken to manually verify functions actually do not ever return before
2536 adding the @code{noreturn} attribute, otherwise subtle code generation
2537 bugs could be introduced. You will not get a warning for @code{main} in
2538 hosted C environments.
2540 @item -Wmissing-format-attribute
2541 @opindex Wmissing-format-attribute
2543 If @option{-Wformat} is enabled, also warn about functions which might be
2544 candidates for @code{format} attributes. Note these are only possible
2545 candidates, not absolute ones. GCC will guess that @code{format}
2546 attributes might be appropriate for any function that calls a function
2547 like @code{vprintf} or @code{vscanf}, but this might not always be the
2548 case, and some functions for which @code{format} attributes are
2549 appropriate may not be detected. This option has no effect unless
2550 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2552 @item -Wno-deprecated-declarations
2553 @opindex Wno-deprecated-declarations
2554 Do not warn about uses of functions, variables, and types marked as
2555 deprecated by using the @code{deprecated} attribute.
2556 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2557 @pxref{Type Attributes}.)
2561 Warn if a structure is given the packed attribute, but the packed
2562 attribute has no effect on the layout or size of the structure.
2563 Such structures may be mis-aligned for little benefit. For
2564 instance, in this code, the variable @code{f.x} in @code{struct bar}
2565 will be misaligned even though @code{struct bar} does not itself
2566 have the packed attribute:
2573 @} __attribute__((packed));
2583 Warn if padding is included in a structure, either to align an element
2584 of the structure or to align the whole structure. Sometimes when this
2585 happens it is possible to rearrange the fields of the structure to
2586 reduce the padding and so make the structure smaller.
2588 @item -Wredundant-decls
2589 @opindex Wredundant-decls
2590 Warn if anything is declared more than once in the same scope, even in
2591 cases where multiple declaration is valid and changes nothing.
2593 @item -Wnested-externs @r{(C only)}
2594 @opindex Wnested-externs
2595 Warn if an @code{extern} declaration is encountered within a function.
2597 @item -Wunreachable-code
2598 @opindex Wunreachable-code
2599 Warn if the compiler detects that code will never be executed.
2601 This option is intended to warn when the compiler detects that at
2602 least a whole line of source code will never be executed, because
2603 some condition is never satisfied or because it is after a
2604 procedure that never returns.
2606 It is possible for this option to produce a warning even though there
2607 are circumstances under which part of the affected line can be executed,
2608 so care should be taken when removing apparently-unreachable code.
2610 For instance, when a function is inlined, a warning may mean that the
2611 line is unreachable in only one inlined copy of the function.
2613 This option is not made part of @option{-Wall} because in a debugging
2614 version of a program there is often substantial code which checks
2615 correct functioning of the program and is, hopefully, unreachable
2616 because the program does work. Another common use of unreachable
2617 code is to provide behavior which is selectable at compile-time.
2621 Warn if a function can not be inlined and it was declared as inline.
2625 @opindex Wno-long-long
2626 Warn if @samp{long long} type is used. This is default. To inhibit
2627 the warning messages, use @option{-Wno-long-long}. Flags
2628 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2629 only when @option{-pedantic} flag is used.
2631 @item -Wdisabled-optimization
2632 @opindex Wdisabled-optimization
2633 Warn if a requested optimization pass is disabled. This warning does
2634 not generally indicate that there is anything wrong with your code; it
2635 merely indicates that GCC's optimizers were unable to handle the code
2636 effectively. Often, the problem is that your code is too big or too
2637 complex; GCC will refuse to optimize programs when the optimization
2638 itself is likely to take inordinate amounts of time.
2642 Make all warnings into errors.
2645 @node Debugging Options
2646 @section Options for Debugging Your Program or GCC
2647 @cindex options, debugging
2648 @cindex debugging information options
2650 GCC has various special options that are used for debugging
2651 either your program or GCC:
2656 Produce debugging information in the operating system's native format
2657 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2660 On most systems that use stabs format, @option{-g} enables use of extra
2661 debugging information that only GDB can use; this extra information
2662 makes debugging work better in GDB but will probably make other debuggers
2664 refuse to read the program. If you want to control for certain whether
2665 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2666 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2667 or @option{-gvms} (see below).
2669 Unlike most other C compilers, GCC allows you to use @option{-g} with
2670 @option{-O}. The shortcuts taken by optimized code may occasionally
2671 produce surprising results: some variables you declared may not exist
2672 at all; flow of control may briefly move where you did not expect it;
2673 some statements may not be executed because they compute constant
2674 results or their values were already at hand; some statements may
2675 execute in different places because they were moved out of loops.
2677 Nevertheless it proves possible to debug optimized output. This makes
2678 it reasonable to use the optimizer for programs that might have bugs.
2680 The following options are useful when GCC is generated with the
2681 capability for more than one debugging format.
2685 Produce debugging information for use by GDB@. This means to use the
2686 most expressive format available (DWARF 2, stabs, or the native format
2687 if neither of those are supported), including GDB extensions if at all
2692 Produce debugging information in stabs format (if that is supported),
2693 without GDB extensions. This is the format used by DBX on most BSD
2694 systems. On MIPS, Alpha and System V Release 4 systems this option
2695 produces stabs debugging output which is not understood by DBX or SDB@.
2696 On System V Release 4 systems this option requires the GNU assembler.
2700 Produce debugging information in stabs format (if that is supported),
2701 using GNU extensions understood only by the GNU debugger (GDB)@. The
2702 use of these extensions is likely to make other debuggers crash or
2703 refuse to read the program.
2707 Produce debugging information in COFF format (if that is supported).
2708 This is the format used by SDB on most System V systems prior to
2713 Produce debugging information in XCOFF format (if that is supported).
2714 This is the format used by the DBX debugger on IBM RS/6000 systems.
2718 Produce debugging information in XCOFF format (if that is supported),
2719 using GNU extensions understood only by the GNU debugger (GDB)@. The
2720 use of these extensions is likely to make other debuggers crash or
2721 refuse to read the program, and may cause assemblers other than the GNU
2722 assembler (GAS) to fail with an error.
2726 Produce debugging information in DWARF version 1 format (if that is
2727 supported). This is the format used by SDB on most System V Release 4
2732 Produce debugging information in DWARF version 1 format (if that is
2733 supported), using GNU extensions understood only by the GNU debugger
2734 (GDB)@. The use of these extensions is likely to make other debuggers
2735 crash or refuse to read the program.
2739 Produce debugging information in DWARF version 2 format (if that is
2740 supported). This is the format used by DBX on IRIX 6.
2744 Produce debugging information in VMS debug format (if that is
2745 supported). This is the format used by DEBUG on VMS systems.
2748 @itemx -ggdb@var{level}
2749 @itemx -gstabs@var{level}
2750 @itemx -gcoff@var{level}
2751 @itemx -gxcoff@var{level}
2752 @itemx -gdwarf@var{level}
2753 @itemx -gdwarf-2@var{level}
2754 @itemx -gvms@var{level}
2755 Request debugging information and also use @var{level} to specify how
2756 much information. The default level is 2.
2758 Level 1 produces minimal information, enough for making backtraces in
2759 parts of the program that you don't plan to debug. This includes
2760 descriptions of functions and external variables, but no information
2761 about local variables and no line numbers.
2763 Level 3 includes extra information, such as all the macro definitions
2764 present in the program. Some debuggers support macro expansion when
2765 you use @option{-g3}.
2770 Generate extra code to write profile information suitable for the
2771 analysis program @code{prof}. You must use this option when compiling
2772 the source files you want data about, and you must also use it when
2775 @cindex @code{gprof}
2778 Generate extra code to write profile information suitable for the
2779 analysis program @code{gprof}. You must use this option when compiling
2780 the source files you want data about, and you must also use it when
2786 Generate extra code to write profile information for basic blocks, which will
2787 record the number of times each basic block is executed, the basic block start
2788 address, and the function name containing the basic block. If @option{-g} is
2789 used, the line number and filename of the start of the basic block will also be
2790 recorded. If not overridden by the machine description, the default action is
2791 to append to the text file @file{bb.out}.
2793 This data could be analyzed by a program like @code{tcov}. Note,
2794 however, that the format of the data is not what @code{tcov} expects.
2795 Eventually GNU @code{gprof} should be extended to process this data.
2799 Makes the compiler print out each function name as it is compiled, and
2800 print some statistics about each pass when it finishes.
2803 @opindex ftime-report
2804 Makes the compiler print some statistics about the time consumed by each
2805 pass when it finishes.
2808 @opindex fmem-report
2809 Makes the compiler print some statistics about permanent memory
2810 allocation when it finishes.
2812 @item -fprofile-arcs
2813 @opindex fprofile-arcs
2814 Instrument @dfn{arcs} during compilation to generate coverage data
2815 or for profile-directed block ordering. During execution the program
2816 records how many times each branch is executed and how many times it is
2817 taken. When the compiled program exits it saves this data to a file
2818 called @file{@var{sourcename}.da} for each source file.
2820 For profile-directed block ordering, compile the program with
2821 @option{-fprofile-arcs} plus optimization and code generation options,
2822 generate the arc profile information by running the program on a
2823 selected workload, and then compile the program again with the same
2824 optimization and code generation options plus
2825 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2826 Control Optimization}).
2828 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2829 when it is used with the @option{-ftest-coverage} option. GCC
2830 supports two methods of determining code coverage: the options that
2831 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2832 write information to text files. The options that support @code{gcov}
2833 do not need to instrument every arc in the program, so a program compiled
2834 with them runs faster than a program compiled with @option{-a}, which
2835 adds instrumentation code to every basic block in the program. The
2836 tradeoff: since @code{gcov} does not have execution counts for all
2837 branches, it must start with the execution counts for the instrumented
2838 branches, and then iterate over the program flow graph until the entire
2839 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2840 a program which uses information from @option{-a} and @option{-ax}.
2842 With @option{-fprofile-arcs}, for each function of your program GCC
2843 creates a program flow graph, then finds a spanning tree for the graph.
2844 Only arcs that are not on the spanning tree have to be instrumented: the
2845 compiler adds code to count the number of times that these arcs are
2846 executed. When an arc is the only exit or only entrance to a block, the
2847 instrumentation code can be added to the block; otherwise, a new basic
2848 block must be created to hold the instrumentation code.
2850 This option makes it possible to estimate branch probabilities and to
2851 calculate basic block execution counts. In general, basic block
2852 execution counts as provided by @option{-a} do not give enough
2853 information to estimate all branch probabilities.
2856 @item -ftest-coverage
2857 @opindex ftest-coverage
2858 Create data files for the @code{gcov} code-coverage utility
2859 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2860 The data file names begin with the name of your source file:
2863 @item @var{sourcename}.bb
2864 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2865 associate basic block execution counts with line numbers.
2867 @item @var{sourcename}.bbg
2868 A list of all arcs in the program flow graph. This allows @code{gcov}
2869 to reconstruct the program flow graph, so that it can compute all basic
2870 block and arc execution counts from the information in the
2871 @code{@var{sourcename}.da} file.
2874 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2875 option adds instrumentation to the program, which then writes
2876 execution counts to another data file:
2879 @item @var{sourcename}.da
2880 Runtime arc execution counts, used in conjunction with the arc
2881 information in the file @code{@var{sourcename}.bbg}.
2884 Coverage data will map better to the source files if
2885 @option{-ftest-coverage} is used without optimization.
2887 @item -d@var{letters}
2889 Says to make debugging dumps during compilation at times specified by
2890 @var{letters}. This is used for debugging the compiler. The file names
2891 for most of the dumps are made by appending a pass number and a word to
2892 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2893 Here are the possible letters for use in @var{letters}, and their meanings:
2898 Annotate the assembler output with miscellaneous debugging information.
2901 Dump after computing branch probabilities, to @file{@var{file}.14.bp}.
2904 Dump after block reordering, to @file{@var{file}.29.bbro}.
2907 Dump after instruction combination, to the file @file{@var{file}.16.combine}.
2910 Dump after the first if conversion, to the file @file{@var{file}.17.ce}.
2913 Dump after delayed branch scheduling, to @file{@var{file}.31.dbr}.
2916 Dump all macro definitions, at the end of preprocessing, in addition to
2920 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2921 @file{@var{file}.07.ussa}.
2924 Dump after the second if conversion, to @file{@var{file}.26.ce2}.
2927 Dump after life analysis, to @file{@var{file}.15.life}.
2930 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.09.addressof}.
2933 Dump after global register allocation, to @file{@var{file}.21.greg}.
2936 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2939 Dump after reg-to-stack conversion, to @file{@var{file}.28.stack}.
2942 Dump after post-reload optimizations, to @file{@var{file}.22.postreload}.
2945 Dump after GCSE, to @file{@var{file}.10.gcse}.
2948 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2951 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2954 Dump after conversion from registers to stack, to @file{@var{file}.32.stack}.
2957 Dump after local register allocation, to @file{@var{file}.20.lreg}.
2960 Dump after loop optimization, to @file{@var{file}.11.loop}.
2963 Dump after performing the machine dependent reorganisation pass, to
2964 @file{@var{file}.30.mach}.
2967 Dump after register renumbering, to @file{@var{file}.25.rnreg}.
2970 Dump after the register move pass, to @file{@var{file}.18.regmove}.
2973 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2976 Dump after the second scheduling pass, to @file{@var{file}.27.sched2}.
2979 Dump after CSE (including the jump optimization that sometimes follows
2980 CSE), to @file{@var{file}.08.cse}.
2983 Dump after the first scheduling pass, to @file{@var{file}.19.sched}.
2986 Dump after the second CSE pass (including the jump optimization that
2987 sometimes follows CSE), to @file{@var{file}.12.cse2}.
2990 Dump after the second flow pass, to @file{@var{file}.23.flow2}.
2993 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2996 Dump after the peephole pass, to @file{@var{file}.24.peephole2}.
2999 Produce all the dumps listed above.
3002 Print statistics on memory usage, at the end of the run, to
3006 Annotate the assembler output with a comment indicating which
3007 pattern and alternative was used. The length of each instruction is
3011 Dump the RTL in the assembler output as a comment before each instruction.
3012 Also turns on @option{-dp} annotation.
3015 For each of the other indicated dump files (except for
3016 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
3017 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3020 Just generate RTL for a function instead of compiling it. Usually used
3024 Dump debugging information during parsing, to standard error.
3027 @item -fdump-unnumbered
3028 @opindex fdump-unnumbered
3029 When doing debugging dumps (see @option{-d} option above), suppress instruction
3030 numbers and line number note output. This makes it more feasible to
3031 use diff on debugging dumps for compiler invocations with different
3032 options, in particular with and without @option{-g}.
3034 @item -fdump-translation-unit @r{(C and C++ only)}
3035 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3036 @opindex fdump-translation-unit
3037 Dump a representation of the tree structure for the entire translation
3038 unit to a file. The file name is made by appending @file{.tu} to the
3039 source file name. If the @samp{-@var{options}} form is used, @var{options}
3040 controls the details of the dump as described for the
3041 @option{-fdump-tree} options.
3043 @item -fdump-class-hierarchy @r{(C++ only)}
3044 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3045 @opindex fdump-class-hierarchy
3046 Dump a representation of each class's hierarchy and virtual function
3047 table layout to a file. The file name is made by appending @file{.class}
3048 to the source file name. If the @samp{-@var{options}} form is used,
3049 @var{options} controls the details of the dump as described for the
3050 @option{-fdump-tree} options.
3052 @item -fdump-tree-@var{switch} @r{(C++ only)}
3053 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3055 Control the dumping at various stages of processing the intermediate
3056 language tree to a file. The file name is generated by appending a switch
3057 specific suffix to the source file name. If the @samp{-@var{options}}
3058 form is used, @var{options} is a list of @samp{-} separated options that
3059 control the details of the dump. Not all options are applicable to all
3060 dumps, those which are not meaningful will be ignored. The following
3061 options are available
3065 Print the address of each node. Usually this is not meaningful as it
3066 changes according to the environment and source file. Its primary use
3067 is for tying up a dump file with a debug environment.
3069 Inhibit dumping of members of a scope or body of a function merely
3070 because that scope has been reached. Only dump such items when they
3071 are directly reachable by some other path.
3073 Turn on all options.
3076 The following tree dumps are possible:
3079 Dump before any tree based optimization, to @file{@var{file}.original}.
3081 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3083 Dump after function inlining, to @file{@var{file}.inlined}.
3086 @item -fpretend-float
3087 @opindex fpretend-float
3088 When running a cross-compiler, pretend that the target machine uses the
3089 same floating point format as the host machine. This causes incorrect
3090 output of the actual floating constants, but the actual instruction
3091 sequence will probably be the same as GCC would make when running on
3096 Store the usual ``temporary'' intermediate files permanently; place them
3097 in the current directory and name them based on the source file. Thus,
3098 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3099 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3100 preprocessed @file{foo.i} output file even though the compiler now
3101 normally uses an integrated preprocessor.
3105 Report the CPU time taken by each subprocess in the compilation
3106 sequence. For C source files, this is the compiler proper and assembler
3107 (plus the linker if linking is done). The output looks like this:
3114 The first number on each line is the ``user time,'' that is time spent
3115 executing the program itself. The second number is ``system time,''
3116 time spent executing operating system routines on behalf of the program.
3117 Both numbers are in seconds.
3119 @item -print-file-name=@var{library}
3120 @opindex print-file-name
3121 Print the full absolute name of the library file @var{library} that
3122 would be used when linking---and don't do anything else. With this
3123 option, GCC does not compile or link anything; it just prints the
3126 @item -print-multi-directory
3127 @opindex print-multi-directory
3128 Print the directory name corresponding to the multilib selected by any
3129 other switches present in the command line. This directory is supposed
3130 to exist in @env{GCC_EXEC_PREFIX}.
3132 @item -print-multi-lib
3133 @opindex print-multi-lib
3134 Print the mapping from multilib directory names to compiler switches
3135 that enable them. The directory name is separated from the switches by
3136 @samp{;}, and each switch starts with an @samp{@@} instead of the
3137 @samp{-}, without spaces between multiple switches. This is supposed to
3138 ease shell-processing.
3140 @item -print-prog-name=@var{program}
3141 @opindex print-prog-name
3142 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3144 @item -print-libgcc-file-name
3145 @opindex print-libgcc-file-name
3146 Same as @option{-print-file-name=libgcc.a}.
3148 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3149 but you do want to link with @file{libgcc.a}. You can do
3152 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3155 @item -print-search-dirs
3156 @opindex print-search-dirs
3157 Print the name of the configured installation directory and a list of
3158 program and library directories gcc will search---and don't do anything else.
3160 This is useful when gcc prints the error message
3161 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3162 To resolve this you either need to put @file{cpp0} and the other compiler
3163 components where gcc expects to find them, or you can set the environment
3164 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3165 Don't forget the trailing '/'.
3166 @xref{Environment Variables}.
3169 @opindex dumpmachine
3170 Print the compiler's target machine (for example,
3171 @samp{i686-pc-linux-gnu})---and don't do anything else.
3174 @opindex dumpversion
3175 Print the compiler version (for example, @samp{3.0})---and don't do
3180 Print the compiler's built-in specs---and don't do anything else. (This
3181 is used when GCC itself is being built.) @xref{Spec Files}.
3184 @node Optimize Options
3185 @section Options That Control Optimization
3186 @cindex optimize options
3187 @cindex options, optimization
3189 These options control various sorts of optimizations:
3196 Optimize. Optimizing compilation takes somewhat more time, and a lot
3197 more memory for a large function.
3199 Without @option{-O}, the compiler's goal is to reduce the cost of
3200 compilation and to make debugging produce the expected results.
3201 Statements are independent: if you stop the program with a breakpoint
3202 between statements, you can then assign a new value to any variable or
3203 change the program counter to any other statement in the function and
3204 get exactly the results you would expect from the source code.
3206 With @option{-O}, the compiler tries to reduce code size and execution
3207 time, without performing any optimizations that take a great deal of
3212 Optimize even more. GCC performs nearly all supported optimizations
3213 that do not involve a space-speed tradeoff. The compiler does not
3214 perform loop unrolling or function inlining when you specify @option{-O2}.
3215 As compared to @option{-O}, this option increases both compilation time
3216 and the performance of the generated code.
3218 @option{-O2} turns on all optional optimizations except for loop unrolling,
3219 function inlining, and register renaming. It also turns on the
3220 @option{-fforce-mem} option on all machines and frame pointer elimination
3221 on machines where doing so does not interfere with debugging.
3223 Please note the warning under @option{-fgcse} about
3224 invoking @option{-O2} on programs that use computed gotos.
3228 Optimize yet more. @option{-O3} turns on all optimizations specified by
3229 @option{-O2} and also turns on the @option{-finline-functions} and
3230 @option{-frename-registers} options.
3238 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3239 do not typically increase code size. It also performs further
3240 optimizations designed to reduce code size.
3242 If you use multiple @option{-O} options, with or without level numbers,
3243 the last such option is the one that is effective.
3246 Options of the form @option{-f@var{flag}} specify machine-independent
3247 flags. Most flags have both positive and negative forms; the negative
3248 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3249 only one of the forms is listed---the one which is not the default.
3250 You can figure out the other form by either removing @samp{no-} or
3255 @opindex ffloat-store
3256 Do not store floating point variables in registers, and inhibit other
3257 options that might change whether a floating point value is taken from a
3260 @cindex floating point precision
3261 This option prevents undesirable excess precision on machines such as
3262 the 68000 where the floating registers (of the 68881) keep more
3263 precision than a @code{double} is supposed to have. Similarly for the
3264 x86 architecture. For most programs, the excess precision does only
3265 good, but a few programs rely on the precise definition of IEEE floating
3266 point. Use @option{-ffloat-store} for such programs, after modifying
3267 them to store all pertinent intermediate computations into variables.
3269 @item -fno-default-inline
3270 @opindex fno-default-inline
3271 Do not make member functions inline by default merely because they are
3272 defined inside the class scope (C++ only). Otherwise, when you specify
3273 @w{@option{-O}}, member functions defined inside class scope are compiled
3274 inline by default; i.e., you don't need to add @samp{inline} in front of
3275 the member function name.
3277 @item -fno-defer-pop
3278 @opindex fno-defer-pop
3279 Always pop the arguments to each function call as soon as that function
3280 returns. For machines which must pop arguments after a function call,
3281 the compiler normally lets arguments accumulate on the stack for several
3282 function calls and pops them all at once.
3286 Force memory operands to be copied into registers before doing
3287 arithmetic on them. This produces better code by making all memory
3288 references potential common subexpressions. When they are not common
3289 subexpressions, instruction combination should eliminate the separate
3290 register-load. The @option{-O2} option turns on this option.
3293 @opindex fforce-addr
3294 Force memory address constants to be copied into registers before
3295 doing arithmetic on them. This may produce better code just as
3296 @option{-fforce-mem} may.
3298 @item -fomit-frame-pointer
3299 @opindex fomit-frame-pointer
3300 Don't keep the frame pointer in a register for functions that
3301 don't need one. This avoids the instructions to save, set up and
3302 restore frame pointers; it also makes an extra register available
3303 in many functions. @strong{It also makes debugging impossible on
3306 On some machines, such as the VAX, this flag has no effect, because
3307 the standard calling sequence automatically handles the frame pointer
3308 and nothing is saved by pretending it doesn't exist. The
3309 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3310 whether a target machine supports this flag. @xref{Registers,,Register
3311 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3313 @item -foptimize-sibling-calls
3314 @opindex foptimize-sibling-calls
3315 Optimize sibling and tail recursive calls.
3319 This option generates traps for signed overflow on addition, subtraction,
3320 multiplication operations.
3324 Don't pay attention to the @code{inline} keyword. Normally this option
3325 is used to keep the compiler from expanding any functions inline.
3326 Note that if you are not optimizing, no functions can be expanded inline.
3328 @item -finline-functions
3329 @opindex finline-functions
3330 Integrate all simple functions into their callers. The compiler
3331 heuristically decides which functions are simple enough to be worth
3332 integrating in this way.
3334 If all calls to a given function are integrated, and the function is
3335 declared @code{static}, then the function is normally not output as
3336 assembler code in its own right.
3338 @item -finline-limit=@var{n}
3339 @opindex finline-limit
3340 By default, gcc limits the size of functions that can be inlined. This flag
3341 allows the control of this limit for functions that are explicitly marked as
3342 inline (ie marked with the inline keyword or defined within the class
3343 definition in c++). @var{n} is the size of functions that can be inlined in
3344 number of pseudo instructions (not counting parameter handling). The default
3345 value of @var{n} is 600.
3346 Increasing this value can result in more inlined code at
3347 the cost of compilation time and memory consumption. Decreasing usually makes
3348 the compilation faster and less code will be inlined (which presumably
3349 means slower programs). This option is particularly useful for programs that
3350 use inlining heavily such as those based on recursive templates with C++.
3352 @emph{Note:} pseudo instruction represents, in this particular context, an
3353 abstract measurement of function's size. In no way, it represents a count
3354 of assembly instructions and as such its exact meaning might change from one
3355 release to an another.
3357 @item -fkeep-inline-functions
3358 @opindex fkeep-inline-functions
3359 Even if all calls to a given function are integrated, and the function
3360 is declared @code{static}, nevertheless output a separate run-time
3361 callable version of the function. This switch does not affect
3362 @code{extern inline} functions.
3364 @item -fkeep-static-consts
3365 @opindex fkeep-static-consts
3366 Emit variables declared @code{static const} when optimization isn't turned
3367 on, even if the variables aren't referenced.
3369 GCC enables this option by default. If you want to force the compiler to
3370 check if the variable was referenced, regardless of whether or not
3371 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3373 @item -fmerge-constants
3374 Attempt to merge identical constants (string constants and floating point
3375 constants) accross compilation units.
3377 This option is default for optimized compilation if assembler and linker
3378 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3380 @item -fmerge-all-constants
3381 Attempt to merge identical constants and identical variables.
3383 This option implies @option{-fmerge-constants}. In addition to
3384 @option{-fmerge-constants} this considers e.g. even constant initialized
3385 arrays or initialized constant variables with integral or floating point
3386 types. Languages like C or C++ require each non-automatic variable to
3387 have distinct location, so using this option will result in non-conforming
3390 @item -fno-function-cse
3391 @opindex fno-function-cse
3392 Do not put function addresses in registers; make each instruction that
3393 calls a constant function contain the function's address explicitly.
3395 This option results in less efficient code, but some strange hacks
3396 that alter the assembler output may be confused by the optimizations
3397 performed when this option is not used.
3401 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3402 @option{-fno-trapping-math}.
3404 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3406 This option should never be turned on by any @option{-O} option since
3407 it can result in incorrect output for programs which depend on
3408 an exact implementation of IEEE or ISO rules/specifications for
3411 @item -fno-math-errno
3412 @opindex fno-math-errno
3413 Do not set ERRNO after calling math functions that are executed
3414 with a single instruction, e.g., sqrt. A program that relies on
3415 IEEE exceptions for math error handling may want to use this flag
3416 for speed while maintaining IEEE arithmetic compatibility.
3418 This option should never be turned on by any @option{-O} option since
3419 it can result in incorrect output for programs which depend on
3420 an exact implementation of IEEE or ISO rules/specifications for
3423 The default is @option{-fmath-errno}.
3425 @item -funsafe-math-optimizations
3426 @opindex funsafe-math-optimizations
3427 Allow optimizations for floating-point arithmetic that (a) assume
3428 that arguments and results are valid and (b) may violate IEEE or
3429 ANSI standards. When used at link-time, it may include libraries
3430 or startup files that change the default FPU control word or other
3431 similar optimizations.
3433 This option should never be turned on by any @option{-O} option since
3434 it can result in incorrect output for programs which depend on
3435 an exact implementation of IEEE or ISO rules/specifications for
3438 The default is @option{-fno-unsafe-math-optimizations}.
3440 @item -fno-trapping-math
3441 @opindex fno-trapping-math
3442 Compile code assuming that floating-point operations cannot generate
3443 user-visible traps. Setting this option may allow faster code
3444 if one relies on ``non-stop'' IEEE arithmetic, for example.
3446 This option should never be turned on by any @option{-O} option since
3447 it can result in incorrect output for programs which depend on
3448 an exact implementation of IEEE or ISO rules/specifications for
3451 The default is @option{-ftrapping-math}.
3454 The following options control specific optimizations. The @option{-O2}
3455 option turns on all of these optimizations except @option{-funroll-loops}
3456 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3457 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3458 but specific machines may handle it differently.
3460 You can use the following flags in the rare cases when ``fine-tuning''
3461 of optimizations to be performed is desired.
3463 Not all of the optimizations performed by GCC have @option{-f} options
3467 @item -fstrength-reduce
3468 @opindex fstrength-reduce
3469 Perform the optimizations of loop strength reduction and
3470 elimination of iteration variables.
3472 @item -fthread-jumps
3473 @opindex fthread-jumps
3474 Perform optimizations where we check to see if a jump branches to a
3475 location where another comparison subsumed by the first is found. If
3476 so, the first branch is redirected to either the destination of the
3477 second branch or a point immediately following it, depending on whether
3478 the condition is known to be true or false.
3480 @item -fcse-follow-jumps
3481 @opindex fcse-follow-jumps
3482 In common subexpression elimination, scan through jump instructions
3483 when the target of the jump is not reached by any other path. For
3484 example, when CSE encounters an @code{if} statement with an
3485 @code{else} clause, CSE will follow the jump when the condition
3488 @item -fcse-skip-blocks
3489 @opindex fcse-skip-blocks
3490 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3491 follow jumps which conditionally skip over blocks. When CSE
3492 encounters a simple @code{if} statement with no else clause,
3493 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3494 body of the @code{if}.
3496 @item -frerun-cse-after-loop
3497 @opindex frerun-cse-after-loop
3498 Re-run common subexpression elimination after loop optimizations has been
3501 @item -frerun-loop-opt
3502 @opindex frerun-loop-opt
3503 Run the loop optimizer twice.
3507 Perform a global common subexpression elimination pass.
3508 This pass also performs global constant and copy propagation.
3510 @emph{Note:} When compiling a program using computed gotos, a GCC
3511 extension, you may get better runtime performance if you disable
3512 the global common subexpression elmination pass by adding
3513 @option{-fno-gcse} to the command line.
3517 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3518 attempt to move loads which are only killed by stores into themselves. This
3519 allows a loop containing a load/store sequence to be changed to a load outside
3520 the loop, and a copy/store within the loop.
3524 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3525 subexpression elimination. This pass will attempt to move stores out of loops.
3526 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3527 can be changed to a load before the loop and a store after the loop.
3529 @item -fdelete-null-pointer-checks
3530 @opindex fdelete-null-pointer-checks
3531 Use global dataflow analysis to identify and eliminate useless checks
3532 for null pointers. The compiler assumes that dereferencing a null
3533 pointer would have halted the program. If a pointer is checked after
3534 it has already been dereferenced, it cannot be null.
3536 In some environments, this assumption is not true, and programs can
3537 safely dereference null pointers. Use
3538 @option{-fno-delete-null-pointer-checks} to disable this optimization
3539 for programs which depend on that behavior.
3541 @item -fexpensive-optimizations
3542 @opindex fexpensive-optimizations
3543 Perform a number of minor optimizations that are relatively expensive.
3545 @item -foptimize-register-move
3547 @opindex foptimize-register-move
3549 Attempt to reassign register numbers in move instructions and as
3550 operands of other simple instructions in order to maximize the amount of
3551 register tying. This is especially helpful on machines with two-operand
3552 instructions. GCC enables this optimization by default with @option{-O2}
3555 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3558 @item -fdelayed-branch
3559 @opindex fdelayed-branch
3560 If supported for the target machine, attempt to reorder instructions
3561 to exploit instruction slots available after delayed branch
3564 @item -fschedule-insns
3565 @opindex fschedule-insns
3566 If supported for the target machine, attempt to reorder instructions to
3567 eliminate execution stalls due to required data being unavailable. This
3568 helps machines that have slow floating point or memory load instructions
3569 by allowing other instructions to be issued until the result of the load
3570 or floating point instruction is required.
3572 @item -fschedule-insns2
3573 @opindex fschedule-insns2
3574 Similar to @option{-fschedule-insns}, but requests an additional pass of
3575 instruction scheduling after register allocation has been done. This is
3576 especially useful on machines with a relatively small number of
3577 registers and where memory load instructions take more than one cycle.
3579 @item -ffunction-sections
3580 @itemx -fdata-sections
3581 @opindex ffunction-sections
3582 @opindex fdata-sections
3583 Place each function or data item into its own section in the output
3584 file if the target supports arbitrary sections. The name of the
3585 function or the name of the data item determines the section's name
3588 Use these options on systems where the linker can perform optimizations
3589 to improve locality of reference in the instruction space. HPPA
3590 processors running HP-UX and Sparc processors running Solaris 2 have
3591 linkers with such optimizations. Other systems using the ELF object format
3592 as well as AIX may have these optimizations in the future.
3594 Only use these options when there are significant benefits from doing
3595 so. When you specify these options, the assembler and linker will
3596 create larger object and executable files and will also be slower.
3597 You will not be able to use @code{gprof} on all systems if you
3598 specify this option and you may have problems with debugging if
3599 you specify both this option and @option{-g}.
3601 @item -fcaller-saves
3602 @opindex fcaller-saves
3603 Enable values to be allocated in registers that will be clobbered by
3604 function calls, by emitting extra instructions to save and restore the
3605 registers around such calls. Such allocation is done only when it
3606 seems to result in better code than would otherwise be produced.
3608 This option is always enabled by default on certain machines, usually
3609 those which have no call-preserved registers to use instead.
3611 For all machines, optimization level 2 and higher enables this flag by
3614 @item -funroll-loops
3615 @opindex funroll-loops
3616 Unroll loops whose number of iterations can be determined at compile
3617 time or upon entry to the loop. @option{-funroll-loops} implies both
3618 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3619 option makes code larger, and may or may not make it run faster.
3621 @item -funroll-all-loops
3622 @opindex funroll-all-loops
3623 Unroll all loops, even if their number of iterations is uncertain when
3624 the loop is entered. This usually makes programs run more slowly.
3625 @option{-funroll-all-loops} implies the same options as
3626 @option{-funroll-loops},
3628 @item -fprefetch-loop-arrays
3629 @opindex fprefetch-loop-arrays
3630 If supported by the target machine, generate instructions to prefetch
3631 memory to improve the performance of loops that access large arrays.
3633 @item -fmove-all-movables
3634 @opindex fmove-all-movables
3635 Forces all invariant computations in loops to be moved
3638 @item -freduce-all-givs
3639 @opindex freduce-all-givs
3640 Forces all general-induction variables in loops to be
3643 @emph{Note:} When compiling programs written in Fortran,
3644 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3645 by default when you use the optimizer.
3647 These options may generate better or worse code; results are highly
3648 dependent on the structure of loops within the source code.
3650 These two options are intended to be removed someday, once
3651 they have helped determine the efficacy of various
3652 approaches to improving loop optimizations.
3654 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3655 know how use of these options affects
3656 the performance of your production code.
3657 We're very interested in code that runs @emph{slower}
3658 when these options are @emph{enabled}.
3661 @itemx -fno-peephole2
3662 @opindex fno-peephole
3663 @opindex fno-peephole2
3664 Disable any machine-specific peephole optimizations. The difference
3665 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3666 are implemented in the compiler; some targets use one, some use the
3667 other, a few use both.
3669 @item -fbranch-probabilities
3670 @opindex fbranch-probabilities
3671 After running a program compiled with @option{-fprofile-arcs}
3672 (@pxref{Debugging Options,, Options for Debugging Your Program or
3673 @command{gcc}}), you can compile it a second time using
3674 @option{-fbranch-probabilities}, to improve optimizations based on
3675 the number of times each branch was taken. When the program
3676 compiled with @option{-fprofile-arcs} exits it saves arc execution
3677 counts to a file called @file{@var{sourcename}.da} for each source
3678 file The information in this data file is very dependent on the
3679 structure of the generated code, so you must use the same source code
3680 and the same optimization options for both compilations.
3682 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3683 note on the first instruction of each basic block, and a
3684 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3685 These can be used to improve optimization. Currently, they are only
3686 used in one place: in @file{reorg.c}, instead of guessing which path a
3687 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3688 exactly determine which path is taken more often.
3690 @item -fno-guess-branch-probability
3691 @opindex fno-guess-branch-probability
3692 Do not guess branch probabilities using a randomized model.
3694 Sometimes gcc will opt to use a randomized model to guess branch
3695 probabilities, when none are available from either profiling feedback
3696 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3697 different runs of the compiler on the same program may produce different
3700 In a hard real-time system, people don't want different runs of the
3701 compiler to produce code that has different behavior; minimizing
3702 non-determinism is of paramount import. This switch allows users to
3703 reduce non-determinism, possibly at the expense of inferior
3706 @item -fstrict-aliasing
3707 @opindex fstrict-aliasing
3708 Allows the compiler to assume the strictest aliasing rules applicable to
3709 the language being compiled. For C (and C++), this activates
3710 optimizations based on the type of expressions. In particular, an
3711 object of one type is assumed never to reside at the same address as an
3712 object of a different type, unless the types are almost the same. For
3713 example, an @code{unsigned int} can alias an @code{int}, but not a
3714 @code{void*} or a @code{double}. A character type may alias any other
3717 Pay special attention to code like this:
3730 The practice of reading from a different union member than the one most
3731 recently written to (called ``type-punning'') is common. Even with
3732 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3733 is accessed through the union type. So, the code above will work as
3734 expected. However, this code might not:
3745 Every language that wishes to perform language-specific alias analysis
3746 should define a function that computes, given an @code{tree}
3747 node, an alias set for the node. Nodes in different alias sets are not
3748 allowed to alias. For an example, see the C front-end function
3749 @code{c_get_alias_set}.
3751 @item -falign-functions
3752 @itemx -falign-functions=@var{n}
3753 @opindex falign-functions
3754 Align the start of functions to the next power-of-two greater than
3755 @var{n}, skipping up to @var{n} bytes. For instance,
3756 @option{-falign-functions=32} aligns functions to the next 32-byte
3757 boundary, but @option{-falign-functions=24} would align to the next
3758 32-byte boundary only if this can be done by skipping 23 bytes or less.
3760 @option{-fno-align-functions} and @option{-falign-functions=1} are
3761 equivalent and mean that functions will not be aligned.
3763 Some assemblers only support this flag when @var{n} is a power of two;
3764 in that case, it is rounded up.
3766 If @var{n} is not specified, use a machine-dependent default.
3768 @item -falign-labels
3769 @itemx -falign-labels=@var{n}
3770 @opindex falign-labels
3771 Align all branch targets to a power-of-two boundary, skipping up to
3772 @var{n} bytes like @option{-falign-functions}. This option can easily
3773 make code slower, because it must insert dummy operations for when the
3774 branch target is reached in the usual flow of the code.
3776 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3777 are greater than this value, then their values are used instead.
3779 If @var{n} is not specified, use a machine-dependent default which is
3780 very likely to be @samp{1}, meaning no alignment.
3783 @itemx -falign-loops=@var{n}
3784 @opindex falign-loops
3785 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3786 like @option{-falign-functions}. The hope is that the loop will be
3787 executed many times, which will make up for any execution of the dummy
3790 If @var{n} is not specified, use a machine-dependent default.
3793 @itemx -falign-jumps=@var{n}
3794 @opindex falign-jumps
3795 Align branch targets to a power-of-two boundary, for branch targets
3796 where the targets can only be reached by jumping, skipping up to @var{n}
3797 bytes like @option{-falign-functions}. In this case, no dummy operations
3800 If @var{n} is not specified, use a machine-dependent default.
3804 Perform optimizations in static single assignment form. Each function's
3805 flow graph is translated into SSA form, optimizations are performed, and
3806 the flow graph is translated back from SSA form. Users should not
3807 specify this option, since it is not yet ready for production use.
3811 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3812 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3816 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3817 Like @option{-fssa}, this is an experimental feature.
3819 @item -fsingle-precision-constant
3820 @opindex fsingle-precision-constant
3821 Treat floating point constant as single precision constant instead of
3822 implicitly converting it to double precision constant.
3824 @item -frename-registers
3825 @opindex frename-registers
3826 Attempt to avoid false dependencies in scheduled code by making use
3827 of registers left over after register allocation. This optimization
3828 will most benefit processors with lots of registers. It can, however,
3829 make debugging impossible, since variables will no longer stay in
3830 a ``home register''.
3832 @item -fno-cprop-registers
3833 @opindex fno-cprop-registers
3834 After register allocation and post-register allocation instruction splitting,
3835 we perform a copy-propagation pass to try to reduce scheduling dependencies
3836 and occasionally eliminate the copy.
3838 @item --param @var{name}=@var{value}
3840 In some places, GCC uses various constants to control the amount of
3841 optimization that is done. For example, GCC will not inline functions
3842 that contain more that a certain number of instructions. You can
3843 control some of these constants on the command-line using the
3844 @option{--param} option.
3846 In each case, the @var{value} is an integer. The allowable choices for
3847 @var{name} are given in the following table:
3850 @item max-delay-slot-insn-search
3851 The maximum number of instructions to consider when looking for an
3852 instruction to fill a delay slot. If more than this arbitrary number of
3853 instructions is searched, the time savings from filling the delay slot
3854 will be minimal so stop searching. Increasing values mean more
3855 aggressive optimization, making the compile time increase with probably
3856 small improvement in executable run time.
3858 @item max-delay-slot-live-search
3859 When trying to fill delay slots, the maximum number of instructions to
3860 consider when searching for a block with valid live register
3861 information. Increasing this arbitrarily chosen value means more
3862 aggressive optimization, increasing the compile time. This parameter
3863 should be removed when the delay slot code is rewritten to maintain the
3866 @item max-gcse-memory
3867 The approximate maximum amount of memory that will be allocated in
3868 order to perform the global common subexpression elimination
3869 optimization. If more memory than specified is required, the
3870 optimization will not be done.
3872 @item max-gcse-passes
3873 The maximum number of passes of GCSE to run.
3875 @item max-pending-list-length
3876 The maximum number of pending dependencies scheduling will allow
3877 before flushing the current state and starting over. Large functions
3878 with few branches or calls can create excessively large lists which
3879 needlessly consume memory and resources.
3881 @item max-inline-insns
3882 If an function contains more than this many instructions, it
3883 will not be inlined. This option is precisely equivalent to
3884 @option{-finline-limit}.
3889 @node Preprocessor Options
3890 @section Options Controlling the Preprocessor
3891 @cindex preprocessor options
3892 @cindex options, preprocessor
3894 These options control the C preprocessor, which is run on each C source
3895 file before actual compilation.
3897 If you use the @option{-E} option, nothing is done except preprocessing.
3898 Some of these options make sense only together with @option{-E} because
3899 they cause the preprocessor output to be unsuitable for actual
3903 @item -include @var{file}
3905 Process @var{file} as input before processing the regular input file.
3906 In effect, the contents of @var{file} are compiled first. Any @option{-D}
3907 and @option{-U} options on the command line are always processed before
3908 @option{-include @var{file}}, regardless of the order in which they are
3909 written. All the @option{-include} and @option{-imacros} options are
3910 processed in the order in which they are written.
3912 @item -imacros @var{file}
3914 Process @var{file} as input, discarding the resulting output, before
3915 processing the regular input file. Because the output generated from
3916 @var{file} is discarded, the only effect of @option{-imacros @var{file}}
3917 is to make the macros defined in @var{file} available for use in the
3918 main input. All the @option{-include} and @option{-imacros} options are
3919 processed in the order in which they are written.
3921 @item -idirafter @var{dir}
3923 @cindex second include path
3924 Add the directory @var{dir} to the second include path. The directories
3925 on the second include path are searched when a header file is not found
3926 in any of the directories in the main include path (the one that
3927 @option{-I} adds to).
3929 @item -iprefix @var{prefix}
3931 Specify @var{prefix} as the prefix for subsequent @option{-iwithprefix}
3934 @item -iwithprefix @var{dir}
3935 @opindex iwithprefix
3936 Add a directory to the second include path. The directory's name is
3937 made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was
3938 specified previously with @option{-iprefix}. If you have not specified a
3939 prefix yet, the directory containing the installed passes of the
3940 compiler is used as the default.
3942 @item -iwithprefixbefore @var{dir}
3943 @opindex iwithprefixbefore
3944 Add a directory to the main include path. The directory's name is made
3945 by concatenating @var{prefix} and @var{dir}, as in the case of
3946 @option{-iwithprefix}.
3948 @item -isystem @var{dir}
3950 Add a directory to the beginning of the second include path, marking it
3951 as a system directory, so that it gets the same special treatment as
3952 is applied to the standard system directories.
3956 Do not search the standard system directories for header files. Only
3957 the directories you have specified with @option{-I} options (and the
3958 current directory, if appropriate) are searched. @xref{Directory
3959 Options}, for information on @option{-I}.
3961 By using both @option{-nostdinc} and @option{-I-}, you can limit the include-file
3962 search path to only those directories you specify explicitly.
3966 When searching for a header file in a directory, remap file names if a
3967 file named @file{header.gcc} exists in that directory. This can be used
3968 to work around limitations of file systems with file name restrictions.
3969 The @file{header.gcc} file should contain a series of lines with two
3970 tokens on each line: the first token is the name to map, and the second
3971 token is the actual name to use.
3975 Do not predefine any nonstandard macros. (Including architecture flags).
3979 Run only the C preprocessor. Preprocess all the C source files
3980 specified and output the results to standard output or to the
3981 specified output file.
3985 Tell the preprocessor not to discard comments. Used with the
3990 Tell the preprocessor not to generate @samp{#line} directives.
3991 Used with the @option{-E} option.
3994 @cindex dependencies, make
3997 Instead of outputting the result of preprocessing, output a rule
3998 suitable for @code{make} describing the dependencies of the main source
3999 file. The preprocessor outputs one @code{make} rule containing the
4000 object file name for that source file, a colon, and the names of all the
4001 included files. Unless overridden explicitly, the object file name
4002 consists of the basename of the source file with any suffix replaced with
4003 object file suffix. If there are many included files then the
4004 rule is split into several lines using @samp{\}-newline.
4006 @option{-M} implies @option{-E}.
4010 Like @option{-M}, but mention only the files included with @samp{#include
4011 "@var{file}"}. System header files included with @samp{#include
4012 <@var{file}>} are omitted.
4016 Like @option{-M} but the dependency information is written to a file
4017 rather than stdout. @code{gcc} will use the same file name and
4018 directory as the object file, but with the suffix @file{.d} instead.
4020 This is in addition to compiling the main file as specified---@option{-MD}
4021 does not inhibit ordinary compilation the way @option{-M} does,
4022 unless you also specify @option{-MG}.
4024 With Mach, you can use the utility @code{md} to merge multiple
4025 dependency files into a single dependency file suitable for using with
4026 the @samp{make} command.
4030 Like @option{-MD} except mention only user header files, not system
4033 @item -MF @var{file}
4035 When used with @option{-M} or @option{-MM}, specifies a file to write the
4036 dependencies to. This allows the preprocessor to write the preprocessed
4037 file to stdout normally. If no @option{-MF} switch is given, CPP sends
4038 the rules to stdout and suppresses normal preprocessed output.
4040 Another way to specify output of a @code{make} rule is by setting
4041 the environment variable @env{DEPENDENCIES_OUTPUT} (@pxref{Environment
4046 When used with @option{-M} or @option{-MM}, @option{-MG} says to treat missing
4047 header files as generated files and assume they live in the same
4048 directory as the source file. It suppresses preprocessed output, as a
4049 missing header file is ordinarily an error.
4051 This feature is used in automatic updating of makefiles.
4055 This option instructs CPP to add a phony target for each dependency
4056 other than the main file, causing each to depend on nothing. These
4057 dummy rules work around errors @code{make} gives if you remove header
4058 files without updating the @code{Makefile} to match.
4060 This is typical output:-
4063 /tmp/test.o: /tmp/test.c /tmp/test.h
4068 @item -MQ @var{target}
4069 @item -MT @var{target}
4072 By default CPP uses the main file name, including any path, and appends
4073 the object suffix, normally ``.o'', to it to obtain the name of the
4074 target for dependency generation. With @option{-MT} you can specify a
4075 target yourself, overriding the default one.
4077 If you want multiple targets, you can specify them as a single argument
4078 to @option{-MT}, or use multiple @option{-MT} options.
4080 The targets you specify are output in the order they appear on the
4081 command line. @option{-MQ} is identical to @option{-MT}, except that the
4082 target name is quoted for Make, but with @option{-MT} it isn't. For
4083 example, @option{-MT '$(objpfx)foo.o'} gives
4086 $(objpfx)foo.o: /tmp/foo.c
4089 but @option{-MQ '$(objpfx)foo.o'} gives
4092 $$(objpfx)foo.o: /tmp/foo.c
4095 The default target is automatically quoted, as if it were given with
4100 Print the name of each header file used, in addition to other normal
4103 @item -A@var{question}(@var{answer})
4105 Assert the answer @var{answer} for @var{question}, in case it is tested
4106 with a preprocessing conditional such as @samp{#if
4107 #@var{question}(@var{answer})}. @option{-A-} disables the standard
4108 assertions that normally describe the target machine.
4112 Define macro @var{macro} with the string @samp{1} as its definition.
4114 @item -D@var{macro}=@var{defn}
4115 Define macro @var{macro} as @var{defn}. All instances of @option{-D} on
4116 the command line are processed before any @option{-U} options.
4118 Any @option{-D} and @option{-U} options on the command line are processed in
4119 order, and always before @option{-imacros @var{file}}, regardless of the
4120 order in which they are written.
4124 Undefine macro @var{macro}. @option{-U} options are evaluated after all
4125 @option{-D} options, but before any @option{-include} and @option{-imacros}
4128 Any @option{-D} and @option{-U} options on the command line are processed in
4129 order, and always before @option{-imacros @var{file}}, regardless of the
4130 order in which they are written.
4134 Tell the preprocessor to output only a list of the macro definitions
4135 that are in effect at the end of preprocessing. Used with the @option{-E}
4140 Tell the preprocessing to pass all macro definitions into the output, in
4141 their proper sequence in the rest of the output.
4145 Like @option{-dD} except that the macro arguments and contents are omitted.
4146 Only @samp{#define @var{name}} is included in the output.
4150 Output @samp{#include} directives in addition to the result of
4153 @item -fpreprocessed
4154 @opindex fpreprocessed
4155 Indicate to the preprocessor that the input file has already been
4156 preprocessed. This suppresses things like macro expansion, trigraph
4157 conversion, escaped newline splicing, and processing of most directives.
4158 The preprocessor still recognizes and removes comments, so that you can
4159 pass a file preprocessed with @option{-C} to the compiler without
4160 problems. In this mode the integrated preprocessor is little more than
4161 a tokenizer for the front ends.
4163 @option{-fpreprocessed} is implicit if the input file has one of the
4164 extensions @samp{i}, @samp{ii} or @samp{mi}. These are the extensions
4165 that GCC uses for preprocessed files created by @option{-save-temps}.
4169 Process ISO standard trigraph sequences. These are three-character
4170 sequences, all starting with @samp{??}, that are defined by ISO C to
4171 stand for single characters. For example, @samp{??/} stands for
4172 @samp{\}, so @samp{'??/n'} is a character constant for a newline. By
4173 default, GCC ignores trigraphs, but in standard-conforming modes it
4174 converts them. See the @option{-std} and @option{-ansi} options.
4176 The nine trigraph sequences are
4179 @expansion{} @samp{[}
4182 @expansion{} @samp{]}
4185 @expansion{} @samp{@{}
4188 @expansion{} @samp{@}}
4191 @expansion{} @samp{#}
4194 @expansion{} @samp{\}
4197 @expansion{} @samp{^}
4200 @expansion{} @samp{|}
4203 @expansion{} @samp{~}
4207 Trigraph support is not popular, so many compilers do not implement it
4208 properly. Portable code should not rely on trigraphs being either
4209 converted or ignored.
4211 @item -Wp,@var{option}
4213 Pass @var{option} as an option to the preprocessor. If @var{option}
4214 contains commas, it is split into multiple options at the commas.
4217 @node Assembler Options
4218 @section Passing Options to the Assembler
4220 @c prevent bad page break with this line
4221 You can pass options to the assembler.
4224 @item -Wa,@var{option}
4226 Pass @var{option} as an option to the assembler. If @var{option}
4227 contains commas, it is split into multiple options at the commas.
4231 @section Options for Linking
4232 @cindex link options
4233 @cindex options, linking
4235 These options come into play when the compiler links object files into
4236 an executable output file. They are meaningless if the compiler is
4237 not doing a link step.
4241 @item @var{object-file-name}
4242 A file name that does not end in a special recognized suffix is
4243 considered to name an object file or library. (Object files are
4244 distinguished from libraries by the linker according to the file
4245 contents.) If linking is done, these object files are used as input
4254 If any of these options is used, then the linker is not run, and
4255 object file names should not be used as arguments. @xref{Overall
4259 @item -l@var{library}
4260 @itemx -l @var{library}
4262 Search the library named @var{library} when linking. (The second
4263 alternative with the library as a separate argument is only for
4264 POSIX compliance and is not recommended.)
4266 It makes a difference where in the command you write this option; the
4267 linker searches and processes libraries and object files in the order they
4268 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4269 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4270 to functions in @samp{z}, those functions may not be loaded.
4272 The linker searches a standard list of directories for the library,
4273 which is actually a file named @file{lib@var{library}.a}. The linker
4274 then uses this file as if it had been specified precisely by name.
4276 The directories searched include several standard system directories
4277 plus any that you specify with @option{-L}.
4279 Normally the files found this way are library files---archive files
4280 whose members are object files. The linker handles an archive file by
4281 scanning through it for members which define symbols that have so far
4282 been referenced but not defined. But if the file that is found is an
4283 ordinary object file, it is linked in the usual fashion. The only
4284 difference between using an @option{-l} option and specifying a file name
4285 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4286 and searches several directories.
4290 You need this special case of the @option{-l} option in order to
4291 link an Objective-C program.
4294 @opindex nostartfiles
4295 Do not use the standard system startup files when linking.
4296 The standard system libraries are used normally, unless @option{-nostdlib}
4297 or @option{-nodefaultlibs} is used.
4299 @item -nodefaultlibs
4300 @opindex nodefaultlibs
4301 Do not use the standard system libraries when linking.
4302 Only the libraries you specify will be passed to the linker.
4303 The standard startup files are used normally, unless @option{-nostartfiles}
4304 is used. The compiler may generate calls to memcmp, memset, and memcpy
4305 for System V (and ISO C) environments or to bcopy and bzero for
4306 BSD environments. These entries are usually resolved by entries in
4307 libc. These entry points should be supplied through some other
4308 mechanism when this option is specified.
4312 Do not use the standard system startup files or libraries when linking.
4313 No startup files and only the libraries you specify will be passed to
4314 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4315 for System V (and ISO C) environments or to bcopy and bzero for
4316 BSD environments. These entries are usually resolved by entries in
4317 libc. These entry points should be supplied through some other
4318 mechanism when this option is specified.
4320 @cindex @option{-lgcc}, use with @option{-nostdlib}
4321 @cindex @option{-nostdlib} and unresolved references
4322 @cindex unresolved references and @option{-nostdlib}
4323 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4324 @cindex @option{-nodefaultlibs} and unresolved references
4325 @cindex unresolved references and @option{-nodefaultlibs}
4326 One of the standard libraries bypassed by @option{-nostdlib} and
4327 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4328 that GCC uses to overcome shortcomings of particular machines, or special
4329 needs for some languages.
4330 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4331 Collection (GCC) Internals},
4332 for more discussion of @file{libgcc.a}.)
4333 In most cases, you need @file{libgcc.a} even when you want to avoid
4334 other standard libraries. In other words, when you specify @option{-nostdlib}
4335 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4336 This ensures that you have no unresolved references to internal GCC
4337 library subroutines. (For example, @samp{__main}, used to ensure C++
4338 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4339 GNU Compiler Collection (GCC) Internals}.)
4343 Remove all symbol table and relocation information from the executable.
4347 On systems that support dynamic linking, this prevents linking with the shared
4348 libraries. On other systems, this option has no effect.
4352 Produce a shared object which can then be linked with other objects to
4353 form an executable. Not all systems support this option. For predictable
4354 results, you must also specify the same set of options that were used to
4355 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4356 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4357 needs to build supplementary stub code for constructors to work. On
4358 multi-libbed systems, @samp{gcc -shared} must select the correct support
4359 libraries to link against. Failing to supply the correct flags may lead
4360 to subtle defects. Supplying them in cases where they are not necessary
4363 @item -shared-libgcc
4364 @itemx -static-libgcc
4365 @opindex shared-libgcc
4366 @opindex static-libgcc
4367 On systems that provide @file{libgcc} as a shared library, these options
4368 force the use of either the shared or static version respectively.
4369 If no shared version of @file{libgcc} was built when the compiler was
4370 configured, these options have no effect.
4372 There are several situations in which an application should use the
4373 shared @file{libgcc} instead of the static version. The most common
4374 of these is when the application wishes to throw and catch exceptions
4375 across different shared libraries. In that case, each of the libraries
4376 as well as the application itself should use the shared @file{libgcc}.
4378 Therefore, whenever you specify the @option{-shared} option, the GCC
4379 driver automatically adds @option{-shared-libgcc}, unless you explicitly
4380 specify @option{-static-libgcc}. The G++ driver automatically adds
4381 @option{-shared-libgcc} when you build a main executable as well because
4382 for C++ programs that is typically the right thing to do.
4383 (Exception-handling will not work reliably otherwise.)
4385 However, when linking a main executable written in C, you must
4386 explicitly say @option{-shared-libgcc} if you want to use the shared
4391 Bind references to global symbols when building a shared object. Warn
4392 about any unresolved references (unless overridden by the link editor
4393 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4396 @item -Xlinker @var{option}
4398 Pass @var{option} as an option to the linker. You can use this to
4399 supply system-specific linker options which GCC does not know how to
4402 If you want to pass an option that takes an argument, you must use
4403 @option{-Xlinker} twice, once for the option and once for the argument.
4404 For example, to pass @option{-assert definitions}, you must write
4405 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4406 @option{-Xlinker "-assert definitions"}, because this passes the entire
4407 string as a single argument, which is not what the linker expects.
4409 @item -Wl,@var{option}
4411 Pass @var{option} as an option to the linker. If @var{option} contains
4412 commas, it is split into multiple options at the commas.
4414 @item -u @var{symbol}
4416 Pretend the symbol @var{symbol} is undefined, to force linking of
4417 library modules to define it. You can use @option{-u} multiple times with
4418 different symbols to force loading of additional library modules.
4421 @node Directory Options
4422 @section Options for Directory Search
4423 @cindex directory options
4424 @cindex options, directory search
4427 These options specify directories to search for header files, for
4428 libraries and for parts of the compiler:
4433 Add the directory @var{dir} to the head of the list of directories to be
4434 searched for header files. This can be used to override a system header
4435 file, substituting your own version, since these directories are
4436 searched before the system header file directories. However, you should
4437 not use this option to add directories that contain vendor-supplied
4438 system header files (use @option{-isystem} for that). If you use more than
4439 one @option{-I} option, the directories are scanned in left-to-right
4440 order; the standard system directories come after.
4442 If a standard system include directory, or a directory specified with
4443 @option{-isystem}, is also specified with @option{-I}, it will be
4444 searched only in the position requested by @option{-I}. Also, it will
4445 not be considered a system include directory. If that directory really
4446 does contain system headers, there is a good chance that they will
4447 break. For instance, if GCC's installation procedure edited the headers
4448 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4449 original, buggy headers to be found instead of the corrected ones. GCC
4450 will issue a warning when a system include directory is hidden in this
4455 Any directories you specify with @option{-I} options before the @option{-I-}
4456 option are searched only for the case of @samp{#include "@var{file}"};
4457 they are not searched for @samp{#include <@var{file}>}.
4459 If additional directories are specified with @option{-I} options after
4460 the @option{-I-}, these directories are searched for all @samp{#include}
4461 directives. (Ordinarily @emph{all} @option{-I} directories are used
4464 In addition, the @option{-I-} option inhibits the use of the current
4465 directory (where the current input file came from) as the first search
4466 directory for @samp{#include "@var{file}"}. There is no way to
4467 override this effect of @option{-I-}. With @option{-I.} you can specify
4468 searching the directory which was current when the compiler was
4469 invoked. That is not exactly the same as what the preprocessor does
4470 by default, but it is often satisfactory.
4472 @option{-I-} does not inhibit the use of the standard system directories
4473 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4478 Add directory @var{dir} to the list of directories to be searched
4481 @item -B@var{prefix}
4483 This option specifies where to find the executables, libraries,
4484 include files, and data files of the compiler itself.
4486 The compiler driver program runs one or more of the subprograms
4487 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4488 @var{prefix} as a prefix for each program it tries to run, both with and
4489 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4491 For each subprogram to be run, the compiler driver first tries the
4492 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4493 was not specified, the driver tries two standard prefixes, which are
4494 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4495 those results in a file name that is found, the unmodified program
4496 name is searched for using the directories specified in your
4497 @env{PATH} environment variable.
4499 The compiler will check to see if the path provided by the @option{-B}
4500 refers to a directory, and if necessary it will add a directory
4501 separator character at the end of the path.
4503 @option{-B} prefixes that effectively specify directory names also apply
4504 to libraries in the linker, because the compiler translates these
4505 options into @option{-L} options for the linker. They also apply to
4506 includes files in the preprocessor, because the compiler translates these
4507 options into @option{-isystem} options for the preprocessor. In this case,
4508 the compiler appends @samp{include} to the prefix.
4510 The run-time support file @file{libgcc.a} can also be searched for using
4511 the @option{-B} prefix, if needed. If it is not found there, the two
4512 standard prefixes above are tried, and that is all. The file is left
4513 out of the link if it is not found by those means.
4515 Another way to specify a prefix much like the @option{-B} prefix is to use
4516 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4519 As a special kludge, if the path provided by @option{-B} is
4520 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4521 9, then it will be replaced by @file{[dir/]include}. This is to help
4522 with boot-strapping the compiler.
4524 @item -specs=@var{file}
4526 Process @var{file} after the compiler reads in the standard @file{specs}
4527 file, in order to override the defaults that the @file{gcc} driver
4528 program uses when determining what switches to pass to @file{cc1},
4529 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4530 @option{-specs=@var{file}} can be specified on the command line, and they
4531 are processed in order, from left to right.
4537 @section Specifying subprocesses and the switches to pass to them
4539 @command{gcc} is a driver program. It performs its job by invoking a
4540 sequence of other programs to do the work of compiling, assembling and
4541 linking. GCC interprets its command-line parameters and uses these to
4542 deduce which programs it should invoke, and which command-line options
4543 it ought to place on their command lines. This behavior is controlled
4544 by @dfn{spec strings}. In most cases there is one spec string for each
4545 program that GCC can invoke, but a few programs have multiple spec
4546 strings to control their behavior. The spec strings built into GCC can
4547 be overridden by using the @option{-specs=} command-line switch to specify
4550 @dfn{Spec files} are plaintext files that are used to construct spec
4551 strings. They consist of a sequence of directives separated by blank
4552 lines. The type of directive is determined by the first non-whitespace
4553 character on the line and it can be one of the following:
4556 @item %@var{command}
4557 Issues a @var{command} to the spec file processor. The commands that can
4561 @item %include <@var{file}>
4563 Search for @var{file} and insert its text at the current point in the
4566 @item %include_noerr <@var{file}>
4567 @cindex %include_noerr
4568 Just like @samp{%include}, but do not generate an error message if the include
4569 file cannot be found.
4571 @item %rename @var{old_name} @var{new_name}
4573 Rename the spec string @var{old_name} to @var{new_name}.
4577 @item *[@var{spec_name}]:
4578 This tells the compiler to create, override or delete the named spec
4579 string. All lines after this directive up to the next directive or
4580 blank line are considered to be the text for the spec string. If this
4581 results in an empty string then the spec will be deleted. (Or, if the
4582 spec did not exist, then nothing will happened.) Otherwise, if the spec
4583 does not currently exist a new spec will be created. If the spec does
4584 exist then its contents will be overridden by the text of this
4585 directive, unless the first character of that text is the @samp{+}
4586 character, in which case the text will be appended to the spec.
4588 @item [@var{suffix}]:
4589 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4590 and up to the next directive or blank line are considered to make up the
4591 spec string for the indicated suffix. When the compiler encounters an
4592 input file with the named suffix, it will processes the spec string in
4593 order to work out how to compile that file. For example:
4600 This says that any input file whose name ends in @samp{.ZZ} should be
4601 passed to the program @samp{z-compile}, which should be invoked with the
4602 command-line switch @option{-input} and with the result of performing the
4603 @samp{%i} substitution. (See below.)
4605 As an alternative to providing a spec string, the text that follows a
4606 suffix directive can be one of the following:
4609 @item @@@var{language}
4610 This says that the suffix is an alias for a known @var{language}. This is
4611 similar to using the @option{-x} command-line switch to GCC to specify a
4612 language explicitly. For example:
4619 Says that .ZZ files are, in fact, C++ source files.
4622 This causes an error messages saying:
4625 @var{name} compiler not installed on this system.
4629 GCC already has an extensive list of suffixes built into it.
4630 This directive will add an entry to the end of the list of suffixes, but
4631 since the list is searched from the end backwards, it is effectively
4632 possible to override earlier entries using this technique.
4636 GCC has the following spec strings built into it. Spec files can
4637 override these strings or create their own. Note that individual
4638 targets can also add their own spec strings to this list.
4641 asm Options to pass to the assembler
4642 asm_final Options to pass to the assembler post-processor
4643 cpp Options to pass to the C preprocessor
4644 cc1 Options to pass to the C compiler
4645 cc1plus Options to pass to the C++ compiler
4646 endfile Object files to include at the end of the link
4647 link Options to pass to the linker
4648 lib Libraries to include on the command line to the linker
4649 libgcc Decides which GCC support library to pass to the linker
4650 linker Sets the name of the linker
4651 predefines Defines to be passed to the C preprocessor
4652 signed_char Defines to pass to CPP to say whether @code{char} is signed
4654 startfile Object files to include at the start of the link
4657 Here is a small example of a spec file:
4663 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4666 This example renames the spec called @samp{lib} to @samp{old_lib} and
4667 then overrides the previous definition of @samp{lib} with a new one.
4668 The new definition adds in some extra command-line options before
4669 including the text of the old definition.
4671 @dfn{Spec strings} are a list of command-line options to be passed to their
4672 corresponding program. In addition, the spec strings can contain
4673 @samp{%}-prefixed sequences to substitute variable text or to
4674 conditionally insert text into the command line. Using these constructs
4675 it is possible to generate quite complex command lines.
4677 Here is a table of all defined @samp{%}-sequences for spec
4678 strings. Note that spaces are not generated automatically around the
4679 results of expanding these sequences. Therefore you can concatenate them
4680 together or combine them with constant text in a single argument.
4684 Substitute one @samp{%} into the program name or argument.
4687 Substitute the name of the input file being processed.
4690 Substitute the basename of the input file being processed.
4691 This is the substring up to (and not including) the last period
4692 and not including the directory.
4695 This is the same as @samp{%b}, but include the file suffix (text after
4699 Marks the argument containing or following the @samp{%d} as a
4700 temporary file name, so that that file will be deleted if GCC exits
4701 successfully. Unlike @samp{%g}, this contributes no text to the
4704 @item %g@var{suffix}
4705 Substitute a file name that has suffix @var{suffix} and is chosen
4706 once per compilation, and mark the argument in the same way as
4707 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4708 name is now chosen in a way that is hard to predict even when previously
4709 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4710 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4711 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4712 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4713 was simply substituted with a file name chosen once per compilation,
4714 without regard to any appended suffix (which was therefore treated
4715 just like ordinary text), making such attacks more likely to succeed.
4717 @item %u@var{suffix}
4718 Like @samp{%g}, but generates a new temporary file name even if
4719 @samp{%u@var{suffix}} was already seen.
4721 @item %U@var{suffix}
4722 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4723 new one if there is no such last file name. In the absence of any
4724 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4725 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4726 would involve the generation of two distinct file names, one
4727 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4728 simply substituted with a file name chosen for the previous @samp{%u},
4729 without regard to any appended suffix.
4731 @item %j@var{SUFFIX}
4732 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4733 writable, and if save-temps is off; otherwise, substitute the name
4734 of a temporary file, just like @samp{%u}. This temporary file is not
4735 meant for communication between processes, but rather as a junk
4738 @item %.@var{SUFFIX}
4739 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4740 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4741 terminated by the next space or %.
4744 Marks the argument containing or following the @samp{%w} as the
4745 designated output file of this compilation. This puts the argument
4746 into the sequence of arguments that @samp{%o} will substitute later.
4749 Substitutes the names of all the output files, with spaces
4750 automatically placed around them. You should write spaces
4751 around the @samp{%o} as well or the results are undefined.
4752 @samp{%o} is for use in the specs for running the linker.
4753 Input files whose names have no recognized suffix are not compiled
4754 at all, but they are included among the output files, so they will
4758 Substitutes the suffix for object files. Note that this is
4759 handled specially when it immediately follows @samp{%g, %u, or %U},
4760 because of the need for those to form complete file names. The
4761 handling is such that @samp{%O} is treated exactly as if it had already
4762 been substituted, except that @samp{%g, %u, and %U} do not currently
4763 support additional @var{suffix} characters following @samp{%O} as they would
4764 following, for example, @samp{.o}.
4767 Substitutes the standard macro predefinitions for the
4768 current target machine. Use this when running @code{cpp}.
4771 Like @samp{%p}, but puts @samp{__} before and after the name of each
4772 predefined macro, except for macros that start with @samp{__} or with
4773 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4777 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4780 Current argument is the name of a library or startup file of some sort.
4781 Search for that file in a standard list of directories and substitute
4782 the full name found.
4785 Print @var{str} as an error message. @var{str} is terminated by a newline.
4786 Use this when inconsistent options are detected.
4789 Output @samp{-} if the input for the current command is coming from a pipe.
4792 Substitute the contents of spec string @var{name} at this point.
4795 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4797 @item %x@{@var{option}@}
4798 Accumulate an option for @samp{%X}.
4801 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4805 Output the accumulated assembler options specified by @option{-Wa}.
4808 Output the accumulated preprocessor options specified by @option{-Wp}.
4811 Substitute the major version number of GCC@.
4812 (For version 2.9.5, this is 2.)
4815 Substitute the minor version number of GCC@.
4816 (For version 2.9.5, this is 9.)
4819 Substitute the patch level number of GCC@.
4820 (For version 2.9.5, this is 5.)
4823 Process the @code{asm} spec. This is used to compute the
4824 switches to be passed to the assembler.
4827 Process the @code{asm_final} spec. This is a spec string for
4828 passing switches to an assembler post-processor, if such a program is
4832 Process the @code{link} spec. This is the spec for computing the
4833 command line passed to the linker. Typically it will make use of the
4834 @samp{%L %G %S %D and %E} sequences.
4837 Dump out a @option{-L} option for each directory that GCC believes might
4838 contain startup files. If the target supports multilibs then the
4839 current multilib directory will be prepended to each of these paths.
4842 Output the multilib directory with directory separators replaced with
4843 @samp{_}. If multilib directories are not set, or the multilib directory is
4844 @file{.} then this option emits nothing.
4847 Process the @code{lib} spec. This is a spec string for deciding which
4848 libraries should be included on the command line to the linker.
4851 Process the @code{libgcc} spec. This is a spec string for deciding
4852 which GCC support library should be included on the command line to the linker.
4855 Process the @code{startfile} spec. This is a spec for deciding which
4856 object files should be the first ones passed to the linker. Typically
4857 this might be a file named @file{crt0.o}.
4860 Process the @code{endfile} spec. This is a spec string that specifies
4861 the last object files that will be passed to the linker.
4864 Process the @code{cpp} spec. This is used to construct the arguments
4865 to be passed to the C preprocessor.
4868 Process the @code{signed_char} spec. This is intended to be used
4869 to tell cpp whether a char is signed. It typically has the definition:
4871 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4875 Process the @code{cc1} spec. This is used to construct the options to be
4876 passed to the actual C compiler (@samp{cc1}).
4879 Process the @code{cc1plus} spec. This is used to construct the options to be
4880 passed to the actual C++ compiler (@samp{cc1plus}).
4883 Substitute the variable part of a matched option. See below.
4884 Note that each comma in the substituted string is replaced by
4888 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4889 If that switch was not specified, this substitutes nothing. Note that
4890 the leading dash is omitted when specifying this option, and it is
4891 automatically inserted if the substitution is performed. Thus the spec
4892 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4893 and would output the command line option @option{-foo}.
4895 @item %W@{@code{S}@}
4896 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4899 @item %@{@code{S}*@}
4900 Substitutes all the switches specified to GCC whose names start
4901 with @code{-S}, but which also take an argument. This is used for
4902 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4903 GCC considers @option{-o foo} as being
4904 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4905 text, including the space. Thus two arguments would be generated.
4907 @item %@{^@code{S}*@}
4908 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4909 argument. Thus %@{^o*@} would only generate one argument, not two.
4911 @item %@{@code{S}*&@code{T}*@}
4912 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4913 (the order of @code{S} and @code{T} in the spec is not significant).
4914 There can be any number of ampersand-separated variables; for each the
4915 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4917 @item %@{<@code{S}@}
4918 Remove all occurrences of @code{-S} from the command line. Note---this
4919 command is position dependent. @samp{%} commands in the spec string
4920 before this option will see @code{-S}, @samp{%} commands in the spec
4921 string after this option will not.
4923 @item %@{@code{S}*:@code{X}@}
4924 Substitutes @code{X} if one or more switches whose names start with
4925 @code{-S} are specified to GCC@. Note that the tail part of the
4926 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4927 for each occurrence of @samp{%*} within @code{X}.
4929 @item %@{@code{S}:@code{X}@}
4930 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4932 @item %@{!@code{S}:@code{X}@}
4933 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4935 @item %@{|@code{S}:@code{X}@}
4936 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4938 @item %@{|!@code{S}:@code{X}@}
4939 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4941 @item %@{.@code{S}:@code{X}@}
4942 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4944 @item %@{!.@code{S}:@code{X}@}
4945 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4947 @item %@{@code{S}|@code{P}:@code{X}@}
4948 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4949 combined with @samp{!} and @samp{.} sequences as well, although they
4950 have a stronger binding than the @samp{|}. For example a spec string
4954 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4957 will output the following command-line options from the following input
4958 command-line options:
4963 -d fred.c -foo -baz -boggle
4964 -d jim.d -bar -baz -boggle
4969 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4970 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4971 or spaces, or even newlines. They are processed as usual, as described
4974 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4975 switches are handled specifically in these
4976 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4977 @option{-W} switch is found later in the command line, the earlier switch
4978 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4979 letter, which passes all matching options.
4981 The character @samp{|} at the beginning of the predicate text is used to indicate
4982 that a command should be piped to the following command, but only if @option{-pipe}
4985 It is built into GCC which switches take arguments and which do not.
4986 (You might think it would be useful to generalize this to allow each
4987 compiler's spec to say which switches take arguments. But this cannot
4988 be done in a consistent fashion. GCC cannot even decide which input
4989 files have been specified without knowing which switches take arguments,
4990 and it must know which input files to compile in order to tell which
4993 GCC also knows implicitly that arguments starting in @option{-l} are to be
4994 treated as compiler output files, and passed to the linker in their
4995 proper position among the other output files.
4997 @c man begin OPTIONS
4999 @node Target Options
5000 @section Specifying Target Machine and Compiler Version
5001 @cindex target options
5002 @cindex cross compiling
5003 @cindex specifying machine version
5004 @cindex specifying compiler version and target machine
5005 @cindex compiler version, specifying
5006 @cindex target machine, specifying
5008 By default, GCC compiles code for the same type of machine that you
5009 are using. However, it can also be installed as a cross-compiler, to
5010 compile for some other type of machine. In fact, several different
5011 configurations of GCC, for different target machines, can be
5012 installed side by side. Then you specify which one to use with the
5015 In addition, older and newer versions of GCC can be installed side
5016 by side. One of them (probably the newest) will be the default, but
5017 you may sometimes wish to use another.
5020 @item -b @var{machine}
5022 The argument @var{machine} specifies the target machine for compilation.
5023 This is useful when you have installed GCC as a cross-compiler.
5025 The value to use for @var{machine} is the same as was specified as the
5026 machine type when configuring GCC as a cross-compiler. For
5027 example, if a cross-compiler was configured with @samp{configure
5028 i386v}, meaning to compile for an 80386 running System V, then you
5029 would specify @option{-b i386v} to run that cross compiler.
5031 When you do not specify @option{-b}, it normally means to compile for
5032 the same type of machine that you are using.
5034 @item -V @var{version}
5036 The argument @var{version} specifies which version of GCC to run.
5037 This is useful when multiple versions are installed. For example,
5038 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5040 The default version, when you do not specify @option{-V}, is the last
5041 version of GCC that you installed.
5044 The @option{-b} and @option{-V} options actually work by controlling part of
5045 the file name used for the executable files and libraries used for
5046 compilation. A given version of GCC, for a given target machine, is
5047 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.
5049 Thus, sites can customize the effect of @option{-b} or @option{-V} either by
5050 changing the names of these directories or adding alternate names (or
5051 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
5052 file @file{80386} is a link to the file @file{i386v}, then @option{-b
5053 80386} becomes an alias for @option{-b i386v}.
5055 In one respect, the @option{-b} or @option{-V} do not completely change
5056 to a different compiler: the top-level driver program @command{gcc}
5057 that you originally invoked continues to run and invoke the other
5058 executables (preprocessor, compiler per se, assembler and linker)
5059 that do the real work. However, since no real work is done in the
5060 driver program, it usually does not matter that the driver program
5061 in use is not the one for the specified target. It is common for the
5062 interface to the other executables to change incompatibly between
5063 compiler versions, so unless the version specified is very close to that
5064 of the driver (for example, @option{-V 3.0} with a driver program from GCC
5065 version 3.0.1), use of @option{-V} may not work; for example, using
5066 @option{-V 2.95.2} will not work with a driver program from GCC 3.0.
5068 The only way that the driver program depends on the target machine is
5069 in the parsing and handling of special machine-specific options.
5070 However, this is controlled by a file which is found, along with the
5071 other executables, in the directory for the specified version and
5072 target machine. As a result, a single installed driver program adapts
5073 to any specified target machine, and sufficiently similar compiler
5076 The driver program executable does control one significant thing,
5077 however: the default version and target machine. Therefore, you can
5078 install different instances of the driver program, compiled for
5079 different targets or versions, under different names.
5081 For example, if the driver for version 2.0 is installed as @command{ogcc}
5082 and that for version 2.1 is installed as @command{gcc}, then the command
5083 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
5084 2.0 by default. However, you can choose either version with either
5085 command with the @option{-V} option.
5087 @node Submodel Options
5088 @section Hardware Models and Configurations
5089 @cindex submodel options
5090 @cindex specifying hardware config
5091 @cindex hardware models and configurations, specifying
5092 @cindex machine dependent options
5094 Earlier we discussed the standard option @option{-b} which chooses among
5095 different installed compilers for completely different target
5096 machines, such as VAX vs.@: 68000 vs.@: 80386.
5098 In addition, each of these target machine types can have its own
5099 special options, starting with @samp{-m}, to choose among various
5100 hardware models or configurations---for example, 68010 vs 68020,
5101 floating coprocessor or none. A single installed version of the
5102 compiler can compile for any model or configuration, according to the
5105 Some configurations of the compiler also support additional special
5106 options, usually for compatibility with other compilers on the same
5109 These options are defined by the macro @code{TARGET_SWITCHES} in the
5110 machine description. The default for the options is also defined by
5111 that macro, which enables you to change the defaults.
5125 * RS/6000 and PowerPC Options::
5128 * i386 and x86-64 Options::
5130 * Intel 960 Options::
5131 * DEC Alpha Options::
5132 * DEC Alpha/VMS Options::
5136 * System V Options::
5137 * TMS320C3x/C4x Options::
5145 * S/390 and zSeries Options::
5149 * Xstormy16 Options::
5153 @node M680x0 Options
5154 @subsection M680x0 Options
5155 @cindex M680x0 options
5157 These are the @samp{-m} options defined for the 68000 series. The default
5158 values for these options depends on which style of 68000 was selected when
5159 the compiler was configured; the defaults for the most common choices are
5167 Generate output for a 68000. This is the default
5168 when the compiler is configured for 68000-based systems.
5170 Use this option for microcontrollers with a 68000 or EC000 core,
5171 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5177 Generate output for a 68020. This is the default
5178 when the compiler is configured for 68020-based systems.
5182 Generate output containing 68881 instructions for floating point.
5183 This is the default for most 68020 systems unless @option{--nfp} was
5184 specified when the compiler was configured.
5188 Generate output for a 68030. This is the default when the compiler is
5189 configured for 68030-based systems.
5193 Generate output for a 68040. This is the default when the compiler is
5194 configured for 68040-based systems.
5196 This option inhibits the use of 68881/68882 instructions that have to be
5197 emulated by software on the 68040. Use this option if your 68040 does not
5198 have code to emulate those instructions.
5202 Generate output for a 68060. This is the default when the compiler is
5203 configured for 68060-based systems.
5205 This option inhibits the use of 68020 and 68881/68882 instructions that
5206 have to be emulated by software on the 68060. Use this option if your 68060
5207 does not have code to emulate those instructions.
5211 Generate output for a CPU32. This is the default
5212 when the compiler is configured for CPU32-based systems.
5214 Use this option for microcontrollers with a
5215 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5216 68336, 68340, 68341, 68349 and 68360.
5220 Generate output for a 520X ``coldfire'' family cpu. This is the default
5221 when the compiler is configured for 520X-based systems.
5223 Use this option for microcontroller with a 5200 core, including
5224 the MCF5202, MCF5203, MCF5204 and MCF5202.
5229 Generate output for a 68040, without using any of the new instructions.
5230 This results in code which can run relatively efficiently on either a
5231 68020/68881 or a 68030 or a 68040. The generated code does use the
5232 68881 instructions that are emulated on the 68040.
5236 Generate output for a 68060, without using any of the new instructions.
5237 This results in code which can run relatively efficiently on either a
5238 68020/68881 or a 68030 or a 68040. The generated code does use the
5239 68881 instructions that are emulated on the 68060.
5243 Generate output containing Sun FPA instructions for floating point.
5246 @opindex msoft-float
5247 Generate output containing library calls for floating point.
5248 @strong{Warning:} the requisite libraries are not available for all m68k
5249 targets. Normally the facilities of the machine's usual C compiler are
5250 used, but this can't be done directly in cross-compilation. You must
5251 make your own arrangements to provide suitable library functions for
5252 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5253 @samp{m68k-*-coff} do provide software floating point support.
5257 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5260 @opindex mnobitfield
5261 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5262 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5266 Do use the bit-field instructions. The @option{-m68020} option implies
5267 @option{-mbitfield}. This is the default if you use a configuration
5268 designed for a 68020.
5272 Use a different function-calling convention, in which functions
5273 that take a fixed number of arguments return with the @code{rtd}
5274 instruction, which pops their arguments while returning. This
5275 saves one instruction in the caller since there is no need to pop
5276 the arguments there.
5278 This calling convention is incompatible with the one normally
5279 used on Unix, so you cannot use it if you need to call libraries
5280 compiled with the Unix compiler.
5282 Also, you must provide function prototypes for all functions that
5283 take variable numbers of arguments (including @code{printf});
5284 otherwise incorrect code will be generated for calls to those
5287 In addition, seriously incorrect code will result if you call a
5288 function with too many arguments. (Normally, extra arguments are
5289 harmlessly ignored.)
5291 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5292 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5295 @itemx -mno-align-int
5297 @opindex mno-align-int
5298 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5299 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5300 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5301 Aligning variables on 32-bit boundaries produces code that runs somewhat
5302 faster on processors with 32-bit busses at the expense of more memory.
5304 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5305 align structures containing the above types differently than
5306 most published application binary interface specifications for the m68k.
5310 Use the pc-relative addressing mode of the 68000 directly, instead of
5311 using a global offset table. At present, this option implies @option{-fpic},
5312 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5313 not presently supported with @option{-mpcrel}, though this could be supported for
5314 68020 and higher processors.
5316 @item -mno-strict-align
5317 @itemx -mstrict-align
5318 @opindex mno-strict-align
5319 @opindex mstrict-align
5320 Do not (do) assume that unaligned memory references will be handled by
5325 @node M68hc1x Options
5326 @subsection M68hc1x Options
5327 @cindex M68hc1x options
5329 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5330 microcontrollers. The default values for these options depends on
5331 which style of microcontroller was selected when the compiler was configured;
5332 the defaults for the most common choices are given below.
5339 Generate output for a 68HC11. This is the default
5340 when the compiler is configured for 68HC11-based systems.
5346 Generate output for a 68HC12. This is the default
5347 when the compiler is configured for 68HC12-based systems.
5350 @opindex mauto-incdec
5351 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5356 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5358 @item -msoft-reg-count=@var{count}
5359 @opindex msoft-reg-count
5360 Specify the number of pseudo-soft registers which are used for the
5361 code generation. The maximum number is 32. Using more pseudo-soft
5362 register may or may not result in better code depending on the program.
5363 The default is 4 for 68HC11 and 2 for 68HC12.
5368 @subsection VAX Options
5371 These @samp{-m} options are defined for the VAX:
5376 Do not output certain jump instructions (@code{aobleq} and so on)
5377 that the Unix assembler for the VAX cannot handle across long
5382 Do output those jump instructions, on the assumption that you
5383 will assemble with the GNU assembler.
5387 Output code for g-format floating point numbers instead of d-format.
5391 @subsection SPARC Options
5392 @cindex SPARC options
5394 These @samp{-m} switches are supported on the SPARC:
5399 @opindex mno-app-regs
5401 Specify @option{-mapp-regs} to generate output using the global registers
5402 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5405 To be fully SVR4 ABI compliant at the cost of some performance loss,
5406 specify @option{-mno-app-regs}. You should compile libraries and system
5407 software with this option.
5412 @opindex mhard-float
5413 Generate output containing floating point instructions. This is the
5419 @opindex msoft-float
5420 Generate output containing library calls for floating point.
5421 @strong{Warning:} the requisite libraries are not available for all SPARC
5422 targets. Normally the facilities of the machine's usual C compiler are
5423 used, but this cannot be done directly in cross-compilation. You must make
5424 your own arrangements to provide suitable library functions for
5425 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5426 @samp{sparclite-*-*} do provide software floating point support.
5428 @option{-msoft-float} changes the calling convention in the output file;
5429 therefore, it is only useful if you compile @emph{all} of a program with
5430 this option. In particular, you need to compile @file{libgcc.a}, the
5431 library that comes with GCC, with @option{-msoft-float} in order for
5434 @item -mhard-quad-float
5435 @opindex mhard-quad-float
5436 Generate output containing quad-word (long double) floating point
5439 @item -msoft-quad-float
5440 @opindex msoft-quad-float
5441 Generate output containing library calls for quad-word (long double)
5442 floating point instructions. The functions called are those specified
5443 in the SPARC ABI@. This is the default.
5445 As of this writing, there are no sparc implementations that have hardware
5446 support for the quad-word floating point instructions. They all invoke
5447 a trap handler for one of these instructions, and then the trap handler
5448 emulates the effect of the instruction. Because of the trap handler overhead,
5449 this is much slower than calling the ABI library routines. Thus the
5450 @option{-msoft-quad-float} option is the default.
5454 @opindex mno-epilogue
5456 With @option{-mepilogue} (the default), the compiler always emits code for
5457 function exit at the end of each function. Any function exit in
5458 the middle of the function (such as a return statement in C) will
5459 generate a jump to the exit code at the end of the function.
5461 With @option{-mno-epilogue}, the compiler tries to emit exit code inline
5462 at every function exit.
5468 With @option{-mflat}, the compiler does not generate save/restore instructions
5469 and will use a ``flat'' or single register window calling convention.
5470 This model uses %i7 as the frame pointer and is compatible with the normal
5471 register window model. Code from either may be intermixed.
5472 The local registers and the input registers (0--5) are still treated as
5473 ``call saved'' registers and will be saved on the stack as necessary.
5475 With @option{-mno-flat} (the default), the compiler emits save/restore
5476 instructions (except for leaf functions) and is the normal mode of operation.
5478 @item -mno-unaligned-doubles
5479 @itemx -munaligned-doubles
5480 @opindex mno-unaligned-doubles
5481 @opindex munaligned-doubles
5482 Assume that doubles have 8 byte alignment. This is the default.
5484 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5485 alignment only if they are contained in another type, or if they have an
5486 absolute address. Otherwise, it assumes they have 4 byte alignment.
5487 Specifying this option avoids some rare compatibility problems with code
5488 generated by other compilers. It is not the default because it results
5489 in a performance loss, especially for floating point code.
5491 @item -mno-faster-structs
5492 @itemx -mfaster-structs
5493 @opindex mno-faster-structs
5494 @opindex mfaster-structs
5495 With @option{-mfaster-structs}, the compiler assumes that structures
5496 should have 8 byte alignment. This enables the use of pairs of
5497 @code{ldd} and @code{std} instructions for copies in structure
5498 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5499 However, the use of this changed alignment directly violates the Sparc
5500 ABI@. Thus, it's intended only for use on targets where the developer
5501 acknowledges that their resulting code will not be directly in line with
5502 the rules of the ABI@.
5508 These two options select variations on the SPARC architecture.
5510 By default (unless specifically configured for the Fujitsu SPARClite),
5511 GCC generates code for the v7 variant of the SPARC architecture.
5513 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5514 code is that the compiler emits the integer multiply and integer
5515 divide instructions which exist in SPARC v8 but not in SPARC v7.
5517 @option{-msparclite} will give you SPARClite code. This adds the integer
5518 multiply, integer divide step and scan (@code{ffs}) instructions which
5519 exist in SPARClite but not in SPARC v7.
5521 These options are deprecated and will be deleted in a future GCC release.
5522 They have been replaced with @option{-mcpu=xxx}.
5527 @opindex msupersparc
5528 These two options select the processor for which the code is optimized.
5530 With @option{-mcypress} (the default), the compiler optimizes code for the
5531 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5532 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5534 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5535 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5536 of the full SPARC v8 instruction set.
5538 These options are deprecated and will be deleted in a future GCC release.
5539 They have been replaced with @option{-mcpu=xxx}.
5541 @item -mcpu=@var{cpu_type}
5543 Set the instruction set, register set, and instruction scheduling parameters
5544 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5545 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5546 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5547 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5549 Default instruction scheduling parameters are used for values that select
5550 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5551 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5553 Here is a list of each supported architecture and their supported
5558 v8: supersparc, hypersparc
5559 sparclite: f930, f934, sparclite86x
5564 @item -mtune=@var{cpu_type}
5566 Set the instruction scheduling parameters for machine type
5567 @var{cpu_type}, but do not set the instruction set or register set that the
5568 option @option{-mcpu=@var{cpu_type}} would.
5570 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5571 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5572 that select a particular cpu implementation. Those are @samp{cypress},
5573 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5574 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5578 These @samp{-m} switches are supported in addition to the above
5579 on the SPARCLET processor.
5582 @item -mlittle-endian
5583 @opindex mlittle-endian
5584 Generate code for a processor running in little-endian mode.
5588 Treat register @code{%g0} as a normal register.
5589 GCC will continue to clobber it as necessary but will not assume
5590 it always reads as 0.
5592 @item -mbroken-saverestore
5593 @opindex mbroken-saverestore
5594 Generate code that does not use non-trivial forms of the @code{save} and
5595 @code{restore} instructions. Early versions of the SPARCLET processor do
5596 not correctly handle @code{save} and @code{restore} instructions used with
5597 arguments. They correctly handle them used without arguments. A @code{save}
5598 instruction used without arguments increments the current window pointer
5599 but does not allocate a new stack frame. It is assumed that the window
5600 overflow trap handler will properly handle this case as will interrupt
5604 These @samp{-m} switches are supported in addition to the above
5605 on SPARC V9 processors in 64-bit environments.
5608 @item -mlittle-endian
5609 @opindex mlittle-endian
5610 Generate code for a processor running in little-endian mode.
5616 Generate code for a 32-bit or 64-bit environment.
5617 The 32-bit environment sets int, long and pointer to 32 bits.
5618 The 64-bit environment sets int to 32 bits and long and pointer
5621 @item -mcmodel=medlow
5622 @opindex mcmodel=medlow
5623 Generate code for the Medium/Low code model: the program must be linked
5624 in the low 32 bits of the address space. Pointers are 64 bits.
5625 Programs can be statically or dynamically linked.
5627 @item -mcmodel=medmid
5628 @opindex mcmodel=medmid
5629 Generate code for the Medium/Middle code model: the program must be linked
5630 in the low 44 bits of the address space, the text segment must be less than
5631 2G bytes, and data segment must be within 2G of the text segment.
5632 Pointers are 64 bits.
5634 @item -mcmodel=medany
5635 @opindex mcmodel=medany
5636 Generate code for the Medium/Anywhere code model: the program may be linked
5637 anywhere in the address space, the text segment must be less than
5638 2G bytes, and data segment must be within 2G of the text segment.
5639 Pointers are 64 bits.
5641 @item -mcmodel=embmedany
5642 @opindex mcmodel=embmedany
5643 Generate code for the Medium/Anywhere code model for embedded systems:
5644 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5645 (determined at link time). Register %g4 points to the base of the
5646 data segment. Pointers are still 64 bits.
5647 Programs are statically linked, PIC is not supported.
5650 @itemx -mno-stack-bias
5651 @opindex mstack-bias
5652 @opindex mno-stack-bias
5653 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5654 frame pointer if present, are offset by @minus{}2047 which must be added back
5655 when making stack frame references.
5656 Otherwise, assume no such offset is present.
5659 @node Convex Options
5660 @subsection Convex Options
5661 @cindex Convex options
5663 These @samp{-m} options are defined for Convex:
5668 Generate output for C1. The code will run on any Convex machine.
5669 The preprocessor symbol @code{__convex__c1__} is defined.
5673 Generate output for C2. Uses instructions not available on C1.
5674 Scheduling and other optimizations are chosen for max performance on C2.
5675 The preprocessor symbol @code{__convex_c2__} is defined.
5679 Generate output for C32xx. Uses instructions not available on C1.
5680 Scheduling and other optimizations are chosen for max performance on C32.
5681 The preprocessor symbol @code{__convex_c32__} is defined.
5685 Generate output for C34xx. Uses instructions not available on C1.
5686 Scheduling and other optimizations are chosen for max performance on C34.
5687 The preprocessor symbol @code{__convex_c34__} is defined.
5691 Generate output for C38xx. Uses instructions not available on C1.
5692 Scheduling and other optimizations are chosen for max performance on C38.
5693 The preprocessor symbol @code{__convex_c38__} is defined.
5697 Generate code which puts an argument count in the word preceding each
5698 argument list. This is compatible with regular CC, and a few programs
5699 may need the argument count word. GDB and other source-level debuggers
5700 do not need it; this info is in the symbol table.
5703 @opindex mnoargcount
5704 Omit the argument count word. This is the default.
5706 @item -mvolatile-cache
5707 @opindex mvolatile-cache
5708 Allow volatile references to be cached. This is the default.
5710 @item -mvolatile-nocache
5711 @opindex mvolatile-nocache
5712 Volatile references bypass the data cache, going all the way to memory.
5713 This is only needed for multi-processor code that does not use standard
5714 synchronization instructions. Making non-volatile references to volatile
5715 locations will not necessarily work.
5719 Type long is 32 bits, the same as type int. This is the default.
5723 Type long is 64 bits, the same as type long long. This option is useless,
5724 because no library support exists for it.
5727 @node AMD29K Options
5728 @subsection AMD29K Options
5729 @cindex AMD29K options
5731 These @samp{-m} options are defined for the AMD Am29000:
5736 @cindex DW bit (29k)
5737 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5738 halfword operations are directly supported by the hardware. This is the
5743 Generate code that assumes the @code{DW} bit is not set.
5747 @cindex byte writes (29k)
5748 Generate code that assumes the system supports byte and halfword write
5749 operations. This is the default.
5753 Generate code that assumes the systems does not support byte and
5754 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5758 @cindex memory model (29k)
5759 Use a small memory model that assumes that all function addresses are
5760 either within a single 256 KB segment or at an absolute address of less
5761 than 256k. This allows the @code{call} instruction to be used instead
5762 of a @code{const}, @code{consth}, @code{calli} sequence.
5766 Use the normal memory model: Generate @code{call} instructions only when
5767 calling functions in the same file and @code{calli} instructions
5768 otherwise. This works if each file occupies less than 256 KB but allows
5769 the entire executable to be larger than 256 KB@. This is the default.
5773 Always use @code{calli} instructions. Specify this option if you expect
5774 a single file to compile into more than 256 KB of code.
5778 @cindex processor selection (29k)
5779 Generate code for the Am29050.
5783 Generate code for the Am29000. This is the default.
5785 @item -mkernel-registers
5786 @opindex mkernel-registers
5787 @cindex kernel and user registers (29k)
5788 Generate references to registers @code{gr64-gr95} instead of to
5789 registers @code{gr96-gr127}. This option can be used when compiling
5790 kernel code that wants a set of global registers disjoint from that used
5793 Note that when this option is used, register names in @samp{-f} flags
5794 must use the normal, user-mode, names.
5796 @item -muser-registers
5797 @opindex muser-registers
5798 Use the normal set of global registers, @code{gr96-gr127}. This is the
5802 @itemx -mno-stack-check
5803 @opindex mstack-check
5804 @opindex mno-stack-check
5805 @cindex stack checks (29k)
5806 Insert (or do not insert) a call to @code{__msp_check} after each stack
5807 adjustment. This is often used for kernel code.
5810 @itemx -mno-storem-bug
5811 @opindex mstorem-bug
5812 @opindex mno-storem-bug
5813 @cindex storem bug (29k)
5814 @option{-mstorem-bug} handles 29k processors which cannot handle the
5815 separation of a mtsrim insn and a storem instruction (most 29000 chips
5816 to date, but not the 29050).
5818 @item -mno-reuse-arg-regs
5819 @itemx -mreuse-arg-regs
5820 @opindex mno-reuse-arg-regs
5821 @opindex mreuse-arg-regs
5822 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5823 registers for copying out arguments. This helps detect calling a function
5824 with fewer arguments than it was declared with.
5826 @item -mno-impure-text
5827 @itemx -mimpure-text
5828 @opindex mno-impure-text
5829 @opindex mimpure-text
5830 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5831 not pass @option{-assert pure-text} to the linker when linking a shared object.
5834 @opindex msoft-float
5835 Generate output containing library calls for floating point.
5836 @strong{Warning:} the requisite libraries are not part of GCC@.
5837 Normally the facilities of the machine's usual C compiler are used, but
5838 this can't be done directly in cross-compilation. You must make your
5839 own arrangements to provide suitable library functions for
5844 Do not generate multm or multmu instructions. This is useful for some embedded
5845 systems which do not have trap handlers for these instructions.
5849 @subsection ARM Options
5852 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5857 @opindex mapcs-frame
5858 Generate a stack frame that is compliant with the ARM Procedure Call
5859 Standard for all functions, even if this is not strictly necessary for
5860 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5861 with this option will cause the stack frames not to be generated for
5862 leaf functions. The default is @option{-mno-apcs-frame}.
5866 This is a synonym for @option{-mapcs-frame}.
5870 Generate code for a processor running with a 26-bit program counter,
5871 and conforming to the function calling standards for the APCS 26-bit
5872 option. This option replaces the @option{-m2} and @option{-m3} options
5873 of previous releases of the compiler.
5877 Generate code for a processor running with a 32-bit program counter,
5878 and conforming to the function calling standards for the APCS 32-bit
5879 option. This option replaces the @option{-m6} option of previous releases
5883 @c not currently implemented
5884 @item -mapcs-stack-check
5885 @opindex mapcs-stack-check
5886 Generate code to check the amount of stack space available upon entry to
5887 every function (that actually uses some stack space). If there is
5888 insufficient space available then either the function
5889 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5890 called, depending upon the amount of stack space required. The run time
5891 system is required to provide these functions. The default is
5892 @option{-mno-apcs-stack-check}, since this produces smaller code.
5894 @c not currently implemented
5896 @opindex mapcs-float
5897 Pass floating point arguments using the float point registers. This is
5898 one of the variants of the APCS@. This option is recommended if the
5899 target hardware has a floating point unit or if a lot of floating point
5900 arithmetic is going to be performed by the code. The default is
5901 @option{-mno-apcs-float}, since integer only code is slightly increased in
5902 size if @option{-mapcs-float} is used.
5904 @c not currently implemented
5905 @item -mapcs-reentrant
5906 @opindex mapcs-reentrant
5907 Generate reentrant, position independent code. The default is
5908 @option{-mno-apcs-reentrant}.
5911 @item -mthumb-interwork
5912 @opindex mthumb-interwork
5913 Generate code which supports calling between the ARM and Thumb
5914 instruction sets. Without this option the two instruction sets cannot
5915 be reliably used inside one program. The default is
5916 @option{-mno-thumb-interwork}, since slightly larger code is generated
5917 when @option{-mthumb-interwork} is specified.
5919 @item -mno-sched-prolog
5920 @opindex mno-sched-prolog
5921 Prevent the reordering of instructions in the function prolog, or the
5922 merging of those instruction with the instructions in the function's
5923 body. This means that all functions will start with a recognizable set
5924 of instructions (or in fact one of a choice from a small set of
5925 different function prologues), and this information can be used to
5926 locate the start if functions inside an executable piece of code. The
5927 default is @option{-msched-prolog}.
5930 @opindex mhard-float
5931 Generate output containing floating point instructions. This is the
5935 @opindex msoft-float
5936 Generate output containing library calls for floating point.
5937 @strong{Warning:} the requisite libraries are not available for all ARM
5938 targets. Normally the facilities of the machine's usual C compiler are
5939 used, but this cannot be done directly in cross-compilation. You must make
5940 your own arrangements to provide suitable library functions for
5943 @option{-msoft-float} changes the calling convention in the output file;
5944 therefore, it is only useful if you compile @emph{all} of a program with
5945 this option. In particular, you need to compile @file{libgcc.a}, the
5946 library that comes with GCC, with @option{-msoft-float} in order for
5949 @item -mlittle-endian
5950 @opindex mlittle-endian
5951 Generate code for a processor running in little-endian mode. This is
5952 the default for all standard configurations.
5955 @opindex mbig-endian
5956 Generate code for a processor running in big-endian mode; the default is
5957 to compile code for a little-endian processor.
5959 @item -mwords-little-endian
5960 @opindex mwords-little-endian
5961 This option only applies when generating code for big-endian processors.
5962 Generate code for a little-endian word order but a big-endian byte
5963 order. That is, a byte order of the form @samp{32107654}. Note: this
5964 option should only be used if you require compatibility with code for
5965 big-endian ARM processors generated by versions of the compiler prior to
5968 @item -malignment-traps
5969 @opindex malignment-traps
5970 Generate code that will not trap if the MMU has alignment traps enabled.
5971 On ARM architectures prior to ARMv4, there were no instructions to
5972 access half-word objects stored in memory. However, when reading from
5973 memory a feature of the ARM architecture allows a word load to be used,
5974 even if the address is unaligned, and the processor core will rotate the
5975 data as it is being loaded. This option tells the compiler that such
5976 misaligned accesses will cause a MMU trap and that it should instead
5977 synthesise the access as a series of byte accesses. The compiler can
5978 still use word accesses to load half-word data if it knows that the
5979 address is aligned to a word boundary.
5981 This option is ignored when compiling for ARM architecture 4 or later,
5982 since these processors have instructions to directly access half-word
5985 @item -mno-alignment-traps
5986 @opindex mno-alignment-traps
5987 Generate code that assumes that the MMU will not trap unaligned
5988 accesses. This produces better code when the target instruction set
5989 does not have half-word memory operations (i.e.@: implementations prior to
5992 Note that you cannot use this option to access unaligned word objects,
5993 since the processor will only fetch one 32-bit aligned object from
5996 The default setting for most targets is @option{-mno-alignment-traps}, since
5997 this produces better code when there are no half-word memory
5998 instructions available.
6000 @item -mshort-load-bytes
6001 @itemx -mno-short-load-words
6002 @opindex mshort-load-bytes
6003 @opindex mno-short-load-words
6004 These are deprecated aliases for @option{-malignment-traps}.
6006 @item -mno-short-load-bytes
6007 @itemx -mshort-load-words
6008 @opindex mno-short-load-bytes
6009 @opindex mshort-load-words
6010 This are deprecated aliases for @option{-mno-alignment-traps}.
6014 This option only applies to RISC iX@. Emulate the native BSD-mode
6015 compiler. This is the default if @option{-ansi} is not specified.
6019 This option only applies to RISC iX@. Emulate the native X/Open-mode
6022 @item -mno-symrename
6023 @opindex mno-symrename
6024 This option only applies to RISC iX@. Do not run the assembler
6025 post-processor, @samp{symrename}, after code has been assembled.
6026 Normally it is necessary to modify some of the standard symbols in
6027 preparation for linking with the RISC iX C library; this option
6028 suppresses this pass. The post-processor is never run when the
6029 compiler is built for cross-compilation.
6031 @item -mcpu=@var{name}
6033 This specifies the name of the target ARM processor. GCC uses this name
6034 to determine what kind of instructions it can emit when generating
6035 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6036 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6037 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6038 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6039 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6040 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6041 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6042 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6043 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
6044 @samp{arm1020t}, @samp{xscale}.
6046 @itemx -mtune=@var{name}
6048 This option is very similar to the @option{-mcpu=} option, except that
6049 instead of specifying the actual target processor type, and hence
6050 restricting which instructions can be used, it specifies that GCC should
6051 tune the performance of the code as if the target were of the type
6052 specified in this option, but still choosing the instructions that it
6053 will generate based on the cpu specified by a @option{-mcpu=} option.
6054 For some ARM implementations better performance can be obtained by using
6057 @item -march=@var{name}
6059 This specifies the name of the target ARM architecture. GCC uses this
6060 name to determine what kind of instructions it can emit when generating
6061 assembly code. This option can be used in conjunction with or instead
6062 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6063 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6064 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
6066 @item -mfpe=@var{number}
6067 @itemx -mfp=@var{number}
6070 This specifies the version of the floating point emulation available on
6071 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6072 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6074 @item -mstructure-size-boundary=@var{n}
6075 @opindex mstructure-size-boundary
6076 The size of all structures and unions will be rounded up to a multiple
6077 of the number of bits set by this option. Permissible values are 8 and
6078 32. The default value varies for different toolchains. For the COFF
6079 targeted toolchain the default value is 8. Specifying the larger number
6080 can produce faster, more efficient code, but can also increase the size
6081 of the program. The two values are potentially incompatible. Code
6082 compiled with one value cannot necessarily expect to work with code or
6083 libraries compiled with the other value, if they exchange information
6084 using structures or unions.
6086 @item -mabort-on-noreturn
6087 @opindex mabort-on-noreturn
6088 Generate a call to the function @code{abort} at the end of a
6089 @code{noreturn} function. It will be executed if the function tries to
6093 @itemx -mno-long-calls
6094 @opindex mlong-calls
6095 @opindex mno-long-calls
6096 Tells the compiler to perform function calls by first loading the
6097 address of the function into a register and then performing a subroutine
6098 call on this register. This switch is needed if the target function
6099 will lie outside of the 64 megabyte addressing range of the offset based
6100 version of subroutine call instruction.
6102 Even if this switch is enabled, not all function calls will be turned
6103 into long calls. The heuristic is that static functions, functions
6104 which have the @samp{short-call} attribute, functions that are inside
6105 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6106 definitions have already been compiled within the current compilation
6107 unit, will not be turned into long calls. The exception to this rule is
6108 that weak function definitions, functions with the @samp{long-call}
6109 attribute or the @samp{section} attribute, and functions that are within
6110 the scope of a @samp{#pragma long_calls} directive, will always be
6111 turned into long calls.
6113 This feature is not enabled by default. Specifying
6114 @option{-mno-long-calls} will restore the default behavior, as will
6115 placing the function calls within the scope of a @samp{#pragma
6116 long_calls_off} directive. Note these switches have no effect on how
6117 the compiler generates code to handle function calls via function
6120 @item -mnop-fun-dllimport
6121 @opindex mnop-fun-dllimport
6122 Disable support for the @code{dllimport} attribute.
6124 @item -msingle-pic-base
6125 @opindex msingle-pic-base
6126 Treat the register used for PIC addressing as read-only, rather than
6127 loading it in the prologue for each function. The run-time system is
6128 responsible for initializing this register with an appropriate value
6129 before execution begins.
6131 @item -mpic-register=@var{reg}
6132 @opindex mpic-register
6133 Specify the register to be used for PIC addressing. The default is R10
6134 unless stack-checking is enabled, when R9 is used.
6136 @item -mpoke-function-name
6137 @opindex mpoke-function-name
6138 Write the name of each function into the text section, directly
6139 preceding the function prologue. The generated code is similar to this:
6143 .ascii "arm_poke_function_name", 0
6146 .word 0xff000000 + (t1 - t0)
6147 arm_poke_function_name
6149 stmfd sp!, @{fp, ip, lr, pc@}
6153 When performing a stack backtrace, code can inspect the value of
6154 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6155 location @code{pc - 12} and the top 8 bits are set, then we know that
6156 there is a function name embedded immediately preceding this location
6157 and has length @code{((pc[-3]) & 0xff000000)}.
6161 Generate code for the 16-bit Thumb instruction set. The default is to
6162 use the 32-bit ARM instruction set.
6165 @opindex mtpcs-frame
6166 Generate a stack frame that is compliant with the Thumb Procedure Call
6167 Standard for all non-leaf functions. (A leaf function is one that does
6168 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6170 @item -mtpcs-leaf-frame
6171 @opindex mtpcs-leaf-frame
6172 Generate a stack frame that is compliant with the Thumb Procedure Call
6173 Standard for all leaf functions. (A leaf function is one that does
6174 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6176 @item -mcallee-super-interworking
6177 @opindex mcallee-super-interworking
6178 Gives all externally visible functions in the file being compiled an ARM
6179 instruction set header which switches to Thumb mode before executing the
6180 rest of the function. This allows these functions to be called from
6181 non-interworking code.
6183 @item -mcaller-super-interworking
6184 @opindex mcaller-super-interworking
6185 Allows calls via function pointers (including virtual functions) to
6186 execute correctly regardless of whether the target code has been
6187 compiled for interworking or not. There is a small overhead in the cost
6188 of executing a function pointer if this option is enabled.
6192 @node MN10200 Options
6193 @subsection MN10200 Options
6194 @cindex MN10200 options
6195 These @option{-m} options are defined for Matsushita MN10200 architectures:
6200 Indicate to the linker that it should perform a relaxation optimization pass
6201 to shorten branches, calls and absolute memory addresses. This option only
6202 has an effect when used on the command line for the final link step.
6204 This option makes symbolic debugging impossible.
6207 @node MN10300 Options
6208 @subsection MN10300 Options
6209 @cindex MN10300 options
6210 These @option{-m} options are defined for Matsushita MN10300 architectures:
6215 Generate code to avoid bugs in the multiply instructions for the MN10300
6216 processors. This is the default.
6219 @opindex mno-mult-bug
6220 Do not generate code to avoid bugs in the multiply instructions for the
6225 Generate code which uses features specific to the AM33 processor.
6229 Do not generate code which uses features specific to the AM33 processor. This
6234 Do not link in the C run-time initialization object file.
6238 Indicate to the linker that it should perform a relaxation optimization pass
6239 to shorten branches, calls and absolute memory addresses. This option only
6240 has an effect when used on the command line for the final link step.
6242 This option makes symbolic debugging impossible.
6246 @node M32R/D Options
6247 @subsection M32R/D Options
6248 @cindex M32R/D options
6250 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6255 Generate code for the M32R/X@.
6259 Generate code for the M32R@. This is the default.
6261 @item -mcode-model=small
6262 @opindex mcode-model=small
6263 Assume all objects live in the lower 16MB of memory (so that their addresses
6264 can be loaded with the @code{ld24} instruction), and assume all subroutines
6265 are reachable with the @code{bl} instruction.
6266 This is the default.
6268 The addressability of a particular object can be set with the
6269 @code{model} attribute.
6271 @item -mcode-model=medium
6272 @opindex mcode-model=medium
6273 Assume objects may be anywhere in the 32-bit address space (the compiler
6274 will generate @code{seth/add3} instructions to load their addresses), and
6275 assume all subroutines are reachable with the @code{bl} instruction.
6277 @item -mcode-model=large
6278 @opindex mcode-model=large
6279 Assume objects may be anywhere in the 32-bit address space (the compiler
6280 will generate @code{seth/add3} instructions to load their addresses), and
6281 assume subroutines may not be reachable with the @code{bl} instruction
6282 (the compiler will generate the much slower @code{seth/add3/jl}
6283 instruction sequence).
6286 @opindex msdata=none
6287 Disable use of the small data area. Variables will be put into
6288 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6289 @code{section} attribute has been specified).
6290 This is the default.
6292 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6293 Objects may be explicitly put in the small data area with the
6294 @code{section} attribute using one of these sections.
6297 @opindex msdata=sdata
6298 Put small global and static data in the small data area, but do not
6299 generate special code to reference them.
6303 Put small global and static data in the small data area, and generate
6304 special instructions to reference them.
6308 @cindex smaller data references
6309 Put global and static objects less than or equal to @var{num} bytes
6310 into the small data or bss sections instead of the normal data or bss
6311 sections. The default value of @var{num} is 8.
6312 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6313 for this option to have any effect.
6315 All modules should be compiled with the same @option{-G @var{num}} value.
6316 Compiling with different values of @var{num} may or may not work; if it
6317 doesn't the linker will give an error message---incorrect code will not be
6323 @subsection M88K Options
6324 @cindex M88k options
6326 These @samp{-m} options are defined for Motorola 88k architectures:
6331 Generate code that works well on both the m88100 and the
6336 Generate code that works best for the m88100, but that also
6341 Generate code that works best for the m88110, and may not run
6346 Obsolete option to be removed from the next revision.
6349 @item -midentify-revision
6350 @opindex midentify-revision
6351 @cindex identifying source, compiler (88k)
6352 Include an @code{ident} directive in the assembler output recording the
6353 source file name, compiler name and version, timestamp, and compilation
6356 @item -mno-underscores
6357 @opindex mno-underscores
6358 @cindex underscores, avoiding (88k)
6359 In assembler output, emit symbol names without adding an underscore
6360 character at the beginning of each name. The default is to use an
6361 underscore as prefix on each name.
6363 @item -mocs-debug-info
6364 @itemx -mno-ocs-debug-info
6365 @opindex mocs-debug-info
6366 @opindex mno-ocs-debug-info
6368 @cindex debugging, 88k OCS
6369 Include (or omit) additional debugging information (about registers used
6370 in each stack frame) as specified in the 88open Object Compatibility
6371 Standard, ``OCS''@. This extra information allows debugging of code that
6372 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6373 Delta 88 SVr3.2 is to include this information; other 88k configurations
6374 omit this information by default.
6376 @item -mocs-frame-position
6377 @opindex mocs-frame-position
6378 @cindex register positions in frame (88k)
6379 When emitting COFF debugging information for automatic variables and
6380 parameters stored on the stack, use the offset from the canonical frame
6381 address, which is the stack pointer (register 31) on entry to the
6382 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6383 @option{-mocs-frame-position}; other 88k configurations have the default
6384 @option{-mno-ocs-frame-position}.
6386 @item -mno-ocs-frame-position
6387 @opindex mno-ocs-frame-position
6388 @cindex register positions in frame (88k)
6389 When emitting COFF debugging information for automatic variables and
6390 parameters stored on the stack, use the offset from the frame pointer
6391 register (register 30). When this option is in effect, the frame
6392 pointer is not eliminated when debugging information is selected by the
6395 @item -moptimize-arg-area
6396 @opindex moptimize-arg-area
6397 @cindex arguments in frame (88k)
6398 Save space by reorganizing the stack frame. This option generates code
6399 that does not agree with the 88open specifications, but uses less
6402 @itemx -mno-optimize-arg-area
6403 @opindex mno-optimize-arg-area
6404 Do not reorganize the stack frame to save space. This is the default.
6405 The generated conforms to the specification, but uses more memory.
6407 @item -mshort-data-@var{num}
6408 @opindex mshort-data
6409 @cindex smaller data references (88k)
6410 @cindex r0-relative references (88k)
6411 Generate smaller data references by making them relative to @code{r0},
6412 which allows loading a value using a single instruction (rather than the
6413 usual two). You control which data references are affected by
6414 specifying @var{num} with this option. For example, if you specify
6415 @option{-mshort-data-512}, then the data references affected are those
6416 involving displacements of less than 512 bytes.
6417 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6420 @item -mserialize-volatile
6421 @opindex mserialize-volatile
6422 @itemx -mno-serialize-volatile
6423 @opindex mno-serialize-volatile
6424 @cindex sequential consistency on 88k
6425 Do, or don't, generate code to guarantee sequential consistency
6426 of volatile memory references. By default, consistency is
6429 The order of memory references made by the MC88110 processor does
6430 not always match the order of the instructions requesting those
6431 references. In particular, a load instruction may execute before
6432 a preceding store instruction. Such reordering violates
6433 sequential consistency of volatile memory references, when there
6434 are multiple processors. When consistency must be guaranteed,
6435 GCC generates special instructions, as needed, to force
6436 execution in the proper order.
6438 The MC88100 processor does not reorder memory references and so
6439 always provides sequential consistency. However, by default, GCC
6440 generates the special instructions to guarantee consistency
6441 even when you use @option{-m88100}, so that the code may be run on an
6442 MC88110 processor. If you intend to run your code only on the
6443 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6445 The extra code generated to guarantee consistency may affect the
6446 performance of your application. If you know that you can safely
6447 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6453 @cindex assembler syntax, 88k
6455 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6456 related to System V release 4 (SVr4). This controls the following:
6460 Which variant of the assembler syntax to emit.
6462 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6463 that is used on System V release 4.
6465 @option{-msvr4} makes GCC issue additional declaration directives used in
6469 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6470 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6471 other m88k configurations.
6473 @item -mversion-03.00
6474 @opindex mversion-03.00
6475 This option is obsolete, and is ignored.
6476 @c ??? which asm syntax better for GAS? option there too?
6478 @item -mno-check-zero-division
6479 @itemx -mcheck-zero-division
6480 @opindex mno-check-zero-division
6481 @opindex mcheck-zero-division
6482 @cindex zero division on 88k
6483 Do, or don't, generate code to guarantee that integer division by
6484 zero will be detected. By default, detection is guaranteed.
6486 Some models of the MC88100 processor fail to trap upon integer
6487 division by zero under certain conditions. By default, when
6488 compiling code that might be run on such a processor, GCC
6489 generates code that explicitly checks for zero-valued divisors
6490 and traps with exception number 503 when one is detected. Use of
6491 @option{-mno-check-zero-division} suppresses such checking for code
6492 generated to run on an MC88100 processor.
6494 GCC assumes that the MC88110 processor correctly detects all instances
6495 of integer division by zero. When @option{-m88110} is specified, no
6496 explicit checks for zero-valued divisors are generated, and both
6497 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6500 @item -muse-div-instruction
6501 @opindex muse-div-instruction
6502 @cindex divide instruction, 88k
6503 Use the div instruction for signed integer division on the
6504 MC88100 processor. By default, the div instruction is not used.
6506 On the MC88100 processor the signed integer division instruction
6507 div) traps to the operating system on a negative operand. The
6508 operating system transparently completes the operation, but at a
6509 large cost in execution time. By default, when compiling code
6510 that might be run on an MC88100 processor, GCC emulates signed
6511 integer division using the unsigned integer division instruction
6512 divu), thereby avoiding the large penalty of a trap to the
6513 operating system. Such emulation has its own, smaller, execution
6514 cost in both time and space. To the extent that your code's
6515 important signed integer division operations are performed on two
6516 nonnegative operands, it may be desirable to use the div
6517 instruction directly.
6519 On the MC88110 processor the div instruction (also known as the
6520 divs instruction) processes negative operands without trapping to
6521 the operating system. When @option{-m88110} is specified,
6522 @option{-muse-div-instruction} is ignored, and the div instruction is used
6523 for signed integer division.
6525 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6526 particular, the behavior of such a division with and without
6527 @option{-muse-div-instruction} may differ.
6529 @item -mtrap-large-shift
6530 @itemx -mhandle-large-shift
6531 @opindex mtrap-large-shift
6532 @opindex mhandle-large-shift
6533 @cindex bit shift overflow (88k)
6534 @cindex large bit shifts (88k)
6535 Include code to detect bit-shifts of more than 31 bits; respectively,
6536 trap such shifts or emit code to handle them properly. By default GCC
6537 makes no special provision for large bit shifts.
6539 @item -mwarn-passed-structs
6540 @opindex mwarn-passed-structs
6541 @cindex structure passing (88k)
6542 Warn when a function passes a struct as an argument or result.
6543 Structure-passing conventions have changed during the evolution of the C
6544 language, and are often the source of portability problems. By default,
6545 GCC issues no such warning.
6548 @c break page here to avoid unsightly interparagraph stretch.
6552 @node RS/6000 and PowerPC Options
6553 @subsection IBM RS/6000 and PowerPC Options
6554 @cindex RS/6000 and PowerPC Options
6555 @cindex IBM RS/6000 and PowerPC Options
6557 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6565 @itemx -mpowerpc-gpopt
6566 @itemx -mno-powerpc-gpopt
6567 @itemx -mpowerpc-gfxopt
6568 @itemx -mno-powerpc-gfxopt
6570 @itemx -mno-powerpc64
6576 @opindex mno-powerpc
6577 @opindex mpowerpc-gpopt
6578 @opindex mno-powerpc-gpopt
6579 @opindex mpowerpc-gfxopt
6580 @opindex mno-powerpc-gfxopt
6582 @opindex mno-powerpc64
6583 GCC supports two related instruction set architectures for the
6584 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6585 instructions supported by the @samp{rios} chip set used in the original
6586 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6587 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6588 the IBM 4xx microprocessors.
6590 Neither architecture is a subset of the other. However there is a
6591 large common subset of instructions supported by both. An MQ
6592 register is included in processors supporting the POWER architecture.
6594 You use these options to specify which instructions are available on the
6595 processor you are using. The default value of these options is
6596 determined when configuring GCC@. Specifying the
6597 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6598 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6599 rather than the options listed above.
6601 The @option{-mpower} option allows GCC to generate instructions that
6602 are found only in the POWER architecture and to use the MQ register.
6603 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6604 to generate instructions that are present in the POWER2 architecture but
6605 not the original POWER architecture.
6607 The @option{-mpowerpc} option allows GCC to generate instructions that
6608 are found only in the 32-bit subset of the PowerPC architecture.
6609 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6610 GCC to use the optional PowerPC architecture instructions in the
6611 General Purpose group, including floating-point square root. Specifying
6612 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6613 use the optional PowerPC architecture instructions in the Graphics
6614 group, including floating-point select.
6616 The @option{-mpowerpc64} option allows GCC to generate the additional
6617 64-bit instructions that are found in the full PowerPC64 architecture
6618 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6619 @option{-mno-powerpc64}.
6621 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6622 will use only the instructions in the common subset of both
6623 architectures plus some special AIX common-mode calls, and will not use
6624 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6625 permits GCC to use any instruction from either architecture and to
6626 allow use of the MQ register; specify this for the Motorola MPC601.
6628 @item -mnew-mnemonics
6629 @itemx -mold-mnemonics
6630 @opindex mnew-mnemonics
6631 @opindex mold-mnemonics
6632 Select which mnemonics to use in the generated assembler code. With
6633 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6634 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6635 assembler mnemonics defined for the POWER architecture. Instructions
6636 defined in only one architecture have only one mnemonic; GCC uses that
6637 mnemonic irrespective of which of these options is specified.
6639 GCC defaults to the mnemonics appropriate for the architecture in
6640 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6641 value of these option. Unless you are building a cross-compiler, you
6642 should normally not specify either @option{-mnew-mnemonics} or
6643 @option{-mold-mnemonics}, but should instead accept the default.
6645 @item -mcpu=@var{cpu_type}
6647 Set architecture type, register usage, choice of mnemonics, and
6648 instruction scheduling parameters for machine type @var{cpu_type}.
6649 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6650 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6651 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6652 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6653 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6654 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6656 @option{-mcpu=common} selects a completely generic processor. Code
6657 generated under this option will run on any POWER or PowerPC processor.
6658 GCC will use only the instructions in the common subset of both
6659 architectures, and will not use the MQ register. GCC assumes a generic
6660 processor model for scheduling purposes.
6662 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6663 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6664 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6665 types, with an appropriate, generic processor model assumed for
6666 scheduling purposes.
6668 The other options specify a specific processor. Code generated under
6669 those options will run best on that processor, and may not run at all on
6672 The @option{-mcpu} options automatically enable or disable other
6673 @option{-m} options as follows:
6677 @option{-mno-power}, @option{-mno-powerc}
6684 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6699 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6702 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6707 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6710 @item -mtune=@var{cpu_type}
6712 Set the instruction scheduling parameters for machine type
6713 @var{cpu_type}, but do not set the architecture type, register usage, or
6714 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6715 values for @var{cpu_type} are used for @option{-mtune} as for
6716 @option{-mcpu}. If both are specified, the code generated will use the
6717 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6718 scheduling parameters set by @option{-mtune}.
6723 @opindex mno-altivec
6724 These switches enable or disable the use of built-in functions that
6725 allow access to the AltiVec instruction set. You may also need to set
6726 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6730 @itemx -mno-fp-in-toc
6731 @itemx -mno-sum-in-toc
6732 @itemx -mminimal-toc
6734 @opindex mno-fp-in-toc
6735 @opindex mno-sum-in-toc
6736 @opindex mminimal-toc
6737 Modify generation of the TOC (Table Of Contents), which is created for
6738 every executable file. The @option{-mfull-toc} option is selected by
6739 default. In that case, GCC will allocate at least one TOC entry for
6740 each unique non-automatic variable reference in your program. GCC
6741 will also place floating-point constants in the TOC@. However, only
6742 16,384 entries are available in the TOC@.
6744 If you receive a linker error message that saying you have overflowed
6745 the available TOC space, you can reduce the amount of TOC space used
6746 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6747 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6748 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6749 generate code to calculate the sum of an address and a constant at
6750 run-time instead of putting that sum into the TOC@. You may specify one
6751 or both of these options. Each causes GCC to produce very slightly
6752 slower and larger code at the expense of conserving TOC space.
6754 If you still run out of space in the TOC even when you specify both of
6755 these options, specify @option{-mminimal-toc} instead. This option causes
6756 GCC to make only one TOC entry for every file. When you specify this
6757 option, GCC will produce code that is slower and larger but which
6758 uses extremely little TOC space. You may wish to use this option
6759 only on files that contain less frequently executed code.
6765 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6766 @code{long} type, and the infrastructure needed to support them.
6767 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6768 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6769 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6774 @opindex mno-xl-call
6775 On AIX, pass floating-point arguments to prototyped functions beyond the
6776 register save area (RSA) on the stack in addition to argument FPRs. The
6777 AIX calling convention was extended but not initially documented to
6778 handle an obscure K&R C case of calling a function that takes the
6779 address of its arguments with fewer arguments than declared. AIX XL
6780 compilers access floating point arguments which do not fit in the
6781 RSA from the stack when a subroutine is compiled without
6782 optimization. Because always storing floating-point arguments on the
6783 stack is inefficient and rarely needed, this option is not enabled by
6784 default and only is necessary when calling subroutines compiled by AIX
6785 XL compilers without optimization.
6789 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6790 application written to use message passing with special startup code to
6791 enable the application to run. The system must have PE installed in the
6792 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6793 must be overridden with the @option{-specs=} option to specify the
6794 appropriate directory location. The Parallel Environment does not
6795 support threads, so the @option{-mpe} option and the @option{-pthread}
6796 option are incompatible.
6800 @opindex msoft-float
6801 @opindex mhard-float
6802 Generate code that does not use (uses) the floating-point register set.
6803 Software floating point emulation is provided if you use the
6804 @option{-msoft-float} option, and pass the option to GCC when linking.
6807 @itemx -mno-multiple
6809 @opindex mno-multiple
6810 Generate code that uses (does not use) the load multiple word
6811 instructions and the store multiple word instructions. These
6812 instructions are generated by default on POWER systems, and not
6813 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6814 endian PowerPC systems, since those instructions do not work when the
6815 processor is in little endian mode. The exceptions are PPC740 and
6816 PPC750 which permit the instructions usage in little endian mode.
6822 Generate code that uses (does not use) the load string instructions
6823 and the store string word instructions to save multiple registers and
6824 do small block moves. These instructions are generated by default on
6825 POWER systems, and not generated on PowerPC systems. Do not use
6826 @option{-mstring} on little endian PowerPC systems, since those
6827 instructions do not work when the processor is in little endian mode.
6828 The exceptions are PPC740 and PPC750 which permit the instructions
6829 usage in little endian mode.
6835 Generate code that uses (does not use) the load or store instructions
6836 that update the base register to the address of the calculated memory
6837 location. These instructions are generated by default. If you use
6838 @option{-mno-update}, there is a small window between the time that the
6839 stack pointer is updated and the address of the previous frame is
6840 stored, which means code that walks the stack frame across interrupts or
6841 signals may get corrupted data.
6844 @itemx -mno-fused-madd
6845 @opindex mfused-madd
6846 @opindex mno-fused-madd
6847 Generate code that uses (does not use) the floating point multiply and
6848 accumulate instructions. These instructions are generated by default if
6849 hardware floating is used.
6851 @item -mno-bit-align
6853 @opindex mno-bit-align
6855 On System V.4 and embedded PowerPC systems do not (do) force structures
6856 and unions that contain bit-fields to be aligned to the base type of the
6859 For example, by default a structure containing nothing but 8
6860 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6861 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6862 the structure would be aligned to a 1 byte boundary and be one byte in
6865 @item -mno-strict-align
6866 @itemx -mstrict-align
6867 @opindex mno-strict-align
6868 @opindex mstrict-align
6869 On System V.4 and embedded PowerPC systems do not (do) assume that
6870 unaligned memory references will be handled by the system.
6873 @itemx -mno-relocatable
6874 @opindex mrelocatable
6875 @opindex mno-relocatable
6876 On embedded PowerPC systems generate code that allows (does not allow)
6877 the program to be relocated to a different address at runtime. If you
6878 use @option{-mrelocatable} on any module, all objects linked together must
6879 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6881 @item -mrelocatable-lib
6882 @itemx -mno-relocatable-lib
6883 @opindex mrelocatable-lib
6884 @opindex mno-relocatable-lib
6885 On embedded PowerPC systems generate code that allows (does not allow)
6886 the program to be relocated to a different address at runtime. Modules
6887 compiled with @option{-mrelocatable-lib} can be linked with either modules
6888 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6889 with modules compiled with the @option{-mrelocatable} options.
6895 On System V.4 and embedded PowerPC systems do not (do) assume that
6896 register 2 contains a pointer to a global area pointing to the addresses
6897 used in the program.
6900 @itemx -mlittle-endian
6902 @opindex mlittle-endian
6903 On System V.4 and embedded PowerPC systems compile code for the
6904 processor in little endian mode. The @option{-mlittle-endian} option is
6905 the same as @option{-mlittle}.
6910 @opindex mbig-endian
6911 On System V.4 and embedded PowerPC systems compile code for the
6912 processor in big endian mode. The @option{-mbig-endian} option is
6913 the same as @option{-mbig}.
6917 On System V.4 and embedded PowerPC systems compile code using calling
6918 conventions that adheres to the March 1995 draft of the System V
6919 Application Binary Interface, PowerPC processor supplement. This is the
6920 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6922 @item -mcall-sysv-eabi
6923 @opindex mcall-sysv-eabi
6924 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6926 @item -mcall-sysv-noeabi
6927 @opindex mcall-sysv-noeabi
6928 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6932 On System V.4 and embedded PowerPC systems compile code using calling
6933 conventions that are similar to those used on AIX@. This is the
6934 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6936 @item -mcall-solaris
6937 @opindex mcall-solaris
6938 On System V.4 and embedded PowerPC systems compile code for the Solaris
6942 @opindex mcall-linux
6943 On System V.4 and embedded PowerPC systems compile code for the
6944 Linux-based GNU system.
6948 On System V.4 and embedded PowerPC systems compile code for the
6949 Hurd-based GNU system.
6952 @opindex mcall-netbsd
6953 On System V.4 and embedded PowerPC systems compile code for the
6954 NetBSD operating system.
6956 @item -maix-struct-return
6957 @opindex maix-struct-return
6958 Return all structures in memory (as specified by the AIX ABI)@.
6960 @item -msvr4-struct-return
6961 @opindex msvr4-struct-return
6962 Return structures smaller than 8 bytes in registers (as specified by the
6966 @opindex mabi=altivec
6967 Extend the current ABI with AltiVec ABI extensions. This does not
6968 change the default ABI, instead it adds the AltiVec ABI extensions to
6972 @itemx -mno-prototype
6974 @opindex mno-prototype
6975 On System V.4 and embedded PowerPC systems assume that all calls to
6976 variable argument functions are properly prototyped. Otherwise, the
6977 compiler must insert an instruction before every non prototyped call to
6978 set or clear bit 6 of the condition code register (@var{CR}) to
6979 indicate whether floating point values were passed in the floating point
6980 registers in case the function takes a variable arguments. With
6981 @option{-mprototype}, only calls to prototyped variable argument functions
6982 will set or clear the bit.
6986 On embedded PowerPC systems, assume that the startup module is called
6987 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6988 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6993 On embedded PowerPC systems, assume that the startup module is called
6994 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6999 On embedded PowerPC systems, assume that the startup module is called
7000 @file{crt0.o} and the standard C libraries are @file{libads.a} and
7004 @opindex myellowknife
7005 On embedded PowerPC systems, assume that the startup module is called
7006 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
7011 On System V.4 and embedded PowerPC systems, specify that you are
7012 compiling for a VxWorks system.
7016 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7017 header to indicate that @samp{eabi} extended relocations are used.
7023 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7024 Embedded Applications Binary Interface (eabi) which is a set of
7025 modifications to the System V.4 specifications. Selecting @option{-meabi}
7026 means that the stack is aligned to an 8 byte boundary, a function
7027 @code{__eabi} is called to from @code{main} to set up the eabi
7028 environment, and the @option{-msdata} option can use both @code{r2} and
7029 @code{r13} to point to two separate small data areas. Selecting
7030 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7031 do not call an initialization function from @code{main}, and the
7032 @option{-msdata} option will only use @code{r13} to point to a single
7033 small data area. The @option{-meabi} option is on by default if you
7034 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7037 @opindex msdata=eabi
7038 On System V.4 and embedded PowerPC systems, put small initialized
7039 @code{const} global and static data in the @samp{.sdata2} section, which
7040 is pointed to by register @code{r2}. Put small initialized
7041 non-@code{const} global and static data in the @samp{.sdata} section,
7042 which is pointed to by register @code{r13}. Put small uninitialized
7043 global and static data in the @samp{.sbss} section, which is adjacent to
7044 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7045 incompatible with the @option{-mrelocatable} option. The
7046 @option{-msdata=eabi} option also sets the @option{-memb} option.
7049 @opindex msdata=sysv
7050 On System V.4 and embedded PowerPC systems, put small global and static
7051 data in the @samp{.sdata} section, which is pointed to by register
7052 @code{r13}. Put small uninitialized global and static data in the
7053 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7054 The @option{-msdata=sysv} option is incompatible with the
7055 @option{-mrelocatable} option.
7057 @item -msdata=default
7059 @opindex msdata=default
7061 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7062 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7063 same as @option{-msdata=sysv}.
7066 @opindex msdata-data
7067 On System V.4 and embedded PowerPC systems, put small global and static
7068 data in the @samp{.sdata} section. Put small uninitialized global and
7069 static data in the @samp{.sbss} section. Do not use register @code{r13}
7070 to address small data however. This is the default behavior unless
7071 other @option{-msdata} options are used.
7075 @opindex msdata=none
7077 On embedded PowerPC systems, put all initialized global and static data
7078 in the @samp{.data} section, and all uninitialized data in the
7079 @samp{.bss} section.
7083 @cindex smaller data references (PowerPC)
7084 @cindex .sdata/.sdata2 references (PowerPC)
7085 On embedded PowerPC systems, put global and static items less than or
7086 equal to @var{num} bytes into the small data or bss sections instead of
7087 the normal data or bss section. By default, @var{num} is 8. The
7088 @option{-G @var{num}} switch is also passed to the linker.
7089 All modules should be compiled with the same @option{-G @var{num}} value.
7092 @itemx -mno-regnames
7094 @opindex mno-regnames
7095 On System V.4 and embedded PowerPC systems do (do not) emit register
7096 names in the assembly language output using symbolic forms.
7100 Adds support for multithreading with the @dfn{pthreads} library.
7101 This option sets flags for both the preprocessor and linker.
7106 @subsection IBM RT Options
7108 @cindex IBM RT options
7110 These @samp{-m} options are defined for the IBM RT PC:
7114 @opindex min-line-mul
7115 Use an in-line code sequence for integer multiplies. This is the
7118 @item -mcall-lib-mul
7119 @opindex mcall-lib-mul
7120 Call @code{lmul$$} for integer multiples.
7122 @item -mfull-fp-blocks
7123 @opindex mfull-fp-blocks
7124 Generate full-size floating point data blocks, including the minimum
7125 amount of scratch space recommended by IBM@. This is the default.
7127 @item -mminimum-fp-blocks
7128 @opindex mminimum-fp-blocks
7129 Do not include extra scratch space in floating point data blocks. This
7130 results in smaller code, but slower execution, since scratch space must
7131 be allocated dynamically.
7133 @cindex @file{varargs.h} and RT PC
7134 @cindex @file{stdarg.h} and RT PC
7135 @item -mfp-arg-in-fpregs
7136 @opindex mfp-arg-in-fpregs
7137 Use a calling sequence incompatible with the IBM calling convention in
7138 which floating point arguments are passed in floating point registers.
7139 Note that @code{varargs.h} and @code{stdarg.h} will not work with
7140 floating point operands if this option is specified.
7142 @item -mfp-arg-in-gregs
7143 @opindex mfp-arg-in-gregs
7144 Use the normal calling convention for floating point arguments. This is
7147 @item -mhc-struct-return
7148 @opindex mhc-struct-return
7149 Return structures of more than one word in memory, rather than in a
7150 register. This provides compatibility with the MetaWare HighC (hc)
7151 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7152 with the Portable C Compiler (pcc).
7154 @item -mnohc-struct-return
7155 @opindex mnohc-struct-return
7156 Return some structures of more than one word in registers, when
7157 convenient. This is the default. For compatibility with the
7158 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7159 option @option{-mhc-struct-return}.
7163 @subsection MIPS Options
7164 @cindex MIPS options
7166 These @samp{-m} options are defined for the MIPS family of computers:
7170 @item -march=@var{cpu-type}
7172 Assume the defaults for the machine type @var{cpu-type} when generating
7173 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7174 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7175 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7176 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7177 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7178 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
7180 @item -mtune=@var{cpu-type}
7182 Assume the defaults for the machine type @var{cpu-type} when scheduling
7183 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7184 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7185 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7186 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7187 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7188 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
7189 @var{cpu-type} will schedule things appropriately for that particular
7190 chip, the compiler will not generate any code that does not meet level 1
7191 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
7192 or @option{-mabi} switch being used.
7194 @item -mcpu=@var{cpu-type}
7196 This is identical to specifying both @option{-march} and @option{-mtune}.
7200 Issue instructions from level 1 of the MIPS ISA@. This is the default.
7201 @samp{r3000} is the default @var{cpu-type} at this ISA level.
7205 Issue instructions from level 2 of the MIPS ISA (branch likely, square
7206 root instructions). @samp{r6000} is the default @var{cpu-type} at this
7211 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
7212 @samp{r4000} is the default @var{cpu-type} at this ISA level.
7216 Issue instructions from level 4 of the MIPS ISA (conditional move,
7217 prefetch, enhanced FPU instructions). @samp{r8000} is the default
7218 @var{cpu-type} at this ISA level.
7222 Assume that 32 32-bit floating point registers are available. This is
7227 Assume that 32 64-bit floating point registers are available. This is
7228 the default when the @option{-mips3} option is used.
7231 @itemx -mno-fused-madd
7232 @opindex mfused-madd
7233 @opindex mno-fused-madd
7234 Generate code that uses (does not use) the floating point multiply and
7235 accumulate instructions, when they are available. These instructions
7236 are generated by default if they are available, but this may be
7237 undesirable if the extra precision causes problems or on certain chips
7238 in the mode where denormals are rounded to zero where denormals
7239 generated by multiply and accumulate instructions cause exceptions
7244 Assume that 32 32-bit general purpose registers are available. This is
7249 Assume that 32 64-bit general purpose registers are available. This is
7250 the default when the @option{-mips3} option is used.
7254 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7255 explanation of the default, and the width of pointers.
7259 Force long types to be 64 bits wide. See @option{-mlong32} for an
7260 explanation of the default, and the width of pointers.
7264 Force long, int, and pointer types to be 32 bits wide.
7266 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
7267 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
7268 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
7269 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
7270 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
7271 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
7272 are 32 bits, and longs are 64 bits wide. The width of pointer types is
7273 the smaller of the width of longs or the width of general purpose
7274 registers (which in turn depends on the ISA)@.
7286 Generate code for the indicated ABI@. The default instruction level is
7287 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
7288 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
7289 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
7294 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7295 add normal debug information. This is the default for all
7296 platforms except for the OSF/1 reference platform, using the OSF/rose
7297 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7298 switches are used, the @file{mips-tfile} program will encapsulate the
7299 stabs within MIPS ECOFF@.
7303 Generate code for the GNU assembler. This is the default on the OSF/1
7304 reference platform, using the OSF/rose object format. Also, this is
7305 the default if the configure option @option{--with-gnu-as} is used.
7307 @item -msplit-addresses
7308 @itemx -mno-split-addresses
7309 @opindex msplit-addresses
7310 @opindex mno-split-addresses
7311 Generate code to load the high and low parts of address constants separately.
7312 This allows GCC to optimize away redundant loads of the high order
7313 bits of addresses. This optimization requires GNU as and GNU ld.
7314 This optimization is enabled by default for some embedded targets where
7315 GNU as and GNU ld are standard.
7321 The @option{-mrnames} switch says to output code using the MIPS software
7322 names for the registers, instead of the hardware names (ie, @var{a0}
7323 instead of @var{$4}). The only known assembler that supports this option
7324 is the Algorithmics assembler.
7330 The @option{-mgpopt} switch says to write all of the data declarations
7331 before the instructions in the text section, this allows the MIPS
7332 assembler to generate one word memory references instead of using two
7333 words for short global or static data items. This is on by default if
7334 optimization is selected.
7340 For each non-inline function processed, the @option{-mstats} switch
7341 causes the compiler to emit one line to the standard error file to
7342 print statistics about the program (number of registers saved, stack
7349 The @option{-mmemcpy} switch makes all block moves call the appropriate
7350 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7351 generating inline code.
7354 @itemx -mno-mips-tfile
7355 @opindex mmips-tfile
7356 @opindex mno-mips-tfile
7357 The @option{-mno-mips-tfile} switch causes the compiler not
7358 postprocess the object file with the @file{mips-tfile} program,
7359 after the MIPS assembler has generated it to add debug support. If
7360 @file{mips-tfile} is not run, then no local variables will be
7361 available to the debugger. In addition, @file{stage2} and
7362 @file{stage3} objects will have the temporary file names passed to the
7363 assembler embedded in the object file, which means the objects will
7364 not compare the same. The @option{-mno-mips-tfile} switch should only
7365 be used when there are bugs in the @file{mips-tfile} program that
7366 prevents compilation.
7369 @opindex msoft-float
7370 Generate output containing library calls for floating point.
7371 @strong{Warning:} the requisite libraries are not part of GCC@.
7372 Normally the facilities of the machine's usual C compiler are used, but
7373 this can't be done directly in cross-compilation. You must make your
7374 own arrangements to provide suitable library functions for
7378 @opindex mhard-float
7379 Generate output containing floating point instructions. This is the
7380 default if you use the unmodified sources.
7383 @itemx -mno-abicalls
7385 @opindex mno-abicalls
7386 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7387 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7388 position independent code.
7391 @itemx -mno-long-calls
7392 @opindex mlong-calls
7393 @opindex mno-long-calls
7394 Do all calls with the @samp{JALR} instruction, which requires
7395 loading up a function's address into a register before the call.
7396 You need to use this switch, if you call outside of the current
7397 512 megabyte segment to functions that are not through pointers.
7400 @itemx -mno-half-pic
7402 @opindex mno-half-pic
7403 Put pointers to extern references into the data section and load them
7404 up, rather than put the references in the text section.
7406 @item -membedded-pic
7407 @itemx -mno-embedded-pic
7408 @opindex membedded-pic
7409 @opindex mno-embedded-pic
7410 Generate PIC code suitable for some embedded systems. All calls are
7411 made using PC relative address, and all data is addressed using the $gp
7412 register. No more than 65536 bytes of global data may be used. This
7413 requires GNU as and GNU ld which do most of the work. This currently
7414 only works on targets which use ECOFF; it does not work with ELF@.
7416 @item -membedded-data
7417 @itemx -mno-embedded-data
7418 @opindex membedded-data
7419 @opindex mno-embedded-data
7420 Allocate variables to the read-only data section first if possible, then
7421 next in the small data section if possible, otherwise in data. This gives
7422 slightly slower code than the default, but reduces the amount of RAM required
7423 when executing, and thus may be preferred for some embedded systems.
7425 @item -muninit-const-in-rodata
7426 @itemx -mno-uninit-const-in-rodata
7427 @opindex muninit-const-in-rodata
7428 @opindex mno-uninit-const-in-rodata
7429 When used together with @option{-membedded-data}, it will always store uninitialized
7430 const variables in the read-only data section.
7432 @item -msingle-float
7433 @itemx -mdouble-float
7434 @opindex msingle-float
7435 @opindex mdouble-float
7436 The @option{-msingle-float} switch tells gcc to assume that the floating
7437 point coprocessor only supports single precision operations, as on the
7438 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7439 double precision operations. This is the default.
7445 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7446 as on the @samp{r4650} chip.
7450 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7451 @option{-mcpu=r4650}.
7457 Enable 16-bit instructions.
7461 Use the entry and exit pseudo ops. This option can only be used with
7466 Compile code for the processor in little endian mode.
7467 The requisite libraries are assumed to exist.
7471 Compile code for the processor in big endian mode.
7472 The requisite libraries are assumed to exist.
7476 @cindex smaller data references (MIPS)
7477 @cindex gp-relative references (MIPS)
7478 Put global and static items less than or equal to @var{num} bytes into
7479 the small data or bss sections instead of the normal data or bss
7480 section. This allows the assembler to emit one word memory reference
7481 instructions based on the global pointer (@var{gp} or @var{$28}),
7482 instead of the normal two words used. By default, @var{num} is 8 when
7483 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7484 @option{-G @var{num}} switch is also passed to the assembler and linker.
7485 All modules should be compiled with the same @option{-G @var{num}}
7490 Tell the MIPS assembler to not run its preprocessor over user
7491 assembler files (with a @samp{.s} suffix) when assembling them.
7495 Pass an option to gas which will cause nops to be inserted if
7496 the read of the destination register of an mfhi or mflo instruction
7497 occurs in the following two instructions.
7501 Do not include the default crt0.
7503 @item -mflush-func=@var{func}
7504 @itemx -mno-flush-func
7505 @opindex mflush-func
7506 Specifies the function to call to flush the I and D caches, or to not
7507 call any such function. If called, the function must take the same
7508 arguments as the common @code{_flush_func()}, that is, the address of the
7509 memory range for which the cache is being flushed, the size of the
7510 memory range, and the number 3 (to flush both caches). The default
7511 depends on the target gcc was configured for, but commonly is either
7512 @samp{_flush_func} or @samp{__cpu_flush}.
7515 These options are defined by the macro
7516 @code{TARGET_SWITCHES} in the machine description. The default for the
7517 options is also defined by that macro, which enables you to change the
7520 @node i386 and x86-64 Options
7521 @subsection Intel 386 and AMD x86-64 Options
7522 @cindex i386 Options
7523 @cindex x86-64 Options
7524 @cindex Intel 386 Options
7525 @cindex AMD x86-64 Options
7527 These @samp{-m} options are defined for the i386 and x86-64 family of
7531 @item -mcpu=@var{cpu-type}
7533 Tune to @var{cpu-type} everything applicable about the generated code, except
7534 for the ABI and the set of available instructions. The choices for
7535 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7536 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7537 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7538 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp}
7539 and @samp{athlon-mp}.
7541 While picking a specific @var{cpu-type} will schedule things appropriately
7542 for that particular chip, the compiler will not generate any code that
7543 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7544 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7545 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7546 AMD chips as opposed to the Intel ones.
7548 @item -march=@var{cpu-type}
7550 Generate instructions for the machine type @var{cpu-type}. The choices
7551 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7552 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7561 @opindex mpentiumpro
7562 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7563 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7564 These synonyms are deprecated.
7566 @item -mfpmath=@var{unit}
7568 generate floating point arithmetics for selected unit @var{unit}. the choices
7573 Use the standard 387 floating point coprocessor present majority of chips and
7574 emulated otherwise. Code compiled with this option will run almost everywhere.
7575 The temporary results are computed in 80bit precesion instead of precision
7576 specified by the type resulting in slightly different results compared to most
7577 of other chips. See @option{-ffloat-store} for more detailed description.
7579 This is the default choice for i386 compiler.
7582 Use scalar floating point instructions present in the SSE instruction set.
7583 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7584 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7585 instruction set supports only single precision arithmetics, thus the double and
7586 extended precision arithmetics is still done using 387. Later version, present
7587 only in Pentium4 and the future AMD x86-64 chips supports double precision
7590 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7591 @option{-msse2} switches to enable SSE extensions and make this option
7592 effective. For x86-64 compiler, these extensions are enabled by default.
7594 The resulting code should be considerably faster in majority of cases and avoid
7595 the numerical instability problems of 387 code, but may break some existing
7596 code that expects temporaries to be 80bit.
7598 This is the default choice for x86-64 compiler.
7601 Attempt to utilize both instruction sets at once. This effectivly double the
7602 amount of available registers and on chips with separate execution units for
7603 387 and SSE the execution resources too. Use this option with care, as it is
7604 still experimental, because gcc register allocator does not model separate
7605 functional units well resulting in instable performance.
7608 @item -masm=@var{dialect}
7609 @opindex masm=@var{dialect}
7610 Output asm instructions using selected @var{dialect}. Supported choices are
7611 @samp{intel} or @samp{att} (the default one).
7616 @opindex mno-ieee-fp
7617 Control whether or not the compiler uses IEEE floating point
7618 comparisons. These handle correctly the case where the result of a
7619 comparison is unordered.
7622 @opindex msoft-float
7623 Generate output containing library calls for floating point.
7624 @strong{Warning:} the requisite libraries are not part of GCC@.
7625 Normally the facilities of the machine's usual C compiler are used, but
7626 this can't be done directly in cross-compilation. You must make your
7627 own arrangements to provide suitable library functions for
7630 On machines where a function returns floating point results in the 80387
7631 register stack, some floating point opcodes may be emitted even if
7632 @option{-msoft-float} is used.
7634 @item -mno-fp-ret-in-387
7635 @opindex mno-fp-ret-in-387
7636 Do not use the FPU registers for return values of functions.
7638 The usual calling convention has functions return values of types
7639 @code{float} and @code{double} in an FPU register, even if there
7640 is no FPU@. The idea is that the operating system should emulate
7643 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7644 in ordinary CPU registers instead.
7646 @item -mno-fancy-math-387
7647 @opindex mno-fancy-math-387
7648 Some 387 emulators do not support the @code{sin}, @code{cos} and
7649 @code{sqrt} instructions for the 387. Specify this option to avoid
7650 generating those instructions. This option is the default on FreeBSD@.
7651 As of revision 2.6.1, these instructions are not generated unless you
7652 also use the @option{-funsafe-math-optimizations} switch.
7654 @item -malign-double
7655 @itemx -mno-align-double
7656 @opindex malign-double
7657 @opindex mno-align-double
7658 Control whether GCC aligns @code{double}, @code{long double}, and
7659 @code{long long} variables on a two word boundary or a one word
7660 boundary. Aligning @code{double} variables on a two word boundary will
7661 produce code that runs somewhat faster on a @samp{Pentium} at the
7662 expense of more memory.
7664 @item -m128bit-long-double
7665 @opindex m128bit-long-double
7666 Control the size of @code{long double} type. i386 application binary interface
7667 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7668 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7669 impossible to reach with 12 byte long doubles in the array accesses.
7671 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7672 structures and arrays containing @code{long double} will change their size as
7673 well as function calling convention for function taking @code{long double}
7676 @item -m96bit-long-double
7677 @opindex m96bit-long-double
7678 Set the size of @code{long double} to 96 bits as required by the i386
7679 application binary interface. This is the default.
7682 @itemx -mno-svr3-shlib
7683 @opindex msvr3-shlib
7684 @opindex mno-svr3-shlib
7685 Control whether GCC places uninitialized local variables into the
7686 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7687 into @code{bss}. These options are meaningful only on System V Release 3.
7691 Use a different function-calling convention, in which functions that
7692 take a fixed number of arguments return with the @code{ret} @var{num}
7693 instruction, which pops their arguments while returning. This saves one
7694 instruction in the caller since there is no need to pop the arguments
7697 You can specify that an individual function is called with this calling
7698 sequence with the function attribute @samp{stdcall}. You can also
7699 override the @option{-mrtd} option by using the function attribute
7700 @samp{cdecl}. @xref{Function Attributes}.
7702 @strong{Warning:} this calling convention is incompatible with the one
7703 normally used on Unix, so you cannot use it if you need to call
7704 libraries compiled with the Unix compiler.
7706 Also, you must provide function prototypes for all functions that
7707 take variable numbers of arguments (including @code{printf});
7708 otherwise incorrect code will be generated for calls to those
7711 In addition, seriously incorrect code will result if you call a
7712 function with too many arguments. (Normally, extra arguments are
7713 harmlessly ignored.)
7715 @item -mregparm=@var{num}
7717 Control how many registers are used to pass integer arguments. By
7718 default, no registers are used to pass arguments, and at most 3
7719 registers can be used. You can control this behavior for a specific
7720 function by using the function attribute @samp{regparm}.
7721 @xref{Function Attributes}.
7723 @strong{Warning:} if you use this switch, and
7724 @var{num} is nonzero, then you must build all modules with the same
7725 value, including any libraries. This includes the system libraries and
7728 @item -mpreferred-stack-boundary=@var{num}
7729 @opindex mpreferred-stack-boundary
7730 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7731 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7732 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7733 size (@option{-Os}), in which case the default is the minimum correct
7734 alignment (4 bytes for x86, and 8 bytes for x86-64).
7736 On Pentium and PentiumPro, @code{double} and @code{long double} values
7737 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7738 suffer significant run time performance penalties. On Pentium III, the
7739 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7740 penalties if it is not 16 byte aligned.
7742 To ensure proper alignment of this values on the stack, the stack boundary
7743 must be as aligned as that required by any value stored on the stack.
7744 Further, every function must be generated such that it keeps the stack
7745 aligned. Thus calling a function compiled with a higher preferred
7746 stack boundary from a function compiled with a lower preferred stack
7747 boundary will most likely misalign the stack. It is recommended that
7748 libraries that use callbacks always use the default setting.
7750 This extra alignment does consume extra stack space, and generally
7751 increases code size. Code that is sensitive to stack space usage, such
7752 as embedded systems and operating system kernels, may want to reduce the
7753 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7769 These switches enable or disable the use of built-in functions that allow
7770 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7772 @xref{X86 Built-in Functions}, for details of the functions enabled
7773 and disabled by these switches.
7776 @itemx -mno-push-args
7778 @opindex mno-push-args
7779 Use PUSH operations to store outgoing parameters. This method is shorter
7780 and usually equally fast as method using SUB/MOV operations and is enabled
7781 by default. In some cases disabling it may improve performance because of
7782 improved scheduling and reduced dependencies.
7784 @item -maccumulate-outgoing-args
7785 @opindex maccumulate-outgoing-args
7786 If enabled, the maximum amount of space required for outgoing arguments will be
7787 computed in the function prologue. This is faster on most modern CPUs
7788 because of reduced dependencies, improved scheduling and reduced stack usage
7789 when preferred stack boundary is not equal to 2. The drawback is a notable
7790 increase in code size. This switch implies @option{-mno-push-args}.
7794 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7795 on thread-safe exception handling must compile and link all code with the
7796 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7797 @option{-D_MT}; when linking, it links in a special thread helper library
7798 @option{-lmingwthrd} which cleans up per thread exception handling data.
7800 @item -mno-align-stringops
7801 @opindex mno-align-stringops
7802 Do not align destination of inlined string operations. This switch reduces
7803 code size and improves performance in case the destination is already aligned,
7804 but gcc don't know about it.
7806 @item -minline-all-stringops
7807 @opindex minline-all-stringops
7808 By default GCC inlines string operations only when destination is known to be
7809 aligned at least to 4 byte boundary. This enables more inlining, increase code
7810 size, but may improve performance of code that depends on fast memcpy, strlen
7811 and memset for short lengths.
7813 @item -momit-leaf-frame-pointer
7814 @opindex momit-leaf-frame-pointer
7815 Don't keep the frame pointer in a register for leaf functions. This
7816 avoids the instructions to save, set up and restore frame pointers and
7817 makes an extra register available in leaf functions. The option
7818 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7819 which might make debugging harder.
7822 These @samp{-m} switches are supported in addition to the above
7823 on AMD x86-64 processors in 64-bit environments.
7830 Generate code for a 32-bit or 64-bit environment.
7831 The 32-bit environment sets int, long and pointer to 32 bits and
7832 generates code that runs on any i386 system.
7833 The 64-bit environment sets int to 32 bits and long and pointer
7834 to 64 bits and generates code for AMD's x86-64 architecture.
7837 @opindex no-red-zone
7838 Do not use a so called red zone for x86-64 code. The red zone is mandated
7839 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7840 stack pointer that will not be modified by signal or interrupt handlers
7841 and therefore can be used for temporary data without adjusting the stack
7842 pointer. The flag @option{-mno-red-zone} disables this red zone.
7846 @subsection HPPA Options
7847 @cindex HPPA Options
7849 These @samp{-m} options are defined for the HPPA family of computers:
7852 @item -march=@var{architecture-type}
7854 Generate code for the specified architecture. The choices for
7855 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7856 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7857 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7858 architecture option for your machine. Code compiled for lower numbered
7859 architectures will run on higher numbered architectures, but not the
7862 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7863 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7867 @itemx -mpa-risc-1-1
7868 @itemx -mpa-risc-2-0
7869 @opindex mpa-risc-1-0
7870 @opindex mpa-risc-1-1
7871 @opindex mpa-risc-2-0
7872 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7875 @opindex mbig-switch
7876 Generate code suitable for big switch tables. Use this option only if
7877 the assembler/linker complain about out of range branches within a switch
7880 @item -mjump-in-delay
7881 @opindex mjump-in-delay
7882 Fill delay slots of function calls with unconditional jump instructions
7883 by modifying the return pointer for the function call to be the target
7884 of the conditional jump.
7886 @item -mdisable-fpregs
7887 @opindex mdisable-fpregs
7888 Prevent floating point registers from being used in any manner. This is
7889 necessary for compiling kernels which perform lazy context switching of
7890 floating point registers. If you use this option and attempt to perform
7891 floating point operations, the compiler will abort.
7893 @item -mdisable-indexing
7894 @opindex mdisable-indexing
7895 Prevent the compiler from using indexing address modes. This avoids some
7896 rather obscure problems when compiling MIG generated code under MACH@.
7898 @item -mno-space-regs
7899 @opindex mno-space-regs
7900 Generate code that assumes the target has no space registers. This allows
7901 GCC to generate faster indirect calls and use unscaled index address modes.
7903 Such code is suitable for level 0 PA systems and kernels.
7905 @item -mfast-indirect-calls
7906 @opindex mfast-indirect-calls
7907 Generate code that assumes calls never cross space boundaries. This
7908 allows GCC to emit code which performs faster indirect calls.
7910 This option will not work in the presence of shared libraries or nested
7913 @item -mlong-load-store
7914 @opindex mlong-load-store
7915 Generate 3-instruction load and store sequences as sometimes required by
7916 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7919 @item -mportable-runtime
7920 @opindex mportable-runtime
7921 Use the portable calling conventions proposed by HP for ELF systems.
7925 Enable the use of assembler directives only GAS understands.
7927 @item -mschedule=@var{cpu-type}
7929 Schedule code according to the constraints for the machine type
7930 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7931 @samp{7100}, @samp{7100LC}, @samp{7200}, and @samp{8000}. Refer to
7932 @file{/usr/lib/sched.models} on an HP-UX system to determine the
7933 proper scheduling option for your machine.
7936 @opindex mlinker-opt
7937 Enable the optimization pass in the HPUX linker. Note this makes symbolic
7938 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
7939 in which they give bogus error messages when linking some programs.
7942 @opindex msoft-float
7943 Generate output containing library calls for floating point.
7944 @strong{Warning:} the requisite libraries are not available for all HPPA
7945 targets. Normally the facilities of the machine's usual C compiler are
7946 used, but this cannot be done directly in cross-compilation. You must make
7947 your own arrangements to provide suitable library functions for
7948 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7949 does provide software floating point support.
7951 @option{-msoft-float} changes the calling convention in the output file;
7952 therefore, it is only useful if you compile @emph{all} of a program with
7953 this option. In particular, you need to compile @file{libgcc.a}, the
7954 library that comes with GCC, with @option{-msoft-float} in order for
7958 @node Intel 960 Options
7959 @subsection Intel 960 Options
7961 These @samp{-m} options are defined for the Intel 960 implementations:
7964 @item -m@var{cpu-type}
7972 Assume the defaults for the machine type @var{cpu-type} for some of
7973 the other options, including instruction scheduling, floating point
7974 support, and addressing modes. The choices for @var{cpu-type} are
7975 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
7976 @samp{sa}, and @samp{sb}.
7983 @opindex msoft-float
7984 The @option{-mnumerics} option indicates that the processor does support
7985 floating-point instructions. The @option{-msoft-float} option indicates
7986 that floating-point support should not be assumed.
7988 @item -mleaf-procedures
7989 @itemx -mno-leaf-procedures
7990 @opindex mleaf-procedures
7991 @opindex mno-leaf-procedures
7992 Do (or do not) attempt to alter leaf procedures to be callable with the
7993 @code{bal} instruction as well as @code{call}. This will result in more
7994 efficient code for explicit calls when the @code{bal} instruction can be
7995 substituted by the assembler or linker, but less efficient code in other
7996 cases, such as calls via function pointers, or using a linker that doesn't
7997 support this optimization.
8000 @itemx -mno-tail-call
8002 @opindex mno-tail-call
8003 Do (or do not) make additional attempts (beyond those of the
8004 machine-independent portions of the compiler) to optimize tail-recursive
8005 calls into branches. You may not want to do this because the detection of
8006 cases where this is not valid is not totally complete. The default is
8007 @option{-mno-tail-call}.
8009 @item -mcomplex-addr
8010 @itemx -mno-complex-addr
8011 @opindex mcomplex-addr
8012 @opindex mno-complex-addr
8013 Assume (or do not assume) that the use of a complex addressing mode is a
8014 win on this implementation of the i960. Complex addressing modes may not
8015 be worthwhile on the K-series, but they definitely are on the C-series.
8016 The default is currently @option{-mcomplex-addr} for all processors except
8020 @itemx -mno-code-align
8021 @opindex mcode-align
8022 @opindex mno-code-align
8023 Align code to 8-byte boundaries for faster fetching (or don't bother).
8024 Currently turned on by default for C-series implementations only.
8027 @item -mclean-linkage
8028 @itemx -mno-clean-linkage
8029 @opindex mclean-linkage
8030 @opindex mno-clean-linkage
8031 These options are not fully implemented.
8035 @itemx -mic2.0-compat
8036 @itemx -mic3.0-compat
8038 @opindex mic2.0-compat
8039 @opindex mic3.0-compat
8040 Enable compatibility with iC960 v2.0 or v3.0.
8044 @opindex masm-compat
8046 Enable compatibility with the iC960 assembler.
8048 @item -mstrict-align
8049 @itemx -mno-strict-align
8050 @opindex mstrict-align
8051 @opindex mno-strict-align
8052 Do not permit (do permit) unaligned accesses.
8056 Enable structure-alignment compatibility with Intel's gcc release version
8057 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8059 @item -mlong-double-64
8060 @opindex mlong-double-64
8061 Implement type @samp{long double} as 64-bit floating point numbers.
8062 Without the option @samp{long double} is implemented by 80-bit
8063 floating point numbers. The only reason we have it because there is
8064 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8065 is only useful for people using soft-float targets. Otherwise, we
8066 should recommend against use of it.
8070 @node DEC Alpha Options
8071 @subsection DEC Alpha Options
8073 These @samp{-m} options are defined for the DEC Alpha implementations:
8076 @item -mno-soft-float
8078 @opindex mno-soft-float
8079 @opindex msoft-float
8080 Use (do not use) the hardware floating-point instructions for
8081 floating-point operations. When @option{-msoft-float} is specified,
8082 functions in @file{libgcc.a} will be used to perform floating-point
8083 operations. Unless they are replaced by routines that emulate the
8084 floating-point operations, or compiled in such a way as to call such
8085 emulations routines, these routines will issue floating-point
8086 operations. If you are compiling for an Alpha without floating-point
8087 operations, you must ensure that the library is built so as not to call
8090 Note that Alpha implementations without floating-point operations are
8091 required to have floating-point registers.
8096 @opindex mno-fp-regs
8097 Generate code that uses (does not use) the floating-point register set.
8098 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8099 register set is not used, floating point operands are passed in integer
8100 registers as if they were integers and floating-point results are passed
8101 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8102 so any function with a floating-point argument or return value called by code
8103 compiled with @option{-mno-fp-regs} must also be compiled with that
8106 A typical use of this option is building a kernel that does not use,
8107 and hence need not save and restore, any floating-point registers.
8111 The Alpha architecture implements floating-point hardware optimized for
8112 maximum performance. It is mostly compliant with the IEEE floating
8113 point standard. However, for full compliance, software assistance is
8114 required. This option generates code fully IEEE compliant code
8115 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8116 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8117 defined during compilation. The resulting code is less efficient but is
8118 able to correctly support denormalized numbers and exceptional IEEE
8119 values such as not-a-number and plus/minus infinity. Other Alpha
8120 compilers call this option @option{-ieee_with_no_inexact}.
8122 @item -mieee-with-inexact
8123 @opindex mieee-with-inexact
8124 This is like @option{-mieee} except the generated code also maintains
8125 the IEEE @var{inexact-flag}. Turning on this option causes the
8126 generated code to implement fully-compliant IEEE math. In addition to
8127 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8128 macro. On some Alpha implementations the resulting code may execute
8129 significantly slower than the code generated by default. Since there is
8130 very little code that depends on the @var{inexact-flag}, you should
8131 normally not specify this option. Other Alpha compilers call this
8132 option @option{-ieee_with_inexact}.
8134 @item -mfp-trap-mode=@var{trap-mode}
8135 @opindex mfp-trap-mode
8136 This option controls what floating-point related traps are enabled.
8137 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8138 The trap mode can be set to one of four values:
8142 This is the default (normal) setting. The only traps that are enabled
8143 are the ones that cannot be disabled in software (e.g., division by zero
8147 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8151 Like @samp{su}, but the instructions are marked to be safe for software
8152 completion (see Alpha architecture manual for details).
8155 Like @samp{su}, but inexact traps are enabled as well.
8158 @item -mfp-rounding-mode=@var{rounding-mode}
8159 @opindex mfp-rounding-mode
8160 Selects the IEEE rounding mode. Other Alpha compilers call this option
8161 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8166 Normal IEEE rounding mode. Floating point numbers are rounded towards
8167 the nearest machine number or towards the even machine number in case
8171 Round towards minus infinity.
8174 Chopped rounding mode. Floating point numbers are rounded towards zero.
8177 Dynamic rounding mode. A field in the floating point control register
8178 (@var{fpcr}, see Alpha architecture reference manual) controls the
8179 rounding mode in effect. The C library initializes this register for
8180 rounding towards plus infinity. Thus, unless your program modifies the
8181 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8184 @item -mtrap-precision=@var{trap-precision}
8185 @opindex mtrap-precision
8186 In the Alpha architecture, floating point traps are imprecise. This
8187 means without software assistance it is impossible to recover from a
8188 floating trap and program execution normally needs to be terminated.
8189 GCC can generate code that can assist operating system trap handlers
8190 in determining the exact location that caused a floating point trap.
8191 Depending on the requirements of an application, different levels of
8192 precisions can be selected:
8196 Program precision. This option is the default and means a trap handler
8197 can only identify which program caused a floating point exception.
8200 Function precision. The trap handler can determine the function that
8201 caused a floating point exception.
8204 Instruction precision. The trap handler can determine the exact
8205 instruction that caused a floating point exception.
8208 Other Alpha compilers provide the equivalent options called
8209 @option{-scope_safe} and @option{-resumption_safe}.
8211 @item -mieee-conformant
8212 @opindex mieee-conformant
8213 This option marks the generated code as IEEE conformant. You must not
8214 use this option unless you also specify @option{-mtrap-precision=i} and either
8215 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8216 is to emit the line @samp{.eflag 48} in the function prologue of the
8217 generated assembly file. Under DEC Unix, this has the effect that
8218 IEEE-conformant math library routines will be linked in.
8220 @item -mbuild-constants
8221 @opindex mbuild-constants
8222 Normally GCC examines a 32- or 64-bit integer constant to
8223 see if it can construct it from smaller constants in two or three
8224 instructions. If it cannot, it will output the constant as a literal and
8225 generate code to load it from the data segment at runtime.
8227 Use this option to require GCC to construct @emph{all} integer constants
8228 using code, even if it takes more instructions (the maximum is six).
8230 You would typically use this option to build a shared library dynamic
8231 loader. Itself a shared library, it must relocate itself in memory
8232 before it can find the variables and constants in its own data segment.
8238 Select whether to generate code to be assembled by the vendor-supplied
8239 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8257 Indicate whether GCC should generate code to use the optional BWX,
8258 CIX, FIX and MAX instruction sets. The default is to use the instruction
8259 sets supported by the CPU type specified via @option{-mcpu=} option or that
8260 of the CPU on which GCC was built if none was specified.
8265 @opindex mfloat-ieee
8266 Generate code that uses (does not use) VAX F and G floating point
8267 arithmetic instead of IEEE single and double precision.
8269 @item -mexplicit-relocs
8270 @itemx -mno-explicit-relocs
8271 @opindex mexplicit-relocs
8272 @opindex mno-explicit-relocs
8273 Older Alpha assemblers provided no way to generate symbol relocations
8274 except via assembler macros. Use of these macros does not allow
8275 optimial instruction scheduling. GNU binutils as of version 2.12
8276 supports a new syntax that allows the compiler to explicitly mark
8277 which relocations should apply to which instructions. This option
8278 is mostly useful for debugging, as GCC detects the capabilities of
8279 the assembler when it is built and sets the default accordingly.
8283 @opindex msmall-data
8284 @opindex mlarge-data
8285 When @option{-mexplicit-relocs} is in effect, static data is
8286 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8287 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8288 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8289 16-bit relocations off of the @code{$gp} register. This limits the
8290 size of the small data area to 64KB, but allows the variables to be
8291 directly accessed via a single instruction.
8293 The default is @option{-mlarge-data}. With this option the data area
8294 is limited to just below 2GB. Programs that require more than 2GB of
8295 data must use @code{malloc} or @code{mmap} to allocate the data in the
8296 heap instead of in the program's data segment.
8298 When generating code for shared libraries, @option{-fpic} implies
8299 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8301 @item -mcpu=@var{cpu_type}
8303 Set the instruction set and instruction scheduling parameters for
8304 machine type @var{cpu_type}. You can specify either the @samp{EV}
8305 style name or the corresponding chip number. GCC supports scheduling
8306 parameters for the EV4, EV5 and EV6 family of processors and will
8307 choose the default values for the instruction set from the processor
8308 you specify. If you do not specify a processor type, GCC will default
8309 to the processor on which the compiler was built.
8311 Supported values for @var{cpu_type} are
8317 Schedules as an EV4 and has no instruction set extensions.
8321 Schedules as an EV5 and has no instruction set extensions.
8325 Schedules as an EV5 and supports the BWX extension.
8330 Schedules as an EV5 and supports the BWX and MAX extensions.
8334 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8338 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8341 @item -mtune=@var{cpu_type}
8343 Set only the instruction scheduling parameters for machine type
8344 @var{cpu_type}. The instruction set is not changed.
8346 @item -mmemory-latency=@var{time}
8347 @opindex mmemory-latency
8348 Sets the latency the scheduler should assume for typical memory
8349 references as seen by the application. This number is highly
8350 dependent on the memory access patterns used by the application
8351 and the size of the external cache on the machine.
8353 Valid options for @var{time} are
8357 A decimal number representing clock cycles.
8363 The compiler contains estimates of the number of clock cycles for
8364 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8365 (also called Dcache, Scache, and Bcache), as well as to main memory.
8366 Note that L3 is only valid for EV5.
8371 @node DEC Alpha/VMS Options
8372 @subsection DEC Alpha/VMS Options
8374 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8377 @item -mvms-return-codes
8378 @opindex mvms-return-codes
8379 Return VMS condition codes from main. The default is to return POSIX
8380 style condition (e.g.@ error) codes.
8383 @node Clipper Options
8384 @subsection Clipper Options
8386 These @samp{-m} options are defined for the Clipper implementations:
8391 Produce code for a C300 Clipper processor. This is the default.
8395 Produce code for a C400 Clipper processor, i.e.@: use floating point
8399 @node H8/300 Options
8400 @subsection H8/300 Options
8402 These @samp{-m} options are defined for the H8/300 implementations:
8407 Shorten some address references at link time, when possible; uses the
8408 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8409 ld.info, Using ld}, for a fuller description.
8413 Generate code for the H8/300H@.
8417 Generate code for the H8/S@.
8421 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8425 Make @code{int} data 32 bits by default.
8429 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8430 The default for the H8/300H and H8/S is to align longs and floats on 4
8432 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8433 This option has no effect on the H8/300.
8437 @subsection SH Options
8439 These @samp{-m} options are defined for the SH implementations:
8444 Generate code for the SH1.
8448 Generate code for the SH2.
8452 Generate code for the SH3.
8456 Generate code for the SH3e.
8460 Generate code for the SH4 without a floating-point unit.
8462 @item -m4-single-only
8463 @opindex m4-single-only
8464 Generate code for the SH4 with a floating-point unit that only
8465 supports single-precision arithmetic.
8469 Generate code for the SH4 assuming the floating-point unit is in
8470 single-precision mode by default.
8474 Generate code for the SH4.
8478 Compile code for the processor in big endian mode.
8482 Compile code for the processor in little endian mode.
8486 Align doubles at 64-bit boundaries. Note that this changes the calling
8487 conventions, and thus some functions from the standard C library will
8488 not work unless you recompile it first with @option{-mdalign}.
8492 Shorten some address references at link time, when possible; uses the
8493 linker option @option{-relax}.
8497 Use 32-bit offsets in @code{switch} tables. The default is to use
8502 Enable the use of the instruction @code{fmovd}.
8506 Comply with the calling conventions defined by Hitachi.
8510 Mark the @code{MAC} register as call-clobbered, even if
8511 @option{-mhitachi} is given.
8515 Increase IEEE-compliance of floating-point code.
8519 Dump instruction size and location in the assembly code.
8523 This option is deprecated. It pads structures to multiple of 4 bytes,
8524 which is incompatible with the SH ABI@.
8528 Optimize for space instead of speed. Implied by @option{-Os}.
8532 When generating position-independent code, emit function calls using
8533 the Global Offset Table instead of the Procedure Linkage Table.
8537 Generate a library function call to invalidate instruction cache
8538 entries, after fixing up a trampoline. This library function call
8539 doesn't assume it can write to the whole memory address space. This
8540 is the default when the target is @code{sh-*-linux*}.
8543 @node System V Options
8544 @subsection Options for System V
8546 These additional options are available on System V Release 4 for
8547 compatibility with other compilers on those systems:
8552 Create a shared object.
8553 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8557 Identify the versions of each tool used by the compiler, in a
8558 @code{.ident} assembler directive in the output.
8562 Refrain from adding @code{.ident} directives to the output file (this is
8565 @item -YP,@var{dirs}
8567 Search the directories @var{dirs}, and no others, for libraries
8568 specified with @option{-l}.
8572 Look in the directory @var{dir} to find the M4 preprocessor.
8573 The assembler uses this option.
8574 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8575 @c the generic assembler that comes with Solaris takes just -Ym.
8578 @node TMS320C3x/C4x Options
8579 @subsection TMS320C3x/C4x Options
8580 @cindex TMS320C3x/C4x Options
8582 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8586 @item -mcpu=@var{cpu_type}
8588 Set the instruction set, register set, and instruction scheduling
8589 parameters for machine type @var{cpu_type}. Supported values for
8590 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8591 @samp{c44}. The default is @samp{c40} to generate code for the
8596 @itemx -msmall-memory
8598 @opindex mbig-memory
8600 @opindex msmall-memory
8602 Generates code for the big or small memory model. The small memory
8603 model assumed that all data fits into one 64K word page. At run-time
8604 the data page (DP) register must be set to point to the 64K page
8605 containing the .bss and .data program sections. The big memory model is
8606 the default and requires reloading of the DP register for every direct
8613 Allow (disallow) allocation of general integer operands into the block
8620 Enable (disable) generation of code using decrement and branch,
8621 DBcond(D), instructions. This is enabled by default for the C4x. To be
8622 on the safe side, this is disabled for the C3x, since the maximum
8623 iteration count on the C3x is @math{2^23 + 1} (but who iterates loops more than
8624 @math{2^23} times on the C3x?). Note that GCC will try to reverse a loop so
8625 that it can utilise the decrement and branch instruction, but will give
8626 up if there is more than one memory reference in the loop. Thus a loop
8627 where the loop counter is decremented can generate slightly more
8628 efficient code, in cases where the RPTB instruction cannot be utilised.
8630 @item -mdp-isr-reload
8632 @opindex mdp-isr-reload
8634 Force the DP register to be saved on entry to an interrupt service
8635 routine (ISR), reloaded to point to the data section, and restored on
8636 exit from the ISR@. This should not be required unless someone has
8637 violated the small memory model by modifying the DP register, say within
8644 For the C3x use the 24-bit MPYI instruction for integer multiplies
8645 instead of a library call to guarantee 32-bit results. Note that if one
8646 of the operands is a constant, then the multiplication will be performed
8647 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8648 then squaring operations are performed inline instead of a library call.
8651 @itemx -mno-fast-fix
8653 @opindex mno-fast-fix
8654 The C3x/C4x FIX instruction to convert a floating point value to an
8655 integer value chooses the nearest integer less than or equal to the
8656 floating point value rather than to the nearest integer. Thus if the
8657 floating point number is negative, the result will be incorrectly
8658 truncated an additional code is necessary to detect and correct this
8659 case. This option can be used to disable generation of the additional
8660 code required to correct the result.
8666 Enable (disable) generation of repeat block sequences using the RPTB
8667 instruction for zero overhead looping. The RPTB construct is only used
8668 for innermost loops that do not call functions or jump across the loop
8669 boundaries. There is no advantage having nested RPTB loops due to the
8670 overhead required to save and restore the RC, RS, and RE registers.
8671 This is enabled by default with @option{-O2}.
8673 @item -mrpts=@var{count}
8677 Enable (disable) the use of the single instruction repeat instruction
8678 RPTS@. If a repeat block contains a single instruction, and the loop
8679 count can be guaranteed to be less than the value @var{count}, GCC will
8680 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8681 then a RPTS will be emitted even if the loop count cannot be determined
8682 at compile time. Note that the repeated instruction following RPTS does
8683 not have to be reloaded from memory each iteration, thus freeing up the
8684 CPU buses for operands. However, since interrupts are blocked by this
8685 instruction, it is disabled by default.
8687 @item -mloop-unsigned
8688 @itemx -mno-loop-unsigned
8689 @opindex mloop-unsigned
8690 @opindex mno-loop-unsigned
8691 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8692 is @math{2^31 + 1} since these instructions test if the iteration count is
8693 negative to terminate the loop. If the iteration count is unsigned
8694 there is a possibility than the @math{2^31 + 1} maximum iteration count may be
8695 exceeded. This switch allows an unsigned iteration count.
8699 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8700 with. This also enforces compatibility with the API employed by the TI
8701 C3x C compiler. For example, long doubles are passed as structures
8702 rather than in floating point registers.
8708 Generate code that uses registers (stack) for passing arguments to functions.
8709 By default, arguments are passed in registers where possible rather
8710 than by pushing arguments on to the stack.
8712 @item -mparallel-insns
8713 @itemx -mno-parallel-insns
8714 @opindex mparallel-insns
8715 @opindex mno-parallel-insns
8716 Allow the generation of parallel instructions. This is enabled by
8717 default with @option{-O2}.
8719 @item -mparallel-mpy
8720 @itemx -mno-parallel-mpy
8721 @opindex mparallel-mpy
8722 @opindex mno-parallel-mpy
8723 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8724 provided @option{-mparallel-insns} is also specified. These instructions have
8725 tight register constraints which can pessimize the code generation
8731 @subsection V850 Options
8732 @cindex V850 Options
8734 These @samp{-m} options are defined for V850 implementations:
8738 @itemx -mno-long-calls
8739 @opindex mlong-calls
8740 @opindex mno-long-calls
8741 Treat all calls as being far away (near). If calls are assumed to be
8742 far away, the compiler will always load the functions address up into a
8743 register, and call indirect through the pointer.
8749 Do not optimize (do optimize) basic blocks that use the same index
8750 pointer 4 or more times to copy pointer into the @code{ep} register, and
8751 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8752 option is on by default if you optimize.
8754 @item -mno-prolog-function
8755 @itemx -mprolog-function
8756 @opindex mno-prolog-function
8757 @opindex mprolog-function
8758 Do not use (do use) external functions to save and restore registers at
8759 the prolog and epilog of a function. The external functions are slower,
8760 but use less code space if more than one function saves the same number
8761 of registers. The @option{-mprolog-function} option is on by default if
8766 Try to make the code as small as possible. At present, this just turns
8767 on the @option{-mep} and @option{-mprolog-function} options.
8771 Put static or global variables whose size is @var{n} bytes or less into
8772 the tiny data area that register @code{ep} points to. The tiny data
8773 area can hold up to 256 bytes in total (128 bytes for byte references).
8777 Put static or global variables whose size is @var{n} bytes or less into
8778 the small data area that register @code{gp} points to. The small data
8779 area can hold up to 64 kilobytes.
8783 Put static or global variables whose size is @var{n} bytes or less into
8784 the first 32 kilobytes of memory.
8788 Specify that the target processor is the V850.
8791 @opindex mbig-switch
8792 Generate code suitable for big switch tables. Use this option only if
8793 the assembler/linker complain about out of range branches within a switch
8798 @subsection ARC Options
8801 These options are defined for ARC implementations:
8806 Compile code for little endian mode. This is the default.
8810 Compile code for big endian mode.
8813 @opindex mmangle-cpu
8814 Prepend the name of the cpu to all public symbol names.
8815 In multiple-processor systems, there are many ARC variants with different
8816 instruction and register set characteristics. This flag prevents code
8817 compiled for one cpu to be linked with code compiled for another.
8818 No facility exists for handling variants that are ``almost identical''.
8819 This is an all or nothing option.
8821 @item -mcpu=@var{cpu}
8823 Compile code for ARC variant @var{cpu}.
8824 Which variants are supported depend on the configuration.
8825 All variants support @option{-mcpu=base}, this is the default.
8827 @item -mtext=@var{text-section}
8828 @itemx -mdata=@var{data-section}
8829 @itemx -mrodata=@var{readonly-data-section}
8833 Put functions, data, and readonly data in @var{text-section},
8834 @var{data-section}, and @var{readonly-data-section} respectively
8835 by default. This can be overridden with the @code{section} attribute.
8836 @xref{Variable Attributes}.
8841 @subsection NS32K Options
8842 @cindex NS32K options
8844 These are the @samp{-m} options defined for the 32000 series. The default
8845 values for these options depends on which style of 32000 was selected when
8846 the compiler was configured; the defaults for the most common choices are
8854 Generate output for a 32032. This is the default
8855 when the compiler is configured for 32032 and 32016 based systems.
8861 Generate output for a 32332. This is the default
8862 when the compiler is configured for 32332-based systems.
8868 Generate output for a 32532. This is the default
8869 when the compiler is configured for 32532-based systems.
8873 Generate output containing 32081 instructions for floating point.
8874 This is the default for all systems.
8878 Generate output containing 32381 instructions for floating point. This
8879 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8880 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8884 Try and generate multiply-add floating point instructions @code{polyF}
8885 and @code{dotF}. This option is only available if the @option{-m32381}
8886 option is in effect. Using these instructions requires changes to
8887 register allocation which generally has a negative impact on
8888 performance. This option should only be enabled when compiling code
8889 particularly likely to make heavy use of multiply-add instructions.
8892 @opindex mnomulti-add
8893 Do not try and generate multiply-add floating point instructions
8894 @code{polyF} and @code{dotF}. This is the default on all platforms.
8897 @opindex msoft-float
8898 Generate output containing library calls for floating point.
8899 @strong{Warning:} the requisite libraries may not be available.
8902 @opindex mnobitfield
8903 Do not use the bit-field instructions. On some machines it is faster to
8904 use shifting and masking operations. This is the default for the pc532.
8908 Do use the bit-field instructions. This is the default for all platforms
8913 Use a different function-calling convention, in which functions
8914 that take a fixed number of arguments return pop their
8915 arguments on return with the @code{ret} instruction.
8917 This calling convention is incompatible with the one normally
8918 used on Unix, so you cannot use it if you need to call libraries
8919 compiled with the Unix compiler.
8921 Also, you must provide function prototypes for all functions that
8922 take variable numbers of arguments (including @code{printf});
8923 otherwise incorrect code will be generated for calls to those
8926 In addition, seriously incorrect code will result if you call a
8927 function with too many arguments. (Normally, extra arguments are
8928 harmlessly ignored.)
8930 This option takes its name from the 680x0 @code{rtd} instruction.
8935 Use a different function-calling convention where the first two arguments
8936 are passed in registers.
8938 This calling convention is incompatible with the one normally
8939 used on Unix, so you cannot use it if you need to call libraries
8940 compiled with the Unix compiler.
8943 @opindex mnoregparam
8944 Do not pass any arguments in registers. This is the default for all
8949 It is OK to use the sb as an index register which is always loaded with
8950 zero. This is the default for the pc532-netbsd target.
8954 The sb register is not available for use or has not been initialized to
8955 zero by the run time system. This is the default for all targets except
8956 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
8957 @option{-fpic} is set.
8961 Many ns32000 series addressing modes use displacements of up to 512MB@.
8962 If an address is above 512MB then displacements from zero can not be used.
8963 This option causes code to be generated which can be loaded above 512MB@.
8964 This may be useful for operating systems or ROM code.
8968 Assume code will be loaded in the first 512MB of virtual address space.
8969 This is the default for all platforms.
8975 @subsection AVR Options
8978 These options are defined for AVR implementations:
8981 @item -mmcu=@var{mcu}
8983 Specify ATMEL AVR instruction set or MCU type.
8985 Instruction set avr1 is for the minimal AVR core, not supported by the C
8986 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8987 attiny11, attiny12, attiny15, attiny28).
8989 Instruction set avr2 (default) is for the classic AVR core with up to
8990 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8991 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8992 at90c8534, at90s8535).
8994 Instruction set avr3 is for the classic AVR core with up to 128K program
8995 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8997 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8998 memory space (MCU types: atmega8, atmega83, atmega85).
9000 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9001 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9002 atmega64, atmega128, at43usb355, at94k).
9006 Output instruction sizes to the asm file.
9008 @item -minit-stack=@var{N}
9009 @opindex minit-stack
9010 Specify the initial stack address, which may be a symbol or numeric value,
9011 @samp{__stack} is the default.
9013 @item -mno-interrupts
9014 @opindex mno-interrupts
9015 Generated code is not compatible with hardware interrupts.
9016 Code size will be smaller.
9018 @item -mcall-prologues
9019 @opindex mcall-prologues
9020 Functions prologues/epilogues expanded as call to appropriate
9021 subroutines. Code size will be smaller.
9023 @item -mno-tablejump
9024 @opindex mno-tablejump
9025 Do not generate tablejump insns which sometimes increase code size.
9028 @opindex mtiny-stack
9029 Change only the low 8 bits of the stack pointer.
9033 @subsection MCore Options
9034 @cindex MCore options
9036 These are the @samp{-m} options defined for the Motorola M*Core
9046 @opindex mno-hardlit
9047 Inline constants into the code stream if it can be done in two
9048 instructions or less.
9056 Use the divide instruction. (Enabled by default).
9058 @item -mrelax-immediate
9059 @itemx -mrelax-immediate
9060 @itemx -mno-relax-immediate
9061 @opindex mrelax-immediate
9062 @opindex mrelax-immediate
9063 @opindex mno-relax-immediate
9064 Allow arbitrary sized immediates in bit operations.
9066 @item -mwide-bitfields
9067 @itemx -mwide-bitfields
9068 @itemx -mno-wide-bitfields
9069 @opindex mwide-bitfields
9070 @opindex mwide-bitfields
9071 @opindex mno-wide-bitfields
9072 Always treat bit-fields as int-sized.
9074 @item -m4byte-functions
9075 @itemx -m4byte-functions
9076 @itemx -mno-4byte-functions
9077 @opindex m4byte-functions
9078 @opindex m4byte-functions
9079 @opindex mno-4byte-functions
9080 Force all functions to be aligned to a four byte boundary.
9082 @item -mcallgraph-data
9083 @itemx -mcallgraph-data
9084 @itemx -mno-callgraph-data
9085 @opindex mcallgraph-data
9086 @opindex mcallgraph-data
9087 @opindex mno-callgraph-data
9088 Emit callgraph information.
9092 @itemx -mno-slow-bytes
9093 @opindex mslow-bytes
9094 @opindex mslow-bytes
9095 @opindex mno-slow-bytes
9096 Prefer word access when reading byte quantities.
9098 @item -mlittle-endian
9099 @itemx -mlittle-endian
9101 @opindex mlittle-endian
9102 @opindex mlittle-endian
9103 @opindex mbig-endian
9104 Generate code for a little endian target.
9112 Generate code for the 210 processor.
9116 @subsection IA-64 Options
9117 @cindex IA-64 Options
9119 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9123 @opindex mbig-endian
9124 Generate code for a big endian target. This is the default for HPUX@.
9126 @item -mlittle-endian
9127 @opindex mlittle-endian
9128 Generate code for a little endian target. This is the default for AIX5
9135 Generate (or don't) code for the GNU assembler. This is the default.
9136 @c Also, this is the default if the configure option @option{--with-gnu-as}
9143 Generate (or don't) code for the GNU linker. This is the default.
9144 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9149 Generate code that does not use a global pointer register. The result
9150 is not position independent code, and violates the IA-64 ABI@.
9152 @item -mvolatile-asm-stop
9153 @itemx -mno-volatile-asm-stop
9154 @opindex mvolatile-asm-stop
9155 @opindex mno-volatile-asm-stop
9156 Generate (or don't) a stop bit immediately before and after volatile asm
9161 Generate code that works around Itanium B step errata.
9163 @item -mregister-names
9164 @itemx -mno-register-names
9165 @opindex mregister-names
9166 @opindex mno-register-names
9167 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9168 the stacked registers. This may make assembler output more readable.
9174 Disable (or enable) optimizations that use the small data section. This may
9175 be useful for working around optimizer bugs.
9178 @opindex mconstant-gp
9179 Generate code that uses a single constant global pointer value. This is
9180 useful when compiling kernel code.
9184 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9185 This is useful when compiling firmware code.
9187 @item -minline-divide-min-latency
9188 @opindex minline-divide-min-latency
9189 Generate code for inline divides using the minimum latency algorithm.
9191 @item -minline-divide-max-throughput
9192 @opindex minline-divide-max-throughput
9193 Generate code for inline divides using the maximum throughput algorithm.
9195 @item -mno-dwarf2-asm
9197 @opindex mno-dwarf2-asm
9198 @opindex mdwarf2-asm
9199 Don't (or do) generate assembler code for the DWARF2 line number debugging
9200 info. This may be useful when not using the GNU assembler.
9202 @item -mfixed-range=@var{register-range}
9203 @opindex mfixed-range
9204 Generate code treating the given register range as fixed registers.
9205 A fixed register is one that the register allocator can not use. This is
9206 useful when compiling kernel code. A register range is specified as
9207 two registers separated by a dash. Multiple register ranges can be
9208 specified separated by a comma.
9212 @subsection D30V Options
9213 @cindex D30V Options
9215 These @samp{-m} options are defined for D30V implementations:
9220 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
9221 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9222 memory, which starts at location @code{0x80000000}.
9226 Same as the @option{-mextmem} switch.
9230 Link the @samp{.text} section into onchip text memory, which starts at
9231 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9232 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9233 into onchip data memory, which starts at location @code{0x20000000}.
9235 @item -mno-asm-optimize
9236 @itemx -masm-optimize
9237 @opindex mno-asm-optimize
9238 @opindex masm-optimize
9239 Disable (enable) passing @option{-O} to the assembler when optimizing.
9240 The assembler uses the @option{-O} option to automatically parallelize
9241 adjacent short instructions where possible.
9243 @item -mbranch-cost=@var{n}
9244 @opindex mbranch-cost
9245 Increase the internal costs of branches to @var{n}. Higher costs means
9246 that the compiler will issue more instructions to avoid doing a branch.
9249 @item -mcond-exec=@var{n}
9251 Specify the maximum number of conditionally executed instructions that
9252 replace a branch. The default is 4.
9255 @node S/390 and zSeries Options
9256 @subsection S/390 and zSeries Options
9257 @cindex S/390 and zSeries Options
9259 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9264 @opindex mhard-float
9265 @opindex msoft-float
9266 Use (do not use) the hardware floating-point instructions and registers
9267 for floating-point operations. When @option{-msoft-float} is specified,
9268 functions in @file{libgcc.a} will be used to perform floating-point
9269 operations. When @option{-mhard-float} is specified, the compiler
9270 generates IEEE floating-point instructions. This is the default.
9273 @itemx -mno-backchain
9275 @opindex mno-backchain
9276 Generate (or do not generate) code which maintains an explicit
9277 backchain within the stack frame that points to the caller's frame.
9278 This is currently needed to allow debugging. The default is to
9279 generate the backchain.
9282 @itemx -mno-small-exec
9283 @opindex msmall-exec
9284 @opindex mno-small-exec
9285 Generate (or do not generate) code using the @code{bras} instruction
9286 to do subroutine calls.
9287 This only works reliably if the total executable size does not
9288 exceed 64k. The default is to use the @code{basr} instruction instead,
9289 which does not have this limitation.
9295 When @option{-m31} is specified, generate code compliant to the
9296 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9297 code compliant to the Linux for zSeries ABI@. This allows GCC in
9298 particular to generate 64-bit instructions. For the @samp{s390}
9299 targets, the default is @option{-m31}, while the @samp{s390x}
9300 targets default to @option{-m64}.
9306 Generate (or do not generate) code using the @code{mvcle} instruction
9307 to perform block moves. When @option{-mno-mvcle} is specifed,
9308 use a @code{mvc} loop instead. This is the default.
9314 Print (or do not print) additional debug information when compiling.
9315 The default is to not print debug information.
9320 @subsection CRIS Options
9321 @cindex CRIS Options
9323 These options are defined specifically for the CRIS ports.
9326 @item -march=@var{architecture-type}
9327 @itemx -mcpu=@var{architecture-type}
9330 Generate code for the specified architecture. The choices for
9331 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9332 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9333 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9336 @item -mtune=@var{architecture-type}
9338 Tune to @var{architecture-type} everything applicable about the generated
9339 code, except for the ABI and the set of available instructions. The
9340 choices for @var{architecture-type} are the same as for
9341 @option{-march=@var{architecture-type}}.
9343 @item -mmax-stack-frame=@var{n}
9344 @opindex mmax-stack-frame
9345 Warn when the stack frame of a function exceeds @var{n} bytes.
9347 @item -melinux-stacksize=@var{n}
9348 @opindex melinux-stacksize
9349 Only available with the @samp{cris-axis-aout} target. Arranges for
9350 indications in the program to the kernel loader that the stack of the
9351 program should be set to @var{n} bytes.
9357 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9358 @option{-march=v3} and @option{-march=v8} respectively.
9362 Enable CRIS-specific verbose debug-related information in the assembly
9363 code. This option also has the effect to turn off the @samp{#NO_APP}
9364 formatted-code indicator to the assembler at the beginning of the
9369 Do not use condition-code results from previous instruction; always emit
9370 compare and test instructions before use of condition codes.
9372 @item -mno-side-effects
9373 @opindex mno-side-effects
9374 Do not emit instructions with side-effects in addressing modes other than
9378 @itemx -mno-stack-align
9380 @itemx -mno-data-align
9381 @itemx -mconst-align
9382 @itemx -mno-const-align
9383 @opindex mstack-align
9384 @opindex mno-stack-align
9385 @opindex mdata-align
9386 @opindex mno-data-align
9387 @opindex mconst-align
9388 @opindex mno-const-align
9389 These options (no-options) arranges (eliminate arrangements) for the
9390 stack-frame, individual data and constants to be aligned for the maximum
9391 single data access size for the chosen CPU model. The default is to
9392 arrange for 32-bit alignment. ABI details such as structure layout are
9393 not affected by these options.
9401 Similar to the stack- data- and const-align options above, these options
9402 arrange for stack-frame, writable data and constants to all be 32-bit,
9403 16-bit or 8-bit aligned. The default is 32-bit alignment.
9405 @item -mno-prologue-epilogue
9406 @itemx -mprologue-epilogue
9407 @opindex mno-prologue-epilogue
9408 @opindex mprologue-epilogue
9409 With @option{-mno-prologue-epilogue}, the normal function prologue and
9410 epilogue that sets up the stack-frame are omitted and no return
9411 instructions or return sequences are generated in the code. Use this
9412 option only together with visual inspection of the compiled code: no
9413 warnings or errors are generated when call-saved registers must be saved,
9414 or storage for local variable needs to be allocated.
9420 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9421 instruction sequences that load addresses for functions from the PLT part
9422 of the GOT rather than (traditional on other architectures) calls to the
9423 PLT. The default is @option{-mgotplt}.
9427 Legacy no-op option only recognized with the cris-axis-aout target.
9431 Legacy no-op option only recognized with the cris-axis-elf and
9432 cris-axis-linux-gnu targets.
9436 Only recognized with the cris-axis-aout target, where it selects a
9437 GNU/linux-like multilib, include files and instruction set for
9442 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9446 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9447 to link with input-output functions from a simulator library. Code,
9448 initialized data and zero-initialized data are allocated consecutively.
9452 Like @option{-sim}, but pass linker options to locate initialized data at
9453 0x40000000 and zero-initialized data at 0x80000000.
9457 @subsection MMIX Options
9458 @cindex MMIX Options
9460 These options are defined for the MMIX:
9464 @itemx -mno-libfuncs
9466 @opindex mno-libfuncs
9467 Specify that intrinsic library functions are being compiled, passing all
9468 values in registers, no matter the size.
9473 @opindex mno-epsilon
9474 Generate floating-point comparison instructions that compare with respect
9475 to the @code{rE} epsilon register.
9477 @item -mabi=mmixware
9479 @opindex mabi-mmixware
9481 Generate code that passes function parameters and return values that (in
9482 the called function) are seen as registers @code{$0} and up, as opposed to
9483 the GNU ABI which uses global registers @code{$231} and up.
9486 @itemx -mno-zero-extend
9487 @opindex mzero-extend
9488 @opindex mno-zero-extend
9489 When reading data from memory in sizes shorter than 64 bits, use (do not
9490 use) zero-extending load instructions by default, rather than
9491 sign-extending ones.
9494 @itemx -mno-knuthdiv
9496 @opindex mno-knuthdiv
9497 Make the result of a division yielding a remainder have the same sign as
9498 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9499 remainder follows the sign of the dividend. Both methods are
9500 arithmetically valid, the latter being almost exclusively used.
9502 @item -mtoplevel-symbols
9503 @itemx -mno-toplevel-symbols
9504 @opindex mtoplevel-symbols
9505 @opindex mno-toplevel-symbols
9506 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9507 code can be used with the @code{PREFIX} assembly directive.
9511 Generate an executable in the ELF format, rather than the default
9512 @samp{mmo} format used by the @command{mmix} simulator.
9514 @item -mbranch-predict
9515 @itemx -mno-branch-predict
9516 @opindex mbranch-predict
9517 @opindex mno-branch-predict
9518 Use (do not use) the probable-branch instructions, when static branch
9519 prediction indicates a probable branch.
9521 @item -mbase-addresses
9522 @itemx -mno-base-addresses
9523 @opindex mbase-addresses
9524 @opindex mno-base-addresses
9525 Generate (do not generate) code that uses @emph{base addresses}. Using a
9526 base address automatically generates a request (handled by the assembler
9527 and the linker) for a constant to be set up in a global register. The
9528 register is used for one or more base address requests within the range 0
9529 to 255 from the value held in the register. The generally leads to short
9530 and fast code, but the number of different data items that can be
9531 addressed is limited. This means that a program that uses lots of static
9532 data may require @option{-mno-base-addresses}.
9535 @node PDP-11 Options
9536 @subsection PDP-11 Options
9537 @cindex PDP-11 Options
9539 These options are defined for the PDP-11:
9544 Use hardware FPP floating point. This is the default. (FIS floating
9545 point on the PDP-11/40 is not supported.)
9548 @opindex msoft-float
9549 Do not use hardware floating point.
9553 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9557 Return floating-point results in memory. This is the default.
9561 Generate code for a PDP-11/40.
9565 Generate code for a PDP-11/45. This is the default.
9569 Generate code for a PDP-11/10.
9571 @item -mbcopy-builtin
9572 @opindex bcopy-builtin
9573 Use inline @code{movstrhi} patterns for copying memory. This is the
9578 Do not use inline @code{movstrhi} patterns for copying memory.
9584 Use 16-bit @code{int}. This is the default.
9590 Use 32-bit @code{int}.
9595 @opindex mno-float32
9596 Use 64-bit @code{float}. This is the default.
9601 @opindex mno-float64
9602 Use 32-bit @code{float}.
9606 Use @code{abshi2} pattern. This is the default.
9610 Do not use @code{abshi2} pattern.
9612 @item -mbranch-expensive
9613 @opindex mbranch-expensive
9614 Pretend that branches are expensive. This is for experimenting with
9615 code generation only.
9617 @item -mbranch-cheap
9618 @opindex mbranch-cheap
9619 Do not pretend that branches are expensive. This is the default.
9623 Generate code for a system with split I&D.
9627 Generate code for a system without split I&D. This is the default.
9631 Use Unix assembler syntax. This is the default when configured for
9636 Use DEC assembler syntax. This is the default when configured for any
9637 PDP-11 target other than @samp{pdp11-*-bsd}.
9640 @node Xstormy16 Options
9641 @subsection Xstormy16 Options
9642 @cindex Xstormy16 Options
9644 These options are defined for Xstormy16:
9649 Choose startup files and linker script suitable for the simulator.
9652 @node Xtensa Options
9653 @subsection Xtensa Options
9654 @cindex Xtensa Options
9656 The Xtensa architecture is designed to support many different
9657 configurations. The compiler's default options can be set to match a
9658 particular Xtensa configuration by copying a configuration file into the
9659 GCC sources when building GCC@. The options below may be used to
9660 override the default options.
9664 @itemx -mlittle-endian
9665 @opindex mbig-endian
9666 @opindex mlittle-endian
9667 Specify big-endian or little-endian byte ordering for the target Xtensa
9673 @opindex mno-density
9674 Enable or disable use of the optional Xtensa code density instructions.
9680 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
9681 will generate MAC16 instructions from standard C code, with the
9682 limitation that it will use neither the MR register file nor any
9683 instruction that operates on the MR registers. When this option is
9684 disabled, GCC will translate 16-bit multiply/accumulate operations to a
9685 combination of core instructions and library calls, depending on whether
9686 any other multiplier options are enabled.
9692 Enable or disable use of the 16-bit integer multiplier option. When
9693 enabled, the compiler will generate 16-bit multiply instructions for
9694 multiplications of 16 bits or smaller in standard C code. When this
9695 option is disabled, the compiler will either use 32-bit multiply or
9696 MAC16 instructions if they are available or generate library calls to
9697 perform the multiply operations using shifts and adds.
9703 Enable or disable use of the 32-bit integer multiplier option. When
9704 enabled, the compiler will generate 32-bit multiply instructions for
9705 multiplications of 32 bits or smaller in standard C code. When this
9706 option is disabled, the compiler will generate library calls to perform
9707 the multiply operations using either shifts and adds or 16-bit multiply
9708 instructions if they are available.
9714 Enable or disable use of the optional normalization shift amount
9715 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
9721 Enable or disable use of the optional minimum and maximum value
9728 Enable or disable use of the optional sign extend (@code{SEXT})
9732 @itemx -mno-booleans
9734 @opindex mno-booleans
9735 Enable or disable support for the boolean register file used by Xtensa
9736 coprocessors. This is not typically useful by itself but may be
9737 required for other options that make use of the boolean registers (e.g.,
9738 the floating-point option).
9742 @opindex mhard-float
9743 @opindex msoft-float
9744 Enable or disable use of the floating-point option. When enabled, GCC
9745 generates floating-point instructions for 32-bit @code{float}
9746 operations. When this option is disabled, GCC generates library calls
9747 to emulate 32-bit floating-point operations using integer instructions.
9748 Regardless of this option, 64-bit @code{double} operations are always
9749 emulated with calls to library functions.
9752 @itemx -mno-fused-madd
9753 @opindex mfused-madd
9754 @opindex mno-fused-madd
9755 Enable or disable use of fused multiply/add and multiply/subtract
9756 instructions in the floating-point option. This has no effect if the
9757 floating-point option is not also enabled. Disabling fused multiply/add
9758 and multiply/subtract instructions forces the compiler to use separate
9759 instructions for the multiply and add/subtract operations. This may be
9760 desirable in some cases where strict IEEE 754-compliant results are
9761 required: the fused multiply add/subtract instructions do not round the
9762 intermediate result, thereby producing results with @emph{more} bits of
9763 precision than specified by the IEEE standard. Disabling fused multiply
9764 add/subtract instructions also ensures that the program output is not
9765 sensitive to the compiler's ability to combine multiply and add/subtract
9768 @item -mserialize-volatile
9769 @itemx -mno-serialize-volatile
9770 @opindex mserialize-volatile
9771 @opindex mno-serialize-volatile
9772 When this option is enabled, GCC inserts @code{MEMW} instructions before
9773 @code{volatile} memory references to guarantee sequential consistency.
9774 The default is @option{-mserialize-volatile}. Use
9775 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
9777 @item -mtext-section-literals
9778 @itemx -mno-text-section-literals
9779 @opindex mtext-section-literals
9780 @opindex mno-text-section-literals
9781 Control the treatment of literal pools. The default is
9782 @option{-mno-text-section-literals}, which places literals in a separate
9783 section in the output file. This allows the literal pool to be placed
9784 in a data RAM/ROM, and it also allows the linker to combine literal
9785 pools from separate object files to remove redundant literals and
9786 improve code size. With @option{-mtext-section-literals}, the literals
9787 are interspersed in the text section in order to keep them as close as
9788 possible to their references. This may be necessary for large assembly
9791 @item -mtarget-align
9792 @itemx -mno-target-align
9793 @opindex mtarget-align
9794 @opindex mno-target-align
9795 When this option is enabled, GCC instructs the assembler to
9796 automatically align instructions to reduce branch penalties at the
9797 expense of some code density. The assembler attempts to widen density
9798 instructions to align branch targets and the instructions following call
9799 instructions. If there are not enough preceding safe density
9800 instructions to align a target, no widening will be performed. The
9801 default is @option{-mtarget-align}. These options do not affect the
9802 treatment of auto-aligned instructions like @code{LOOP}, which the
9803 assembler will always align, either by widening density instructions or
9804 by inserting no-op instructions.
9807 @itemx -mno-longcalls
9809 @opindex mno-longcalls
9810 When this option is enabled, GCC instructs the assembler to translate
9811 direct calls to indirect calls unless it can determine that the target
9812 of a direct call is in the range allowed by the call instruction. This
9813 translation typically occurs for calls to functions in other source
9814 files. Specifically, the assembler translates a direct @code{CALL}
9815 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
9816 The default is @option{-mno-longcalls}. This option should be used in
9817 programs where the call target can potentially be out of range. This
9818 option is implemented in the assembler, not the compiler, so the
9819 assembly code generated by GCC will still show direct call
9820 instructions---look at the disassembled object code to see the actual
9821 instructions. Note that the assembler will use an indirect call for
9822 every cross-file call, not just those that really will be out of range.
9825 @node Code Gen Options
9826 @section Options for Code Generation Conventions
9827 @cindex code generation conventions
9828 @cindex options, code generation
9829 @cindex run-time options
9831 These machine-independent options control the interface conventions
9832 used in code generation.
9834 Most of them have both positive and negative forms; the negative form
9835 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9836 one of the forms is listed---the one which is not the default. You
9837 can figure out the other form by either removing @samp{no-} or adding
9842 @opindex fexceptions
9843 Enable exception handling. Generates extra code needed to propagate
9844 exceptions. For some targets, this implies GCC will generate frame
9845 unwind information for all functions, which can produce significant data
9846 size overhead, although it does not affect execution. If you do not
9847 specify this option, GCC will enable it by default for languages like
9848 C++ which normally require exception handling, and disable it for
9849 languages like C that do not normally require it. However, you may need
9850 to enable this option when compiling C code that needs to interoperate
9851 properly with exception handlers written in C++. You may also wish to
9852 disable this option if you are compiling older C++ programs that don't
9853 use exception handling.
9855 @item -fnon-call-exceptions
9856 @opindex fnon-call-exceptions
9857 Generate code that allows trapping instructions to throw exceptions.
9858 Note that this requires platform-specific runtime support that does
9859 not exist everywhere. Moreover, it only allows @emph{trapping}
9860 instructions to throw exceptions, i.e.@: memory references or floating
9861 point instructions. It does not allow exceptions to be thrown from
9862 arbitrary signal handlers such as @code{SIGALRM}.
9864 @item -funwind-tables
9865 @opindex funwind-tables
9866 Similar to @option{-fexceptions}, except that it will just generate any needed
9867 static data, but will not affect the generated code in any other way.
9868 You will normally not enable this option; instead, a language processor
9869 that needs this handling would enable it on your behalf.
9871 @item -fasynchronous-unwind-tables
9872 @opindex funwind-tables
9873 Generate unwind table in dwarf2 format, if supported by target machine. The
9874 table is exact at each instruction boundary, so it can be used for stack
9875 unwinding from asynchronous events (such as debugger or garbage collector).
9877 @item -fpcc-struct-return
9878 @opindex fpcc-struct-return
9879 Return ``short'' @code{struct} and @code{union} values in memory like
9880 longer ones, rather than in registers. This convention is less
9881 efficient, but it has the advantage of allowing intercallability between
9882 GCC-compiled files and files compiled with other compilers.
9884 The precise convention for returning structures in memory depends
9885 on the target configuration macros.
9887 Short structures and unions are those whose size and alignment match
9888 that of some integer type.
9890 @item -freg-struct-return
9891 @opindex freg-struct-return
9892 Return @code{struct} and @code{union} values in registers when possible.
9893 This is more efficient for small structures than
9894 @option{-fpcc-struct-return}.
9896 If you specify neither @option{-fpcc-struct-return} nor
9897 @option{-freg-struct-return}, GCC defaults to whichever convention is
9898 standard for the target. If there is no standard convention, GCC
9899 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9900 the principal compiler. In those cases, we can choose the standard, and
9901 we chose the more efficient register return alternative.
9904 @opindex fshort-enums
9905 Allocate to an @code{enum} type only as many bytes as it needs for the
9906 declared range of possible values. Specifically, the @code{enum} type
9907 will be equivalent to the smallest integer type which has enough room.
9909 @item -fshort-double
9910 @opindex fshort-double
9911 Use the same size for @code{double} as for @code{float}.
9914 @opindex fshared-data
9915 Requests that the data and non-@code{const} variables of this
9916 compilation be shared data rather than private data. The distinction
9917 makes sense only on certain operating systems, where shared data is
9918 shared between processes running the same program, while private data
9919 exists in one copy per process.
9923 In C, allocate even uninitialized global variables in the data section of the
9924 object file, rather than generating them as common blocks. This has the
9925 effect that if the same variable is declared (without @code{extern}) in
9926 two different compilations, you will get an error when you link them.
9927 The only reason this might be useful is if you wish to verify that the
9928 program will work on other systems which always work this way.
9932 Ignore the @samp{#ident} directive.
9934 @item -fno-gnu-linker
9935 @opindex fno-gnu-linker
9936 Do not output global initializations (such as C++ constructors and
9937 destructors) in the form used by the GNU linker (on systems where the GNU
9938 linker is the standard method of handling them). Use this option when
9939 you want to use a non-GNU linker, which also requires using the
9940 @command{collect2} program to make sure the system linker includes
9941 constructors and destructors. (@command{collect2} is included in the GCC
9942 distribution.) For systems which @emph{must} use @command{collect2}, the
9943 compiler driver @command{gcc} is configured to do this automatically.
9945 @item -finhibit-size-directive
9946 @opindex finhibit-size-directive
9947 Don't output a @code{.size} assembler directive, or anything else that
9948 would cause trouble if the function is split in the middle, and the
9949 two halves are placed at locations far apart in memory. This option is
9950 used when compiling @file{crtstuff.c}; you should not need to use it
9954 @opindex fverbose-asm
9955 Put extra commentary information in the generated assembly code to
9956 make it more readable. This option is generally only of use to those
9957 who actually need to read the generated assembly code (perhaps while
9958 debugging the compiler itself).
9960 @option{-fno-verbose-asm}, the default, causes the
9961 extra information to be omitted and is useful when comparing two assembler
9966 Consider all memory references through pointers to be volatile.
9968 @item -fvolatile-global
9969 @opindex fvolatile-global
9970 Consider all memory references to extern and global data items to
9971 be volatile. GCC does not consider static data items to be volatile
9972 because of this switch.
9974 @item -fvolatile-static
9975 @opindex fvolatile-static
9976 Consider all memory references to static data to be volatile.
9980 @cindex global offset table
9982 Generate position-independent code (PIC) suitable for use in a shared
9983 library, if supported for the target machine. Such code accesses all
9984 constant addresses through a global offset table (GOT)@. The dynamic
9985 loader resolves the GOT entries when the program starts (the dynamic
9986 loader is not part of GCC; it is part of the operating system). If
9987 the GOT size for the linked executable exceeds a machine-specific
9988 maximum size, you get an error message from the linker indicating that
9989 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9990 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9991 on the m68k and RS/6000. The 386 has no such limit.)
9993 Position-independent code requires special support, and therefore works
9994 only on certain machines. For the 386, GCC supports PIC for System V
9995 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9996 position-independent.
10000 If supported for the target machine, emit position-independent code,
10001 suitable for dynamic linking and avoiding any limit on the size of the
10002 global offset table. This option makes a difference on the m68k, m88k,
10005 Position-independent code requires special support, and therefore works
10006 only on certain machines.
10008 @item -ffixed-@var{reg}
10010 Treat the register named @var{reg} as a fixed register; generated code
10011 should never refer to it (except perhaps as a stack pointer, frame
10012 pointer or in some other fixed role).
10014 @var{reg} must be the name of a register. The register names accepted
10015 are machine-specific and are defined in the @code{REGISTER_NAMES}
10016 macro in the machine description macro file.
10018 This flag does not have a negative form, because it specifies a
10021 @item -fcall-used-@var{reg}
10022 @opindex fcall-used
10023 Treat the register named @var{reg} as an allocable register that is
10024 clobbered by function calls. It may be allocated for temporaries or
10025 variables that do not live across a call. Functions compiled this way
10026 will not save and restore the register @var{reg}.
10028 It is an error to used this flag with the frame pointer or stack pointer.
10029 Use of this flag for other registers that have fixed pervasive roles in
10030 the machine's execution model will produce disastrous results.
10032 This flag does not have a negative form, because it specifies a
10035 @item -fcall-saved-@var{reg}
10036 @opindex fcall-saved
10037 Treat the register named @var{reg} as an allocable register saved by
10038 functions. It may be allocated even for temporaries or variables that
10039 live across a call. Functions compiled this way will save and restore
10040 the register @var{reg} if they use it.
10042 It is an error to used this flag with the frame pointer or stack pointer.
10043 Use of this flag for other registers that have fixed pervasive roles in
10044 the machine's execution model will produce disastrous results.
10046 A different sort of disaster will result from the use of this flag for
10047 a register in which function values may be returned.
10049 This flag does not have a negative form, because it specifies a
10052 @item -fpack-struct
10053 @opindex fpack-struct
10054 Pack all structure members together without holes. Usually you would
10055 not want to use this option, since it makes the code suboptimal, and
10056 the offsets of structure members won't agree with system libraries.
10058 @item -finstrument-functions
10059 @opindex finstrument-functions
10060 Generate instrumentation calls for entry and exit to functions. Just
10061 after function entry and just before function exit, the following
10062 profiling functions will be called with the address of the current
10063 function and its call site. (On some platforms,
10064 @code{__builtin_return_address} does not work beyond the current
10065 function, so the call site information may not be available to the
10066 profiling functions otherwise.)
10069 void __cyg_profile_func_enter (void *this_fn,
10071 void __cyg_profile_func_exit (void *this_fn,
10075 The first argument is the address of the start of the current function,
10076 which may be looked up exactly in the symbol table.
10078 This instrumentation is also done for functions expanded inline in other
10079 functions. The profiling calls will indicate where, conceptually, the
10080 inline function is entered and exited. This means that addressable
10081 versions of such functions must be available. If all your uses of a
10082 function are expanded inline, this may mean an additional expansion of
10083 code size. If you use @samp{extern inline} in your C code, an
10084 addressable version of such functions must be provided. (This is
10085 normally the case anyways, but if you get lucky and the optimizer always
10086 expands the functions inline, you might have gotten away without
10087 providing static copies.)
10089 A function may be given the attribute @code{no_instrument_function}, in
10090 which case this instrumentation will not be done. This can be used, for
10091 example, for the profiling functions listed above, high-priority
10092 interrupt routines, and any functions from which the profiling functions
10093 cannot safely be called (perhaps signal handlers, if the profiling
10094 routines generate output or allocate memory).
10096 @item -fstack-check
10097 @opindex fstack-check
10098 Generate code to verify that you do not go beyond the boundary of the
10099 stack. You should specify this flag if you are running in an
10100 environment with multiple threads, but only rarely need to specify it in
10101 a single-threaded environment since stack overflow is automatically
10102 detected on nearly all systems if there is only one stack.
10104 Note that this switch does not actually cause checking to be done; the
10105 operating system must do that. The switch causes generation of code
10106 to ensure that the operating system sees the stack being extended.
10108 @item -fstack-limit-register=@var{reg}
10109 @itemx -fstack-limit-symbol=@var{sym}
10110 @itemx -fno-stack-limit
10111 @opindex fstack-limit-register
10112 @opindex fstack-limit-symbol
10113 @opindex fno-stack-limit
10114 Generate code to ensure that the stack does not grow beyond a certain value,
10115 either the value of a register or the address of a symbol. If the stack
10116 would grow beyond the value, a signal is raised. For most targets,
10117 the signal is raised before the stack overruns the boundary, so
10118 it is possible to catch the signal without taking special precautions.
10120 For instance, if the stack starts at absolute address @samp{0x80000000}
10121 and grows downwards, you can use the flags
10122 @option{-fstack-limit-symbol=__stack_limit} and
10123 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
10124 of 128KB@. Note that this may only work with the GNU linker.
10126 @cindex aliasing of parameters
10127 @cindex parameters, aliased
10128 @item -fargument-alias
10129 @itemx -fargument-noalias
10130 @itemx -fargument-noalias-global
10131 @opindex fargument-alias
10132 @opindex fargument-noalias
10133 @opindex fargument-noalias-global
10134 Specify the possible relationships among parameters and between
10135 parameters and global data.
10137 @option{-fargument-alias} specifies that arguments (parameters) may
10138 alias each other and may alias global storage.@*
10139 @option{-fargument-noalias} specifies that arguments do not alias
10140 each other, but may alias global storage.@*
10141 @option{-fargument-noalias-global} specifies that arguments do not
10142 alias each other and do not alias global storage.
10144 Each language will automatically use whatever option is required by
10145 the language standard. You should not need to use these options yourself.
10147 @item -fleading-underscore
10148 @opindex fleading-underscore
10149 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
10150 change the way C symbols are represented in the object file. One use
10151 is to help link with legacy assembly code.
10153 Be warned that you should know what you are doing when invoking this
10154 option, and that not all targets provide complete support for it.
10159 @node Environment Variables
10160 @section Environment Variables Affecting GCC
10161 @cindex environment variables
10163 @c man begin ENVIRONMENT
10165 This section describes several environment variables that affect how GCC
10166 operates. Some of them work by specifying directories or prefixes to use
10167 when searching for various kinds of files. Some are used to specify other
10168 aspects of the compilation environment.
10170 Note that you can also specify places to search using options such as
10171 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
10172 take precedence over places specified using environment variables, which
10173 in turn take precedence over those specified by the configuration of GCC@.
10174 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
10175 GNU Compiler Collection (GCC) Internals}.
10180 @c @itemx LC_COLLATE
10182 @c @itemx LC_MONETARY
10183 @c @itemx LC_NUMERIC
10188 @c @findex LC_COLLATE
10189 @findex LC_MESSAGES
10190 @c @findex LC_MONETARY
10191 @c @findex LC_NUMERIC
10195 These environment variables control the way that GCC uses
10196 localization information that allow GCC to work with different
10197 national conventions. GCC inspects the locale categories
10198 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
10199 so. These locale categories can be set to any value supported by your
10200 installation. A typical value is @samp{en_UK} for English in the United
10203 The @env{LC_CTYPE} environment variable specifies character
10204 classification. GCC uses it to determine the character boundaries in
10205 a string; this is needed for some multibyte encodings that contain quote
10206 and escape characters that would otherwise be interpreted as a string
10209 The @env{LC_MESSAGES} environment variable specifies the language to
10210 use in diagnostic messages.
10212 If the @env{LC_ALL} environment variable is set, it overrides the value
10213 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10214 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10215 environment variable. If none of these variables are set, GCC
10216 defaults to traditional C English behavior.
10220 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10221 files. GCC uses temporary files to hold the output of one stage of
10222 compilation which is to be used as input to the next stage: for example,
10223 the output of the preprocessor, which is the input to the compiler
10226 @item GCC_EXEC_PREFIX
10227 @findex GCC_EXEC_PREFIX
10228 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10229 names of the subprograms executed by the compiler. No slash is added
10230 when this prefix is combined with the name of a subprogram, but you can
10231 specify a prefix that ends with a slash if you wish.
10233 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10234 an appropriate prefix to use based on the pathname it was invoked with.
10236 If GCC cannot find the subprogram using the specified prefix, it
10237 tries looking in the usual places for the subprogram.
10239 The default value of @env{GCC_EXEC_PREFIX} is
10240 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10241 of @code{prefix} when you ran the @file{configure} script.
10243 Other prefixes specified with @option{-B} take precedence over this prefix.
10245 This prefix is also used for finding files such as @file{crt0.o} that are
10248 In addition, the prefix is used in an unusual way in finding the
10249 directories to search for header files. For each of the standard
10250 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10251 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10252 replacing that beginning with the specified prefix to produce an
10253 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10254 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10255 These alternate directories are searched first; the standard directories
10258 @item COMPILER_PATH
10259 @findex COMPILER_PATH
10260 The value of @env{COMPILER_PATH} is a colon-separated list of
10261 directories, much like @env{PATH}. GCC tries the directories thus
10262 specified when searching for subprograms, if it can't find the
10263 subprograms using @env{GCC_EXEC_PREFIX}.
10266 @findex LIBRARY_PATH
10267 The value of @env{LIBRARY_PATH} is a colon-separated list of
10268 directories, much like @env{PATH}. When configured as a native compiler,
10269 GCC tries the directories thus specified when searching for special
10270 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10271 using GCC also uses these directories when searching for ordinary
10272 libraries for the @option{-l} option (but directories specified with
10273 @option{-L} come first).
10275 @item C_INCLUDE_PATH
10276 @itemx CPLUS_INCLUDE_PATH
10277 @itemx OBJC_INCLUDE_PATH
10278 @findex C_INCLUDE_PATH
10279 @findex CPLUS_INCLUDE_PATH
10280 @findex OBJC_INCLUDE_PATH
10281 @c @itemx OBJCPLUS_INCLUDE_PATH
10282 These environment variables pertain to particular languages. Each
10283 variable's value is a colon-separated list of directories, much like
10284 @env{PATH}. When GCC searches for header files, it tries the
10285 directories listed in the variable for the language you are using, after
10286 the directories specified with @option{-I} but before the standard header
10289 @item DEPENDENCIES_OUTPUT
10290 @findex DEPENDENCIES_OUTPUT
10291 @cindex dependencies for make as output
10292 If this variable is set, its value specifies how to output dependencies
10293 for Make based on the header files processed by the compiler. This
10294 output looks much like the output from the @option{-M} option
10295 (@pxref{Preprocessor Options}), but it goes to a separate file, and is
10296 in addition to the usual results of compilation.
10298 The value of @env{DEPENDENCIES_OUTPUT} can be just a file name, in
10299 which case the Make rules are written to that file, guessing the target
10300 name from the source file name. Or the value can have the form
10301 @samp{@var{file} @var{target}}, in which case the rules are written to
10302 file @var{file} using @var{target} as the target name.
10306 @cindex locale definition
10307 This variable is used to pass locale information to the compiler. One way in
10308 which this information is used is to determine the character set to be used
10309 when character literals, string literals and comments are parsed in C and C++.
10310 When the compiler is configured to allow multibyte characters,
10311 the following values for @env{LANG} are recognized:
10315 Recognize JIS characters.
10317 Recognize SJIS characters.
10319 Recognize EUCJP characters.
10322 If @env{LANG} is not defined, or if it has some other value, then the
10323 compiler will use mblen and mbtowc as defined by the default locale to
10324 recognize and translate multibyte characters.
10329 @node Running Protoize
10330 @section Running Protoize
10332 The program @code{protoize} is an optional part of GCC@. You can use
10333 it to add prototypes to a program, thus converting the program to ISO
10334 C in one respect. The companion program @code{unprotoize} does the
10335 reverse: it removes argument types from any prototypes that are found.
10337 When you run these programs, you must specify a set of source files as
10338 command line arguments. The conversion programs start out by compiling
10339 these files to see what functions they define. The information gathered
10340 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10342 After scanning comes actual conversion. The specified files are all
10343 eligible to be converted; any files they include (whether sources or
10344 just headers) are eligible as well.
10346 But not all the eligible files are converted. By default,
10347 @code{protoize} and @code{unprotoize} convert only source and header
10348 files in the current directory. You can specify additional directories
10349 whose files should be converted with the @option{-d @var{directory}}
10350 option. You can also specify particular files to exclude with the
10351 @option{-x @var{file}} option. A file is converted if it is eligible, its
10352 directory name matches one of the specified directory names, and its
10353 name within the directory has not been excluded.
10355 Basic conversion with @code{protoize} consists of rewriting most
10356 function definitions and function declarations to specify the types of
10357 the arguments. The only ones not rewritten are those for varargs
10360 @code{protoize} optionally inserts prototype declarations at the
10361 beginning of the source file, to make them available for any calls that
10362 precede the function's definition. Or it can insert prototype
10363 declarations with block scope in the blocks where undeclared functions
10366 Basic conversion with @code{unprotoize} consists of rewriting most
10367 function declarations to remove any argument types, and rewriting
10368 function definitions to the old-style pre-ISO form.
10370 Both conversion programs print a warning for any function declaration or
10371 definition that they can't convert. You can suppress these warnings
10374 The output from @code{protoize} or @code{unprotoize} replaces the
10375 original source file. The original file is renamed to a name ending
10376 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10377 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10378 for DOS) file already exists, then the source file is simply discarded.
10380 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10381 scan the program and collect information about the functions it uses.
10382 So neither of these programs will work until GCC is installed.
10384 Here is a table of the options you can use with @code{protoize} and
10385 @code{unprotoize}. Each option works with both programs unless
10389 @item -B @var{directory}
10390 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10391 usual directory (normally @file{/usr/local/lib}). This file contains
10392 prototype information about standard system functions. This option
10393 applies only to @code{protoize}.
10395 @item -c @var{compilation-options}
10396 Use @var{compilation-options} as the options when running @code{gcc} to
10397 produce the @samp{.X} files. The special option @option{-aux-info} is
10398 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10400 Note that the compilation options must be given as a single argument to
10401 @code{protoize} or @code{unprotoize}. If you want to specify several
10402 @code{gcc} options, you must quote the entire set of compilation options
10403 to make them a single word in the shell.
10405 There are certain @code{gcc} arguments that you cannot use, because they
10406 would produce the wrong kind of output. These include @option{-g},
10407 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10408 the @var{compilation-options}, they are ignored.
10411 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10412 systems) instead of @samp{.c}. This is convenient if you are converting
10413 a C program to C++. This option applies only to @code{protoize}.
10416 Add explicit global declarations. This means inserting explicit
10417 declarations at the beginning of each source file for each function
10418 that is called in the file and was not declared. These declarations
10419 precede the first function definition that contains a call to an
10420 undeclared function. This option applies only to @code{protoize}.
10422 @item -i @var{string}
10423 Indent old-style parameter declarations with the string @var{string}.
10424 This option applies only to @code{protoize}.
10426 @code{unprotoize} converts prototyped function definitions to old-style
10427 function definitions, where the arguments are declared between the
10428 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10429 uses five spaces as the indentation. If you want to indent with just
10430 one space instead, use @option{-i " "}.
10433 Keep the @samp{.X} files. Normally, they are deleted after conversion
10437 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10438 a prototype declaration for each function in each block which calls the
10439 function without any declaration. This option applies only to
10443 Make no real changes. This mode just prints information about the conversions
10444 that would have been done without @option{-n}.
10447 Make no @samp{.save} files. The original files are simply deleted.
10448 Use this option with caution.
10450 @item -p @var{program}
10451 Use the program @var{program} as the compiler. Normally, the name
10452 @file{gcc} is used.
10455 Work quietly. Most warnings are suppressed.
10458 Print the version number, just like @option{-v} for @code{gcc}.
10461 If you need special compiler options to compile one of your program's
10462 source files, then you should generate that file's @samp{.X} file
10463 specially, by running @code{gcc} on that source file with the
10464 appropriate options and the option @option{-aux-info}. Then run
10465 @code{protoize} on the entire set of files. @code{protoize} will use
10466 the existing @samp{.X} file because it is newer than the source file.
10470 gcc -Dfoo=bar file1.c -aux-info file1.X
10475 You need to include the special files along with the rest in the
10476 @code{protoize} command, even though their @samp{.X} files already
10477 exist, because otherwise they won't get converted.
10479 @xref{Protoize Caveats}, for more information on how to use
10480 @code{protoize} successfully.