1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002 Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
8 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
9 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
11 Permission is granted to copy, distribute and/or modify this document
12 under the terms of the GNU Free Documentation License, Version 1.1 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU project C and C++ compiler
33 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
39 [@option{-o} @var{outfile}] @var{infile}@dots{}
41 Only the most useful options are listed here; see below for the
42 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
45 gpl(7), gfdl(7), fsf-funding(7),
46 cpp(1), gcov(1), g77(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
47 and the Info entries for @file{gcc}, @file{cpp}, @file{g77}, @file{as},
48 @file{ld}, @file{binutils} and @file{gdb}.
51 For instructions on reporting bugs, see
52 @w{@uref{http://gcc.gnu.org/bugs.html}}. Use of the @command{gccbug}
53 script to report bugs is recommended.
56 See the Info entry for @command{gcc}, or
57 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
58 for contributors to GCC@.
63 @chapter GCC Command Options
64 @cindex GCC command options
65 @cindex command options
66 @cindex options, GCC command
68 @c man begin DESCRIPTION
70 When you invoke GCC, it normally does preprocessing, compilation,
71 assembly and linking. The ``overall options'' allow you to stop this
72 process at an intermediate stage. For example, the @option{-c} option
73 says not to run the linker. Then the output consists of object files
74 output by the assembler.
76 Other options are passed on to one stage of processing. Some options
77 control the preprocessor and others the compiler itself. Yet other
78 options control the assembler and linker; most of these are not
79 documented here, since you rarely need to use any of them.
81 @cindex C compilation options
82 Most of the command line options that you can use with GCC are useful
83 for C programs; when an option is only useful with another language
84 (usually C++), the explanation says so explicitly. If the description
85 for a particular option does not mention a source language, you can use
86 that option with all supported languages.
88 @cindex C++ compilation options
89 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
90 options for compiling C++ programs.
92 @cindex grouping options
93 @cindex options, grouping
94 The @command{gcc} program accepts options and file names as operands. Many
95 options have multi-letter names; therefore multiple single-letter options
96 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
99 @cindex order of options
100 @cindex options, order
101 You can mix options and other arguments. For the most part, the order
102 you use doesn't matter. Order does matter when you use several options
103 of the same kind; for example, if you specify @option{-L} more than once,
104 the directories are searched in the order specified.
106 Many options have long names starting with @samp{-f} or with
107 @samp{-W}---for example, @option{-fforce-mem},
108 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
109 these have both positive and negative forms; the negative form of
110 @option{-ffoo} would be @option{-fno-foo}. This manual documents
111 only one of these two forms, whichever one is not the default.
115 @xref{Option Index}, for an index to GCC's options.
118 * Option Summary:: Brief list of all options, without explanations.
119 * Overall Options:: Controlling the kind of output:
120 an executable, object files, assembler files,
121 or preprocessed source.
122 * Invoking G++:: Compiling C++ programs.
123 * C Dialect Options:: Controlling the variant of C language compiled.
124 * C++ Dialect Options:: Variations on C++.
125 * Objective-C Dialect Options:: Variations on Objective-C.
126 * Language Independent Options:: Controlling how diagnostics should be
128 * Warning Options:: How picky should the compiler be?
129 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
130 * Optimize Options:: How much optimization?
131 * Preprocessor Options:: Controlling header files and macro definitions.
132 Also, getting dependency information for Make.
133 * Assembler Options:: Passing options to the assembler.
134 * Link Options:: Specifying libraries and so on.
135 * Directory Options:: Where to find header files and libraries.
136 Where to find the compiler executable files.
137 * Spec Files:: How to pass switches to sub-processes.
138 * Target Options:: Running a cross-compiler, or an old version of GCC.
139 * Submodel Options:: Specifying minor hardware or convention variations,
140 such as 68010 vs 68020.
141 * Code Gen Options:: Specifying conventions for function calls, data layout
143 * Environment Variables:: Env vars that affect GCC.
144 * Running Protoize:: Automatically adding or removing function prototypes.
150 @section Option Summary
152 Here is a summary of all the options, grouped by type. Explanations are
153 in the following sections.
156 @item Overall Options
157 @xref{Overall Options,,Options Controlling the Kind of Output}.
159 -c -S -E -o @var{file} -pipe -pass-exit-codes -x @var{language} @gol
160 -v --target-help --help}
162 @item C Language Options
163 @xref{C Dialect Options,,Options Controlling C Dialect}.
165 -ansi -std=@var{standard} -aux-info @var{filename} @gol
166 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
167 -fhosted -ffreestanding @gol
168 -trigraphs -traditional -traditional-cpp @gol
169 -fallow-single-precision -fcond-mismatch @gol
170 -fsigned-bitfields -fsigned-char @gol
171 -funsigned-bitfields -funsigned-char @gol
172 -fwritable-strings -fshort-wchar}
174 @item C++ Language Options
175 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
177 -fno-access-control -fcheck-new -fconserve-space @gol
178 -fno-const-strings -fdollars-in-identifiers @gol
179 -fno-elide-constructors @gol
180 -fno-enforce-eh-specs -fexternal-templates @gol
181 -falt-external-templates @gol
182 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
183 -fno-implicit-templates @gol
184 -fno-implicit-inline-templates @gol
185 -fno-implement-inlines -fms-extensions @gol
186 -fno-nonansi-builtins -fno-operator-names @gol
187 -fno-optional-diags -fpermissive @gol
188 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
189 -fuse-cxa-atexit -fvtable-gc -fno-weak -nostdinc++ @gol
190 -fno-default-inline -Wctor-dtor-privacy @gol
191 -Wnon-virtual-dtor -Wreorder @gol
192 -Weffc++ -Wno-deprecated @gol
193 -Wno-non-template-friend -Wold-style-cast @gol
194 -Woverloaded-virtual -Wno-pmf-conversions @gol
195 -Wsign-promo -Wsynth}
197 @item Objective-C Language Options
198 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
200 -fconstant-string-class=@var{class-name} @gol
201 -fgnu-runtime -fnext-runtime -gen-decls @gol
202 -Wno-protocol -Wselector}
204 @item Language Independent Options
205 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
207 -fmessage-length=@var{n} @gol
208 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
210 @item Warning Options
211 @xref{Warning Options,,Options to Request or Suppress Warnings}.
213 -fsyntax-only -pedantic -pedantic-errors @gol
214 -w -W -Wall -Waggregate-return @gol
215 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
216 -Wconversion -Wno-deprecated-declarations @gol
217 -Wdisabled-optimization -Wdiv-by-zero -Werror @gol
218 -Wfloat-equal -Wformat -Wformat=2 @gol
219 -Wformat-nonliteral -Wformat-security @gol
220 -Wimplicit -Wimplicit-int @gol
221 -Wimplicit-function-declaration @gol
222 -Werror-implicit-function-declaration @gol
223 -Wimport -Winline @gol
224 -Wlarger-than-@var{len} -Wlong-long @gol
225 -Wmain -Wmissing-braces -Wmissing-declarations @gol
226 -Wmissing-format-attribute -Wmissing-noreturn @gol
227 -Wmultichar -Wno-format-extra-args -Wno-format-y2k @gol
228 -Wno-import -Wpacked -Wpadded @gol
229 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
230 -Wreturn-type -Wsequence-point -Wshadow @gol
231 -Wsign-compare -Wswitch -Wsystem-headers @gol
232 -Wtrigraphs -Wundef -Wuninitialized @gol
233 -Wunknown-pragmas -Wunreachable-code @gol
234 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
235 -Wunused-value -Wunused-variable -Wwrite-strings}
237 @item C-only Warning Options
239 -Wbad-function-cast -Wmissing-prototypes -Wnested-externs @gol
240 -Wstrict-prototypes -Wtraditional}
242 @item Debugging Options
243 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
245 -d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
247 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
248 -fdump-tree-original@r{[}-@var{n}@r{]} -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
249 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
250 -fmem-report -fpretend-float @gol
251 -fprofile-arcs -ftest-coverage -ftime-report @gol
252 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
253 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
254 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
255 -print-multi-directory -print-multi-lib @gol
256 -print-prog-name=@var{program} -print-search-dirs -Q @gol
259 @item Optimization Options
260 @xref{Optimize Options,,Options that Control Optimization}.
262 -falign-functions=@var{n} -falign-jumps=@var{n} @gol
263 -falign-labels=@var{n} -falign-loops=@var{n} @gol
264 -fbranch-probabilities -fcaller-saves -fcprop-registers @gol
265 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
266 -fdelayed-branch -fdelete-null-pointer-checks @gol
267 -fexpensive-optimizations -ffast-math -ffloat-store @gol
268 -fforce-addr -fforce-mem -ffunction-sections @gol
269 -fgcse -fgcse-lm -fgcse-sm @gol
270 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
271 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
272 -fmove-all-movables -fno-default-inline -fno-defer-pop @gol
273 -fno-function-cse -fno-guess-branch-probability @gol
274 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
275 -funsafe-math-optimizations -fno-trapping-math @gol
276 -fomit-frame-pointer -foptimize-register-move @gol
277 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
278 -freduce-all-givs -fregmove -frename-registers @gol
279 -frerun-cse-after-loop -frerun-loop-opt @gol
280 -fschedule-insns -fschedule-insns2 @gol
281 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
282 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol
283 -funroll-all-loops -funroll-loops @gol
284 --param @var{name}=@var{value}
285 -O -O0 -O1 -O2 -O3 -Os}
287 @item Preprocessor Options
288 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
290 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
291 -C -dD -dI -dM -dN @gol
292 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
293 -idirafter @var{dir} @gol
294 -include @var{file} -imacros @var{file} @gol
295 -iprefix @var{file} -iwithprefix @var{dir} @gol
296 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
297 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
298 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
300 @item Assembler Option
301 @xref{Assembler Options,,Passing Options to the Assembler}.
306 @xref{Link Options,,Options for Linking}.
308 @var{object-file-name} -l@var{library} @gol
309 -nostartfiles -nodefaultlibs -nostdlib @gol
310 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
311 -Wl,@var{option} -Xlinker @var{option} @gol
314 @item Directory Options
315 @xref{Directory Options,,Options for Directory Search}.
317 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
320 @c I wrote this xref this way to avoid overfull hbox. -- rms
321 @xref{Target Options}.
323 -b @var{machine} -V @var{version}}
325 @item Machine Dependent Options
326 @xref{Submodel Options,,Hardware Models and Configurations}.
328 @emph{M680x0 Options}
330 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
331 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
332 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
333 -malign-int -mstrict-align}
335 @emph{M68hc1x Options}
337 -m6811 -m6812 -m68hc11 -m68hc12 @gol
338 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
346 -mcpu=@var{cpu-type} @gol
347 -mtune=@var{cpu-type} @gol
348 -mcmodel=@var{code-model} @gol
350 -mapp-regs -mbroken-saverestore -mcypress @gol
351 -mepilogue -mfaster-structs -mflat @gol
352 -mfpu -mhard-float -mhard-quad-float @gol
353 -mimpure-text -mlive-g0 -mno-app-regs @gol
354 -mno-epilogue -mno-faster-structs -mno-flat -mno-fpu @gol
355 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
356 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
357 -msupersparc -munaligned-doubles -mv8}
359 @emph{Convex Options}
361 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
362 -margcount -mnoargcount @gol
363 -mlong32 -mlong64 @gol
364 -mvolatile-cache -mvolatile-nocache}
366 @emph{AMD29K Options}
368 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
369 -mlarge -mnormal -msmall @gol
370 -mkernel-registers -mno-reuse-arg-regs @gol
371 -mno-stack-check -mno-storem-bug @gol
372 -mreuse-arg-regs -msoft-float -mstack-check @gol
373 -mstorem-bug -muser-registers}
377 -mapcs-frame -mno-apcs-frame @gol
378 -mapcs-26 -mapcs-32 @gol
379 -mapcs-stack-check -mno-apcs-stack-check @gol
380 -mapcs-float -mno-apcs-float @gol
381 -mapcs-reentrant -mno-apcs-reentrant @gol
382 -msched-prolog -mno-sched-prolog @gol
383 -mlittle-endian -mbig-endian -mwords-little-endian @gol
384 -malignment-traps -mno-alignment-traps @gol
385 -msoft-float -mhard-float -mfpe @gol
386 -mthumb-interwork -mno-thumb-interwork @gol
387 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
388 -mstructure-size-boundary=@var{n} @gol
389 -mbsd -mxopen -mno-symrename @gol
390 -mabort-on-noreturn @gol
391 -mlong-calls -mno-long-calls @gol
392 -msingle-pic-base -mno-single-pic-base @gol
393 -mpic-register=@var{reg} @gol
394 -mnop-fun-dllimport @gol
395 -mpoke-function-name @gol
397 -mtpcs-frame -mtpcs-leaf-frame @gol
398 -mcaller-super-interworking -mcallee-super-interworking }
400 @emph{MN10200 Options}
404 @emph{MN10300 Options}
406 -mmult-bug -mno-mult-bug @gol
407 -mam33 -mno-am33 @gol
410 @emph{M32R/D Options}
412 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
417 -m88000 -m88100 -m88110 -mbig-pic @gol
418 -mcheck-zero-division -mhandle-large-shift @gol
419 -midentify-revision -mno-check-zero-division @gol
420 -mno-ocs-debug-info -mno-ocs-frame-position @gol
421 -mno-optimize-arg-area -mno-serialize-volatile @gol
422 -mno-underscores -mocs-debug-info @gol
423 -mocs-frame-position -moptimize-arg-area @gol
424 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
425 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
426 -mversion-03.00 -mwarn-passed-structs}
428 @emph{RS/6000 and PowerPC Options}
430 -mcpu=@var{cpu-type} @gol
431 -mtune=@var{cpu-type} @gol
432 -mpower -mno-power -mpower2 -mno-power2 @gol
433 -mpowerpc -mpowerpc64 -mno-powerpc @gol
434 -maltivec -mno-altivec @gol
435 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
436 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
437 -mnew-mnemonics -mold-mnemonics @gol
438 -mfull-toc -mminimal-toc -mno-fop-in-toc -mno-sum-in-toc @gol
439 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
440 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
441 -mstring -mno-string -mupdate -mno-update @gol
442 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
443 -mstrict-align -mno-strict-align -mrelocatable @gol
444 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
445 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
446 -mcall-aix -mcall-sysv -mcall-netbsd @gol
447 -maix-struct-return -msvr4-struct-return
449 -mprototype -mno-prototype @gol
450 -msim -mmvme -mads -myellowknife -memb -msdata @gol
451 -msdata=@var{opt} -mvxworks -G @var{num} -pthread}
455 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
456 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
457 -mminimum-fp-blocks -mnohc-struct-return}
461 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
462 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
463 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
464 -mgas -mgp32 -mgp64 @gol
465 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
466 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
467 -mmips-as -mmips-tfile -mno-abicalls @gol
468 -mno-embedded-data -mno-uninit-const-in-rodata @gol
469 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
470 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
471 -mrnames -msoft-float @gol
472 -m4650 -msingle-float -mmad @gol
473 -mstats -EL -EB -G @var{num} -nocpp @gol
474 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
475 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
477 @emph{i386 and x86-64 Options}
479 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
480 -masm=@var{dialect} -mno-fancy-math-387 @gol
481 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
482 -mno-wide-multiply -mrtd -malign-double @gol
483 -mpreferred-stack-boundary=@var{num} @gol
484 -mmmx -msse -msse2 -msse-math -m3dnow @gol
485 -mthreads -mno-align-stringops -minline-all-stringops @gol
486 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
487 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
493 -march=@var{architecture-type} @gol
494 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
495 -mfast-indirect-calls -mgas -mjump-in-delay @gol
496 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
497 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
498 -mno-jump-in-delay -mno-long-load-store @gol
499 -mno-portable-runtime -mno-soft-float @gol
500 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
501 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
502 -mschedule=@var{cpu-type} -mspace-regs}
504 @emph{Intel 960 Options}
506 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
507 -mcode-align -mcomplex-addr -mleaf-procedures @gol
508 -mic-compat -mic2.0-compat -mic3.0-compat @gol
509 -mintel-asm -mno-clean-linkage -mno-code-align @gol
510 -mno-complex-addr -mno-leaf-procedures @gol
511 -mno-old-align -mno-strict-align -mno-tail-call @gol
512 -mnumerics -mold-align -msoft-float -mstrict-align @gol
515 @emph{DEC Alpha Options}
517 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
518 -mieee -mieee-with-inexact -mieee-conformant @gol
519 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
520 -mtrap-precision=@var{mode} -mbuild-constants @gol
521 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
522 -mbwx -mmax -mfix -mcix @gol
523 -mfloat-vax -mfloat-ieee @gol
524 -mexplicit-relocs -msmall-data -mlarge-data @gol
525 -mmemory-latency=@var{time}}
527 @emph{DEC Alpha/VMS Options}
531 @emph{Clipper Options}
535 @emph{H8/300 Options}
537 -mrelax -mh -ms -mint32 -malign-300}
541 -m1 -m2 -m3 -m3e @gol
542 -m4-nofpu -m4-single-only -m4-single -m4 @gol
543 -mb -ml -mdalign -mrelax @gol
544 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
545 -mieee -misize -mpadstruct -mspace @gol
546 -mprefergot -musermode}
548 @emph{System V Options}
550 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
555 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
556 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
558 @emph{TMS320C3x/C4x Options}
560 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
561 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
562 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
563 -mparallel-insns -mparallel-mpy -mpreserve-float}
567 -mlong-calls -mno-long-calls -mep -mno-ep @gol
568 -mprolog-function -mno-prolog-function -mspace @gol
569 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
574 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
575 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
576 -mregparam -mnoregparam -msb -mnosb @gol
577 -mbitfield -mnobitfield -mhimem -mnohimem}
581 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
582 -mcall-prologues -mno-tablejump -mtiny-stack}
586 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
587 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
588 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
589 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
590 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
594 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
595 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
596 -melf -mbranch-predict -mno-branch-predict}
600 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
601 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
602 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
603 -minline-divide-max-throughput -mno-dwarf2-asm @gol
604 -mfixed-range=@var{register-range}}
608 -mextmem -mextmemory -monchip -mno-asm-optimize -masm-optimize @gol
609 -mbranch-cost=@var{n} -mcond-exec=@var{n}}
611 @emph{S/390 and zSeries Options}
613 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
614 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
615 -m64 -m31 -mdebug -mno-debug}
619 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
620 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
621 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
622 -mstack-align -mdata-align -mconst-align @gol
623 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
624 -melf -maout -melinux -mlinux -sim -sim2}
626 @emph{PDP-11 Options}
628 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
629 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
630 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
631 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
632 -mbranch-expensive -mbranch-cheap @gol
633 -msplit -mno-split -munix-asm -mdec-asm}
635 @emph{Xstormy16 Options}
639 @emph{Xtensa Options}
641 -mbig-endian -mlittle-endian @gol
642 -mdensity -mno-density @gol
643 -mmac16 -mno-mac16 @gol
644 -mmul16 -mno-mul16 @gol
645 -mmul32 -mno-mul32 @gol
647 -mminmax -mno-minmax @gol
648 -msext -mno-sext @gol
649 -mbooleans -mno-booleans @gol
650 -mhard-float -msoft-float @gol
651 -mfused-madd -mno-fused-madd @gol
652 -mserialize-volatile -mno-serialize-volatile @gol
653 -mtext-section-literals -mno-text-section-literals @gol
654 -mtarget-align -mno-target-align @gol
655 -mlongcalls -mno-longcalls}
657 @item Code Generation Options
658 @xref{Code Gen Options,,Options for Code Generation Conventions}.
660 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
661 -ffixed-@var{reg} -fexceptions @gol
662 -fnon-call-exceptions -funwind-tables @gol
663 -fasynchronous-unwind-tables @gol
664 -finhibit-size-directive -finstrument-functions @gol
665 -fno-common -fno-ident -fno-gnu-linker @gol
666 -fpcc-struct-return -fpic -fPIC @gol
667 -freg-struct-return -fshared-data -fshort-enums @gol
668 -fshort-double -fvolatile @gol
669 -fvolatile-global -fvolatile-static @gol
670 -fverbose-asm -fpack-struct -fstack-check @gol
671 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
672 -fargument-alias -fargument-noalias @gol
673 -fargument-noalias-global -fleading-underscore}
677 * Overall Options:: Controlling the kind of output:
678 an executable, object files, assembler files,
679 or preprocessed source.
680 * C Dialect Options:: Controlling the variant of C language compiled.
681 * C++ Dialect Options:: Variations on C++.
682 * Objective-C Dialect Options:: Variations on Objective-C.
683 * Language Independent Options:: Controlling how diagnostics should be
685 * Warning Options:: How picky should the compiler be?
686 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
687 * Optimize Options:: How much optimization?
688 * Preprocessor Options:: Controlling header files and macro definitions.
689 Also, getting dependency information for Make.
690 * Assembler Options:: Passing options to the assembler.
691 * Link Options:: Specifying libraries and so on.
692 * Directory Options:: Where to find header files and libraries.
693 Where to find the compiler executable files.
694 * Spec Files:: How to pass switches to sub-processes.
695 * Target Options:: Running a cross-compiler, or an old version of GCC.
698 @node Overall Options
699 @section Options Controlling the Kind of Output
701 Compilation can involve up to four stages: preprocessing, compilation
702 proper, assembly and linking, always in that order. The first three
703 stages apply to an individual source file, and end by producing an
704 object file; linking combines all the object files (those newly
705 compiled, and those specified as input) into an executable file.
707 @cindex file name suffix
708 For any given input file, the file name suffix determines what kind of
713 C source code which must be preprocessed.
716 C source code which should not be preprocessed.
719 C++ source code which should not be preprocessed.
722 Objective-C source code. Note that you must link with the library
723 @file{libobjc.a} to make an Objective-C program work.
726 Objective-C source code which should not be preprocessed.
729 C header file (not to be compiled or linked).
733 @itemx @var{file}.cxx
734 @itemx @var{file}.cpp
735 @itemx @var{file}.c++
737 C++ source code which must be preprocessed. Note that in @samp{.cxx},
738 the last two letters must both be literally @samp{x}. Likewise,
739 @samp{.C} refers to a literal capital C@.
742 @itemx @var{file}.for
743 @itemx @var{file}.FOR
744 Fortran source code which should not be preprocessed.
747 @itemx @var{file}.fpp
748 @itemx @var{file}.FPP
749 Fortran source code which must be preprocessed (with the traditional
753 Fortran source code which must be preprocessed with a RATFOR
754 preprocessor (not included with GCC)@.
756 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
757 Using and Porting GNU Fortran}, for more details of the handling of
760 @c FIXME: Descriptions of Java file types.
767 Ada source code file which contains a library unit declaration (a
768 declaration of a package, subprogram, or generic, or a generic
769 instantiation), or a library unit renaming declaration (a package,
770 generic, or subprogram renaming declaration). Such files are also
773 @itemx @var{file}.adb
774 Ada source code file containing a library unit body (a subprogram or
775 package body). Such files are also called @dfn{bodies}.
777 @c GCC also knows about some suffixes for languages not yet included:
783 @itemx @var{file}.chi
784 CHILL source code (preprocessed with the traditional preprocessor).
790 Assembler code which must be preprocessed.
793 An object file to be fed straight into linking.
794 Any file name with no recognized suffix is treated this way.
798 You can specify the input language explicitly with the @option{-x} option:
801 @item -x @var{language}
802 Specify explicitly the @var{language} for the following input files
803 (rather than letting the compiler choose a default based on the file
804 name suffix). This option applies to all following input files until
805 the next @option{-x} option. Possible values for @var{language} are:
807 c c-header cpp-output
809 objective-c objc-cpp-output
810 assembler assembler-with-cpp
813 f77 f77-cpp-input ratfor
818 Turn off any specification of a language, so that subsequent files are
819 handled according to their file name suffixes (as they are if @option{-x}
820 has not been used at all).
822 @item -pass-exit-codes
823 @opindex pass-exit-codes
824 Normally the @command{gcc} program will exit with the code of 1 if any
825 phase of the compiler returns a non-success return code. If you specify
826 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
827 numerically highest error produced by any phase that returned an error
831 If you only want some of the stages of compilation, you can use
832 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
833 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
834 @command{gcc} is to stop. Note that some combinations (for example,
835 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
840 Compile or assemble the source files, but do not link. The linking
841 stage simply is not done. The ultimate output is in the form of an
842 object file for each source file.
844 By default, the object file name for a source file is made by replacing
845 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
847 Unrecognized input files, not requiring compilation or assembly, are
852 Stop after the stage of compilation proper; do not assemble. The output
853 is in the form of an assembler code file for each non-assembler input
856 By default, the assembler file name for a source file is made by
857 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
859 Input files that don't require compilation are ignored.
863 Stop after the preprocessing stage; do not run the compiler proper. The
864 output is in the form of preprocessed source code, which is sent to the
867 Input files which don't require preprocessing are ignored.
869 @cindex output file option
872 Place output in file @var{file}. This applies regardless to whatever
873 sort of output is being produced, whether it be an executable file,
874 an object file, an assembler file or preprocessed C code.
876 Since only one output file can be specified, it does not make sense to
877 use @option{-o} when compiling more than one input file, unless you are
878 producing an executable file as output.
880 If @option{-o} is not specified, the default is to put an executable file
881 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
882 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
883 all preprocessed C source on standard output.
887 Print (on standard error output) the commands executed to run the stages
888 of compilation. Also print the version number of the compiler driver
889 program and of the preprocessor and the compiler proper.
893 Use pipes rather than temporary files for communication between the
894 various stages of compilation. This fails to work on some systems where
895 the assembler is unable to read from a pipe; but the GNU assembler has
900 Print (on the standard output) a description of the command line options
901 understood by @command{gcc}. If the @option{-v} option is also specified
902 then @option{--help} will also be passed on to the various processes
903 invoked by @command{gcc}, so that they can display the command line options
904 they accept. If the @option{-W} option is also specified then command
905 line options which have no documentation associated with them will also
910 Print (on the standard output) a description of target specific command
911 line options for each tool.
915 @section Compiling C++ Programs
917 @cindex suffixes for C++ source
918 @cindex C++ source file suffixes
919 C++ source files conventionally use one of the suffixes @samp{.C},
920 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
921 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
922 files with these names and compiles them as C++ programs even if you
923 call the compiler the same way as for compiling C programs (usually with
924 the name @command{gcc}).
928 However, C++ programs often require class libraries as well as a
929 compiler that understands the C++ language---and under some
930 circumstances, you might want to compile programs from standard input,
931 or otherwise without a suffix that flags them as C++ programs.
932 @command{g++} is a program that calls GCC with the default language
933 set to C++, and automatically specifies linking against the C++
934 library. On many systems, @command{g++} is also
935 installed with the name @command{c++}.
937 @cindex invoking @command{g++}
938 When you compile C++ programs, you may specify many of the same
939 command-line options that you use for compiling programs in any
940 language; or command-line options meaningful for C and related
941 languages; or options that are meaningful only for C++ programs.
942 @xref{C Dialect Options,,Options Controlling C Dialect}, for
943 explanations of options for languages related to C@.
944 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
945 explanations of options that are meaningful only for C++ programs.
947 @node C Dialect Options
948 @section Options Controlling C Dialect
949 @cindex dialect options
950 @cindex language dialect options
951 @cindex options, dialect
953 The following options control the dialect of C (or languages derived
954 from C, such as C++ and Objective-C) that the compiler accepts:
961 In C mode, support all ISO C89 programs. In C++ mode,
962 remove GNU extensions that conflict with ISO C++.
964 This turns off certain features of GCC that are incompatible with ISO
965 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
966 such as the @code{asm} and @code{typeof} keywords, and
967 predefined macros such as @code{unix} and @code{vax} that identify the
968 type of system you are using. It also enables the undesirable and
969 rarely used ISO trigraph feature. For the C compiler,
970 it disables recognition of C++ style @samp{//} comments as well as
971 the @code{inline} keyword.
973 The alternate keywords @code{__asm__}, @code{__extension__},
974 @code{__inline__} and @code{__typeof__} continue to work despite
975 @option{-ansi}. You would not want to use them in an ISO C program, of
976 course, but it is useful to put them in header files that might be included
977 in compilations done with @option{-ansi}. Alternate predefined macros
978 such as @code{__unix__} and @code{__vax__} are also available, with or
979 without @option{-ansi}.
981 The @option{-ansi} option does not cause non-ISO programs to be
982 rejected gratuitously. For that, @option{-pedantic} is required in
983 addition to @option{-ansi}. @xref{Warning Options}.
985 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
986 option is used. Some header files may notice this macro and refrain
987 from declaring certain functions or defining certain macros that the
988 ISO standard doesn't call for; this is to avoid interfering with any
989 programs that might use these names for other things.
991 Functions which would normally be built in but do not have semantics
992 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
993 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
994 built-in functions provided by GCC}, for details of the functions
999 Determine the language standard. This option is currently only
1000 supported when compiling C@. A value for this option must be provided;
1006 ISO C89 (same as @option{-ansi}).
1008 @item iso9899:199409
1009 ISO C89 as modified in amendment 1.
1015 ISO C99. Note that this standard is not yet fully supported; see
1016 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1017 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1020 Default, ISO C89 plus GNU extensions (including some C99 features).
1024 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1025 this will become the default. The name @samp{gnu9x} is deprecated.
1029 Even when this option is not specified, you can still use some of the
1030 features of newer standards in so far as they do not conflict with
1031 previous C standards. For example, you may use @code{__restrict__} even
1032 when @option{-std=c99} is not specified.
1034 The @option{-std} options specifying some version of ISO C have the same
1035 effects as @option{-ansi}, except that features that were not in ISO C89
1036 but are in the specified version (for example, @samp{//} comments and
1037 the @code{inline} keyword in ISO C99) are not disabled.
1039 @xref{Standards,,Language Standards Supported by GCC}, for details of
1040 these standard versions.
1042 @item -aux-info @var{filename}
1044 Output to the given filename prototyped declarations for all functions
1045 declared and/or defined in a translation unit, including those in header
1046 files. This option is silently ignored in any language other than C@.
1048 Besides declarations, the file indicates, in comments, the origin of
1049 each declaration (source file and line), whether the declaration was
1050 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1051 @samp{O} for old, respectively, in the first character after the line
1052 number and the colon), and whether it came from a declaration or a
1053 definition (@samp{C} or @samp{F}, respectively, in the following
1054 character). In the case of function definitions, a K&R-style list of
1055 arguments followed by their declarations is also provided, inside
1056 comments, after the declaration.
1060 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1061 keyword, so that code can use these words as identifiers. You can use
1062 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1063 instead. @option{-ansi} implies @option{-fno-asm}.
1065 In C++, this switch only affects the @code{typeof} keyword, since
1066 @code{asm} and @code{inline} are standard keywords. You may want to
1067 use the @option{-fno-gnu-keywords} flag instead, which has the same
1068 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1069 switch only affects the @code{asm} and @code{typeof} keywords, since
1070 @code{inline} is a standard keyword in ISO C99.
1073 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1074 @opindex fno-builtin
1075 @cindex built-in functions
1076 Don't recognize built-in functions that do not begin with
1077 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1078 functions provided by GCC}, for details of the functions affected,
1079 including those which are not built-in functions when @option{-ansi} or
1080 @option{-std} options for strict ISO C conformance are used because they
1081 do not have an ISO standard meaning.
1083 GCC normally generates special code to handle certain built-in functions
1084 more efficiently; for instance, calls to @code{alloca} may become single
1085 instructions that adjust the stack directly, and calls to @code{memcpy}
1086 may become inline copy loops. The resulting code is often both smaller
1087 and faster, but since the function calls no longer appear as such, you
1088 cannot set a breakpoint on those calls, nor can you change the behavior
1089 of the functions by linking with a different library.
1091 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1092 option has no effect. Therefore, in C++, the only way to get the
1093 optimization benefits of built-in functions is to call the function
1094 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1095 built-in functions to implement many functions (like
1096 @code{std::strchr}), so that you automatically get efficient code.
1098 With the @option{-fno-builtin-@var{function}} option, not available
1099 when compiling C++, only the built-in function @var{function} is
1100 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1101 function is named this is not built-in in this version of GCC, this
1102 option is ignored. There is no corresponding
1103 @option{-fbuiltin-@var{function}} option; if you wish to enable
1104 built-in functions selectively when using @option{-fno-builtin} or
1105 @option{-ffreestanding}, you may define macros such as:
1108 #define abs(n) __builtin_abs ((n))
1109 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1114 @cindex hosted environment
1116 Assert that compilation takes place in a hosted environment. This implies
1117 @option{-fbuiltin}. A hosted environment is one in which the
1118 entire standard library is available, and in which @code{main} has a return
1119 type of @code{int}. Examples are nearly everything except a kernel.
1120 This is equivalent to @option{-fno-freestanding}.
1122 @item -ffreestanding
1123 @opindex ffreestanding
1124 @cindex hosted environment
1126 Assert that compilation takes place in a freestanding environment. This
1127 implies @option{-fno-builtin}. A freestanding environment
1128 is one in which the standard library may not exist, and program startup may
1129 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1130 This is equivalent to @option{-fno-hosted}.
1132 @xref{Standards,,Language Standards Supported by GCC}, for details of
1133 freestanding and hosted environments.
1137 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1138 options for strict ISO C conformance) implies @option{-trigraphs}.
1140 @cindex traditional C language
1141 @cindex C language, traditional
1143 @opindex traditional
1144 Attempt to support some aspects of traditional C compilers.
1149 All @code{extern} declarations take effect globally even if they
1150 are written inside of a function definition. This includes implicit
1151 declarations of functions.
1154 The newer keywords @code{typeof}, @code{inline}, @code{signed}, @code{const}
1155 and @code{volatile} are not recognized. (You can still use the
1156 alternative keywords such as @code{__typeof__}, @code{__inline__}, and
1160 Comparisons between pointers and integers are always allowed.
1163 Integer types @code{unsigned short} and @code{unsigned char} promote
1164 to @code{unsigned int}.
1167 Out-of-range floating point literals are not an error.
1170 Certain constructs which ISO regards as a single invalid preprocessing
1171 number, such as @samp{0xe-0xd}, are treated as expressions instead.
1174 String ``constants'' are not necessarily constant; they are stored in
1175 writable space, and identical looking constants are allocated
1176 separately. (This is the same as the effect of
1177 @option{-fwritable-strings}.)
1179 @cindex @code{longjmp} and automatic variables
1181 All automatic variables not declared @code{register} are preserved by
1182 @code{longjmp}. Ordinarily, GNU C follows ISO C: automatic variables
1183 not declared @code{volatile} may be clobbered.
1188 @cindex escape sequences, traditional
1189 The character escape sequences @samp{\x} and @samp{\a} evaluate as the
1190 literal characters @samp{x} and @samp{a} respectively. Without
1191 @w{@option{-traditional}}, @samp{\x} is a prefix for the hexadecimal
1192 representation of a character, and @samp{\a} produces a bell.
1195 This option is deprecated and may be removed.
1197 You may wish to use @option{-fno-builtin} as well as @option{-traditional}
1198 if your program uses names that are normally GNU C built-in functions for
1199 other purposes of its own.
1201 You cannot use @option{-traditional} if you include any header files that
1202 rely on ISO C features. Some vendors are starting to ship systems with
1203 ISO C header files and you cannot use @option{-traditional} on such
1204 systems to compile files that include any system headers.
1206 The @option{-traditional} option also enables @option{-traditional-cpp}.
1208 @item -traditional-cpp
1209 @opindex traditional-cpp
1210 Attempt to support some aspects of traditional C preprocessors.
1211 See the GNU CPP manual for details.
1213 @item -fcond-mismatch
1214 @opindex fcond-mismatch
1215 Allow conditional expressions with mismatched types in the second and
1216 third arguments. The value of such an expression is void. This option
1217 is not supported for C++.
1219 @item -funsigned-char
1220 @opindex funsigned-char
1221 Let the type @code{char} be unsigned, like @code{unsigned char}.
1223 Each kind of machine has a default for what @code{char} should
1224 be. It is either like @code{unsigned char} by default or like
1225 @code{signed char} by default.
1227 Ideally, a portable program should always use @code{signed char} or
1228 @code{unsigned char} when it depends on the signedness of an object.
1229 But many programs have been written to use plain @code{char} and
1230 expect it to be signed, or expect it to be unsigned, depending on the
1231 machines they were written for. This option, and its inverse, let you
1232 make such a program work with the opposite default.
1234 The type @code{char} is always a distinct type from each of
1235 @code{signed char} or @code{unsigned char}, even though its behavior
1236 is always just like one of those two.
1239 @opindex fsigned-char
1240 Let the type @code{char} be signed, like @code{signed char}.
1242 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1243 the negative form of @option{-funsigned-char}. Likewise, the option
1244 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1246 @item -fsigned-bitfields
1247 @itemx -funsigned-bitfields
1248 @itemx -fno-signed-bitfields
1249 @itemx -fno-unsigned-bitfields
1250 @opindex fsigned-bitfields
1251 @opindex funsigned-bitfields
1252 @opindex fno-signed-bitfields
1253 @opindex fno-unsigned-bitfields
1254 These options control whether a bit-field is signed or unsigned, when the
1255 declaration does not use either @code{signed} or @code{unsigned}. By
1256 default, such a bit-field is signed, because this is consistent: the
1257 basic integer types such as @code{int} are signed types.
1259 However, when @option{-traditional} is used, bit-fields are all unsigned
1262 @item -fwritable-strings
1263 @opindex fwritable-strings
1264 Store string constants in the writable data segment and don't uniquize
1265 them. This is for compatibility with old programs which assume they can
1266 write into string constants. The option @option{-traditional} also has
1269 Writing into string constants is a very bad idea; ``constants'' should
1272 @item -fallow-single-precision
1273 @opindex fallow-single-precision
1274 Do not promote single precision math operations to double precision,
1275 even when compiling with @option{-traditional}.
1277 Traditional K&R C promotes all floating point operations to double
1278 precision, regardless of the sizes of the operands. On the
1279 architecture for which you are compiling, single precision may be faster
1280 than double precision. If you must use @option{-traditional}, but want
1281 to use single precision operations when the operands are single
1282 precision, use this option. This option has no effect when compiling
1283 with ISO or GNU C conventions (the default).
1286 @opindex fshort-wchar
1287 Override the underlying type for @samp{wchar_t} to be @samp{short
1288 unsigned int} instead of the default for the target. This option is
1289 useful for building programs to run under WINE@.
1292 @node C++ Dialect Options
1293 @section Options Controlling C++ Dialect
1295 @cindex compiler options, C++
1296 @cindex C++ options, command line
1297 @cindex options, C++
1298 This section describes the command-line options that are only meaningful
1299 for C++ programs; but you can also use most of the GNU compiler options
1300 regardless of what language your program is in. For example, you
1301 might compile a file @code{firstClass.C} like this:
1304 g++ -g -frepo -O -c firstClass.C
1308 In this example, only @option{-frepo} is an option meant
1309 only for C++ programs; you can use the other options with any
1310 language supported by GCC@.
1312 Here is a list of options that are @emph{only} for compiling C++ programs:
1315 @item -fno-access-control
1316 @opindex fno-access-control
1317 Turn off all access checking. This switch is mainly useful for working
1318 around bugs in the access control code.
1322 Check that the pointer returned by @code{operator new} is non-null
1323 before attempting to modify the storage allocated. The current Working
1324 Paper requires that @code{operator new} never return a null pointer, so
1325 this check is normally unnecessary.
1327 An alternative to using this option is to specify that your
1328 @code{operator new} does not throw any exceptions; if you declare it
1329 @samp{throw()}, G++ will check the return value. See also @samp{new
1332 @item -fconserve-space
1333 @opindex fconserve-space
1334 Put uninitialized or runtime-initialized global variables into the
1335 common segment, as C does. This saves space in the executable at the
1336 cost of not diagnosing duplicate definitions. If you compile with this
1337 flag and your program mysteriously crashes after @code{main()} has
1338 completed, you may have an object that is being destroyed twice because
1339 two definitions were merged.
1341 This option is no longer useful on most targets, now that support has
1342 been added for putting variables into BSS without making them common.
1344 @item -fno-const-strings
1345 @opindex fno-const-strings
1346 Give string constants type @code{char *} instead of type @code{const
1347 char *}. By default, G++ uses type @code{const char *} as required by
1348 the standard. Even if you use @option{-fno-const-strings}, you cannot
1349 actually modify the value of a string constant, unless you also use
1350 @option{-fwritable-strings}.
1352 This option might be removed in a future release of G++. For maximum
1353 portability, you should structure your code so that it works with
1354 string constants that have type @code{const char *}.
1356 @item -fdollars-in-identifiers
1357 @opindex fdollars-in-identifiers
1358 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1359 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1360 @samp{$} by default on most target systems, but there are a few exceptions.)
1361 Traditional C allowed the character @samp{$} to form part of
1362 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1364 @item -fno-elide-constructors
1365 @opindex fno-elide-constructors
1366 The C++ standard allows an implementation to omit creating a temporary
1367 which is only used to initialize another object of the same type.
1368 Specifying this option disables that optimization, and forces G++ to
1369 call the copy constructor in all cases.
1371 @item -fno-enforce-eh-specs
1372 @opindex fno-enforce-eh-specs
1373 Don't check for violation of exception specifications at runtime. This
1374 option violates the C++ standard, but may be useful for reducing code
1375 size in production builds, much like defining @samp{NDEBUG}. The compiler
1376 will still optimize based on the exception specifications.
1378 @item -fexternal-templates
1379 @opindex fexternal-templates
1381 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1382 template instantiation; template instances are emitted or not according
1383 to the location of the template definition. @xref{Template
1384 Instantiation}, for more information.
1386 This option is deprecated.
1388 @item -falt-external-templates
1389 @opindex falt-external-templates
1390 Similar to @option{-fexternal-templates}, but template instances are
1391 emitted or not according to the place where they are first instantiated.
1392 @xref{Template Instantiation}, for more information.
1394 This option is deprecated.
1397 @itemx -fno-for-scope
1399 @opindex fno-for-scope
1400 If @option{-ffor-scope} is specified, the scope of variables declared in
1401 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1402 as specified by the C++ standard.
1403 If @option{-fno-for-scope} is specified, the scope of variables declared in
1404 a @i{for-init-statement} extends to the end of the enclosing scope,
1405 as was the case in old versions of G++, and other (traditional)
1406 implementations of C++.
1408 The default if neither flag is given to follow the standard,
1409 but to allow and give a warning for old-style code that would
1410 otherwise be invalid, or have different behavior.
1412 @item -fno-gnu-keywords
1413 @opindex fno-gnu-keywords
1414 Do not recognize @code{typeof} as a keyword, so that code can use this
1415 word as an identifier. You can use the keyword @code{__typeof__} instead.
1416 @option{-ansi} implies @option{-fno-gnu-keywords}.
1418 @item -fno-implicit-templates
1419 @opindex fno-implicit-templates
1420 Never emit code for non-inline templates which are instantiated
1421 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1422 @xref{Template Instantiation}, for more information.
1424 @item -fno-implicit-inline-templates
1425 @opindex fno-implicit-inline-templates
1426 Don't emit code for implicit instantiations of inline templates, either.
1427 The default is to handle inlines differently so that compiles with and
1428 without optimization will need the same set of explicit instantiations.
1430 @item -fno-implement-inlines
1431 @opindex fno-implement-inlines
1432 To save space, do not emit out-of-line copies of inline functions
1433 controlled by @samp{#pragma implementation}. This will cause linker
1434 errors if these functions are not inlined everywhere they are called.
1436 @item -fms-extensions
1437 @opindex fms-extensions
1438 Disable pedantic warnings about constructs used in MFC, such as implicit
1439 int and getting a pointer to member function via non-standard syntax.
1441 @item -fno-nonansi-builtins
1442 @opindex fno-nonansi-builtins
1443 Disable built-in declarations of functions that are not mandated by
1444 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1445 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1447 @item -fno-operator-names
1448 @opindex fno-operator-names
1449 Do not treat the operator name keywords @code{and}, @code{bitand},
1450 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1451 synonyms as keywords.
1453 @item -fno-optional-diags
1454 @opindex fno-optional-diags
1455 Disable diagnostics that the standard says a compiler does not need to
1456 issue. Currently, the only such diagnostic issued by G++ is the one for
1457 a name having multiple meanings within a class.
1460 @opindex fpermissive
1461 Downgrade messages about nonconformant code from errors to warnings. By
1462 default, G++ effectively sets @option{-pedantic-errors} without
1463 @option{-pedantic}; this option reverses that. This behavior and this
1464 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1468 Enable automatic template instantiation at link time. This option also
1469 implies @option{-fno-implicit-templates}. @xref{Template
1470 Instantiation}, for more information.
1474 Disable generation of information about every class with virtual
1475 functions for use by the C++ runtime type identification features
1476 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1477 of the language, you can save some space by using this flag. Note that
1478 exception handling uses the same information, but it will generate it as
1483 Emit statistics about front-end processing at the end of the compilation.
1484 This information is generally only useful to the G++ development team.
1486 @item -ftemplate-depth-@var{n}
1487 @opindex ftemplate-depth
1488 Set the maximum instantiation depth for template classes to @var{n}.
1489 A limit on the template instantiation depth is needed to detect
1490 endless recursions during template class instantiation. ANSI/ISO C++
1491 conforming programs must not rely on a maximum depth greater than 17.
1493 @item -fuse-cxa-atexit
1494 @opindex fuse-cxa-atexit
1495 Register destructors for objects with static storage duration with the
1496 @code{__cxa_atexit} function rather than the @code{atexit} function.
1497 This option is required for fully standards-compliant handling of static
1498 destructors, but will only work if your C library supports
1499 @code{__cxa_atexit}.
1503 Emit special relocations for vtables and virtual function references
1504 so that the linker can identify unused virtual functions and zero out
1505 vtable slots that refer to them. This is most useful with
1506 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1507 also discard the functions themselves.
1509 This optimization requires GNU as and GNU ld. Not all systems support
1510 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1514 Do not use weak symbol support, even if it is provided by the linker.
1515 By default, G++ will use weak symbols if they are available. This
1516 option exists only for testing, and should not be used by end-users;
1517 it will result in inferior code and has no benefits. This option may
1518 be removed in a future release of G++.
1522 Do not search for header files in the standard directories specific to
1523 C++, but do still search the other standard directories. (This option
1524 is used when building the C++ library.)
1527 In addition, these optimization, warning, and code generation options
1528 have meanings only for C++ programs:
1531 @item -fno-default-inline
1532 @opindex fno-default-inline
1533 Do not assume @samp{inline} for functions defined inside a class scope.
1534 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1535 functions will have linkage like inline functions; they just won't be
1538 @item -Wctor-dtor-privacy @r{(C++ only)}
1539 @opindex Wctor-dtor-privacy
1540 Warn when a class seems unusable, because all the constructors or
1541 destructors in a class are private and the class has no friends or
1542 public static member functions.
1544 @item -Wnon-virtual-dtor @r{(C++ only)}
1545 @opindex Wnon-virtual-dtor
1546 Warn when a class declares a non-virtual destructor that should probably
1547 be virtual, because it looks like the class will be used polymorphically.
1549 @item -Wreorder @r{(C++ only)}
1551 @cindex reordering, warning
1552 @cindex warning for reordering of member initializers
1553 Warn when the order of member initializers given in the code does not
1554 match the order in which they must be executed. For instance:
1560 A(): j (0), i (1) @{ @}
1564 Here the compiler will warn that the member initializers for @samp{i}
1565 and @samp{j} will be rearranged to match the declaration order of the
1569 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1572 @item -Weffc++ @r{(C++ only)}
1574 Warn about violations of various style guidelines from Scott Meyers'
1575 @cite{Effective C++} books. If you use this option, you should be aware
1576 that the standard library headers do not obey all of these guidelines;
1577 you can use @samp{grep -v} to filter out those warnings.
1579 @item -Wno-deprecated @r{(C++ only)}
1580 @opindex Wno-deprecated
1581 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1583 @item -Wno-non-template-friend @r{(C++ only)}
1584 @opindex Wno-non-template-friend
1585 Disable warnings when non-templatized friend functions are declared
1586 within a template. With the advent of explicit template specification
1587 support in G++, if the name of the friend is an unqualified-id (i.e.,
1588 @samp{friend foo(int)}), the C++ language specification demands that the
1589 friend declare or define an ordinary, nontemplate function. (Section
1590 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1591 could be interpreted as a particular specialization of a templatized
1592 function. Because this non-conforming behavior is no longer the default
1593 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1594 check existing code for potential trouble spots, and is on by default.
1595 This new compiler behavior can be turned off with
1596 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1597 but disables the helpful warning.
1599 @item -Wold-style-cast @r{(C++ only)}
1600 @opindex Wold-style-cast
1601 Warn if an old-style (C-style) cast to a non-void type is used within
1602 a C++ program. The new-style casts (@samp{static_cast},
1603 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1604 unintended effects, and much easier to grep for.
1606 @item -Woverloaded-virtual @r{(C++ only)}
1607 @opindex Woverloaded-virtual
1608 @cindex overloaded virtual fn, warning
1609 @cindex warning for overloaded virtual fn
1610 Warn when a function declaration hides virtual functions from a
1611 base class. For example, in:
1618 struct B: public A @{
1623 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1631 will fail to compile.
1633 @item -Wno-pmf-conversions @r{(C++ only)}
1634 @opindex Wno-pmf-conversions
1635 Disable the diagnostic for converting a bound pointer to member function
1638 @item -Wsign-promo @r{(C++ only)}
1639 @opindex Wsign-promo
1640 Warn when overload resolution chooses a promotion from unsigned or
1641 enumeral type to a signed type over a conversion to an unsigned type of
1642 the same size. Previous versions of G++ would try to preserve
1643 unsignedness, but the standard mandates the current behavior.
1645 @item -Wsynth @r{(C++ only)}
1647 @cindex warning for synthesized methods
1648 @cindex synthesized methods, warning
1649 Warn when G++'s synthesis behavior does not match that of cfront. For
1655 A& operator = (int);
1665 In this example, G++ will synthesize a default @samp{A& operator =
1666 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1669 @node Objective-C Dialect Options
1670 @section Options Controlling Objective-C Dialect
1672 @cindex compiler options, Objective-C
1673 @cindex Objective-C options, command line
1674 @cindex options, Objective-C
1675 This section describes the command-line options that are only meaningful
1676 for Objective-C programs; but you can also use most of the GNU compiler
1677 options regardless of what language your program is in. For example,
1678 you might compile a file @code{some_class.m} like this:
1681 gcc -g -fgnu-runtime -O -c some_class.m
1685 In this example, only @option{-fgnu-runtime} is an option meant only for
1686 Objective-C programs; you can use the other options with any language
1689 Here is a list of options that are @emph{only} for compiling Objective-C
1693 @item -fconstant-string-class=@var{class-name}
1694 @opindex fconstant-string-class
1695 Use @var{class-name} as the name of the class to instantiate for each
1696 literal string specified with the syntax @code{@@"@dots{}"}. The default
1697 class name is @code{NXConstantString}.
1700 @opindex fgnu-runtime
1701 Generate object code compatible with the standard GNU Objective-C
1702 runtime. This is the default for most types of systems.
1704 @item -fnext-runtime
1705 @opindex fnext-runtime
1706 Generate output compatible with the NeXT runtime. This is the default
1707 for NeXT-based systems, including Darwin and Mac OS X@.
1711 Dump interface declarations for all classes seen in the source file to a
1712 file named @file{@var{sourcename}.decl}.
1715 @opindex Wno-protocol
1716 Do not warn if methods required by a protocol are not implemented
1717 in the class adopting it.
1721 Warn if a selector has multiple methods of different types defined.
1723 @c not documented because only avail via -Wp
1724 @c @item -print-objc-runtime-info
1728 @node Language Independent Options
1729 @section Options to Control Diagnostic Messages Formatting
1730 @cindex options to control diagnostics formatting
1731 @cindex diagnostic messages
1732 @cindex message formatting
1734 Traditionally, diagnostic messages have been formatted irrespective of
1735 the output device's aspect (e.g.@: its width, @dots{}). The options described
1736 below can be used to control the diagnostic messages formatting
1737 algorithm, e.g.@: how many characters per line, how often source location
1738 information should be reported. Right now, only the C++ front end can
1739 honor these options. However it is expected, in the near future, that
1740 the remaining front ends would be able to digest them correctly.
1743 @item -fmessage-length=@var{n}
1744 @opindex fmessage-length
1745 Try to format error messages so that they fit on lines of about @var{n}
1746 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1747 the front ends supported by GCC@. If @var{n} is zero, then no
1748 line-wrapping will be done; each error message will appear on a single
1751 @opindex fdiagnostics-show-location
1752 @item -fdiagnostics-show-location=once
1753 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1754 reporter to emit @emph{once} source location information; that is, in
1755 case the message is too long to fit on a single physical line and has to
1756 be wrapped, the source location won't be emitted (as prefix) again,
1757 over and over, in subsequent continuation lines. This is the default
1760 @item -fdiagnostics-show-location=every-line
1761 Only meaningful in line-wrapping mode. Instructs the diagnostic
1762 messages reporter to emit the same source location information (as
1763 prefix) for physical lines that result from the process of breaking
1764 a message which is too long to fit on a single line.
1768 @node Warning Options
1769 @section Options to Request or Suppress Warnings
1770 @cindex options to control warnings
1771 @cindex warning messages
1772 @cindex messages, warning
1773 @cindex suppressing warnings
1775 Warnings are diagnostic messages that report constructions which
1776 are not inherently erroneous but which are risky or suggest there
1777 may have been an error.
1779 You can request many specific warnings with options beginning @samp{-W},
1780 for example @option{-Wimplicit} to request warnings on implicit
1781 declarations. Each of these specific warning options also has a
1782 negative form beginning @samp{-Wno-} to turn off warnings;
1783 for example, @option{-Wno-implicit}. This manual lists only one of the
1784 two forms, whichever is not the default.
1786 These options control the amount and kinds of warnings produced by GCC:
1789 @cindex syntax checking
1791 @opindex fsyntax-only
1792 Check the code for syntax errors, but don't do anything beyond that.
1796 Issue all the warnings demanded by strict ISO C and ISO C++;
1797 reject all programs that use forbidden extensions, and some other
1798 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1799 version of the ISO C standard specified by any @option{-std} option used.
1801 Valid ISO C and ISO C++ programs should compile properly with or without
1802 this option (though a rare few will require @option{-ansi} or a
1803 @option{-std} option specifying the required version of ISO C)@. However,
1804 without this option, certain GNU extensions and traditional C and C++
1805 features are supported as well. With this option, they are rejected.
1807 @option{-pedantic} does not cause warning messages for use of the
1808 alternate keywords whose names begin and end with @samp{__}. Pedantic
1809 warnings are also disabled in the expression that follows
1810 @code{__extension__}. However, only system header files should use
1811 these escape routes; application programs should avoid them.
1812 @xref{Alternate Keywords}.
1814 Some users try to use @option{-pedantic} to check programs for strict ISO
1815 C conformance. They soon find that it does not do quite what they want:
1816 it finds some non-ISO practices, but not all---only those for which
1817 ISO C @emph{requires} a diagnostic, and some others for which
1818 diagnostics have been added.
1820 A feature to report any failure to conform to ISO C might be useful in
1821 some instances, but would require considerable additional work and would
1822 be quite different from @option{-pedantic}. We don't have plans to
1823 support such a feature in the near future.
1825 Where the standard specified with @option{-std} represents a GNU
1826 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1827 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1828 extended dialect is based. Warnings from @option{-pedantic} are given
1829 where they are required by the base standard. (It would not make sense
1830 for such warnings to be given only for features not in the specified GNU
1831 C dialect, since by definition the GNU dialects of C include all
1832 features the compiler supports with the given option, and there would be
1833 nothing to warn about.)
1835 @item -pedantic-errors
1836 @opindex pedantic-errors
1837 Like @option{-pedantic}, except that errors are produced rather than
1842 Inhibit all warning messages.
1846 Inhibit warning messages about the use of @samp{#import}.
1848 @item -Wchar-subscripts
1849 @opindex Wchar-subscripts
1850 Warn if an array subscript has type @code{char}. This is a common cause
1851 of error, as programmers often forget that this type is signed on some
1856 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1857 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1861 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1862 the arguments supplied have types appropriate to the format string
1863 specified, and that the conversions specified in the format string make
1864 sense. This includes standard functions, and others specified by format
1865 attributes (@pxref{Function Attributes}), in the @code{printf},
1866 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1867 not in the C standard) families.
1869 The formats are checked against the format features supported by GNU
1870 libc version 2.2. These include all ISO C89 and C99 features, as well
1871 as features from the Single Unix Specification and some BSD and GNU
1872 extensions. Other library implementations may not support all these
1873 features; GCC does not support warning about features that go beyond a
1874 particular library's limitations. However, if @option{-pedantic} is used
1875 with @option{-Wformat}, warnings will be given about format features not
1876 in the selected standard version (but not for @code{strfmon} formats,
1877 since those are not in any version of the C standard). @xref{C Dialect
1878 Options,,Options Controlling C Dialect}.
1880 @option{-Wformat} is included in @option{-Wall}. For more control over some
1881 aspects of format checking, the options @option{-Wno-format-y2k},
1882 @option{-Wno-format-extra-args}, @option{-Wformat-nonliteral},
1883 @option{-Wformat-security} and @option{-Wformat=2} are available, but are
1884 not included in @option{-Wall}.
1886 @item -Wno-format-y2k
1887 @opindex Wno-format-y2k
1888 If @option{-Wformat} is specified, do not warn about @code{strftime}
1889 formats which may yield only a two-digit year.
1891 @item -Wno-format-extra-args
1892 @opindex Wno-format-extra-args
1893 If @option{-Wformat} is specified, do not warn about excess arguments to a
1894 @code{printf} or @code{scanf} format function. The C standard specifies
1895 that such arguments are ignored.
1897 Where the unused arguments lie between used arguments that are
1898 specified with @samp{$} operand number specifications, normally
1899 warnings are still given, since the implementation could not know what
1900 type to pass to @code{va_arg} to skip the unused arguments. However,
1901 in the case of @code{scanf} formats, this option will suppress the
1902 warning if the unused arguments are all pointers, since the Single
1903 Unix Specification says that such unused arguments are allowed.
1905 @item -Wformat-nonliteral
1906 @opindex Wformat-nonliteral
1907 If @option{-Wformat} is specified, also warn if the format string is not a
1908 string literal and so cannot be checked, unless the format function
1909 takes its format arguments as a @code{va_list}.
1911 @item -Wformat-security
1912 @opindex Wformat-security
1913 If @option{-Wformat} is specified, also warn about uses of format
1914 functions that represent possible security problems. At present, this
1915 warns about calls to @code{printf} and @code{scanf} functions where the
1916 format string is not a string literal and there are no format arguments,
1917 as in @code{printf (foo);}. This may be a security hole if the format
1918 string came from untrusted input and contains @samp{%n}. (This is
1919 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1920 in future warnings may be added to @option{-Wformat-security} that are not
1921 included in @option{-Wformat-nonliteral}.)
1925 Enable @option{-Wformat} plus format checks not included in
1926 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1927 -Wformat-nonliteral -Wformat-security}.
1929 @item -Wimplicit-int
1930 @opindex Wimplicit-int
1931 Warn when a declaration does not specify a type.
1933 @item -Wimplicit-function-declaration
1934 @itemx -Werror-implicit-function-declaration
1935 @opindex Wimplicit-function-declaration
1936 @opindex Werror-implicit-function-declaration
1937 Give a warning (or error) whenever a function is used before being
1942 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1946 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1947 function with external linkage, returning int, taking either zero
1948 arguments, two, or three arguments of appropriate types.
1950 @item -Wmissing-braces
1951 @opindex Wmissing-braces
1952 Warn if an aggregate or union initializer is not fully bracketed. In
1953 the following example, the initializer for @samp{a} is not fully
1954 bracketed, but that for @samp{b} is fully bracketed.
1957 int a[2][2] = @{ 0, 1, 2, 3 @};
1958 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1962 @opindex Wparentheses
1963 Warn if parentheses are omitted in certain contexts, such
1964 as when there is an assignment in a context where a truth value
1965 is expected, or when operators are nested whose precedence people
1966 often get confused about.
1968 Also warn about constructions where there may be confusion to which
1969 @code{if} statement an @code{else} branch belongs. Here is an example of
1984 In C, every @code{else} branch belongs to the innermost possible @code{if}
1985 statement, which in this example is @code{if (b)}. This is often not
1986 what the programmer expected, as illustrated in the above example by
1987 indentation the programmer chose. When there is the potential for this
1988 confusion, GCC will issue a warning when this flag is specified.
1989 To eliminate the warning, add explicit braces around the innermost
1990 @code{if} statement so there is no way the @code{else} could belong to
1991 the enclosing @code{if}. The resulting code would look like this:
2007 @item -Wsequence-point
2008 @opindex Wsequence-point
2009 Warn about code that may have undefined semantics because of violations
2010 of sequence point rules in the C standard.
2012 The C standard defines the order in which expressions in a C program are
2013 evaluated in terms of @dfn{sequence points}, which represent a partial
2014 ordering between the execution of parts of the program: those executed
2015 before the sequence point, and those executed after it. These occur
2016 after the evaluation of a full expression (one which is not part of a
2017 larger expression), after the evaluation of the first operand of a
2018 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2019 function is called (but after the evaluation of its arguments and the
2020 expression denoting the called function), and in certain other places.
2021 Other than as expressed by the sequence point rules, the order of
2022 evaluation of subexpressions of an expression is not specified. All
2023 these rules describe only a partial order rather than a total order,
2024 since, for example, if two functions are called within one expression
2025 with no sequence point between them, the order in which the functions
2026 are called is not specified. However, the standards committee have
2027 ruled that function calls do not overlap.
2029 It is not specified when between sequence points modifications to the
2030 values of objects take effect. Programs whose behavior depends on this
2031 have undefined behavior; the C standard specifies that ``Between the
2032 previous and next sequence point an object shall have its stored value
2033 modified at most once by the evaluation of an expression. Furthermore,
2034 the prior value shall be read only to determine the value to be
2035 stored.''. If a program breaks these rules, the results on any
2036 particular implementation are entirely unpredictable.
2038 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2039 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2040 diagnosed by this option, and it may give an occasional false positive
2041 result, but in general it has been found fairly effective at detecting
2042 this sort of problem in programs.
2044 The present implementation of this option only works for C programs. A
2045 future implementation may also work for C++ programs.
2047 The C standard is worded confusingly, therefore there is some debate
2048 over the precise meaning of the sequence point rules in subtle cases.
2049 Links to discussions of the problem, including proposed formal
2050 definitions, may be found on our readings page, at
2051 @w{@uref{http://gcc.gnu.org/readings.html}}.
2054 @opindex Wreturn-type
2055 Warn whenever a function is defined with a return-type that defaults to
2056 @code{int}. Also warn about any @code{return} statement with no
2057 return-value in a function whose return-type is not @code{void}.
2059 For C++, a function without return type always produces a diagnostic
2060 message, even when @option{-Wno-return-type} is specified. The only
2061 exceptions are @samp{main} and functions defined in system headers.
2065 Warn whenever a @code{switch} statement has an index of enumeral type
2066 and lacks a @code{case} for one or more of the named codes of that
2067 enumeration. (The presence of a @code{default} label prevents this
2068 warning.) @code{case} labels outside the enumeration range also
2069 provoke warnings when this option is used.
2073 Warn if any trigraphs are encountered that might change the meaning of
2074 the program (trigraphs within comments are not warned about).
2076 @item -Wunused-function
2077 @opindex Wunused-function
2078 Warn whenever a static function is declared but not defined or a
2079 non\-inline static function is unused.
2081 @item -Wunused-label
2082 @opindex Wunused-label
2083 Warn whenever a label is declared but not used.
2085 To suppress this warning use the @samp{unused} attribute
2086 (@pxref{Variable Attributes}).
2088 @item -Wunused-parameter
2089 @opindex Wunused-parameter
2090 Warn whenever a function parameter is unused aside from its declaration.
2092 To suppress this warning use the @samp{unused} attribute
2093 (@pxref{Variable Attributes}).
2095 @item -Wunused-variable
2096 @opindex Wunused-variable
2097 Warn whenever a local variable or non-constant static variable is unused
2098 aside from its declaration
2100 To suppress this warning use the @samp{unused} attribute
2101 (@pxref{Variable Attributes}).
2103 @item -Wunused-value
2104 @opindex Wunused-value
2105 Warn whenever a statement computes a result that is explicitly not used.
2107 To suppress this warning cast the expression to @samp{void}.
2111 All all the above @option{-Wunused} options combined.
2113 In order to get a warning about an unused function parameter, you must
2114 either specify @samp{-W -Wunused} or separately specify
2115 @option{-Wunused-parameter}.
2117 @item -Wuninitialized
2118 @opindex Wuninitialized
2119 Warn if an automatic variable is used without first being initialized or
2120 if a variable may be clobbered by a @code{setjmp} call.
2122 These warnings are possible only in optimizing compilation,
2123 because they require data flow information that is computed only
2124 when optimizing. If you don't specify @option{-O}, you simply won't
2127 These warnings occur only for variables that are candidates for
2128 register allocation. Therefore, they do not occur for a variable that
2129 is declared @code{volatile}, or whose address is taken, or whose size
2130 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2131 structures, unions or arrays, even when they are in registers.
2133 Note that there may be no warning about a variable that is used only
2134 to compute a value that itself is never used, because such
2135 computations may be deleted by data flow analysis before the warnings
2138 These warnings are made optional because GCC is not smart
2139 enough to see all the reasons why the code might be correct
2140 despite appearing to have an error. Here is one example of how
2161 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2162 always initialized, but GCC doesn't know this. Here is
2163 another common case:
2168 if (change_y) save_y = y, y = new_y;
2170 if (change_y) y = save_y;
2175 This has no bug because @code{save_y} is used only if it is set.
2177 @cindex @code{longjmp} warnings
2178 This option also warns when a non-volatile automatic variable might be
2179 changed by a call to @code{longjmp}. These warnings as well are possible
2180 only in optimizing compilation.
2182 The compiler sees only the calls to @code{setjmp}. It cannot know
2183 where @code{longjmp} will be called; in fact, a signal handler could
2184 call it at any point in the code. As a result, you may get a warning
2185 even when there is in fact no problem because @code{longjmp} cannot
2186 in fact be called at the place which would cause a problem.
2188 Some spurious warnings can be avoided if you declare all the functions
2189 you use that never return as @code{noreturn}. @xref{Function
2192 @item -Wreorder @r{(C++ only)}
2194 @cindex reordering, warning
2195 @cindex warning for reordering of member initializers
2196 Warn when the order of member initializers given in the code does not
2197 match the order in which they must be executed. For instance:
2199 @item -Wunknown-pragmas
2200 @opindex Wunknown-pragmas
2201 @cindex warning for unknown pragmas
2202 @cindex unknown pragmas, warning
2203 @cindex pragmas, warning of unknown
2204 Warn when a #pragma directive is encountered which is not understood by
2205 GCC@. If this command line option is used, warnings will even be issued
2206 for unknown pragmas in system header files. This is not the case if
2207 the warnings were only enabled by the @option{-Wall} command line option.
2211 All of the above @samp{-W} options combined. This enables all the
2212 warnings about constructions that some users consider questionable, and
2213 that are easy to avoid (or modify to prevent the warning), even in
2214 conjunction with macros.
2217 @opindex Wno-div-by-zero
2218 @opindex Wdiv-by-zero
2219 Warn about compile-time integer division by zero. This is default. To
2220 inhibit the warning messages, use @option{-Wno-div-by-zero}. Floating
2221 point division by zero is not warned about, as it can be a legitimate
2222 way of obtaining infinities and NaNs.
2225 @opindex Wno-multichar
2227 Warn if a multicharacter constant (@samp{'FOOF'}) is used. This is
2228 default. To inhibit the warning messages, use @option{-Wno-multichar}.
2229 Usually they indicate a typo in the user's code, as they have
2230 implementation-defined values, and should not be used in portable code.
2232 @item -Wsystem-headers
2233 @opindex Wsystem-headers
2234 @cindex warnings from system headers
2235 @cindex system headers, warnings from
2236 Print warning messages for constructs found in system header files.
2237 Warnings from system headers are normally suppressed, on the assumption
2238 that they usually do not indicate real problems and would only make the
2239 compiler output harder to read. Using this command line option tells
2240 GCC to emit warnings from system headers as if they occurred in user
2241 code. However, note that using @option{-Wall} in conjunction with this
2242 option will @emph{not} warn about unknown pragmas in system
2243 headers---for that, @option{-Wunknown-pragmas} must also be used.
2246 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2247 Some of them warn about constructions that users generally do not
2248 consider questionable, but which occasionally you might wish to check
2249 for; others warn about constructions that are necessary or hard to avoid
2250 in some cases, and there is no simple way to modify the code to suppress
2256 Print extra warning messages for these events:
2260 A function can return either with or without a value. (Falling
2261 off the end of the function body is considered returning without
2262 a value.) For example, this function would evoke such a
2276 An expression-statement or the left-hand side of a comma expression
2277 contains no side effects.
2278 To suppress the warning, cast the unused expression to void.
2279 For example, an expression such as @samp{x[i,j]} will cause a warning,
2280 but @samp{x[(void)i,j]} will not.
2283 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2286 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2287 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2288 that of ordinary mathematical notation.
2291 Storage-class specifiers like @code{static} are not the first things in
2292 a declaration. According to the C Standard, this usage is obsolescent.
2295 The return type of a function has a type qualifier such as @code{const}.
2296 Such a type qualifier has no effect, since the value returned by a
2297 function is not an lvalue. (But don't warn about the GNU extension of
2298 @code{volatile void} return types. That extension will be warned about
2299 if @option{-pedantic} is specified.)
2302 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2306 A comparison between signed and unsigned values could produce an
2307 incorrect result when the signed value is converted to unsigned.
2308 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2311 An aggregate has a partly bracketed initializer.
2312 For example, the following code would evoke such a warning,
2313 because braces are missing around the initializer for @code{x.h}:
2316 struct s @{ int f, g; @};
2317 struct t @{ struct s h; int i; @};
2318 struct t x = @{ 1, 2, 3 @};
2322 An aggregate has an initializer which does not initialize all members.
2323 For example, the following code would cause such a warning, because
2324 @code{x.h} would be implicitly initialized to zero:
2327 struct s @{ int f, g, h; @};
2328 struct s x = @{ 3, 4 @};
2333 @opindex Wfloat-equal
2334 Warn if floating point values are used in equality comparisons.
2336 The idea behind this is that sometimes it is convenient (for the
2337 programmer) to consider floating-point values as approximations to
2338 infinitely precise real numbers. If you are doing this, then you need
2339 to compute (by analysing the code, or in some other way) the maximum or
2340 likely maximum error that the computation introduces, and allow for it
2341 when performing comparisons (and when producing output, but that's a
2342 different problem). In particular, instead of testing for equality, you
2343 would check to see whether the two values have ranges that overlap; and
2344 this is done with the relational operators, so equality comparisons are
2347 @item -Wtraditional @r{(C only)}
2348 @opindex Wtraditional
2349 Warn about certain constructs that behave differently in traditional and
2350 ISO C@. Also warn about ISO C constructs that have no traditional C
2351 equivalent, and/or problematic constructs which should be avoided.
2355 Macro parameters that appear within string literals in the macro body.
2356 In traditional C macro replacement takes place within string literals,
2357 but does not in ISO C@.
2360 In traditional C, some preprocessor directives did not exist.
2361 Traditional preprocessors would only consider a line to be a directive
2362 if the @samp{#} appeared in column 1 on the line. Therefore
2363 @option{-Wtraditional} warns about directives that traditional C
2364 understands but would ignore because the @samp{#} does not appear as the
2365 first character on the line. It also suggests you hide directives like
2366 @samp{#pragma} not understood by traditional C by indenting them. Some
2367 traditional implementations would not recognize @samp{#elif}, so it
2368 suggests avoiding it altogether.
2371 A function-like macro that appears without arguments.
2374 The unary plus operator.
2377 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2378 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2379 constants.) Note, these suffixes appear in macros defined in the system
2380 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2381 Use of these macros in user code might normally lead to spurious
2382 warnings, however gcc's integrated preprocessor has enough context to
2383 avoid warning in these cases.
2386 A function declared external in one block and then used after the end of
2390 A @code{switch} statement has an operand of type @code{long}.
2393 A non-@code{static} function declaration follows a @code{static} one.
2394 This construct is not accepted by some traditional C compilers.
2397 The ISO type of an integer constant has a different width or
2398 signedness from its traditional type. This warning is only issued if
2399 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2400 typically represent bit patterns, are not warned about.
2403 Usage of ISO string concatenation is detected.
2406 Initialization of automatic aggregates.
2409 Identifier conflicts with labels. Traditional C lacks a separate
2410 namespace for labels.
2413 Initialization of unions. If the initializer is zero, the warning is
2414 omitted. This is done under the assumption that the zero initializer in
2415 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2416 initializer warnings and relies on default initialization to zero in the
2420 Conversions by prototypes between fixed/floating point values and vice
2421 versa. The absence of these prototypes when compiling with traditional
2422 C would cause serious problems. This is a subset of the possible
2423 conversion warnings, for the full set use @option{-Wconversion}.
2428 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2432 Warn whenever a local variable shadows another local variable, parameter or
2433 global variable or whenever a built-in function is shadowed.
2435 @item -Wlarger-than-@var{len}
2436 @opindex Wlarger-than
2437 Warn whenever an object of larger than @var{len} bytes is defined.
2439 @item -Wpointer-arith
2440 @opindex Wpointer-arith
2441 Warn about anything that depends on the ``size of'' a function type or
2442 of @code{void}. GNU C assigns these types a size of 1, for
2443 convenience in calculations with @code{void *} pointers and pointers
2446 @item -Wbad-function-cast @r{(C only)}
2447 @opindex Wbad-function-cast
2448 Warn whenever a function call is cast to a non-matching type.
2449 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2453 Warn whenever a pointer is cast so as to remove a type qualifier from
2454 the target type. For example, warn if a @code{const char *} is cast
2455 to an ordinary @code{char *}.
2458 @opindex Wcast-align
2459 Warn whenever a pointer is cast such that the required alignment of the
2460 target is increased. For example, warn if a @code{char *} is cast to
2461 an @code{int *} on machines where integers can only be accessed at
2462 two- or four-byte boundaries.
2464 @item -Wwrite-strings
2465 @opindex Wwrite-strings
2466 When compiling C, give string constants the type @code{const
2467 char[@var{length}]} so that
2468 copying the address of one into a non-@code{const} @code{char *}
2469 pointer will get a warning; when compiling C++, warn about the
2470 deprecated conversion from string constants to @code{char *}.
2471 These warnings will help you find at
2472 compile time code that can try to write into a string constant, but
2473 only if you have been very careful about using @code{const} in
2474 declarations and prototypes. Otherwise, it will just be a nuisance;
2475 this is why we did not make @option{-Wall} request these warnings.
2478 @opindex Wconversion
2479 Warn if a prototype causes a type conversion that is different from what
2480 would happen to the same argument in the absence of a prototype. This
2481 includes conversions of fixed point to floating and vice versa, and
2482 conversions changing the width or signedness of a fixed point argument
2483 except when the same as the default promotion.
2485 Also, warn if a negative integer constant expression is implicitly
2486 converted to an unsigned type. For example, warn about the assignment
2487 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2488 casts like @code{(unsigned) -1}.
2490 @item -Wsign-compare
2491 @opindex Wsign-compare
2492 @cindex warning for comparison of signed and unsigned values
2493 @cindex comparison of signed and unsigned values, warning
2494 @cindex signed and unsigned values, comparison warning
2495 Warn when a comparison between signed and unsigned values could produce
2496 an incorrect result when the signed value is converted to unsigned.
2497 This warning is also enabled by @option{-W}; to get the other warnings
2498 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2500 @item -Waggregate-return
2501 @opindex Waggregate-return
2502 Warn if any functions that return structures or unions are defined or
2503 called. (In languages where you can return an array, this also elicits
2506 @item -Wstrict-prototypes @r{(C only)}
2507 @opindex Wstrict-prototypes
2508 Warn if a function is declared or defined without specifying the
2509 argument types. (An old-style function definition is permitted without
2510 a warning if preceded by a declaration which specifies the argument
2513 @item -Wmissing-prototypes @r{(C only)}
2514 @opindex Wmissing-prototypes
2515 Warn if a global function is defined without a previous prototype
2516 declaration. This warning is issued even if the definition itself
2517 provides a prototype. The aim is to detect global functions that fail
2518 to be declared in header files.
2520 @item -Wmissing-declarations
2521 @opindex Wmissing-declarations
2522 Warn if a global function is defined without a previous declaration.
2523 Do so even if the definition itself provides a prototype.
2524 Use this option to detect global functions that are not declared in
2527 @item -Wmissing-noreturn
2528 @opindex Wmissing-noreturn
2529 Warn about functions which might be candidates for attribute @code{noreturn}.
2530 Note these are only possible candidates, not absolute ones. Care should
2531 be taken to manually verify functions actually do not ever return before
2532 adding the @code{noreturn} attribute, otherwise subtle code generation
2533 bugs could be introduced. You will not get a warning for @code{main} in
2534 hosted C environments.
2536 @item -Wmissing-format-attribute
2537 @opindex Wmissing-format-attribute
2539 If @option{-Wformat} is enabled, also warn about functions which might be
2540 candidates for @code{format} attributes. Note these are only possible
2541 candidates, not absolute ones. GCC will guess that @code{format}
2542 attributes might be appropriate for any function that calls a function
2543 like @code{vprintf} or @code{vscanf}, but this might not always be the
2544 case, and some functions for which @code{format} attributes are
2545 appropriate may not be detected. This option has no effect unless
2546 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2548 @item -Wno-deprecated-declarations
2549 @opindex Wno-deprecated-declarations
2550 Do not warn about uses of functions, variables, and types marked as
2551 deprecated by using the @code{deprecated} attribute.
2552 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2553 @pxref{Type Attributes}.)
2557 Warn if a structure is given the packed attribute, but the packed
2558 attribute has no effect on the layout or size of the structure.
2559 Such structures may be mis-aligned for little benefit. For
2560 instance, in this code, the variable @code{f.x} in @code{struct bar}
2561 will be misaligned even though @code{struct bar} does not itself
2562 have the packed attribute:
2569 @} __attribute__((packed));
2579 Warn if padding is included in a structure, either to align an element
2580 of the structure or to align the whole structure. Sometimes when this
2581 happens it is possible to rearrange the fields of the structure to
2582 reduce the padding and so make the structure smaller.
2584 @item -Wredundant-decls
2585 @opindex Wredundant-decls
2586 Warn if anything is declared more than once in the same scope, even in
2587 cases where multiple declaration is valid and changes nothing.
2589 @item -Wnested-externs @r{(C only)}
2590 @opindex Wnested-externs
2591 Warn if an @code{extern} declaration is encountered within a function.
2593 @item -Wunreachable-code
2594 @opindex Wunreachable-code
2595 Warn if the compiler detects that code will never be executed.
2597 This option is intended to warn when the compiler detects that at
2598 least a whole line of source code will never be executed, because
2599 some condition is never satisfied or because it is after a
2600 procedure that never returns.
2602 It is possible for this option to produce a warning even though there
2603 are circumstances under which part of the affected line can be executed,
2604 so care should be taken when removing apparently-unreachable code.
2606 For instance, when a function is inlined, a warning may mean that the
2607 line is unreachable in only one inlined copy of the function.
2609 This option is not made part of @option{-Wall} because in a debugging
2610 version of a program there is often substantial code which checks
2611 correct functioning of the program and is, hopefully, unreachable
2612 because the program does work. Another common use of unreachable
2613 code is to provide behavior which is selectable at compile-time.
2617 Warn if a function can not be inlined and it was declared as inline.
2621 @opindex Wno-long-long
2622 Warn if @samp{long long} type is used. This is default. To inhibit
2623 the warning messages, use @option{-Wno-long-long}. Flags
2624 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2625 only when @option{-pedantic} flag is used.
2627 @item -Wdisabled-optimization
2628 @opindex Wdisabled-optimization
2629 Warn if a requested optimization pass is disabled. This warning does
2630 not generally indicate that there is anything wrong with your code; it
2631 merely indicates that GCC's optimizers were unable to handle the code
2632 effectively. Often, the problem is that your code is too big or too
2633 complex; GCC will refuse to optimize programs when the optimization
2634 itself is likely to take inordinate amounts of time.
2638 Make all warnings into errors.
2641 @node Debugging Options
2642 @section Options for Debugging Your Program or GCC
2643 @cindex options, debugging
2644 @cindex debugging information options
2646 GCC has various special options that are used for debugging
2647 either your program or GCC:
2652 Produce debugging information in the operating system's native format
2653 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2656 On most systems that use stabs format, @option{-g} enables use of extra
2657 debugging information that only GDB can use; this extra information
2658 makes debugging work better in GDB but will probably make other debuggers
2660 refuse to read the program. If you want to control for certain whether
2661 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2662 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2663 or @option{-gvms} (see below).
2665 Unlike most other C compilers, GCC allows you to use @option{-g} with
2666 @option{-O}. The shortcuts taken by optimized code may occasionally
2667 produce surprising results: some variables you declared may not exist
2668 at all; flow of control may briefly move where you did not expect it;
2669 some statements may not be executed because they compute constant
2670 results or their values were already at hand; some statements may
2671 execute in different places because they were moved out of loops.
2673 Nevertheless it proves possible to debug optimized output. This makes
2674 it reasonable to use the optimizer for programs that might have bugs.
2676 The following options are useful when GCC is generated with the
2677 capability for more than one debugging format.
2681 Produce debugging information for use by GDB@. This means to use the
2682 most expressive format available (DWARF 2, stabs, or the native format
2683 if neither of those are supported), including GDB extensions if at all
2688 Produce debugging information in stabs format (if that is supported),
2689 without GDB extensions. This is the format used by DBX on most BSD
2690 systems. On MIPS, Alpha and System V Release 4 systems this option
2691 produces stabs debugging output which is not understood by DBX or SDB@.
2692 On System V Release 4 systems this option requires the GNU assembler.
2696 Produce debugging information in stabs format (if that is supported),
2697 using GNU extensions understood only by the GNU debugger (GDB)@. The
2698 use of these extensions is likely to make other debuggers crash or
2699 refuse to read the program.
2703 Produce debugging information in COFF format (if that is supported).
2704 This is the format used by SDB on most System V systems prior to
2709 Produce debugging information in XCOFF format (if that is supported).
2710 This is the format used by the DBX debugger on IBM RS/6000 systems.
2714 Produce debugging information in XCOFF format (if that is supported),
2715 using GNU extensions understood only by the GNU debugger (GDB)@. The
2716 use of these extensions is likely to make other debuggers crash or
2717 refuse to read the program, and may cause assemblers other than the GNU
2718 assembler (GAS) to fail with an error.
2722 Produce debugging information in DWARF version 1 format (if that is
2723 supported). This is the format used by SDB on most System V Release 4
2728 Produce debugging information in DWARF version 1 format (if that is
2729 supported), using GNU extensions understood only by the GNU debugger
2730 (GDB)@. The use of these extensions is likely to make other debuggers
2731 crash or refuse to read the program.
2735 Produce debugging information in DWARF version 2 format (if that is
2736 supported). This is the format used by DBX on IRIX 6.
2740 Produce debugging information in VMS debug format (if that is
2741 supported). This is the format used by DEBUG on VMS systems.
2744 @itemx -ggdb@var{level}
2745 @itemx -gstabs@var{level}
2746 @itemx -gcoff@var{level}
2747 @itemx -gxcoff@var{level}
2748 @itemx -gdwarf@var{level}
2749 @itemx -gdwarf-2@var{level}
2750 @itemx -gvms@var{level}
2751 Request debugging information and also use @var{level} to specify how
2752 much information. The default level is 2.
2754 Level 1 produces minimal information, enough for making backtraces in
2755 parts of the program that you don't plan to debug. This includes
2756 descriptions of functions and external variables, but no information
2757 about local variables and no line numbers.
2759 Level 3 includes extra information, such as all the macro definitions
2760 present in the program. Some debuggers support macro expansion when
2761 you use @option{-g3}.
2766 Generate extra code to write profile information suitable for the
2767 analysis program @code{prof}. You must use this option when compiling
2768 the source files you want data about, and you must also use it when
2771 @cindex @code{gprof}
2774 Generate extra code to write profile information suitable for the
2775 analysis program @code{gprof}. You must use this option when compiling
2776 the source files you want data about, and you must also use it when
2782 Generate extra code to write profile information for basic blocks, which will
2783 record the number of times each basic block is executed, the basic block start
2784 address, and the function name containing the basic block. If @option{-g} is
2785 used, the line number and filename of the start of the basic block will also be
2786 recorded. If not overridden by the machine description, the default action is
2787 to append to the text file @file{bb.out}.
2789 This data could be analyzed by a program like @code{tcov}. Note,
2790 however, that the format of the data is not what @code{tcov} expects.
2791 Eventually GNU @code{gprof} should be extended to process this data.
2795 Makes the compiler print out each function name as it is compiled, and
2796 print some statistics about each pass when it finishes.
2799 @opindex ftime-report
2800 Makes the compiler print some statistics about the time consumed by each
2801 pass when it finishes.
2804 @opindex fmem-report
2805 Makes the compiler print some statistics about permanent memory
2806 allocation when it finishes.
2808 @item -fprofile-arcs
2809 @opindex fprofile-arcs
2810 Instrument @dfn{arcs} during compilation to generate coverage data
2811 or for profile-directed block ordering. During execution the program
2812 records how many times each branch is executed and how many times it is
2813 taken. When the compiled program exits it saves this data to a file
2814 called @file{@var{sourcename}.da} for each source file.
2816 For profile-directed block ordering, compile the program with
2817 @option{-fprofile-arcs} plus optimization and code generation options,
2818 generate the arc profile information by running the program on a
2819 selected workload, and then compile the program again with the same
2820 optimization and code generation options plus
2821 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2822 Control Optimization}).
2824 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2825 when it is used with the @option{-ftest-coverage} option. GCC
2826 supports two methods of determining code coverage: the options that
2827 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2828 write information to text files. The options that support @code{gcov}
2829 do not need to instrument every arc in the program, so a program compiled
2830 with them runs faster than a program compiled with @option{-a}, which
2831 adds instrumentation code to every basic block in the program. The
2832 tradeoff: since @code{gcov} does not have execution counts for all
2833 branches, it must start with the execution counts for the instrumented
2834 branches, and then iterate over the program flow graph until the entire
2835 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2836 a program which uses information from @option{-a} and @option{-ax}.
2838 With @option{-fprofile-arcs}, for each function of your program GCC
2839 creates a program flow graph, then finds a spanning tree for the graph.
2840 Only arcs that are not on the spanning tree have to be instrumented: the
2841 compiler adds code to count the number of times that these arcs are
2842 executed. When an arc is the only exit or only entrance to a block, the
2843 instrumentation code can be added to the block; otherwise, a new basic
2844 block must be created to hold the instrumentation code.
2846 This option makes it possible to estimate branch probabilities and to
2847 calculate basic block execution counts. In general, basic block
2848 execution counts as provided by @option{-a} do not give enough
2849 information to estimate all branch probabilities.
2852 @item -ftest-coverage
2853 @opindex ftest-coverage
2854 Create data files for the @code{gcov} code-coverage utility
2855 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2856 The data file names begin with the name of your source file:
2859 @item @var{sourcename}.bb
2860 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2861 associate basic block execution counts with line numbers.
2863 @item @var{sourcename}.bbg
2864 A list of all arcs in the program flow graph. This allows @code{gcov}
2865 to reconstruct the program flow graph, so that it can compute all basic
2866 block and arc execution counts from the information in the
2867 @code{@var{sourcename}.da} file.
2870 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2871 option adds instrumentation to the program, which then writes
2872 execution counts to another data file:
2875 @item @var{sourcename}.da
2876 Runtime arc execution counts, used in conjunction with the arc
2877 information in the file @code{@var{sourcename}.bbg}.
2880 Coverage data will map better to the source files if
2881 @option{-ftest-coverage} is used without optimization.
2883 @item -d@var{letters}
2885 Says to make debugging dumps during compilation at times specified by
2886 @var{letters}. This is used for debugging the compiler. The file names
2887 for most of the dumps are made by appending a pass number and a word to
2888 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2889 Here are the possible letters for use in @var{letters}, and their meanings:
2894 Annotate the assembler output with miscellaneous debugging information.
2897 Dump after computing branch probabilities, to @file{@var{file}.14.bp}.
2900 Dump after block reordering, to @file{@var{file}.29.bbro}.
2903 Dump after instruction combination, to the file @file{@var{file}.16.combine}.
2906 Dump after the first if conversion, to the file @file{@var{file}.17.ce}.
2909 Dump after delayed branch scheduling, to @file{@var{file}.31.dbr}.
2912 Dump all macro definitions, at the end of preprocessing, in addition to
2916 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2917 @file{@var{file}.07.ussa}.
2920 Dump after the second if conversion, to @file{@var{file}.26.ce2}.
2923 Dump after life analysis, to @file{@var{file}.15.life}.
2926 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.09.addressof}.
2929 Dump after global register allocation, to @file{@var{file}.21.greg}.
2932 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2935 Dump after reg-to-stack conversion, to @file{@var{file}.28.stack}.
2938 Dump after post-reload optimizations, to @file{@var{file}.22.postreload}.
2941 Dump after GCSE, to @file{@var{file}.10.gcse}.
2944 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2947 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2950 Dump after conversion from registers to stack, to @file{@var{file}.32.stack}.
2953 Dump after local register allocation, to @file{@var{file}.20.lreg}.
2956 Dump after loop optimization, to @file{@var{file}.11.loop}.
2959 Dump after performing the machine dependent reorganisation pass, to
2960 @file{@var{file}.30.mach}.
2963 Dump after register renumbering, to @file{@var{file}.25.rnreg}.
2966 Dump after the register move pass, to @file{@var{file}.18.regmove}.
2969 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2972 Dump after the second scheduling pass, to @file{@var{file}.27.sched2}.
2975 Dump after CSE (including the jump optimization that sometimes follows
2976 CSE), to @file{@var{file}.08.cse}.
2979 Dump after the first scheduling pass, to @file{@var{file}.19.sched}.
2982 Dump after the second CSE pass (including the jump optimization that
2983 sometimes follows CSE), to @file{@var{file}.12.cse2}.
2986 Dump after the second flow pass, to @file{@var{file}.23.flow2}.
2989 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2992 Dump after the peephole pass, to @file{@var{file}.24.peephole2}.
2995 Produce all the dumps listed above.
2998 Print statistics on memory usage, at the end of the run, to
3002 Annotate the assembler output with a comment indicating which
3003 pattern and alternative was used. The length of each instruction is
3007 Dump the RTL in the assembler output as a comment before each instruction.
3008 Also turns on @option{-dp} annotation.
3011 For each of the other indicated dump files (except for
3012 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
3013 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3016 Just generate RTL for a function instead of compiling it. Usually used
3020 Dump debugging information during parsing, to standard error.
3023 @item -fdump-unnumbered
3024 @opindex fdump-unnumbered
3025 When doing debugging dumps (see @option{-d} option above), suppress instruction
3026 numbers and line number note output. This makes it more feasible to
3027 use diff on debugging dumps for compiler invocations with different
3028 options, in particular with and without @option{-g}.
3030 @item -fdump-class-hierarchy @r{(C++ only)}
3031 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3032 @opindex fdump-class-hierarchy
3033 Dump a representation of each class's hierarchy and virtual function
3034 table layout to a file. The file name is made by appending @file{.class}
3035 to the source file name. If the @samp{-@var{options}} form is used,
3036 @var{options} controls the details of the dump as described for the
3037 @option{-fdump-tree} options.
3039 @item -fdump-tree-@var{switch} @r{(C++ only)}
3040 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3042 Control the dumping at various stages of processing the intermediate
3043 language tree to a file. The file name is generated by appending a switch
3044 specific suffix to the source file name. If the @samp{-@var{options}}
3045 form is used, @var{options} is a list of @samp{-} separated options that
3046 control the details of the dump. Not all options are applicable to all
3047 dumps, those which are not meaningful will be ignored. The following
3048 options are available
3052 Print the address of each node. Usually this is not meaningful as it
3053 changes according to the environment and source file. Its primary use
3054 is for tying up a dump file with a debug environment.
3056 Inhibit dumping of members of a scope or body of a function merely
3057 because that scope has been reached. Only dump such items when they
3058 are directly reachable by some other path.
3060 Turn on all options.
3063 The following tree dumps are possible:
3066 Dump before any tree based optimization, to @file{@var{file}.original}.
3068 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3070 Dump after function inlining, to @file{@var{file}.inlined}.
3073 @item -fpretend-float
3074 @opindex fpretend-float
3075 When running a cross-compiler, pretend that the target machine uses the
3076 same floating point format as the host machine. This causes incorrect
3077 output of the actual floating constants, but the actual instruction
3078 sequence will probably be the same as GCC would make when running on
3083 Store the usual ``temporary'' intermediate files permanently; place them
3084 in the current directory and name them based on the source file. Thus,
3085 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3086 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3087 preprocessed @file{foo.i} output file even though the compiler now
3088 normally uses an integrated preprocessor.
3092 Report the CPU time taken by each subprocess in the compilation
3093 sequence. For C source files, this is the compiler proper and assembler
3094 (plus the linker if linking is done). The output looks like this:
3101 The first number on each line is the ``user time,'' that is time spent
3102 executing the program itself. The second number is ``system time,''
3103 time spent executing operating system routines on behalf of the program.
3104 Both numbers are in seconds.
3106 @item -print-file-name=@var{library}
3107 @opindex print-file-name
3108 Print the full absolute name of the library file @var{library} that
3109 would be used when linking---and don't do anything else. With this
3110 option, GCC does not compile or link anything; it just prints the
3113 @item -print-multi-directory
3114 @opindex print-multi-directory
3115 Print the directory name corresponding to the multilib selected by any
3116 other switches present in the command line. This directory is supposed
3117 to exist in @env{GCC_EXEC_PREFIX}.
3119 @item -print-multi-lib
3120 @opindex print-multi-lib
3121 Print the mapping from multilib directory names to compiler switches
3122 that enable them. The directory name is separated from the switches by
3123 @samp{;}, and each switch starts with an @samp{@@} instead of the
3124 @samp{-}, without spaces between multiple switches. This is supposed to
3125 ease shell-processing.
3127 @item -print-prog-name=@var{program}
3128 @opindex print-prog-name
3129 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3131 @item -print-libgcc-file-name
3132 @opindex print-libgcc-file-name
3133 Same as @option{-print-file-name=libgcc.a}.
3135 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3136 but you do want to link with @file{libgcc.a}. You can do
3139 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3142 @item -print-search-dirs
3143 @opindex print-search-dirs
3144 Print the name of the configured installation directory and a list of
3145 program and library directories gcc will search---and don't do anything else.
3147 This is useful when gcc prints the error message
3148 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3149 To resolve this you either need to put @file{cpp0} and the other compiler
3150 components where gcc expects to find them, or you can set the environment
3151 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3152 Don't forget the trailing '/'.
3153 @xref{Environment Variables}.
3156 @opindex dumpmachine
3157 Print the compiler's target machine (for example,
3158 @samp{i686-pc-linux-gnu})---and don't do anything else.
3161 @opindex dumpversion
3162 Print the compiler version (for example, @samp{3.0})---and don't do
3167 Print the compiler's built-in specs---and don't do anything else. (This
3168 is used when GCC itself is being built.) @xref{Spec Files}.
3171 @node Optimize Options
3172 @section Options That Control Optimization
3173 @cindex optimize options
3174 @cindex options, optimization
3176 These options control various sorts of optimizations:
3183 Optimize. Optimizing compilation takes somewhat more time, and a lot
3184 more memory for a large function.
3186 Without @option{-O}, the compiler's goal is to reduce the cost of
3187 compilation and to make debugging produce the expected results.
3188 Statements are independent: if you stop the program with a breakpoint
3189 between statements, you can then assign a new value to any variable or
3190 change the program counter to any other statement in the function and
3191 get exactly the results you would expect from the source code.
3193 With @option{-O}, the compiler tries to reduce code size and execution
3194 time, without performing any optimizations that take a great deal of
3199 Optimize even more. GCC performs nearly all supported optimizations
3200 that do not involve a space-speed tradeoff. The compiler does not
3201 perform loop unrolling or function inlining when you specify @option{-O2}.
3202 As compared to @option{-O}, this option increases both compilation time
3203 and the performance of the generated code.
3205 @option{-O2} turns on all optional optimizations except for loop unrolling,
3206 function inlining, and register renaming. It also turns on the
3207 @option{-fforce-mem} option on all machines and frame pointer elimination
3208 on machines where doing so does not interfere with debugging.
3210 Please note the warning under @option{-fgcse} about
3211 invoking @option{-O2} on programs that use computed gotos.
3215 Optimize yet more. @option{-O3} turns on all optimizations specified by
3216 @option{-O2} and also turns on the @option{-finline-functions} and
3217 @option{-frename-registers} options.
3225 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3226 do not typically increase code size. It also performs further
3227 optimizations designed to reduce code size.
3229 If you use multiple @option{-O} options, with or without level numbers,
3230 the last such option is the one that is effective.
3233 Options of the form @option{-f@var{flag}} specify machine-independent
3234 flags. Most flags have both positive and negative forms; the negative
3235 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3236 only one of the forms is listed---the one which is not the default.
3237 You can figure out the other form by either removing @samp{no-} or
3242 @opindex ffloat-store
3243 Do not store floating point variables in registers, and inhibit other
3244 options that might change whether a floating point value is taken from a
3247 @cindex floating point precision
3248 This option prevents undesirable excess precision on machines such as
3249 the 68000 where the floating registers (of the 68881) keep more
3250 precision than a @code{double} is supposed to have. Similarly for the
3251 x86 architecture. For most programs, the excess precision does only
3252 good, but a few programs rely on the precise definition of IEEE floating
3253 point. Use @option{-ffloat-store} for such programs, after modifying
3254 them to store all pertinent intermediate computations into variables.
3256 @item -fno-default-inline
3257 @opindex fno-default-inline
3258 Do not make member functions inline by default merely because they are
3259 defined inside the class scope (C++ only). Otherwise, when you specify
3260 @w{@option{-O}}, member functions defined inside class scope are compiled
3261 inline by default; i.e., you don't need to add @samp{inline} in front of
3262 the member function name.
3264 @item -fno-defer-pop
3265 @opindex fno-defer-pop
3266 Always pop the arguments to each function call as soon as that function
3267 returns. For machines which must pop arguments after a function call,
3268 the compiler normally lets arguments accumulate on the stack for several
3269 function calls and pops them all at once.
3273 Force memory operands to be copied into registers before doing
3274 arithmetic on them. This produces better code by making all memory
3275 references potential common subexpressions. When they are not common
3276 subexpressions, instruction combination should eliminate the separate
3277 register-load. The @option{-O2} option turns on this option.
3280 @opindex fforce-addr
3281 Force memory address constants to be copied into registers before
3282 doing arithmetic on them. This may produce better code just as
3283 @option{-fforce-mem} may.
3285 @item -fomit-frame-pointer
3286 @opindex fomit-frame-pointer
3287 Don't keep the frame pointer in a register for functions that
3288 don't need one. This avoids the instructions to save, set up and
3289 restore frame pointers; it also makes an extra register available
3290 in many functions. @strong{It also makes debugging impossible on
3293 On some machines, such as the VAX, this flag has no effect, because
3294 the standard calling sequence automatically handles the frame pointer
3295 and nothing is saved by pretending it doesn't exist. The
3296 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3297 whether a target machine supports this flag. @xref{Registers,,Register
3298 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3300 @item -foptimize-sibling-calls
3301 @opindex foptimize-sibling-calls
3302 Optimize sibling and tail recursive calls.
3306 This option generates traps for signed overflow on addition, subtraction,
3307 multiplication operations.
3311 Don't pay attention to the @code{inline} keyword. Normally this option
3312 is used to keep the compiler from expanding any functions inline.
3313 Note that if you are not optimizing, no functions can be expanded inline.
3315 @item -finline-functions
3316 @opindex finline-functions
3317 Integrate all simple functions into their callers. The compiler
3318 heuristically decides which functions are simple enough to be worth
3319 integrating in this way.
3321 If all calls to a given function are integrated, and the function is
3322 declared @code{static}, then the function is normally not output as
3323 assembler code in its own right.
3325 @item -finline-limit=@var{n}
3326 @opindex finline-limit
3327 By default, gcc limits the size of functions that can be inlined. This flag
3328 allows the control of this limit for functions that are explicitly marked as
3329 inline (ie marked with the inline keyword or defined within the class
3330 definition in c++). @var{n} is the size of functions that can be inlined in
3331 number of pseudo instructions (not counting parameter handling). The default
3332 value of @var{n} is 600.
3333 Increasing this value can result in more inlined code at
3334 the cost of compilation time and memory consumption. Decreasing usually makes
3335 the compilation faster and less code will be inlined (which presumably
3336 means slower programs). This option is particularly useful for programs that
3337 use inlining heavily such as those based on recursive templates with C++.
3339 @emph{Note:} pseudo instruction represents, in this particular context, an
3340 abstract measurement of function's size. In no way, it represents a count
3341 of assembly instructions and as such its exact meaning might change from one
3342 release to an another.
3344 @item -fkeep-inline-functions
3345 @opindex fkeep-inline-functions
3346 Even if all calls to a given function are integrated, and the function
3347 is declared @code{static}, nevertheless output a separate run-time
3348 callable version of the function. This switch does not affect
3349 @code{extern inline} functions.
3351 @item -fkeep-static-consts
3352 @opindex fkeep-static-consts
3353 Emit variables declared @code{static const} when optimization isn't turned
3354 on, even if the variables aren't referenced.
3356 GCC enables this option by default. If you want to force the compiler to
3357 check if the variable was referenced, regardless of whether or not
3358 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3360 @item -fmerge-constants
3361 Attempt to merge identical constants (string constants and floating point
3362 constants) accross compilation units.
3364 This option is default for optimized compilation if assembler and linker
3365 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3367 @item -fmerge-all-constants
3368 Attempt to merge identical constants and identical variables.
3370 This option implies @option{-fmerge-constants}. In addition to
3371 @option{-fmerge-constants} this considers e.g. even constant initialized
3372 arrays or initialized constant variables with integral or floating point
3373 types. Languages like C or C++ require each non-automatic variable to
3374 have distinct location, so using this option will result in non-conforming
3377 @item -fno-function-cse
3378 @opindex fno-function-cse
3379 Do not put function addresses in registers; make each instruction that
3380 calls a constant function contain the function's address explicitly.
3382 This option results in less efficient code, but some strange hacks
3383 that alter the assembler output may be confused by the optimizations
3384 performed when this option is not used.
3388 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3389 @option{-fno-trapping-math}.
3391 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3393 This option should never be turned on by any @option{-O} option since
3394 it can result in incorrect output for programs which depend on
3395 an exact implementation of IEEE or ISO rules/specifications for
3398 @item -fno-math-errno
3399 @opindex fno-math-errno
3400 Do not set ERRNO after calling math functions that are executed
3401 with a single instruction, e.g., sqrt. A program that relies on
3402 IEEE exceptions for math error handling may want to use this flag
3403 for speed while maintaining IEEE arithmetic compatibility.
3405 This option should never be turned on by any @option{-O} option since
3406 it can result in incorrect output for programs which depend on
3407 an exact implementation of IEEE or ISO rules/specifications for
3410 The default is @option{-fmath-errno}.
3412 @item -funsafe-math-optimizations
3413 @opindex funsafe-math-optimizations
3414 Allow optimizations for floating-point arithmetic that (a) assume
3415 that arguments and results are valid and (b) may violate IEEE or
3416 ANSI standards. When used at link-time, it may include libraries
3417 or startup files that change the default FPU control word or other
3418 similar optimizations.
3420 This option should never be turned on by any @option{-O} option since
3421 it can result in incorrect output for programs which depend on
3422 an exact implementation of IEEE or ISO rules/specifications for
3425 The default is @option{-fno-unsafe-math-optimizations}.
3427 @item -fno-trapping-math
3428 @opindex fno-trapping-math
3429 Compile code assuming that floating-point operations cannot generate
3430 user-visible traps. Setting this option may allow faster code
3431 if one relies on ``non-stop'' IEEE arithmetic, for example.
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{-ftrapping-math}.
3441 The following options control specific optimizations. The @option{-O2}
3442 option turns on all of these optimizations except @option{-funroll-loops}
3443 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3444 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3445 but specific machines may handle it differently.
3447 You can use the following flags in the rare cases when ``fine-tuning''
3448 of optimizations to be performed is desired.
3450 Not all of the optimizations performed by GCC have @option{-f} options
3454 @item -fstrength-reduce
3455 @opindex fstrength-reduce
3456 Perform the optimizations of loop strength reduction and
3457 elimination of iteration variables.
3459 @item -fthread-jumps
3460 @opindex fthread-jumps
3461 Perform optimizations where we check to see if a jump branches to a
3462 location where another comparison subsumed by the first is found. If
3463 so, the first branch is redirected to either the destination of the
3464 second branch or a point immediately following it, depending on whether
3465 the condition is known to be true or false.
3467 @item -fcse-follow-jumps
3468 @opindex fcse-follow-jumps
3469 In common subexpression elimination, scan through jump instructions
3470 when the target of the jump is not reached by any other path. For
3471 example, when CSE encounters an @code{if} statement with an
3472 @code{else} clause, CSE will follow the jump when the condition
3475 @item -fcse-skip-blocks
3476 @opindex fcse-skip-blocks
3477 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3478 follow jumps which conditionally skip over blocks. When CSE
3479 encounters a simple @code{if} statement with no else clause,
3480 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3481 body of the @code{if}.
3483 @item -frerun-cse-after-loop
3484 @opindex frerun-cse-after-loop
3485 Re-run common subexpression elimination after loop optimizations has been
3488 @item -frerun-loop-opt
3489 @opindex frerun-loop-opt
3490 Run the loop optimizer twice.
3494 Perform a global common subexpression elimination pass.
3495 This pass also performs global constant and copy propagation.
3497 @emph{Note:} When compiling a program using computed gotos, a GCC
3498 extension, you may get better runtime performance if you disable
3499 the global common subexpression elmination pass by adding
3500 @option{-fno-gcse} to the command line.
3504 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3505 attempt to move loads which are only killed by stores into themselves. This
3506 allows a loop containing a load/store sequence to be changed to a load outside
3507 the loop, and a copy/store within the loop.
3511 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3512 subexpression elimination. This pass will attempt to move stores out of loops.
3513 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3514 can be changed to a load before the loop and a store after the loop.
3516 @item -fdelete-null-pointer-checks
3517 @opindex fdelete-null-pointer-checks
3518 Use global dataflow analysis to identify and eliminate useless checks
3519 for null pointers. The compiler assumes that dereferencing a null
3520 pointer would have halted the program. If a pointer is checked after
3521 it has already been dereferenced, it cannot be null.
3523 In some environments, this assumption is not true, and programs can
3524 safely dereference null pointers. Use
3525 @option{-fno-delete-null-pointer-checks} to disable this optimization
3526 for programs which depend on that behavior.
3528 @item -fexpensive-optimizations
3529 @opindex fexpensive-optimizations
3530 Perform a number of minor optimizations that are relatively expensive.
3532 @item -foptimize-register-move
3534 @opindex foptimize-register-move
3536 Attempt to reassign register numbers in move instructions and as
3537 operands of other simple instructions in order to maximize the amount of
3538 register tying. This is especially helpful on machines with two-operand
3539 instructions. GCC enables this optimization by default with @option{-O2}
3542 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3545 @item -fdelayed-branch
3546 @opindex fdelayed-branch
3547 If supported for the target machine, attempt to reorder instructions
3548 to exploit instruction slots available after delayed branch
3551 @item -fschedule-insns
3552 @opindex fschedule-insns
3553 If supported for the target machine, attempt to reorder instructions to
3554 eliminate execution stalls due to required data being unavailable. This
3555 helps machines that have slow floating point or memory load instructions
3556 by allowing other instructions to be issued until the result of the load
3557 or floating point instruction is required.
3559 @item -fschedule-insns2
3560 @opindex fschedule-insns2
3561 Similar to @option{-fschedule-insns}, but requests an additional pass of
3562 instruction scheduling after register allocation has been done. This is
3563 especially useful on machines with a relatively small number of
3564 registers and where memory load instructions take more than one cycle.
3566 @item -ffunction-sections
3567 @itemx -fdata-sections
3568 @opindex ffunction-sections
3569 @opindex fdata-sections
3570 Place each function or data item into its own section in the output
3571 file if the target supports arbitrary sections. The name of the
3572 function or the name of the data item determines the section's name
3575 Use these options on systems where the linker can perform optimizations
3576 to improve locality of reference in the instruction space. HPPA
3577 processors running HP-UX and Sparc processors running Solaris 2 have
3578 linkers with such optimizations. Other systems using the ELF object format
3579 as well as AIX may have these optimizations in the future.
3581 Only use these options when there are significant benefits from doing
3582 so. When you specify these options, the assembler and linker will
3583 create larger object and executable files and will also be slower.
3584 You will not be able to use @code{gprof} on all systems if you
3585 specify this option and you may have problems with debugging if
3586 you specify both this option and @option{-g}.
3588 @item -fcaller-saves
3589 @opindex fcaller-saves
3590 Enable values to be allocated in registers that will be clobbered by
3591 function calls, by emitting extra instructions to save and restore the
3592 registers around such calls. Such allocation is done only when it
3593 seems to result in better code than would otherwise be produced.
3595 This option is always enabled by default on certain machines, usually
3596 those which have no call-preserved registers to use instead.
3598 For all machines, optimization level 2 and higher enables this flag by
3601 @item -funroll-loops
3602 @opindex funroll-loops
3603 Unroll loops whose number of iterations can be determined at compile
3604 time or upon entry to the loop. @option{-funroll-loops} implies both
3605 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3606 option makes code larger, and may or may not make it run faster.
3608 @item -funroll-all-loops
3609 @opindex funroll-all-loops
3610 Unroll all loops, even if their number of iterations is uncertain when
3611 the loop is entered. This usually makes programs run more slowly.
3612 @option{-funroll-all-loops} implies the same options as
3613 @option{-funroll-loops},
3615 @item -fprefetch-loop-arrays
3616 @opindex fprefetch-loop-arrays
3617 If supported by the target machine, generate instructions to prefetch
3618 memory to improve the performance of loops that access large arrays.
3620 @item -fmove-all-movables
3621 @opindex fmove-all-movables
3622 Forces all invariant computations in loops to be moved
3625 @item -freduce-all-givs
3626 @opindex freduce-all-givs
3627 Forces all general-induction variables in loops to be
3630 @emph{Note:} When compiling programs written in Fortran,
3631 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3632 by default when you use the optimizer.
3634 These options may generate better or worse code; results are highly
3635 dependent on the structure of loops within the source code.
3637 These two options are intended to be removed someday, once
3638 they have helped determine the efficacy of various
3639 approaches to improving loop optimizations.
3641 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3642 know how use of these options affects
3643 the performance of your production code.
3644 We're very interested in code that runs @emph{slower}
3645 when these options are @emph{enabled}.
3648 @itemx -fno-peephole2
3649 @opindex fno-peephole
3650 @opindex fno-peephole2
3651 Disable any machine-specific peephole optimizations. The difference
3652 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3653 are implemented in the compiler; some targets use one, some use the
3654 other, a few use both.
3656 @item -fbranch-probabilities
3657 @opindex fbranch-probabilities
3658 After running a program compiled with @option{-fprofile-arcs}
3659 (@pxref{Debugging Options,, Options for Debugging Your Program or
3660 @command{gcc}}), you can compile it a second time using
3661 @option{-fbranch-probabilities}, to improve optimizations based on
3662 the number of times each branch was taken. When the program
3663 compiled with @option{-fprofile-arcs} exits it saves arc execution
3664 counts to a file called @file{@var{sourcename}.da} for each source
3665 file The information in this data file is very dependent on the
3666 structure of the generated code, so you must use the same source code
3667 and the same optimization options for both compilations.
3669 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3670 note on the first instruction of each basic block, and a
3671 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3672 These can be used to improve optimization. Currently, they are only
3673 used in one place: in @file{reorg.c}, instead of guessing which path a
3674 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3675 exactly determine which path is taken more often.
3677 @item -fno-guess-branch-probability
3678 @opindex fno-guess-branch-probability
3679 Do not guess branch probabilities using a randomized model.
3681 Sometimes gcc will opt to use a randomized model to guess branch
3682 probabilities, when none are available from either profiling feedback
3683 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3684 different runs of the compiler on the same program may produce different
3687 In a hard real-time system, people don't want different runs of the
3688 compiler to produce code that has different behavior; minimizing
3689 non-determinism is of paramount import. This switch allows users to
3690 reduce non-determinism, possibly at the expense of inferior
3693 @item -fstrict-aliasing
3694 @opindex fstrict-aliasing
3695 Allows the compiler to assume the strictest aliasing rules applicable to
3696 the language being compiled. For C (and C++), this activates
3697 optimizations based on the type of expressions. In particular, an
3698 object of one type is assumed never to reside at the same address as an
3699 object of a different type, unless the types are almost the same. For
3700 example, an @code{unsigned int} can alias an @code{int}, but not a
3701 @code{void*} or a @code{double}. A character type may alias any other
3704 Pay special attention to code like this:
3717 The practice of reading from a different union member than the one most
3718 recently written to (called ``type-punning'') is common. Even with
3719 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3720 is accessed through the union type. So, the code above will work as
3721 expected. However, this code might not:
3732 Every language that wishes to perform language-specific alias analysis
3733 should define a function that computes, given an @code{tree}
3734 node, an alias set for the node. Nodes in different alias sets are not
3735 allowed to alias. For an example, see the C front-end function
3736 @code{c_get_alias_set}.
3738 @item -falign-functions
3739 @itemx -falign-functions=@var{n}
3740 @opindex falign-functions
3741 Align the start of functions to the next power-of-two greater than
3742 @var{n}, skipping up to @var{n} bytes. For instance,
3743 @option{-falign-functions=32} aligns functions to the next 32-byte
3744 boundary, but @option{-falign-functions=24} would align to the next
3745 32-byte boundary only if this can be done by skipping 23 bytes or less.
3747 @option{-fno-align-functions} and @option{-falign-functions=1} are
3748 equivalent and mean that functions will not be aligned.
3750 Some assemblers only support this flag when @var{n} is a power of two;
3751 in that case, it is rounded up.
3753 If @var{n} is not specified, use a machine-dependent default.
3755 @item -falign-labels
3756 @itemx -falign-labels=@var{n}
3757 @opindex falign-labels
3758 Align all branch targets to a power-of-two boundary, skipping up to
3759 @var{n} bytes like @option{-falign-functions}. This option can easily
3760 make code slower, because it must insert dummy operations for when the
3761 branch target is reached in the usual flow of the code.
3763 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3764 are greater than this value, then their values are used instead.
3766 If @var{n} is not specified, use a machine-dependent default which is
3767 very likely to be @samp{1}, meaning no alignment.
3770 @itemx -falign-loops=@var{n}
3771 @opindex falign-loops
3772 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3773 like @option{-falign-functions}. The hope is that the loop will be
3774 executed many times, which will make up for any execution of the dummy
3777 If @var{n} is not specified, use a machine-dependent default.
3780 @itemx -falign-jumps=@var{n}
3781 @opindex falign-jumps
3782 Align branch targets to a power-of-two boundary, for branch targets
3783 where the targets can only be reached by jumping, skipping up to @var{n}
3784 bytes like @option{-falign-functions}. In this case, no dummy operations
3787 If @var{n} is not specified, use a machine-dependent default.
3791 Perform optimizations in static single assignment form. Each function's
3792 flow graph is translated into SSA form, optimizations are performed, and
3793 the flow graph is translated back from SSA form. Users should not
3794 specify this option, since it is not yet ready for production use.
3798 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3799 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3803 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3804 Like @option{-fssa}, this is an experimental feature.
3806 @item -fsingle-precision-constant
3807 @opindex fsingle-precision-constant
3808 Treat floating point constant as single precision constant instead of
3809 implicitly converting it to double precision constant.
3811 @item -frename-registers
3812 @opindex frename-registers
3813 Attempt to avoid false dependencies in scheduled code by making use
3814 of registers left over after register allocation. This optimization
3815 will most benefit processors with lots of registers. It can, however,
3816 make debugging impossible, since variables will no longer stay in
3817 a ``home register''.
3819 @item -fno-cprop-registers
3820 @opindex fno-cprop-registers
3821 After register allocation and post-register allocation instruction splitting,
3822 we perform a copy-propagation pass to try to reduce scheduling dependencies
3823 and occasionally eliminate the copy.
3825 @item --param @var{name}=@var{value}
3827 In some places, GCC uses various constants to control the amount of
3828 optimization that is done. For example, GCC will not inline functions
3829 that contain more that a certain number of instructions. You can
3830 control some of these constants on the command-line using the
3831 @option{--param} option.
3833 In each case, the @var{value} is an integer. The allowable choices for
3834 @var{name} are given in the following table:
3837 @item max-delay-slot-insn-search
3838 The maximum number of instructions to consider when looking for an
3839 instruction to fill a delay slot. If more than this arbitrary number of
3840 instructions is searched, the time savings from filling the delay slot
3841 will be minimal so stop searching. Increasing values mean more
3842 aggressive optimization, making the compile time increase with probably
3843 small improvement in executable run time.
3845 @item max-delay-slot-live-search
3846 When trying to fill delay slots, the maximum number of instructions to
3847 consider when searching for a block with valid live register
3848 information. Increasing this arbitrarily chosen value means more
3849 aggressive optimization, increasing the compile time. This parameter
3850 should be removed when the delay slot code is rewritten to maintain the
3853 @item max-gcse-memory
3854 The approximate maximum amount of memory that will be allocated in
3855 order to perform the global common subexpression elimination
3856 optimization. If more memory than specified is required, the
3857 optimization will not be done.
3859 @item max-gcse-passes
3860 The maximum number of passes of GCSE to run.
3862 @item max-pending-list-length
3863 The maximum number of pending dependencies scheduling will allow
3864 before flushing the current state and starting over. Large functions
3865 with few branches or calls can create excessively large lists which
3866 needlessly consume memory and resources.
3868 @item max-inline-insns
3869 If an function contains more than this many instructions, it
3870 will not be inlined. This option is precisely equivalent to
3871 @option{-finline-limit}.
3876 @node Preprocessor Options
3877 @section Options Controlling the Preprocessor
3878 @cindex preprocessor options
3879 @cindex options, preprocessor
3881 These options control the C preprocessor, which is run on each C source
3882 file before actual compilation.
3884 If you use the @option{-E} option, nothing is done except preprocessing.
3885 Some of these options make sense only together with @option{-E} because
3886 they cause the preprocessor output to be unsuitable for actual
3890 @item -include @var{file}
3892 Process @var{file} as input before processing the regular input file.
3893 In effect, the contents of @var{file} are compiled first. Any @option{-D}
3894 and @option{-U} options on the command line are always processed before
3895 @option{-include @var{file}}, regardless of the order in which they are
3896 written. All the @option{-include} and @option{-imacros} options are
3897 processed in the order in which they are written.
3899 @item -imacros @var{file}
3901 Process @var{file} as input, discarding the resulting output, before
3902 processing the regular input file. Because the output generated from
3903 @var{file} is discarded, the only effect of @option{-imacros @var{file}}
3904 is to make the macros defined in @var{file} available for use in the
3905 main input. All the @option{-include} and @option{-imacros} options are
3906 processed in the order in which they are written.
3908 @item -idirafter @var{dir}
3910 @cindex second include path
3911 Add the directory @var{dir} to the second include path. The directories
3912 on the second include path are searched when a header file is not found
3913 in any of the directories in the main include path (the one that
3914 @option{-I} adds to).
3916 @item -iprefix @var{prefix}
3918 Specify @var{prefix} as the prefix for subsequent @option{-iwithprefix}
3921 @item -iwithprefix @var{dir}
3922 @opindex iwithprefix
3923 Add a directory to the second include path. The directory's name is
3924 made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was
3925 specified previously with @option{-iprefix}. If you have not specified a
3926 prefix yet, the directory containing the installed passes of the
3927 compiler is used as the default.
3929 @item -iwithprefixbefore @var{dir}
3930 @opindex iwithprefixbefore
3931 Add a directory to the main include path. The directory's name is made
3932 by concatenating @var{prefix} and @var{dir}, as in the case of
3933 @option{-iwithprefix}.
3935 @item -isystem @var{dir}
3937 Add a directory to the beginning of the second include path, marking it
3938 as a system directory, so that it gets the same special treatment as
3939 is applied to the standard system directories.
3943 Do not search the standard system directories for header files. Only
3944 the directories you have specified with @option{-I} options (and the
3945 current directory, if appropriate) are searched. @xref{Directory
3946 Options}, for information on @option{-I}.
3948 By using both @option{-nostdinc} and @option{-I-}, you can limit the include-file
3949 search path to only those directories you specify explicitly.
3953 When searching for a header file in a directory, remap file names if a
3954 file named @file{header.gcc} exists in that directory. This can be used
3955 to work around limitations of file systems with file name restrictions.
3956 The @file{header.gcc} file should contain a series of lines with two
3957 tokens on each line: the first token is the name to map, and the second
3958 token is the actual name to use.
3962 Do not predefine any nonstandard macros. (Including architecture flags).
3966 Run only the C preprocessor. Preprocess all the C source files
3967 specified and output the results to standard output or to the
3968 specified output file.
3972 Tell the preprocessor not to discard comments. Used with the
3977 Tell the preprocessor not to generate @samp{#line} directives.
3978 Used with the @option{-E} option.
3981 @cindex dependencies, make
3984 Instead of outputting the result of preprocessing, output a rule
3985 suitable for @code{make} describing the dependencies of the main source
3986 file. The preprocessor outputs one @code{make} rule containing the
3987 object file name for that source file, a colon, and the names of all the
3988 included files. Unless overridden explicitly, the object file name
3989 consists of the basename of the source file with any suffix replaced with
3990 object file suffix. If there are many included files then the
3991 rule is split into several lines using @samp{\}-newline.
3993 @option{-M} implies @option{-E}.
3997 Like @option{-M}, but mention only the files included with @samp{#include
3998 "@var{file}"}. System header files included with @samp{#include
3999 <@var{file}>} are omitted.
4003 Like @option{-M} but the dependency information is written to a file
4004 rather than stdout. @code{gcc} will use the same file name and
4005 directory as the object file, but with the suffix @file{.d} instead.
4007 This is in addition to compiling the main file as specified---@option{-MD}
4008 does not inhibit ordinary compilation the way @option{-M} does,
4009 unless you also specify @option{-MG}.
4011 With Mach, you can use the utility @code{md} to merge multiple
4012 dependency files into a single dependency file suitable for using with
4013 the @samp{make} command.
4017 Like @option{-MD} except mention only user header files, not system
4020 @item -MF @var{file}
4022 When used with @option{-M} or @option{-MM}, specifies a file to write the
4023 dependencies to. This allows the preprocessor to write the preprocessed
4024 file to stdout normally. If no @option{-MF} switch is given, CPP sends
4025 the rules to stdout and suppresses normal preprocessed output.
4027 Another way to specify output of a @code{make} rule is by setting
4028 the environment variable @env{DEPENDENCIES_OUTPUT} (@pxref{Environment
4033 When used with @option{-M} or @option{-MM}, @option{-MG} says to treat missing
4034 header files as generated files and assume they live in the same
4035 directory as the source file. It suppresses preprocessed output, as a
4036 missing header file is ordinarily an error.
4038 This feature is used in automatic updating of makefiles.
4042 This option instructs CPP to add a phony target for each dependency
4043 other than the main file, causing each to depend on nothing. These
4044 dummy rules work around errors @code{make} gives if you remove header
4045 files without updating the @code{Makefile} to match.
4047 This is typical output:-
4050 /tmp/test.o: /tmp/test.c /tmp/test.h
4055 @item -MQ @var{target}
4056 @item -MT @var{target}
4059 By default CPP uses the main file name, including any path, and appends
4060 the object suffix, normally ``.o'', to it to obtain the name of the
4061 target for dependency generation. With @option{-MT} you can specify a
4062 target yourself, overriding the default one.
4064 If you want multiple targets, you can specify them as a single argument
4065 to @option{-MT}, or use multiple @option{-MT} options.
4067 The targets you specify are output in the order they appear on the
4068 command line. @option{-MQ} is identical to @option{-MT}, except that the
4069 target name is quoted for Make, but with @option{-MT} it isn't. For
4070 example, @option{-MT '$(objpfx)foo.o'} gives
4073 $(objpfx)foo.o: /tmp/foo.c
4076 but @option{-MQ '$(objpfx)foo.o'} gives
4079 $$(objpfx)foo.o: /tmp/foo.c
4082 The default target is automatically quoted, as if it were given with
4087 Print the name of each header file used, in addition to other normal
4090 @item -A@var{question}(@var{answer})
4092 Assert the answer @var{answer} for @var{question}, in case it is tested
4093 with a preprocessing conditional such as @samp{#if
4094 #@var{question}(@var{answer})}. @option{-A-} disables the standard
4095 assertions that normally describe the target machine.
4099 Define macro @var{macro} with the string @samp{1} as its definition.
4101 @item -D@var{macro}=@var{defn}
4102 Define macro @var{macro} as @var{defn}. All instances of @option{-D} on
4103 the command line are processed before any @option{-U} options.
4105 Any @option{-D} and @option{-U} options on the command line are processed in
4106 order, and always before @option{-imacros @var{file}}, regardless of the
4107 order in which they are written.
4111 Undefine macro @var{macro}. @option{-U} options are evaluated after all
4112 @option{-D} options, but before any @option{-include} and @option{-imacros}
4115 Any @option{-D} and @option{-U} options on the command line are processed in
4116 order, and always before @option{-imacros @var{file}}, regardless of the
4117 order in which they are written.
4121 Tell the preprocessor to output only a list of the macro definitions
4122 that are in effect at the end of preprocessing. Used with the @option{-E}
4127 Tell the preprocessing to pass all macro definitions into the output, in
4128 their proper sequence in the rest of the output.
4132 Like @option{-dD} except that the macro arguments and contents are omitted.
4133 Only @samp{#define @var{name}} is included in the output.
4137 Output @samp{#include} directives in addition to the result of
4140 @item -fpreprocessed
4141 @opindex fpreprocessed
4142 Indicate to the preprocessor that the input file has already been
4143 preprocessed. This suppresses things like macro expansion, trigraph
4144 conversion, escaped newline splicing, and processing of most directives.
4145 The preprocessor still recognizes and removes comments, so that you can
4146 pass a file preprocessed with @option{-C} to the compiler without
4147 problems. In this mode the integrated preprocessor is little more than
4148 a tokenizer for the front ends.
4150 @option{-fpreprocessed} is implicit if the input file has one of the
4151 extensions @samp{i}, @samp{ii} or @samp{mi}. These are the extensions
4152 that GCC uses for preprocessed files created by @option{-save-temps}.
4156 Process ISO standard trigraph sequences. These are three-character
4157 sequences, all starting with @samp{??}, that are defined by ISO C to
4158 stand for single characters. For example, @samp{??/} stands for
4159 @samp{\}, so @samp{'??/n'} is a character constant for a newline. By
4160 default, GCC ignores trigraphs, but in standard-conforming modes it
4161 converts them. See the @option{-std} and @option{-ansi} options.
4163 The nine trigraph sequences are
4166 @expansion{} @samp{[}
4169 @expansion{} @samp{]}
4172 @expansion{} @samp{@{}
4175 @expansion{} @samp{@}}
4178 @expansion{} @samp{#}
4181 @expansion{} @samp{\}
4184 @expansion{} @samp{^}
4187 @expansion{} @samp{|}
4190 @expansion{} @samp{~}
4194 Trigraph support is not popular, so many compilers do not implement it
4195 properly. Portable code should not rely on trigraphs being either
4196 converted or ignored.
4198 @item -Wp,@var{option}
4200 Pass @var{option} as an option to the preprocessor. If @var{option}
4201 contains commas, it is split into multiple options at the commas.
4204 @node Assembler Options
4205 @section Passing Options to the Assembler
4207 @c prevent bad page break with this line
4208 You can pass options to the assembler.
4211 @item -Wa,@var{option}
4213 Pass @var{option} as an option to the assembler. If @var{option}
4214 contains commas, it is split into multiple options at the commas.
4218 @section Options for Linking
4219 @cindex link options
4220 @cindex options, linking
4222 These options come into play when the compiler links object files into
4223 an executable output file. They are meaningless if the compiler is
4224 not doing a link step.
4228 @item @var{object-file-name}
4229 A file name that does not end in a special recognized suffix is
4230 considered to name an object file or library. (Object files are
4231 distinguished from libraries by the linker according to the file
4232 contents.) If linking is done, these object files are used as input
4241 If any of these options is used, then the linker is not run, and
4242 object file names should not be used as arguments. @xref{Overall
4246 @item -l@var{library}
4247 @itemx -l @var{library}
4249 Search the library named @var{library} when linking. (The second
4250 alternative with the library as a separate argument is only for
4251 POSIX compliance and is not recommended.)
4253 It makes a difference where in the command you write this option; the
4254 linker searches and processes libraries and object files in the order they
4255 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4256 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4257 to functions in @samp{z}, those functions may not be loaded.
4259 The linker searches a standard list of directories for the library,
4260 which is actually a file named @file{lib@var{library}.a}. The linker
4261 then uses this file as if it had been specified precisely by name.
4263 The directories searched include several standard system directories
4264 plus any that you specify with @option{-L}.
4266 Normally the files found this way are library files---archive files
4267 whose members are object files. The linker handles an archive file by
4268 scanning through it for members which define symbols that have so far
4269 been referenced but not defined. But if the file that is found is an
4270 ordinary object file, it is linked in the usual fashion. The only
4271 difference between using an @option{-l} option and specifying a file name
4272 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4273 and searches several directories.
4277 You need this special case of the @option{-l} option in order to
4278 link an Objective-C program.
4281 @opindex nostartfiles
4282 Do not use the standard system startup files when linking.
4283 The standard system libraries are used normally, unless @option{-nostdlib}
4284 or @option{-nodefaultlibs} is used.
4286 @item -nodefaultlibs
4287 @opindex nodefaultlibs
4288 Do not use the standard system libraries when linking.
4289 Only the libraries you specify will be passed to the linker.
4290 The standard startup files are used normally, unless @option{-nostartfiles}
4291 is used. The compiler may generate calls to memcmp, memset, and memcpy
4292 for System V (and ISO C) environments or to bcopy and bzero for
4293 BSD environments. These entries are usually resolved by entries in
4294 libc. These entry points should be supplied through some other
4295 mechanism when this option is specified.
4299 Do not use the standard system startup files or libraries when linking.
4300 No startup files and only the libraries you specify will be passed to
4301 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4302 for System V (and ISO C) environments or to bcopy and bzero for
4303 BSD environments. These entries are usually resolved by entries in
4304 libc. These entry points should be supplied through some other
4305 mechanism when this option is specified.
4307 @cindex @option{-lgcc}, use with @option{-nostdlib}
4308 @cindex @option{-nostdlib} and unresolved references
4309 @cindex unresolved references and @option{-nostdlib}
4310 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4311 @cindex @option{-nodefaultlibs} and unresolved references
4312 @cindex unresolved references and @option{-nodefaultlibs}
4313 One of the standard libraries bypassed by @option{-nostdlib} and
4314 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4315 that GCC uses to overcome shortcomings of particular machines, or special
4316 needs for some languages.
4317 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4318 Collection (GCC) Internals},
4319 for more discussion of @file{libgcc.a}.)
4320 In most cases, you need @file{libgcc.a} even when you want to avoid
4321 other standard libraries. In other words, when you specify @option{-nostdlib}
4322 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4323 This ensures that you have no unresolved references to internal GCC
4324 library subroutines. (For example, @samp{__main}, used to ensure C++
4325 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4326 GNU Compiler Collection (GCC) Internals}.)
4330 Remove all symbol table and relocation information from the executable.
4334 On systems that support dynamic linking, this prevents linking with the shared
4335 libraries. On other systems, this option has no effect.
4339 Produce a shared object which can then be linked with other objects to
4340 form an executable. Not all systems support this option. For predictable
4341 results, you must also specify the same set of options that were used to
4342 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4343 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4344 needs to build supplementary stub code for constructors to work. On
4345 multi-libbed systems, @samp{gcc -shared} must select the correct support
4346 libraries to link against. Failing to supply the correct flags may lead
4347 to subtle defects. Supplying them in cases where they are not necessary
4350 @item -shared-libgcc
4351 @itemx -static-libgcc
4352 @opindex shared-libgcc
4353 @opindex static-libgcc
4354 On systems that provide @file{libgcc} as a shared library, these options
4355 force the use of either the shared or static version respectively.
4356 If no shared version of @file{libgcc} was built when the compiler was
4357 configured, these options have no effect.
4359 There are several situations in which an application should use the
4360 shared @file{libgcc} instead of the static version. The most common
4361 of these is when the application wishes to throw and catch exceptions
4362 across different shared libraries. In that case, each of the libraries
4363 as well as the application itself should use the shared @file{libgcc}.
4365 Therefore, whenever you specify the @option{-shared} option, the GCC
4366 driver automatically adds @option{-shared-libgcc}, unless you explicitly
4367 specify @option{-static-libgcc}. The G++ driver automatically adds
4368 @option{-shared-libgcc} when you build a main executable as well because
4369 for C++ programs that is typically the right thing to do.
4370 (Exception-handling will not work reliably otherwise.)
4372 However, when linking a main executable written in C, you must
4373 explicitly say @option{-shared-libgcc} if you want to use the shared
4378 Bind references to global symbols when building a shared object. Warn
4379 about any unresolved references (unless overridden by the link editor
4380 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4383 @item -Xlinker @var{option}
4385 Pass @var{option} as an option to the linker. You can use this to
4386 supply system-specific linker options which GCC does not know how to
4389 If you want to pass an option that takes an argument, you must use
4390 @option{-Xlinker} twice, once for the option and once for the argument.
4391 For example, to pass @option{-assert definitions}, you must write
4392 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4393 @option{-Xlinker "-assert definitions"}, because this passes the entire
4394 string as a single argument, which is not what the linker expects.
4396 @item -Wl,@var{option}
4398 Pass @var{option} as an option to the linker. If @var{option} contains
4399 commas, it is split into multiple options at the commas.
4401 @item -u @var{symbol}
4403 Pretend the symbol @var{symbol} is undefined, to force linking of
4404 library modules to define it. You can use @option{-u} multiple times with
4405 different symbols to force loading of additional library modules.
4408 @node Directory Options
4409 @section Options for Directory Search
4410 @cindex directory options
4411 @cindex options, directory search
4414 These options specify directories to search for header files, for
4415 libraries and for parts of the compiler:
4420 Add the directory @var{dir} to the head of the list of directories to be
4421 searched for header files. This can be used to override a system header
4422 file, substituting your own version, since these directories are
4423 searched before the system header file directories. However, you should
4424 not use this option to add directories that contain vendor-supplied
4425 system header files (use @option{-isystem} for that). If you use more than
4426 one @option{-I} option, the directories are scanned in left-to-right
4427 order; the standard system directories come after.
4429 If a standard system include directory, or a directory specified with
4430 @option{-isystem}, is also specified with @option{-I}, it will be
4431 searched only in the position requested by @option{-I}. Also, it will
4432 not be considered a system include directory. If that directory really
4433 does contain system headers, there is a good chance that they will
4434 break. For instance, if GCC's installation procedure edited the headers
4435 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4436 original, buggy headers to be found instead of the corrected ones. GCC
4437 will issue a warning when a system include directory is hidden in this
4442 Any directories you specify with @option{-I} options before the @option{-I-}
4443 option are searched only for the case of @samp{#include "@var{file}"};
4444 they are not searched for @samp{#include <@var{file}>}.
4446 If additional directories are specified with @option{-I} options after
4447 the @option{-I-}, these directories are searched for all @samp{#include}
4448 directives. (Ordinarily @emph{all} @option{-I} directories are used
4451 In addition, the @option{-I-} option inhibits the use of the current
4452 directory (where the current input file came from) as the first search
4453 directory for @samp{#include "@var{file}"}. There is no way to
4454 override this effect of @option{-I-}. With @option{-I.} you can specify
4455 searching the directory which was current when the compiler was
4456 invoked. That is not exactly the same as what the preprocessor does
4457 by default, but it is often satisfactory.
4459 @option{-I-} does not inhibit the use of the standard system directories
4460 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4465 Add directory @var{dir} to the list of directories to be searched
4468 @item -B@var{prefix}
4470 This option specifies where to find the executables, libraries,
4471 include files, and data files of the compiler itself.
4473 The compiler driver program runs one or more of the subprograms
4474 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4475 @var{prefix} as a prefix for each program it tries to run, both with and
4476 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4478 For each subprogram to be run, the compiler driver first tries the
4479 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4480 was not specified, the driver tries two standard prefixes, which are
4481 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4482 those results in a file name that is found, the unmodified program
4483 name is searched for using the directories specified in your
4484 @env{PATH} environment variable.
4486 The compiler will check to see if the path provided by the @option{-B}
4487 refers to a directory, and if necessary it will add a directory
4488 separator character at the end of the path.
4490 @option{-B} prefixes that effectively specify directory names also apply
4491 to libraries in the linker, because the compiler translates these
4492 options into @option{-L} options for the linker. They also apply to
4493 includes files in the preprocessor, because the compiler translates these
4494 options into @option{-isystem} options for the preprocessor. In this case,
4495 the compiler appends @samp{include} to the prefix.
4497 The run-time support file @file{libgcc.a} can also be searched for using
4498 the @option{-B} prefix, if needed. If it is not found there, the two
4499 standard prefixes above are tried, and that is all. The file is left
4500 out of the link if it is not found by those means.
4502 Another way to specify a prefix much like the @option{-B} prefix is to use
4503 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4506 As a special kludge, if the path provided by @option{-B} is
4507 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4508 9, then it will be replaced by @file{[dir/]include}. This is to help
4509 with boot-strapping the compiler.
4511 @item -specs=@var{file}
4513 Process @var{file} after the compiler reads in the standard @file{specs}
4514 file, in order to override the defaults that the @file{gcc} driver
4515 program uses when determining what switches to pass to @file{cc1},
4516 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4517 @option{-specs=@var{file}} can be specified on the command line, and they
4518 are processed in order, from left to right.
4524 @section Specifying subprocesses and the switches to pass to them
4526 @command{gcc} is a driver program. It performs its job by invoking a
4527 sequence of other programs to do the work of compiling, assembling and
4528 linking. GCC interprets its command-line parameters and uses these to
4529 deduce which programs it should invoke, and which command-line options
4530 it ought to place on their command lines. This behavior is controlled
4531 by @dfn{spec strings}. In most cases there is one spec string for each
4532 program that GCC can invoke, but a few programs have multiple spec
4533 strings to control their behavior. The spec strings built into GCC can
4534 be overridden by using the @option{-specs=} command-line switch to specify
4537 @dfn{Spec files} are plaintext files that are used to construct spec
4538 strings. They consist of a sequence of directives separated by blank
4539 lines. The type of directive is determined by the first non-whitespace
4540 character on the line and it can be one of the following:
4543 @item %@var{command}
4544 Issues a @var{command} to the spec file processor. The commands that can
4548 @item %include <@var{file}>
4550 Search for @var{file} and insert its text at the current point in the
4553 @item %include_noerr <@var{file}>
4554 @cindex %include_noerr
4555 Just like @samp{%include}, but do not generate an error message if the include
4556 file cannot be found.
4558 @item %rename @var{old_name} @var{new_name}
4560 Rename the spec string @var{old_name} to @var{new_name}.
4564 @item *[@var{spec_name}]:
4565 This tells the compiler to create, override or delete the named spec
4566 string. All lines after this directive up to the next directive or
4567 blank line are considered to be the text for the spec string. If this
4568 results in an empty string then the spec will be deleted. (Or, if the
4569 spec did not exist, then nothing will happened.) Otherwise, if the spec
4570 does not currently exist a new spec will be created. If the spec does
4571 exist then its contents will be overridden by the text of this
4572 directive, unless the first character of that text is the @samp{+}
4573 character, in which case the text will be appended to the spec.
4575 @item [@var{suffix}]:
4576 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4577 and up to the next directive or blank line are considered to make up the
4578 spec string for the indicated suffix. When the compiler encounters an
4579 input file with the named suffix, it will processes the spec string in
4580 order to work out how to compile that file. For example:
4587 This says that any input file whose name ends in @samp{.ZZ} should be
4588 passed to the program @samp{z-compile}, which should be invoked with the
4589 command-line switch @option{-input} and with the result of performing the
4590 @samp{%i} substitution. (See below.)
4592 As an alternative to providing a spec string, the text that follows a
4593 suffix directive can be one of the following:
4596 @item @@@var{language}
4597 This says that the suffix is an alias for a known @var{language}. This is
4598 similar to using the @option{-x} command-line switch to GCC to specify a
4599 language explicitly. For example:
4606 Says that .ZZ files are, in fact, C++ source files.
4609 This causes an error messages saying:
4612 @var{name} compiler not installed on this system.
4616 GCC already has an extensive list of suffixes built into it.
4617 This directive will add an entry to the end of the list of suffixes, but
4618 since the list is searched from the end backwards, it is effectively
4619 possible to override earlier entries using this technique.
4623 GCC has the following spec strings built into it. Spec files can
4624 override these strings or create their own. Note that individual
4625 targets can also add their own spec strings to this list.
4628 asm Options to pass to the assembler
4629 asm_final Options to pass to the assembler post-processor
4630 cpp Options to pass to the C preprocessor
4631 cc1 Options to pass to the C compiler
4632 cc1plus Options to pass to the C++ compiler
4633 endfile Object files to include at the end of the link
4634 link Options to pass to the linker
4635 lib Libraries to include on the command line to the linker
4636 libgcc Decides which GCC support library to pass to the linker
4637 linker Sets the name of the linker
4638 predefines Defines to be passed to the C preprocessor
4639 signed_char Defines to pass to CPP to say whether @code{char} is signed
4641 startfile Object files to include at the start of the link
4644 Here is a small example of a spec file:
4650 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4653 This example renames the spec called @samp{lib} to @samp{old_lib} and
4654 then overrides the previous definition of @samp{lib} with a new one.
4655 The new definition adds in some extra command-line options before
4656 including the text of the old definition.
4658 @dfn{Spec strings} are a list of command-line options to be passed to their
4659 corresponding program. In addition, the spec strings can contain
4660 @samp{%}-prefixed sequences to substitute variable text or to
4661 conditionally insert text into the command line. Using these constructs
4662 it is possible to generate quite complex command lines.
4664 Here is a table of all defined @samp{%}-sequences for spec
4665 strings. Note that spaces are not generated automatically around the
4666 results of expanding these sequences. Therefore you can concatenate them
4667 together or combine them with constant text in a single argument.
4671 Substitute one @samp{%} into the program name or argument.
4674 Substitute the name of the input file being processed.
4677 Substitute the basename of the input file being processed.
4678 This is the substring up to (and not including) the last period
4679 and not including the directory.
4682 This is the same as @samp{%b}, but include the file suffix (text after
4686 Marks the argument containing or following the @samp{%d} as a
4687 temporary file name, so that that file will be deleted if GCC exits
4688 successfully. Unlike @samp{%g}, this contributes no text to the
4691 @item %g@var{suffix}
4692 Substitute a file name that has suffix @var{suffix} and is chosen
4693 once per compilation, and mark the argument in the same way as
4694 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4695 name is now chosen in a way that is hard to predict even when previously
4696 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4697 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4698 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4699 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4700 was simply substituted with a file name chosen once per compilation,
4701 without regard to any appended suffix (which was therefore treated
4702 just like ordinary text), making such attacks more likely to succeed.
4704 @item %u@var{suffix}
4705 Like @samp{%g}, but generates a new temporary file name even if
4706 @samp{%u@var{suffix}} was already seen.
4708 @item %U@var{suffix}
4709 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4710 new one if there is no such last file name. In the absence of any
4711 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4712 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4713 would involve the generation of two distinct file names, one
4714 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4715 simply substituted with a file name chosen for the previous @samp{%u},
4716 without regard to any appended suffix.
4718 @item %j@var{SUFFIX}
4719 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4720 writable, and if save-temps is off; otherwise, substitute the name
4721 of a temporary file, just like @samp{%u}. This temporary file is not
4722 meant for communication between processes, but rather as a junk
4725 @item %.@var{SUFFIX}
4726 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4727 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4728 terminated by the next space or %.
4731 Marks the argument containing or following the @samp{%w} as the
4732 designated output file of this compilation. This puts the argument
4733 into the sequence of arguments that @samp{%o} will substitute later.
4736 Substitutes the names of all the output files, with spaces
4737 automatically placed around them. You should write spaces
4738 around the @samp{%o} as well or the results are undefined.
4739 @samp{%o} is for use in the specs for running the linker.
4740 Input files whose names have no recognized suffix are not compiled
4741 at all, but they are included among the output files, so they will
4745 Substitutes the suffix for object files. Note that this is
4746 handled specially when it immediately follows @samp{%g, %u, or %U},
4747 because of the need for those to form complete file names. The
4748 handling is such that @samp{%O} is treated exactly as if it had already
4749 been substituted, except that @samp{%g, %u, and %U} do not currently
4750 support additional @var{suffix} characters following @samp{%O} as they would
4751 following, for example, @samp{.o}.
4754 Substitutes the standard macro predefinitions for the
4755 current target machine. Use this when running @code{cpp}.
4758 Like @samp{%p}, but puts @samp{__} before and after the name of each
4759 predefined macro, except for macros that start with @samp{__} or with
4760 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4764 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4767 Current argument is the name of a library or startup file of some sort.
4768 Search for that file in a standard list of directories and substitute
4769 the full name found.
4772 Print @var{str} as an error message. @var{str} is terminated by a newline.
4773 Use this when inconsistent options are detected.
4776 Output @samp{-} if the input for the current command is coming from a pipe.
4779 Substitute the contents of spec string @var{name} at this point.
4782 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4784 @item %x@{@var{option}@}
4785 Accumulate an option for @samp{%X}.
4788 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4792 Output the accumulated assembler options specified by @option{-Wa}.
4795 Output the accumulated preprocessor options specified by @option{-Wp}.
4798 Substitute the major version number of GCC@.
4799 (For version 2.9.5, this is 2.)
4802 Substitute the minor version number of GCC@.
4803 (For version 2.9.5, this is 9.)
4806 Substitute the patch level number of GCC@.
4807 (For version 2.9.5, this is 5.)
4810 Process the @code{asm} spec. This is used to compute the
4811 switches to be passed to the assembler.
4814 Process the @code{asm_final} spec. This is a spec string for
4815 passing switches to an assembler post-processor, if such a program is
4819 Process the @code{link} spec. This is the spec for computing the
4820 command line passed to the linker. Typically it will make use of the
4821 @samp{%L %G %S %D and %E} sequences.
4824 Dump out a @option{-L} option for each directory that GCC believes might
4825 contain startup files. If the target supports multilibs then the
4826 current multilib directory will be prepended to each of these paths.
4829 Output the multilib directory with directory separators replaced with
4830 @samp{_}. If multilib directories are not set, or the multilib directory is
4831 @file{.} then this option emits nothing.
4834 Process the @code{lib} spec. This is a spec string for deciding which
4835 libraries should be included on the command line to the linker.
4838 Process the @code{libgcc} spec. This is a spec string for deciding
4839 which GCC support library should be included on the command line to the linker.
4842 Process the @code{startfile} spec. This is a spec for deciding which
4843 object files should be the first ones passed to the linker. Typically
4844 this might be a file named @file{crt0.o}.
4847 Process the @code{endfile} spec. This is a spec string that specifies
4848 the last object files that will be passed to the linker.
4851 Process the @code{cpp} spec. This is used to construct the arguments
4852 to be passed to the C preprocessor.
4855 Process the @code{signed_char} spec. This is intended to be used
4856 to tell cpp whether a char is signed. It typically has the definition:
4858 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4862 Process the @code{cc1} spec. This is used to construct the options to be
4863 passed to the actual C compiler (@samp{cc1}).
4866 Process the @code{cc1plus} spec. This is used to construct the options to be
4867 passed to the actual C++ compiler (@samp{cc1plus}).
4870 Substitute the variable part of a matched option. See below.
4871 Note that each comma in the substituted string is replaced by
4875 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4876 If that switch was not specified, this substitutes nothing. Note that
4877 the leading dash is omitted when specifying this option, and it is
4878 automatically inserted if the substitution is performed. Thus the spec
4879 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4880 and would output the command line option @option{-foo}.
4882 @item %W@{@code{S}@}
4883 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4886 @item %@{@code{S}*@}
4887 Substitutes all the switches specified to GCC whose names start
4888 with @code{-S}, but which also take an argument. This is used for
4889 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4890 GCC considers @option{-o foo} as being
4891 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4892 text, including the space. Thus two arguments would be generated.
4894 @item %@{^@code{S}*@}
4895 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4896 argument. Thus %@{^o*@} would only generate one argument, not two.
4898 @item %@{@code{S}*&@code{T}*@}
4899 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4900 (the order of @code{S} and @code{T} in the spec is not significant).
4901 There can be any number of ampersand-separated variables; for each the
4902 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4904 @item %@{<@code{S}@}
4905 Remove all occurrences of @code{-S} from the command line. Note---this
4906 command is position dependent. @samp{%} commands in the spec string
4907 before this option will see @code{-S}, @samp{%} commands in the spec
4908 string after this option will not.
4910 @item %@{@code{S}*:@code{X}@}
4911 Substitutes @code{X} if one or more switches whose names start with
4912 @code{-S} are specified to GCC@. Note that the tail part of the
4913 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4914 for each occurrence of @samp{%*} within @code{X}.
4916 @item %@{@code{S}:@code{X}@}
4917 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4919 @item %@{!@code{S}:@code{X}@}
4920 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4922 @item %@{|@code{S}:@code{X}@}
4923 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4925 @item %@{|!@code{S}:@code{X}@}
4926 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4928 @item %@{.@code{S}:@code{X}@}
4929 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4931 @item %@{!.@code{S}:@code{X}@}
4932 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4934 @item %@{@code{S}|@code{P}:@code{X}@}
4935 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4936 combined with @samp{!} and @samp{.} sequences as well, although they
4937 have a stronger binding than the @samp{|}. For example a spec string
4941 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4944 will output the following command-line options from the following input
4945 command-line options:
4950 -d fred.c -foo -baz -boggle
4951 -d jim.d -bar -baz -boggle
4956 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4957 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4958 or spaces, or even newlines. They are processed as usual, as described
4961 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4962 switches are handled specifically in these
4963 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4964 @option{-W} switch is found later in the command line, the earlier switch
4965 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4966 letter, which passes all matching options.
4968 The character @samp{|} at the beginning of the predicate text is used to indicate
4969 that a command should be piped to the following command, but only if @option{-pipe}
4972 It is built into GCC which switches take arguments and which do not.
4973 (You might think it would be useful to generalize this to allow each
4974 compiler's spec to say which switches take arguments. But this cannot
4975 be done in a consistent fashion. GCC cannot even decide which input
4976 files have been specified without knowing which switches take arguments,
4977 and it must know which input files to compile in order to tell which
4980 GCC also knows implicitly that arguments starting in @option{-l} are to be
4981 treated as compiler output files, and passed to the linker in their
4982 proper position among the other output files.
4984 @c man begin OPTIONS
4986 @node Target Options
4987 @section Specifying Target Machine and Compiler Version
4988 @cindex target options
4989 @cindex cross compiling
4990 @cindex specifying machine version
4991 @cindex specifying compiler version and target machine
4992 @cindex compiler version, specifying
4993 @cindex target machine, specifying
4995 By default, GCC compiles code for the same type of machine that you
4996 are using. However, it can also be installed as a cross-compiler, to
4997 compile for some other type of machine. In fact, several different
4998 configurations of GCC, for different target machines, can be
4999 installed side by side. Then you specify which one to use with the
5002 In addition, older and newer versions of GCC can be installed side
5003 by side. One of them (probably the newest) will be the default, but
5004 you may sometimes wish to use another.
5007 @item -b @var{machine}
5009 The argument @var{machine} specifies the target machine for compilation.
5010 This is useful when you have installed GCC as a cross-compiler.
5012 The value to use for @var{machine} is the same as was specified as the
5013 machine type when configuring GCC as a cross-compiler. For
5014 example, if a cross-compiler was configured with @samp{configure
5015 i386v}, meaning to compile for an 80386 running System V, then you
5016 would specify @option{-b i386v} to run that cross compiler.
5018 When you do not specify @option{-b}, it normally means to compile for
5019 the same type of machine that you are using.
5021 @item -V @var{version}
5023 The argument @var{version} specifies which version of GCC to run.
5024 This is useful when multiple versions are installed. For example,
5025 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5027 The default version, when you do not specify @option{-V}, is the last
5028 version of GCC that you installed.
5031 The @option{-b} and @option{-V} options actually work by controlling part of
5032 the file name used for the executable files and libraries used for
5033 compilation. A given version of GCC, for a given target machine, is
5034 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.
5036 Thus, sites can customize the effect of @option{-b} or @option{-V} either by
5037 changing the names of these directories or adding alternate names (or
5038 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
5039 file @file{80386} is a link to the file @file{i386v}, then @option{-b
5040 80386} becomes an alias for @option{-b i386v}.
5042 In one respect, the @option{-b} or @option{-V} do not completely change
5043 to a different compiler: the top-level driver program @command{gcc}
5044 that you originally invoked continues to run and invoke the other
5045 executables (preprocessor, compiler per se, assembler and linker)
5046 that do the real work. However, since no real work is done in the
5047 driver program, it usually does not matter that the driver program
5048 in use is not the one for the specified target. It is common for the
5049 interface to the other executables to change incompatibly between
5050 compiler versions, so unless the version specified is very close to that
5051 of the driver (for example, @option{-V 3.0} with a driver program from GCC
5052 version 3.0.1), use of @option{-V} may not work; for example, using
5053 @option{-V 2.95.2} will not work with a driver program from GCC 3.0.
5055 The only way that the driver program depends on the target machine is
5056 in the parsing and handling of special machine-specific options.
5057 However, this is controlled by a file which is found, along with the
5058 other executables, in the directory for the specified version and
5059 target machine. As a result, a single installed driver program adapts
5060 to any specified target machine, and sufficiently similar compiler
5063 The driver program executable does control one significant thing,
5064 however: the default version and target machine. Therefore, you can
5065 install different instances of the driver program, compiled for
5066 different targets or versions, under different names.
5068 For example, if the driver for version 2.0 is installed as @command{ogcc}
5069 and that for version 2.1 is installed as @command{gcc}, then the command
5070 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
5071 2.0 by default. However, you can choose either version with either
5072 command with the @option{-V} option.
5074 @node Submodel Options
5075 @section Hardware Models and Configurations
5076 @cindex submodel options
5077 @cindex specifying hardware config
5078 @cindex hardware models and configurations, specifying
5079 @cindex machine dependent options
5081 Earlier we discussed the standard option @option{-b} which chooses among
5082 different installed compilers for completely different target
5083 machines, such as VAX vs.@: 68000 vs.@: 80386.
5085 In addition, each of these target machine types can have its own
5086 special options, starting with @samp{-m}, to choose among various
5087 hardware models or configurations---for example, 68010 vs 68020,
5088 floating coprocessor or none. A single installed version of the
5089 compiler can compile for any model or configuration, according to the
5092 Some configurations of the compiler also support additional special
5093 options, usually for compatibility with other compilers on the same
5096 These options are defined by the macro @code{TARGET_SWITCHES} in the
5097 machine description. The default for the options is also defined by
5098 that macro, which enables you to change the defaults.
5112 * RS/6000 and PowerPC Options::
5115 * i386 and x86-64 Options::
5117 * Intel 960 Options::
5118 * DEC Alpha Options::
5119 * DEC Alpha/VMS Options::
5123 * System V Options::
5124 * TMS320C3x/C4x Options::
5132 * S/390 and zSeries Options::
5136 * Xstormy16 Options::
5140 @node M680x0 Options
5141 @subsection M680x0 Options
5142 @cindex M680x0 options
5144 These are the @samp{-m} options defined for the 68000 series. The default
5145 values for these options depends on which style of 68000 was selected when
5146 the compiler was configured; the defaults for the most common choices are
5154 Generate output for a 68000. This is the default
5155 when the compiler is configured for 68000-based systems.
5157 Use this option for microcontrollers with a 68000 or EC000 core,
5158 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5164 Generate output for a 68020. This is the default
5165 when the compiler is configured for 68020-based systems.
5169 Generate output containing 68881 instructions for floating point.
5170 This is the default for most 68020 systems unless @option{--nfp} was
5171 specified when the compiler was configured.
5175 Generate output for a 68030. This is the default when the compiler is
5176 configured for 68030-based systems.
5180 Generate output for a 68040. This is the default when the compiler is
5181 configured for 68040-based systems.
5183 This option inhibits the use of 68881/68882 instructions that have to be
5184 emulated by software on the 68040. Use this option if your 68040 does not
5185 have code to emulate those instructions.
5189 Generate output for a 68060. This is the default when the compiler is
5190 configured for 68060-based systems.
5192 This option inhibits the use of 68020 and 68881/68882 instructions that
5193 have to be emulated by software on the 68060. Use this option if your 68060
5194 does not have code to emulate those instructions.
5198 Generate output for a CPU32. This is the default
5199 when the compiler is configured for CPU32-based systems.
5201 Use this option for microcontrollers with a
5202 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5203 68336, 68340, 68341, 68349 and 68360.
5207 Generate output for a 520X ``coldfire'' family cpu. This is the default
5208 when the compiler is configured for 520X-based systems.
5210 Use this option for microcontroller with a 5200 core, including
5211 the MCF5202, MCF5203, MCF5204 and MCF5202.
5216 Generate output for a 68040, without using any of the new instructions.
5217 This results in code which can run relatively efficiently on either a
5218 68020/68881 or a 68030 or a 68040. The generated code does use the
5219 68881 instructions that are emulated on the 68040.
5223 Generate output for a 68060, without using any of the new instructions.
5224 This results in code which can run relatively efficiently on either a
5225 68020/68881 or a 68030 or a 68040. The generated code does use the
5226 68881 instructions that are emulated on the 68060.
5230 Generate output containing Sun FPA instructions for floating point.
5233 @opindex msoft-float
5234 Generate output containing library calls for floating point.
5235 @strong{Warning:} the requisite libraries are not available for all m68k
5236 targets. Normally the facilities of the machine's usual C compiler are
5237 used, but this can't be done directly in cross-compilation. You must
5238 make your own arrangements to provide suitable library functions for
5239 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5240 @samp{m68k-*-coff} do provide software floating point support.
5244 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5247 @opindex mnobitfield
5248 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5249 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5253 Do use the bit-field instructions. The @option{-m68020} option implies
5254 @option{-mbitfield}. This is the default if you use a configuration
5255 designed for a 68020.
5259 Use a different function-calling convention, in which functions
5260 that take a fixed number of arguments return with the @code{rtd}
5261 instruction, which pops their arguments while returning. This
5262 saves one instruction in the caller since there is no need to pop
5263 the arguments there.
5265 This calling convention is incompatible with the one normally
5266 used on Unix, so you cannot use it if you need to call libraries
5267 compiled with the Unix compiler.
5269 Also, you must provide function prototypes for all functions that
5270 take variable numbers of arguments (including @code{printf});
5271 otherwise incorrect code will be generated for calls to those
5274 In addition, seriously incorrect code will result if you call a
5275 function with too many arguments. (Normally, extra arguments are
5276 harmlessly ignored.)
5278 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5279 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5282 @itemx -mno-align-int
5284 @opindex mno-align-int
5285 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5286 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5287 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5288 Aligning variables on 32-bit boundaries produces code that runs somewhat
5289 faster on processors with 32-bit busses at the expense of more memory.
5291 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5292 align structures containing the above types differently than
5293 most published application binary interface specifications for the m68k.
5297 Use the pc-relative addressing mode of the 68000 directly, instead of
5298 using a global offset table. At present, this option implies @option{-fpic},
5299 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5300 not presently supported with @option{-mpcrel}, though this could be supported for
5301 68020 and higher processors.
5303 @item -mno-strict-align
5304 @itemx -mstrict-align
5305 @opindex mno-strict-align
5306 @opindex mstrict-align
5307 Do not (do) assume that unaligned memory references will be handled by
5312 @node M68hc1x Options
5313 @subsection M68hc1x Options
5314 @cindex M68hc1x options
5316 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5317 microcontrollers. The default values for these options depends on
5318 which style of microcontroller was selected when the compiler was configured;
5319 the defaults for the most common choices are given below.
5326 Generate output for a 68HC11. This is the default
5327 when the compiler is configured for 68HC11-based systems.
5333 Generate output for a 68HC12. This is the default
5334 when the compiler is configured for 68HC12-based systems.
5337 @opindex mauto-incdec
5338 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5343 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5345 @item -msoft-reg-count=@var{count}
5346 @opindex msoft-reg-count
5347 Specify the number of pseudo-soft registers which are used for the
5348 code generation. The maximum number is 32. Using more pseudo-soft
5349 register may or may not result in better code depending on the program.
5350 The default is 4 for 68HC11 and 2 for 68HC12.
5355 @subsection VAX Options
5358 These @samp{-m} options are defined for the VAX:
5363 Do not output certain jump instructions (@code{aobleq} and so on)
5364 that the Unix assembler for the VAX cannot handle across long
5369 Do output those jump instructions, on the assumption that you
5370 will assemble with the GNU assembler.
5374 Output code for g-format floating point numbers instead of d-format.
5378 @subsection SPARC Options
5379 @cindex SPARC options
5381 These @samp{-m} switches are supported on the SPARC:
5386 @opindex mno-app-regs
5388 Specify @option{-mapp-regs} to generate output using the global registers
5389 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5392 To be fully SVR4 ABI compliant at the cost of some performance loss,
5393 specify @option{-mno-app-regs}. You should compile libraries and system
5394 software with this option.
5399 @opindex mhard-float
5400 Generate output containing floating point instructions. This is the
5406 @opindex msoft-float
5407 Generate output containing library calls for floating point.
5408 @strong{Warning:} the requisite libraries are not available for all SPARC
5409 targets. Normally the facilities of the machine's usual C compiler are
5410 used, but this cannot be done directly in cross-compilation. You must make
5411 your own arrangements to provide suitable library functions for
5412 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5413 @samp{sparclite-*-*} do provide software floating point support.
5415 @option{-msoft-float} changes the calling convention in the output file;
5416 therefore, it is only useful if you compile @emph{all} of a program with
5417 this option. In particular, you need to compile @file{libgcc.a}, the
5418 library that comes with GCC, with @option{-msoft-float} in order for
5421 @item -mhard-quad-float
5422 @opindex mhard-quad-float
5423 Generate output containing quad-word (long double) floating point
5426 @item -msoft-quad-float
5427 @opindex msoft-quad-float
5428 Generate output containing library calls for quad-word (long double)
5429 floating point instructions. The functions called are those specified
5430 in the SPARC ABI@. This is the default.
5432 As of this writing, there are no sparc implementations that have hardware
5433 support for the quad-word floating point instructions. They all invoke
5434 a trap handler for one of these instructions, and then the trap handler
5435 emulates the effect of the instruction. Because of the trap handler overhead,
5436 this is much slower than calling the ABI library routines. Thus the
5437 @option{-msoft-quad-float} option is the default.
5441 @opindex mno-epilogue
5443 With @option{-mepilogue} (the default), the compiler always emits code for
5444 function exit at the end of each function. Any function exit in
5445 the middle of the function (such as a return statement in C) will
5446 generate a jump to the exit code at the end of the function.
5448 With @option{-mno-epilogue}, the compiler tries to emit exit code inline
5449 at every function exit.
5455 With @option{-mflat}, the compiler does not generate save/restore instructions
5456 and will use a ``flat'' or single register window calling convention.
5457 This model uses %i7 as the frame pointer and is compatible with the normal
5458 register window model. Code from either may be intermixed.
5459 The local registers and the input registers (0--5) are still treated as
5460 ``call saved'' registers and will be saved on the stack as necessary.
5462 With @option{-mno-flat} (the default), the compiler emits save/restore
5463 instructions (except for leaf functions) and is the normal mode of operation.
5465 @item -mno-unaligned-doubles
5466 @itemx -munaligned-doubles
5467 @opindex mno-unaligned-doubles
5468 @opindex munaligned-doubles
5469 Assume that doubles have 8 byte alignment. This is the default.
5471 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5472 alignment only if they are contained in another type, or if they have an
5473 absolute address. Otherwise, it assumes they have 4 byte alignment.
5474 Specifying this option avoids some rare compatibility problems with code
5475 generated by other compilers. It is not the default because it results
5476 in a performance loss, especially for floating point code.
5478 @item -mno-faster-structs
5479 @itemx -mfaster-structs
5480 @opindex mno-faster-structs
5481 @opindex mfaster-structs
5482 With @option{-mfaster-structs}, the compiler assumes that structures
5483 should have 8 byte alignment. This enables the use of pairs of
5484 @code{ldd} and @code{std} instructions for copies in structure
5485 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5486 However, the use of this changed alignment directly violates the Sparc
5487 ABI@. Thus, it's intended only for use on targets where the developer
5488 acknowledges that their resulting code will not be directly in line with
5489 the rules of the ABI@.
5495 These two options select variations on the SPARC architecture.
5497 By default (unless specifically configured for the Fujitsu SPARClite),
5498 GCC generates code for the v7 variant of the SPARC architecture.
5500 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5501 code is that the compiler emits the integer multiply and integer
5502 divide instructions which exist in SPARC v8 but not in SPARC v7.
5504 @option{-msparclite} will give you SPARClite code. This adds the integer
5505 multiply, integer divide step and scan (@code{ffs}) instructions which
5506 exist in SPARClite but not in SPARC v7.
5508 These options are deprecated and will be deleted in a future GCC release.
5509 They have been replaced with @option{-mcpu=xxx}.
5514 @opindex msupersparc
5515 These two options select the processor for which the code is optimized.
5517 With @option{-mcypress} (the default), the compiler optimizes code for the
5518 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5519 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5521 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5522 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5523 of the full SPARC v8 instruction set.
5525 These options are deprecated and will be deleted in a future GCC release.
5526 They have been replaced with @option{-mcpu=xxx}.
5528 @item -mcpu=@var{cpu_type}
5530 Set the instruction set, register set, and instruction scheduling parameters
5531 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5532 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5533 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5534 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5536 Default instruction scheduling parameters are used for values that select
5537 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5538 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5540 Here is a list of each supported architecture and their supported
5545 v8: supersparc, hypersparc
5546 sparclite: f930, f934, sparclite86x
5551 @item -mtune=@var{cpu_type}
5553 Set the instruction scheduling parameters for machine type
5554 @var{cpu_type}, but do not set the instruction set or register set that the
5555 option @option{-mcpu=@var{cpu_type}} would.
5557 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5558 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5559 that select a particular cpu implementation. Those are @samp{cypress},
5560 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5561 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5565 These @samp{-m} switches are supported in addition to the above
5566 on the SPARCLET processor.
5569 @item -mlittle-endian
5570 @opindex mlittle-endian
5571 Generate code for a processor running in little-endian mode.
5575 Treat register @code{%g0} as a normal register.
5576 GCC will continue to clobber it as necessary but will not assume
5577 it always reads as 0.
5579 @item -mbroken-saverestore
5580 @opindex mbroken-saverestore
5581 Generate code that does not use non-trivial forms of the @code{save} and
5582 @code{restore} instructions. Early versions of the SPARCLET processor do
5583 not correctly handle @code{save} and @code{restore} instructions used with
5584 arguments. They correctly handle them used without arguments. A @code{save}
5585 instruction used without arguments increments the current window pointer
5586 but does not allocate a new stack frame. It is assumed that the window
5587 overflow trap handler will properly handle this case as will interrupt
5591 These @samp{-m} switches are supported in addition to the above
5592 on SPARC V9 processors in 64-bit environments.
5595 @item -mlittle-endian
5596 @opindex mlittle-endian
5597 Generate code for a processor running in little-endian mode.
5603 Generate code for a 32-bit or 64-bit environment.
5604 The 32-bit environment sets int, long and pointer to 32 bits.
5605 The 64-bit environment sets int to 32 bits and long and pointer
5608 @item -mcmodel=medlow
5609 @opindex mcmodel=medlow
5610 Generate code for the Medium/Low code model: the program must be linked
5611 in the low 32 bits of the address space. Pointers are 64 bits.
5612 Programs can be statically or dynamically linked.
5614 @item -mcmodel=medmid
5615 @opindex mcmodel=medmid
5616 Generate code for the Medium/Middle code model: the program must be linked
5617 in the low 44 bits of the address space, the text segment must be less than
5618 2G bytes, and data segment must be within 2G of the text segment.
5619 Pointers are 64 bits.
5621 @item -mcmodel=medany
5622 @opindex mcmodel=medany
5623 Generate code for the Medium/Anywhere code model: the program may be linked
5624 anywhere in the address space, the text segment must be less than
5625 2G bytes, and data segment must be within 2G of the text segment.
5626 Pointers are 64 bits.
5628 @item -mcmodel=embmedany
5629 @opindex mcmodel=embmedany
5630 Generate code for the Medium/Anywhere code model for embedded systems:
5631 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5632 (determined at link time). Register %g4 points to the base of the
5633 data segment. Pointers are still 64 bits.
5634 Programs are statically linked, PIC is not supported.
5637 @itemx -mno-stack-bias
5638 @opindex mstack-bias
5639 @opindex mno-stack-bias
5640 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5641 frame pointer if present, are offset by @minus{}2047 which must be added back
5642 when making stack frame references.
5643 Otherwise, assume no such offset is present.
5646 @node Convex Options
5647 @subsection Convex Options
5648 @cindex Convex options
5650 These @samp{-m} options are defined for Convex:
5655 Generate output for C1. The code will run on any Convex machine.
5656 The preprocessor symbol @code{__convex__c1__} is defined.
5660 Generate output for C2. Uses instructions not available on C1.
5661 Scheduling and other optimizations are chosen for max performance on C2.
5662 The preprocessor symbol @code{__convex_c2__} is defined.
5666 Generate output for C32xx. Uses instructions not available on C1.
5667 Scheduling and other optimizations are chosen for max performance on C32.
5668 The preprocessor symbol @code{__convex_c32__} is defined.
5672 Generate output for C34xx. Uses instructions not available on C1.
5673 Scheduling and other optimizations are chosen for max performance on C34.
5674 The preprocessor symbol @code{__convex_c34__} is defined.
5678 Generate output for C38xx. Uses instructions not available on C1.
5679 Scheduling and other optimizations are chosen for max performance on C38.
5680 The preprocessor symbol @code{__convex_c38__} is defined.
5684 Generate code which puts an argument count in the word preceding each
5685 argument list. This is compatible with regular CC, and a few programs
5686 may need the argument count word. GDB and other source-level debuggers
5687 do not need it; this info is in the symbol table.
5690 @opindex mnoargcount
5691 Omit the argument count word. This is the default.
5693 @item -mvolatile-cache
5694 @opindex mvolatile-cache
5695 Allow volatile references to be cached. This is the default.
5697 @item -mvolatile-nocache
5698 @opindex mvolatile-nocache
5699 Volatile references bypass the data cache, going all the way to memory.
5700 This is only needed for multi-processor code that does not use standard
5701 synchronization instructions. Making non-volatile references to volatile
5702 locations will not necessarily work.
5706 Type long is 32 bits, the same as type int. This is the default.
5710 Type long is 64 bits, the same as type long long. This option is useless,
5711 because no library support exists for it.
5714 @node AMD29K Options
5715 @subsection AMD29K Options
5716 @cindex AMD29K options
5718 These @samp{-m} options are defined for the AMD Am29000:
5723 @cindex DW bit (29k)
5724 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5725 halfword operations are directly supported by the hardware. This is the
5730 Generate code that assumes the @code{DW} bit is not set.
5734 @cindex byte writes (29k)
5735 Generate code that assumes the system supports byte and halfword write
5736 operations. This is the default.
5740 Generate code that assumes the systems does not support byte and
5741 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5745 @cindex memory model (29k)
5746 Use a small memory model that assumes that all function addresses are
5747 either within a single 256 KB segment or at an absolute address of less
5748 than 256k. This allows the @code{call} instruction to be used instead
5749 of a @code{const}, @code{consth}, @code{calli} sequence.
5753 Use the normal memory model: Generate @code{call} instructions only when
5754 calling functions in the same file and @code{calli} instructions
5755 otherwise. This works if each file occupies less than 256 KB but allows
5756 the entire executable to be larger than 256 KB@. This is the default.
5760 Always use @code{calli} instructions. Specify this option if you expect
5761 a single file to compile into more than 256 KB of code.
5765 @cindex processor selection (29k)
5766 Generate code for the Am29050.
5770 Generate code for the Am29000. This is the default.
5772 @item -mkernel-registers
5773 @opindex mkernel-registers
5774 @cindex kernel and user registers (29k)
5775 Generate references to registers @code{gr64-gr95} instead of to
5776 registers @code{gr96-gr127}. This option can be used when compiling
5777 kernel code that wants a set of global registers disjoint from that used
5780 Note that when this option is used, register names in @samp{-f} flags
5781 must use the normal, user-mode, names.
5783 @item -muser-registers
5784 @opindex muser-registers
5785 Use the normal set of global registers, @code{gr96-gr127}. This is the
5789 @itemx -mno-stack-check
5790 @opindex mstack-check
5791 @opindex mno-stack-check
5792 @cindex stack checks (29k)
5793 Insert (or do not insert) a call to @code{__msp_check} after each stack
5794 adjustment. This is often used for kernel code.
5797 @itemx -mno-storem-bug
5798 @opindex mstorem-bug
5799 @opindex mno-storem-bug
5800 @cindex storem bug (29k)
5801 @option{-mstorem-bug} handles 29k processors which cannot handle the
5802 separation of a mtsrim insn and a storem instruction (most 29000 chips
5803 to date, but not the 29050).
5805 @item -mno-reuse-arg-regs
5806 @itemx -mreuse-arg-regs
5807 @opindex mno-reuse-arg-regs
5808 @opindex mreuse-arg-regs
5809 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5810 registers for copying out arguments. This helps detect calling a function
5811 with fewer arguments than it was declared with.
5813 @item -mno-impure-text
5814 @itemx -mimpure-text
5815 @opindex mno-impure-text
5816 @opindex mimpure-text
5817 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5818 not pass @option{-assert pure-text} to the linker when linking a shared object.
5821 @opindex msoft-float
5822 Generate output containing library calls for floating point.
5823 @strong{Warning:} the requisite libraries are not part of GCC@.
5824 Normally the facilities of the machine's usual C compiler are used, but
5825 this can't be done directly in cross-compilation. You must make your
5826 own arrangements to provide suitable library functions for
5831 Do not generate multm or multmu instructions. This is useful for some embedded
5832 systems which do not have trap handlers for these instructions.
5836 @subsection ARM Options
5839 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5844 @opindex mapcs-frame
5845 Generate a stack frame that is compliant with the ARM Procedure Call
5846 Standard for all functions, even if this is not strictly necessary for
5847 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5848 with this option will cause the stack frames not to be generated for
5849 leaf functions. The default is @option{-mno-apcs-frame}.
5853 This is a synonym for @option{-mapcs-frame}.
5857 Generate code for a processor running with a 26-bit program counter,
5858 and conforming to the function calling standards for the APCS 26-bit
5859 option. This option replaces the @option{-m2} and @option{-m3} options
5860 of previous releases of the compiler.
5864 Generate code for a processor running with a 32-bit program counter,
5865 and conforming to the function calling standards for the APCS 32-bit
5866 option. This option replaces the @option{-m6} option of previous releases
5870 @c not currently implemented
5871 @item -mapcs-stack-check
5872 @opindex mapcs-stack-check
5873 Generate code to check the amount of stack space available upon entry to
5874 every function (that actually uses some stack space). If there is
5875 insufficient space available then either the function
5876 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5877 called, depending upon the amount of stack space required. The run time
5878 system is required to provide these functions. The default is
5879 @option{-mno-apcs-stack-check}, since this produces smaller code.
5881 @c not currently implemented
5883 @opindex mapcs-float
5884 Pass floating point arguments using the float point registers. This is
5885 one of the variants of the APCS@. This option is recommended if the
5886 target hardware has a floating point unit or if a lot of floating point
5887 arithmetic is going to be performed by the code. The default is
5888 @option{-mno-apcs-float}, since integer only code is slightly increased in
5889 size if @option{-mapcs-float} is used.
5891 @c not currently implemented
5892 @item -mapcs-reentrant
5893 @opindex mapcs-reentrant
5894 Generate reentrant, position independent code. The default is
5895 @option{-mno-apcs-reentrant}.
5898 @item -mthumb-interwork
5899 @opindex mthumb-interwork
5900 Generate code which supports calling between the ARM and Thumb
5901 instruction sets. Without this option the two instruction sets cannot
5902 be reliably used inside one program. The default is
5903 @option{-mno-thumb-interwork}, since slightly larger code is generated
5904 when @option{-mthumb-interwork} is specified.
5906 @item -mno-sched-prolog
5907 @opindex mno-sched-prolog
5908 Prevent the reordering of instructions in the function prolog, or the
5909 merging of those instruction with the instructions in the function's
5910 body. This means that all functions will start with a recognizable set
5911 of instructions (or in fact one of a choice from a small set of
5912 different function prologues), and this information can be used to
5913 locate the start if functions inside an executable piece of code. The
5914 default is @option{-msched-prolog}.
5917 @opindex mhard-float
5918 Generate output containing floating point instructions. This is the
5922 @opindex msoft-float
5923 Generate output containing library calls for floating point.
5924 @strong{Warning:} the requisite libraries are not available for all ARM
5925 targets. Normally the facilities of the machine's usual C compiler are
5926 used, but this cannot be done directly in cross-compilation. You must make
5927 your own arrangements to provide suitable library functions for
5930 @option{-msoft-float} changes the calling convention in the output file;
5931 therefore, it is only useful if you compile @emph{all} of a program with
5932 this option. In particular, you need to compile @file{libgcc.a}, the
5933 library that comes with GCC, with @option{-msoft-float} in order for
5936 @item -mlittle-endian
5937 @opindex mlittle-endian
5938 Generate code for a processor running in little-endian mode. This is
5939 the default for all standard configurations.
5942 @opindex mbig-endian
5943 Generate code for a processor running in big-endian mode; the default is
5944 to compile code for a little-endian processor.
5946 @item -mwords-little-endian
5947 @opindex mwords-little-endian
5948 This option only applies when generating code for big-endian processors.
5949 Generate code for a little-endian word order but a big-endian byte
5950 order. That is, a byte order of the form @samp{32107654}. Note: this
5951 option should only be used if you require compatibility with code for
5952 big-endian ARM processors generated by versions of the compiler prior to
5955 @item -malignment-traps
5956 @opindex malignment-traps
5957 Generate code that will not trap if the MMU has alignment traps enabled.
5958 On ARM architectures prior to ARMv4, there were no instructions to
5959 access half-word objects stored in memory. However, when reading from
5960 memory a feature of the ARM architecture allows a word load to be used,
5961 even if the address is unaligned, and the processor core will rotate the
5962 data as it is being loaded. This option tells the compiler that such
5963 misaligned accesses will cause a MMU trap and that it should instead
5964 synthesise the access as a series of byte accesses. The compiler can
5965 still use word accesses to load half-word data if it knows that the
5966 address is aligned to a word boundary.
5968 This option is ignored when compiling for ARM architecture 4 or later,
5969 since these processors have instructions to directly access half-word
5972 @item -mno-alignment-traps
5973 @opindex mno-alignment-traps
5974 Generate code that assumes that the MMU will not trap unaligned
5975 accesses. This produces better code when the target instruction set
5976 does not have half-word memory operations (i.e.@: implementations prior to
5979 Note that you cannot use this option to access unaligned word objects,
5980 since the processor will only fetch one 32-bit aligned object from
5983 The default setting for most targets is @option{-mno-alignment-traps}, since
5984 this produces better code when there are no half-word memory
5985 instructions available.
5987 @item -mshort-load-bytes
5988 @itemx -mno-short-load-words
5989 @opindex mshort-load-bytes
5990 @opindex mno-short-load-words
5991 These are deprecated aliases for @option{-malignment-traps}.
5993 @item -mno-short-load-bytes
5994 @itemx -mshort-load-words
5995 @opindex mno-short-load-bytes
5996 @opindex mshort-load-words
5997 This are deprecated aliases for @option{-mno-alignment-traps}.
6001 This option only applies to RISC iX@. Emulate the native BSD-mode
6002 compiler. This is the default if @option{-ansi} is not specified.
6006 This option only applies to RISC iX@. Emulate the native X/Open-mode
6009 @item -mno-symrename
6010 @opindex mno-symrename
6011 This option only applies to RISC iX@. Do not run the assembler
6012 post-processor, @samp{symrename}, after code has been assembled.
6013 Normally it is necessary to modify some of the standard symbols in
6014 preparation for linking with the RISC iX C library; this option
6015 suppresses this pass. The post-processor is never run when the
6016 compiler is built for cross-compilation.
6018 @item -mcpu=@var{name}
6020 This specifies the name of the target ARM processor. GCC uses this name
6021 to determine what kind of instructions it can emit when generating
6022 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6023 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6024 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6025 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6026 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6027 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6028 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6029 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6030 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
6031 @samp{arm1020t}, @samp{xscale}.
6033 @itemx -mtune=@var{name}
6035 This option is very similar to the @option{-mcpu=} option, except that
6036 instead of specifying the actual target processor type, and hence
6037 restricting which instructions can be used, it specifies that GCC should
6038 tune the performance of the code as if the target were of the type
6039 specified in this option, but still choosing the instructions that it
6040 will generate based on the cpu specified by a @option{-mcpu=} option.
6041 For some ARM implementations better performance can be obtained by using
6044 @item -march=@var{name}
6046 This specifies the name of the target ARM architecture. GCC uses this
6047 name to determine what kind of instructions it can emit when generating
6048 assembly code. This option can be used in conjunction with or instead
6049 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6050 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6051 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
6053 @item -mfpe=@var{number}
6054 @itemx -mfp=@var{number}
6057 This specifies the version of the floating point emulation available on
6058 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6059 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6061 @item -mstructure-size-boundary=@var{n}
6062 @opindex mstructure-size-boundary
6063 The size of all structures and unions will be rounded up to a multiple
6064 of the number of bits set by this option. Permissible values are 8 and
6065 32. The default value varies for different toolchains. For the COFF
6066 targeted toolchain the default value is 8. Specifying the larger number
6067 can produce faster, more efficient code, but can also increase the size
6068 of the program. The two values are potentially incompatible. Code
6069 compiled with one value cannot necessarily expect to work with code or
6070 libraries compiled with the other value, if they exchange information
6071 using structures or unions.
6073 @item -mabort-on-noreturn
6074 @opindex mabort-on-noreturn
6075 Generate a call to the function @code{abort} at the end of a
6076 @code{noreturn} function. It will be executed if the function tries to
6080 @itemx -mno-long-calls
6081 @opindex mlong-calls
6082 @opindex mno-long-calls
6083 Tells the compiler to perform function calls by first loading the
6084 address of the function into a register and then performing a subroutine
6085 call on this register. This switch is needed if the target function
6086 will lie outside of the 64 megabyte addressing range of the offset based
6087 version of subroutine call instruction.
6089 Even if this switch is enabled, not all function calls will be turned
6090 into long calls. The heuristic is that static functions, functions
6091 which have the @samp{short-call} attribute, functions that are inside
6092 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6093 definitions have already been compiled within the current compilation
6094 unit, will not be turned into long calls. The exception to this rule is
6095 that weak function definitions, functions with the @samp{long-call}
6096 attribute or the @samp{section} attribute, and functions that are within
6097 the scope of a @samp{#pragma long_calls} directive, will always be
6098 turned into long calls.
6100 This feature is not enabled by default. Specifying
6101 @option{-mno-long-calls} will restore the default behavior, as will
6102 placing the function calls within the scope of a @samp{#pragma
6103 long_calls_off} directive. Note these switches have no effect on how
6104 the compiler generates code to handle function calls via function
6107 @item -mnop-fun-dllimport
6108 @opindex mnop-fun-dllimport
6109 Disable support for the @code{dllimport} attribute.
6111 @item -msingle-pic-base
6112 @opindex msingle-pic-base
6113 Treat the register used for PIC addressing as read-only, rather than
6114 loading it in the prologue for each function. The run-time system is
6115 responsible for initializing this register with an appropriate value
6116 before execution begins.
6118 @item -mpic-register=@var{reg}
6119 @opindex mpic-register
6120 Specify the register to be used for PIC addressing. The default is R10
6121 unless stack-checking is enabled, when R9 is used.
6123 @item -mpoke-function-name
6124 @opindex mpoke-function-name
6125 Write the name of each function into the text section, directly
6126 preceding the function prologue. The generated code is similar to this:
6130 .ascii "arm_poke_function_name", 0
6133 .word 0xff000000 + (t1 - t0)
6134 arm_poke_function_name
6136 stmfd sp!, @{fp, ip, lr, pc@}
6140 When performing a stack backtrace, code can inspect the value of
6141 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6142 location @code{pc - 12} and the top 8 bits are set, then we know that
6143 there is a function name embedded immediately preceding this location
6144 and has length @code{((pc[-3]) & 0xff000000)}.
6148 Generate code for the 16-bit Thumb instruction set. The default is to
6149 use the 32-bit ARM instruction set.
6152 @opindex mtpcs-frame
6153 Generate a stack frame that is compliant with the Thumb Procedure Call
6154 Standard for all non-leaf functions. (A leaf function is one that does
6155 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6157 @item -mtpcs-leaf-frame
6158 @opindex mtpcs-leaf-frame
6159 Generate a stack frame that is compliant with the Thumb Procedure Call
6160 Standard for all leaf functions. (A leaf function is one that does
6161 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6163 @item -mcallee-super-interworking
6164 @opindex mcallee-super-interworking
6165 Gives all externally visible functions in the file being compiled an ARM
6166 instruction set header which switches to Thumb mode before executing the
6167 rest of the function. This allows these functions to be called from
6168 non-interworking code.
6170 @item -mcaller-super-interworking
6171 @opindex mcaller-super-interworking
6172 Allows calls via function pointers (including virtual functions) to
6173 execute correctly regardless of whether the target code has been
6174 compiled for interworking or not. There is a small overhead in the cost
6175 of executing a function pointer if this option is enabled.
6179 @node MN10200 Options
6180 @subsection MN10200 Options
6181 @cindex MN10200 options
6182 These @option{-m} options are defined for Matsushita MN10200 architectures:
6187 Indicate to the linker that it should perform a relaxation optimization pass
6188 to shorten branches, calls and absolute memory addresses. This option only
6189 has an effect when used on the command line for the final link step.
6191 This option makes symbolic debugging impossible.
6194 @node MN10300 Options
6195 @subsection MN10300 Options
6196 @cindex MN10300 options
6197 These @option{-m} options are defined for Matsushita MN10300 architectures:
6202 Generate code to avoid bugs in the multiply instructions for the MN10300
6203 processors. This is the default.
6206 @opindex mno-mult-bug
6207 Do not generate code to avoid bugs in the multiply instructions for the
6212 Generate code which uses features specific to the AM33 processor.
6216 Do not generate code which uses features specific to the AM33 processor. This
6221 Do not link in the C run-time initialization object file.
6225 Indicate to the linker that it should perform a relaxation optimization pass
6226 to shorten branches, calls and absolute memory addresses. This option only
6227 has an effect when used on the command line for the final link step.
6229 This option makes symbolic debugging impossible.
6233 @node M32R/D Options
6234 @subsection M32R/D Options
6235 @cindex M32R/D options
6237 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6242 Generate code for the M32R/X@.
6246 Generate code for the M32R@. This is the default.
6248 @item -mcode-model=small
6249 @opindex mcode-model=small
6250 Assume all objects live in the lower 16MB of memory (so that their addresses
6251 can be loaded with the @code{ld24} instruction), and assume all subroutines
6252 are reachable with the @code{bl} instruction.
6253 This is the default.
6255 The addressability of a particular object can be set with the
6256 @code{model} attribute.
6258 @item -mcode-model=medium
6259 @opindex mcode-model=medium
6260 Assume objects may be anywhere in the 32-bit address space (the compiler
6261 will generate @code{seth/add3} instructions to load their addresses), and
6262 assume all subroutines are reachable with the @code{bl} instruction.
6264 @item -mcode-model=large
6265 @opindex mcode-model=large
6266 Assume objects may be anywhere in the 32-bit address space (the compiler
6267 will generate @code{seth/add3} instructions to load their addresses), and
6268 assume subroutines may not be reachable with the @code{bl} instruction
6269 (the compiler will generate the much slower @code{seth/add3/jl}
6270 instruction sequence).
6273 @opindex msdata=none
6274 Disable use of the small data area. Variables will be put into
6275 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6276 @code{section} attribute has been specified).
6277 This is the default.
6279 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6280 Objects may be explicitly put in the small data area with the
6281 @code{section} attribute using one of these sections.
6284 @opindex msdata=sdata
6285 Put small global and static data in the small data area, but do not
6286 generate special code to reference them.
6290 Put small global and static data in the small data area, and generate
6291 special instructions to reference them.
6295 @cindex smaller data references
6296 Put global and static objects less than or equal to @var{num} bytes
6297 into the small data or bss sections instead of the normal data or bss
6298 sections. The default value of @var{num} is 8.
6299 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6300 for this option to have any effect.
6302 All modules should be compiled with the same @option{-G @var{num}} value.
6303 Compiling with different values of @var{num} may or may not work; if it
6304 doesn't the linker will give an error message---incorrect code will not be
6310 @subsection M88K Options
6311 @cindex M88k options
6313 These @samp{-m} options are defined for Motorola 88k architectures:
6318 Generate code that works well on both the m88100 and the
6323 Generate code that works best for the m88100, but that also
6328 Generate code that works best for the m88110, and may not run
6333 Obsolete option to be removed from the next revision.
6336 @item -midentify-revision
6337 @opindex midentify-revision
6338 @cindex identifying source, compiler (88k)
6339 Include an @code{ident} directive in the assembler output recording the
6340 source file name, compiler name and version, timestamp, and compilation
6343 @item -mno-underscores
6344 @opindex mno-underscores
6345 @cindex underscores, avoiding (88k)
6346 In assembler output, emit symbol names without adding an underscore
6347 character at the beginning of each name. The default is to use an
6348 underscore as prefix on each name.
6350 @item -mocs-debug-info
6351 @itemx -mno-ocs-debug-info
6352 @opindex mocs-debug-info
6353 @opindex mno-ocs-debug-info
6355 @cindex debugging, 88k OCS
6356 Include (or omit) additional debugging information (about registers used
6357 in each stack frame) as specified in the 88open Object Compatibility
6358 Standard, ``OCS''@. This extra information allows debugging of code that
6359 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6360 Delta 88 SVr3.2 is to include this information; other 88k configurations
6361 omit this information by default.
6363 @item -mocs-frame-position
6364 @opindex mocs-frame-position
6365 @cindex register positions in frame (88k)
6366 When emitting COFF debugging information for automatic variables and
6367 parameters stored on the stack, use the offset from the canonical frame
6368 address, which is the stack pointer (register 31) on entry to the
6369 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6370 @option{-mocs-frame-position}; other 88k configurations have the default
6371 @option{-mno-ocs-frame-position}.
6373 @item -mno-ocs-frame-position
6374 @opindex mno-ocs-frame-position
6375 @cindex register positions in frame (88k)
6376 When emitting COFF debugging information for automatic variables and
6377 parameters stored on the stack, use the offset from the frame pointer
6378 register (register 30). When this option is in effect, the frame
6379 pointer is not eliminated when debugging information is selected by the
6382 @item -moptimize-arg-area
6383 @opindex moptimize-arg-area
6384 @cindex arguments in frame (88k)
6385 Save space by reorganizing the stack frame. This option generates code
6386 that does not agree with the 88open specifications, but uses less
6389 @itemx -mno-optimize-arg-area
6390 @opindex mno-optimize-arg-area
6391 Do not reorganize the stack frame to save space. This is the default.
6392 The generated conforms to the specification, but uses more memory.
6394 @item -mshort-data-@var{num}
6395 @opindex mshort-data
6396 @cindex smaller data references (88k)
6397 @cindex r0-relative references (88k)
6398 Generate smaller data references by making them relative to @code{r0},
6399 which allows loading a value using a single instruction (rather than the
6400 usual two). You control which data references are affected by
6401 specifying @var{num} with this option. For example, if you specify
6402 @option{-mshort-data-512}, then the data references affected are those
6403 involving displacements of less than 512 bytes.
6404 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6407 @item -mserialize-volatile
6408 @opindex mserialize-volatile
6409 @itemx -mno-serialize-volatile
6410 @opindex mno-serialize-volatile
6411 @cindex sequential consistency on 88k
6412 Do, or don't, generate code to guarantee sequential consistency
6413 of volatile memory references. By default, consistency is
6416 The order of memory references made by the MC88110 processor does
6417 not always match the order of the instructions requesting those
6418 references. In particular, a load instruction may execute before
6419 a preceding store instruction. Such reordering violates
6420 sequential consistency of volatile memory references, when there
6421 are multiple processors. When consistency must be guaranteed,
6422 GCC generates special instructions, as needed, to force
6423 execution in the proper order.
6425 The MC88100 processor does not reorder memory references and so
6426 always provides sequential consistency. However, by default, GCC
6427 generates the special instructions to guarantee consistency
6428 even when you use @option{-m88100}, so that the code may be run on an
6429 MC88110 processor. If you intend to run your code only on the
6430 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6432 The extra code generated to guarantee consistency may affect the
6433 performance of your application. If you know that you can safely
6434 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6440 @cindex assembler syntax, 88k
6442 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6443 related to System V release 4 (SVr4). This controls the following:
6447 Which variant of the assembler syntax to emit.
6449 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6450 that is used on System V release 4.
6452 @option{-msvr4} makes GCC issue additional declaration directives used in
6456 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6457 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6458 other m88k configurations.
6460 @item -mversion-03.00
6461 @opindex mversion-03.00
6462 This option is obsolete, and is ignored.
6463 @c ??? which asm syntax better for GAS? option there too?
6465 @item -mno-check-zero-division
6466 @itemx -mcheck-zero-division
6467 @opindex mno-check-zero-division
6468 @opindex mcheck-zero-division
6469 @cindex zero division on 88k
6470 Do, or don't, generate code to guarantee that integer division by
6471 zero will be detected. By default, detection is guaranteed.
6473 Some models of the MC88100 processor fail to trap upon integer
6474 division by zero under certain conditions. By default, when
6475 compiling code that might be run on such a processor, GCC
6476 generates code that explicitly checks for zero-valued divisors
6477 and traps with exception number 503 when one is detected. Use of
6478 @option{-mno-check-zero-division} suppresses such checking for code
6479 generated to run on an MC88100 processor.
6481 GCC assumes that the MC88110 processor correctly detects all instances
6482 of integer division by zero. When @option{-m88110} is specified, no
6483 explicit checks for zero-valued divisors are generated, and both
6484 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6487 @item -muse-div-instruction
6488 @opindex muse-div-instruction
6489 @cindex divide instruction, 88k
6490 Use the div instruction for signed integer division on the
6491 MC88100 processor. By default, the div instruction is not used.
6493 On the MC88100 processor the signed integer division instruction
6494 div) traps to the operating system on a negative operand. The
6495 operating system transparently completes the operation, but at a
6496 large cost in execution time. By default, when compiling code
6497 that might be run on an MC88100 processor, GCC emulates signed
6498 integer division using the unsigned integer division instruction
6499 divu), thereby avoiding the large penalty of a trap to the
6500 operating system. Such emulation has its own, smaller, execution
6501 cost in both time and space. To the extent that your code's
6502 important signed integer division operations are performed on two
6503 nonnegative operands, it may be desirable to use the div
6504 instruction directly.
6506 On the MC88110 processor the div instruction (also known as the
6507 divs instruction) processes negative operands without trapping to
6508 the operating system. When @option{-m88110} is specified,
6509 @option{-muse-div-instruction} is ignored, and the div instruction is used
6510 for signed integer division.
6512 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6513 particular, the behavior of such a division with and without
6514 @option{-muse-div-instruction} may differ.
6516 @item -mtrap-large-shift
6517 @itemx -mhandle-large-shift
6518 @opindex mtrap-large-shift
6519 @opindex mhandle-large-shift
6520 @cindex bit shift overflow (88k)
6521 @cindex large bit shifts (88k)
6522 Include code to detect bit-shifts of more than 31 bits; respectively,
6523 trap such shifts or emit code to handle them properly. By default GCC
6524 makes no special provision for large bit shifts.
6526 @item -mwarn-passed-structs
6527 @opindex mwarn-passed-structs
6528 @cindex structure passing (88k)
6529 Warn when a function passes a struct as an argument or result.
6530 Structure-passing conventions have changed during the evolution of the C
6531 language, and are often the source of portability problems. By default,
6532 GCC issues no such warning.
6535 @c break page here to avoid unsightly interparagraph stretch.
6539 @node RS/6000 and PowerPC Options
6540 @subsection IBM RS/6000 and PowerPC Options
6541 @cindex RS/6000 and PowerPC Options
6542 @cindex IBM RS/6000 and PowerPC Options
6544 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6552 @itemx -mpowerpc-gpopt
6553 @itemx -mno-powerpc-gpopt
6554 @itemx -mpowerpc-gfxopt
6555 @itemx -mno-powerpc-gfxopt
6557 @itemx -mno-powerpc64
6563 @opindex mno-powerpc
6564 @opindex mpowerpc-gpopt
6565 @opindex mno-powerpc-gpopt
6566 @opindex mpowerpc-gfxopt
6567 @opindex mno-powerpc-gfxopt
6569 @opindex mno-powerpc64
6570 GCC supports two related instruction set architectures for the
6571 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6572 instructions supported by the @samp{rios} chip set used in the original
6573 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6574 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6575 the IBM 4xx microprocessors.
6577 Neither architecture is a subset of the other. However there is a
6578 large common subset of instructions supported by both. An MQ
6579 register is included in processors supporting the POWER architecture.
6581 You use these options to specify which instructions are available on the
6582 processor you are using. The default value of these options is
6583 determined when configuring GCC@. Specifying the
6584 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6585 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6586 rather than the options listed above.
6588 The @option{-mpower} option allows GCC to generate instructions that
6589 are found only in the POWER architecture and to use the MQ register.
6590 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6591 to generate instructions that are present in the POWER2 architecture but
6592 not the original POWER architecture.
6594 The @option{-mpowerpc} option allows GCC to generate instructions that
6595 are found only in the 32-bit subset of the PowerPC architecture.
6596 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6597 GCC to use the optional PowerPC architecture instructions in the
6598 General Purpose group, including floating-point square root. Specifying
6599 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6600 use the optional PowerPC architecture instructions in the Graphics
6601 group, including floating-point select.
6603 The @option{-mpowerpc64} option allows GCC to generate the additional
6604 64-bit instructions that are found in the full PowerPC64 architecture
6605 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6606 @option{-mno-powerpc64}.
6608 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6609 will use only the instructions in the common subset of both
6610 architectures plus some special AIX common-mode calls, and will not use
6611 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6612 permits GCC to use any instruction from either architecture and to
6613 allow use of the MQ register; specify this for the Motorola MPC601.
6615 @item -mnew-mnemonics
6616 @itemx -mold-mnemonics
6617 @opindex mnew-mnemonics
6618 @opindex mold-mnemonics
6619 Select which mnemonics to use in the generated assembler code. With
6620 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6621 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6622 assembler mnemonics defined for the POWER architecture. Instructions
6623 defined in only one architecture have only one mnemonic; GCC uses that
6624 mnemonic irrespective of which of these options is specified.
6626 GCC defaults to the mnemonics appropriate for the architecture in
6627 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6628 value of these option. Unless you are building a cross-compiler, you
6629 should normally not specify either @option{-mnew-mnemonics} or
6630 @option{-mold-mnemonics}, but should instead accept the default.
6632 @item -mcpu=@var{cpu_type}
6634 Set architecture type, register usage, choice of mnemonics, and
6635 instruction scheduling parameters for machine type @var{cpu_type}.
6636 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6637 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6638 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6639 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6640 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6641 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6643 @option{-mcpu=common} selects a completely generic processor. Code
6644 generated under this option will run on any POWER or PowerPC processor.
6645 GCC will use only the instructions in the common subset of both
6646 architectures, and will not use the MQ register. GCC assumes a generic
6647 processor model for scheduling purposes.
6649 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6650 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6651 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6652 types, with an appropriate, generic processor model assumed for
6653 scheduling purposes.
6655 The other options specify a specific processor. Code generated under
6656 those options will run best on that processor, and may not run at all on
6659 The @option{-mcpu} options automatically enable or disable other
6660 @option{-m} options as follows:
6664 @option{-mno-power}, @option{-mno-powerc}
6671 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6686 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6689 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6694 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6697 @item -mtune=@var{cpu_type}
6699 Set the instruction scheduling parameters for machine type
6700 @var{cpu_type}, but do not set the architecture type, register usage, or
6701 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6702 values for @var{cpu_type} are used for @option{-mtune} as for
6703 @option{-mcpu}. If both are specified, the code generated will use the
6704 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6705 scheduling parameters set by @option{-mtune}.
6710 @opindex mno-altivec
6711 These switches enable or disable the use of built-in functions that
6712 allow access to the AltiVec instruction set. You may also need to set
6713 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6717 @itemx -mno-fp-in-toc
6718 @itemx -mno-sum-in-toc
6719 @itemx -mminimal-toc
6721 @opindex mno-fp-in-toc
6722 @opindex mno-sum-in-toc
6723 @opindex mminimal-toc
6724 Modify generation of the TOC (Table Of Contents), which is created for
6725 every executable file. The @option{-mfull-toc} option is selected by
6726 default. In that case, GCC will allocate at least one TOC entry for
6727 each unique non-automatic variable reference in your program. GCC
6728 will also place floating-point constants in the TOC@. However, only
6729 16,384 entries are available in the TOC@.
6731 If you receive a linker error message that saying you have overflowed
6732 the available TOC space, you can reduce the amount of TOC space used
6733 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6734 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6735 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6736 generate code to calculate the sum of an address and a constant at
6737 run-time instead of putting that sum into the TOC@. You may specify one
6738 or both of these options. Each causes GCC to produce very slightly
6739 slower and larger code at the expense of conserving TOC space.
6741 If you still run out of space in the TOC even when you specify both of
6742 these options, specify @option{-mminimal-toc} instead. This option causes
6743 GCC to make only one TOC entry for every file. When you specify this
6744 option, GCC will produce code that is slower and larger but which
6745 uses extremely little TOC space. You may wish to use this option
6746 only on files that contain less frequently executed code.
6752 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6753 @code{long} type, and the infrastructure needed to support them.
6754 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6755 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6756 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6761 @opindex mno-xl-call
6762 On AIX, pass floating-point arguments to prototyped functions beyond the
6763 register save area (RSA) on the stack in addition to argument FPRs. The
6764 AIX calling convention was extended but not initially documented to
6765 handle an obscure K&R C case of calling a function that takes the
6766 address of its arguments with fewer arguments than declared. AIX XL
6767 compilers access floating point arguments which do not fit in the
6768 RSA from the stack when a subroutine is compiled without
6769 optimization. Because always storing floating-point arguments on the
6770 stack is inefficient and rarely needed, this option is not enabled by
6771 default and only is necessary when calling subroutines compiled by AIX
6772 XL compilers without optimization.
6776 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6777 application written to use message passing with special startup code to
6778 enable the application to run. The system must have PE installed in the
6779 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6780 must be overridden with the @option{-specs=} option to specify the
6781 appropriate directory location. The Parallel Environment does not
6782 support threads, so the @option{-mpe} option and the @option{-pthread}
6783 option are incompatible.
6787 @opindex msoft-float
6788 @opindex mhard-float
6789 Generate code that does not use (uses) the floating-point register set.
6790 Software floating point emulation is provided if you use the
6791 @option{-msoft-float} option, and pass the option to GCC when linking.
6794 @itemx -mno-multiple
6796 @opindex mno-multiple
6797 Generate code that uses (does not use) the load multiple word
6798 instructions and the store multiple word instructions. These
6799 instructions are generated by default on POWER systems, and not
6800 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6801 endian PowerPC systems, since those instructions do not work when the
6802 processor is in little endian mode. The exceptions are PPC740 and
6803 PPC750 which permit the instructions usage in little endian mode.
6809 Generate code that uses (does not use) the load string instructions
6810 and the store string word instructions to save multiple registers and
6811 do small block moves. These instructions are generated by default on
6812 POWER systems, and not generated on PowerPC systems. Do not use
6813 @option{-mstring} on little endian PowerPC systems, since those
6814 instructions do not work when the processor is in little endian mode.
6815 The exceptions are PPC740 and PPC750 which permit the instructions
6816 usage in little endian mode.
6822 Generate code that uses (does not use) the load or store instructions
6823 that update the base register to the address of the calculated memory
6824 location. These instructions are generated by default. If you use
6825 @option{-mno-update}, there is a small window between the time that the
6826 stack pointer is updated and the address of the previous frame is
6827 stored, which means code that walks the stack frame across interrupts or
6828 signals may get corrupted data.
6831 @itemx -mno-fused-madd
6832 @opindex mfused-madd
6833 @opindex mno-fused-madd
6834 Generate code that uses (does not use) the floating point multiply and
6835 accumulate instructions. These instructions are generated by default if
6836 hardware floating is used.
6838 @item -mno-bit-align
6840 @opindex mno-bit-align
6842 On System V.4 and embedded PowerPC systems do not (do) force structures
6843 and unions that contain bit-fields to be aligned to the base type of the
6846 For example, by default a structure containing nothing but 8
6847 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6848 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6849 the structure would be aligned to a 1 byte boundary and be one byte in
6852 @item -mno-strict-align
6853 @itemx -mstrict-align
6854 @opindex mno-strict-align
6855 @opindex mstrict-align
6856 On System V.4 and embedded PowerPC systems do not (do) assume that
6857 unaligned memory references will be handled by the system.
6860 @itemx -mno-relocatable
6861 @opindex mrelocatable
6862 @opindex mno-relocatable
6863 On embedded PowerPC systems generate code that allows (does not allow)
6864 the program to be relocated to a different address at runtime. If you
6865 use @option{-mrelocatable} on any module, all objects linked together must
6866 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6868 @item -mrelocatable-lib
6869 @itemx -mno-relocatable-lib
6870 @opindex mrelocatable-lib
6871 @opindex mno-relocatable-lib
6872 On embedded PowerPC systems generate code that allows (does not allow)
6873 the program to be relocated to a different address at runtime. Modules
6874 compiled with @option{-mrelocatable-lib} can be linked with either modules
6875 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6876 with modules compiled with the @option{-mrelocatable} options.
6882 On System V.4 and embedded PowerPC systems do not (do) assume that
6883 register 2 contains a pointer to a global area pointing to the addresses
6884 used in the program.
6887 @itemx -mlittle-endian
6889 @opindex mlittle-endian
6890 On System V.4 and embedded PowerPC systems compile code for the
6891 processor in little endian mode. The @option{-mlittle-endian} option is
6892 the same as @option{-mlittle}.
6897 @opindex mbig-endian
6898 On System V.4 and embedded PowerPC systems compile code for the
6899 processor in big endian mode. The @option{-mbig-endian} option is
6900 the same as @option{-mbig}.
6904 On System V.4 and embedded PowerPC systems compile code using calling
6905 conventions that adheres to the March 1995 draft of the System V
6906 Application Binary Interface, PowerPC processor supplement. This is the
6907 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6909 @item -mcall-sysv-eabi
6910 @opindex mcall-sysv-eabi
6911 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6913 @item -mcall-sysv-noeabi
6914 @opindex mcall-sysv-noeabi
6915 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6919 On System V.4 and embedded PowerPC systems compile code using calling
6920 conventions that are similar to those used on AIX@. This is the
6921 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6923 @item -mcall-solaris
6924 @opindex mcall-solaris
6925 On System V.4 and embedded PowerPC systems compile code for the Solaris
6929 @opindex mcall-linux
6930 On System V.4 and embedded PowerPC systems compile code for the
6931 Linux-based GNU system.
6935 On System V.4 and embedded PowerPC systems compile code for the
6936 Hurd-based GNU system.
6939 @opindex mcall-netbsd
6940 On System V.4 and embedded PowerPC systems compile code for the
6941 NetBSD operating system.
6943 @item -maix-struct-return
6944 @opindex maix-struct-return
6945 Return all structures in memory (as specified by the AIX ABI)@.
6947 @item -msvr4-struct-return
6948 @opindex msvr4-struct-return
6949 Return structures smaller than 8 bytes in registers (as specified by the
6953 @opindex mabi=altivec
6954 Extend the current ABI with AltiVec ABI extensions. This does not
6955 change the default ABI, instead it adds the AltiVec ABI extensions to
6959 @itemx -mno-prototype
6961 @opindex mno-prototype
6962 On System V.4 and embedded PowerPC systems assume that all calls to
6963 variable argument functions are properly prototyped. Otherwise, the
6964 compiler must insert an instruction before every non prototyped call to
6965 set or clear bit 6 of the condition code register (@var{CR}) to
6966 indicate whether floating point values were passed in the floating point
6967 registers in case the function takes a variable arguments. With
6968 @option{-mprototype}, only calls to prototyped variable argument functions
6969 will set or clear the bit.
6973 On embedded PowerPC systems, assume that the startup module is called
6974 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6975 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6980 On embedded PowerPC systems, assume that the startup module is called
6981 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6986 On embedded PowerPC systems, assume that the startup module is called
6987 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6991 @opindex myellowknife
6992 On embedded PowerPC systems, assume that the startup module is called
6993 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6998 On System V.4 and embedded PowerPC systems, specify that you are
6999 compiling for a VxWorks system.
7003 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7004 header to indicate that @samp{eabi} extended relocations are used.
7010 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7011 Embedded Applications Binary Interface (eabi) which is a set of
7012 modifications to the System V.4 specifications. Selecting @option{-meabi}
7013 means that the stack is aligned to an 8 byte boundary, a function
7014 @code{__eabi} is called to from @code{main} to set up the eabi
7015 environment, and the @option{-msdata} option can use both @code{r2} and
7016 @code{r13} to point to two separate small data areas. Selecting
7017 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7018 do not call an initialization function from @code{main}, and the
7019 @option{-msdata} option will only use @code{r13} to point to a single
7020 small data area. The @option{-meabi} option is on by default if you
7021 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7024 @opindex msdata=eabi
7025 On System V.4 and embedded PowerPC systems, put small initialized
7026 @code{const} global and static data in the @samp{.sdata2} section, which
7027 is pointed to by register @code{r2}. Put small initialized
7028 non-@code{const} global and static data in the @samp{.sdata} section,
7029 which is pointed to by register @code{r13}. Put small uninitialized
7030 global and static data in the @samp{.sbss} section, which is adjacent to
7031 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7032 incompatible with the @option{-mrelocatable} option. The
7033 @option{-msdata=eabi} option also sets the @option{-memb} option.
7036 @opindex msdata=sysv
7037 On System V.4 and embedded PowerPC systems, put small global and static
7038 data in the @samp{.sdata} section, which is pointed to by register
7039 @code{r13}. Put small uninitialized global and static data in the
7040 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7041 The @option{-msdata=sysv} option is incompatible with the
7042 @option{-mrelocatable} option.
7044 @item -msdata=default
7046 @opindex msdata=default
7048 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7049 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7050 same as @option{-msdata=sysv}.
7053 @opindex msdata-data
7054 On System V.4 and embedded PowerPC systems, put small global and static
7055 data in the @samp{.sdata} section. Put small uninitialized global and
7056 static data in the @samp{.sbss} section. Do not use register @code{r13}
7057 to address small data however. This is the default behavior unless
7058 other @option{-msdata} options are used.
7062 @opindex msdata=none
7064 On embedded PowerPC systems, put all initialized global and static data
7065 in the @samp{.data} section, and all uninitialized data in the
7066 @samp{.bss} section.
7070 @cindex smaller data references (PowerPC)
7071 @cindex .sdata/.sdata2 references (PowerPC)
7072 On embedded PowerPC systems, put global and static items less than or
7073 equal to @var{num} bytes into the small data or bss sections instead of
7074 the normal data or bss section. By default, @var{num} is 8. The
7075 @option{-G @var{num}} switch is also passed to the linker.
7076 All modules should be compiled with the same @option{-G @var{num}} value.
7079 @itemx -mno-regnames
7081 @opindex mno-regnames
7082 On System V.4 and embedded PowerPC systems do (do not) emit register
7083 names in the assembly language output using symbolic forms.
7087 Adds support for multithreading with the @dfn{pthreads} library.
7088 This option sets flags for both the preprocessor and linker.
7093 @subsection IBM RT Options
7095 @cindex IBM RT options
7097 These @samp{-m} options are defined for the IBM RT PC:
7101 @opindex min-line-mul
7102 Use an in-line code sequence for integer multiplies. This is the
7105 @item -mcall-lib-mul
7106 @opindex mcall-lib-mul
7107 Call @code{lmul$$} for integer multiples.
7109 @item -mfull-fp-blocks
7110 @opindex mfull-fp-blocks
7111 Generate full-size floating point data blocks, including the minimum
7112 amount of scratch space recommended by IBM@. This is the default.
7114 @item -mminimum-fp-blocks
7115 @opindex mminimum-fp-blocks
7116 Do not include extra scratch space in floating point data blocks. This
7117 results in smaller code, but slower execution, since scratch space must
7118 be allocated dynamically.
7120 @cindex @file{varargs.h} and RT PC
7121 @cindex @file{stdarg.h} and RT PC
7122 @item -mfp-arg-in-fpregs
7123 @opindex mfp-arg-in-fpregs
7124 Use a calling sequence incompatible with the IBM calling convention in
7125 which floating point arguments are passed in floating point registers.
7126 Note that @code{varargs.h} and @code{stdarg.h} will not work with
7127 floating point operands if this option is specified.
7129 @item -mfp-arg-in-gregs
7130 @opindex mfp-arg-in-gregs
7131 Use the normal calling convention for floating point arguments. This is
7134 @item -mhc-struct-return
7135 @opindex mhc-struct-return
7136 Return structures of more than one word in memory, rather than in a
7137 register. This provides compatibility with the MetaWare HighC (hc)
7138 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7139 with the Portable C Compiler (pcc).
7141 @item -mnohc-struct-return
7142 @opindex mnohc-struct-return
7143 Return some structures of more than one word in registers, when
7144 convenient. This is the default. For compatibility with the
7145 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7146 option @option{-mhc-struct-return}.
7150 @subsection MIPS Options
7151 @cindex MIPS options
7153 These @samp{-m} options are defined for the MIPS family of computers:
7157 @item -march=@var{cpu-type}
7159 Assume the defaults for the machine type @var{cpu-type} when generating
7160 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7161 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7162 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7163 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7164 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7165 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
7167 @item -mtune=@var{cpu-type}
7169 Assume the defaults for the machine type @var{cpu-type} when scheduling
7170 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
7171 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
7172 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
7173 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
7174 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
7175 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
7176 @var{cpu-type} will schedule things appropriately for that particular
7177 chip, the compiler will not generate any code that does not meet level 1
7178 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
7179 or @option{-mabi} switch being used.
7181 @item -mcpu=@var{cpu-type}
7183 This is identical to specifying both @option{-march} and @option{-mtune}.
7187 Issue instructions from level 1 of the MIPS ISA@. This is the default.
7188 @samp{r3000} is the default @var{cpu-type} at this ISA level.
7192 Issue instructions from level 2 of the MIPS ISA (branch likely, square
7193 root instructions). @samp{r6000} is the default @var{cpu-type} at this
7198 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
7199 @samp{r4000} is the default @var{cpu-type} at this ISA level.
7203 Issue instructions from level 4 of the MIPS ISA (conditional move,
7204 prefetch, enhanced FPU instructions). @samp{r8000} is the default
7205 @var{cpu-type} at this ISA level.
7209 Assume that 32 32-bit floating point registers are available. This is
7214 Assume that 32 64-bit floating point registers are available. This is
7215 the default when the @option{-mips3} option is used.
7218 @itemx -mno-fused-madd
7219 @opindex mfused-madd
7220 @opindex mno-fused-madd
7221 Generate code that uses (does not use) the floating point multiply and
7222 accumulate instructions, when they are available. These instructions
7223 are generated by default if they are available, but this may be
7224 undesirable if the extra precision causes problems or on certain chips
7225 in the mode where denormals are rounded to zero where denormals
7226 generated by multiply and accumulate instructions cause exceptions
7231 Assume that 32 32-bit general purpose registers are available. This is
7236 Assume that 32 64-bit general purpose registers are available. This is
7237 the default when the @option{-mips3} option is used.
7241 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7242 explanation of the default, and the width of pointers.
7246 Force long types to be 64 bits wide. See @option{-mlong32} for an
7247 explanation of the default, and the width of pointers.
7251 Force long, int, and pointer types to be 32 bits wide.
7253 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
7254 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
7255 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
7256 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
7257 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
7258 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
7259 are 32 bits, and longs are 64 bits wide. The width of pointer types is
7260 the smaller of the width of longs or the width of general purpose
7261 registers (which in turn depends on the ISA)@.
7273 Generate code for the indicated ABI@. The default instruction level is
7274 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
7275 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
7276 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
7281 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7282 add normal debug information. This is the default for all
7283 platforms except for the OSF/1 reference platform, using the OSF/rose
7284 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7285 switches are used, the @file{mips-tfile} program will encapsulate the
7286 stabs within MIPS ECOFF@.
7290 Generate code for the GNU assembler. This is the default on the OSF/1
7291 reference platform, using the OSF/rose object format. Also, this is
7292 the default if the configure option @option{--with-gnu-as} is used.
7294 @item -msplit-addresses
7295 @itemx -mno-split-addresses
7296 @opindex msplit-addresses
7297 @opindex mno-split-addresses
7298 Generate code to load the high and low parts of address constants separately.
7299 This allows GCC to optimize away redundant loads of the high order
7300 bits of addresses. This optimization requires GNU as and GNU ld.
7301 This optimization is enabled by default for some embedded targets where
7302 GNU as and GNU ld are standard.
7308 The @option{-mrnames} switch says to output code using the MIPS software
7309 names for the registers, instead of the hardware names (ie, @var{a0}
7310 instead of @var{$4}). The only known assembler that supports this option
7311 is the Algorithmics assembler.
7317 The @option{-mgpopt} switch says to write all of the data declarations
7318 before the instructions in the text section, this allows the MIPS
7319 assembler to generate one word memory references instead of using two
7320 words for short global or static data items. This is on by default if
7321 optimization is selected.
7327 For each non-inline function processed, the @option{-mstats} switch
7328 causes the compiler to emit one line to the standard error file to
7329 print statistics about the program (number of registers saved, stack
7336 The @option{-mmemcpy} switch makes all block moves call the appropriate
7337 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7338 generating inline code.
7341 @itemx -mno-mips-tfile
7342 @opindex mmips-tfile
7343 @opindex mno-mips-tfile
7344 The @option{-mno-mips-tfile} switch causes the compiler not
7345 postprocess the object file with the @file{mips-tfile} program,
7346 after the MIPS assembler has generated it to add debug support. If
7347 @file{mips-tfile} is not run, then no local variables will be
7348 available to the debugger. In addition, @file{stage2} and
7349 @file{stage3} objects will have the temporary file names passed to the
7350 assembler embedded in the object file, which means the objects will
7351 not compare the same. The @option{-mno-mips-tfile} switch should only
7352 be used when there are bugs in the @file{mips-tfile} program that
7353 prevents compilation.
7356 @opindex msoft-float
7357 Generate output containing library calls for floating point.
7358 @strong{Warning:} the requisite libraries are not part of GCC@.
7359 Normally the facilities of the machine's usual C compiler are used, but
7360 this can't be done directly in cross-compilation. You must make your
7361 own arrangements to provide suitable library functions for
7365 @opindex mhard-float
7366 Generate output containing floating point instructions. This is the
7367 default if you use the unmodified sources.
7370 @itemx -mno-abicalls
7372 @opindex mno-abicalls
7373 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7374 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7375 position independent code.
7378 @itemx -mno-long-calls
7379 @opindex mlong-calls
7380 @opindex mno-long-calls
7381 Do all calls with the @samp{JALR} instruction, which requires
7382 loading up a function's address into a register before the call.
7383 You need to use this switch, if you call outside of the current
7384 512 megabyte segment to functions that are not through pointers.
7387 @itemx -mno-half-pic
7389 @opindex mno-half-pic
7390 Put pointers to extern references into the data section and load them
7391 up, rather than put the references in the text section.
7393 @item -membedded-pic
7394 @itemx -mno-embedded-pic
7395 @opindex membedded-pic
7396 @opindex mno-embedded-pic
7397 Generate PIC code suitable for some embedded systems. All calls are
7398 made using PC relative address, and all data is addressed using the $gp
7399 register. No more than 65536 bytes of global data may be used. This
7400 requires GNU as and GNU ld which do most of the work. This currently
7401 only works on targets which use ECOFF; it does not work with ELF@.
7403 @item -membedded-data
7404 @itemx -mno-embedded-data
7405 @opindex membedded-data
7406 @opindex mno-embedded-data
7407 Allocate variables to the read-only data section first if possible, then
7408 next in the small data section if possible, otherwise in data. This gives
7409 slightly slower code than the default, but reduces the amount of RAM required
7410 when executing, and thus may be preferred for some embedded systems.
7412 @item -muninit-const-in-rodata
7413 @itemx -mno-uninit-const-in-rodata
7414 @opindex muninit-const-in-rodata
7415 @opindex mno-uninit-const-in-rodata
7416 When used together with @option{-membedded-data}, it will always store uninitialized
7417 const variables in the read-only data section.
7419 @item -msingle-float
7420 @itemx -mdouble-float
7421 @opindex msingle-float
7422 @opindex mdouble-float
7423 The @option{-msingle-float} switch tells gcc to assume that the floating
7424 point coprocessor only supports single precision operations, as on the
7425 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7426 double precision operations. This is the default.
7432 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7433 as on the @samp{r4650} chip.
7437 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7438 @option{-mcpu=r4650}.
7444 Enable 16-bit instructions.
7448 Use the entry and exit pseudo ops. This option can only be used with
7453 Compile code for the processor in little endian mode.
7454 The requisite libraries are assumed to exist.
7458 Compile code for the processor in big endian mode.
7459 The requisite libraries are assumed to exist.
7463 @cindex smaller data references (MIPS)
7464 @cindex gp-relative references (MIPS)
7465 Put global and static items less than or equal to @var{num} bytes into
7466 the small data or bss sections instead of the normal data or bss
7467 section. This allows the assembler to emit one word memory reference
7468 instructions based on the global pointer (@var{gp} or @var{$28}),
7469 instead of the normal two words used. By default, @var{num} is 8 when
7470 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7471 @option{-G @var{num}} switch is also passed to the assembler and linker.
7472 All modules should be compiled with the same @option{-G @var{num}}
7477 Tell the MIPS assembler to not run its preprocessor over user
7478 assembler files (with a @samp{.s} suffix) when assembling them.
7482 Pass an option to gas which will cause nops to be inserted if
7483 the read of the destination register of an mfhi or mflo instruction
7484 occurs in the following two instructions.
7488 Do not include the default crt0.
7490 @item -mflush-func=@var{func}
7491 @itemx -mno-flush-func
7492 @opindex mflush-func
7493 Specifies the function to call to flush the I and D caches, or to not
7494 call any such function. If called, the function must take the same
7495 arguments as the common @code{_flush_func()}, that is, the address of the
7496 memory range for which the cache is being flushed, the size of the
7497 memory range, and the number 3 (to flush both caches). The default
7498 depends on the target gcc was configured for, but commonly is either
7499 @samp{_flush_func} or @samp{__cpu_flush}.
7502 These options are defined by the macro
7503 @code{TARGET_SWITCHES} in the machine description. The default for the
7504 options is also defined by that macro, which enables you to change the
7507 @node i386 and x86-64 Options
7508 @subsection Intel 386 and AMD x86-64 Options
7509 @cindex i386 Options
7510 @cindex x86-64 Options
7511 @cindex Intel 386 Options
7512 @cindex AMD x86-64 Options
7514 These @samp{-m} options are defined for the i386 and x86-64 family of
7518 @item -mcpu=@var{cpu-type}
7520 Tune to @var{cpu-type} everything applicable about the generated code, except
7521 for the ABI and the set of available instructions. The choices for
7522 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7523 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7524 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7525 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp}
7526 and @samp{athlon-mp}.
7528 While picking a specific @var{cpu-type} will schedule things appropriately
7529 for that particular chip, the compiler will not generate any code that
7530 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7531 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7532 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7533 AMD chips as opposed to the Intel ones.
7535 @item -march=@var{cpu-type}
7537 Generate instructions for the machine type @var{cpu-type}. The choices
7538 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7539 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7548 @opindex mpentiumpro
7549 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7550 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7551 These synonyms are deprecated.
7553 @item -mfpmath=@var{unit}
7555 generate floating point arithmetics for selected unit @var{unit}. the choices
7560 Use the standard 387 floating point coprocessor present majority of chips and
7561 emulated otherwise. Code compiled with this option will run almost everywhere.
7562 The temporary results are computed in 80bit precesion instead of precision
7563 specified by the type resulting in slightly different results compared to most
7564 of other chips. See @option{-ffloat-store} for more detailed description.
7566 This is the default choice for i386 compiler.
7569 Use scalar floating point instructions present in the SSE instruction set.
7570 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7571 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7572 instruction set supports only single precision arithmetics, thus the double and
7573 extended precision arithmetics is still done using 387. Later version, present
7574 only in Pentium4 and the future AMD x86-64 chips supports double precision
7577 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7578 @option{-msse2} switches to enable SSE extensions and make this option
7579 effective. For x86-64 compiler, these extensions are enabled by default.
7581 The resulting code should be considerably faster in majority of cases and avoid
7582 the numerical instability problems of 387 code, but may break some existing
7583 code that expects temporaries to be 80bit.
7585 This is the default choice for x86-64 compiler.
7588 Attempt to utilize both instruction sets at once. This effectivly double the
7589 amount of available registers and on chips with separate execution units for
7590 387 and SSE the execution resources too. Use this option with care, as it is
7591 still experimental, because gcc register allocator does not model separate
7592 functional units well resulting in instable performance.
7595 @item -masm=@var{dialect}
7596 @opindex masm=@var{dialect}
7597 Output asm instructions using selected @var{dialect}. Supported choices are
7598 @samp{intel} or @samp{att} (the default one).
7603 @opindex mno-ieee-fp
7604 Control whether or not the compiler uses IEEE floating point
7605 comparisons. These handle correctly the case where the result of a
7606 comparison is unordered.
7609 @opindex msoft-float
7610 Generate output containing library calls for floating point.
7611 @strong{Warning:} the requisite libraries are not part of GCC@.
7612 Normally the facilities of the machine's usual C compiler are used, but
7613 this can't be done directly in cross-compilation. You must make your
7614 own arrangements to provide suitable library functions for
7617 On machines where a function returns floating point results in the 80387
7618 register stack, some floating point opcodes may be emitted even if
7619 @option{-msoft-float} is used.
7621 @item -mno-fp-ret-in-387
7622 @opindex mno-fp-ret-in-387
7623 Do not use the FPU registers for return values of functions.
7625 The usual calling convention has functions return values of types
7626 @code{float} and @code{double} in an FPU register, even if there
7627 is no FPU@. The idea is that the operating system should emulate
7630 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7631 in ordinary CPU registers instead.
7633 @item -mno-fancy-math-387
7634 @opindex mno-fancy-math-387
7635 Some 387 emulators do not support the @code{sin}, @code{cos} and
7636 @code{sqrt} instructions for the 387. Specify this option to avoid
7637 generating those instructions. This option is the default on FreeBSD@.
7638 As of revision 2.6.1, these instructions are not generated unless you
7639 also use the @option{-funsafe-math-optimizations} switch.
7641 @item -malign-double
7642 @itemx -mno-align-double
7643 @opindex malign-double
7644 @opindex mno-align-double
7645 Control whether GCC aligns @code{double}, @code{long double}, and
7646 @code{long long} variables on a two word boundary or a one word
7647 boundary. Aligning @code{double} variables on a two word boundary will
7648 produce code that runs somewhat faster on a @samp{Pentium} at the
7649 expense of more memory.
7651 @item -m128bit-long-double
7652 @opindex m128bit-long-double
7653 Control the size of @code{long double} type. i386 application binary interface
7654 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7655 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7656 impossible to reach with 12 byte long doubles in the array accesses.
7658 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7659 structures and arrays containing @code{long double} will change their size as
7660 well as function calling convention for function taking @code{long double}
7663 @item -m96bit-long-double
7664 @opindex m96bit-long-double
7665 Set the size of @code{long double} to 96 bits as required by the i386
7666 application binary interface. This is the default.
7669 @itemx -mno-svr3-shlib
7670 @opindex msvr3-shlib
7671 @opindex mno-svr3-shlib
7672 Control whether GCC places uninitialized local variables into the
7673 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7674 into @code{bss}. These options are meaningful only on System V Release 3.
7678 Use a different function-calling convention, in which functions that
7679 take a fixed number of arguments return with the @code{ret} @var{num}
7680 instruction, which pops their arguments while returning. This saves one
7681 instruction in the caller since there is no need to pop the arguments
7684 You can specify that an individual function is called with this calling
7685 sequence with the function attribute @samp{stdcall}. You can also
7686 override the @option{-mrtd} option by using the function attribute
7687 @samp{cdecl}. @xref{Function Attributes}.
7689 @strong{Warning:} this calling convention is incompatible with the one
7690 normally used on Unix, so you cannot use it if you need to call
7691 libraries compiled with the Unix compiler.
7693 Also, you must provide function prototypes for all functions that
7694 take variable numbers of arguments (including @code{printf});
7695 otherwise incorrect code will be generated for calls to those
7698 In addition, seriously incorrect code will result if you call a
7699 function with too many arguments. (Normally, extra arguments are
7700 harmlessly ignored.)
7702 @item -mregparm=@var{num}
7704 Control how many registers are used to pass integer arguments. By
7705 default, no registers are used to pass arguments, and at most 3
7706 registers can be used. You can control this behavior for a specific
7707 function by using the function attribute @samp{regparm}.
7708 @xref{Function Attributes}.
7710 @strong{Warning:} if you use this switch, and
7711 @var{num} is nonzero, then you must build all modules with the same
7712 value, including any libraries. This includes the system libraries and
7715 @item -mpreferred-stack-boundary=@var{num}
7716 @opindex mpreferred-stack-boundary
7717 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7718 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7719 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7720 size (@option{-Os}), in which case the default is the minimum correct
7721 alignment (4 bytes for x86, and 8 bytes for x86-64).
7723 On Pentium and PentiumPro, @code{double} and @code{long double} values
7724 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7725 suffer significant run time performance penalties. On Pentium III, the
7726 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7727 penalties if it is not 16 byte aligned.
7729 To ensure proper alignment of this values on the stack, the stack boundary
7730 must be as aligned as that required by any value stored on the stack.
7731 Further, every function must be generated such that it keeps the stack
7732 aligned. Thus calling a function compiled with a higher preferred
7733 stack boundary from a function compiled with a lower preferred stack
7734 boundary will most likely misalign the stack. It is recommended that
7735 libraries that use callbacks always use the default setting.
7737 This extra alignment does consume extra stack space, and generally
7738 increases code size. Code that is sensitive to stack space usage, such
7739 as embedded systems and operating system kernels, may want to reduce the
7740 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7756 These switches enable or disable the use of built-in functions that allow
7757 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7759 @xref{X86 Built-in Functions}, for details of the functions enabled
7760 and disabled by these switches.
7763 @itemx -mno-push-args
7765 @opindex mno-push-args
7766 Use PUSH operations to store outgoing parameters. This method is shorter
7767 and usually equally fast as method using SUB/MOV operations and is enabled
7768 by default. In some cases disabling it may improve performance because of
7769 improved scheduling and reduced dependencies.
7771 @item -maccumulate-outgoing-args
7772 @opindex maccumulate-outgoing-args
7773 If enabled, the maximum amount of space required for outgoing arguments will be
7774 computed in the function prologue. This is faster on most modern CPUs
7775 because of reduced dependencies, improved scheduling and reduced stack usage
7776 when preferred stack boundary is not equal to 2. The drawback is a notable
7777 increase in code size. This switch implies @option{-mno-push-args}.
7781 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7782 on thread-safe exception handling must compile and link all code with the
7783 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7784 @option{-D_MT}; when linking, it links in a special thread helper library
7785 @option{-lmingwthrd} which cleans up per thread exception handling data.
7787 @item -mno-align-stringops
7788 @opindex mno-align-stringops
7789 Do not align destination of inlined string operations. This switch reduces
7790 code size and improves performance in case the destination is already aligned,
7791 but gcc don't know about it.
7793 @item -minline-all-stringops
7794 @opindex minline-all-stringops
7795 By default GCC inlines string operations only when destination is known to be
7796 aligned at least to 4 byte boundary. This enables more inlining, increase code
7797 size, but may improve performance of code that depends on fast memcpy, strlen
7798 and memset for short lengths.
7800 @item -momit-leaf-frame-pointer
7801 @opindex momit-leaf-frame-pointer
7802 Don't keep the frame pointer in a register for leaf functions. This
7803 avoids the instructions to save, set up and restore frame pointers and
7804 makes an extra register available in leaf functions. The option
7805 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7806 which might make debugging harder.
7809 These @samp{-m} switches are supported in addition to the above
7810 on AMD x86-64 processors in 64-bit environments.
7817 Generate code for a 32-bit or 64-bit environment.
7818 The 32-bit environment sets int, long and pointer to 32 bits and
7819 generates code that runs on any i386 system.
7820 The 64-bit environment sets int to 32 bits and long and pointer
7821 to 64 bits and generates code for AMD's x86-64 architecture.
7824 @opindex no-red-zone
7825 Do not use a so called red zone for x86-64 code. The red zone is mandated
7826 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7827 stack pointer that will not be modified by signal or interrupt handlers
7828 and therefore can be used for temporary data without adjusting the stack
7829 pointer. The flag @option{-mno-red-zone} disables this red zone.
7833 @subsection HPPA Options
7834 @cindex HPPA Options
7836 These @samp{-m} options are defined for the HPPA family of computers:
7839 @item -march=@var{architecture-type}
7841 Generate code for the specified architecture. The choices for
7842 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7843 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7844 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7845 architecture option for your machine. Code compiled for lower numbered
7846 architectures will run on higher numbered architectures, but not the
7849 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7850 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7854 @itemx -mpa-risc-1-1
7855 @itemx -mpa-risc-2-0
7856 @opindex mpa-risc-1-0
7857 @opindex mpa-risc-1-1
7858 @opindex mpa-risc-2-0
7859 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7862 @opindex mbig-switch
7863 Generate code suitable for big switch tables. Use this option only if
7864 the assembler/linker complain about out of range branches within a switch
7867 @item -mjump-in-delay
7868 @opindex mjump-in-delay
7869 Fill delay slots of function calls with unconditional jump instructions
7870 by modifying the return pointer for the function call to be the target
7871 of the conditional jump.
7873 @item -mdisable-fpregs
7874 @opindex mdisable-fpregs
7875 Prevent floating point registers from being used in any manner. This is
7876 necessary for compiling kernels which perform lazy context switching of
7877 floating point registers. If you use this option and attempt to perform
7878 floating point operations, the compiler will abort.
7880 @item -mdisable-indexing
7881 @opindex mdisable-indexing
7882 Prevent the compiler from using indexing address modes. This avoids some
7883 rather obscure problems when compiling MIG generated code under MACH@.
7885 @item -mno-space-regs
7886 @opindex mno-space-regs
7887 Generate code that assumes the target has no space registers. This allows
7888 GCC to generate faster indirect calls and use unscaled index address modes.
7890 Such code is suitable for level 0 PA systems and kernels.
7892 @item -mfast-indirect-calls
7893 @opindex mfast-indirect-calls
7894 Generate code that assumes calls never cross space boundaries. This
7895 allows GCC to emit code which performs faster indirect calls.
7897 This option will not work in the presence of shared libraries or nested
7900 @item -mlong-load-store
7901 @opindex mlong-load-store
7902 Generate 3-instruction load and store sequences as sometimes required by
7903 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7906 @item -mportable-runtime
7907 @opindex mportable-runtime
7908 Use the portable calling conventions proposed by HP for ELF systems.
7912 Enable the use of assembler directives only GAS understands.
7914 @item -mschedule=@var{cpu-type}
7916 Schedule code according to the constraints for the machine type
7917 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7918 @samp{7100}, @samp{7100LC}, @samp{7200}, and @samp{8000}. Refer to
7919 @file{/usr/lib/sched.models} on an HP-UX system to determine the
7920 proper scheduling option for your machine.
7923 @opindex mlinker-opt
7924 Enable the optimization pass in the HPUX linker. Note this makes symbolic
7925 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
7926 in which they give bogus error messages when linking some programs.
7929 @opindex msoft-float
7930 Generate output containing library calls for floating point.
7931 @strong{Warning:} the requisite libraries are not available for all HPPA
7932 targets. Normally the facilities of the machine's usual C compiler are
7933 used, but this cannot be done directly in cross-compilation. You must make
7934 your own arrangements to provide suitable library functions for
7935 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7936 does provide software floating point support.
7938 @option{-msoft-float} changes the calling convention in the output file;
7939 therefore, it is only useful if you compile @emph{all} of a program with
7940 this option. In particular, you need to compile @file{libgcc.a}, the
7941 library that comes with GCC, with @option{-msoft-float} in order for
7945 @node Intel 960 Options
7946 @subsection Intel 960 Options
7948 These @samp{-m} options are defined for the Intel 960 implementations:
7951 @item -m@var{cpu-type}
7959 Assume the defaults for the machine type @var{cpu-type} for some of
7960 the other options, including instruction scheduling, floating point
7961 support, and addressing modes. The choices for @var{cpu-type} are
7962 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
7963 @samp{sa}, and @samp{sb}.
7970 @opindex msoft-float
7971 The @option{-mnumerics} option indicates that the processor does support
7972 floating-point instructions. The @option{-msoft-float} option indicates
7973 that floating-point support should not be assumed.
7975 @item -mleaf-procedures
7976 @itemx -mno-leaf-procedures
7977 @opindex mleaf-procedures
7978 @opindex mno-leaf-procedures
7979 Do (or do not) attempt to alter leaf procedures to be callable with the
7980 @code{bal} instruction as well as @code{call}. This will result in more
7981 efficient code for explicit calls when the @code{bal} instruction can be
7982 substituted by the assembler or linker, but less efficient code in other
7983 cases, such as calls via function pointers, or using a linker that doesn't
7984 support this optimization.
7987 @itemx -mno-tail-call
7989 @opindex mno-tail-call
7990 Do (or do not) make additional attempts (beyond those of the
7991 machine-independent portions of the compiler) to optimize tail-recursive
7992 calls into branches. You may not want to do this because the detection of
7993 cases where this is not valid is not totally complete. The default is
7994 @option{-mno-tail-call}.
7996 @item -mcomplex-addr
7997 @itemx -mno-complex-addr
7998 @opindex mcomplex-addr
7999 @opindex mno-complex-addr
8000 Assume (or do not assume) that the use of a complex addressing mode is a
8001 win on this implementation of the i960. Complex addressing modes may not
8002 be worthwhile on the K-series, but they definitely are on the C-series.
8003 The default is currently @option{-mcomplex-addr} for all processors except
8007 @itemx -mno-code-align
8008 @opindex mcode-align
8009 @opindex mno-code-align
8010 Align code to 8-byte boundaries for faster fetching (or don't bother).
8011 Currently turned on by default for C-series implementations only.
8014 @item -mclean-linkage
8015 @itemx -mno-clean-linkage
8016 @opindex mclean-linkage
8017 @opindex mno-clean-linkage
8018 These options are not fully implemented.
8022 @itemx -mic2.0-compat
8023 @itemx -mic3.0-compat
8025 @opindex mic2.0-compat
8026 @opindex mic3.0-compat
8027 Enable compatibility with iC960 v2.0 or v3.0.
8031 @opindex masm-compat
8033 Enable compatibility with the iC960 assembler.
8035 @item -mstrict-align
8036 @itemx -mno-strict-align
8037 @opindex mstrict-align
8038 @opindex mno-strict-align
8039 Do not permit (do permit) unaligned accesses.
8043 Enable structure-alignment compatibility with Intel's gcc release version
8044 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8046 @item -mlong-double-64
8047 @opindex mlong-double-64
8048 Implement type @samp{long double} as 64-bit floating point numbers.
8049 Without the option @samp{long double} is implemented by 80-bit
8050 floating point numbers. The only reason we have it because there is
8051 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8052 is only useful for people using soft-float targets. Otherwise, we
8053 should recommend against use of it.
8057 @node DEC Alpha Options
8058 @subsection DEC Alpha Options
8060 These @samp{-m} options are defined for the DEC Alpha implementations:
8063 @item -mno-soft-float
8065 @opindex mno-soft-float
8066 @opindex msoft-float
8067 Use (do not use) the hardware floating-point instructions for
8068 floating-point operations. When @option{-msoft-float} is specified,
8069 functions in @file{libgcc.a} will be used to perform floating-point
8070 operations. Unless they are replaced by routines that emulate the
8071 floating-point operations, or compiled in such a way as to call such
8072 emulations routines, these routines will issue floating-point
8073 operations. If you are compiling for an Alpha without floating-point
8074 operations, you must ensure that the library is built so as not to call
8077 Note that Alpha implementations without floating-point operations are
8078 required to have floating-point registers.
8083 @opindex mno-fp-regs
8084 Generate code that uses (does not use) the floating-point register set.
8085 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8086 register set is not used, floating point operands are passed in integer
8087 registers as if they were integers and floating-point results are passed
8088 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8089 so any function with a floating-point argument or return value called by code
8090 compiled with @option{-mno-fp-regs} must also be compiled with that
8093 A typical use of this option is building a kernel that does not use,
8094 and hence need not save and restore, any floating-point registers.
8098 The Alpha architecture implements floating-point hardware optimized for
8099 maximum performance. It is mostly compliant with the IEEE floating
8100 point standard. However, for full compliance, software assistance is
8101 required. This option generates code fully IEEE compliant code
8102 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8103 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8104 defined during compilation. The resulting code is less efficient but is
8105 able to correctly support denormalized numbers and exceptional IEEE
8106 values such as not-a-number and plus/minus infinity. Other Alpha
8107 compilers call this option @option{-ieee_with_no_inexact}.
8109 @item -mieee-with-inexact
8110 @opindex mieee-with-inexact
8111 This is like @option{-mieee} except the generated code also maintains
8112 the IEEE @var{inexact-flag}. Turning on this option causes the
8113 generated code to implement fully-compliant IEEE math. In addition to
8114 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8115 macro. On some Alpha implementations the resulting code may execute
8116 significantly slower than the code generated by default. Since there is
8117 very little code that depends on the @var{inexact-flag}, you should
8118 normally not specify this option. Other Alpha compilers call this
8119 option @option{-ieee_with_inexact}.
8121 @item -mfp-trap-mode=@var{trap-mode}
8122 @opindex mfp-trap-mode
8123 This option controls what floating-point related traps are enabled.
8124 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8125 The trap mode can be set to one of four values:
8129 This is the default (normal) setting. The only traps that are enabled
8130 are the ones that cannot be disabled in software (e.g., division by zero
8134 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8138 Like @samp{su}, but the instructions are marked to be safe for software
8139 completion (see Alpha architecture manual for details).
8142 Like @samp{su}, but inexact traps are enabled as well.
8145 @item -mfp-rounding-mode=@var{rounding-mode}
8146 @opindex mfp-rounding-mode
8147 Selects the IEEE rounding mode. Other Alpha compilers call this option
8148 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8153 Normal IEEE rounding mode. Floating point numbers are rounded towards
8154 the nearest machine number or towards the even machine number in case
8158 Round towards minus infinity.
8161 Chopped rounding mode. Floating point numbers are rounded towards zero.
8164 Dynamic rounding mode. A field in the floating point control register
8165 (@var{fpcr}, see Alpha architecture reference manual) controls the
8166 rounding mode in effect. The C library initializes this register for
8167 rounding towards plus infinity. Thus, unless your program modifies the
8168 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8171 @item -mtrap-precision=@var{trap-precision}
8172 @opindex mtrap-precision
8173 In the Alpha architecture, floating point traps are imprecise. This
8174 means without software assistance it is impossible to recover from a
8175 floating trap and program execution normally needs to be terminated.
8176 GCC can generate code that can assist operating system trap handlers
8177 in determining the exact location that caused a floating point trap.
8178 Depending on the requirements of an application, different levels of
8179 precisions can be selected:
8183 Program precision. This option is the default and means a trap handler
8184 can only identify which program caused a floating point exception.
8187 Function precision. The trap handler can determine the function that
8188 caused a floating point exception.
8191 Instruction precision. The trap handler can determine the exact
8192 instruction that caused a floating point exception.
8195 Other Alpha compilers provide the equivalent options called
8196 @option{-scope_safe} and @option{-resumption_safe}.
8198 @item -mieee-conformant
8199 @opindex mieee-conformant
8200 This option marks the generated code as IEEE conformant. You must not
8201 use this option unless you also specify @option{-mtrap-precision=i} and either
8202 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8203 is to emit the line @samp{.eflag 48} in the function prologue of the
8204 generated assembly file. Under DEC Unix, this has the effect that
8205 IEEE-conformant math library routines will be linked in.
8207 @item -mbuild-constants
8208 @opindex mbuild-constants
8209 Normally GCC examines a 32- or 64-bit integer constant to
8210 see if it can construct it from smaller constants in two or three
8211 instructions. If it cannot, it will output the constant as a literal and
8212 generate code to load it from the data segment at runtime.
8214 Use this option to require GCC to construct @emph{all} integer constants
8215 using code, even if it takes more instructions (the maximum is six).
8217 You would typically use this option to build a shared library dynamic
8218 loader. Itself a shared library, it must relocate itself in memory
8219 before it can find the variables and constants in its own data segment.
8225 Select whether to generate code to be assembled by the vendor-supplied
8226 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8244 Indicate whether GCC should generate code to use the optional BWX,
8245 CIX, FIX and MAX instruction sets. The default is to use the instruction
8246 sets supported by the CPU type specified via @option{-mcpu=} option or that
8247 of the CPU on which GCC was built if none was specified.
8252 @opindex mfloat-ieee
8253 Generate code that uses (does not use) VAX F and G floating point
8254 arithmetic instead of IEEE single and double precision.
8256 @item -mexplicit-relocs
8257 @itemx -mno-explicit-relocs
8258 @opindex mexplicit-relocs
8259 @opindex mno-explicit-relocs
8260 Older Alpha assemblers provided no way to generate symbol relocations
8261 except via assembler macros. Use of these macros does not allow
8262 optimial instruction scheduling. GNU binutils as of version 2.12
8263 supports a new syntax that allows the compiler to explicitly mark
8264 which relocations should apply to which instructions. This option
8265 is mostly useful for debugging, as GCC detects the capabilities of
8266 the assembler when it is built and sets the default accordingly.
8270 @opindex msmall-data
8271 @opindex mlarge-data
8272 When @option{-mexplicit-relocs} is in effect, static data is
8273 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8274 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8275 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8276 16-bit relocations off of the @code{$gp} register. This limits the
8277 size of the small data area to 64KB, but allows the variables to be
8278 directly accessed via a single instruction.
8280 The default is @option{-mlarge-data}. With this option the data area
8281 is limited to just below 2GB. Programs that require more than 2GB of
8282 data must use @code{malloc} or @code{mmap} to allocate the data in the
8283 heap instead of in the program's data segment.
8285 When generating code for shared libraries, @option{-fpic} implies
8286 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8288 @item -mcpu=@var{cpu_type}
8290 Set the instruction set and instruction scheduling parameters for
8291 machine type @var{cpu_type}. You can specify either the @samp{EV}
8292 style name or the corresponding chip number. GCC supports scheduling
8293 parameters for the EV4, EV5 and EV6 family of processors and will
8294 choose the default values for the instruction set from the processor
8295 you specify. If you do not specify a processor type, GCC will default
8296 to the processor on which the compiler was built.
8298 Supported values for @var{cpu_type} are
8304 Schedules as an EV4 and has no instruction set extensions.
8308 Schedules as an EV5 and has no instruction set extensions.
8312 Schedules as an EV5 and supports the BWX extension.
8317 Schedules as an EV5 and supports the BWX and MAX extensions.
8321 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8325 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8328 @item -mtune=@var{cpu_type}
8330 Set only the instruction scheduling parameters for machine type
8331 @var{cpu_type}. The instruction set is not changed.
8333 @item -mmemory-latency=@var{time}
8334 @opindex mmemory-latency
8335 Sets the latency the scheduler should assume for typical memory
8336 references as seen by the application. This number is highly
8337 dependent on the memory access patterns used by the application
8338 and the size of the external cache on the machine.
8340 Valid options for @var{time} are
8344 A decimal number representing clock cycles.
8350 The compiler contains estimates of the number of clock cycles for
8351 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8352 (also called Dcache, Scache, and Bcache), as well as to main memory.
8353 Note that L3 is only valid for EV5.
8358 @node DEC Alpha/VMS Options
8359 @subsection DEC Alpha/VMS Options
8361 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8364 @item -mvms-return-codes
8365 @opindex mvms-return-codes
8366 Return VMS condition codes from main. The default is to return POSIX
8367 style condition (e.g.@ error) codes.
8370 @node Clipper Options
8371 @subsection Clipper Options
8373 These @samp{-m} options are defined for the Clipper implementations:
8378 Produce code for a C300 Clipper processor. This is the default.
8382 Produce code for a C400 Clipper processor, i.e.@: use floating point
8386 @node H8/300 Options
8387 @subsection H8/300 Options
8389 These @samp{-m} options are defined for the H8/300 implementations:
8394 Shorten some address references at link time, when possible; uses the
8395 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8396 ld.info, Using ld}, for a fuller description.
8400 Generate code for the H8/300H@.
8404 Generate code for the H8/S@.
8408 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8412 Make @code{int} data 32 bits by default.
8416 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8417 The default for the H8/300H and H8/S is to align longs and floats on 4
8419 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8420 This option has no effect on the H8/300.
8424 @subsection SH Options
8426 These @samp{-m} options are defined for the SH implementations:
8431 Generate code for the SH1.
8435 Generate code for the SH2.
8439 Generate code for the SH3.
8443 Generate code for the SH3e.
8447 Generate code for the SH4 without a floating-point unit.
8449 @item -m4-single-only
8450 @opindex m4-single-only
8451 Generate code for the SH4 with a floating-point unit that only
8452 supports single-precision arithmetic.
8456 Generate code for the SH4 assuming the floating-point unit is in
8457 single-precision mode by default.
8461 Generate code for the SH4.
8465 Compile code for the processor in big endian mode.
8469 Compile code for the processor in little endian mode.
8473 Align doubles at 64-bit boundaries. Note that this changes the calling
8474 conventions, and thus some functions from the standard C library will
8475 not work unless you recompile it first with @option{-mdalign}.
8479 Shorten some address references at link time, when possible; uses the
8480 linker option @option{-relax}.
8484 Use 32-bit offsets in @code{switch} tables. The default is to use
8489 Enable the use of the instruction @code{fmovd}.
8493 Comply with the calling conventions defined by Hitachi.
8497 Mark the @code{MAC} register as call-clobbered, even if
8498 @option{-mhitachi} is given.
8502 Increase IEEE-compliance of floating-point code.
8506 Dump instruction size and location in the assembly code.
8510 This option is deprecated. It pads structures to multiple of 4 bytes,
8511 which is incompatible with the SH ABI@.
8515 Optimize for space instead of speed. Implied by @option{-Os}.
8519 When generating position-independent code, emit function calls using
8520 the Global Offset Table instead of the Procedure Linkage Table.
8524 Generate a library function call to invalidate instruction cache
8525 entries, after fixing up a trampoline. This library function call
8526 doesn't assume it can write to the whole memory address space. This
8527 is the default when the target is @code{sh-*-linux*}.
8530 @node System V Options
8531 @subsection Options for System V
8533 These additional options are available on System V Release 4 for
8534 compatibility with other compilers on those systems:
8539 Create a shared object.
8540 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8544 Identify the versions of each tool used by the compiler, in a
8545 @code{.ident} assembler directive in the output.
8549 Refrain from adding @code{.ident} directives to the output file (this is
8552 @item -YP,@var{dirs}
8554 Search the directories @var{dirs}, and no others, for libraries
8555 specified with @option{-l}.
8559 Look in the directory @var{dir} to find the M4 preprocessor.
8560 The assembler uses this option.
8561 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8562 @c the generic assembler that comes with Solaris takes just -Ym.
8565 @node TMS320C3x/C4x Options
8566 @subsection TMS320C3x/C4x Options
8567 @cindex TMS320C3x/C4x Options
8569 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8573 @item -mcpu=@var{cpu_type}
8575 Set the instruction set, register set, and instruction scheduling
8576 parameters for machine type @var{cpu_type}. Supported values for
8577 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8578 @samp{c44}. The default is @samp{c40} to generate code for the
8583 @itemx -msmall-memory
8585 @opindex mbig-memory
8587 @opindex msmall-memory
8589 Generates code for the big or small memory model. The small memory
8590 model assumed that all data fits into one 64K word page. At run-time
8591 the data page (DP) register must be set to point to the 64K page
8592 containing the .bss and .data program sections. The big memory model is
8593 the default and requires reloading of the DP register for every direct
8600 Allow (disallow) allocation of general integer operands into the block
8607 Enable (disable) generation of code using decrement and branch,
8608 DBcond(D), instructions. This is enabled by default for the C4x. To be
8609 on the safe side, this is disabled for the C3x, since the maximum
8610 iteration count on the C3x is @math{2^23 + 1} (but who iterates loops more than
8611 @math{2^23} times on the C3x?). Note that GCC will try to reverse a loop so
8612 that it can utilise the decrement and branch instruction, but will give
8613 up if there is more than one memory reference in the loop. Thus a loop
8614 where the loop counter is decremented can generate slightly more
8615 efficient code, in cases where the RPTB instruction cannot be utilised.
8617 @item -mdp-isr-reload
8619 @opindex mdp-isr-reload
8621 Force the DP register to be saved on entry to an interrupt service
8622 routine (ISR), reloaded to point to the data section, and restored on
8623 exit from the ISR@. This should not be required unless someone has
8624 violated the small memory model by modifying the DP register, say within
8631 For the C3x use the 24-bit MPYI instruction for integer multiplies
8632 instead of a library call to guarantee 32-bit results. Note that if one
8633 of the operands is a constant, then the multiplication will be performed
8634 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8635 then squaring operations are performed inline instead of a library call.
8638 @itemx -mno-fast-fix
8640 @opindex mno-fast-fix
8641 The C3x/C4x FIX instruction to convert a floating point value to an
8642 integer value chooses the nearest integer less than or equal to the
8643 floating point value rather than to the nearest integer. Thus if the
8644 floating point number is negative, the result will be incorrectly
8645 truncated an additional code is necessary to detect and correct this
8646 case. This option can be used to disable generation of the additional
8647 code required to correct the result.
8653 Enable (disable) generation of repeat block sequences using the RPTB
8654 instruction for zero overhead looping. The RPTB construct is only used
8655 for innermost loops that do not call functions or jump across the loop
8656 boundaries. There is no advantage having nested RPTB loops due to the
8657 overhead required to save and restore the RC, RS, and RE registers.
8658 This is enabled by default with @option{-O2}.
8660 @item -mrpts=@var{count}
8664 Enable (disable) the use of the single instruction repeat instruction
8665 RPTS@. If a repeat block contains a single instruction, and the loop
8666 count can be guaranteed to be less than the value @var{count}, GCC will
8667 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8668 then a RPTS will be emitted even if the loop count cannot be determined
8669 at compile time. Note that the repeated instruction following RPTS does
8670 not have to be reloaded from memory each iteration, thus freeing up the
8671 CPU buses for operands. However, since interrupts are blocked by this
8672 instruction, it is disabled by default.
8674 @item -mloop-unsigned
8675 @itemx -mno-loop-unsigned
8676 @opindex mloop-unsigned
8677 @opindex mno-loop-unsigned
8678 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8679 is @math{2^31 + 1} since these instructions test if the iteration count is
8680 negative to terminate the loop. If the iteration count is unsigned
8681 there is a possibility than the @math{2^31 + 1} maximum iteration count may be
8682 exceeded. This switch allows an unsigned iteration count.
8686 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8687 with. This also enforces compatibility with the API employed by the TI
8688 C3x C compiler. For example, long doubles are passed as structures
8689 rather than in floating point registers.
8695 Generate code that uses registers (stack) for passing arguments to functions.
8696 By default, arguments are passed in registers where possible rather
8697 than by pushing arguments on to the stack.
8699 @item -mparallel-insns
8700 @itemx -mno-parallel-insns
8701 @opindex mparallel-insns
8702 @opindex mno-parallel-insns
8703 Allow the generation of parallel instructions. This is enabled by
8704 default with @option{-O2}.
8706 @item -mparallel-mpy
8707 @itemx -mno-parallel-mpy
8708 @opindex mparallel-mpy
8709 @opindex mno-parallel-mpy
8710 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8711 provided @option{-mparallel-insns} is also specified. These instructions have
8712 tight register constraints which can pessimize the code generation
8718 @subsection V850 Options
8719 @cindex V850 Options
8721 These @samp{-m} options are defined for V850 implementations:
8725 @itemx -mno-long-calls
8726 @opindex mlong-calls
8727 @opindex mno-long-calls
8728 Treat all calls as being far away (near). If calls are assumed to be
8729 far away, the compiler will always load the functions address up into a
8730 register, and call indirect through the pointer.
8736 Do not optimize (do optimize) basic blocks that use the same index
8737 pointer 4 or more times to copy pointer into the @code{ep} register, and
8738 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8739 option is on by default if you optimize.
8741 @item -mno-prolog-function
8742 @itemx -mprolog-function
8743 @opindex mno-prolog-function
8744 @opindex mprolog-function
8745 Do not use (do use) external functions to save and restore registers at
8746 the prolog and epilog of a function. The external functions are slower,
8747 but use less code space if more than one function saves the same number
8748 of registers. The @option{-mprolog-function} option is on by default if
8753 Try to make the code as small as possible. At present, this just turns
8754 on the @option{-mep} and @option{-mprolog-function} options.
8758 Put static or global variables whose size is @var{n} bytes or less into
8759 the tiny data area that register @code{ep} points to. The tiny data
8760 area can hold up to 256 bytes in total (128 bytes for byte references).
8764 Put static or global variables whose size is @var{n} bytes or less into
8765 the small data area that register @code{gp} points to. The small data
8766 area can hold up to 64 kilobytes.
8770 Put static or global variables whose size is @var{n} bytes or less into
8771 the first 32 kilobytes of memory.
8775 Specify that the target processor is the V850.
8778 @opindex mbig-switch
8779 Generate code suitable for big switch tables. Use this option only if
8780 the assembler/linker complain about out of range branches within a switch
8785 @subsection ARC Options
8788 These options are defined for ARC implementations:
8793 Compile code for little endian mode. This is the default.
8797 Compile code for big endian mode.
8800 @opindex mmangle-cpu
8801 Prepend the name of the cpu to all public symbol names.
8802 In multiple-processor systems, there are many ARC variants with different
8803 instruction and register set characteristics. This flag prevents code
8804 compiled for one cpu to be linked with code compiled for another.
8805 No facility exists for handling variants that are ``almost identical''.
8806 This is an all or nothing option.
8808 @item -mcpu=@var{cpu}
8810 Compile code for ARC variant @var{cpu}.
8811 Which variants are supported depend on the configuration.
8812 All variants support @option{-mcpu=base}, this is the default.
8814 @item -mtext=@var{text-section}
8815 @itemx -mdata=@var{data-section}
8816 @itemx -mrodata=@var{readonly-data-section}
8820 Put functions, data, and readonly data in @var{text-section},
8821 @var{data-section}, and @var{readonly-data-section} respectively
8822 by default. This can be overridden with the @code{section} attribute.
8823 @xref{Variable Attributes}.
8828 @subsection NS32K Options
8829 @cindex NS32K options
8831 These are the @samp{-m} options defined for the 32000 series. The default
8832 values for these options depends on which style of 32000 was selected when
8833 the compiler was configured; the defaults for the most common choices are
8841 Generate output for a 32032. This is the default
8842 when the compiler is configured for 32032 and 32016 based systems.
8848 Generate output for a 32332. This is the default
8849 when the compiler is configured for 32332-based systems.
8855 Generate output for a 32532. This is the default
8856 when the compiler is configured for 32532-based systems.
8860 Generate output containing 32081 instructions for floating point.
8861 This is the default for all systems.
8865 Generate output containing 32381 instructions for floating point. This
8866 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8867 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8871 Try and generate multiply-add floating point instructions @code{polyF}
8872 and @code{dotF}. This option is only available if the @option{-m32381}
8873 option is in effect. Using these instructions requires changes to
8874 register allocation which generally has a negative impact on
8875 performance. This option should only be enabled when compiling code
8876 particularly likely to make heavy use of multiply-add instructions.
8879 @opindex mnomulti-add
8880 Do not try and generate multiply-add floating point instructions
8881 @code{polyF} and @code{dotF}. This is the default on all platforms.
8884 @opindex msoft-float
8885 Generate output containing library calls for floating point.
8886 @strong{Warning:} the requisite libraries may not be available.
8889 @opindex mnobitfield
8890 Do not use the bit-field instructions. On some machines it is faster to
8891 use shifting and masking operations. This is the default for the pc532.
8895 Do use the bit-field instructions. This is the default for all platforms
8900 Use a different function-calling convention, in which functions
8901 that take a fixed number of arguments return pop their
8902 arguments on return with the @code{ret} instruction.
8904 This calling convention is incompatible with the one normally
8905 used on Unix, so you cannot use it if you need to call libraries
8906 compiled with the Unix compiler.
8908 Also, you must provide function prototypes for all functions that
8909 take variable numbers of arguments (including @code{printf});
8910 otherwise incorrect code will be generated for calls to those
8913 In addition, seriously incorrect code will result if you call a
8914 function with too many arguments. (Normally, extra arguments are
8915 harmlessly ignored.)
8917 This option takes its name from the 680x0 @code{rtd} instruction.
8922 Use a different function-calling convention where the first two arguments
8923 are passed in registers.
8925 This calling convention is incompatible with the one normally
8926 used on Unix, so you cannot use it if you need to call libraries
8927 compiled with the Unix compiler.
8930 @opindex mnoregparam
8931 Do not pass any arguments in registers. This is the default for all
8936 It is OK to use the sb as an index register which is always loaded with
8937 zero. This is the default for the pc532-netbsd target.
8941 The sb register is not available for use or has not been initialized to
8942 zero by the run time system. This is the default for all targets except
8943 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
8944 @option{-fpic} is set.
8948 Many ns32000 series addressing modes use displacements of up to 512MB@.
8949 If an address is above 512MB then displacements from zero can not be used.
8950 This option causes code to be generated which can be loaded above 512MB@.
8951 This may be useful for operating systems or ROM code.
8955 Assume code will be loaded in the first 512MB of virtual address space.
8956 This is the default for all platforms.
8962 @subsection AVR Options
8965 These options are defined for AVR implementations:
8968 @item -mmcu=@var{mcu}
8970 Specify ATMEL AVR instruction set or MCU type.
8972 Instruction set avr1 is for the minimal AVR core, not supported by the C
8973 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8974 attiny11, attiny12, attiny15, attiny28).
8976 Instruction set avr2 (default) is for the classic AVR core with up to
8977 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8978 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8979 at90c8534, at90s8535).
8981 Instruction set avr3 is for the classic AVR core with up to 128K program
8982 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8984 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8985 memory space (MCU types: atmega8, atmega83, atmega85).
8987 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8988 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8989 atmega64, atmega128, at43usb355, at94k).
8993 Output instruction sizes to the asm file.
8995 @item -minit-stack=@var{N}
8996 @opindex minit-stack
8997 Specify the initial stack address, which may be a symbol or numeric value,
8998 @samp{__stack} is the default.
9000 @item -mno-interrupts
9001 @opindex mno-interrupts
9002 Generated code is not compatible with hardware interrupts.
9003 Code size will be smaller.
9005 @item -mcall-prologues
9006 @opindex mcall-prologues
9007 Functions prologues/epilogues expanded as call to appropriate
9008 subroutines. Code size will be smaller.
9010 @item -mno-tablejump
9011 @opindex mno-tablejump
9012 Do not generate tablejump insns which sometimes increase code size.
9015 @opindex mtiny-stack
9016 Change only the low 8 bits of the stack pointer.
9020 @subsection MCore Options
9021 @cindex MCore options
9023 These are the @samp{-m} options defined for the Motorola M*Core
9033 @opindex mno-hardlit
9034 Inline constants into the code stream if it can be done in two
9035 instructions or less.
9043 Use the divide instruction. (Enabled by default).
9045 @item -mrelax-immediate
9046 @itemx -mrelax-immediate
9047 @itemx -mno-relax-immediate
9048 @opindex mrelax-immediate
9049 @opindex mrelax-immediate
9050 @opindex mno-relax-immediate
9051 Allow arbitrary sized immediates in bit operations.
9053 @item -mwide-bitfields
9054 @itemx -mwide-bitfields
9055 @itemx -mno-wide-bitfields
9056 @opindex mwide-bitfields
9057 @opindex mwide-bitfields
9058 @opindex mno-wide-bitfields
9059 Always treat bit-fields as int-sized.
9061 @item -m4byte-functions
9062 @itemx -m4byte-functions
9063 @itemx -mno-4byte-functions
9064 @opindex m4byte-functions
9065 @opindex m4byte-functions
9066 @opindex mno-4byte-functions
9067 Force all functions to be aligned to a four byte boundary.
9069 @item -mcallgraph-data
9070 @itemx -mcallgraph-data
9071 @itemx -mno-callgraph-data
9072 @opindex mcallgraph-data
9073 @opindex mcallgraph-data
9074 @opindex mno-callgraph-data
9075 Emit callgraph information.
9079 @itemx -mno-slow-bytes
9080 @opindex mslow-bytes
9081 @opindex mslow-bytes
9082 @opindex mno-slow-bytes
9083 Prefer word access when reading byte quantities.
9085 @item -mlittle-endian
9086 @itemx -mlittle-endian
9088 @opindex mlittle-endian
9089 @opindex mlittle-endian
9090 @opindex mbig-endian
9091 Generate code for a little endian target.
9099 Generate code for the 210 processor.
9103 @subsection IA-64 Options
9104 @cindex IA-64 Options
9106 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9110 @opindex mbig-endian
9111 Generate code for a big endian target. This is the default for HPUX@.
9113 @item -mlittle-endian
9114 @opindex mlittle-endian
9115 Generate code for a little endian target. This is the default for AIX5
9122 Generate (or don't) code for the GNU assembler. This is the default.
9123 @c Also, this is the default if the configure option @option{--with-gnu-as}
9130 Generate (or don't) code for the GNU linker. This is the default.
9131 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9136 Generate code that does not use a global pointer register. The result
9137 is not position independent code, and violates the IA-64 ABI@.
9139 @item -mvolatile-asm-stop
9140 @itemx -mno-volatile-asm-stop
9141 @opindex mvolatile-asm-stop
9142 @opindex mno-volatile-asm-stop
9143 Generate (or don't) a stop bit immediately before and after volatile asm
9148 Generate code that works around Itanium B step errata.
9150 @item -mregister-names
9151 @itemx -mno-register-names
9152 @opindex mregister-names
9153 @opindex mno-register-names
9154 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9155 the stacked registers. This may make assembler output more readable.
9161 Disable (or enable) optimizations that use the small data section. This may
9162 be useful for working around optimizer bugs.
9165 @opindex mconstant-gp
9166 Generate code that uses a single constant global pointer value. This is
9167 useful when compiling kernel code.
9171 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9172 This is useful when compiling firmware code.
9174 @item -minline-divide-min-latency
9175 @opindex minline-divide-min-latency
9176 Generate code for inline divides using the minimum latency algorithm.
9178 @item -minline-divide-max-throughput
9179 @opindex minline-divide-max-throughput
9180 Generate code for inline divides using the maximum throughput algorithm.
9182 @item -mno-dwarf2-asm
9184 @opindex mno-dwarf2-asm
9185 @opindex mdwarf2-asm
9186 Don't (or do) generate assembler code for the DWARF2 line number debugging
9187 info. This may be useful when not using the GNU assembler.
9189 @item -mfixed-range=@var{register-range}
9190 @opindex mfixed-range
9191 Generate code treating the given register range as fixed registers.
9192 A fixed register is one that the register allocator can not use. This is
9193 useful when compiling kernel code. A register range is specified as
9194 two registers separated by a dash. Multiple register ranges can be
9195 specified separated by a comma.
9199 @subsection D30V Options
9200 @cindex D30V Options
9202 These @samp{-m} options are defined for D30V implementations:
9207 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
9208 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9209 memory, which starts at location @code{0x80000000}.
9213 Same as the @option{-mextmem} switch.
9217 Link the @samp{.text} section into onchip text memory, which starts at
9218 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9219 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9220 into onchip data memory, which starts at location @code{0x20000000}.
9222 @item -mno-asm-optimize
9223 @itemx -masm-optimize
9224 @opindex mno-asm-optimize
9225 @opindex masm-optimize
9226 Disable (enable) passing @option{-O} to the assembler when optimizing.
9227 The assembler uses the @option{-O} option to automatically parallelize
9228 adjacent short instructions where possible.
9230 @item -mbranch-cost=@var{n}
9231 @opindex mbranch-cost
9232 Increase the internal costs of branches to @var{n}. Higher costs means
9233 that the compiler will issue more instructions to avoid doing a branch.
9236 @item -mcond-exec=@var{n}
9238 Specify the maximum number of conditionally executed instructions that
9239 replace a branch. The default is 4.
9242 @node S/390 and zSeries Options
9243 @subsection S/390 and zSeries Options
9244 @cindex S/390 and zSeries Options
9246 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9251 @opindex mhard-float
9252 @opindex msoft-float
9253 Use (do not use) the hardware floating-point instructions and registers
9254 for floating-point operations. When @option{-msoft-float} is specified,
9255 functions in @file{libgcc.a} will be used to perform floating-point
9256 operations. When @option{-mhard-float} is specified, the compiler
9257 generates IEEE floating-point instructions. This is the default.
9260 @itemx -mno-backchain
9262 @opindex mno-backchain
9263 Generate (or do not generate) code which maintains an explicit
9264 backchain within the stack frame that points to the caller's frame.
9265 This is currently needed to allow debugging. The default is to
9266 generate the backchain.
9269 @itemx -mno-small-exec
9270 @opindex msmall-exec
9271 @opindex mno-small-exec
9272 Generate (or do not generate) code using the @code{bras} instruction
9273 to do subroutine calls.
9274 This only works reliably if the total executable size does not
9275 exceed 64k. The default is to use the @code{basr} instruction instead,
9276 which does not have this limitation.
9282 When @option{-m31} is specified, generate code compliant to the
9283 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9284 code compliant to the Linux for zSeries ABI@. This allows GCC in
9285 particular to generate 64-bit instructions. For the @samp{s390}
9286 targets, the default is @option{-m31}, while the @samp{s390x}
9287 targets default to @option{-m64}.
9293 Generate (or do not generate) code using the @code{mvcle} instruction
9294 to perform block moves. When @option{-mno-mvcle} is specifed,
9295 use a @code{mvc} loop instead. This is the default.
9301 Print (or do not print) additional debug information when compiling.
9302 The default is to not print debug information.
9307 @subsection CRIS Options
9308 @cindex CRIS Options
9310 These options are defined specifically for the CRIS ports.
9313 @item -march=@var{architecture-type}
9314 @itemx -mcpu=@var{architecture-type}
9317 Generate code for the specified architecture. The choices for
9318 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9319 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9320 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9323 @item -mtune=@var{architecture-type}
9325 Tune to @var{architecture-type} everything applicable about the generated
9326 code, except for the ABI and the set of available instructions. The
9327 choices for @var{architecture-type} are the same as for
9328 @option{-march=@var{architecture-type}}.
9330 @item -mmax-stack-frame=@var{n}
9331 @opindex mmax-stack-frame
9332 Warn when the stack frame of a function exceeds @var{n} bytes.
9334 @item -melinux-stacksize=@var{n}
9335 @opindex melinux-stacksize
9336 Only available with the @samp{cris-axis-aout} target. Arranges for
9337 indications in the program to the kernel loader that the stack of the
9338 program should be set to @var{n} bytes.
9344 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9345 @option{-march=v3} and @option{-march=v8} respectively.
9349 Enable CRIS-specific verbose debug-related information in the assembly
9350 code. This option also has the effect to turn off the @samp{#NO_APP}
9351 formatted-code indicator to the assembler at the beginning of the
9356 Do not use condition-code results from previous instruction; always emit
9357 compare and test instructions before use of condition codes.
9359 @item -mno-side-effects
9360 @opindex mno-side-effects
9361 Do not emit instructions with side-effects in addressing modes other than
9365 @itemx -mno-stack-align
9367 @itemx -mno-data-align
9368 @itemx -mconst-align
9369 @itemx -mno-const-align
9370 @opindex mstack-align
9371 @opindex mno-stack-align
9372 @opindex mdata-align
9373 @opindex mno-data-align
9374 @opindex mconst-align
9375 @opindex mno-const-align
9376 These options (no-options) arranges (eliminate arrangements) for the
9377 stack-frame, individual data and constants to be aligned for the maximum
9378 single data access size for the chosen CPU model. The default is to
9379 arrange for 32-bit alignment. ABI details such as structure layout are
9380 not affected by these options.
9388 Similar to the stack- data- and const-align options above, these options
9389 arrange for stack-frame, writable data and constants to all be 32-bit,
9390 16-bit or 8-bit aligned. The default is 32-bit alignment.
9392 @item -mno-prologue-epilogue
9393 @itemx -mprologue-epilogue
9394 @opindex mno-prologue-epilogue
9395 @opindex mprologue-epilogue
9396 With @option{-mno-prologue-epilogue}, the normal function prologue and
9397 epilogue that sets up the stack-frame are omitted and no return
9398 instructions or return sequences are generated in the code. Use this
9399 option only together with visual inspection of the compiled code: no
9400 warnings or errors are generated when call-saved registers must be saved,
9401 or storage for local variable needs to be allocated.
9407 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9408 instruction sequences that load addresses for functions from the PLT part
9409 of the GOT rather than (traditional on other architectures) calls to the
9410 PLT. The default is @option{-mgotplt}.
9414 Legacy no-op option only recognized with the cris-axis-aout target.
9418 Legacy no-op option only recognized with the cris-axis-elf and
9419 cris-axis-linux-gnu targets.
9423 Only recognized with the cris-axis-aout target, where it selects a
9424 GNU/linux-like multilib, include files and instruction set for
9429 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9433 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9434 to link with input-output functions from a simulator library. Code,
9435 initialized data and zero-initialized data are allocated consecutively.
9439 Like @option{-sim}, but pass linker options to locate initialized data at
9440 0x40000000 and zero-initialized data at 0x80000000.
9444 @subsection MMIX Options
9445 @cindex MMIX Options
9447 These options are defined for the MMIX:
9451 @itemx -mno-libfuncs
9453 @opindex mno-libfuncs
9454 Specify that intrinsic library functions are being compiled, passing all
9455 values in registers, no matter the size.
9460 @opindex mno-epsilon
9461 Generate floating-point comparison instructions that compare with respect
9462 to the @code{rE} epsilon register.
9464 @item -mabi=mmixware
9466 @opindex mabi-mmixware
9468 Generate code that passes function parameters and return values that (in
9469 the called function) are seen as registers @code{$0} and up, as opposed to
9470 the GNU ABI which uses global registers @code{$231} and up.
9473 @itemx -mno-zero-extend
9474 @opindex mzero-extend
9475 @opindex mno-zero-extend
9476 When reading data from memory in sizes shorter than 64 bits, use (do not
9477 use) zero-extending load instructions by default, rather than
9478 sign-extending ones.
9481 @itemx -mno-knuthdiv
9483 @opindex mno-knuthdiv
9484 Make the result of a division yielding a remainder have the same sign as
9485 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9486 remainder follows the sign of the dividend. Both methods are
9487 arithmetically valid, the latter being almost exclusively used.
9489 @item -mtoplevel-symbols
9490 @itemx -mno-toplevel-symbols
9491 @opindex mtoplevel-symbols
9492 @opindex mno-toplevel-symbols
9493 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9494 code can be used with the @code{PREFIX} assembly directive.
9498 Generate an executable in the ELF format, rather than the default
9499 @samp{mmo} format used by the @command{mmix} simulator.
9501 @item -mbranch-predict
9502 @itemx -mno-branch-predict
9503 @opindex mbranch-predict
9504 @opindex mno-branch-predict
9505 Use (do not use) the probable-branch instructions, when static branch
9506 prediction indicates a probable branch.
9509 @node PDP-11 Options
9510 @subsection PDP-11 Options
9511 @cindex PDP-11 Options
9513 These options are defined for the PDP-11:
9518 Use hardware FPP floating point. This is the default. (FIS floating
9519 point on the PDP-11/40 is not supported.)
9522 @opindex msoft-float
9523 Do not use hardware floating point.
9527 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9531 Return floating-point results in memory. This is the default.
9535 Generate code for a PDP-11/40.
9539 Generate code for a PDP-11/45. This is the default.
9543 Generate code for a PDP-11/10.
9545 @item -mbcopy-builtin
9546 @opindex bcopy-builtin
9547 Use inline @code{movstrhi} patterns for copying memory. This is the
9552 Do not use inline @code{movstrhi} patterns for copying memory.
9558 Use 16-bit @code{int}. This is the default.
9564 Use 32-bit @code{int}.
9569 @opindex mno-float32
9570 Use 64-bit @code{float}. This is the default.
9575 @opindex mno-float64
9576 Use 32-bit @code{float}.
9580 Use @code{abshi2} pattern. This is the default.
9584 Do not use @code{abshi2} pattern.
9586 @item -mbranch-expensive
9587 @opindex mbranch-expensive
9588 Pretend that branches are expensive. This is for experimenting with
9589 code generation only.
9591 @item -mbranch-cheap
9592 @opindex mbranch-cheap
9593 Do not pretend that branches are expensive. This is the default.
9597 Generate code for a system with split I&D.
9601 Generate code for a system without split I&D. This is the default.
9605 Use Unix assembler syntax. This is the default when configured for
9610 Use DEC assembler syntax. This is the default when configured for any
9611 PDP-11 target other than @samp{pdp11-*-bsd}.
9614 @node Xstormy16 Options
9615 @subsection Xstormy16 Options
9616 @cindex Xstormy16 Options
9618 These options are defined for Xstormy16:
9623 Choose startup files and linker script suitable for the simulator.
9626 @node Xtensa Options
9627 @subsection Xtensa Options
9628 @cindex Xtensa Options
9630 The Xtensa architecture is designed to support many different
9631 configurations. The compiler's default options can be set to match a
9632 particular Xtensa configuration by copying a configuration file into the
9633 GCC sources when building GCC@. The options below may be used to
9634 override the default options.
9638 @itemx -mlittle-endian
9639 @opindex mbig-endian
9640 @opindex mlittle-endian
9641 Specify big-endian or little-endian byte ordering for the target Xtensa
9647 @opindex mno-density
9648 Enable or disable use of the optional Xtensa code density instructions.
9654 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
9655 will generate MAC16 instructions from standard C code, with the
9656 limitation that it will use neither the MR register file nor any
9657 instruction that operates on the MR registers. When this option is
9658 disabled, GCC will translate 16-bit multiply/accumulate operations to a
9659 combination of core instructions and library calls, depending on whether
9660 any other multiplier options are enabled.
9666 Enable or disable use of the 16-bit integer multiplier option. When
9667 enabled, the compiler will generate 16-bit multiply instructions for
9668 multiplications of 16 bits or smaller in standard C code. When this
9669 option is disabled, the compiler will either use 32-bit multiply or
9670 MAC16 instructions if they are available or generate library calls to
9671 perform the multiply operations using shifts and adds.
9677 Enable or disable use of the 32-bit integer multiplier option. When
9678 enabled, the compiler will generate 32-bit multiply instructions for
9679 multiplications of 32 bits or smaller in standard C code. When this
9680 option is disabled, the compiler will generate library calls to perform
9681 the multiply operations using either shifts and adds or 16-bit multiply
9682 instructions if they are available.
9688 Enable or disable use of the optional normalization shift amount
9689 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
9695 Enable or disable use of the optional minimum and maximum value
9702 Enable or disable use of the optional sign extend (@code{SEXT})
9706 @itemx -mno-booleans
9708 @opindex mno-booleans
9709 Enable or disable support for the boolean register file used by Xtensa
9710 coprocessors. This is not typically useful by itself but may be
9711 required for other options that make use of the boolean registers (e.g.,
9712 the floating-point option).
9716 @opindex mhard-float
9717 @opindex msoft-float
9718 Enable or disable use of the floating-point option. When enabled, GCC
9719 generates floating-point instructions for 32-bit @code{float}
9720 operations. When this option is disabled, GCC generates library calls
9721 to emulate 32-bit floating-point operations using integer instructions.
9722 Regardless of this option, 64-bit @code{double} operations are always
9723 emulated with calls to library functions.
9726 @itemx -mno-fused-madd
9727 @opindex mfused-madd
9728 @opindex mno-fused-madd
9729 Enable or disable use of fused multiply/add and multiply/subtract
9730 instructions in the floating-point option. This has no effect if the
9731 floating-point option is not also enabled. Disabling fused multiply/add
9732 and multiply/subtract instructions forces the compiler to use separate
9733 instructions for the multiply and add/subtract operations. This may be
9734 desirable in some cases where strict IEEE 754-compliant results are
9735 required: the fused multiply add/subtract instructions do not round the
9736 intermediate result, thereby producing results with @emph{more} bits of
9737 precision than specified by the IEEE standard. Disabling fused multiply
9738 add/subtract instructions also ensures that the program output is not
9739 sensitive to the compiler's ability to combine multiply and add/subtract
9742 @item -mserialize-volatile
9743 @itemx -mno-serialize-volatile
9744 @opindex mserialize-volatile
9745 @opindex mno-serialize-volatile
9746 When this option is enabled, GCC inserts @code{MEMW} instructions before
9747 @code{volatile} memory references to guarantee sequential consistency.
9748 The default is @option{-mserialize-volatile}. Use
9749 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
9751 @item -mtext-section-literals
9752 @itemx -mno-text-section-literals
9753 @opindex mtext-section-literals
9754 @opindex mno-text-section-literals
9755 Control the treatment of literal pools. The default is
9756 @option{-mno-text-section-literals}, which places literals in a separate
9757 section in the output file. This allows the literal pool to be placed
9758 in a data RAM/ROM, and it also allows the linker to combine literal
9759 pools from separate object files to remove redundant literals and
9760 improve code size. With @option{-mtext-section-literals}, the literals
9761 are interspersed in the text section in order to keep them as close as
9762 possible to their references. This may be necessary for large assembly
9765 @item -mtarget-align
9766 @itemx -mno-target-align
9767 @opindex mtarget-align
9768 @opindex mno-target-align
9769 When this option is enabled, GCC instructs the assembler to
9770 automatically align instructions to reduce branch penalties at the
9771 expense of some code density. The assembler attempts to widen density
9772 instructions to align branch targets and the instructions following call
9773 instructions. If there are not enough preceding safe density
9774 instructions to align a target, no widening will be performed. The
9775 default is @option{-mtarget-align}. These options do not affect the
9776 treatment of auto-aligned instructions like @code{LOOP}, which the
9777 assembler will always align, either by widening density instructions or
9778 by inserting no-op instructions.
9781 @itemx -mno-longcalls
9783 @opindex mno-longcalls
9784 When this option is enabled, GCC instructs the assembler to translate
9785 direct calls to indirect calls unless it can determine that the target
9786 of a direct call is in the range allowed by the call instruction. This
9787 translation typically occurs for calls to functions in other source
9788 files. Specifically, the assembler translates a direct @code{CALL}
9789 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
9790 The default is @option{-mno-longcalls}. This option should be used in
9791 programs where the call target can potentially be out of range. This
9792 option is implemented in the assembler, not the compiler, so the
9793 assembly code generated by GCC will still show direct call
9794 instructions---look at the disassembled object code to see the actual
9795 instructions. Note that the assembler will use an indirect call for
9796 every cross-file call, not just those that really will be out of range.
9799 @node Code Gen Options
9800 @section Options for Code Generation Conventions
9801 @cindex code generation conventions
9802 @cindex options, code generation
9803 @cindex run-time options
9805 These machine-independent options control the interface conventions
9806 used in code generation.
9808 Most of them have both positive and negative forms; the negative form
9809 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9810 one of the forms is listed---the one which is not the default. You
9811 can figure out the other form by either removing @samp{no-} or adding
9816 @opindex fexceptions
9817 Enable exception handling. Generates extra code needed to propagate
9818 exceptions. For some targets, this implies GCC will generate frame
9819 unwind information for all functions, which can produce significant data
9820 size overhead, although it does not affect execution. If you do not
9821 specify this option, GCC will enable it by default for languages like
9822 C++ which normally require exception handling, and disable it for
9823 languages like C that do not normally require it. However, you may need
9824 to enable this option when compiling C code that needs to interoperate
9825 properly with exception handlers written in C++. You may also wish to
9826 disable this option if you are compiling older C++ programs that don't
9827 use exception handling.
9829 @item -fnon-call-exceptions
9830 @opindex fnon-call-exceptions
9831 Generate code that allows trapping instructions to throw exceptions.
9832 Note that this requires platform-specific runtime support that does
9833 not exist everywhere. Moreover, it only allows @emph{trapping}
9834 instructions to throw exceptions, i.e.@: memory references or floating
9835 point instructions. It does not allow exceptions to be thrown from
9836 arbitrary signal handlers such as @code{SIGALRM}.
9838 @item -funwind-tables
9839 @opindex funwind-tables
9840 Similar to @option{-fexceptions}, except that it will just generate any needed
9841 static data, but will not affect the generated code in any other way.
9842 You will normally not enable this option; instead, a language processor
9843 that needs this handling would enable it on your behalf.
9845 @item -fasynchronous-unwind-tables
9846 @opindex funwind-tables
9847 Generate unwind table in dwarf2 format, if supported by target machine. The
9848 table is exact at each instruction boundary, so it can be used for stack
9849 unwinding from asynchronous events (such as debugger or garbage collector).
9851 @item -fpcc-struct-return
9852 @opindex fpcc-struct-return
9853 Return ``short'' @code{struct} and @code{union} values in memory like
9854 longer ones, rather than in registers. This convention is less
9855 efficient, but it has the advantage of allowing intercallability between
9856 GCC-compiled files and files compiled with other compilers.
9858 The precise convention for returning structures in memory depends
9859 on the target configuration macros.
9861 Short structures and unions are those whose size and alignment match
9862 that of some integer type.
9864 @item -freg-struct-return
9865 @opindex freg-struct-return
9866 Return @code{struct} and @code{union} values in registers when possible.
9867 This is more efficient for small structures than
9868 @option{-fpcc-struct-return}.
9870 If you specify neither @option{-fpcc-struct-return} nor
9871 @option{-freg-struct-return}, GCC defaults to whichever convention is
9872 standard for the target. If there is no standard convention, GCC
9873 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9874 the principal compiler. In those cases, we can choose the standard, and
9875 we chose the more efficient register return alternative.
9878 @opindex fshort-enums
9879 Allocate to an @code{enum} type only as many bytes as it needs for the
9880 declared range of possible values. Specifically, the @code{enum} type
9881 will be equivalent to the smallest integer type which has enough room.
9883 @item -fshort-double
9884 @opindex fshort-double
9885 Use the same size for @code{double} as for @code{float}.
9888 @opindex fshared-data
9889 Requests that the data and non-@code{const} variables of this
9890 compilation be shared data rather than private data. The distinction
9891 makes sense only on certain operating systems, where shared data is
9892 shared between processes running the same program, while private data
9893 exists in one copy per process.
9897 In C, allocate even uninitialized global variables in the data section of the
9898 object file, rather than generating them as common blocks. This has the
9899 effect that if the same variable is declared (without @code{extern}) in
9900 two different compilations, you will get an error when you link them.
9901 The only reason this might be useful is if you wish to verify that the
9902 program will work on other systems which always work this way.
9906 Ignore the @samp{#ident} directive.
9908 @item -fno-gnu-linker
9909 @opindex fno-gnu-linker
9910 Do not output global initializations (such as C++ constructors and
9911 destructors) in the form used by the GNU linker (on systems where the GNU
9912 linker is the standard method of handling them). Use this option when
9913 you want to use a non-GNU linker, which also requires using the
9914 @command{collect2} program to make sure the system linker includes
9915 constructors and destructors. (@command{collect2} is included in the GCC
9916 distribution.) For systems which @emph{must} use @command{collect2}, the
9917 compiler driver @command{gcc} is configured to do this automatically.
9919 @item -finhibit-size-directive
9920 @opindex finhibit-size-directive
9921 Don't output a @code{.size} assembler directive, or anything else that
9922 would cause trouble if the function is split in the middle, and the
9923 two halves are placed at locations far apart in memory. This option is
9924 used when compiling @file{crtstuff.c}; you should not need to use it
9928 @opindex fverbose-asm
9929 Put extra commentary information in the generated assembly code to
9930 make it more readable. This option is generally only of use to those
9931 who actually need to read the generated assembly code (perhaps while
9932 debugging the compiler itself).
9934 @option{-fno-verbose-asm}, the default, causes the
9935 extra information to be omitted and is useful when comparing two assembler
9940 Consider all memory references through pointers to be volatile.
9942 @item -fvolatile-global
9943 @opindex fvolatile-global
9944 Consider all memory references to extern and global data items to
9945 be volatile. GCC does not consider static data items to be volatile
9946 because of this switch.
9948 @item -fvolatile-static
9949 @opindex fvolatile-static
9950 Consider all memory references to static data to be volatile.
9954 @cindex global offset table
9956 Generate position-independent code (PIC) suitable for use in a shared
9957 library, if supported for the target machine. Such code accesses all
9958 constant addresses through a global offset table (GOT)@. The dynamic
9959 loader resolves the GOT entries when the program starts (the dynamic
9960 loader is not part of GCC; it is part of the operating system). If
9961 the GOT size for the linked executable exceeds a machine-specific
9962 maximum size, you get an error message from the linker indicating that
9963 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9964 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9965 on the m68k and RS/6000. The 386 has no such limit.)
9967 Position-independent code requires special support, and therefore works
9968 only on certain machines. For the 386, GCC supports PIC for System V
9969 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9970 position-independent.
9974 If supported for the target machine, emit position-independent code,
9975 suitable for dynamic linking and avoiding any limit on the size of the
9976 global offset table. This option makes a difference on the m68k, m88k,
9979 Position-independent code requires special support, and therefore works
9980 only on certain machines.
9982 @item -ffixed-@var{reg}
9984 Treat the register named @var{reg} as a fixed register; generated code
9985 should never refer to it (except perhaps as a stack pointer, frame
9986 pointer or in some other fixed role).
9988 @var{reg} must be the name of a register. The register names accepted
9989 are machine-specific and are defined in the @code{REGISTER_NAMES}
9990 macro in the machine description macro file.
9992 This flag does not have a negative form, because it specifies a
9995 @item -fcall-used-@var{reg}
9997 Treat the register named @var{reg} as an allocable register that is
9998 clobbered by function calls. It may be allocated for temporaries or
9999 variables that do not live across a call. Functions compiled this way
10000 will not save and restore the register @var{reg}.
10002 It is an error to used this flag with the frame pointer or stack pointer.
10003 Use of this flag for other registers that have fixed pervasive roles in
10004 the machine's execution model will produce disastrous results.
10006 This flag does not have a negative form, because it specifies a
10009 @item -fcall-saved-@var{reg}
10010 @opindex fcall-saved
10011 Treat the register named @var{reg} as an allocable register saved by
10012 functions. It may be allocated even for temporaries or variables that
10013 live across a call. Functions compiled this way will save and restore
10014 the register @var{reg} if they use it.
10016 It is an error to used this flag with the frame pointer or stack pointer.
10017 Use of this flag for other registers that have fixed pervasive roles in
10018 the machine's execution model will produce disastrous results.
10020 A different sort of disaster will result from the use of this flag for
10021 a register in which function values may be returned.
10023 This flag does not have a negative form, because it specifies a
10026 @item -fpack-struct
10027 @opindex fpack-struct
10028 Pack all structure members together without holes. Usually you would
10029 not want to use this option, since it makes the code suboptimal, and
10030 the offsets of structure members won't agree with system libraries.
10032 @item -finstrument-functions
10033 @opindex finstrument-functions
10034 Generate instrumentation calls for entry and exit to functions. Just
10035 after function entry and just before function exit, the following
10036 profiling functions will be called with the address of the current
10037 function and its call site. (On some platforms,
10038 @code{__builtin_return_address} does not work beyond the current
10039 function, so the call site information may not be available to the
10040 profiling functions otherwise.)
10043 void __cyg_profile_func_enter (void *this_fn,
10045 void __cyg_profile_func_exit (void *this_fn,
10049 The first argument is the address of the start of the current function,
10050 which may be looked up exactly in the symbol table.
10052 This instrumentation is also done for functions expanded inline in other
10053 functions. The profiling calls will indicate where, conceptually, the
10054 inline function is entered and exited. This means that addressable
10055 versions of such functions must be available. If all your uses of a
10056 function are expanded inline, this may mean an additional expansion of
10057 code size. If you use @samp{extern inline} in your C code, an
10058 addressable version of such functions must be provided. (This is
10059 normally the case anyways, but if you get lucky and the optimizer always
10060 expands the functions inline, you might have gotten away without
10061 providing static copies.)
10063 A function may be given the attribute @code{no_instrument_function}, in
10064 which case this instrumentation will not be done. This can be used, for
10065 example, for the profiling functions listed above, high-priority
10066 interrupt routines, and any functions from which the profiling functions
10067 cannot safely be called (perhaps signal handlers, if the profiling
10068 routines generate output or allocate memory).
10070 @item -fstack-check
10071 @opindex fstack-check
10072 Generate code to verify that you do not go beyond the boundary of the
10073 stack. You should specify this flag if you are running in an
10074 environment with multiple threads, but only rarely need to specify it in
10075 a single-threaded environment since stack overflow is automatically
10076 detected on nearly all systems if there is only one stack.
10078 Note that this switch does not actually cause checking to be done; the
10079 operating system must do that. The switch causes generation of code
10080 to ensure that the operating system sees the stack being extended.
10082 @item -fstack-limit-register=@var{reg}
10083 @itemx -fstack-limit-symbol=@var{sym}
10084 @itemx -fno-stack-limit
10085 @opindex fstack-limit-register
10086 @opindex fstack-limit-symbol
10087 @opindex fno-stack-limit
10088 Generate code to ensure that the stack does not grow beyond a certain value,
10089 either the value of a register or the address of a symbol. If the stack
10090 would grow beyond the value, a signal is raised. For most targets,
10091 the signal is raised before the stack overruns the boundary, so
10092 it is possible to catch the signal without taking special precautions.
10094 For instance, if the stack starts at absolute address @samp{0x80000000}
10095 and grows downwards, you can use the flags
10096 @option{-fstack-limit-symbol=__stack_limit} and
10097 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
10098 of 128KB@. Note that this may only work with the GNU linker.
10100 @cindex aliasing of parameters
10101 @cindex parameters, aliased
10102 @item -fargument-alias
10103 @itemx -fargument-noalias
10104 @itemx -fargument-noalias-global
10105 @opindex fargument-alias
10106 @opindex fargument-noalias
10107 @opindex fargument-noalias-global
10108 Specify the possible relationships among parameters and between
10109 parameters and global data.
10111 @option{-fargument-alias} specifies that arguments (parameters) may
10112 alias each other and may alias global storage.@*
10113 @option{-fargument-noalias} specifies that arguments do not alias
10114 each other, but may alias global storage.@*
10115 @option{-fargument-noalias-global} specifies that arguments do not
10116 alias each other and do not alias global storage.
10118 Each language will automatically use whatever option is required by
10119 the language standard. You should not need to use these options yourself.
10121 @item -fleading-underscore
10122 @opindex fleading-underscore
10123 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
10124 change the way C symbols are represented in the object file. One use
10125 is to help link with legacy assembly code.
10127 Be warned that you should know what you are doing when invoking this
10128 option, and that not all targets provide complete support for it.
10133 @node Environment Variables
10134 @section Environment Variables Affecting GCC
10135 @cindex environment variables
10137 @c man begin ENVIRONMENT
10139 This section describes several environment variables that affect how GCC
10140 operates. Some of them work by specifying directories or prefixes to use
10141 when searching for various kinds of files. Some are used to specify other
10142 aspects of the compilation environment.
10144 Note that you can also specify places to search using options such as
10145 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
10146 take precedence over places specified using environment variables, which
10147 in turn take precedence over those specified by the configuration of GCC@.
10148 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
10149 GNU Compiler Collection (GCC) Internals}.
10154 @c @itemx LC_COLLATE
10156 @c @itemx LC_MONETARY
10157 @c @itemx LC_NUMERIC
10162 @c @findex LC_COLLATE
10163 @findex LC_MESSAGES
10164 @c @findex LC_MONETARY
10165 @c @findex LC_NUMERIC
10169 These environment variables control the way that GCC uses
10170 localization information that allow GCC to work with different
10171 national conventions. GCC inspects the locale categories
10172 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
10173 so. These locale categories can be set to any value supported by your
10174 installation. A typical value is @samp{en_UK} for English in the United
10177 The @env{LC_CTYPE} environment variable specifies character
10178 classification. GCC uses it to determine the character boundaries in
10179 a string; this is needed for some multibyte encodings that contain quote
10180 and escape characters that would otherwise be interpreted as a string
10183 The @env{LC_MESSAGES} environment variable specifies the language to
10184 use in diagnostic messages.
10186 If the @env{LC_ALL} environment variable is set, it overrides the value
10187 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10188 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10189 environment variable. If none of these variables are set, GCC
10190 defaults to traditional C English behavior.
10194 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10195 files. GCC uses temporary files to hold the output of one stage of
10196 compilation which is to be used as input to the next stage: for example,
10197 the output of the preprocessor, which is the input to the compiler
10200 @item GCC_EXEC_PREFIX
10201 @findex GCC_EXEC_PREFIX
10202 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10203 names of the subprograms executed by the compiler. No slash is added
10204 when this prefix is combined with the name of a subprogram, but you can
10205 specify a prefix that ends with a slash if you wish.
10207 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10208 an appropriate prefix to use based on the pathname it was invoked with.
10210 If GCC cannot find the subprogram using the specified prefix, it
10211 tries looking in the usual places for the subprogram.
10213 The default value of @env{GCC_EXEC_PREFIX} is
10214 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10215 of @code{prefix} when you ran the @file{configure} script.
10217 Other prefixes specified with @option{-B} take precedence over this prefix.
10219 This prefix is also used for finding files such as @file{crt0.o} that are
10222 In addition, the prefix is used in an unusual way in finding the
10223 directories to search for header files. For each of the standard
10224 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10225 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10226 replacing that beginning with the specified prefix to produce an
10227 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10228 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10229 These alternate directories are searched first; the standard directories
10232 @item COMPILER_PATH
10233 @findex COMPILER_PATH
10234 The value of @env{COMPILER_PATH} is a colon-separated list of
10235 directories, much like @env{PATH}. GCC tries the directories thus
10236 specified when searching for subprograms, if it can't find the
10237 subprograms using @env{GCC_EXEC_PREFIX}.
10240 @findex LIBRARY_PATH
10241 The value of @env{LIBRARY_PATH} is a colon-separated list of
10242 directories, much like @env{PATH}. When configured as a native compiler,
10243 GCC tries the directories thus specified when searching for special
10244 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10245 using GCC also uses these directories when searching for ordinary
10246 libraries for the @option{-l} option (but directories specified with
10247 @option{-L} come first).
10249 @item C_INCLUDE_PATH
10250 @itemx CPLUS_INCLUDE_PATH
10251 @itemx OBJC_INCLUDE_PATH
10252 @findex C_INCLUDE_PATH
10253 @findex CPLUS_INCLUDE_PATH
10254 @findex OBJC_INCLUDE_PATH
10255 @c @itemx OBJCPLUS_INCLUDE_PATH
10256 These environment variables pertain to particular languages. Each
10257 variable's value is a colon-separated list of directories, much like
10258 @env{PATH}. When GCC searches for header files, it tries the
10259 directories listed in the variable for the language you are using, after
10260 the directories specified with @option{-I} but before the standard header
10263 @item DEPENDENCIES_OUTPUT
10264 @findex DEPENDENCIES_OUTPUT
10265 @cindex dependencies for make as output
10266 If this variable is set, its value specifies how to output dependencies
10267 for Make based on the header files processed by the compiler. This
10268 output looks much like the output from the @option{-M} option
10269 (@pxref{Preprocessor Options}), but it goes to a separate file, and is
10270 in addition to the usual results of compilation.
10272 The value of @env{DEPENDENCIES_OUTPUT} can be just a file name, in
10273 which case the Make rules are written to that file, guessing the target
10274 name from the source file name. Or the value can have the form
10275 @samp{@var{file} @var{target}}, in which case the rules are written to
10276 file @var{file} using @var{target} as the target name.
10280 @cindex locale definition
10281 This variable is used to pass locale information to the compiler. One way in
10282 which this information is used is to determine the character set to be used
10283 when character literals, string literals and comments are parsed in C and C++.
10284 When the compiler is configured to allow multibyte characters,
10285 the following values for @env{LANG} are recognized:
10289 Recognize JIS characters.
10291 Recognize SJIS characters.
10293 Recognize EUCJP characters.
10296 If @env{LANG} is not defined, or if it has some other value, then the
10297 compiler will use mblen and mbtowc as defined by the default locale to
10298 recognize and translate multibyte characters.
10303 @node Running Protoize
10304 @section Running Protoize
10306 The program @code{protoize} is an optional part of GCC@. You can use
10307 it to add prototypes to a program, thus converting the program to ISO
10308 C in one respect. The companion program @code{unprotoize} does the
10309 reverse: it removes argument types from any prototypes that are found.
10311 When you run these programs, you must specify a set of source files as
10312 command line arguments. The conversion programs start out by compiling
10313 these files to see what functions they define. The information gathered
10314 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10316 After scanning comes actual conversion. The specified files are all
10317 eligible to be converted; any files they include (whether sources or
10318 just headers) are eligible as well.
10320 But not all the eligible files are converted. By default,
10321 @code{protoize} and @code{unprotoize} convert only source and header
10322 files in the current directory. You can specify additional directories
10323 whose files should be converted with the @option{-d @var{directory}}
10324 option. You can also specify particular files to exclude with the
10325 @option{-x @var{file}} option. A file is converted if it is eligible, its
10326 directory name matches one of the specified directory names, and its
10327 name within the directory has not been excluded.
10329 Basic conversion with @code{protoize} consists of rewriting most
10330 function definitions and function declarations to specify the types of
10331 the arguments. The only ones not rewritten are those for varargs
10334 @code{protoize} optionally inserts prototype declarations at the
10335 beginning of the source file, to make them available for any calls that
10336 precede the function's definition. Or it can insert prototype
10337 declarations with block scope in the blocks where undeclared functions
10340 Basic conversion with @code{unprotoize} consists of rewriting most
10341 function declarations to remove any argument types, and rewriting
10342 function definitions to the old-style pre-ISO form.
10344 Both conversion programs print a warning for any function declaration or
10345 definition that they can't convert. You can suppress these warnings
10348 The output from @code{protoize} or @code{unprotoize} replaces the
10349 original source file. The original file is renamed to a name ending
10350 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10351 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10352 for DOS) file already exists, then the source file is simply discarded.
10354 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10355 scan the program and collect information about the functions it uses.
10356 So neither of these programs will work until GCC is installed.
10358 Here is a table of the options you can use with @code{protoize} and
10359 @code{unprotoize}. Each option works with both programs unless
10363 @item -B @var{directory}
10364 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10365 usual directory (normally @file{/usr/local/lib}). This file contains
10366 prototype information about standard system functions. This option
10367 applies only to @code{protoize}.
10369 @item -c @var{compilation-options}
10370 Use @var{compilation-options} as the options when running @code{gcc} to
10371 produce the @samp{.X} files. The special option @option{-aux-info} is
10372 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10374 Note that the compilation options must be given as a single argument to
10375 @code{protoize} or @code{unprotoize}. If you want to specify several
10376 @code{gcc} options, you must quote the entire set of compilation options
10377 to make them a single word in the shell.
10379 There are certain @code{gcc} arguments that you cannot use, because they
10380 would produce the wrong kind of output. These include @option{-g},
10381 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10382 the @var{compilation-options}, they are ignored.
10385 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10386 systems) instead of @samp{.c}. This is convenient if you are converting
10387 a C program to C++. This option applies only to @code{protoize}.
10390 Add explicit global declarations. This means inserting explicit
10391 declarations at the beginning of each source file for each function
10392 that is called in the file and was not declared. These declarations
10393 precede the first function definition that contains a call to an
10394 undeclared function. This option applies only to @code{protoize}.
10396 @item -i @var{string}
10397 Indent old-style parameter declarations with the string @var{string}.
10398 This option applies only to @code{protoize}.
10400 @code{unprotoize} converts prototyped function definitions to old-style
10401 function definitions, where the arguments are declared between the
10402 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10403 uses five spaces as the indentation. If you want to indent with just
10404 one space instead, use @option{-i " "}.
10407 Keep the @samp{.X} files. Normally, they are deleted after conversion
10411 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10412 a prototype declaration for each function in each block which calls the
10413 function without any declaration. This option applies only to
10417 Make no real changes. This mode just prints information about the conversions
10418 that would have been done without @option{-n}.
10421 Make no @samp{.save} files. The original files are simply deleted.
10422 Use this option with caution.
10424 @item -p @var{program}
10425 Use the program @var{program} as the compiler. Normally, the name
10426 @file{gcc} is used.
10429 Work quietly. Most warnings are suppressed.
10432 Print the version number, just like @option{-v} for @code{gcc}.
10435 If you need special compiler options to compile one of your program's
10436 source files, then you should generate that file's @samp{.X} file
10437 specially, by running @code{gcc} on that source file with the
10438 appropriate options and the option @option{-aux-info}. Then run
10439 @code{protoize} on the entire set of files. @code{protoize} will use
10440 the existing @samp{.X} file because it is newer than the source file.
10444 gcc -Dfoo=bar file1.c -aux-info file1.X
10449 You need to include the special files along with the rest in the
10450 @code{protoize} command, even though their @samp{.X} files already
10451 exist, because otherwise they won't get converted.
10453 @xref{Protoize Caveats}, for more information on how to use
10454 @code{protoize} successfully.