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 -Wno-div-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 -Wno-multichar -Wno-format-extra-args -Wno-format-y2k @gol
228 -Wno-import -Wpacked -Wpadded @gol
229 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
230 -Wreturn-type -Wsequence-point -Wshadow @gol
231 -Wsign-compare -Wswitch -Wsystem-headers @gol
232 -Wtrigraphs -Wundef -Wuninitialized @gol
233 -Wunknown-pragmas -Wunreachable-code @gol
234 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
235 -Wunused-value -Wunused-variable -Wwrite-strings}
237 @item C-only Warning Options
239 -Wbad-function-cast -Wmissing-prototypes -Wnested-externs @gol
240 -Wstrict-prototypes -Wtraditional}
242 @item Debugging Options
243 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
245 -d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
246 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
247 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
248 -fdump-tree-original@r{[}-@var{n}@r{]} -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
249 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
251 -fprofile-arcs -ftest-coverage -ftime-report @gol
252 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
253 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
254 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
255 -print-multi-directory -print-multi-lib @gol
256 -print-prog-name=@var{program} -print-search-dirs -Q @gol
259 @item Optimization Options
260 @xref{Optimize Options,,Options that Control Optimization}.
262 -falign-functions=@var{n} -falign-jumps=@var{n} @gol
263 -falign-labels=@var{n} -falign-loops=@var{n} @gol
264 -fbranch-probabilities -fcaller-saves -fcprop-registers @gol
265 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
266 -fdelayed-branch -fdelete-null-pointer-checks @gol
267 -fexpensive-optimizations -ffast-math -ffloat-store @gol
268 -fforce-addr -fforce-mem -ffunction-sections @gol
269 -fgcse -fgcse-lm -fgcse-sm @gol
270 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
271 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
272 -fmove-all-movables -fno-default-inline -fno-defer-pop @gol
273 -fno-function-cse -fno-guess-branch-probability @gol
274 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
275 -funsafe-math-optimizations -fno-trapping-math @gol
276 -fno-zero-initialized-in-bss @gol
277 -fomit-frame-pointer -foptimize-register-move @gol
278 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
279 -freduce-all-givs -fregmove -frename-registers @gol
280 -frerun-cse-after-loop -frerun-loop-opt @gol
281 -fschedule-insns -fschedule-insns2 @gol
282 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
283 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol
284 -funroll-all-loops -funroll-loops @gol
285 --param @var{name}=@var{value}
286 -O -O0 -O1 -O2 -O3 -Os}
288 @item Preprocessor Options
289 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
291 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
292 -C -dD -dI -dM -dN @gol
293 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
294 -idirafter @var{dir} @gol
295 -include @var{file} -imacros @var{file} @gol
296 -iprefix @var{file} -iwithprefix @var{dir} @gol
297 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
298 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
299 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
301 @item Assembler Option
302 @xref{Assembler Options,,Passing Options to the Assembler}.
307 @xref{Link Options,,Options for Linking}.
309 @var{object-file-name} -l@var{library} @gol
310 -nostartfiles -nodefaultlibs -nostdlib @gol
311 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
312 -Wl,@var{option} -Xlinker @var{option} @gol
315 @item Directory Options
316 @xref{Directory Options,,Options for Directory Search}.
318 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
321 @c I wrote this xref this way to avoid overfull hbox. -- rms
322 @xref{Target Options}.
324 -b @var{machine} -V @var{version}}
326 @item Machine Dependent Options
327 @xref{Submodel Options,,Hardware Models and Configurations}.
329 @emph{M680x0 Options}
331 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
332 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
333 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
334 -malign-int -mstrict-align}
336 @emph{M68hc1x Options}
338 -m6811 -m6812 -m68hc11 -m68hc12 @gol
339 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
347 -mcpu=@var{cpu-type} @gol
348 -mtune=@var{cpu-type} @gol
349 -mcmodel=@var{code-model} @gol
351 -mapp-regs -mbroken-saverestore -mcypress @gol
352 -mepilogue -mfaster-structs -mflat @gol
353 -mfpu -mhard-float -mhard-quad-float @gol
354 -mimpure-text -mlive-g0 -mno-app-regs @gol
355 -mno-epilogue -mno-faster-structs -mno-flat -mno-fpu @gol
356 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
357 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
358 -msupersparc -munaligned-doubles -mv8}
360 @emph{Convex Options}
362 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
363 -margcount -mnoargcount @gol
364 -mlong32 -mlong64 @gol
365 -mvolatile-cache -mvolatile-nocache}
367 @emph{AMD29K Options}
369 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
370 -mlarge -mnormal -msmall @gol
371 -mkernel-registers -mno-reuse-arg-regs @gol
372 -mno-stack-check -mno-storem-bug @gol
373 -mreuse-arg-regs -msoft-float -mstack-check @gol
374 -mstorem-bug -muser-registers}
378 -mapcs-frame -mno-apcs-frame @gol
379 -mapcs-26 -mapcs-32 @gol
380 -mapcs-stack-check -mno-apcs-stack-check @gol
381 -mapcs-float -mno-apcs-float @gol
382 -mapcs-reentrant -mno-apcs-reentrant @gol
383 -msched-prolog -mno-sched-prolog @gol
384 -mlittle-endian -mbig-endian -mwords-little-endian @gol
385 -malignment-traps -mno-alignment-traps @gol
386 -msoft-float -mhard-float -mfpe @gol
387 -mthumb-interwork -mno-thumb-interwork @gol
388 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
389 -mstructure-size-boundary=@var{n} @gol
390 -mbsd -mxopen -mno-symrename @gol
391 -mabort-on-noreturn @gol
392 -mlong-calls -mno-long-calls @gol
393 -msingle-pic-base -mno-single-pic-base @gol
394 -mpic-register=@var{reg} @gol
395 -mnop-fun-dllimport @gol
396 -mpoke-function-name @gol
398 -mtpcs-frame -mtpcs-leaf-frame @gol
399 -mcaller-super-interworking -mcallee-super-interworking }
401 @emph{MN10200 Options}
405 @emph{MN10300 Options}
407 -mmult-bug -mno-mult-bug @gol
408 -mam33 -mno-am33 @gol
411 @emph{M32R/D Options}
413 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
418 -m88000 -m88100 -m88110 -mbig-pic @gol
419 -mcheck-zero-division -mhandle-large-shift @gol
420 -midentify-revision -mno-check-zero-division @gol
421 -mno-ocs-debug-info -mno-ocs-frame-position @gol
422 -mno-optimize-arg-area -mno-serialize-volatile @gol
423 -mno-underscores -mocs-debug-info @gol
424 -mocs-frame-position -moptimize-arg-area @gol
425 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
426 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
427 -mversion-03.00 -mwarn-passed-structs}
429 @emph{RS/6000 and PowerPC Options}
431 -mcpu=@var{cpu-type} @gol
432 -mtune=@var{cpu-type} @gol
433 -mpower -mno-power -mpower2 -mno-power2 @gol
434 -mpowerpc -mpowerpc64 -mno-powerpc @gol
435 -maltivec -mno-altivec @gol
436 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
437 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
438 -mnew-mnemonics -mold-mnemonics @gol
439 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
440 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
441 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
442 -mstring -mno-string -mupdate -mno-update @gol
443 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
444 -mstrict-align -mno-strict-align -mrelocatable @gol
445 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
446 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
447 -mcall-aix -mcall-sysv -mcall-netbsd @gol
448 -maix-struct-return -msvr4-struct-return
449 -mabi=altivec -mabi=no-altivec @gol
450 -mprototype -mno-prototype @gol
451 -msim -mmvme -mads -myellowknife -memb -msdata @gol
452 -msdata=@var{opt} -mvxworks -G @var{num} -pthread}
456 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
457 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
458 -mminimum-fp-blocks -mnohc-struct-return}
462 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
463 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
464 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
465 -mgas -mgp32 -mgp64 @gol
466 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
467 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
468 -mmips-as -mmips-tfile -mno-abicalls @gol
469 -mno-embedded-data -mno-uninit-const-in-rodata @gol
470 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
471 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
472 -mrnames -msoft-float @gol
473 -m4650 -msingle-float -mmad @gol
474 -mstats -EL -EB -G @var{num} -nocpp @gol
475 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
476 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
478 @emph{i386 and x86-64 Options}
480 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
481 -masm=@var{dialect} -mno-fancy-math-387 @gol
482 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
483 -mno-wide-multiply -mrtd -malign-double @gol
484 -mpreferred-stack-boundary=@var{num} @gol
485 -mmmx -msse -msse2 -msse-math -m3dnow @gol
486 -mthreads -mno-align-stringops -minline-all-stringops @gol
487 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
488 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
494 -march=@var{architecture-type} @gol
495 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
496 -mfast-indirect-calls -mgas -mjump-in-delay @gol
497 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
498 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
499 -mno-jump-in-delay -mno-long-load-store @gol
500 -mno-portable-runtime -mno-soft-float @gol
501 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
502 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
503 -mschedule=@var{cpu-type} -mspace-regs}
505 @emph{Intel 960 Options}
507 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
508 -mcode-align -mcomplex-addr -mleaf-procedures @gol
509 -mic-compat -mic2.0-compat -mic3.0-compat @gol
510 -mintel-asm -mno-clean-linkage -mno-code-align @gol
511 -mno-complex-addr -mno-leaf-procedures @gol
512 -mno-old-align -mno-strict-align -mno-tail-call @gol
513 -mnumerics -mold-align -msoft-float -mstrict-align @gol
516 @emph{DEC Alpha Options}
518 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
519 -mieee -mieee-with-inexact -mieee-conformant @gol
520 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
521 -mtrap-precision=@var{mode} -mbuild-constants @gol
522 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
523 -mbwx -mmax -mfix -mcix @gol
524 -mfloat-vax -mfloat-ieee @gol
525 -mexplicit-relocs -msmall-data -mlarge-data @gol
526 -mmemory-latency=@var{time}}
528 @emph{DEC Alpha/VMS Options}
532 @emph{Clipper Options}
536 @emph{H8/300 Options}
538 -mrelax -mh -ms -mint32 -malign-300}
542 -m1 -m2 -m3 -m3e @gol
543 -m4-nofpu -m4-single-only -m4-single -m4 @gol
544 -m5-64media -m5-64media-nofpu @gol
545 -m5-32media -m5-32media-nofpu @gol
546 -m5-compact -m5-compact-nofpu @gol
547 -mb -ml -mdalign -mrelax @gol
548 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
549 -mieee -misize -mpadstruct -mspace @gol
550 -mprefergot -musermode}
552 @emph{System V Options}
554 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
559 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
560 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
562 @emph{TMS320C3x/C4x Options}
564 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
565 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
566 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
567 -mparallel-insns -mparallel-mpy -mpreserve-float}
571 -mlong-calls -mno-long-calls -mep -mno-ep @gol
572 -mprolog-function -mno-prolog-function -mspace @gol
573 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
578 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
579 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
580 -mregparam -mnoregparam -msb -mnosb @gol
581 -mbitfield -mnobitfield -mhimem -mnohimem}
585 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
586 -mcall-prologues -mno-tablejump -mtiny-stack}
590 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
591 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
592 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
593 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
594 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
598 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
599 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
600 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
605 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
606 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
607 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
608 -minline-divide-max-throughput -mno-dwarf2-asm @gol
609 -mfixed-range=@var{register-range}}
613 -mextmem -mextmemory -monchip -mno-asm-optimize -masm-optimize @gol
614 -mbranch-cost=@var{n} -mcond-exec=@var{n}}
616 @emph{S/390 and zSeries Options}
618 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
619 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
620 -m64 -m31 -mdebug -mno-debug}
624 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
625 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
626 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
627 -mstack-align -mdata-align -mconst-align @gol
628 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
629 -melf -maout -melinux -mlinux -sim -sim2}
631 @emph{PDP-11 Options}
633 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
634 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
635 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
636 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
637 -mbranch-expensive -mbranch-cheap @gol
638 -msplit -mno-split -munix-asm -mdec-asm}
640 @emph{Xstormy16 Options}
644 @emph{Xtensa Options}
646 -mbig-endian -mlittle-endian @gol
647 -mdensity -mno-density @gol
648 -mmac16 -mno-mac16 @gol
649 -mmul16 -mno-mul16 @gol
650 -mmul32 -mno-mul32 @gol
652 -mminmax -mno-minmax @gol
653 -msext -mno-sext @gol
654 -mbooleans -mno-booleans @gol
655 -mhard-float -msoft-float @gol
656 -mfused-madd -mno-fused-madd @gol
657 -mserialize-volatile -mno-serialize-volatile @gol
658 -mtext-section-literals -mno-text-section-literals @gol
659 -mtarget-align -mno-target-align @gol
660 -mlongcalls -mno-longcalls}
662 @item Code Generation Options
663 @xref{Code Gen Options,,Options for Code Generation Conventions}.
665 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
666 -ffixed-@var{reg} -fexceptions @gol
667 -fnon-call-exceptions -funwind-tables @gol
668 -fasynchronous-unwind-tables @gol
669 -finhibit-size-directive -finstrument-functions @gol
670 -fno-common -fno-ident -fno-gnu-linker @gol
671 -fpcc-struct-return -fpic -fPIC @gol
672 -freg-struct-return -fshared-data -fshort-enums @gol
673 -fshort-double -fvolatile @gol
674 -fvolatile-global -fvolatile-static @gol
675 -fverbose-asm -fpack-struct -fstack-check @gol
676 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
677 -fargument-alias -fargument-noalias @gol
678 -fargument-noalias-global -fleading-underscore}
682 * Overall Options:: Controlling the kind of output:
683 an executable, object files, assembler files,
684 or preprocessed source.
685 * C Dialect Options:: Controlling the variant of C language compiled.
686 * C++ Dialect Options:: Variations on C++.
687 * Objective-C Dialect Options:: Variations on Objective-C.
688 * Language Independent Options:: Controlling how diagnostics should be
690 * Warning Options:: How picky should the compiler be?
691 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
692 * Optimize Options:: How much optimization?
693 * Preprocessor Options:: Controlling header files and macro definitions.
694 Also, getting dependency information for Make.
695 * Assembler Options:: Passing options to the assembler.
696 * Link Options:: Specifying libraries and so on.
697 * Directory Options:: Where to find header files and libraries.
698 Where to find the compiler executable files.
699 * Spec Files:: How to pass switches to sub-processes.
700 * Target Options:: Running a cross-compiler, or an old version of GCC.
703 @node Overall Options
704 @section Options Controlling the Kind of Output
706 Compilation can involve up to four stages: preprocessing, compilation
707 proper, assembly and linking, always in that order. The first three
708 stages apply to an individual source file, and end by producing an
709 object file; linking combines all the object files (those newly
710 compiled, and those specified as input) into an executable file.
712 @cindex file name suffix
713 For any given input file, the file name suffix determines what kind of
718 C source code which must be preprocessed.
721 C source code which should not be preprocessed.
724 C++ source code which should not be preprocessed.
727 Objective-C source code. Note that you must link with the library
728 @file{libobjc.a} to make an Objective-C program work.
731 Objective-C source code which should not be preprocessed.
734 C header file (not to be compiled or linked).
738 @itemx @var{file}.cxx
739 @itemx @var{file}.cpp
740 @itemx @var{file}.c++
742 C++ source code which must be preprocessed. Note that in @samp{.cxx},
743 the last two letters must both be literally @samp{x}. Likewise,
744 @samp{.C} refers to a literal capital C@.
747 @itemx @var{file}.for
748 @itemx @var{file}.FOR
749 Fortran source code which should not be preprocessed.
752 @itemx @var{file}.fpp
753 @itemx @var{file}.FPP
754 Fortran source code which must be preprocessed (with the traditional
758 Fortran source code which must be preprocessed with a RATFOR
759 preprocessor (not included with GCC)@.
761 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
762 Using and Porting GNU Fortran}, for more details of the handling of
765 @c FIXME: Descriptions of Java file types.
772 Ada source code file which contains a library unit declaration (a
773 declaration of a package, subprogram, or generic, or a generic
774 instantiation), or a library unit renaming declaration (a package,
775 generic, or subprogram renaming declaration). Such files are also
778 @itemx @var{file}.adb
779 Ada source code file containing a library unit body (a subprogram or
780 package body). Such files are also called @dfn{bodies}.
782 @c GCC also knows about some suffixes for languages not yet included:
788 @itemx @var{file}.chi
789 CHILL source code (preprocessed with the traditional preprocessor).
795 Assembler code which must be preprocessed.
798 An object file to be fed straight into linking.
799 Any file name with no recognized suffix is treated this way.
803 You can specify the input language explicitly with the @option{-x} option:
806 @item -x @var{language}
807 Specify explicitly the @var{language} for the following input files
808 (rather than letting the compiler choose a default based on the file
809 name suffix). This option applies to all following input files until
810 the next @option{-x} option. Possible values for @var{language} are:
812 c c-header cpp-output
814 objective-c objc-cpp-output
815 assembler assembler-with-cpp
818 f77 f77-cpp-input ratfor
823 Turn off any specification of a language, so that subsequent files are
824 handled according to their file name suffixes (as they are if @option{-x}
825 has not been used at all).
827 @item -pass-exit-codes
828 @opindex pass-exit-codes
829 Normally the @command{gcc} program will exit with the code of 1 if any
830 phase of the compiler returns a non-success return code. If you specify
831 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
832 numerically highest error produced by any phase that returned an error
836 If you only want some of the stages of compilation, you can use
837 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
838 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
839 @command{gcc} is to stop. Note that some combinations (for example,
840 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
845 Compile or assemble the source files, but do not link. The linking
846 stage simply is not done. The ultimate output is in the form of an
847 object file for each source file.
849 By default, the object file name for a source file is made by replacing
850 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
852 Unrecognized input files, not requiring compilation or assembly, are
857 Stop after the stage of compilation proper; do not assemble. The output
858 is in the form of an assembler code file for each non-assembler input
861 By default, the assembler file name for a source file is made by
862 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
864 Input files that don't require compilation are ignored.
868 Stop after the preprocessing stage; do not run the compiler proper. The
869 output is in the form of preprocessed source code, which is sent to the
872 Input files which don't require preprocessing are ignored.
874 @cindex output file option
877 Place output in file @var{file}. This applies regardless to whatever
878 sort of output is being produced, whether it be an executable file,
879 an object file, an assembler file or preprocessed C code.
881 Since only one output file can be specified, it does not make sense to
882 use @option{-o} when compiling more than one input file, unless you are
883 producing an executable file as output.
885 If @option{-o} is not specified, the default is to put an executable file
886 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
887 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
888 all preprocessed C source on standard output.
892 Print (on standard error output) the commands executed to run the stages
893 of compilation. Also print the version number of the compiler driver
894 program and of the preprocessor and the compiler proper.
898 Like @option{-v} except the commands are not executed and all command
899 arguments are quoted. This is useful for shell scripts to capture the
900 driver-generated command lines.
904 Use pipes rather than temporary files for communication between the
905 various stages of compilation. This fails to work on some systems where
906 the assembler is unable to read from a pipe; but the GNU assembler has
911 Print (on the standard output) a description of the command line options
912 understood by @command{gcc}. If the @option{-v} option is also specified
913 then @option{--help} will also be passed on to the various processes
914 invoked by @command{gcc}, so that they can display the command line options
915 they accept. If the @option{-W} option is also specified then command
916 line options which have no documentation associated with them will also
921 Print (on the standard output) a description of target specific command
922 line options for each tool.
926 @section Compiling C++ Programs
928 @cindex suffixes for C++ source
929 @cindex C++ source file suffixes
930 C++ source files conventionally use one of the suffixes @samp{.C},
931 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
932 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
933 files with these names and compiles them as C++ programs even if you
934 call the compiler the same way as for compiling C programs (usually with
935 the name @command{gcc}).
939 However, C++ programs often require class libraries as well as a
940 compiler that understands the C++ language---and under some
941 circumstances, you might want to compile programs from standard input,
942 or otherwise without a suffix that flags them as C++ programs.
943 @command{g++} is a program that calls GCC with the default language
944 set to C++, and automatically specifies linking against the C++
945 library. On many systems, @command{g++} is also
946 installed with the name @command{c++}.
948 @cindex invoking @command{g++}
949 When you compile C++ programs, you may specify many of the same
950 command-line options that you use for compiling programs in any
951 language; or command-line options meaningful for C and related
952 languages; or options that are meaningful only for C++ programs.
953 @xref{C Dialect Options,,Options Controlling C Dialect}, for
954 explanations of options for languages related to C@.
955 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
956 explanations of options that are meaningful only for C++ programs.
958 @node C Dialect Options
959 @section Options Controlling C Dialect
960 @cindex dialect options
961 @cindex language dialect options
962 @cindex options, dialect
964 The following options control the dialect of C (or languages derived
965 from C, such as C++ and Objective-C) that the compiler accepts:
972 In C mode, support all ISO C89 programs. In C++ mode,
973 remove GNU extensions that conflict with ISO C++.
975 This turns off certain features of GCC that are incompatible with ISO
976 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
977 such as the @code{asm} and @code{typeof} keywords, and
978 predefined macros such as @code{unix} and @code{vax} that identify the
979 type of system you are using. It also enables the undesirable and
980 rarely used ISO trigraph feature. For the C compiler,
981 it disables recognition of C++ style @samp{//} comments as well as
982 the @code{inline} keyword.
984 The alternate keywords @code{__asm__}, @code{__extension__},
985 @code{__inline__} and @code{__typeof__} continue to work despite
986 @option{-ansi}. You would not want to use them in an ISO C program, of
987 course, but it is useful to put them in header files that might be included
988 in compilations done with @option{-ansi}. Alternate predefined macros
989 such as @code{__unix__} and @code{__vax__} are also available, with or
990 without @option{-ansi}.
992 The @option{-ansi} option does not cause non-ISO programs to be
993 rejected gratuitously. For that, @option{-pedantic} is required in
994 addition to @option{-ansi}. @xref{Warning Options}.
996 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
997 option is used. Some header files may notice this macro and refrain
998 from declaring certain functions or defining certain macros that the
999 ISO standard doesn't call for; this is to avoid interfering with any
1000 programs that might use these names for other things.
1002 Functions which would normally be built in but do not have semantics
1003 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1004 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1005 built-in functions provided by GCC}, for details of the functions
1010 Determine the language standard. This option is currently only
1011 supported when compiling C@. A value for this option must be provided;
1017 ISO C89 (same as @option{-ansi}).
1019 @item iso9899:199409
1020 ISO C89 as modified in amendment 1.
1026 ISO C99. Note that this standard is not yet fully supported; see
1027 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1028 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1031 Default, ISO C89 plus GNU extensions (including some C99 features).
1035 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1036 this will become the default. The name @samp{gnu9x} is deprecated.
1040 Even when this option is not specified, you can still use some of the
1041 features of newer standards in so far as they do not conflict with
1042 previous C standards. For example, you may use @code{__restrict__} even
1043 when @option{-std=c99} is not specified.
1045 The @option{-std} options specifying some version of ISO C have the same
1046 effects as @option{-ansi}, except that features that were not in ISO C89
1047 but are in the specified version (for example, @samp{//} comments and
1048 the @code{inline} keyword in ISO C99) are not disabled.
1050 @xref{Standards,,Language Standards Supported by GCC}, for details of
1051 these standard versions.
1053 @item -aux-info @var{filename}
1055 Output to the given filename prototyped declarations for all functions
1056 declared and/or defined in a translation unit, including those in header
1057 files. This option is silently ignored in any language other than C@.
1059 Besides declarations, the file indicates, in comments, the origin of
1060 each declaration (source file and line), whether the declaration was
1061 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1062 @samp{O} for old, respectively, in the first character after the line
1063 number and the colon), and whether it came from a declaration or a
1064 definition (@samp{C} or @samp{F}, respectively, in the following
1065 character). In the case of function definitions, a K&R-style list of
1066 arguments followed by their declarations is also provided, inside
1067 comments, after the declaration.
1071 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1072 keyword, so that code can use these words as identifiers. You can use
1073 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1074 instead. @option{-ansi} implies @option{-fno-asm}.
1076 In C++, this switch only affects the @code{typeof} keyword, since
1077 @code{asm} and @code{inline} are standard keywords. You may want to
1078 use the @option{-fno-gnu-keywords} flag instead, which has the same
1079 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1080 switch only affects the @code{asm} and @code{typeof} keywords, since
1081 @code{inline} is a standard keyword in ISO C99.
1084 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1085 @opindex fno-builtin
1086 @cindex built-in functions
1087 Don't recognize built-in functions that do not begin with
1088 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1089 functions provided by GCC}, for details of the functions affected,
1090 including those which are not built-in functions when @option{-ansi} or
1091 @option{-std} options for strict ISO C conformance are used because they
1092 do not have an ISO standard meaning.
1094 GCC normally generates special code to handle certain built-in functions
1095 more efficiently; for instance, calls to @code{alloca} may become single
1096 instructions that adjust the stack directly, and calls to @code{memcpy}
1097 may become inline copy loops. The resulting code is often both smaller
1098 and faster, but since the function calls no longer appear as such, you
1099 cannot set a breakpoint on those calls, nor can you change the behavior
1100 of the functions by linking with a different library.
1102 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1103 option has no effect. Therefore, in C++, the only way to get the
1104 optimization benefits of built-in functions is to call the function
1105 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1106 built-in functions to implement many functions (like
1107 @code{std::strchr}), so that you automatically get efficient code.
1109 With the @option{-fno-builtin-@var{function}} option, not available
1110 when compiling C++, only the built-in function @var{function} is
1111 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1112 function is named this is not built-in in this version of GCC, this
1113 option is ignored. There is no corresponding
1114 @option{-fbuiltin-@var{function}} option; if you wish to enable
1115 built-in functions selectively when using @option{-fno-builtin} or
1116 @option{-ffreestanding}, you may define macros such as:
1119 #define abs(n) __builtin_abs ((n))
1120 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1125 @cindex hosted environment
1127 Assert that compilation takes place in a hosted environment. This implies
1128 @option{-fbuiltin}. A hosted environment is one in which the
1129 entire standard library is available, and in which @code{main} has a return
1130 type of @code{int}. Examples are nearly everything except a kernel.
1131 This is equivalent to @option{-fno-freestanding}.
1133 @item -ffreestanding
1134 @opindex ffreestanding
1135 @cindex hosted environment
1137 Assert that compilation takes place in a freestanding environment. This
1138 implies @option{-fno-builtin}. A freestanding environment
1139 is one in which the standard library may not exist, and program startup may
1140 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1141 This is equivalent to @option{-fno-hosted}.
1143 @xref{Standards,,Language Standards Supported by GCC}, for details of
1144 freestanding and hosted environments.
1148 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1149 options for strict ISO C conformance) implies @option{-trigraphs}.
1151 @cindex traditional C language
1152 @cindex C language, traditional
1154 @itemx -traditional-cpp
1155 @opindex traditional-cpp
1156 @opindex traditional
1157 Formerly, these options caused GCC to attempt to emulate a pre-standard
1158 C compiler. They are now only supported with the @option{-E} switch.
1159 The preprocessor continues to support a pre-standard mode. See the GNU
1160 CPP manual for details.
1162 @item -fcond-mismatch
1163 @opindex fcond-mismatch
1164 Allow conditional expressions with mismatched types in the second and
1165 third arguments. The value of such an expression is void. This option
1166 is not supported for C++.
1168 @item -funsigned-char
1169 @opindex funsigned-char
1170 Let the type @code{char} be unsigned, like @code{unsigned char}.
1172 Each kind of machine has a default for what @code{char} should
1173 be. It is either like @code{unsigned char} by default or like
1174 @code{signed char} by default.
1176 Ideally, a portable program should always use @code{signed char} or
1177 @code{unsigned char} when it depends on the signedness of an object.
1178 But many programs have been written to use plain @code{char} and
1179 expect it to be signed, or expect it to be unsigned, depending on the
1180 machines they were written for. This option, and its inverse, let you
1181 make such a program work with the opposite default.
1183 The type @code{char} is always a distinct type from each of
1184 @code{signed char} or @code{unsigned char}, even though its behavior
1185 is always just like one of those two.
1188 @opindex fsigned-char
1189 Let the type @code{char} be signed, like @code{signed char}.
1191 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1192 the negative form of @option{-funsigned-char}. Likewise, the option
1193 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1195 @item -fsigned-bitfields
1196 @itemx -funsigned-bitfields
1197 @itemx -fno-signed-bitfields
1198 @itemx -fno-unsigned-bitfields
1199 @opindex fsigned-bitfields
1200 @opindex funsigned-bitfields
1201 @opindex fno-signed-bitfields
1202 @opindex fno-unsigned-bitfields
1203 These options control whether a bit-field is signed or unsigned, when the
1204 declaration does not use either @code{signed} or @code{unsigned}. By
1205 default, such a bit-field is signed, because this is consistent: the
1206 basic integer types such as @code{int} are signed types.
1208 @item -fwritable-strings
1209 @opindex fwritable-strings
1210 Store string constants in the writable data segment and don't uniquize
1211 them. This is for compatibility with old programs which assume they can
1212 write into string constants.
1214 Writing into string constants is a very bad idea; ``constants'' should
1218 @opindex fshort-wchar
1219 Override the underlying type for @samp{wchar_t} to be @samp{short
1220 unsigned int} instead of the default for the target. This option is
1221 useful for building programs to run under WINE@.
1224 @node C++ Dialect Options
1225 @section Options Controlling C++ Dialect
1227 @cindex compiler options, C++
1228 @cindex C++ options, command line
1229 @cindex options, C++
1230 This section describes the command-line options that are only meaningful
1231 for C++ programs; but you can also use most of the GNU compiler options
1232 regardless of what language your program is in. For example, you
1233 might compile a file @code{firstClass.C} like this:
1236 g++ -g -frepo -O -c firstClass.C
1240 In this example, only @option{-frepo} is an option meant
1241 only for C++ programs; you can use the other options with any
1242 language supported by GCC@.
1244 Here is a list of options that are @emph{only} for compiling C++ programs:
1247 @item -fno-access-control
1248 @opindex fno-access-control
1249 Turn off all access checking. This switch is mainly useful for working
1250 around bugs in the access control code.
1254 Check that the pointer returned by @code{operator new} is non-null
1255 before attempting to modify the storage allocated. The current Working
1256 Paper requires that @code{operator new} never return a null pointer, so
1257 this check is normally unnecessary.
1259 An alternative to using this option is to specify that your
1260 @code{operator new} does not throw any exceptions; if you declare it
1261 @samp{throw()}, G++ will check the return value. See also @samp{new
1264 @item -fconserve-space
1265 @opindex fconserve-space
1266 Put uninitialized or runtime-initialized global variables into the
1267 common segment, as C does. This saves space in the executable at the
1268 cost of not diagnosing duplicate definitions. If you compile with this
1269 flag and your program mysteriously crashes after @code{main()} has
1270 completed, you may have an object that is being destroyed twice because
1271 two definitions were merged.
1273 This option is no longer useful on most targets, now that support has
1274 been added for putting variables into BSS without making them common.
1276 @item -fno-const-strings
1277 @opindex fno-const-strings
1278 Give string constants type @code{char *} instead of type @code{const
1279 char *}. By default, G++ uses type @code{const char *} as required by
1280 the standard. Even if you use @option{-fno-const-strings}, you cannot
1281 actually modify the value of a string constant, unless you also use
1282 @option{-fwritable-strings}.
1284 This option might be removed in a future release of G++. For maximum
1285 portability, you should structure your code so that it works with
1286 string constants that have type @code{const char *}.
1288 @item -fdollars-in-identifiers
1289 @opindex fdollars-in-identifiers
1290 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1291 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1292 @samp{$} by default on most target systems, but there are a few exceptions.)
1293 Traditional C allowed the character @samp{$} to form part of
1294 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1296 @item -fno-elide-constructors
1297 @opindex fno-elide-constructors
1298 The C++ standard allows an implementation to omit creating a temporary
1299 which is only used to initialize another object of the same type.
1300 Specifying this option disables that optimization, and forces G++ to
1301 call the copy constructor in all cases.
1303 @item -fno-enforce-eh-specs
1304 @opindex fno-enforce-eh-specs
1305 Don't check for violation of exception specifications at runtime. This
1306 option violates the C++ standard, but may be useful for reducing code
1307 size in production builds, much like defining @samp{NDEBUG}. The compiler
1308 will still optimize based on the exception specifications.
1310 @item -fexternal-templates
1311 @opindex fexternal-templates
1313 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1314 template instantiation; template instances are emitted or not according
1315 to the location of the template definition. @xref{Template
1316 Instantiation}, for more information.
1318 This option is deprecated.
1320 @item -falt-external-templates
1321 @opindex falt-external-templates
1322 Similar to @option{-fexternal-templates}, but template instances are
1323 emitted or not according to the place where they are first instantiated.
1324 @xref{Template Instantiation}, for more information.
1326 This option is deprecated.
1329 @itemx -fno-for-scope
1331 @opindex fno-for-scope
1332 If @option{-ffor-scope} is specified, the scope of variables declared in
1333 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1334 as specified by the C++ standard.
1335 If @option{-fno-for-scope} is specified, the scope of variables declared in
1336 a @i{for-init-statement} extends to the end of the enclosing scope,
1337 as was the case in old versions of G++, and other (traditional)
1338 implementations of C++.
1340 The default if neither flag is given to follow the standard,
1341 but to allow and give a warning for old-style code that would
1342 otherwise be invalid, or have different behavior.
1344 @item -fno-gnu-keywords
1345 @opindex fno-gnu-keywords
1346 Do not recognize @code{typeof} as a keyword, so that code can use this
1347 word as an identifier. You can use the keyword @code{__typeof__} instead.
1348 @option{-ansi} implies @option{-fno-gnu-keywords}.
1350 @item -fno-implicit-templates
1351 @opindex fno-implicit-templates
1352 Never emit code for non-inline templates which are instantiated
1353 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1354 @xref{Template Instantiation}, for more information.
1356 @item -fno-implicit-inline-templates
1357 @opindex fno-implicit-inline-templates
1358 Don't emit code for implicit instantiations of inline templates, either.
1359 The default is to handle inlines differently so that compiles with and
1360 without optimization will need the same set of explicit instantiations.
1362 @item -fno-implement-inlines
1363 @opindex fno-implement-inlines
1364 To save space, do not emit out-of-line copies of inline functions
1365 controlled by @samp{#pragma implementation}. This will cause linker
1366 errors if these functions are not inlined everywhere they are called.
1368 @item -fms-extensions
1369 @opindex fms-extensions
1370 Disable pedantic warnings about constructs used in MFC, such as implicit
1371 int and getting a pointer to member function via non-standard syntax.
1373 @item -fno-nonansi-builtins
1374 @opindex fno-nonansi-builtins
1375 Disable built-in declarations of functions that are not mandated by
1376 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1377 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1379 @item -fno-operator-names
1380 @opindex fno-operator-names
1381 Do not treat the operator name keywords @code{and}, @code{bitand},
1382 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1383 synonyms as keywords.
1385 @item -fno-optional-diags
1386 @opindex fno-optional-diags
1387 Disable diagnostics that the standard says a compiler does not need to
1388 issue. Currently, the only such diagnostic issued by G++ is the one for
1389 a name having multiple meanings within a class.
1392 @opindex fpermissive
1393 Downgrade messages about nonconformant code from errors to warnings. By
1394 default, G++ effectively sets @option{-pedantic-errors} without
1395 @option{-pedantic}; this option reverses that. This behavior and this
1396 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1400 Enable automatic template instantiation at link time. This option also
1401 implies @option{-fno-implicit-templates}. @xref{Template
1402 Instantiation}, for more information.
1406 Disable generation of information about every class with virtual
1407 functions for use by the C++ runtime type identification features
1408 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1409 of the language, you can save some space by using this flag. Note that
1410 exception handling uses the same information, but it will generate it as
1415 Emit statistics about front-end processing at the end of the compilation.
1416 This information is generally only useful to the G++ development team.
1418 @item -ftemplate-depth-@var{n}
1419 @opindex ftemplate-depth
1420 Set the maximum instantiation depth for template classes to @var{n}.
1421 A limit on the template instantiation depth is needed to detect
1422 endless recursions during template class instantiation. ANSI/ISO C++
1423 conforming programs must not rely on a maximum depth greater than 17.
1425 @item -fuse-cxa-atexit
1426 @opindex fuse-cxa-atexit
1427 Register destructors for objects with static storage duration with the
1428 @code{__cxa_atexit} function rather than the @code{atexit} function.
1429 This option is required for fully standards-compliant handling of static
1430 destructors, but will only work if your C library supports
1431 @code{__cxa_atexit}.
1435 Emit special relocations for vtables and virtual function references
1436 so that the linker can identify unused virtual functions and zero out
1437 vtable slots that refer to them. This is most useful with
1438 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1439 also discard the functions themselves.
1441 This optimization requires GNU as and GNU ld. Not all systems support
1442 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1446 Do not use weak symbol support, even if it is provided by the linker.
1447 By default, G++ will use weak symbols if they are available. This
1448 option exists only for testing, and should not be used by end-users;
1449 it will result in inferior code and has no benefits. This option may
1450 be removed in a future release of G++.
1454 Do not search for header files in the standard directories specific to
1455 C++, but do still search the other standard directories. (This option
1456 is used when building the C++ library.)
1459 In addition, these optimization, warning, and code generation options
1460 have meanings only for C++ programs:
1463 @item -fno-default-inline
1464 @opindex fno-default-inline
1465 Do not assume @samp{inline} for functions defined inside a class scope.
1466 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1467 functions will have linkage like inline functions; they just won't be
1470 @item -Wctor-dtor-privacy @r{(C++ only)}
1471 @opindex Wctor-dtor-privacy
1472 Warn when a class seems unusable, because all the constructors or
1473 destructors in a class are private and the class has no friends or
1474 public static member functions.
1476 @item -Wnon-virtual-dtor @r{(C++ only)}
1477 @opindex Wnon-virtual-dtor
1478 Warn when a class declares a non-virtual destructor that should probably
1479 be virtual, because it looks like the class will be used polymorphically.
1481 @item -Wreorder @r{(C++ only)}
1483 @cindex reordering, warning
1484 @cindex warning for reordering of member initializers
1485 Warn when the order of member initializers given in the code does not
1486 match the order in which they must be executed. For instance:
1492 A(): j (0), i (1) @{ @}
1496 Here the compiler will warn that the member initializers for @samp{i}
1497 and @samp{j} will be rearranged to match the declaration order of the
1501 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1504 @item -Weffc++ @r{(C++ only)}
1506 Warn about violations of the following style guidelines from Scott Meyers'
1507 @cite{Effective C++} book:
1511 Item 11: Define a copy constructor and an assignment operator for classes
1512 with dynamically allocated memory.
1515 Item 12: Prefer initialization to assignment in constructors.
1518 Item 14: Make destructors virtual in base classes.
1521 Item 15: Have @code{operator=} return a reference to @code{*this}.
1524 Item 23: Don't try to return a reference when you must return an object.
1528 and about violations of the following style guidelines from Scott Meyers'
1529 @cite{More Effective C++} book:
1533 Item 6: Distinguish between prefix and postfix forms of increment and
1534 decrement operators.
1537 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1541 If you use this option, you should be aware that the standard library
1542 headers do not obey all of these guidelines; you can use @samp{grep -v}
1543 to filter out those warnings.
1545 @item -Wno-deprecated @r{(C++ only)}
1546 @opindex Wno-deprecated
1547 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1549 @item -Wno-non-template-friend @r{(C++ only)}
1550 @opindex Wno-non-template-friend
1551 Disable warnings when non-templatized friend functions are declared
1552 within a template. With the advent of explicit template specification
1553 support in G++, if the name of the friend is an unqualified-id (i.e.,
1554 @samp{friend foo(int)}), the C++ language specification demands that the
1555 friend declare or define an ordinary, nontemplate function. (Section
1556 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1557 could be interpreted as a particular specialization of a templatized
1558 function. Because this non-conforming behavior is no longer the default
1559 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1560 check existing code for potential trouble spots, and is on by default.
1561 This new compiler behavior can be turned off with
1562 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1563 but disables the helpful warning.
1565 @item -Wold-style-cast @r{(C++ only)}
1566 @opindex Wold-style-cast
1567 Warn if an old-style (C-style) cast to a non-void type is used within
1568 a C++ program. The new-style casts (@samp{static_cast},
1569 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1570 unintended effects, and much easier to grep for.
1572 @item -Woverloaded-virtual @r{(C++ only)}
1573 @opindex Woverloaded-virtual
1574 @cindex overloaded virtual fn, warning
1575 @cindex warning for overloaded virtual fn
1576 Warn when a function declaration hides virtual functions from a
1577 base class. For example, in:
1584 struct B: public A @{
1589 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1597 will fail to compile.
1599 @item -Wno-pmf-conversions @r{(C++ only)}
1600 @opindex Wno-pmf-conversions
1601 Disable the diagnostic for converting a bound pointer to member function
1604 @item -Wsign-promo @r{(C++ only)}
1605 @opindex Wsign-promo
1606 Warn when overload resolution chooses a promotion from unsigned or
1607 enumeral type to a signed type over a conversion to an unsigned type of
1608 the same size. Previous versions of G++ would try to preserve
1609 unsignedness, but the standard mandates the current behavior.
1611 @item -Wsynth @r{(C++ only)}
1613 @cindex warning for synthesized methods
1614 @cindex synthesized methods, warning
1615 Warn when G++'s synthesis behavior does not match that of cfront. For
1621 A& operator = (int);
1631 In this example, G++ will synthesize a default @samp{A& operator =
1632 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1635 @node Objective-C Dialect Options
1636 @section Options Controlling Objective-C Dialect
1638 @cindex compiler options, Objective-C
1639 @cindex Objective-C options, command line
1640 @cindex options, Objective-C
1641 This section describes the command-line options that are only meaningful
1642 for Objective-C programs; but you can also use most of the GNU compiler
1643 options regardless of what language your program is in. For example,
1644 you might compile a file @code{some_class.m} like this:
1647 gcc -g -fgnu-runtime -O -c some_class.m
1651 In this example, only @option{-fgnu-runtime} is an option meant only for
1652 Objective-C programs; you can use the other options with any language
1655 Here is a list of options that are @emph{only} for compiling Objective-C
1659 @item -fconstant-string-class=@var{class-name}
1660 @opindex fconstant-string-class
1661 Use @var{class-name} as the name of the class to instantiate for each
1662 literal string specified with the syntax @code{@@"@dots{}"}. The default
1663 class name is @code{NXConstantString}.
1666 @opindex fgnu-runtime
1667 Generate object code compatible with the standard GNU Objective-C
1668 runtime. This is the default for most types of systems.
1670 @item -fnext-runtime
1671 @opindex fnext-runtime
1672 Generate output compatible with the NeXT runtime. This is the default
1673 for NeXT-based systems, including Darwin and Mac OS X@.
1677 Dump interface declarations for all classes seen in the source file to a
1678 file named @file{@var{sourcename}.decl}.
1681 @opindex Wno-protocol
1682 Do not warn if methods required by a protocol are not implemented
1683 in the class adopting it.
1687 Warn if a selector has multiple methods of different types defined.
1689 @c not documented because only avail via -Wp
1690 @c @item -print-objc-runtime-info
1694 @node Language Independent Options
1695 @section Options to Control Diagnostic Messages Formatting
1696 @cindex options to control diagnostics formatting
1697 @cindex diagnostic messages
1698 @cindex message formatting
1700 Traditionally, diagnostic messages have been formatted irrespective of
1701 the output device's aspect (e.g.@: its width, @dots{}). The options described
1702 below can be used to control the diagnostic messages formatting
1703 algorithm, e.g.@: how many characters per line, how often source location
1704 information should be reported. Right now, only the C++ front end can
1705 honor these options. However it is expected, in the near future, that
1706 the remaining front ends would be able to digest them correctly.
1709 @item -fmessage-length=@var{n}
1710 @opindex fmessage-length
1711 Try to format error messages so that they fit on lines of about @var{n}
1712 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1713 the front ends supported by GCC@. If @var{n} is zero, then no
1714 line-wrapping will be done; each error message will appear on a single
1717 @opindex fdiagnostics-show-location
1718 @item -fdiagnostics-show-location=once
1719 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1720 reporter to emit @emph{once} source location information; that is, in
1721 case the message is too long to fit on a single physical line and has to
1722 be wrapped, the source location won't be emitted (as prefix) again,
1723 over and over, in subsequent continuation lines. This is the default
1726 @item -fdiagnostics-show-location=every-line
1727 Only meaningful in line-wrapping mode. Instructs the diagnostic
1728 messages reporter to emit the same source location information (as
1729 prefix) for physical lines that result from the process of breaking
1730 a message which is too long to fit on a single line.
1734 @node Warning Options
1735 @section Options to Request or Suppress Warnings
1736 @cindex options to control warnings
1737 @cindex warning messages
1738 @cindex messages, warning
1739 @cindex suppressing warnings
1741 Warnings are diagnostic messages that report constructions which
1742 are not inherently erroneous but which are risky or suggest there
1743 may have been an error.
1745 You can request many specific warnings with options beginning @samp{-W},
1746 for example @option{-Wimplicit} to request warnings on implicit
1747 declarations. Each of these specific warning options also has a
1748 negative form beginning @samp{-Wno-} to turn off warnings;
1749 for example, @option{-Wno-implicit}. This manual lists only one of the
1750 two forms, whichever is not the default.
1752 These options control the amount and kinds of warnings produced by GCC:
1755 @cindex syntax checking
1757 @opindex fsyntax-only
1758 Check the code for syntax errors, but don't do anything beyond that.
1762 Issue all the warnings demanded by strict ISO C and ISO C++;
1763 reject all programs that use forbidden extensions, and some other
1764 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1765 version of the ISO C standard specified by any @option{-std} option used.
1767 Valid ISO C and ISO C++ programs should compile properly with or without
1768 this option (though a rare few will require @option{-ansi} or a
1769 @option{-std} option specifying the required version of ISO C)@. However,
1770 without this option, certain GNU extensions and traditional C and C++
1771 features are supported as well. With this option, they are rejected.
1773 @option{-pedantic} does not cause warning messages for use of the
1774 alternate keywords whose names begin and end with @samp{__}. Pedantic
1775 warnings are also disabled in the expression that follows
1776 @code{__extension__}. However, only system header files should use
1777 these escape routes; application programs should avoid them.
1778 @xref{Alternate Keywords}.
1780 Some users try to use @option{-pedantic} to check programs for strict ISO
1781 C conformance. They soon find that it does not do quite what they want:
1782 it finds some non-ISO practices, but not all---only those for which
1783 ISO C @emph{requires} a diagnostic, and some others for which
1784 diagnostics have been added.
1786 A feature to report any failure to conform to ISO C might be useful in
1787 some instances, but would require considerable additional work and would
1788 be quite different from @option{-pedantic}. We don't have plans to
1789 support such a feature in the near future.
1791 Where the standard specified with @option{-std} represents a GNU
1792 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1793 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1794 extended dialect is based. Warnings from @option{-pedantic} are given
1795 where they are required by the base standard. (It would not make sense
1796 for such warnings to be given only for features not in the specified GNU
1797 C dialect, since by definition the GNU dialects of C include all
1798 features the compiler supports with the given option, and there would be
1799 nothing to warn about.)
1801 @item -pedantic-errors
1802 @opindex pedantic-errors
1803 Like @option{-pedantic}, except that errors are produced rather than
1808 Inhibit all warning messages.
1812 Inhibit warning messages about the use of @samp{#import}.
1814 @item -Wchar-subscripts
1815 @opindex Wchar-subscripts
1816 Warn if an array subscript has type @code{char}. This is a common cause
1817 of error, as programmers often forget that this type is signed on some
1822 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1823 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1827 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1828 the arguments supplied have types appropriate to the format string
1829 specified, and that the conversions specified in the format string make
1830 sense. This includes standard functions, and others specified by format
1831 attributes (@pxref{Function Attributes}), in the @code{printf},
1832 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1833 not in the C standard) families.
1835 The formats are checked against the format features supported by GNU
1836 libc version 2.2. These include all ISO C89 and C99 features, as well
1837 as features from the Single Unix Specification and some BSD and GNU
1838 extensions. Other library implementations may not support all these
1839 features; GCC does not support warning about features that go beyond a
1840 particular library's limitations. However, if @option{-pedantic} is used
1841 with @option{-Wformat}, warnings will be given about format features not
1842 in the selected standard version (but not for @code{strfmon} formats,
1843 since those are not in any version of the C standard). @xref{C Dialect
1844 Options,,Options Controlling C Dialect}.
1846 @option{-Wformat} is included in @option{-Wall}. For more control over some
1847 aspects of format checking, the options @option{-Wno-format-y2k},
1848 @option{-Wno-format-extra-args}, @option{-Wformat-nonliteral},
1849 @option{-Wformat-security} and @option{-Wformat=2} are available, but are
1850 not included in @option{-Wall}.
1852 @item -Wno-format-y2k
1853 @opindex Wno-format-y2k
1854 If @option{-Wformat} is specified, do not warn about @code{strftime}
1855 formats which may yield only a two-digit year.
1857 @item -Wno-format-extra-args
1858 @opindex Wno-format-extra-args
1859 If @option{-Wformat} is specified, do not warn about excess arguments to a
1860 @code{printf} or @code{scanf} format function. The C standard specifies
1861 that such arguments are ignored.
1863 Where the unused arguments lie between used arguments that are
1864 specified with @samp{$} operand number specifications, normally
1865 warnings are still given, since the implementation could not know what
1866 type to pass to @code{va_arg} to skip the unused arguments. However,
1867 in the case of @code{scanf} formats, this option will suppress the
1868 warning if the unused arguments are all pointers, since the Single
1869 Unix Specification says that such unused arguments are allowed.
1871 @item -Wformat-nonliteral
1872 @opindex Wformat-nonliteral
1873 If @option{-Wformat} is specified, also warn if the format string is not a
1874 string literal and so cannot be checked, unless the format function
1875 takes its format arguments as a @code{va_list}.
1877 @item -Wformat-security
1878 @opindex Wformat-security
1879 If @option{-Wformat} is specified, also warn about uses of format
1880 functions that represent possible security problems. At present, this
1881 warns about calls to @code{printf} and @code{scanf} functions where the
1882 format string is not a string literal and there are no format arguments,
1883 as in @code{printf (foo);}. This may be a security hole if the format
1884 string came from untrusted input and contains @samp{%n}. (This is
1885 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1886 in future warnings may be added to @option{-Wformat-security} that are not
1887 included in @option{-Wformat-nonliteral}.)
1891 Enable @option{-Wformat} plus format checks not included in
1892 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1893 -Wformat-nonliteral -Wformat-security}.
1895 @item -Wimplicit-int
1896 @opindex Wimplicit-int
1897 Warn when a declaration does not specify a type.
1899 @item -Wimplicit-function-declaration
1900 @itemx -Werror-implicit-function-declaration
1901 @opindex Wimplicit-function-declaration
1902 @opindex Werror-implicit-function-declaration
1903 Give a warning (or error) whenever a function is used before being
1908 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1912 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1913 function with external linkage, returning int, taking either zero
1914 arguments, two, or three arguments of appropriate types.
1916 @item -Wmissing-braces
1917 @opindex Wmissing-braces
1918 Warn if an aggregate or union initializer is not fully bracketed. In
1919 the following example, the initializer for @samp{a} is not fully
1920 bracketed, but that for @samp{b} is fully bracketed.
1923 int a[2][2] = @{ 0, 1, 2, 3 @};
1924 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1928 @opindex Wparentheses
1929 Warn if parentheses are omitted in certain contexts, such
1930 as when there is an assignment in a context where a truth value
1931 is expected, or when operators are nested whose precedence people
1932 often get confused about.
1934 Also warn about constructions where there may be confusion to which
1935 @code{if} statement an @code{else} branch belongs. Here is an example of
1950 In C, every @code{else} branch belongs to the innermost possible @code{if}
1951 statement, which in this example is @code{if (b)}. This is often not
1952 what the programmer expected, as illustrated in the above example by
1953 indentation the programmer chose. When there is the potential for this
1954 confusion, GCC will issue a warning when this flag is specified.
1955 To eliminate the warning, add explicit braces around the innermost
1956 @code{if} statement so there is no way the @code{else} could belong to
1957 the enclosing @code{if}. The resulting code would look like this:
1973 @item -Wsequence-point
1974 @opindex Wsequence-point
1975 Warn about code that may have undefined semantics because of violations
1976 of sequence point rules in the C standard.
1978 The C standard defines the order in which expressions in a C program are
1979 evaluated in terms of @dfn{sequence points}, which represent a partial
1980 ordering between the execution of parts of the program: those executed
1981 before the sequence point, and those executed after it. These occur
1982 after the evaluation of a full expression (one which is not part of a
1983 larger expression), after the evaluation of the first operand of a
1984 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
1985 function is called (but after the evaluation of its arguments and the
1986 expression denoting the called function), and in certain other places.
1987 Other than as expressed by the sequence point rules, the order of
1988 evaluation of subexpressions of an expression is not specified. All
1989 these rules describe only a partial order rather than a total order,
1990 since, for example, if two functions are called within one expression
1991 with no sequence point between them, the order in which the functions
1992 are called is not specified. However, the standards committee have
1993 ruled that function calls do not overlap.
1995 It is not specified when between sequence points modifications to the
1996 values of objects take effect. Programs whose behavior depends on this
1997 have undefined behavior; the C standard specifies that ``Between the
1998 previous and next sequence point an object shall have its stored value
1999 modified at most once by the evaluation of an expression. Furthermore,
2000 the prior value shall be read only to determine the value to be
2001 stored.''. If a program breaks these rules, the results on any
2002 particular implementation are entirely unpredictable.
2004 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2005 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2006 diagnosed by this option, and it may give an occasional false positive
2007 result, but in general it has been found fairly effective at detecting
2008 this sort of problem in programs.
2010 The present implementation of this option only works for C programs. A
2011 future implementation may also work for C++ programs.
2013 The C standard is worded confusingly, therefore there is some debate
2014 over the precise meaning of the sequence point rules in subtle cases.
2015 Links to discussions of the problem, including proposed formal
2016 definitions, may be found on our readings page, at
2017 @w{@uref{http://gcc.gnu.org/readings.html}}.
2020 @opindex Wreturn-type
2021 Warn whenever a function is defined with a return-type that defaults to
2022 @code{int}. Also warn about any @code{return} statement with no
2023 return-value in a function whose return-type is not @code{void}.
2025 For C++, a function without return type always produces a diagnostic
2026 message, even when @option{-Wno-return-type} is specified. The only
2027 exceptions are @samp{main} and functions defined in system headers.
2031 Warn whenever a @code{switch} statement has an index of enumeral type
2032 and lacks a @code{case} for one or more of the named codes of that
2033 enumeration. (The presence of a @code{default} label prevents this
2034 warning.) @code{case} labels outside the enumeration range also
2035 provoke warnings when this option is used.
2039 Warn if any trigraphs are encountered that might change the meaning of
2040 the program (trigraphs within comments are not warned about).
2042 @item -Wunused-function
2043 @opindex Wunused-function
2044 Warn whenever a static function is declared but not defined or a
2045 non\-inline static function is unused.
2047 @item -Wunused-label
2048 @opindex Wunused-label
2049 Warn whenever a label is declared but not used.
2051 To suppress this warning use the @samp{unused} attribute
2052 (@pxref{Variable Attributes}).
2054 @item -Wunused-parameter
2055 @opindex Wunused-parameter
2056 Warn whenever a function parameter is unused aside from its declaration.
2058 To suppress this warning use the @samp{unused} attribute
2059 (@pxref{Variable Attributes}).
2061 @item -Wunused-variable
2062 @opindex Wunused-variable
2063 Warn whenever a local variable or non-constant static variable is unused
2064 aside from its declaration
2066 To suppress this warning use the @samp{unused} attribute
2067 (@pxref{Variable Attributes}).
2069 @item -Wunused-value
2070 @opindex Wunused-value
2071 Warn whenever a statement computes a result that is explicitly not used.
2073 To suppress this warning cast the expression to @samp{void}.
2077 All all the above @option{-Wunused} options combined.
2079 In order to get a warning about an unused function parameter, you must
2080 either specify @samp{-W -Wunused} or separately specify
2081 @option{-Wunused-parameter}.
2083 @item -Wuninitialized
2084 @opindex Wuninitialized
2085 Warn if an automatic variable is used without first being initialized or
2086 if a variable may be clobbered by a @code{setjmp} call.
2088 These warnings are possible only in optimizing compilation,
2089 because they require data flow information that is computed only
2090 when optimizing. If you don't specify @option{-O}, you simply won't
2093 These warnings occur only for variables that are candidates for
2094 register allocation. Therefore, they do not occur for a variable that
2095 is declared @code{volatile}, or whose address is taken, or whose size
2096 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2097 structures, unions or arrays, even when they are in registers.
2099 Note that there may be no warning about a variable that is used only
2100 to compute a value that itself is never used, because such
2101 computations may be deleted by data flow analysis before the warnings
2104 These warnings are made optional because GCC is not smart
2105 enough to see all the reasons why the code might be correct
2106 despite appearing to have an error. Here is one example of how
2127 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2128 always initialized, but GCC doesn't know this. Here is
2129 another common case:
2134 if (change_y) save_y = y, y = new_y;
2136 if (change_y) y = save_y;
2141 This has no bug because @code{save_y} is used only if it is set.
2143 @cindex @code{longjmp} warnings
2144 This option also warns when a non-volatile automatic variable might be
2145 changed by a call to @code{longjmp}. These warnings as well are possible
2146 only in optimizing compilation.
2148 The compiler sees only the calls to @code{setjmp}. It cannot know
2149 where @code{longjmp} will be called; in fact, a signal handler could
2150 call it at any point in the code. As a result, you may get a warning
2151 even when there is in fact no problem because @code{longjmp} cannot
2152 in fact be called at the place which would cause a problem.
2154 Some spurious warnings can be avoided if you declare all the functions
2155 you use that never return as @code{noreturn}. @xref{Function
2158 @item -Wreorder @r{(C++ only)}
2160 @cindex reordering, warning
2161 @cindex warning for reordering of member initializers
2162 Warn when the order of member initializers given in the code does not
2163 match the order in which they must be executed. For instance:
2165 @item -Wunknown-pragmas
2166 @opindex Wunknown-pragmas
2167 @cindex warning for unknown pragmas
2168 @cindex unknown pragmas, warning
2169 @cindex pragmas, warning of unknown
2170 Warn when a #pragma directive is encountered which is not understood by
2171 GCC@. If this command line option is used, warnings will even be issued
2172 for unknown pragmas in system header files. This is not the case if
2173 the warnings were only enabled by the @option{-Wall} command line option.
2177 All of the above @samp{-W} options combined. This enables all the
2178 warnings about constructions that some users consider questionable, and
2179 that are easy to avoid (or modify to prevent the warning), even in
2180 conjunction with macros.
2183 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2184 Some of them warn about constructions that users generally do not
2185 consider questionable, but which occasionally you might wish to check
2186 for; others warn about constructions that are necessary or hard to avoid
2187 in some cases, and there is no simple way to modify the code to suppress
2193 Print extra warning messages for these events:
2197 A function can return either with or without a value. (Falling
2198 off the end of the function body is considered returning without
2199 a value.) For example, this function would evoke such a
2213 An expression-statement or the left-hand side of a comma expression
2214 contains no side effects.
2215 To suppress the warning, cast the unused expression to void.
2216 For example, an expression such as @samp{x[i,j]} will cause a warning,
2217 but @samp{x[(void)i,j]} will not.
2220 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2223 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2224 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2225 that of ordinary mathematical notation.
2228 Storage-class specifiers like @code{static} are not the first things in
2229 a declaration. According to the C Standard, this usage is obsolescent.
2232 The return type of a function has a type qualifier such as @code{const}.
2233 Such a type qualifier has no effect, since the value returned by a
2234 function is not an lvalue. (But don't warn about the GNU extension of
2235 @code{volatile void} return types. That extension will be warned about
2236 if @option{-pedantic} is specified.)
2239 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2243 A comparison between signed and unsigned values could produce an
2244 incorrect result when the signed value is converted to unsigned.
2245 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2248 An aggregate has a partly bracketed initializer.
2249 For example, the following code would evoke such a warning,
2250 because braces are missing around the initializer for @code{x.h}:
2253 struct s @{ int f, g; @};
2254 struct t @{ struct s h; int i; @};
2255 struct t x = @{ 1, 2, 3 @};
2259 An aggregate has an initializer which does not initialize all members.
2260 For example, the following code would cause such a warning, because
2261 @code{x.h} would be implicitly initialized to zero:
2264 struct s @{ int f, g, h; @};
2265 struct s x = @{ 3, 4 @};
2269 @item -Wno-div-by-zero
2270 @opindex Wno-div-by-zero
2271 @opindex Wdiv-by-zero
2272 Do not warn about compile-time integer division by zero. Floating point
2273 division by zero is not warned about, as it can be a legitimate way of
2274 obtaining infinities and NaNs.
2276 @item -Wsystem-headers
2277 @opindex Wsystem-headers
2278 @cindex warnings from system headers
2279 @cindex system headers, warnings from
2280 Print warning messages for constructs found in system header files.
2281 Warnings from system headers are normally suppressed, on the assumption
2282 that they usually do not indicate real problems and would only make the
2283 compiler output harder to read. Using this command line option tells
2284 GCC to emit warnings from system headers as if they occurred in user
2285 code. However, note that using @option{-Wall} in conjunction with this
2286 option will @emph{not} warn about unknown pragmas in system
2287 headers---for that, @option{-Wunknown-pragmas} must also be used.
2290 @opindex Wfloat-equal
2291 Warn if floating point values are used in equality comparisons.
2293 The idea behind this is that sometimes it is convenient (for the
2294 programmer) to consider floating-point values as approximations to
2295 infinitely precise real numbers. If you are doing this, then you need
2296 to compute (by analysing the code, or in some other way) the maximum or
2297 likely maximum error that the computation introduces, and allow for it
2298 when performing comparisons (and when producing output, but that's a
2299 different problem). In particular, instead of testing for equality, you
2300 would check to see whether the two values have ranges that overlap; and
2301 this is done with the relational operators, so equality comparisons are
2304 @item -Wtraditional @r{(C only)}
2305 @opindex Wtraditional
2306 Warn about certain constructs that behave differently in traditional and
2307 ISO C@. Also warn about ISO C constructs that have no traditional C
2308 equivalent, and/or problematic constructs which should be avoided.
2312 Macro parameters that appear within string literals in the macro body.
2313 In traditional C macro replacement takes place within string literals,
2314 but does not in ISO C@.
2317 In traditional C, some preprocessor directives did not exist.
2318 Traditional preprocessors would only consider a line to be a directive
2319 if the @samp{#} appeared in column 1 on the line. Therefore
2320 @option{-Wtraditional} warns about directives that traditional C
2321 understands but would ignore because the @samp{#} does not appear as the
2322 first character on the line. It also suggests you hide directives like
2323 @samp{#pragma} not understood by traditional C by indenting them. Some
2324 traditional implementations would not recognize @samp{#elif}, so it
2325 suggests avoiding it altogether.
2328 A function-like macro that appears without arguments.
2331 The unary plus operator.
2334 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2335 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2336 constants.) Note, these suffixes appear in macros defined in the system
2337 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2338 Use of these macros in user code might normally lead to spurious
2339 warnings, however gcc's integrated preprocessor has enough context to
2340 avoid warning in these cases.
2343 A function declared external in one block and then used after the end of
2347 A @code{switch} statement has an operand of type @code{long}.
2350 A non-@code{static} function declaration follows a @code{static} one.
2351 This construct is not accepted by some traditional C compilers.
2354 The ISO type of an integer constant has a different width or
2355 signedness from its traditional type. This warning is only issued if
2356 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2357 typically represent bit patterns, are not warned about.
2360 Usage of ISO string concatenation is detected.
2363 Initialization of automatic aggregates.
2366 Identifier conflicts with labels. Traditional C lacks a separate
2367 namespace for labels.
2370 Initialization of unions. If the initializer is zero, the warning is
2371 omitted. This is done under the assumption that the zero initializer in
2372 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2373 initializer warnings and relies on default initialization to zero in the
2377 Conversions by prototypes between fixed/floating point values and vice
2378 versa. The absence of these prototypes when compiling with traditional
2379 C would cause serious problems. This is a subset of the possible
2380 conversion warnings, for the full set use @option{-Wconversion}.
2385 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2389 Warn whenever a local variable shadows another local variable, parameter or
2390 global variable or whenever a built-in function is shadowed.
2392 @item -Wlarger-than-@var{len}
2393 @opindex Wlarger-than
2394 Warn whenever an object of larger than @var{len} bytes is defined.
2396 @item -Wpointer-arith
2397 @opindex Wpointer-arith
2398 Warn about anything that depends on the ``size of'' a function type or
2399 of @code{void}. GNU C assigns these types a size of 1, for
2400 convenience in calculations with @code{void *} pointers and pointers
2403 @item -Wbad-function-cast @r{(C only)}
2404 @opindex Wbad-function-cast
2405 Warn whenever a function call is cast to a non-matching type.
2406 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2410 Warn whenever a pointer is cast so as to remove a type qualifier from
2411 the target type. For example, warn if a @code{const char *} is cast
2412 to an ordinary @code{char *}.
2415 @opindex Wcast-align
2416 Warn whenever a pointer is cast such that the required alignment of the
2417 target is increased. For example, warn if a @code{char *} is cast to
2418 an @code{int *} on machines where integers can only be accessed at
2419 two- or four-byte boundaries.
2421 @item -Wwrite-strings
2422 @opindex Wwrite-strings
2423 When compiling C, give string constants the type @code{const
2424 char[@var{length}]} so that
2425 copying the address of one into a non-@code{const} @code{char *}
2426 pointer will get a warning; when compiling C++, warn about the
2427 deprecated conversion from string constants to @code{char *}.
2428 These warnings will help you find at
2429 compile time code that can try to write into a string constant, but
2430 only if you have been very careful about using @code{const} in
2431 declarations and prototypes. Otherwise, it will just be a nuisance;
2432 this is why we did not make @option{-Wall} request these warnings.
2435 @opindex Wconversion
2436 Warn if a prototype causes a type conversion that is different from what
2437 would happen to the same argument in the absence of a prototype. This
2438 includes conversions of fixed point to floating and vice versa, and
2439 conversions changing the width or signedness of a fixed point argument
2440 except when the same as the default promotion.
2442 Also, warn if a negative integer constant expression is implicitly
2443 converted to an unsigned type. For example, warn about the assignment
2444 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2445 casts like @code{(unsigned) -1}.
2447 @item -Wsign-compare
2448 @opindex Wsign-compare
2449 @cindex warning for comparison of signed and unsigned values
2450 @cindex comparison of signed and unsigned values, warning
2451 @cindex signed and unsigned values, comparison warning
2452 Warn when a comparison between signed and unsigned values could produce
2453 an incorrect result when the signed value is converted to unsigned.
2454 This warning is also enabled by @option{-W}; to get the other warnings
2455 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2457 @item -Waggregate-return
2458 @opindex Waggregate-return
2459 Warn if any functions that return structures or unions are defined or
2460 called. (In languages where you can return an array, this also elicits
2463 @item -Wstrict-prototypes @r{(C only)}
2464 @opindex Wstrict-prototypes
2465 Warn if a function is declared or defined without specifying the
2466 argument types. (An old-style function definition is permitted without
2467 a warning if preceded by a declaration which specifies the argument
2470 @item -Wmissing-prototypes @r{(C only)}
2471 @opindex Wmissing-prototypes
2472 Warn if a global function is defined without a previous prototype
2473 declaration. This warning is issued even if the definition itself
2474 provides a prototype. The aim is to detect global functions that fail
2475 to be declared in header files.
2477 @item -Wmissing-declarations
2478 @opindex Wmissing-declarations
2479 Warn if a global function is defined without a previous declaration.
2480 Do so even if the definition itself provides a prototype.
2481 Use this option to detect global functions that are not declared in
2484 @item -Wmissing-noreturn
2485 @opindex Wmissing-noreturn
2486 Warn about functions which might be candidates for attribute @code{noreturn}.
2487 Note these are only possible candidates, not absolute ones. Care should
2488 be taken to manually verify functions actually do not ever return before
2489 adding the @code{noreturn} attribute, otherwise subtle code generation
2490 bugs could be introduced. You will not get a warning for @code{main} in
2491 hosted C environments.
2493 @item -Wmissing-format-attribute
2494 @opindex Wmissing-format-attribute
2496 If @option{-Wformat} is enabled, also warn about functions which might be
2497 candidates for @code{format} attributes. Note these are only possible
2498 candidates, not absolute ones. GCC will guess that @code{format}
2499 attributes might be appropriate for any function that calls a function
2500 like @code{vprintf} or @code{vscanf}, but this might not always be the
2501 case, and some functions for which @code{format} attributes are
2502 appropriate may not be detected. This option has no effect unless
2503 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2505 @item -Wno-multichar
2506 @opindex Wno-multichar
2508 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2509 Usually they indicate a typo in the user's code, as they have
2510 implementation-defined values, and should not be used in portable code.
2512 @item -Wno-deprecated-declarations
2513 @opindex Wno-deprecated-declarations
2514 Do not warn about uses of functions, variables, and types marked as
2515 deprecated by using the @code{deprecated} attribute.
2516 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2517 @pxref{Type Attributes}.)
2521 Warn if a structure is given the packed attribute, but the packed
2522 attribute has no effect on the layout or size of the structure.
2523 Such structures may be mis-aligned for little benefit. For
2524 instance, in this code, the variable @code{f.x} in @code{struct bar}
2525 will be misaligned even though @code{struct bar} does not itself
2526 have the packed attribute:
2533 @} __attribute__((packed));
2543 Warn if padding is included in a structure, either to align an element
2544 of the structure or to align the whole structure. Sometimes when this
2545 happens it is possible to rearrange the fields of the structure to
2546 reduce the padding and so make the structure smaller.
2548 @item -Wredundant-decls
2549 @opindex Wredundant-decls
2550 Warn if anything is declared more than once in the same scope, even in
2551 cases where multiple declaration is valid and changes nothing.
2553 @item -Wnested-externs @r{(C only)}
2554 @opindex Wnested-externs
2555 Warn if an @code{extern} declaration is encountered within a function.
2557 @item -Wunreachable-code
2558 @opindex Wunreachable-code
2559 Warn if the compiler detects that code will never be executed.
2561 This option is intended to warn when the compiler detects that at
2562 least a whole line of source code will never be executed, because
2563 some condition is never satisfied or because it is after a
2564 procedure that never returns.
2566 It is possible for this option to produce a warning even though there
2567 are circumstances under which part of the affected line can be executed,
2568 so care should be taken when removing apparently-unreachable code.
2570 For instance, when a function is inlined, a warning may mean that the
2571 line is unreachable in only one inlined copy of the function.
2573 This option is not made part of @option{-Wall} because in a debugging
2574 version of a program there is often substantial code which checks
2575 correct functioning of the program and is, hopefully, unreachable
2576 because the program does work. Another common use of unreachable
2577 code is to provide behavior which is selectable at compile-time.
2581 Warn if a function can not be inlined and it was declared as inline.
2585 @opindex Wno-long-long
2586 Warn if @samp{long long} type is used. This is default. To inhibit
2587 the warning messages, use @option{-Wno-long-long}. Flags
2588 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2589 only when @option{-pedantic} flag is used.
2591 @item -Wdisabled-optimization
2592 @opindex Wdisabled-optimization
2593 Warn if a requested optimization pass is disabled. This warning does
2594 not generally indicate that there is anything wrong with your code; it
2595 merely indicates that GCC's optimizers were unable to handle the code
2596 effectively. Often, the problem is that your code is too big or too
2597 complex; GCC will refuse to optimize programs when the optimization
2598 itself is likely to take inordinate amounts of time.
2602 Make all warnings into errors.
2605 @node Debugging Options
2606 @section Options for Debugging Your Program or GCC
2607 @cindex options, debugging
2608 @cindex debugging information options
2610 GCC has various special options that are used for debugging
2611 either your program or GCC:
2616 Produce debugging information in the operating system's native format
2617 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2620 On most systems that use stabs format, @option{-g} enables use of extra
2621 debugging information that only GDB can use; this extra information
2622 makes debugging work better in GDB but will probably make other debuggers
2624 refuse to read the program. If you want to control for certain whether
2625 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2626 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2627 or @option{-gvms} (see below).
2629 Unlike most other C compilers, GCC allows you to use @option{-g} with
2630 @option{-O}. The shortcuts taken by optimized code may occasionally
2631 produce surprising results: some variables you declared may not exist
2632 at all; flow of control may briefly move where you did not expect it;
2633 some statements may not be executed because they compute constant
2634 results or their values were already at hand; some statements may
2635 execute in different places because they were moved out of loops.
2637 Nevertheless it proves possible to debug optimized output. This makes
2638 it reasonable to use the optimizer for programs that might have bugs.
2640 The following options are useful when GCC is generated with the
2641 capability for more than one debugging format.
2645 Produce debugging information for use by GDB@. This means to use the
2646 most expressive format available (DWARF 2, stabs, or the native format
2647 if neither of those are supported), including GDB extensions if at all
2652 Produce debugging information in stabs format (if that is supported),
2653 without GDB extensions. This is the format used by DBX on most BSD
2654 systems. On MIPS, Alpha and System V Release 4 systems this option
2655 produces stabs debugging output which is not understood by DBX or SDB@.
2656 On System V Release 4 systems this option requires the GNU assembler.
2660 Produce debugging information in stabs format (if that is supported),
2661 using GNU extensions understood only by the GNU debugger (GDB)@. The
2662 use of these extensions is likely to make other debuggers crash or
2663 refuse to read the program.
2667 Produce debugging information in COFF format (if that is supported).
2668 This is the format used by SDB on most System V systems prior to
2673 Produce debugging information in XCOFF format (if that is supported).
2674 This is the format used by the DBX debugger on IBM RS/6000 systems.
2678 Produce debugging information in XCOFF format (if that is supported),
2679 using GNU extensions understood only by the GNU debugger (GDB)@. The
2680 use of these extensions is likely to make other debuggers crash or
2681 refuse to read the program, and may cause assemblers other than the GNU
2682 assembler (GAS) to fail with an error.
2686 Produce debugging information in DWARF version 1 format (if that is
2687 supported). This is the format used by SDB on most System V Release 4
2692 Produce debugging information in DWARF version 1 format (if that is
2693 supported), using GNU extensions understood only by the GNU debugger
2694 (GDB)@. The use of these extensions is likely to make other debuggers
2695 crash or refuse to read the program.
2699 Produce debugging information in DWARF version 2 format (if that is
2700 supported). This is the format used by DBX on IRIX 6.
2704 Produce debugging information in VMS debug format (if that is
2705 supported). This is the format used by DEBUG on VMS systems.
2708 @itemx -ggdb@var{level}
2709 @itemx -gstabs@var{level}
2710 @itemx -gcoff@var{level}
2711 @itemx -gxcoff@var{level}
2712 @itemx -gdwarf@var{level}
2713 @itemx -gdwarf-2@var{level}
2714 @itemx -gvms@var{level}
2715 Request debugging information and also use @var{level} to specify how
2716 much information. The default level is 2.
2718 Level 1 produces minimal information, enough for making backtraces in
2719 parts of the program that you don't plan to debug. This includes
2720 descriptions of functions and external variables, but no information
2721 about local variables and no line numbers.
2723 Level 3 includes extra information, such as all the macro definitions
2724 present in the program. Some debuggers support macro expansion when
2725 you use @option{-g3}.
2730 Generate extra code to write profile information suitable for the
2731 analysis program @code{prof}. You must use this option when compiling
2732 the source files you want data about, and you must also use it when
2735 @cindex @code{gprof}
2738 Generate extra code to write profile information suitable for the
2739 analysis program @code{gprof}. You must use this option when compiling
2740 the source files you want data about, and you must also use it when
2746 Generate extra code to write profile information for basic blocks, which will
2747 record the number of times each basic block is executed, the basic block start
2748 address, and the function name containing the basic block. If @option{-g} is
2749 used, the line number and filename of the start of the basic block will also be
2750 recorded. If not overridden by the machine description, the default action is
2751 to append to the text file @file{bb.out}.
2753 This data could be analyzed by a program like @code{tcov}. Note,
2754 however, that the format of the data is not what @code{tcov} expects.
2755 Eventually GNU @code{gprof} should be extended to process this data.
2759 Makes the compiler print out each function name as it is compiled, and
2760 print some statistics about each pass when it finishes.
2763 @opindex ftime-report
2764 Makes the compiler print some statistics about the time consumed by each
2765 pass when it finishes.
2768 @opindex fmem-report
2769 Makes the compiler print some statistics about permanent memory
2770 allocation when it finishes.
2772 @item -fprofile-arcs
2773 @opindex fprofile-arcs
2774 Instrument @dfn{arcs} during compilation to generate coverage data
2775 or for profile-directed block ordering. During execution the program
2776 records how many times each branch is executed and how many times it is
2777 taken. When the compiled program exits it saves this data to a file
2778 called @file{@var{sourcename}.da} for each source file.
2780 For profile-directed block ordering, compile the program with
2781 @option{-fprofile-arcs} plus optimization and code generation options,
2782 generate the arc profile information by running the program on a
2783 selected workload, and then compile the program again with the same
2784 optimization and code generation options plus
2785 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2786 Control Optimization}).
2788 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2789 when it is used with the @option{-ftest-coverage} option. GCC
2790 supports two methods of determining code coverage: the options that
2791 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2792 write information to text files. The options that support @code{gcov}
2793 do not need to instrument every arc in the program, so a program compiled
2794 with them runs faster than a program compiled with @option{-a}, which
2795 adds instrumentation code to every basic block in the program. The
2796 tradeoff: since @code{gcov} does not have execution counts for all
2797 branches, it must start with the execution counts for the instrumented
2798 branches, and then iterate over the program flow graph until the entire
2799 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2800 a program which uses information from @option{-a} and @option{-ax}.
2802 With @option{-fprofile-arcs}, for each function of your program GCC
2803 creates a program flow graph, then finds a spanning tree for the graph.
2804 Only arcs that are not on the spanning tree have to be instrumented: the
2805 compiler adds code to count the number of times that these arcs are
2806 executed. When an arc is the only exit or only entrance to a block, the
2807 instrumentation code can be added to the block; otherwise, a new basic
2808 block must be created to hold the instrumentation code.
2810 This option makes it possible to estimate branch probabilities and to
2811 calculate basic block execution counts. In general, basic block
2812 execution counts as provided by @option{-a} do not give enough
2813 information to estimate all branch probabilities.
2816 @item -ftest-coverage
2817 @opindex ftest-coverage
2818 Create data files for the @code{gcov} code-coverage utility
2819 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2820 The data file names begin with the name of your source file:
2823 @item @var{sourcename}.bb
2824 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2825 associate basic block execution counts with line numbers.
2827 @item @var{sourcename}.bbg
2828 A list of all arcs in the program flow graph. This allows @code{gcov}
2829 to reconstruct the program flow graph, so that it can compute all basic
2830 block and arc execution counts from the information in the
2831 @code{@var{sourcename}.da} file.
2834 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2835 option adds instrumentation to the program, which then writes
2836 execution counts to another data file:
2839 @item @var{sourcename}.da
2840 Runtime arc execution counts, used in conjunction with the arc
2841 information in the file @code{@var{sourcename}.bbg}.
2844 Coverage data will map better to the source files if
2845 @option{-ftest-coverage} is used without optimization.
2847 @item -d@var{letters}
2849 Says to make debugging dumps during compilation at times specified by
2850 @var{letters}. This is used for debugging the compiler. The file names
2851 for most of the dumps are made by appending a pass number and a word to
2852 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2853 Here are the possible letters for use in @var{letters}, and their meanings:
2858 Annotate the assembler output with miscellaneous debugging information.
2861 Dump after computing branch probabilities, to @file{@var{file}.15.bp}.
2864 Dump after block reordering, to @file{@var{file}.30.bbro}.
2867 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
2870 Dump after the first if conversion, to the file @file{@var{file}.18.ce}.
2873 Dump after delayed branch scheduling, to @file{@var{file}.32.dbr}.
2876 Dump all macro definitions, at the end of preprocessing, in addition to
2880 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2881 @file{@var{file}.07.ussa}.
2884 Dump after the second if conversion, to @file{@var{file}.27.ce2}.
2887 Dump after life analysis, to @file{@var{file}.16.life}.
2890 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.10.addressof}.
2893 Dump after global register allocation, to @file{@var{file}.22.greg}.
2896 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2899 Dump after reg-to-stack conversion, to @file{@var{file}.29.stack}.
2902 Dump after post-reload optimizations, to @file{@var{file}.23.postreload}.
2905 Dump after GCSE, to @file{@var{file}.11.gcse}.
2908 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2911 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2914 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
2917 Dump after local register allocation, to @file{@var{file}.21.lreg}.
2920 Dump after loop optimization, to @file{@var{file}.12.loop}.
2923 Dump after performing the machine dependent reorganisation pass, to
2924 @file{@var{file}.31.mach}.
2927 Dump after register renumbering, to @file{@var{file}.26.rnreg}.
2930 Dump after the register move pass, to @file{@var{file}.19.regmove}.
2933 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2936 Dump after the second scheduling pass, to @file{@var{file}.28.sched2}.
2939 Dump after CSE (including the jump optimization that sometimes follows
2940 CSE), to @file{@var{file}.09.cse}.
2943 Dump after the first scheduling pass, to @file{@var{file}.20.sched}.
2946 Dump after the second CSE pass (including the jump optimization that
2947 sometimes follows CSE), to @file{@var{file}.13.cse2}.
2950 Dump after null pointer ellimination pass ti @file{@var{file}.08.null}.
2953 Dump after the second flow pass, to @file{@var{file}.24.flow2}.
2956 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2959 Dump after the peephole pass, to @file{@var{file}.25.peephole2}.
2962 Produce all the dumps listed above.
2965 Print statistics on memory usage, at the end of the run, to
2969 Annotate the assembler output with a comment indicating which
2970 pattern and alternative was used. The length of each instruction is
2974 Dump the RTL in the assembler output as a comment before each instruction.
2975 Also turns on @option{-dp} annotation.
2978 For each of the other indicated dump files (except for
2979 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
2980 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
2983 Just generate RTL for a function instead of compiling it. Usually used
2987 Dump debugging information during parsing, to standard error.
2990 @item -fdump-unnumbered
2991 @opindex fdump-unnumbered
2992 When doing debugging dumps (see @option{-d} option above), suppress instruction
2993 numbers and line number note output. This makes it more feasible to
2994 use diff on debugging dumps for compiler invocations with different
2995 options, in particular with and without @option{-g}.
2997 @item -fdump-translation-unit @r{(C and C++ only)}
2998 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
2999 @opindex fdump-translation-unit
3000 Dump a representation of the tree structure for the entire translation
3001 unit to a file. The file name is made by appending @file{.tu} to the
3002 source file name. If the @samp{-@var{options}} form is used, @var{options}
3003 controls the details of the dump as described for the
3004 @option{-fdump-tree} options.
3006 @item -fdump-class-hierarchy @r{(C++ only)}
3007 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3008 @opindex fdump-class-hierarchy
3009 Dump a representation of each class's hierarchy and virtual function
3010 table layout to a file. The file name is made by appending @file{.class}
3011 to the source file name. If the @samp{-@var{options}} form is used,
3012 @var{options} controls the details of the dump as described for the
3013 @option{-fdump-tree} options.
3015 @item -fdump-tree-@var{switch} @r{(C++ only)}
3016 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3018 Control the dumping at various stages of processing the intermediate
3019 language tree to a file. The file name is generated by appending a switch
3020 specific suffix to the source file name. If the @samp{-@var{options}}
3021 form is used, @var{options} is a list of @samp{-} separated options that
3022 control the details of the dump. Not all options are applicable to all
3023 dumps, those which are not meaningful will be ignored. The following
3024 options are available
3028 Print the address of each node. Usually this is not meaningful as it
3029 changes according to the environment and source file. Its primary use
3030 is for tying up a dump file with a debug environment.
3032 Inhibit dumping of members of a scope or body of a function merely
3033 because that scope has been reached. Only dump such items when they
3034 are directly reachable by some other path.
3036 Turn on all options.
3039 The following tree dumps are possible:
3042 Dump before any tree based optimization, to @file{@var{file}.original}.
3044 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3046 Dump after function inlining, to @file{@var{file}.inlined}.
3051 Store the usual ``temporary'' intermediate files permanently; place them
3052 in the current directory and name them based on the source file. Thus,
3053 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3054 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3055 preprocessed @file{foo.i} output file even though the compiler now
3056 normally uses an integrated preprocessor.
3060 Report the CPU time taken by each subprocess in the compilation
3061 sequence. For C source files, this is the compiler proper and assembler
3062 (plus the linker if linking is done). The output looks like this:
3069 The first number on each line is the ``user time,'' that is time spent
3070 executing the program itself. The second number is ``system time,''
3071 time spent executing operating system routines on behalf of the program.
3072 Both numbers are in seconds.
3074 @item -print-file-name=@var{library}
3075 @opindex print-file-name
3076 Print the full absolute name of the library file @var{library} that
3077 would be used when linking---and don't do anything else. With this
3078 option, GCC does not compile or link anything; it just prints the
3081 @item -print-multi-directory
3082 @opindex print-multi-directory
3083 Print the directory name corresponding to the multilib selected by any
3084 other switches present in the command line. This directory is supposed
3085 to exist in @env{GCC_EXEC_PREFIX}.
3087 @item -print-multi-lib
3088 @opindex print-multi-lib
3089 Print the mapping from multilib directory names to compiler switches
3090 that enable them. The directory name is separated from the switches by
3091 @samp{;}, and each switch starts with an @samp{@@} instead of the
3092 @samp{-}, without spaces between multiple switches. This is supposed to
3093 ease shell-processing.
3095 @item -print-prog-name=@var{program}
3096 @opindex print-prog-name
3097 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3099 @item -print-libgcc-file-name
3100 @opindex print-libgcc-file-name
3101 Same as @option{-print-file-name=libgcc.a}.
3103 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3104 but you do want to link with @file{libgcc.a}. You can do
3107 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3110 @item -print-search-dirs
3111 @opindex print-search-dirs
3112 Print the name of the configured installation directory and a list of
3113 program and library directories gcc will search---and don't do anything else.
3115 This is useful when gcc prints the error message
3116 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3117 To resolve this you either need to put @file{cpp0} and the other compiler
3118 components where gcc expects to find them, or you can set the environment
3119 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3120 Don't forget the trailing '/'.
3121 @xref{Environment Variables}.
3124 @opindex dumpmachine
3125 Print the compiler's target machine (for example,
3126 @samp{i686-pc-linux-gnu})---and don't do anything else.
3129 @opindex dumpversion
3130 Print the compiler version (for example, @samp{3.0})---and don't do
3135 Print the compiler's built-in specs---and don't do anything else. (This
3136 is used when GCC itself is being built.) @xref{Spec Files}.
3139 @node Optimize Options
3140 @section Options That Control Optimization
3141 @cindex optimize options
3142 @cindex options, optimization
3144 These options control various sorts of optimizations:
3151 Optimize. Optimizing compilation takes somewhat more time, and a lot
3152 more memory for a large function.
3154 Without @option{-O}, the compiler's goal is to reduce the cost of
3155 compilation and to make debugging produce the expected results.
3156 Statements are independent: if you stop the program with a breakpoint
3157 between statements, you can then assign a new value to any variable or
3158 change the program counter to any other statement in the function and
3159 get exactly the results you would expect from the source code.
3161 With @option{-O}, the compiler tries to reduce code size and execution
3162 time, without performing any optimizations that take a great deal of
3167 Optimize even more. GCC performs nearly all supported optimizations
3168 that do not involve a space-speed tradeoff. The compiler does not
3169 perform loop unrolling or function inlining when you specify @option{-O2}.
3170 As compared to @option{-O}, this option increases both compilation time
3171 and the performance of the generated code.
3173 @option{-O2} turns on all optional optimizations except for loop unrolling,
3174 function inlining, and register renaming. It also turns on the
3175 @option{-fforce-mem} option on all machines and frame pointer elimination
3176 on machines where doing so does not interfere with debugging.
3178 Please note the warning under @option{-fgcse} about
3179 invoking @option{-O2} on programs that use computed gotos.
3183 Optimize yet more. @option{-O3} turns on all optimizations specified by
3184 @option{-O2} and also turns on the @option{-finline-functions} and
3185 @option{-frename-registers} options.
3193 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3194 do not typically increase code size. It also performs further
3195 optimizations designed to reduce code size.
3197 If you use multiple @option{-O} options, with or without level numbers,
3198 the last such option is the one that is effective.
3201 Options of the form @option{-f@var{flag}} specify machine-independent
3202 flags. Most flags have both positive and negative forms; the negative
3203 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3204 only one of the forms is listed---the one which is not the default.
3205 You can figure out the other form by either removing @samp{no-} or
3210 @opindex ffloat-store
3211 Do not store floating point variables in registers, and inhibit other
3212 options that might change whether a floating point value is taken from a
3215 @cindex floating point precision
3216 This option prevents undesirable excess precision on machines such as
3217 the 68000 where the floating registers (of the 68881) keep more
3218 precision than a @code{double} is supposed to have. Similarly for the
3219 x86 architecture. For most programs, the excess precision does only
3220 good, but a few programs rely on the precise definition of IEEE floating
3221 point. Use @option{-ffloat-store} for such programs, after modifying
3222 them to store all pertinent intermediate computations into variables.
3224 @item -fno-default-inline
3225 @opindex fno-default-inline
3226 Do not make member functions inline by default merely because they are
3227 defined inside the class scope (C++ only). Otherwise, when you specify
3228 @w{@option{-O}}, member functions defined inside class scope are compiled
3229 inline by default; i.e., you don't need to add @samp{inline} in front of
3230 the member function name.
3232 @item -fno-defer-pop
3233 @opindex fno-defer-pop
3234 Always pop the arguments to each function call as soon as that function
3235 returns. For machines which must pop arguments after a function call,
3236 the compiler normally lets arguments accumulate on the stack for several
3237 function calls and pops them all at once.
3241 Force memory operands to be copied into registers before doing
3242 arithmetic on them. This produces better code by making all memory
3243 references potential common subexpressions. When they are not common
3244 subexpressions, instruction combination should eliminate the separate
3245 register-load. The @option{-O2} option turns on this option.
3248 @opindex fforce-addr
3249 Force memory address constants to be copied into registers before
3250 doing arithmetic on them. This may produce better code just as
3251 @option{-fforce-mem} may.
3253 @item -fomit-frame-pointer
3254 @opindex fomit-frame-pointer
3255 Don't keep the frame pointer in a register for functions that
3256 don't need one. This avoids the instructions to save, set up and
3257 restore frame pointers; it also makes an extra register available
3258 in many functions. @strong{It also makes debugging impossible on
3261 On some machines, such as the VAX, this flag has no effect, because
3262 the standard calling sequence automatically handles the frame pointer
3263 and nothing is saved by pretending it doesn't exist. The
3264 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3265 whether a target machine supports this flag. @xref{Registers,,Register
3266 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3268 @item -foptimize-sibling-calls
3269 @opindex foptimize-sibling-calls
3270 Optimize sibling and tail recursive calls.
3274 This option generates traps for signed overflow on addition, subtraction,
3275 multiplication operations.
3279 Don't pay attention to the @code{inline} keyword. Normally this option
3280 is used to keep the compiler from expanding any functions inline.
3281 Note that if you are not optimizing, no functions can be expanded inline.
3283 @item -finline-functions
3284 @opindex finline-functions
3285 Integrate all simple functions into their callers. The compiler
3286 heuristically decides which functions are simple enough to be worth
3287 integrating in this way.
3289 If all calls to a given function are integrated, and the function is
3290 declared @code{static}, then the function is normally not output as
3291 assembler code in its own right.
3293 @item -finline-limit=@var{n}
3294 @opindex finline-limit
3295 By default, gcc limits the size of functions that can be inlined. This flag
3296 allows the control of this limit for functions that are explicitly marked as
3297 inline (ie marked with the inline keyword or defined within the class
3298 definition in c++). @var{n} is the size of functions that can be inlined in
3299 number of pseudo instructions (not counting parameter handling). The default
3300 value of @var{n} is 600.
3301 Increasing this value can result in more inlined code at
3302 the cost of compilation time and memory consumption. Decreasing usually makes
3303 the compilation faster and less code will be inlined (which presumably
3304 means slower programs). This option is particularly useful for programs that
3305 use inlining heavily such as those based on recursive templates with C++.
3307 @emph{Note:} pseudo instruction represents, in this particular context, an
3308 abstract measurement of function's size. In no way, it represents a count
3309 of assembly instructions and as such its exact meaning might change from one
3310 release to an another.
3312 @item -fkeep-inline-functions
3313 @opindex fkeep-inline-functions
3314 Even if all calls to a given function are integrated, and the function
3315 is declared @code{static}, nevertheless output a separate run-time
3316 callable version of the function. This switch does not affect
3317 @code{extern inline} functions.
3319 @item -fkeep-static-consts
3320 @opindex fkeep-static-consts
3321 Emit variables declared @code{static const} when optimization isn't turned
3322 on, even if the variables aren't referenced.
3324 GCC enables this option by default. If you want to force the compiler to
3325 check if the variable was referenced, regardless of whether or not
3326 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3328 @item -fmerge-constants
3329 Attempt to merge identical constants (string constants and floating point
3330 constants) accross compilation units.
3332 This option is default for optimized compilation if assembler and linker
3333 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3335 @item -fmerge-all-constants
3336 Attempt to merge identical constants and identical variables.
3338 This option implies @option{-fmerge-constants}. In addition to
3339 @option{-fmerge-constants} this considers e.g. even constant initialized
3340 arrays or initialized constant variables with integral or floating point
3341 types. Languages like C or C++ require each non-automatic variable to
3342 have distinct location, so using this option will result in non-conforming
3345 @item -fno-function-cse
3346 @opindex fno-function-cse
3347 Do not put function addresses in registers; make each instruction that
3348 calls a constant function contain the function's address explicitly.
3350 This option results in less efficient code, but some strange hacks
3351 that alter the assembler output may be confused by the optimizations
3352 performed when this option is not used.
3356 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3357 @option{-fno-trapping-math}.
3359 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3361 This option should never be turned on by any @option{-O} option since
3362 it can result in incorrect output for programs which depend on
3363 an exact implementation of IEEE or ISO rules/specifications for
3366 @item -fno-math-errno
3367 @opindex fno-math-errno
3368 Do not set ERRNO after calling math functions that are executed
3369 with a single instruction, e.g., sqrt. A program that relies on
3370 IEEE exceptions for math error handling may want to use this flag
3371 for speed while maintaining IEEE arithmetic compatibility.
3373 This option should never be turned on by any @option{-O} option since
3374 it can result in incorrect output for programs which depend on
3375 an exact implementation of IEEE or ISO rules/specifications for
3378 The default is @option{-fmath-errno}.
3380 @item -funsafe-math-optimizations
3381 @opindex funsafe-math-optimizations
3382 Allow optimizations for floating-point arithmetic that (a) assume
3383 that arguments and results are valid and (b) may violate IEEE or
3384 ANSI standards. When used at link-time, it may include libraries
3385 or startup files that change the default FPU control word or other
3386 similar optimizations.
3388 This option should never be turned on by any @option{-O} option since
3389 it can result in incorrect output for programs which depend on
3390 an exact implementation of IEEE or ISO rules/specifications for
3393 The default is @option{-fno-unsafe-math-optimizations}.
3395 @item -fno-trapping-math
3396 @opindex fno-trapping-math
3397 Compile code assuming that floating-point operations cannot generate
3398 user-visible traps. Setting this option may allow faster code
3399 if one relies on ``non-stop'' IEEE arithmetic, for example.
3401 This option should never be turned on by any @option{-O} option since
3402 it can result in incorrect output for programs which depend on
3403 an exact implementation of IEEE or ISO rules/specifications for
3406 The default is @option{-ftrapping-math}.
3408 @item -fno-zero-initialized-in-bss
3409 @opindex fno-zero-initialized-in-bss
3410 If the target supports a BSS section, GCC by default puts variables that
3411 are initialized to zero into BSS@. This can save space in the resulting
3414 This option turns off this behavior because some programs explicitly
3415 rely on variables going to the data section. E.g., so that the
3416 resulting executable can find the beginning of that section and/or make
3417 assumptions based on that.
3419 The default is @option{-fzero-initialized-in-bss}.
3422 The following options control specific optimizations. The @option{-O2}
3423 option turns on all of these optimizations except @option{-funroll-loops}
3424 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3425 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3426 but specific machines may handle it differently.
3428 You can use the following flags in the rare cases when ``fine-tuning''
3429 of optimizations to be performed is desired.
3431 Not all of the optimizations performed by GCC have @option{-f} options
3435 @item -fstrength-reduce
3436 @opindex fstrength-reduce
3437 Perform the optimizations of loop strength reduction and
3438 elimination of iteration variables.
3440 @item -fthread-jumps
3441 @opindex fthread-jumps
3442 Perform optimizations where we check to see if a jump branches to a
3443 location where another comparison subsumed by the first is found. If
3444 so, the first branch is redirected to either the destination of the
3445 second branch or a point immediately following it, depending on whether
3446 the condition is known to be true or false.
3448 @item -fcse-follow-jumps
3449 @opindex fcse-follow-jumps
3450 In common subexpression elimination, scan through jump instructions
3451 when the target of the jump is not reached by any other path. For
3452 example, when CSE encounters an @code{if} statement with an
3453 @code{else} clause, CSE will follow the jump when the condition
3456 @item -fcse-skip-blocks
3457 @opindex fcse-skip-blocks
3458 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3459 follow jumps which conditionally skip over blocks. When CSE
3460 encounters a simple @code{if} statement with no else clause,
3461 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3462 body of the @code{if}.
3464 @item -frerun-cse-after-loop
3465 @opindex frerun-cse-after-loop
3466 Re-run common subexpression elimination after loop optimizations has been
3469 @item -frerun-loop-opt
3470 @opindex frerun-loop-opt
3471 Run the loop optimizer twice.
3475 Perform a global common subexpression elimination pass.
3476 This pass also performs global constant and copy propagation.
3478 @emph{Note:} When compiling a program using computed gotos, a GCC
3479 extension, you may get better runtime performance if you disable
3480 the global common subexpression elmination pass by adding
3481 @option{-fno-gcse} to the command line.
3485 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3486 attempt to move loads which are only killed by stores into themselves. This
3487 allows a loop containing a load/store sequence to be changed to a load outside
3488 the loop, and a copy/store within the loop.
3492 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3493 subexpression elimination. This pass will attempt to move stores out of loops.
3494 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3495 can be changed to a load before the loop and a store after the loop.
3497 @item -fdelete-null-pointer-checks
3498 @opindex fdelete-null-pointer-checks
3499 Use global dataflow analysis to identify and eliminate useless checks
3500 for null pointers. The compiler assumes that dereferencing a null
3501 pointer would have halted the program. If a pointer is checked after
3502 it has already been dereferenced, it cannot be null.
3504 In some environments, this assumption is not true, and programs can
3505 safely dereference null pointers. Use
3506 @option{-fno-delete-null-pointer-checks} to disable this optimization
3507 for programs which depend on that behavior.
3509 @item -fexpensive-optimizations
3510 @opindex fexpensive-optimizations
3511 Perform a number of minor optimizations that are relatively expensive.
3513 @item -foptimize-register-move
3515 @opindex foptimize-register-move
3517 Attempt to reassign register numbers in move instructions and as
3518 operands of other simple instructions in order to maximize the amount of
3519 register tying. This is especially helpful on machines with two-operand
3520 instructions. GCC enables this optimization by default with @option{-O2}
3523 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3526 @item -fdelayed-branch
3527 @opindex fdelayed-branch
3528 If supported for the target machine, attempt to reorder instructions
3529 to exploit instruction slots available after delayed branch
3532 @item -fschedule-insns
3533 @opindex fschedule-insns
3534 If supported for the target machine, attempt to reorder instructions to
3535 eliminate execution stalls due to required data being unavailable. This
3536 helps machines that have slow floating point or memory load instructions
3537 by allowing other instructions to be issued until the result of the load
3538 or floating point instruction is required.
3540 @item -fschedule-insns2
3541 @opindex fschedule-insns2
3542 Similar to @option{-fschedule-insns}, but requests an additional pass of
3543 instruction scheduling after register allocation has been done. This is
3544 especially useful on machines with a relatively small number of
3545 registers and where memory load instructions take more than one cycle.
3547 @item -ffunction-sections
3548 @itemx -fdata-sections
3549 @opindex ffunction-sections
3550 @opindex fdata-sections
3551 Place each function or data item into its own section in the output
3552 file if the target supports arbitrary sections. The name of the
3553 function or the name of the data item determines the section's name
3556 Use these options on systems where the linker can perform optimizations
3557 to improve locality of reference in the instruction space. HPPA
3558 processors running HP-UX and Sparc processors running Solaris 2 have
3559 linkers with such optimizations. Other systems using the ELF object format
3560 as well as AIX may have these optimizations in the future.
3562 Only use these options when there are significant benefits from doing
3563 so. When you specify these options, the assembler and linker will
3564 create larger object and executable files and will also be slower.
3565 You will not be able to use @code{gprof} on all systems if you
3566 specify this option and you may have problems with debugging if
3567 you specify both this option and @option{-g}.
3569 @item -fcaller-saves
3570 @opindex fcaller-saves
3571 Enable values to be allocated in registers that will be clobbered by
3572 function calls, by emitting extra instructions to save and restore the
3573 registers around such calls. Such allocation is done only when it
3574 seems to result in better code than would otherwise be produced.
3576 This option is always enabled by default on certain machines, usually
3577 those which have no call-preserved registers to use instead.
3579 For all machines, optimization level 2 and higher enables this flag by
3582 @item -funroll-loops
3583 @opindex funroll-loops
3584 Unroll loops whose number of iterations can be determined at compile
3585 time or upon entry to the loop. @option{-funroll-loops} implies both
3586 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3587 option makes code larger, and may or may not make it run faster.
3589 @item -funroll-all-loops
3590 @opindex funroll-all-loops
3591 Unroll all loops, even if their number of iterations is uncertain when
3592 the loop is entered. This usually makes programs run more slowly.
3593 @option{-funroll-all-loops} implies the same options as
3594 @option{-funroll-loops},
3596 @item -fprefetch-loop-arrays
3597 @opindex fprefetch-loop-arrays
3598 If supported by the target machine, generate instructions to prefetch
3599 memory to improve the performance of loops that access large arrays.
3601 @item -fmove-all-movables
3602 @opindex fmove-all-movables
3603 Forces all invariant computations in loops to be moved
3606 @item -freduce-all-givs
3607 @opindex freduce-all-givs
3608 Forces all general-induction variables in loops to be
3611 @emph{Note:} When compiling programs written in Fortran,
3612 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3613 by default when you use the optimizer.
3615 These options may generate better or worse code; results are highly
3616 dependent on the structure of loops within the source code.
3618 These two options are intended to be removed someday, once
3619 they have helped determine the efficacy of various
3620 approaches to improving loop optimizations.
3622 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3623 know how use of these options affects
3624 the performance of your production code.
3625 We're very interested in code that runs @emph{slower}
3626 when these options are @emph{enabled}.
3629 @itemx -fno-peephole2
3630 @opindex fno-peephole
3631 @opindex fno-peephole2
3632 Disable any machine-specific peephole optimizations. The difference
3633 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3634 are implemented in the compiler; some targets use one, some use the
3635 other, a few use both.
3637 @item -fbranch-probabilities
3638 @opindex fbranch-probabilities
3639 After running a program compiled with @option{-fprofile-arcs}
3640 (@pxref{Debugging Options,, Options for Debugging Your Program or
3641 @command{gcc}}), you can compile it a second time using
3642 @option{-fbranch-probabilities}, to improve optimizations based on
3643 the number of times each branch was taken. When the program
3644 compiled with @option{-fprofile-arcs} exits it saves arc execution
3645 counts to a file called @file{@var{sourcename}.da} for each source
3646 file The information in this data file is very dependent on the
3647 structure of the generated code, so you must use the same source code
3648 and the same optimization options for both compilations.
3650 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3651 note on the first instruction of each basic block, and a
3652 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3653 These can be used to improve optimization. Currently, they are only
3654 used in one place: in @file{reorg.c}, instead of guessing which path a
3655 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3656 exactly determine which path is taken more often.
3658 @item -fno-guess-branch-probability
3659 @opindex fno-guess-branch-probability
3660 Do not guess branch probabilities using a randomized model.
3662 Sometimes gcc will opt to use a randomized model to guess branch
3663 probabilities, when none are available from either profiling feedback
3664 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3665 different runs of the compiler on the same program may produce different
3668 In a hard real-time system, people don't want different runs of the
3669 compiler to produce code that has different behavior; minimizing
3670 non-determinism is of paramount import. This switch allows users to
3671 reduce non-determinism, possibly at the expense of inferior
3674 @item -fstrict-aliasing
3675 @opindex fstrict-aliasing
3676 Allows the compiler to assume the strictest aliasing rules applicable to
3677 the language being compiled. For C (and C++), this activates
3678 optimizations based on the type of expressions. In particular, an
3679 object of one type is assumed never to reside at the same address as an
3680 object of a different type, unless the types are almost the same. For
3681 example, an @code{unsigned int} can alias an @code{int}, but not a
3682 @code{void*} or a @code{double}. A character type may alias any other
3685 Pay special attention to code like this:
3698 The practice of reading from a different union member than the one most
3699 recently written to (called ``type-punning'') is common. Even with
3700 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3701 is accessed through the union type. So, the code above will work as
3702 expected. However, this code might not:
3713 Every language that wishes to perform language-specific alias analysis
3714 should define a function that computes, given an @code{tree}
3715 node, an alias set for the node. Nodes in different alias sets are not
3716 allowed to alias. For an example, see the C front-end function
3717 @code{c_get_alias_set}.
3719 @item -falign-functions
3720 @itemx -falign-functions=@var{n}
3721 @opindex falign-functions
3722 Align the start of functions to the next power-of-two greater than
3723 @var{n}, skipping up to @var{n} bytes. For instance,
3724 @option{-falign-functions=32} aligns functions to the next 32-byte
3725 boundary, but @option{-falign-functions=24} would align to the next
3726 32-byte boundary only if this can be done by skipping 23 bytes or less.
3728 @option{-fno-align-functions} and @option{-falign-functions=1} are
3729 equivalent and mean that functions will not be aligned.
3731 Some assemblers only support this flag when @var{n} is a power of two;
3732 in that case, it is rounded up.
3734 If @var{n} is not specified, use a machine-dependent default.
3736 @item -falign-labels
3737 @itemx -falign-labels=@var{n}
3738 @opindex falign-labels
3739 Align all branch targets to a power-of-two boundary, skipping up to
3740 @var{n} bytes like @option{-falign-functions}. This option can easily
3741 make code slower, because it must insert dummy operations for when the
3742 branch target is reached in the usual flow of the code.
3744 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3745 are greater than this value, then their values are used instead.
3747 If @var{n} is not specified, use a machine-dependent default which is
3748 very likely to be @samp{1}, meaning no alignment.
3751 @itemx -falign-loops=@var{n}
3752 @opindex falign-loops
3753 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3754 like @option{-falign-functions}. The hope is that the loop will be
3755 executed many times, which will make up for any execution of the dummy
3758 If @var{n} is not specified, use a machine-dependent default.
3761 @itemx -falign-jumps=@var{n}
3762 @opindex falign-jumps
3763 Align branch targets to a power-of-two boundary, for branch targets
3764 where the targets can only be reached by jumping, skipping up to @var{n}
3765 bytes like @option{-falign-functions}. In this case, no dummy operations
3768 If @var{n} is not specified, use a machine-dependent default.
3772 Perform optimizations in static single assignment form. Each function's
3773 flow graph is translated into SSA form, optimizations are performed, and
3774 the flow graph is translated back from SSA form. Users should not
3775 specify this option, since it is not yet ready for production use.
3779 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3780 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3784 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3785 Like @option{-fssa}, this is an experimental feature.
3787 @item -fsingle-precision-constant
3788 @opindex fsingle-precision-constant
3789 Treat floating point constant as single precision constant instead of
3790 implicitly converting it to double precision constant.
3792 @item -frename-registers
3793 @opindex frename-registers
3794 Attempt to avoid false dependencies in scheduled code by making use
3795 of registers left over after register allocation. This optimization
3796 will most benefit processors with lots of registers. It can, however,
3797 make debugging impossible, since variables will no longer stay in
3798 a ``home register''.
3800 @item -fno-cprop-registers
3801 @opindex fno-cprop-registers
3802 After register allocation and post-register allocation instruction splitting,
3803 we perform a copy-propagation pass to try to reduce scheduling dependencies
3804 and occasionally eliminate the copy.
3806 @item --param @var{name}=@var{value}
3808 In some places, GCC uses various constants to control the amount of
3809 optimization that is done. For example, GCC will not inline functions
3810 that contain more that a certain number of instructions. You can
3811 control some of these constants on the command-line using the
3812 @option{--param} option.
3814 In each case, the @var{value} is an integer. The allowable choices for
3815 @var{name} are given in the following table:
3818 @item max-delay-slot-insn-search
3819 The maximum number of instructions to consider when looking for an
3820 instruction to fill a delay slot. If more than this arbitrary number of
3821 instructions is searched, the time savings from filling the delay slot
3822 will be minimal so stop searching. Increasing values mean more
3823 aggressive optimization, making the compile time increase with probably
3824 small improvement in executable run time.
3826 @item max-delay-slot-live-search
3827 When trying to fill delay slots, the maximum number of instructions to
3828 consider when searching for a block with valid live register
3829 information. Increasing this arbitrarily chosen value means more
3830 aggressive optimization, increasing the compile time. This parameter
3831 should be removed when the delay slot code is rewritten to maintain the
3834 @item max-gcse-memory
3835 The approximate maximum amount of memory that will be allocated in
3836 order to perform the global common subexpression elimination
3837 optimization. If more memory than specified is required, the
3838 optimization will not be done.
3840 @item max-gcse-passes
3841 The maximum number of passes of GCSE to run.
3843 @item max-pending-list-length
3844 The maximum number of pending dependencies scheduling will allow
3845 before flushing the current state and starting over. Large functions
3846 with few branches or calls can create excessively large lists which
3847 needlessly consume memory and resources.
3849 @item max-inline-insns
3850 If an function contains more than this many instructions, it
3851 will not be inlined. This option is precisely equivalent to
3852 @option{-finline-limit}.
3854 @item max-unrolled-insns
3855 The maximum number of instructions that a loop should have if that loop
3856 is unrolled, and if the loop is unrolled, it determines how many times
3857 the loop code is unrolled.
3862 @node Preprocessor Options
3863 @section Options Controlling the Preprocessor
3864 @cindex preprocessor options
3865 @cindex options, preprocessor
3867 These options control the C preprocessor, which is run on each C source
3868 file before actual compilation.
3870 If you use the @option{-E} option, nothing is done except preprocessing.
3871 Some of these options make sense only together with @option{-E} because
3872 they cause the preprocessor output to be unsuitable for actual
3876 You can use @option{-Wp,@var{option}} to bypass the compiler driver
3877 and pass @var{option} directly through to the preprocessor. If
3878 @var{option} contains commas, it is split into multiple options at the
3879 commas. However, many options are modified, translated or interpreted
3880 by the compiler driver before being passed to the preprocessor, and
3881 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
3882 interface is undocumented and subject to change, so whenever possible
3883 you should avoid using @option{-Wp} and let the driver handle the
3886 @include cppopts.texi
3888 @node Assembler Options
3889 @section Passing Options to the Assembler
3891 @c prevent bad page break with this line
3892 You can pass options to the assembler.
3895 @item -Wa,@var{option}
3897 Pass @var{option} as an option to the assembler. If @var{option}
3898 contains commas, it is split into multiple options at the commas.
3902 @section Options for Linking
3903 @cindex link options
3904 @cindex options, linking
3906 These options come into play when the compiler links object files into
3907 an executable output file. They are meaningless if the compiler is
3908 not doing a link step.
3912 @item @var{object-file-name}
3913 A file name that does not end in a special recognized suffix is
3914 considered to name an object file or library. (Object files are
3915 distinguished from libraries by the linker according to the file
3916 contents.) If linking is done, these object files are used as input
3925 If any of these options is used, then the linker is not run, and
3926 object file names should not be used as arguments. @xref{Overall
3930 @item -l@var{library}
3931 @itemx -l @var{library}
3933 Search the library named @var{library} when linking. (The second
3934 alternative with the library as a separate argument is only for
3935 POSIX compliance and is not recommended.)
3937 It makes a difference where in the command you write this option; the
3938 linker searches and processes libraries and object files in the order they
3939 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
3940 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
3941 to functions in @samp{z}, those functions may not be loaded.
3943 The linker searches a standard list of directories for the library,
3944 which is actually a file named @file{lib@var{library}.a}. The linker
3945 then uses this file as if it had been specified precisely by name.
3947 The directories searched include several standard system directories
3948 plus any that you specify with @option{-L}.
3950 Normally the files found this way are library files---archive files
3951 whose members are object files. The linker handles an archive file by
3952 scanning through it for members which define symbols that have so far
3953 been referenced but not defined. But if the file that is found is an
3954 ordinary object file, it is linked in the usual fashion. The only
3955 difference between using an @option{-l} option and specifying a file name
3956 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
3957 and searches several directories.
3961 You need this special case of the @option{-l} option in order to
3962 link an Objective-C program.
3965 @opindex nostartfiles
3966 Do not use the standard system startup files when linking.
3967 The standard system libraries are used normally, unless @option{-nostdlib}
3968 or @option{-nodefaultlibs} is used.
3970 @item -nodefaultlibs
3971 @opindex nodefaultlibs
3972 Do not use the standard system libraries when linking.
3973 Only the libraries you specify will be passed to the linker.
3974 The standard startup files are used normally, unless @option{-nostartfiles}
3975 is used. The compiler may generate calls to memcmp, memset, and memcpy
3976 for System V (and ISO C) environments or to bcopy and bzero for
3977 BSD environments. These entries are usually resolved by entries in
3978 libc. These entry points should be supplied through some other
3979 mechanism when this option is specified.
3983 Do not use the standard system startup files or libraries when linking.
3984 No startup files and only the libraries you specify will be passed to
3985 the linker. The compiler may generate calls to memcmp, memset, and memcpy
3986 for System V (and ISO C) environments or to bcopy and bzero for
3987 BSD environments. These entries are usually resolved by entries in
3988 libc. These entry points should be supplied through some other
3989 mechanism when this option is specified.
3991 @cindex @option{-lgcc}, use with @option{-nostdlib}
3992 @cindex @option{-nostdlib} and unresolved references
3993 @cindex unresolved references and @option{-nostdlib}
3994 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
3995 @cindex @option{-nodefaultlibs} and unresolved references
3996 @cindex unresolved references and @option{-nodefaultlibs}
3997 One of the standard libraries bypassed by @option{-nostdlib} and
3998 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
3999 that GCC uses to overcome shortcomings of particular machines, or special
4000 needs for some languages.
4001 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4002 Collection (GCC) Internals},
4003 for more discussion of @file{libgcc.a}.)
4004 In most cases, you need @file{libgcc.a} even when you want to avoid
4005 other standard libraries. In other words, when you specify @option{-nostdlib}
4006 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4007 This ensures that you have no unresolved references to internal GCC
4008 library subroutines. (For example, @samp{__main}, used to ensure C++
4009 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4010 GNU Compiler Collection (GCC) Internals}.)
4014 Remove all symbol table and relocation information from the executable.
4018 On systems that support dynamic linking, this prevents linking with the shared
4019 libraries. On other systems, this option has no effect.
4023 Produce a shared object which can then be linked with other objects to
4024 form an executable. Not all systems support this option. For predictable
4025 results, you must also specify the same set of options that were used to
4026 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4027 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4028 needs to build supplementary stub code for constructors to work. On
4029 multi-libbed systems, @samp{gcc -shared} must select the correct support
4030 libraries to link against. Failing to supply the correct flags may lead
4031 to subtle defects. Supplying them in cases where they are not necessary
4034 @item -shared-libgcc
4035 @itemx -static-libgcc
4036 @opindex shared-libgcc
4037 @opindex static-libgcc
4038 On systems that provide @file{libgcc} as a shared library, these options
4039 force the use of either the shared or static version respectively.
4040 If no shared version of @file{libgcc} was built when the compiler was
4041 configured, these options have no effect.
4043 There are several situations in which an application should use the
4044 shared @file{libgcc} instead of the static version. The most common
4045 of these is when the application wishes to throw and catch exceptions
4046 across different shared libraries. In that case, each of the libraries
4047 as well as the application itself should use the shared @file{libgcc}.
4049 Therefore, the G++ and GCJ drivers automatically add
4050 @option{-shared-libgcc} whenever you build a shared library or a main
4051 executable, because C++ and Java programs typically use exceptions, so
4052 this is the right thing to do.
4054 If, instead, you use the GCC driver to create shared libraries, you may
4055 find that they will not always be linked with the shared @file{libgcc}.
4056 If GCC finds, at its configuration time, that you have a GNU linker that
4057 does not support option @option{--eh-frame-hdr}, it will link the shared
4058 version of @file{libgcc} into shared libraries by default. Otherwise,
4059 it will take advantage of the linker and optimize away the linking with
4060 the shared version of @file{libgcc}, linking with the static version of
4061 libgcc by default. This allows exceptions to propagate through such
4062 shared libraries, without incurring relocation costs at library load
4065 However, if a library or main executable is supposed to throw or catch
4066 exceptions, you must link it using the G++ or GCJ driver, as appropriate
4067 for the languages used in the program, or using the option
4068 @option{-shared-libgcc}, such that it is linked with the shared
4073 Bind references to global symbols when building a shared object. Warn
4074 about any unresolved references (unless overridden by the link editor
4075 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4078 @item -Xlinker @var{option}
4080 Pass @var{option} as an option to the linker. You can use this to
4081 supply system-specific linker options which GCC does not know how to
4084 If you want to pass an option that takes an argument, you must use
4085 @option{-Xlinker} twice, once for the option and once for the argument.
4086 For example, to pass @option{-assert definitions}, you must write
4087 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4088 @option{-Xlinker "-assert definitions"}, because this passes the entire
4089 string as a single argument, which is not what the linker expects.
4091 @item -Wl,@var{option}
4093 Pass @var{option} as an option to the linker. If @var{option} contains
4094 commas, it is split into multiple options at the commas.
4096 @item -u @var{symbol}
4098 Pretend the symbol @var{symbol} is undefined, to force linking of
4099 library modules to define it. You can use @option{-u} multiple times with
4100 different symbols to force loading of additional library modules.
4103 @node Directory Options
4104 @section Options for Directory Search
4105 @cindex directory options
4106 @cindex options, directory search
4109 These options specify directories to search for header files, for
4110 libraries and for parts of the compiler:
4115 Add the directory @var{dir} to the head of the list of directories to be
4116 searched for header files. This can be used to override a system header
4117 file, substituting your own version, since these directories are
4118 searched before the system header file directories. However, you should
4119 not use this option to add directories that contain vendor-supplied
4120 system header files (use @option{-isystem} for that). If you use more than
4121 one @option{-I} option, the directories are scanned in left-to-right
4122 order; the standard system directories come after.
4124 If a standard system include directory, or a directory specified with
4125 @option{-isystem}, is also specified with @option{-I}, it will be
4126 searched only in the position requested by @option{-I}. Also, it will
4127 not be considered a system include directory. If that directory really
4128 does contain system headers, there is a good chance that they will
4129 break. For instance, if GCC's installation procedure edited the headers
4130 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4131 original, buggy headers to be found instead of the corrected ones. GCC
4132 will issue a warning when a system include directory is hidden in this
4137 Any directories you specify with @option{-I} options before the @option{-I-}
4138 option are searched only for the case of @samp{#include "@var{file}"};
4139 they are not searched for @samp{#include <@var{file}>}.
4141 If additional directories are specified with @option{-I} options after
4142 the @option{-I-}, these directories are searched for all @samp{#include}
4143 directives. (Ordinarily @emph{all} @option{-I} directories are used
4146 In addition, the @option{-I-} option inhibits the use of the current
4147 directory (where the current input file came from) as the first search
4148 directory for @samp{#include "@var{file}"}. There is no way to
4149 override this effect of @option{-I-}. With @option{-I.} you can specify
4150 searching the directory which was current when the compiler was
4151 invoked. That is not exactly the same as what the preprocessor does
4152 by default, but it is often satisfactory.
4154 @option{-I-} does not inhibit the use of the standard system directories
4155 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4160 Add directory @var{dir} to the list of directories to be searched
4163 @item -B@var{prefix}
4165 This option specifies where to find the executables, libraries,
4166 include files, and data files of the compiler itself.
4168 The compiler driver program runs one or more of the subprograms
4169 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4170 @var{prefix} as a prefix for each program it tries to run, both with and
4171 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4173 For each subprogram to be run, the compiler driver first tries the
4174 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4175 was not specified, the driver tries two standard prefixes, which are
4176 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4177 those results in a file name that is found, the unmodified program
4178 name is searched for using the directories specified in your
4179 @env{PATH} environment variable.
4181 The compiler will check to see if the path provided by the @option{-B}
4182 refers to a directory, and if necessary it will add a directory
4183 separator character at the end of the path.
4185 @option{-B} prefixes that effectively specify directory names also apply
4186 to libraries in the linker, because the compiler translates these
4187 options into @option{-L} options for the linker. They also apply to
4188 includes files in the preprocessor, because the compiler translates these
4189 options into @option{-isystem} options for the preprocessor. In this case,
4190 the compiler appends @samp{include} to the prefix.
4192 The run-time support file @file{libgcc.a} can also be searched for using
4193 the @option{-B} prefix, if needed. If it is not found there, the two
4194 standard prefixes above are tried, and that is all. The file is left
4195 out of the link if it is not found by those means.
4197 Another way to specify a prefix much like the @option{-B} prefix is to use
4198 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4201 As a special kludge, if the path provided by @option{-B} is
4202 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4203 9, then it will be replaced by @file{[dir/]include}. This is to help
4204 with boot-strapping the compiler.
4206 @item -specs=@var{file}
4208 Process @var{file} after the compiler reads in the standard @file{specs}
4209 file, in order to override the defaults that the @file{gcc} driver
4210 program uses when determining what switches to pass to @file{cc1},
4211 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4212 @option{-specs=@var{file}} can be specified on the command line, and they
4213 are processed in order, from left to right.
4219 @section Specifying subprocesses and the switches to pass to them
4221 @command{gcc} is a driver program. It performs its job by invoking a
4222 sequence of other programs to do the work of compiling, assembling and
4223 linking. GCC interprets its command-line parameters and uses these to
4224 deduce which programs it should invoke, and which command-line options
4225 it ought to place on their command lines. This behavior is controlled
4226 by @dfn{spec strings}. In most cases there is one spec string for each
4227 program that GCC can invoke, but a few programs have multiple spec
4228 strings to control their behavior. The spec strings built into GCC can
4229 be overridden by using the @option{-specs=} command-line switch to specify
4232 @dfn{Spec files} are plaintext files that are used to construct spec
4233 strings. They consist of a sequence of directives separated by blank
4234 lines. The type of directive is determined by the first non-whitespace
4235 character on the line and it can be one of the following:
4238 @item %@var{command}
4239 Issues a @var{command} to the spec file processor. The commands that can
4243 @item %include <@var{file}>
4245 Search for @var{file} and insert its text at the current point in the
4248 @item %include_noerr <@var{file}>
4249 @cindex %include_noerr
4250 Just like @samp{%include}, but do not generate an error message if the include
4251 file cannot be found.
4253 @item %rename @var{old_name} @var{new_name}
4255 Rename the spec string @var{old_name} to @var{new_name}.
4259 @item *[@var{spec_name}]:
4260 This tells the compiler to create, override or delete the named spec
4261 string. All lines after this directive up to the next directive or
4262 blank line are considered to be the text for the spec string. If this
4263 results in an empty string then the spec will be deleted. (Or, if the
4264 spec did not exist, then nothing will happened.) Otherwise, if the spec
4265 does not currently exist a new spec will be created. If the spec does
4266 exist then its contents will be overridden by the text of this
4267 directive, unless the first character of that text is the @samp{+}
4268 character, in which case the text will be appended to the spec.
4270 @item [@var{suffix}]:
4271 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4272 and up to the next directive or blank line are considered to make up the
4273 spec string for the indicated suffix. When the compiler encounters an
4274 input file with the named suffix, it will processes the spec string in
4275 order to work out how to compile that file. For example:
4282 This says that any input file whose name ends in @samp{.ZZ} should be
4283 passed to the program @samp{z-compile}, which should be invoked with the
4284 command-line switch @option{-input} and with the result of performing the
4285 @samp{%i} substitution. (See below.)
4287 As an alternative to providing a spec string, the text that follows a
4288 suffix directive can be one of the following:
4291 @item @@@var{language}
4292 This says that the suffix is an alias for a known @var{language}. This is
4293 similar to using the @option{-x} command-line switch to GCC to specify a
4294 language explicitly. For example:
4301 Says that .ZZ files are, in fact, C++ source files.
4304 This causes an error messages saying:
4307 @var{name} compiler not installed on this system.
4311 GCC already has an extensive list of suffixes built into it.
4312 This directive will add an entry to the end of the list of suffixes, but
4313 since the list is searched from the end backwards, it is effectively
4314 possible to override earlier entries using this technique.
4318 GCC has the following spec strings built into it. Spec files can
4319 override these strings or create their own. Note that individual
4320 targets can also add their own spec strings to this list.
4323 asm Options to pass to the assembler
4324 asm_final Options to pass to the assembler post-processor
4325 cpp Options to pass to the C preprocessor
4326 cc1 Options to pass to the C compiler
4327 cc1plus Options to pass to the C++ compiler
4328 endfile Object files to include at the end of the link
4329 link Options to pass to the linker
4330 lib Libraries to include on the command line to the linker
4331 libgcc Decides which GCC support library to pass to the linker
4332 linker Sets the name of the linker
4333 predefines Defines to be passed to the C preprocessor
4334 signed_char Defines to pass to CPP to say whether @code{char} is signed
4336 startfile Object files to include at the start of the link
4339 Here is a small example of a spec file:
4345 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4348 This example renames the spec called @samp{lib} to @samp{old_lib} and
4349 then overrides the previous definition of @samp{lib} with a new one.
4350 The new definition adds in some extra command-line options before
4351 including the text of the old definition.
4353 @dfn{Spec strings} are a list of command-line options to be passed to their
4354 corresponding program. In addition, the spec strings can contain
4355 @samp{%}-prefixed sequences to substitute variable text or to
4356 conditionally insert text into the command line. Using these constructs
4357 it is possible to generate quite complex command lines.
4359 Here is a table of all defined @samp{%}-sequences for spec
4360 strings. Note that spaces are not generated automatically around the
4361 results of expanding these sequences. Therefore you can concatenate them
4362 together or combine them with constant text in a single argument.
4366 Substitute one @samp{%} into the program name or argument.
4369 Substitute the name of the input file being processed.
4372 Substitute the basename of the input file being processed.
4373 This is the substring up to (and not including) the last period
4374 and not including the directory.
4377 This is the same as @samp{%b}, but include the file suffix (text after
4381 Marks the argument containing or following the @samp{%d} as a
4382 temporary file name, so that that file will be deleted if GCC exits
4383 successfully. Unlike @samp{%g}, this contributes no text to the
4386 @item %g@var{suffix}
4387 Substitute a file name that has suffix @var{suffix} and is chosen
4388 once per compilation, and mark the argument in the same way as
4389 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4390 name is now chosen in a way that is hard to predict even when previously
4391 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4392 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4393 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4394 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4395 was simply substituted with a file name chosen once per compilation,
4396 without regard to any appended suffix (which was therefore treated
4397 just like ordinary text), making such attacks more likely to succeed.
4399 @item %u@var{suffix}
4400 Like @samp{%g}, but generates a new temporary file name even if
4401 @samp{%u@var{suffix}} was already seen.
4403 @item %U@var{suffix}
4404 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4405 new one if there is no such last file name. In the absence of any
4406 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4407 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4408 would involve the generation of two distinct file names, one
4409 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4410 simply substituted with a file name chosen for the previous @samp{%u},
4411 without regard to any appended suffix.
4413 @item %j@var{SUFFIX}
4414 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4415 writable, and if save-temps is off; otherwise, substitute the name
4416 of a temporary file, just like @samp{%u}. This temporary file is not
4417 meant for communication between processes, but rather as a junk
4420 @item %.@var{SUFFIX}
4421 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4422 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4423 terminated by the next space or %.
4426 Marks the argument containing or following the @samp{%w} as the
4427 designated output file of this compilation. This puts the argument
4428 into the sequence of arguments that @samp{%o} will substitute later.
4431 Substitutes the names of all the output files, with spaces
4432 automatically placed around them. You should write spaces
4433 around the @samp{%o} as well or the results are undefined.
4434 @samp{%o} is for use in the specs for running the linker.
4435 Input files whose names have no recognized suffix are not compiled
4436 at all, but they are included among the output files, so they will
4440 Substitutes the suffix for object files. Note that this is
4441 handled specially when it immediately follows @samp{%g, %u, or %U},
4442 because of the need for those to form complete file names. The
4443 handling is such that @samp{%O} is treated exactly as if it had already
4444 been substituted, except that @samp{%g, %u, and %U} do not currently
4445 support additional @var{suffix} characters following @samp{%O} as they would
4446 following, for example, @samp{.o}.
4449 Substitutes the standard macro predefinitions for the
4450 current target machine. Use this when running @code{cpp}.
4453 Like @samp{%p}, but puts @samp{__} before and after the name of each
4454 predefined macro, except for macros that start with @samp{__} or with
4455 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4459 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4462 Current argument is the name of a library or startup file of some sort.
4463 Search for that file in a standard list of directories and substitute
4464 the full name found.
4467 Print @var{str} as an error message. @var{str} is terminated by a newline.
4468 Use this when inconsistent options are detected.
4471 Output @samp{-} if the input for the current command is coming from a pipe.
4474 Substitute the contents of spec string @var{name} at this point.
4477 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4479 @item %x@{@var{option}@}
4480 Accumulate an option for @samp{%X}.
4483 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4487 Output the accumulated assembler options specified by @option{-Wa}.
4490 Output the accumulated preprocessor options specified by @option{-Wp}.
4493 Substitute the major version number of GCC@.
4494 (For version 2.9.5, this is 2.)
4497 Substitute the minor version number of GCC@.
4498 (For version 2.9.5, this is 9.)
4501 Substitute the patch level number of GCC@.
4502 (For version 2.9.5, this is 5.)
4505 Process the @code{asm} spec. This is used to compute the
4506 switches to be passed to the assembler.
4509 Process the @code{asm_final} spec. This is a spec string for
4510 passing switches to an assembler post-processor, if such a program is
4514 Process the @code{link} spec. This is the spec for computing the
4515 command line passed to the linker. Typically it will make use of the
4516 @samp{%L %G %S %D and %E} sequences.
4519 Dump out a @option{-L} option for each directory that GCC believes might
4520 contain startup files. If the target supports multilibs then the
4521 current multilib directory will be prepended to each of these paths.
4524 Output the multilib directory with directory separators replaced with
4525 @samp{_}. If multilib directories are not set, or the multilib directory is
4526 @file{.} then this option emits nothing.
4529 Process the @code{lib} spec. This is a spec string for deciding which
4530 libraries should be included on the command line to the linker.
4533 Process the @code{libgcc} spec. This is a spec string for deciding
4534 which GCC support library should be included on the command line to the linker.
4537 Process the @code{startfile} spec. This is a spec for deciding which
4538 object files should be the first ones passed to the linker. Typically
4539 this might be a file named @file{crt0.o}.
4542 Process the @code{endfile} spec. This is a spec string that specifies
4543 the last object files that will be passed to the linker.
4546 Process the @code{cpp} spec. This is used to construct the arguments
4547 to be passed to the C preprocessor.
4550 Process the @code{signed_char} spec. This is intended to be used
4551 to tell cpp whether a char is signed. It typically has the definition:
4553 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4557 Process the @code{cc1} spec. This is used to construct the options to be
4558 passed to the actual C compiler (@samp{cc1}).
4561 Process the @code{cc1plus} spec. This is used to construct the options to be
4562 passed to the actual C++ compiler (@samp{cc1plus}).
4565 Substitute the variable part of a matched option. See below.
4566 Note that each comma in the substituted string is replaced by
4570 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4571 If that switch was not specified, this substitutes nothing. Note that
4572 the leading dash is omitted when specifying this option, and it is
4573 automatically inserted if the substitution is performed. Thus the spec
4574 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4575 and would output the command line option @option{-foo}.
4577 @item %W@{@code{S}@}
4578 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4581 @item %@{@code{S}*@}
4582 Substitutes all the switches specified to GCC whose names start
4583 with @code{-S}, but which also take an argument. This is used for
4584 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4585 GCC considers @option{-o foo} as being
4586 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4587 text, including the space. Thus two arguments would be generated.
4589 @item %@{^@code{S}*@}
4590 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4591 argument. Thus %@{^o*@} would only generate one argument, not two.
4593 @item %@{@code{S}*&@code{T}*@}
4594 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4595 (the order of @code{S} and @code{T} in the spec is not significant).
4596 There can be any number of ampersand-separated variables; for each the
4597 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4599 @item %@{<@code{S}@}
4600 Remove all occurrences of @code{-S} from the command line. Note---this
4601 command is position dependent. @samp{%} commands in the spec string
4602 before this option will see @code{-S}, @samp{%} commands in the spec
4603 string after this option will not.
4605 @item %@{@code{S}*:@code{X}@}
4606 Substitutes @code{X} if one or more switches whose names start with
4607 @code{-S} are specified to GCC@. Note that the tail part of the
4608 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4609 for each occurrence of @samp{%*} within @code{X}.
4611 @item %@{@code{S}:@code{X}@}
4612 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4614 @item %@{!@code{S}:@code{X}@}
4615 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4617 @item %@{|@code{S}:@code{X}@}
4618 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4620 @item %@{|!@code{S}:@code{X}@}
4621 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4623 @item %@{.@code{S}:@code{X}@}
4624 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4626 @item %@{!.@code{S}:@code{X}@}
4627 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4629 @item %@{@code{S}|@code{P}:@code{X}@}
4630 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4631 combined with @samp{!} and @samp{.} sequences as well, although they
4632 have a stronger binding than the @samp{|}. For example a spec string
4636 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4639 will output the following command-line options from the following input
4640 command-line options:
4645 -d fred.c -foo -baz -boggle
4646 -d jim.d -bar -baz -boggle
4651 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4652 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4653 or spaces, or even newlines. They are processed as usual, as described
4656 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4657 switches are handled specifically in these
4658 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4659 @option{-W} switch is found later in the command line, the earlier switch
4660 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4661 letter, which passes all matching options.
4663 The character @samp{|} at the beginning of the predicate text is used to indicate
4664 that a command should be piped to the following command, but only if @option{-pipe}
4667 It is built into GCC which switches take arguments and which do not.
4668 (You might think it would be useful to generalize this to allow each
4669 compiler's spec to say which switches take arguments. But this cannot
4670 be done in a consistent fashion. GCC cannot even decide which input
4671 files have been specified without knowing which switches take arguments,
4672 and it must know which input files to compile in order to tell which
4675 GCC also knows implicitly that arguments starting in @option{-l} are to be
4676 treated as compiler output files, and passed to the linker in their
4677 proper position among the other output files.
4679 @c man begin OPTIONS
4681 @node Target Options
4682 @section Specifying Target Machine and Compiler Version
4683 @cindex target options
4684 @cindex cross compiling
4685 @cindex specifying machine version
4686 @cindex specifying compiler version and target machine
4687 @cindex compiler version, specifying
4688 @cindex target machine, specifying
4690 By default, GCC compiles code for the same type of machine that you
4691 are using. However, it can also be installed as a cross-compiler, to
4692 compile for some other type of machine. In fact, several different
4693 configurations of GCC, for different target machines, can be
4694 installed side by side. Then you specify which one to use with the
4697 In addition, older and newer versions of GCC can be installed side
4698 by side. One of them (probably the newest) will be the default, but
4699 you may sometimes wish to use another.
4702 @item -b @var{machine}
4704 The argument @var{machine} specifies the target machine for compilation.
4705 This is useful when you have installed GCC as a cross-compiler.
4707 The value to use for @var{machine} is the same as was specified as the
4708 machine type when configuring GCC as a cross-compiler. For
4709 example, if a cross-compiler was configured with @samp{configure
4710 i386v}, meaning to compile for an 80386 running System V, then you
4711 would specify @option{-b i386v} to run that cross compiler.
4713 When you do not specify @option{-b}, it normally means to compile for
4714 the same type of machine that you are using.
4716 @item -V @var{version}
4718 The argument @var{version} specifies which version of GCC to run.
4719 This is useful when multiple versions are installed. For example,
4720 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
4722 The default version, when you do not specify @option{-V}, is the last
4723 version of GCC that you installed.
4726 The @option{-b} and @option{-V} options actually work by controlling part of
4727 the file name used for the executable files and libraries used for
4728 compilation. A given version of GCC, for a given target machine, is
4729 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.
4731 Thus, sites can customize the effect of @option{-b} or @option{-V} either by
4732 changing the names of these directories or adding alternate names (or
4733 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
4734 file @file{80386} is a link to the file @file{i386v}, then @option{-b
4735 80386} becomes an alias for @option{-b i386v}.
4737 In one respect, the @option{-b} or @option{-V} do not completely change
4738 to a different compiler: the top-level driver program @command{gcc}
4739 that you originally invoked continues to run and invoke the other
4740 executables (preprocessor, compiler per se, assembler and linker)
4741 that do the real work. However, since no real work is done in the
4742 driver program, it usually does not matter that the driver program
4743 in use is not the one for the specified target. It is common for the
4744 interface to the other executables to change incompatibly between
4745 compiler versions, so unless the version specified is very close to that
4746 of the driver (for example, @option{-V 3.0} with a driver program from GCC
4747 version 3.0.1), use of @option{-V} may not work; for example, using
4748 @option{-V 2.95.2} will not work with a driver program from GCC 3.0.
4750 The only way that the driver program depends on the target machine is
4751 in the parsing and handling of special machine-specific options.
4752 However, this is controlled by a file which is found, along with the
4753 other executables, in the directory for the specified version and
4754 target machine. As a result, a single installed driver program adapts
4755 to any specified target machine, and sufficiently similar compiler
4758 The driver program executable does control one significant thing,
4759 however: the default version and target machine. Therefore, you can
4760 install different instances of the driver program, compiled for
4761 different targets or versions, under different names.
4763 For example, if the driver for version 2.0 is installed as @command{ogcc}
4764 and that for version 2.1 is installed as @command{gcc}, then the command
4765 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
4766 2.0 by default. However, you can choose either version with either
4767 command with the @option{-V} option.
4769 @node Submodel Options
4770 @section Hardware Models and Configurations
4771 @cindex submodel options
4772 @cindex specifying hardware config
4773 @cindex hardware models and configurations, specifying
4774 @cindex machine dependent options
4776 Earlier we discussed the standard option @option{-b} which chooses among
4777 different installed compilers for completely different target
4778 machines, such as VAX vs.@: 68000 vs.@: 80386.
4780 In addition, each of these target machine types can have its own
4781 special options, starting with @samp{-m}, to choose among various
4782 hardware models or configurations---for example, 68010 vs 68020,
4783 floating coprocessor or none. A single installed version of the
4784 compiler can compile for any model or configuration, according to the
4787 Some configurations of the compiler also support additional special
4788 options, usually for compatibility with other compilers on the same
4791 These options are defined by the macro @code{TARGET_SWITCHES} in the
4792 machine description. The default for the options is also defined by
4793 that macro, which enables you to change the defaults.
4807 * RS/6000 and PowerPC Options::
4810 * i386 and x86-64 Options::
4812 * Intel 960 Options::
4813 * DEC Alpha Options::
4814 * DEC Alpha/VMS Options::
4818 * System V Options::
4819 * TMS320C3x/C4x Options::
4827 * S/390 and zSeries Options::
4831 * Xstormy16 Options::
4835 @node M680x0 Options
4836 @subsection M680x0 Options
4837 @cindex M680x0 options
4839 These are the @samp{-m} options defined for the 68000 series. The default
4840 values for these options depends on which style of 68000 was selected when
4841 the compiler was configured; the defaults for the most common choices are
4849 Generate output for a 68000. This is the default
4850 when the compiler is configured for 68000-based systems.
4852 Use this option for microcontrollers with a 68000 or EC000 core,
4853 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
4859 Generate output for a 68020. This is the default
4860 when the compiler is configured for 68020-based systems.
4864 Generate output containing 68881 instructions for floating point.
4865 This is the default for most 68020 systems unless @option{--nfp} was
4866 specified when the compiler was configured.
4870 Generate output for a 68030. This is the default when the compiler is
4871 configured for 68030-based systems.
4875 Generate output for a 68040. This is the default when the compiler is
4876 configured for 68040-based systems.
4878 This option inhibits the use of 68881/68882 instructions that have to be
4879 emulated by software on the 68040. Use this option if your 68040 does not
4880 have code to emulate those instructions.
4884 Generate output for a 68060. This is the default when the compiler is
4885 configured for 68060-based systems.
4887 This option inhibits the use of 68020 and 68881/68882 instructions that
4888 have to be emulated by software on the 68060. Use this option if your 68060
4889 does not have code to emulate those instructions.
4893 Generate output for a CPU32. This is the default
4894 when the compiler is configured for CPU32-based systems.
4896 Use this option for microcontrollers with a
4897 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
4898 68336, 68340, 68341, 68349 and 68360.
4902 Generate output for a 520X ``coldfire'' family cpu. This is the default
4903 when the compiler is configured for 520X-based systems.
4905 Use this option for microcontroller with a 5200 core, including
4906 the MCF5202, MCF5203, MCF5204 and MCF5202.
4911 Generate output for a 68040, without using any of the new instructions.
4912 This results in code which can run relatively efficiently on either a
4913 68020/68881 or a 68030 or a 68040. The generated code does use the
4914 68881 instructions that are emulated on the 68040.
4918 Generate output for a 68060, without using any of the new instructions.
4919 This results in code which can run relatively efficiently on either a
4920 68020/68881 or a 68030 or a 68040. The generated code does use the
4921 68881 instructions that are emulated on the 68060.
4925 Generate output containing Sun FPA instructions for floating point.
4928 @opindex msoft-float
4929 Generate output containing library calls for floating point.
4930 @strong{Warning:} the requisite libraries are not available for all m68k
4931 targets. Normally the facilities of the machine's usual C compiler are
4932 used, but this can't be done directly in cross-compilation. You must
4933 make your own arrangements to provide suitable library functions for
4934 cross-compilation. The embedded targets @samp{m68k-*-aout} and
4935 @samp{m68k-*-coff} do provide software floating point support.
4939 Consider type @code{int} to be 16 bits wide, like @code{short int}.
4942 @opindex mnobitfield
4943 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
4944 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
4948 Do use the bit-field instructions. The @option{-m68020} option implies
4949 @option{-mbitfield}. This is the default if you use a configuration
4950 designed for a 68020.
4954 Use a different function-calling convention, in which functions
4955 that take a fixed number of arguments return with the @code{rtd}
4956 instruction, which pops their arguments while returning. This
4957 saves one instruction in the caller since there is no need to pop
4958 the arguments there.
4960 This calling convention is incompatible with the one normally
4961 used on Unix, so you cannot use it if you need to call libraries
4962 compiled with the Unix compiler.
4964 Also, you must provide function prototypes for all functions that
4965 take variable numbers of arguments (including @code{printf});
4966 otherwise incorrect code will be generated for calls to those
4969 In addition, seriously incorrect code will result if you call a
4970 function with too many arguments. (Normally, extra arguments are
4971 harmlessly ignored.)
4973 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
4974 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
4977 @itemx -mno-align-int
4979 @opindex mno-align-int
4980 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
4981 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
4982 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
4983 Aligning variables on 32-bit boundaries produces code that runs somewhat
4984 faster on processors with 32-bit busses at the expense of more memory.
4986 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
4987 align structures containing the above types differently than
4988 most published application binary interface specifications for the m68k.
4992 Use the pc-relative addressing mode of the 68000 directly, instead of
4993 using a global offset table. At present, this option implies @option{-fpic},
4994 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
4995 not presently supported with @option{-mpcrel}, though this could be supported for
4996 68020 and higher processors.
4998 @item -mno-strict-align
4999 @itemx -mstrict-align
5000 @opindex mno-strict-align
5001 @opindex mstrict-align
5002 Do not (do) assume that unaligned memory references will be handled by
5007 @node M68hc1x Options
5008 @subsection M68hc1x Options
5009 @cindex M68hc1x options
5011 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5012 microcontrollers. The default values for these options depends on
5013 which style of microcontroller was selected when the compiler was configured;
5014 the defaults for the most common choices are given below.
5021 Generate output for a 68HC11. This is the default
5022 when the compiler is configured for 68HC11-based systems.
5028 Generate output for a 68HC12. This is the default
5029 when the compiler is configured for 68HC12-based systems.
5032 @opindex mauto-incdec
5033 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5038 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5040 @item -msoft-reg-count=@var{count}
5041 @opindex msoft-reg-count
5042 Specify the number of pseudo-soft registers which are used for the
5043 code generation. The maximum number is 32. Using more pseudo-soft
5044 register may or may not result in better code depending on the program.
5045 The default is 4 for 68HC11 and 2 for 68HC12.
5050 @subsection VAX Options
5053 These @samp{-m} options are defined for the VAX:
5058 Do not output certain jump instructions (@code{aobleq} and so on)
5059 that the Unix assembler for the VAX cannot handle across long
5064 Do output those jump instructions, on the assumption that you
5065 will assemble with the GNU assembler.
5069 Output code for g-format floating point numbers instead of d-format.
5073 @subsection SPARC Options
5074 @cindex SPARC options
5076 These @samp{-m} switches are supported on the SPARC:
5081 @opindex mno-app-regs
5083 Specify @option{-mapp-regs} to generate output using the global registers
5084 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5087 To be fully SVR4 ABI compliant at the cost of some performance loss,
5088 specify @option{-mno-app-regs}. You should compile libraries and system
5089 software with this option.
5094 @opindex mhard-float
5095 Generate output containing floating point instructions. This is the
5101 @opindex msoft-float
5102 Generate output containing library calls for floating point.
5103 @strong{Warning:} the requisite libraries are not available for all SPARC
5104 targets. Normally the facilities of the machine's usual C compiler are
5105 used, but this cannot be done directly in cross-compilation. You must make
5106 your own arrangements to provide suitable library functions for
5107 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5108 @samp{sparclite-*-*} do provide software floating point support.
5110 @option{-msoft-float} changes the calling convention in the output file;
5111 therefore, it is only useful if you compile @emph{all} of a program with
5112 this option. In particular, you need to compile @file{libgcc.a}, the
5113 library that comes with GCC, with @option{-msoft-float} in order for
5116 @item -mhard-quad-float
5117 @opindex mhard-quad-float
5118 Generate output containing quad-word (long double) floating point
5121 @item -msoft-quad-float
5122 @opindex msoft-quad-float
5123 Generate output containing library calls for quad-word (long double)
5124 floating point instructions. The functions called are those specified
5125 in the SPARC ABI@. This is the default.
5127 As of this writing, there are no sparc implementations that have hardware
5128 support for the quad-word floating point instructions. They all invoke
5129 a trap handler for one of these instructions, and then the trap handler
5130 emulates the effect of the instruction. Because of the trap handler overhead,
5131 this is much slower than calling the ABI library routines. Thus the
5132 @option{-msoft-quad-float} option is the default.
5136 @opindex mno-epilogue
5138 With @option{-mepilogue} (the default), the compiler always emits code for
5139 function exit at the end of each function. Any function exit in
5140 the middle of the function (such as a return statement in C) will
5141 generate a jump to the exit code at the end of the function.
5143 With @option{-mno-epilogue}, the compiler tries to emit exit code inline
5144 at every function exit.
5150 With @option{-mflat}, the compiler does not generate save/restore instructions
5151 and will use a ``flat'' or single register window calling convention.
5152 This model uses %i7 as the frame pointer and is compatible with the normal
5153 register window model. Code from either may be intermixed.
5154 The local registers and the input registers (0--5) are still treated as
5155 ``call saved'' registers and will be saved on the stack as necessary.
5157 With @option{-mno-flat} (the default), the compiler emits save/restore
5158 instructions (except for leaf functions) and is the normal mode of operation.
5160 @item -mno-unaligned-doubles
5161 @itemx -munaligned-doubles
5162 @opindex mno-unaligned-doubles
5163 @opindex munaligned-doubles
5164 Assume that doubles have 8 byte alignment. This is the default.
5166 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5167 alignment only if they are contained in another type, or if they have an
5168 absolute address. Otherwise, it assumes they have 4 byte alignment.
5169 Specifying this option avoids some rare compatibility problems with code
5170 generated by other compilers. It is not the default because it results
5171 in a performance loss, especially for floating point code.
5173 @item -mno-faster-structs
5174 @itemx -mfaster-structs
5175 @opindex mno-faster-structs
5176 @opindex mfaster-structs
5177 With @option{-mfaster-structs}, the compiler assumes that structures
5178 should have 8 byte alignment. This enables the use of pairs of
5179 @code{ldd} and @code{std} instructions for copies in structure
5180 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5181 However, the use of this changed alignment directly violates the Sparc
5182 ABI@. Thus, it's intended only for use on targets where the developer
5183 acknowledges that their resulting code will not be directly in line with
5184 the rules of the ABI@.
5190 These two options select variations on the SPARC architecture.
5192 By default (unless specifically configured for the Fujitsu SPARClite),
5193 GCC generates code for the v7 variant of the SPARC architecture.
5195 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5196 code is that the compiler emits the integer multiply and integer
5197 divide instructions which exist in SPARC v8 but not in SPARC v7.
5199 @option{-msparclite} will give you SPARClite code. This adds the integer
5200 multiply, integer divide step and scan (@code{ffs}) instructions which
5201 exist in SPARClite but not in SPARC v7.
5203 These options are deprecated and will be deleted in a future GCC release.
5204 They have been replaced with @option{-mcpu=xxx}.
5209 @opindex msupersparc
5210 These two options select the processor for which the code is optimized.
5212 With @option{-mcypress} (the default), the compiler optimizes code for the
5213 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5214 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5216 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5217 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5218 of the full SPARC v8 instruction set.
5220 These options are deprecated and will be deleted in a future GCC release.
5221 They have been replaced with @option{-mcpu=xxx}.
5223 @item -mcpu=@var{cpu_type}
5225 Set the instruction set, register set, and instruction scheduling parameters
5226 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5227 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5228 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5229 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5231 Default instruction scheduling parameters are used for values that select
5232 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5233 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5235 Here is a list of each supported architecture and their supported
5240 v8: supersparc, hypersparc
5241 sparclite: f930, f934, sparclite86x
5246 @item -mtune=@var{cpu_type}
5248 Set the instruction scheduling parameters for machine type
5249 @var{cpu_type}, but do not set the instruction set or register set that the
5250 option @option{-mcpu=@var{cpu_type}} would.
5252 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5253 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5254 that select a particular cpu implementation. Those are @samp{cypress},
5255 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5256 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5260 These @samp{-m} switches are supported in addition to the above
5261 on the SPARCLET processor.
5264 @item -mlittle-endian
5265 @opindex mlittle-endian
5266 Generate code for a processor running in little-endian mode.
5270 Treat register @code{%g0} as a normal register.
5271 GCC will continue to clobber it as necessary but will not assume
5272 it always reads as 0.
5274 @item -mbroken-saverestore
5275 @opindex mbroken-saverestore
5276 Generate code that does not use non-trivial forms of the @code{save} and
5277 @code{restore} instructions. Early versions of the SPARCLET processor do
5278 not correctly handle @code{save} and @code{restore} instructions used with
5279 arguments. They correctly handle them used without arguments. A @code{save}
5280 instruction used without arguments increments the current window pointer
5281 but does not allocate a new stack frame. It is assumed that the window
5282 overflow trap handler will properly handle this case as will interrupt
5286 These @samp{-m} switches are supported in addition to the above
5287 on SPARC V9 processors in 64-bit environments.
5290 @item -mlittle-endian
5291 @opindex mlittle-endian
5292 Generate code for a processor running in little-endian mode.
5298 Generate code for a 32-bit or 64-bit environment.
5299 The 32-bit environment sets int, long and pointer to 32 bits.
5300 The 64-bit environment sets int to 32 bits and long and pointer
5303 @item -mcmodel=medlow
5304 @opindex mcmodel=medlow
5305 Generate code for the Medium/Low code model: the program must be linked
5306 in the low 32 bits of the address space. Pointers are 64 bits.
5307 Programs can be statically or dynamically linked.
5309 @item -mcmodel=medmid
5310 @opindex mcmodel=medmid
5311 Generate code for the Medium/Middle code model: the program must be linked
5312 in the low 44 bits of the address space, the text segment must be less than
5313 2G bytes, and data segment must be within 2G of the text segment.
5314 Pointers are 64 bits.
5316 @item -mcmodel=medany
5317 @opindex mcmodel=medany
5318 Generate code for the Medium/Anywhere code model: the program may be linked
5319 anywhere in the address space, the text segment must be less than
5320 2G bytes, and data segment must be within 2G of the text segment.
5321 Pointers are 64 bits.
5323 @item -mcmodel=embmedany
5324 @opindex mcmodel=embmedany
5325 Generate code for the Medium/Anywhere code model for embedded systems:
5326 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5327 (determined at link time). Register %g4 points to the base of the
5328 data segment. Pointers are still 64 bits.
5329 Programs are statically linked, PIC is not supported.
5332 @itemx -mno-stack-bias
5333 @opindex mstack-bias
5334 @opindex mno-stack-bias
5335 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5336 frame pointer if present, are offset by @minus{}2047 which must be added back
5337 when making stack frame references.
5338 Otherwise, assume no such offset is present.
5341 @node Convex Options
5342 @subsection Convex Options
5343 @cindex Convex options
5345 These @samp{-m} options are defined for Convex:
5350 Generate output for C1. The code will run on any Convex machine.
5351 The preprocessor symbol @code{__convex__c1__} is defined.
5355 Generate output for C2. Uses instructions not available on C1.
5356 Scheduling and other optimizations are chosen for max performance on C2.
5357 The preprocessor symbol @code{__convex_c2__} is defined.
5361 Generate output for C32xx. Uses instructions not available on C1.
5362 Scheduling and other optimizations are chosen for max performance on C32.
5363 The preprocessor symbol @code{__convex_c32__} is defined.
5367 Generate output for C34xx. Uses instructions not available on C1.
5368 Scheduling and other optimizations are chosen for max performance on C34.
5369 The preprocessor symbol @code{__convex_c34__} is defined.
5373 Generate output for C38xx. Uses instructions not available on C1.
5374 Scheduling and other optimizations are chosen for max performance on C38.
5375 The preprocessor symbol @code{__convex_c38__} is defined.
5379 Generate code which puts an argument count in the word preceding each
5380 argument list. This is compatible with regular CC, and a few programs
5381 may need the argument count word. GDB and other source-level debuggers
5382 do not need it; this info is in the symbol table.
5385 @opindex mnoargcount
5386 Omit the argument count word. This is the default.
5388 @item -mvolatile-cache
5389 @opindex mvolatile-cache
5390 Allow volatile references to be cached. This is the default.
5392 @item -mvolatile-nocache
5393 @opindex mvolatile-nocache
5394 Volatile references bypass the data cache, going all the way to memory.
5395 This is only needed for multi-processor code that does not use standard
5396 synchronization instructions. Making non-volatile references to volatile
5397 locations will not necessarily work.
5401 Type long is 32 bits, the same as type int. This is the default.
5405 Type long is 64 bits, the same as type long long. This option is useless,
5406 because no library support exists for it.
5409 @node AMD29K Options
5410 @subsection AMD29K Options
5411 @cindex AMD29K options
5413 These @samp{-m} options are defined for the AMD Am29000:
5418 @cindex DW bit (29k)
5419 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5420 halfword operations are directly supported by the hardware. This is the
5425 Generate code that assumes the @code{DW} bit is not set.
5429 @cindex byte writes (29k)
5430 Generate code that assumes the system supports byte and halfword write
5431 operations. This is the default.
5435 Generate code that assumes the systems does not support byte and
5436 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5440 @cindex memory model (29k)
5441 Use a small memory model that assumes that all function addresses are
5442 either within a single 256 KB segment or at an absolute address of less
5443 than 256k. This allows the @code{call} instruction to be used instead
5444 of a @code{const}, @code{consth}, @code{calli} sequence.
5448 Use the normal memory model: Generate @code{call} instructions only when
5449 calling functions in the same file and @code{calli} instructions
5450 otherwise. This works if each file occupies less than 256 KB but allows
5451 the entire executable to be larger than 256 KB@. This is the default.
5455 Always use @code{calli} instructions. Specify this option if you expect
5456 a single file to compile into more than 256 KB of code.
5460 @cindex processor selection (29k)
5461 Generate code for the Am29050.
5465 Generate code for the Am29000. This is the default.
5467 @item -mkernel-registers
5468 @opindex mkernel-registers
5469 @cindex kernel and user registers (29k)
5470 Generate references to registers @code{gr64-gr95} instead of to
5471 registers @code{gr96-gr127}. This option can be used when compiling
5472 kernel code that wants a set of global registers disjoint from that used
5475 Note that when this option is used, register names in @samp{-f} flags
5476 must use the normal, user-mode, names.
5478 @item -muser-registers
5479 @opindex muser-registers
5480 Use the normal set of global registers, @code{gr96-gr127}. This is the
5484 @itemx -mno-stack-check
5485 @opindex mstack-check
5486 @opindex mno-stack-check
5487 @cindex stack checks (29k)
5488 Insert (or do not insert) a call to @code{__msp_check} after each stack
5489 adjustment. This is often used for kernel code.
5492 @itemx -mno-storem-bug
5493 @opindex mstorem-bug
5494 @opindex mno-storem-bug
5495 @cindex storem bug (29k)
5496 @option{-mstorem-bug} handles 29k processors which cannot handle the
5497 separation of a mtsrim insn and a storem instruction (most 29000 chips
5498 to date, but not the 29050).
5500 @item -mno-reuse-arg-regs
5501 @itemx -mreuse-arg-regs
5502 @opindex mno-reuse-arg-regs
5503 @opindex mreuse-arg-regs
5504 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5505 registers for copying out arguments. This helps detect calling a function
5506 with fewer arguments than it was declared with.
5508 @item -mno-impure-text
5509 @itemx -mimpure-text
5510 @opindex mno-impure-text
5511 @opindex mimpure-text
5512 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5513 not pass @option{-assert pure-text} to the linker when linking a shared object.
5516 @opindex msoft-float
5517 Generate output containing library calls for floating point.
5518 @strong{Warning:} the requisite libraries are not part of GCC@.
5519 Normally the facilities of the machine's usual C compiler are used, but
5520 this can't be done directly in cross-compilation. You must make your
5521 own arrangements to provide suitable library functions for
5526 Do not generate multm or multmu instructions. This is useful for some embedded
5527 systems which do not have trap handlers for these instructions.
5531 @subsection ARM Options
5534 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5539 @opindex mapcs-frame
5540 Generate a stack frame that is compliant with the ARM Procedure Call
5541 Standard for all functions, even if this is not strictly necessary for
5542 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5543 with this option will cause the stack frames not to be generated for
5544 leaf functions. The default is @option{-mno-apcs-frame}.
5548 This is a synonym for @option{-mapcs-frame}.
5552 Generate code for a processor running with a 26-bit program counter,
5553 and conforming to the function calling standards for the APCS 26-bit
5554 option. This option replaces the @option{-m2} and @option{-m3} options
5555 of previous releases of the compiler.
5559 Generate code for a processor running with a 32-bit program counter,
5560 and conforming to the function calling standards for the APCS 32-bit
5561 option. This option replaces the @option{-m6} option of previous releases
5565 @c not currently implemented
5566 @item -mapcs-stack-check
5567 @opindex mapcs-stack-check
5568 Generate code to check the amount of stack space available upon entry to
5569 every function (that actually uses some stack space). If there is
5570 insufficient space available then either the function
5571 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5572 called, depending upon the amount of stack space required. The run time
5573 system is required to provide these functions. The default is
5574 @option{-mno-apcs-stack-check}, since this produces smaller code.
5576 @c not currently implemented
5578 @opindex mapcs-float
5579 Pass floating point arguments using the float point registers. This is
5580 one of the variants of the APCS@. This option is recommended if the
5581 target hardware has a floating point unit or if a lot of floating point
5582 arithmetic is going to be performed by the code. The default is
5583 @option{-mno-apcs-float}, since integer only code is slightly increased in
5584 size if @option{-mapcs-float} is used.
5586 @c not currently implemented
5587 @item -mapcs-reentrant
5588 @opindex mapcs-reentrant
5589 Generate reentrant, position independent code. The default is
5590 @option{-mno-apcs-reentrant}.
5593 @item -mthumb-interwork
5594 @opindex mthumb-interwork
5595 Generate code which supports calling between the ARM and Thumb
5596 instruction sets. Without this option the two instruction sets cannot
5597 be reliably used inside one program. The default is
5598 @option{-mno-thumb-interwork}, since slightly larger code is generated
5599 when @option{-mthumb-interwork} is specified.
5601 @item -mno-sched-prolog
5602 @opindex mno-sched-prolog
5603 Prevent the reordering of instructions in the function prolog, or the
5604 merging of those instruction with the instructions in the function's
5605 body. This means that all functions will start with a recognizable set
5606 of instructions (or in fact one of a choice from a small set of
5607 different function prologues), and this information can be used to
5608 locate the start if functions inside an executable piece of code. The
5609 default is @option{-msched-prolog}.
5612 @opindex mhard-float
5613 Generate output containing floating point instructions. This is the
5617 @opindex msoft-float
5618 Generate output containing library calls for floating point.
5619 @strong{Warning:} the requisite libraries are not available for all ARM
5620 targets. Normally the facilities of the machine's usual C compiler are
5621 used, but this cannot be done directly in cross-compilation. You must make
5622 your own arrangements to provide suitable library functions for
5625 @option{-msoft-float} changes the calling convention in the output file;
5626 therefore, it is only useful if you compile @emph{all} of a program with
5627 this option. In particular, you need to compile @file{libgcc.a}, the
5628 library that comes with GCC, with @option{-msoft-float} in order for
5631 @item -mlittle-endian
5632 @opindex mlittle-endian
5633 Generate code for a processor running in little-endian mode. This is
5634 the default for all standard configurations.
5637 @opindex mbig-endian
5638 Generate code for a processor running in big-endian mode; the default is
5639 to compile code for a little-endian processor.
5641 @item -mwords-little-endian
5642 @opindex mwords-little-endian
5643 This option only applies when generating code for big-endian processors.
5644 Generate code for a little-endian word order but a big-endian byte
5645 order. That is, a byte order of the form @samp{32107654}. Note: this
5646 option should only be used if you require compatibility with code for
5647 big-endian ARM processors generated by versions of the compiler prior to
5650 @item -malignment-traps
5651 @opindex malignment-traps
5652 Generate code that will not trap if the MMU has alignment traps enabled.
5653 On ARM architectures prior to ARMv4, there were no instructions to
5654 access half-word objects stored in memory. However, when reading from
5655 memory a feature of the ARM architecture allows a word load to be used,
5656 even if the address is unaligned, and the processor core will rotate the
5657 data as it is being loaded. This option tells the compiler that such
5658 misaligned accesses will cause a MMU trap and that it should instead
5659 synthesise the access as a series of byte accesses. The compiler can
5660 still use word accesses to load half-word data if it knows that the
5661 address is aligned to a word boundary.
5663 This option is ignored when compiling for ARM architecture 4 or later,
5664 since these processors have instructions to directly access half-word
5667 @item -mno-alignment-traps
5668 @opindex mno-alignment-traps
5669 Generate code that assumes that the MMU will not trap unaligned
5670 accesses. This produces better code when the target instruction set
5671 does not have half-word memory operations (i.e.@: implementations prior to
5674 Note that you cannot use this option to access unaligned word objects,
5675 since the processor will only fetch one 32-bit aligned object from
5678 The default setting for most targets is @option{-mno-alignment-traps}, since
5679 this produces better code when there are no half-word memory
5680 instructions available.
5682 @item -mshort-load-bytes
5683 @itemx -mno-short-load-words
5684 @opindex mshort-load-bytes
5685 @opindex mno-short-load-words
5686 These are deprecated aliases for @option{-malignment-traps}.
5688 @item -mno-short-load-bytes
5689 @itemx -mshort-load-words
5690 @opindex mno-short-load-bytes
5691 @opindex mshort-load-words
5692 This are deprecated aliases for @option{-mno-alignment-traps}.
5696 This option only applies to RISC iX@. Emulate the native BSD-mode
5697 compiler. This is the default if @option{-ansi} is not specified.
5701 This option only applies to RISC iX@. Emulate the native X/Open-mode
5704 @item -mno-symrename
5705 @opindex mno-symrename
5706 This option only applies to RISC iX@. Do not run the assembler
5707 post-processor, @samp{symrename}, after code has been assembled.
5708 Normally it is necessary to modify some of the standard symbols in
5709 preparation for linking with the RISC iX C library; this option
5710 suppresses this pass. The post-processor is never run when the
5711 compiler is built for cross-compilation.
5713 @item -mcpu=@var{name}
5715 This specifies the name of the target ARM processor. GCC uses this name
5716 to determine what kind of instructions it can emit when generating
5717 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
5718 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
5719 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
5720 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
5721 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
5722 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
5723 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
5724 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
5725 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
5726 @samp{arm1020t}, @samp{xscale}.
5728 @itemx -mtune=@var{name}
5730 This option is very similar to the @option{-mcpu=} option, except that
5731 instead of specifying the actual target processor type, and hence
5732 restricting which instructions can be used, it specifies that GCC should
5733 tune the performance of the code as if the target were of the type
5734 specified in this option, but still choosing the instructions that it
5735 will generate based on the cpu specified by a @option{-mcpu=} option.
5736 For some ARM implementations better performance can be obtained by using
5739 @item -march=@var{name}
5741 This specifies the name of the target ARM architecture. GCC uses this
5742 name to determine what kind of instructions it can emit when generating
5743 assembly code. This option can be used in conjunction with or instead
5744 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
5745 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
5746 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
5748 @item -mfpe=@var{number}
5749 @itemx -mfp=@var{number}
5752 This specifies the version of the floating point emulation available on
5753 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
5754 for @option{-mfpe=}, for compatibility with older versions of GCC@.
5756 @item -mstructure-size-boundary=@var{n}
5757 @opindex mstructure-size-boundary
5758 The size of all structures and unions will be rounded up to a multiple
5759 of the number of bits set by this option. Permissible values are 8 and
5760 32. The default value varies for different toolchains. For the COFF
5761 targeted toolchain the default value is 8. Specifying the larger number
5762 can produce faster, more efficient code, but can also increase the size
5763 of the program. The two values are potentially incompatible. Code
5764 compiled with one value cannot necessarily expect to work with code or
5765 libraries compiled with the other value, if they exchange information
5766 using structures or unions.
5768 @item -mabort-on-noreturn
5769 @opindex mabort-on-noreturn
5770 Generate a call to the function @code{abort} at the end of a
5771 @code{noreturn} function. It will be executed if the function tries to
5775 @itemx -mno-long-calls
5776 @opindex mlong-calls
5777 @opindex mno-long-calls
5778 Tells the compiler to perform function calls by first loading the
5779 address of the function into a register and then performing a subroutine
5780 call on this register. This switch is needed if the target function
5781 will lie outside of the 64 megabyte addressing range of the offset based
5782 version of subroutine call instruction.
5784 Even if this switch is enabled, not all function calls will be turned
5785 into long calls. The heuristic is that static functions, functions
5786 which have the @samp{short-call} attribute, functions that are inside
5787 the scope of a @samp{#pragma no_long_calls} directive and functions whose
5788 definitions have already been compiled within the current compilation
5789 unit, will not be turned into long calls. The exception to this rule is
5790 that weak function definitions, functions with the @samp{long-call}
5791 attribute or the @samp{section} attribute, and functions that are within
5792 the scope of a @samp{#pragma long_calls} directive, will always be
5793 turned into long calls.
5795 This feature is not enabled by default. Specifying
5796 @option{-mno-long-calls} will restore the default behavior, as will
5797 placing the function calls within the scope of a @samp{#pragma
5798 long_calls_off} directive. Note these switches have no effect on how
5799 the compiler generates code to handle function calls via function
5802 @item -mnop-fun-dllimport
5803 @opindex mnop-fun-dllimport
5804 Disable support for the @code{dllimport} attribute.
5806 @item -msingle-pic-base
5807 @opindex msingle-pic-base
5808 Treat the register used for PIC addressing as read-only, rather than
5809 loading it in the prologue for each function. The run-time system is
5810 responsible for initializing this register with an appropriate value
5811 before execution begins.
5813 @item -mpic-register=@var{reg}
5814 @opindex mpic-register
5815 Specify the register to be used for PIC addressing. The default is R10
5816 unless stack-checking is enabled, when R9 is used.
5818 @item -mpoke-function-name
5819 @opindex mpoke-function-name
5820 Write the name of each function into the text section, directly
5821 preceding the function prologue. The generated code is similar to this:
5825 .ascii "arm_poke_function_name", 0
5828 .word 0xff000000 + (t1 - t0)
5829 arm_poke_function_name
5831 stmfd sp!, @{fp, ip, lr, pc@}
5835 When performing a stack backtrace, code can inspect the value of
5836 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
5837 location @code{pc - 12} and the top 8 bits are set, then we know that
5838 there is a function name embedded immediately preceding this location
5839 and has length @code{((pc[-3]) & 0xff000000)}.
5843 Generate code for the 16-bit Thumb instruction set. The default is to
5844 use the 32-bit ARM instruction set.
5847 @opindex mtpcs-frame
5848 Generate a stack frame that is compliant with the Thumb Procedure Call
5849 Standard for all non-leaf functions. (A leaf function is one that does
5850 not call any other functions.) The default is @option{-mno-tpcs-frame}.
5852 @item -mtpcs-leaf-frame
5853 @opindex mtpcs-leaf-frame
5854 Generate a stack frame that is compliant with the Thumb Procedure Call
5855 Standard for all leaf functions. (A leaf function is one that does
5856 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
5858 @item -mcallee-super-interworking
5859 @opindex mcallee-super-interworking
5860 Gives all externally visible functions in the file being compiled an ARM
5861 instruction set header which switches to Thumb mode before executing the
5862 rest of the function. This allows these functions to be called from
5863 non-interworking code.
5865 @item -mcaller-super-interworking
5866 @opindex mcaller-super-interworking
5867 Allows calls via function pointers (including virtual functions) to
5868 execute correctly regardless of whether the target code has been
5869 compiled for interworking or not. There is a small overhead in the cost
5870 of executing a function pointer if this option is enabled.
5874 @node MN10200 Options
5875 @subsection MN10200 Options
5876 @cindex MN10200 options
5877 These @option{-m} options are defined for Matsushita MN10200 architectures:
5882 Indicate to the linker that it should perform a relaxation optimization pass
5883 to shorten branches, calls and absolute memory addresses. This option only
5884 has an effect when used on the command line for the final link step.
5886 This option makes symbolic debugging impossible.
5889 @node MN10300 Options
5890 @subsection MN10300 Options
5891 @cindex MN10300 options
5892 These @option{-m} options are defined for Matsushita MN10300 architectures:
5897 Generate code to avoid bugs in the multiply instructions for the MN10300
5898 processors. This is the default.
5901 @opindex mno-mult-bug
5902 Do not generate code to avoid bugs in the multiply instructions for the
5907 Generate code which uses features specific to the AM33 processor.
5911 Do not generate code which uses features specific to the AM33 processor. This
5916 Do not link in the C run-time initialization object file.
5920 Indicate to the linker that it should perform a relaxation optimization pass
5921 to shorten branches, calls and absolute memory addresses. This option only
5922 has an effect when used on the command line for the final link step.
5924 This option makes symbolic debugging impossible.
5928 @node M32R/D Options
5929 @subsection M32R/D Options
5930 @cindex M32R/D options
5932 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
5937 Generate code for the M32R/X@.
5941 Generate code for the M32R@. This is the default.
5943 @item -mcode-model=small
5944 @opindex mcode-model=small
5945 Assume all objects live in the lower 16MB of memory (so that their addresses
5946 can be loaded with the @code{ld24} instruction), and assume all subroutines
5947 are reachable with the @code{bl} instruction.
5948 This is the default.
5950 The addressability of a particular object can be set with the
5951 @code{model} attribute.
5953 @item -mcode-model=medium
5954 @opindex mcode-model=medium
5955 Assume objects may be anywhere in the 32-bit address space (the compiler
5956 will generate @code{seth/add3} instructions to load their addresses), and
5957 assume all subroutines are reachable with the @code{bl} instruction.
5959 @item -mcode-model=large
5960 @opindex mcode-model=large
5961 Assume objects may be anywhere in the 32-bit address space (the compiler
5962 will generate @code{seth/add3} instructions to load their addresses), and
5963 assume subroutines may not be reachable with the @code{bl} instruction
5964 (the compiler will generate the much slower @code{seth/add3/jl}
5965 instruction sequence).
5968 @opindex msdata=none
5969 Disable use of the small data area. Variables will be put into
5970 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
5971 @code{section} attribute has been specified).
5972 This is the default.
5974 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
5975 Objects may be explicitly put in the small data area with the
5976 @code{section} attribute using one of these sections.
5979 @opindex msdata=sdata
5980 Put small global and static data in the small data area, but do not
5981 generate special code to reference them.
5985 Put small global and static data in the small data area, and generate
5986 special instructions to reference them.
5990 @cindex smaller data references
5991 Put global and static objects less than or equal to @var{num} bytes
5992 into the small data or bss sections instead of the normal data or bss
5993 sections. The default value of @var{num} is 8.
5994 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
5995 for this option to have any effect.
5997 All modules should be compiled with the same @option{-G @var{num}} value.
5998 Compiling with different values of @var{num} may or may not work; if it
5999 doesn't the linker will give an error message---incorrect code will not be
6005 @subsection M88K Options
6006 @cindex M88k options
6008 These @samp{-m} options are defined for Motorola 88k architectures:
6013 Generate code that works well on both the m88100 and the
6018 Generate code that works best for the m88100, but that also
6023 Generate code that works best for the m88110, and may not run
6028 Obsolete option to be removed from the next revision.
6031 @item -midentify-revision
6032 @opindex midentify-revision
6033 @cindex identifying source, compiler (88k)
6034 Include an @code{ident} directive in the assembler output recording the
6035 source file name, compiler name and version, timestamp, and compilation
6038 @item -mno-underscores
6039 @opindex mno-underscores
6040 @cindex underscores, avoiding (88k)
6041 In assembler output, emit symbol names without adding an underscore
6042 character at the beginning of each name. The default is to use an
6043 underscore as prefix on each name.
6045 @item -mocs-debug-info
6046 @itemx -mno-ocs-debug-info
6047 @opindex mocs-debug-info
6048 @opindex mno-ocs-debug-info
6050 @cindex debugging, 88k OCS
6051 Include (or omit) additional debugging information (about registers used
6052 in each stack frame) as specified in the 88open Object Compatibility
6053 Standard, ``OCS''@. This extra information allows debugging of code that
6054 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6055 Delta 88 SVr3.2 is to include this information; other 88k configurations
6056 omit this information by default.
6058 @item -mocs-frame-position
6059 @opindex mocs-frame-position
6060 @cindex register positions in frame (88k)
6061 When emitting COFF debugging information for automatic variables and
6062 parameters stored on the stack, use the offset from the canonical frame
6063 address, which is the stack pointer (register 31) on entry to the
6064 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6065 @option{-mocs-frame-position}; other 88k configurations have the default
6066 @option{-mno-ocs-frame-position}.
6068 @item -mno-ocs-frame-position
6069 @opindex mno-ocs-frame-position
6070 @cindex register positions in frame (88k)
6071 When emitting COFF debugging information for automatic variables and
6072 parameters stored on the stack, use the offset from the frame pointer
6073 register (register 30). When this option is in effect, the frame
6074 pointer is not eliminated when debugging information is selected by the
6077 @item -moptimize-arg-area
6078 @opindex moptimize-arg-area
6079 @cindex arguments in frame (88k)
6080 Save space by reorganizing the stack frame. This option generates code
6081 that does not agree with the 88open specifications, but uses less
6084 @itemx -mno-optimize-arg-area
6085 @opindex mno-optimize-arg-area
6086 Do not reorganize the stack frame to save space. This is the default.
6087 The generated conforms to the specification, but uses more memory.
6089 @item -mshort-data-@var{num}
6090 @opindex mshort-data
6091 @cindex smaller data references (88k)
6092 @cindex r0-relative references (88k)
6093 Generate smaller data references by making them relative to @code{r0},
6094 which allows loading a value using a single instruction (rather than the
6095 usual two). You control which data references are affected by
6096 specifying @var{num} with this option. For example, if you specify
6097 @option{-mshort-data-512}, then the data references affected are those
6098 involving displacements of less than 512 bytes.
6099 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6102 @item -mserialize-volatile
6103 @opindex mserialize-volatile
6104 @itemx -mno-serialize-volatile
6105 @opindex mno-serialize-volatile
6106 @cindex sequential consistency on 88k
6107 Do, or don't, generate code to guarantee sequential consistency
6108 of volatile memory references. By default, consistency is
6111 The order of memory references made by the MC88110 processor does
6112 not always match the order of the instructions requesting those
6113 references. In particular, a load instruction may execute before
6114 a preceding store instruction. Such reordering violates
6115 sequential consistency of volatile memory references, when there
6116 are multiple processors. When consistency must be guaranteed,
6117 GCC generates special instructions, as needed, to force
6118 execution in the proper order.
6120 The MC88100 processor does not reorder memory references and so
6121 always provides sequential consistency. However, by default, GCC
6122 generates the special instructions to guarantee consistency
6123 even when you use @option{-m88100}, so that the code may be run on an
6124 MC88110 processor. If you intend to run your code only on the
6125 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6127 The extra code generated to guarantee consistency may affect the
6128 performance of your application. If you know that you can safely
6129 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6135 @cindex assembler syntax, 88k
6137 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6138 related to System V release 4 (SVr4). This controls the following:
6142 Which variant of the assembler syntax to emit.
6144 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6145 that is used on System V release 4.
6147 @option{-msvr4} makes GCC issue additional declaration directives used in
6151 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6152 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6153 other m88k configurations.
6155 @item -mversion-03.00
6156 @opindex mversion-03.00
6157 This option is obsolete, and is ignored.
6158 @c ??? which asm syntax better for GAS? option there too?
6160 @item -mno-check-zero-division
6161 @itemx -mcheck-zero-division
6162 @opindex mno-check-zero-division
6163 @opindex mcheck-zero-division
6164 @cindex zero division on 88k
6165 Do, or don't, generate code to guarantee that integer division by
6166 zero will be detected. By default, detection is guaranteed.
6168 Some models of the MC88100 processor fail to trap upon integer
6169 division by zero under certain conditions. By default, when
6170 compiling code that might be run on such a processor, GCC
6171 generates code that explicitly checks for zero-valued divisors
6172 and traps with exception number 503 when one is detected. Use of
6173 @option{-mno-check-zero-division} suppresses such checking for code
6174 generated to run on an MC88100 processor.
6176 GCC assumes that the MC88110 processor correctly detects all instances
6177 of integer division by zero. When @option{-m88110} is specified, no
6178 explicit checks for zero-valued divisors are generated, and both
6179 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6182 @item -muse-div-instruction
6183 @opindex muse-div-instruction
6184 @cindex divide instruction, 88k
6185 Use the div instruction for signed integer division on the
6186 MC88100 processor. By default, the div instruction is not used.
6188 On the MC88100 processor the signed integer division instruction
6189 div) traps to the operating system on a negative operand. The
6190 operating system transparently completes the operation, but at a
6191 large cost in execution time. By default, when compiling code
6192 that might be run on an MC88100 processor, GCC emulates signed
6193 integer division using the unsigned integer division instruction
6194 divu), thereby avoiding the large penalty of a trap to the
6195 operating system. Such emulation has its own, smaller, execution
6196 cost in both time and space. To the extent that your code's
6197 important signed integer division operations are performed on two
6198 nonnegative operands, it may be desirable to use the div
6199 instruction directly.
6201 On the MC88110 processor the div instruction (also known as the
6202 divs instruction) processes negative operands without trapping to
6203 the operating system. When @option{-m88110} is specified,
6204 @option{-muse-div-instruction} is ignored, and the div instruction is used
6205 for signed integer division.
6207 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6208 particular, the behavior of such a division with and without
6209 @option{-muse-div-instruction} may differ.
6211 @item -mtrap-large-shift
6212 @itemx -mhandle-large-shift
6213 @opindex mtrap-large-shift
6214 @opindex mhandle-large-shift
6215 @cindex bit shift overflow (88k)
6216 @cindex large bit shifts (88k)
6217 Include code to detect bit-shifts of more than 31 bits; respectively,
6218 trap such shifts or emit code to handle them properly. By default GCC
6219 makes no special provision for large bit shifts.
6221 @item -mwarn-passed-structs
6222 @opindex mwarn-passed-structs
6223 @cindex structure passing (88k)
6224 Warn when a function passes a struct as an argument or result.
6225 Structure-passing conventions have changed during the evolution of the C
6226 language, and are often the source of portability problems. By default,
6227 GCC issues no such warning.
6230 @c break page here to avoid unsightly interparagraph stretch.
6234 @node RS/6000 and PowerPC Options
6235 @subsection IBM RS/6000 and PowerPC Options
6236 @cindex RS/6000 and PowerPC Options
6237 @cindex IBM RS/6000 and PowerPC Options
6239 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6247 @itemx -mpowerpc-gpopt
6248 @itemx -mno-powerpc-gpopt
6249 @itemx -mpowerpc-gfxopt
6250 @itemx -mno-powerpc-gfxopt
6252 @itemx -mno-powerpc64
6258 @opindex mno-powerpc
6259 @opindex mpowerpc-gpopt
6260 @opindex mno-powerpc-gpopt
6261 @opindex mpowerpc-gfxopt
6262 @opindex mno-powerpc-gfxopt
6264 @opindex mno-powerpc64
6265 GCC supports two related instruction set architectures for the
6266 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6267 instructions supported by the @samp{rios} chip set used in the original
6268 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6269 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6270 the IBM 4xx microprocessors.
6272 Neither architecture is a subset of the other. However there is a
6273 large common subset of instructions supported by both. An MQ
6274 register is included in processors supporting the POWER architecture.
6276 You use these options to specify which instructions are available on the
6277 processor you are using. The default value of these options is
6278 determined when configuring GCC@. Specifying the
6279 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6280 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6281 rather than the options listed above.
6283 The @option{-mpower} option allows GCC to generate instructions that
6284 are found only in the POWER architecture and to use the MQ register.
6285 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6286 to generate instructions that are present in the POWER2 architecture but
6287 not the original POWER architecture.
6289 The @option{-mpowerpc} option allows GCC to generate instructions that
6290 are found only in the 32-bit subset of the PowerPC architecture.
6291 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6292 GCC to use the optional PowerPC architecture instructions in the
6293 General Purpose group, including floating-point square root. Specifying
6294 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6295 use the optional PowerPC architecture instructions in the Graphics
6296 group, including floating-point select.
6298 The @option{-mpowerpc64} option allows GCC to generate the additional
6299 64-bit instructions that are found in the full PowerPC64 architecture
6300 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6301 @option{-mno-powerpc64}.
6303 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6304 will use only the instructions in the common subset of both
6305 architectures plus some special AIX common-mode calls, and will not use
6306 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6307 permits GCC to use any instruction from either architecture and to
6308 allow use of the MQ register; specify this for the Motorola MPC601.
6310 @item -mnew-mnemonics
6311 @itemx -mold-mnemonics
6312 @opindex mnew-mnemonics
6313 @opindex mold-mnemonics
6314 Select which mnemonics to use in the generated assembler code. With
6315 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6316 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6317 assembler mnemonics defined for the POWER architecture. Instructions
6318 defined in only one architecture have only one mnemonic; GCC uses that
6319 mnemonic irrespective of which of these options is specified.
6321 GCC defaults to the mnemonics appropriate for the architecture in
6322 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6323 value of these option. Unless you are building a cross-compiler, you
6324 should normally not specify either @option{-mnew-mnemonics} or
6325 @option{-mold-mnemonics}, but should instead accept the default.
6327 @item -mcpu=@var{cpu_type}
6329 Set architecture type, register usage, choice of mnemonics, and
6330 instruction scheduling parameters for machine type @var{cpu_type}.
6331 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6332 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6333 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6334 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6335 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6336 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6338 @option{-mcpu=common} selects a completely generic processor. Code
6339 generated under this option will run on any POWER or PowerPC processor.
6340 GCC will use only the instructions in the common subset of both
6341 architectures, and will not use the MQ register. GCC assumes a generic
6342 processor model for scheduling purposes.
6344 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6345 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6346 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6347 types, with an appropriate, generic processor model assumed for
6348 scheduling purposes.
6350 The other options specify a specific processor. Code generated under
6351 those options will run best on that processor, and may not run at all on
6354 The @option{-mcpu} options automatically enable or disable other
6355 @option{-m} options as follows:
6359 @option{-mno-power}, @option{-mno-powerc}
6366 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6381 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6384 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6389 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6392 @item -mtune=@var{cpu_type}
6394 Set the instruction scheduling parameters for machine type
6395 @var{cpu_type}, but do not set the architecture type, register usage, or
6396 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6397 values for @var{cpu_type} are used for @option{-mtune} as for
6398 @option{-mcpu}. If both are specified, the code generated will use the
6399 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6400 scheduling parameters set by @option{-mtune}.
6405 @opindex mno-altivec
6406 These switches enable or disable the use of built-in functions that
6407 allow access to the AltiVec instruction set. You may also need to set
6408 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6412 @itemx -mno-fp-in-toc
6413 @itemx -mno-sum-in-toc
6414 @itemx -mminimal-toc
6416 @opindex mno-fp-in-toc
6417 @opindex mno-sum-in-toc
6418 @opindex mminimal-toc
6419 Modify generation of the TOC (Table Of Contents), which is created for
6420 every executable file. The @option{-mfull-toc} option is selected by
6421 default. In that case, GCC will allocate at least one TOC entry for
6422 each unique non-automatic variable reference in your program. GCC
6423 will also place floating-point constants in the TOC@. However, only
6424 16,384 entries are available in the TOC@.
6426 If you receive a linker error message that saying you have overflowed
6427 the available TOC space, you can reduce the amount of TOC space used
6428 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6429 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6430 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6431 generate code to calculate the sum of an address and a constant at
6432 run-time instead of putting that sum into the TOC@. You may specify one
6433 or both of these options. Each causes GCC to produce very slightly
6434 slower and larger code at the expense of conserving TOC space.
6436 If you still run out of space in the TOC even when you specify both of
6437 these options, specify @option{-mminimal-toc} instead. This option causes
6438 GCC to make only one TOC entry for every file. When you specify this
6439 option, GCC will produce code that is slower and larger but which
6440 uses extremely little TOC space. You may wish to use this option
6441 only on files that contain less frequently executed code.
6447 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6448 @code{long} type, and the infrastructure needed to support them.
6449 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6450 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6451 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6456 @opindex mno-xl-call
6457 On AIX, pass floating-point arguments to prototyped functions beyond the
6458 register save area (RSA) on the stack in addition to argument FPRs. The
6459 AIX calling convention was extended but not initially documented to
6460 handle an obscure K&R C case of calling a function that takes the
6461 address of its arguments with fewer arguments than declared. AIX XL
6462 compilers access floating point arguments which do not fit in the
6463 RSA from the stack when a subroutine is compiled without
6464 optimization. Because always storing floating-point arguments on the
6465 stack is inefficient and rarely needed, this option is not enabled by
6466 default and only is necessary when calling subroutines compiled by AIX
6467 XL compilers without optimization.
6471 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6472 application written to use message passing with special startup code to
6473 enable the application to run. The system must have PE installed in the
6474 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6475 must be overridden with the @option{-specs=} option to specify the
6476 appropriate directory location. The Parallel Environment does not
6477 support threads, so the @option{-mpe} option and the @option{-pthread}
6478 option are incompatible.
6482 @opindex msoft-float
6483 @opindex mhard-float
6484 Generate code that does not use (uses) the floating-point register set.
6485 Software floating point emulation is provided if you use the
6486 @option{-msoft-float} option, and pass the option to GCC when linking.
6489 @itemx -mno-multiple
6491 @opindex mno-multiple
6492 Generate code that uses (does not use) the load multiple word
6493 instructions and the store multiple word instructions. These
6494 instructions are generated by default on POWER systems, and not
6495 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6496 endian PowerPC systems, since those instructions do not work when the
6497 processor is in little endian mode. The exceptions are PPC740 and
6498 PPC750 which permit the instructions usage in little endian mode.
6504 Generate code that uses (does not use) the load string instructions
6505 and the store string word instructions to save multiple registers and
6506 do small block moves. These instructions are generated by default on
6507 POWER systems, and not generated on PowerPC systems. Do not use
6508 @option{-mstring} on little endian PowerPC systems, since those
6509 instructions do not work when the processor is in little endian mode.
6510 The exceptions are PPC740 and PPC750 which permit the instructions
6511 usage in little endian mode.
6517 Generate code that uses (does not use) the load or store instructions
6518 that update the base register to the address of the calculated memory
6519 location. These instructions are generated by default. If you use
6520 @option{-mno-update}, there is a small window between the time that the
6521 stack pointer is updated and the address of the previous frame is
6522 stored, which means code that walks the stack frame across interrupts or
6523 signals may get corrupted data.
6526 @itemx -mno-fused-madd
6527 @opindex mfused-madd
6528 @opindex mno-fused-madd
6529 Generate code that uses (does not use) the floating point multiply and
6530 accumulate instructions. These instructions are generated by default if
6531 hardware floating is used.
6533 @item -mno-bit-align
6535 @opindex mno-bit-align
6537 On System V.4 and embedded PowerPC systems do not (do) force structures
6538 and unions that contain bit-fields to be aligned to the base type of the
6541 For example, by default a structure containing nothing but 8
6542 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6543 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6544 the structure would be aligned to a 1 byte boundary and be one byte in
6547 @item -mno-strict-align
6548 @itemx -mstrict-align
6549 @opindex mno-strict-align
6550 @opindex mstrict-align
6551 On System V.4 and embedded PowerPC systems do not (do) assume that
6552 unaligned memory references will be handled by the system.
6555 @itemx -mno-relocatable
6556 @opindex mrelocatable
6557 @opindex mno-relocatable
6558 On embedded PowerPC systems generate code that allows (does not allow)
6559 the program to be relocated to a different address at runtime. If you
6560 use @option{-mrelocatable} on any module, all objects linked together must
6561 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6563 @item -mrelocatable-lib
6564 @itemx -mno-relocatable-lib
6565 @opindex mrelocatable-lib
6566 @opindex mno-relocatable-lib
6567 On embedded PowerPC systems generate code that allows (does not allow)
6568 the program to be relocated to a different address at runtime. Modules
6569 compiled with @option{-mrelocatable-lib} can be linked with either modules
6570 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6571 with modules compiled with the @option{-mrelocatable} options.
6577 On System V.4 and embedded PowerPC systems do not (do) assume that
6578 register 2 contains a pointer to a global area pointing to the addresses
6579 used in the program.
6582 @itemx -mlittle-endian
6584 @opindex mlittle-endian
6585 On System V.4 and embedded PowerPC systems compile code for the
6586 processor in little endian mode. The @option{-mlittle-endian} option is
6587 the same as @option{-mlittle}.
6592 @opindex mbig-endian
6593 On System V.4 and embedded PowerPC systems compile code for the
6594 processor in big endian mode. The @option{-mbig-endian} option is
6595 the same as @option{-mbig}.
6599 On System V.4 and embedded PowerPC systems compile code using calling
6600 conventions that adheres to the March 1995 draft of the System V
6601 Application Binary Interface, PowerPC processor supplement. This is the
6602 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6604 @item -mcall-sysv-eabi
6605 @opindex mcall-sysv-eabi
6606 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6608 @item -mcall-sysv-noeabi
6609 @opindex mcall-sysv-noeabi
6610 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6614 On System V.4 and embedded PowerPC systems compile code using calling
6615 conventions that are similar to those used on AIX@. This is the
6616 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6618 @item -mcall-solaris
6619 @opindex mcall-solaris
6620 On System V.4 and embedded PowerPC systems compile code for the Solaris
6624 @opindex mcall-linux
6625 On System V.4 and embedded PowerPC systems compile code for the
6626 Linux-based GNU system.
6630 On System V.4 and embedded PowerPC systems compile code for the
6631 Hurd-based GNU system.
6634 @opindex mcall-netbsd
6635 On System V.4 and embedded PowerPC systems compile code for the
6636 NetBSD operating system.
6638 @item -maix-struct-return
6639 @opindex maix-struct-return
6640 Return all structures in memory (as specified by the AIX ABI)@.
6642 @item -msvr4-struct-return
6643 @opindex msvr4-struct-return
6644 Return structures smaller than 8 bytes in registers (as specified by the
6648 @opindex mabi=altivec
6649 Extend the current ABI with AltiVec ABI extensions. This does not
6650 change the default ABI, instead it adds the AltiVec ABI extensions to
6653 @item -mabi=no-altivec
6654 @opindex mabi=no-altivec
6655 Disable AltiVec ABI extensions for the current ABI.
6658 @itemx -mno-prototype
6660 @opindex mno-prototype
6661 On System V.4 and embedded PowerPC systems assume that all calls to
6662 variable argument functions are properly prototyped. Otherwise, the
6663 compiler must insert an instruction before every non prototyped call to
6664 set or clear bit 6 of the condition code register (@var{CR}) to
6665 indicate whether floating point values were passed in the floating point
6666 registers in case the function takes a variable arguments. With
6667 @option{-mprototype}, only calls to prototyped variable argument functions
6668 will set or clear the bit.
6672 On embedded PowerPC systems, assume that the startup module is called
6673 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6674 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6679 On embedded PowerPC systems, assume that the startup module is called
6680 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6685 On embedded PowerPC systems, assume that the startup module is called
6686 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6690 @opindex myellowknife
6691 On embedded PowerPC systems, assume that the startup module is called
6692 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6697 On System V.4 and embedded PowerPC systems, specify that you are
6698 compiling for a VxWorks system.
6702 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
6703 header to indicate that @samp{eabi} extended relocations are used.
6709 On System V.4 and embedded PowerPC systems do (do not) adhere to the
6710 Embedded Applications Binary Interface (eabi) which is a set of
6711 modifications to the System V.4 specifications. Selecting @option{-meabi}
6712 means that the stack is aligned to an 8 byte boundary, a function
6713 @code{__eabi} is called to from @code{main} to set up the eabi
6714 environment, and the @option{-msdata} option can use both @code{r2} and
6715 @code{r13} to point to two separate small data areas. Selecting
6716 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
6717 do not call an initialization function from @code{main}, and the
6718 @option{-msdata} option will only use @code{r13} to point to a single
6719 small data area. The @option{-meabi} option is on by default if you
6720 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
6723 @opindex msdata=eabi
6724 On System V.4 and embedded PowerPC systems, put small initialized
6725 @code{const} global and static data in the @samp{.sdata2} section, which
6726 is pointed to by register @code{r2}. Put small initialized
6727 non-@code{const} global and static data in the @samp{.sdata} section,
6728 which is pointed to by register @code{r13}. Put small uninitialized
6729 global and static data in the @samp{.sbss} section, which is adjacent to
6730 the @samp{.sdata} section. The @option{-msdata=eabi} option is
6731 incompatible with the @option{-mrelocatable} option. The
6732 @option{-msdata=eabi} option also sets the @option{-memb} option.
6735 @opindex msdata=sysv
6736 On System V.4 and embedded PowerPC systems, put small global and static
6737 data in the @samp{.sdata} section, which is pointed to by register
6738 @code{r13}. Put small uninitialized global and static data in the
6739 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
6740 The @option{-msdata=sysv} option is incompatible with the
6741 @option{-mrelocatable} option.
6743 @item -msdata=default
6745 @opindex msdata=default
6747 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
6748 compile code the same as @option{-msdata=eabi}, otherwise compile code the
6749 same as @option{-msdata=sysv}.
6752 @opindex msdata-data
6753 On System V.4 and embedded PowerPC systems, put small global and static
6754 data in the @samp{.sdata} section. Put small uninitialized global and
6755 static data in the @samp{.sbss} section. Do not use register @code{r13}
6756 to address small data however. This is the default behavior unless
6757 other @option{-msdata} options are used.
6761 @opindex msdata=none
6763 On embedded PowerPC systems, put all initialized global and static data
6764 in the @samp{.data} section, and all uninitialized data in the
6765 @samp{.bss} section.
6769 @cindex smaller data references (PowerPC)
6770 @cindex .sdata/.sdata2 references (PowerPC)
6771 On embedded PowerPC systems, put global and static items less than or
6772 equal to @var{num} bytes into the small data or bss sections instead of
6773 the normal data or bss section. By default, @var{num} is 8. The
6774 @option{-G @var{num}} switch is also passed to the linker.
6775 All modules should be compiled with the same @option{-G @var{num}} value.
6778 @itemx -mno-regnames
6780 @opindex mno-regnames
6781 On System V.4 and embedded PowerPC systems do (do not) emit register
6782 names in the assembly language output using symbolic forms.
6786 Adds support for multithreading with the @dfn{pthreads} library.
6787 This option sets flags for both the preprocessor and linker.
6792 @subsection IBM RT Options
6794 @cindex IBM RT options
6796 These @samp{-m} options are defined for the IBM RT PC:
6800 @opindex min-line-mul
6801 Use an in-line code sequence for integer multiplies. This is the
6804 @item -mcall-lib-mul
6805 @opindex mcall-lib-mul
6806 Call @code{lmul$$} for integer multiples.
6808 @item -mfull-fp-blocks
6809 @opindex mfull-fp-blocks
6810 Generate full-size floating point data blocks, including the minimum
6811 amount of scratch space recommended by IBM@. This is the default.
6813 @item -mminimum-fp-blocks
6814 @opindex mminimum-fp-blocks
6815 Do not include extra scratch space in floating point data blocks. This
6816 results in smaller code, but slower execution, since scratch space must
6817 be allocated dynamically.
6819 @cindex @file{varargs.h} and RT PC
6820 @cindex @file{stdarg.h} and RT PC
6821 @item -mfp-arg-in-fpregs
6822 @opindex mfp-arg-in-fpregs
6823 Use a calling sequence incompatible with the IBM calling convention in
6824 which floating point arguments are passed in floating point registers.
6825 Note that @code{varargs.h} and @code{stdarg.h} will not work with
6826 floating point operands if this option is specified.
6828 @item -mfp-arg-in-gregs
6829 @opindex mfp-arg-in-gregs
6830 Use the normal calling convention for floating point arguments. This is
6833 @item -mhc-struct-return
6834 @opindex mhc-struct-return
6835 Return structures of more than one word in memory, rather than in a
6836 register. This provides compatibility with the MetaWare HighC (hc)
6837 compiler. Use the option @option{-fpcc-struct-return} for compatibility
6838 with the Portable C Compiler (pcc).
6840 @item -mnohc-struct-return
6841 @opindex mnohc-struct-return
6842 Return some structures of more than one word in registers, when
6843 convenient. This is the default. For compatibility with the
6844 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
6845 option @option{-mhc-struct-return}.
6849 @subsection MIPS Options
6850 @cindex MIPS options
6852 These @samp{-m} options are defined for the MIPS family of computers:
6856 @item -march=@var{cpu-type}
6858 Assume the defaults for the machine type @var{cpu-type} when generating
6859 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
6860 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
6861 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
6862 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
6863 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
6864 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
6866 @item -mtune=@var{cpu-type}
6868 Assume the defaults for the machine type @var{cpu-type} when scheduling
6869 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
6870 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
6871 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
6872 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
6873 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
6874 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
6875 @var{cpu-type} will schedule things appropriately for that particular
6876 chip, the compiler will not generate any code that does not meet level 1
6877 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
6878 or @option{-mabi} switch being used.
6880 @item -mcpu=@var{cpu-type}
6882 This is identical to specifying both @option{-march} and @option{-mtune}.
6886 Issue instructions from level 1 of the MIPS ISA@. This is the default.
6887 @samp{r3000} is the default @var{cpu-type} at this ISA level.
6891 Issue instructions from level 2 of the MIPS ISA (branch likely, square
6892 root instructions). @samp{r6000} is the default @var{cpu-type} at this
6897 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
6898 @samp{r4000} is the default @var{cpu-type} at this ISA level.
6902 Issue instructions from level 4 of the MIPS ISA (conditional move,
6903 prefetch, enhanced FPU instructions). @samp{r8000} is the default
6904 @var{cpu-type} at this ISA level.
6908 Assume that 32 32-bit floating point registers are available. This is
6913 Assume that 32 64-bit floating point registers are available. This is
6914 the default when the @option{-mips3} option is used.
6917 @itemx -mno-fused-madd
6918 @opindex mfused-madd
6919 @opindex mno-fused-madd
6920 Generate code that uses (does not use) the floating point multiply and
6921 accumulate instructions, when they are available. These instructions
6922 are generated by default if they are available, but this may be
6923 undesirable if the extra precision causes problems or on certain chips
6924 in the mode where denormals are rounded to zero where denormals
6925 generated by multiply and accumulate instructions cause exceptions
6930 Assume that 32 32-bit general purpose registers are available. This is
6935 Assume that 32 64-bit general purpose registers are available. This is
6936 the default when the @option{-mips3} option is used.
6940 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
6941 explanation of the default, and the width of pointers.
6945 Force long types to be 64 bits wide. See @option{-mlong32} for an
6946 explanation of the default, and the width of pointers.
6950 Force long, int, and pointer types to be 32 bits wide.
6952 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
6953 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
6954 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
6955 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
6956 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
6957 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
6958 are 32 bits, and longs are 64 bits wide. The width of pointer types is
6959 the smaller of the width of longs or the width of general purpose
6960 registers (which in turn depends on the ISA)@.
6972 Generate code for the indicated ABI@. The default instruction level is
6973 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
6974 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
6975 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
6980 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
6981 add normal debug information. This is the default for all
6982 platforms except for the OSF/1 reference platform, using the OSF/rose
6983 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
6984 switches are used, the @file{mips-tfile} program will encapsulate the
6985 stabs within MIPS ECOFF@.
6989 Generate code for the GNU assembler. This is the default on the OSF/1
6990 reference platform, using the OSF/rose object format. Also, this is
6991 the default if the configure option @option{--with-gnu-as} is used.
6993 @item -msplit-addresses
6994 @itemx -mno-split-addresses
6995 @opindex msplit-addresses
6996 @opindex mno-split-addresses
6997 Generate code to load the high and low parts of address constants separately.
6998 This allows GCC to optimize away redundant loads of the high order
6999 bits of addresses. This optimization requires GNU as and GNU ld.
7000 This optimization is enabled by default for some embedded targets where
7001 GNU as and GNU ld are standard.
7007 The @option{-mrnames} switch says to output code using the MIPS software
7008 names for the registers, instead of the hardware names (ie, @var{a0}
7009 instead of @var{$4}). The only known assembler that supports this option
7010 is the Algorithmics assembler.
7016 The @option{-mgpopt} switch says to write all of the data declarations
7017 before the instructions in the text section, this allows the MIPS
7018 assembler to generate one word memory references instead of using two
7019 words for short global or static data items. This is on by default if
7020 optimization is selected.
7026 For each non-inline function processed, the @option{-mstats} switch
7027 causes the compiler to emit one line to the standard error file to
7028 print statistics about the program (number of registers saved, stack
7035 The @option{-mmemcpy} switch makes all block moves call the appropriate
7036 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7037 generating inline code.
7040 @itemx -mno-mips-tfile
7041 @opindex mmips-tfile
7042 @opindex mno-mips-tfile
7043 The @option{-mno-mips-tfile} switch causes the compiler not
7044 postprocess the object file with the @file{mips-tfile} program,
7045 after the MIPS assembler has generated it to add debug support. If
7046 @file{mips-tfile} is not run, then no local variables will be
7047 available to the debugger. In addition, @file{stage2} and
7048 @file{stage3} objects will have the temporary file names passed to the
7049 assembler embedded in the object file, which means the objects will
7050 not compare the same. The @option{-mno-mips-tfile} switch should only
7051 be used when there are bugs in the @file{mips-tfile} program that
7052 prevents compilation.
7055 @opindex msoft-float
7056 Generate output containing library calls for floating point.
7057 @strong{Warning:} the requisite libraries are not part of GCC@.
7058 Normally the facilities of the machine's usual C compiler are used, but
7059 this can't be done directly in cross-compilation. You must make your
7060 own arrangements to provide suitable library functions for
7064 @opindex mhard-float
7065 Generate output containing floating point instructions. This is the
7066 default if you use the unmodified sources.
7069 @itemx -mno-abicalls
7071 @opindex mno-abicalls
7072 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7073 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7074 position independent code.
7077 @itemx -mno-long-calls
7078 @opindex mlong-calls
7079 @opindex mno-long-calls
7080 Do all calls with the @samp{JALR} instruction, which requires
7081 loading up a function's address into a register before the call.
7082 You need to use this switch, if you call outside of the current
7083 512 megabyte segment to functions that are not through pointers.
7086 @itemx -mno-half-pic
7088 @opindex mno-half-pic
7089 Put pointers to extern references into the data section and load them
7090 up, rather than put the references in the text section.
7092 @item -membedded-pic
7093 @itemx -mno-embedded-pic
7094 @opindex membedded-pic
7095 @opindex mno-embedded-pic
7096 Generate PIC code suitable for some embedded systems. All calls are
7097 made using PC relative address, and all data is addressed using the $gp
7098 register. No more than 65536 bytes of global data may be used. This
7099 requires GNU as and GNU ld which do most of the work. This currently
7100 only works on targets which use ECOFF; it does not work with ELF@.
7102 @item -membedded-data
7103 @itemx -mno-embedded-data
7104 @opindex membedded-data
7105 @opindex mno-embedded-data
7106 Allocate variables to the read-only data section first if possible, then
7107 next in the small data section if possible, otherwise in data. This gives
7108 slightly slower code than the default, but reduces the amount of RAM required
7109 when executing, and thus may be preferred for some embedded systems.
7111 @item -muninit-const-in-rodata
7112 @itemx -mno-uninit-const-in-rodata
7113 @opindex muninit-const-in-rodata
7114 @opindex mno-uninit-const-in-rodata
7115 When used together with @option{-membedded-data}, it will always store uninitialized
7116 const variables in the read-only data section.
7118 @item -msingle-float
7119 @itemx -mdouble-float
7120 @opindex msingle-float
7121 @opindex mdouble-float
7122 The @option{-msingle-float} switch tells gcc to assume that the floating
7123 point coprocessor only supports single precision operations, as on the
7124 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7125 double precision operations. This is the default.
7131 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7132 as on the @samp{r4650} chip.
7136 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7137 @option{-mcpu=r4650}.
7143 Enable 16-bit instructions.
7147 Use the entry and exit pseudo ops. This option can only be used with
7152 Compile code for the processor in little endian mode.
7153 The requisite libraries are assumed to exist.
7157 Compile code for the processor in big endian mode.
7158 The requisite libraries are assumed to exist.
7162 @cindex smaller data references (MIPS)
7163 @cindex gp-relative references (MIPS)
7164 Put global and static items less than or equal to @var{num} bytes into
7165 the small data or bss sections instead of the normal data or bss
7166 section. This allows the assembler to emit one word memory reference
7167 instructions based on the global pointer (@var{gp} or @var{$28}),
7168 instead of the normal two words used. By default, @var{num} is 8 when
7169 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7170 @option{-G @var{num}} switch is also passed to the assembler and linker.
7171 All modules should be compiled with the same @option{-G @var{num}}
7176 Tell the MIPS assembler to not run its preprocessor over user
7177 assembler files (with a @samp{.s} suffix) when assembling them.
7181 Pass an option to gas which will cause nops to be inserted if
7182 the read of the destination register of an mfhi or mflo instruction
7183 occurs in the following two instructions.
7187 Do not include the default crt0.
7189 @item -mflush-func=@var{func}
7190 @itemx -mno-flush-func
7191 @opindex mflush-func
7192 Specifies the function to call to flush the I and D caches, or to not
7193 call any such function. If called, the function must take the same
7194 arguments as the common @code{_flush_func()}, that is, the address of the
7195 memory range for which the cache is being flushed, the size of the
7196 memory range, and the number 3 (to flush both caches). The default
7197 depends on the target gcc was configured for, but commonly is either
7198 @samp{_flush_func} or @samp{__cpu_flush}.
7201 These options are defined by the macro
7202 @code{TARGET_SWITCHES} in the machine description. The default for the
7203 options is also defined by that macro, which enables you to change the
7206 @node i386 and x86-64 Options
7207 @subsection Intel 386 and AMD x86-64 Options
7208 @cindex i386 Options
7209 @cindex x86-64 Options
7210 @cindex Intel 386 Options
7211 @cindex AMD x86-64 Options
7213 These @samp{-m} options are defined for the i386 and x86-64 family of
7217 @item -mcpu=@var{cpu-type}
7219 Tune to @var{cpu-type} everything applicable about the generated code, except
7220 for the ABI and the set of available instructions. The choices for
7221 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7222 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7223 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7224 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp}
7225 and @samp{athlon-mp}.
7227 While picking a specific @var{cpu-type} will schedule things appropriately
7228 for that particular chip, the compiler will not generate any code that
7229 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7230 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7231 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7232 AMD chips as opposed to the Intel ones.
7234 @item -march=@var{cpu-type}
7236 Generate instructions for the machine type @var{cpu-type}. The choices
7237 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7238 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7247 @opindex mpentiumpro
7248 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7249 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7250 These synonyms are deprecated.
7252 @item -mfpmath=@var{unit}
7254 generate floating point arithmetics for selected unit @var{unit}. the choices
7259 Use the standard 387 floating point coprocessor present majority of chips and
7260 emulated otherwise. Code compiled with this option will run almost everywhere.
7261 The temporary results are computed in 80bit precesion instead of precision
7262 specified by the type resulting in slightly different results compared to most
7263 of other chips. See @option{-ffloat-store} for more detailed description.
7265 This is the default choice for i386 compiler.
7268 Use scalar floating point instructions present in the SSE instruction set.
7269 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7270 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7271 instruction set supports only single precision arithmetics, thus the double and
7272 extended precision arithmetics is still done using 387. Later version, present
7273 only in Pentium4 and the future AMD x86-64 chips supports double precision
7276 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7277 @option{-msse2} switches to enable SSE extensions and make this option
7278 effective. For x86-64 compiler, these extensions are enabled by default.
7280 The resulting code should be considerably faster in majority of cases and avoid
7281 the numerical instability problems of 387 code, but may break some existing
7282 code that expects temporaries to be 80bit.
7284 This is the default choice for x86-64 compiler.
7287 Attempt to utilize both instruction sets at once. This effectivly double the
7288 amount of available registers and on chips with separate execution units for
7289 387 and SSE the execution resources too. Use this option with care, as it is
7290 still experimental, because gcc register allocator does not model separate
7291 functional units well resulting in instable performance.
7294 @item -masm=@var{dialect}
7295 @opindex masm=@var{dialect}
7296 Output asm instructions using selected @var{dialect}. Supported choices are
7297 @samp{intel} or @samp{att} (the default one).
7302 @opindex mno-ieee-fp
7303 Control whether or not the compiler uses IEEE floating point
7304 comparisons. These handle correctly the case where the result of a
7305 comparison is unordered.
7308 @opindex msoft-float
7309 Generate output containing library calls for floating point.
7310 @strong{Warning:} the requisite libraries are not part of GCC@.
7311 Normally the facilities of the machine's usual C compiler are used, but
7312 this can't be done directly in cross-compilation. You must make your
7313 own arrangements to provide suitable library functions for
7316 On machines where a function returns floating point results in the 80387
7317 register stack, some floating point opcodes may be emitted even if
7318 @option{-msoft-float} is used.
7320 @item -mno-fp-ret-in-387
7321 @opindex mno-fp-ret-in-387
7322 Do not use the FPU registers for return values of functions.
7324 The usual calling convention has functions return values of types
7325 @code{float} and @code{double} in an FPU register, even if there
7326 is no FPU@. The idea is that the operating system should emulate
7329 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7330 in ordinary CPU registers instead.
7332 @item -mno-fancy-math-387
7333 @opindex mno-fancy-math-387
7334 Some 387 emulators do not support the @code{sin}, @code{cos} and
7335 @code{sqrt} instructions for the 387. Specify this option to avoid
7336 generating those instructions. This option is the default on FreeBSD@.
7337 As of revision 2.6.1, these instructions are not generated unless you
7338 also use the @option{-funsafe-math-optimizations} switch.
7340 @item -malign-double
7341 @itemx -mno-align-double
7342 @opindex malign-double
7343 @opindex mno-align-double
7344 Control whether GCC aligns @code{double}, @code{long double}, and
7345 @code{long long} variables on a two word boundary or a one word
7346 boundary. Aligning @code{double} variables on a two word boundary will
7347 produce code that runs somewhat faster on a @samp{Pentium} at the
7348 expense of more memory.
7350 @item -m128bit-long-double
7351 @opindex m128bit-long-double
7352 Control the size of @code{long double} type. i386 application binary interface
7353 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7354 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7355 impossible to reach with 12 byte long doubles in the array accesses.
7357 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7358 structures and arrays containing @code{long double} will change their size as
7359 well as function calling convention for function taking @code{long double}
7362 @item -m96bit-long-double
7363 @opindex m96bit-long-double
7364 Set the size of @code{long double} to 96 bits as required by the i386
7365 application binary interface. This is the default.
7368 @itemx -mno-svr3-shlib
7369 @opindex msvr3-shlib
7370 @opindex mno-svr3-shlib
7371 Control whether GCC places uninitialized local variables into the
7372 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7373 into @code{bss}. These options are meaningful only on System V Release 3.
7377 Use a different function-calling convention, in which functions that
7378 take a fixed number of arguments return with the @code{ret} @var{num}
7379 instruction, which pops their arguments while returning. This saves one
7380 instruction in the caller since there is no need to pop the arguments
7383 You can specify that an individual function is called with this calling
7384 sequence with the function attribute @samp{stdcall}. You can also
7385 override the @option{-mrtd} option by using the function attribute
7386 @samp{cdecl}. @xref{Function Attributes}.
7388 @strong{Warning:} this calling convention is incompatible with the one
7389 normally used on Unix, so you cannot use it if you need to call
7390 libraries compiled with the Unix compiler.
7392 Also, you must provide function prototypes for all functions that
7393 take variable numbers of arguments (including @code{printf});
7394 otherwise incorrect code will be generated for calls to those
7397 In addition, seriously incorrect code will result if you call a
7398 function with too many arguments. (Normally, extra arguments are
7399 harmlessly ignored.)
7401 @item -mregparm=@var{num}
7403 Control how many registers are used to pass integer arguments. By
7404 default, no registers are used to pass arguments, and at most 3
7405 registers can be used. You can control this behavior for a specific
7406 function by using the function attribute @samp{regparm}.
7407 @xref{Function Attributes}.
7409 @strong{Warning:} if you use this switch, and
7410 @var{num} is nonzero, then you must build all modules with the same
7411 value, including any libraries. This includes the system libraries and
7414 @item -mpreferred-stack-boundary=@var{num}
7415 @opindex mpreferred-stack-boundary
7416 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7417 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7418 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7419 size (@option{-Os}), in which case the default is the minimum correct
7420 alignment (4 bytes for x86, and 8 bytes for x86-64).
7422 On Pentium and PentiumPro, @code{double} and @code{long double} values
7423 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7424 suffer significant run time performance penalties. On Pentium III, the
7425 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7426 penalties if it is not 16 byte aligned.
7428 To ensure proper alignment of this values on the stack, the stack boundary
7429 must be as aligned as that required by any value stored on the stack.
7430 Further, every function must be generated such that it keeps the stack
7431 aligned. Thus calling a function compiled with a higher preferred
7432 stack boundary from a function compiled with a lower preferred stack
7433 boundary will most likely misalign the stack. It is recommended that
7434 libraries that use callbacks always use the default setting.
7436 This extra alignment does consume extra stack space, and generally
7437 increases code size. Code that is sensitive to stack space usage, such
7438 as embedded systems and operating system kernels, may want to reduce the
7439 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7455 These switches enable or disable the use of built-in functions that allow
7456 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7458 @xref{X86 Built-in Functions}, for details of the functions enabled
7459 and disabled by these switches.
7462 @itemx -mno-push-args
7464 @opindex mno-push-args
7465 Use PUSH operations to store outgoing parameters. This method is shorter
7466 and usually equally fast as method using SUB/MOV operations and is enabled
7467 by default. In some cases disabling it may improve performance because of
7468 improved scheduling and reduced dependencies.
7470 @item -maccumulate-outgoing-args
7471 @opindex maccumulate-outgoing-args
7472 If enabled, the maximum amount of space required for outgoing arguments will be
7473 computed in the function prologue. This is faster on most modern CPUs
7474 because of reduced dependencies, improved scheduling and reduced stack usage
7475 when preferred stack boundary is not equal to 2. The drawback is a notable
7476 increase in code size. This switch implies @option{-mno-push-args}.
7480 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7481 on thread-safe exception handling must compile and link all code with the
7482 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7483 @option{-D_MT}; when linking, it links in a special thread helper library
7484 @option{-lmingwthrd} which cleans up per thread exception handling data.
7486 @item -mno-align-stringops
7487 @opindex mno-align-stringops
7488 Do not align destination of inlined string operations. This switch reduces
7489 code size and improves performance in case the destination is already aligned,
7490 but gcc don't know about it.
7492 @item -minline-all-stringops
7493 @opindex minline-all-stringops
7494 By default GCC inlines string operations only when destination is known to be
7495 aligned at least to 4 byte boundary. This enables more inlining, increase code
7496 size, but may improve performance of code that depends on fast memcpy, strlen
7497 and memset for short lengths.
7499 @item -momit-leaf-frame-pointer
7500 @opindex momit-leaf-frame-pointer
7501 Don't keep the frame pointer in a register for leaf functions. This
7502 avoids the instructions to save, set up and restore frame pointers and
7503 makes an extra register available in leaf functions. The option
7504 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7505 which might make debugging harder.
7508 These @samp{-m} switches are supported in addition to the above
7509 on AMD x86-64 processors in 64-bit environments.
7516 Generate code for a 32-bit or 64-bit environment.
7517 The 32-bit environment sets int, long and pointer to 32 bits and
7518 generates code that runs on any i386 system.
7519 The 64-bit environment sets int to 32 bits and long and pointer
7520 to 64 bits and generates code for AMD's x86-64 architecture.
7523 @opindex no-red-zone
7524 Do not use a so called red zone for x86-64 code. The red zone is mandated
7525 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7526 stack pointer that will not be modified by signal or interrupt handlers
7527 and therefore can be used for temporary data without adjusting the stack
7528 pointer. The flag @option{-mno-red-zone} disables this red zone.
7532 @subsection HPPA Options
7533 @cindex HPPA Options
7535 These @samp{-m} options are defined for the HPPA family of computers:
7538 @item -march=@var{architecture-type}
7540 Generate code for the specified architecture. The choices for
7541 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7542 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7543 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7544 architecture option for your machine. Code compiled for lower numbered
7545 architectures will run on higher numbered architectures, but not the
7548 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7549 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7553 @itemx -mpa-risc-1-1
7554 @itemx -mpa-risc-2-0
7555 @opindex mpa-risc-1-0
7556 @opindex mpa-risc-1-1
7557 @opindex mpa-risc-2-0
7558 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7561 @opindex mbig-switch
7562 Generate code suitable for big switch tables. Use this option only if
7563 the assembler/linker complain about out of range branches within a switch
7566 @item -mjump-in-delay
7567 @opindex mjump-in-delay
7568 Fill delay slots of function calls with unconditional jump instructions
7569 by modifying the return pointer for the function call to be the target
7570 of the conditional jump.
7572 @item -mdisable-fpregs
7573 @opindex mdisable-fpregs
7574 Prevent floating point registers from being used in any manner. This is
7575 necessary for compiling kernels which perform lazy context switching of
7576 floating point registers. If you use this option and attempt to perform
7577 floating point operations, the compiler will abort.
7579 @item -mdisable-indexing
7580 @opindex mdisable-indexing
7581 Prevent the compiler from using indexing address modes. This avoids some
7582 rather obscure problems when compiling MIG generated code under MACH@.
7584 @item -mno-space-regs
7585 @opindex mno-space-regs
7586 Generate code that assumes the target has no space registers. This allows
7587 GCC to generate faster indirect calls and use unscaled index address modes.
7589 Such code is suitable for level 0 PA systems and kernels.
7591 @item -mfast-indirect-calls
7592 @opindex mfast-indirect-calls
7593 Generate code that assumes calls never cross space boundaries. This
7594 allows GCC to emit code which performs faster indirect calls.
7596 This option will not work in the presence of shared libraries or nested
7599 @item -mlong-load-store
7600 @opindex mlong-load-store
7601 Generate 3-instruction load and store sequences as sometimes required by
7602 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7605 @item -mportable-runtime
7606 @opindex mportable-runtime
7607 Use the portable calling conventions proposed by HP for ELF systems.
7611 Enable the use of assembler directives only GAS understands.
7613 @item -mschedule=@var{cpu-type}
7615 Schedule code according to the constraints for the machine type
7616 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7617 @samp{7100}, @samp{7100LC}, @samp{7200}, and @samp{8000}. Refer to
7618 @file{/usr/lib/sched.models} on an HP-UX system to determine the
7619 proper scheduling option for your machine.
7622 @opindex mlinker-opt
7623 Enable the optimization pass in the HPUX linker. Note this makes symbolic
7624 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
7625 in which they give bogus error messages when linking some programs.
7628 @opindex msoft-float
7629 Generate output containing library calls for floating point.
7630 @strong{Warning:} the requisite libraries are not available for all HPPA
7631 targets. Normally the facilities of the machine's usual C compiler are
7632 used, but this cannot be done directly in cross-compilation. You must make
7633 your own arrangements to provide suitable library functions for
7634 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7635 does provide software floating point support.
7637 @option{-msoft-float} changes the calling convention in the output file;
7638 therefore, it is only useful if you compile @emph{all} of a program with
7639 this option. In particular, you need to compile @file{libgcc.a}, the
7640 library that comes with GCC, with @option{-msoft-float} in order for
7644 @node Intel 960 Options
7645 @subsection Intel 960 Options
7647 These @samp{-m} options are defined for the Intel 960 implementations:
7650 @item -m@var{cpu-type}
7658 Assume the defaults for the machine type @var{cpu-type} for some of
7659 the other options, including instruction scheduling, floating point
7660 support, and addressing modes. The choices for @var{cpu-type} are
7661 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
7662 @samp{sa}, and @samp{sb}.
7669 @opindex msoft-float
7670 The @option{-mnumerics} option indicates that the processor does support
7671 floating-point instructions. The @option{-msoft-float} option indicates
7672 that floating-point support should not be assumed.
7674 @item -mleaf-procedures
7675 @itemx -mno-leaf-procedures
7676 @opindex mleaf-procedures
7677 @opindex mno-leaf-procedures
7678 Do (or do not) attempt to alter leaf procedures to be callable with the
7679 @code{bal} instruction as well as @code{call}. This will result in more
7680 efficient code for explicit calls when the @code{bal} instruction can be
7681 substituted by the assembler or linker, but less efficient code in other
7682 cases, such as calls via function pointers, or using a linker that doesn't
7683 support this optimization.
7686 @itemx -mno-tail-call
7688 @opindex mno-tail-call
7689 Do (or do not) make additional attempts (beyond those of the
7690 machine-independent portions of the compiler) to optimize tail-recursive
7691 calls into branches. You may not want to do this because the detection of
7692 cases where this is not valid is not totally complete. The default is
7693 @option{-mno-tail-call}.
7695 @item -mcomplex-addr
7696 @itemx -mno-complex-addr
7697 @opindex mcomplex-addr
7698 @opindex mno-complex-addr
7699 Assume (or do not assume) that the use of a complex addressing mode is a
7700 win on this implementation of the i960. Complex addressing modes may not
7701 be worthwhile on the K-series, but they definitely are on the C-series.
7702 The default is currently @option{-mcomplex-addr} for all processors except
7706 @itemx -mno-code-align
7707 @opindex mcode-align
7708 @opindex mno-code-align
7709 Align code to 8-byte boundaries for faster fetching (or don't bother).
7710 Currently turned on by default for C-series implementations only.
7713 @item -mclean-linkage
7714 @itemx -mno-clean-linkage
7715 @opindex mclean-linkage
7716 @opindex mno-clean-linkage
7717 These options are not fully implemented.
7721 @itemx -mic2.0-compat
7722 @itemx -mic3.0-compat
7724 @opindex mic2.0-compat
7725 @opindex mic3.0-compat
7726 Enable compatibility with iC960 v2.0 or v3.0.
7730 @opindex masm-compat
7732 Enable compatibility with the iC960 assembler.
7734 @item -mstrict-align
7735 @itemx -mno-strict-align
7736 @opindex mstrict-align
7737 @opindex mno-strict-align
7738 Do not permit (do permit) unaligned accesses.
7742 Enable structure-alignment compatibility with Intel's gcc release version
7743 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
7745 @item -mlong-double-64
7746 @opindex mlong-double-64
7747 Implement type @samp{long double} as 64-bit floating point numbers.
7748 Without the option @samp{long double} is implemented by 80-bit
7749 floating point numbers. The only reason we have it because there is
7750 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
7751 is only useful for people using soft-float targets. Otherwise, we
7752 should recommend against use of it.
7756 @node DEC Alpha Options
7757 @subsection DEC Alpha Options
7759 These @samp{-m} options are defined for the DEC Alpha implementations:
7762 @item -mno-soft-float
7764 @opindex mno-soft-float
7765 @opindex msoft-float
7766 Use (do not use) the hardware floating-point instructions for
7767 floating-point operations. When @option{-msoft-float} is specified,
7768 functions in @file{libgcc.a} will be used to perform floating-point
7769 operations. Unless they are replaced by routines that emulate the
7770 floating-point operations, or compiled in such a way as to call such
7771 emulations routines, these routines will issue floating-point
7772 operations. If you are compiling for an Alpha without floating-point
7773 operations, you must ensure that the library is built so as not to call
7776 Note that Alpha implementations without floating-point operations are
7777 required to have floating-point registers.
7782 @opindex mno-fp-regs
7783 Generate code that uses (does not use) the floating-point register set.
7784 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
7785 register set is not used, floating point operands are passed in integer
7786 registers as if they were integers and floating-point results are passed
7787 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
7788 so any function with a floating-point argument or return value called by code
7789 compiled with @option{-mno-fp-regs} must also be compiled with that
7792 A typical use of this option is building a kernel that does not use,
7793 and hence need not save and restore, any floating-point registers.
7797 The Alpha architecture implements floating-point hardware optimized for
7798 maximum performance. It is mostly compliant with the IEEE floating
7799 point standard. However, for full compliance, software assistance is
7800 required. This option generates code fully IEEE compliant code
7801 @emph{except} that the @var{inexact-flag} is not maintained (see below).
7802 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
7803 defined during compilation. The resulting code is less efficient but is
7804 able to correctly support denormalized numbers and exceptional IEEE
7805 values such as not-a-number and plus/minus infinity. Other Alpha
7806 compilers call this option @option{-ieee_with_no_inexact}.
7808 @item -mieee-with-inexact
7809 @opindex mieee-with-inexact
7810 This is like @option{-mieee} except the generated code also maintains
7811 the IEEE @var{inexact-flag}. Turning on this option causes the
7812 generated code to implement fully-compliant IEEE math. In addition to
7813 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
7814 macro. On some Alpha implementations the resulting code may execute
7815 significantly slower than the code generated by default. Since there is
7816 very little code that depends on the @var{inexact-flag}, you should
7817 normally not specify this option. Other Alpha compilers call this
7818 option @option{-ieee_with_inexact}.
7820 @item -mfp-trap-mode=@var{trap-mode}
7821 @opindex mfp-trap-mode
7822 This option controls what floating-point related traps are enabled.
7823 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
7824 The trap mode can be set to one of four values:
7828 This is the default (normal) setting. The only traps that are enabled
7829 are the ones that cannot be disabled in software (e.g., division by zero
7833 In addition to the traps enabled by @samp{n}, underflow traps are enabled
7837 Like @samp{su}, but the instructions are marked to be safe for software
7838 completion (see Alpha architecture manual for details).
7841 Like @samp{su}, but inexact traps are enabled as well.
7844 @item -mfp-rounding-mode=@var{rounding-mode}
7845 @opindex mfp-rounding-mode
7846 Selects the IEEE rounding mode. Other Alpha compilers call this option
7847 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
7852 Normal IEEE rounding mode. Floating point numbers are rounded towards
7853 the nearest machine number or towards the even machine number in case
7857 Round towards minus infinity.
7860 Chopped rounding mode. Floating point numbers are rounded towards zero.
7863 Dynamic rounding mode. A field in the floating point control register
7864 (@var{fpcr}, see Alpha architecture reference manual) controls the
7865 rounding mode in effect. The C library initializes this register for
7866 rounding towards plus infinity. Thus, unless your program modifies the
7867 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
7870 @item -mtrap-precision=@var{trap-precision}
7871 @opindex mtrap-precision
7872 In the Alpha architecture, floating point traps are imprecise. This
7873 means without software assistance it is impossible to recover from a
7874 floating trap and program execution normally needs to be terminated.
7875 GCC can generate code that can assist operating system trap handlers
7876 in determining the exact location that caused a floating point trap.
7877 Depending on the requirements of an application, different levels of
7878 precisions can be selected:
7882 Program precision. This option is the default and means a trap handler
7883 can only identify which program caused a floating point exception.
7886 Function precision. The trap handler can determine the function that
7887 caused a floating point exception.
7890 Instruction precision. The trap handler can determine the exact
7891 instruction that caused a floating point exception.
7894 Other Alpha compilers provide the equivalent options called
7895 @option{-scope_safe} and @option{-resumption_safe}.
7897 @item -mieee-conformant
7898 @opindex mieee-conformant
7899 This option marks the generated code as IEEE conformant. You must not
7900 use this option unless you also specify @option{-mtrap-precision=i} and either
7901 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
7902 is to emit the line @samp{.eflag 48} in the function prologue of the
7903 generated assembly file. Under DEC Unix, this has the effect that
7904 IEEE-conformant math library routines will be linked in.
7906 @item -mbuild-constants
7907 @opindex mbuild-constants
7908 Normally GCC examines a 32- or 64-bit integer constant to
7909 see if it can construct it from smaller constants in two or three
7910 instructions. If it cannot, it will output the constant as a literal and
7911 generate code to load it from the data segment at runtime.
7913 Use this option to require GCC to construct @emph{all} integer constants
7914 using code, even if it takes more instructions (the maximum is six).
7916 You would typically use this option to build a shared library dynamic
7917 loader. Itself a shared library, it must relocate itself in memory
7918 before it can find the variables and constants in its own data segment.
7924 Select whether to generate code to be assembled by the vendor-supplied
7925 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
7943 Indicate whether GCC should generate code to use the optional BWX,
7944 CIX, FIX and MAX instruction sets. The default is to use the instruction
7945 sets supported by the CPU type specified via @option{-mcpu=} option or that
7946 of the CPU on which GCC was built if none was specified.
7951 @opindex mfloat-ieee
7952 Generate code that uses (does not use) VAX F and G floating point
7953 arithmetic instead of IEEE single and double precision.
7955 @item -mexplicit-relocs
7956 @itemx -mno-explicit-relocs
7957 @opindex mexplicit-relocs
7958 @opindex mno-explicit-relocs
7959 Older Alpha assemblers provided no way to generate symbol relocations
7960 except via assembler macros. Use of these macros does not allow
7961 optimial instruction scheduling. GNU binutils as of version 2.12
7962 supports a new syntax that allows the compiler to explicitly mark
7963 which relocations should apply to which instructions. This option
7964 is mostly useful for debugging, as GCC detects the capabilities of
7965 the assembler when it is built and sets the default accordingly.
7969 @opindex msmall-data
7970 @opindex mlarge-data
7971 When @option{-mexplicit-relocs} is in effect, static data is
7972 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
7973 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
7974 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
7975 16-bit relocations off of the @code{$gp} register. This limits the
7976 size of the small data area to 64KB, but allows the variables to be
7977 directly accessed via a single instruction.
7979 The default is @option{-mlarge-data}. With this option the data area
7980 is limited to just below 2GB. Programs that require more than 2GB of
7981 data must use @code{malloc} or @code{mmap} to allocate the data in the
7982 heap instead of in the program's data segment.
7984 When generating code for shared libraries, @option{-fpic} implies
7985 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
7987 @item -mcpu=@var{cpu_type}
7989 Set the instruction set and instruction scheduling parameters for
7990 machine type @var{cpu_type}. You can specify either the @samp{EV}
7991 style name or the corresponding chip number. GCC supports scheduling
7992 parameters for the EV4, EV5 and EV6 family of processors and will
7993 choose the default values for the instruction set from the processor
7994 you specify. If you do not specify a processor type, GCC will default
7995 to the processor on which the compiler was built.
7997 Supported values for @var{cpu_type} are
8003 Schedules as an EV4 and has no instruction set extensions.
8007 Schedules as an EV5 and has no instruction set extensions.
8011 Schedules as an EV5 and supports the BWX extension.
8016 Schedules as an EV5 and supports the BWX and MAX extensions.
8020 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8024 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8027 @item -mtune=@var{cpu_type}
8029 Set only the instruction scheduling parameters for machine type
8030 @var{cpu_type}. The instruction set is not changed.
8032 @item -mmemory-latency=@var{time}
8033 @opindex mmemory-latency
8034 Sets the latency the scheduler should assume for typical memory
8035 references as seen by the application. This number is highly
8036 dependent on the memory access patterns used by the application
8037 and the size of the external cache on the machine.
8039 Valid options for @var{time} are
8043 A decimal number representing clock cycles.
8049 The compiler contains estimates of the number of clock cycles for
8050 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8051 (also called Dcache, Scache, and Bcache), as well as to main memory.
8052 Note that L3 is only valid for EV5.
8057 @node DEC Alpha/VMS Options
8058 @subsection DEC Alpha/VMS Options
8060 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8063 @item -mvms-return-codes
8064 @opindex mvms-return-codes
8065 Return VMS condition codes from main. The default is to return POSIX
8066 style condition (e.g.@ error) codes.
8069 @node Clipper Options
8070 @subsection Clipper Options
8072 These @samp{-m} options are defined for the Clipper implementations:
8077 Produce code for a C300 Clipper processor. This is the default.
8081 Produce code for a C400 Clipper processor, i.e.@: use floating point
8085 @node H8/300 Options
8086 @subsection H8/300 Options
8088 These @samp{-m} options are defined for the H8/300 implementations:
8093 Shorten some address references at link time, when possible; uses the
8094 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8095 ld.info, Using ld}, for a fuller description.
8099 Generate code for the H8/300H@.
8103 Generate code for the H8/S@.
8107 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8111 Make @code{int} data 32 bits by default.
8115 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8116 The default for the H8/300H and H8/S is to align longs and floats on 4
8118 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8119 This option has no effect on the H8/300.
8123 @subsection SH Options
8125 These @samp{-m} options are defined for the SH implementations:
8130 Generate code for the SH1.
8134 Generate code for the SH2.
8138 Generate code for the SH3.
8142 Generate code for the SH3e.
8146 Generate code for the SH4 without a floating-point unit.
8148 @item -m4-single-only
8149 @opindex m4-single-only
8150 Generate code for the SH4 with a floating-point unit that only
8151 supports single-precision arithmetic.
8155 Generate code for the SH4 assuming the floating-point unit is in
8156 single-precision mode by default.
8160 Generate code for the SH4.
8164 Compile code for the processor in big endian mode.
8168 Compile code for the processor in little endian mode.
8172 Align doubles at 64-bit boundaries. Note that this changes the calling
8173 conventions, and thus some functions from the standard C library will
8174 not work unless you recompile it first with @option{-mdalign}.
8178 Shorten some address references at link time, when possible; uses the
8179 linker option @option{-relax}.
8183 Use 32-bit offsets in @code{switch} tables. The default is to use
8188 Enable the use of the instruction @code{fmovd}.
8192 Comply with the calling conventions defined by Hitachi.
8196 Mark the @code{MAC} register as call-clobbered, even if
8197 @option{-mhitachi} is given.
8201 Increase IEEE-compliance of floating-point code.
8205 Dump instruction size and location in the assembly code.
8209 This option is deprecated. It pads structures to multiple of 4 bytes,
8210 which is incompatible with the SH ABI@.
8214 Optimize for space instead of speed. Implied by @option{-Os}.
8218 When generating position-independent code, emit function calls using
8219 the Global Offset Table instead of the Procedure Linkage Table.
8223 Generate a library function call to invalidate instruction cache
8224 entries, after fixing up a trampoline. This library function call
8225 doesn't assume it can write to the whole memory address space. This
8226 is the default when the target is @code{sh-*-linux*}.
8229 @node System V Options
8230 @subsection Options for System V
8232 These additional options are available on System V Release 4 for
8233 compatibility with other compilers on those systems:
8238 Create a shared object.
8239 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8243 Identify the versions of each tool used by the compiler, in a
8244 @code{.ident} assembler directive in the output.
8248 Refrain from adding @code{.ident} directives to the output file (this is
8251 @item -YP,@var{dirs}
8253 Search the directories @var{dirs}, and no others, for libraries
8254 specified with @option{-l}.
8258 Look in the directory @var{dir} to find the M4 preprocessor.
8259 The assembler uses this option.
8260 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8261 @c the generic assembler that comes with Solaris takes just -Ym.
8264 @node TMS320C3x/C4x Options
8265 @subsection TMS320C3x/C4x Options
8266 @cindex TMS320C3x/C4x Options
8268 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8272 @item -mcpu=@var{cpu_type}
8274 Set the instruction set, register set, and instruction scheduling
8275 parameters for machine type @var{cpu_type}. Supported values for
8276 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8277 @samp{c44}. The default is @samp{c40} to generate code for the
8282 @itemx -msmall-memory
8284 @opindex mbig-memory
8286 @opindex msmall-memory
8288 Generates code for the big or small memory model. The small memory
8289 model assumed that all data fits into one 64K word page. At run-time
8290 the data page (DP) register must be set to point to the 64K page
8291 containing the .bss and .data program sections. The big memory model is
8292 the default and requires reloading of the DP register for every direct
8299 Allow (disallow) allocation of general integer operands into the block
8306 Enable (disable) generation of code using decrement and branch,
8307 DBcond(D), instructions. This is enabled by default for the C4x. To be
8308 on the safe side, this is disabled for the C3x, since the maximum
8309 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
8310 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
8311 that it can utilise the decrement and branch instruction, but will give
8312 up if there is more than one memory reference in the loop. Thus a loop
8313 where the loop counter is decremented can generate slightly more
8314 efficient code, in cases where the RPTB instruction cannot be utilised.
8316 @item -mdp-isr-reload
8318 @opindex mdp-isr-reload
8320 Force the DP register to be saved on entry to an interrupt service
8321 routine (ISR), reloaded to point to the data section, and restored on
8322 exit from the ISR@. This should not be required unless someone has
8323 violated the small memory model by modifying the DP register, say within
8330 For the C3x use the 24-bit MPYI instruction for integer multiplies
8331 instead of a library call to guarantee 32-bit results. Note that if one
8332 of the operands is a constant, then the multiplication will be performed
8333 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8334 then squaring operations are performed inline instead of a library call.
8337 @itemx -mno-fast-fix
8339 @opindex mno-fast-fix
8340 The C3x/C4x FIX instruction to convert a floating point value to an
8341 integer value chooses the nearest integer less than or equal to the
8342 floating point value rather than to the nearest integer. Thus if the
8343 floating point number is negative, the result will be incorrectly
8344 truncated an additional code is necessary to detect and correct this
8345 case. This option can be used to disable generation of the additional
8346 code required to correct the result.
8352 Enable (disable) generation of repeat block sequences using the RPTB
8353 instruction for zero overhead looping. The RPTB construct is only used
8354 for innermost loops that do not call functions or jump across the loop
8355 boundaries. There is no advantage having nested RPTB loops due to the
8356 overhead required to save and restore the RC, RS, and RE registers.
8357 This is enabled by default with @option{-O2}.
8359 @item -mrpts=@var{count}
8363 Enable (disable) the use of the single instruction repeat instruction
8364 RPTS@. If a repeat block contains a single instruction, and the loop
8365 count can be guaranteed to be less than the value @var{count}, GCC will
8366 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8367 then a RPTS will be emitted even if the loop count cannot be determined
8368 at compile time. Note that the repeated instruction following RPTS does
8369 not have to be reloaded from memory each iteration, thus freeing up the
8370 CPU buses for operands. However, since interrupts are blocked by this
8371 instruction, it is disabled by default.
8373 @item -mloop-unsigned
8374 @itemx -mno-loop-unsigned
8375 @opindex mloop-unsigned
8376 @opindex mno-loop-unsigned
8377 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8378 is @math{2^{31} + 1} since these instructions test if the iteration count is
8379 negative to terminate the loop. If the iteration count is unsigned
8380 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
8381 exceeded. This switch allows an unsigned iteration count.
8385 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8386 with. This also enforces compatibility with the API employed by the TI
8387 C3x C compiler. For example, long doubles are passed as structures
8388 rather than in floating point registers.
8394 Generate code that uses registers (stack) for passing arguments to functions.
8395 By default, arguments are passed in registers where possible rather
8396 than by pushing arguments on to the stack.
8398 @item -mparallel-insns
8399 @itemx -mno-parallel-insns
8400 @opindex mparallel-insns
8401 @opindex mno-parallel-insns
8402 Allow the generation of parallel instructions. This is enabled by
8403 default with @option{-O2}.
8405 @item -mparallel-mpy
8406 @itemx -mno-parallel-mpy
8407 @opindex mparallel-mpy
8408 @opindex mno-parallel-mpy
8409 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8410 provided @option{-mparallel-insns} is also specified. These instructions have
8411 tight register constraints which can pessimize the code generation
8417 @subsection V850 Options
8418 @cindex V850 Options
8420 These @samp{-m} options are defined for V850 implementations:
8424 @itemx -mno-long-calls
8425 @opindex mlong-calls
8426 @opindex mno-long-calls
8427 Treat all calls as being far away (near). If calls are assumed to be
8428 far away, the compiler will always load the functions address up into a
8429 register, and call indirect through the pointer.
8435 Do not optimize (do optimize) basic blocks that use the same index
8436 pointer 4 or more times to copy pointer into the @code{ep} register, and
8437 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8438 option is on by default if you optimize.
8440 @item -mno-prolog-function
8441 @itemx -mprolog-function
8442 @opindex mno-prolog-function
8443 @opindex mprolog-function
8444 Do not use (do use) external functions to save and restore registers at
8445 the prolog and epilog of a function. The external functions are slower,
8446 but use less code space if more than one function saves the same number
8447 of registers. The @option{-mprolog-function} option is on by default if
8452 Try to make the code as small as possible. At present, this just turns
8453 on the @option{-mep} and @option{-mprolog-function} options.
8457 Put static or global variables whose size is @var{n} bytes or less into
8458 the tiny data area that register @code{ep} points to. The tiny data
8459 area can hold up to 256 bytes in total (128 bytes for byte references).
8463 Put static or global variables whose size is @var{n} bytes or less into
8464 the small data area that register @code{gp} points to. The small data
8465 area can hold up to 64 kilobytes.
8469 Put static or global variables whose size is @var{n} bytes or less into
8470 the first 32 kilobytes of memory.
8474 Specify that the target processor is the V850.
8477 @opindex mbig-switch
8478 Generate code suitable for big switch tables. Use this option only if
8479 the assembler/linker complain about out of range branches within a switch
8484 @subsection ARC Options
8487 These options are defined for ARC implementations:
8492 Compile code for little endian mode. This is the default.
8496 Compile code for big endian mode.
8499 @opindex mmangle-cpu
8500 Prepend the name of the cpu to all public symbol names.
8501 In multiple-processor systems, there are many ARC variants with different
8502 instruction and register set characteristics. This flag prevents code
8503 compiled for one cpu to be linked with code compiled for another.
8504 No facility exists for handling variants that are ``almost identical''.
8505 This is an all or nothing option.
8507 @item -mcpu=@var{cpu}
8509 Compile code for ARC variant @var{cpu}.
8510 Which variants are supported depend on the configuration.
8511 All variants support @option{-mcpu=base}, this is the default.
8513 @item -mtext=@var{text-section}
8514 @itemx -mdata=@var{data-section}
8515 @itemx -mrodata=@var{readonly-data-section}
8519 Put functions, data, and readonly data in @var{text-section},
8520 @var{data-section}, and @var{readonly-data-section} respectively
8521 by default. This can be overridden with the @code{section} attribute.
8522 @xref{Variable Attributes}.
8527 @subsection NS32K Options
8528 @cindex NS32K options
8530 These are the @samp{-m} options defined for the 32000 series. The default
8531 values for these options depends on which style of 32000 was selected when
8532 the compiler was configured; the defaults for the most common choices are
8540 Generate output for a 32032. This is the default
8541 when the compiler is configured for 32032 and 32016 based systems.
8547 Generate output for a 32332. This is the default
8548 when the compiler is configured for 32332-based systems.
8554 Generate output for a 32532. This is the default
8555 when the compiler is configured for 32532-based systems.
8559 Generate output containing 32081 instructions for floating point.
8560 This is the default for all systems.
8564 Generate output containing 32381 instructions for floating point. This
8565 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8566 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8570 Try and generate multiply-add floating point instructions @code{polyF}
8571 and @code{dotF}. This option is only available if the @option{-m32381}
8572 option is in effect. Using these instructions requires changes to
8573 register allocation which generally has a negative impact on
8574 performance. This option should only be enabled when compiling code
8575 particularly likely to make heavy use of multiply-add instructions.
8578 @opindex mnomulti-add
8579 Do not try and generate multiply-add floating point instructions
8580 @code{polyF} and @code{dotF}. This is the default on all platforms.
8583 @opindex msoft-float
8584 Generate output containing library calls for floating point.
8585 @strong{Warning:} the requisite libraries may not be available.
8588 @opindex mnobitfield
8589 Do not use the bit-field instructions. On some machines it is faster to
8590 use shifting and masking operations. This is the default for the pc532.
8594 Do use the bit-field instructions. This is the default for all platforms
8599 Use a different function-calling convention, in which functions
8600 that take a fixed number of arguments return pop their
8601 arguments on return with the @code{ret} instruction.
8603 This calling convention is incompatible with the one normally
8604 used on Unix, so you cannot use it if you need to call libraries
8605 compiled with the Unix compiler.
8607 Also, you must provide function prototypes for all functions that
8608 take variable numbers of arguments (including @code{printf});
8609 otherwise incorrect code will be generated for calls to those
8612 In addition, seriously incorrect code will result if you call a
8613 function with too many arguments. (Normally, extra arguments are
8614 harmlessly ignored.)
8616 This option takes its name from the 680x0 @code{rtd} instruction.
8621 Use a different function-calling convention where the first two arguments
8622 are passed in registers.
8624 This calling convention is incompatible with the one normally
8625 used on Unix, so you cannot use it if you need to call libraries
8626 compiled with the Unix compiler.
8629 @opindex mnoregparam
8630 Do not pass any arguments in registers. This is the default for all
8635 It is OK to use the sb as an index register which is always loaded with
8636 zero. This is the default for the pc532-netbsd target.
8640 The sb register is not available for use or has not been initialized to
8641 zero by the run time system. This is the default for all targets except
8642 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
8643 @option{-fpic} is set.
8647 Many ns32000 series addressing modes use displacements of up to 512MB@.
8648 If an address is above 512MB then displacements from zero can not be used.
8649 This option causes code to be generated which can be loaded above 512MB@.
8650 This may be useful for operating systems or ROM code.
8654 Assume code will be loaded in the first 512MB of virtual address space.
8655 This is the default for all platforms.
8661 @subsection AVR Options
8664 These options are defined for AVR implementations:
8667 @item -mmcu=@var{mcu}
8669 Specify ATMEL AVR instruction set or MCU type.
8671 Instruction set avr1 is for the minimal AVR core, not supported by the C
8672 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8673 attiny11, attiny12, attiny15, attiny28).
8675 Instruction set avr2 (default) is for the classic AVR core with up to
8676 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8677 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8678 at90c8534, at90s8535).
8680 Instruction set avr3 is for the classic AVR core with up to 128K program
8681 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8683 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8684 memory space (MCU types: atmega8, atmega83, atmega85).
8686 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8687 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8688 atmega64, atmega128, at43usb355, at94k).
8692 Output instruction sizes to the asm file.
8694 @item -minit-stack=@var{N}
8695 @opindex minit-stack
8696 Specify the initial stack address, which may be a symbol or numeric value,
8697 @samp{__stack} is the default.
8699 @item -mno-interrupts
8700 @opindex mno-interrupts
8701 Generated code is not compatible with hardware interrupts.
8702 Code size will be smaller.
8704 @item -mcall-prologues
8705 @opindex mcall-prologues
8706 Functions prologues/epilogues expanded as call to appropriate
8707 subroutines. Code size will be smaller.
8709 @item -mno-tablejump
8710 @opindex mno-tablejump
8711 Do not generate tablejump insns which sometimes increase code size.
8714 @opindex mtiny-stack
8715 Change only the low 8 bits of the stack pointer.
8719 @subsection MCore Options
8720 @cindex MCore options
8722 These are the @samp{-m} options defined for the Motorola M*Core
8732 @opindex mno-hardlit
8733 Inline constants into the code stream if it can be done in two
8734 instructions or less.
8742 Use the divide instruction. (Enabled by default).
8744 @item -mrelax-immediate
8745 @itemx -mrelax-immediate
8746 @itemx -mno-relax-immediate
8747 @opindex mrelax-immediate
8748 @opindex mrelax-immediate
8749 @opindex mno-relax-immediate
8750 Allow arbitrary sized immediates in bit operations.
8752 @item -mwide-bitfields
8753 @itemx -mwide-bitfields
8754 @itemx -mno-wide-bitfields
8755 @opindex mwide-bitfields
8756 @opindex mwide-bitfields
8757 @opindex mno-wide-bitfields
8758 Always treat bit-fields as int-sized.
8760 @item -m4byte-functions
8761 @itemx -m4byte-functions
8762 @itemx -mno-4byte-functions
8763 @opindex m4byte-functions
8764 @opindex m4byte-functions
8765 @opindex mno-4byte-functions
8766 Force all functions to be aligned to a four byte boundary.
8768 @item -mcallgraph-data
8769 @itemx -mcallgraph-data
8770 @itemx -mno-callgraph-data
8771 @opindex mcallgraph-data
8772 @opindex mcallgraph-data
8773 @opindex mno-callgraph-data
8774 Emit callgraph information.
8778 @itemx -mno-slow-bytes
8779 @opindex mslow-bytes
8780 @opindex mslow-bytes
8781 @opindex mno-slow-bytes
8782 Prefer word access when reading byte quantities.
8784 @item -mlittle-endian
8785 @itemx -mlittle-endian
8787 @opindex mlittle-endian
8788 @opindex mlittle-endian
8789 @opindex mbig-endian
8790 Generate code for a little endian target.
8798 Generate code for the 210 processor.
8802 @subsection IA-64 Options
8803 @cindex IA-64 Options
8805 These are the @samp{-m} options defined for the Intel IA-64 architecture.
8809 @opindex mbig-endian
8810 Generate code for a big endian target. This is the default for HPUX@.
8812 @item -mlittle-endian
8813 @opindex mlittle-endian
8814 Generate code for a little endian target. This is the default for AIX5
8821 Generate (or don't) code for the GNU assembler. This is the default.
8822 @c Also, this is the default if the configure option @option{--with-gnu-as}
8829 Generate (or don't) code for the GNU linker. This is the default.
8830 @c Also, this is the default if the configure option @option{--with-gnu-ld}
8835 Generate code that does not use a global pointer register. The result
8836 is not position independent code, and violates the IA-64 ABI@.
8838 @item -mvolatile-asm-stop
8839 @itemx -mno-volatile-asm-stop
8840 @opindex mvolatile-asm-stop
8841 @opindex mno-volatile-asm-stop
8842 Generate (or don't) a stop bit immediately before and after volatile asm
8847 Generate code that works around Itanium B step errata.
8849 @item -mregister-names
8850 @itemx -mno-register-names
8851 @opindex mregister-names
8852 @opindex mno-register-names
8853 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
8854 the stacked registers. This may make assembler output more readable.
8860 Disable (or enable) optimizations that use the small data section. This may
8861 be useful for working around optimizer bugs.
8864 @opindex mconstant-gp
8865 Generate code that uses a single constant global pointer value. This is
8866 useful when compiling kernel code.
8870 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
8871 This is useful when compiling firmware code.
8873 @item -minline-divide-min-latency
8874 @opindex minline-divide-min-latency
8875 Generate code for inline divides using the minimum latency algorithm.
8877 @item -minline-divide-max-throughput
8878 @opindex minline-divide-max-throughput
8879 Generate code for inline divides using the maximum throughput algorithm.
8881 @item -mno-dwarf2-asm
8883 @opindex mno-dwarf2-asm
8884 @opindex mdwarf2-asm
8885 Don't (or do) generate assembler code for the DWARF2 line number debugging
8886 info. This may be useful when not using the GNU assembler.
8888 @item -mfixed-range=@var{register-range}
8889 @opindex mfixed-range
8890 Generate code treating the given register range as fixed registers.
8891 A fixed register is one that the register allocator can not use. This is
8892 useful when compiling kernel code. A register range is specified as
8893 two registers separated by a dash. Multiple register ranges can be
8894 specified separated by a comma.
8898 @subsection D30V Options
8899 @cindex D30V Options
8901 These @samp{-m} options are defined for D30V implementations:
8906 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
8907 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
8908 memory, which starts at location @code{0x80000000}.
8912 Same as the @option{-mextmem} switch.
8916 Link the @samp{.text} section into onchip text memory, which starts at
8917 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
8918 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
8919 into onchip data memory, which starts at location @code{0x20000000}.
8921 @item -mno-asm-optimize
8922 @itemx -masm-optimize
8923 @opindex mno-asm-optimize
8924 @opindex masm-optimize
8925 Disable (enable) passing @option{-O} to the assembler when optimizing.
8926 The assembler uses the @option{-O} option to automatically parallelize
8927 adjacent short instructions where possible.
8929 @item -mbranch-cost=@var{n}
8930 @opindex mbranch-cost
8931 Increase the internal costs of branches to @var{n}. Higher costs means
8932 that the compiler will issue more instructions to avoid doing a branch.
8935 @item -mcond-exec=@var{n}
8937 Specify the maximum number of conditionally executed instructions that
8938 replace a branch. The default is 4.
8941 @node S/390 and zSeries Options
8942 @subsection S/390 and zSeries Options
8943 @cindex S/390 and zSeries Options
8945 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
8950 @opindex mhard-float
8951 @opindex msoft-float
8952 Use (do not use) the hardware floating-point instructions and registers
8953 for floating-point operations. When @option{-msoft-float} is specified,
8954 functions in @file{libgcc.a} will be used to perform floating-point
8955 operations. When @option{-mhard-float} is specified, the compiler
8956 generates IEEE floating-point instructions. This is the default.
8959 @itemx -mno-backchain
8961 @opindex mno-backchain
8962 Generate (or do not generate) code which maintains an explicit
8963 backchain within the stack frame that points to the caller's frame.
8964 This is currently needed to allow debugging. The default is to
8965 generate the backchain.
8968 @itemx -mno-small-exec
8969 @opindex msmall-exec
8970 @opindex mno-small-exec
8971 Generate (or do not generate) code using the @code{bras} instruction
8972 to do subroutine calls.
8973 This only works reliably if the total executable size does not
8974 exceed 64k. The default is to use the @code{basr} instruction instead,
8975 which does not have this limitation.
8981 When @option{-m31} is specified, generate code compliant to the
8982 Linux for S/390 ABI@. When @option{-m64} is specified, generate
8983 code compliant to the Linux for zSeries ABI@. This allows GCC in
8984 particular to generate 64-bit instructions. For the @samp{s390}
8985 targets, the default is @option{-m31}, while the @samp{s390x}
8986 targets default to @option{-m64}.
8992 Generate (or do not generate) code using the @code{mvcle} instruction
8993 to perform block moves. When @option{-mno-mvcle} is specifed,
8994 use a @code{mvc} loop instead. This is the default.
9000 Print (or do not print) additional debug information when compiling.
9001 The default is to not print debug information.
9006 @subsection CRIS Options
9007 @cindex CRIS Options
9009 These options are defined specifically for the CRIS ports.
9012 @item -march=@var{architecture-type}
9013 @itemx -mcpu=@var{architecture-type}
9016 Generate code for the specified architecture. The choices for
9017 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9018 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9019 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9022 @item -mtune=@var{architecture-type}
9024 Tune to @var{architecture-type} everything applicable about the generated
9025 code, except for the ABI and the set of available instructions. The
9026 choices for @var{architecture-type} are the same as for
9027 @option{-march=@var{architecture-type}}.
9029 @item -mmax-stack-frame=@var{n}
9030 @opindex mmax-stack-frame
9031 Warn when the stack frame of a function exceeds @var{n} bytes.
9033 @item -melinux-stacksize=@var{n}
9034 @opindex melinux-stacksize
9035 Only available with the @samp{cris-axis-aout} target. Arranges for
9036 indications in the program to the kernel loader that the stack of the
9037 program should be set to @var{n} bytes.
9043 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9044 @option{-march=v3} and @option{-march=v8} respectively.
9048 Enable CRIS-specific verbose debug-related information in the assembly
9049 code. This option also has the effect to turn off the @samp{#NO_APP}
9050 formatted-code indicator to the assembler at the beginning of the
9055 Do not use condition-code results from previous instruction; always emit
9056 compare and test instructions before use of condition codes.
9058 @item -mno-side-effects
9059 @opindex mno-side-effects
9060 Do not emit instructions with side-effects in addressing modes other than
9064 @itemx -mno-stack-align
9066 @itemx -mno-data-align
9067 @itemx -mconst-align
9068 @itemx -mno-const-align
9069 @opindex mstack-align
9070 @opindex mno-stack-align
9071 @opindex mdata-align
9072 @opindex mno-data-align
9073 @opindex mconst-align
9074 @opindex mno-const-align
9075 These options (no-options) arranges (eliminate arrangements) for the
9076 stack-frame, individual data and constants to be aligned for the maximum
9077 single data access size for the chosen CPU model. The default is to
9078 arrange for 32-bit alignment. ABI details such as structure layout are
9079 not affected by these options.
9087 Similar to the stack- data- and const-align options above, these options
9088 arrange for stack-frame, writable data and constants to all be 32-bit,
9089 16-bit or 8-bit aligned. The default is 32-bit alignment.
9091 @item -mno-prologue-epilogue
9092 @itemx -mprologue-epilogue
9093 @opindex mno-prologue-epilogue
9094 @opindex mprologue-epilogue
9095 With @option{-mno-prologue-epilogue}, the normal function prologue and
9096 epilogue that sets up the stack-frame are omitted and no return
9097 instructions or return sequences are generated in the code. Use this
9098 option only together with visual inspection of the compiled code: no
9099 warnings or errors are generated when call-saved registers must be saved,
9100 or storage for local variable needs to be allocated.
9106 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9107 instruction sequences that load addresses for functions from the PLT part
9108 of the GOT rather than (traditional on other architectures) calls to the
9109 PLT. The default is @option{-mgotplt}.
9113 Legacy no-op option only recognized with the cris-axis-aout target.
9117 Legacy no-op option only recognized with the cris-axis-elf and
9118 cris-axis-linux-gnu targets.
9122 Only recognized with the cris-axis-aout target, where it selects a
9123 GNU/linux-like multilib, include files and instruction set for
9128 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9132 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9133 to link with input-output functions from a simulator library. Code,
9134 initialized data and zero-initialized data are allocated consecutively.
9138 Like @option{-sim}, but pass linker options to locate initialized data at
9139 0x40000000 and zero-initialized data at 0x80000000.
9143 @subsection MMIX Options
9144 @cindex MMIX Options
9146 These options are defined for the MMIX:
9150 @itemx -mno-libfuncs
9152 @opindex mno-libfuncs
9153 Specify that intrinsic library functions are being compiled, passing all
9154 values in registers, no matter the size.
9159 @opindex mno-epsilon
9160 Generate floating-point comparison instructions that compare with respect
9161 to the @code{rE} epsilon register.
9163 @item -mabi=mmixware
9165 @opindex mabi-mmixware
9167 Generate code that passes function parameters and return values that (in
9168 the called function) are seen as registers @code{$0} and up, as opposed to
9169 the GNU ABI which uses global registers @code{$231} and up.
9172 @itemx -mno-zero-extend
9173 @opindex mzero-extend
9174 @opindex mno-zero-extend
9175 When reading data from memory in sizes shorter than 64 bits, use (do not
9176 use) zero-extending load instructions by default, rather than
9177 sign-extending ones.
9180 @itemx -mno-knuthdiv
9182 @opindex mno-knuthdiv
9183 Make the result of a division yielding a remainder have the same sign as
9184 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9185 remainder follows the sign of the dividend. Both methods are
9186 arithmetically valid, the latter being almost exclusively used.
9188 @item -mtoplevel-symbols
9189 @itemx -mno-toplevel-symbols
9190 @opindex mtoplevel-symbols
9191 @opindex mno-toplevel-symbols
9192 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9193 code can be used with the @code{PREFIX} assembly directive.
9197 Generate an executable in the ELF format, rather than the default
9198 @samp{mmo} format used by the @command{mmix} simulator.
9200 @item -mbranch-predict
9201 @itemx -mno-branch-predict
9202 @opindex mbranch-predict
9203 @opindex mno-branch-predict
9204 Use (do not use) the probable-branch instructions, when static branch
9205 prediction indicates a probable branch.
9207 @item -mbase-addresses
9208 @itemx -mno-base-addresses
9209 @opindex mbase-addresses
9210 @opindex mno-base-addresses
9211 Generate (do not generate) code that uses @emph{base addresses}. Using a
9212 base address automatically generates a request (handled by the assembler
9213 and the linker) for a constant to be set up in a global register. The
9214 register is used for one or more base address requests within the range 0
9215 to 255 from the value held in the register. The generally leads to short
9216 and fast code, but the number of different data items that can be
9217 addressed is limited. This means that a program that uses lots of static
9218 data may require @option{-mno-base-addresses}.
9221 @node PDP-11 Options
9222 @subsection PDP-11 Options
9223 @cindex PDP-11 Options
9225 These options are defined for the PDP-11:
9230 Use hardware FPP floating point. This is the default. (FIS floating
9231 point on the PDP-11/40 is not supported.)
9234 @opindex msoft-float
9235 Do not use hardware floating point.
9239 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9243 Return floating-point results in memory. This is the default.
9247 Generate code for a PDP-11/40.
9251 Generate code for a PDP-11/45. This is the default.
9255 Generate code for a PDP-11/10.
9257 @item -mbcopy-builtin
9258 @opindex bcopy-builtin
9259 Use inline @code{movstrhi} patterns for copying memory. This is the
9264 Do not use inline @code{movstrhi} patterns for copying memory.
9270 Use 16-bit @code{int}. This is the default.
9276 Use 32-bit @code{int}.
9281 @opindex mno-float32
9282 Use 64-bit @code{float}. This is the default.
9287 @opindex mno-float64
9288 Use 32-bit @code{float}.
9292 Use @code{abshi2} pattern. This is the default.
9296 Do not use @code{abshi2} pattern.
9298 @item -mbranch-expensive
9299 @opindex mbranch-expensive
9300 Pretend that branches are expensive. This is for experimenting with
9301 code generation only.
9303 @item -mbranch-cheap
9304 @opindex mbranch-cheap
9305 Do not pretend that branches are expensive. This is the default.
9309 Generate code for a system with split I&D.
9313 Generate code for a system without split I&D. This is the default.
9317 Use Unix assembler syntax. This is the default when configured for
9322 Use DEC assembler syntax. This is the default when configured for any
9323 PDP-11 target other than @samp{pdp11-*-bsd}.
9326 @node Xstormy16 Options
9327 @subsection Xstormy16 Options
9328 @cindex Xstormy16 Options
9330 These options are defined for Xstormy16:
9335 Choose startup files and linker script suitable for the simulator.
9338 @node Xtensa Options
9339 @subsection Xtensa Options
9340 @cindex Xtensa Options
9342 The Xtensa architecture is designed to support many different
9343 configurations. The compiler's default options can be set to match a
9344 particular Xtensa configuration by copying a configuration file into the
9345 GCC sources when building GCC@. The options below may be used to
9346 override the default options.
9350 @itemx -mlittle-endian
9351 @opindex mbig-endian
9352 @opindex mlittle-endian
9353 Specify big-endian or little-endian byte ordering for the target Xtensa
9359 @opindex mno-density
9360 Enable or disable use of the optional Xtensa code density instructions.
9366 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
9367 will generate MAC16 instructions from standard C code, with the
9368 limitation that it will use neither the MR register file nor any
9369 instruction that operates on the MR registers. When this option is
9370 disabled, GCC will translate 16-bit multiply/accumulate operations to a
9371 combination of core instructions and library calls, depending on whether
9372 any other multiplier options are enabled.
9378 Enable or disable use of the 16-bit integer multiplier option. When
9379 enabled, the compiler will generate 16-bit multiply instructions for
9380 multiplications of 16 bits or smaller in standard C code. When this
9381 option is disabled, the compiler will either use 32-bit multiply or
9382 MAC16 instructions if they are available or generate library calls to
9383 perform the multiply operations using shifts and adds.
9389 Enable or disable use of the 32-bit integer multiplier option. When
9390 enabled, the compiler will generate 32-bit multiply instructions for
9391 multiplications of 32 bits or smaller in standard C code. When this
9392 option is disabled, the compiler will generate library calls to perform
9393 the multiply operations using either shifts and adds or 16-bit multiply
9394 instructions if they are available.
9400 Enable or disable use of the optional normalization shift amount
9401 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
9407 Enable or disable use of the optional minimum and maximum value
9414 Enable or disable use of the optional sign extend (@code{SEXT})
9418 @itemx -mno-booleans
9420 @opindex mno-booleans
9421 Enable or disable support for the boolean register file used by Xtensa
9422 coprocessors. This is not typically useful by itself but may be
9423 required for other options that make use of the boolean registers (e.g.,
9424 the floating-point option).
9428 @opindex mhard-float
9429 @opindex msoft-float
9430 Enable or disable use of the floating-point option. When enabled, GCC
9431 generates floating-point instructions for 32-bit @code{float}
9432 operations. When this option is disabled, GCC generates library calls
9433 to emulate 32-bit floating-point operations using integer instructions.
9434 Regardless of this option, 64-bit @code{double} operations are always
9435 emulated with calls to library functions.
9438 @itemx -mno-fused-madd
9439 @opindex mfused-madd
9440 @opindex mno-fused-madd
9441 Enable or disable use of fused multiply/add and multiply/subtract
9442 instructions in the floating-point option. This has no effect if the
9443 floating-point option is not also enabled. Disabling fused multiply/add
9444 and multiply/subtract instructions forces the compiler to use separate
9445 instructions for the multiply and add/subtract operations. This may be
9446 desirable in some cases where strict IEEE 754-compliant results are
9447 required: the fused multiply add/subtract instructions do not round the
9448 intermediate result, thereby producing results with @emph{more} bits of
9449 precision than specified by the IEEE standard. Disabling fused multiply
9450 add/subtract instructions also ensures that the program output is not
9451 sensitive to the compiler's ability to combine multiply and add/subtract
9454 @item -mserialize-volatile
9455 @itemx -mno-serialize-volatile
9456 @opindex mserialize-volatile
9457 @opindex mno-serialize-volatile
9458 When this option is enabled, GCC inserts @code{MEMW} instructions before
9459 @code{volatile} memory references to guarantee sequential consistency.
9460 The default is @option{-mserialize-volatile}. Use
9461 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
9463 @item -mtext-section-literals
9464 @itemx -mno-text-section-literals
9465 @opindex mtext-section-literals
9466 @opindex mno-text-section-literals
9467 Control the treatment of literal pools. The default is
9468 @option{-mno-text-section-literals}, which places literals in a separate
9469 section in the output file. This allows the literal pool to be placed
9470 in a data RAM/ROM, and it also allows the linker to combine literal
9471 pools from separate object files to remove redundant literals and
9472 improve code size. With @option{-mtext-section-literals}, the literals
9473 are interspersed in the text section in order to keep them as close as
9474 possible to their references. This may be necessary for large assembly
9477 @item -mtarget-align
9478 @itemx -mno-target-align
9479 @opindex mtarget-align
9480 @opindex mno-target-align
9481 When this option is enabled, GCC instructs the assembler to
9482 automatically align instructions to reduce branch penalties at the
9483 expense of some code density. The assembler attempts to widen density
9484 instructions to align branch targets and the instructions following call
9485 instructions. If there are not enough preceding safe density
9486 instructions to align a target, no widening will be performed. The
9487 default is @option{-mtarget-align}. These options do not affect the
9488 treatment of auto-aligned instructions like @code{LOOP}, which the
9489 assembler will always align, either by widening density instructions or
9490 by inserting no-op instructions.
9493 @itemx -mno-longcalls
9495 @opindex mno-longcalls
9496 When this option is enabled, GCC instructs the assembler to translate
9497 direct calls to indirect calls unless it can determine that the target
9498 of a direct call is in the range allowed by the call instruction. This
9499 translation typically occurs for calls to functions in other source
9500 files. Specifically, the assembler translates a direct @code{CALL}
9501 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
9502 The default is @option{-mno-longcalls}. This option should be used in
9503 programs where the call target can potentially be out of range. This
9504 option is implemented in the assembler, not the compiler, so the
9505 assembly code generated by GCC will still show direct call
9506 instructions---look at the disassembled object code to see the actual
9507 instructions. Note that the assembler will use an indirect call for
9508 every cross-file call, not just those that really will be out of range.
9511 @node Code Gen Options
9512 @section Options for Code Generation Conventions
9513 @cindex code generation conventions
9514 @cindex options, code generation
9515 @cindex run-time options
9517 These machine-independent options control the interface conventions
9518 used in code generation.
9520 Most of them have both positive and negative forms; the negative form
9521 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9522 one of the forms is listed---the one which is not the default. You
9523 can figure out the other form by either removing @samp{no-} or adding
9528 @opindex fexceptions
9529 Enable exception handling. Generates extra code needed to propagate
9530 exceptions. For some targets, this implies GCC will generate frame
9531 unwind information for all functions, which can produce significant data
9532 size overhead, although it does not affect execution. If you do not
9533 specify this option, GCC will enable it by default for languages like
9534 C++ which normally require exception handling, and disable it for
9535 languages like C that do not normally require it. However, you may need
9536 to enable this option when compiling C code that needs to interoperate
9537 properly with exception handlers written in C++. You may also wish to
9538 disable this option if you are compiling older C++ programs that don't
9539 use exception handling.
9541 @item -fnon-call-exceptions
9542 @opindex fnon-call-exceptions
9543 Generate code that allows trapping instructions to throw exceptions.
9544 Note that this requires platform-specific runtime support that does
9545 not exist everywhere. Moreover, it only allows @emph{trapping}
9546 instructions to throw exceptions, i.e.@: memory references or floating
9547 point instructions. It does not allow exceptions to be thrown from
9548 arbitrary signal handlers such as @code{SIGALRM}.
9550 @item -funwind-tables
9551 @opindex funwind-tables
9552 Similar to @option{-fexceptions}, except that it will just generate any needed
9553 static data, but will not affect the generated code in any other way.
9554 You will normally not enable this option; instead, a language processor
9555 that needs this handling would enable it on your behalf.
9557 @item -fasynchronous-unwind-tables
9558 @opindex funwind-tables
9559 Generate unwind table in dwarf2 format, if supported by target machine. The
9560 table is exact at each instruction boundary, so it can be used for stack
9561 unwinding from asynchronous events (such as debugger or garbage collector).
9563 @item -fpcc-struct-return
9564 @opindex fpcc-struct-return
9565 Return ``short'' @code{struct} and @code{union} values in memory like
9566 longer ones, rather than in registers. This convention is less
9567 efficient, but it has the advantage of allowing intercallability between
9568 GCC-compiled files and files compiled with other compilers.
9570 The precise convention for returning structures in memory depends
9571 on the target configuration macros.
9573 Short structures and unions are those whose size and alignment match
9574 that of some integer type.
9576 @item -freg-struct-return
9577 @opindex freg-struct-return
9578 Return @code{struct} and @code{union} values in registers when possible.
9579 This is more efficient for small structures than
9580 @option{-fpcc-struct-return}.
9582 If you specify neither @option{-fpcc-struct-return} nor
9583 @option{-freg-struct-return}, GCC defaults to whichever convention is
9584 standard for the target. If there is no standard convention, GCC
9585 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9586 the principal compiler. In those cases, we can choose the standard, and
9587 we chose the more efficient register return alternative.
9590 @opindex fshort-enums
9591 Allocate to an @code{enum} type only as many bytes as it needs for the
9592 declared range of possible values. Specifically, the @code{enum} type
9593 will be equivalent to the smallest integer type which has enough room.
9595 @item -fshort-double
9596 @opindex fshort-double
9597 Use the same size for @code{double} as for @code{float}.
9600 @opindex fshared-data
9601 Requests that the data and non-@code{const} variables of this
9602 compilation be shared data rather than private data. The distinction
9603 makes sense only on certain operating systems, where shared data is
9604 shared between processes running the same program, while private data
9605 exists in one copy per process.
9609 In C, allocate even uninitialized global variables in the data section of the
9610 object file, rather than generating them as common blocks. This has the
9611 effect that if the same variable is declared (without @code{extern}) in
9612 two different compilations, you will get an error when you link them.
9613 The only reason this might be useful is if you wish to verify that the
9614 program will work on other systems which always work this way.
9618 Ignore the @samp{#ident} directive.
9620 @item -fno-gnu-linker
9621 @opindex fno-gnu-linker
9622 Do not output global initializations (such as C++ constructors and
9623 destructors) in the form used by the GNU linker (on systems where the GNU
9624 linker is the standard method of handling them). Use this option when
9625 you want to use a non-GNU linker, which also requires using the
9626 @command{collect2} program to make sure the system linker includes
9627 constructors and destructors. (@command{collect2} is included in the GCC
9628 distribution.) For systems which @emph{must} use @command{collect2}, the
9629 compiler driver @command{gcc} is configured to do this automatically.
9631 @item -finhibit-size-directive
9632 @opindex finhibit-size-directive
9633 Don't output a @code{.size} assembler directive, or anything else that
9634 would cause trouble if the function is split in the middle, and the
9635 two halves are placed at locations far apart in memory. This option is
9636 used when compiling @file{crtstuff.c}; you should not need to use it
9640 @opindex fverbose-asm
9641 Put extra commentary information in the generated assembly code to
9642 make it more readable. This option is generally only of use to those
9643 who actually need to read the generated assembly code (perhaps while
9644 debugging the compiler itself).
9646 @option{-fno-verbose-asm}, the default, causes the
9647 extra information to be omitted and is useful when comparing two assembler
9652 Consider all memory references through pointers to be volatile.
9654 @item -fvolatile-global
9655 @opindex fvolatile-global
9656 Consider all memory references to extern and global data items to
9657 be volatile. GCC does not consider static data items to be volatile
9658 because of this switch.
9660 @item -fvolatile-static
9661 @opindex fvolatile-static
9662 Consider all memory references to static data to be volatile.
9666 @cindex global offset table
9668 Generate position-independent code (PIC) suitable for use in a shared
9669 library, if supported for the target machine. Such code accesses all
9670 constant addresses through a global offset table (GOT)@. The dynamic
9671 loader resolves the GOT entries when the program starts (the dynamic
9672 loader is not part of GCC; it is part of the operating system). If
9673 the GOT size for the linked executable exceeds a machine-specific
9674 maximum size, you get an error message from the linker indicating that
9675 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9676 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9677 on the m68k and RS/6000. The 386 has no such limit.)
9679 Position-independent code requires special support, and therefore works
9680 only on certain machines. For the 386, GCC supports PIC for System V
9681 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9682 position-independent.
9686 If supported for the target machine, emit position-independent code,
9687 suitable for dynamic linking and avoiding any limit on the size of the
9688 global offset table. This option makes a difference on the m68k, m88k,
9691 Position-independent code requires special support, and therefore works
9692 only on certain machines.
9694 @item -ffixed-@var{reg}
9696 Treat the register named @var{reg} as a fixed register; generated code
9697 should never refer to it (except perhaps as a stack pointer, frame
9698 pointer or in some other fixed role).
9700 @var{reg} must be the name of a register. The register names accepted
9701 are machine-specific and are defined in the @code{REGISTER_NAMES}
9702 macro in the machine description macro file.
9704 This flag does not have a negative form, because it specifies a
9707 @item -fcall-used-@var{reg}
9709 Treat the register named @var{reg} as an allocable register that is
9710 clobbered by function calls. It may be allocated for temporaries or
9711 variables that do not live across a call. Functions compiled this way
9712 will not save and restore the register @var{reg}.
9714 It is an error to used this flag with the frame pointer or stack pointer.
9715 Use of this flag for other registers that have fixed pervasive roles in
9716 the machine's execution model will produce disastrous results.
9718 This flag does not have a negative form, because it specifies a
9721 @item -fcall-saved-@var{reg}
9722 @opindex fcall-saved
9723 Treat the register named @var{reg} as an allocable register saved by
9724 functions. It may be allocated even for temporaries or variables that
9725 live across a call. Functions compiled this way will save and restore
9726 the register @var{reg} if they use it.
9728 It is an error to used this flag with the frame pointer or stack pointer.
9729 Use of this flag for other registers that have fixed pervasive roles in
9730 the machine's execution model will produce disastrous results.
9732 A different sort of disaster will result from the use of this flag for
9733 a register in which function values may be returned.
9735 This flag does not have a negative form, because it specifies a
9739 @opindex fpack-struct
9740 Pack all structure members together without holes. Usually you would
9741 not want to use this option, since it makes the code suboptimal, and
9742 the offsets of structure members won't agree with system libraries.
9744 @item -finstrument-functions
9745 @opindex finstrument-functions
9746 Generate instrumentation calls for entry and exit to functions. Just
9747 after function entry and just before function exit, the following
9748 profiling functions will be called with the address of the current
9749 function and its call site. (On some platforms,
9750 @code{__builtin_return_address} does not work beyond the current
9751 function, so the call site information may not be available to the
9752 profiling functions otherwise.)
9755 void __cyg_profile_func_enter (void *this_fn,
9757 void __cyg_profile_func_exit (void *this_fn,
9761 The first argument is the address of the start of the current function,
9762 which may be looked up exactly in the symbol table.
9764 This instrumentation is also done for functions expanded inline in other
9765 functions. The profiling calls will indicate where, conceptually, the
9766 inline function is entered and exited. This means that addressable
9767 versions of such functions must be available. If all your uses of a
9768 function are expanded inline, this may mean an additional expansion of
9769 code size. If you use @samp{extern inline} in your C code, an
9770 addressable version of such functions must be provided. (This is
9771 normally the case anyways, but if you get lucky and the optimizer always
9772 expands the functions inline, you might have gotten away without
9773 providing static copies.)
9775 A function may be given the attribute @code{no_instrument_function}, in
9776 which case this instrumentation will not be done. This can be used, for
9777 example, for the profiling functions listed above, high-priority
9778 interrupt routines, and any functions from which the profiling functions
9779 cannot safely be called (perhaps signal handlers, if the profiling
9780 routines generate output or allocate memory).
9783 @opindex fstack-check
9784 Generate code to verify that you do not go beyond the boundary of the
9785 stack. You should specify this flag if you are running in an
9786 environment with multiple threads, but only rarely need to specify it in
9787 a single-threaded environment since stack overflow is automatically
9788 detected on nearly all systems if there is only one stack.
9790 Note that this switch does not actually cause checking to be done; the
9791 operating system must do that. The switch causes generation of code
9792 to ensure that the operating system sees the stack being extended.
9794 @item -fstack-limit-register=@var{reg}
9795 @itemx -fstack-limit-symbol=@var{sym}
9796 @itemx -fno-stack-limit
9797 @opindex fstack-limit-register
9798 @opindex fstack-limit-symbol
9799 @opindex fno-stack-limit
9800 Generate code to ensure that the stack does not grow beyond a certain value,
9801 either the value of a register or the address of a symbol. If the stack
9802 would grow beyond the value, a signal is raised. For most targets,
9803 the signal is raised before the stack overruns the boundary, so
9804 it is possible to catch the signal without taking special precautions.
9806 For instance, if the stack starts at absolute address @samp{0x80000000}
9807 and grows downwards, you can use the flags
9808 @option{-fstack-limit-symbol=__stack_limit} and
9809 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
9810 of 128KB@. Note that this may only work with the GNU linker.
9812 @cindex aliasing of parameters
9813 @cindex parameters, aliased
9814 @item -fargument-alias
9815 @itemx -fargument-noalias
9816 @itemx -fargument-noalias-global
9817 @opindex fargument-alias
9818 @opindex fargument-noalias
9819 @opindex fargument-noalias-global
9820 Specify the possible relationships among parameters and between
9821 parameters and global data.
9823 @option{-fargument-alias} specifies that arguments (parameters) may
9824 alias each other and may alias global storage.@*
9825 @option{-fargument-noalias} specifies that arguments do not alias
9826 each other, but may alias global storage.@*
9827 @option{-fargument-noalias-global} specifies that arguments do not
9828 alias each other and do not alias global storage.
9830 Each language will automatically use whatever option is required by
9831 the language standard. You should not need to use these options yourself.
9833 @item -fleading-underscore
9834 @opindex fleading-underscore
9835 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
9836 change the way C symbols are represented in the object file. One use
9837 is to help link with legacy assembly code.
9839 Be warned that you should know what you are doing when invoking this
9840 option, and that not all targets provide complete support for it.
9845 @node Environment Variables
9846 @section Environment Variables Affecting GCC
9847 @cindex environment variables
9849 @c man begin ENVIRONMENT
9851 This section describes several environment variables that affect how GCC
9852 operates. Some of them work by specifying directories or prefixes to use
9853 when searching for various kinds of files. Some are used to specify other
9854 aspects of the compilation environment.
9856 Note that you can also specify places to search using options such as
9857 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
9858 take precedence over places specified using environment variables, which
9859 in turn take precedence over those specified by the configuration of GCC@.
9860 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
9861 GNU Compiler Collection (GCC) Internals}.
9866 @c @itemx LC_COLLATE
9868 @c @itemx LC_MONETARY
9869 @c @itemx LC_NUMERIC
9874 @c @findex LC_COLLATE
9876 @c @findex LC_MONETARY
9877 @c @findex LC_NUMERIC
9881 These environment variables control the way that GCC uses
9882 localization information that allow GCC to work with different
9883 national conventions. GCC inspects the locale categories
9884 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
9885 so. These locale categories can be set to any value supported by your
9886 installation. A typical value is @samp{en_UK} for English in the United
9889 The @env{LC_CTYPE} environment variable specifies character
9890 classification. GCC uses it to determine the character boundaries in
9891 a string; this is needed for some multibyte encodings that contain quote
9892 and escape characters that would otherwise be interpreted as a string
9895 The @env{LC_MESSAGES} environment variable specifies the language to
9896 use in diagnostic messages.
9898 If the @env{LC_ALL} environment variable is set, it overrides the value
9899 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
9900 and @env{LC_MESSAGES} default to the value of the @env{LANG}
9901 environment variable. If none of these variables are set, GCC
9902 defaults to traditional C English behavior.
9906 If @env{TMPDIR} is set, it specifies the directory to use for temporary
9907 files. GCC uses temporary files to hold the output of one stage of
9908 compilation which is to be used as input to the next stage: for example,
9909 the output of the preprocessor, which is the input to the compiler
9912 @item GCC_EXEC_PREFIX
9913 @findex GCC_EXEC_PREFIX
9914 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
9915 names of the subprograms executed by the compiler. No slash is added
9916 when this prefix is combined with the name of a subprogram, but you can
9917 specify a prefix that ends with a slash if you wish.
9919 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
9920 an appropriate prefix to use based on the pathname it was invoked with.
9922 If GCC cannot find the subprogram using the specified prefix, it
9923 tries looking in the usual places for the subprogram.
9925 The default value of @env{GCC_EXEC_PREFIX} is
9926 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
9927 of @code{prefix} when you ran the @file{configure} script.
9929 Other prefixes specified with @option{-B} take precedence over this prefix.
9931 This prefix is also used for finding files such as @file{crt0.o} that are
9934 In addition, the prefix is used in an unusual way in finding the
9935 directories to search for header files. For each of the standard
9936 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
9937 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
9938 replacing that beginning with the specified prefix to produce an
9939 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
9940 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
9941 These alternate directories are searched first; the standard directories
9945 @findex COMPILER_PATH
9946 The value of @env{COMPILER_PATH} is a colon-separated list of
9947 directories, much like @env{PATH}. GCC tries the directories thus
9948 specified when searching for subprograms, if it can't find the
9949 subprograms using @env{GCC_EXEC_PREFIX}.
9952 @findex LIBRARY_PATH
9953 The value of @env{LIBRARY_PATH} is a colon-separated list of
9954 directories, much like @env{PATH}. When configured as a native compiler,
9955 GCC tries the directories thus specified when searching for special
9956 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
9957 using GCC also uses these directories when searching for ordinary
9958 libraries for the @option{-l} option (but directories specified with
9959 @option{-L} come first).
9963 @cindex locale definition
9964 This variable is used to pass locale information to the compiler. One way in
9965 which this information is used is to determine the character set to be used
9966 when character literals, string literals and comments are parsed in C and C++.
9967 When the compiler is configured to allow multibyte characters,
9968 the following values for @env{LANG} are recognized:
9972 Recognize JIS characters.
9974 Recognize SJIS characters.
9976 Recognize EUCJP characters.
9979 If @env{LANG} is not defined, or if it has some other value, then the
9980 compiler will use mblen and mbtowc as defined by the default locale to
9981 recognize and translate multibyte characters.
9985 Some additional environments variables affect the behavior of the
9988 @include cppenv.texi
9992 @node Running Protoize
9993 @section Running Protoize
9995 The program @code{protoize} is an optional part of GCC@. You can use
9996 it to add prototypes to a program, thus converting the program to ISO
9997 C in one respect. The companion program @code{unprotoize} does the
9998 reverse: it removes argument types from any prototypes that are found.
10000 When you run these programs, you must specify a set of source files as
10001 command line arguments. The conversion programs start out by compiling
10002 these files to see what functions they define. The information gathered
10003 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10005 After scanning comes actual conversion. The specified files are all
10006 eligible to be converted; any files they include (whether sources or
10007 just headers) are eligible as well.
10009 But not all the eligible files are converted. By default,
10010 @code{protoize} and @code{unprotoize} convert only source and header
10011 files in the current directory. You can specify additional directories
10012 whose files should be converted with the @option{-d @var{directory}}
10013 option. You can also specify particular files to exclude with the
10014 @option{-x @var{file}} option. A file is converted if it is eligible, its
10015 directory name matches one of the specified directory names, and its
10016 name within the directory has not been excluded.
10018 Basic conversion with @code{protoize} consists of rewriting most
10019 function definitions and function declarations to specify the types of
10020 the arguments. The only ones not rewritten are those for varargs
10023 @code{protoize} optionally inserts prototype declarations at the
10024 beginning of the source file, to make them available for any calls that
10025 precede the function's definition. Or it can insert prototype
10026 declarations with block scope in the blocks where undeclared functions
10029 Basic conversion with @code{unprotoize} consists of rewriting most
10030 function declarations to remove any argument types, and rewriting
10031 function definitions to the old-style pre-ISO form.
10033 Both conversion programs print a warning for any function declaration or
10034 definition that they can't convert. You can suppress these warnings
10037 The output from @code{protoize} or @code{unprotoize} replaces the
10038 original source file. The original file is renamed to a name ending
10039 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10040 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10041 for DOS) file already exists, then the source file is simply discarded.
10043 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10044 scan the program and collect information about the functions it uses.
10045 So neither of these programs will work until GCC is installed.
10047 Here is a table of the options you can use with @code{protoize} and
10048 @code{unprotoize}. Each option works with both programs unless
10052 @item -B @var{directory}
10053 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10054 usual directory (normally @file{/usr/local/lib}). This file contains
10055 prototype information about standard system functions. This option
10056 applies only to @code{protoize}.
10058 @item -c @var{compilation-options}
10059 Use @var{compilation-options} as the options when running @code{gcc} to
10060 produce the @samp{.X} files. The special option @option{-aux-info} is
10061 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10063 Note that the compilation options must be given as a single argument to
10064 @code{protoize} or @code{unprotoize}. If you want to specify several
10065 @code{gcc} options, you must quote the entire set of compilation options
10066 to make them a single word in the shell.
10068 There are certain @code{gcc} arguments that you cannot use, because they
10069 would produce the wrong kind of output. These include @option{-g},
10070 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10071 the @var{compilation-options}, they are ignored.
10074 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10075 systems) instead of @samp{.c}. This is convenient if you are converting
10076 a C program to C++. This option applies only to @code{protoize}.
10079 Add explicit global declarations. This means inserting explicit
10080 declarations at the beginning of each source file for each function
10081 that is called in the file and was not declared. These declarations
10082 precede the first function definition that contains a call to an
10083 undeclared function. This option applies only to @code{protoize}.
10085 @item -i @var{string}
10086 Indent old-style parameter declarations with the string @var{string}.
10087 This option applies only to @code{protoize}.
10089 @code{unprotoize} converts prototyped function definitions to old-style
10090 function definitions, where the arguments are declared between the
10091 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10092 uses five spaces as the indentation. If you want to indent with just
10093 one space instead, use @option{-i " "}.
10096 Keep the @samp{.X} files. Normally, they are deleted after conversion
10100 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10101 a prototype declaration for each function in each block which calls the
10102 function without any declaration. This option applies only to
10106 Make no real changes. This mode just prints information about the conversions
10107 that would have been done without @option{-n}.
10110 Make no @samp{.save} files. The original files are simply deleted.
10111 Use this option with caution.
10113 @item -p @var{program}
10114 Use the program @var{program} as the compiler. Normally, the name
10115 @file{gcc} is used.
10118 Work quietly. Most warnings are suppressed.
10121 Print the version number, just like @option{-v} for @code{gcc}.
10124 If you need special compiler options to compile one of your program's
10125 source files, then you should generate that file's @samp{.X} file
10126 specially, by running @code{gcc} on that source file with the
10127 appropriate options and the option @option{-aux-info}. Then run
10128 @code{protoize} on the entire set of files. @code{protoize} will use
10129 the existing @samp{.X} file because it is newer than the source file.
10133 gcc -Dfoo=bar file1.c -aux-info file1.X
10138 You need to include the special files along with the rest in the
10139 @code{protoize} command, even though their @samp{.X} files already
10140 exist, because otherwise they won't get converted.
10142 @xref{Protoize Caveats}, for more information on how to use
10143 @code{protoize} successfully.