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 -Wno-endif-labels @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 -Wnonnull -Wpacked -Wpadded @gol
229 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
230 -Wreturn-type -Wsequence-point -Wshadow @gol
231 -Wsign-compare -Wswitch -Wswitch-default -Wswitch-enum @gol
232 -Wsystem-headers -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 -floop-optimize -fcrossjumping @gol
270 -fif-conversion -fif-conversion2 @gol
271 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
272 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
273 -fmove-all-movables -fno-default-inline -fno-defer-pop @gol
274 -fno-function-cse -fno-guess-branch-probability @gol
275 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
276 -funsafe-math-optimizations -fno-trapping-math @gol
277 -fno-zero-initialized-in-bss @gol
278 -fomit-frame-pointer -foptimize-register-move @gol
279 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
280 -freduce-all-givs -fregmove -frename-registers @gol
281 -freorder-blocks -freorder-functions @gol
282 -frerun-cse-after-loop -frerun-loop-opt @gol
283 -fschedule-insns -fschedule-insns2 @gol
284 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
285 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps -ftrapv @gol
286 -funroll-all-loops -funroll-loops @gol
287 --param @var{name}=@var{value}
288 -O -O0 -O1 -O2 -O3 -Os}
290 @item Preprocessor Options
291 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
293 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
294 -C -dD -dI -dM -dN @gol
295 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
296 -idirafter @var{dir} @gol
297 -include @var{file} -imacros @var{file} @gol
298 -iprefix @var{file} -iwithprefix @var{dir} @gol
299 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
300 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
301 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
303 @item Assembler Option
304 @xref{Assembler Options,,Passing Options to the Assembler}.
309 @xref{Link Options,,Options for Linking}.
311 @var{object-file-name} -l@var{library} @gol
312 -nostartfiles -nodefaultlibs -nostdlib @gol
313 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
314 -Wl,@var{option} -Xlinker @var{option} @gol
317 @item Directory Options
318 @xref{Directory Options,,Options for Directory Search}.
320 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
323 @c I wrote this xref this way to avoid overfull hbox. -- rms
324 @xref{Target Options}.
326 -V @var{version} -b @var{machine}}
328 @item Machine Dependent Options
329 @xref{Submodel Options,,Hardware Models and Configurations}.
331 @emph{M680x0 Options}
333 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
334 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
335 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
336 -malign-int -mstrict-align}
338 @emph{M68hc1x Options}
340 -m6811 -m6812 -m68hc11 -m68hc12 @gol
341 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
349 -mcpu=@var{cpu-type} @gol
350 -mtune=@var{cpu-type} @gol
351 -mcmodel=@var{code-model} @gol
353 -mapp-regs -mbroken-saverestore -mcypress @gol
354 -mfaster-structs -mflat @gol
355 -mfpu -mhard-float -mhard-quad-float @gol
356 -mimpure-text -mlive-g0 -mno-app-regs @gol
357 -mno-faster-structs -mno-flat -mno-fpu @gol
358 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
359 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
360 -msupersparc -munaligned-doubles -mv8}
362 @emph{Convex Options}
364 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
365 -margcount -mnoargcount @gol
366 -mlong32 -mlong64 @gol
367 -mvolatile-cache -mvolatile-nocache}
369 @emph{AMD29K Options}
371 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
372 -mlarge -mnormal -msmall @gol
373 -mkernel-registers -mno-reuse-arg-regs @gol
374 -mno-stack-check -mno-storem-bug @gol
375 -mreuse-arg-regs -msoft-float -mstack-check @gol
376 -mstorem-bug -muser-registers}
380 -mapcs-frame -mno-apcs-frame @gol
381 -mapcs-26 -mapcs-32 @gol
382 -mapcs-stack-check -mno-apcs-stack-check @gol
383 -mapcs-float -mno-apcs-float @gol
384 -mapcs-reentrant -mno-apcs-reentrant @gol
385 -msched-prolog -mno-sched-prolog @gol
386 -mlittle-endian -mbig-endian -mwords-little-endian @gol
387 -malignment-traps -mno-alignment-traps @gol
388 -msoft-float -mhard-float -mfpe @gol
389 -mthumb-interwork -mno-thumb-interwork @gol
390 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
391 -mstructure-size-boundary=@var{n} @gol
392 -mbsd -mxopen -mno-symrename @gol
393 -mabort-on-noreturn @gol
394 -mlong-calls -mno-long-calls @gol
395 -msingle-pic-base -mno-single-pic-base @gol
396 -mpic-register=@var{reg} @gol
397 -mnop-fun-dllimport @gol
398 -mpoke-function-name @gol
400 -mtpcs-frame -mtpcs-leaf-frame @gol
401 -mcaller-super-interworking -mcallee-super-interworking }
403 @emph{MN10200 Options}
407 @emph{MN10300 Options}
409 -mmult-bug -mno-mult-bug @gol
410 -mam33 -mno-am33 @gol
413 @emph{M32R/D Options}
415 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
420 -m88000 -m88100 -m88110 -mbig-pic @gol
421 -mcheck-zero-division -mhandle-large-shift @gol
422 -midentify-revision -mno-check-zero-division @gol
423 -mno-ocs-debug-info -mno-ocs-frame-position @gol
424 -mno-optimize-arg-area -mno-serialize-volatile @gol
425 -mno-underscores -mocs-debug-info @gol
426 -mocs-frame-position -moptimize-arg-area @gol
427 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
428 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
429 -mversion-03.00 -mwarn-passed-structs}
431 @emph{RS/6000 and PowerPC Options}
433 -mcpu=@var{cpu-type} @gol
434 -mtune=@var{cpu-type} @gol
435 -mpower -mno-power -mpower2 -mno-power2 @gol
436 -mpowerpc -mpowerpc64 -mno-powerpc @gol
437 -maltivec -mno-altivec @gol
438 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
439 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
440 -mnew-mnemonics -mold-mnemonics @gol
441 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
442 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
443 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
444 -mstring -mno-string -mupdate -mno-update @gol
445 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
446 -mstrict-align -mno-strict-align -mrelocatable @gol
447 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
448 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
449 -mcall-aix -mcall-sysv -mcall-netbsd @gol
450 -maix-struct-return -msvr4-struct-return
451 -mabi=altivec -mabi=no-altivec @gol
452 -mprototype -mno-prototype @gol
453 -msim -mmvme -mads -myellowknife -memb -msdata @gol
454 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
458 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
459 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
460 -mminimum-fp-blocks -mnohc-struct-return}
464 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
465 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
466 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
467 -mgas -mgp32 -mgp64 @gol
468 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
469 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
470 -mmips-as -mmips-tfile -mno-abicalls @gol
471 -mno-embedded-data -mno-uninit-const-in-rodata @gol
472 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
473 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
474 -mrnames -msoft-float @gol
475 -m4650 -msingle-float -mmad @gol
476 -mstats -EL -EB -G @var{num} -nocpp @gol
477 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
478 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
480 @emph{i386 and x86-64 Options}
482 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
483 -masm=@var{dialect} -mno-fancy-math-387 @gol
484 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
485 -mno-wide-multiply -mrtd -malign-double @gol
486 -mpreferred-stack-boundary=@var{num} @gol
487 -mmmx -msse -msse2 -msse-math -m3dnow @gol
488 -mthreads -mno-align-stringops -minline-all-stringops @gol
489 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
490 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
496 -march=@var{architecture-type} @gol
497 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
498 -mfast-indirect-calls -mgas -mjump-in-delay @gol
499 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
500 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
501 -mno-jump-in-delay -mno-long-load-store @gol
502 -mno-portable-runtime -mno-soft-float @gol
503 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
504 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
505 -mschedule=@var{cpu-type} -mspace-regs}
507 @emph{Intel 960 Options}
509 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
510 -mcode-align -mcomplex-addr -mleaf-procedures @gol
511 -mic-compat -mic2.0-compat -mic3.0-compat @gol
512 -mintel-asm -mno-clean-linkage -mno-code-align @gol
513 -mno-complex-addr -mno-leaf-procedures @gol
514 -mno-old-align -mno-strict-align -mno-tail-call @gol
515 -mnumerics -mold-align -msoft-float -mstrict-align @gol
518 @emph{DEC Alpha Options}
520 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
521 -mieee -mieee-with-inexact -mieee-conformant @gol
522 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
523 -mtrap-precision=@var{mode} -mbuild-constants @gol
524 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
525 -mbwx -mmax -mfix -mcix @gol
526 -mfloat-vax -mfloat-ieee @gol
527 -mexplicit-relocs -msmall-data -mlarge-data @gol
528 -mmemory-latency=@var{time}}
530 @emph{DEC Alpha/VMS Options}
534 @emph{Clipper Options}
538 @emph{H8/300 Options}
540 -mrelax -mh -ms -mint32 -malign-300}
544 -m1 -m2 -m3 -m3e @gol
545 -m4-nofpu -m4-single-only -m4-single -m4 @gol
546 -m5-64media -m5-64media-nofpu @gol
547 -m5-32media -m5-32media-nofpu @gol
548 -m5-compact -m5-compact-nofpu @gol
549 -mb -ml -mdalign -mrelax @gol
550 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
551 -mieee -misize -mpadstruct -mspace @gol
552 -mprefergot -musermode}
554 @emph{System V Options}
556 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
561 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
562 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
564 @emph{TMS320C3x/C4x Options}
566 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
567 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
568 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
569 -mparallel-insns -mparallel-mpy -mpreserve-float}
573 -mlong-calls -mno-long-calls -mep -mno-ep @gol
574 -mprolog-function -mno-prolog-function -mspace @gol
575 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
580 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
581 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
582 -mregparam -mnoregparam -msb -mnosb @gol
583 -mbitfield -mnobitfield -mhimem -mnohimem}
587 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
588 -mcall-prologues -mno-tablejump -mtiny-stack}
592 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
593 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
594 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
595 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
596 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
600 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
601 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
602 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
607 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
608 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
609 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
610 -minline-divide-max-throughput -mno-dwarf2-asm @gol
611 -mfixed-range=@var{register-range}}
615 -mextmem -mextmemory -monchip -mno-asm-optimize -masm-optimize @gol
616 -mbranch-cost=@var{n} -mcond-exec=@var{n}}
618 @emph{S/390 and zSeries Options}
620 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
621 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
622 -m64 -m31 -mdebug -mno-debug}
626 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
627 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
628 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
629 -mstack-align -mdata-align -mconst-align @gol
630 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
631 -melf -maout -melinux -mlinux -sim -sim2}
633 @emph{PDP-11 Options}
635 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
636 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
637 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
638 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
639 -mbranch-expensive -mbranch-cheap @gol
640 -msplit -mno-split -munix-asm -mdec-asm}
642 @emph{Xstormy16 Options}
646 @emph{Xtensa Options}
648 -mbig-endian -mlittle-endian @gol
649 -mdensity -mno-density @gol
650 -mmac16 -mno-mac16 @gol
651 -mmul16 -mno-mul16 @gol
652 -mmul32 -mno-mul32 @gol
654 -mminmax -mno-minmax @gol
655 -msext -mno-sext @gol
656 -mbooleans -mno-booleans @gol
657 -mhard-float -msoft-float @gol
658 -mfused-madd -mno-fused-madd @gol
659 -mserialize-volatile -mno-serialize-volatile @gol
660 -mtext-section-literals -mno-text-section-literals @gol
661 -mtarget-align -mno-target-align @gol
662 -mlongcalls -mno-longcalls}
664 @item Code Generation Options
665 @xref{Code Gen Options,,Options for Code Generation Conventions}.
667 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
668 -ffixed-@var{reg} -fexceptions @gol
669 -fnon-call-exceptions -funwind-tables @gol
670 -fasynchronous-unwind-tables @gol
671 -finhibit-size-directive -finstrument-functions @gol
672 -fno-common -fno-ident -fno-gnu-linker @gol
673 -fpcc-struct-return -fpic -fPIC @gol
674 -freg-struct-return -fshared-data -fshort-enums @gol
675 -fshort-double -fvolatile @gol
676 -fvolatile-global -fvolatile-static @gol
677 -fverbose-asm -fpack-struct -fstack-check @gol
678 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
679 -fargument-alias -fargument-noalias @gol
680 -fargument-noalias-global -fleading-underscore -ftls-model=@var{model}}
684 * Overall Options:: Controlling the kind of output:
685 an executable, object files, assembler files,
686 or preprocessed source.
687 * C Dialect Options:: Controlling the variant of C language compiled.
688 * C++ Dialect Options:: Variations on C++.
689 * Objective-C Dialect Options:: Variations on Objective-C.
690 * Language Independent Options:: Controlling how diagnostics should be
692 * Warning Options:: How picky should the compiler be?
693 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
694 * Optimize Options:: How much optimization?
695 * Preprocessor Options:: Controlling header files and macro definitions.
696 Also, getting dependency information for Make.
697 * Assembler Options:: Passing options to the assembler.
698 * Link Options:: Specifying libraries and so on.
699 * Directory Options:: Where to find header files and libraries.
700 Where to find the compiler executable files.
701 * Spec Files:: How to pass switches to sub-processes.
702 * Target Options:: Running a cross-compiler, or an old version of GCC.
705 @node Overall Options
706 @section Options Controlling the Kind of Output
708 Compilation can involve up to four stages: preprocessing, compilation
709 proper, assembly and linking, always in that order. The first three
710 stages apply to an individual source file, and end by producing an
711 object file; linking combines all the object files (those newly
712 compiled, and those specified as input) into an executable file.
714 @cindex file name suffix
715 For any given input file, the file name suffix determines what kind of
720 C source code which must be preprocessed.
723 C source code which should not be preprocessed.
726 C++ source code which should not be preprocessed.
729 Objective-C source code. Note that you must link with the library
730 @file{libobjc.a} to make an Objective-C program work.
733 Objective-C source code which should not be preprocessed.
736 C header file (not to be compiled or linked).
740 @itemx @var{file}.cxx
741 @itemx @var{file}.cpp
742 @itemx @var{file}.c++
744 C++ source code which must be preprocessed. Note that in @samp{.cxx},
745 the last two letters must both be literally @samp{x}. Likewise,
746 @samp{.C} refers to a literal capital C@.
749 @itemx @var{file}.for
750 @itemx @var{file}.FOR
751 Fortran source code which should not be preprocessed.
754 @itemx @var{file}.fpp
755 @itemx @var{file}.FPP
756 Fortran source code which must be preprocessed (with the traditional
760 Fortran source code which must be preprocessed with a RATFOR
761 preprocessor (not included with GCC)@.
763 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
764 Using and Porting GNU Fortran}, for more details of the handling of
767 @c FIXME: Descriptions of Java file types.
774 Ada source code file which contains a library unit declaration (a
775 declaration of a package, subprogram, or generic, or a generic
776 instantiation), or a library unit renaming declaration (a package,
777 generic, or subprogram renaming declaration). Such files are also
780 @itemx @var{file}.adb
781 Ada source code file containing a library unit body (a subprogram or
782 package body). Such files are also called @dfn{bodies}.
784 @c GCC also knows about some suffixes for languages not yet included:
793 Assembler code which must be preprocessed.
796 An object file to be fed straight into linking.
797 Any file name with no recognized suffix is treated this way.
801 You can specify the input language explicitly with the @option{-x} option:
804 @item -x @var{language}
805 Specify explicitly the @var{language} for the following input files
806 (rather than letting the compiler choose a default based on the file
807 name suffix). This option applies to all following input files until
808 the next @option{-x} option. Possible values for @var{language} are:
810 c c-header cpp-output
812 objective-c objc-cpp-output
813 assembler assembler-with-cpp
815 f77 f77-cpp-input ratfor
821 Turn off any specification of a language, so that subsequent files are
822 handled according to their file name suffixes (as they are if @option{-x}
823 has not been used at all).
825 @item -pass-exit-codes
826 @opindex pass-exit-codes
827 Normally the @command{gcc} program will exit with the code of 1 if any
828 phase of the compiler returns a non-success return code. If you specify
829 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
830 numerically highest error produced by any phase that returned an error
834 If you only want some of the stages of compilation, you can use
835 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
836 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
837 @command{gcc} is to stop. Note that some combinations (for example,
838 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
843 Compile or assemble the source files, but do not link. The linking
844 stage simply is not done. The ultimate output is in the form of an
845 object file for each source file.
847 By default, the object file name for a source file is made by replacing
848 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
850 Unrecognized input files, not requiring compilation or assembly, are
855 Stop after the stage of compilation proper; do not assemble. The output
856 is in the form of an assembler code file for each non-assembler input
859 By default, the assembler file name for a source file is made by
860 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
862 Input files that don't require compilation are ignored.
866 Stop after the preprocessing stage; do not run the compiler proper. The
867 output is in the form of preprocessed source code, which is sent to the
870 Input files which don't require preprocessing are ignored.
872 @cindex output file option
875 Place output in file @var{file}. This applies regardless to whatever
876 sort of output is being produced, whether it be an executable file,
877 an object file, an assembler file or preprocessed C code.
879 Since only one output file can be specified, it does not make sense to
880 use @option{-o} when compiling more than one input file, unless you are
881 producing an executable file as output.
883 If @option{-o} is not specified, the default is to put an executable file
884 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
885 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
886 all preprocessed C source on standard output.
890 Print (on standard error output) the commands executed to run the stages
891 of compilation. Also print the version number of the compiler driver
892 program and of the preprocessor and the compiler proper.
896 Like @option{-v} except the commands are not executed and all command
897 arguments are quoted. This is useful for shell scripts to capture the
898 driver-generated command lines.
902 Use pipes rather than temporary files for communication between the
903 various stages of compilation. This fails to work on some systems where
904 the assembler is unable to read from a pipe; but the GNU assembler has
909 Print (on the standard output) a description of the command line options
910 understood by @command{gcc}. If the @option{-v} option is also specified
911 then @option{--help} will also be passed on to the various processes
912 invoked by @command{gcc}, so that they can display the command line options
913 they accept. If the @option{-W} option is also specified then command
914 line options which have no documentation associated with them will also
919 Print (on the standard output) a description of target specific command
920 line options for each tool.
924 @section Compiling C++ Programs
926 @cindex suffixes for C++ source
927 @cindex C++ source file suffixes
928 C++ source files conventionally use one of the suffixes @samp{.C},
929 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
930 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
931 files with these names and compiles them as C++ programs even if you
932 call the compiler the same way as for compiling C programs (usually with
933 the name @command{gcc}).
937 However, C++ programs often require class libraries as well as a
938 compiler that understands the C++ language---and under some
939 circumstances, you might want to compile programs from standard input,
940 or otherwise without a suffix that flags them as C++ programs.
941 @command{g++} is a program that calls GCC with the default language
942 set to C++, and automatically specifies linking against the C++
943 library. On many systems, @command{g++} is also
944 installed with the name @command{c++}.
946 @cindex invoking @command{g++}
947 When you compile C++ programs, you may specify many of the same
948 command-line options that you use for compiling programs in any
949 language; or command-line options meaningful for C and related
950 languages; or options that are meaningful only for C++ programs.
951 @xref{C Dialect Options,,Options Controlling C Dialect}, for
952 explanations of options for languages related to C@.
953 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
954 explanations of options that are meaningful only for C++ programs.
956 @node C Dialect Options
957 @section Options Controlling C Dialect
958 @cindex dialect options
959 @cindex language dialect options
960 @cindex options, dialect
962 The following options control the dialect of C (or languages derived
963 from C, such as C++ and Objective-C) that the compiler accepts:
970 In C mode, support all ISO C89 programs. In C++ mode,
971 remove GNU extensions that conflict with ISO C++.
973 This turns off certain features of GCC that are incompatible with ISO
974 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
975 such as the @code{asm} and @code{typeof} keywords, and
976 predefined macros such as @code{unix} and @code{vax} that identify the
977 type of system you are using. It also enables the undesirable and
978 rarely used ISO trigraph feature. For the C compiler,
979 it disables recognition of C++ style @samp{//} comments as well as
980 the @code{inline} keyword.
982 The alternate keywords @code{__asm__}, @code{__extension__},
983 @code{__inline__} and @code{__typeof__} continue to work despite
984 @option{-ansi}. You would not want to use them in an ISO C program, of
985 course, but it is useful to put them in header files that might be included
986 in compilations done with @option{-ansi}. Alternate predefined macros
987 such as @code{__unix__} and @code{__vax__} are also available, with or
988 without @option{-ansi}.
990 The @option{-ansi} option does not cause non-ISO programs to be
991 rejected gratuitously. For that, @option{-pedantic} is required in
992 addition to @option{-ansi}. @xref{Warning Options}.
994 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
995 option is used. Some header files may notice this macro and refrain
996 from declaring certain functions or defining certain macros that the
997 ISO standard doesn't call for; this is to avoid interfering with any
998 programs that might use these names for other things.
1000 Functions which would normally be built in but do not have semantics
1001 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1002 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1003 built-in functions provided by GCC}, for details of the functions
1008 Determine the language standard. This option is currently only
1009 supported when compiling C@. A value for this option must be provided;
1015 ISO C89 (same as @option{-ansi}).
1017 @item iso9899:199409
1018 ISO C89 as modified in amendment 1.
1024 ISO C99. Note that this standard is not yet fully supported; see
1025 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1026 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1029 Default, ISO C89 plus GNU extensions (including some C99 features).
1033 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1034 this will become the default. The name @samp{gnu9x} is deprecated.
1038 Even when this option is not specified, you can still use some of the
1039 features of newer standards in so far as they do not conflict with
1040 previous C standards. For example, you may use @code{__restrict__} even
1041 when @option{-std=c99} is not specified.
1043 The @option{-std} options specifying some version of ISO C have the same
1044 effects as @option{-ansi}, except that features that were not in ISO C89
1045 but are in the specified version (for example, @samp{//} comments and
1046 the @code{inline} keyword in ISO C99) are not disabled.
1048 @xref{Standards,,Language Standards Supported by GCC}, for details of
1049 these standard versions.
1051 @item -aux-info @var{filename}
1053 Output to the given filename prototyped declarations for all functions
1054 declared and/or defined in a translation unit, including those in header
1055 files. This option is silently ignored in any language other than C@.
1057 Besides declarations, the file indicates, in comments, the origin of
1058 each declaration (source file and line), whether the declaration was
1059 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1060 @samp{O} for old, respectively, in the first character after the line
1061 number and the colon), and whether it came from a declaration or a
1062 definition (@samp{C} or @samp{F}, respectively, in the following
1063 character). In the case of function definitions, a K&R-style list of
1064 arguments followed by their declarations is also provided, inside
1065 comments, after the declaration.
1069 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1070 keyword, so that code can use these words as identifiers. You can use
1071 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1072 instead. @option{-ansi} implies @option{-fno-asm}.
1074 In C++, this switch only affects the @code{typeof} keyword, since
1075 @code{asm} and @code{inline} are standard keywords. You may want to
1076 use the @option{-fno-gnu-keywords} flag instead, which has the same
1077 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1078 switch only affects the @code{asm} and @code{typeof} keywords, since
1079 @code{inline} is a standard keyword in ISO C99.
1082 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1083 @opindex fno-builtin
1084 @cindex built-in functions
1085 Don't recognize built-in functions that do not begin with
1086 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1087 functions provided by GCC}, for details of the functions affected,
1088 including those which are not built-in functions when @option{-ansi} or
1089 @option{-std} options for strict ISO C conformance are used because they
1090 do not have an ISO standard meaning.
1092 GCC normally generates special code to handle certain built-in functions
1093 more efficiently; for instance, calls to @code{alloca} may become single
1094 instructions that adjust the stack directly, and calls to @code{memcpy}
1095 may become inline copy loops. The resulting code is often both smaller
1096 and faster, but since the function calls no longer appear as such, you
1097 cannot set a breakpoint on those calls, nor can you change the behavior
1098 of the functions by linking with a different library.
1100 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1101 option has no effect. Therefore, in C++, the only way to get the
1102 optimization benefits of built-in functions is to call the function
1103 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1104 built-in functions to implement many functions (like
1105 @code{std::strchr}), so that you automatically get efficient code.
1107 With the @option{-fno-builtin-@var{function}} option, not available
1108 when compiling C++, only the built-in function @var{function} is
1109 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1110 function is named this is not built-in in this version of GCC, this
1111 option is ignored. There is no corresponding
1112 @option{-fbuiltin-@var{function}} option; if you wish to enable
1113 built-in functions selectively when using @option{-fno-builtin} or
1114 @option{-ffreestanding}, you may define macros such as:
1117 #define abs(n) __builtin_abs ((n))
1118 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1123 @cindex hosted environment
1125 Assert that compilation takes place in a hosted environment. This implies
1126 @option{-fbuiltin}. A hosted environment is one in which the
1127 entire standard library is available, and in which @code{main} has a return
1128 type of @code{int}. Examples are nearly everything except a kernel.
1129 This is equivalent to @option{-fno-freestanding}.
1131 @item -ffreestanding
1132 @opindex ffreestanding
1133 @cindex hosted environment
1135 Assert that compilation takes place in a freestanding environment. This
1136 implies @option{-fno-builtin}. A freestanding environment
1137 is one in which the standard library may not exist, and program startup may
1138 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1139 This is equivalent to @option{-fno-hosted}.
1141 @xref{Standards,,Language Standards Supported by GCC}, for details of
1142 freestanding and hosted environments.
1146 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1147 options for strict ISO C conformance) implies @option{-trigraphs}.
1149 @cindex traditional C language
1150 @cindex C language, traditional
1152 @itemx -traditional-cpp
1153 @opindex traditional-cpp
1154 @opindex traditional
1155 Formerly, these options caused GCC to attempt to emulate a pre-standard
1156 C compiler. They are now only supported with the @option{-E} switch.
1157 The preprocessor continues to support a pre-standard mode. See the GNU
1158 CPP manual for details.
1160 @item -fcond-mismatch
1161 @opindex fcond-mismatch
1162 Allow conditional expressions with mismatched types in the second and
1163 third arguments. The value of such an expression is void. This option
1164 is not supported for C++.
1166 @item -funsigned-char
1167 @opindex funsigned-char
1168 Let the type @code{char} be unsigned, like @code{unsigned char}.
1170 Each kind of machine has a default for what @code{char} should
1171 be. It is either like @code{unsigned char} by default or like
1172 @code{signed char} by default.
1174 Ideally, a portable program should always use @code{signed char} or
1175 @code{unsigned char} when it depends on the signedness of an object.
1176 But many programs have been written to use plain @code{char} and
1177 expect it to be signed, or expect it to be unsigned, depending on the
1178 machines they were written for. This option, and its inverse, let you
1179 make such a program work with the opposite default.
1181 The type @code{char} is always a distinct type from each of
1182 @code{signed char} or @code{unsigned char}, even though its behavior
1183 is always just like one of those two.
1186 @opindex fsigned-char
1187 Let the type @code{char} be signed, like @code{signed char}.
1189 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1190 the negative form of @option{-funsigned-char}. Likewise, the option
1191 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1193 @item -fsigned-bitfields
1194 @itemx -funsigned-bitfields
1195 @itemx -fno-signed-bitfields
1196 @itemx -fno-unsigned-bitfields
1197 @opindex fsigned-bitfields
1198 @opindex funsigned-bitfields
1199 @opindex fno-signed-bitfields
1200 @opindex fno-unsigned-bitfields
1201 These options control whether a bit-field is signed or unsigned, when the
1202 declaration does not use either @code{signed} or @code{unsigned}. By
1203 default, such a bit-field is signed, because this is consistent: the
1204 basic integer types such as @code{int} are signed types.
1206 @item -fwritable-strings
1207 @opindex fwritable-strings
1208 Store string constants in the writable data segment and don't uniquize
1209 them. This is for compatibility with old programs which assume they can
1210 write into string constants.
1212 Writing into string constants is a very bad idea; ``constants'' should
1216 @opindex fshort-wchar
1217 Override the underlying type for @samp{wchar_t} to be @samp{short
1218 unsigned int} instead of the default for the target. This option is
1219 useful for building programs to run under WINE@.
1222 @node C++ Dialect Options
1223 @section Options Controlling C++ Dialect
1225 @cindex compiler options, C++
1226 @cindex C++ options, command line
1227 @cindex options, C++
1228 This section describes the command-line options that are only meaningful
1229 for C++ programs; but you can also use most of the GNU compiler options
1230 regardless of what language your program is in. For example, you
1231 might compile a file @code{firstClass.C} like this:
1234 g++ -g -frepo -O -c firstClass.C
1238 In this example, only @option{-frepo} is an option meant
1239 only for C++ programs; you can use the other options with any
1240 language supported by GCC@.
1242 Here is a list of options that are @emph{only} for compiling C++ programs:
1245 @item -fno-access-control
1246 @opindex fno-access-control
1247 Turn off all access checking. This switch is mainly useful for working
1248 around bugs in the access control code.
1252 Check that the pointer returned by @code{operator new} is non-null
1253 before attempting to modify the storage allocated. The current Working
1254 Paper requires that @code{operator new} never return a null pointer, so
1255 this check is normally unnecessary.
1257 An alternative to using this option is to specify that your
1258 @code{operator new} does not throw any exceptions; if you declare it
1259 @samp{throw()}, G++ will check the return value. See also @samp{new
1262 @item -fconserve-space
1263 @opindex fconserve-space
1264 Put uninitialized or runtime-initialized global variables into the
1265 common segment, as C does. This saves space in the executable at the
1266 cost of not diagnosing duplicate definitions. If you compile with this
1267 flag and your program mysteriously crashes after @code{main()} has
1268 completed, you may have an object that is being destroyed twice because
1269 two definitions were merged.
1271 This option is no longer useful on most targets, now that support has
1272 been added for putting variables into BSS without making them common.
1274 @item -fno-const-strings
1275 @opindex fno-const-strings
1276 Give string constants type @code{char *} instead of type @code{const
1277 char *}. By default, G++ uses type @code{const char *} as required by
1278 the standard. Even if you use @option{-fno-const-strings}, you cannot
1279 actually modify the value of a string constant, unless you also use
1280 @option{-fwritable-strings}.
1282 This option might be removed in a future release of G++. For maximum
1283 portability, you should structure your code so that it works with
1284 string constants that have type @code{const char *}.
1286 @item -fdollars-in-identifiers
1287 @opindex fdollars-in-identifiers
1288 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1289 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1290 @samp{$} by default on most target systems, but there are a few exceptions.)
1291 Traditional C allowed the character @samp{$} to form part of
1292 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1294 @item -fno-elide-constructors
1295 @opindex fno-elide-constructors
1296 The C++ standard allows an implementation to omit creating a temporary
1297 which is only used to initialize another object of the same type.
1298 Specifying this option disables that optimization, and forces G++ to
1299 call the copy constructor in all cases.
1301 @item -fno-enforce-eh-specs
1302 @opindex fno-enforce-eh-specs
1303 Don't check for violation of exception specifications at runtime. This
1304 option violates the C++ standard, but may be useful for reducing code
1305 size in production builds, much like defining @samp{NDEBUG}. The compiler
1306 will still optimize based on the exception specifications.
1308 @item -fexternal-templates
1309 @opindex fexternal-templates
1311 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1312 template instantiation; template instances are emitted or not according
1313 to the location of the template definition. @xref{Template
1314 Instantiation}, for more information.
1316 This option is deprecated.
1318 @item -falt-external-templates
1319 @opindex falt-external-templates
1320 Similar to @option{-fexternal-templates}, but template instances are
1321 emitted or not according to the place where they are first instantiated.
1322 @xref{Template Instantiation}, for more information.
1324 This option is deprecated.
1327 @itemx -fno-for-scope
1329 @opindex fno-for-scope
1330 If @option{-ffor-scope} is specified, the scope of variables declared in
1331 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1332 as specified by the C++ standard.
1333 If @option{-fno-for-scope} is specified, the scope of variables declared in
1334 a @i{for-init-statement} extends to the end of the enclosing scope,
1335 as was the case in old versions of G++, and other (traditional)
1336 implementations of C++.
1338 The default if neither flag is given to follow the standard,
1339 but to allow and give a warning for old-style code that would
1340 otherwise be invalid, or have different behavior.
1342 @item -fno-gnu-keywords
1343 @opindex fno-gnu-keywords
1344 Do not recognize @code{typeof} as a keyword, so that code can use this
1345 word as an identifier. You can use the keyword @code{__typeof__} instead.
1346 @option{-ansi} implies @option{-fno-gnu-keywords}.
1348 @item -fno-implicit-templates
1349 @opindex fno-implicit-templates
1350 Never emit code for non-inline templates which are instantiated
1351 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1352 @xref{Template Instantiation}, for more information.
1354 @item -fno-implicit-inline-templates
1355 @opindex fno-implicit-inline-templates
1356 Don't emit code for implicit instantiations of inline templates, either.
1357 The default is to handle inlines differently so that compiles with and
1358 without optimization will need the same set of explicit instantiations.
1360 @item -fno-implement-inlines
1361 @opindex fno-implement-inlines
1362 To save space, do not emit out-of-line copies of inline functions
1363 controlled by @samp{#pragma implementation}. This will cause linker
1364 errors if these functions are not inlined everywhere they are called.
1366 @item -fms-extensions
1367 @opindex fms-extensions
1368 Disable pedantic warnings about constructs used in MFC, such as implicit
1369 int and getting a pointer to member function via non-standard syntax.
1371 @item -fno-nonansi-builtins
1372 @opindex fno-nonansi-builtins
1373 Disable built-in declarations of functions that are not mandated by
1374 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1375 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1377 @item -fno-operator-names
1378 @opindex fno-operator-names
1379 Do not treat the operator name keywords @code{and}, @code{bitand},
1380 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1381 synonyms as keywords.
1383 @item -fno-optional-diags
1384 @opindex fno-optional-diags
1385 Disable diagnostics that the standard says a compiler does not need to
1386 issue. Currently, the only such diagnostic issued by G++ is the one for
1387 a name having multiple meanings within a class.
1390 @opindex fpermissive
1391 Downgrade messages about nonconformant code from errors to warnings. By
1392 default, G++ effectively sets @option{-pedantic-errors} without
1393 @option{-pedantic}; this option reverses that. This behavior and this
1394 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1398 Enable automatic template instantiation at link time. This option also
1399 implies @option{-fno-implicit-templates}. @xref{Template
1400 Instantiation}, for more information.
1404 Disable generation of information about every class with virtual
1405 functions for use by the C++ runtime type identification features
1406 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1407 of the language, you can save some space by using this flag. Note that
1408 exception handling uses the same information, but it will generate it as
1413 Emit statistics about front-end processing at the end of the compilation.
1414 This information is generally only useful to the G++ development team.
1416 @item -ftemplate-depth-@var{n}
1417 @opindex ftemplate-depth
1418 Set the maximum instantiation depth for template classes to @var{n}.
1419 A limit on the template instantiation depth is needed to detect
1420 endless recursions during template class instantiation. ANSI/ISO C++
1421 conforming programs must not rely on a maximum depth greater than 17.
1423 @item -fuse-cxa-atexit
1424 @opindex fuse-cxa-atexit
1425 Register destructors for objects with static storage duration with the
1426 @code{__cxa_atexit} function rather than the @code{atexit} function.
1427 This option is required for fully standards-compliant handling of static
1428 destructors, but will only work if your C library supports
1429 @code{__cxa_atexit}.
1433 Emit special relocations for vtables and virtual function references
1434 so that the linker can identify unused virtual functions and zero out
1435 vtable slots that refer to them. This is most useful with
1436 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1437 also discard the functions themselves.
1439 This optimization requires GNU as and GNU ld. Not all systems support
1440 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1444 Do not use weak symbol support, even if it is provided by the linker.
1445 By default, G++ will use weak symbols if they are available. This
1446 option exists only for testing, and should not be used by end-users;
1447 it will result in inferior code and has no benefits. This option may
1448 be removed in a future release of G++.
1452 Do not search for header files in the standard directories specific to
1453 C++, but do still search the other standard directories. (This option
1454 is used when building the C++ library.)
1457 In addition, these optimization, warning, and code generation options
1458 have meanings only for C++ programs:
1461 @item -fno-default-inline
1462 @opindex fno-default-inline
1463 Do not assume @samp{inline} for functions defined inside a class scope.
1464 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1465 functions will have linkage like inline functions; they just won't be
1468 @item -Wctor-dtor-privacy @r{(C++ only)}
1469 @opindex Wctor-dtor-privacy
1470 Warn when a class seems unusable, because all the constructors or
1471 destructors in a class are private and the class has no friends or
1472 public static member functions.
1474 @item -Wnon-virtual-dtor @r{(C++ only)}
1475 @opindex Wnon-virtual-dtor
1476 Warn when a class declares a non-virtual destructor that should probably
1477 be virtual, because it looks like the class will be used polymorphically.
1479 @item -Wreorder @r{(C++ only)}
1481 @cindex reordering, warning
1482 @cindex warning for reordering of member initializers
1483 Warn when the order of member initializers given in the code does not
1484 match the order in which they must be executed. For instance:
1490 A(): j (0), i (1) @{ @}
1494 Here the compiler will warn that the member initializers for @samp{i}
1495 and @samp{j} will be rearranged to match the declaration order of the
1499 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1502 @item -Weffc++ @r{(C++ only)}
1504 Warn about violations of the following style guidelines from Scott Meyers'
1505 @cite{Effective C++} book:
1509 Item 11: Define a copy constructor and an assignment operator for classes
1510 with dynamically allocated memory.
1513 Item 12: Prefer initialization to assignment in constructors.
1516 Item 14: Make destructors virtual in base classes.
1519 Item 15: Have @code{operator=} return a reference to @code{*this}.
1522 Item 23: Don't try to return a reference when you must return an object.
1526 and about violations of the following style guidelines from Scott Meyers'
1527 @cite{More Effective C++} book:
1531 Item 6: Distinguish between prefix and postfix forms of increment and
1532 decrement operators.
1535 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1539 If you use this option, you should be aware that the standard library
1540 headers do not obey all of these guidelines; you can use @samp{grep -v}
1541 to filter out those warnings.
1543 @item -Wno-deprecated @r{(C++ only)}
1544 @opindex Wno-deprecated
1545 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1547 @item -Wno-non-template-friend @r{(C++ only)}
1548 @opindex Wno-non-template-friend
1549 Disable warnings when non-templatized friend functions are declared
1550 within a template. With the advent of explicit template specification
1551 support in G++, if the name of the friend is an unqualified-id (i.e.,
1552 @samp{friend foo(int)}), the C++ language specification demands that the
1553 friend declare or define an ordinary, nontemplate function. (Section
1554 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1555 could be interpreted as a particular specialization of a templatized
1556 function. Because this non-conforming behavior is no longer the default
1557 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1558 check existing code for potential trouble spots, and is on by default.
1559 This new compiler behavior can be turned off with
1560 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1561 but disables the helpful warning.
1563 @item -Wold-style-cast @r{(C++ only)}
1564 @opindex Wold-style-cast
1565 Warn if an old-style (C-style) cast to a non-void type is used within
1566 a C++ program. The new-style casts (@samp{static_cast},
1567 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1568 unintended effects, and much easier to grep for.
1570 @item -Woverloaded-virtual @r{(C++ only)}
1571 @opindex Woverloaded-virtual
1572 @cindex overloaded virtual fn, warning
1573 @cindex warning for overloaded virtual fn
1574 Warn when a function declaration hides virtual functions from a
1575 base class. For example, in:
1582 struct B: public A @{
1587 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1595 will fail to compile.
1597 @item -Wno-pmf-conversions @r{(C++ only)}
1598 @opindex Wno-pmf-conversions
1599 Disable the diagnostic for converting a bound pointer to member function
1602 @item -Wsign-promo @r{(C++ only)}
1603 @opindex Wsign-promo
1604 Warn when overload resolution chooses a promotion from unsigned or
1605 enumeral type to a signed type over a conversion to an unsigned type of
1606 the same size. Previous versions of G++ would try to preserve
1607 unsignedness, but the standard mandates the current behavior.
1609 @item -Wsynth @r{(C++ only)}
1611 @cindex warning for synthesized methods
1612 @cindex synthesized methods, warning
1613 Warn when G++'s synthesis behavior does not match that of cfront. For
1619 A& operator = (int);
1629 In this example, G++ will synthesize a default @samp{A& operator =
1630 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1633 @node Objective-C Dialect Options
1634 @section Options Controlling Objective-C Dialect
1636 @cindex compiler options, Objective-C
1637 @cindex Objective-C options, command line
1638 @cindex options, Objective-C
1639 This section describes the command-line options that are only meaningful
1640 for Objective-C programs; but you can also use most of the GNU compiler
1641 options regardless of what language your program is in. For example,
1642 you might compile a file @code{some_class.m} like this:
1645 gcc -g -fgnu-runtime -O -c some_class.m
1649 In this example, only @option{-fgnu-runtime} is an option meant only for
1650 Objective-C programs; you can use the other options with any language
1653 Here is a list of options that are @emph{only} for compiling Objective-C
1657 @item -fconstant-string-class=@var{class-name}
1658 @opindex fconstant-string-class
1659 Use @var{class-name} as the name of the class to instantiate for each
1660 literal string specified with the syntax @code{@@"@dots{}"}. The default
1661 class name is @code{NXConstantString}.
1664 @opindex fgnu-runtime
1665 Generate object code compatible with the standard GNU Objective-C
1666 runtime. This is the default for most types of systems.
1668 @item -fnext-runtime
1669 @opindex fnext-runtime
1670 Generate output compatible with the NeXT runtime. This is the default
1671 for NeXT-based systems, including Darwin and Mac OS X@.
1675 Dump interface declarations for all classes seen in the source file to a
1676 file named @file{@var{sourcename}.decl}.
1679 @opindex Wno-protocol
1680 Do not warn if methods required by a protocol are not implemented
1681 in the class adopting it.
1685 Warn if a selector has multiple methods of different types defined.
1687 @c not documented because only avail via -Wp
1688 @c @item -print-objc-runtime-info
1692 @node Language Independent Options
1693 @section Options to Control Diagnostic Messages Formatting
1694 @cindex options to control diagnostics formatting
1695 @cindex diagnostic messages
1696 @cindex message formatting
1698 Traditionally, diagnostic messages have been formatted irrespective of
1699 the output device's aspect (e.g.@: its width, @dots{}). The options described
1700 below can be used to control the diagnostic messages formatting
1701 algorithm, e.g.@: how many characters per line, how often source location
1702 information should be reported. Right now, only the C++ front end can
1703 honor these options. However it is expected, in the near future, that
1704 the remaining front ends would be able to digest them correctly.
1707 @item -fmessage-length=@var{n}
1708 @opindex fmessage-length
1709 Try to format error messages so that they fit on lines of about @var{n}
1710 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1711 the front ends supported by GCC@. If @var{n} is zero, then no
1712 line-wrapping will be done; each error message will appear on a single
1715 @opindex fdiagnostics-show-location
1716 @item -fdiagnostics-show-location=once
1717 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1718 reporter to emit @emph{once} source location information; that is, in
1719 case the message is too long to fit on a single physical line and has to
1720 be wrapped, the source location won't be emitted (as prefix) again,
1721 over and over, in subsequent continuation lines. This is the default
1724 @item -fdiagnostics-show-location=every-line
1725 Only meaningful in line-wrapping mode. Instructs the diagnostic
1726 messages reporter to emit the same source location information (as
1727 prefix) for physical lines that result from the process of breaking
1728 a message which is too long to fit on a single line.
1732 @node Warning Options
1733 @section Options to Request or Suppress Warnings
1734 @cindex options to control warnings
1735 @cindex warning messages
1736 @cindex messages, warning
1737 @cindex suppressing warnings
1739 Warnings are diagnostic messages that report constructions which
1740 are not inherently erroneous but which are risky or suggest there
1741 may have been an error.
1743 You can request many specific warnings with options beginning @samp{-W},
1744 for example @option{-Wimplicit} to request warnings on implicit
1745 declarations. Each of these specific warning options also has a
1746 negative form beginning @samp{-Wno-} to turn off warnings;
1747 for example, @option{-Wno-implicit}. This manual lists only one of the
1748 two forms, whichever is not the default.
1750 The following options control the amount and kinds of warnings produced
1751 by GCC; for further, language-specific options also refer to
1752 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
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 Since @option{-Wformat} also checks for null format arguments for
1847 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
1849 @option{-Wformat} is included in @option{-Wall}. For more control over some
1850 aspects of format checking, the options @option{-Wno-format-y2k},
1851 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
1852 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
1853 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
1855 @item -Wno-format-y2k
1856 @opindex Wno-format-y2k
1857 If @option{-Wformat} is specified, do not warn about @code{strftime}
1858 formats which may yield only a two-digit year.
1860 @item -Wno-format-extra-args
1861 @opindex Wno-format-extra-args
1862 If @option{-Wformat} is specified, do not warn about excess arguments to a
1863 @code{printf} or @code{scanf} format function. The C standard specifies
1864 that such arguments are ignored.
1866 Where the unused arguments lie between used arguments that are
1867 specified with @samp{$} operand number specifications, normally
1868 warnings are still given, since the implementation could not know what
1869 type to pass to @code{va_arg} to skip the unused arguments. However,
1870 in the case of @code{scanf} formats, this option will suppress the
1871 warning if the unused arguments are all pointers, since the Single
1872 Unix Specification says that such unused arguments are allowed.
1874 @item -Wno-format-zero-length
1875 @opindex Wno-format-zero-length
1876 If @option{-Wformat} is specified, do not warn about zero-length formats.
1877 The C standard specifies that zero-length formats are allowed.
1879 @item -Wformat-nonliteral
1880 @opindex Wformat-nonliteral
1881 If @option{-Wformat} is specified, also warn if the format string is not a
1882 string literal and so cannot be checked, unless the format function
1883 takes its format arguments as a @code{va_list}.
1885 @item -Wformat-security
1886 @opindex Wformat-security
1887 If @option{-Wformat} is specified, also warn about uses of format
1888 functions that represent possible security problems. At present, this
1889 warns about calls to @code{printf} and @code{scanf} functions where the
1890 format string is not a string literal and there are no format arguments,
1891 as in @code{printf (foo);}. This may be a security hole if the format
1892 string came from untrusted input and contains @samp{%n}. (This is
1893 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1894 in future warnings may be added to @option{-Wformat-security} that are not
1895 included in @option{-Wformat-nonliteral}.)
1899 Enable @option{-Wformat} plus format checks not included in
1900 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1901 -Wformat-nonliteral -Wformat-security}.
1905 Enable warning about passing a null pointer for arguments marked as
1906 requiring a non-null value by the @code{nonnull} function attribute.
1908 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
1909 can be disabled with the @option{-Wno-nonnull} option.
1911 @item -Wimplicit-int
1912 @opindex Wimplicit-int
1913 Warn when a declaration does not specify a type.
1915 @item -Wimplicit-function-declaration
1916 @itemx -Werror-implicit-function-declaration
1917 @opindex Wimplicit-function-declaration
1918 @opindex Werror-implicit-function-declaration
1919 Give a warning (or error) whenever a function is used before being
1924 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1928 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1929 function with external linkage, returning int, taking either zero
1930 arguments, two, or three arguments of appropriate types.
1932 @item -Wmissing-braces
1933 @opindex Wmissing-braces
1934 Warn if an aggregate or union initializer is not fully bracketed. In
1935 the following example, the initializer for @samp{a} is not fully
1936 bracketed, but that for @samp{b} is fully bracketed.
1939 int a[2][2] = @{ 0, 1, 2, 3 @};
1940 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1944 @opindex Wparentheses
1945 Warn if parentheses are omitted in certain contexts, such
1946 as when there is an assignment in a context where a truth value
1947 is expected, or when operators are nested whose precedence people
1948 often get confused about.
1950 Also warn about constructions where there may be confusion to which
1951 @code{if} statement an @code{else} branch belongs. Here is an example of
1966 In C, every @code{else} branch belongs to the innermost possible @code{if}
1967 statement, which in this example is @code{if (b)}. This is often not
1968 what the programmer expected, as illustrated in the above example by
1969 indentation the programmer chose. When there is the potential for this
1970 confusion, GCC will issue a warning when this flag is specified.
1971 To eliminate the warning, add explicit braces around the innermost
1972 @code{if} statement so there is no way the @code{else} could belong to
1973 the enclosing @code{if}. The resulting code would look like this:
1989 @item -Wsequence-point
1990 @opindex Wsequence-point
1991 Warn about code that may have undefined semantics because of violations
1992 of sequence point rules in the C standard.
1994 The C standard defines the order in which expressions in a C program are
1995 evaluated in terms of @dfn{sequence points}, which represent a partial
1996 ordering between the execution of parts of the program: those executed
1997 before the sequence point, and those executed after it. These occur
1998 after the evaluation of a full expression (one which is not part of a
1999 larger expression), after the evaluation of the first operand of a
2000 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2001 function is called (but after the evaluation of its arguments and the
2002 expression denoting the called function), and in certain other places.
2003 Other than as expressed by the sequence point rules, the order of
2004 evaluation of subexpressions of an expression is not specified. All
2005 these rules describe only a partial order rather than a total order,
2006 since, for example, if two functions are called within one expression
2007 with no sequence point between them, the order in which the functions
2008 are called is not specified. However, the standards committee have
2009 ruled that function calls do not overlap.
2011 It is not specified when between sequence points modifications to the
2012 values of objects take effect. Programs whose behavior depends on this
2013 have undefined behavior; the C standard specifies that ``Between the
2014 previous and next sequence point an object shall have its stored value
2015 modified at most once by the evaluation of an expression. Furthermore,
2016 the prior value shall be read only to determine the value to be
2017 stored.''. If a program breaks these rules, the results on any
2018 particular implementation are entirely unpredictable.
2020 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2021 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2022 diagnosed by this option, and it may give an occasional false positive
2023 result, but in general it has been found fairly effective at detecting
2024 this sort of problem in programs.
2026 The present implementation of this option only works for C programs. A
2027 future implementation may also work for C++ programs.
2029 The C standard is worded confusingly, therefore there is some debate
2030 over the precise meaning of the sequence point rules in subtle cases.
2031 Links to discussions of the problem, including proposed formal
2032 definitions, may be found on our readings page, at
2033 @w{@uref{http://gcc.gnu.org/readings.html}}.
2036 @opindex Wreturn-type
2037 Warn whenever a function is defined with a return-type that defaults to
2038 @code{int}. Also warn about any @code{return} statement with no
2039 return-value in a function whose return-type is not @code{void}.
2041 For C++, a function without return type always produces a diagnostic
2042 message, even when @option{-Wno-return-type} is specified. The only
2043 exceptions are @samp{main} and functions defined in system headers.
2047 Warn whenever a @code{switch} statement has an index of enumeral type
2048 and lacks a @code{case} for one or more of the named codes of that
2049 enumeration. (The presence of a @code{default} label prevents this
2050 warning.) @code{case} labels outside the enumeration range also
2051 provoke warnings when this option is used.
2053 @item -Wswitch-default
2054 @opindex Wswitch-switch
2055 Warn whenever a @code{switch} statement does not have a @code{default}
2059 @opindex Wswitch-enum
2060 Warn whenever a @code{switch} statement has an index of enumeral type
2061 and lacks a @code{case} for one or more of the named codes of that
2062 enumeration. @code{case} labels outside the enumeration range also
2063 provoke warnings when this option is used.
2067 Warn if any trigraphs are encountered that might change the meaning of
2068 the program (trigraphs within comments are not warned about).
2070 @item -Wunused-function
2071 @opindex Wunused-function
2072 Warn whenever a static function is declared but not defined or a
2073 non\-inline static function is unused.
2075 @item -Wunused-label
2076 @opindex Wunused-label
2077 Warn whenever a label is declared but not used.
2079 To suppress this warning use the @samp{unused} attribute
2080 (@pxref{Variable Attributes}).
2082 @item -Wunused-parameter
2083 @opindex Wunused-parameter
2084 Warn whenever a function parameter is unused aside from its declaration.
2086 To suppress this warning use the @samp{unused} attribute
2087 (@pxref{Variable Attributes}).
2089 @item -Wunused-variable
2090 @opindex Wunused-variable
2091 Warn whenever a local variable or non-constant static variable is unused
2092 aside from its declaration
2094 To suppress this warning use the @samp{unused} attribute
2095 (@pxref{Variable Attributes}).
2097 @item -Wunused-value
2098 @opindex Wunused-value
2099 Warn whenever a statement computes a result that is explicitly not used.
2101 To suppress this warning cast the expression to @samp{void}.
2105 All all the above @option{-Wunused} options combined.
2107 In order to get a warning about an unused function parameter, you must
2108 either specify @samp{-W -Wunused} or separately specify
2109 @option{-Wunused-parameter}.
2111 @item -Wuninitialized
2112 @opindex Wuninitialized
2113 Warn if an automatic variable is used without first being initialized or
2114 if a variable may be clobbered by a @code{setjmp} call.
2116 These warnings are possible only in optimizing compilation,
2117 because they require data flow information that is computed only
2118 when optimizing. If you don't specify @option{-O}, you simply won't
2121 These warnings occur only for variables that are candidates for
2122 register allocation. Therefore, they do not occur for a variable that
2123 is declared @code{volatile}, or whose address is taken, or whose size
2124 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2125 structures, unions or arrays, even when they are in registers.
2127 Note that there may be no warning about a variable that is used only
2128 to compute a value that itself is never used, because such
2129 computations may be deleted by data flow analysis before the warnings
2132 These warnings are made optional because GCC is not smart
2133 enough to see all the reasons why the code might be correct
2134 despite appearing to have an error. Here is one example of how
2155 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2156 always initialized, but GCC doesn't know this. Here is
2157 another common case:
2162 if (change_y) save_y = y, y = new_y;
2164 if (change_y) y = save_y;
2169 This has no bug because @code{save_y} is used only if it is set.
2171 @cindex @code{longjmp} warnings
2172 This option also warns when a non-volatile automatic variable might be
2173 changed by a call to @code{longjmp}. These warnings as well are possible
2174 only in optimizing compilation.
2176 The compiler sees only the calls to @code{setjmp}. It cannot know
2177 where @code{longjmp} will be called; in fact, a signal handler could
2178 call it at any point in the code. As a result, you may get a warning
2179 even when there is in fact no problem because @code{longjmp} cannot
2180 in fact be called at the place which would cause a problem.
2182 Some spurious warnings can be avoided if you declare all the functions
2183 you use that never return as @code{noreturn}. @xref{Function
2186 @item -Wreorder @r{(C++ only)}
2188 @cindex reordering, warning
2189 @cindex warning for reordering of member initializers
2190 Warn when the order of member initializers given in the code does not
2191 match the order in which they must be executed. For instance:
2193 @item -Wunknown-pragmas
2194 @opindex Wunknown-pragmas
2195 @cindex warning for unknown pragmas
2196 @cindex unknown pragmas, warning
2197 @cindex pragmas, warning of unknown
2198 Warn when a #pragma directive is encountered which is not understood by
2199 GCC@. If this command line option is used, warnings will even be issued
2200 for unknown pragmas in system header files. This is not the case if
2201 the warnings were only enabled by the @option{-Wall} command line option.
2205 All of the above @samp{-W} options combined. This enables all the
2206 warnings about constructions that some users consider questionable, and
2207 that are easy to avoid (or modify to prevent the warning), even in
2208 conjunction with macros.
2211 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2212 Some of them warn about constructions that users generally do not
2213 consider questionable, but which occasionally you might wish to check
2214 for; others warn about constructions that are necessary or hard to avoid
2215 in some cases, and there is no simple way to modify the code to suppress
2221 Print extra warning messages for these events:
2225 A function can return either with or without a value. (Falling
2226 off the end of the function body is considered returning without
2227 a value.) For example, this function would evoke such a
2241 An expression-statement or the left-hand side of a comma expression
2242 contains no side effects.
2243 To suppress the warning, cast the unused expression to void.
2244 For example, an expression such as @samp{x[i,j]} will cause a warning,
2245 but @samp{x[(void)i,j]} will not.
2248 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2251 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2252 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2253 that of ordinary mathematical notation.
2256 Storage-class specifiers like @code{static} are not the first things in
2257 a declaration. According to the C Standard, this usage is obsolescent.
2260 The return type of a function has a type qualifier such as @code{const}.
2261 Such a type qualifier has no effect, since the value returned by a
2262 function is not an lvalue. (But don't warn about the GNU extension of
2263 @code{volatile void} return types. That extension will be warned about
2264 if @option{-pedantic} is specified.)
2267 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2271 A comparison between signed and unsigned values could produce an
2272 incorrect result when the signed value is converted to unsigned.
2273 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2276 An aggregate has a partly bracketed initializer.
2277 For example, the following code would evoke such a warning,
2278 because braces are missing around the initializer for @code{x.h}:
2281 struct s @{ int f, g; @};
2282 struct t @{ struct s h; int i; @};
2283 struct t x = @{ 1, 2, 3 @};
2287 An aggregate has an initializer which does not initialize all members.
2288 For example, the following code would cause such a warning, because
2289 @code{x.h} would be implicitly initialized to zero:
2292 struct s @{ int f, g, h; @};
2293 struct s x = @{ 3, 4 @};
2297 @item -Wno-div-by-zero
2298 @opindex Wno-div-by-zero
2299 @opindex Wdiv-by-zero
2300 Do not warn about compile-time integer division by zero. Floating point
2301 division by zero is not warned about, as it can be a legitimate way of
2302 obtaining infinities and NaNs.
2304 @item -Wsystem-headers
2305 @opindex Wsystem-headers
2306 @cindex warnings from system headers
2307 @cindex system headers, warnings from
2308 Print warning messages for constructs found in system header files.
2309 Warnings from system headers are normally suppressed, on the assumption
2310 that they usually do not indicate real problems and would only make the
2311 compiler output harder to read. Using this command line option tells
2312 GCC to emit warnings from system headers as if they occurred in user
2313 code. However, note that using @option{-Wall} in conjunction with this
2314 option will @emph{not} warn about unknown pragmas in system
2315 headers---for that, @option{-Wunknown-pragmas} must also be used.
2318 @opindex Wfloat-equal
2319 Warn if floating point values are used in equality comparisons.
2321 The idea behind this is that sometimes it is convenient (for the
2322 programmer) to consider floating-point values as approximations to
2323 infinitely precise real numbers. If you are doing this, then you need
2324 to compute (by analysing the code, or in some other way) the maximum or
2325 likely maximum error that the computation introduces, and allow for it
2326 when performing comparisons (and when producing output, but that's a
2327 different problem). In particular, instead of testing for equality, you
2328 would check to see whether the two values have ranges that overlap; and
2329 this is done with the relational operators, so equality comparisons are
2332 @item -Wtraditional @r{(C only)}
2333 @opindex Wtraditional
2334 Warn about certain constructs that behave differently in traditional and
2335 ISO C@. Also warn about ISO C constructs that have no traditional C
2336 equivalent, and/or problematic constructs which should be avoided.
2340 Macro parameters that appear within string literals in the macro body.
2341 In traditional C macro replacement takes place within string literals,
2342 but does not in ISO C@.
2345 In traditional C, some preprocessor directives did not exist.
2346 Traditional preprocessors would only consider a line to be a directive
2347 if the @samp{#} appeared in column 1 on the line. Therefore
2348 @option{-Wtraditional} warns about directives that traditional C
2349 understands but would ignore because the @samp{#} does not appear as the
2350 first character on the line. It also suggests you hide directives like
2351 @samp{#pragma} not understood by traditional C by indenting them. Some
2352 traditional implementations would not recognize @samp{#elif}, so it
2353 suggests avoiding it altogether.
2356 A function-like macro that appears without arguments.
2359 The unary plus operator.
2362 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2363 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2364 constants.) Note, these suffixes appear in macros defined in the system
2365 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2366 Use of these macros in user code might normally lead to spurious
2367 warnings, however gcc's integrated preprocessor has enough context to
2368 avoid warning in these cases.
2371 A function declared external in one block and then used after the end of
2375 A @code{switch} statement has an operand of type @code{long}.
2378 A non-@code{static} function declaration follows a @code{static} one.
2379 This construct is not accepted by some traditional C compilers.
2382 The ISO type of an integer constant has a different width or
2383 signedness from its traditional type. This warning is only issued if
2384 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2385 typically represent bit patterns, are not warned about.
2388 Usage of ISO string concatenation is detected.
2391 Initialization of automatic aggregates.
2394 Identifier conflicts with labels. Traditional C lacks a separate
2395 namespace for labels.
2398 Initialization of unions. If the initializer is zero, the warning is
2399 omitted. This is done under the assumption that the zero initializer in
2400 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2401 initializer warnings and relies on default initialization to zero in the
2405 Conversions by prototypes between fixed/floating point values and vice
2406 versa. The absence of these prototypes when compiling with traditional
2407 C would cause serious problems. This is a subset of the possible
2408 conversion warnings, for the full set use @option{-Wconversion}.
2413 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2415 @item -Wendif-labels
2416 @opindex Wendif-labels
2417 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2421 Warn whenever a local variable shadows another local variable, parameter or
2422 global variable or whenever a built-in function is shadowed.
2424 @item -Wlarger-than-@var{len}
2425 @opindex Wlarger-than
2426 Warn whenever an object of larger than @var{len} bytes is defined.
2428 @item -Wpointer-arith
2429 @opindex Wpointer-arith
2430 Warn about anything that depends on the ``size of'' a function type or
2431 of @code{void}. GNU C assigns these types a size of 1, for
2432 convenience in calculations with @code{void *} pointers and pointers
2435 @item -Wbad-function-cast @r{(C only)}
2436 @opindex Wbad-function-cast
2437 Warn whenever a function call is cast to a non-matching type.
2438 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2442 Warn whenever a pointer is cast so as to remove a type qualifier from
2443 the target type. For example, warn if a @code{const char *} is cast
2444 to an ordinary @code{char *}.
2447 @opindex Wcast-align
2448 Warn whenever a pointer is cast such that the required alignment of the
2449 target is increased. For example, warn if a @code{char *} is cast to
2450 an @code{int *} on machines where integers can only be accessed at
2451 two- or four-byte boundaries.
2453 @item -Wwrite-strings
2454 @opindex Wwrite-strings
2455 When compiling C, give string constants the type @code{const
2456 char[@var{length}]} so that
2457 copying the address of one into a non-@code{const} @code{char *}
2458 pointer will get a warning; when compiling C++, warn about the
2459 deprecated conversion from string constants to @code{char *}.
2460 These warnings will help you find at
2461 compile time code that can try to write into a string constant, but
2462 only if you have been very careful about using @code{const} in
2463 declarations and prototypes. Otherwise, it will just be a nuisance;
2464 this is why we did not make @option{-Wall} request these warnings.
2467 @opindex Wconversion
2468 Warn if a prototype causes a type conversion that is different from what
2469 would happen to the same argument in the absence of a prototype. This
2470 includes conversions of fixed point to floating and vice versa, and
2471 conversions changing the width or signedness of a fixed point argument
2472 except when the same as the default promotion.
2474 Also, warn if a negative integer constant expression is implicitly
2475 converted to an unsigned type. For example, warn about the assignment
2476 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2477 casts like @code{(unsigned) -1}.
2479 @item -Wsign-compare
2480 @opindex Wsign-compare
2481 @cindex warning for comparison of signed and unsigned values
2482 @cindex comparison of signed and unsigned values, warning
2483 @cindex signed and unsigned values, comparison warning
2484 Warn when a comparison between signed and unsigned values could produce
2485 an incorrect result when the signed value is converted to unsigned.
2486 This warning is also enabled by @option{-W}; to get the other warnings
2487 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2489 @item -Waggregate-return
2490 @opindex Waggregate-return
2491 Warn if any functions that return structures or unions are defined or
2492 called. (In languages where you can return an array, this also elicits
2495 @item -Wstrict-prototypes @r{(C only)}
2496 @opindex Wstrict-prototypes
2497 Warn if a function is declared or defined without specifying the
2498 argument types. (An old-style function definition is permitted without
2499 a warning if preceded by a declaration which specifies the argument
2502 @item -Wmissing-prototypes @r{(C only)}
2503 @opindex Wmissing-prototypes
2504 Warn if a global function is defined without a previous prototype
2505 declaration. This warning is issued even if the definition itself
2506 provides a prototype. The aim is to detect global functions that fail
2507 to be declared in header files.
2509 @item -Wmissing-declarations
2510 @opindex Wmissing-declarations
2511 Warn if a global function is defined without a previous declaration.
2512 Do so even if the definition itself provides a prototype.
2513 Use this option to detect global functions that are not declared in
2516 @item -Wmissing-noreturn
2517 @opindex Wmissing-noreturn
2518 Warn about functions which might be candidates for attribute @code{noreturn}.
2519 Note these are only possible candidates, not absolute ones. Care should
2520 be taken to manually verify functions actually do not ever return before
2521 adding the @code{noreturn} attribute, otherwise subtle code generation
2522 bugs could be introduced. You will not get a warning for @code{main} in
2523 hosted C environments.
2525 @item -Wmissing-format-attribute
2526 @opindex Wmissing-format-attribute
2528 If @option{-Wformat} is enabled, also warn about functions which might be
2529 candidates for @code{format} attributes. Note these are only possible
2530 candidates, not absolute ones. GCC will guess that @code{format}
2531 attributes might be appropriate for any function that calls a function
2532 like @code{vprintf} or @code{vscanf}, but this might not always be the
2533 case, and some functions for which @code{format} attributes are
2534 appropriate may not be detected. This option has no effect unless
2535 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2537 @item -Wno-multichar
2538 @opindex Wno-multichar
2540 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2541 Usually they indicate a typo in the user's code, as they have
2542 implementation-defined values, and should not be used in portable code.
2544 @item -Wno-deprecated-declarations
2545 @opindex Wno-deprecated-declarations
2546 Do not warn about uses of functions, variables, and types marked as
2547 deprecated by using the @code{deprecated} attribute.
2548 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2549 @pxref{Type Attributes}.)
2553 Warn if a structure is given the packed attribute, but the packed
2554 attribute has no effect on the layout or size of the structure.
2555 Such structures may be mis-aligned for little benefit. For
2556 instance, in this code, the variable @code{f.x} in @code{struct bar}
2557 will be misaligned even though @code{struct bar} does not itself
2558 have the packed attribute:
2565 @} __attribute__((packed));
2575 Warn if padding is included in a structure, either to align an element
2576 of the structure or to align the whole structure. Sometimes when this
2577 happens it is possible to rearrange the fields of the structure to
2578 reduce the padding and so make the structure smaller.
2580 @item -Wredundant-decls
2581 @opindex Wredundant-decls
2582 Warn if anything is declared more than once in the same scope, even in
2583 cases where multiple declaration is valid and changes nothing.
2585 @item -Wnested-externs @r{(C only)}
2586 @opindex Wnested-externs
2587 Warn if an @code{extern} declaration is encountered within a function.
2589 @item -Wunreachable-code
2590 @opindex Wunreachable-code
2591 Warn if the compiler detects that code will never be executed.
2593 This option is intended to warn when the compiler detects that at
2594 least a whole line of source code will never be executed, because
2595 some condition is never satisfied or because it is after a
2596 procedure that never returns.
2598 It is possible for this option to produce a warning even though there
2599 are circumstances under which part of the affected line can be executed,
2600 so care should be taken when removing apparently-unreachable code.
2602 For instance, when a function is inlined, a warning may mean that the
2603 line is unreachable in only one inlined copy of the function.
2605 This option is not made part of @option{-Wall} because in a debugging
2606 version of a program there is often substantial code which checks
2607 correct functioning of the program and is, hopefully, unreachable
2608 because the program does work. Another common use of unreachable
2609 code is to provide behavior which is selectable at compile-time.
2613 Warn if a function can not be inlined and it was declared as inline.
2617 @opindex Wno-long-long
2618 Warn if @samp{long long} type is used. This is default. To inhibit
2619 the warning messages, use @option{-Wno-long-long}. Flags
2620 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2621 only when @option{-pedantic} flag is used.
2623 @item -Wdisabled-optimization
2624 @opindex Wdisabled-optimization
2625 Warn if a requested optimization pass is disabled. This warning does
2626 not generally indicate that there is anything wrong with your code; it
2627 merely indicates that GCC's optimizers were unable to handle the code
2628 effectively. Often, the problem is that your code is too big or too
2629 complex; GCC will refuse to optimize programs when the optimization
2630 itself is likely to take inordinate amounts of time.
2634 Make all warnings into errors.
2637 @node Debugging Options
2638 @section Options for Debugging Your Program or GCC
2639 @cindex options, debugging
2640 @cindex debugging information options
2642 GCC has various special options that are used for debugging
2643 either your program or GCC:
2648 Produce debugging information in the operating system's native format
2649 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2652 On most systems that use stabs format, @option{-g} enables use of extra
2653 debugging information that only GDB can use; this extra information
2654 makes debugging work better in GDB but will probably make other debuggers
2656 refuse to read the program. If you want to control for certain whether
2657 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2658 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2659 or @option{-gvms} (see below).
2661 Unlike most other C compilers, GCC allows you to use @option{-g} with
2662 @option{-O}. The shortcuts taken by optimized code may occasionally
2663 produce surprising results: some variables you declared may not exist
2664 at all; flow of control may briefly move where you did not expect it;
2665 some statements may not be executed because they compute constant
2666 results or their values were already at hand; some statements may
2667 execute in different places because they were moved out of loops.
2669 Nevertheless it proves possible to debug optimized output. This makes
2670 it reasonable to use the optimizer for programs that might have bugs.
2672 The following options are useful when GCC is generated with the
2673 capability for more than one debugging format.
2677 Produce debugging information for use by GDB@. This means to use the
2678 most expressive format available (DWARF 2, stabs, or the native format
2679 if neither of those are supported), including GDB extensions if at all
2684 Produce debugging information in stabs format (if that is supported),
2685 without GDB extensions. This is the format used by DBX on most BSD
2686 systems. On MIPS, Alpha and System V Release 4 systems this option
2687 produces stabs debugging output which is not understood by DBX or SDB@.
2688 On System V Release 4 systems this option requires the GNU assembler.
2692 Produce debugging information in stabs format (if that is supported),
2693 using GNU extensions understood only by the GNU debugger (GDB)@. The
2694 use of these extensions is likely to make other debuggers crash or
2695 refuse to read the program.
2699 Produce debugging information in COFF format (if that is supported).
2700 This is the format used by SDB on most System V systems prior to
2705 Produce debugging information in XCOFF format (if that is supported).
2706 This is the format used by the DBX debugger on IBM RS/6000 systems.
2710 Produce debugging information in XCOFF format (if that is supported),
2711 using GNU extensions understood only by the GNU debugger (GDB)@. The
2712 use of these extensions is likely to make other debuggers crash or
2713 refuse to read the program, and may cause assemblers other than the GNU
2714 assembler (GAS) to fail with an error.
2718 Produce debugging information in DWARF version 1 format (if that is
2719 supported). This is the format used by SDB on most System V Release 4
2724 Produce debugging information in DWARF version 1 format (if that is
2725 supported), using GNU extensions understood only by the GNU debugger
2726 (GDB)@. The use of these extensions is likely to make other debuggers
2727 crash or refuse to read the program.
2731 Produce debugging information in DWARF version 2 format (if that is
2732 supported). This is the format used by DBX on IRIX 6.
2736 Produce debugging information in VMS debug format (if that is
2737 supported). This is the format used by DEBUG on VMS systems.
2740 @itemx -ggdb@var{level}
2741 @itemx -gstabs@var{level}
2742 @itemx -gcoff@var{level}
2743 @itemx -gxcoff@var{level}
2744 @itemx -gvms@var{level}
2745 Request debugging information and also use @var{level} to specify how
2746 much information. The default level is 2.
2748 Level 1 produces minimal information, enough for making backtraces in
2749 parts of the program that you don't plan to debug. This includes
2750 descriptions of functions and external variables, but no information
2751 about local variables and no line numbers.
2753 Level 3 includes extra information, such as all the macro definitions
2754 present in the program. Some debuggers support macro expansion when
2755 you use @option{-g3}.
2757 Note that in order to avoid confusion between DWARF1 debug level 2,
2758 and DWARF2, neither @option{-gdwarf} nor @option{-gdwarf-2} accept
2759 a concatenated debug level. Instead use an additional @option{-g@var{level}}
2760 option to change the debug level for DWARF1 or DWARF2.
2765 Generate extra code to write profile information suitable for the
2766 analysis program @code{prof}. You must use this option when compiling
2767 the source files you want data about, and you must also use it when
2770 @cindex @code{gprof}
2773 Generate extra code to write profile information suitable for the
2774 analysis program @code{gprof}. You must use this option when compiling
2775 the source files you want data about, and you must also use it when
2781 Generate extra code to write profile information for basic blocks, which will
2782 record the number of times each basic block is executed, the basic block start
2783 address, and the function name containing the basic block. If @option{-g} is
2784 used, the line number and filename of the start of the basic block will also be
2785 recorded. If not overridden by the machine description, the default action is
2786 to append to the text file @file{bb.out}.
2788 This data could be analyzed by a program like @code{tcov}. Note,
2789 however, that the format of the data is not what @code{tcov} expects.
2790 Eventually GNU @code{gprof} should be extended to process this data.
2794 Makes the compiler print out each function name as it is compiled, and
2795 print some statistics about each pass when it finishes.
2798 @opindex ftime-report
2799 Makes the compiler print some statistics about the time consumed by each
2800 pass when it finishes.
2803 @opindex fmem-report
2804 Makes the compiler print some statistics about permanent memory
2805 allocation when it finishes.
2807 @item -fprofile-arcs
2808 @opindex fprofile-arcs
2809 Instrument @dfn{arcs} during compilation to generate coverage data
2810 or for profile-directed block ordering. During execution the program
2811 records how many times each branch is executed and how many times it is
2812 taken. When the compiled program exits it saves this data to a file
2813 called @file{@var{sourcename}.da} for each source file.
2815 For profile-directed block ordering, compile the program with
2816 @option{-fprofile-arcs} plus optimization and code generation options,
2817 generate the arc profile information by running the program on a
2818 selected workload, and then compile the program again with the same
2819 optimization and code generation options plus
2820 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2821 Control Optimization}).
2823 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2824 when it is used with the @option{-ftest-coverage} option. GCC
2825 supports two methods of determining code coverage: the options that
2826 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2827 write information to text files. The options that support @code{gcov}
2828 do not need to instrument every arc in the program, so a program compiled
2829 with them runs faster than a program compiled with @option{-a}, which
2830 adds instrumentation code to every basic block in the program. The
2831 tradeoff: since @code{gcov} does not have execution counts for all
2832 branches, it must start with the execution counts for the instrumented
2833 branches, and then iterate over the program flow graph until the entire
2834 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2835 a program which uses information from @option{-a} and @option{-ax}.
2837 With @option{-fprofile-arcs}, for each function of your program GCC
2838 creates a program flow graph, then finds a spanning tree for the graph.
2839 Only arcs that are not on the spanning tree have to be instrumented: the
2840 compiler adds code to count the number of times that these arcs are
2841 executed. When an arc is the only exit or only entrance to a block, the
2842 instrumentation code can be added to the block; otherwise, a new basic
2843 block must be created to hold the instrumentation code.
2845 This option makes it possible to estimate branch probabilities and to
2846 calculate basic block execution counts. In general, basic block
2847 execution counts as provided by @option{-a} do not give enough
2848 information to estimate all branch probabilities.
2851 @item -ftest-coverage
2852 @opindex ftest-coverage
2853 Create data files for the @code{gcov} code-coverage utility
2854 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2855 The data file names begin with the name of your source file:
2858 @item @var{sourcename}.bb
2859 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2860 associate basic block execution counts with line numbers.
2862 @item @var{sourcename}.bbg
2863 A list of all arcs in the program flow graph. This allows @code{gcov}
2864 to reconstruct the program flow graph, so that it can compute all basic
2865 block and arc execution counts from the information in the
2866 @code{@var{sourcename}.da} file.
2869 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2870 option adds instrumentation to the program, which then writes
2871 execution counts to another data file:
2874 @item @var{sourcename}.da
2875 Runtime arc execution counts, used in conjunction with the arc
2876 information in the file @code{@var{sourcename}.bbg}.
2879 Coverage data will map better to the source files if
2880 @option{-ftest-coverage} is used without optimization.
2882 @item -d@var{letters}
2884 Says to make debugging dumps during compilation at times specified by
2885 @var{letters}. This is used for debugging the compiler. The file names
2886 for most of the dumps are made by appending a pass number and a word to
2887 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2888 Here are the possible letters for use in @var{letters}, and their meanings:
2893 Annotate the assembler output with miscellaneous debugging information.
2896 Dump after computing branch probabilities, to @file{@var{file}.15.bp}.
2899 Dump after block reordering, to @file{@var{file}.30.bbro}.
2902 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
2905 Dump after the first if conversion, to the file @file{@var{file}.18.ce}.
2908 Dump after delayed branch scheduling, to @file{@var{file}.32.dbr}.
2911 Dump all macro definitions, at the end of preprocessing, in addition to
2915 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2916 @file{@var{file}.07.ussa}.
2919 Dump after the second if conversion, to @file{@var{file}.27.ce2}.
2922 Dump after life analysis, to @file{@var{file}.16.life}.
2925 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.10.addressof}.
2928 Dump after global register allocation, to @file{@var{file}.22.greg}.
2931 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2934 Dump after reg-to-stack conversion, to @file{@var{file}.29.stack}.
2937 Dump after post-reload optimizations, to @file{@var{file}.23.postreload}.
2940 Dump after GCSE, to @file{@var{file}.11.gcse}.
2943 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2946 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2949 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
2952 Dump after local register allocation, to @file{@var{file}.21.lreg}.
2955 Dump after loop optimization, to @file{@var{file}.12.loop}.
2958 Dump after performing the machine dependent reorganisation pass, to
2959 @file{@var{file}.31.mach}.
2962 Dump after register renumbering, to @file{@var{file}.26.rnreg}.
2965 Dump after the register move pass, to @file{@var{file}.19.regmove}.
2968 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2971 Dump after the second scheduling pass, to @file{@var{file}.28.sched2}.
2974 Dump after CSE (including the jump optimization that sometimes follows
2975 CSE), to @file{@var{file}.09.cse}.
2978 Dump after the first scheduling pass, to @file{@var{file}.20.sched}.
2981 Dump after the second CSE pass (including the jump optimization that
2982 sometimes follows CSE), to @file{@var{file}.13.cse2}.
2985 Dump after null pointer ellimination pass ti @file{@var{file}.08.null}.
2988 Dump after the second flow pass, to @file{@var{file}.24.flow2}.
2991 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2994 Dump after the peephole pass, to @file{@var{file}.25.peephole2}.
2997 Produce all the dumps listed above.
3000 Print statistics on memory usage, at the end of the run, to
3004 Annotate the assembler output with a comment indicating which
3005 pattern and alternative was used. The length of each instruction is
3009 Dump the RTL in the assembler output as a comment before each instruction.
3010 Also turns on @option{-dp} annotation.
3013 For each of the other indicated dump files (except for
3014 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
3015 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3018 Just generate RTL for a function instead of compiling it. Usually used
3022 Dump debugging information during parsing, to standard error.
3025 @item -fdump-unnumbered
3026 @opindex fdump-unnumbered
3027 When doing debugging dumps (see @option{-d} option above), suppress instruction
3028 numbers and line number note output. This makes it more feasible to
3029 use diff on debugging dumps for compiler invocations with different
3030 options, in particular with and without @option{-g}.
3032 @item -fdump-translation-unit @r{(C and C++ only)}
3033 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3034 @opindex fdump-translation-unit
3035 Dump a representation of the tree structure for the entire translation
3036 unit to a file. The file name is made by appending @file{.tu} to the
3037 source file name. If the @samp{-@var{options}} form is used, @var{options}
3038 controls the details of the dump as described for the
3039 @option{-fdump-tree} options.
3041 @item -fdump-class-hierarchy @r{(C++ only)}
3042 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3043 @opindex fdump-class-hierarchy
3044 Dump a representation of each class's hierarchy and virtual function
3045 table layout to a file. The file name is made by appending @file{.class}
3046 to the source file name. If the @samp{-@var{options}} form is used,
3047 @var{options} controls the details of the dump as described for the
3048 @option{-fdump-tree} options.
3050 @item -fdump-tree-@var{switch} @r{(C++ only)}
3051 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3053 Control the dumping at various stages of processing the intermediate
3054 language tree to a file. The file name is generated by appending a switch
3055 specific suffix to the source file name. If the @samp{-@var{options}}
3056 form is used, @var{options} is a list of @samp{-} separated options that
3057 control the details of the dump. Not all options are applicable to all
3058 dumps, those which are not meaningful will be ignored. The following
3059 options are available
3063 Print the address of each node. Usually this is not meaningful as it
3064 changes according to the environment and source file. Its primary use
3065 is for tying up a dump file with a debug environment.
3067 Inhibit dumping of members of a scope or body of a function merely
3068 because that scope has been reached. Only dump such items when they
3069 are directly reachable by some other path.
3071 Turn on all options.
3074 The following tree dumps are possible:
3077 Dump before any tree based optimization, to @file{@var{file}.original}.
3079 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3081 Dump after function inlining, to @file{@var{file}.inlined}.
3086 Store the usual ``temporary'' intermediate files permanently; place them
3087 in the current directory and name them based on the source file. Thus,
3088 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3089 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3090 preprocessed @file{foo.i} output file even though the compiler now
3091 normally uses an integrated preprocessor.
3095 Report the CPU time taken by each subprocess in the compilation
3096 sequence. For C source files, this is the compiler proper and assembler
3097 (plus the linker if linking is done). The output looks like this:
3104 The first number on each line is the ``user time,'' that is time spent
3105 executing the program itself. The second number is ``system time,''
3106 time spent executing operating system routines on behalf of the program.
3107 Both numbers are in seconds.
3109 @item -print-file-name=@var{library}
3110 @opindex print-file-name
3111 Print the full absolute name of the library file @var{library} that
3112 would be used when linking---and don't do anything else. With this
3113 option, GCC does not compile or link anything; it just prints the
3116 @item -print-multi-directory
3117 @opindex print-multi-directory
3118 Print the directory name corresponding to the multilib selected by any
3119 other switches present in the command line. This directory is supposed
3120 to exist in @env{GCC_EXEC_PREFIX}.
3122 @item -print-multi-lib
3123 @opindex print-multi-lib
3124 Print the mapping from multilib directory names to compiler switches
3125 that enable them. The directory name is separated from the switches by
3126 @samp{;}, and each switch starts with an @samp{@@} instead of the
3127 @samp{-}, without spaces between multiple switches. This is supposed to
3128 ease shell-processing.
3130 @item -print-prog-name=@var{program}
3131 @opindex print-prog-name
3132 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3134 @item -print-libgcc-file-name
3135 @opindex print-libgcc-file-name
3136 Same as @option{-print-file-name=libgcc.a}.
3138 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3139 but you do want to link with @file{libgcc.a}. You can do
3142 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3145 @item -print-search-dirs
3146 @opindex print-search-dirs
3147 Print the name of the configured installation directory and a list of
3148 program and library directories gcc will search---and don't do anything else.
3150 This is useful when gcc prints the error message
3151 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3152 To resolve this you either need to put @file{cpp0} and the other compiler
3153 components where gcc expects to find them, or you can set the environment
3154 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3155 Don't forget the trailing '/'.
3156 @xref{Environment Variables}.
3159 @opindex dumpmachine
3160 Print the compiler's target machine (for example,
3161 @samp{i686-pc-linux-gnu})---and don't do anything else.
3164 @opindex dumpversion
3165 Print the compiler version (for example, @samp{3.0})---and don't do
3170 Print the compiler's built-in specs---and don't do anything else. (This
3171 is used when GCC itself is being built.) @xref{Spec Files}.
3174 @node Optimize Options
3175 @section Options That Control Optimization
3176 @cindex optimize options
3177 @cindex options, optimization
3179 These options control various sorts of optimizations:
3186 Optimize. Optimizing compilation takes somewhat more time, and a lot
3187 more memory for a large function.
3189 Without @option{-O}, the compiler's goal is to reduce the cost of
3190 compilation and to make debugging produce the expected results.
3191 Statements are independent: if you stop the program with a breakpoint
3192 between statements, you can then assign a new value to any variable or
3193 change the program counter to any other statement in the function and
3194 get exactly the results you would expect from the source code.
3196 With @option{-O}, the compiler tries to reduce code size and execution
3197 time, without performing any optimizations that take a great deal of
3202 Optimize even more. GCC performs nearly all supported optimizations
3203 that do not involve a space-speed tradeoff. The compiler does not
3204 perform loop unrolling or function inlining when you specify @option{-O2}.
3205 As compared to @option{-O}, this option increases both compilation time
3206 and the performance of the generated code.
3208 @option{-O2} turns on all optional optimizations except for loop
3209 unrolling, function inlining, and register renaming. It also turns on
3210 the @option{-fforce-mem} and @option{-fstrict-aliasing} option on all
3211 machines and frame pointer elimination on machines where doing so does
3212 not interfere with debugging.
3214 Please note the warning under @option{-fgcse} about
3215 invoking @option{-O2} on programs that use computed gotos.
3219 Optimize yet more. @option{-O3} turns on all optimizations specified by
3220 @option{-O2} and also turns on the @option{-finline-functions} and
3221 @option{-frename-registers} options.
3229 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3230 do not typically increase code size. It also performs further
3231 optimizations designed to reduce code size.
3233 If you use multiple @option{-O} options, with or without level numbers,
3234 the last such option is the one that is effective.
3237 Options of the form @option{-f@var{flag}} specify machine-independent
3238 flags. Most flags have both positive and negative forms; the negative
3239 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3240 only one of the forms is listed---the one which is not the default.
3241 You can figure out the other form by either removing @samp{no-} or
3246 @opindex ffloat-store
3247 Do not store floating point variables in registers, and inhibit other
3248 options that might change whether a floating point value is taken from a
3251 @cindex floating point precision
3252 This option prevents undesirable excess precision on machines such as
3253 the 68000 where the floating registers (of the 68881) keep more
3254 precision than a @code{double} is supposed to have. Similarly for the
3255 x86 architecture. For most programs, the excess precision does only
3256 good, but a few programs rely on the precise definition of IEEE floating
3257 point. Use @option{-ffloat-store} for such programs, after modifying
3258 them to store all pertinent intermediate computations into variables.
3260 @item -fno-default-inline
3261 @opindex fno-default-inline
3262 Do not make member functions inline by default merely because they are
3263 defined inside the class scope (C++ only). Otherwise, when you specify
3264 @w{@option{-O}}, member functions defined inside class scope are compiled
3265 inline by default; i.e., you don't need to add @samp{inline} in front of
3266 the member function name.
3268 @item -fno-defer-pop
3269 @opindex fno-defer-pop
3270 Always pop the arguments to each function call as soon as that function
3271 returns. For machines which must pop arguments after a function call,
3272 the compiler normally lets arguments accumulate on the stack for several
3273 function calls and pops them all at once.
3277 Force memory operands to be copied into registers before doing
3278 arithmetic on them. This produces better code by making all memory
3279 references potential common subexpressions. When they are not common
3280 subexpressions, instruction combination should eliminate the separate
3281 register-load. The @option{-O2} option turns on this option.
3284 @opindex fforce-addr
3285 Force memory address constants to be copied into registers before
3286 doing arithmetic on them. This may produce better code just as
3287 @option{-fforce-mem} may.
3289 @item -fomit-frame-pointer
3290 @opindex fomit-frame-pointer
3291 Don't keep the frame pointer in a register for functions that
3292 don't need one. This avoids the instructions to save, set up and
3293 restore frame pointers; it also makes an extra register available
3294 in many functions. @strong{It also makes debugging impossible on
3297 On some machines, such as the VAX, this flag has no effect, because
3298 the standard calling sequence automatically handles the frame pointer
3299 and nothing is saved by pretending it doesn't exist. The
3300 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3301 whether a target machine supports this flag. @xref{Registers,,Register
3302 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3304 @item -foptimize-sibling-calls
3305 @opindex foptimize-sibling-calls
3306 Optimize sibling and tail recursive calls.
3310 This option generates traps for signed overflow on addition, subtraction,
3311 multiplication operations.
3315 Don't pay attention to the @code{inline} keyword. Normally this option
3316 is used to keep the compiler from expanding any functions inline.
3317 Note that if you are not optimizing, no functions can be expanded inline.
3319 @item -finline-functions
3320 @opindex finline-functions
3321 Integrate all simple functions into their callers. The compiler
3322 heuristically decides which functions are simple enough to be worth
3323 integrating in this way.
3325 If all calls to a given function are integrated, and the function is
3326 declared @code{static}, then the function is normally not output as
3327 assembler code in its own right.
3329 @item -finline-limit=@var{n}
3330 @opindex finline-limit
3331 By default, gcc limits the size of functions that can be inlined. This flag
3332 allows the control of this limit for functions that are explicitly marked as
3333 inline (ie marked with the inline keyword or defined within the class
3334 definition in c++). @var{n} is the size of functions that can be inlined in
3335 number of pseudo instructions (not counting parameter handling). The default
3336 value of @var{n} is 600.
3337 Increasing this value can result in more inlined code at
3338 the cost of compilation time and memory consumption. Decreasing usually makes
3339 the compilation faster and less code will be inlined (which presumably
3340 means slower programs). This option is particularly useful for programs that
3341 use inlining heavily such as those based on recursive templates with C++.
3343 @emph{Note:} pseudo instruction represents, in this particular context, an
3344 abstract measurement of function's size. In no way, it represents a count
3345 of assembly instructions and as such its exact meaning might change from one
3346 release to an another.
3348 @item -fkeep-inline-functions
3349 @opindex fkeep-inline-functions
3350 Even if all calls to a given function are integrated, and the function
3351 is declared @code{static}, nevertheless output a separate run-time
3352 callable version of the function. This switch does not affect
3353 @code{extern inline} functions.
3355 @item -fkeep-static-consts
3356 @opindex fkeep-static-consts
3357 Emit variables declared @code{static const} when optimization isn't turned
3358 on, even if the variables aren't referenced.
3360 GCC enables this option by default. If you want to force the compiler to
3361 check if the variable was referenced, regardless of whether or not
3362 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3364 @item -fmerge-constants
3365 Attempt to merge identical constants (string constants and floating point
3366 constants) accross compilation units.
3368 This option is default for optimized compilation if assembler and linker
3369 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3371 @item -fmerge-all-constants
3372 Attempt to merge identical constants and identical variables.
3374 This option implies @option{-fmerge-constants}. In addition to
3375 @option{-fmerge-constants} this considers e.g. even constant initialized
3376 arrays or initialized constant variables with integral or floating point
3377 types. Languages like C or C++ require each non-automatic variable to
3378 have distinct location, so using this option will result in non-conforming
3381 @item -fno-function-cse
3382 @opindex fno-function-cse
3383 Do not put function addresses in registers; make each instruction that
3384 calls a constant function contain the function's address explicitly.
3386 This option results in less efficient code, but some strange hacks
3387 that alter the assembler output may be confused by the optimizations
3388 performed when this option is not used.
3392 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3393 @option{-fno-trapping-math}.
3395 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3397 This option should never be turned on by any @option{-O} option since
3398 it can result in incorrect output for programs which depend on
3399 an exact implementation of IEEE or ISO rules/specifications for
3402 @item -fno-math-errno
3403 @opindex fno-math-errno
3404 Do not set ERRNO after calling math functions that are executed
3405 with a single instruction, e.g., sqrt. A program that relies on
3406 IEEE exceptions for math error handling may want to use this flag
3407 for speed while maintaining IEEE arithmetic compatibility.
3409 This option should never be turned on by any @option{-O} option since
3410 it can result in incorrect output for programs which depend on
3411 an exact implementation of IEEE or ISO rules/specifications for
3414 The default is @option{-fmath-errno}.
3416 @item -funsafe-math-optimizations
3417 @opindex funsafe-math-optimizations
3418 Allow optimizations for floating-point arithmetic that (a) assume
3419 that arguments and results are valid and (b) may violate IEEE or
3420 ANSI standards. When used at link-time, it may include libraries
3421 or startup files that change the default FPU control word or other
3422 similar optimizations.
3424 This option should never be turned on by any @option{-O} option since
3425 it can result in incorrect output for programs which depend on
3426 an exact implementation of IEEE or ISO rules/specifications for
3429 The default is @option{-fno-unsafe-math-optimizations}.
3431 @item -fno-trapping-math
3432 @opindex fno-trapping-math
3433 Compile code assuming that floating-point operations cannot generate
3434 user-visible traps. Setting this option may allow faster code
3435 if one relies on ``non-stop'' IEEE arithmetic, for example.
3437 This option should never be turned on by any @option{-O} option since
3438 it can result in incorrect output for programs which depend on
3439 an exact implementation of IEEE or ISO rules/specifications for
3442 The default is @option{-ftrapping-math}.
3444 @item -fno-zero-initialized-in-bss
3445 @opindex fno-zero-initialized-in-bss
3446 If the target supports a BSS section, GCC by default puts variables that
3447 are initialized to zero into BSS@. This can save space in the resulting
3450 This option turns off this behavior because some programs explicitly
3451 rely on variables going to the data section. E.g., so that the
3452 resulting executable can find the beginning of that section and/or make
3453 assumptions based on that.
3455 The default is @option{-fzero-initialized-in-bss}.
3458 The following options control specific optimizations. The @option{-O2}
3459 option turns on all of these optimizations except @option{-funroll-loops}
3460 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3461 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3462 but specific machines may handle it differently.
3464 You can use the following flags in the rare cases when ``fine-tuning''
3465 of optimizations to be performed is desired.
3467 Not all of the optimizations performed by GCC have @option{-f} options
3471 @item -fstrength-reduce
3472 @opindex fstrength-reduce
3473 Perform the optimizations of loop strength reduction and
3474 elimination of iteration variables.
3476 @item -fthread-jumps
3477 @opindex fthread-jumps
3478 Perform optimizations where we check to see if a jump branches to a
3479 location where another comparison subsumed by the first is found. If
3480 so, the first branch is redirected to either the destination of the
3481 second branch or a point immediately following it, depending on whether
3482 the condition is known to be true or false.
3484 @item -fcse-follow-jumps
3485 @opindex fcse-follow-jumps
3486 In common subexpression elimination, scan through jump instructions
3487 when the target of the jump is not reached by any other path. For
3488 example, when CSE encounters an @code{if} statement with an
3489 @code{else} clause, CSE will follow the jump when the condition
3492 @item -fcse-skip-blocks
3493 @opindex fcse-skip-blocks
3494 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3495 follow jumps which conditionally skip over blocks. When CSE
3496 encounters a simple @code{if} statement with no else clause,
3497 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3498 body of the @code{if}.
3500 @item -frerun-cse-after-loop
3501 @opindex frerun-cse-after-loop
3502 Re-run common subexpression elimination after loop optimizations has been
3505 @item -frerun-loop-opt
3506 @opindex frerun-loop-opt
3507 Run the loop optimizer twice.
3511 Perform a global common subexpression elimination pass.
3512 This pass also performs global constant and copy propagation.
3514 @emph{Note:} When compiling a program using computed gotos, a GCC
3515 extension, you may get better runtime performance if you disable
3516 the global common subexpression elmination pass by adding
3517 @option{-fno-gcse} to the command line.
3521 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3522 attempt to move loads which are only killed by stores into themselves. This
3523 allows a loop containing a load/store sequence to be changed to a load outside
3524 the loop, and a copy/store within the loop.
3528 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3529 subexpression elimination. This pass will attempt to move stores out of loops.
3530 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3531 can be changed to a load before the loop and a store after the loop.
3533 @item -floop-optimize
3534 @opindex floop-optimize
3535 Perform loop optimizations: move constant expressions out of loops, simplify
3536 exit test conditions and optionally do strength-reduction and loop unrolling as
3539 @item -fcrossjumping
3540 @opindex crossjumping
3541 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
3542 resulting code may or may not perform better than without cross-jumping.
3544 @item -fif-conversion
3545 @opindex if-conversion
3546 Attempt to transform conditional jumps into branch-less equivalents. This
3547 include use of conditional moves, min, max, set flags and abs instructions, and
3548 some tricks doable by standard arithmetics. The use of conditional execution
3549 on chips where it is available is controlled by @code{if-conversion2}.
3551 @item -fif-conversion2
3552 @opindex if-conversion2
3553 Use conditional execution (where available) to transform conditional jumps into
3554 branch-less equivalents.
3556 @item -fdelete-null-pointer-checks
3557 @opindex fdelete-null-pointer-checks
3558 Use global dataflow analysis to identify and eliminate useless checks
3559 for null pointers. The compiler assumes that dereferencing a null
3560 pointer would have halted the program. If a pointer is checked after
3561 it has already been dereferenced, it cannot be null.
3563 In some environments, this assumption is not true, and programs can
3564 safely dereference null pointers. Use
3565 @option{-fno-delete-null-pointer-checks} to disable this optimization
3566 for programs which depend on that behavior.
3568 @item -fexpensive-optimizations
3569 @opindex fexpensive-optimizations
3570 Perform a number of minor optimizations that are relatively expensive.
3572 @item -foptimize-register-move
3574 @opindex foptimize-register-move
3576 Attempt to reassign register numbers in move instructions and as
3577 operands of other simple instructions in order to maximize the amount of
3578 register tying. This is especially helpful on machines with two-operand
3579 instructions. GCC enables this optimization by default with @option{-O2}
3582 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3585 @item -fdelayed-branch
3586 @opindex fdelayed-branch
3587 If supported for the target machine, attempt to reorder instructions
3588 to exploit instruction slots available after delayed branch
3591 @item -fschedule-insns
3592 @opindex fschedule-insns
3593 If supported for the target machine, attempt to reorder instructions to
3594 eliminate execution stalls due to required data being unavailable. This
3595 helps machines that have slow floating point or memory load instructions
3596 by allowing other instructions to be issued until the result of the load
3597 or floating point instruction is required.
3599 @item -fschedule-insns2
3600 @opindex fschedule-insns2
3601 Similar to @option{-fschedule-insns}, but requests an additional pass of
3602 instruction scheduling after register allocation has been done. This is
3603 especially useful on machines with a relatively small number of
3604 registers and where memory load instructions take more than one cycle.
3606 @item -ffunction-sections
3607 @itemx -fdata-sections
3608 @opindex ffunction-sections
3609 @opindex fdata-sections
3610 Place each function or data item into its own section in the output
3611 file if the target supports arbitrary sections. The name of the
3612 function or the name of the data item determines the section's name
3615 Use these options on systems where the linker can perform optimizations
3616 to improve locality of reference in the instruction space. HPPA
3617 processors running HP-UX and Sparc processors running Solaris 2 have
3618 linkers with such optimizations. Other systems using the ELF object format
3619 as well as AIX may have these optimizations in the future.
3621 Only use these options when there are significant benefits from doing
3622 so. When you specify these options, the assembler and linker will
3623 create larger object and executable files and will also be slower.
3624 You will not be able to use @code{gprof} on all systems if you
3625 specify this option and you may have problems with debugging if
3626 you specify both this option and @option{-g}.
3628 @item -fcaller-saves
3629 @opindex fcaller-saves
3630 Enable values to be allocated in registers that will be clobbered by
3631 function calls, by emitting extra instructions to save and restore the
3632 registers around such calls. Such allocation is done only when it
3633 seems to result in better code than would otherwise be produced.
3635 This option is always enabled by default on certain machines, usually
3636 those which have no call-preserved registers to use instead.
3638 For all machines, optimization level 2 and higher enables this flag by
3643 Perform tail duplication to enlarge superblock size. This transformation
3644 simplifies the control flow of the function allowing other optimizations to do
3647 @item -funroll-loops
3648 @opindex funroll-loops
3649 Unroll loops whose number of iterations can be determined at compile
3650 time or upon entry to the loop. @option{-funroll-loops} implies both
3651 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3652 option makes code larger, and may or may not make it run faster.
3654 @item -funroll-all-loops
3655 @opindex funroll-all-loops
3656 Unroll all loops, even if their number of iterations is uncertain when
3657 the loop is entered. This usually makes programs run more slowly.
3658 @option{-funroll-all-loops} implies the same options as
3659 @option{-funroll-loops},
3661 @item -fprefetch-loop-arrays
3662 @opindex fprefetch-loop-arrays
3663 If supported by the target machine, generate instructions to prefetch
3664 memory to improve the performance of loops that access large arrays.
3666 @item -fmove-all-movables
3667 @opindex fmove-all-movables
3668 Forces all invariant computations in loops to be moved
3671 @item -freduce-all-givs
3672 @opindex freduce-all-givs
3673 Forces all general-induction variables in loops to be
3676 @emph{Note:} When compiling programs written in Fortran,
3677 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3678 by default when you use the optimizer.
3680 These options may generate better or worse code; results are highly
3681 dependent on the structure of loops within the source code.
3683 These two options are intended to be removed someday, once
3684 they have helped determine the efficacy of various
3685 approaches to improving loop optimizations.
3687 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3688 know how use of these options affects
3689 the performance of your production code.
3690 We're very interested in code that runs @emph{slower}
3691 when these options are @emph{enabled}.
3694 @itemx -fno-peephole2
3695 @opindex fno-peephole
3696 @opindex fno-peephole2
3697 Disable any machine-specific peephole optimizations. The difference
3698 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3699 are implemented in the compiler; some targets use one, some use the
3700 other, a few use both.
3702 @item -fbranch-probabilities
3703 @opindex fbranch-probabilities
3704 After running a program compiled with @option{-fprofile-arcs}
3705 (@pxref{Debugging Options,, Options for Debugging Your Program or
3706 @command{gcc}}), you can compile it a second time using
3707 @option{-fbranch-probabilities}, to improve optimizations based on
3708 the number of times each branch was taken. When the program
3709 compiled with @option{-fprofile-arcs} exits it saves arc execution
3710 counts to a file called @file{@var{sourcename}.da} for each source
3711 file The information in this data file is very dependent on the
3712 structure of the generated code, so you must use the same source code
3713 and the same optimization options for both compilations.
3715 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3716 note on the first instruction of each basic block, and a
3717 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3718 These can be used to improve optimization. Currently, they are only
3719 used in one place: in @file{reorg.c}, instead of guessing which path a
3720 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3721 exactly determine which path is taken more often.
3723 @item -fno-guess-branch-probability
3724 @opindex fno-guess-branch-probability
3725 Do not guess branch probabilities using a randomized model.
3727 Sometimes gcc will opt to use a randomized model to guess branch
3728 probabilities, when none are available from either profiling feedback
3729 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3730 different runs of the compiler on the same program may produce different
3733 In a hard real-time system, people don't want different runs of the
3734 compiler to produce code that has different behavior; minimizing
3735 non-determinism is of paramount import. This switch allows users to
3736 reduce non-determinism, possibly at the expense of inferior
3739 @item -freorder-blocks
3740 @opindex freorder-blocks
3741 Reorder basic blocks in the compiled function in order to reduce number of
3742 taken branches and improve code locality.
3744 @item -freorder-functions
3745 @opindex freorder-functions
3746 Reorder basic blocks in the compiled function in order to reduce number of
3747 taken branches and improve code locality. This is implemented by using special
3748 subsections @code{text.hot} for most frequently executed functions and
3749 @code{text.unlikely} for unlikely executed functions. Reordering is done by
3750 the linker so object file format must support named sections and linker must
3751 place them in resonable way.
3753 Also profile feedback must be available in to make this option effective. See
3754 @option{-fprofile-arcs} for details.
3756 @item -fstrict-aliasing
3757 @opindex fstrict-aliasing
3758 Allows the compiler to assume the strictest aliasing rules applicable to
3759 the language being compiled. For C (and C++), this activates
3760 optimizations based on the type of expressions. In particular, an
3761 object of one type is assumed never to reside at the same address as an
3762 object of a different type, unless the types are almost the same. For
3763 example, an @code{unsigned int} can alias an @code{int}, but not a
3764 @code{void*} or a @code{double}. A character type may alias any other
3767 Pay special attention to code like this:
3780 The practice of reading from a different union member than the one most
3781 recently written to (called ``type-punning'') is common. Even with
3782 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3783 is accessed through the union type. So, the code above will work as
3784 expected. However, this code might not:
3795 Every language that wishes to perform language-specific alias analysis
3796 should define a function that computes, given an @code{tree}
3797 node, an alias set for the node. Nodes in different alias sets are not
3798 allowed to alias. For an example, see the C front-end function
3799 @code{c_get_alias_set}.
3801 For all machines, optimization level 2 and higher enables this flag by
3804 @item -falign-functions
3805 @itemx -falign-functions=@var{n}
3806 @opindex falign-functions
3807 Align the start of functions to the next power-of-two greater than
3808 @var{n}, skipping up to @var{n} bytes. For instance,
3809 @option{-falign-functions=32} aligns functions to the next 32-byte
3810 boundary, but @option{-falign-functions=24} would align to the next
3811 32-byte boundary only if this can be done by skipping 23 bytes or less.
3813 @option{-fno-align-functions} and @option{-falign-functions=1} are
3814 equivalent and mean that functions will not be aligned.
3816 Some assemblers only support this flag when @var{n} is a power of two;
3817 in that case, it is rounded up.
3819 If @var{n} is not specified, use a machine-dependent default.
3821 @item -falign-labels
3822 @itemx -falign-labels=@var{n}
3823 @opindex falign-labels
3824 Align all branch targets to a power-of-two boundary, skipping up to
3825 @var{n} bytes like @option{-falign-functions}. This option can easily
3826 make code slower, because it must insert dummy operations for when the
3827 branch target is reached in the usual flow of the code.
3829 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3830 are greater than this value, then their values are used instead.
3832 If @var{n} is not specified, use a machine-dependent default which is
3833 very likely to be @samp{1}, meaning no alignment.
3836 @itemx -falign-loops=@var{n}
3837 @opindex falign-loops
3838 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3839 like @option{-falign-functions}. The hope is that the loop will be
3840 executed many times, which will make up for any execution of the dummy
3843 If @var{n} is not specified, use a machine-dependent default.
3846 @itemx -falign-jumps=@var{n}
3847 @opindex falign-jumps
3848 Align branch targets to a power-of-two boundary, for branch targets
3849 where the targets can only be reached by jumping, skipping up to @var{n}
3850 bytes like @option{-falign-functions}. In this case, no dummy operations
3853 If @var{n} is not specified, use a machine-dependent default.
3857 Perform optimizations in static single assignment form. Each function's
3858 flow graph is translated into SSA form, optimizations are performed, and
3859 the flow graph is translated back from SSA form. Users should not
3860 specify this option, since it is not yet ready for production use.
3864 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3865 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3869 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3870 Like @option{-fssa}, this is an experimental feature.
3872 @item -fsingle-precision-constant
3873 @opindex fsingle-precision-constant
3874 Treat floating point constant as single precision constant instead of
3875 implicitly converting it to double precision constant.
3877 @item -frename-registers
3878 @opindex frename-registers
3879 Attempt to avoid false dependencies in scheduled code by making use
3880 of registers left over after register allocation. This optimization
3881 will most benefit processors with lots of registers. It can, however,
3882 make debugging impossible, since variables will no longer stay in
3883 a ``home register''.
3885 @item -fno-cprop-registers
3886 @opindex fno-cprop-registers
3887 After register allocation and post-register allocation instruction splitting,
3888 we perform a copy-propagation pass to try to reduce scheduling dependencies
3889 and occasionally eliminate the copy.
3891 @item --param @var{name}=@var{value}
3893 In some places, GCC uses various constants to control the amount of
3894 optimization that is done. For example, GCC will not inline functions
3895 that contain more that a certain number of instructions. You can
3896 control some of these constants on the command-line using the
3897 @option{--param} option.
3899 In each case, the @var{value} is an integer. The allowable choices for
3900 @var{name} are given in the following table:
3903 @item max-delay-slot-insn-search
3904 The maximum number of instructions to consider when looking for an
3905 instruction to fill a delay slot. If more than this arbitrary number of
3906 instructions is searched, the time savings from filling the delay slot
3907 will be minimal so stop searching. Increasing values mean more
3908 aggressive optimization, making the compile time increase with probably
3909 small improvement in executable run time.
3911 @item max-delay-slot-live-search
3912 When trying to fill delay slots, the maximum number of instructions to
3913 consider when searching for a block with valid live register
3914 information. Increasing this arbitrarily chosen value means more
3915 aggressive optimization, increasing the compile time. This parameter
3916 should be removed when the delay slot code is rewritten to maintain the
3919 @item max-gcse-memory
3920 The approximate maximum amount of memory that will be allocated in
3921 order to perform the global common subexpression elimination
3922 optimization. If more memory than specified is required, the
3923 optimization will not be done.
3925 @item max-gcse-passes
3926 The maximum number of passes of GCSE to run.
3928 @item max-pending-list-length
3929 The maximum number of pending dependencies scheduling will allow
3930 before flushing the current state and starting over. Large functions
3931 with few branches or calls can create excessively large lists which
3932 needlessly consume memory and resources.
3934 @item max-inline-insns
3935 If an function contains more than this many instructions, it
3936 will not be inlined. This option is precisely equivalent to
3937 @option{-finline-limit}.
3939 @item max-unrolled-insns
3940 The maximum number of instructions that a loop should have if that loop
3941 is unrolled, and if the loop is unrolled, it determines how many times
3942 the loop code is unrolled.
3944 @item hot-bb-count-fraction
3945 Select fraction of the maximal count of repetitions of basic block in program
3946 given basic block needs to have to be considered hot.
3948 @item hot-bb-frequency-fraction
3949 Select fraction of the maximal frequency of executions of basic block in
3950 function given basic block needs to have to be considered hot
3952 @item tracer-dynamic-coverage
3953 @itemx tracer-dynamic-coverage-feedback
3955 This value is used to limit superblock formation once given percentage of
3956 executed instructions is covered. This limits unnecesary code size expansion.
3958 The @option{tracer-dynamic-coverage-feedback} is used only when profile
3959 feedback is available. The real profiles (as opposed to statically estimated
3960 ones) are much less balanced allowing the threshold to be larger value.
3962 @item tracer-max-code-growth
3963 Stop tail duplication once code growth has reached given percentage. This is
3964 rather hokey argument, as most of the duplicates will be elliminated later in
3965 cross jumping, so it may be set to much higher values than is the desired code
3968 @item tracer-min-branch-ratio
3970 Stop reverse growth when the reverse probability of best edge is less than this
3971 threshold (in percent).
3973 @item tracer-min-branch-ratio
3974 @itemx tracer-min-branch-ratio-feedback
3976 Stop forward growth if the best edge do have probability lower than this
3979 Similary to @option{tracer-dynamic-coverage} two values are present, one for
3980 compilation for profile feedback and one for compilation without. The value
3981 for compilation with profile feedback needs to be more conservative (higher) in
3982 order to make tracer effective.
3987 @node Preprocessor Options
3988 @section Options Controlling the Preprocessor
3989 @cindex preprocessor options
3990 @cindex options, preprocessor
3992 These options control the C preprocessor, which is run on each C source
3993 file before actual compilation.
3995 If you use the @option{-E} option, nothing is done except preprocessing.
3996 Some of these options make sense only together with @option{-E} because
3997 they cause the preprocessor output to be unsuitable for actual
4001 You can use @option{-Wp,@var{option}} to bypass the compiler driver
4002 and pass @var{option} directly through to the preprocessor. If
4003 @var{option} contains commas, it is split into multiple options at the
4004 commas. However, many options are modified, translated or interpreted
4005 by the compiler driver before being passed to the preprocessor, and
4006 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
4007 interface is undocumented and subject to change, so whenever possible
4008 you should avoid using @option{-Wp} and let the driver handle the
4011 @include cppopts.texi
4013 @node Assembler Options
4014 @section Passing Options to the Assembler
4016 @c prevent bad page break with this line
4017 You can pass options to the assembler.
4020 @item -Wa,@var{option}
4022 Pass @var{option} as an option to the assembler. If @var{option}
4023 contains commas, it is split into multiple options at the commas.
4027 @section Options for Linking
4028 @cindex link options
4029 @cindex options, linking
4031 These options come into play when the compiler links object files into
4032 an executable output file. They are meaningless if the compiler is
4033 not doing a link step.
4037 @item @var{object-file-name}
4038 A file name that does not end in a special recognized suffix is
4039 considered to name an object file or library. (Object files are
4040 distinguished from libraries by the linker according to the file
4041 contents.) If linking is done, these object files are used as input
4050 If any of these options is used, then the linker is not run, and
4051 object file names should not be used as arguments. @xref{Overall
4055 @item -l@var{library}
4056 @itemx -l @var{library}
4058 Search the library named @var{library} when linking. (The second
4059 alternative with the library as a separate argument is only for
4060 POSIX compliance and is not recommended.)
4062 It makes a difference where in the command you write this option; the
4063 linker searches and processes libraries and object files in the order they
4064 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4065 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4066 to functions in @samp{z}, those functions may not be loaded.
4068 The linker searches a standard list of directories for the library,
4069 which is actually a file named @file{lib@var{library}.a}. The linker
4070 then uses this file as if it had been specified precisely by name.
4072 The directories searched include several standard system directories
4073 plus any that you specify with @option{-L}.
4075 Normally the files found this way are library files---archive files
4076 whose members are object files. The linker handles an archive file by
4077 scanning through it for members which define symbols that have so far
4078 been referenced but not defined. But if the file that is found is an
4079 ordinary object file, it is linked in the usual fashion. The only
4080 difference between using an @option{-l} option and specifying a file name
4081 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4082 and searches several directories.
4086 You need this special case of the @option{-l} option in order to
4087 link an Objective-C program.
4090 @opindex nostartfiles
4091 Do not use the standard system startup files when linking.
4092 The standard system libraries are used normally, unless @option{-nostdlib}
4093 or @option{-nodefaultlibs} is used.
4095 @item -nodefaultlibs
4096 @opindex nodefaultlibs
4097 Do not use the standard system libraries when linking.
4098 Only the libraries you specify will be passed to the linker.
4099 The standard startup files are used normally, unless @option{-nostartfiles}
4100 is used. The compiler may generate calls to memcmp, memset, and memcpy
4101 for System V (and ISO C) environments or to bcopy and bzero for
4102 BSD environments. These entries are usually resolved by entries in
4103 libc. These entry points should be supplied through some other
4104 mechanism when this option is specified.
4108 Do not use the standard system startup files or libraries when linking.
4109 No startup files and only the libraries you specify will be passed to
4110 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4111 for System V (and ISO C) environments or to bcopy and bzero for
4112 BSD environments. These entries are usually resolved by entries in
4113 libc. These entry points should be supplied through some other
4114 mechanism when this option is specified.
4116 @cindex @option{-lgcc}, use with @option{-nostdlib}
4117 @cindex @option{-nostdlib} and unresolved references
4118 @cindex unresolved references and @option{-nostdlib}
4119 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4120 @cindex @option{-nodefaultlibs} and unresolved references
4121 @cindex unresolved references and @option{-nodefaultlibs}
4122 One of the standard libraries bypassed by @option{-nostdlib} and
4123 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4124 that GCC uses to overcome shortcomings of particular machines, or special
4125 needs for some languages.
4126 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4127 Collection (GCC) Internals},
4128 for more discussion of @file{libgcc.a}.)
4129 In most cases, you need @file{libgcc.a} even when you want to avoid
4130 other standard libraries. In other words, when you specify @option{-nostdlib}
4131 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4132 This ensures that you have no unresolved references to internal GCC
4133 library subroutines. (For example, @samp{__main}, used to ensure C++
4134 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4135 GNU Compiler Collection (GCC) Internals}.)
4139 Remove all symbol table and relocation information from the executable.
4143 On systems that support dynamic linking, this prevents linking with the shared
4144 libraries. On other systems, this option has no effect.
4148 Produce a shared object which can then be linked with other objects to
4149 form an executable. Not all systems support this option. For predictable
4150 results, you must also specify the same set of options that were used to
4151 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4152 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4153 needs to build supplementary stub code for constructors to work. On
4154 multi-libbed systems, @samp{gcc -shared} must select the correct support
4155 libraries to link against. Failing to supply the correct flags may lead
4156 to subtle defects. Supplying them in cases where they are not necessary
4159 @item -shared-libgcc
4160 @itemx -static-libgcc
4161 @opindex shared-libgcc
4162 @opindex static-libgcc
4163 On systems that provide @file{libgcc} as a shared library, these options
4164 force the use of either the shared or static version respectively.
4165 If no shared version of @file{libgcc} was built when the compiler was
4166 configured, these options have no effect.
4168 There are several situations in which an application should use the
4169 shared @file{libgcc} instead of the static version. The most common
4170 of these is when the application wishes to throw and catch exceptions
4171 across different shared libraries. In that case, each of the libraries
4172 as well as the application itself should use the shared @file{libgcc}.
4174 Therefore, the G++ and GCJ drivers automatically add
4175 @option{-shared-libgcc} whenever you build a shared library or a main
4176 executable, because C++ and Java programs typically use exceptions, so
4177 this is the right thing to do.
4179 If, instead, you use the GCC driver to create shared libraries, you may
4180 find that they will not always be linked with the shared @file{libgcc}.
4181 If GCC finds, at its configuration time, that you have a GNU linker that
4182 does not support option @option{--eh-frame-hdr}, it will link the shared
4183 version of @file{libgcc} into shared libraries by default. Otherwise,
4184 it will take advantage of the linker and optimize away the linking with
4185 the shared version of @file{libgcc}, linking with the static version of
4186 libgcc by default. This allows exceptions to propagate through such
4187 shared libraries, without incurring relocation costs at library load
4190 However, if a library or main executable is supposed to throw or catch
4191 exceptions, you must link it using the G++ or GCJ driver, as appropriate
4192 for the languages used in the program, or using the option
4193 @option{-shared-libgcc}, such that it is linked with the shared
4198 Bind references to global symbols when building a shared object. Warn
4199 about any unresolved references (unless overridden by the link editor
4200 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4203 @item -Xlinker @var{option}
4205 Pass @var{option} as an option to the linker. You can use this to
4206 supply system-specific linker options which GCC does not know how to
4209 If you want to pass an option that takes an argument, you must use
4210 @option{-Xlinker} twice, once for the option and once for the argument.
4211 For example, to pass @option{-assert definitions}, you must write
4212 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4213 @option{-Xlinker "-assert definitions"}, because this passes the entire
4214 string as a single argument, which is not what the linker expects.
4216 @item -Wl,@var{option}
4218 Pass @var{option} as an option to the linker. If @var{option} contains
4219 commas, it is split into multiple options at the commas.
4221 @item -u @var{symbol}
4223 Pretend the symbol @var{symbol} is undefined, to force linking of
4224 library modules to define it. You can use @option{-u} multiple times with
4225 different symbols to force loading of additional library modules.
4228 @node Directory Options
4229 @section Options for Directory Search
4230 @cindex directory options
4231 @cindex options, directory search
4234 These options specify directories to search for header files, for
4235 libraries and for parts of the compiler:
4240 Add the directory @var{dir} to the head of the list of directories to be
4241 searched for header files. This can be used to override a system header
4242 file, substituting your own version, since these directories are
4243 searched before the system header file directories. However, you should
4244 not use this option to add directories that contain vendor-supplied
4245 system header files (use @option{-isystem} for that). If you use more than
4246 one @option{-I} option, the directories are scanned in left-to-right
4247 order; the standard system directories come after.
4249 If a standard system include directory, or a directory specified with
4250 @option{-isystem}, is also specified with @option{-I}, it will be
4251 searched only in the position requested by @option{-I}. Also, it will
4252 not be considered a system include directory. If that directory really
4253 does contain system headers, there is a good chance that they will
4254 break. For instance, if GCC's installation procedure edited the headers
4255 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4256 original, buggy headers to be found instead of the corrected ones. GCC
4257 will issue a warning when a system include directory is hidden in this
4262 Any directories you specify with @option{-I} options before the @option{-I-}
4263 option are searched only for the case of @samp{#include "@var{file}"};
4264 they are not searched for @samp{#include <@var{file}>}.
4266 If additional directories are specified with @option{-I} options after
4267 the @option{-I-}, these directories are searched for all @samp{#include}
4268 directives. (Ordinarily @emph{all} @option{-I} directories are used
4271 In addition, the @option{-I-} option inhibits the use of the current
4272 directory (where the current input file came from) as the first search
4273 directory for @samp{#include "@var{file}"}. There is no way to
4274 override this effect of @option{-I-}. With @option{-I.} you can specify
4275 searching the directory which was current when the compiler was
4276 invoked. That is not exactly the same as what the preprocessor does
4277 by default, but it is often satisfactory.
4279 @option{-I-} does not inhibit the use of the standard system directories
4280 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4285 Add directory @var{dir} to the list of directories to be searched
4288 @item -B@var{prefix}
4290 This option specifies where to find the executables, libraries,
4291 include files, and data files of the compiler itself.
4293 The compiler driver program runs one or more of the subprograms
4294 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4295 @var{prefix} as a prefix for each program it tries to run, both with and
4296 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4298 For each subprogram to be run, the compiler driver first tries the
4299 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4300 was not specified, the driver tries two standard prefixes, which are
4301 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4302 those results in a file name that is found, the unmodified program
4303 name is searched for using the directories specified in your
4304 @env{PATH} environment variable.
4306 The compiler will check to see if the path provided by the @option{-B}
4307 refers to a directory, and if necessary it will add a directory
4308 separator character at the end of the path.
4310 @option{-B} prefixes that effectively specify directory names also apply
4311 to libraries in the linker, because the compiler translates these
4312 options into @option{-L} options for the linker. They also apply to
4313 includes files in the preprocessor, because the compiler translates these
4314 options into @option{-isystem} options for the preprocessor. In this case,
4315 the compiler appends @samp{include} to the prefix.
4317 The run-time support file @file{libgcc.a} can also be searched for using
4318 the @option{-B} prefix, if needed. If it is not found there, the two
4319 standard prefixes above are tried, and that is all. The file is left
4320 out of the link if it is not found by those means.
4322 Another way to specify a prefix much like the @option{-B} prefix is to use
4323 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4326 As a special kludge, if the path provided by @option{-B} is
4327 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4328 9, then it will be replaced by @file{[dir/]include}. This is to help
4329 with boot-strapping the compiler.
4331 @item -specs=@var{file}
4333 Process @var{file} after the compiler reads in the standard @file{specs}
4334 file, in order to override the defaults that the @file{gcc} driver
4335 program uses when determining what switches to pass to @file{cc1},
4336 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4337 @option{-specs=@var{file}} can be specified on the command line, and they
4338 are processed in order, from left to right.
4344 @section Specifying subprocesses and the switches to pass to them
4346 @command{gcc} is a driver program. It performs its job by invoking a
4347 sequence of other programs to do the work of compiling, assembling and
4348 linking. GCC interprets its command-line parameters and uses these to
4349 deduce which programs it should invoke, and which command-line options
4350 it ought to place on their command lines. This behavior is controlled
4351 by @dfn{spec strings}. In most cases there is one spec string for each
4352 program that GCC can invoke, but a few programs have multiple spec
4353 strings to control their behavior. The spec strings built into GCC can
4354 be overridden by using the @option{-specs=} command-line switch to specify
4357 @dfn{Spec files} are plaintext files that are used to construct spec
4358 strings. They consist of a sequence of directives separated by blank
4359 lines. The type of directive is determined by the first non-whitespace
4360 character on the line and it can be one of the following:
4363 @item %@var{command}
4364 Issues a @var{command} to the spec file processor. The commands that can
4368 @item %include <@var{file}>
4370 Search for @var{file} and insert its text at the current point in the
4373 @item %include_noerr <@var{file}>
4374 @cindex %include_noerr
4375 Just like @samp{%include}, but do not generate an error message if the include
4376 file cannot be found.
4378 @item %rename @var{old_name} @var{new_name}
4380 Rename the spec string @var{old_name} to @var{new_name}.
4384 @item *[@var{spec_name}]:
4385 This tells the compiler to create, override or delete the named spec
4386 string. All lines after this directive up to the next directive or
4387 blank line are considered to be the text for the spec string. If this
4388 results in an empty string then the spec will be deleted. (Or, if the
4389 spec did not exist, then nothing will happened.) Otherwise, if the spec
4390 does not currently exist a new spec will be created. If the spec does
4391 exist then its contents will be overridden by the text of this
4392 directive, unless the first character of that text is the @samp{+}
4393 character, in which case the text will be appended to the spec.
4395 @item [@var{suffix}]:
4396 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4397 and up to the next directive or blank line are considered to make up the
4398 spec string for the indicated suffix. When the compiler encounters an
4399 input file with the named suffix, it will processes the spec string in
4400 order to work out how to compile that file. For example:
4407 This says that any input file whose name ends in @samp{.ZZ} should be
4408 passed to the program @samp{z-compile}, which should be invoked with the
4409 command-line switch @option{-input} and with the result of performing the
4410 @samp{%i} substitution. (See below.)
4412 As an alternative to providing a spec string, the text that follows a
4413 suffix directive can be one of the following:
4416 @item @@@var{language}
4417 This says that the suffix is an alias for a known @var{language}. This is
4418 similar to using the @option{-x} command-line switch to GCC to specify a
4419 language explicitly. For example:
4426 Says that .ZZ files are, in fact, C++ source files.
4429 This causes an error messages saying:
4432 @var{name} compiler not installed on this system.
4436 GCC already has an extensive list of suffixes built into it.
4437 This directive will add an entry to the end of the list of suffixes, but
4438 since the list is searched from the end backwards, it is effectively
4439 possible to override earlier entries using this technique.
4443 GCC has the following spec strings built into it. Spec files can
4444 override these strings or create their own. Note that individual
4445 targets can also add their own spec strings to this list.
4448 asm Options to pass to the assembler
4449 asm_final Options to pass to the assembler post-processor
4450 cpp Options to pass to the C preprocessor
4451 cc1 Options to pass to the C compiler
4452 cc1plus Options to pass to the C++ compiler
4453 endfile Object files to include at the end of the link
4454 link Options to pass to the linker
4455 lib Libraries to include on the command line to the linker
4456 libgcc Decides which GCC support library to pass to the linker
4457 linker Sets the name of the linker
4458 predefines Defines to be passed to the C preprocessor
4459 signed_char Defines to pass to CPP to say whether @code{char} is signed
4461 startfile Object files to include at the start of the link
4464 Here is a small example of a spec file:
4470 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4473 This example renames the spec called @samp{lib} to @samp{old_lib} and
4474 then overrides the previous definition of @samp{lib} with a new one.
4475 The new definition adds in some extra command-line options before
4476 including the text of the old definition.
4478 @dfn{Spec strings} are a list of command-line options to be passed to their
4479 corresponding program. In addition, the spec strings can contain
4480 @samp{%}-prefixed sequences to substitute variable text or to
4481 conditionally insert text into the command line. Using these constructs
4482 it is possible to generate quite complex command lines.
4484 Here is a table of all defined @samp{%}-sequences for spec
4485 strings. Note that spaces are not generated automatically around the
4486 results of expanding these sequences. Therefore you can concatenate them
4487 together or combine them with constant text in a single argument.
4491 Substitute one @samp{%} into the program name or argument.
4494 Substitute the name of the input file being processed.
4497 Substitute the basename of the input file being processed.
4498 This is the substring up to (and not including) the last period
4499 and not including the directory.
4502 This is the same as @samp{%b}, but include the file suffix (text after
4506 Marks the argument containing or following the @samp{%d} as a
4507 temporary file name, so that that file will be deleted if GCC exits
4508 successfully. Unlike @samp{%g}, this contributes no text to the
4511 @item %g@var{suffix}
4512 Substitute a file name that has suffix @var{suffix} and is chosen
4513 once per compilation, and mark the argument in the same way as
4514 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4515 name is now chosen in a way that is hard to predict even when previously
4516 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4517 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4518 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4519 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4520 was simply substituted with a file name chosen once per compilation,
4521 without regard to any appended suffix (which was therefore treated
4522 just like ordinary text), making such attacks more likely to succeed.
4524 @item %u@var{suffix}
4525 Like @samp{%g}, but generates a new temporary file name even if
4526 @samp{%u@var{suffix}} was already seen.
4528 @item %U@var{suffix}
4529 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4530 new one if there is no such last file name. In the absence of any
4531 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4532 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4533 would involve the generation of two distinct file names, one
4534 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4535 simply substituted with a file name chosen for the previous @samp{%u},
4536 without regard to any appended suffix.
4538 @item %j@var{SUFFIX}
4539 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4540 writable, and if save-temps is off; otherwise, substitute the name
4541 of a temporary file, just like @samp{%u}. This temporary file is not
4542 meant for communication between processes, but rather as a junk
4545 @item %.@var{SUFFIX}
4546 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4547 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4548 terminated by the next space or %.
4551 Marks the argument containing or following the @samp{%w} as the
4552 designated output file of this compilation. This puts the argument
4553 into the sequence of arguments that @samp{%o} will substitute later.
4556 Substitutes the names of all the output files, with spaces
4557 automatically placed around them. You should write spaces
4558 around the @samp{%o} as well or the results are undefined.
4559 @samp{%o} is for use in the specs for running the linker.
4560 Input files whose names have no recognized suffix are not compiled
4561 at all, but they are included among the output files, so they will
4565 Substitutes the suffix for object files. Note that this is
4566 handled specially when it immediately follows @samp{%g, %u, or %U},
4567 because of the need for those to form complete file names. The
4568 handling is such that @samp{%O} is treated exactly as if it had already
4569 been substituted, except that @samp{%g, %u, and %U} do not currently
4570 support additional @var{suffix} characters following @samp{%O} as they would
4571 following, for example, @samp{.o}.
4574 Substitutes the standard macro predefinitions for the
4575 current target machine. Use this when running @code{cpp}.
4578 Like @samp{%p}, but puts @samp{__} before and after the name of each
4579 predefined macro, except for macros that start with @samp{__} or with
4580 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4584 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4587 Current argument is the name of a library or startup file of some sort.
4588 Search for that file in a standard list of directories and substitute
4589 the full name found.
4592 Print @var{str} as an error message. @var{str} is terminated by a newline.
4593 Use this when inconsistent options are detected.
4596 Output @samp{-} if the input for the current command is coming from a pipe.
4599 Substitute the contents of spec string @var{name} at this point.
4602 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4604 @item %x@{@var{option}@}
4605 Accumulate an option for @samp{%X}.
4608 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4612 Output the accumulated assembler options specified by @option{-Wa}.
4615 Output the accumulated preprocessor options specified by @option{-Wp}.
4618 Substitute the major version number of GCC@.
4619 (For version 2.9.5, this is 2.)
4622 Substitute the minor version number of GCC@.
4623 (For version 2.9.5, this is 9.)
4626 Substitute the patch level number of GCC@.
4627 (For version 2.9.5, this is 5.)
4630 Process the @code{asm} spec. This is used to compute the
4631 switches to be passed to the assembler.
4634 Process the @code{asm_final} spec. This is a spec string for
4635 passing switches to an assembler post-processor, if such a program is
4639 Process the @code{link} spec. This is the spec for computing the
4640 command line passed to the linker. Typically it will make use of the
4641 @samp{%L %G %S %D and %E} sequences.
4644 Dump out a @option{-L} option for each directory that GCC believes might
4645 contain startup files. If the target supports multilibs then the
4646 current multilib directory will be prepended to each of these paths.
4649 Output the multilib directory with directory separators replaced with
4650 @samp{_}. If multilib directories are not set, or the multilib directory is
4651 @file{.} then this option emits nothing.
4654 Process the @code{lib} spec. This is a spec string for deciding which
4655 libraries should be included on the command line to the linker.
4658 Process the @code{libgcc} spec. This is a spec string for deciding
4659 which GCC support library should be included on the command line to the linker.
4662 Process the @code{startfile} spec. This is a spec for deciding which
4663 object files should be the first ones passed to the linker. Typically
4664 this might be a file named @file{crt0.o}.
4667 Process the @code{endfile} spec. This is a spec string that specifies
4668 the last object files that will be passed to the linker.
4671 Process the @code{cpp} spec. This is used to construct the arguments
4672 to be passed to the C preprocessor.
4675 Process the @code{signed_char} spec. This is intended to be used
4676 to tell cpp whether a char is signed. It typically has the definition:
4678 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4682 Process the @code{cc1} spec. This is used to construct the options to be
4683 passed to the actual C compiler (@samp{cc1}).
4686 Process the @code{cc1plus} spec. This is used to construct the options to be
4687 passed to the actual C++ compiler (@samp{cc1plus}).
4690 Substitute the variable part of a matched option. See below.
4691 Note that each comma in the substituted string is replaced by
4695 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4696 If that switch was not specified, this substitutes nothing. Note that
4697 the leading dash is omitted when specifying this option, and it is
4698 automatically inserted if the substitution is performed. Thus the spec
4699 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4700 and would output the command line option @option{-foo}.
4702 @item %W@{@code{S}@}
4703 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4706 @item %@{@code{S}*@}
4707 Substitutes all the switches specified to GCC whose names start
4708 with @code{-S}, but which also take an argument. This is used for
4709 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4710 GCC considers @option{-o foo} as being
4711 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4712 text, including the space. Thus two arguments would be generated.
4714 @item %@{^@code{S}*@}
4715 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4716 argument. Thus %@{^o*@} would only generate one argument, not two.
4718 @item %@{@code{S}*&@code{T}*@}
4719 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4720 (the order of @code{S} and @code{T} in the spec is not significant).
4721 There can be any number of ampersand-separated variables; for each the
4722 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4724 @item %@{<@code{S}@}
4725 Remove all occurrences of @code{-S} from the command line. Note---this
4726 command is position dependent. @samp{%} commands in the spec string
4727 before this option will see @code{-S}, @samp{%} commands in the spec
4728 string after this option will not.
4730 @item %@{@code{S}*:@code{X}@}
4731 Substitutes @code{X} if one or more switches whose names start with
4732 @code{-S} are specified to GCC@. Note that the tail part of the
4733 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4734 for each occurrence of @samp{%*} within @code{X}.
4736 @item %@{@code{S}:@code{X}@}
4737 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4739 @item %@{!@code{S}:@code{X}@}
4740 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4742 @item %@{|@code{S}:@code{X}@}
4743 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4745 @item %@{|!@code{S}:@code{X}@}
4746 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4748 @item %@{.@code{S}:@code{X}@}
4749 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4751 @item %@{!.@code{S}:@code{X}@}
4752 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4754 @item %@{@code{S}|@code{P}:@code{X}@}
4755 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4756 combined with @samp{!} and @samp{.} sequences as well, although they
4757 have a stronger binding than the @samp{|}. For example a spec string
4761 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4764 will output the following command-line options from the following input
4765 command-line options:
4770 -d fred.c -foo -baz -boggle
4771 -d jim.d -bar -baz -boggle
4776 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4777 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4778 or spaces, or even newlines. They are processed as usual, as described
4781 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4782 switches are handled specifically in these
4783 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4784 @option{-W} switch is found later in the command line, the earlier switch
4785 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4786 letter, which passes all matching options.
4788 The character @samp{|} at the beginning of the predicate text is used to indicate
4789 that a command should be piped to the following command, but only if @option{-pipe}
4792 It is built into GCC which switches take arguments and which do not.
4793 (You might think it would be useful to generalize this to allow each
4794 compiler's spec to say which switches take arguments. But this cannot
4795 be done in a consistent fashion. GCC cannot even decide which input
4796 files have been specified without knowing which switches take arguments,
4797 and it must know which input files to compile in order to tell which
4800 GCC also knows implicitly that arguments starting in @option{-l} are to be
4801 treated as compiler output files, and passed to the linker in their
4802 proper position among the other output files.
4804 @c man begin OPTIONS
4806 @node Target Options
4807 @section Specifying Target Machine and Compiler Version
4808 @cindex target options
4809 @cindex cross compiling
4810 @cindex specifying machine version
4811 @cindex specifying compiler version and target machine
4812 @cindex compiler version, specifying
4813 @cindex target machine, specifying
4815 The usual way to run GCC is to run the executable called @file{gcc}, or
4816 @file{<machine>-gcc} when cross-compiling, or
4817 @file{<machine>-gcc-<version>} to run a version other than the one that
4818 was installed last. Sometimes this is inconvenient, so GCC provides
4819 options that will switch to another cross-compiler or version.
4822 @item -b @var{machine}
4824 The argument @var{machine} specifies the target machine for compilation.
4826 The value to use for @var{machine} is the same as was specified as the
4827 machine type when configuring GCC as a cross-compiler. For
4828 example, if a cross-compiler was configured with @samp{configure
4829 i386v}, meaning to compile for an 80386 running System V, then you
4830 would specify @option{-b i386v} to run that cross compiler.
4832 @item -V @var{version}
4834 The argument @var{version} specifies which version of GCC to run.
4835 This is useful when multiple versions are installed. For example,
4836 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
4839 The @option{-V} and @option{-b} options work by running the
4840 @file{<machine>-gcc-<version>} executable, so there's no real reason to
4841 use them if you can just run that directly.
4843 @node Submodel Options
4844 @section Hardware Models and Configurations
4845 @cindex submodel options
4846 @cindex specifying hardware config
4847 @cindex hardware models and configurations, specifying
4848 @cindex machine dependent options
4850 Earlier we discussed the standard option @option{-b} which chooses among
4851 different installed compilers for completely different target
4852 machines, such as VAX vs.@: 68000 vs.@: 80386.
4854 In addition, each of these target machine types can have its own
4855 special options, starting with @samp{-m}, to choose among various
4856 hardware models or configurations---for example, 68010 vs 68020,
4857 floating coprocessor or none. A single installed version of the
4858 compiler can compile for any model or configuration, according to the
4861 Some configurations of the compiler also support additional special
4862 options, usually for compatibility with other compilers on the same
4865 These options are defined by the macro @code{TARGET_SWITCHES} in the
4866 machine description. The default for the options is also defined by
4867 that macro, which enables you to change the defaults.
4881 * RS/6000 and PowerPC Options::
4884 * i386 and x86-64 Options::
4886 * Intel 960 Options::
4887 * DEC Alpha Options::
4888 * DEC Alpha/VMS Options::
4892 * System V Options::
4893 * TMS320C3x/C4x Options::
4901 * S/390 and zSeries Options::
4905 * Xstormy16 Options::
4909 @node M680x0 Options
4910 @subsection M680x0 Options
4911 @cindex M680x0 options
4913 These are the @samp{-m} options defined for the 68000 series. The default
4914 values for these options depends on which style of 68000 was selected when
4915 the compiler was configured; the defaults for the most common choices are
4923 Generate output for a 68000. This is the default
4924 when the compiler is configured for 68000-based systems.
4926 Use this option for microcontrollers with a 68000 or EC000 core,
4927 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
4933 Generate output for a 68020. This is the default
4934 when the compiler is configured for 68020-based systems.
4938 Generate output containing 68881 instructions for floating point.
4939 This is the default for most 68020 systems unless @option{--nfp} was
4940 specified when the compiler was configured.
4944 Generate output for a 68030. This is the default when the compiler is
4945 configured for 68030-based systems.
4949 Generate output for a 68040. This is the default when the compiler is
4950 configured for 68040-based systems.
4952 This option inhibits the use of 68881/68882 instructions that have to be
4953 emulated by software on the 68040. Use this option if your 68040 does not
4954 have code to emulate those instructions.
4958 Generate output for a 68060. This is the default when the compiler is
4959 configured for 68060-based systems.
4961 This option inhibits the use of 68020 and 68881/68882 instructions that
4962 have to be emulated by software on the 68060. Use this option if your 68060
4963 does not have code to emulate those instructions.
4967 Generate output for a CPU32. This is the default
4968 when the compiler is configured for CPU32-based systems.
4970 Use this option for microcontrollers with a
4971 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
4972 68336, 68340, 68341, 68349 and 68360.
4976 Generate output for a 520X ``coldfire'' family cpu. This is the default
4977 when the compiler is configured for 520X-based systems.
4979 Use this option for microcontroller with a 5200 core, including
4980 the MCF5202, MCF5203, MCF5204 and MCF5202.
4985 Generate output for a 68040, without using any of the new instructions.
4986 This results in code which can run relatively efficiently on either a
4987 68020/68881 or a 68030 or a 68040. The generated code does use the
4988 68881 instructions that are emulated on the 68040.
4992 Generate output for a 68060, without using any of the new instructions.
4993 This results in code which can run relatively efficiently on either a
4994 68020/68881 or a 68030 or a 68040. The generated code does use the
4995 68881 instructions that are emulated on the 68060.
4999 Generate output containing Sun FPA instructions for floating point.
5002 @opindex msoft-float
5003 Generate output containing library calls for floating point.
5004 @strong{Warning:} the requisite libraries are not available for all m68k
5005 targets. Normally the facilities of the machine's usual C compiler are
5006 used, but this can't be done directly in cross-compilation. You must
5007 make your own arrangements to provide suitable library functions for
5008 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5009 @samp{m68k-*-coff} do provide software floating point support.
5013 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5016 @opindex mnobitfield
5017 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5018 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5022 Do use the bit-field instructions. The @option{-m68020} option implies
5023 @option{-mbitfield}. This is the default if you use a configuration
5024 designed for a 68020.
5028 Use a different function-calling convention, in which functions
5029 that take a fixed number of arguments return with the @code{rtd}
5030 instruction, which pops their arguments while returning. This
5031 saves one instruction in the caller since there is no need to pop
5032 the arguments there.
5034 This calling convention is incompatible with the one normally
5035 used on Unix, so you cannot use it if you need to call libraries
5036 compiled with the Unix compiler.
5038 Also, you must provide function prototypes for all functions that
5039 take variable numbers of arguments (including @code{printf});
5040 otherwise incorrect code will be generated for calls to those
5043 In addition, seriously incorrect code will result if you call a
5044 function with too many arguments. (Normally, extra arguments are
5045 harmlessly ignored.)
5047 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5048 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5051 @itemx -mno-align-int
5053 @opindex mno-align-int
5054 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5055 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5056 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5057 Aligning variables on 32-bit boundaries produces code that runs somewhat
5058 faster on processors with 32-bit busses at the expense of more memory.
5060 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5061 align structures containing the above types differently than
5062 most published application binary interface specifications for the m68k.
5066 Use the pc-relative addressing mode of the 68000 directly, instead of
5067 using a global offset table. At present, this option implies @option{-fpic},
5068 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5069 not presently supported with @option{-mpcrel}, though this could be supported for
5070 68020 and higher processors.
5072 @item -mno-strict-align
5073 @itemx -mstrict-align
5074 @opindex mno-strict-align
5075 @opindex mstrict-align
5076 Do not (do) assume that unaligned memory references will be handled by
5081 @node M68hc1x Options
5082 @subsection M68hc1x Options
5083 @cindex M68hc1x options
5085 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5086 microcontrollers. The default values for these options depends on
5087 which style of microcontroller was selected when the compiler was configured;
5088 the defaults for the most common choices are given below.
5095 Generate output for a 68HC11. This is the default
5096 when the compiler is configured for 68HC11-based systems.
5102 Generate output for a 68HC12. This is the default
5103 when the compiler is configured for 68HC12-based systems.
5106 @opindex mauto-incdec
5107 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5112 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5114 @item -msoft-reg-count=@var{count}
5115 @opindex msoft-reg-count
5116 Specify the number of pseudo-soft registers which are used for the
5117 code generation. The maximum number is 32. Using more pseudo-soft
5118 register may or may not result in better code depending on the program.
5119 The default is 4 for 68HC11 and 2 for 68HC12.
5124 @subsection VAX Options
5127 These @samp{-m} options are defined for the VAX:
5132 Do not output certain jump instructions (@code{aobleq} and so on)
5133 that the Unix assembler for the VAX cannot handle across long
5138 Do output those jump instructions, on the assumption that you
5139 will assemble with the GNU assembler.
5143 Output code for g-format floating point numbers instead of d-format.
5147 @subsection SPARC Options
5148 @cindex SPARC options
5150 These @samp{-m} switches are supported on the SPARC:
5155 @opindex mno-app-regs
5157 Specify @option{-mapp-regs} to generate output using the global registers
5158 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5161 To be fully SVR4 ABI compliant at the cost of some performance loss,
5162 specify @option{-mno-app-regs}. You should compile libraries and system
5163 software with this option.
5168 @opindex mhard-float
5169 Generate output containing floating point instructions. This is the
5175 @opindex msoft-float
5176 Generate output containing library calls for floating point.
5177 @strong{Warning:} the requisite libraries are not available for all SPARC
5178 targets. Normally the facilities of the machine's usual C compiler are
5179 used, but this cannot be done directly in cross-compilation. You must make
5180 your own arrangements to provide suitable library functions for
5181 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5182 @samp{sparclite-*-*} do provide software floating point support.
5184 @option{-msoft-float} changes the calling convention in the output file;
5185 therefore, it is only useful if you compile @emph{all} of a program with
5186 this option. In particular, you need to compile @file{libgcc.a}, the
5187 library that comes with GCC, with @option{-msoft-float} in order for
5190 @item -mhard-quad-float
5191 @opindex mhard-quad-float
5192 Generate output containing quad-word (long double) floating point
5195 @item -msoft-quad-float
5196 @opindex msoft-quad-float
5197 Generate output containing library calls for quad-word (long double)
5198 floating point instructions. The functions called are those specified
5199 in the SPARC ABI@. This is the default.
5201 As of this writing, there are no sparc implementations that have hardware
5202 support for the quad-word floating point instructions. They all invoke
5203 a trap handler for one of these instructions, and then the trap handler
5204 emulates the effect of the instruction. Because of the trap handler overhead,
5205 this is much slower than calling the ABI library routines. Thus the
5206 @option{-msoft-quad-float} option is the default.
5212 With @option{-mflat}, the compiler does not generate save/restore instructions
5213 and will use a ``flat'' or single register window calling convention.
5214 This model uses %i7 as the frame pointer and is compatible with the normal
5215 register window model. Code from either may be intermixed.
5216 The local registers and the input registers (0--5) are still treated as
5217 ``call saved'' registers and will be saved on the stack as necessary.
5219 With @option{-mno-flat} (the default), the compiler emits save/restore
5220 instructions (except for leaf functions) and is the normal mode of operation.
5222 @item -mno-unaligned-doubles
5223 @itemx -munaligned-doubles
5224 @opindex mno-unaligned-doubles
5225 @opindex munaligned-doubles
5226 Assume that doubles have 8 byte alignment. This is the default.
5228 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5229 alignment only if they are contained in another type, or if they have an
5230 absolute address. Otherwise, it assumes they have 4 byte alignment.
5231 Specifying this option avoids some rare compatibility problems with code
5232 generated by other compilers. It is not the default because it results
5233 in a performance loss, especially for floating point code.
5235 @item -mno-faster-structs
5236 @itemx -mfaster-structs
5237 @opindex mno-faster-structs
5238 @opindex mfaster-structs
5239 With @option{-mfaster-structs}, the compiler assumes that structures
5240 should have 8 byte alignment. This enables the use of pairs of
5241 @code{ldd} and @code{std} instructions for copies in structure
5242 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5243 However, the use of this changed alignment directly violates the Sparc
5244 ABI@. Thus, it's intended only for use on targets where the developer
5245 acknowledges that their resulting code will not be directly in line with
5246 the rules of the ABI@.
5252 These two options select variations on the SPARC architecture.
5254 By default (unless specifically configured for the Fujitsu SPARClite),
5255 GCC generates code for the v7 variant of the SPARC architecture.
5257 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5258 code is that the compiler emits the integer multiply and integer
5259 divide instructions which exist in SPARC v8 but not in SPARC v7.
5261 @option{-msparclite} will give you SPARClite code. This adds the integer
5262 multiply, integer divide step and scan (@code{ffs}) instructions which
5263 exist in SPARClite but not in SPARC v7.
5265 These options are deprecated and will be deleted in a future GCC release.
5266 They have been replaced with @option{-mcpu=xxx}.
5271 @opindex msupersparc
5272 These two options select the processor for which the code is optimized.
5274 With @option{-mcypress} (the default), the compiler optimizes code for the
5275 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5276 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5278 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5279 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5280 of the full SPARC v8 instruction set.
5282 These options are deprecated and will be deleted in a future GCC release.
5283 They have been replaced with @option{-mcpu=xxx}.
5285 @item -mcpu=@var{cpu_type}
5287 Set the instruction set, register set, and instruction scheduling parameters
5288 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5289 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5290 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5291 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5293 Default instruction scheduling parameters are used for values that select
5294 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5295 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5297 Here is a list of each supported architecture and their supported
5302 v8: supersparc, hypersparc
5303 sparclite: f930, f934, sparclite86x
5308 @item -mtune=@var{cpu_type}
5310 Set the instruction scheduling parameters for machine type
5311 @var{cpu_type}, but do not set the instruction set or register set that the
5312 option @option{-mcpu=@var{cpu_type}} would.
5314 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5315 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5316 that select a particular cpu implementation. Those are @samp{cypress},
5317 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5318 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5322 These @samp{-m} switches are supported in addition to the above
5323 on the SPARCLET processor.
5326 @item -mlittle-endian
5327 @opindex mlittle-endian
5328 Generate code for a processor running in little-endian mode.
5332 Treat register @code{%g0} as a normal register.
5333 GCC will continue to clobber it as necessary but will not assume
5334 it always reads as 0.
5336 @item -mbroken-saverestore
5337 @opindex mbroken-saverestore
5338 Generate code that does not use non-trivial forms of the @code{save} and
5339 @code{restore} instructions. Early versions of the SPARCLET processor do
5340 not correctly handle @code{save} and @code{restore} instructions used with
5341 arguments. They correctly handle them used without arguments. A @code{save}
5342 instruction used without arguments increments the current window pointer
5343 but does not allocate a new stack frame. It is assumed that the window
5344 overflow trap handler will properly handle this case as will interrupt
5348 These @samp{-m} switches are supported in addition to the above
5349 on SPARC V9 processors in 64-bit environments.
5352 @item -mlittle-endian
5353 @opindex mlittle-endian
5354 Generate code for a processor running in little-endian mode.
5360 Generate code for a 32-bit or 64-bit environment.
5361 The 32-bit environment sets int, long and pointer to 32 bits.
5362 The 64-bit environment sets int to 32 bits and long and pointer
5365 @item -mcmodel=medlow
5366 @opindex mcmodel=medlow
5367 Generate code for the Medium/Low code model: the program must be linked
5368 in the low 32 bits of the address space. Pointers are 64 bits.
5369 Programs can be statically or dynamically linked.
5371 @item -mcmodel=medmid
5372 @opindex mcmodel=medmid
5373 Generate code for the Medium/Middle code model: the program must be linked
5374 in the low 44 bits of the address space, the text segment must be less than
5375 2G bytes, and data segment must be within 2G of the text segment.
5376 Pointers are 64 bits.
5378 @item -mcmodel=medany
5379 @opindex mcmodel=medany
5380 Generate code for the Medium/Anywhere code model: the program may be linked
5381 anywhere in the address space, the text segment must be less than
5382 2G bytes, and data segment must be within 2G of the text segment.
5383 Pointers are 64 bits.
5385 @item -mcmodel=embmedany
5386 @opindex mcmodel=embmedany
5387 Generate code for the Medium/Anywhere code model for embedded systems:
5388 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5389 (determined at link time). Register %g4 points to the base of the
5390 data segment. Pointers are still 64 bits.
5391 Programs are statically linked, PIC is not supported.
5394 @itemx -mno-stack-bias
5395 @opindex mstack-bias
5396 @opindex mno-stack-bias
5397 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5398 frame pointer if present, are offset by @minus{}2047 which must be added back
5399 when making stack frame references.
5400 Otherwise, assume no such offset is present.
5403 @node Convex Options
5404 @subsection Convex Options
5405 @cindex Convex options
5407 These @samp{-m} options are defined for Convex:
5412 Generate output for C1. The code will run on any Convex machine.
5413 The preprocessor symbol @code{__convex__c1__} is defined.
5417 Generate output for C2. Uses instructions not available on C1.
5418 Scheduling and other optimizations are chosen for max performance on C2.
5419 The preprocessor symbol @code{__convex_c2__} is defined.
5423 Generate output for C32xx. Uses instructions not available on C1.
5424 Scheduling and other optimizations are chosen for max performance on C32.
5425 The preprocessor symbol @code{__convex_c32__} is defined.
5429 Generate output for C34xx. Uses instructions not available on C1.
5430 Scheduling and other optimizations are chosen for max performance on C34.
5431 The preprocessor symbol @code{__convex_c34__} is defined.
5435 Generate output for C38xx. Uses instructions not available on C1.
5436 Scheduling and other optimizations are chosen for max performance on C38.
5437 The preprocessor symbol @code{__convex_c38__} is defined.
5441 Generate code which puts an argument count in the word preceding each
5442 argument list. This is compatible with regular CC, and a few programs
5443 may need the argument count word. GDB and other source-level debuggers
5444 do not need it; this info is in the symbol table.
5447 @opindex mnoargcount
5448 Omit the argument count word. This is the default.
5450 @item -mvolatile-cache
5451 @opindex mvolatile-cache
5452 Allow volatile references to be cached. This is the default.
5454 @item -mvolatile-nocache
5455 @opindex mvolatile-nocache
5456 Volatile references bypass the data cache, going all the way to memory.
5457 This is only needed for multi-processor code that does not use standard
5458 synchronization instructions. Making non-volatile references to volatile
5459 locations will not necessarily work.
5463 Type long is 32 bits, the same as type int. This is the default.
5467 Type long is 64 bits, the same as type long long. This option is useless,
5468 because no library support exists for it.
5471 @node AMD29K Options
5472 @subsection AMD29K Options
5473 @cindex AMD29K options
5475 These @samp{-m} options are defined for the AMD Am29000:
5480 @cindex DW bit (29k)
5481 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5482 halfword operations are directly supported by the hardware. This is the
5487 Generate code that assumes the @code{DW} bit is not set.
5491 @cindex byte writes (29k)
5492 Generate code that assumes the system supports byte and halfword write
5493 operations. This is the default.
5497 Generate code that assumes the systems does not support byte and
5498 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5502 @cindex memory model (29k)
5503 Use a small memory model that assumes that all function addresses are
5504 either within a single 256 KB segment or at an absolute address of less
5505 than 256k. This allows the @code{call} instruction to be used instead
5506 of a @code{const}, @code{consth}, @code{calli} sequence.
5510 Use the normal memory model: Generate @code{call} instructions only when
5511 calling functions in the same file and @code{calli} instructions
5512 otherwise. This works if each file occupies less than 256 KB but allows
5513 the entire executable to be larger than 256 KB@. This is the default.
5517 Always use @code{calli} instructions. Specify this option if you expect
5518 a single file to compile into more than 256 KB of code.
5522 @cindex processor selection (29k)
5523 Generate code for the Am29050.
5527 Generate code for the Am29000. This is the default.
5529 @item -mkernel-registers
5530 @opindex mkernel-registers
5531 @cindex kernel and user registers (29k)
5532 Generate references to registers @code{gr64-gr95} instead of to
5533 registers @code{gr96-gr127}. This option can be used when compiling
5534 kernel code that wants a set of global registers disjoint from that used
5537 Note that when this option is used, register names in @samp{-f} flags
5538 must use the normal, user-mode, names.
5540 @item -muser-registers
5541 @opindex muser-registers
5542 Use the normal set of global registers, @code{gr96-gr127}. This is the
5546 @itemx -mno-stack-check
5547 @opindex mstack-check
5548 @opindex mno-stack-check
5549 @cindex stack checks (29k)
5550 Insert (or do not insert) a call to @code{__msp_check} after each stack
5551 adjustment. This is often used for kernel code.
5554 @itemx -mno-storem-bug
5555 @opindex mstorem-bug
5556 @opindex mno-storem-bug
5557 @cindex storem bug (29k)
5558 @option{-mstorem-bug} handles 29k processors which cannot handle the
5559 separation of a mtsrim insn and a storem instruction (most 29000 chips
5560 to date, but not the 29050).
5562 @item -mno-reuse-arg-regs
5563 @itemx -mreuse-arg-regs
5564 @opindex mno-reuse-arg-regs
5565 @opindex mreuse-arg-regs
5566 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5567 registers for copying out arguments. This helps detect calling a function
5568 with fewer arguments than it was declared with.
5570 @item -mno-impure-text
5571 @itemx -mimpure-text
5572 @opindex mno-impure-text
5573 @opindex mimpure-text
5574 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5575 not pass @option{-assert pure-text} to the linker when linking a shared object.
5578 @opindex msoft-float
5579 Generate output containing library calls for floating point.
5580 @strong{Warning:} the requisite libraries are not part of GCC@.
5581 Normally the facilities of the machine's usual C compiler are used, but
5582 this can't be done directly in cross-compilation. You must make your
5583 own arrangements to provide suitable library functions for
5588 Do not generate multm or multmu instructions. This is useful for some embedded
5589 systems which do not have trap handlers for these instructions.
5593 @subsection ARM Options
5596 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5601 @opindex mapcs-frame
5602 Generate a stack frame that is compliant with the ARM Procedure Call
5603 Standard for all functions, even if this is not strictly necessary for
5604 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5605 with this option will cause the stack frames not to be generated for
5606 leaf functions. The default is @option{-mno-apcs-frame}.
5610 This is a synonym for @option{-mapcs-frame}.
5614 Generate code for a processor running with a 26-bit program counter,
5615 and conforming to the function calling standards for the APCS 26-bit
5616 option. This option replaces the @option{-m2} and @option{-m3} options
5617 of previous releases of the compiler.
5621 Generate code for a processor running with a 32-bit program counter,
5622 and conforming to the function calling standards for the APCS 32-bit
5623 option. This option replaces the @option{-m6} option of previous releases
5627 @c not currently implemented
5628 @item -mapcs-stack-check
5629 @opindex mapcs-stack-check
5630 Generate code to check the amount of stack space available upon entry to
5631 every function (that actually uses some stack space). If there is
5632 insufficient space available then either the function
5633 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5634 called, depending upon the amount of stack space required. The run time
5635 system is required to provide these functions. The default is
5636 @option{-mno-apcs-stack-check}, since this produces smaller code.
5638 @c not currently implemented
5640 @opindex mapcs-float
5641 Pass floating point arguments using the float point registers. This is
5642 one of the variants of the APCS@. This option is recommended if the
5643 target hardware has a floating point unit or if a lot of floating point
5644 arithmetic is going to be performed by the code. The default is
5645 @option{-mno-apcs-float}, since integer only code is slightly increased in
5646 size if @option{-mapcs-float} is used.
5648 @c not currently implemented
5649 @item -mapcs-reentrant
5650 @opindex mapcs-reentrant
5651 Generate reentrant, position independent code. The default is
5652 @option{-mno-apcs-reentrant}.
5655 @item -mthumb-interwork
5656 @opindex mthumb-interwork
5657 Generate code which supports calling between the ARM and Thumb
5658 instruction sets. Without this option the two instruction sets cannot
5659 be reliably used inside one program. The default is
5660 @option{-mno-thumb-interwork}, since slightly larger code is generated
5661 when @option{-mthumb-interwork} is specified.
5663 @item -mno-sched-prolog
5664 @opindex mno-sched-prolog
5665 Prevent the reordering of instructions in the function prolog, or the
5666 merging of those instruction with the instructions in the function's
5667 body. This means that all functions will start with a recognizable set
5668 of instructions (or in fact one of a choice from a small set of
5669 different function prologues), and this information can be used to
5670 locate the start if functions inside an executable piece of code. The
5671 default is @option{-msched-prolog}.
5674 @opindex mhard-float
5675 Generate output containing floating point instructions. This is the
5679 @opindex msoft-float
5680 Generate output containing library calls for floating point.
5681 @strong{Warning:} the requisite libraries are not available for all ARM
5682 targets. Normally the facilities of the machine's usual C compiler are
5683 used, but this cannot be done directly in cross-compilation. You must make
5684 your own arrangements to provide suitable library functions for
5687 @option{-msoft-float} changes the calling convention in the output file;
5688 therefore, it is only useful if you compile @emph{all} of a program with
5689 this option. In particular, you need to compile @file{libgcc.a}, the
5690 library that comes with GCC, with @option{-msoft-float} in order for
5693 @item -mlittle-endian
5694 @opindex mlittle-endian
5695 Generate code for a processor running in little-endian mode. This is
5696 the default for all standard configurations.
5699 @opindex mbig-endian
5700 Generate code for a processor running in big-endian mode; the default is
5701 to compile code for a little-endian processor.
5703 @item -mwords-little-endian
5704 @opindex mwords-little-endian
5705 This option only applies when generating code for big-endian processors.
5706 Generate code for a little-endian word order but a big-endian byte
5707 order. That is, a byte order of the form @samp{32107654}. Note: this
5708 option should only be used if you require compatibility with code for
5709 big-endian ARM processors generated by versions of the compiler prior to
5712 @item -malignment-traps
5713 @opindex malignment-traps
5714 Generate code that will not trap if the MMU has alignment traps enabled.
5715 On ARM architectures prior to ARMv4, there were no instructions to
5716 access half-word objects stored in memory. However, when reading from
5717 memory a feature of the ARM architecture allows a word load to be used,
5718 even if the address is unaligned, and the processor core will rotate the
5719 data as it is being loaded. This option tells the compiler that such
5720 misaligned accesses will cause a MMU trap and that it should instead
5721 synthesise the access as a series of byte accesses. The compiler can
5722 still use word accesses to load half-word data if it knows that the
5723 address is aligned to a word boundary.
5725 This option is ignored when compiling for ARM architecture 4 or later,
5726 since these processors have instructions to directly access half-word
5729 @item -mno-alignment-traps
5730 @opindex mno-alignment-traps
5731 Generate code that assumes that the MMU will not trap unaligned
5732 accesses. This produces better code when the target instruction set
5733 does not have half-word memory operations (i.e.@: implementations prior to
5736 Note that you cannot use this option to access unaligned word objects,
5737 since the processor will only fetch one 32-bit aligned object from
5740 The default setting for most targets is @option{-mno-alignment-traps}, since
5741 this produces better code when there are no half-word memory
5742 instructions available.
5744 @item -mshort-load-bytes
5745 @itemx -mno-short-load-words
5746 @opindex mshort-load-bytes
5747 @opindex mno-short-load-words
5748 These are deprecated aliases for @option{-malignment-traps}.
5750 @item -mno-short-load-bytes
5751 @itemx -mshort-load-words
5752 @opindex mno-short-load-bytes
5753 @opindex mshort-load-words
5754 This are deprecated aliases for @option{-mno-alignment-traps}.
5758 This option only applies to RISC iX@. Emulate the native BSD-mode
5759 compiler. This is the default if @option{-ansi} is not specified.
5763 This option only applies to RISC iX@. Emulate the native X/Open-mode
5766 @item -mno-symrename
5767 @opindex mno-symrename
5768 This option only applies to RISC iX@. Do not run the assembler
5769 post-processor, @samp{symrename}, after code has been assembled.
5770 Normally it is necessary to modify some of the standard symbols in
5771 preparation for linking with the RISC iX C library; this option
5772 suppresses this pass. The post-processor is never run when the
5773 compiler is built for cross-compilation.
5775 @item -mcpu=@var{name}
5777 This specifies the name of the target ARM processor. GCC uses this name
5778 to determine what kind of instructions it can emit when generating
5779 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
5780 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
5781 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
5782 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
5783 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
5784 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
5785 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
5786 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
5787 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
5788 @samp{arm1020t}, @samp{xscale}.
5790 @itemx -mtune=@var{name}
5792 This option is very similar to the @option{-mcpu=} option, except that
5793 instead of specifying the actual target processor type, and hence
5794 restricting which instructions can be used, it specifies that GCC should
5795 tune the performance of the code as if the target were of the type
5796 specified in this option, but still choosing the instructions that it
5797 will generate based on the cpu specified by a @option{-mcpu=} option.
5798 For some ARM implementations better performance can be obtained by using
5801 @item -march=@var{name}
5803 This specifies the name of the target ARM architecture. GCC uses this
5804 name to determine what kind of instructions it can emit when generating
5805 assembly code. This option can be used in conjunction with or instead
5806 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
5807 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
5808 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
5810 @item -mfpe=@var{number}
5811 @itemx -mfp=@var{number}
5814 This specifies the version of the floating point emulation available on
5815 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
5816 for @option{-mfpe=}, for compatibility with older versions of GCC@.
5818 @item -mstructure-size-boundary=@var{n}
5819 @opindex mstructure-size-boundary
5820 The size of all structures and unions will be rounded up to a multiple
5821 of the number of bits set by this option. Permissible values are 8 and
5822 32. The default value varies for different toolchains. For the COFF
5823 targeted toolchain the default value is 8. Specifying the larger number
5824 can produce faster, more efficient code, but can also increase the size
5825 of the program. The two values are potentially incompatible. Code
5826 compiled with one value cannot necessarily expect to work with code or
5827 libraries compiled with the other value, if they exchange information
5828 using structures or unions.
5830 @item -mabort-on-noreturn
5831 @opindex mabort-on-noreturn
5832 Generate a call to the function @code{abort} at the end of a
5833 @code{noreturn} function. It will be executed if the function tries to
5837 @itemx -mno-long-calls
5838 @opindex mlong-calls
5839 @opindex mno-long-calls
5840 Tells the compiler to perform function calls by first loading the
5841 address of the function into a register and then performing a subroutine
5842 call on this register. This switch is needed if the target function
5843 will lie outside of the 64 megabyte addressing range of the offset based
5844 version of subroutine call instruction.
5846 Even if this switch is enabled, not all function calls will be turned
5847 into long calls. The heuristic is that static functions, functions
5848 which have the @samp{short-call} attribute, functions that are inside
5849 the scope of a @samp{#pragma no_long_calls} directive and functions whose
5850 definitions have already been compiled within the current compilation
5851 unit, will not be turned into long calls. The exception to this rule is
5852 that weak function definitions, functions with the @samp{long-call}
5853 attribute or the @samp{section} attribute, and functions that are within
5854 the scope of a @samp{#pragma long_calls} directive, will always be
5855 turned into long calls.
5857 This feature is not enabled by default. Specifying
5858 @option{-mno-long-calls} will restore the default behavior, as will
5859 placing the function calls within the scope of a @samp{#pragma
5860 long_calls_off} directive. Note these switches have no effect on how
5861 the compiler generates code to handle function calls via function
5864 @item -mnop-fun-dllimport
5865 @opindex mnop-fun-dllimport
5866 Disable support for the @code{dllimport} attribute.
5868 @item -msingle-pic-base
5869 @opindex msingle-pic-base
5870 Treat the register used for PIC addressing as read-only, rather than
5871 loading it in the prologue for each function. The run-time system is
5872 responsible for initializing this register with an appropriate value
5873 before execution begins.
5875 @item -mpic-register=@var{reg}
5876 @opindex mpic-register
5877 Specify the register to be used for PIC addressing. The default is R10
5878 unless stack-checking is enabled, when R9 is used.
5880 @item -mpoke-function-name
5881 @opindex mpoke-function-name
5882 Write the name of each function into the text section, directly
5883 preceding the function prologue. The generated code is similar to this:
5887 .ascii "arm_poke_function_name", 0
5890 .word 0xff000000 + (t1 - t0)
5891 arm_poke_function_name
5893 stmfd sp!, @{fp, ip, lr, pc@}
5897 When performing a stack backtrace, code can inspect the value of
5898 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
5899 location @code{pc - 12} and the top 8 bits are set, then we know that
5900 there is a function name embedded immediately preceding this location
5901 and has length @code{((pc[-3]) & 0xff000000)}.
5905 Generate code for the 16-bit Thumb instruction set. The default is to
5906 use the 32-bit ARM instruction set.
5909 @opindex mtpcs-frame
5910 Generate a stack frame that is compliant with the Thumb Procedure Call
5911 Standard for all non-leaf functions. (A leaf function is one that does
5912 not call any other functions.) The default is @option{-mno-tpcs-frame}.
5914 @item -mtpcs-leaf-frame
5915 @opindex mtpcs-leaf-frame
5916 Generate a stack frame that is compliant with the Thumb Procedure Call
5917 Standard for all leaf functions. (A leaf function is one that does
5918 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
5920 @item -mcallee-super-interworking
5921 @opindex mcallee-super-interworking
5922 Gives all externally visible functions in the file being compiled an ARM
5923 instruction set header which switches to Thumb mode before executing the
5924 rest of the function. This allows these functions to be called from
5925 non-interworking code.
5927 @item -mcaller-super-interworking
5928 @opindex mcaller-super-interworking
5929 Allows calls via function pointers (including virtual functions) to
5930 execute correctly regardless of whether the target code has been
5931 compiled for interworking or not. There is a small overhead in the cost
5932 of executing a function pointer if this option is enabled.
5936 @node MN10200 Options
5937 @subsection MN10200 Options
5938 @cindex MN10200 options
5939 These @option{-m} options are defined for Matsushita MN10200 architectures:
5944 Indicate to the linker that it should perform a relaxation optimization pass
5945 to shorten branches, calls and absolute memory addresses. This option only
5946 has an effect when used on the command line for the final link step.
5948 This option makes symbolic debugging impossible.
5951 @node MN10300 Options
5952 @subsection MN10300 Options
5953 @cindex MN10300 options
5954 These @option{-m} options are defined for Matsushita MN10300 architectures:
5959 Generate code to avoid bugs in the multiply instructions for the MN10300
5960 processors. This is the default.
5963 @opindex mno-mult-bug
5964 Do not generate code to avoid bugs in the multiply instructions for the
5969 Generate code which uses features specific to the AM33 processor.
5973 Do not generate code which uses features specific to the AM33 processor. This
5978 Do not link in the C run-time initialization object file.
5982 Indicate to the linker that it should perform a relaxation optimization pass
5983 to shorten branches, calls and absolute memory addresses. This option only
5984 has an effect when used on the command line for the final link step.
5986 This option makes symbolic debugging impossible.
5990 @node M32R/D Options
5991 @subsection M32R/D Options
5992 @cindex M32R/D options
5994 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
5999 Generate code for the M32R/X@.
6003 Generate code for the M32R@. This is the default.
6005 @item -mcode-model=small
6006 @opindex mcode-model=small
6007 Assume all objects live in the lower 16MB of memory (so that their addresses
6008 can be loaded with the @code{ld24} instruction), and assume all subroutines
6009 are reachable with the @code{bl} instruction.
6010 This is the default.
6012 The addressability of a particular object can be set with the
6013 @code{model} attribute.
6015 @item -mcode-model=medium
6016 @opindex mcode-model=medium
6017 Assume objects may be anywhere in the 32-bit address space (the compiler
6018 will generate @code{seth/add3} instructions to load their addresses), and
6019 assume all subroutines are reachable with the @code{bl} instruction.
6021 @item -mcode-model=large
6022 @opindex mcode-model=large
6023 Assume objects may be anywhere in the 32-bit address space (the compiler
6024 will generate @code{seth/add3} instructions to load their addresses), and
6025 assume subroutines may not be reachable with the @code{bl} instruction
6026 (the compiler will generate the much slower @code{seth/add3/jl}
6027 instruction sequence).
6030 @opindex msdata=none
6031 Disable use of the small data area. Variables will be put into
6032 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6033 @code{section} attribute has been specified).
6034 This is the default.
6036 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6037 Objects may be explicitly put in the small data area with the
6038 @code{section} attribute using one of these sections.
6041 @opindex msdata=sdata
6042 Put small global and static data in the small data area, but do not
6043 generate special code to reference them.
6047 Put small global and static data in the small data area, and generate
6048 special instructions to reference them.
6052 @cindex smaller data references
6053 Put global and static objects less than or equal to @var{num} bytes
6054 into the small data or bss sections instead of the normal data or bss
6055 sections. The default value of @var{num} is 8.
6056 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6057 for this option to have any effect.
6059 All modules should be compiled with the same @option{-G @var{num}} value.
6060 Compiling with different values of @var{num} may or may not work; if it
6061 doesn't the linker will give an error message---incorrect code will not be
6067 @subsection M88K Options
6068 @cindex M88k options
6070 These @samp{-m} options are defined for Motorola 88k architectures:
6075 Generate code that works well on both the m88100 and the
6080 Generate code that works best for the m88100, but that also
6085 Generate code that works best for the m88110, and may not run
6090 Obsolete option to be removed from the next revision.
6093 @item -midentify-revision
6094 @opindex midentify-revision
6095 @cindex identifying source, compiler (88k)
6096 Include an @code{ident} directive in the assembler output recording the
6097 source file name, compiler name and version, timestamp, and compilation
6100 @item -mno-underscores
6101 @opindex mno-underscores
6102 @cindex underscores, avoiding (88k)
6103 In assembler output, emit symbol names without adding an underscore
6104 character at the beginning of each name. The default is to use an
6105 underscore as prefix on each name.
6107 @item -mocs-debug-info
6108 @itemx -mno-ocs-debug-info
6109 @opindex mocs-debug-info
6110 @opindex mno-ocs-debug-info
6112 @cindex debugging, 88k OCS
6113 Include (or omit) additional debugging information (about registers used
6114 in each stack frame) as specified in the 88open Object Compatibility
6115 Standard, ``OCS''@. This extra information allows debugging of code that
6116 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6117 Delta 88 SVr3.2 is to include this information; other 88k configurations
6118 omit this information by default.
6120 @item -mocs-frame-position
6121 @opindex mocs-frame-position
6122 @cindex register positions in frame (88k)
6123 When emitting COFF debugging information for automatic variables and
6124 parameters stored on the stack, use the offset from the canonical frame
6125 address, which is the stack pointer (register 31) on entry to the
6126 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6127 @option{-mocs-frame-position}; other 88k configurations have the default
6128 @option{-mno-ocs-frame-position}.
6130 @item -mno-ocs-frame-position
6131 @opindex mno-ocs-frame-position
6132 @cindex register positions in frame (88k)
6133 When emitting COFF debugging information for automatic variables and
6134 parameters stored on the stack, use the offset from the frame pointer
6135 register (register 30). When this option is in effect, the frame
6136 pointer is not eliminated when debugging information is selected by the
6139 @item -moptimize-arg-area
6140 @opindex moptimize-arg-area
6141 @cindex arguments in frame (88k)
6142 Save space by reorganizing the stack frame. This option generates code
6143 that does not agree with the 88open specifications, but uses less
6146 @itemx -mno-optimize-arg-area
6147 @opindex mno-optimize-arg-area
6148 Do not reorganize the stack frame to save space. This is the default.
6149 The generated conforms to the specification, but uses more memory.
6151 @item -mshort-data-@var{num}
6152 @opindex mshort-data
6153 @cindex smaller data references (88k)
6154 @cindex r0-relative references (88k)
6155 Generate smaller data references by making them relative to @code{r0},
6156 which allows loading a value using a single instruction (rather than the
6157 usual two). You control which data references are affected by
6158 specifying @var{num} with this option. For example, if you specify
6159 @option{-mshort-data-512}, then the data references affected are those
6160 involving displacements of less than 512 bytes.
6161 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6164 @item -mserialize-volatile
6165 @opindex mserialize-volatile
6166 @itemx -mno-serialize-volatile
6167 @opindex mno-serialize-volatile
6168 @cindex sequential consistency on 88k
6169 Do, or don't, generate code to guarantee sequential consistency
6170 of volatile memory references. By default, consistency is
6173 The order of memory references made by the MC88110 processor does
6174 not always match the order of the instructions requesting those
6175 references. In particular, a load instruction may execute before
6176 a preceding store instruction. Such reordering violates
6177 sequential consistency of volatile memory references, when there
6178 are multiple processors. When consistency must be guaranteed,
6179 GCC generates special instructions, as needed, to force
6180 execution in the proper order.
6182 The MC88100 processor does not reorder memory references and so
6183 always provides sequential consistency. However, by default, GCC
6184 generates the special instructions to guarantee consistency
6185 even when you use @option{-m88100}, so that the code may be run on an
6186 MC88110 processor. If you intend to run your code only on the
6187 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6189 The extra code generated to guarantee consistency may affect the
6190 performance of your application. If you know that you can safely
6191 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6197 @cindex assembler syntax, 88k
6199 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6200 related to System V release 4 (SVr4). This controls the following:
6204 Which variant of the assembler syntax to emit.
6206 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6207 that is used on System V release 4.
6209 @option{-msvr4} makes GCC issue additional declaration directives used in
6213 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6214 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6215 other m88k configurations.
6217 @item -mversion-03.00
6218 @opindex mversion-03.00
6219 This option is obsolete, and is ignored.
6220 @c ??? which asm syntax better for GAS? option there too?
6222 @item -mno-check-zero-division
6223 @itemx -mcheck-zero-division
6224 @opindex mno-check-zero-division
6225 @opindex mcheck-zero-division
6226 @cindex zero division on 88k
6227 Do, or don't, generate code to guarantee that integer division by
6228 zero will be detected. By default, detection is guaranteed.
6230 Some models of the MC88100 processor fail to trap upon integer
6231 division by zero under certain conditions. By default, when
6232 compiling code that might be run on such a processor, GCC
6233 generates code that explicitly checks for zero-valued divisors
6234 and traps with exception number 503 when one is detected. Use of
6235 @option{-mno-check-zero-division} suppresses such checking for code
6236 generated to run on an MC88100 processor.
6238 GCC assumes that the MC88110 processor correctly detects all instances
6239 of integer division by zero. When @option{-m88110} is specified, no
6240 explicit checks for zero-valued divisors are generated, and both
6241 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6244 @item -muse-div-instruction
6245 @opindex muse-div-instruction
6246 @cindex divide instruction, 88k
6247 Use the div instruction for signed integer division on the
6248 MC88100 processor. By default, the div instruction is not used.
6250 On the MC88100 processor the signed integer division instruction
6251 div) traps to the operating system on a negative operand. The
6252 operating system transparently completes the operation, but at a
6253 large cost in execution time. By default, when compiling code
6254 that might be run on an MC88100 processor, GCC emulates signed
6255 integer division using the unsigned integer division instruction
6256 divu), thereby avoiding the large penalty of a trap to the
6257 operating system. Such emulation has its own, smaller, execution
6258 cost in both time and space. To the extent that your code's
6259 important signed integer division operations are performed on two
6260 nonnegative operands, it may be desirable to use the div
6261 instruction directly.
6263 On the MC88110 processor the div instruction (also known as the
6264 divs instruction) processes negative operands without trapping to
6265 the operating system. When @option{-m88110} is specified,
6266 @option{-muse-div-instruction} is ignored, and the div instruction is used
6267 for signed integer division.
6269 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6270 particular, the behavior of such a division with and without
6271 @option{-muse-div-instruction} may differ.
6273 @item -mtrap-large-shift
6274 @itemx -mhandle-large-shift
6275 @opindex mtrap-large-shift
6276 @opindex mhandle-large-shift
6277 @cindex bit shift overflow (88k)
6278 @cindex large bit shifts (88k)
6279 Include code to detect bit-shifts of more than 31 bits; respectively,
6280 trap such shifts or emit code to handle them properly. By default GCC
6281 makes no special provision for large bit shifts.
6283 @item -mwarn-passed-structs
6284 @opindex mwarn-passed-structs
6285 @cindex structure passing (88k)
6286 Warn when a function passes a struct as an argument or result.
6287 Structure-passing conventions have changed during the evolution of the C
6288 language, and are often the source of portability problems. By default,
6289 GCC issues no such warning.
6292 @c break page here to avoid unsightly interparagraph stretch.
6296 @node RS/6000 and PowerPC Options
6297 @subsection IBM RS/6000 and PowerPC Options
6298 @cindex RS/6000 and PowerPC Options
6299 @cindex IBM RS/6000 and PowerPC Options
6301 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6309 @itemx -mpowerpc-gpopt
6310 @itemx -mno-powerpc-gpopt
6311 @itemx -mpowerpc-gfxopt
6312 @itemx -mno-powerpc-gfxopt
6314 @itemx -mno-powerpc64
6320 @opindex mno-powerpc
6321 @opindex mpowerpc-gpopt
6322 @opindex mno-powerpc-gpopt
6323 @opindex mpowerpc-gfxopt
6324 @opindex mno-powerpc-gfxopt
6326 @opindex mno-powerpc64
6327 GCC supports two related instruction set architectures for the
6328 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6329 instructions supported by the @samp{rios} chip set used in the original
6330 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6331 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6332 the IBM 4xx microprocessors.
6334 Neither architecture is a subset of the other. However there is a
6335 large common subset of instructions supported by both. An MQ
6336 register is included in processors supporting the POWER architecture.
6338 You use these options to specify which instructions are available on the
6339 processor you are using. The default value of these options is
6340 determined when configuring GCC@. Specifying the
6341 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6342 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6343 rather than the options listed above.
6345 The @option{-mpower} option allows GCC to generate instructions that
6346 are found only in the POWER architecture and to use the MQ register.
6347 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6348 to generate instructions that are present in the POWER2 architecture but
6349 not the original POWER architecture.
6351 The @option{-mpowerpc} option allows GCC to generate instructions that
6352 are found only in the 32-bit subset of the PowerPC architecture.
6353 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6354 GCC to use the optional PowerPC architecture instructions in the
6355 General Purpose group, including floating-point square root. Specifying
6356 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6357 use the optional PowerPC architecture instructions in the Graphics
6358 group, including floating-point select.
6360 The @option{-mpowerpc64} option allows GCC to generate the additional
6361 64-bit instructions that are found in the full PowerPC64 architecture
6362 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6363 @option{-mno-powerpc64}.
6365 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6366 will use only the instructions in the common subset of both
6367 architectures plus some special AIX common-mode calls, and will not use
6368 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6369 permits GCC to use any instruction from either architecture and to
6370 allow use of the MQ register; specify this for the Motorola MPC601.
6372 @item -mnew-mnemonics
6373 @itemx -mold-mnemonics
6374 @opindex mnew-mnemonics
6375 @opindex mold-mnemonics
6376 Select which mnemonics to use in the generated assembler code. With
6377 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6378 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6379 assembler mnemonics defined for the POWER architecture. Instructions
6380 defined in only one architecture have only one mnemonic; GCC uses that
6381 mnemonic irrespective of which of these options is specified.
6383 GCC defaults to the mnemonics appropriate for the architecture in
6384 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6385 value of these option. Unless you are building a cross-compiler, you
6386 should normally not specify either @option{-mnew-mnemonics} or
6387 @option{-mold-mnemonics}, but should instead accept the default.
6389 @item -mcpu=@var{cpu_type}
6391 Set architecture type, register usage, choice of mnemonics, and
6392 instruction scheduling parameters for machine type @var{cpu_type}.
6393 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6394 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6395 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6396 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6397 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6398 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6400 @option{-mcpu=common} selects a completely generic processor. Code
6401 generated under this option will run on any POWER or PowerPC processor.
6402 GCC will use only the instructions in the common subset of both
6403 architectures, and will not use the MQ register. GCC assumes a generic
6404 processor model for scheduling purposes.
6406 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6407 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6408 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6409 types, with an appropriate, generic processor model assumed for
6410 scheduling purposes.
6412 The other options specify a specific processor. Code generated under
6413 those options will run best on that processor, and may not run at all on
6416 The @option{-mcpu} options automatically enable or disable other
6417 @option{-m} options as follows:
6421 @option{-mno-power}, @option{-mno-powerc}
6428 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6443 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6446 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6451 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6454 @item -mtune=@var{cpu_type}
6456 Set the instruction scheduling parameters for machine type
6457 @var{cpu_type}, but do not set the architecture type, register usage, or
6458 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6459 values for @var{cpu_type} are used for @option{-mtune} as for
6460 @option{-mcpu}. If both are specified, the code generated will use the
6461 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6462 scheduling parameters set by @option{-mtune}.
6467 @opindex mno-altivec
6468 These switches enable or disable the use of built-in functions that
6469 allow access to the AltiVec instruction set. You may also need to set
6470 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6474 @itemx -mno-fp-in-toc
6475 @itemx -mno-sum-in-toc
6476 @itemx -mminimal-toc
6478 @opindex mno-fp-in-toc
6479 @opindex mno-sum-in-toc
6480 @opindex mminimal-toc
6481 Modify generation of the TOC (Table Of Contents), which is created for
6482 every executable file. The @option{-mfull-toc} option is selected by
6483 default. In that case, GCC will allocate at least one TOC entry for
6484 each unique non-automatic variable reference in your program. GCC
6485 will also place floating-point constants in the TOC@. However, only
6486 16,384 entries are available in the TOC@.
6488 If you receive a linker error message that saying you have overflowed
6489 the available TOC space, you can reduce the amount of TOC space used
6490 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6491 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6492 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6493 generate code to calculate the sum of an address and a constant at
6494 run-time instead of putting that sum into the TOC@. You may specify one
6495 or both of these options. Each causes GCC to produce very slightly
6496 slower and larger code at the expense of conserving TOC space.
6498 If you still run out of space in the TOC even when you specify both of
6499 these options, specify @option{-mminimal-toc} instead. This option causes
6500 GCC to make only one TOC entry for every file. When you specify this
6501 option, GCC will produce code that is slower and larger but which
6502 uses extremely little TOC space. You may wish to use this option
6503 only on files that contain less frequently executed code.
6509 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6510 @code{long} type, and the infrastructure needed to support them.
6511 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6512 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6513 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6518 @opindex mno-xl-call
6519 On AIX, pass floating-point arguments to prototyped functions beyond the
6520 register save area (RSA) on the stack in addition to argument FPRs. The
6521 AIX calling convention was extended but not initially documented to
6522 handle an obscure K&R C case of calling a function that takes the
6523 address of its arguments with fewer arguments than declared. AIX XL
6524 compilers access floating point arguments which do not fit in the
6525 RSA from the stack when a subroutine is compiled without
6526 optimization. Because always storing floating-point arguments on the
6527 stack is inefficient and rarely needed, this option is not enabled by
6528 default and only is necessary when calling subroutines compiled by AIX
6529 XL compilers without optimization.
6533 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6534 application written to use message passing with special startup code to
6535 enable the application to run. The system must have PE installed in the
6536 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6537 must be overridden with the @option{-specs=} option to specify the
6538 appropriate directory location. The Parallel Environment does not
6539 support threads, so the @option{-mpe} option and the @option{-pthread}
6540 option are incompatible.
6544 @opindex msoft-float
6545 @opindex mhard-float
6546 Generate code that does not use (uses) the floating-point register set.
6547 Software floating point emulation is provided if you use the
6548 @option{-msoft-float} option, and pass the option to GCC when linking.
6551 @itemx -mno-multiple
6553 @opindex mno-multiple
6554 Generate code that uses (does not use) the load multiple word
6555 instructions and the store multiple word instructions. These
6556 instructions are generated by default on POWER systems, and not
6557 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6558 endian PowerPC systems, since those instructions do not work when the
6559 processor is in little endian mode. The exceptions are PPC740 and
6560 PPC750 which permit the instructions usage in little endian mode.
6566 Generate code that uses (does not use) the load string instructions
6567 and the store string word instructions to save multiple registers and
6568 do small block moves. These instructions are generated by default on
6569 POWER systems, and not generated on PowerPC systems. Do not use
6570 @option{-mstring} on little endian PowerPC systems, since those
6571 instructions do not work when the processor is in little endian mode.
6572 The exceptions are PPC740 and PPC750 which permit the instructions
6573 usage in little endian mode.
6579 Generate code that uses (does not use) the load or store instructions
6580 that update the base register to the address of the calculated memory
6581 location. These instructions are generated by default. If you use
6582 @option{-mno-update}, there is a small window between the time that the
6583 stack pointer is updated and the address of the previous frame is
6584 stored, which means code that walks the stack frame across interrupts or
6585 signals may get corrupted data.
6588 @itemx -mno-fused-madd
6589 @opindex mfused-madd
6590 @opindex mno-fused-madd
6591 Generate code that uses (does not use) the floating point multiply and
6592 accumulate instructions. These instructions are generated by default if
6593 hardware floating is used.
6595 @item -mno-bit-align
6597 @opindex mno-bit-align
6599 On System V.4 and embedded PowerPC systems do not (do) force structures
6600 and unions that contain bit-fields to be aligned to the base type of the
6603 For example, by default a structure containing nothing but 8
6604 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6605 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6606 the structure would be aligned to a 1 byte boundary and be one byte in
6609 @item -mno-strict-align
6610 @itemx -mstrict-align
6611 @opindex mno-strict-align
6612 @opindex mstrict-align
6613 On System V.4 and embedded PowerPC systems do not (do) assume that
6614 unaligned memory references will be handled by the system.
6617 @itemx -mno-relocatable
6618 @opindex mrelocatable
6619 @opindex mno-relocatable
6620 On embedded PowerPC systems generate code that allows (does not allow)
6621 the program to be relocated to a different address at runtime. If you
6622 use @option{-mrelocatable} on any module, all objects linked together must
6623 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6625 @item -mrelocatable-lib
6626 @itemx -mno-relocatable-lib
6627 @opindex mrelocatable-lib
6628 @opindex mno-relocatable-lib
6629 On embedded PowerPC systems generate code that allows (does not allow)
6630 the program to be relocated to a different address at runtime. Modules
6631 compiled with @option{-mrelocatable-lib} can be linked with either modules
6632 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6633 with modules compiled with the @option{-mrelocatable} options.
6639 On System V.4 and embedded PowerPC systems do not (do) assume that
6640 register 2 contains a pointer to a global area pointing to the addresses
6641 used in the program.
6644 @itemx -mlittle-endian
6646 @opindex mlittle-endian
6647 On System V.4 and embedded PowerPC systems compile code for the
6648 processor in little endian mode. The @option{-mlittle-endian} option is
6649 the same as @option{-mlittle}.
6654 @opindex mbig-endian
6655 On System V.4 and embedded PowerPC systems compile code for the
6656 processor in big endian mode. The @option{-mbig-endian} option is
6657 the same as @option{-mbig}.
6661 On System V.4 and embedded PowerPC systems compile code using calling
6662 conventions that adheres to the March 1995 draft of the System V
6663 Application Binary Interface, PowerPC processor supplement. This is the
6664 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6666 @item -mcall-sysv-eabi
6667 @opindex mcall-sysv-eabi
6668 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6670 @item -mcall-sysv-noeabi
6671 @opindex mcall-sysv-noeabi
6672 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6676 On System V.4 and embedded PowerPC systems compile code using calling
6677 conventions that are similar to those used on AIX@. This is the
6678 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6680 @item -mcall-solaris
6681 @opindex mcall-solaris
6682 On System V.4 and embedded PowerPC systems compile code for the Solaris
6686 @opindex mcall-linux
6687 On System V.4 and embedded PowerPC systems compile code for the
6688 Linux-based GNU system.
6692 On System V.4 and embedded PowerPC systems compile code for the
6693 Hurd-based GNU system.
6696 @opindex mcall-netbsd
6697 On System V.4 and embedded PowerPC systems compile code for the
6698 NetBSD operating system.
6700 @item -maix-struct-return
6701 @opindex maix-struct-return
6702 Return all structures in memory (as specified by the AIX ABI)@.
6704 @item -msvr4-struct-return
6705 @opindex msvr4-struct-return
6706 Return structures smaller than 8 bytes in registers (as specified by the
6710 @opindex mabi=altivec
6711 Extend the current ABI with AltiVec ABI extensions. This does not
6712 change the default ABI, instead it adds the AltiVec ABI extensions to
6715 @item -mabi=no-altivec
6716 @opindex mabi=no-altivec
6717 Disable AltiVec ABI extensions for the current ABI.
6720 @itemx -mno-prototype
6722 @opindex mno-prototype
6723 On System V.4 and embedded PowerPC systems assume that all calls to
6724 variable argument functions are properly prototyped. Otherwise, the
6725 compiler must insert an instruction before every non prototyped call to
6726 set or clear bit 6 of the condition code register (@var{CR}) to
6727 indicate whether floating point values were passed in the floating point
6728 registers in case the function takes a variable arguments. With
6729 @option{-mprototype}, only calls to prototyped variable argument functions
6730 will set or clear the bit.
6734 On embedded PowerPC systems, assume that the startup module is called
6735 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6736 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6741 On embedded PowerPC systems, assume that the startup module is called
6742 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6747 On embedded PowerPC systems, assume that the startup module is called
6748 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6752 @opindex myellowknife
6753 On embedded PowerPC systems, assume that the startup module is called
6754 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6759 On System V.4 and embedded PowerPC systems, specify that you are
6760 compiling for a VxWorks system.
6764 Specify that you are compiling for the WindISS simulation environment.
6768 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
6769 header to indicate that @samp{eabi} extended relocations are used.
6775 On System V.4 and embedded PowerPC systems do (do not) adhere to the
6776 Embedded Applications Binary Interface (eabi) which is a set of
6777 modifications to the System V.4 specifications. Selecting @option{-meabi}
6778 means that the stack is aligned to an 8 byte boundary, a function
6779 @code{__eabi} is called to from @code{main} to set up the eabi
6780 environment, and the @option{-msdata} option can use both @code{r2} and
6781 @code{r13} to point to two separate small data areas. Selecting
6782 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
6783 do not call an initialization function from @code{main}, and the
6784 @option{-msdata} option will only use @code{r13} to point to a single
6785 small data area. The @option{-meabi} option is on by default if you
6786 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
6789 @opindex msdata=eabi
6790 On System V.4 and embedded PowerPC systems, put small initialized
6791 @code{const} global and static data in the @samp{.sdata2} section, which
6792 is pointed to by register @code{r2}. Put small initialized
6793 non-@code{const} global and static data in the @samp{.sdata} section,
6794 which is pointed to by register @code{r13}. Put small uninitialized
6795 global and static data in the @samp{.sbss} section, which is adjacent to
6796 the @samp{.sdata} section. The @option{-msdata=eabi} option is
6797 incompatible with the @option{-mrelocatable} option. The
6798 @option{-msdata=eabi} option also sets the @option{-memb} option.
6801 @opindex msdata=sysv
6802 On System V.4 and embedded PowerPC systems, put small global and static
6803 data in the @samp{.sdata} section, which is pointed to by register
6804 @code{r13}. Put small uninitialized global and static data in the
6805 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
6806 The @option{-msdata=sysv} option is incompatible with the
6807 @option{-mrelocatable} option.
6809 @item -msdata=default
6811 @opindex msdata=default
6813 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
6814 compile code the same as @option{-msdata=eabi}, otherwise compile code the
6815 same as @option{-msdata=sysv}.
6818 @opindex msdata-data
6819 On System V.4 and embedded PowerPC systems, put small global and static
6820 data in the @samp{.sdata} section. Put small uninitialized global and
6821 static data in the @samp{.sbss} section. Do not use register @code{r13}
6822 to address small data however. This is the default behavior unless
6823 other @option{-msdata} options are used.
6827 @opindex msdata=none
6829 On embedded PowerPC systems, put all initialized global and static data
6830 in the @samp{.data} section, and all uninitialized data in the
6831 @samp{.bss} section.
6835 @cindex smaller data references (PowerPC)
6836 @cindex .sdata/.sdata2 references (PowerPC)
6837 On embedded PowerPC systems, put global and static items less than or
6838 equal to @var{num} bytes into the small data or bss sections instead of
6839 the normal data or bss section. By default, @var{num} is 8. The
6840 @option{-G @var{num}} switch is also passed to the linker.
6841 All modules should be compiled with the same @option{-G @var{num}} value.
6844 @itemx -mno-regnames
6846 @opindex mno-regnames
6847 On System V.4 and embedded PowerPC systems do (do not) emit register
6848 names in the assembly language output using symbolic forms.
6851 @itemx -mno-longcall
6853 @opindex mno-longcall
6854 Default to making all function calls via pointers, so that functions
6855 which reside further than 64 megabytes (67,108,864 bytes) from the
6856 current location can be called. This setting can be overridden by the
6857 @code{shortcall} function attribute, or by @code{#pragma longcall(0)}.
6859 Some linkers are capable of detecting out-of-range calls and generating
6860 glue code on the fly. On these systems, long calls are unnecessary and
6861 generate slower code. As of this writing, the AIX linker can do this,
6862 as can the GNU linker for PowerPC/64. It is planned to add this feature
6863 to the GNU linker for 32-bit PowerPC systems as well.
6865 In the future, we may cause GCC to ignore all longcall specifications
6866 when the linker is known to generate glue.
6870 Adds support for multithreading with the @dfn{pthreads} library.
6871 This option sets flags for both the preprocessor and linker.
6876 @subsection IBM RT Options
6878 @cindex IBM RT options
6880 These @samp{-m} options are defined for the IBM RT PC:
6884 @opindex min-line-mul
6885 Use an in-line code sequence for integer multiplies. This is the
6888 @item -mcall-lib-mul
6889 @opindex mcall-lib-mul
6890 Call @code{lmul$$} for integer multiples.
6892 @item -mfull-fp-blocks
6893 @opindex mfull-fp-blocks
6894 Generate full-size floating point data blocks, including the minimum
6895 amount of scratch space recommended by IBM@. This is the default.
6897 @item -mminimum-fp-blocks
6898 @opindex mminimum-fp-blocks
6899 Do not include extra scratch space in floating point data blocks. This
6900 results in smaller code, but slower execution, since scratch space must
6901 be allocated dynamically.
6903 @cindex @file{varargs.h} and RT PC
6904 @cindex @file{stdarg.h} and RT PC
6905 @item -mfp-arg-in-fpregs
6906 @opindex mfp-arg-in-fpregs
6907 Use a calling sequence incompatible with the IBM calling convention in
6908 which floating point arguments are passed in floating point registers.
6909 Note that @code{varargs.h} and @code{stdarg.h} will not work with
6910 floating point operands if this option is specified.
6912 @item -mfp-arg-in-gregs
6913 @opindex mfp-arg-in-gregs
6914 Use the normal calling convention for floating point arguments. This is
6917 @item -mhc-struct-return
6918 @opindex mhc-struct-return
6919 Return structures of more than one word in memory, rather than in a
6920 register. This provides compatibility with the MetaWare HighC (hc)
6921 compiler. Use the option @option{-fpcc-struct-return} for compatibility
6922 with the Portable C Compiler (pcc).
6924 @item -mnohc-struct-return
6925 @opindex mnohc-struct-return
6926 Return some structures of more than one word in registers, when
6927 convenient. This is the default. For compatibility with the
6928 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
6929 option @option{-mhc-struct-return}.
6933 @subsection MIPS Options
6934 @cindex MIPS options
6936 These @samp{-m} options are defined for the MIPS family of computers:
6940 @item -march=@var{cpu-type}
6942 Assume the defaults for the machine type @var{cpu-type} when generating
6943 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
6944 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
6945 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
6946 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
6947 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
6948 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
6950 @item -mtune=@var{cpu-type}
6952 Assume the defaults for the machine type @var{cpu-type} when scheduling
6953 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
6954 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
6955 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
6956 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
6957 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
6958 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
6959 @var{cpu-type} will schedule things appropriately for that particular
6960 chip, the compiler will not generate any code that does not meet level 1
6961 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
6962 or @option{-mabi} switch being used.
6964 @item -mcpu=@var{cpu-type}
6966 This is identical to specifying both @option{-march} and @option{-mtune}.
6970 Issue instructions from level 1 of the MIPS ISA@. This is the default.
6971 @samp{r3000} is the default @var{cpu-type} at this ISA level.
6975 Issue instructions from level 2 of the MIPS ISA (branch likely, square
6976 root instructions). @samp{r6000} is the default @var{cpu-type} at this
6981 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
6982 @samp{r4000} is the default @var{cpu-type} at this ISA level.
6986 Issue instructions from level 4 of the MIPS ISA (conditional move,
6987 prefetch, enhanced FPU instructions). @samp{r8000} is the default
6988 @var{cpu-type} at this ISA level.
6992 Assume that 32 32-bit floating point registers are available. This is
6997 Assume that 32 64-bit floating point registers are available. This is
6998 the default when the @option{-mips3} option is used.
7001 @itemx -mno-fused-madd
7002 @opindex mfused-madd
7003 @opindex mno-fused-madd
7004 Generate code that uses (does not use) the floating point multiply and
7005 accumulate instructions, when they are available. These instructions
7006 are generated by default if they are available, but this may be
7007 undesirable if the extra precision causes problems or on certain chips
7008 in the mode where denormals are rounded to zero where denormals
7009 generated by multiply and accumulate instructions cause exceptions
7014 Assume that 32 32-bit general purpose registers are available. This is
7019 Assume that 32 64-bit general purpose registers are available. This is
7020 the default when the @option{-mips3} option is used.
7024 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7025 explanation of the default, and the width of pointers.
7029 Force long types to be 64 bits wide. See @option{-mlong32} for an
7030 explanation of the default, and the width of pointers.
7034 Force long, int, and pointer types to be 32 bits wide.
7036 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
7037 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
7038 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
7039 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
7040 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
7041 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
7042 are 32 bits, and longs are 64 bits wide. The width of pointer types is
7043 the smaller of the width of longs or the width of general purpose
7044 registers (which in turn depends on the ISA)@.
7056 Generate code for the indicated ABI@. The default instruction level is
7057 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
7058 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
7059 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
7064 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7065 add normal debug information. This is the default for all
7066 platforms except for the OSF/1 reference platform, using the OSF/rose
7067 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7068 switches are used, the @file{mips-tfile} program will encapsulate the
7069 stabs within MIPS ECOFF@.
7073 Generate code for the GNU assembler. This is the default on the OSF/1
7074 reference platform, using the OSF/rose object format. Also, this is
7075 the default if the configure option @option{--with-gnu-as} is used.
7077 @item -msplit-addresses
7078 @itemx -mno-split-addresses
7079 @opindex msplit-addresses
7080 @opindex mno-split-addresses
7081 Generate code to load the high and low parts of address constants separately.
7082 This allows GCC to optimize away redundant loads of the high order
7083 bits of addresses. This optimization requires GNU as and GNU ld.
7084 This optimization is enabled by default for some embedded targets where
7085 GNU as and GNU ld are standard.
7091 The @option{-mrnames} switch says to output code using the MIPS software
7092 names for the registers, instead of the hardware names (ie, @var{a0}
7093 instead of @var{$4}). The only known assembler that supports this option
7094 is the Algorithmics assembler.
7100 The @option{-mgpopt} switch says to write all of the data declarations
7101 before the instructions in the text section, this allows the MIPS
7102 assembler to generate one word memory references instead of using two
7103 words for short global or static data items. This is on by default if
7104 optimization is selected.
7110 For each non-inline function processed, the @option{-mstats} switch
7111 causes the compiler to emit one line to the standard error file to
7112 print statistics about the program (number of registers saved, stack
7119 The @option{-mmemcpy} switch makes all block moves call the appropriate
7120 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7121 generating inline code.
7124 @itemx -mno-mips-tfile
7125 @opindex mmips-tfile
7126 @opindex mno-mips-tfile
7127 The @option{-mno-mips-tfile} switch causes the compiler not
7128 postprocess the object file with the @file{mips-tfile} program,
7129 after the MIPS assembler has generated it to add debug support. If
7130 @file{mips-tfile} is not run, then no local variables will be
7131 available to the debugger. In addition, @file{stage2} and
7132 @file{stage3} objects will have the temporary file names passed to the
7133 assembler embedded in the object file, which means the objects will
7134 not compare the same. The @option{-mno-mips-tfile} switch should only
7135 be used when there are bugs in the @file{mips-tfile} program that
7136 prevents compilation.
7139 @opindex msoft-float
7140 Generate output containing library calls for floating point.
7141 @strong{Warning:} the requisite libraries are not part of GCC@.
7142 Normally the facilities of the machine's usual C compiler are used, but
7143 this can't be done directly in cross-compilation. You must make your
7144 own arrangements to provide suitable library functions for
7148 @opindex mhard-float
7149 Generate output containing floating point instructions. This is the
7150 default if you use the unmodified sources.
7153 @itemx -mno-abicalls
7155 @opindex mno-abicalls
7156 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7157 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7158 position independent code.
7161 @itemx -mno-long-calls
7162 @opindex mlong-calls
7163 @opindex mno-long-calls
7164 Do all calls with the @samp{JALR} instruction, which requires
7165 loading up a function's address into a register before the call.
7166 You need to use this switch, if you call outside of the current
7167 512 megabyte segment to functions that are not through pointers.
7170 @itemx -mno-half-pic
7172 @opindex mno-half-pic
7173 Put pointers to extern references into the data section and load them
7174 up, rather than put the references in the text section.
7176 @item -membedded-pic
7177 @itemx -mno-embedded-pic
7178 @opindex membedded-pic
7179 @opindex mno-embedded-pic
7180 Generate PIC code suitable for some embedded systems. All calls are
7181 made using PC relative address, and all data is addressed using the $gp
7182 register. No more than 65536 bytes of global data may be used. This
7183 requires GNU as and GNU ld which do most of the work. This currently
7184 only works on targets which use ECOFF; it does not work with ELF@.
7186 @item -membedded-data
7187 @itemx -mno-embedded-data
7188 @opindex membedded-data
7189 @opindex mno-embedded-data
7190 Allocate variables to the read-only data section first if possible, then
7191 next in the small data section if possible, otherwise in data. This gives
7192 slightly slower code than the default, but reduces the amount of RAM required
7193 when executing, and thus may be preferred for some embedded systems.
7195 @item -muninit-const-in-rodata
7196 @itemx -mno-uninit-const-in-rodata
7197 @opindex muninit-const-in-rodata
7198 @opindex mno-uninit-const-in-rodata
7199 When used together with @option{-membedded-data}, it will always store uninitialized
7200 const variables in the read-only data section.
7202 @item -msingle-float
7203 @itemx -mdouble-float
7204 @opindex msingle-float
7205 @opindex mdouble-float
7206 The @option{-msingle-float} switch tells gcc to assume that the floating
7207 point coprocessor only supports single precision operations, as on the
7208 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7209 double precision operations. This is the default.
7215 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7216 as on the @samp{r4650} chip.
7220 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7221 @option{-mcpu=r4650}.
7227 Enable 16-bit instructions.
7231 Use the entry and exit pseudo ops. This option can only be used with
7236 Compile code for the processor in little endian mode.
7237 The requisite libraries are assumed to exist.
7241 Compile code for the processor in big endian mode.
7242 The requisite libraries are assumed to exist.
7246 @cindex smaller data references (MIPS)
7247 @cindex gp-relative references (MIPS)
7248 Put global and static items less than or equal to @var{num} bytes into
7249 the small data or bss sections instead of the normal data or bss
7250 section. This allows the assembler to emit one word memory reference
7251 instructions based on the global pointer (@var{gp} or @var{$28}),
7252 instead of the normal two words used. By default, @var{num} is 8 when
7253 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7254 @option{-G @var{num}} switch is also passed to the assembler and linker.
7255 All modules should be compiled with the same @option{-G @var{num}}
7260 Tell the MIPS assembler to not run its preprocessor over user
7261 assembler files (with a @samp{.s} suffix) when assembling them.
7265 Pass an option to gas which will cause nops to be inserted if
7266 the read of the destination register of an mfhi or mflo instruction
7267 occurs in the following two instructions.
7271 Do not include the default crt0.
7273 @item -mflush-func=@var{func}
7274 @itemx -mno-flush-func
7275 @opindex mflush-func
7276 Specifies the function to call to flush the I and D caches, or to not
7277 call any such function. If called, the function must take the same
7278 arguments as the common @code{_flush_func()}, that is, the address of the
7279 memory range for which the cache is being flushed, the size of the
7280 memory range, and the number 3 (to flush both caches). The default
7281 depends on the target gcc was configured for, but commonly is either
7282 @samp{_flush_func} or @samp{__cpu_flush}.
7285 These options are defined by the macro
7286 @code{TARGET_SWITCHES} in the machine description. The default for the
7287 options is also defined by that macro, which enables you to change the
7290 @node i386 and x86-64 Options
7291 @subsection Intel 386 and AMD x86-64 Options
7292 @cindex i386 Options
7293 @cindex x86-64 Options
7294 @cindex Intel 386 Options
7295 @cindex AMD x86-64 Options
7297 These @samp{-m} options are defined for the i386 and x86-64 family of
7301 @item -mcpu=@var{cpu-type}
7303 Tune to @var{cpu-type} everything applicable about the generated code, except
7304 for the ABI and the set of available instructions. The choices for
7305 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7306 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7307 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7308 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp}
7309 and @samp{athlon-mp}.
7311 While picking a specific @var{cpu-type} will schedule things appropriately
7312 for that particular chip, the compiler will not generate any code that
7313 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7314 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7315 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7316 AMD chips as opposed to the Intel ones.
7318 @item -march=@var{cpu-type}
7320 Generate instructions for the machine type @var{cpu-type}. The choices
7321 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7322 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7331 @opindex mpentiumpro
7332 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7333 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7334 These synonyms are deprecated.
7336 @item -mfpmath=@var{unit}
7338 generate floating point arithmetics for selected unit @var{unit}. the choices
7343 Use the standard 387 floating point coprocessor present majority of chips and
7344 emulated otherwise. Code compiled with this option will run almost everywhere.
7345 The temporary results are computed in 80bit precesion instead of precision
7346 specified by the type resulting in slightly different results compared to most
7347 of other chips. See @option{-ffloat-store} for more detailed description.
7349 This is the default choice for i386 compiler.
7352 Use scalar floating point instructions present in the SSE instruction set.
7353 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7354 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7355 instruction set supports only single precision arithmetics, thus the double and
7356 extended precision arithmetics is still done using 387. Later version, present
7357 only in Pentium4 and the future AMD x86-64 chips supports double precision
7360 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7361 @option{-msse2} switches to enable SSE extensions and make this option
7362 effective. For x86-64 compiler, these extensions are enabled by default.
7364 The resulting code should be considerably faster in majority of cases and avoid
7365 the numerical instability problems of 387 code, but may break some existing
7366 code that expects temporaries to be 80bit.
7368 This is the default choice for x86-64 compiler.
7371 Attempt to utilize both instruction sets at once. This effectivly double the
7372 amount of available registers and on chips with separate execution units for
7373 387 and SSE the execution resources too. Use this option with care, as it is
7374 still experimental, because gcc register allocator does not model separate
7375 functional units well resulting in instable performance.
7378 @item -masm=@var{dialect}
7379 @opindex masm=@var{dialect}
7380 Output asm instructions using selected @var{dialect}. Supported choices are
7381 @samp{intel} or @samp{att} (the default one).
7386 @opindex mno-ieee-fp
7387 Control whether or not the compiler uses IEEE floating point
7388 comparisons. These handle correctly the case where the result of a
7389 comparison is unordered.
7392 @opindex msoft-float
7393 Generate output containing library calls for floating point.
7394 @strong{Warning:} the requisite libraries are not part of GCC@.
7395 Normally the facilities of the machine's usual C compiler are used, but
7396 this can't be done directly in cross-compilation. You must make your
7397 own arrangements to provide suitable library functions for
7400 On machines where a function returns floating point results in the 80387
7401 register stack, some floating point opcodes may be emitted even if
7402 @option{-msoft-float} is used.
7404 @item -mno-fp-ret-in-387
7405 @opindex mno-fp-ret-in-387
7406 Do not use the FPU registers for return values of functions.
7408 The usual calling convention has functions return values of types
7409 @code{float} and @code{double} in an FPU register, even if there
7410 is no FPU@. The idea is that the operating system should emulate
7413 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7414 in ordinary CPU registers instead.
7416 @item -mno-fancy-math-387
7417 @opindex mno-fancy-math-387
7418 Some 387 emulators do not support the @code{sin}, @code{cos} and
7419 @code{sqrt} instructions for the 387. Specify this option to avoid
7420 generating those instructions. This option is the default on FreeBSD,
7421 OpenBSD and NetBSD@. This option is overridden when @option{-march}
7422 indicates that the target cpu will always have an FPU and so the
7423 instruction will not need emulation. As of revision 2.6.1, these
7424 instructions are not generated unless you also use the
7425 @option{-funsafe-math-optimizations} switch.
7427 @item -malign-double
7428 @itemx -mno-align-double
7429 @opindex malign-double
7430 @opindex mno-align-double
7431 Control whether GCC aligns @code{double}, @code{long double}, and
7432 @code{long long} variables on a two word boundary or a one word
7433 boundary. Aligning @code{double} variables on a two word boundary will
7434 produce code that runs somewhat faster on a @samp{Pentium} at the
7435 expense of more memory.
7437 @strong{Warning:} if you use the @samp{-malign-double} switch,
7438 structures containing the above types will be aligned differently than
7439 the published application binary interface specifications for the 386.
7441 @item -m128bit-long-double
7442 @opindex m128bit-long-double
7443 Control the size of @code{long double} type. i386 application binary interface
7444 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7445 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7446 impossible to reach with 12 byte long doubles in the array accesses.
7448 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7449 structures and arrays containing @code{long double} will change their size as
7450 well as function calling convention for function taking @code{long double}
7453 @item -m96bit-long-double
7454 @opindex m96bit-long-double
7455 Set the size of @code{long double} to 96 bits as required by the i386
7456 application binary interface. This is the default.
7459 @itemx -mno-svr3-shlib
7460 @opindex msvr3-shlib
7461 @opindex mno-svr3-shlib
7462 Control whether GCC places uninitialized local variables into the
7463 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7464 into @code{bss}. These options are meaningful only on System V Release 3.
7468 Use a different function-calling convention, in which functions that
7469 take a fixed number of arguments return with the @code{ret} @var{num}
7470 instruction, which pops their arguments while returning. This saves one
7471 instruction in the caller since there is no need to pop the arguments
7474 You can specify that an individual function is called with this calling
7475 sequence with the function attribute @samp{stdcall}. You can also
7476 override the @option{-mrtd} option by using the function attribute
7477 @samp{cdecl}. @xref{Function Attributes}.
7479 @strong{Warning:} this calling convention is incompatible with the one
7480 normally used on Unix, so you cannot use it if you need to call
7481 libraries compiled with the Unix compiler.
7483 Also, you must provide function prototypes for all functions that
7484 take variable numbers of arguments (including @code{printf});
7485 otherwise incorrect code will be generated for calls to those
7488 In addition, seriously incorrect code will result if you call a
7489 function with too many arguments. (Normally, extra arguments are
7490 harmlessly ignored.)
7492 @item -mregparm=@var{num}
7494 Control how many registers are used to pass integer arguments. By
7495 default, no registers are used to pass arguments, and at most 3
7496 registers can be used. You can control this behavior for a specific
7497 function by using the function attribute @samp{regparm}.
7498 @xref{Function Attributes}.
7500 @strong{Warning:} if you use this switch, and
7501 @var{num} is nonzero, then you must build all modules with the same
7502 value, including any libraries. This includes the system libraries and
7505 @item -mpreferred-stack-boundary=@var{num}
7506 @opindex mpreferred-stack-boundary
7507 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7508 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7509 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7510 size (@option{-Os}), in which case the default is the minimum correct
7511 alignment (4 bytes for x86, and 8 bytes for x86-64).
7513 On Pentium and PentiumPro, @code{double} and @code{long double} values
7514 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7515 suffer significant run time performance penalties. On Pentium III, the
7516 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7517 penalties if it is not 16 byte aligned.
7519 To ensure proper alignment of this values on the stack, the stack boundary
7520 must be as aligned as that required by any value stored on the stack.
7521 Further, every function must be generated such that it keeps the stack
7522 aligned. Thus calling a function compiled with a higher preferred
7523 stack boundary from a function compiled with a lower preferred stack
7524 boundary will most likely misalign the stack. It is recommended that
7525 libraries that use callbacks always use the default setting.
7527 This extra alignment does consume extra stack space, and generally
7528 increases code size. Code that is sensitive to stack space usage, such
7529 as embedded systems and operating system kernels, may want to reduce the
7530 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7546 These switches enable or disable the use of built-in functions that allow
7547 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7549 @xref{X86 Built-in Functions}, for details of the functions enabled
7550 and disabled by these switches.
7553 @itemx -mno-push-args
7555 @opindex mno-push-args
7556 Use PUSH operations to store outgoing parameters. This method is shorter
7557 and usually equally fast as method using SUB/MOV operations and is enabled
7558 by default. In some cases disabling it may improve performance because of
7559 improved scheduling and reduced dependencies.
7561 @item -maccumulate-outgoing-args
7562 @opindex maccumulate-outgoing-args
7563 If enabled, the maximum amount of space required for outgoing arguments will be
7564 computed in the function prologue. This is faster on most modern CPUs
7565 because of reduced dependencies, improved scheduling and reduced stack usage
7566 when preferred stack boundary is not equal to 2. The drawback is a notable
7567 increase in code size. This switch implies @option{-mno-push-args}.
7571 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7572 on thread-safe exception handling must compile and link all code with the
7573 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7574 @option{-D_MT}; when linking, it links in a special thread helper library
7575 @option{-lmingwthrd} which cleans up per thread exception handling data.
7577 @item -mno-align-stringops
7578 @opindex mno-align-stringops
7579 Do not align destination of inlined string operations. This switch reduces
7580 code size and improves performance in case the destination is already aligned,
7581 but gcc don't know about it.
7583 @item -minline-all-stringops
7584 @opindex minline-all-stringops
7585 By default GCC inlines string operations only when destination is known to be
7586 aligned at least to 4 byte boundary. This enables more inlining, increase code
7587 size, but may improve performance of code that depends on fast memcpy, strlen
7588 and memset for short lengths.
7590 @item -momit-leaf-frame-pointer
7591 @opindex momit-leaf-frame-pointer
7592 Don't keep the frame pointer in a register for leaf functions. This
7593 avoids the instructions to save, set up and restore frame pointers and
7594 makes an extra register available in leaf functions. The option
7595 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7596 which might make debugging harder.
7599 These @samp{-m} switches are supported in addition to the above
7600 on AMD x86-64 processors in 64-bit environments.
7607 Generate code for a 32-bit or 64-bit environment.
7608 The 32-bit environment sets int, long and pointer to 32 bits and
7609 generates code that runs on any i386 system.
7610 The 64-bit environment sets int to 32 bits and long and pointer
7611 to 64 bits and generates code for AMD's x86-64 architecture.
7614 @opindex no-red-zone
7615 Do not use a so called red zone for x86-64 code. The red zone is mandated
7616 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7617 stack pointer that will not be modified by signal or interrupt handlers
7618 and therefore can be used for temporary data without adjusting the stack
7619 pointer. The flag @option{-mno-red-zone} disables this red zone.
7623 @subsection HPPA Options
7624 @cindex HPPA Options
7626 These @samp{-m} options are defined for the HPPA family of computers:
7629 @item -march=@var{architecture-type}
7631 Generate code for the specified architecture. The choices for
7632 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7633 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7634 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7635 architecture option for your machine. Code compiled for lower numbered
7636 architectures will run on higher numbered architectures, but not the
7639 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7640 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7644 @itemx -mpa-risc-1-1
7645 @itemx -mpa-risc-2-0
7646 @opindex mpa-risc-1-0
7647 @opindex mpa-risc-1-1
7648 @opindex mpa-risc-2-0
7649 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7652 @opindex mbig-switch
7653 Generate code suitable for big switch tables. Use this option only if
7654 the assembler/linker complain about out of range branches within a switch
7657 @item -mjump-in-delay
7658 @opindex mjump-in-delay
7659 Fill delay slots of function calls with unconditional jump instructions
7660 by modifying the return pointer for the function call to be the target
7661 of the conditional jump.
7663 @item -mdisable-fpregs
7664 @opindex mdisable-fpregs
7665 Prevent floating point registers from being used in any manner. This is
7666 necessary for compiling kernels which perform lazy context switching of
7667 floating point registers. If you use this option and attempt to perform
7668 floating point operations, the compiler will abort.
7670 @item -mdisable-indexing
7671 @opindex mdisable-indexing
7672 Prevent the compiler from using indexing address modes. This avoids some
7673 rather obscure problems when compiling MIG generated code under MACH@.
7675 @item -mno-space-regs
7676 @opindex mno-space-regs
7677 Generate code that assumes the target has no space registers. This allows
7678 GCC to generate faster indirect calls and use unscaled index address modes.
7680 Such code is suitable for level 0 PA systems and kernels.
7682 @item -mfast-indirect-calls
7683 @opindex mfast-indirect-calls
7684 Generate code that assumes calls never cross space boundaries. This
7685 allows GCC to emit code which performs faster indirect calls.
7687 This option will not work in the presence of shared libraries or nested
7690 @item -mlong-load-store
7691 @opindex mlong-load-store
7692 Generate 3-instruction load and store sequences as sometimes required by
7693 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7696 @item -mportable-runtime
7697 @opindex mportable-runtime
7698 Use the portable calling conventions proposed by HP for ELF systems.
7702 Enable the use of assembler directives only GAS understands.
7704 @item -mschedule=@var{cpu-type}
7706 Schedule code according to the constraints for the machine type
7707 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7708 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
7709 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
7710 proper scheduling option for your machine. The default scheduling is
7714 @opindex mlinker-opt
7715 Enable the optimization pass in the HPUX linker. Note this makes symbolic
7716 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
7717 in which they give bogus error messages when linking some programs.
7720 @opindex msoft-float
7721 Generate output containing library calls for floating point.
7722 @strong{Warning:} the requisite libraries are not available for all HPPA
7723 targets. Normally the facilities of the machine's usual C compiler are
7724 used, but this cannot be done directly in cross-compilation. You must make
7725 your own arrangements to provide suitable library functions for
7726 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7727 does provide software floating point support.
7729 @option{-msoft-float} changes the calling convention in the output file;
7730 therefore, it is only useful if you compile @emph{all} of a program with
7731 this option. In particular, you need to compile @file{libgcc.a}, the
7732 library that comes with GCC, with @option{-msoft-float} in order for
7736 @node Intel 960 Options
7737 @subsection Intel 960 Options
7739 These @samp{-m} options are defined for the Intel 960 implementations:
7742 @item -m@var{cpu-type}
7750 Assume the defaults for the machine type @var{cpu-type} for some of
7751 the other options, including instruction scheduling, floating point
7752 support, and addressing modes. The choices for @var{cpu-type} are
7753 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
7754 @samp{sa}, and @samp{sb}.
7761 @opindex msoft-float
7762 The @option{-mnumerics} option indicates that the processor does support
7763 floating-point instructions. The @option{-msoft-float} option indicates
7764 that floating-point support should not be assumed.
7766 @item -mleaf-procedures
7767 @itemx -mno-leaf-procedures
7768 @opindex mleaf-procedures
7769 @opindex mno-leaf-procedures
7770 Do (or do not) attempt to alter leaf procedures to be callable with the
7771 @code{bal} instruction as well as @code{call}. This will result in more
7772 efficient code for explicit calls when the @code{bal} instruction can be
7773 substituted by the assembler or linker, but less efficient code in other
7774 cases, such as calls via function pointers, or using a linker that doesn't
7775 support this optimization.
7778 @itemx -mno-tail-call
7780 @opindex mno-tail-call
7781 Do (or do not) make additional attempts (beyond those of the
7782 machine-independent portions of the compiler) to optimize tail-recursive
7783 calls into branches. You may not want to do this because the detection of
7784 cases where this is not valid is not totally complete. The default is
7785 @option{-mno-tail-call}.
7787 @item -mcomplex-addr
7788 @itemx -mno-complex-addr
7789 @opindex mcomplex-addr
7790 @opindex mno-complex-addr
7791 Assume (or do not assume) that the use of a complex addressing mode is a
7792 win on this implementation of the i960. Complex addressing modes may not
7793 be worthwhile on the K-series, but they definitely are on the C-series.
7794 The default is currently @option{-mcomplex-addr} for all processors except
7798 @itemx -mno-code-align
7799 @opindex mcode-align
7800 @opindex mno-code-align
7801 Align code to 8-byte boundaries for faster fetching (or don't bother).
7802 Currently turned on by default for C-series implementations only.
7805 @item -mclean-linkage
7806 @itemx -mno-clean-linkage
7807 @opindex mclean-linkage
7808 @opindex mno-clean-linkage
7809 These options are not fully implemented.
7813 @itemx -mic2.0-compat
7814 @itemx -mic3.0-compat
7816 @opindex mic2.0-compat
7817 @opindex mic3.0-compat
7818 Enable compatibility with iC960 v2.0 or v3.0.
7822 @opindex masm-compat
7824 Enable compatibility with the iC960 assembler.
7826 @item -mstrict-align
7827 @itemx -mno-strict-align
7828 @opindex mstrict-align
7829 @opindex mno-strict-align
7830 Do not permit (do permit) unaligned accesses.
7834 Enable structure-alignment compatibility with Intel's gcc release version
7835 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
7837 @item -mlong-double-64
7838 @opindex mlong-double-64
7839 Implement type @samp{long double} as 64-bit floating point numbers.
7840 Without the option @samp{long double} is implemented by 80-bit
7841 floating point numbers. The only reason we have it because there is
7842 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
7843 is only useful for people using soft-float targets. Otherwise, we
7844 should recommend against use of it.
7848 @node DEC Alpha Options
7849 @subsection DEC Alpha Options
7851 These @samp{-m} options are defined for the DEC Alpha implementations:
7854 @item -mno-soft-float
7856 @opindex mno-soft-float
7857 @opindex msoft-float
7858 Use (do not use) the hardware floating-point instructions for
7859 floating-point operations. When @option{-msoft-float} is specified,
7860 functions in @file{libgcc.a} will be used to perform floating-point
7861 operations. Unless they are replaced by routines that emulate the
7862 floating-point operations, or compiled in such a way as to call such
7863 emulations routines, these routines will issue floating-point
7864 operations. If you are compiling for an Alpha without floating-point
7865 operations, you must ensure that the library is built so as not to call
7868 Note that Alpha implementations without floating-point operations are
7869 required to have floating-point registers.
7874 @opindex mno-fp-regs
7875 Generate code that uses (does not use) the floating-point register set.
7876 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
7877 register set is not used, floating point operands are passed in integer
7878 registers as if they were integers and floating-point results are passed
7879 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
7880 so any function with a floating-point argument or return value called by code
7881 compiled with @option{-mno-fp-regs} must also be compiled with that
7884 A typical use of this option is building a kernel that does not use,
7885 and hence need not save and restore, any floating-point registers.
7889 The Alpha architecture implements floating-point hardware optimized for
7890 maximum performance. It is mostly compliant with the IEEE floating
7891 point standard. However, for full compliance, software assistance is
7892 required. This option generates code fully IEEE compliant code
7893 @emph{except} that the @var{inexact-flag} is not maintained (see below).
7894 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
7895 defined during compilation. The resulting code is less efficient but is
7896 able to correctly support denormalized numbers and exceptional IEEE
7897 values such as not-a-number and plus/minus infinity. Other Alpha
7898 compilers call this option @option{-ieee_with_no_inexact}.
7900 @item -mieee-with-inexact
7901 @opindex mieee-with-inexact
7902 This is like @option{-mieee} except the generated code also maintains
7903 the IEEE @var{inexact-flag}. Turning on this option causes the
7904 generated code to implement fully-compliant IEEE math. In addition to
7905 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
7906 macro. On some Alpha implementations the resulting code may execute
7907 significantly slower than the code generated by default. Since there is
7908 very little code that depends on the @var{inexact-flag}, you should
7909 normally not specify this option. Other Alpha compilers call this
7910 option @option{-ieee_with_inexact}.
7912 @item -mfp-trap-mode=@var{trap-mode}
7913 @opindex mfp-trap-mode
7914 This option controls what floating-point related traps are enabled.
7915 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
7916 The trap mode can be set to one of four values:
7920 This is the default (normal) setting. The only traps that are enabled
7921 are the ones that cannot be disabled in software (e.g., division by zero
7925 In addition to the traps enabled by @samp{n}, underflow traps are enabled
7929 Like @samp{su}, but the instructions are marked to be safe for software
7930 completion (see Alpha architecture manual for details).
7933 Like @samp{su}, but inexact traps are enabled as well.
7936 @item -mfp-rounding-mode=@var{rounding-mode}
7937 @opindex mfp-rounding-mode
7938 Selects the IEEE rounding mode. Other Alpha compilers call this option
7939 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
7944 Normal IEEE rounding mode. Floating point numbers are rounded towards
7945 the nearest machine number or towards the even machine number in case
7949 Round towards minus infinity.
7952 Chopped rounding mode. Floating point numbers are rounded towards zero.
7955 Dynamic rounding mode. A field in the floating point control register
7956 (@var{fpcr}, see Alpha architecture reference manual) controls the
7957 rounding mode in effect. The C library initializes this register for
7958 rounding towards plus infinity. Thus, unless your program modifies the
7959 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
7962 @item -mtrap-precision=@var{trap-precision}
7963 @opindex mtrap-precision
7964 In the Alpha architecture, floating point traps are imprecise. This
7965 means without software assistance it is impossible to recover from a
7966 floating trap and program execution normally needs to be terminated.
7967 GCC can generate code that can assist operating system trap handlers
7968 in determining the exact location that caused a floating point trap.
7969 Depending on the requirements of an application, different levels of
7970 precisions can be selected:
7974 Program precision. This option is the default and means a trap handler
7975 can only identify which program caused a floating point exception.
7978 Function precision. The trap handler can determine the function that
7979 caused a floating point exception.
7982 Instruction precision. The trap handler can determine the exact
7983 instruction that caused a floating point exception.
7986 Other Alpha compilers provide the equivalent options called
7987 @option{-scope_safe} and @option{-resumption_safe}.
7989 @item -mieee-conformant
7990 @opindex mieee-conformant
7991 This option marks the generated code as IEEE conformant. You must not
7992 use this option unless you also specify @option{-mtrap-precision=i} and either
7993 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
7994 is to emit the line @samp{.eflag 48} in the function prologue of the
7995 generated assembly file. Under DEC Unix, this has the effect that
7996 IEEE-conformant math library routines will be linked in.
7998 @item -mbuild-constants
7999 @opindex mbuild-constants
8000 Normally GCC examines a 32- or 64-bit integer constant to
8001 see if it can construct it from smaller constants in two or three
8002 instructions. If it cannot, it will output the constant as a literal and
8003 generate code to load it from the data segment at runtime.
8005 Use this option to require GCC to construct @emph{all} integer constants
8006 using code, even if it takes more instructions (the maximum is six).
8008 You would typically use this option to build a shared library dynamic
8009 loader. Itself a shared library, it must relocate itself in memory
8010 before it can find the variables and constants in its own data segment.
8016 Select whether to generate code to be assembled by the vendor-supplied
8017 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8035 Indicate whether GCC should generate code to use the optional BWX,
8036 CIX, FIX and MAX instruction sets. The default is to use the instruction
8037 sets supported by the CPU type specified via @option{-mcpu=} option or that
8038 of the CPU on which GCC was built if none was specified.
8043 @opindex mfloat-ieee
8044 Generate code that uses (does not use) VAX F and G floating point
8045 arithmetic instead of IEEE single and double precision.
8047 @item -mexplicit-relocs
8048 @itemx -mno-explicit-relocs
8049 @opindex mexplicit-relocs
8050 @opindex mno-explicit-relocs
8051 Older Alpha assemblers provided no way to generate symbol relocations
8052 except via assembler macros. Use of these macros does not allow
8053 optimial instruction scheduling. GNU binutils as of version 2.12
8054 supports a new syntax that allows the compiler to explicitly mark
8055 which relocations should apply to which instructions. This option
8056 is mostly useful for debugging, as GCC detects the capabilities of
8057 the assembler when it is built and sets the default accordingly.
8061 @opindex msmall-data
8062 @opindex mlarge-data
8063 When @option{-mexplicit-relocs} is in effect, static data is
8064 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8065 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8066 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8067 16-bit relocations off of the @code{$gp} register. This limits the
8068 size of the small data area to 64KB, but allows the variables to be
8069 directly accessed via a single instruction.
8071 The default is @option{-mlarge-data}. With this option the data area
8072 is limited to just below 2GB. Programs that require more than 2GB of
8073 data must use @code{malloc} or @code{mmap} to allocate the data in the
8074 heap instead of in the program's data segment.
8076 When generating code for shared libraries, @option{-fpic} implies
8077 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8079 @item -mcpu=@var{cpu_type}
8081 Set the instruction set and instruction scheduling parameters for
8082 machine type @var{cpu_type}. You can specify either the @samp{EV}
8083 style name or the corresponding chip number. GCC supports scheduling
8084 parameters for the EV4, EV5 and EV6 family of processors and will
8085 choose the default values for the instruction set from the processor
8086 you specify. If you do not specify a processor type, GCC will default
8087 to the processor on which the compiler was built.
8089 Supported values for @var{cpu_type} are
8095 Schedules as an EV4 and has no instruction set extensions.
8099 Schedules as an EV5 and has no instruction set extensions.
8103 Schedules as an EV5 and supports the BWX extension.
8108 Schedules as an EV5 and supports the BWX and MAX extensions.
8112 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8116 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8119 @item -mtune=@var{cpu_type}
8121 Set only the instruction scheduling parameters for machine type
8122 @var{cpu_type}. The instruction set is not changed.
8124 @item -mmemory-latency=@var{time}
8125 @opindex mmemory-latency
8126 Sets the latency the scheduler should assume for typical memory
8127 references as seen by the application. This number is highly
8128 dependent on the memory access patterns used by the application
8129 and the size of the external cache on the machine.
8131 Valid options for @var{time} are
8135 A decimal number representing clock cycles.
8141 The compiler contains estimates of the number of clock cycles for
8142 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8143 (also called Dcache, Scache, and Bcache), as well as to main memory.
8144 Note that L3 is only valid for EV5.
8149 @node DEC Alpha/VMS Options
8150 @subsection DEC Alpha/VMS Options
8152 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8155 @item -mvms-return-codes
8156 @opindex mvms-return-codes
8157 Return VMS condition codes from main. The default is to return POSIX
8158 style condition (e.g.@ error) codes.
8161 @node Clipper Options
8162 @subsection Clipper Options
8164 These @samp{-m} options are defined for the Clipper implementations:
8169 Produce code for a C300 Clipper processor. This is the default.
8173 Produce code for a C400 Clipper processor, i.e.@: use floating point
8177 @node H8/300 Options
8178 @subsection H8/300 Options
8180 These @samp{-m} options are defined for the H8/300 implementations:
8185 Shorten some address references at link time, when possible; uses the
8186 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8187 ld.info, Using ld}, for a fuller description.
8191 Generate code for the H8/300H@.
8195 Generate code for the H8/S@.
8199 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8203 Make @code{int} data 32 bits by default.
8207 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8208 The default for the H8/300H and H8/S is to align longs and floats on 4
8210 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8211 This option has no effect on the H8/300.
8215 @subsection SH Options
8217 These @samp{-m} options are defined for the SH implementations:
8222 Generate code for the SH1.
8226 Generate code for the SH2.
8230 Generate code for the SH3.
8234 Generate code for the SH3e.
8238 Generate code for the SH4 without a floating-point unit.
8240 @item -m4-single-only
8241 @opindex m4-single-only
8242 Generate code for the SH4 with a floating-point unit that only
8243 supports single-precision arithmetic.
8247 Generate code for the SH4 assuming the floating-point unit is in
8248 single-precision mode by default.
8252 Generate code for the SH4.
8256 Compile code for the processor in big endian mode.
8260 Compile code for the processor in little endian mode.
8264 Align doubles at 64-bit boundaries. Note that this changes the calling
8265 conventions, and thus some functions from the standard C library will
8266 not work unless you recompile it first with @option{-mdalign}.
8270 Shorten some address references at link time, when possible; uses the
8271 linker option @option{-relax}.
8275 Use 32-bit offsets in @code{switch} tables. The default is to use
8280 Enable the use of the instruction @code{fmovd}.
8284 Comply with the calling conventions defined by Hitachi.
8288 Mark the @code{MAC} register as call-clobbered, even if
8289 @option{-mhitachi} is given.
8293 Increase IEEE-compliance of floating-point code.
8297 Dump instruction size and location in the assembly code.
8301 This option is deprecated. It pads structures to multiple of 4 bytes,
8302 which is incompatible with the SH ABI@.
8306 Optimize for space instead of speed. Implied by @option{-Os}.
8310 When generating position-independent code, emit function calls using
8311 the Global Offset Table instead of the Procedure Linkage Table.
8315 Generate a library function call to invalidate instruction cache
8316 entries, after fixing up a trampoline. This library function call
8317 doesn't assume it can write to the whole memory address space. This
8318 is the default when the target is @code{sh-*-linux*}.
8321 @node System V Options
8322 @subsection Options for System V
8324 These additional options are available on System V Release 4 for
8325 compatibility with other compilers on those systems:
8330 Create a shared object.
8331 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8335 Identify the versions of each tool used by the compiler, in a
8336 @code{.ident} assembler directive in the output.
8340 Refrain from adding @code{.ident} directives to the output file (this is
8343 @item -YP,@var{dirs}
8345 Search the directories @var{dirs}, and no others, for libraries
8346 specified with @option{-l}.
8350 Look in the directory @var{dir} to find the M4 preprocessor.
8351 The assembler uses this option.
8352 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8353 @c the generic assembler that comes with Solaris takes just -Ym.
8356 @node TMS320C3x/C4x Options
8357 @subsection TMS320C3x/C4x Options
8358 @cindex TMS320C3x/C4x Options
8360 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8364 @item -mcpu=@var{cpu_type}
8366 Set the instruction set, register set, and instruction scheduling
8367 parameters for machine type @var{cpu_type}. Supported values for
8368 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8369 @samp{c44}. The default is @samp{c40} to generate code for the
8374 @itemx -msmall-memory
8376 @opindex mbig-memory
8378 @opindex msmall-memory
8380 Generates code for the big or small memory model. The small memory
8381 model assumed that all data fits into one 64K word page. At run-time
8382 the data page (DP) register must be set to point to the 64K page
8383 containing the .bss and .data program sections. The big memory model is
8384 the default and requires reloading of the DP register for every direct
8391 Allow (disallow) allocation of general integer operands into the block
8398 Enable (disable) generation of code using decrement and branch,
8399 DBcond(D), instructions. This is enabled by default for the C4x. To be
8400 on the safe side, this is disabled for the C3x, since the maximum
8401 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
8402 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
8403 that it can utilise the decrement and branch instruction, but will give
8404 up if there is more than one memory reference in the loop. Thus a loop
8405 where the loop counter is decremented can generate slightly more
8406 efficient code, in cases where the RPTB instruction cannot be utilised.
8408 @item -mdp-isr-reload
8410 @opindex mdp-isr-reload
8412 Force the DP register to be saved on entry to an interrupt service
8413 routine (ISR), reloaded to point to the data section, and restored on
8414 exit from the ISR@. This should not be required unless someone has
8415 violated the small memory model by modifying the DP register, say within
8422 For the C3x use the 24-bit MPYI instruction for integer multiplies
8423 instead of a library call to guarantee 32-bit results. Note that if one
8424 of the operands is a constant, then the multiplication will be performed
8425 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8426 then squaring operations are performed inline instead of a library call.
8429 @itemx -mno-fast-fix
8431 @opindex mno-fast-fix
8432 The C3x/C4x FIX instruction to convert a floating point value to an
8433 integer value chooses the nearest integer less than or equal to the
8434 floating point value rather than to the nearest integer. Thus if the
8435 floating point number is negative, the result will be incorrectly
8436 truncated an additional code is necessary to detect and correct this
8437 case. This option can be used to disable generation of the additional
8438 code required to correct the result.
8444 Enable (disable) generation of repeat block sequences using the RPTB
8445 instruction for zero overhead looping. The RPTB construct is only used
8446 for innermost loops that do not call functions or jump across the loop
8447 boundaries. There is no advantage having nested RPTB loops due to the
8448 overhead required to save and restore the RC, RS, and RE registers.
8449 This is enabled by default with @option{-O2}.
8451 @item -mrpts=@var{count}
8455 Enable (disable) the use of the single instruction repeat instruction
8456 RPTS@. If a repeat block contains a single instruction, and the loop
8457 count can be guaranteed to be less than the value @var{count}, GCC will
8458 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8459 then a RPTS will be emitted even if the loop count cannot be determined
8460 at compile time. Note that the repeated instruction following RPTS does
8461 not have to be reloaded from memory each iteration, thus freeing up the
8462 CPU buses for operands. However, since interrupts are blocked by this
8463 instruction, it is disabled by default.
8465 @item -mloop-unsigned
8466 @itemx -mno-loop-unsigned
8467 @opindex mloop-unsigned
8468 @opindex mno-loop-unsigned
8469 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8470 is @math{2^{31} + 1} since these instructions test if the iteration count is
8471 negative to terminate the loop. If the iteration count is unsigned
8472 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
8473 exceeded. This switch allows an unsigned iteration count.
8477 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8478 with. This also enforces compatibility with the API employed by the TI
8479 C3x C compiler. For example, long doubles are passed as structures
8480 rather than in floating point registers.
8486 Generate code that uses registers (stack) for passing arguments to functions.
8487 By default, arguments are passed in registers where possible rather
8488 than by pushing arguments on to the stack.
8490 @item -mparallel-insns
8491 @itemx -mno-parallel-insns
8492 @opindex mparallel-insns
8493 @opindex mno-parallel-insns
8494 Allow the generation of parallel instructions. This is enabled by
8495 default with @option{-O2}.
8497 @item -mparallel-mpy
8498 @itemx -mno-parallel-mpy
8499 @opindex mparallel-mpy
8500 @opindex mno-parallel-mpy
8501 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8502 provided @option{-mparallel-insns} is also specified. These instructions have
8503 tight register constraints which can pessimize the code generation
8509 @subsection V850 Options
8510 @cindex V850 Options
8512 These @samp{-m} options are defined for V850 implementations:
8516 @itemx -mno-long-calls
8517 @opindex mlong-calls
8518 @opindex mno-long-calls
8519 Treat all calls as being far away (near). If calls are assumed to be
8520 far away, the compiler will always load the functions address up into a
8521 register, and call indirect through the pointer.
8527 Do not optimize (do optimize) basic blocks that use the same index
8528 pointer 4 or more times to copy pointer into the @code{ep} register, and
8529 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8530 option is on by default if you optimize.
8532 @item -mno-prolog-function
8533 @itemx -mprolog-function
8534 @opindex mno-prolog-function
8535 @opindex mprolog-function
8536 Do not use (do use) external functions to save and restore registers at
8537 the prolog and epilog of a function. The external functions are slower,
8538 but use less code space if more than one function saves the same number
8539 of registers. The @option{-mprolog-function} option is on by default if
8544 Try to make the code as small as possible. At present, this just turns
8545 on the @option{-mep} and @option{-mprolog-function} options.
8549 Put static or global variables whose size is @var{n} bytes or less into
8550 the tiny data area that register @code{ep} points to. The tiny data
8551 area can hold up to 256 bytes in total (128 bytes for byte references).
8555 Put static or global variables whose size is @var{n} bytes or less into
8556 the small data area that register @code{gp} points to. The small data
8557 area can hold up to 64 kilobytes.
8561 Put static or global variables whose size is @var{n} bytes or less into
8562 the first 32 kilobytes of memory.
8566 Specify that the target processor is the V850.
8569 @opindex mbig-switch
8570 Generate code suitable for big switch tables. Use this option only if
8571 the assembler/linker complain about out of range branches within a switch
8576 @subsection ARC Options
8579 These options are defined for ARC implementations:
8584 Compile code for little endian mode. This is the default.
8588 Compile code for big endian mode.
8591 @opindex mmangle-cpu
8592 Prepend the name of the cpu to all public symbol names.
8593 In multiple-processor systems, there are many ARC variants with different
8594 instruction and register set characteristics. This flag prevents code
8595 compiled for one cpu to be linked with code compiled for another.
8596 No facility exists for handling variants that are ``almost identical''.
8597 This is an all or nothing option.
8599 @item -mcpu=@var{cpu}
8601 Compile code for ARC variant @var{cpu}.
8602 Which variants are supported depend on the configuration.
8603 All variants support @option{-mcpu=base}, this is the default.
8605 @item -mtext=@var{text-section}
8606 @itemx -mdata=@var{data-section}
8607 @itemx -mrodata=@var{readonly-data-section}
8611 Put functions, data, and readonly data in @var{text-section},
8612 @var{data-section}, and @var{readonly-data-section} respectively
8613 by default. This can be overridden with the @code{section} attribute.
8614 @xref{Variable Attributes}.
8619 @subsection NS32K Options
8620 @cindex NS32K options
8622 These are the @samp{-m} options defined for the 32000 series. The default
8623 values for these options depends on which style of 32000 was selected when
8624 the compiler was configured; the defaults for the most common choices are
8632 Generate output for a 32032. This is the default
8633 when the compiler is configured for 32032 and 32016 based systems.
8639 Generate output for a 32332. This is the default
8640 when the compiler is configured for 32332-based systems.
8646 Generate output for a 32532. This is the default
8647 when the compiler is configured for 32532-based systems.
8651 Generate output containing 32081 instructions for floating point.
8652 This is the default for all systems.
8656 Generate output containing 32381 instructions for floating point. This
8657 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8658 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8662 Try and generate multiply-add floating point instructions @code{polyF}
8663 and @code{dotF}. This option is only available if the @option{-m32381}
8664 option is in effect. Using these instructions requires changes to
8665 register allocation which generally has a negative impact on
8666 performance. This option should only be enabled when compiling code
8667 particularly likely to make heavy use of multiply-add instructions.
8670 @opindex mnomulti-add
8671 Do not try and generate multiply-add floating point instructions
8672 @code{polyF} and @code{dotF}. This is the default on all platforms.
8675 @opindex msoft-float
8676 Generate output containing library calls for floating point.
8677 @strong{Warning:} the requisite libraries may not be available.
8680 @opindex mnobitfield
8681 Do not use the bit-field instructions. On some machines it is faster to
8682 use shifting and masking operations. This is the default for the pc532.
8686 Do use the bit-field instructions. This is the default for all platforms
8691 Use a different function-calling convention, in which functions
8692 that take a fixed number of arguments return pop their
8693 arguments on return with the @code{ret} instruction.
8695 This calling convention is incompatible with the one normally
8696 used on Unix, so you cannot use it if you need to call libraries
8697 compiled with the Unix compiler.
8699 Also, you must provide function prototypes for all functions that
8700 take variable numbers of arguments (including @code{printf});
8701 otherwise incorrect code will be generated for calls to those
8704 In addition, seriously incorrect code will result if you call a
8705 function with too many arguments. (Normally, extra arguments are
8706 harmlessly ignored.)
8708 This option takes its name from the 680x0 @code{rtd} instruction.
8713 Use a different function-calling convention where the first two arguments
8714 are passed in registers.
8716 This calling convention is incompatible with the one normally
8717 used on Unix, so you cannot use it if you need to call libraries
8718 compiled with the Unix compiler.
8721 @opindex mnoregparam
8722 Do not pass any arguments in registers. This is the default for all
8727 It is OK to use the sb as an index register which is always loaded with
8728 zero. This is the default for the pc532-netbsd target.
8732 The sb register is not available for use or has not been initialized to
8733 zero by the run time system. This is the default for all targets except
8734 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
8735 @option{-fpic} is set.
8739 Many ns32000 series addressing modes use displacements of up to 512MB@.
8740 If an address is above 512MB then displacements from zero can not be used.
8741 This option causes code to be generated which can be loaded above 512MB@.
8742 This may be useful for operating systems or ROM code.
8746 Assume code will be loaded in the first 512MB of virtual address space.
8747 This is the default for all platforms.
8753 @subsection AVR Options
8756 These options are defined for AVR implementations:
8759 @item -mmcu=@var{mcu}
8761 Specify ATMEL AVR instruction set or MCU type.
8763 Instruction set avr1 is for the minimal AVR core, not supported by the C
8764 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8765 attiny11, attiny12, attiny15, attiny28).
8767 Instruction set avr2 (default) is for the classic AVR core with up to
8768 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8769 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8770 at90c8534, at90s8535).
8772 Instruction set avr3 is for the classic AVR core with up to 128K program
8773 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8775 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8776 memory space (MCU types: atmega8, atmega83, atmega85).
8778 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8779 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8780 atmega64, atmega128, at43usb355, at94k).
8784 Output instruction sizes to the asm file.
8786 @item -minit-stack=@var{N}
8787 @opindex minit-stack
8788 Specify the initial stack address, which may be a symbol or numeric value,
8789 @samp{__stack} is the default.
8791 @item -mno-interrupts
8792 @opindex mno-interrupts
8793 Generated code is not compatible with hardware interrupts.
8794 Code size will be smaller.
8796 @item -mcall-prologues
8797 @opindex mcall-prologues
8798 Functions prologues/epilogues expanded as call to appropriate
8799 subroutines. Code size will be smaller.
8801 @item -mno-tablejump
8802 @opindex mno-tablejump
8803 Do not generate tablejump insns which sometimes increase code size.
8806 @opindex mtiny-stack
8807 Change only the low 8 bits of the stack pointer.
8811 @subsection MCore Options
8812 @cindex MCore options
8814 These are the @samp{-m} options defined for the Motorola M*Core
8824 @opindex mno-hardlit
8825 Inline constants into the code stream if it can be done in two
8826 instructions or less.
8834 Use the divide instruction. (Enabled by default).
8836 @item -mrelax-immediate
8837 @itemx -mrelax-immediate
8838 @itemx -mno-relax-immediate
8839 @opindex mrelax-immediate
8840 @opindex mrelax-immediate
8841 @opindex mno-relax-immediate
8842 Allow arbitrary sized immediates in bit operations.
8844 @item -mwide-bitfields
8845 @itemx -mwide-bitfields
8846 @itemx -mno-wide-bitfields
8847 @opindex mwide-bitfields
8848 @opindex mwide-bitfields
8849 @opindex mno-wide-bitfields
8850 Always treat bit-fields as int-sized.
8852 @item -m4byte-functions
8853 @itemx -m4byte-functions
8854 @itemx -mno-4byte-functions
8855 @opindex m4byte-functions
8856 @opindex m4byte-functions
8857 @opindex mno-4byte-functions
8858 Force all functions to be aligned to a four byte boundary.
8860 @item -mcallgraph-data
8861 @itemx -mcallgraph-data
8862 @itemx -mno-callgraph-data
8863 @opindex mcallgraph-data
8864 @opindex mcallgraph-data
8865 @opindex mno-callgraph-data
8866 Emit callgraph information.
8870 @itemx -mno-slow-bytes
8871 @opindex mslow-bytes
8872 @opindex mslow-bytes
8873 @opindex mno-slow-bytes
8874 Prefer word access when reading byte quantities.
8876 @item -mlittle-endian
8877 @itemx -mlittle-endian
8879 @opindex mlittle-endian
8880 @opindex mlittle-endian
8881 @opindex mbig-endian
8882 Generate code for a little endian target.
8890 Generate code for the 210 processor.
8894 @subsection IA-64 Options
8895 @cindex IA-64 Options
8897 These are the @samp{-m} options defined for the Intel IA-64 architecture.
8901 @opindex mbig-endian
8902 Generate code for a big endian target. This is the default for HPUX@.
8904 @item -mlittle-endian
8905 @opindex mlittle-endian
8906 Generate code for a little endian target. This is the default for AIX5
8913 Generate (or don't) code for the GNU assembler. This is the default.
8914 @c Also, this is the default if the configure option @option{--with-gnu-as}
8921 Generate (or don't) code for the GNU linker. This is the default.
8922 @c Also, this is the default if the configure option @option{--with-gnu-ld}
8927 Generate code that does not use a global pointer register. The result
8928 is not position independent code, and violates the IA-64 ABI@.
8930 @item -mvolatile-asm-stop
8931 @itemx -mno-volatile-asm-stop
8932 @opindex mvolatile-asm-stop
8933 @opindex mno-volatile-asm-stop
8934 Generate (or don't) a stop bit immediately before and after volatile asm
8939 Generate code that works around Itanium B step errata.
8941 @item -mregister-names
8942 @itemx -mno-register-names
8943 @opindex mregister-names
8944 @opindex mno-register-names
8945 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
8946 the stacked registers. This may make assembler output more readable.
8952 Disable (or enable) optimizations that use the small data section. This may
8953 be useful for working around optimizer bugs.
8956 @opindex mconstant-gp
8957 Generate code that uses a single constant global pointer value. This is
8958 useful when compiling kernel code.
8962 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
8963 This is useful when compiling firmware code.
8965 @item -minline-divide-min-latency
8966 @opindex minline-divide-min-latency
8967 Generate code for inline divides using the minimum latency algorithm.
8969 @item -minline-divide-max-throughput
8970 @opindex minline-divide-max-throughput
8971 Generate code for inline divides using the maximum throughput algorithm.
8973 @item -mno-dwarf2-asm
8975 @opindex mno-dwarf2-asm
8976 @opindex mdwarf2-asm
8977 Don't (or do) generate assembler code for the DWARF2 line number debugging
8978 info. This may be useful when not using the GNU assembler.
8980 @item -mfixed-range=@var{register-range}
8981 @opindex mfixed-range
8982 Generate code treating the given register range as fixed registers.
8983 A fixed register is one that the register allocator can not use. This is
8984 useful when compiling kernel code. A register range is specified as
8985 two registers separated by a dash. Multiple register ranges can be
8986 specified separated by a comma.
8990 @subsection D30V Options
8991 @cindex D30V Options
8993 These @samp{-m} options are defined for D30V implementations:
8998 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
8999 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9000 memory, which starts at location @code{0x80000000}.
9004 Same as the @option{-mextmem} switch.
9008 Link the @samp{.text} section into onchip text memory, which starts at
9009 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9010 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9011 into onchip data memory, which starts at location @code{0x20000000}.
9013 @item -mno-asm-optimize
9014 @itemx -masm-optimize
9015 @opindex mno-asm-optimize
9016 @opindex masm-optimize
9017 Disable (enable) passing @option{-O} to the assembler when optimizing.
9018 The assembler uses the @option{-O} option to automatically parallelize
9019 adjacent short instructions where possible.
9021 @item -mbranch-cost=@var{n}
9022 @opindex mbranch-cost
9023 Increase the internal costs of branches to @var{n}. Higher costs means
9024 that the compiler will issue more instructions to avoid doing a branch.
9027 @item -mcond-exec=@var{n}
9029 Specify the maximum number of conditionally executed instructions that
9030 replace a branch. The default is 4.
9033 @node S/390 and zSeries Options
9034 @subsection S/390 and zSeries Options
9035 @cindex S/390 and zSeries Options
9037 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9042 @opindex mhard-float
9043 @opindex msoft-float
9044 Use (do not use) the hardware floating-point instructions and registers
9045 for floating-point operations. When @option{-msoft-float} is specified,
9046 functions in @file{libgcc.a} will be used to perform floating-point
9047 operations. When @option{-mhard-float} is specified, the compiler
9048 generates IEEE floating-point instructions. This is the default.
9051 @itemx -mno-backchain
9053 @opindex mno-backchain
9054 Generate (or do not generate) code which maintains an explicit
9055 backchain within the stack frame that points to the caller's frame.
9056 This is currently needed to allow debugging. The default is to
9057 generate the backchain.
9060 @itemx -mno-small-exec
9061 @opindex msmall-exec
9062 @opindex mno-small-exec
9063 Generate (or do not generate) code using the @code{bras} instruction
9064 to do subroutine calls.
9065 This only works reliably if the total executable size does not
9066 exceed 64k. The default is to use the @code{basr} instruction instead,
9067 which does not have this limitation.
9073 When @option{-m31} is specified, generate code compliant to the
9074 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9075 code compliant to the Linux for zSeries ABI@. This allows GCC in
9076 particular to generate 64-bit instructions. For the @samp{s390}
9077 targets, the default is @option{-m31}, while the @samp{s390x}
9078 targets default to @option{-m64}.
9084 Generate (or do not generate) code using the @code{mvcle} instruction
9085 to perform block moves. When @option{-mno-mvcle} is specifed,
9086 use a @code{mvc} loop instead. This is the default.
9092 Print (or do not print) additional debug information when compiling.
9093 The default is to not print debug information.
9098 @subsection CRIS Options
9099 @cindex CRIS Options
9101 These options are defined specifically for the CRIS ports.
9104 @item -march=@var{architecture-type}
9105 @itemx -mcpu=@var{architecture-type}
9108 Generate code for the specified architecture. The choices for
9109 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9110 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9111 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9114 @item -mtune=@var{architecture-type}
9116 Tune to @var{architecture-type} everything applicable about the generated
9117 code, except for the ABI and the set of available instructions. The
9118 choices for @var{architecture-type} are the same as for
9119 @option{-march=@var{architecture-type}}.
9121 @item -mmax-stack-frame=@var{n}
9122 @opindex mmax-stack-frame
9123 Warn when the stack frame of a function exceeds @var{n} bytes.
9125 @item -melinux-stacksize=@var{n}
9126 @opindex melinux-stacksize
9127 Only available with the @samp{cris-axis-aout} target. Arranges for
9128 indications in the program to the kernel loader that the stack of the
9129 program should be set to @var{n} bytes.
9135 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9136 @option{-march=v3} and @option{-march=v8} respectively.
9140 Enable CRIS-specific verbose debug-related information in the assembly
9141 code. This option also has the effect to turn off the @samp{#NO_APP}
9142 formatted-code indicator to the assembler at the beginning of the
9147 Do not use condition-code results from previous instruction; always emit
9148 compare and test instructions before use of condition codes.
9150 @item -mno-side-effects
9151 @opindex mno-side-effects
9152 Do not emit instructions with side-effects in addressing modes other than
9156 @itemx -mno-stack-align
9158 @itemx -mno-data-align
9159 @itemx -mconst-align
9160 @itemx -mno-const-align
9161 @opindex mstack-align
9162 @opindex mno-stack-align
9163 @opindex mdata-align
9164 @opindex mno-data-align
9165 @opindex mconst-align
9166 @opindex mno-const-align
9167 These options (no-options) arranges (eliminate arrangements) for the
9168 stack-frame, individual data and constants to be aligned for the maximum
9169 single data access size for the chosen CPU model. The default is to
9170 arrange for 32-bit alignment. ABI details such as structure layout are
9171 not affected by these options.
9179 Similar to the stack- data- and const-align options above, these options
9180 arrange for stack-frame, writable data and constants to all be 32-bit,
9181 16-bit or 8-bit aligned. The default is 32-bit alignment.
9183 @item -mno-prologue-epilogue
9184 @itemx -mprologue-epilogue
9185 @opindex mno-prologue-epilogue
9186 @opindex mprologue-epilogue
9187 With @option{-mno-prologue-epilogue}, the normal function prologue and
9188 epilogue that sets up the stack-frame are omitted and no return
9189 instructions or return sequences are generated in the code. Use this
9190 option only together with visual inspection of the compiled code: no
9191 warnings or errors are generated when call-saved registers must be saved,
9192 or storage for local variable needs to be allocated.
9198 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9199 instruction sequences that load addresses for functions from the PLT part
9200 of the GOT rather than (traditional on other architectures) calls to the
9201 PLT. The default is @option{-mgotplt}.
9205 Legacy no-op option only recognized with the cris-axis-aout target.
9209 Legacy no-op option only recognized with the cris-axis-elf and
9210 cris-axis-linux-gnu targets.
9214 Only recognized with the cris-axis-aout target, where it selects a
9215 GNU/linux-like multilib, include files and instruction set for
9220 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9224 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9225 to link with input-output functions from a simulator library. Code,
9226 initialized data and zero-initialized data are allocated consecutively.
9230 Like @option{-sim}, but pass linker options to locate initialized data at
9231 0x40000000 and zero-initialized data at 0x80000000.
9235 @subsection MMIX Options
9236 @cindex MMIX Options
9238 These options are defined for the MMIX:
9242 @itemx -mno-libfuncs
9244 @opindex mno-libfuncs
9245 Specify that intrinsic library functions are being compiled, passing all
9246 values in registers, no matter the size.
9251 @opindex mno-epsilon
9252 Generate floating-point comparison instructions that compare with respect
9253 to the @code{rE} epsilon register.
9255 @item -mabi=mmixware
9257 @opindex mabi-mmixware
9259 Generate code that passes function parameters and return values that (in
9260 the called function) are seen as registers @code{$0} and up, as opposed to
9261 the GNU ABI which uses global registers @code{$231} and up.
9264 @itemx -mno-zero-extend
9265 @opindex mzero-extend
9266 @opindex mno-zero-extend
9267 When reading data from memory in sizes shorter than 64 bits, use (do not
9268 use) zero-extending load instructions by default, rather than
9269 sign-extending ones.
9272 @itemx -mno-knuthdiv
9274 @opindex mno-knuthdiv
9275 Make the result of a division yielding a remainder have the same sign as
9276 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9277 remainder follows the sign of the dividend. Both methods are
9278 arithmetically valid, the latter being almost exclusively used.
9280 @item -mtoplevel-symbols
9281 @itemx -mno-toplevel-symbols
9282 @opindex mtoplevel-symbols
9283 @opindex mno-toplevel-symbols
9284 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9285 code can be used with the @code{PREFIX} assembly directive.
9289 Generate an executable in the ELF format, rather than the default
9290 @samp{mmo} format used by the @command{mmix} simulator.
9292 @item -mbranch-predict
9293 @itemx -mno-branch-predict
9294 @opindex mbranch-predict
9295 @opindex mno-branch-predict
9296 Use (do not use) the probable-branch instructions, when static branch
9297 prediction indicates a probable branch.
9299 @item -mbase-addresses
9300 @itemx -mno-base-addresses
9301 @opindex mbase-addresses
9302 @opindex mno-base-addresses
9303 Generate (do not generate) code that uses @emph{base addresses}. Using a
9304 base address automatically generates a request (handled by the assembler
9305 and the linker) for a constant to be set up in a global register. The
9306 register is used for one or more base address requests within the range 0
9307 to 255 from the value held in the register. The generally leads to short
9308 and fast code, but the number of different data items that can be
9309 addressed is limited. This means that a program that uses lots of static
9310 data may require @option{-mno-base-addresses}.
9313 @node PDP-11 Options
9314 @subsection PDP-11 Options
9315 @cindex PDP-11 Options
9317 These options are defined for the PDP-11:
9322 Use hardware FPP floating point. This is the default. (FIS floating
9323 point on the PDP-11/40 is not supported.)
9326 @opindex msoft-float
9327 Do not use hardware floating point.
9331 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9335 Return floating-point results in memory. This is the default.
9339 Generate code for a PDP-11/40.
9343 Generate code for a PDP-11/45. This is the default.
9347 Generate code for a PDP-11/10.
9349 @item -mbcopy-builtin
9350 @opindex bcopy-builtin
9351 Use inline @code{movstrhi} patterns for copying memory. This is the
9356 Do not use inline @code{movstrhi} patterns for copying memory.
9362 Use 16-bit @code{int}. This is the default.
9368 Use 32-bit @code{int}.
9373 @opindex mno-float32
9374 Use 64-bit @code{float}. This is the default.
9379 @opindex mno-float64
9380 Use 32-bit @code{float}.
9384 Use @code{abshi2} pattern. This is the default.
9388 Do not use @code{abshi2} pattern.
9390 @item -mbranch-expensive
9391 @opindex mbranch-expensive
9392 Pretend that branches are expensive. This is for experimenting with
9393 code generation only.
9395 @item -mbranch-cheap
9396 @opindex mbranch-cheap
9397 Do not pretend that branches are expensive. This is the default.
9401 Generate code for a system with split I&D.
9405 Generate code for a system without split I&D. This is the default.
9409 Use Unix assembler syntax. This is the default when configured for
9414 Use DEC assembler syntax. This is the default when configured for any
9415 PDP-11 target other than @samp{pdp11-*-bsd}.
9418 @node Xstormy16 Options
9419 @subsection Xstormy16 Options
9420 @cindex Xstormy16 Options
9422 These options are defined for Xstormy16:
9427 Choose startup files and linker script suitable for the simulator.
9430 @node Xtensa Options
9431 @subsection Xtensa Options
9432 @cindex Xtensa Options
9434 The Xtensa architecture is designed to support many different
9435 configurations. The compiler's default options can be set to match a
9436 particular Xtensa configuration by copying a configuration file into the
9437 GCC sources when building GCC@. The options below may be used to
9438 override the default options.
9442 @itemx -mlittle-endian
9443 @opindex mbig-endian
9444 @opindex mlittle-endian
9445 Specify big-endian or little-endian byte ordering for the target Xtensa
9451 @opindex mno-density
9452 Enable or disable use of the optional Xtensa code density instructions.
9458 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
9459 will generate MAC16 instructions from standard C code, with the
9460 limitation that it will use neither the MR register file nor any
9461 instruction that operates on the MR registers. When this option is
9462 disabled, GCC will translate 16-bit multiply/accumulate operations to a
9463 combination of core instructions and library calls, depending on whether
9464 any other multiplier options are enabled.
9470 Enable or disable use of the 16-bit integer multiplier option. When
9471 enabled, the compiler will generate 16-bit multiply instructions for
9472 multiplications of 16 bits or smaller in standard C code. When this
9473 option is disabled, the compiler will either use 32-bit multiply or
9474 MAC16 instructions if they are available or generate library calls to
9475 perform the multiply operations using shifts and adds.
9481 Enable or disable use of the 32-bit integer multiplier option. When
9482 enabled, the compiler will generate 32-bit multiply instructions for
9483 multiplications of 32 bits or smaller in standard C code. When this
9484 option is disabled, the compiler will generate library calls to perform
9485 the multiply operations using either shifts and adds or 16-bit multiply
9486 instructions if they are available.
9492 Enable or disable use of the optional normalization shift amount
9493 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
9499 Enable or disable use of the optional minimum and maximum value
9506 Enable or disable use of the optional sign extend (@code{SEXT})
9510 @itemx -mno-booleans
9512 @opindex mno-booleans
9513 Enable or disable support for the boolean register file used by Xtensa
9514 coprocessors. This is not typically useful by itself but may be
9515 required for other options that make use of the boolean registers (e.g.,
9516 the floating-point option).
9520 @opindex mhard-float
9521 @opindex msoft-float
9522 Enable or disable use of the floating-point option. When enabled, GCC
9523 generates floating-point instructions for 32-bit @code{float}
9524 operations. When this option is disabled, GCC generates library calls
9525 to emulate 32-bit floating-point operations using integer instructions.
9526 Regardless of this option, 64-bit @code{double} operations are always
9527 emulated with calls to library functions.
9530 @itemx -mno-fused-madd
9531 @opindex mfused-madd
9532 @opindex mno-fused-madd
9533 Enable or disable use of fused multiply/add and multiply/subtract
9534 instructions in the floating-point option. This has no effect if the
9535 floating-point option is not also enabled. Disabling fused multiply/add
9536 and multiply/subtract instructions forces the compiler to use separate
9537 instructions for the multiply and add/subtract operations. This may be
9538 desirable in some cases where strict IEEE 754-compliant results are
9539 required: the fused multiply add/subtract instructions do not round the
9540 intermediate result, thereby producing results with @emph{more} bits of
9541 precision than specified by the IEEE standard. Disabling fused multiply
9542 add/subtract instructions also ensures that the program output is not
9543 sensitive to the compiler's ability to combine multiply and add/subtract
9546 @item -mserialize-volatile
9547 @itemx -mno-serialize-volatile
9548 @opindex mserialize-volatile
9549 @opindex mno-serialize-volatile
9550 When this option is enabled, GCC inserts @code{MEMW} instructions before
9551 @code{volatile} memory references to guarantee sequential consistency.
9552 The default is @option{-mserialize-volatile}. Use
9553 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
9555 @item -mtext-section-literals
9556 @itemx -mno-text-section-literals
9557 @opindex mtext-section-literals
9558 @opindex mno-text-section-literals
9559 Control the treatment of literal pools. The default is
9560 @option{-mno-text-section-literals}, which places literals in a separate
9561 section in the output file. This allows the literal pool to be placed
9562 in a data RAM/ROM, and it also allows the linker to combine literal
9563 pools from separate object files to remove redundant literals and
9564 improve code size. With @option{-mtext-section-literals}, the literals
9565 are interspersed in the text section in order to keep them as close as
9566 possible to their references. This may be necessary for large assembly
9569 @item -mtarget-align
9570 @itemx -mno-target-align
9571 @opindex mtarget-align
9572 @opindex mno-target-align
9573 When this option is enabled, GCC instructs the assembler to
9574 automatically align instructions to reduce branch penalties at the
9575 expense of some code density. The assembler attempts to widen density
9576 instructions to align branch targets and the instructions following call
9577 instructions. If there are not enough preceding safe density
9578 instructions to align a target, no widening will be performed. The
9579 default is @option{-mtarget-align}. These options do not affect the
9580 treatment of auto-aligned instructions like @code{LOOP}, which the
9581 assembler will always align, either by widening density instructions or
9582 by inserting no-op instructions.
9585 @itemx -mno-longcalls
9587 @opindex mno-longcalls
9588 When this option is enabled, GCC instructs the assembler to translate
9589 direct calls to indirect calls unless it can determine that the target
9590 of a direct call is in the range allowed by the call instruction. This
9591 translation typically occurs for calls to functions in other source
9592 files. Specifically, the assembler translates a direct @code{CALL}
9593 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
9594 The default is @option{-mno-longcalls}. This option should be used in
9595 programs where the call target can potentially be out of range. This
9596 option is implemented in the assembler, not the compiler, so the
9597 assembly code generated by GCC will still show direct call
9598 instructions---look at the disassembled object code to see the actual
9599 instructions. Note that the assembler will use an indirect call for
9600 every cross-file call, not just those that really will be out of range.
9603 @node Code Gen Options
9604 @section Options for Code Generation Conventions
9605 @cindex code generation conventions
9606 @cindex options, code generation
9607 @cindex run-time options
9609 These machine-independent options control the interface conventions
9610 used in code generation.
9612 Most of them have both positive and negative forms; the negative form
9613 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9614 one of the forms is listed---the one which is not the default. You
9615 can figure out the other form by either removing @samp{no-} or adding
9620 @opindex fexceptions
9621 Enable exception handling. Generates extra code needed to propagate
9622 exceptions. For some targets, this implies GCC will generate frame
9623 unwind information for all functions, which can produce significant data
9624 size overhead, although it does not affect execution. If you do not
9625 specify this option, GCC will enable it by default for languages like
9626 C++ which normally require exception handling, and disable it for
9627 languages like C that do not normally require it. However, you may need
9628 to enable this option when compiling C code that needs to interoperate
9629 properly with exception handlers written in C++. You may also wish to
9630 disable this option if you are compiling older C++ programs that don't
9631 use exception handling.
9633 @item -fnon-call-exceptions
9634 @opindex fnon-call-exceptions
9635 Generate code that allows trapping instructions to throw exceptions.
9636 Note that this requires platform-specific runtime support that does
9637 not exist everywhere. Moreover, it only allows @emph{trapping}
9638 instructions to throw exceptions, i.e.@: memory references or floating
9639 point instructions. It does not allow exceptions to be thrown from
9640 arbitrary signal handlers such as @code{SIGALRM}.
9642 @item -funwind-tables
9643 @opindex funwind-tables
9644 Similar to @option{-fexceptions}, except that it will just generate any needed
9645 static data, but will not affect the generated code in any other way.
9646 You will normally not enable this option; instead, a language processor
9647 that needs this handling would enable it on your behalf.
9649 @item -fasynchronous-unwind-tables
9650 @opindex funwind-tables
9651 Generate unwind table in dwarf2 format, if supported by target machine. The
9652 table is exact at each instruction boundary, so it can be used for stack
9653 unwinding from asynchronous events (such as debugger or garbage collector).
9655 @item -fpcc-struct-return
9656 @opindex fpcc-struct-return
9657 Return ``short'' @code{struct} and @code{union} values in memory like
9658 longer ones, rather than in registers. This convention is less
9659 efficient, but it has the advantage of allowing intercallability between
9660 GCC-compiled files and files compiled with other compilers.
9662 The precise convention for returning structures in memory depends
9663 on the target configuration macros.
9665 Short structures and unions are those whose size and alignment match
9666 that of some integer type.
9668 @item -freg-struct-return
9669 @opindex freg-struct-return
9670 Return @code{struct} and @code{union} values in registers when possible.
9671 This is more efficient for small structures than
9672 @option{-fpcc-struct-return}.
9674 If you specify neither @option{-fpcc-struct-return} nor
9675 @option{-freg-struct-return}, GCC defaults to whichever convention is
9676 standard for the target. If there is no standard convention, GCC
9677 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9678 the principal compiler. In those cases, we can choose the standard, and
9679 we chose the more efficient register return alternative.
9682 @opindex fshort-enums
9683 Allocate to an @code{enum} type only as many bytes as it needs for the
9684 declared range of possible values. Specifically, the @code{enum} type
9685 will be equivalent to the smallest integer type which has enough room.
9687 @item -fshort-double
9688 @opindex fshort-double
9689 Use the same size for @code{double} as for @code{float}.
9692 @opindex fshared-data
9693 Requests that the data and non-@code{const} variables of this
9694 compilation be shared data rather than private data. The distinction
9695 makes sense only on certain operating systems, where shared data is
9696 shared between processes running the same program, while private data
9697 exists in one copy per process.
9701 In C, allocate even uninitialized global variables in the data section of the
9702 object file, rather than generating them as common blocks. This has the
9703 effect that if the same variable is declared (without @code{extern}) in
9704 two different compilations, you will get an error when you link them.
9705 The only reason this might be useful is if you wish to verify that the
9706 program will work on other systems which always work this way.
9710 Ignore the @samp{#ident} directive.
9712 @item -fno-gnu-linker
9713 @opindex fno-gnu-linker
9714 Do not output global initializations (such as C++ constructors and
9715 destructors) in the form used by the GNU linker (on systems where the GNU
9716 linker is the standard method of handling them). Use this option when
9717 you want to use a non-GNU linker, which also requires using the
9718 @command{collect2} program to make sure the system linker includes
9719 constructors and destructors. (@command{collect2} is included in the GCC
9720 distribution.) For systems which @emph{must} use @command{collect2}, the
9721 compiler driver @command{gcc} is configured to do this automatically.
9723 @item -finhibit-size-directive
9724 @opindex finhibit-size-directive
9725 Don't output a @code{.size} assembler directive, or anything else that
9726 would cause trouble if the function is split in the middle, and the
9727 two halves are placed at locations far apart in memory. This option is
9728 used when compiling @file{crtstuff.c}; you should not need to use it
9732 @opindex fverbose-asm
9733 Put extra commentary information in the generated assembly code to
9734 make it more readable. This option is generally only of use to those
9735 who actually need to read the generated assembly code (perhaps while
9736 debugging the compiler itself).
9738 @option{-fno-verbose-asm}, the default, causes the
9739 extra information to be omitted and is useful when comparing two assembler
9744 Consider all memory references through pointers to be volatile.
9746 @item -fvolatile-global
9747 @opindex fvolatile-global
9748 Consider all memory references to extern and global data items to
9749 be volatile. GCC does not consider static data items to be volatile
9750 because of this switch.
9752 @item -fvolatile-static
9753 @opindex fvolatile-static
9754 Consider all memory references to static data to be volatile.
9758 @cindex global offset table
9760 Generate position-independent code (PIC) suitable for use in a shared
9761 library, if supported for the target machine. Such code accesses all
9762 constant addresses through a global offset table (GOT)@. The dynamic
9763 loader resolves the GOT entries when the program starts (the dynamic
9764 loader is not part of GCC; it is part of the operating system). If
9765 the GOT size for the linked executable exceeds a machine-specific
9766 maximum size, you get an error message from the linker indicating that
9767 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9768 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9769 on the m68k and RS/6000. The 386 has no such limit.)
9771 Position-independent code requires special support, and therefore works
9772 only on certain machines. For the 386, GCC supports PIC for System V
9773 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9774 position-independent.
9778 If supported for the target machine, emit position-independent code,
9779 suitable for dynamic linking and avoiding any limit on the size of the
9780 global offset table. This option makes a difference on the m68k, m88k,
9783 Position-independent code requires special support, and therefore works
9784 only on certain machines.
9786 @item -ffixed-@var{reg}
9788 Treat the register named @var{reg} as a fixed register; generated code
9789 should never refer to it (except perhaps as a stack pointer, frame
9790 pointer or in some other fixed role).
9792 @var{reg} must be the name of a register. The register names accepted
9793 are machine-specific and are defined in the @code{REGISTER_NAMES}
9794 macro in the machine description macro file.
9796 This flag does not have a negative form, because it specifies a
9799 @item -fcall-used-@var{reg}
9801 Treat the register named @var{reg} as an allocable register that is
9802 clobbered by function calls. It may be allocated for temporaries or
9803 variables that do not live across a call. Functions compiled this way
9804 will not save and restore the register @var{reg}.
9806 It is an error to used this flag with the frame pointer or stack pointer.
9807 Use of this flag for other registers that have fixed pervasive roles in
9808 the machine's execution model will produce disastrous results.
9810 This flag does not have a negative form, because it specifies a
9813 @item -fcall-saved-@var{reg}
9814 @opindex fcall-saved
9815 Treat the register named @var{reg} as an allocable register saved by
9816 functions. It may be allocated even for temporaries or variables that
9817 live across a call. Functions compiled this way will save and restore
9818 the register @var{reg} if they use it.
9820 It is an error to used this flag with the frame pointer or stack pointer.
9821 Use of this flag for other registers that have fixed pervasive roles in
9822 the machine's execution model will produce disastrous results.
9824 A different sort of disaster will result from the use of this flag for
9825 a register in which function values may be returned.
9827 This flag does not have a negative form, because it specifies a
9831 @opindex fpack-struct
9832 Pack all structure members together without holes. Usually you would
9833 not want to use this option, since it makes the code suboptimal, and
9834 the offsets of structure members won't agree with system libraries.
9836 @item -finstrument-functions
9837 @opindex finstrument-functions
9838 Generate instrumentation calls for entry and exit to functions. Just
9839 after function entry and just before function exit, the following
9840 profiling functions will be called with the address of the current
9841 function and its call site. (On some platforms,
9842 @code{__builtin_return_address} does not work beyond the current
9843 function, so the call site information may not be available to the
9844 profiling functions otherwise.)
9847 void __cyg_profile_func_enter (void *this_fn,
9849 void __cyg_profile_func_exit (void *this_fn,
9853 The first argument is the address of the start of the current function,
9854 which may be looked up exactly in the symbol table.
9856 This instrumentation is also done for functions expanded inline in other
9857 functions. The profiling calls will indicate where, conceptually, the
9858 inline function is entered and exited. This means that addressable
9859 versions of such functions must be available. If all your uses of a
9860 function are expanded inline, this may mean an additional expansion of
9861 code size. If you use @samp{extern inline} in your C code, an
9862 addressable version of such functions must be provided. (This is
9863 normally the case anyways, but if you get lucky and the optimizer always
9864 expands the functions inline, you might have gotten away without
9865 providing static copies.)
9867 A function may be given the attribute @code{no_instrument_function}, in
9868 which case this instrumentation will not be done. This can be used, for
9869 example, for the profiling functions listed above, high-priority
9870 interrupt routines, and any functions from which the profiling functions
9871 cannot safely be called (perhaps signal handlers, if the profiling
9872 routines generate output or allocate memory).
9875 @opindex fstack-check
9876 Generate code to verify that you do not go beyond the boundary of the
9877 stack. You should specify this flag if you are running in an
9878 environment with multiple threads, but only rarely need to specify it in
9879 a single-threaded environment since stack overflow is automatically
9880 detected on nearly all systems if there is only one stack.
9882 Note that this switch does not actually cause checking to be done; the
9883 operating system must do that. The switch causes generation of code
9884 to ensure that the operating system sees the stack being extended.
9886 @item -fstack-limit-register=@var{reg}
9887 @itemx -fstack-limit-symbol=@var{sym}
9888 @itemx -fno-stack-limit
9889 @opindex fstack-limit-register
9890 @opindex fstack-limit-symbol
9891 @opindex fno-stack-limit
9892 Generate code to ensure that the stack does not grow beyond a certain value,
9893 either the value of a register or the address of a symbol. If the stack
9894 would grow beyond the value, a signal is raised. For most targets,
9895 the signal is raised before the stack overruns the boundary, so
9896 it is possible to catch the signal without taking special precautions.
9898 For instance, if the stack starts at absolute address @samp{0x80000000}
9899 and grows downwards, you can use the flags
9900 @option{-fstack-limit-symbol=__stack_limit} and
9901 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
9902 of 128KB@. Note that this may only work with the GNU linker.
9904 @cindex aliasing of parameters
9905 @cindex parameters, aliased
9906 @item -fargument-alias
9907 @itemx -fargument-noalias
9908 @itemx -fargument-noalias-global
9909 @opindex fargument-alias
9910 @opindex fargument-noalias
9911 @opindex fargument-noalias-global
9912 Specify the possible relationships among parameters and between
9913 parameters and global data.
9915 @option{-fargument-alias} specifies that arguments (parameters) may
9916 alias each other and may alias global storage.@*
9917 @option{-fargument-noalias} specifies that arguments do not alias
9918 each other, but may alias global storage.@*
9919 @option{-fargument-noalias-global} specifies that arguments do not
9920 alias each other and do not alias global storage.
9922 Each language will automatically use whatever option is required by
9923 the language standard. You should not need to use these options yourself.
9925 @item -fleading-underscore
9926 @opindex fleading-underscore
9927 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
9928 change the way C symbols are represented in the object file. One use
9929 is to help link with legacy assembly code.
9931 Be warned that you should know what you are doing when invoking this
9932 option, and that not all targets provide complete support for it.
9934 @item -ftls-model=@var{model}
9935 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
9936 The @var{model} argument should be one of @code{global-dynamic},
9937 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
9939 The default without @option{-fpic} is @code{initial-exec}; with
9940 @option{-fpic} the default is @code{global-dynamic}.
9945 @node Environment Variables
9946 @section Environment Variables Affecting GCC
9947 @cindex environment variables
9949 @c man begin ENVIRONMENT
9951 This section describes several environment variables that affect how GCC
9952 operates. Some of them work by specifying directories or prefixes to use
9953 when searching for various kinds of files. Some are used to specify other
9954 aspects of the compilation environment.
9956 Note that you can also specify places to search using options such as
9957 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
9958 take precedence over places specified using environment variables, which
9959 in turn take precedence over those specified by the configuration of GCC@.
9960 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
9961 GNU Compiler Collection (GCC) Internals}.
9966 @c @itemx LC_COLLATE
9968 @c @itemx LC_MONETARY
9969 @c @itemx LC_NUMERIC
9974 @c @findex LC_COLLATE
9976 @c @findex LC_MONETARY
9977 @c @findex LC_NUMERIC
9981 These environment variables control the way that GCC uses
9982 localization information that allow GCC to work with different
9983 national conventions. GCC inspects the locale categories
9984 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
9985 so. These locale categories can be set to any value supported by your
9986 installation. A typical value is @samp{en_UK} for English in the United
9989 The @env{LC_CTYPE} environment variable specifies character
9990 classification. GCC uses it to determine the character boundaries in
9991 a string; this is needed for some multibyte encodings that contain quote
9992 and escape characters that would otherwise be interpreted as a string
9995 The @env{LC_MESSAGES} environment variable specifies the language to
9996 use in diagnostic messages.
9998 If the @env{LC_ALL} environment variable is set, it overrides the value
9999 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10000 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10001 environment variable. If none of these variables are set, GCC
10002 defaults to traditional C English behavior.
10006 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10007 files. GCC uses temporary files to hold the output of one stage of
10008 compilation which is to be used as input to the next stage: for example,
10009 the output of the preprocessor, which is the input to the compiler
10012 @item GCC_EXEC_PREFIX
10013 @findex GCC_EXEC_PREFIX
10014 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10015 names of the subprograms executed by the compiler. No slash is added
10016 when this prefix is combined with the name of a subprogram, but you can
10017 specify a prefix that ends with a slash if you wish.
10019 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10020 an appropriate prefix to use based on the pathname it was invoked with.
10022 If GCC cannot find the subprogram using the specified prefix, it
10023 tries looking in the usual places for the subprogram.
10025 The default value of @env{GCC_EXEC_PREFIX} is
10026 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10027 of @code{prefix} when you ran the @file{configure} script.
10029 Other prefixes specified with @option{-B} take precedence over this prefix.
10031 This prefix is also used for finding files such as @file{crt0.o} that are
10034 In addition, the prefix is used in an unusual way in finding the
10035 directories to search for header files. For each of the standard
10036 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10037 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10038 replacing that beginning with the specified prefix to produce an
10039 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10040 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10041 These alternate directories are searched first; the standard directories
10044 @item COMPILER_PATH
10045 @findex COMPILER_PATH
10046 The value of @env{COMPILER_PATH} is a colon-separated list of
10047 directories, much like @env{PATH}. GCC tries the directories thus
10048 specified when searching for subprograms, if it can't find the
10049 subprograms using @env{GCC_EXEC_PREFIX}.
10052 @findex LIBRARY_PATH
10053 The value of @env{LIBRARY_PATH} is a colon-separated list of
10054 directories, much like @env{PATH}. When configured as a native compiler,
10055 GCC tries the directories thus specified when searching for special
10056 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10057 using GCC also uses these directories when searching for ordinary
10058 libraries for the @option{-l} option (but directories specified with
10059 @option{-L} come first).
10063 @cindex locale definition
10064 This variable is used to pass locale information to the compiler. One way in
10065 which this information is used is to determine the character set to be used
10066 when character literals, string literals and comments are parsed in C and C++.
10067 When the compiler is configured to allow multibyte characters,
10068 the following values for @env{LANG} are recognized:
10072 Recognize JIS characters.
10074 Recognize SJIS characters.
10076 Recognize EUCJP characters.
10079 If @env{LANG} is not defined, or if it has some other value, then the
10080 compiler will use mblen and mbtowc as defined by the default locale to
10081 recognize and translate multibyte characters.
10085 Some additional environments variables affect the behavior of the
10088 @include cppenv.texi
10092 @node Running Protoize
10093 @section Running Protoize
10095 The program @code{protoize} is an optional part of GCC@. You can use
10096 it to add prototypes to a program, thus converting the program to ISO
10097 C in one respect. The companion program @code{unprotoize} does the
10098 reverse: it removes argument types from any prototypes that are found.
10100 When you run these programs, you must specify a set of source files as
10101 command line arguments. The conversion programs start out by compiling
10102 these files to see what functions they define. The information gathered
10103 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10105 After scanning comes actual conversion. The specified files are all
10106 eligible to be converted; any files they include (whether sources or
10107 just headers) are eligible as well.
10109 But not all the eligible files are converted. By default,
10110 @code{protoize} and @code{unprotoize} convert only source and header
10111 files in the current directory. You can specify additional directories
10112 whose files should be converted with the @option{-d @var{directory}}
10113 option. You can also specify particular files to exclude with the
10114 @option{-x @var{file}} option. A file is converted if it is eligible, its
10115 directory name matches one of the specified directory names, and its
10116 name within the directory has not been excluded.
10118 Basic conversion with @code{protoize} consists of rewriting most
10119 function definitions and function declarations to specify the types of
10120 the arguments. The only ones not rewritten are those for varargs
10123 @code{protoize} optionally inserts prototype declarations at the
10124 beginning of the source file, to make them available for any calls that
10125 precede the function's definition. Or it can insert prototype
10126 declarations with block scope in the blocks where undeclared functions
10129 Basic conversion with @code{unprotoize} consists of rewriting most
10130 function declarations to remove any argument types, and rewriting
10131 function definitions to the old-style pre-ISO form.
10133 Both conversion programs print a warning for any function declaration or
10134 definition that they can't convert. You can suppress these warnings
10137 The output from @code{protoize} or @code{unprotoize} replaces the
10138 original source file. The original file is renamed to a name ending
10139 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10140 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10141 for DOS) file already exists, then the source file is simply discarded.
10143 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10144 scan the program and collect information about the functions it uses.
10145 So neither of these programs will work until GCC is installed.
10147 Here is a table of the options you can use with @code{protoize} and
10148 @code{unprotoize}. Each option works with both programs unless
10152 @item -B @var{directory}
10153 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10154 usual directory (normally @file{/usr/local/lib}). This file contains
10155 prototype information about standard system functions. This option
10156 applies only to @code{protoize}.
10158 @item -c @var{compilation-options}
10159 Use @var{compilation-options} as the options when running @code{gcc} to
10160 produce the @samp{.X} files. The special option @option{-aux-info} is
10161 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10163 Note that the compilation options must be given as a single argument to
10164 @code{protoize} or @code{unprotoize}. If you want to specify several
10165 @code{gcc} options, you must quote the entire set of compilation options
10166 to make them a single word in the shell.
10168 There are certain @code{gcc} arguments that you cannot use, because they
10169 would produce the wrong kind of output. These include @option{-g},
10170 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10171 the @var{compilation-options}, they are ignored.
10174 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10175 systems) instead of @samp{.c}. This is convenient if you are converting
10176 a C program to C++. This option applies only to @code{protoize}.
10179 Add explicit global declarations. This means inserting explicit
10180 declarations at the beginning of each source file for each function
10181 that is called in the file and was not declared. These declarations
10182 precede the first function definition that contains a call to an
10183 undeclared function. This option applies only to @code{protoize}.
10185 @item -i @var{string}
10186 Indent old-style parameter declarations with the string @var{string}.
10187 This option applies only to @code{protoize}.
10189 @code{unprotoize} converts prototyped function definitions to old-style
10190 function definitions, where the arguments are declared between the
10191 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10192 uses five spaces as the indentation. If you want to indent with just
10193 one space instead, use @option{-i " "}.
10196 Keep the @samp{.X} files. Normally, they are deleted after conversion
10200 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10201 a prototype declaration for each function in each block which calls the
10202 function without any declaration. This option applies only to
10206 Make no real changes. This mode just prints information about the conversions
10207 that would have been done without @option{-n}.
10210 Make no @samp{.save} files. The original files are simply deleted.
10211 Use this option with caution.
10213 @item -p @var{program}
10214 Use the program @var{program} as the compiler. Normally, the name
10215 @file{gcc} is used.
10218 Work quietly. Most warnings are suppressed.
10221 Print the version number, just like @option{-v} for @code{gcc}.
10224 If you need special compiler options to compile one of your program's
10225 source files, then you should generate that file's @samp{.X} file
10226 specially, by running @code{gcc} on that source file with the
10227 appropriate options and the option @option{-aux-info}. Then run
10228 @code{protoize} on the entire set of files. @code{protoize} will use
10229 the existing @samp{.X} file because it is newer than the source file.
10233 gcc -Dfoo=bar file1.c -aux-info file1.X
10238 You need to include the special files along with the rest in the
10239 @code{protoize} command, even though their @samp{.X} files already
10240 exist, because otherwise they won't get converted.
10242 @xref{Protoize Caveats}, for more information on how to use
10243 @code{protoize} successfully.