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 -fms-extensions @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
174 @item C++ Language Options
175 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
177 -fabi-version=@var{n} -fno-access-control -fcheck-new @gol
178 -fconserve-space -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 -Wabi -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 -Wundeclared-selector}
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 -Wstrict-aliasing -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 -fnew-ra -fno-default-inline @gol
275 -fno-function-cse -fno-guess-branch-probability @gol
276 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
277 -funsafe-math-optimizations -ffinite-math-only @gol
278 -fno-trapping-math -fno-zero-initialized-in-bss @gol
279 -fomit-frame-pointer -foptimize-register-move @gol
280 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
281 -freduce-all-givs -fregmove -frename-registers @gol
282 -freorder-blocks -freorder-functions @gol
283 -frerun-cse-after-loop -frerun-loop-opt @gol
284 -fschedule-insns -fschedule-insns2 -fsignaling-nans @gol
285 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
286 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
287 -ftrapv -funroll-all-loops -funroll-loops @gol
288 --param @var{name}=@var{value}
289 -O -O0 -O1 -O2 -O3 -Os}
291 @item Preprocessor Options
292 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
294 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
295 -C -dD -dI -dM -dN @gol
296 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
297 -idirafter @var{dir} @gol
298 -include @var{file} -imacros @var{file} @gol
299 -iprefix @var{file} -iwithprefix @var{dir} @gol
300 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
301 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
302 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
304 @item Assembler Option
305 @xref{Assembler Options,,Passing Options to the Assembler}.
310 @xref{Link Options,,Options for Linking}.
312 @var{object-file-name} -l@var{library} @gol
313 -nostartfiles -nodefaultlibs -nostdlib @gol
314 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
315 -Wl,@var{option} -Xlinker @var{option} @gol
318 @item Directory Options
319 @xref{Directory Options,,Options for Directory Search}.
321 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
324 @c I wrote this xref this way to avoid overfull hbox. -- rms
325 @xref{Target Options}.
327 -V @var{version} -b @var{machine}}
329 @item Machine Dependent Options
330 @xref{Submodel Options,,Hardware Models and Configurations}.
332 @emph{M680x0 Options}
334 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
335 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
336 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
337 -malign-int -mstrict-align}
339 @emph{M68hc1x Options}
341 -m6811 -m6812 -m68hc11 -m68hc12 @gol
342 -mauto-incdec -minmax -mlong-calls -mshort @gol
343 -msoft-reg-count=@var{count}}
351 -mcpu=@var{cpu-type} @gol
352 -mtune=@var{cpu-type} @gol
353 -mcmodel=@var{code-model} @gol
355 -mapp-regs -mbroken-saverestore -mcypress @gol
356 -mfaster-structs -mflat @gol
357 -mfpu -mhard-float -mhard-quad-float @gol
358 -mimpure-text -mlive-g0 -mno-app-regs @gol
359 -mno-faster-structs -mno-flat -mno-fpu @gol
360 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
361 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
362 -msupersparc -munaligned-doubles -mv8}
364 @emph{Convex Options}
366 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
367 -margcount -mnoargcount @gol
368 -mlong32 -mlong64 @gol
369 -mvolatile-cache -mvolatile-nocache}
371 @emph{AMD29K Options}
373 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
374 -mlarge -mnormal -msmall @gol
375 -mkernel-registers -mno-reuse-arg-regs @gol
376 -mno-stack-check -mno-storem-bug @gol
377 -mreuse-arg-regs -msoft-float -mstack-check @gol
378 -mstorem-bug -muser-registers}
382 -mapcs-frame -mno-apcs-frame @gol
383 -mapcs-26 -mapcs-32 @gol
384 -mapcs-stack-check -mno-apcs-stack-check @gol
385 -mapcs-float -mno-apcs-float @gol
386 -mapcs-reentrant -mno-apcs-reentrant @gol
387 -msched-prolog -mno-sched-prolog @gol
388 -mlittle-endian -mbig-endian -mwords-little-endian @gol
389 -malignment-traps -mno-alignment-traps @gol
390 -msoft-float -mhard-float -mfpe @gol
391 -mthumb-interwork -mno-thumb-interwork @gol
392 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
393 -mstructure-size-boundary=@var{n} @gol
394 -mbsd -mxopen -mno-symrename @gol
395 -mabort-on-noreturn @gol
396 -mlong-calls -mno-long-calls @gol
397 -msingle-pic-base -mno-single-pic-base @gol
398 -mpic-register=@var{reg} @gol
399 -mnop-fun-dllimport @gol
400 -mpoke-function-name @gol
402 -mtpcs-frame -mtpcs-leaf-frame @gol
403 -mcaller-super-interworking -mcallee-super-interworking }
405 @emph{MN10200 Options}
409 @emph{MN10300 Options}
411 -mmult-bug -mno-mult-bug @gol
412 -mam33 -mno-am33 @gol
415 @emph{M32R/D Options}
417 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
422 -m88000 -m88100 -m88110 -mbig-pic @gol
423 -mcheck-zero-division -mhandle-large-shift @gol
424 -midentify-revision -mno-check-zero-division @gol
425 -mno-ocs-debug-info -mno-ocs-frame-position @gol
426 -mno-optimize-arg-area -mno-serialize-volatile @gol
427 -mno-underscores -mocs-debug-info @gol
428 -mocs-frame-position -moptimize-arg-area @gol
429 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
430 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
431 -mversion-03.00 -mwarn-passed-structs}
433 @emph{RS/6000 and PowerPC Options}
435 -mcpu=@var{cpu-type} @gol
436 -mtune=@var{cpu-type} @gol
437 -mpower -mno-power -mpower2 -mno-power2 @gol
438 -mpowerpc -mpowerpc64 -mno-powerpc @gol
439 -maltivec -mno-altivec @gol
440 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
441 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
442 -mnew-mnemonics -mold-mnemonics @gol
443 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
444 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
445 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
446 -mstring -mno-string -mupdate -mno-update @gol
447 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
448 -mstrict-align -mno-strict-align -mrelocatable @gol
449 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
450 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
451 -mcall-aix -mcall-sysv -mcall-netbsd @gol
452 -maix-struct-return -msvr4-struct-return @gol
453 -mabi=altivec -mabi=no-altivec @gol
454 -mabi=spe -mabi=no-spe @gol
455 -misel=yes -misel=no @gol
456 -mprototype -mno-prototype @gol
457 -msim -mmvme -mads -myellowknife -memb -msdata @gol
458 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
462 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
463 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
464 -mminimum-fp-blocks -mnohc-struct-return}
468 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
469 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
470 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
471 -mgas -mgp32 -mgp64 @gol
472 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
473 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
474 -mmips-as -mmips-tfile -mno-abicalls @gol
475 -mno-embedded-data -mno-uninit-const-in-rodata @gol
476 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
477 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
478 -mrnames -msoft-float @gol
479 -m4650 -msingle-float -mmad @gol
480 -mstats -EL -EB -G @var{num} -nocpp @gol
481 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
482 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func @gol
483 -mbranch-likely -mno-branch-likely}
485 @emph{i386 and x86-64 Options}
487 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
488 -masm=@var{dialect} -mno-fancy-math-387 @gol
489 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
490 -mno-wide-multiply -mrtd -malign-double @gol
491 -mpreferred-stack-boundary=@var{num} @gol
492 -mmmx -msse -msse2 -msse-math -m3dnow @gol
493 -mthreads -mno-align-stringops -minline-all-stringops @gol
494 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
495 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
497 -mcmodel=@var{code-model} @gol
502 -march=@var{architecture-type} @gol
503 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
504 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
505 -mjump-in-delay -mlinker-opt @gol
506 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
507 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
508 -mno-jump-in-delay -mno-long-load-store @gol
509 -mno-portable-runtime -mno-soft-float @gol
510 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
511 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
512 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio}
514 @emph{Intel 960 Options}
516 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
517 -mcode-align -mcomplex-addr -mleaf-procedures @gol
518 -mic-compat -mic2.0-compat -mic3.0-compat @gol
519 -mintel-asm -mno-clean-linkage -mno-code-align @gol
520 -mno-complex-addr -mno-leaf-procedures @gol
521 -mno-old-align -mno-strict-align -mno-tail-call @gol
522 -mnumerics -mold-align -msoft-float -mstrict-align @gol
525 @emph{DEC Alpha Options}
527 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
528 -mieee -mieee-with-inexact -mieee-conformant @gol
529 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
530 -mtrap-precision=@var{mode} -mbuild-constants @gol
531 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
532 -mbwx -mmax -mfix -mcix @gol
533 -mfloat-vax -mfloat-ieee @gol
534 -mexplicit-relocs -msmall-data -mlarge-data @gol
535 -mmemory-latency=@var{time}}
537 @emph{DEC Alpha/VMS Options}
541 @emph{Clipper Options}
545 @emph{H8/300 Options}
547 -mrelax -mh -ms -mint32 -malign-300}
551 -m1 -m2 -m3 -m3e @gol
552 -m4-nofpu -m4-single-only -m4-single -m4 @gol
553 -m5-64media -m5-64media-nofpu @gol
554 -m5-32media -m5-32media-nofpu @gol
555 -m5-compact -m5-compact-nofpu @gol
556 -mb -ml -mdalign -mrelax @gol
557 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
558 -mieee -misize -mpadstruct -mspace @gol
559 -mprefergot -musermode}
561 @emph{System V Options}
563 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
568 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
569 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
571 @emph{TMS320C3x/C4x Options}
573 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
574 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
575 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
576 -mparallel-insns -mparallel-mpy -mpreserve-float}
580 -mlong-calls -mno-long-calls -mep -mno-ep @gol
581 -mprolog-function -mno-prolog-function -mspace @gol
582 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
583 -mapp-regs -mno-app-regs @gol
584 -mdisable-callt -mno-disable-callt @gol
590 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
591 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
592 -mregparam -mnoregparam -msb -mnosb @gol
593 -mbitfield -mnobitfield -mhimem -mnohimem}
597 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
598 -mcall-prologues -mno-tablejump -mtiny-stack}
602 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
603 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
604 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
605 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
606 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
610 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
611 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
612 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
613 -mno-base-addresses -msingle-exit -mno-single-exit}
617 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
618 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
619 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
620 -minline-divide-max-throughput -mno-dwarf2-asm @gol
621 -mfixed-range=@var{register-range}}
625 -mextmem -mextmemory -monchip -mno-asm-optimize @gol
626 -masm-optimize -mbranch-cost=@var{n} -mcond-exec=@var{n}}
628 @emph{S/390 and zSeries Options}
630 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
631 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
632 -m64 -m31 -mdebug -mno-debug}
636 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
637 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
638 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
639 -mstack-align -mdata-align -mconst-align @gol
640 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
641 -melf -maout -melinux -mlinux -sim -sim2}
643 @emph{PDP-11 Options}
645 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
646 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
647 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
648 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
649 -mbranch-expensive -mbranch-cheap @gol
650 -msplit -mno-split -munix-asm -mdec-asm}
652 @emph{Xstormy16 Options}
656 @emph{Xtensa Options}
658 -mbig-endian -mlittle-endian @gol
659 -mdensity -mno-density @gol
660 -mmac16 -mno-mac16 @gol
661 -mmul16 -mno-mul16 @gol
662 -mmul32 -mno-mul32 @gol
664 -mminmax -mno-minmax @gol
665 -msext -mno-sext @gol
666 -mbooleans -mno-booleans @gol
667 -mhard-float -msoft-float @gol
668 -mfused-madd -mno-fused-madd @gol
669 -mserialize-volatile -mno-serialize-volatile @gol
670 -mtext-section-literals -mno-text-section-literals @gol
671 -mtarget-align -mno-target-align @gol
672 -mlongcalls -mno-longcalls}
674 @item Code Generation Options
675 @xref{Code Gen Options,,Options for Code Generation Conventions}.
677 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
678 -ffixed-@var{reg} -fexceptions @gol
679 -fnon-call-exceptions -funwind-tables @gol
680 -fasynchronous-unwind-tables @gol
681 -finhibit-size-directive -finstrument-functions @gol
682 -fno-common -fno-ident -fno-gnu-linker @gol
683 -fpcc-struct-return -fpic -fPIC @gol
684 -freg-struct-return -fshared-data -fshort-enums @gol
685 -fshort-double -fshort-wchar -fvolatile @gol
686 -fvolatile-global -fvolatile-static @gol
687 -fverbose-asm -fpack-struct -fstack-check @gol
688 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
689 -fargument-alias -fargument-noalias @gol
690 -fargument-noalias-global -fleading-underscore @gol
691 -ftls-model=@var{model}}
695 * Overall Options:: Controlling the kind of output:
696 an executable, object files, assembler files,
697 or preprocessed source.
698 * C Dialect Options:: Controlling the variant of C language compiled.
699 * C++ Dialect Options:: Variations on C++.
700 * Objective-C Dialect Options:: Variations on Objective-C.
701 * Language Independent Options:: Controlling how diagnostics should be
703 * Warning Options:: How picky should the compiler be?
704 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
705 * Optimize Options:: How much optimization?
706 * Preprocessor Options:: Controlling header files and macro definitions.
707 Also, getting dependency information for Make.
708 * Assembler Options:: Passing options to the assembler.
709 * Link Options:: Specifying libraries and so on.
710 * Directory Options:: Where to find header files and libraries.
711 Where to find the compiler executable files.
712 * Spec Files:: How to pass switches to sub-processes.
713 * Target Options:: Running a cross-compiler, or an old version of GCC.
716 @node Overall Options
717 @section Options Controlling the Kind of Output
719 Compilation can involve up to four stages: preprocessing, compilation
720 proper, assembly and linking, always in that order. The first three
721 stages apply to an individual source file, and end by producing an
722 object file; linking combines all the object files (those newly
723 compiled, and those specified as input) into an executable file.
725 @cindex file name suffix
726 For any given input file, the file name suffix determines what kind of
731 C source code which must be preprocessed.
734 C source code which should not be preprocessed.
737 C++ source code which should not be preprocessed.
740 Objective-C source code. Note that you must link with the library
741 @file{libobjc.a} to make an Objective-C program work.
744 Objective-C source code which should not be preprocessed.
747 C header file (not to be compiled or linked).
751 @itemx @var{file}.cxx
752 @itemx @var{file}.cpp
753 @itemx @var{file}.c++
755 C++ source code which must be preprocessed. Note that in @samp{.cxx},
756 the last two letters must both be literally @samp{x}. Likewise,
757 @samp{.C} refers to a literal capital C@.
760 @itemx @var{file}.for
761 @itemx @var{file}.FOR
762 Fortran source code which should not be preprocessed.
765 @itemx @var{file}.fpp
766 @itemx @var{file}.FPP
767 Fortran source code which must be preprocessed (with the traditional
771 Fortran source code which must be preprocessed with a RATFOR
772 preprocessor (not included with GCC)@.
774 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
775 Using and Porting GNU Fortran}, for more details of the handling of
778 @c FIXME: Descriptions of Java file types.
785 Ada source code file which contains a library unit declaration (a
786 declaration of a package, subprogram, or generic, or a generic
787 instantiation), or a library unit renaming declaration (a package,
788 generic, or subprogram renaming declaration). Such files are also
791 @itemx @var{file}.adb
792 Ada source code file containing a library unit body (a subprogram or
793 package body). Such files are also called @dfn{bodies}.
795 @c GCC also knows about some suffixes for languages not yet included:
804 Assembler code which must be preprocessed.
807 An object file to be fed straight into linking.
808 Any file name with no recognized suffix is treated this way.
812 You can specify the input language explicitly with the @option{-x} option:
815 @item -x @var{language}
816 Specify explicitly the @var{language} for the following input files
817 (rather than letting the compiler choose a default based on the file
818 name suffix). This option applies to all following input files until
819 the next @option{-x} option. Possible values for @var{language} are:
821 c c-header cpp-output
823 objective-c objc-cpp-output
824 assembler assembler-with-cpp
826 f77 f77-cpp-input ratfor
832 Turn off any specification of a language, so that subsequent files are
833 handled according to their file name suffixes (as they are if @option{-x}
834 has not been used at all).
836 @item -pass-exit-codes
837 @opindex pass-exit-codes
838 Normally the @command{gcc} program will exit with the code of 1 if any
839 phase of the compiler returns a non-success return code. If you specify
840 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
841 numerically highest error produced by any phase that returned an error
845 If you only want some of the stages of compilation, you can use
846 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
847 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
848 @command{gcc} is to stop. Note that some combinations (for example,
849 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
854 Compile or assemble the source files, but do not link. The linking
855 stage simply is not done. The ultimate output is in the form of an
856 object file for each source file.
858 By default, the object file name for a source file is made by replacing
859 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
861 Unrecognized input files, not requiring compilation or assembly, are
866 Stop after the stage of compilation proper; do not assemble. The output
867 is in the form of an assembler code file for each non-assembler input
870 By default, the assembler file name for a source file is made by
871 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
873 Input files that don't require compilation are ignored.
877 Stop after the preprocessing stage; do not run the compiler proper. The
878 output is in the form of preprocessed source code, which is sent to the
881 Input files which don't require preprocessing are ignored.
883 @cindex output file option
886 Place output in file @var{file}. This applies regardless to whatever
887 sort of output is being produced, whether it be an executable file,
888 an object file, an assembler file or preprocessed C code.
890 Since only one output file can be specified, it does not make sense to
891 use @option{-o} when compiling more than one input file, unless you are
892 producing an executable file as output.
894 If @option{-o} is not specified, the default is to put an executable file
895 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
896 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
897 all preprocessed C source on standard output.
901 Print (on standard error output) the commands executed to run the stages
902 of compilation. Also print the version number of the compiler driver
903 program and of the preprocessor and the compiler proper.
907 Like @option{-v} except the commands are not executed and all command
908 arguments are quoted. This is useful for shell scripts to capture the
909 driver-generated command lines.
913 Use pipes rather than temporary files for communication between the
914 various stages of compilation. This fails to work on some systems where
915 the assembler is unable to read from a pipe; but the GNU assembler has
920 Print (on the standard output) a description of the command line options
921 understood by @command{gcc}. If the @option{-v} option is also specified
922 then @option{--help} will also be passed on to the various processes
923 invoked by @command{gcc}, so that they can display the command line options
924 they accept. If the @option{-W} option is also specified then command
925 line options which have no documentation associated with them will also
930 Print (on the standard output) a description of target specific command
931 line options for each tool.
935 @section Compiling C++ Programs
937 @cindex suffixes for C++ source
938 @cindex C++ source file suffixes
939 C++ source files conventionally use one of the suffixes @samp{.C},
940 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
941 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
942 files with these names and compiles them as C++ programs even if you
943 call the compiler the same way as for compiling C programs (usually with
944 the name @command{gcc}).
948 However, C++ programs often require class libraries as well as a
949 compiler that understands the C++ language---and under some
950 circumstances, you might want to compile programs from standard input,
951 or otherwise without a suffix that flags them as C++ programs.
952 @command{g++} is a program that calls GCC with the default language
953 set to C++, and automatically specifies linking against the C++
954 library. On many systems, @command{g++} is also
955 installed with the name @command{c++}.
957 @cindex invoking @command{g++}
958 When you compile C++ programs, you may specify many of the same
959 command-line options that you use for compiling programs in any
960 language; or command-line options meaningful for C and related
961 languages; or options that are meaningful only for C++ programs.
962 @xref{C Dialect Options,,Options Controlling C Dialect}, for
963 explanations of options for languages related to C@.
964 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
965 explanations of options that are meaningful only for C++ programs.
967 @node C Dialect Options
968 @section Options Controlling C Dialect
969 @cindex dialect options
970 @cindex language dialect options
971 @cindex options, dialect
973 The following options control the dialect of C (or languages derived
974 from C, such as C++ and Objective-C) that the compiler accepts:
981 In C mode, support all ISO C90 programs. In C++ mode,
982 remove GNU extensions that conflict with ISO C++.
984 This turns off certain features of GCC that are incompatible with ISO
985 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
986 such as the @code{asm} and @code{typeof} keywords, and
987 predefined macros such as @code{unix} and @code{vax} that identify the
988 type of system you are using. It also enables the undesirable and
989 rarely used ISO trigraph feature. For the C compiler,
990 it disables recognition of C++ style @samp{//} comments as well as
991 the @code{inline} keyword.
993 The alternate keywords @code{__asm__}, @code{__extension__},
994 @code{__inline__} and @code{__typeof__} continue to work despite
995 @option{-ansi}. You would not want to use them in an ISO C program, of
996 course, but it is useful to put them in header files that might be included
997 in compilations done with @option{-ansi}. Alternate predefined macros
998 such as @code{__unix__} and @code{__vax__} are also available, with or
999 without @option{-ansi}.
1001 The @option{-ansi} option does not cause non-ISO programs to be
1002 rejected gratuitously. For that, @option{-pedantic} is required in
1003 addition to @option{-ansi}. @xref{Warning Options}.
1005 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1006 option is used. Some header files may notice this macro and refrain
1007 from declaring certain functions or defining certain macros that the
1008 ISO standard doesn't call for; this is to avoid interfering with any
1009 programs that might use these names for other things.
1011 Functions which would normally be built in but do not have semantics
1012 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1013 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1014 built-in functions provided by GCC}, for details of the functions
1019 Determine the language standard. This option is currently only
1020 supported when compiling C or C++. A value for this option must be
1021 provided; possible values are
1026 ISO C90 (same as @option{-ansi}).
1028 @item iso9899:199409
1029 ISO C90 as modified in amendment 1.
1035 ISO C99. Note that this standard is not yet fully supported; see
1036 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1037 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1040 Default, ISO C90 plus GNU extensions (including some C99 features).
1044 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1045 this will become the default. The name @samp{gnu9x} is deprecated.
1048 The 1998 ISO C++ standard plus amendments.
1051 The same as @option{-std=c++98} plus GNU extensions. This is the
1052 default for C++ code.
1055 Even when this option is not specified, you can still use some of the
1056 features of newer standards in so far as they do not conflict with
1057 previous C standards. For example, you may use @code{__restrict__} even
1058 when @option{-std=c99} is not specified.
1060 The @option{-std} options specifying some version of ISO C have the same
1061 effects as @option{-ansi}, except that features that were not in ISO C90
1062 but are in the specified version (for example, @samp{//} comments and
1063 the @code{inline} keyword in ISO C99) are not disabled.
1065 @xref{Standards,,Language Standards Supported by GCC}, for details of
1066 these standard versions.
1068 @item -aux-info @var{filename}
1070 Output to the given filename prototyped declarations for all functions
1071 declared and/or defined in a translation unit, including those in header
1072 files. This option is silently ignored in any language other than C@.
1074 Besides declarations, the file indicates, in comments, the origin of
1075 each declaration (source file and line), whether the declaration was
1076 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1077 @samp{O} for old, respectively, in the first character after the line
1078 number and the colon), and whether it came from a declaration or a
1079 definition (@samp{C} or @samp{F}, respectively, in the following
1080 character). In the case of function definitions, a K&R-style list of
1081 arguments followed by their declarations is also provided, inside
1082 comments, after the declaration.
1086 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1087 keyword, so that code can use these words as identifiers. You can use
1088 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1089 instead. @option{-ansi} implies @option{-fno-asm}.
1091 In C++, this switch only affects the @code{typeof} keyword, since
1092 @code{asm} and @code{inline} are standard keywords. You may want to
1093 use the @option{-fno-gnu-keywords} flag instead, which has the same
1094 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1095 switch only affects the @code{asm} and @code{typeof} keywords, since
1096 @code{inline} is a standard keyword in ISO C99.
1099 @itemx -fno-builtin-@var{function}
1100 @opindex fno-builtin
1101 @cindex built-in functions
1102 Don't recognize built-in functions that do not begin with
1103 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1104 functions provided by GCC}, for details of the functions affected,
1105 including those which are not built-in functions when @option{-ansi} or
1106 @option{-std} options for strict ISO C conformance are used because they
1107 do not have an ISO standard meaning.
1109 GCC normally generates special code to handle certain built-in functions
1110 more efficiently; for instance, calls to @code{alloca} may become single
1111 instructions that adjust the stack directly, and calls to @code{memcpy}
1112 may become inline copy loops. The resulting code is often both smaller
1113 and faster, but since the function calls no longer appear as such, you
1114 cannot set a breakpoint on those calls, nor can you change the behavior
1115 of the functions by linking with a different library.
1117 With the @option{-fno-builtin-@var{function}} option
1118 only the built-in function @var{function} is
1119 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1120 function is named this is not built-in in this version of GCC, this
1121 option is ignored. There is no corresponding
1122 @option{-fbuiltin-@var{function}} option; if you wish to enable
1123 built-in functions selectively when using @option{-fno-builtin} or
1124 @option{-ffreestanding}, you may define macros such as:
1127 #define abs(n) __builtin_abs ((n))
1128 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1133 @cindex hosted environment
1135 Assert that compilation takes place in a hosted environment. This implies
1136 @option{-fbuiltin}. A hosted environment is one in which the
1137 entire standard library is available, and in which @code{main} has a return
1138 type of @code{int}. Examples are nearly everything except a kernel.
1139 This is equivalent to @option{-fno-freestanding}.
1141 @item -ffreestanding
1142 @opindex ffreestanding
1143 @cindex hosted environment
1145 Assert that compilation takes place in a freestanding environment. This
1146 implies @option{-fno-builtin}. A freestanding environment
1147 is one in which the standard library may not exist, and program startup may
1148 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1149 This is equivalent to @option{-fno-hosted}.
1151 @xref{Standards,,Language Standards Supported by GCC}, for details of
1152 freestanding and hosted environments.
1154 @item -fms-extensions
1155 @opindex fms-extensions
1156 Accept some non-standard constructs used in Microsoft header files.
1160 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1161 options for strict ISO C conformance) implies @option{-trigraphs}.
1163 @cindex traditional C language
1164 @cindex C language, traditional
1166 @itemx -traditional-cpp
1167 @opindex traditional-cpp
1168 @opindex traditional
1169 Formerly, these options caused GCC to attempt to emulate a pre-standard
1170 C compiler. They are now only supported with the @option{-E} switch.
1171 The preprocessor continues to support a pre-standard mode. See the GNU
1172 CPP manual for details.
1174 @item -fcond-mismatch
1175 @opindex fcond-mismatch
1176 Allow conditional expressions with mismatched types in the second and
1177 third arguments. The value of such an expression is void. This option
1178 is not supported for C++.
1180 @item -funsigned-char
1181 @opindex funsigned-char
1182 Let the type @code{char} be unsigned, like @code{unsigned char}.
1184 Each kind of machine has a default for what @code{char} should
1185 be. It is either like @code{unsigned char} by default or like
1186 @code{signed char} by default.
1188 Ideally, a portable program should always use @code{signed char} or
1189 @code{unsigned char} when it depends on the signedness of an object.
1190 But many programs have been written to use plain @code{char} and
1191 expect it to be signed, or expect it to be unsigned, depending on the
1192 machines they were written for. This option, and its inverse, let you
1193 make such a program work with the opposite default.
1195 The type @code{char} is always a distinct type from each of
1196 @code{signed char} or @code{unsigned char}, even though its behavior
1197 is always just like one of those two.
1200 @opindex fsigned-char
1201 Let the type @code{char} be signed, like @code{signed char}.
1203 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1204 the negative form of @option{-funsigned-char}. Likewise, the option
1205 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1207 @item -fsigned-bitfields
1208 @itemx -funsigned-bitfields
1209 @itemx -fno-signed-bitfields
1210 @itemx -fno-unsigned-bitfields
1211 @opindex fsigned-bitfields
1212 @opindex funsigned-bitfields
1213 @opindex fno-signed-bitfields
1214 @opindex fno-unsigned-bitfields
1215 These options control whether a bit-field is signed or unsigned, when the
1216 declaration does not use either @code{signed} or @code{unsigned}. By
1217 default, such a bit-field is signed, because this is consistent: the
1218 basic integer types such as @code{int} are signed types.
1220 @item -fwritable-strings
1221 @opindex fwritable-strings
1222 Store string constants in the writable data segment and don't uniquize
1223 them. This is for compatibility with old programs which assume they can
1224 write into string constants.
1226 Writing into string constants is a very bad idea; ``constants'' should
1230 @node C++ Dialect Options
1231 @section Options Controlling C++ Dialect
1233 @cindex compiler options, C++
1234 @cindex C++ options, command line
1235 @cindex options, C++
1236 This section describes the command-line options that are only meaningful
1237 for C++ programs; but you can also use most of the GNU compiler options
1238 regardless of what language your program is in. For example, you
1239 might compile a file @code{firstClass.C} like this:
1242 g++ -g -frepo -O -c firstClass.C
1246 In this example, only @option{-frepo} is an option meant
1247 only for C++ programs; you can use the other options with any
1248 language supported by GCC@.
1250 Here is a list of options that are @emph{only} for compiling C++ programs:
1254 @item -fabi-version=@var{n}
1255 @opindex fabi-version
1256 Use version @var{n} of the C++ ABI. Version 1 is the version of the C++
1257 ABI that first appeared in G++ 3.2. Version 0 will always be the
1258 version that conforms most closely to the C++ ABI specification.
1259 Therefore, the ABI obtained using version 0 will change as ABI bugs are
1262 The default is version 1.
1264 @item -fno-access-control
1265 @opindex fno-access-control
1266 Turn off all access checking. This switch is mainly useful for working
1267 around bugs in the access control code.
1271 Check that the pointer returned by @code{operator new} is non-null
1272 before attempting to modify the storage allocated. The current Working
1273 Paper requires that @code{operator new} never return a null pointer, so
1274 this check is normally unnecessary.
1276 An alternative to using this option is to specify that your
1277 @code{operator new} does not throw any exceptions; if you declare it
1278 @samp{throw()}, G++ will check the return value. See also @samp{new
1281 @item -fconserve-space
1282 @opindex fconserve-space
1283 Put uninitialized or runtime-initialized global variables into the
1284 common segment, as C does. This saves space in the executable at the
1285 cost of not diagnosing duplicate definitions. If you compile with this
1286 flag and your program mysteriously crashes after @code{main()} has
1287 completed, you may have an object that is being destroyed twice because
1288 two definitions were merged.
1290 This option is no longer useful on most targets, now that support has
1291 been added for putting variables into BSS without making them common.
1293 @item -fno-const-strings
1294 @opindex fno-const-strings
1295 Give string constants type @code{char *} instead of type @code{const
1296 char *}. By default, G++ uses type @code{const char *} as required by
1297 the standard. Even if you use @option{-fno-const-strings}, you cannot
1298 actually modify the value of a string constant, unless you also use
1299 @option{-fwritable-strings}.
1301 This option might be removed in a future release of G++. For maximum
1302 portability, you should structure your code so that it works with
1303 string constants that have type @code{const char *}.
1305 @item -fdollars-in-identifiers
1306 @opindex fdollars-in-identifiers
1307 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1308 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1309 @samp{$} by default on most target systems, but there are a few exceptions.)
1310 Traditional C allowed the character @samp{$} to form part of
1311 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1313 @item -fno-elide-constructors
1314 @opindex fno-elide-constructors
1315 The C++ standard allows an implementation to omit creating a temporary
1316 which is only used to initialize another object of the same type.
1317 Specifying this option disables that optimization, and forces G++ to
1318 call the copy constructor in all cases.
1320 @item -fno-enforce-eh-specs
1321 @opindex fno-enforce-eh-specs
1322 Don't check for violation of exception specifications at runtime. This
1323 option violates the C++ standard, but may be useful for reducing code
1324 size in production builds, much like defining @samp{NDEBUG}. The compiler
1325 will still optimize based on the exception specifications.
1327 @item -fexternal-templates
1328 @opindex fexternal-templates
1330 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1331 template instantiation; template instances are emitted or not according
1332 to the location of the template definition. @xref{Template
1333 Instantiation}, for more information.
1335 This option is deprecated.
1337 @item -falt-external-templates
1338 @opindex falt-external-templates
1339 Similar to @option{-fexternal-templates}, but template instances are
1340 emitted or not according to the place where they are first instantiated.
1341 @xref{Template Instantiation}, for more information.
1343 This option is deprecated.
1346 @itemx -fno-for-scope
1348 @opindex fno-for-scope
1349 If @option{-ffor-scope} is specified, the scope of variables declared in
1350 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1351 as specified by the C++ standard.
1352 If @option{-fno-for-scope} is specified, the scope of variables declared in
1353 a @i{for-init-statement} extends to the end of the enclosing scope,
1354 as was the case in old versions of G++, and other (traditional)
1355 implementations of C++.
1357 The default if neither flag is given to follow the standard,
1358 but to allow and give a warning for old-style code that would
1359 otherwise be invalid, or have different behavior.
1361 @item -fno-gnu-keywords
1362 @opindex fno-gnu-keywords
1363 Do not recognize @code{typeof} as a keyword, so that code can use this
1364 word as an identifier. You can use the keyword @code{__typeof__} instead.
1365 @option{-ansi} implies @option{-fno-gnu-keywords}.
1367 @item -fno-implicit-templates
1368 @opindex fno-implicit-templates
1369 Never emit code for non-inline templates which are instantiated
1370 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1371 @xref{Template Instantiation}, for more information.
1373 @item -fno-implicit-inline-templates
1374 @opindex fno-implicit-inline-templates
1375 Don't emit code for implicit instantiations of inline templates, either.
1376 The default is to handle inlines differently so that compiles with and
1377 without optimization will need the same set of explicit instantiations.
1379 @item -fno-implement-inlines
1380 @opindex fno-implement-inlines
1381 To save space, do not emit out-of-line copies of inline functions
1382 controlled by @samp{#pragma implementation}. This will cause linker
1383 errors if these functions are not inlined everywhere they are called.
1385 @item -fms-extensions
1386 @opindex fms-extensions
1387 Disable pedantic warnings about constructs used in MFC, such as implicit
1388 int and getting a pointer to member function via non-standard syntax.
1390 @item -fno-nonansi-builtins
1391 @opindex fno-nonansi-builtins
1392 Disable built-in declarations of functions that are not mandated by
1393 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1394 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1396 @item -fno-operator-names
1397 @opindex fno-operator-names
1398 Do not treat the operator name keywords @code{and}, @code{bitand},
1399 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1400 synonyms as keywords.
1402 @item -fno-optional-diags
1403 @opindex fno-optional-diags
1404 Disable diagnostics that the standard says a compiler does not need to
1405 issue. Currently, the only such diagnostic issued by G++ is the one for
1406 a name having multiple meanings within a class.
1409 @opindex fpermissive
1410 Downgrade messages about nonconformant code from errors to warnings. By
1411 default, G++ effectively sets @option{-pedantic-errors} without
1412 @option{-pedantic}; this option reverses that. This behavior and this
1413 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1417 Enable automatic template instantiation at link time. This option also
1418 implies @option{-fno-implicit-templates}. @xref{Template
1419 Instantiation}, for more information.
1423 Disable generation of information about every class with virtual
1424 functions for use by the C++ runtime type identification features
1425 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1426 of the language, you can save some space by using this flag. Note that
1427 exception handling uses the same information, but it will generate it as
1432 Emit statistics about front-end processing at the end of the compilation.
1433 This information is generally only useful to the G++ development team.
1435 @item -ftemplate-depth-@var{n}
1436 @opindex ftemplate-depth
1437 Set the maximum instantiation depth for template classes to @var{n}.
1438 A limit on the template instantiation depth is needed to detect
1439 endless recursions during template class instantiation. ANSI/ISO C++
1440 conforming programs must not rely on a maximum depth greater than 17.
1442 @item -fuse-cxa-atexit
1443 @opindex fuse-cxa-atexit
1444 Register destructors for objects with static storage duration with the
1445 @code{__cxa_atexit} function rather than the @code{atexit} function.
1446 This option is required for fully standards-compliant handling of static
1447 destructors, but will only work if your C library supports
1448 @code{__cxa_atexit}.
1452 Emit special relocations for vtables and virtual function references
1453 so that the linker can identify unused virtual functions and zero out
1454 vtable slots that refer to them. This is most useful with
1455 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1456 also discard the functions themselves.
1458 This optimization requires GNU as and GNU ld. Not all systems support
1459 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1463 Do not use weak symbol support, even if it is provided by the linker.
1464 By default, G++ will use weak symbols if they are available. This
1465 option exists only for testing, and should not be used by end-users;
1466 it will result in inferior code and has no benefits. This option may
1467 be removed in a future release of G++.
1471 Do not search for header files in the standard directories specific to
1472 C++, but do still search the other standard directories. (This option
1473 is used when building the C++ library.)
1476 In addition, these optimization, warning, and code generation options
1477 have meanings only for C++ programs:
1480 @item -fno-default-inline
1481 @opindex fno-default-inline
1482 Do not assume @samp{inline} for functions defined inside a class scope.
1483 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1484 functions will have linkage like inline functions; they just won't be
1487 @item -Wabi @r{(C++ only)}
1489 Warn when G++ generates code that is probably not compatible with the
1490 vendor-neutral C++ ABI. Although an effort has been made to warn about
1491 all such cases, there are probably some cases that are not warned about,
1492 even though G++ is generating incompatible code. There may also be
1493 cases where warnings are emitted even though the code that is generated
1496 You should rewrite your code to avoid these warnings if you are
1497 concerned about the fact that code generated by G++ may not be binary
1498 compatible with code generated by other compilers.
1500 The known incompatibilites at this point include:
1505 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1506 pack data into the same byte as a base class. For example:
1509 struct A @{ virtual void f(); int f1 : 1; @};
1510 struct B : public A @{ int f2 : 1; @};
1514 In this case, G++ will place @code{B::f2} into the same byte
1515 as@code{A::f1}; other compilers will not. You can avoid this problem
1516 by explicitly padding @code{A} so that its size is a multiple of the
1517 byte size on your platform; that will cause G++ and other compilers to
1518 layout @code{B} identically.
1521 Incorrect handling of tail-padding for virtual bases. G++ does not use
1522 tail padding when laying out virtual bases. For example:
1525 struct A @{ virtual void f(); char c1; @};
1526 struct B @{ B(); char c2; @};
1527 struct C : public A, public virtual B @{@};
1531 In this case, G++ will not place @code{B} into the tail-padding for
1532 @code{A}; other compilers will. You can avoid this problem by
1533 explicitly padding @code{A} so that its size is a multiple of its
1534 alignment (ignoring virtual base classes); that will cause G++ and other
1535 compilers to layout @code{C} identically.
1538 Incorrect handling of bit-fields with declared widths greater than that
1539 of their underlying types, when the bit-fields appear in a union. For
1543 union U @{ int i : 4096; @};
1547 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1548 union too small by the number of bits in an @code{int}.
1551 Empty classes can be placed at incorrect offsets. For example:
1561 struct C : public B, public A @{@};
1565 G++ will place the @code{A} base class of @code{C} at a non-zero offset;
1566 it should be placed at offset zero. G++ mistakenly believes that the
1567 @code{A} data member of @code{B} is already at offset zero.
1570 Names of template functions whose types involve @code{typename} or
1571 template template parameters can be mangled incorrectly.
1574 template <typename Q>
1575 void f(typename Q::X) @{@}
1577 template <template <typename> class Q>
1578 void f(typename Q<int>::X) @{@}
1582 Instantions of these templates may be mangled incorrectly.
1586 @item -Wctor-dtor-privacy @r{(C++ only)}
1587 @opindex Wctor-dtor-privacy
1588 Warn when a class seems unusable, because all the constructors or
1589 destructors in a class are private and the class has no friends or
1590 public static member functions. This warning is enabled by default.
1592 @item -Wnon-virtual-dtor @r{(C++ only)}
1593 @opindex Wnon-virtual-dtor
1594 Warn when a class declares a non-virtual destructor that should probably
1595 be virtual, because it looks like the class will be used polymorphically.
1596 This warning is enabled by @option{-Wall}.
1598 @item -Wreorder @r{(C++ only)}
1600 @cindex reordering, warning
1601 @cindex warning for reordering of member initializers
1602 Warn when the order of member initializers given in the code does not
1603 match the order in which they must be executed. For instance:
1609 A(): j (0), i (1) @{ @}
1613 Here the compiler will warn that the member initializers for @samp{i}
1614 and @samp{j} will be rearranged to match the declaration order of the
1615 members. This warning is enabled by @option{-Wall}.
1618 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1621 @item -Weffc++ @r{(C++ only)}
1623 Warn about violations of the following style guidelines from Scott Meyers'
1624 @cite{Effective C++} book:
1628 Item 11: Define a copy constructor and an assignment operator for classes
1629 with dynamically allocated memory.
1632 Item 12: Prefer initialization to assignment in constructors.
1635 Item 14: Make destructors virtual in base classes.
1638 Item 15: Have @code{operator=} return a reference to @code{*this}.
1641 Item 23: Don't try to return a reference when you must return an object.
1645 and about violations of the following style guidelines from Scott Meyers'
1646 @cite{More Effective C++} book:
1650 Item 6: Distinguish between prefix and postfix forms of increment and
1651 decrement operators.
1654 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1658 If you use this option, you should be aware that the standard library
1659 headers do not obey all of these guidelines; you can use @samp{grep -v}
1660 to filter out those warnings.
1662 @item -Wno-deprecated @r{(C++ only)}
1663 @opindex Wno-deprecated
1664 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1666 @item -Wno-non-template-friend @r{(C++ only)}
1667 @opindex Wno-non-template-friend
1668 Disable warnings when non-templatized friend functions are declared
1669 within a template. With the advent of explicit template specification
1670 support in G++, if the name of the friend is an unqualified-id (i.e.,
1671 @samp{friend foo(int)}), the C++ language specification demands that the
1672 friend declare or define an ordinary, nontemplate function. (Section
1673 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1674 could be interpreted as a particular specialization of a templatized
1675 function. Because this non-conforming behavior is no longer the default
1676 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1677 check existing code for potential trouble spots, and is on by default.
1678 This new compiler behavior can be turned off with
1679 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1680 but disables the helpful warning.
1682 @item -Wold-style-cast @r{(C++ only)}
1683 @opindex Wold-style-cast
1684 Warn if an old-style (C-style) cast to a non-void type is used within
1685 a C++ program. The new-style casts (@samp{static_cast},
1686 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1687 unintended effects, and much easier to grep for.
1689 @item -Woverloaded-virtual @r{(C++ only)}
1690 @opindex Woverloaded-virtual
1691 @cindex overloaded virtual fn, warning
1692 @cindex warning for overloaded virtual fn
1693 Warn when a function declaration hides virtual functions from a
1694 base class. For example, in:
1701 struct B: public A @{
1706 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1714 will fail to compile.
1716 @item -Wno-pmf-conversions @r{(C++ only)}
1717 @opindex Wno-pmf-conversions
1718 Disable the diagnostic for converting a bound pointer to member function
1721 @item -Wsign-promo @r{(C++ only)}
1722 @opindex Wsign-promo
1723 Warn when overload resolution chooses a promotion from unsigned or
1724 enumeral type to a signed type over a conversion to an unsigned type of
1725 the same size. Previous versions of G++ would try to preserve
1726 unsignedness, but the standard mandates the current behavior.
1728 @item -Wsynth @r{(C++ only)}
1730 @cindex warning for synthesized methods
1731 @cindex synthesized methods, warning
1732 Warn when G++'s synthesis behavior does not match that of cfront. For
1738 A& operator = (int);
1748 In this example, G++ will synthesize a default @samp{A& operator =
1749 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1752 @node Objective-C Dialect Options
1753 @section Options Controlling Objective-C Dialect
1755 @cindex compiler options, Objective-C
1756 @cindex Objective-C options, command line
1757 @cindex options, Objective-C
1758 This section describes the command-line options that are only meaningful
1759 for Objective-C programs; but you can also use most of the GNU compiler
1760 options regardless of what language your program is in. For example,
1761 you might compile a file @code{some_class.m} like this:
1764 gcc -g -fgnu-runtime -O -c some_class.m
1768 In this example, only @option{-fgnu-runtime} is an option meant only for
1769 Objective-C programs; you can use the other options with any language
1772 Here is a list of options that are @emph{only} for compiling Objective-C
1776 @item -fconstant-string-class=@var{class-name}
1777 @opindex fconstant-string-class
1778 Use @var{class-name} as the name of the class to instantiate for each
1779 literal string specified with the syntax @code{@@"@dots{}"}. The default
1780 class name is @code{NXConstantString}.
1783 @opindex fgnu-runtime
1784 Generate object code compatible with the standard GNU Objective-C
1785 runtime. This is the default for most types of systems.
1787 @item -fnext-runtime
1788 @opindex fnext-runtime
1789 Generate output compatible with the NeXT runtime. This is the default
1790 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1791 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1796 Dump interface declarations for all classes seen in the source file to a
1797 file named @file{@var{sourcename}.decl}.
1800 @opindex Wno-protocol
1801 If a class is declared to implement a protocol, a warning is issued for
1802 every method in the protocol that is not implemented by the class. The
1803 default behavior is to issue a warning for every method not explicitly
1804 implemented in the class, even if a method implementation is inherited
1805 from the superclass. If you use the @code{-Wno-protocol} option, then
1806 methods inherited from the superclass are considered to be implemented,
1807 and no warning is issued for them.
1811 Warn if multiple methods of different types for the same selector are
1812 found during compilation. The check is performed on the list of methods
1813 in the final stage of compilation. Additionally, a check is performed
1814 that for each selector appearing in a @code{@@selector(@dots{})}
1815 expression, a corresponding method with that selector has been found
1816 during compilation. Because these checks scan the method table only at
1817 the end of compilation, these warnings are not produced if the final
1818 stage of compilation is not reached, for example because an error is
1819 found during compilation, or because the @code{-fsyntax-only} option is
1822 @item -Wundeclared-selector
1823 @opindex Wundeclared-selector
1824 Warn if a @code{@@selector(@dots{})} expression referring to an
1825 undeclared selector is found. A selector is considered undeclared if no
1826 method with that name has been declared (explicitly, in an
1827 @code{@@interface} or @code{@@protocol} declaration, or implicitly, in
1828 an @code{@@implementation} section) before the
1829 @code{@@selector(@dots{})} expression. This option always performs its
1830 checks as soon as a @code{@@selector(@dots{})} expression is found
1831 (while @code{-Wselector} only performs its checks in the final stage of
1832 compilation), and so additionally enforces the coding style convention
1833 that methods and selectors must be declared before being used.
1835 @c not documented because only avail via -Wp
1836 @c @item -print-objc-runtime-info
1840 @node Language Independent Options
1841 @section Options to Control Diagnostic Messages Formatting
1842 @cindex options to control diagnostics formatting
1843 @cindex diagnostic messages
1844 @cindex message formatting
1846 Traditionally, diagnostic messages have been formatted irrespective of
1847 the output device's aspect (e.g.@: its width, @dots{}). The options described
1848 below can be used to control the diagnostic messages formatting
1849 algorithm, e.g.@: how many characters per line, how often source location
1850 information should be reported. Right now, only the C++ front end can
1851 honor these options. However it is expected, in the near future, that
1852 the remaining front ends would be able to digest them correctly.
1855 @item -fmessage-length=@var{n}
1856 @opindex fmessage-length
1857 Try to format error messages so that they fit on lines of about @var{n}
1858 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1859 the front ends supported by GCC@. If @var{n} is zero, then no
1860 line-wrapping will be done; each error message will appear on a single
1863 @opindex fdiagnostics-show-location
1864 @item -fdiagnostics-show-location=once
1865 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1866 reporter to emit @emph{once} source location information; that is, in
1867 case the message is too long to fit on a single physical line and has to
1868 be wrapped, the source location won't be emitted (as prefix) again,
1869 over and over, in subsequent continuation lines. This is the default
1872 @item -fdiagnostics-show-location=every-line
1873 Only meaningful in line-wrapping mode. Instructs the diagnostic
1874 messages reporter to emit the same source location information (as
1875 prefix) for physical lines that result from the process of breaking
1876 a message which is too long to fit on a single line.
1880 @node Warning Options
1881 @section Options to Request or Suppress Warnings
1882 @cindex options to control warnings
1883 @cindex warning messages
1884 @cindex messages, warning
1885 @cindex suppressing warnings
1887 Warnings are diagnostic messages that report constructions which
1888 are not inherently erroneous but which are risky or suggest there
1889 may have been an error.
1891 You can request many specific warnings with options beginning @samp{-W},
1892 for example @option{-Wimplicit} to request warnings on implicit
1893 declarations. Each of these specific warning options also has a
1894 negative form beginning @samp{-Wno-} to turn off warnings;
1895 for example, @option{-Wno-implicit}. This manual lists only one of the
1896 two forms, whichever is not the default.
1898 The following options control the amount and kinds of warnings produced
1899 by GCC; for further, language-specific options also refer to
1900 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
1903 @cindex syntax checking
1905 @opindex fsyntax-only
1906 Check the code for syntax errors, but don't do anything beyond that.
1910 Issue all the warnings demanded by strict ISO C and ISO C++;
1911 reject all programs that use forbidden extensions, and some other
1912 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1913 version of the ISO C standard specified by any @option{-std} option used.
1915 Valid ISO C and ISO C++ programs should compile properly with or without
1916 this option (though a rare few will require @option{-ansi} or a
1917 @option{-std} option specifying the required version of ISO C)@. However,
1918 without this option, certain GNU extensions and traditional C and C++
1919 features are supported as well. With this option, they are rejected.
1921 @option{-pedantic} does not cause warning messages for use of the
1922 alternate keywords whose names begin and end with @samp{__}. Pedantic
1923 warnings are also disabled in the expression that follows
1924 @code{__extension__}. However, only system header files should use
1925 these escape routes; application programs should avoid them.
1926 @xref{Alternate Keywords}.
1928 Some users try to use @option{-pedantic} to check programs for strict ISO
1929 C conformance. They soon find that it does not do quite what they want:
1930 it finds some non-ISO practices, but not all---only those for which
1931 ISO C @emph{requires} a diagnostic, and some others for which
1932 diagnostics have been added.
1934 A feature to report any failure to conform to ISO C might be useful in
1935 some instances, but would require considerable additional work and would
1936 be quite different from @option{-pedantic}. We don't have plans to
1937 support such a feature in the near future.
1939 Where the standard specified with @option{-std} represents a GNU
1940 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1941 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1942 extended dialect is based. Warnings from @option{-pedantic} are given
1943 where they are required by the base standard. (It would not make sense
1944 for such warnings to be given only for features not in the specified GNU
1945 C dialect, since by definition the GNU dialects of C include all
1946 features the compiler supports with the given option, and there would be
1947 nothing to warn about.)
1949 @item -pedantic-errors
1950 @opindex pedantic-errors
1951 Like @option{-pedantic}, except that errors are produced rather than
1956 Inhibit all warning messages.
1960 Inhibit warning messages about the use of @samp{#import}.
1962 @item -Wchar-subscripts
1963 @opindex Wchar-subscripts
1964 Warn if an array subscript has type @code{char}. This is a common cause
1965 of error, as programmers often forget that this type is signed on some
1970 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1971 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1975 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1976 the arguments supplied have types appropriate to the format string
1977 specified, and that the conversions specified in the format string make
1978 sense. This includes standard functions, and others specified by format
1979 attributes (@pxref{Function Attributes}), in the @code{printf},
1980 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1981 not in the C standard) families.
1983 The formats are checked against the format features supported by GNU
1984 libc version 2.2. These include all ISO C90 and C99 features, as well
1985 as features from the Single Unix Specification and some BSD and GNU
1986 extensions. Other library implementations may not support all these
1987 features; GCC does not support warning about features that go beyond a
1988 particular library's limitations. However, if @option{-pedantic} is used
1989 with @option{-Wformat}, warnings will be given about format features not
1990 in the selected standard version (but not for @code{strfmon} formats,
1991 since those are not in any version of the C standard). @xref{C Dialect
1992 Options,,Options Controlling C Dialect}.
1994 Since @option{-Wformat} also checks for null format arguments for
1995 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
1997 @option{-Wformat} is included in @option{-Wall}. For more control over some
1998 aspects of format checking, the options @option{-Wno-format-y2k},
1999 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2000 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2001 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2003 @item -Wno-format-y2k
2004 @opindex Wno-format-y2k
2005 If @option{-Wformat} is specified, do not warn about @code{strftime}
2006 formats which may yield only a two-digit year.
2008 @item -Wno-format-extra-args
2009 @opindex Wno-format-extra-args
2010 If @option{-Wformat} is specified, do not warn about excess arguments to a
2011 @code{printf} or @code{scanf} format function. The C standard specifies
2012 that such arguments are ignored.
2014 Where the unused arguments lie between used arguments that are
2015 specified with @samp{$} operand number specifications, normally
2016 warnings are still given, since the implementation could not know what
2017 type to pass to @code{va_arg} to skip the unused arguments. However,
2018 in the case of @code{scanf} formats, this option will suppress the
2019 warning if the unused arguments are all pointers, since the Single
2020 Unix Specification says that such unused arguments are allowed.
2022 @item -Wno-format-zero-length
2023 @opindex Wno-format-zero-length
2024 If @option{-Wformat} is specified, do not warn about zero-length formats.
2025 The C standard specifies that zero-length formats are allowed.
2027 @item -Wformat-nonliteral
2028 @opindex Wformat-nonliteral
2029 If @option{-Wformat} is specified, also warn if the format string is not a
2030 string literal and so cannot be checked, unless the format function
2031 takes its format arguments as a @code{va_list}.
2033 @item -Wformat-security
2034 @opindex Wformat-security
2035 If @option{-Wformat} is specified, also warn about uses of format
2036 functions that represent possible security problems. At present, this
2037 warns about calls to @code{printf} and @code{scanf} functions where the
2038 format string is not a string literal and there are no format arguments,
2039 as in @code{printf (foo);}. This may be a security hole if the format
2040 string came from untrusted input and contains @samp{%n}. (This is
2041 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2042 in future warnings may be added to @option{-Wformat-security} that are not
2043 included in @option{-Wformat-nonliteral}.)
2047 Enable @option{-Wformat} plus format checks not included in
2048 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2049 -Wformat-nonliteral -Wformat-security}.
2053 Enable warning about passing a null pointer for arguments marked as
2054 requiring a non-null value by the @code{nonnull} function attribute.
2056 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2057 can be disabled with the @option{-Wno-nonnull} option.
2059 @item -Wimplicit-int
2060 @opindex Wimplicit-int
2061 Warn when a declaration does not specify a type.
2063 @item -Wimplicit-function-declaration
2064 @itemx -Werror-implicit-function-declaration
2065 @opindex Wimplicit-function-declaration
2066 @opindex Werror-implicit-function-declaration
2067 Give a warning (or error) whenever a function is used before being
2072 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2076 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2077 function with external linkage, returning int, taking either zero
2078 arguments, two, or three arguments of appropriate types.
2080 @item -Wmissing-braces
2081 @opindex Wmissing-braces
2082 Warn if an aggregate or union initializer is not fully bracketed. In
2083 the following example, the initializer for @samp{a} is not fully
2084 bracketed, but that for @samp{b} is fully bracketed.
2087 int a[2][2] = @{ 0, 1, 2, 3 @};
2088 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2092 @opindex Wparentheses
2093 Warn if parentheses are omitted in certain contexts, such
2094 as when there is an assignment in a context where a truth value
2095 is expected, or when operators are nested whose precedence people
2096 often get confused about.
2098 Also warn about constructions where there may be confusion to which
2099 @code{if} statement an @code{else} branch belongs. Here is an example of
2114 In C, every @code{else} branch belongs to the innermost possible @code{if}
2115 statement, which in this example is @code{if (b)}. This is often not
2116 what the programmer expected, as illustrated in the above example by
2117 indentation the programmer chose. When there is the potential for this
2118 confusion, GCC will issue a warning when this flag is specified.
2119 To eliminate the warning, add explicit braces around the innermost
2120 @code{if} statement so there is no way the @code{else} could belong to
2121 the enclosing @code{if}. The resulting code would look like this:
2137 @item -Wsequence-point
2138 @opindex Wsequence-point
2139 Warn about code that may have undefined semantics because of violations
2140 of sequence point rules in the C standard.
2142 The C standard defines the order in which expressions in a C program are
2143 evaluated in terms of @dfn{sequence points}, which represent a partial
2144 ordering between the execution of parts of the program: those executed
2145 before the sequence point, and those executed after it. These occur
2146 after the evaluation of a full expression (one which is not part of a
2147 larger expression), after the evaluation of the first operand of a
2148 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2149 function is called (but after the evaluation of its arguments and the
2150 expression denoting the called function), and in certain other places.
2151 Other than as expressed by the sequence point rules, the order of
2152 evaluation of subexpressions of an expression is not specified. All
2153 these rules describe only a partial order rather than a total order,
2154 since, for example, if two functions are called within one expression
2155 with no sequence point between them, the order in which the functions
2156 are called is not specified. However, the standards committee have
2157 ruled that function calls do not overlap.
2159 It is not specified when between sequence points modifications to the
2160 values of objects take effect. Programs whose behavior depends on this
2161 have undefined behavior; the C standard specifies that ``Between the
2162 previous and next sequence point an object shall have its stored value
2163 modified at most once by the evaluation of an expression. Furthermore,
2164 the prior value shall be read only to determine the value to be
2165 stored.''. If a program breaks these rules, the results on any
2166 particular implementation are entirely unpredictable.
2168 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2169 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2170 diagnosed by this option, and it may give an occasional false positive
2171 result, but in general it has been found fairly effective at detecting
2172 this sort of problem in programs.
2174 The present implementation of this option only works for C programs. A
2175 future implementation may also work for C++ programs.
2177 The C standard is worded confusingly, therefore there is some debate
2178 over the precise meaning of the sequence point rules in subtle cases.
2179 Links to discussions of the problem, including proposed formal
2180 definitions, may be found on our readings page, at
2181 @w{@uref{http://gcc.gnu.org/readings.html}}.
2184 @opindex Wreturn-type
2185 Warn whenever a function is defined with a return-type that defaults to
2186 @code{int}. Also warn about any @code{return} statement with no
2187 return-value in a function whose return-type is not @code{void}.
2189 For C++, a function without return type always produces a diagnostic
2190 message, even when @option{-Wno-return-type} is specified. The only
2191 exceptions are @samp{main} and functions defined in system headers.
2195 Warn whenever a @code{switch} statement has an index of enumeral type
2196 and lacks a @code{case} for one or more of the named codes of that
2197 enumeration. (The presence of a @code{default} label prevents this
2198 warning.) @code{case} labels outside the enumeration range also
2199 provoke warnings when this option is used.
2201 @item -Wswitch-default
2202 @opindex Wswitch-switch
2203 Warn whenever a @code{switch} statement does not have a @code{default}
2207 @opindex Wswitch-enum
2208 Warn whenever a @code{switch} statement has an index of enumeral type
2209 and lacks a @code{case} for one or more of the named codes of that
2210 enumeration. @code{case} labels outside the enumeration range also
2211 provoke warnings when this option is used.
2215 Warn if any trigraphs are encountered that might change the meaning of
2216 the program (trigraphs within comments are not warned about).
2218 @item -Wunused-function
2219 @opindex Wunused-function
2220 Warn whenever a static function is declared but not defined or a
2221 non\-inline static function is unused.
2223 @item -Wunused-label
2224 @opindex Wunused-label
2225 Warn whenever a label is declared but not used.
2227 To suppress this warning use the @samp{unused} attribute
2228 (@pxref{Variable Attributes}).
2230 @item -Wunused-parameter
2231 @opindex Wunused-parameter
2232 Warn whenever a function parameter is unused aside from its declaration.
2234 To suppress this warning use the @samp{unused} attribute
2235 (@pxref{Variable Attributes}).
2237 @item -Wunused-variable
2238 @opindex Wunused-variable
2239 Warn whenever a local variable or non-constant static variable is unused
2240 aside from its declaration
2242 To suppress this warning use the @samp{unused} attribute
2243 (@pxref{Variable Attributes}).
2245 @item -Wunused-value
2246 @opindex Wunused-value
2247 Warn whenever a statement computes a result that is explicitly not used.
2249 To suppress this warning cast the expression to @samp{void}.
2253 All the above @option{-Wunused} options combined.
2255 In order to get a warning about an unused function parameter, you must
2256 either specify @samp{-W -Wunused} or separately specify
2257 @option{-Wunused-parameter}.
2259 @item -Wuninitialized
2260 @opindex Wuninitialized
2261 Warn if an automatic variable is used without first being initialized or
2262 if a variable may be clobbered by a @code{setjmp} call.
2264 These warnings are possible only in optimizing compilation,
2265 because they require data flow information that is computed only
2266 when optimizing. If you don't specify @option{-O}, you simply won't
2269 These warnings occur only for variables that are candidates for
2270 register allocation. Therefore, they do not occur for a variable that
2271 is declared @code{volatile}, or whose address is taken, or whose size
2272 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2273 structures, unions or arrays, even when they are in registers.
2275 Note that there may be no warning about a variable that is used only
2276 to compute a value that itself is never used, because such
2277 computations may be deleted by data flow analysis before the warnings
2280 These warnings are made optional because GCC is not smart
2281 enough to see all the reasons why the code might be correct
2282 despite appearing to have an error. Here is one example of how
2303 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2304 always initialized, but GCC doesn't know this. Here is
2305 another common case:
2310 if (change_y) save_y = y, y = new_y;
2312 if (change_y) y = save_y;
2317 This has no bug because @code{save_y} is used only if it is set.
2319 @cindex @code{longjmp} warnings
2320 This option also warns when a non-volatile automatic variable might be
2321 changed by a call to @code{longjmp}. These warnings as well are possible
2322 only in optimizing compilation.
2324 The compiler sees only the calls to @code{setjmp}. It cannot know
2325 where @code{longjmp} will be called; in fact, a signal handler could
2326 call it at any point in the code. As a result, you may get a warning
2327 even when there is in fact no problem because @code{longjmp} cannot
2328 in fact be called at the place which would cause a problem.
2330 Some spurious warnings can be avoided if you declare all the functions
2331 you use that never return as @code{noreturn}. @xref{Function
2334 @item -Wunknown-pragmas
2335 @opindex Wunknown-pragmas
2336 @cindex warning for unknown pragmas
2337 @cindex unknown pragmas, warning
2338 @cindex pragmas, warning of unknown
2339 Warn when a #pragma directive is encountered which is not understood by
2340 GCC@. If this command line option is used, warnings will even be issued
2341 for unknown pragmas in system header files. This is not the case if
2342 the warnings were only enabled by the @option{-Wall} command line option.
2344 @item -Wstrict-aliasing
2345 @opindex Wstrict-aliasing
2346 This option is only active when @option{-fstrict-aliasing} is active.
2347 It warns about code which might break the strict aliasing rules that the
2348 compiler is using for optimization. The warning does not catch all
2349 cases, but does attempt to catch the more common pitfalls. It is
2350 included in @option{-Wall}.
2354 All of the above @samp{-W} options combined. This enables all the
2355 warnings about constructions that some users consider questionable, and
2356 that are easy to avoid (or modify to prevent the warning), even in
2357 conjunction with macros. This also enables some language-specific
2358 warnings described in @ref{C++ Dialect Options} and
2359 @ref{Objective-C Dialect Options}.
2362 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2363 Some of them warn about constructions that users generally do not
2364 consider questionable, but which occasionally you might wish to check
2365 for; others warn about constructions that are necessary or hard to avoid
2366 in some cases, and there is no simple way to modify the code to suppress
2372 Print extra warning messages for these events:
2376 A function can return either with or without a value. (Falling
2377 off the end of the function body is considered returning without
2378 a value.) For example, this function would evoke such a
2392 An expression-statement or the left-hand side of a comma expression
2393 contains no side effects.
2394 To suppress the warning, cast the unused expression to void.
2395 For example, an expression such as @samp{x[i,j]} will cause a warning,
2396 but @samp{x[(void)i,j]} will not.
2399 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2402 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2403 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2404 that of ordinary mathematical notation.
2407 Storage-class specifiers like @code{static} are not the first things in
2408 a declaration. According to the C Standard, this usage is obsolescent.
2411 The return type of a function has a type qualifier such as @code{const}.
2412 Such a type qualifier has no effect, since the value returned by a
2413 function is not an lvalue. (But don't warn about the GNU extension of
2414 @code{volatile void} return types. That extension will be warned about
2415 if @option{-pedantic} is specified.)
2418 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2422 A comparison between signed and unsigned values could produce an
2423 incorrect result when the signed value is converted to unsigned.
2424 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2427 An aggregate has a partly bracketed initializer.
2428 For example, the following code would evoke such a warning,
2429 because braces are missing around the initializer for @code{x.h}:
2432 struct s @{ int f, g; @};
2433 struct t @{ struct s h; int i; @};
2434 struct t x = @{ 1, 2, 3 @};
2438 An aggregate has an initializer which does not initialize all members.
2439 For example, the following code would cause such a warning, because
2440 @code{x.h} would be implicitly initialized to zero:
2443 struct s @{ int f, g, h; @};
2444 struct s x = @{ 3, 4 @};
2448 @item -Wno-div-by-zero
2449 @opindex Wno-div-by-zero
2450 @opindex Wdiv-by-zero
2451 Do not warn about compile-time integer division by zero. Floating point
2452 division by zero is not warned about, as it can be a legitimate way of
2453 obtaining infinities and NaNs.
2455 @item -Wsystem-headers
2456 @opindex Wsystem-headers
2457 @cindex warnings from system headers
2458 @cindex system headers, warnings from
2459 Print warning messages for constructs found in system header files.
2460 Warnings from system headers are normally suppressed, on the assumption
2461 that they usually do not indicate real problems and would only make the
2462 compiler output harder to read. Using this command line option tells
2463 GCC to emit warnings from system headers as if they occurred in user
2464 code. However, note that using @option{-Wall} in conjunction with this
2465 option will @emph{not} warn about unknown pragmas in system
2466 headers---for that, @option{-Wunknown-pragmas} must also be used.
2469 @opindex Wfloat-equal
2470 Warn if floating point values are used in equality comparisons.
2472 The idea behind this is that sometimes it is convenient (for the
2473 programmer) to consider floating-point values as approximations to
2474 infinitely precise real numbers. If you are doing this, then you need
2475 to compute (by analysing the code, or in some other way) the maximum or
2476 likely maximum error that the computation introduces, and allow for it
2477 when performing comparisons (and when producing output, but that's a
2478 different problem). In particular, instead of testing for equality, you
2479 would check to see whether the two values have ranges that overlap; and
2480 this is done with the relational operators, so equality comparisons are
2483 @item -Wtraditional @r{(C only)}
2484 @opindex Wtraditional
2485 Warn about certain constructs that behave differently in traditional and
2486 ISO C@. Also warn about ISO C constructs that have no traditional C
2487 equivalent, and/or problematic constructs which should be avoided.
2491 Macro parameters that appear within string literals in the macro body.
2492 In traditional C macro replacement takes place within string literals,
2493 but does not in ISO C@.
2496 In traditional C, some preprocessor directives did not exist.
2497 Traditional preprocessors would only consider a line to be a directive
2498 if the @samp{#} appeared in column 1 on the line. Therefore
2499 @option{-Wtraditional} warns about directives that traditional C
2500 understands but would ignore because the @samp{#} does not appear as the
2501 first character on the line. It also suggests you hide directives like
2502 @samp{#pragma} not understood by traditional C by indenting them. Some
2503 traditional implementations would not recognize @samp{#elif}, so it
2504 suggests avoiding it altogether.
2507 A function-like macro that appears without arguments.
2510 The unary plus operator.
2513 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2514 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2515 constants.) Note, these suffixes appear in macros defined in the system
2516 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2517 Use of these macros in user code might normally lead to spurious
2518 warnings, however gcc's integrated preprocessor has enough context to
2519 avoid warning in these cases.
2522 A function declared external in one block and then used after the end of
2526 A @code{switch} statement has an operand of type @code{long}.
2529 A non-@code{static} function declaration follows a @code{static} one.
2530 This construct is not accepted by some traditional C compilers.
2533 The ISO type of an integer constant has a different width or
2534 signedness from its traditional type. This warning is only issued if
2535 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2536 typically represent bit patterns, are not warned about.
2539 Usage of ISO string concatenation is detected.
2542 Initialization of automatic aggregates.
2545 Identifier conflicts with labels. Traditional C lacks a separate
2546 namespace for labels.
2549 Initialization of unions. If the initializer is zero, the warning is
2550 omitted. This is done under the assumption that the zero initializer in
2551 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2552 initializer warnings and relies on default initialization to zero in the
2556 Conversions by prototypes between fixed/floating point values and vice
2557 versa. The absence of these prototypes when compiling with traditional
2558 C would cause serious problems. This is a subset of the possible
2559 conversion warnings, for the full set use @option{-Wconversion}.
2562 Use of ISO C style function definitions. This warning intentionally is
2563 @emph{not} issued for prototype declarations or variadic functions
2564 because these ISO C features will appear in your code when using
2565 libiberty's traditional C compatibility macros, @code{PARAMS} and
2566 @code{VPARAMS}. This warning is also bypassed for nested functions
2567 because that feature is already a gcc extension and thus not relevant to
2568 traditional C compatibility.
2573 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2575 @item -Wendif-labels
2576 @opindex Wendif-labels
2577 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2581 Warn whenever a local variable shadows another local variable, parameter or
2582 global variable or whenever a built-in function is shadowed.
2584 @item -Wlarger-than-@var{len}
2585 @opindex Wlarger-than
2586 Warn whenever an object of larger than @var{len} bytes is defined.
2588 @item -Wpointer-arith
2589 @opindex Wpointer-arith
2590 Warn about anything that depends on the ``size of'' a function type or
2591 of @code{void}. GNU C assigns these types a size of 1, for
2592 convenience in calculations with @code{void *} pointers and pointers
2595 @item -Wbad-function-cast @r{(C only)}
2596 @opindex Wbad-function-cast
2597 Warn whenever a function call is cast to a non-matching type.
2598 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2602 Warn whenever a pointer is cast so as to remove a type qualifier from
2603 the target type. For example, warn if a @code{const char *} is cast
2604 to an ordinary @code{char *}.
2607 @opindex Wcast-align
2608 Warn whenever a pointer is cast such that the required alignment of the
2609 target is increased. For example, warn if a @code{char *} is cast to
2610 an @code{int *} on machines where integers can only be accessed at
2611 two- or four-byte boundaries.
2613 @item -Wwrite-strings
2614 @opindex Wwrite-strings
2615 When compiling C, give string constants the type @code{const
2616 char[@var{length}]} so that
2617 copying the address of one into a non-@code{const} @code{char *}
2618 pointer will get a warning; when compiling C++, warn about the
2619 deprecated conversion from string constants to @code{char *}.
2620 These warnings will help you find at
2621 compile time code that can try to write into a string constant, but
2622 only if you have been very careful about using @code{const} in
2623 declarations and prototypes. Otherwise, it will just be a nuisance;
2624 this is why we did not make @option{-Wall} request these warnings.
2627 @opindex Wconversion
2628 Warn if a prototype causes a type conversion that is different from what
2629 would happen to the same argument in the absence of a prototype. This
2630 includes conversions of fixed point to floating and vice versa, and
2631 conversions changing the width or signedness of a fixed point argument
2632 except when the same as the default promotion.
2634 Also, warn if a negative integer constant expression is implicitly
2635 converted to an unsigned type. For example, warn about the assignment
2636 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2637 casts like @code{(unsigned) -1}.
2639 @item -Wsign-compare
2640 @opindex Wsign-compare
2641 @cindex warning for comparison of signed and unsigned values
2642 @cindex comparison of signed and unsigned values, warning
2643 @cindex signed and unsigned values, comparison warning
2644 Warn when a comparison between signed and unsigned values could produce
2645 an incorrect result when the signed value is converted to unsigned.
2646 This warning is also enabled by @option{-W}; to get the other warnings
2647 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2649 @item -Waggregate-return
2650 @opindex Waggregate-return
2651 Warn if any functions that return structures or unions are defined or
2652 called. (In languages where you can return an array, this also elicits
2655 @item -Wstrict-prototypes @r{(C only)}
2656 @opindex Wstrict-prototypes
2657 Warn if a function is declared or defined without specifying the
2658 argument types. (An old-style function definition is permitted without
2659 a warning if preceded by a declaration which specifies the argument
2662 @item -Wmissing-prototypes @r{(C only)}
2663 @opindex Wmissing-prototypes
2664 Warn if a global function is defined without a previous prototype
2665 declaration. This warning is issued even if the definition itself
2666 provides a prototype. The aim is to detect global functions that fail
2667 to be declared in header files.
2669 @item -Wmissing-declarations
2670 @opindex Wmissing-declarations
2671 Warn if a global function is defined without a previous declaration.
2672 Do so even if the definition itself provides a prototype.
2673 Use this option to detect global functions that are not declared in
2676 @item -Wmissing-noreturn
2677 @opindex Wmissing-noreturn
2678 Warn about functions which might be candidates for attribute @code{noreturn}.
2679 Note these are only possible candidates, not absolute ones. Care should
2680 be taken to manually verify functions actually do not ever return before
2681 adding the @code{noreturn} attribute, otherwise subtle code generation
2682 bugs could be introduced. You will not get a warning for @code{main} in
2683 hosted C environments.
2685 @item -Wmissing-format-attribute
2686 @opindex Wmissing-format-attribute
2688 If @option{-Wformat} is enabled, also warn about functions which might be
2689 candidates for @code{format} attributes. Note these are only possible
2690 candidates, not absolute ones. GCC will guess that @code{format}
2691 attributes might be appropriate for any function that calls a function
2692 like @code{vprintf} or @code{vscanf}, but this might not always be the
2693 case, and some functions for which @code{format} attributes are
2694 appropriate may not be detected. This option has no effect unless
2695 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2697 @item -Wno-multichar
2698 @opindex Wno-multichar
2700 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2701 Usually they indicate a typo in the user's code, as they have
2702 implementation-defined values, and should not be used in portable code.
2704 @item -Wno-deprecated-declarations
2705 @opindex Wno-deprecated-declarations
2706 Do not warn about uses of functions, variables, and types marked as
2707 deprecated by using the @code{deprecated} attribute.
2708 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2709 @pxref{Type Attributes}.)
2713 Warn if a structure is given the packed attribute, but the packed
2714 attribute has no effect on the layout or size of the structure.
2715 Such structures may be mis-aligned for little benefit. For
2716 instance, in this code, the variable @code{f.x} in @code{struct bar}
2717 will be misaligned even though @code{struct bar} does not itself
2718 have the packed attribute:
2725 @} __attribute__((packed));
2735 Warn if padding is included in a structure, either to align an element
2736 of the structure or to align the whole structure. Sometimes when this
2737 happens it is possible to rearrange the fields of the structure to
2738 reduce the padding and so make the structure smaller.
2740 @item -Wredundant-decls
2741 @opindex Wredundant-decls
2742 Warn if anything is declared more than once in the same scope, even in
2743 cases where multiple declaration is valid and changes nothing.
2745 @item -Wnested-externs @r{(C only)}
2746 @opindex Wnested-externs
2747 Warn if an @code{extern} declaration is encountered within a function.
2749 @item -Wunreachable-code
2750 @opindex Wunreachable-code
2751 Warn if the compiler detects that code will never be executed.
2753 This option is intended to warn when the compiler detects that at
2754 least a whole line of source code will never be executed, because
2755 some condition is never satisfied or because it is after a
2756 procedure that never returns.
2758 It is possible for this option to produce a warning even though there
2759 are circumstances under which part of the affected line can be executed,
2760 so care should be taken when removing apparently-unreachable code.
2762 For instance, when a function is inlined, a warning may mean that the
2763 line is unreachable in only one inlined copy of the function.
2765 This option is not made part of @option{-Wall} because in a debugging
2766 version of a program there is often substantial code which checks
2767 correct functioning of the program and is, hopefully, unreachable
2768 because the program does work. Another common use of unreachable
2769 code is to provide behavior which is selectable at compile-time.
2773 Warn if a function can not be inlined and it was declared as inline.
2777 @opindex Wno-long-long
2778 Warn if @samp{long long} type is used. This is default. To inhibit
2779 the warning messages, use @option{-Wno-long-long}. Flags
2780 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2781 only when @option{-pedantic} flag is used.
2783 @item -Wdisabled-optimization
2784 @opindex Wdisabled-optimization
2785 Warn if a requested optimization pass is disabled. This warning does
2786 not generally indicate that there is anything wrong with your code; it
2787 merely indicates that GCC's optimizers were unable to handle the code
2788 effectively. Often, the problem is that your code is too big or too
2789 complex; GCC will refuse to optimize programs when the optimization
2790 itself is likely to take inordinate amounts of time.
2794 Make all warnings into errors.
2797 @node Debugging Options
2798 @section Options for Debugging Your Program or GCC
2799 @cindex options, debugging
2800 @cindex debugging information options
2802 GCC has various special options that are used for debugging
2803 either your program or GCC:
2808 Produce debugging information in the operating system's native format
2809 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2812 On most systems that use stabs format, @option{-g} enables use of extra
2813 debugging information that only GDB can use; this extra information
2814 makes debugging work better in GDB but will probably make other debuggers
2816 refuse to read the program. If you want to control for certain whether
2817 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2818 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2819 or @option{-gvms} (see below).
2821 Unlike most other C compilers, GCC allows you to use @option{-g} with
2822 @option{-O}. The shortcuts taken by optimized code may occasionally
2823 produce surprising results: some variables you declared may not exist
2824 at all; flow of control may briefly move where you did not expect it;
2825 some statements may not be executed because they compute constant
2826 results or their values were already at hand; some statements may
2827 execute in different places because they were moved out of loops.
2829 Nevertheless it proves possible to debug optimized output. This makes
2830 it reasonable to use the optimizer for programs that might have bugs.
2832 The following options are useful when GCC is generated with the
2833 capability for more than one debugging format.
2837 Produce debugging information for use by GDB@. This means to use the
2838 most expressive format available (DWARF 2, stabs, or the native format
2839 if neither of those are supported), including GDB extensions if at all
2844 Produce debugging information in stabs format (if that is supported),
2845 without GDB extensions. This is the format used by DBX on most BSD
2846 systems. On MIPS, Alpha and System V Release 4 systems this option
2847 produces stabs debugging output which is not understood by DBX or SDB@.
2848 On System V Release 4 systems this option requires the GNU assembler.
2852 Produce debugging information in stabs format (if that is supported),
2853 using GNU extensions understood only by the GNU debugger (GDB)@. The
2854 use of these extensions is likely to make other debuggers crash or
2855 refuse to read the program.
2859 Produce debugging information in COFF format (if that is supported).
2860 This is the format used by SDB on most System V systems prior to
2865 Produce debugging information in XCOFF format (if that is supported).
2866 This is the format used by the DBX debugger on IBM RS/6000 systems.
2870 Produce debugging information in XCOFF format (if that is supported),
2871 using GNU extensions understood only by the GNU debugger (GDB)@. The
2872 use of these extensions is likely to make other debuggers crash or
2873 refuse to read the program, and may cause assemblers other than the GNU
2874 assembler (GAS) to fail with an error.
2878 Produce debugging information in DWARF version 1 format (if that is
2879 supported). This is the format used by SDB on most System V Release 4
2882 This option is deprecated.
2886 Produce debugging information in DWARF version 1 format (if that is
2887 supported), using GNU extensions understood only by the GNU debugger
2888 (GDB)@. The use of these extensions is likely to make other debuggers
2889 crash or refuse to read the program.
2891 This option is deprecated.
2895 Produce debugging information in DWARF version 2 format (if that is
2896 supported). This is the format used by DBX on IRIX 6.
2900 Produce debugging information in VMS debug format (if that is
2901 supported). This is the format used by DEBUG on VMS systems.
2904 @itemx -ggdb@var{level}
2905 @itemx -gstabs@var{level}
2906 @itemx -gcoff@var{level}
2907 @itemx -gxcoff@var{level}
2908 @itemx -gvms@var{level}
2909 Request debugging information and also use @var{level} to specify how
2910 much information. The default level is 2.
2912 Level 1 produces minimal information, enough for making backtraces in
2913 parts of the program that you don't plan to debug. This includes
2914 descriptions of functions and external variables, but no information
2915 about local variables and no line numbers.
2917 Level 3 includes extra information, such as all the macro definitions
2918 present in the program. Some debuggers support macro expansion when
2919 you use @option{-g3}.
2921 Note that in order to avoid confusion between DWARF1 debug level 2,
2922 and DWARF2, neither @option{-gdwarf} nor @option{-gdwarf-2} accept
2923 a concatenated debug level. Instead use an additional @option{-g@var{level}}
2924 option to change the debug level for DWARF1 or DWARF2.
2929 Generate extra code to write profile information suitable for the
2930 analysis program @code{prof}. You must use this option when compiling
2931 the source files you want data about, and you must also use it when
2934 @cindex @code{gprof}
2937 Generate extra code to write profile information suitable for the
2938 analysis program @code{gprof}. You must use this option when compiling
2939 the source files you want data about, and you must also use it when
2944 Makes the compiler print out each function name as it is compiled, and
2945 print some statistics about each pass when it finishes.
2948 @opindex ftime-report
2949 Makes the compiler print some statistics about the time consumed by each
2950 pass when it finishes.
2953 @opindex fmem-report
2954 Makes the compiler print some statistics about permanent memory
2955 allocation when it finishes.
2957 @item -fprofile-arcs
2958 @opindex fprofile-arcs
2959 Instrument @dfn{arcs} during compilation to generate coverage data or
2960 for profile-directed block ordering. During execution the program
2961 records how many times each branch is executed and how many times it is
2962 taken. When the compiled program exits it saves this data to a file
2963 called @file{@var{auxname}.da} for each source file. @var{auxname} is
2964 generated from the name of the output file, if explicitly specified and
2965 it is not the final executable, otherwise it is the basename of the
2966 source file. In both cases any suffix is removed (e.g. @file{foo.da}
2967 for input file @file{dir/foo.c}, or @file{dir/foo.da} for output file
2968 specified as @option{-o dir/foo.o}).
2970 For profile-directed block ordering, compile the program with
2971 @option{-fprofile-arcs} plus optimization and code generation options,
2972 generate the arc profile information by running the program on a
2973 selected workload, and then compile the program again with the same
2974 optimization and code generation options plus
2975 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2976 Control Optimization}).
2978 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2979 when it is used with the @option{-ftest-coverage} option.
2981 With @option{-fprofile-arcs}, for each function of your program GCC
2982 creates a program flow graph, then finds a spanning tree for the graph.
2983 Only arcs that are not on the spanning tree have to be instrumented: the
2984 compiler adds code to count the number of times that these arcs are
2985 executed. When an arc is the only exit or only entrance to a block, the
2986 instrumentation code can be added to the block; otherwise, a new basic
2987 block must be created to hold the instrumentation code.
2990 @item -ftest-coverage
2991 @opindex ftest-coverage
2992 Create data files for the @code{gcov} code-coverage utility
2993 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}). See
2994 @option{-fprofile-arcs} option above for a description of @var{auxname}.
2997 @item @var{auxname}.bb
2998 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2999 associate basic block execution counts with line numbers.
3001 @item @var{auxname}.bbg
3002 A list of all arcs in the program flow graph. This allows @code{gcov}
3003 to reconstruct the program flow graph, so that it can compute all basic
3004 block and arc execution counts from the information in the
3005 @file{@var{auxname}.da} file.
3008 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
3009 option adds instrumentation to the program, which then writes
3010 execution counts to another data file:
3013 @item @var{auxname}.da
3014 Runtime arc execution counts, used in conjunction with the arc
3015 information in the file @file{@var{auxname}.bbg}.
3018 Coverage data will map better to the source files if
3019 @option{-ftest-coverage} is used without optimization.
3021 @item -d@var{letters}
3023 Says to make debugging dumps during compilation at times specified by
3024 @var{letters}. This is used for debugging the compiler. The file names
3025 for most of the dumps are made by appending a pass number and a word to
3026 the @var{dumpname}. @var{dumpname} is generated from the name of the
3027 output file, if explicitly specified and it is not an executable,
3028 otherwise it is the basename of the source file. In both cases any
3029 suffix is removed (e.g. @file{foo.00.rtl} or @file{foo.01.sibling}).
3030 Here are the possible letters for use in @var{letters}, and their
3036 Annotate the assembler output with miscellaneous debugging information.
3039 Dump after computing branch probabilities, to @file{@var{file}.15.bp}.
3042 Dump after block reordering, to @file{@var{file}.30.bbro}.
3045 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
3048 Dump after the first if conversion, to the file @file{@var{file}.18.ce}.
3051 Dump after delayed branch scheduling, to @file{@var{file}.32.dbr}.
3054 Dump all macro definitions, at the end of preprocessing, in addition to
3058 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
3059 @file{@var{file}.07.ussa}.
3062 Dump after the second if conversion, to @file{@var{file}.27.ce2}.
3065 Dump after life analysis, to @file{@var{file}.16.life}.
3068 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.10.addressof}.
3071 Dump after global register allocation, to @file{@var{file}.22.greg}.
3074 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
3077 Dump after reg-to-stack conversion, to @file{@var{file}.29.stack}.
3080 Dump after post-reload optimizations, to @file{@var{file}.23.postreload}.
3083 Dump after GCSE, to @file{@var{file}.11.gcse}.
3086 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
3089 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
3092 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
3095 Dump after local register allocation, to @file{@var{file}.21.lreg}.
3098 Dump after loop optimization, to @file{@var{file}.12.loop}.
3101 Dump after performing the machine dependent reorganisation pass, to
3102 @file{@var{file}.31.mach}.
3105 Dump after register renumbering, to @file{@var{file}.26.rnreg}.
3108 Dump after the register move pass, to @file{@var{file}.19.regmove}.
3111 Dump after RTL generation, to @file{@var{file}.00.rtl}.
3114 Dump after the second scheduling pass, to @file{@var{file}.28.sched2}.
3117 Dump after CSE (including the jump optimization that sometimes follows
3118 CSE), to @file{@var{file}.09.cse}.
3121 Dump after the first scheduling pass, to @file{@var{file}.20.sched}.
3124 Dump after the second CSE pass (including the jump optimization that
3125 sometimes follows CSE), to @file{@var{file}.13.cse2}.
3128 Dump after null pointer ellimination pass ti @file{@var{file}.08.null}.
3131 Dump after the second flow pass, to @file{@var{file}.24.flow2}.
3134 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
3137 Dump after the peephole pass, to @file{@var{file}.25.peephole2}.
3140 Produce all the dumps listed above.
3143 Print statistics on memory usage, at the end of the run, to
3147 Annotate the assembler output with a comment indicating which
3148 pattern and alternative was used. The length of each instruction is
3152 Dump the RTL in the assembler output as a comment before each instruction.
3153 Also turns on @option{-dp} annotation.
3156 For each of the other indicated dump files (except for
3157 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
3158 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3161 Just generate RTL for a function instead of compiling it. Usually used
3165 Dump debugging information during parsing, to standard error.
3168 @item -fdump-unnumbered
3169 @opindex fdump-unnumbered
3170 When doing debugging dumps (see @option{-d} option above), suppress instruction
3171 numbers and line number note output. This makes it more feasible to
3172 use diff on debugging dumps for compiler invocations with different
3173 options, in particular with and without @option{-g}.
3175 @item -fdump-translation-unit @r{(C and C++ only)}
3176 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3177 @opindex fdump-translation-unit
3178 Dump a representation of the tree structure for the entire translation
3179 unit to a file. The file name is made by appending @file{.tu} to the
3180 source file name. If the @samp{-@var{options}} form is used, @var{options}
3181 controls the details of the dump as described for the
3182 @option{-fdump-tree} options.
3184 @item -fdump-class-hierarchy @r{(C++ only)}
3185 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3186 @opindex fdump-class-hierarchy
3187 Dump a representation of each class's hierarchy and virtual function
3188 table layout to a file. The file name is made by appending @file{.class}
3189 to the source file name. If the @samp{-@var{options}} form is used,
3190 @var{options} controls the details of the dump as described for the
3191 @option{-fdump-tree} options.
3193 @item -fdump-tree-@var{switch} @r{(C++ only)}
3194 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3196 Control the dumping at various stages of processing the intermediate
3197 language tree to a file. The file name is generated by appending a switch
3198 specific suffix to the source file name. If the @samp{-@var{options}}
3199 form is used, @var{options} is a list of @samp{-} separated options that
3200 control the details of the dump. Not all options are applicable to all
3201 dumps, those which are not meaningful will be ignored. The following
3202 options are available
3206 Print the address of each node. Usually this is not meaningful as it
3207 changes according to the environment and source file. Its primary use
3208 is for tying up a dump file with a debug environment.
3210 Inhibit dumping of members of a scope or body of a function merely
3211 because that scope has been reached. Only dump such items when they
3212 are directly reachable by some other path.
3214 Turn on all options.
3217 The following tree dumps are possible:
3220 Dump before any tree based optimization, to @file{@var{file}.original}.
3222 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3224 Dump after function inlining, to @file{@var{file}.inlined}.
3229 Store the usual ``temporary'' intermediate files permanently; place them
3230 in the current directory and name them based on the source file. Thus,
3231 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3232 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3233 preprocessed @file{foo.i} output file even though the compiler now
3234 normally uses an integrated preprocessor.
3238 Report the CPU time taken by each subprocess in the compilation
3239 sequence. For C source files, this is the compiler proper and assembler
3240 (plus the linker if linking is done). The output looks like this:
3247 The first number on each line is the ``user time,'' that is time spent
3248 executing the program itself. The second number is ``system time,''
3249 time spent executing operating system routines on behalf of the program.
3250 Both numbers are in seconds.
3252 @item -print-file-name=@var{library}
3253 @opindex print-file-name
3254 Print the full absolute name of the library file @var{library} that
3255 would be used when linking---and don't do anything else. With this
3256 option, GCC does not compile or link anything; it just prints the
3259 @item -print-multi-directory
3260 @opindex print-multi-directory
3261 Print the directory name corresponding to the multilib selected by any
3262 other switches present in the command line. This directory is supposed
3263 to exist in @env{GCC_EXEC_PREFIX}.
3265 @item -print-multi-lib
3266 @opindex print-multi-lib
3267 Print the mapping from multilib directory names to compiler switches
3268 that enable them. The directory name is separated from the switches by
3269 @samp{;}, and each switch starts with an @samp{@@} instead of the
3270 @samp{-}, without spaces between multiple switches. This is supposed to
3271 ease shell-processing.
3273 @item -print-prog-name=@var{program}
3274 @opindex print-prog-name
3275 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3277 @item -print-libgcc-file-name
3278 @opindex print-libgcc-file-name
3279 Same as @option{-print-file-name=libgcc.a}.
3281 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3282 but you do want to link with @file{libgcc.a}. You can do
3285 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3288 @item -print-search-dirs
3289 @opindex print-search-dirs
3290 Print the name of the configured installation directory and a list of
3291 program and library directories gcc will search---and don't do anything else.
3293 This is useful when gcc prints the error message
3294 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3295 To resolve this you either need to put @file{cpp0} and the other compiler
3296 components where gcc expects to find them, or you can set the environment
3297 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3298 Don't forget the trailing '/'.
3299 @xref{Environment Variables}.
3302 @opindex dumpmachine
3303 Print the compiler's target machine (for example,
3304 @samp{i686-pc-linux-gnu})---and don't do anything else.
3307 @opindex dumpversion
3308 Print the compiler version (for example, @samp{3.0})---and don't do
3313 Print the compiler's built-in specs---and don't do anything else. (This
3314 is used when GCC itself is being built.) @xref{Spec Files}.
3317 @node Optimize Options
3318 @section Options That Control Optimization
3319 @cindex optimize options
3320 @cindex options, optimization
3322 These options control various sorts of optimizations:
3329 Optimize. Optimizing compilation takes somewhat more time, and a lot
3330 more memory for a large function.
3332 Without @option{-O}, the compiler's goal is to reduce the cost of
3333 compilation and to make debugging produce the expected results.
3334 Statements are independent: if you stop the program with a breakpoint
3335 between statements, you can then assign a new value to any variable or
3336 change the program counter to any other statement in the function and
3337 get exactly the results you would expect from the source code.
3339 With @option{-O}, the compiler tries to reduce code size and execution
3340 time, without performing any optimizations that take a great deal of
3345 Optimize even more. GCC performs nearly all supported optimizations
3346 that do not involve a space-speed tradeoff. The compiler does not
3347 perform loop unrolling or function inlining when you specify @option{-O2}.
3348 As compared to @option{-O}, this option increases both compilation time
3349 and the performance of the generated code.
3351 @option{-O2} turns on all optional optimizations except for loop
3352 unrolling, function inlining, and register renaming. It also turns on
3353 the @option{-fforce-mem} and @option{-fstrict-aliasing} option on all
3354 machines and frame pointer elimination on machines where doing so does
3355 not interfere with debugging.
3357 Please note the warning under @option{-fgcse} about
3358 invoking @option{-O2} on programs that use computed gotos.
3362 Optimize yet more. @option{-O3} turns on all optimizations specified by
3363 @option{-O2} and also turns on the @option{-finline-functions} and
3364 @option{-frename-registers} options.
3372 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3373 do not typically increase code size. It also performs further
3374 optimizations designed to reduce code size.
3376 If you use multiple @option{-O} options, with or without level numbers,
3377 the last such option is the one that is effective.
3380 Options of the form @option{-f@var{flag}} specify machine-independent
3381 flags. Most flags have both positive and negative forms; the negative
3382 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3383 only one of the forms is listed---the one which is not the default.
3384 You can figure out the other form by either removing @samp{no-} or
3389 @opindex ffloat-store
3390 Do not store floating point variables in registers, and inhibit other
3391 options that might change whether a floating point value is taken from a
3394 @cindex floating point precision
3395 This option prevents undesirable excess precision on machines such as
3396 the 68000 where the floating registers (of the 68881) keep more
3397 precision than a @code{double} is supposed to have. Similarly for the
3398 x86 architecture. For most programs, the excess precision does only
3399 good, but a few programs rely on the precise definition of IEEE floating
3400 point. Use @option{-ffloat-store} for such programs, after modifying
3401 them to store all pertinent intermediate computations into variables.
3403 @item -fno-default-inline
3404 @opindex fno-default-inline
3405 Do not make member functions inline by default merely because they are
3406 defined inside the class scope (C++ only). Otherwise, when you specify
3407 @w{@option{-O}}, member functions defined inside class scope are compiled
3408 inline by default; i.e., you don't need to add @samp{inline} in front of
3409 the member function name.
3411 @item -fno-defer-pop
3412 @opindex fno-defer-pop
3413 Always pop the arguments to each function call as soon as that function
3414 returns. For machines which must pop arguments after a function call,
3415 the compiler normally lets arguments accumulate on the stack for several
3416 function calls and pops them all at once.
3420 Force memory operands to be copied into registers before doing
3421 arithmetic on them. This produces better code by making all memory
3422 references potential common subexpressions. When they are not common
3423 subexpressions, instruction combination should eliminate the separate
3424 register-load. The @option{-O2} option turns on this option.
3427 @opindex fforce-addr
3428 Force memory address constants to be copied into registers before
3429 doing arithmetic on them. This may produce better code just as
3430 @option{-fforce-mem} may.
3432 @item -fomit-frame-pointer
3433 @opindex fomit-frame-pointer
3434 Don't keep the frame pointer in a register for functions that
3435 don't need one. This avoids the instructions to save, set up and
3436 restore frame pointers; it also makes an extra register available
3437 in many functions. @strong{It also makes debugging impossible on
3440 On some machines, such as the VAX, this flag has no effect, because
3441 the standard calling sequence automatically handles the frame pointer
3442 and nothing is saved by pretending it doesn't exist. The
3443 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3444 whether a target machine supports this flag. @xref{Registers,,Register
3445 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3447 @item -foptimize-sibling-calls
3448 @opindex foptimize-sibling-calls
3449 Optimize sibling and tail recursive calls.
3453 This option generates traps for signed overflow on addition, subtraction,
3454 multiplication operations.
3458 Don't pay attention to the @code{inline} keyword. Normally this option
3459 is used to keep the compiler from expanding any functions inline.
3460 Note that if you are not optimizing, no functions can be expanded inline.
3462 @item -finline-functions
3463 @opindex finline-functions
3464 Integrate all simple functions into their callers. The compiler
3465 heuristically decides which functions are simple enough to be worth
3466 integrating in this way.
3468 If all calls to a given function are integrated, and the function is
3469 declared @code{static}, then the function is normally not output as
3470 assembler code in its own right.
3472 @item -finline-limit=@var{n}
3473 @opindex finline-limit
3474 By default, gcc limits the size of functions that can be inlined. This flag
3475 allows the control of this limit for functions that are explicitly marked as
3476 inline (ie marked with the inline keyword or defined within the class
3477 definition in c++). @var{n} is the size of functions that can be inlined in
3478 number of pseudo instructions (not counting parameter handling). The default
3479 value of @var{n} is 600.
3480 Increasing this value can result in more inlined code at
3481 the cost of compilation time and memory consumption. Decreasing usually makes
3482 the compilation faster and less code will be inlined (which presumably
3483 means slower programs). This option is particularly useful for programs that
3484 use inlining heavily such as those based on recursive templates with C++.
3486 @emph{Note:} pseudo instruction represents, in this particular context, an
3487 abstract measurement of function's size. In no way, it represents a count
3488 of assembly instructions and as such its exact meaning might change from one
3489 release to an another.
3491 @item -fkeep-inline-functions
3492 @opindex fkeep-inline-functions
3493 Even if all calls to a given function are integrated, and the function
3494 is declared @code{static}, nevertheless output a separate run-time
3495 callable version of the function. This switch does not affect
3496 @code{extern inline} functions.
3498 @item -fkeep-static-consts
3499 @opindex fkeep-static-consts
3500 Emit variables declared @code{static const} when optimization isn't turned
3501 on, even if the variables aren't referenced.
3503 GCC enables this option by default. If you want to force the compiler to
3504 check if the variable was referenced, regardless of whether or not
3505 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3507 @item -fmerge-constants
3508 Attempt to merge identical constants (string constants and floating point
3509 constants) accross compilation units.
3511 This option is default for optimized compilation if assembler and linker
3512 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3514 @item -fmerge-all-constants
3515 Attempt to merge identical constants and identical variables.
3517 This option implies @option{-fmerge-constants}. In addition to
3518 @option{-fmerge-constants} this considers e.g. even constant initialized
3519 arrays or initialized constant variables with integral or floating point
3520 types. Languages like C or C++ require each non-automatic variable to
3521 have distinct location, so using this option will result in non-conforming
3526 Use a graph coloring register allocator. Currently this option is meant
3527 for testing, so we are interested to hear about miscompilations with
3530 @item -fno-function-cse
3531 @opindex fno-function-cse
3532 Do not put function addresses in registers; make each instruction that
3533 calls a constant function contain the function's address explicitly.
3535 This option results in less efficient code, but some strange hacks
3536 that alter the assembler output may be confused by the optimizations
3537 performed when this option is not used.
3541 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
3542 @option{-fno-trapping-math}, @option{-ffinite-math-only} and @*
3543 @option{-fno-signaling-nans}.
3545 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3547 This option should never be turned on by any @option{-O} option since
3548 it can result in incorrect output for programs which depend on
3549 an exact implementation of IEEE or ISO rules/specifications for
3552 @item -fno-math-errno
3553 @opindex fno-math-errno
3554 Do not set ERRNO after calling math functions that are executed
3555 with a single instruction, e.g., sqrt. A program that relies on
3556 IEEE exceptions for math error handling may want to use this flag
3557 for speed while maintaining IEEE arithmetic compatibility.
3559 This option should never be turned on by any @option{-O} option since
3560 it can result in incorrect output for programs which depend on
3561 an exact implementation of IEEE or ISO rules/specifications for
3564 The default is @option{-fmath-errno}.
3566 @item -funsafe-math-optimizations
3567 @opindex funsafe-math-optimizations
3568 Allow optimizations for floating-point arithmetic that (a) assume
3569 that arguments and results are valid and (b) may violate IEEE or
3570 ANSI standards. When used at link-time, it may include libraries
3571 or startup files that change the default FPU control word or other
3572 similar optimizations.
3574 This option should never be turned on by any @option{-O} option since
3575 it can result in incorrect output for programs which depend on
3576 an exact implementation of IEEE or ISO rules/specifications for
3579 The default is @option{-fno-unsafe-math-optimizations}.
3581 @item -ffinite-math-only
3582 @opindex ffinite-math-only
3583 Allow optimizations for floating-point arithmetic that assume
3584 that arguments and results are not NaNs or +-Infs.
3586 This option should never be turned on by any @option{-O} option since
3587 it can result in incorrect output for programs which depend on
3588 an exact implementation of IEEE or ISO rules/specifications.
3590 The default is @option{-fno-finite-math-only}.
3592 @item -fno-trapping-math
3593 @opindex fno-trapping-math
3594 Compile code assuming that floating-point operations cannot generate
3595 user-visible traps. These traps include division by zero, overflow,
3596 underflow, inexact result and invalid operation. This option implies
3597 @option{-fno-signaling-nans}. Setting this option may allow faster
3598 code if one relies on ``non-stop'' IEEE arithmetic, for example.
3600 This option should never be turned on by any @option{-O} option since
3601 it can result in incorrect output for programs which depend on
3602 an exact implementation of IEEE or ISO rules/specifications for
3605 The default is @option{-ftrapping-math}.
3607 @item -fsignaling-nans
3608 @opindex fsignaling-nans
3609 Compile code assuming that IEEE signaling NaNs may generate user-visible
3610 traps during floating-point operations. Setting this option disables
3611 optimizations that may change the number of exceptions visible with
3612 signaling NaNs. This option implies @option{-ftrapping-math}.
3614 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
3617 The default is @option{-fno-signaling-nans}.
3619 This option is experimental and does not currently guarantee to
3620 disable all GCC optimizations that affect signaling NaN behavior.
3622 @item -fno-zero-initialized-in-bss
3623 @opindex fno-zero-initialized-in-bss
3624 If the target supports a BSS section, GCC by default puts variables that
3625 are initialized to zero into BSS@. This can save space in the resulting
3628 This option turns off this behavior because some programs explicitly
3629 rely on variables going to the data section. E.g., so that the
3630 resulting executable can find the beginning of that section and/or make
3631 assumptions based on that.
3633 The default is @option{-fzero-initialized-in-bss}.
3636 The following options control specific optimizations. The @option{-O2}
3637 option turns on all of these optimizations except @option{-funroll-loops}
3638 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3639 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3640 but specific machines may handle it differently.
3642 You can use the following flags in the rare cases when ``fine-tuning''
3643 of optimizations to be performed is desired.
3645 Not all of the optimizations performed by GCC have @option{-f} options
3649 @item -fstrength-reduce
3650 @opindex fstrength-reduce
3651 Perform the optimizations of loop strength reduction and
3652 elimination of iteration variables.
3654 @item -fthread-jumps
3655 @opindex fthread-jumps
3656 Perform optimizations where we check to see if a jump branches to a
3657 location where another comparison subsumed by the first is found. If
3658 so, the first branch is redirected to either the destination of the
3659 second branch or a point immediately following it, depending on whether
3660 the condition is known to be true or false.
3662 @item -fcse-follow-jumps
3663 @opindex fcse-follow-jumps
3664 In common subexpression elimination, scan through jump instructions
3665 when the target of the jump is not reached by any other path. For
3666 example, when CSE encounters an @code{if} statement with an
3667 @code{else} clause, CSE will follow the jump when the condition
3670 @item -fcse-skip-blocks
3671 @opindex fcse-skip-blocks
3672 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3673 follow jumps which conditionally skip over blocks. When CSE
3674 encounters a simple @code{if} statement with no else clause,
3675 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3676 body of the @code{if}.
3678 @item -frerun-cse-after-loop
3679 @opindex frerun-cse-after-loop
3680 Re-run common subexpression elimination after loop optimizations has been
3683 @item -frerun-loop-opt
3684 @opindex frerun-loop-opt
3685 Run the loop optimizer twice.
3689 Perform a global common subexpression elimination pass.
3690 This pass also performs global constant and copy propagation.
3692 @emph{Note:} When compiling a program using computed gotos, a GCC
3693 extension, you may get better runtime performance if you disable
3694 the global common subexpression elmination pass by adding
3695 @option{-fno-gcse} to the command line.
3699 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3700 attempt to move loads which are only killed by stores into themselves. This
3701 allows a loop containing a load/store sequence to be changed to a load outside
3702 the loop, and a copy/store within the loop.
3706 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3707 subexpression elimination. This pass will attempt to move stores out of loops.
3708 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3709 can be changed to a load before the loop and a store after the loop.
3711 @item -floop-optimize
3712 @opindex floop-optimize
3713 Perform loop optimizations: move constant expressions out of loops, simplify
3714 exit test conditions and optionally do strength-reduction and loop unrolling as
3717 @item -fcrossjumping
3718 @opindex crossjumping
3719 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
3720 resulting code may or may not perform better than without cross-jumping.
3722 @item -fif-conversion
3723 @opindex if-conversion
3724 Attempt to transform conditional jumps into branch-less equivalents. This
3725 include use of conditional moves, min, max, set flags and abs instructions, and
3726 some tricks doable by standard arithmetics. The use of conditional execution
3727 on chips where it is available is controlled by @code{if-conversion2}.
3729 @item -fif-conversion2
3730 @opindex if-conversion2
3731 Use conditional execution (where available) to transform conditional jumps into
3732 branch-less equivalents.
3734 @item -fdelete-null-pointer-checks
3735 @opindex fdelete-null-pointer-checks
3736 Use global dataflow analysis to identify and eliminate useless checks
3737 for null pointers. The compiler assumes that dereferencing a null
3738 pointer would have halted the program. If a pointer is checked after
3739 it has already been dereferenced, it cannot be null.
3741 In some environments, this assumption is not true, and programs can
3742 safely dereference null pointers. Use
3743 @option{-fno-delete-null-pointer-checks} to disable this optimization
3744 for programs which depend on that behavior.
3746 @item -fexpensive-optimizations
3747 @opindex fexpensive-optimizations
3748 Perform a number of minor optimizations that are relatively expensive.
3750 @item -foptimize-register-move
3752 @opindex foptimize-register-move
3754 Attempt to reassign register numbers in move instructions and as
3755 operands of other simple instructions in order to maximize the amount of
3756 register tying. This is especially helpful on machines with two-operand
3757 instructions. GCC enables this optimization by default with @option{-O2}
3760 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3763 @item -fdelayed-branch
3764 @opindex fdelayed-branch
3765 If supported for the target machine, attempt to reorder instructions
3766 to exploit instruction slots available after delayed branch
3769 @item -fschedule-insns
3770 @opindex fschedule-insns
3771 If supported for the target machine, attempt to reorder instructions to
3772 eliminate execution stalls due to required data being unavailable. This
3773 helps machines that have slow floating point or memory load instructions
3774 by allowing other instructions to be issued until the result of the load
3775 or floating point instruction is required.
3777 @item -fschedule-insns2
3778 @opindex fschedule-insns2
3779 Similar to @option{-fschedule-insns}, but requests an additional pass of
3780 instruction scheduling after register allocation has been done. This is
3781 especially useful on machines with a relatively small number of
3782 registers and where memory load instructions take more than one cycle.
3784 @item -ffunction-sections
3785 @itemx -fdata-sections
3786 @opindex ffunction-sections
3787 @opindex fdata-sections
3788 Place each function or data item into its own section in the output
3789 file if the target supports arbitrary sections. The name of the
3790 function or the name of the data item determines the section's name
3793 Use these options on systems where the linker can perform optimizations
3794 to improve locality of reference in the instruction space. HPPA
3795 processors running HP-UX and SPARC processors running Solaris 2 have
3796 linkers with such optimizations. Other systems using the ELF object format
3797 as well as AIX may have these optimizations in the future.
3799 Only use these options when there are significant benefits from doing
3800 so. When you specify these options, the assembler and linker will
3801 create larger object and executable files and will also be slower.
3802 You will not be able to use @code{gprof} on all systems if you
3803 specify this option and you may have problems with debugging if
3804 you specify both this option and @option{-g}.
3806 @item -fcaller-saves
3807 @opindex fcaller-saves
3808 Enable values to be allocated in registers that will be clobbered by
3809 function calls, by emitting extra instructions to save and restore the
3810 registers around such calls. Such allocation is done only when it
3811 seems to result in better code than would otherwise be produced.
3813 This option is always enabled by default on certain machines, usually
3814 those which have no call-preserved registers to use instead.
3816 For all machines, optimization level 2 and higher enables this flag by
3821 Perform tail duplication to enlarge superblock size. This transformation
3822 simplifies the control flow of the function allowing other optimizations to do
3825 @item -funroll-loops
3826 @opindex funroll-loops
3827 Unroll loops whose number of iterations can be determined at compile
3828 time or upon entry to the loop. @option{-funroll-loops} implies both
3829 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3830 option makes code larger, and may or may not make it run faster.
3832 @item -funroll-all-loops
3833 @opindex funroll-all-loops
3834 Unroll all loops, even if their number of iterations is uncertain when
3835 the loop is entered. This usually makes programs run more slowly.
3836 @option{-funroll-all-loops} implies the same options as
3837 @option{-funroll-loops},
3839 @item -fprefetch-loop-arrays
3840 @opindex fprefetch-loop-arrays
3841 If supported by the target machine, generate instructions to prefetch
3842 memory to improve the performance of loops that access large arrays.
3844 @item -fmove-all-movables
3845 @opindex fmove-all-movables
3846 Forces all invariant computations in loops to be moved
3849 @item -freduce-all-givs
3850 @opindex freduce-all-givs
3851 Forces all general-induction variables in loops to be
3854 @emph{Note:} When compiling programs written in Fortran,
3855 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3856 by default when you use the optimizer.
3858 These options may generate better or worse code; results are highly
3859 dependent on the structure of loops within the source code.
3861 These two options are intended to be removed someday, once
3862 they have helped determine the efficacy of various
3863 approaches to improving loop optimizations.
3865 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3866 know how use of these options affects
3867 the performance of your production code.
3868 We're very interested in code that runs @emph{slower}
3869 when these options are @emph{enabled}.
3872 @itemx -fno-peephole2
3873 @opindex fno-peephole
3874 @opindex fno-peephole2
3875 Disable any machine-specific peephole optimizations. The difference
3876 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3877 are implemented in the compiler; some targets use one, some use the
3878 other, a few use both.
3880 @item -fbranch-probabilities
3881 @opindex fbranch-probabilities
3882 After running a program compiled with @option{-fprofile-arcs}
3883 (@pxref{Debugging Options,, Options for Debugging Your Program or
3884 @command{gcc}}), you can compile it a second time using
3885 @option{-fbranch-probabilities}, to improve optimizations based on
3886 the number of times each branch was taken. When the program
3887 compiled with @option{-fprofile-arcs} exits it saves arc execution
3888 counts to a file called @file{@var{sourcename}.da} for each source
3889 file The information in this data file is very dependent on the
3890 structure of the generated code, so you must use the same source code
3891 and the same optimization options for both compilations.
3893 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3894 note on the first instruction of each basic block, and a
3895 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3896 These can be used to improve optimization. Currently, they are only
3897 used in one place: in @file{reorg.c}, instead of guessing which path a
3898 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3899 exactly determine which path is taken more often.
3901 @item -fno-guess-branch-probability
3902 @opindex fno-guess-branch-probability
3903 Do not guess branch probabilities using a randomized model.
3905 Sometimes gcc will opt to use a randomized model to guess branch
3906 probabilities, when none are available from either profiling feedback
3907 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3908 different runs of the compiler on the same program may produce different
3911 In a hard real-time system, people don't want different runs of the
3912 compiler to produce code that has different behavior; minimizing
3913 non-determinism is of paramount import. This switch allows users to
3914 reduce non-determinism, possibly at the expense of inferior
3917 @item -freorder-blocks
3918 @opindex freorder-blocks
3919 Reorder basic blocks in the compiled function in order to reduce number of
3920 taken branches and improve code locality.
3922 @item -freorder-functions
3923 @opindex freorder-functions
3924 Reorder basic blocks in the compiled function in order to reduce number of
3925 taken branches and improve code locality. This is implemented by using special
3926 subsections @code{text.hot} for most frequently executed functions and
3927 @code{text.unlikely} for unlikely executed functions. Reordering is done by
3928 the linker so object file format must support named sections and linker must
3929 place them in resonable way.
3931 Also profile feedback must be available in to make this option effective. See
3932 @option{-fprofile-arcs} for details.
3934 @item -fstrict-aliasing
3935 @opindex fstrict-aliasing
3936 Allows the compiler to assume the strictest aliasing rules applicable to
3937 the language being compiled. For C (and C++), this activates
3938 optimizations based on the type of expressions. In particular, an
3939 object of one type is assumed never to reside at the same address as an
3940 object of a different type, unless the types are almost the same. For
3941 example, an @code{unsigned int} can alias an @code{int}, but not a
3942 @code{void*} or a @code{double}. A character type may alias any other
3945 Pay special attention to code like this:
3958 The practice of reading from a different union member than the one most
3959 recently written to (called ``type-punning'') is common. Even with
3960 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3961 is accessed through the union type. So, the code above will work as
3962 expected. However, this code might not:
3973 Every language that wishes to perform language-specific alias analysis
3974 should define a function that computes, given an @code{tree}
3975 node, an alias set for the node. Nodes in different alias sets are not
3976 allowed to alias. For an example, see the C front-end function
3977 @code{c_get_alias_set}.
3979 For all machines, optimization level 2 and higher enables this flag by
3982 @item -falign-functions
3983 @itemx -falign-functions=@var{n}
3984 @opindex falign-functions
3985 Align the start of functions to the next power-of-two greater than
3986 @var{n}, skipping up to @var{n} bytes. For instance,
3987 @option{-falign-functions=32} aligns functions to the next 32-byte
3988 boundary, but @option{-falign-functions=24} would align to the next
3989 32-byte boundary only if this can be done by skipping 23 bytes or less.
3991 @option{-fno-align-functions} and @option{-falign-functions=1} are
3992 equivalent and mean that functions will not be aligned.
3994 Some assemblers only support this flag when @var{n} is a power of two;
3995 in that case, it is rounded up.
3997 If @var{n} is not specified, use a machine-dependent default.
3999 @item -falign-labels
4000 @itemx -falign-labels=@var{n}
4001 @opindex falign-labels
4002 Align all branch targets to a power-of-two boundary, skipping up to
4003 @var{n} bytes like @option{-falign-functions}. This option can easily
4004 make code slower, because it must insert dummy operations for when the
4005 branch target is reached in the usual flow of the code.
4007 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4008 are greater than this value, then their values are used instead.
4010 If @var{n} is not specified, use a machine-dependent default which is
4011 very likely to be @samp{1}, meaning no alignment.
4014 @itemx -falign-loops=@var{n}
4015 @opindex falign-loops
4016 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4017 like @option{-falign-functions}. The hope is that the loop will be
4018 executed many times, which will make up for any execution of the dummy
4021 If @var{n} is not specified, use a machine-dependent default.
4024 @itemx -falign-jumps=@var{n}
4025 @opindex falign-jumps
4026 Align branch targets to a power-of-two boundary, for branch targets
4027 where the targets can only be reached by jumping, skipping up to @var{n}
4028 bytes like @option{-falign-functions}. In this case, no dummy operations
4031 If @var{n} is not specified, use a machine-dependent default.
4035 Perform optimizations in static single assignment form. Each function's
4036 flow graph is translated into SSA form, optimizations are performed, and
4037 the flow graph is translated back from SSA form. Users should not
4038 specify this option, since it is not yet ready for production use.
4042 Perform Sparse Conditional Constant Propagation in SSA form. Requires
4043 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
4047 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
4048 Like @option{-fssa}, this is an experimental feature.
4050 @item -fsingle-precision-constant
4051 @opindex fsingle-precision-constant
4052 Treat floating point constant as single precision constant instead of
4053 implicitly converting it to double precision constant.
4055 @item -frename-registers
4056 @opindex frename-registers
4057 Attempt to avoid false dependencies in scheduled code by making use
4058 of registers left over after register allocation. This optimization
4059 will most benefit processors with lots of registers. It can, however,
4060 make debugging impossible, since variables will no longer stay in
4061 a ``home register''.
4063 @item -fno-cprop-registers
4064 @opindex fno-cprop-registers
4065 After register allocation and post-register allocation instruction splitting,
4066 we perform a copy-propagation pass to try to reduce scheduling dependencies
4067 and occasionally eliminate the copy.
4069 @item --param @var{name}=@var{value}
4071 In some places, GCC uses various constants to control the amount of
4072 optimization that is done. For example, GCC will not inline functions
4073 that contain more that a certain number of instructions. You can
4074 control some of these constants on the command-line using the
4075 @option{--param} option.
4077 In each case, the @var{value} is an integer. The allowable choices for
4078 @var{name} are given in the following table:
4081 @item max-delay-slot-insn-search
4082 The maximum number of instructions to consider when looking for an
4083 instruction to fill a delay slot. If more than this arbitrary number of
4084 instructions is searched, the time savings from filling the delay slot
4085 will be minimal so stop searching. Increasing values mean more
4086 aggressive optimization, making the compile time increase with probably
4087 small improvement in executable run time.
4089 @item max-delay-slot-live-search
4090 When trying to fill delay slots, the maximum number of instructions to
4091 consider when searching for a block with valid live register
4092 information. Increasing this arbitrarily chosen value means more
4093 aggressive optimization, increasing the compile time. This parameter
4094 should be removed when the delay slot code is rewritten to maintain the
4097 @item max-gcse-memory
4098 The approximate maximum amount of memory that will be allocated in
4099 order to perform the global common subexpression elimination
4100 optimization. If more memory than specified is required, the
4101 optimization will not be done.
4103 @item max-gcse-passes
4104 The maximum number of passes of GCSE to run.
4106 @item max-pending-list-length
4107 The maximum number of pending dependencies scheduling will allow
4108 before flushing the current state and starting over. Large functions
4109 with few branches or calls can create excessively large lists which
4110 needlessly consume memory and resources.
4112 @item max-inline-insns
4113 If an function contains more than this many instructions, it
4114 will not be inlined. This option is precisely equivalent to
4115 @option{-finline-limit}.
4117 @item max-unrolled-insns
4118 The maximum number of instructions that a loop should have if that loop
4119 is unrolled, and if the loop is unrolled, it determines how many times
4120 the loop code is unrolled.
4122 @item hot-bb-count-fraction
4123 Select fraction of the maximal count of repetitions of basic block in program
4124 given basic block needs to have to be considered hot.
4126 @item hot-bb-frequency-fraction
4127 Select fraction of the maximal frequency of executions of basic block in
4128 function given basic block needs to have to be considered hot
4130 @item tracer-dynamic-coverage
4131 @itemx tracer-dynamic-coverage-feedback
4133 This value is used to limit superblock formation once given percentage of
4134 executed instructions is covered. This limits unnecesary code size expansion.
4136 The @option{tracer-dynamic-coverage-feedback} is used only when profile
4137 feedback is available. The real profiles (as opposed to statically estimated
4138 ones) are much less balanced allowing the threshold to be larger value.
4140 @item tracer-max-code-growth
4141 Stop tail duplication once code growth has reached given percentage. This is
4142 rather hokey argument, as most of the duplicates will be elliminated later in
4143 cross jumping, so it may be set to much higher values than is the desired code
4146 @item tracer-min-branch-ratio
4148 Stop reverse growth when the reverse probability of best edge is less than this
4149 threshold (in percent).
4151 @item tracer-min-branch-ratio
4152 @itemx tracer-min-branch-ratio-feedback
4154 Stop forward growth if the best edge do have probability lower than this
4157 Similary to @option{tracer-dynamic-coverage} two values are present, one for
4158 compilation for profile feedback and one for compilation without. The value
4159 for compilation with profile feedback needs to be more conservative (higher) in
4160 order to make tracer effective.
4165 @node Preprocessor Options
4166 @section Options Controlling the Preprocessor
4167 @cindex preprocessor options
4168 @cindex options, preprocessor
4170 These options control the C preprocessor, which is run on each C source
4171 file before actual compilation.
4173 If you use the @option{-E} option, nothing is done except preprocessing.
4174 Some of these options make sense only together with @option{-E} because
4175 they cause the preprocessor output to be unsuitable for actual
4179 You can use @option{-Wp,@var{option}} to bypass the compiler driver
4180 and pass @var{option} directly through to the preprocessor. If
4181 @var{option} contains commas, it is split into multiple options at the
4182 commas. However, many options are modified, translated or interpreted
4183 by the compiler driver before being passed to the preprocessor, and
4184 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
4185 interface is undocumented and subject to change, so whenever possible
4186 you should avoid using @option{-Wp} and let the driver handle the
4189 @include cppopts.texi
4191 @node Assembler Options
4192 @section Passing Options to the Assembler
4194 @c prevent bad page break with this line
4195 You can pass options to the assembler.
4198 @item -Wa,@var{option}
4200 Pass @var{option} as an option to the assembler. If @var{option}
4201 contains commas, it is split into multiple options at the commas.
4205 @section Options for Linking
4206 @cindex link options
4207 @cindex options, linking
4209 These options come into play when the compiler links object files into
4210 an executable output file. They are meaningless if the compiler is
4211 not doing a link step.
4215 @item @var{object-file-name}
4216 A file name that does not end in a special recognized suffix is
4217 considered to name an object file or library. (Object files are
4218 distinguished from libraries by the linker according to the file
4219 contents.) If linking is done, these object files are used as input
4228 If any of these options is used, then the linker is not run, and
4229 object file names should not be used as arguments. @xref{Overall
4233 @item -l@var{library}
4234 @itemx -l @var{library}
4236 Search the library named @var{library} when linking. (The second
4237 alternative with the library as a separate argument is only for
4238 POSIX compliance and is not recommended.)
4240 It makes a difference where in the command you write this option; the
4241 linker searches and processes libraries and object files in the order they
4242 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4243 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4244 to functions in @samp{z}, those functions may not be loaded.
4246 The linker searches a standard list of directories for the library,
4247 which is actually a file named @file{lib@var{library}.a}. The linker
4248 then uses this file as if it had been specified precisely by name.
4250 The directories searched include several standard system directories
4251 plus any that you specify with @option{-L}.
4253 Normally the files found this way are library files---archive files
4254 whose members are object files. The linker handles an archive file by
4255 scanning through it for members which define symbols that have so far
4256 been referenced but not defined. But if the file that is found is an
4257 ordinary object file, it is linked in the usual fashion. The only
4258 difference between using an @option{-l} option and specifying a file name
4259 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4260 and searches several directories.
4264 You need this special case of the @option{-l} option in order to
4265 link an Objective-C program.
4268 @opindex nostartfiles
4269 Do not use the standard system startup files when linking.
4270 The standard system libraries are used normally, unless @option{-nostdlib}
4271 or @option{-nodefaultlibs} is used.
4273 @item -nodefaultlibs
4274 @opindex nodefaultlibs
4275 Do not use the standard system libraries when linking.
4276 Only the libraries you specify will be passed to the linker.
4277 The standard startup files are used normally, unless @option{-nostartfiles}
4278 is used. The compiler may generate calls to memcmp, memset, and memcpy
4279 for System V (and ISO C) environments or to bcopy and bzero for
4280 BSD environments. These entries are usually resolved by entries in
4281 libc. These entry points should be supplied through some other
4282 mechanism when this option is specified.
4286 Do not use the standard system startup files or libraries when linking.
4287 No startup files and only the libraries you specify will be passed to
4288 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4289 for System V (and ISO C) environments or to bcopy and bzero for
4290 BSD environments. These entries are usually resolved by entries in
4291 libc. These entry points should be supplied through some other
4292 mechanism when this option is specified.
4294 @cindex @option{-lgcc}, use with @option{-nostdlib}
4295 @cindex @option{-nostdlib} and unresolved references
4296 @cindex unresolved references and @option{-nostdlib}
4297 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4298 @cindex @option{-nodefaultlibs} and unresolved references
4299 @cindex unresolved references and @option{-nodefaultlibs}
4300 One of the standard libraries bypassed by @option{-nostdlib} and
4301 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4302 that GCC uses to overcome shortcomings of particular machines, or special
4303 needs for some languages.
4304 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4305 Collection (GCC) Internals},
4306 for more discussion of @file{libgcc.a}.)
4307 In most cases, you need @file{libgcc.a} even when you want to avoid
4308 other standard libraries. In other words, when you specify @option{-nostdlib}
4309 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4310 This ensures that you have no unresolved references to internal GCC
4311 library subroutines. (For example, @samp{__main}, used to ensure C++
4312 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4313 GNU Compiler Collection (GCC) Internals}.)
4317 Remove all symbol table and relocation information from the executable.
4321 On systems that support dynamic linking, this prevents linking with the shared
4322 libraries. On other systems, this option has no effect.
4326 Produce a shared object which can then be linked with other objects to
4327 form an executable. Not all systems support this option. For predictable
4328 results, you must also specify the same set of options that were used to
4329 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4330 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4331 needs to build supplementary stub code for constructors to work. On
4332 multi-libbed systems, @samp{gcc -shared} must select the correct support
4333 libraries to link against. Failing to supply the correct flags may lead
4334 to subtle defects. Supplying them in cases where they are not necessary
4337 @item -shared-libgcc
4338 @itemx -static-libgcc
4339 @opindex shared-libgcc
4340 @opindex static-libgcc
4341 On systems that provide @file{libgcc} as a shared library, these options
4342 force the use of either the shared or static version respectively.
4343 If no shared version of @file{libgcc} was built when the compiler was
4344 configured, these options have no effect.
4346 There are several situations in which an application should use the
4347 shared @file{libgcc} instead of the static version. The most common
4348 of these is when the application wishes to throw and catch exceptions
4349 across different shared libraries. In that case, each of the libraries
4350 as well as the application itself should use the shared @file{libgcc}.
4352 Therefore, the G++ and GCJ drivers automatically add
4353 @option{-shared-libgcc} whenever you build a shared library or a main
4354 executable, because C++ and Java programs typically use exceptions, so
4355 this is the right thing to do.
4357 If, instead, you use the GCC driver to create shared libraries, you may
4358 find that they will not always be linked with the shared @file{libgcc}.
4359 If GCC finds, at its configuration time, that you have a GNU linker that
4360 does not support option @option{--eh-frame-hdr}, it will link the shared
4361 version of @file{libgcc} into shared libraries by default. Otherwise,
4362 it will take advantage of the linker and optimize away the linking with
4363 the shared version of @file{libgcc}, linking with the static version of
4364 libgcc by default. This allows exceptions to propagate through such
4365 shared libraries, without incurring relocation costs at library load
4368 However, if a library or main executable is supposed to throw or catch
4369 exceptions, you must link it using the G++ or GCJ driver, as appropriate
4370 for the languages used in the program, or using the option
4371 @option{-shared-libgcc}, such that it is linked with the shared
4376 Bind references to global symbols when building a shared object. Warn
4377 about any unresolved references (unless overridden by the link editor
4378 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4381 @item -Xlinker @var{option}
4383 Pass @var{option} as an option to the linker. You can use this to
4384 supply system-specific linker options which GCC does not know how to
4387 If you want to pass an option that takes an argument, you must use
4388 @option{-Xlinker} twice, once for the option and once for the argument.
4389 For example, to pass @option{-assert definitions}, you must write
4390 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4391 @option{-Xlinker "-assert definitions"}, because this passes the entire
4392 string as a single argument, which is not what the linker expects.
4394 @item -Wl,@var{option}
4396 Pass @var{option} as an option to the linker. If @var{option} contains
4397 commas, it is split into multiple options at the commas.
4399 @item -u @var{symbol}
4401 Pretend the symbol @var{symbol} is undefined, to force linking of
4402 library modules to define it. You can use @option{-u} multiple times with
4403 different symbols to force loading of additional library modules.
4406 @node Directory Options
4407 @section Options for Directory Search
4408 @cindex directory options
4409 @cindex options, directory search
4412 These options specify directories to search for header files, for
4413 libraries and for parts of the compiler:
4418 Add the directory @var{dir} to the head of the list of directories to be
4419 searched for header files. This can be used to override a system header
4420 file, substituting your own version, since these directories are
4421 searched before the system header file directories. However, you should
4422 not use this option to add directories that contain vendor-supplied
4423 system header files (use @option{-isystem} for that). If you use more than
4424 one @option{-I} option, the directories are scanned in left-to-right
4425 order; the standard system directories come after.
4427 If a standard system include directory, or a directory specified with
4428 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
4429 option will be ignored. The directory will still be searched but as a
4430 system directory at its normal position in the system include chain.
4431 This is to ensure that GCC's procedure to fix buggy system headers and
4432 the ordering for the include_next directive are not inadvertantly changed.
4433 If you really need to change the search order for system directories,
4434 use the @option{-nostdinc} and/or @option{-isystem} options.
4438 Any directories you specify with @option{-I} options before the @option{-I-}
4439 option are searched only for the case of @samp{#include "@var{file}"};
4440 they are not searched for @samp{#include <@var{file}>}.
4442 If additional directories are specified with @option{-I} options after
4443 the @option{-I-}, these directories are searched for all @samp{#include}
4444 directives. (Ordinarily @emph{all} @option{-I} directories are used
4447 In addition, the @option{-I-} option inhibits the use of the current
4448 directory (where the current input file came from) as the first search
4449 directory for @samp{#include "@var{file}"}. There is no way to
4450 override this effect of @option{-I-}. With @option{-I.} you can specify
4451 searching the directory which was current when the compiler was
4452 invoked. That is not exactly the same as what the preprocessor does
4453 by default, but it is often satisfactory.
4455 @option{-I-} does not inhibit the use of the standard system directories
4456 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4461 Add directory @var{dir} to the list of directories to be searched
4464 @item -B@var{prefix}
4466 This option specifies where to find the executables, libraries,
4467 include files, and data files of the compiler itself.
4469 The compiler driver program runs one or more of the subprograms
4470 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4471 @var{prefix} as a prefix for each program it tries to run, both with and
4472 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4474 For each subprogram to be run, the compiler driver first tries the
4475 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4476 was not specified, the driver tries two standard prefixes, which are
4477 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4478 those results in a file name that is found, the unmodified program
4479 name is searched for using the directories specified in your
4480 @env{PATH} environment variable.
4482 The compiler will check to see if the path provided by the @option{-B}
4483 refers to a directory, and if necessary it will add a directory
4484 separator character at the end of the path.
4486 @option{-B} prefixes that effectively specify directory names also apply
4487 to libraries in the linker, because the compiler translates these
4488 options into @option{-L} options for the linker. They also apply to
4489 includes files in the preprocessor, because the compiler translates these
4490 options into @option{-isystem} options for the preprocessor. In this case,
4491 the compiler appends @samp{include} to the prefix.
4493 The run-time support file @file{libgcc.a} can also be searched for using
4494 the @option{-B} prefix, if needed. If it is not found there, the two
4495 standard prefixes above are tried, and that is all. The file is left
4496 out of the link if it is not found by those means.
4498 Another way to specify a prefix much like the @option{-B} prefix is to use
4499 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4502 As a special kludge, if the path provided by @option{-B} is
4503 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4504 9, then it will be replaced by @file{[dir/]include}. This is to help
4505 with boot-strapping the compiler.
4507 @item -specs=@var{file}
4509 Process @var{file} after the compiler reads in the standard @file{specs}
4510 file, in order to override the defaults that the @file{gcc} driver
4511 program uses when determining what switches to pass to @file{cc1},
4512 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4513 @option{-specs=@var{file}} can be specified on the command line, and they
4514 are processed in order, from left to right.
4520 @section Specifying subprocesses and the switches to pass to them
4522 @command{gcc} is a driver program. It performs its job by invoking a
4523 sequence of other programs to do the work of compiling, assembling and
4524 linking. GCC interprets its command-line parameters and uses these to
4525 deduce which programs it should invoke, and which command-line options
4526 it ought to place on their command lines. This behavior is controlled
4527 by @dfn{spec strings}. In most cases there is one spec string for each
4528 program that GCC can invoke, but a few programs have multiple spec
4529 strings to control their behavior. The spec strings built into GCC can
4530 be overridden by using the @option{-specs=} command-line switch to specify
4533 @dfn{Spec files} are plaintext files that are used to construct spec
4534 strings. They consist of a sequence of directives separated by blank
4535 lines. The type of directive is determined by the first non-whitespace
4536 character on the line and it can be one of the following:
4539 @item %@var{command}
4540 Issues a @var{command} to the spec file processor. The commands that can
4544 @item %include <@var{file}>
4546 Search for @var{file} and insert its text at the current point in the
4549 @item %include_noerr <@var{file}>
4550 @cindex %include_noerr
4551 Just like @samp{%include}, but do not generate an error message if the include
4552 file cannot be found.
4554 @item %rename @var{old_name} @var{new_name}
4556 Rename the spec string @var{old_name} to @var{new_name}.
4560 @item *[@var{spec_name}]:
4561 This tells the compiler to create, override or delete the named spec
4562 string. All lines after this directive up to the next directive or
4563 blank line are considered to be the text for the spec string. If this
4564 results in an empty string then the spec will be deleted. (Or, if the
4565 spec did not exist, then nothing will happened.) Otherwise, if the spec
4566 does not currently exist a new spec will be created. If the spec does
4567 exist then its contents will be overridden by the text of this
4568 directive, unless the first character of that text is the @samp{+}
4569 character, in which case the text will be appended to the spec.
4571 @item [@var{suffix}]:
4572 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4573 and up to the next directive or blank line are considered to make up the
4574 spec string for the indicated suffix. When the compiler encounters an
4575 input file with the named suffix, it will processes the spec string in
4576 order to work out how to compile that file. For example:
4583 This says that any input file whose name ends in @samp{.ZZ} should be
4584 passed to the program @samp{z-compile}, which should be invoked with the
4585 command-line switch @option{-input} and with the result of performing the
4586 @samp{%i} substitution. (See below.)
4588 As an alternative to providing a spec string, the text that follows a
4589 suffix directive can be one of the following:
4592 @item @@@var{language}
4593 This says that the suffix is an alias for a known @var{language}. This is
4594 similar to using the @option{-x} command-line switch to GCC to specify a
4595 language explicitly. For example:
4602 Says that .ZZ files are, in fact, C++ source files.
4605 This causes an error messages saying:
4608 @var{name} compiler not installed on this system.
4612 GCC already has an extensive list of suffixes built into it.
4613 This directive will add an entry to the end of the list of suffixes, but
4614 since the list is searched from the end backwards, it is effectively
4615 possible to override earlier entries using this technique.
4619 GCC has the following spec strings built into it. Spec files can
4620 override these strings or create their own. Note that individual
4621 targets can also add their own spec strings to this list.
4624 asm Options to pass to the assembler
4625 asm_final Options to pass to the assembler post-processor
4626 cpp Options to pass to the C preprocessor
4627 cc1 Options to pass to the C compiler
4628 cc1plus Options to pass to the C++ compiler
4629 endfile Object files to include at the end of the link
4630 link Options to pass to the linker
4631 lib Libraries to include on the command line to the linker
4632 libgcc Decides which GCC support library to pass to the linker
4633 linker Sets the name of the linker
4634 predefines Defines to be passed to the C preprocessor
4635 signed_char Defines to pass to CPP to say whether @code{char} is signed
4637 startfile Object files to include at the start of the link
4640 Here is a small example of a spec file:
4646 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4649 This example renames the spec called @samp{lib} to @samp{old_lib} and
4650 then overrides the previous definition of @samp{lib} with a new one.
4651 The new definition adds in some extra command-line options before
4652 including the text of the old definition.
4654 @dfn{Spec strings} are a list of command-line options to be passed to their
4655 corresponding program. In addition, the spec strings can contain
4656 @samp{%}-prefixed sequences to substitute variable text or to
4657 conditionally insert text into the command line. Using these constructs
4658 it is possible to generate quite complex command lines.
4660 Here is a table of all defined @samp{%}-sequences for spec
4661 strings. Note that spaces are not generated automatically around the
4662 results of expanding these sequences. Therefore you can concatenate them
4663 together or combine them with constant text in a single argument.
4667 Substitute one @samp{%} into the program name or argument.
4670 Substitute the name of the input file being processed.
4673 Substitute the basename of the input file being processed.
4674 This is the substring up to (and not including) the last period
4675 and not including the directory.
4678 This is the same as @samp{%b}, but include the file suffix (text after
4682 Marks the argument containing or following the @samp{%d} as a
4683 temporary file name, so that that file will be deleted if GCC exits
4684 successfully. Unlike @samp{%g}, this contributes no text to the
4687 @item %g@var{suffix}
4688 Substitute a file name that has suffix @var{suffix} and is chosen
4689 once per compilation, and mark the argument in the same way as
4690 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4691 name is now chosen in a way that is hard to predict even when previously
4692 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4693 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4694 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4695 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4696 was simply substituted with a file name chosen once per compilation,
4697 without regard to any appended suffix (which was therefore treated
4698 just like ordinary text), making such attacks more likely to succeed.
4700 @item %u@var{suffix}
4701 Like @samp{%g}, but generates a new temporary file name even if
4702 @samp{%u@var{suffix}} was already seen.
4704 @item %U@var{suffix}
4705 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4706 new one if there is no such last file name. In the absence of any
4707 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4708 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4709 would involve the generation of two distinct file names, one
4710 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4711 simply substituted with a file name chosen for the previous @samp{%u},
4712 without regard to any appended suffix.
4714 @item %j@var{SUFFIX}
4715 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4716 writable, and if save-temps is off; otherwise, substitute the name
4717 of a temporary file, just like @samp{%u}. This temporary file is not
4718 meant for communication between processes, but rather as a junk
4721 @item %.@var{SUFFIX}
4722 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4723 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4724 terminated by the next space or %.
4727 Marks the argument containing or following the @samp{%w} as the
4728 designated output file of this compilation. This puts the argument
4729 into the sequence of arguments that @samp{%o} will substitute later.
4732 Substitutes the names of all the output files, with spaces
4733 automatically placed around them. You should write spaces
4734 around the @samp{%o} as well or the results are undefined.
4735 @samp{%o} is for use in the specs for running the linker.
4736 Input files whose names have no recognized suffix are not compiled
4737 at all, but they are included among the output files, so they will
4741 Substitutes the suffix for object files. Note that this is
4742 handled specially when it immediately follows @samp{%g, %u, or %U},
4743 because of the need for those to form complete file names. The
4744 handling is such that @samp{%O} is treated exactly as if it had already
4745 been substituted, except that @samp{%g, %u, and %U} do not currently
4746 support additional @var{suffix} characters following @samp{%O} as they would
4747 following, for example, @samp{.o}.
4750 Substitutes the standard macro predefinitions for the
4751 current target machine. Use this when running @code{cpp}.
4754 Like @samp{%p}, but puts @samp{__} before and after the name of each
4755 predefined macro, except for macros that start with @samp{__} or with
4756 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4760 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4763 Current argument is the name of a library or startup file of some sort.
4764 Search for that file in a standard list of directories and substitute
4765 the full name found.
4768 Print @var{str} as an error message. @var{str} is terminated by a newline.
4769 Use this when inconsistent options are detected.
4772 Output @samp{-} if the input for the current command is coming from a pipe.
4775 Substitute the contents of spec string @var{name} at this point.
4778 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4780 @item %x@{@var{option}@}
4781 Accumulate an option for @samp{%X}.
4784 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4788 Output the accumulated assembler options specified by @option{-Wa}.
4791 Output the accumulated preprocessor options specified by @option{-Wp}.
4794 Substitute the major version number of GCC@.
4795 (For version 2.9.5, this is 2.)
4798 Substitute the minor version number of GCC@.
4799 (For version 2.9.5, this is 9.)
4802 Substitute the patch level number of GCC@.
4803 (For version 2.9.5, this is 5.)
4806 Process the @code{asm} spec. This is used to compute the
4807 switches to be passed to the assembler.
4810 Process the @code{asm_final} spec. This is a spec string for
4811 passing switches to an assembler post-processor, if such a program is
4815 Process the @code{link} spec. This is the spec for computing the
4816 command line passed to the linker. Typically it will make use of the
4817 @samp{%L %G %S %D and %E} sequences.
4820 Dump out a @option{-L} option for each directory that GCC believes might
4821 contain startup files. If the target supports multilibs then the
4822 current multilib directory will be prepended to each of these paths.
4825 Output the multilib directory with directory separators replaced with
4826 @samp{_}. If multilib directories are not set, or the multilib directory is
4827 @file{.} then this option emits nothing.
4830 Process the @code{lib} spec. This is a spec string for deciding which
4831 libraries should be included on the command line to the linker.
4834 Process the @code{libgcc} spec. This is a spec string for deciding
4835 which GCC support library should be included on the command line to the linker.
4838 Process the @code{startfile} spec. This is a spec for deciding which
4839 object files should be the first ones passed to the linker. Typically
4840 this might be a file named @file{crt0.o}.
4843 Process the @code{endfile} spec. This is a spec string that specifies
4844 the last object files that will be passed to the linker.
4847 Process the @code{cpp} spec. This is used to construct the arguments
4848 to be passed to the C preprocessor.
4851 Process the @code{signed_char} spec. This is intended to be used
4852 to tell cpp whether a char is signed. It typically has the definition:
4854 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4858 Process the @code{cc1} spec. This is used to construct the options to be
4859 passed to the actual C compiler (@samp{cc1}).
4862 Process the @code{cc1plus} spec. This is used to construct the options to be
4863 passed to the actual C++ compiler (@samp{cc1plus}).
4866 Substitute the variable part of a matched option. See below.
4867 Note that each comma in the substituted string is replaced by
4871 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4872 If that switch was not specified, this substitutes nothing. Note that
4873 the leading dash is omitted when specifying this option, and it is
4874 automatically inserted if the substitution is performed. Thus the spec
4875 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4876 and would output the command line option @option{-foo}.
4878 @item %W@{@code{S}@}
4879 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4882 @item %@{@code{S}*@}
4883 Substitutes all the switches specified to GCC whose names start
4884 with @code{-S}, but which also take an argument. This is used for
4885 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4886 GCC considers @option{-o foo} as being
4887 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4888 text, including the space. Thus two arguments would be generated.
4890 @item %@{^@code{S}*@}
4891 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4892 argument. Thus %@{^o*@} would only generate one argument, not two.
4894 @item %@{@code{S}*&@code{T}*@}
4895 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4896 (the order of @code{S} and @code{T} in the spec is not significant).
4897 There can be any number of ampersand-separated variables; for each the
4898 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4900 @item %@{<@code{S}@}
4901 Remove all occurrences of @code{-S} from the command line. Note---this
4902 command is position dependent. @samp{%} commands in the spec string
4903 before this option will see @code{-S}, @samp{%} commands in the spec
4904 string after this option will not.
4906 @item %@{@code{S}*:@code{X}@}
4907 Substitutes @code{X} if one or more switches whose names start with
4908 @code{-S} are specified to GCC@. Note that the tail part of the
4909 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4910 for each occurrence of @samp{%*} within @code{X}.
4912 @item %@{@code{S}:@code{X}@}
4913 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4915 @item %@{!@code{S}:@code{X}@}
4916 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4918 @item %@{|@code{S}:@code{X}@}
4919 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4921 @item %@{|!@code{S}:@code{X}@}
4922 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4924 @item %@{.@code{S}:@code{X}@}
4925 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4927 @item %@{!.@code{S}:@code{X}@}
4928 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4930 @item %@{@code{S}|@code{P}:@code{X}@}
4931 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4932 combined with @samp{!} and @samp{.} sequences as well, although they
4933 have a stronger binding than the @samp{|}. For example a spec string
4937 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4940 will output the following command-line options from the following input
4941 command-line options:
4946 -d fred.c -foo -baz -boggle
4947 -d jim.d -bar -baz -boggle
4952 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4953 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4954 or spaces, or even newlines. They are processed as usual, as described
4957 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4958 switches are handled specifically in these
4959 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4960 @option{-W} switch is found later in the command line, the earlier switch
4961 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4962 letter, which passes all matching options.
4964 The character @samp{|} at the beginning of the predicate text is used to indicate
4965 that a command should be piped to the following command, but only if @option{-pipe}
4968 It is built into GCC which switches take arguments and which do not.
4969 (You might think it would be useful to generalize this to allow each
4970 compiler's spec to say which switches take arguments. But this cannot
4971 be done in a consistent fashion. GCC cannot even decide which input
4972 files have been specified without knowing which switches take arguments,
4973 and it must know which input files to compile in order to tell which
4976 GCC also knows implicitly that arguments starting in @option{-l} are to be
4977 treated as compiler output files, and passed to the linker in their
4978 proper position among the other output files.
4980 @c man begin OPTIONS
4982 @node Target Options
4983 @section Specifying Target Machine and Compiler Version
4984 @cindex target options
4985 @cindex cross compiling
4986 @cindex specifying machine version
4987 @cindex specifying compiler version and target machine
4988 @cindex compiler version, specifying
4989 @cindex target machine, specifying
4991 The usual way to run GCC is to run the executable called @file{gcc}, or
4992 @file{<machine>-gcc} when cross-compiling, or
4993 @file{<machine>-gcc-<version>} to run a version other than the one that
4994 was installed last. Sometimes this is inconvenient, so GCC provides
4995 options that will switch to another cross-compiler or version.
4998 @item -b @var{machine}
5000 The argument @var{machine} specifies the target machine for compilation.
5002 The value to use for @var{machine} is the same as was specified as the
5003 machine type when configuring GCC as a cross-compiler. For
5004 example, if a cross-compiler was configured with @samp{configure
5005 i386v}, meaning to compile for an 80386 running System V, then you
5006 would specify @option{-b i386v} to run that cross compiler.
5008 @item -V @var{version}
5010 The argument @var{version} specifies which version of GCC to run.
5011 This is useful when multiple versions are installed. For example,
5012 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5015 The @option{-V} and @option{-b} options work by running the
5016 @file{<machine>-gcc-<version>} executable, so there's no real reason to
5017 use them if you can just run that directly.
5019 @node Submodel Options
5020 @section Hardware Models and Configurations
5021 @cindex submodel options
5022 @cindex specifying hardware config
5023 @cindex hardware models and configurations, specifying
5024 @cindex machine dependent options
5026 Earlier we discussed the standard option @option{-b} which chooses among
5027 different installed compilers for completely different target
5028 machines, such as VAX vs.@: 68000 vs.@: 80386.
5030 In addition, each of these target machine types can have its own
5031 special options, starting with @samp{-m}, to choose among various
5032 hardware models or configurations---for example, 68010 vs 68020,
5033 floating coprocessor or none. A single installed version of the
5034 compiler can compile for any model or configuration, according to the
5037 Some configurations of the compiler also support additional special
5038 options, usually for compatibility with other compilers on the same
5041 These options are defined by the macro @code{TARGET_SWITCHES} in the
5042 machine description. The default for the options is also defined by
5043 that macro, which enables you to change the defaults.
5057 * RS/6000 and PowerPC Options::
5060 * i386 and x86-64 Options::
5062 * Intel 960 Options::
5063 * DEC Alpha Options::
5064 * DEC Alpha/VMS Options::
5068 * System V Options::
5069 * TMS320C3x/C4x Options::
5077 * S/390 and zSeries Options::
5081 * Xstormy16 Options::
5085 @node M680x0 Options
5086 @subsection M680x0 Options
5087 @cindex M680x0 options
5089 These are the @samp{-m} options defined for the 68000 series. The default
5090 values for these options depends on which style of 68000 was selected when
5091 the compiler was configured; the defaults for the most common choices are
5099 Generate output for a 68000. This is the default
5100 when the compiler is configured for 68000-based systems.
5102 Use this option for microcontrollers with a 68000 or EC000 core,
5103 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5109 Generate output for a 68020. This is the default
5110 when the compiler is configured for 68020-based systems.
5114 Generate output containing 68881 instructions for floating point.
5115 This is the default for most 68020 systems unless @option{--nfp} was
5116 specified when the compiler was configured.
5120 Generate output for a 68030. This is the default when the compiler is
5121 configured for 68030-based systems.
5125 Generate output for a 68040. This is the default when the compiler is
5126 configured for 68040-based systems.
5128 This option inhibits the use of 68881/68882 instructions that have to be
5129 emulated by software on the 68040. Use this option if your 68040 does not
5130 have code to emulate those instructions.
5134 Generate output for a 68060. This is the default when the compiler is
5135 configured for 68060-based systems.
5137 This option inhibits the use of 68020 and 68881/68882 instructions that
5138 have to be emulated by software on the 68060. Use this option if your 68060
5139 does not have code to emulate those instructions.
5143 Generate output for a CPU32. This is the default
5144 when the compiler is configured for CPU32-based systems.
5146 Use this option for microcontrollers with a
5147 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5148 68336, 68340, 68341, 68349 and 68360.
5152 Generate output for a 520X ``coldfire'' family cpu. This is the default
5153 when the compiler is configured for 520X-based systems.
5155 Use this option for microcontroller with a 5200 core, including
5156 the MCF5202, MCF5203, MCF5204 and MCF5202.
5161 Generate output for a 68040, without using any of the new instructions.
5162 This results in code which can run relatively efficiently on either a
5163 68020/68881 or a 68030 or a 68040. The generated code does use the
5164 68881 instructions that are emulated on the 68040.
5168 Generate output for a 68060, without using any of the new instructions.
5169 This results in code which can run relatively efficiently on either a
5170 68020/68881 or a 68030 or a 68040. The generated code does use the
5171 68881 instructions that are emulated on the 68060.
5175 Generate output containing Sun FPA instructions for floating point.
5178 @opindex msoft-float
5179 Generate output containing library calls for floating point.
5180 @strong{Warning:} the requisite libraries are not available for all m68k
5181 targets. Normally the facilities of the machine's usual C compiler are
5182 used, but this can't be done directly in cross-compilation. You must
5183 make your own arrangements to provide suitable library functions for
5184 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5185 @samp{m68k-*-coff} do provide software floating point support.
5189 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5192 @opindex mnobitfield
5193 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5194 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5198 Do use the bit-field instructions. The @option{-m68020} option implies
5199 @option{-mbitfield}. This is the default if you use a configuration
5200 designed for a 68020.
5204 Use a different function-calling convention, in which functions
5205 that take a fixed number of arguments return with the @code{rtd}
5206 instruction, which pops their arguments while returning. This
5207 saves one instruction in the caller since there is no need to pop
5208 the arguments there.
5210 This calling convention is incompatible with the one normally
5211 used on Unix, so you cannot use it if you need to call libraries
5212 compiled with the Unix compiler.
5214 Also, you must provide function prototypes for all functions that
5215 take variable numbers of arguments (including @code{printf});
5216 otherwise incorrect code will be generated for calls to those
5219 In addition, seriously incorrect code will result if you call a
5220 function with too many arguments. (Normally, extra arguments are
5221 harmlessly ignored.)
5223 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5224 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5227 @itemx -mno-align-int
5229 @opindex mno-align-int
5230 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5231 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5232 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5233 Aligning variables on 32-bit boundaries produces code that runs somewhat
5234 faster on processors with 32-bit busses at the expense of more memory.
5236 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5237 align structures containing the above types differently than
5238 most published application binary interface specifications for the m68k.
5242 Use the pc-relative addressing mode of the 68000 directly, instead of
5243 using a global offset table. At present, this option implies @option{-fpic},
5244 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5245 not presently supported with @option{-mpcrel}, though this could be supported for
5246 68020 and higher processors.
5248 @item -mno-strict-align
5249 @itemx -mstrict-align
5250 @opindex mno-strict-align
5251 @opindex mstrict-align
5252 Do not (do) assume that unaligned memory references will be handled by
5257 @node M68hc1x Options
5258 @subsection M68hc1x Options
5259 @cindex M68hc1x options
5261 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5262 microcontrollers. The default values for these options depends on
5263 which style of microcontroller was selected when the compiler was configured;
5264 the defaults for the most common choices are given below.
5271 Generate output for a 68HC11. This is the default
5272 when the compiler is configured for 68HC11-based systems.
5278 Generate output for a 68HC12. This is the default
5279 when the compiler is configured for 68HC12-based systems.
5282 @opindex mauto-incdec
5283 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5290 Enable the use of 68HC12 min and max instructions.
5293 @itemx -mno-long-calls
5294 @opindex mlong-calls
5295 @opindex mno-long-calls
5296 Treat all calls as being far away (near). If calls are assumed to be
5297 far away, the compiler will use the @code{call} instruction to
5298 call a function and the @code{rtc} instruction for returning.
5302 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5304 @item -msoft-reg-count=@var{count}
5305 @opindex msoft-reg-count
5306 Specify the number of pseudo-soft registers which are used for the
5307 code generation. The maximum number is 32. Using more pseudo-soft
5308 register may or may not result in better code depending on the program.
5309 The default is 4 for 68HC11 and 2 for 68HC12.
5314 @subsection VAX Options
5317 These @samp{-m} options are defined for the VAX:
5322 Do not output certain jump instructions (@code{aobleq} and so on)
5323 that the Unix assembler for the VAX cannot handle across long
5328 Do output those jump instructions, on the assumption that you
5329 will assemble with the GNU assembler.
5333 Output code for g-format floating point numbers instead of d-format.
5337 @subsection SPARC Options
5338 @cindex SPARC options
5340 These @samp{-m} switches are supported on the SPARC:
5345 @opindex mno-app-regs
5347 Specify @option{-mapp-regs} to generate output using the global registers
5348 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5351 To be fully SVR4 ABI compliant at the cost of some performance loss,
5352 specify @option{-mno-app-regs}. You should compile libraries and system
5353 software with this option.
5358 @opindex mhard-float
5359 Generate output containing floating point instructions. This is the
5365 @opindex msoft-float
5366 Generate output containing library calls for floating point.
5367 @strong{Warning:} the requisite libraries are not available for all SPARC
5368 targets. Normally the facilities of the machine's usual C compiler are
5369 used, but this cannot be done directly in cross-compilation. You must make
5370 your own arrangements to provide suitable library functions for
5371 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5372 @samp{sparclite-*-*} do provide software floating point support.
5374 @option{-msoft-float} changes the calling convention in the output file;
5375 therefore, it is only useful if you compile @emph{all} of a program with
5376 this option. In particular, you need to compile @file{libgcc.a}, the
5377 library that comes with GCC, with @option{-msoft-float} in order for
5380 @item -mhard-quad-float
5381 @opindex mhard-quad-float
5382 Generate output containing quad-word (long double) floating point
5385 @item -msoft-quad-float
5386 @opindex msoft-quad-float
5387 Generate output containing library calls for quad-word (long double)
5388 floating point instructions. The functions called are those specified
5389 in the SPARC ABI@. This is the default.
5391 As of this writing, there are no sparc implementations that have hardware
5392 support for the quad-word floating point instructions. They all invoke
5393 a trap handler for one of these instructions, and then the trap handler
5394 emulates the effect of the instruction. Because of the trap handler overhead,
5395 this is much slower than calling the ABI library routines. Thus the
5396 @option{-msoft-quad-float} option is the default.
5402 With @option{-mflat}, the compiler does not generate save/restore instructions
5403 and will use a ``flat'' or single register window calling convention.
5404 This model uses %i7 as the frame pointer and is compatible with the normal
5405 register window model. Code from either may be intermixed.
5406 The local registers and the input registers (0--5) are still treated as
5407 ``call saved'' registers and will be saved on the stack as necessary.
5409 With @option{-mno-flat} (the default), the compiler emits save/restore
5410 instructions (except for leaf functions) and is the normal mode of operation.
5412 @item -mno-unaligned-doubles
5413 @itemx -munaligned-doubles
5414 @opindex mno-unaligned-doubles
5415 @opindex munaligned-doubles
5416 Assume that doubles have 8 byte alignment. This is the default.
5418 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5419 alignment only if they are contained in another type, or if they have an
5420 absolute address. Otherwise, it assumes they have 4 byte alignment.
5421 Specifying this option avoids some rare compatibility problems with code
5422 generated by other compilers. It is not the default because it results
5423 in a performance loss, especially for floating point code.
5425 @item -mno-faster-structs
5426 @itemx -mfaster-structs
5427 @opindex mno-faster-structs
5428 @opindex mfaster-structs
5429 With @option{-mfaster-structs}, the compiler assumes that structures
5430 should have 8 byte alignment. This enables the use of pairs of
5431 @code{ldd} and @code{std} instructions for copies in structure
5432 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5433 However, the use of this changed alignment directly violates the SPARC
5434 ABI@. Thus, it's intended only for use on targets where the developer
5435 acknowledges that their resulting code will not be directly in line with
5436 the rules of the ABI@.
5442 These two options select variations on the SPARC architecture.
5444 By default (unless specifically configured for the Fujitsu SPARClite),
5445 GCC generates code for the v7 variant of the SPARC architecture.
5447 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5448 code is that the compiler emits the integer multiply and integer
5449 divide instructions which exist in SPARC v8 but not in SPARC v7.
5451 @option{-msparclite} will give you SPARClite code. This adds the integer
5452 multiply, integer divide step and scan (@code{ffs}) instructions which
5453 exist in SPARClite but not in SPARC v7.
5455 These options are deprecated and will be deleted in a future GCC release.
5456 They have been replaced with @option{-mcpu=xxx}.
5461 @opindex msupersparc
5462 These two options select the processor for which the code is optimized.
5464 With @option{-mcypress} (the default), the compiler optimizes code for the
5465 Cypress CY7C602 chip, as used in the SPARCStation/SPARCServer 3xx series.
5466 This is also appropriate for the older SPARCStation 1, 2, IPX etc.
5468 With @option{-msupersparc} the compiler optimizes code for the SuperSPARC cpu, as
5469 used in the SPARCStation 10, 1000 and 2000 series. This flag also enables use
5470 of the full SPARC v8 instruction set.
5472 These options are deprecated and will be deleted in a future GCC release.
5473 They have been replaced with @option{-mcpu=xxx}.
5475 @item -mcpu=@var{cpu_type}
5477 Set the instruction set, register set, and instruction scheduling parameters
5478 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5479 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5480 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5481 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5483 Default instruction scheduling parameters are used for values that select
5484 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5485 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5487 Here is a list of each supported architecture and their supported
5492 v8: supersparc, hypersparc
5493 sparclite: f930, f934, sparclite86x
5498 @item -mtune=@var{cpu_type}
5500 Set the instruction scheduling parameters for machine type
5501 @var{cpu_type}, but do not set the instruction set or register set that the
5502 option @option{-mcpu=@var{cpu_type}} would.
5504 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5505 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5506 that select a particular cpu implementation. Those are @samp{cypress},
5507 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5508 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5512 These @samp{-m} switches are supported in addition to the above
5513 on the SPARCLET processor.
5516 @item -mlittle-endian
5517 @opindex mlittle-endian
5518 Generate code for a processor running in little-endian mode.
5522 Treat register @code{%g0} as a normal register.
5523 GCC will continue to clobber it as necessary but will not assume
5524 it always reads as 0.
5526 @item -mbroken-saverestore
5527 @opindex mbroken-saverestore
5528 Generate code that does not use non-trivial forms of the @code{save} and
5529 @code{restore} instructions. Early versions of the SPARCLET processor do
5530 not correctly handle @code{save} and @code{restore} instructions used with
5531 arguments. They correctly handle them used without arguments. A @code{save}
5532 instruction used without arguments increments the current window pointer
5533 but does not allocate a new stack frame. It is assumed that the window
5534 overflow trap handler will properly handle this case as will interrupt
5538 These @samp{-m} switches are supported in addition to the above
5539 on SPARC V9 processors in 64-bit environments.
5542 @item -mlittle-endian
5543 @opindex mlittle-endian
5544 Generate code for a processor running in little-endian mode.
5550 Generate code for a 32-bit or 64-bit environment.
5551 The 32-bit environment sets int, long and pointer to 32 bits.
5552 The 64-bit environment sets int to 32 bits and long and pointer
5555 @item -mcmodel=medlow
5556 @opindex mcmodel=medlow
5557 Generate code for the Medium/Low code model: the program must be linked
5558 in the low 32 bits of the address space. Pointers are 64 bits.
5559 Programs can be statically or dynamically linked.
5561 @item -mcmodel=medmid
5562 @opindex mcmodel=medmid
5563 Generate code for the Medium/Middle code model: the program must be linked
5564 in the low 44 bits of the address space, the text segment must be less than
5565 2G bytes, and data segment must be within 2G of the text segment.
5566 Pointers are 64 bits.
5568 @item -mcmodel=medany
5569 @opindex mcmodel=medany
5570 Generate code for the Medium/Anywhere code model: the program may be linked
5571 anywhere in the address space, the text segment must be less than
5572 2G bytes, and data segment must be within 2G of the text segment.
5573 Pointers are 64 bits.
5575 @item -mcmodel=embmedany
5576 @opindex mcmodel=embmedany
5577 Generate code for the Medium/Anywhere code model for embedded systems:
5578 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5579 (determined at link time). Register %g4 points to the base of the
5580 data segment. Pointers are still 64 bits.
5581 Programs are statically linked, PIC is not supported.
5584 @itemx -mno-stack-bias
5585 @opindex mstack-bias
5586 @opindex mno-stack-bias
5587 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5588 frame pointer if present, are offset by @minus{}2047 which must be added back
5589 when making stack frame references.
5590 Otherwise, assume no such offset is present.
5593 @node Convex Options
5594 @subsection Convex Options
5595 @cindex Convex options
5597 These @samp{-m} options are defined for Convex:
5602 Generate output for C1. The code will run on any Convex machine.
5603 The preprocessor symbol @code{__convex__c1__} is defined.
5607 Generate output for C2. Uses instructions not available on C1.
5608 Scheduling and other optimizations are chosen for max performance on C2.
5609 The preprocessor symbol @code{__convex_c2__} is defined.
5613 Generate output for C32xx. Uses instructions not available on C1.
5614 Scheduling and other optimizations are chosen for max performance on C32.
5615 The preprocessor symbol @code{__convex_c32__} is defined.
5619 Generate output for C34xx. Uses instructions not available on C1.
5620 Scheduling and other optimizations are chosen for max performance on C34.
5621 The preprocessor symbol @code{__convex_c34__} is defined.
5625 Generate output for C38xx. Uses instructions not available on C1.
5626 Scheduling and other optimizations are chosen for max performance on C38.
5627 The preprocessor symbol @code{__convex_c38__} is defined.
5631 Generate code which puts an argument count in the word preceding each
5632 argument list. This is compatible with regular CC, and a few programs
5633 may need the argument count word. GDB and other source-level debuggers
5634 do not need it; this info is in the symbol table.
5637 @opindex mnoargcount
5638 Omit the argument count word. This is the default.
5640 @item -mvolatile-cache
5641 @opindex mvolatile-cache
5642 Allow volatile references to be cached. This is the default.
5644 @item -mvolatile-nocache
5645 @opindex mvolatile-nocache
5646 Volatile references bypass the data cache, going all the way to memory.
5647 This is only needed for multi-processor code that does not use standard
5648 synchronization instructions. Making non-volatile references to volatile
5649 locations will not necessarily work.
5653 Type long is 32 bits, the same as type int. This is the default.
5657 Type long is 64 bits, the same as type long long. This option is useless,
5658 because no library support exists for it.
5661 @node AMD29K Options
5662 @subsection AMD29K Options
5663 @cindex AMD29K options
5665 These @samp{-m} options are defined for the AMD Am29000:
5670 @cindex DW bit (29k)
5671 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5672 halfword operations are directly supported by the hardware. This is the
5677 Generate code that assumes the @code{DW} bit is not set.
5681 @cindex byte writes (29k)
5682 Generate code that assumes the system supports byte and halfword write
5683 operations. This is the default.
5687 Generate code that assumes the systems does not support byte and
5688 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5692 @cindex memory model (29k)
5693 Use a small memory model that assumes that all function addresses are
5694 either within a single 256 KB segment or at an absolute address of less
5695 than 256k. This allows the @code{call} instruction to be used instead
5696 of a @code{const}, @code{consth}, @code{calli} sequence.
5700 Use the normal memory model: Generate @code{call} instructions only when
5701 calling functions in the same file and @code{calli} instructions
5702 otherwise. This works if each file occupies less than 256 KB but allows
5703 the entire executable to be larger than 256 KB@. This is the default.
5707 Always use @code{calli} instructions. Specify this option if you expect
5708 a single file to compile into more than 256 KB of code.
5712 @cindex processor selection (29k)
5713 Generate code for the Am29050.
5717 Generate code for the Am29000. This is the default.
5719 @item -mkernel-registers
5720 @opindex mkernel-registers
5721 @cindex kernel and user registers (29k)
5722 Generate references to registers @code{gr64-gr95} instead of to
5723 registers @code{gr96-gr127}. This option can be used when compiling
5724 kernel code that wants a set of global registers disjoint from that used
5727 Note that when this option is used, register names in @samp{-f} flags
5728 must use the normal, user-mode, names.
5730 @item -muser-registers
5731 @opindex muser-registers
5732 Use the normal set of global registers, @code{gr96-gr127}. This is the
5736 @itemx -mno-stack-check
5737 @opindex mstack-check
5738 @opindex mno-stack-check
5739 @cindex stack checks (29k)
5740 Insert (or do not insert) a call to @code{__msp_check} after each stack
5741 adjustment. This is often used for kernel code.
5744 @itemx -mno-storem-bug
5745 @opindex mstorem-bug
5746 @opindex mno-storem-bug
5747 @cindex storem bug (29k)
5748 @option{-mstorem-bug} handles 29k processors which cannot handle the
5749 separation of a mtsrim insn and a storem instruction (most 29000 chips
5750 to date, but not the 29050).
5752 @item -mno-reuse-arg-regs
5753 @itemx -mreuse-arg-regs
5754 @opindex mno-reuse-arg-regs
5755 @opindex mreuse-arg-regs
5756 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5757 registers for copying out arguments. This helps detect calling a function
5758 with fewer arguments than it was declared with.
5760 @item -mno-impure-text
5761 @itemx -mimpure-text
5762 @opindex mno-impure-text
5763 @opindex mimpure-text
5764 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5765 not pass @option{-assert pure-text} to the linker when linking a shared object.
5768 @opindex msoft-float
5769 Generate output containing library calls for floating point.
5770 @strong{Warning:} the requisite libraries are not part of GCC@.
5771 Normally the facilities of the machine's usual C compiler are used, but
5772 this can't be done directly in cross-compilation. You must make your
5773 own arrangements to provide suitable library functions for
5778 Do not generate multm or multmu instructions. This is useful for some embedded
5779 systems which do not have trap handlers for these instructions.
5783 @subsection ARM Options
5786 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5791 @opindex mapcs-frame
5792 Generate a stack frame that is compliant with the ARM Procedure Call
5793 Standard for all functions, even if this is not strictly necessary for
5794 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5795 with this option will cause the stack frames not to be generated for
5796 leaf functions. The default is @option{-mno-apcs-frame}.
5800 This is a synonym for @option{-mapcs-frame}.
5804 Generate code for a processor running with a 26-bit program counter,
5805 and conforming to the function calling standards for the APCS 26-bit
5806 option. This option replaces the @option{-m2} and @option{-m3} options
5807 of previous releases of the compiler.
5811 Generate code for a processor running with a 32-bit program counter,
5812 and conforming to the function calling standards for the APCS 32-bit
5813 option. This option replaces the @option{-m6} option of previous releases
5817 @c not currently implemented
5818 @item -mapcs-stack-check
5819 @opindex mapcs-stack-check
5820 Generate code to check the amount of stack space available upon entry to
5821 every function (that actually uses some stack space). If there is
5822 insufficient space available then either the function
5823 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5824 called, depending upon the amount of stack space required. The run time
5825 system is required to provide these functions. The default is
5826 @option{-mno-apcs-stack-check}, since this produces smaller code.
5828 @c not currently implemented
5830 @opindex mapcs-float
5831 Pass floating point arguments using the float point registers. This is
5832 one of the variants of the APCS@. This option is recommended if the
5833 target hardware has a floating point unit or if a lot of floating point
5834 arithmetic is going to be performed by the code. The default is
5835 @option{-mno-apcs-float}, since integer only code is slightly increased in
5836 size if @option{-mapcs-float} is used.
5838 @c not currently implemented
5839 @item -mapcs-reentrant
5840 @opindex mapcs-reentrant
5841 Generate reentrant, position independent code. The default is
5842 @option{-mno-apcs-reentrant}.
5845 @item -mthumb-interwork
5846 @opindex mthumb-interwork
5847 Generate code which supports calling between the ARM and Thumb
5848 instruction sets. Without this option the two instruction sets cannot
5849 be reliably used inside one program. The default is
5850 @option{-mno-thumb-interwork}, since slightly larger code is generated
5851 when @option{-mthumb-interwork} is specified.
5853 @item -mno-sched-prolog
5854 @opindex mno-sched-prolog
5855 Prevent the reordering of instructions in the function prolog, or the
5856 merging of those instruction with the instructions in the function's
5857 body. This means that all functions will start with a recognizable set
5858 of instructions (or in fact one of a choice from a small set of
5859 different function prologues), and this information can be used to
5860 locate the start if functions inside an executable piece of code. The
5861 default is @option{-msched-prolog}.
5864 @opindex mhard-float
5865 Generate output containing floating point instructions. This is the
5869 @opindex msoft-float
5870 Generate output containing library calls for floating point.
5871 @strong{Warning:} the requisite libraries are not available for all ARM
5872 targets. Normally the facilities of the machine's usual C compiler are
5873 used, but this cannot be done directly in cross-compilation. You must make
5874 your own arrangements to provide suitable library functions for
5877 @option{-msoft-float} changes the calling convention in the output file;
5878 therefore, it is only useful if you compile @emph{all} of a program with
5879 this option. In particular, you need to compile @file{libgcc.a}, the
5880 library that comes with GCC, with @option{-msoft-float} in order for
5883 @item -mlittle-endian
5884 @opindex mlittle-endian
5885 Generate code for a processor running in little-endian mode. This is
5886 the default for all standard configurations.
5889 @opindex mbig-endian
5890 Generate code for a processor running in big-endian mode; the default is
5891 to compile code for a little-endian processor.
5893 @item -mwords-little-endian
5894 @opindex mwords-little-endian
5895 This option only applies when generating code for big-endian processors.
5896 Generate code for a little-endian word order but a big-endian byte
5897 order. That is, a byte order of the form @samp{32107654}. Note: this
5898 option should only be used if you require compatibility with code for
5899 big-endian ARM processors generated by versions of the compiler prior to
5902 @item -malignment-traps
5903 @opindex malignment-traps
5904 Generate code that will not trap if the MMU has alignment traps enabled.
5905 On ARM architectures prior to ARMv4, there were no instructions to
5906 access half-word objects stored in memory. However, when reading from
5907 memory a feature of the ARM architecture allows a word load to be used,
5908 even if the address is unaligned, and the processor core will rotate the
5909 data as it is being loaded. This option tells the compiler that such
5910 misaligned accesses will cause a MMU trap and that it should instead
5911 synthesize the access as a series of byte accesses. The compiler can
5912 still use word accesses to load half-word data if it knows that the
5913 address is aligned to a word boundary.
5915 This option is ignored when compiling for ARM architecture 4 or later,
5916 since these processors have instructions to directly access half-word
5919 @item -mno-alignment-traps
5920 @opindex mno-alignment-traps
5921 Generate code that assumes that the MMU will not trap unaligned
5922 accesses. This produces better code when the target instruction set
5923 does not have half-word memory operations (i.e.@: implementations prior to
5926 Note that you cannot use this option to access unaligned word objects,
5927 since the processor will only fetch one 32-bit aligned object from
5930 The default setting for most targets is @option{-mno-alignment-traps}, since
5931 this produces better code when there are no half-word memory
5932 instructions available.
5934 @item -mshort-load-bytes
5935 @itemx -mno-short-load-words
5936 @opindex mshort-load-bytes
5937 @opindex mno-short-load-words
5938 These are deprecated aliases for @option{-malignment-traps}.
5940 @item -mno-short-load-bytes
5941 @itemx -mshort-load-words
5942 @opindex mno-short-load-bytes
5943 @opindex mshort-load-words
5944 This are deprecated aliases for @option{-mno-alignment-traps}.
5948 This option only applies to RISC iX@. Emulate the native BSD-mode
5949 compiler. This is the default if @option{-ansi} is not specified.
5953 This option only applies to RISC iX@. Emulate the native X/Open-mode
5956 @item -mno-symrename
5957 @opindex mno-symrename
5958 This option only applies to RISC iX@. Do not run the assembler
5959 post-processor, @samp{symrename}, after code has been assembled.
5960 Normally it is necessary to modify some of the standard symbols in
5961 preparation for linking with the RISC iX C library; this option
5962 suppresses this pass. The post-processor is never run when the
5963 compiler is built for cross-compilation.
5965 @item -mcpu=@var{name}
5967 This specifies the name of the target ARM processor. GCC uses this name
5968 to determine what kind of instructions it can emit when generating
5969 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
5970 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
5971 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
5972 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
5973 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
5974 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
5975 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
5976 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
5977 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
5978 @samp{arm1020t}, @samp{xscale}.
5980 @itemx -mtune=@var{name}
5982 This option is very similar to the @option{-mcpu=} option, except that
5983 instead of specifying the actual target processor type, and hence
5984 restricting which instructions can be used, it specifies that GCC should
5985 tune the performance of the code as if the target were of the type
5986 specified in this option, but still choosing the instructions that it
5987 will generate based on the cpu specified by a @option{-mcpu=} option.
5988 For some ARM implementations better performance can be obtained by using
5991 @item -march=@var{name}
5993 This specifies the name of the target ARM architecture. GCC uses this
5994 name to determine what kind of instructions it can emit when generating
5995 assembly code. This option can be used in conjunction with or instead
5996 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
5997 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
5998 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
6000 @item -mfpe=@var{number}
6001 @itemx -mfp=@var{number}
6004 This specifies the version of the floating point emulation available on
6005 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6006 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6008 @item -mstructure-size-boundary=@var{n}
6009 @opindex mstructure-size-boundary
6010 The size of all structures and unions will be rounded up to a multiple
6011 of the number of bits set by this option. Permissible values are 8 and
6012 32. The default value varies for different toolchains. For the COFF
6013 targeted toolchain the default value is 8. Specifying the larger number
6014 can produce faster, more efficient code, but can also increase the size
6015 of the program. The two values are potentially incompatible. Code
6016 compiled with one value cannot necessarily expect to work with code or
6017 libraries compiled with the other value, if they exchange information
6018 using structures or unions.
6020 @item -mabort-on-noreturn
6021 @opindex mabort-on-noreturn
6022 Generate a call to the function @code{abort} at the end of a
6023 @code{noreturn} function. It will be executed if the function tries to
6027 @itemx -mno-long-calls
6028 @opindex mlong-calls
6029 @opindex mno-long-calls
6030 Tells the compiler to perform function calls by first loading the
6031 address of the function into a register and then performing a subroutine
6032 call on this register. This switch is needed if the target function
6033 will lie outside of the 64 megabyte addressing range of the offset based
6034 version of subroutine call instruction.
6036 Even if this switch is enabled, not all function calls will be turned
6037 into long calls. The heuristic is that static functions, functions
6038 which have the @samp{short-call} attribute, functions that are inside
6039 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6040 definitions have already been compiled within the current compilation
6041 unit, will not be turned into long calls. The exception to this rule is
6042 that weak function definitions, functions with the @samp{long-call}
6043 attribute or the @samp{section} attribute, and functions that are within
6044 the scope of a @samp{#pragma long_calls} directive, will always be
6045 turned into long calls.
6047 This feature is not enabled by default. Specifying
6048 @option{-mno-long-calls} will restore the default behavior, as will
6049 placing the function calls within the scope of a @samp{#pragma
6050 long_calls_off} directive. Note these switches have no effect on how
6051 the compiler generates code to handle function calls via function
6054 @item -mnop-fun-dllimport
6055 @opindex mnop-fun-dllimport
6056 Disable support for the @code{dllimport} attribute.
6058 @item -msingle-pic-base
6059 @opindex msingle-pic-base
6060 Treat the register used for PIC addressing as read-only, rather than
6061 loading it in the prologue for each function. The run-time system is
6062 responsible for initializing this register with an appropriate value
6063 before execution begins.
6065 @item -mpic-register=@var{reg}
6066 @opindex mpic-register
6067 Specify the register to be used for PIC addressing. The default is R10
6068 unless stack-checking is enabled, when R9 is used.
6070 @item -mpoke-function-name
6071 @opindex mpoke-function-name
6072 Write the name of each function into the text section, directly
6073 preceding the function prologue. The generated code is similar to this:
6077 .ascii "arm_poke_function_name", 0
6080 .word 0xff000000 + (t1 - t0)
6081 arm_poke_function_name
6083 stmfd sp!, @{fp, ip, lr, pc@}
6087 When performing a stack backtrace, code can inspect the value of
6088 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6089 location @code{pc - 12} and the top 8 bits are set, then we know that
6090 there is a function name embedded immediately preceding this location
6091 and has length @code{((pc[-3]) & 0xff000000)}.
6095 Generate code for the 16-bit Thumb instruction set. The default is to
6096 use the 32-bit ARM instruction set.
6099 @opindex mtpcs-frame
6100 Generate a stack frame that is compliant with the Thumb Procedure Call
6101 Standard for all non-leaf functions. (A leaf function is one that does
6102 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6104 @item -mtpcs-leaf-frame
6105 @opindex mtpcs-leaf-frame
6106 Generate a stack frame that is compliant with the Thumb Procedure Call
6107 Standard for all leaf functions. (A leaf function is one that does
6108 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6110 @item -mcallee-super-interworking
6111 @opindex mcallee-super-interworking
6112 Gives all externally visible functions in the file being compiled an ARM
6113 instruction set header which switches to Thumb mode before executing the
6114 rest of the function. This allows these functions to be called from
6115 non-interworking code.
6117 @item -mcaller-super-interworking
6118 @opindex mcaller-super-interworking
6119 Allows calls via function pointers (including virtual functions) to
6120 execute correctly regardless of whether the target code has been
6121 compiled for interworking or not. There is a small overhead in the cost
6122 of executing a function pointer if this option is enabled.
6126 @node MN10200 Options
6127 @subsection MN10200 Options
6128 @cindex MN10200 options
6129 These @option{-m} options are defined for Matsushita MN10200 architectures:
6134 Indicate to the linker that it should perform a relaxation optimization pass
6135 to shorten branches, calls and absolute memory addresses. This option only
6136 has an effect when used on the command line for the final link step.
6138 This option makes symbolic debugging impossible.
6141 @node MN10300 Options
6142 @subsection MN10300 Options
6143 @cindex MN10300 options
6144 These @option{-m} options are defined for Matsushita MN10300 architectures:
6149 Generate code to avoid bugs in the multiply instructions for the MN10300
6150 processors. This is the default.
6153 @opindex mno-mult-bug
6154 Do not generate code to avoid bugs in the multiply instructions for the
6159 Generate code which uses features specific to the AM33 processor.
6163 Do not generate code which uses features specific to the AM33 processor. This
6168 Do not link in the C run-time initialization object file.
6172 Indicate to the linker that it should perform a relaxation optimization pass
6173 to shorten branches, calls and absolute memory addresses. This option only
6174 has an effect when used on the command line for the final link step.
6176 This option makes symbolic debugging impossible.
6180 @node M32R/D Options
6181 @subsection M32R/D Options
6182 @cindex M32R/D options
6184 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6189 Generate code for the M32R/X@.
6193 Generate code for the M32R@. This is the default.
6195 @item -mcode-model=small
6196 @opindex mcode-model=small
6197 Assume all objects live in the lower 16MB of memory (so that their addresses
6198 can be loaded with the @code{ld24} instruction), and assume all subroutines
6199 are reachable with the @code{bl} instruction.
6200 This is the default.
6202 The addressability of a particular object can be set with the
6203 @code{model} attribute.
6205 @item -mcode-model=medium
6206 @opindex mcode-model=medium
6207 Assume objects may be anywhere in the 32-bit address space (the compiler
6208 will generate @code{seth/add3} instructions to load their addresses), and
6209 assume all subroutines are reachable with the @code{bl} instruction.
6211 @item -mcode-model=large
6212 @opindex mcode-model=large
6213 Assume objects may be anywhere in the 32-bit address space (the compiler
6214 will generate @code{seth/add3} instructions to load their addresses), and
6215 assume subroutines may not be reachable with the @code{bl} instruction
6216 (the compiler will generate the much slower @code{seth/add3/jl}
6217 instruction sequence).
6220 @opindex msdata=none
6221 Disable use of the small data area. Variables will be put into
6222 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6223 @code{section} attribute has been specified).
6224 This is the default.
6226 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6227 Objects may be explicitly put in the small data area with the
6228 @code{section} attribute using one of these sections.
6231 @opindex msdata=sdata
6232 Put small global and static data in the small data area, but do not
6233 generate special code to reference them.
6237 Put small global and static data in the small data area, and generate
6238 special instructions to reference them.
6242 @cindex smaller data references
6243 Put global and static objects less than or equal to @var{num} bytes
6244 into the small data or bss sections instead of the normal data or bss
6245 sections. The default value of @var{num} is 8.
6246 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6247 for this option to have any effect.
6249 All modules should be compiled with the same @option{-G @var{num}} value.
6250 Compiling with different values of @var{num} may or may not work; if it
6251 doesn't the linker will give an error message---incorrect code will not be
6257 @subsection M88K Options
6258 @cindex M88k options
6260 These @samp{-m} options are defined for Motorola 88k architectures:
6265 Generate code that works well on both the m88100 and the
6270 Generate code that works best for the m88100, but that also
6275 Generate code that works best for the m88110, and may not run
6280 Obsolete option to be removed from the next revision.
6283 @item -midentify-revision
6284 @opindex midentify-revision
6285 @cindex identifying source, compiler (88k)
6286 Include an @code{ident} directive in the assembler output recording the
6287 source file name, compiler name and version, timestamp, and compilation
6290 @item -mno-underscores
6291 @opindex mno-underscores
6292 @cindex underscores, avoiding (88k)
6293 In assembler output, emit symbol names without adding an underscore
6294 character at the beginning of each name. The default is to use an
6295 underscore as prefix on each name.
6297 @item -mocs-debug-info
6298 @itemx -mno-ocs-debug-info
6299 @opindex mocs-debug-info
6300 @opindex mno-ocs-debug-info
6302 @cindex debugging, 88k OCS
6303 Include (or omit) additional debugging information (about registers used
6304 in each stack frame) as specified in the 88open Object Compatibility
6305 Standard, ``OCS''@. This extra information allows debugging of code that
6306 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6307 Delta 88 SVr3.2 is to include this information; other 88k configurations
6308 omit this information by default.
6310 @item -mocs-frame-position
6311 @opindex mocs-frame-position
6312 @cindex register positions in frame (88k)
6313 When emitting COFF debugging information for automatic variables and
6314 parameters stored on the stack, use the offset from the canonical frame
6315 address, which is the stack pointer (register 31) on entry to the
6316 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6317 @option{-mocs-frame-position}; other 88k configurations have the default
6318 @option{-mno-ocs-frame-position}.
6320 @item -mno-ocs-frame-position
6321 @opindex mno-ocs-frame-position
6322 @cindex register positions in frame (88k)
6323 When emitting COFF debugging information for automatic variables and
6324 parameters stored on the stack, use the offset from the frame pointer
6325 register (register 30). When this option is in effect, the frame
6326 pointer is not eliminated when debugging information is selected by the
6329 @item -moptimize-arg-area
6330 @opindex moptimize-arg-area
6331 @cindex arguments in frame (88k)
6332 Save space by reorganizing the stack frame. This option generates code
6333 that does not agree with the 88open specifications, but uses less
6336 @itemx -mno-optimize-arg-area
6337 @opindex mno-optimize-arg-area
6338 Do not reorganize the stack frame to save space. This is the default.
6339 The generated conforms to the specification, but uses more memory.
6341 @item -mshort-data-@var{num}
6342 @opindex mshort-data
6343 @cindex smaller data references (88k)
6344 @cindex r0-relative references (88k)
6345 Generate smaller data references by making them relative to @code{r0},
6346 which allows loading a value using a single instruction (rather than the
6347 usual two). You control which data references are affected by
6348 specifying @var{num} with this option. For example, if you specify
6349 @option{-mshort-data-512}, then the data references affected are those
6350 involving displacements of less than 512 bytes.
6351 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6354 @item -mserialize-volatile
6355 @opindex mserialize-volatile
6356 @itemx -mno-serialize-volatile
6357 @opindex mno-serialize-volatile
6358 @cindex sequential consistency on 88k
6359 Do, or don't, generate code to guarantee sequential consistency
6360 of volatile memory references. By default, consistency is
6363 The order of memory references made by the MC88110 processor does
6364 not always match the order of the instructions requesting those
6365 references. In particular, a load instruction may execute before
6366 a preceding store instruction. Such reordering violates
6367 sequential consistency of volatile memory references, when there
6368 are multiple processors. When consistency must be guaranteed,
6369 GCC generates special instructions, as needed, to force
6370 execution in the proper order.
6372 The MC88100 processor does not reorder memory references and so
6373 always provides sequential consistency. However, by default, GCC
6374 generates the special instructions to guarantee consistency
6375 even when you use @option{-m88100}, so that the code may be run on an
6376 MC88110 processor. If you intend to run your code only on the
6377 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6379 The extra code generated to guarantee consistency may affect the
6380 performance of your application. If you know that you can safely
6381 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6387 @cindex assembler syntax, 88k
6389 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6390 related to System V release 4 (SVr4). This controls the following:
6394 Which variant of the assembler syntax to emit.
6396 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6397 that is used on System V release 4.
6399 @option{-msvr4} makes GCC issue additional declaration directives used in
6403 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6404 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6405 other m88k configurations.
6407 @item -mversion-03.00
6408 @opindex mversion-03.00
6409 This option is obsolete, and is ignored.
6410 @c ??? which asm syntax better for GAS? option there too?
6412 @item -mno-check-zero-division
6413 @itemx -mcheck-zero-division
6414 @opindex mno-check-zero-division
6415 @opindex mcheck-zero-division
6416 @cindex zero division on 88k
6417 Do, or don't, generate code to guarantee that integer division by
6418 zero will be detected. By default, detection is guaranteed.
6420 Some models of the MC88100 processor fail to trap upon integer
6421 division by zero under certain conditions. By default, when
6422 compiling code that might be run on such a processor, GCC
6423 generates code that explicitly checks for zero-valued divisors
6424 and traps with exception number 503 when one is detected. Use of
6425 @option{-mno-check-zero-division} suppresses such checking for code
6426 generated to run on an MC88100 processor.
6428 GCC assumes that the MC88110 processor correctly detects all instances
6429 of integer division by zero. When @option{-m88110} is specified, no
6430 explicit checks for zero-valued divisors are generated, and both
6431 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6434 @item -muse-div-instruction
6435 @opindex muse-div-instruction
6436 @cindex divide instruction, 88k
6437 Use the div instruction for signed integer division on the
6438 MC88100 processor. By default, the div instruction is not used.
6440 On the MC88100 processor the signed integer division instruction
6441 div) traps to the operating system on a negative operand. The
6442 operating system transparently completes the operation, but at a
6443 large cost in execution time. By default, when compiling code
6444 that might be run on an MC88100 processor, GCC emulates signed
6445 integer division using the unsigned integer division instruction
6446 divu), thereby avoiding the large penalty of a trap to the
6447 operating system. Such emulation has its own, smaller, execution
6448 cost in both time and space. To the extent that your code's
6449 important signed integer division operations are performed on two
6450 nonnegative operands, it may be desirable to use the div
6451 instruction directly.
6453 On the MC88110 processor the div instruction (also known as the
6454 divs instruction) processes negative operands without trapping to
6455 the operating system. When @option{-m88110} is specified,
6456 @option{-muse-div-instruction} is ignored, and the div instruction is used
6457 for signed integer division.
6459 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6460 particular, the behavior of such a division with and without
6461 @option{-muse-div-instruction} may differ.
6463 @item -mtrap-large-shift
6464 @itemx -mhandle-large-shift
6465 @opindex mtrap-large-shift
6466 @opindex mhandle-large-shift
6467 @cindex bit shift overflow (88k)
6468 @cindex large bit shifts (88k)
6469 Include code to detect bit-shifts of more than 31 bits; respectively,
6470 trap such shifts or emit code to handle them properly. By default GCC
6471 makes no special provision for large bit shifts.
6473 @item -mwarn-passed-structs
6474 @opindex mwarn-passed-structs
6475 @cindex structure passing (88k)
6476 Warn when a function passes a struct as an argument or result.
6477 Structure-passing conventions have changed during the evolution of the C
6478 language, and are often the source of portability problems. By default,
6479 GCC issues no such warning.
6482 @c break page here to avoid unsightly interparagraph stretch.
6486 @node RS/6000 and PowerPC Options
6487 @subsection IBM RS/6000 and PowerPC Options
6488 @cindex RS/6000 and PowerPC Options
6489 @cindex IBM RS/6000 and PowerPC Options
6491 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6499 @itemx -mpowerpc-gpopt
6500 @itemx -mno-powerpc-gpopt
6501 @itemx -mpowerpc-gfxopt
6502 @itemx -mno-powerpc-gfxopt
6504 @itemx -mno-powerpc64
6510 @opindex mno-powerpc
6511 @opindex mpowerpc-gpopt
6512 @opindex mno-powerpc-gpopt
6513 @opindex mpowerpc-gfxopt
6514 @opindex mno-powerpc-gfxopt
6516 @opindex mno-powerpc64
6517 GCC supports two related instruction set architectures for the
6518 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6519 instructions supported by the @samp{rios} chip set used in the original
6520 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6521 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6522 the IBM 4xx microprocessors.
6524 Neither architecture is a subset of the other. However there is a
6525 large common subset of instructions supported by both. An MQ
6526 register is included in processors supporting the POWER architecture.
6528 You use these options to specify which instructions are available on the
6529 processor you are using. The default value of these options is
6530 determined when configuring GCC@. Specifying the
6531 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6532 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6533 rather than the options listed above.
6535 The @option{-mpower} option allows GCC to generate instructions that
6536 are found only in the POWER architecture and to use the MQ register.
6537 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6538 to generate instructions that are present in the POWER2 architecture but
6539 not the original POWER architecture.
6541 The @option{-mpowerpc} option allows GCC to generate instructions that
6542 are found only in the 32-bit subset of the PowerPC architecture.
6543 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6544 GCC to use the optional PowerPC architecture instructions in the
6545 General Purpose group, including floating-point square root. Specifying
6546 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6547 use the optional PowerPC architecture instructions in the Graphics
6548 group, including floating-point select.
6550 The @option{-mpowerpc64} option allows GCC to generate the additional
6551 64-bit instructions that are found in the full PowerPC64 architecture
6552 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6553 @option{-mno-powerpc64}.
6555 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6556 will use only the instructions in the common subset of both
6557 architectures plus some special AIX common-mode calls, and will not use
6558 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6559 permits GCC to use any instruction from either architecture and to
6560 allow use of the MQ register; specify this for the Motorola MPC601.
6562 @item -mnew-mnemonics
6563 @itemx -mold-mnemonics
6564 @opindex mnew-mnemonics
6565 @opindex mold-mnemonics
6566 Select which mnemonics to use in the generated assembler code. With
6567 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6568 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6569 assembler mnemonics defined for the POWER architecture. Instructions
6570 defined in only one architecture have only one mnemonic; GCC uses that
6571 mnemonic irrespective of which of these options is specified.
6573 GCC defaults to the mnemonics appropriate for the architecture in
6574 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6575 value of these option. Unless you are building a cross-compiler, you
6576 should normally not specify either @option{-mnew-mnemonics} or
6577 @option{-mold-mnemonics}, but should instead accept the default.
6579 @item -mcpu=@var{cpu_type}
6581 Set architecture type, register usage, choice of mnemonics, and
6582 instruction scheduling parameters for machine type @var{cpu_type}.
6583 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6584 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6585 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6586 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6587 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6588 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6590 @option{-mcpu=common} selects a completely generic processor. Code
6591 generated under this option will run on any POWER or PowerPC processor.
6592 GCC will use only the instructions in the common subset of both
6593 architectures, and will not use the MQ register. GCC assumes a generic
6594 processor model for scheduling purposes.
6596 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6597 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6598 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6599 types, with an appropriate, generic processor model assumed for
6600 scheduling purposes.
6602 The other options specify a specific processor. Code generated under
6603 those options will run best on that processor, and may not run at all on
6606 The @option{-mcpu} options automatically enable or disable other
6607 @option{-m} options as follows:
6611 @option{-mno-power}, @option{-mno-powerc}
6618 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6633 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6636 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6641 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6644 @item -mtune=@var{cpu_type}
6646 Set the instruction scheduling parameters for machine type
6647 @var{cpu_type}, but do not set the architecture type, register usage, or
6648 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6649 values for @var{cpu_type} are used for @option{-mtune} as for
6650 @option{-mcpu}. If both are specified, the code generated will use the
6651 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6652 scheduling parameters set by @option{-mtune}.
6657 @opindex mno-altivec
6658 These switches enable or disable the use of built-in functions that
6659 allow access to the AltiVec instruction set. You may also need to set
6660 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6665 Extend the current ABI with SPE ABI extensions. This does not change
6666 the default ABI, instead it adds the SPE ABI extensions to the current
6670 @opindex mabi=no-spe
6671 Disable Booke SPE ABI extensions for the current ABI.
6673 @item -misel=@var{yes/no}
6676 This switch enables or disables the generation of ISEL instructions.
6679 @itemx -mno-fp-in-toc
6680 @itemx -mno-sum-in-toc
6681 @itemx -mminimal-toc
6683 @opindex mno-fp-in-toc
6684 @opindex mno-sum-in-toc
6685 @opindex mminimal-toc
6686 Modify generation of the TOC (Table Of Contents), which is created for
6687 every executable file. The @option{-mfull-toc} option is selected by
6688 default. In that case, GCC will allocate at least one TOC entry for
6689 each unique non-automatic variable reference in your program. GCC
6690 will also place floating-point constants in the TOC@. However, only
6691 16,384 entries are available in the TOC@.
6693 If you receive a linker error message that saying you have overflowed
6694 the available TOC space, you can reduce the amount of TOC space used
6695 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6696 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6697 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6698 generate code to calculate the sum of an address and a constant at
6699 run-time instead of putting that sum into the TOC@. You may specify one
6700 or both of these options. Each causes GCC to produce very slightly
6701 slower and larger code at the expense of conserving TOC space.
6703 If you still run out of space in the TOC even when you specify both of
6704 these options, specify @option{-mminimal-toc} instead. This option causes
6705 GCC to make only one TOC entry for every file. When you specify this
6706 option, GCC will produce code that is slower and larger but which
6707 uses extremely little TOC space. You may wish to use this option
6708 only on files that contain less frequently executed code.
6714 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6715 @code{long} type, and the infrastructure needed to support them.
6716 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6717 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6718 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6723 @opindex mno-xl-call
6724 On AIX, pass floating-point arguments to prototyped functions beyond the
6725 register save area (RSA) on the stack in addition to argument FPRs. The
6726 AIX calling convention was extended but not initially documented to
6727 handle an obscure K&R C case of calling a function that takes the
6728 address of its arguments with fewer arguments than declared. AIX XL
6729 compilers access floating point arguments which do not fit in the
6730 RSA from the stack when a subroutine is compiled without
6731 optimization. Because always storing floating-point arguments on the
6732 stack is inefficient and rarely needed, this option is not enabled by
6733 default and only is necessary when calling subroutines compiled by AIX
6734 XL compilers without optimization.
6738 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6739 application written to use message passing with special startup code to
6740 enable the application to run. The system must have PE installed in the
6741 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6742 must be overridden with the @option{-specs=} option to specify the
6743 appropriate directory location. The Parallel Environment does not
6744 support threads, so the @option{-mpe} option and the @option{-pthread}
6745 option are incompatible.
6749 @opindex msoft-float
6750 @opindex mhard-float
6751 Generate code that does not use (uses) the floating-point register set.
6752 Software floating point emulation is provided if you use the
6753 @option{-msoft-float} option, and pass the option to GCC when linking.
6756 @itemx -mno-multiple
6758 @opindex mno-multiple
6759 Generate code that uses (does not use) the load multiple word
6760 instructions and the store multiple word instructions. These
6761 instructions are generated by default on POWER systems, and not
6762 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6763 endian PowerPC systems, since those instructions do not work when the
6764 processor is in little endian mode. The exceptions are PPC740 and
6765 PPC750 which permit the instructions usage in little endian mode.
6771 Generate code that uses (does not use) the load string instructions
6772 and the store string word instructions to save multiple registers and
6773 do small block moves. These instructions are generated by default on
6774 POWER systems, and not generated on PowerPC systems. Do not use
6775 @option{-mstring} on little endian PowerPC systems, since those
6776 instructions do not work when the processor is in little endian mode.
6777 The exceptions are PPC740 and PPC750 which permit the instructions
6778 usage in little endian mode.
6784 Generate code that uses (does not use) the load or store instructions
6785 that update the base register to the address of the calculated memory
6786 location. These instructions are generated by default. If you use
6787 @option{-mno-update}, there is a small window between the time that the
6788 stack pointer is updated and the address of the previous frame is
6789 stored, which means code that walks the stack frame across interrupts or
6790 signals may get corrupted data.
6793 @itemx -mno-fused-madd
6794 @opindex mfused-madd
6795 @opindex mno-fused-madd
6796 Generate code that uses (does not use) the floating point multiply and
6797 accumulate instructions. These instructions are generated by default if
6798 hardware floating is used.
6800 @item -mno-bit-align
6802 @opindex mno-bit-align
6804 On System V.4 and embedded PowerPC systems do not (do) force structures
6805 and unions that contain bit-fields to be aligned to the base type of the
6808 For example, by default a structure containing nothing but 8
6809 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6810 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6811 the structure would be aligned to a 1 byte boundary and be one byte in
6814 @item -mno-strict-align
6815 @itemx -mstrict-align
6816 @opindex mno-strict-align
6817 @opindex mstrict-align
6818 On System V.4 and embedded PowerPC systems do not (do) assume that
6819 unaligned memory references will be handled by the system.
6822 @itemx -mno-relocatable
6823 @opindex mrelocatable
6824 @opindex mno-relocatable
6825 On embedded PowerPC systems generate code that allows (does not allow)
6826 the program to be relocated to a different address at runtime. If you
6827 use @option{-mrelocatable} on any module, all objects linked together must
6828 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6830 @item -mrelocatable-lib
6831 @itemx -mno-relocatable-lib
6832 @opindex mrelocatable-lib
6833 @opindex mno-relocatable-lib
6834 On embedded PowerPC systems generate code that allows (does not allow)
6835 the program to be relocated to a different address at runtime. Modules
6836 compiled with @option{-mrelocatable-lib} can be linked with either modules
6837 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6838 with modules compiled with the @option{-mrelocatable} options.
6844 On System V.4 and embedded PowerPC systems do not (do) assume that
6845 register 2 contains a pointer to a global area pointing to the addresses
6846 used in the program.
6849 @itemx -mlittle-endian
6851 @opindex mlittle-endian
6852 On System V.4 and embedded PowerPC systems compile code for the
6853 processor in little endian mode. The @option{-mlittle-endian} option is
6854 the same as @option{-mlittle}.
6859 @opindex mbig-endian
6860 On System V.4 and embedded PowerPC systems compile code for the
6861 processor in big endian mode. The @option{-mbig-endian} option is
6862 the same as @option{-mbig}.
6866 On System V.4 and embedded PowerPC systems compile code using calling
6867 conventions that adheres to the March 1995 draft of the System V
6868 Application Binary Interface, PowerPC processor supplement. This is the
6869 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6871 @item -mcall-sysv-eabi
6872 @opindex mcall-sysv-eabi
6873 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6875 @item -mcall-sysv-noeabi
6876 @opindex mcall-sysv-noeabi
6877 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6881 On System V.4 and embedded PowerPC systems compile code using calling
6882 conventions that are similar to those used on AIX@. This is the
6883 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6885 @item -mcall-solaris
6886 @opindex mcall-solaris
6887 On System V.4 and embedded PowerPC systems compile code for the Solaris
6891 @opindex mcall-linux
6892 On System V.4 and embedded PowerPC systems compile code for the
6893 Linux-based GNU system.
6897 On System V.4 and embedded PowerPC systems compile code for the
6898 Hurd-based GNU system.
6901 @opindex mcall-netbsd
6902 On System V.4 and embedded PowerPC systems compile code for the
6903 NetBSD operating system.
6905 @item -maix-struct-return
6906 @opindex maix-struct-return
6907 Return all structures in memory (as specified by the AIX ABI)@.
6909 @item -msvr4-struct-return
6910 @opindex msvr4-struct-return
6911 Return structures smaller than 8 bytes in registers (as specified by the
6915 @opindex mabi=altivec
6916 Extend the current ABI with AltiVec ABI extensions. This does not
6917 change the default ABI, instead it adds the AltiVec ABI extensions to
6920 @item -mabi=no-altivec
6921 @opindex mabi=no-altivec
6922 Disable AltiVec ABI extensions for the current ABI.
6925 @itemx -mno-prototype
6927 @opindex mno-prototype
6928 On System V.4 and embedded PowerPC systems assume that all calls to
6929 variable argument functions are properly prototyped. Otherwise, the
6930 compiler must insert an instruction before every non prototyped call to
6931 set or clear bit 6 of the condition code register (@var{CR}) to
6932 indicate whether floating point values were passed in the floating point
6933 registers in case the function takes a variable arguments. With
6934 @option{-mprototype}, only calls to prototyped variable argument functions
6935 will set or clear the bit.
6939 On embedded PowerPC systems, assume that the startup module is called
6940 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6941 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6946 On embedded PowerPC systems, assume that the startup module is called
6947 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6952 On embedded PowerPC systems, assume that the startup module is called
6953 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6957 @opindex myellowknife
6958 On embedded PowerPC systems, assume that the startup module is called
6959 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6964 On System V.4 and embedded PowerPC systems, specify that you are
6965 compiling for a VxWorks system.
6969 Specify that you are compiling for the WindISS simulation environment.
6973 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
6974 header to indicate that @samp{eabi} extended relocations are used.
6980 On System V.4 and embedded PowerPC systems do (do not) adhere to the
6981 Embedded Applications Binary Interface (eabi) which is a set of
6982 modifications to the System V.4 specifications. Selecting @option{-meabi}
6983 means that the stack is aligned to an 8 byte boundary, a function
6984 @code{__eabi} is called to from @code{main} to set up the eabi
6985 environment, and the @option{-msdata} option can use both @code{r2} and
6986 @code{r13} to point to two separate small data areas. Selecting
6987 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
6988 do not call an initialization function from @code{main}, and the
6989 @option{-msdata} option will only use @code{r13} to point to a single
6990 small data area. The @option{-meabi} option is on by default if you
6991 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
6994 @opindex msdata=eabi
6995 On System V.4 and embedded PowerPC systems, put small initialized
6996 @code{const} global and static data in the @samp{.sdata2} section, which
6997 is pointed to by register @code{r2}. Put small initialized
6998 non-@code{const} global and static data in the @samp{.sdata} section,
6999 which is pointed to by register @code{r13}. Put small uninitialized
7000 global and static data in the @samp{.sbss} section, which is adjacent to
7001 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7002 incompatible with the @option{-mrelocatable} option. The
7003 @option{-msdata=eabi} option also sets the @option{-memb} option.
7006 @opindex msdata=sysv
7007 On System V.4 and embedded PowerPC systems, put small global and static
7008 data in the @samp{.sdata} section, which is pointed to by register
7009 @code{r13}. Put small uninitialized global and static data in the
7010 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7011 The @option{-msdata=sysv} option is incompatible with the
7012 @option{-mrelocatable} option.
7014 @item -msdata=default
7016 @opindex msdata=default
7018 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7019 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7020 same as @option{-msdata=sysv}.
7023 @opindex msdata-data
7024 On System V.4 and embedded PowerPC systems, put small global and static
7025 data in the @samp{.sdata} section. Put small uninitialized global and
7026 static data in the @samp{.sbss} section. Do not use register @code{r13}
7027 to address small data however. This is the default behavior unless
7028 other @option{-msdata} options are used.
7032 @opindex msdata=none
7034 On embedded PowerPC systems, put all initialized global and static data
7035 in the @samp{.data} section, and all uninitialized data in the
7036 @samp{.bss} section.
7040 @cindex smaller data references (PowerPC)
7041 @cindex .sdata/.sdata2 references (PowerPC)
7042 On embedded PowerPC systems, put global and static items less than or
7043 equal to @var{num} bytes into the small data or bss sections instead of
7044 the normal data or bss section. By default, @var{num} is 8. The
7045 @option{-G @var{num}} switch is also passed to the linker.
7046 All modules should be compiled with the same @option{-G @var{num}} value.
7049 @itemx -mno-regnames
7051 @opindex mno-regnames
7052 On System V.4 and embedded PowerPC systems do (do not) emit register
7053 names in the assembly language output using symbolic forms.
7056 @itemx -mno-longcall
7058 @opindex mno-longcall
7059 Default to making all function calls via pointers, so that functions
7060 which reside further than 64 megabytes (67,108,864 bytes) from the
7061 current location can be called. This setting can be overridden by the
7062 @code{shortcall} function attribute, or by @code{#pragma longcall(0)}.
7064 Some linkers are capable of detecting out-of-range calls and generating
7065 glue code on the fly. On these systems, long calls are unnecessary and
7066 generate slower code. As of this writing, the AIX linker can do this,
7067 as can the GNU linker for PowerPC/64. It is planned to add this feature
7068 to the GNU linker for 32-bit PowerPC systems as well.
7070 In the future, we may cause GCC to ignore all longcall specifications
7071 when the linker is known to generate glue.
7075 Adds support for multithreading with the @dfn{pthreads} library.
7076 This option sets flags for both the preprocessor and linker.
7081 @subsection IBM RT Options
7083 @cindex IBM RT options
7085 These @samp{-m} options are defined for the IBM RT PC:
7089 @opindex min-line-mul
7090 Use an in-line code sequence for integer multiplies. This is the
7093 @item -mcall-lib-mul
7094 @opindex mcall-lib-mul
7095 Call @code{lmul$$} for integer multiples.
7097 @item -mfull-fp-blocks
7098 @opindex mfull-fp-blocks
7099 Generate full-size floating point data blocks, including the minimum
7100 amount of scratch space recommended by IBM@. This is the default.
7102 @item -mminimum-fp-blocks
7103 @opindex mminimum-fp-blocks
7104 Do not include extra scratch space in floating point data blocks. This
7105 results in smaller code, but slower execution, since scratch space must
7106 be allocated dynamically.
7108 @cindex @file{stdarg.h} and RT PC
7109 @item -mfp-arg-in-fpregs
7110 @opindex mfp-arg-in-fpregs
7111 Use a calling sequence incompatible with the IBM calling convention in
7112 which floating point arguments are passed in floating point registers.
7113 Note that @code{stdarg.h} will not work with floating point operands
7114 if this option is specified.
7116 @item -mfp-arg-in-gregs
7117 @opindex mfp-arg-in-gregs
7118 Use the normal calling convention for floating point arguments. This is
7121 @item -mhc-struct-return
7122 @opindex mhc-struct-return
7123 Return structures of more than one word in memory, rather than in a
7124 register. This provides compatibility with the MetaWare HighC (hc)
7125 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7126 with the Portable C Compiler (pcc).
7128 @item -mnohc-struct-return
7129 @opindex mnohc-struct-return
7130 Return some structures of more than one word in registers, when
7131 convenient. This is the default. For compatibility with the
7132 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7133 option @option{-mhc-struct-return}.
7137 @subsection MIPS Options
7138 @cindex MIPS options
7140 These @samp{-m} options are defined for the MIPS family of computers:
7144 @item -march=@var{arch}
7146 Generate code that will run on @var{arch}, which can be the name of a
7147 generic MIPS ISA, or the name of a particular processor. The ISA names
7148 are: @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4}, @samp{mips32}
7149 and @samp{mips64}. The processor names are: @samp{r2000},
7150 @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{vr4100}, @samp{vr4300},
7151 @samp{r4400}, @samp{r4600}, @samp{r4650}, @samp{vr5000}, @samp{r6000},
7152 @samp{r8000}, @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
7153 @samp{orion}, and @samp{sb1}. The special value @samp{from-abi} selects the
7154 most compatible architecture for the selected ABI (that is,
7155 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
7157 In processor names, a final @samp{000} can be abbreviated as @samp{k}
7158 (for example, @samp{-march=r2k}). Prefixes are optional, and
7159 @samp{vr} may be written @samp{r}.
7161 GCC defines two macros based on the value of this option. The first
7162 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
7163 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
7164 where @var{foo} is the capitialized value of @samp{_MIPS_ARCH}@.
7165 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
7166 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
7168 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
7169 above. In other words, it will have the full prefix and will not
7170 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
7171 the macro names the resolved architecture (either @samp{"mips1"} or
7172 @samp{"mips3"}). It names the default architecture when no
7173 @option{-march} option is given.
7175 @item -mtune=@var{arch}
7177 Optimize for @var{arch}. Among other things, this option controls
7178 the way instructions are scheduled, and the perceived cost of arithmetic
7179 operations. The list of @var{arch} values is the same as for
7182 When this option is not used, GCC will optimize for the processor
7183 specified by @option{-march}. By using @option{-march} and
7184 @option{-mtune} together, it is possible to generate code that will
7185 run on a family of processors, but optimize the code for one
7186 particular member of that family.
7188 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
7189 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
7190 @samp{-march} ones described above.
7194 Equivalent to @samp{-march=mips1}.
7198 Equivalent to @samp{-march=mips2}.
7202 Equivalent to @samp{-march=mips3}.
7206 Equivalent to @samp{-march=mips4}.
7210 Equivalent to @samp{-march=mips32}.
7214 Equivalent to @samp{-march=mips64}.
7217 @itemx -mno-fused-madd
7218 @opindex mfused-madd
7219 @opindex mno-fused-madd
7220 Generate code that uses (does not use) the floating point multiply and
7221 accumulate instructions, when they are available. These instructions
7222 are generated by default if they are available, but this may be
7223 undesirable if the extra precision causes problems or on certain chips
7224 in the mode where denormals are rounded to zero where denormals
7225 generated by multiply and accumulate instructions cause exceptions
7230 Assume that floating point registers are 32 bits wide.
7234 Assume that floating point registers are 64 bits wide.
7238 Assume that general purpose registers are 32 bits wide.
7242 Assume that general purpose registers are 64 bits wide.
7246 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7247 explanation of the default, and the width of pointers.
7251 Force long types to be 64 bits wide. See @option{-mlong32} for an
7252 explanation of the default, and the width of pointers.
7256 Force long, int, and pointer types to be 32 bits wide.
7258 The default size of ints, longs and pointers depends on the ABI@. All
7259 the supported ABIs use 32-bit ints. The n64 ABI uses 64-bit longs, as
7260 does the 64-bit Cygnus EABI; the others use 32-bit longs. Pointers
7261 are the same size as longs, or the same size as integer registers,
7262 whichever is smaller.
7276 Generate code for the given ABI@.
7278 Note that there are two embedded ABIs: @option{-mabi=eabi}
7279 selects the one defined by Cygnus while @option{-meabi=meabi}
7280 selects the one defined by MIPS@. Both these ABIs have
7281 32-bit and 64-bit variants. Normally, GCC will generate
7282 64-bit code when you select a 64-bit architecture, but you
7283 can use @option{-mgp32} to get 32-bit code instead.
7287 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7288 add normal debug information. This is the default for all
7289 platforms except for the OSF/1 reference platform, using the OSF/rose
7290 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7291 switches are used, the @file{mips-tfile} program will encapsulate the
7292 stabs within MIPS ECOFF@.
7296 Generate code for the GNU assembler. This is the default on the OSF/1
7297 reference platform, using the OSF/rose object format. Also, this is
7298 the default if the configure option @option{--with-gnu-as} is used.
7300 @item -msplit-addresses
7301 @itemx -mno-split-addresses
7302 @opindex msplit-addresses
7303 @opindex mno-split-addresses
7304 Generate code to load the high and low parts of address constants separately.
7305 This allows GCC to optimize away redundant loads of the high order
7306 bits of addresses. This optimization requires GNU as and GNU ld.
7307 This optimization is enabled by default for some embedded targets where
7308 GNU as and GNU ld are standard.
7314 The @option{-mrnames} switch says to output code using the MIPS software
7315 names for the registers, instead of the hardware names (ie, @var{a0}
7316 instead of @var{$4}). The only known assembler that supports this option
7317 is the Algorithmics assembler.
7323 The @option{-mgpopt} switch says to write all of the data declarations
7324 before the instructions in the text section, this allows the MIPS
7325 assembler to generate one word memory references instead of using two
7326 words for short global or static data items. This is on by default if
7327 optimization is selected.
7333 For each non-inline function processed, the @option{-mstats} switch
7334 causes the compiler to emit one line to the standard error file to
7335 print statistics about the program (number of registers saved, stack
7342 The @option{-mmemcpy} switch makes all block moves call the appropriate
7343 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7344 generating inline code.
7347 @itemx -mno-mips-tfile
7348 @opindex mmips-tfile
7349 @opindex mno-mips-tfile
7350 The @option{-mno-mips-tfile} switch causes the compiler not
7351 postprocess the object file with the @file{mips-tfile} program,
7352 after the MIPS assembler has generated it to add debug support. If
7353 @file{mips-tfile} is not run, then no local variables will be
7354 available to the debugger. In addition, @file{stage2} and
7355 @file{stage3} objects will have the temporary file names passed to the
7356 assembler embedded in the object file, which means the objects will
7357 not compare the same. The @option{-mno-mips-tfile} switch should only
7358 be used when there are bugs in the @file{mips-tfile} program that
7359 prevents compilation.
7362 @opindex msoft-float
7363 Generate output containing library calls for floating point.
7364 @strong{Warning:} the requisite libraries are not part of GCC@.
7365 Normally the facilities of the machine's usual C compiler are used, but
7366 this can't be done directly in cross-compilation. You must make your
7367 own arrangements to provide suitable library functions for
7371 @opindex mhard-float
7372 Generate output containing floating point instructions. This is the
7373 default if you use the unmodified sources.
7376 @itemx -mno-abicalls
7378 @opindex mno-abicalls
7379 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7380 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7381 position independent code.
7384 @itemx -mno-long-calls
7385 @opindex mlong-calls
7386 @opindex mno-long-calls
7387 Do all calls with the @samp{JALR} instruction, which requires
7388 loading up a function's address into a register before the call.
7389 You need to use this switch, if you call outside of the current
7390 512 megabyte segment to functions that are not through pointers.
7393 @itemx -mno-half-pic
7395 @opindex mno-half-pic
7396 Put pointers to extern references into the data section and load them
7397 up, rather than put the references in the text section.
7399 @item -membedded-pic
7400 @itemx -mno-embedded-pic
7401 @opindex membedded-pic
7402 @opindex mno-embedded-pic
7403 Generate PIC code suitable for some embedded systems. All calls are
7404 made using PC relative address, and all data is addressed using the $gp
7405 register. No more than 65536 bytes of global data may be used. This
7406 requires GNU as and GNU ld which do most of the work. This currently
7407 only works on targets which use ECOFF; it does not work with ELF@.
7409 @item -membedded-data
7410 @itemx -mno-embedded-data
7411 @opindex membedded-data
7412 @opindex mno-embedded-data
7413 Allocate variables to the read-only data section first if possible, then
7414 next in the small data section if possible, otherwise in data. This gives
7415 slightly slower code than the default, but reduces the amount of RAM required
7416 when executing, and thus may be preferred for some embedded systems.
7418 @item -muninit-const-in-rodata
7419 @itemx -mno-uninit-const-in-rodata
7420 @opindex muninit-const-in-rodata
7421 @opindex mno-uninit-const-in-rodata
7422 When used together with @option{-membedded-data}, it will always store uninitialized
7423 const variables in the read-only data section.
7425 @item -msingle-float
7426 @itemx -mdouble-float
7427 @opindex msingle-float
7428 @opindex mdouble-float
7429 The @option{-msingle-float} switch tells gcc to assume that the floating
7430 point coprocessor only supports single precision operations, as on the
7431 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7432 double precision operations. This is the default.
7438 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7439 as on the @samp{r4650} chip.
7443 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7444 @option{-mcpu=r4650}.
7450 Enable 16-bit instructions.
7454 Use the entry and exit pseudo ops. This option can only be used with
7459 Compile code for the processor in little endian mode.
7460 The requisite libraries are assumed to exist.
7464 Compile code for the processor in big endian mode.
7465 The requisite libraries are assumed to exist.
7469 @cindex smaller data references (MIPS)
7470 @cindex gp-relative references (MIPS)
7471 Put global and static items less than or equal to @var{num} bytes into
7472 the small data or bss sections instead of the normal data or bss
7473 section. This allows the assembler to emit one word memory reference
7474 instructions based on the global pointer (@var{gp} or @var{$28}),
7475 instead of the normal two words used. By default, @var{num} is 8 when
7476 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7477 @option{-G @var{num}} switch is also passed to the assembler and linker.
7478 All modules should be compiled with the same @option{-G @var{num}}
7483 Tell the MIPS assembler to not run its preprocessor over user
7484 assembler files (with a @samp{.s} suffix) when assembling them.
7488 Pass an option to gas which will cause nops to be inserted if
7489 the read of the destination register of an mfhi or mflo instruction
7490 occurs in the following two instructions.
7494 Do not include the default crt0.
7496 @item -mflush-func=@var{func}
7497 @itemx -mno-flush-func
7498 @opindex mflush-func
7499 Specifies the function to call to flush the I and D caches, or to not
7500 call any such function. If called, the function must take the same
7501 arguments as the common @code{_flush_func()}, that is, the address of the
7502 memory range for which the cache is being flushed, the size of the
7503 memory range, and the number 3 (to flush both caches). The default
7504 depends on the target gcc was configured for, but commonly is either
7505 @samp{_flush_func} or @samp{__cpu_flush}.
7507 @item -mbranch-likely
7508 @itemx -mno-branch-likely
7509 @opindex mbranch-likely
7510 @opindex mno-branch-likely
7511 Enable or disable use of Branch Likely instructions, regardless of the
7512 default for the selected architecture. By default, Branch Likely
7513 instructions may be generated if they are supported by the selected
7514 architecture. An exception is for the MIPS32 and MIPS64 architectures
7515 and processors which implement those architectures; for those, Branch
7516 Likely instructions will not be generated by default because the MIPS32
7517 and MIPS64 architectures specifically deprecate their use.
7520 @node i386 and x86-64 Options
7521 @subsection Intel 386 and AMD x86-64 Options
7522 @cindex i386 Options
7523 @cindex x86-64 Options
7524 @cindex Intel 386 Options
7525 @cindex AMD x86-64 Options
7527 These @samp{-m} options are defined for the i386 and x86-64 family of
7531 @item -mcpu=@var{cpu-type}
7533 Tune to @var{cpu-type} everything applicable about the generated code, except
7534 for the ABI and the set of available instructions. The choices for
7535 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7536 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7537 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7538 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp},
7539 @samp{athlon-mp}, @samp{winchip-c6}, @samp{winchip2} and @samp{c3}.
7541 While picking a specific @var{cpu-type} will schedule things appropriately
7542 for that particular chip, the compiler will not generate any code that
7543 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7544 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7545 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7546 AMD chips as opposed to the Intel ones.
7548 @item -march=@var{cpu-type}
7550 Generate instructions for the machine type @var{cpu-type}. The choices
7551 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7552 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7561 @opindex mpentiumpro
7562 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7563 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7564 These synonyms are deprecated.
7566 @item -mfpmath=@var{unit}
7568 generate floating point arithmetics for selected unit @var{unit}. the choices
7573 Use the standard 387 floating point coprocessor present majority of chips and
7574 emulated otherwise. Code compiled with this option will run almost everywhere.
7575 The temporary results are computed in 80bit precesion instead of precision
7576 specified by the type resulting in slightly different results compared to most
7577 of other chips. See @option{-ffloat-store} for more detailed description.
7579 This is the default choice for i386 compiler.
7582 Use scalar floating point instructions present in the SSE instruction set.
7583 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7584 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7585 instruction set supports only single precision arithmetics, thus the double and
7586 extended precision arithmetics is still done using 387. Later version, present
7587 only in Pentium4 and the future AMD x86-64 chips supports double precision
7590 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7591 @option{-msse2} switches to enable SSE extensions and make this option
7592 effective. For x86-64 compiler, these extensions are enabled by default.
7594 The resulting code should be considerably faster in majority of cases and avoid
7595 the numerical instability problems of 387 code, but may break some existing
7596 code that expects temporaries to be 80bit.
7598 This is the default choice for x86-64 compiler.
7601 Attempt to utilize both instruction sets at once. This effectivly double the
7602 amount of available registers and on chips with separate execution units for
7603 387 and SSE the execution resources too. Use this option with care, as it is
7604 still experimental, because gcc register allocator does not model separate
7605 functional units well resulting in instable performance.
7608 @item -masm=@var{dialect}
7609 @opindex masm=@var{dialect}
7610 Output asm instructions using selected @var{dialect}. Supported choices are
7611 @samp{intel} or @samp{att} (the default one).
7616 @opindex mno-ieee-fp
7617 Control whether or not the compiler uses IEEE floating point
7618 comparisons. These handle correctly the case where the result of a
7619 comparison is unordered.
7622 @opindex msoft-float
7623 Generate output containing library calls for floating point.
7624 @strong{Warning:} the requisite libraries are not part of GCC@.
7625 Normally the facilities of the machine's usual C compiler are used, but
7626 this can't be done directly in cross-compilation. You must make your
7627 own arrangements to provide suitable library functions for
7630 On machines where a function returns floating point results in the 80387
7631 register stack, some floating point opcodes may be emitted even if
7632 @option{-msoft-float} is used.
7634 @item -mno-fp-ret-in-387
7635 @opindex mno-fp-ret-in-387
7636 Do not use the FPU registers for return values of functions.
7638 The usual calling convention has functions return values of types
7639 @code{float} and @code{double} in an FPU register, even if there
7640 is no FPU@. The idea is that the operating system should emulate
7643 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7644 in ordinary CPU registers instead.
7646 @item -mno-fancy-math-387
7647 @opindex mno-fancy-math-387
7648 Some 387 emulators do not support the @code{sin}, @code{cos} and
7649 @code{sqrt} instructions for the 387. Specify this option to avoid
7650 generating those instructions. This option is the default on FreeBSD,
7651 OpenBSD and NetBSD@. This option is overridden when @option{-march}
7652 indicates that the target cpu will always have an FPU and so the
7653 instruction will not need emulation. As of revision 2.6.1, these
7654 instructions are not generated unless you also use the
7655 @option{-funsafe-math-optimizations} switch.
7657 @item -malign-double
7658 @itemx -mno-align-double
7659 @opindex malign-double
7660 @opindex mno-align-double
7661 Control whether GCC aligns @code{double}, @code{long double}, and
7662 @code{long long} variables on a two word boundary or a one word
7663 boundary. Aligning @code{double} variables on a two word boundary will
7664 produce code that runs somewhat faster on a @samp{Pentium} at the
7665 expense of more memory.
7667 @strong{Warning:} if you use the @samp{-malign-double} switch,
7668 structures containing the above types will be aligned differently than
7669 the published application binary interface specifications for the 386.
7671 @item -m128bit-long-double
7672 @opindex m128bit-long-double
7673 Control the size of @code{long double} type. i386 application binary interface
7674 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7675 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7676 impossible to reach with 12 byte long doubles in the array accesses.
7678 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7679 structures and arrays containing @code{long double} will change their size as
7680 well as function calling convention for function taking @code{long double}
7683 @item -m96bit-long-double
7684 @opindex m96bit-long-double
7685 Set the size of @code{long double} to 96 bits as required by the i386
7686 application binary interface. This is the default.
7689 @itemx -mno-svr3-shlib
7690 @opindex msvr3-shlib
7691 @opindex mno-svr3-shlib
7692 Control whether GCC places uninitialized local variables into the
7693 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7694 into @code{bss}. These options are meaningful only on System V Release 3.
7698 Use a different function-calling convention, in which functions that
7699 take a fixed number of arguments return with the @code{ret} @var{num}
7700 instruction, which pops their arguments while returning. This saves one
7701 instruction in the caller since there is no need to pop the arguments
7704 You can specify that an individual function is called with this calling
7705 sequence with the function attribute @samp{stdcall}. You can also
7706 override the @option{-mrtd} option by using the function attribute
7707 @samp{cdecl}. @xref{Function Attributes}.
7709 @strong{Warning:} this calling convention is incompatible with the one
7710 normally used on Unix, so you cannot use it if you need to call
7711 libraries compiled with the Unix compiler.
7713 Also, you must provide function prototypes for all functions that
7714 take variable numbers of arguments (including @code{printf});
7715 otherwise incorrect code will be generated for calls to those
7718 In addition, seriously incorrect code will result if you call a
7719 function with too many arguments. (Normally, extra arguments are
7720 harmlessly ignored.)
7722 @item -mregparm=@var{num}
7724 Control how many registers are used to pass integer arguments. By
7725 default, no registers are used to pass arguments, and at most 3
7726 registers can be used. You can control this behavior for a specific
7727 function by using the function attribute @samp{regparm}.
7728 @xref{Function Attributes}.
7730 @strong{Warning:} if you use this switch, and
7731 @var{num} is nonzero, then you must build all modules with the same
7732 value, including any libraries. This includes the system libraries and
7735 @item -mpreferred-stack-boundary=@var{num}
7736 @opindex mpreferred-stack-boundary
7737 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7738 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7739 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7740 size (@option{-Os}), in which case the default is the minimum correct
7741 alignment (4 bytes for x86, and 8 bytes for x86-64).
7743 On Pentium and PentiumPro, @code{double} and @code{long double} values
7744 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7745 suffer significant run time performance penalties. On Pentium III, the
7746 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7747 penalties if it is not 16 byte aligned.
7749 To ensure proper alignment of this values on the stack, the stack boundary
7750 must be as aligned as that required by any value stored on the stack.
7751 Further, every function must be generated such that it keeps the stack
7752 aligned. Thus calling a function compiled with a higher preferred
7753 stack boundary from a function compiled with a lower preferred stack
7754 boundary will most likely misalign the stack. It is recommended that
7755 libraries that use callbacks always use the default setting.
7757 This extra alignment does consume extra stack space, and generally
7758 increases code size. Code that is sensitive to stack space usage, such
7759 as embedded systems and operating system kernels, may want to reduce the
7760 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7776 These switches enable or disable the use of built-in functions that allow
7777 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7779 @xref{X86 Built-in Functions}, for details of the functions enabled
7780 and disabled by these switches.
7783 @itemx -mno-push-args
7785 @opindex mno-push-args
7786 Use PUSH operations to store outgoing parameters. This method is shorter
7787 and usually equally fast as method using SUB/MOV operations and is enabled
7788 by default. In some cases disabling it may improve performance because of
7789 improved scheduling and reduced dependencies.
7791 @item -maccumulate-outgoing-args
7792 @opindex maccumulate-outgoing-args
7793 If enabled, the maximum amount of space required for outgoing arguments will be
7794 computed in the function prologue. This is faster on most modern CPUs
7795 because of reduced dependencies, improved scheduling and reduced stack usage
7796 when preferred stack boundary is not equal to 2. The drawback is a notable
7797 increase in code size. This switch implies @option{-mno-push-args}.
7801 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7802 on thread-safe exception handling must compile and link all code with the
7803 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7804 @option{-D_MT}; when linking, it links in a special thread helper library
7805 @option{-lmingwthrd} which cleans up per thread exception handling data.
7807 @item -mno-align-stringops
7808 @opindex mno-align-stringops
7809 Do not align destination of inlined string operations. This switch reduces
7810 code size and improves performance in case the destination is already aligned,
7811 but gcc don't know about it.
7813 @item -minline-all-stringops
7814 @opindex minline-all-stringops
7815 By default GCC inlines string operations only when destination is known to be
7816 aligned at least to 4 byte boundary. This enables more inlining, increase code
7817 size, but may improve performance of code that depends on fast memcpy, strlen
7818 and memset for short lengths.
7820 @item -momit-leaf-frame-pointer
7821 @opindex momit-leaf-frame-pointer
7822 Don't keep the frame pointer in a register for leaf functions. This
7823 avoids the instructions to save, set up and restore frame pointers and
7824 makes an extra register available in leaf functions. The option
7825 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7826 which might make debugging harder.
7829 These @samp{-m} switches are supported in addition to the above
7830 on AMD x86-64 processors in 64-bit environments.
7837 Generate code for a 32-bit or 64-bit environment.
7838 The 32-bit environment sets int, long and pointer to 32 bits and
7839 generates code that runs on any i386 system.
7840 The 64-bit environment sets int to 32 bits and long and pointer
7841 to 64 bits and generates code for AMD's x86-64 architecture.
7844 @opindex no-red-zone
7845 Do not use a so called red zone for x86-64 code. The red zone is mandated
7846 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7847 stack pointer that will not be modified by signal or interrupt handlers
7848 and therefore can be used for temporary data without adjusting the stack
7849 pointer. The flag @option{-mno-red-zone} disables this red zone.
7851 @item -mcmodel=small
7852 @opindex mcmodel=small
7853 Generate code for the small code model: the program and its symbols must
7854 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
7855 Programs can be statically or dynamically linked. This is the default
7858 @item -mcmodel=kernel
7859 @opindex mcmodel=kernel
7860 Generate code for the kernel code model. The kernel runs in the
7861 negative 2 GB of the address space.
7862 This model has to be used for Linux kernel code.
7864 @item -mcmodel=medium
7865 @opindex mcmodel=medium
7866 Generate code for the medium model: The program is linked in the lower 2
7867 GB of the address space but symbols can be located anywhere in the
7868 address space. Programs can be statically or dynamically linked, but
7869 building of shared libraries are not supported with the medium model.
7871 @item -mcmodel=large
7872 @opindex mcmodel=large
7873 Generate code for the large model: This model makes no assumptions
7874 about addresses and sizes of sections. Currently GCC does not implement
7879 @subsection HPPA Options
7880 @cindex HPPA Options
7882 These @samp{-m} options are defined for the HPPA family of computers:
7885 @item -march=@var{architecture-type}
7887 Generate code for the specified architecture. The choices for
7888 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7889 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7890 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7891 architecture option for your machine. Code compiled for lower numbered
7892 architectures will run on higher numbered architectures, but not the
7895 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7896 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7900 @itemx -mpa-risc-1-1
7901 @itemx -mpa-risc-2-0
7902 @opindex mpa-risc-1-0
7903 @opindex mpa-risc-1-1
7904 @opindex mpa-risc-2-0
7905 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7908 @opindex mbig-switch
7909 Generate code suitable for big switch tables. Use this option only if
7910 the assembler/linker complain about out of range branches within a switch
7913 @item -mjump-in-delay
7914 @opindex mjump-in-delay
7915 Fill delay slots of function calls with unconditional jump instructions
7916 by modifying the return pointer for the function call to be the target
7917 of the conditional jump.
7919 @item -mdisable-fpregs
7920 @opindex mdisable-fpregs
7921 Prevent floating point registers from being used in any manner. This is
7922 necessary for compiling kernels which perform lazy context switching of
7923 floating point registers. If you use this option and attempt to perform
7924 floating point operations, the compiler will abort.
7926 @item -mdisable-indexing
7927 @opindex mdisable-indexing
7928 Prevent the compiler from using indexing address modes. This avoids some
7929 rather obscure problems when compiling MIG generated code under MACH@.
7931 @item -mno-space-regs
7932 @opindex mno-space-regs
7933 Generate code that assumes the target has no space registers. This allows
7934 GCC to generate faster indirect calls and use unscaled index address modes.
7936 Such code is suitable for level 0 PA systems and kernels.
7938 @item -mfast-indirect-calls
7939 @opindex mfast-indirect-calls
7940 Generate code that assumes calls never cross space boundaries. This
7941 allows GCC to emit code which performs faster indirect calls.
7943 This option will not work in the presence of shared libraries or nested
7946 @item -mlong-load-store
7947 @opindex mlong-load-store
7948 Generate 3-instruction load and store sequences as sometimes required by
7949 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7952 @item -mportable-runtime
7953 @opindex mportable-runtime
7954 Use the portable calling conventions proposed by HP for ELF systems.
7958 Enable the use of assembler directives only GAS understands.
7960 @item -mschedule=@var{cpu-type}
7962 Schedule code according to the constraints for the machine type
7963 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7964 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
7965 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
7966 proper scheduling option for your machine. The default scheduling is
7970 @opindex mlinker-opt
7971 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
7972 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
7973 linkers in which they give bogus error messages when linking some programs.
7976 @opindex msoft-float
7977 Generate output containing library calls for floating point.
7978 @strong{Warning:} the requisite libraries are not available for all HPPA
7979 targets. Normally the facilities of the machine's usual C compiler are
7980 used, but this cannot be done directly in cross-compilation. You must make
7981 your own arrangements to provide suitable library functions for
7982 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7983 does provide software floating point support.
7985 @option{-msoft-float} changes the calling convention in the output file;
7986 therefore, it is only useful if you compile @emph{all} of a program with
7987 this option. In particular, you need to compile @file{libgcc.a}, the
7988 library that comes with GCC, with @option{-msoft-float} in order for
7993 Generate the predefine, @code{_SIO}, for server IO. The default is
7994 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
7995 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO. These
7996 options are available under HP-UX and HI-UX.
8000 Use GNU ld specific options. This passes @option{-shared} to ld when
8001 building a shared library. It is the default when GCC is configured,
8002 explicitly or implicitly, with the GNU linker. This option does not
8003 have any affect on which ld is called, it only changes what parameters
8004 are passed to that ld. The ld that is called is determined by the
8005 @option{--with-ld} configure option, gcc's program search path, and
8006 finally by the user's @env{PATH}. The linker used by GCC can be printed
8007 using @samp{which `gcc -print-prog-name=ld`}.
8011 Use HP ld specific options. This passes @option{-b} to ld when building
8012 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8013 links. It is the default when GCC is configured, explicitly or
8014 implicitly, with the HP linker. This option does not have any affect on
8015 which ld is called, it only changes what parameters are passed to that
8016 ld. The ld that is called is determined by the @option{--with-ld}
8017 configure option, gcc's program search path, and finally by the user's
8018 @env{PATH}. The linker used by GCC can be printed using @samp{which
8019 `gcc -print-prog-name=ld`}.
8023 @node Intel 960 Options
8024 @subsection Intel 960 Options
8026 These @samp{-m} options are defined for the Intel 960 implementations:
8029 @item -m@var{cpu-type}
8037 Assume the defaults for the machine type @var{cpu-type} for some of
8038 the other options, including instruction scheduling, floating point
8039 support, and addressing modes. The choices for @var{cpu-type} are
8040 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
8041 @samp{sa}, and @samp{sb}.
8048 @opindex msoft-float
8049 The @option{-mnumerics} option indicates that the processor does support
8050 floating-point instructions. The @option{-msoft-float} option indicates
8051 that floating-point support should not be assumed.
8053 @item -mleaf-procedures
8054 @itemx -mno-leaf-procedures
8055 @opindex mleaf-procedures
8056 @opindex mno-leaf-procedures
8057 Do (or do not) attempt to alter leaf procedures to be callable with the
8058 @code{bal} instruction as well as @code{call}. This will result in more
8059 efficient code for explicit calls when the @code{bal} instruction can be
8060 substituted by the assembler or linker, but less efficient code in other
8061 cases, such as calls via function pointers, or using a linker that doesn't
8062 support this optimization.
8065 @itemx -mno-tail-call
8067 @opindex mno-tail-call
8068 Do (or do not) make additional attempts (beyond those of the
8069 machine-independent portions of the compiler) to optimize tail-recursive
8070 calls into branches. You may not want to do this because the detection of
8071 cases where this is not valid is not totally complete. The default is
8072 @option{-mno-tail-call}.
8074 @item -mcomplex-addr
8075 @itemx -mno-complex-addr
8076 @opindex mcomplex-addr
8077 @opindex mno-complex-addr
8078 Assume (or do not assume) that the use of a complex addressing mode is a
8079 win on this implementation of the i960. Complex addressing modes may not
8080 be worthwhile on the K-series, but they definitely are on the C-series.
8081 The default is currently @option{-mcomplex-addr} for all processors except
8085 @itemx -mno-code-align
8086 @opindex mcode-align
8087 @opindex mno-code-align
8088 Align code to 8-byte boundaries for faster fetching (or don't bother).
8089 Currently turned on by default for C-series implementations only.
8092 @item -mclean-linkage
8093 @itemx -mno-clean-linkage
8094 @opindex mclean-linkage
8095 @opindex mno-clean-linkage
8096 These options are not fully implemented.
8100 @itemx -mic2.0-compat
8101 @itemx -mic3.0-compat
8103 @opindex mic2.0-compat
8104 @opindex mic3.0-compat
8105 Enable compatibility with iC960 v2.0 or v3.0.
8109 @opindex masm-compat
8111 Enable compatibility with the iC960 assembler.
8113 @item -mstrict-align
8114 @itemx -mno-strict-align
8115 @opindex mstrict-align
8116 @opindex mno-strict-align
8117 Do not permit (do permit) unaligned accesses.
8121 Enable structure-alignment compatibility with Intel's gcc release version
8122 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8124 @item -mlong-double-64
8125 @opindex mlong-double-64
8126 Implement type @samp{long double} as 64-bit floating point numbers.
8127 Without the option @samp{long double} is implemented by 80-bit
8128 floating point numbers. The only reason we have it because there is
8129 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8130 is only useful for people using soft-float targets. Otherwise, we
8131 should recommend against use of it.
8135 @node DEC Alpha Options
8136 @subsection DEC Alpha Options
8138 These @samp{-m} options are defined for the DEC Alpha implementations:
8141 @item -mno-soft-float
8143 @opindex mno-soft-float
8144 @opindex msoft-float
8145 Use (do not use) the hardware floating-point instructions for
8146 floating-point operations. When @option{-msoft-float} is specified,
8147 functions in @file{libgcc.a} will be used to perform floating-point
8148 operations. Unless they are replaced by routines that emulate the
8149 floating-point operations, or compiled in such a way as to call such
8150 emulations routines, these routines will issue floating-point
8151 operations. If you are compiling for an Alpha without floating-point
8152 operations, you must ensure that the library is built so as not to call
8155 Note that Alpha implementations without floating-point operations are
8156 required to have floating-point registers.
8161 @opindex mno-fp-regs
8162 Generate code that uses (does not use) the floating-point register set.
8163 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8164 register set is not used, floating point operands are passed in integer
8165 registers as if they were integers and floating-point results are passed
8166 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8167 so any function with a floating-point argument or return value called by code
8168 compiled with @option{-mno-fp-regs} must also be compiled with that
8171 A typical use of this option is building a kernel that does not use,
8172 and hence need not save and restore, any floating-point registers.
8176 The Alpha architecture implements floating-point hardware optimized for
8177 maximum performance. It is mostly compliant with the IEEE floating
8178 point standard. However, for full compliance, software assistance is
8179 required. This option generates code fully IEEE compliant code
8180 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8181 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8182 defined during compilation. The resulting code is less efficient but is
8183 able to correctly support denormalized numbers and exceptional IEEE
8184 values such as not-a-number and plus/minus infinity. Other Alpha
8185 compilers call this option @option{-ieee_with_no_inexact}.
8187 @item -mieee-with-inexact
8188 @opindex mieee-with-inexact
8189 This is like @option{-mieee} except the generated code also maintains
8190 the IEEE @var{inexact-flag}. Turning on this option causes the
8191 generated code to implement fully-compliant IEEE math. In addition to
8192 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8193 macro. On some Alpha implementations the resulting code may execute
8194 significantly slower than the code generated by default. Since there is
8195 very little code that depends on the @var{inexact-flag}, you should
8196 normally not specify this option. Other Alpha compilers call this
8197 option @option{-ieee_with_inexact}.
8199 @item -mfp-trap-mode=@var{trap-mode}
8200 @opindex mfp-trap-mode
8201 This option controls what floating-point related traps are enabled.
8202 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8203 The trap mode can be set to one of four values:
8207 This is the default (normal) setting. The only traps that are enabled
8208 are the ones that cannot be disabled in software (e.g., division by zero
8212 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8216 Like @samp{su}, but the instructions are marked to be safe for software
8217 completion (see Alpha architecture manual for details).
8220 Like @samp{su}, but inexact traps are enabled as well.
8223 @item -mfp-rounding-mode=@var{rounding-mode}
8224 @opindex mfp-rounding-mode
8225 Selects the IEEE rounding mode. Other Alpha compilers call this option
8226 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8231 Normal IEEE rounding mode. Floating point numbers are rounded towards
8232 the nearest machine number or towards the even machine number in case
8236 Round towards minus infinity.
8239 Chopped rounding mode. Floating point numbers are rounded towards zero.
8242 Dynamic rounding mode. A field in the floating point control register
8243 (@var{fpcr}, see Alpha architecture reference manual) controls the
8244 rounding mode in effect. The C library initializes this register for
8245 rounding towards plus infinity. Thus, unless your program modifies the
8246 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8249 @item -mtrap-precision=@var{trap-precision}
8250 @opindex mtrap-precision
8251 In the Alpha architecture, floating point traps are imprecise. This
8252 means without software assistance it is impossible to recover from a
8253 floating trap and program execution normally needs to be terminated.
8254 GCC can generate code that can assist operating system trap handlers
8255 in determining the exact location that caused a floating point trap.
8256 Depending on the requirements of an application, different levels of
8257 precisions can be selected:
8261 Program precision. This option is the default and means a trap handler
8262 can only identify which program caused a floating point exception.
8265 Function precision. The trap handler can determine the function that
8266 caused a floating point exception.
8269 Instruction precision. The trap handler can determine the exact
8270 instruction that caused a floating point exception.
8273 Other Alpha compilers provide the equivalent options called
8274 @option{-scope_safe} and @option{-resumption_safe}.
8276 @item -mieee-conformant
8277 @opindex mieee-conformant
8278 This option marks the generated code as IEEE conformant. You must not
8279 use this option unless you also specify @option{-mtrap-precision=i} and either
8280 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8281 is to emit the line @samp{.eflag 48} in the function prologue of the
8282 generated assembly file. Under DEC Unix, this has the effect that
8283 IEEE-conformant math library routines will be linked in.
8285 @item -mbuild-constants
8286 @opindex mbuild-constants
8287 Normally GCC examines a 32- or 64-bit integer constant to
8288 see if it can construct it from smaller constants in two or three
8289 instructions. If it cannot, it will output the constant as a literal and
8290 generate code to load it from the data segment at runtime.
8292 Use this option to require GCC to construct @emph{all} integer constants
8293 using code, even if it takes more instructions (the maximum is six).
8295 You would typically use this option to build a shared library dynamic
8296 loader. Itself a shared library, it must relocate itself in memory
8297 before it can find the variables and constants in its own data segment.
8303 Select whether to generate code to be assembled by the vendor-supplied
8304 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8322 Indicate whether GCC should generate code to use the optional BWX,
8323 CIX, FIX and MAX instruction sets. The default is to use the instruction
8324 sets supported by the CPU type specified via @option{-mcpu=} option or that
8325 of the CPU on which GCC was built if none was specified.
8330 @opindex mfloat-ieee
8331 Generate code that uses (does not use) VAX F and G floating point
8332 arithmetic instead of IEEE single and double precision.
8334 @item -mexplicit-relocs
8335 @itemx -mno-explicit-relocs
8336 @opindex mexplicit-relocs
8337 @opindex mno-explicit-relocs
8338 Older Alpha assemblers provided no way to generate symbol relocations
8339 except via assembler macros. Use of these macros does not allow
8340 optimial instruction scheduling. GNU binutils as of version 2.12
8341 supports a new syntax that allows the compiler to explicitly mark
8342 which relocations should apply to which instructions. This option
8343 is mostly useful for debugging, as GCC detects the capabilities of
8344 the assembler when it is built and sets the default accordingly.
8348 @opindex msmall-data
8349 @opindex mlarge-data
8350 When @option{-mexplicit-relocs} is in effect, static data is
8351 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8352 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8353 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8354 16-bit relocations off of the @code{$gp} register. This limits the
8355 size of the small data area to 64KB, but allows the variables to be
8356 directly accessed via a single instruction.
8358 The default is @option{-mlarge-data}. With this option the data area
8359 is limited to just below 2GB. Programs that require more than 2GB of
8360 data must use @code{malloc} or @code{mmap} to allocate the data in the
8361 heap instead of in the program's data segment.
8363 When generating code for shared libraries, @option{-fpic} implies
8364 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8366 @item -mcpu=@var{cpu_type}
8368 Set the instruction set and instruction scheduling parameters for
8369 machine type @var{cpu_type}. You can specify either the @samp{EV}
8370 style name or the corresponding chip number. GCC supports scheduling
8371 parameters for the EV4, EV5 and EV6 family of processors and will
8372 choose the default values for the instruction set from the processor
8373 you specify. If you do not specify a processor type, GCC will default
8374 to the processor on which the compiler was built.
8376 Supported values for @var{cpu_type} are
8382 Schedules as an EV4 and has no instruction set extensions.
8386 Schedules as an EV5 and has no instruction set extensions.
8390 Schedules as an EV5 and supports the BWX extension.
8395 Schedules as an EV5 and supports the BWX and MAX extensions.
8399 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8403 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8406 @item -mtune=@var{cpu_type}
8408 Set only the instruction scheduling parameters for machine type
8409 @var{cpu_type}. The instruction set is not changed.
8411 @item -mmemory-latency=@var{time}
8412 @opindex mmemory-latency
8413 Sets the latency the scheduler should assume for typical memory
8414 references as seen by the application. This number is highly
8415 dependent on the memory access patterns used by the application
8416 and the size of the external cache on the machine.
8418 Valid options for @var{time} are
8422 A decimal number representing clock cycles.
8428 The compiler contains estimates of the number of clock cycles for
8429 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8430 (also called Dcache, Scache, and Bcache), as well as to main memory.
8431 Note that L3 is only valid for EV5.
8436 @node DEC Alpha/VMS Options
8437 @subsection DEC Alpha/VMS Options
8439 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8442 @item -mvms-return-codes
8443 @opindex mvms-return-codes
8444 Return VMS condition codes from main. The default is to return POSIX
8445 style condition (e.g.@ error) codes.
8448 @node Clipper Options
8449 @subsection Clipper Options
8451 These @samp{-m} options are defined for the Clipper implementations:
8456 Produce code for a C300 Clipper processor. This is the default.
8460 Produce code for a C400 Clipper processor, i.e.@: use floating point
8464 @node H8/300 Options
8465 @subsection H8/300 Options
8467 These @samp{-m} options are defined for the H8/300 implementations:
8472 Shorten some address references at link time, when possible; uses the
8473 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8474 ld.info, Using ld}, for a fuller description.
8478 Generate code for the H8/300H@.
8482 Generate code for the H8S@.
8486 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
8490 Make @code{int} data 32 bits by default.
8494 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
8495 The default for the H8/300H and H8S is to align longs and floats on 4
8497 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8498 This option has no effect on the H8/300.
8502 @subsection SH Options
8504 These @samp{-m} options are defined for the SH implementations:
8509 Generate code for the SH1.
8513 Generate code for the SH2.
8517 Generate code for the SH3.
8521 Generate code for the SH3e.
8525 Generate code for the SH4 without a floating-point unit.
8527 @item -m4-single-only
8528 @opindex m4-single-only
8529 Generate code for the SH4 with a floating-point unit that only
8530 supports single-precision arithmetic.
8534 Generate code for the SH4 assuming the floating-point unit is in
8535 single-precision mode by default.
8539 Generate code for the SH4.
8543 Compile code for the processor in big endian mode.
8547 Compile code for the processor in little endian mode.
8551 Align doubles at 64-bit boundaries. Note that this changes the calling
8552 conventions, and thus some functions from the standard C library will
8553 not work unless you recompile it first with @option{-mdalign}.
8557 Shorten some address references at link time, when possible; uses the
8558 linker option @option{-relax}.
8562 Use 32-bit offsets in @code{switch} tables. The default is to use
8567 Enable the use of the instruction @code{fmovd}.
8571 Comply with the calling conventions defined by Hitachi.
8575 Mark the @code{MAC} register as call-clobbered, even if
8576 @option{-mhitachi} is given.
8580 Increase IEEE-compliance of floating-point code.
8584 Dump instruction size and location in the assembly code.
8588 This option is deprecated. It pads structures to multiple of 4 bytes,
8589 which is incompatible with the SH ABI@.
8593 Optimize for space instead of speed. Implied by @option{-Os}.
8597 When generating position-independent code, emit function calls using
8598 the Global Offset Table instead of the Procedure Linkage Table.
8602 Generate a library function call to invalidate instruction cache
8603 entries, after fixing up a trampoline. This library function call
8604 doesn't assume it can write to the whole memory address space. This
8605 is the default when the target is @code{sh-*-linux*}.
8608 @node System V Options
8609 @subsection Options for System V
8611 These additional options are available on System V Release 4 for
8612 compatibility with other compilers on those systems:
8617 Create a shared object.
8618 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8622 Identify the versions of each tool used by the compiler, in a
8623 @code{.ident} assembler directive in the output.
8627 Refrain from adding @code{.ident} directives to the output file (this is
8630 @item -YP,@var{dirs}
8632 Search the directories @var{dirs}, and no others, for libraries
8633 specified with @option{-l}.
8637 Look in the directory @var{dir} to find the M4 preprocessor.
8638 The assembler uses this option.
8639 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8640 @c the generic assembler that comes with Solaris takes just -Ym.
8643 @node TMS320C3x/C4x Options
8644 @subsection TMS320C3x/C4x Options
8645 @cindex TMS320C3x/C4x Options
8647 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8651 @item -mcpu=@var{cpu_type}
8653 Set the instruction set, register set, and instruction scheduling
8654 parameters for machine type @var{cpu_type}. Supported values for
8655 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8656 @samp{c44}. The default is @samp{c40} to generate code for the
8661 @itemx -msmall-memory
8663 @opindex mbig-memory
8665 @opindex msmall-memory
8667 Generates code for the big or small memory model. The small memory
8668 model assumed that all data fits into one 64K word page. At run-time
8669 the data page (DP) register must be set to point to the 64K page
8670 containing the .bss and .data program sections. The big memory model is
8671 the default and requires reloading of the DP register for every direct
8678 Allow (disallow) allocation of general integer operands into the block
8685 Enable (disable) generation of code using decrement and branch,
8686 DBcond(D), instructions. This is enabled by default for the C4x. To be
8687 on the safe side, this is disabled for the C3x, since the maximum
8688 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
8689 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
8690 that it can utilize the decrement and branch instruction, but will give
8691 up if there is more than one memory reference in the loop. Thus a loop
8692 where the loop counter is decremented can generate slightly more
8693 efficient code, in cases where the RPTB instruction cannot be utilized.
8695 @item -mdp-isr-reload
8697 @opindex mdp-isr-reload
8699 Force the DP register to be saved on entry to an interrupt service
8700 routine (ISR), reloaded to point to the data section, and restored on
8701 exit from the ISR@. This should not be required unless someone has
8702 violated the small memory model by modifying the DP register, say within
8709 For the C3x use the 24-bit MPYI instruction for integer multiplies
8710 instead of a library call to guarantee 32-bit results. Note that if one
8711 of the operands is a constant, then the multiplication will be performed
8712 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8713 then squaring operations are performed inline instead of a library call.
8716 @itemx -mno-fast-fix
8718 @opindex mno-fast-fix
8719 The C3x/C4x FIX instruction to convert a floating point value to an
8720 integer value chooses the nearest integer less than or equal to the
8721 floating point value rather than to the nearest integer. Thus if the
8722 floating point number is negative, the result will be incorrectly
8723 truncated an additional code is necessary to detect and correct this
8724 case. This option can be used to disable generation of the additional
8725 code required to correct the result.
8731 Enable (disable) generation of repeat block sequences using the RPTB
8732 instruction for zero overhead looping. The RPTB construct is only used
8733 for innermost loops that do not call functions or jump across the loop
8734 boundaries. There is no advantage having nested RPTB loops due to the
8735 overhead required to save and restore the RC, RS, and RE registers.
8736 This is enabled by default with @option{-O2}.
8738 @item -mrpts=@var{count}
8742 Enable (disable) the use of the single instruction repeat instruction
8743 RPTS@. If a repeat block contains a single instruction, and the loop
8744 count can be guaranteed to be less than the value @var{count}, GCC will
8745 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8746 then a RPTS will be emitted even if the loop count cannot be determined
8747 at compile time. Note that the repeated instruction following RPTS does
8748 not have to be reloaded from memory each iteration, thus freeing up the
8749 CPU buses for operands. However, since interrupts are blocked by this
8750 instruction, it is disabled by default.
8752 @item -mloop-unsigned
8753 @itemx -mno-loop-unsigned
8754 @opindex mloop-unsigned
8755 @opindex mno-loop-unsigned
8756 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8757 is @math{2^{31} + 1} since these instructions test if the iteration count is
8758 negative to terminate the loop. If the iteration count is unsigned
8759 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
8760 exceeded. This switch allows an unsigned iteration count.
8764 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8765 with. This also enforces compatibility with the API employed by the TI
8766 C3x C compiler. For example, long doubles are passed as structures
8767 rather than in floating point registers.
8773 Generate code that uses registers (stack) for passing arguments to functions.
8774 By default, arguments are passed in registers where possible rather
8775 than by pushing arguments on to the stack.
8777 @item -mparallel-insns
8778 @itemx -mno-parallel-insns
8779 @opindex mparallel-insns
8780 @opindex mno-parallel-insns
8781 Allow the generation of parallel instructions. This is enabled by
8782 default with @option{-O2}.
8784 @item -mparallel-mpy
8785 @itemx -mno-parallel-mpy
8786 @opindex mparallel-mpy
8787 @opindex mno-parallel-mpy
8788 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8789 provided @option{-mparallel-insns} is also specified. These instructions have
8790 tight register constraints which can pessimize the code generation
8796 @subsection V850 Options
8797 @cindex V850 Options
8799 These @samp{-m} options are defined for V850 implementations:
8803 @itemx -mno-long-calls
8804 @opindex mlong-calls
8805 @opindex mno-long-calls
8806 Treat all calls as being far away (near). If calls are assumed to be
8807 far away, the compiler will always load the functions address up into a
8808 register, and call indirect through the pointer.
8814 Do not optimize (do optimize) basic blocks that use the same index
8815 pointer 4 or more times to copy pointer into the @code{ep} register, and
8816 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8817 option is on by default if you optimize.
8819 @item -mno-prolog-function
8820 @itemx -mprolog-function
8821 @opindex mno-prolog-function
8822 @opindex mprolog-function
8823 Do not use (do use) external functions to save and restore registers at
8824 the prolog and epilog of a function. The external functions are slower,
8825 but use less code space if more than one function saves the same number
8826 of registers. The @option{-mprolog-function} option is on by default if
8831 Try to make the code as small as possible. At present, this just turns
8832 on the @option{-mep} and @option{-mprolog-function} options.
8836 Put static or global variables whose size is @var{n} bytes or less into
8837 the tiny data area that register @code{ep} points to. The tiny data
8838 area can hold up to 256 bytes in total (128 bytes for byte references).
8842 Put static or global variables whose size is @var{n} bytes or less into
8843 the small data area that register @code{gp} points to. The small data
8844 area can hold up to 64 kilobytes.
8848 Put static or global variables whose size is @var{n} bytes or less into
8849 the first 32 kilobytes of memory.
8853 Specify that the target processor is the V850.
8856 @opindex mbig-switch
8857 Generate code suitable for big switch tables. Use this option only if
8858 the assembler/linker complain about out of range branches within a switch
8863 This option will cause r2 and r5 to be used in the code generated by
8864 the compiler. This setting is the default.
8867 @opindex -mno-app-regs
8868 This option will cause r2 and r5 to be treated as fixed registers.
8872 Specify that the target processor is the V850E. The preprocessor
8873 constant @samp{__v850e__} will be defined if this option is used.
8875 If neither @option{-mv850} nor @option{-mv850e} are defined
8876 then a default target processor will be chosen and the relevant
8877 @samp{__v850*__} preprocessor constant will be defined.
8879 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
8880 defined, regardless of which processor variant is the target.
8882 @item -mdisable-callt
8883 @opindex -mdisable-callt
8884 This option will suppress generation of the CALLT instruction for the
8885 v850e flavors of the v850 architecture. The default is
8886 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
8891 @subsection ARC Options
8894 These options are defined for ARC implementations:
8899 Compile code for little endian mode. This is the default.
8903 Compile code for big endian mode.
8906 @opindex mmangle-cpu
8907 Prepend the name of the cpu to all public symbol names.
8908 In multiple-processor systems, there are many ARC variants with different
8909 instruction and register set characteristics. This flag prevents code
8910 compiled for one cpu to be linked with code compiled for another.
8911 No facility exists for handling variants that are ``almost identical''.
8912 This is an all or nothing option.
8914 @item -mcpu=@var{cpu}
8916 Compile code for ARC variant @var{cpu}.
8917 Which variants are supported depend on the configuration.
8918 All variants support @option{-mcpu=base}, this is the default.
8920 @item -mtext=@var{text-section}
8921 @itemx -mdata=@var{data-section}
8922 @itemx -mrodata=@var{readonly-data-section}
8926 Put functions, data, and readonly data in @var{text-section},
8927 @var{data-section}, and @var{readonly-data-section} respectively
8928 by default. This can be overridden with the @code{section} attribute.
8929 @xref{Variable Attributes}.
8934 @subsection NS32K Options
8935 @cindex NS32K options
8937 These are the @samp{-m} options defined for the 32000 series. The default
8938 values for these options depends on which style of 32000 was selected when
8939 the compiler was configured; the defaults for the most common choices are
8947 Generate output for a 32032. This is the default
8948 when the compiler is configured for 32032 and 32016 based systems.
8954 Generate output for a 32332. This is the default
8955 when the compiler is configured for 32332-based systems.
8961 Generate output for a 32532. This is the default
8962 when the compiler is configured for 32532-based systems.
8966 Generate output containing 32081 instructions for floating point.
8967 This is the default for all systems.
8971 Generate output containing 32381 instructions for floating point. This
8972 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8973 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8977 Try and generate multiply-add floating point instructions @code{polyF}
8978 and @code{dotF}. This option is only available if the @option{-m32381}
8979 option is in effect. Using these instructions requires changes to
8980 register allocation which generally has a negative impact on
8981 performance. This option should only be enabled when compiling code
8982 particularly likely to make heavy use of multiply-add instructions.
8985 @opindex mnomulti-add
8986 Do not try and generate multiply-add floating point instructions
8987 @code{polyF} and @code{dotF}. This is the default on all platforms.
8990 @opindex msoft-float
8991 Generate output containing library calls for floating point.
8992 @strong{Warning:} the requisite libraries may not be available.
8994 @item -mieee-compare
8995 @itemx -mno-ieee-compare
8996 @opindex mieee-compare
8997 @opindex mno-ieee-compare
8998 Control whether or not the compiler uses IEEE floating point
8999 comparisons. These handle correctly the case where the result of a
9000 comparison is unordered.
9001 @strong{Warning:} the requisite kernel support may not be available.
9004 @opindex mnobitfield
9005 Do not use the bit-field instructions. On some machines it is faster to
9006 use shifting and masking operations. This is the default for the pc532.
9010 Do use the bit-field instructions. This is the default for all platforms
9015 Use a different function-calling convention, in which functions
9016 that take a fixed number of arguments return pop their
9017 arguments on return with the @code{ret} instruction.
9019 This calling convention is incompatible with the one normally
9020 used on Unix, so you cannot use it if you need to call libraries
9021 compiled with the Unix compiler.
9023 Also, you must provide function prototypes for all functions that
9024 take variable numbers of arguments (including @code{printf});
9025 otherwise incorrect code will be generated for calls to those
9028 In addition, seriously incorrect code will result if you call a
9029 function with too many arguments. (Normally, extra arguments are
9030 harmlessly ignored.)
9032 This option takes its name from the 680x0 @code{rtd} instruction.
9037 Use a different function-calling convention where the first two arguments
9038 are passed in registers.
9040 This calling convention is incompatible with the one normally
9041 used on Unix, so you cannot use it if you need to call libraries
9042 compiled with the Unix compiler.
9045 @opindex mnoregparam
9046 Do not pass any arguments in registers. This is the default for all
9051 It is OK to use the sb as an index register which is always loaded with
9052 zero. This is the default for the pc532-netbsd target.
9056 The sb register is not available for use or has not been initialized to
9057 zero by the run time system. This is the default for all targets except
9058 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
9059 @option{-fpic} is set.
9063 Many ns32000 series addressing modes use displacements of up to 512MB@.
9064 If an address is above 512MB then displacements from zero can not be used.
9065 This option causes code to be generated which can be loaded above 512MB@.
9066 This may be useful for operating systems or ROM code.
9070 Assume code will be loaded in the first 512MB of virtual address space.
9071 This is the default for all platforms.
9077 @subsection AVR Options
9080 These options are defined for AVR implementations:
9083 @item -mmcu=@var{mcu}
9085 Specify ATMEL AVR instruction set or MCU type.
9087 Instruction set avr1 is for the minimal AVR core, not supported by the C
9088 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
9089 attiny11, attiny12, attiny15, attiny28).
9091 Instruction set avr2 (default) is for the classic AVR core with up to
9092 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
9093 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
9094 at90c8534, at90s8535).
9096 Instruction set avr3 is for the classic AVR core with up to 128K program
9097 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
9099 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9100 memory space (MCU types: atmega8, atmega83, atmega85).
9102 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9103 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9104 atmega64, atmega128, at43usb355, at94k).
9108 Output instruction sizes to the asm file.
9110 @item -minit-stack=@var{N}
9111 @opindex minit-stack
9112 Specify the initial stack address, which may be a symbol or numeric value,
9113 @samp{__stack} is the default.
9115 @item -mno-interrupts
9116 @opindex mno-interrupts
9117 Generated code is not compatible with hardware interrupts.
9118 Code size will be smaller.
9120 @item -mcall-prologues
9121 @opindex mcall-prologues
9122 Functions prologues/epilogues expanded as call to appropriate
9123 subroutines. Code size will be smaller.
9125 @item -mno-tablejump
9126 @opindex mno-tablejump
9127 Do not generate tablejump insns which sometimes increase code size.
9130 @opindex mtiny-stack
9131 Change only the low 8 bits of the stack pointer.
9135 @subsection MCore Options
9136 @cindex MCore options
9138 These are the @samp{-m} options defined for the Motorola M*Core
9148 @opindex mno-hardlit
9149 Inline constants into the code stream if it can be done in two
9150 instructions or less.
9158 Use the divide instruction. (Enabled by default).
9160 @item -mrelax-immediate
9161 @itemx -mrelax-immediate
9162 @itemx -mno-relax-immediate
9163 @opindex mrelax-immediate
9164 @opindex mrelax-immediate
9165 @opindex mno-relax-immediate
9166 Allow arbitrary sized immediates in bit operations.
9168 @item -mwide-bitfields
9169 @itemx -mwide-bitfields
9170 @itemx -mno-wide-bitfields
9171 @opindex mwide-bitfields
9172 @opindex mwide-bitfields
9173 @opindex mno-wide-bitfields
9174 Always treat bit-fields as int-sized.
9176 @item -m4byte-functions
9177 @itemx -m4byte-functions
9178 @itemx -mno-4byte-functions
9179 @opindex m4byte-functions
9180 @opindex m4byte-functions
9181 @opindex mno-4byte-functions
9182 Force all functions to be aligned to a four byte boundary.
9184 @item -mcallgraph-data
9185 @itemx -mcallgraph-data
9186 @itemx -mno-callgraph-data
9187 @opindex mcallgraph-data
9188 @opindex mcallgraph-data
9189 @opindex mno-callgraph-data
9190 Emit callgraph information.
9194 @itemx -mno-slow-bytes
9195 @opindex mslow-bytes
9196 @opindex mslow-bytes
9197 @opindex mno-slow-bytes
9198 Prefer word access when reading byte quantities.
9200 @item -mlittle-endian
9201 @itemx -mlittle-endian
9203 @opindex mlittle-endian
9204 @opindex mlittle-endian
9205 @opindex mbig-endian
9206 Generate code for a little endian target.
9214 Generate code for the 210 processor.
9218 @subsection IA-64 Options
9219 @cindex IA-64 Options
9221 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9225 @opindex mbig-endian
9226 Generate code for a big endian target. This is the default for HP-UX@.
9228 @item -mlittle-endian
9229 @opindex mlittle-endian
9230 Generate code for a little endian target. This is the default for AIX5
9237 Generate (or don't) code for the GNU assembler. This is the default.
9238 @c Also, this is the default if the configure option @option{--with-gnu-as}
9245 Generate (or don't) code for the GNU linker. This is the default.
9246 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9251 Generate code that does not use a global pointer register. The result
9252 is not position independent code, and violates the IA-64 ABI@.
9254 @item -mvolatile-asm-stop
9255 @itemx -mno-volatile-asm-stop
9256 @opindex mvolatile-asm-stop
9257 @opindex mno-volatile-asm-stop
9258 Generate (or don't) a stop bit immediately before and after volatile asm
9263 Generate code that works around Itanium B step errata.
9265 @item -mregister-names
9266 @itemx -mno-register-names
9267 @opindex mregister-names
9268 @opindex mno-register-names
9269 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9270 the stacked registers. This may make assembler output more readable.
9276 Disable (or enable) optimizations that use the small data section. This may
9277 be useful for working around optimizer bugs.
9280 @opindex mconstant-gp
9281 Generate code that uses a single constant global pointer value. This is
9282 useful when compiling kernel code.
9286 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9287 This is useful when compiling firmware code.
9289 @item -minline-divide-min-latency
9290 @opindex minline-divide-min-latency
9291 Generate code for inline divides using the minimum latency algorithm.
9293 @item -minline-divide-max-throughput
9294 @opindex minline-divide-max-throughput
9295 Generate code for inline divides using the maximum throughput algorithm.
9297 @item -mno-dwarf2-asm
9299 @opindex mno-dwarf2-asm
9300 @opindex mdwarf2-asm
9301 Don't (or do) generate assembler code for the DWARF2 line number debugging
9302 info. This may be useful when not using the GNU assembler.
9304 @item -mfixed-range=@var{register-range}
9305 @opindex mfixed-range
9306 Generate code treating the given register range as fixed registers.
9307 A fixed register is one that the register allocator can not use. This is
9308 useful when compiling kernel code. A register range is specified as
9309 two registers separated by a dash. Multiple register ranges can be
9310 specified separated by a comma.
9314 @subsection D30V Options
9315 @cindex D30V Options
9317 These @samp{-m} options are defined for D30V implementations:
9322 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
9323 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9324 memory, which starts at location @code{0x80000000}.
9328 Same as the @option{-mextmem} switch.
9332 Link the @samp{.text} section into onchip text memory, which starts at
9333 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9334 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9335 into onchip data memory, which starts at location @code{0x20000000}.
9337 @item -mno-asm-optimize
9338 @itemx -masm-optimize
9339 @opindex mno-asm-optimize
9340 @opindex masm-optimize
9341 Disable (enable) passing @option{-O} to the assembler when optimizing.
9342 The assembler uses the @option{-O} option to automatically parallelize
9343 adjacent short instructions where possible.
9345 @item -mbranch-cost=@var{n}
9346 @opindex mbranch-cost
9347 Increase the internal costs of branches to @var{n}. Higher costs means
9348 that the compiler will issue more instructions to avoid doing a branch.
9351 @item -mcond-exec=@var{n}
9353 Specify the maximum number of conditionally executed instructions that
9354 replace a branch. The default is 4.
9357 @node S/390 and zSeries Options
9358 @subsection S/390 and zSeries Options
9359 @cindex S/390 and zSeries Options
9361 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9366 @opindex mhard-float
9367 @opindex msoft-float
9368 Use (do not use) the hardware floating-point instructions and registers
9369 for floating-point operations. When @option{-msoft-float} is specified,
9370 functions in @file{libgcc.a} will be used to perform floating-point
9371 operations. When @option{-mhard-float} is specified, the compiler
9372 generates IEEE floating-point instructions. This is the default.
9375 @itemx -mno-backchain
9377 @opindex mno-backchain
9378 Generate (or do not generate) code which maintains an explicit
9379 backchain within the stack frame that points to the caller's frame.
9380 This is currently needed to allow debugging. The default is to
9381 generate the backchain.
9384 @itemx -mno-small-exec
9385 @opindex msmall-exec
9386 @opindex mno-small-exec
9387 Generate (or do not generate) code using the @code{bras} instruction
9388 to do subroutine calls.
9389 This only works reliably if the total executable size does not
9390 exceed 64k. The default is to use the @code{basr} instruction instead,
9391 which does not have this limitation.
9397 When @option{-m31} is specified, generate code compliant to the
9398 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9399 code compliant to the Linux for zSeries ABI@. This allows GCC in
9400 particular to generate 64-bit instructions. For the @samp{s390}
9401 targets, the default is @option{-m31}, while the @samp{s390x}
9402 targets default to @option{-m64}.
9408 Generate (or do not generate) code using the @code{mvcle} instruction
9409 to perform block moves. When @option{-mno-mvcle} is specifed,
9410 use a @code{mvc} loop instead. This is the default.
9416 Print (or do not print) additional debug information when compiling.
9417 The default is to not print debug information.
9422 @subsection CRIS Options
9423 @cindex CRIS Options
9425 These options are defined specifically for the CRIS ports.
9428 @item -march=@var{architecture-type}
9429 @itemx -mcpu=@var{architecture-type}
9432 Generate code for the specified architecture. The choices for
9433 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9434 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9435 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9438 @item -mtune=@var{architecture-type}
9440 Tune to @var{architecture-type} everything applicable about the generated
9441 code, except for the ABI and the set of available instructions. The
9442 choices for @var{architecture-type} are the same as for
9443 @option{-march=@var{architecture-type}}.
9445 @item -mmax-stack-frame=@var{n}
9446 @opindex mmax-stack-frame
9447 Warn when the stack frame of a function exceeds @var{n} bytes.
9449 @item -melinux-stacksize=@var{n}
9450 @opindex melinux-stacksize
9451 Only available with the @samp{cris-axis-aout} target. Arranges for
9452 indications in the program to the kernel loader that the stack of the
9453 program should be set to @var{n} bytes.
9459 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9460 @option{-march=v3} and @option{-march=v8} respectively.
9464 Enable CRIS-specific verbose debug-related information in the assembly
9465 code. This option also has the effect to turn off the @samp{#NO_APP}
9466 formatted-code indicator to the assembler at the beginning of the
9471 Do not use condition-code results from previous instruction; always emit
9472 compare and test instructions before use of condition codes.
9474 @item -mno-side-effects
9475 @opindex mno-side-effects
9476 Do not emit instructions with side-effects in addressing modes other than
9480 @itemx -mno-stack-align
9482 @itemx -mno-data-align
9483 @itemx -mconst-align
9484 @itemx -mno-const-align
9485 @opindex mstack-align
9486 @opindex mno-stack-align
9487 @opindex mdata-align
9488 @opindex mno-data-align
9489 @opindex mconst-align
9490 @opindex mno-const-align
9491 These options (no-options) arranges (eliminate arrangements) for the
9492 stack-frame, individual data and constants to be aligned for the maximum
9493 single data access size for the chosen CPU model. The default is to
9494 arrange for 32-bit alignment. ABI details such as structure layout are
9495 not affected by these options.
9503 Similar to the stack- data- and const-align options above, these options
9504 arrange for stack-frame, writable data and constants to all be 32-bit,
9505 16-bit or 8-bit aligned. The default is 32-bit alignment.
9507 @item -mno-prologue-epilogue
9508 @itemx -mprologue-epilogue
9509 @opindex mno-prologue-epilogue
9510 @opindex mprologue-epilogue
9511 With @option{-mno-prologue-epilogue}, the normal function prologue and
9512 epilogue that sets up the stack-frame are omitted and no return
9513 instructions or return sequences are generated in the code. Use this
9514 option only together with visual inspection of the compiled code: no
9515 warnings or errors are generated when call-saved registers must be saved,
9516 or storage for local variable needs to be allocated.
9522 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9523 instruction sequences that load addresses for functions from the PLT part
9524 of the GOT rather than (traditional on other architectures) calls to the
9525 PLT. The default is @option{-mgotplt}.
9529 Legacy no-op option only recognized with the cris-axis-aout target.
9533 Legacy no-op option only recognized with the cris-axis-elf and
9534 cris-axis-linux-gnu targets.
9538 Only recognized with the cris-axis-aout target, where it selects a
9539 GNU/linux-like multilib, include files and instruction set for
9544 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9548 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9549 to link with input-output functions from a simulator library. Code,
9550 initialized data and zero-initialized data are allocated consecutively.
9554 Like @option{-sim}, but pass linker options to locate initialized data at
9555 0x40000000 and zero-initialized data at 0x80000000.
9559 @subsection MMIX Options
9560 @cindex MMIX Options
9562 These options are defined for the MMIX:
9566 @itemx -mno-libfuncs
9568 @opindex mno-libfuncs
9569 Specify that intrinsic library functions are being compiled, passing all
9570 values in registers, no matter the size.
9575 @opindex mno-epsilon
9576 Generate floating-point comparison instructions that compare with respect
9577 to the @code{rE} epsilon register.
9579 @item -mabi=mmixware
9581 @opindex mabi-mmixware
9583 Generate code that passes function parameters and return values that (in
9584 the called function) are seen as registers @code{$0} and up, as opposed to
9585 the GNU ABI which uses global registers @code{$231} and up.
9588 @itemx -mno-zero-extend
9589 @opindex mzero-extend
9590 @opindex mno-zero-extend
9591 When reading data from memory in sizes shorter than 64 bits, use (do not
9592 use) zero-extending load instructions by default, rather than
9593 sign-extending ones.
9596 @itemx -mno-knuthdiv
9598 @opindex mno-knuthdiv
9599 Make the result of a division yielding a remainder have the same sign as
9600 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9601 remainder follows the sign of the dividend. Both methods are
9602 arithmetically valid, the latter being almost exclusively used.
9604 @item -mtoplevel-symbols
9605 @itemx -mno-toplevel-symbols
9606 @opindex mtoplevel-symbols
9607 @opindex mno-toplevel-symbols
9608 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9609 code can be used with the @code{PREFIX} assembly directive.
9613 Generate an executable in the ELF format, rather than the default
9614 @samp{mmo} format used by the @command{mmix} simulator.
9616 @item -mbranch-predict
9617 @itemx -mno-branch-predict
9618 @opindex mbranch-predict
9619 @opindex mno-branch-predict
9620 Use (do not use) the probable-branch instructions, when static branch
9621 prediction indicates a probable branch.
9623 @item -mbase-addresses
9624 @itemx -mno-base-addresses
9625 @opindex mbase-addresses
9626 @opindex mno-base-addresses
9627 Generate (do not generate) code that uses @emph{base addresses}. Using a
9628 base address automatically generates a request (handled by the assembler
9629 and the linker) for a constant to be set up in a global register. The
9630 register is used for one or more base address requests within the range 0
9631 to 255 from the value held in the register. The generally leads to short
9632 and fast code, but the number of different data items that can be
9633 addressed is limited. This means that a program that uses lots of static
9634 data may require @option{-mno-base-addresses}.
9637 @itemx -mno-single-exit
9638 @opindex msingle-exit
9639 @opindex mno-single-exit
9640 Force (do not force) generated code to have a single exit point in each
9644 @node PDP-11 Options
9645 @subsection PDP-11 Options
9646 @cindex PDP-11 Options
9648 These options are defined for the PDP-11:
9653 Use hardware FPP floating point. This is the default. (FIS floating
9654 point on the PDP-11/40 is not supported.)
9657 @opindex msoft-float
9658 Do not use hardware floating point.
9662 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9666 Return floating-point results in memory. This is the default.
9670 Generate code for a PDP-11/40.
9674 Generate code for a PDP-11/45. This is the default.
9678 Generate code for a PDP-11/10.
9680 @item -mbcopy-builtin
9681 @opindex bcopy-builtin
9682 Use inline @code{movstrhi} patterns for copying memory. This is the
9687 Do not use inline @code{movstrhi} patterns for copying memory.
9693 Use 16-bit @code{int}. This is the default.
9699 Use 32-bit @code{int}.
9704 @opindex mno-float32
9705 Use 64-bit @code{float}. This is the default.
9710 @opindex mno-float64
9711 Use 32-bit @code{float}.
9715 Use @code{abshi2} pattern. This is the default.
9719 Do not use @code{abshi2} pattern.
9721 @item -mbranch-expensive
9722 @opindex mbranch-expensive
9723 Pretend that branches are expensive. This is for experimenting with
9724 code generation only.
9726 @item -mbranch-cheap
9727 @opindex mbranch-cheap
9728 Do not pretend that branches are expensive. This is the default.
9732 Generate code for a system with split I&D.
9736 Generate code for a system without split I&D. This is the default.
9740 Use Unix assembler syntax. This is the default when configured for
9745 Use DEC assembler syntax. This is the default when configured for any
9746 PDP-11 target other than @samp{pdp11-*-bsd}.
9749 @node Xstormy16 Options
9750 @subsection Xstormy16 Options
9751 @cindex Xstormy16 Options
9753 These options are defined for Xstormy16:
9758 Choose startup files and linker script suitable for the simulator.
9761 @node Xtensa Options
9762 @subsection Xtensa Options
9763 @cindex Xtensa Options
9765 The Xtensa architecture is designed to support many different
9766 configurations. The compiler's default options can be set to match a
9767 particular Xtensa configuration by copying a configuration file into the
9768 GCC sources when building GCC@. The options below may be used to
9769 override the default options.
9773 @itemx -mlittle-endian
9774 @opindex mbig-endian
9775 @opindex mlittle-endian
9776 Specify big-endian or little-endian byte ordering for the target Xtensa
9782 @opindex mno-density
9783 Enable or disable use of the optional Xtensa code density instructions.
9789 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
9790 will generate MAC16 instructions from standard C code, with the
9791 limitation that it will use neither the MR register file nor any
9792 instruction that operates on the MR registers. When this option is
9793 disabled, GCC will translate 16-bit multiply/accumulate operations to a
9794 combination of core instructions and library calls, depending on whether
9795 any other multiplier options are enabled.
9801 Enable or disable use of the 16-bit integer multiplier option. When
9802 enabled, the compiler will generate 16-bit multiply instructions for
9803 multiplications of 16 bits or smaller in standard C code. When this
9804 option is disabled, the compiler will either use 32-bit multiply or
9805 MAC16 instructions if they are available or generate library calls to
9806 perform the multiply operations using shifts and adds.
9812 Enable or disable use of the 32-bit integer multiplier option. When
9813 enabled, the compiler will generate 32-bit multiply instructions for
9814 multiplications of 32 bits or smaller in standard C code. When this
9815 option is disabled, the compiler will generate library calls to perform
9816 the multiply operations using either shifts and adds or 16-bit multiply
9817 instructions if they are available.
9823 Enable or disable use of the optional normalization shift amount
9824 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
9830 Enable or disable use of the optional minimum and maximum value
9837 Enable or disable use of the optional sign extend (@code{SEXT})
9841 @itemx -mno-booleans
9843 @opindex mno-booleans
9844 Enable or disable support for the boolean register file used by Xtensa
9845 coprocessors. This is not typically useful by itself but may be
9846 required for other options that make use of the boolean registers (e.g.,
9847 the floating-point option).
9851 @opindex mhard-float
9852 @opindex msoft-float
9853 Enable or disable use of the floating-point option. When enabled, GCC
9854 generates floating-point instructions for 32-bit @code{float}
9855 operations. When this option is disabled, GCC generates library calls
9856 to emulate 32-bit floating-point operations using integer instructions.
9857 Regardless of this option, 64-bit @code{double} operations are always
9858 emulated with calls to library functions.
9861 @itemx -mno-fused-madd
9862 @opindex mfused-madd
9863 @opindex mno-fused-madd
9864 Enable or disable use of fused multiply/add and multiply/subtract
9865 instructions in the floating-point option. This has no effect if the
9866 floating-point option is not also enabled. Disabling fused multiply/add
9867 and multiply/subtract instructions forces the compiler to use separate
9868 instructions for the multiply and add/subtract operations. This may be
9869 desirable in some cases where strict IEEE 754-compliant results are
9870 required: the fused multiply add/subtract instructions do not round the
9871 intermediate result, thereby producing results with @emph{more} bits of
9872 precision than specified by the IEEE standard. Disabling fused multiply
9873 add/subtract instructions also ensures that the program output is not
9874 sensitive to the compiler's ability to combine multiply and add/subtract
9877 @item -mserialize-volatile
9878 @itemx -mno-serialize-volatile
9879 @opindex mserialize-volatile
9880 @opindex mno-serialize-volatile
9881 When this option is enabled, GCC inserts @code{MEMW} instructions before
9882 @code{volatile} memory references to guarantee sequential consistency.
9883 The default is @option{-mserialize-volatile}. Use
9884 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
9886 @item -mtext-section-literals
9887 @itemx -mno-text-section-literals
9888 @opindex mtext-section-literals
9889 @opindex mno-text-section-literals
9890 Control the treatment of literal pools. The default is
9891 @option{-mno-text-section-literals}, which places literals in a separate
9892 section in the output file. This allows the literal pool to be placed
9893 in a data RAM/ROM, and it also allows the linker to combine literal
9894 pools from separate object files to remove redundant literals and
9895 improve code size. With @option{-mtext-section-literals}, the literals
9896 are interspersed in the text section in order to keep them as close as
9897 possible to their references. This may be necessary for large assembly
9900 @item -mtarget-align
9901 @itemx -mno-target-align
9902 @opindex mtarget-align
9903 @opindex mno-target-align
9904 When this option is enabled, GCC instructs the assembler to
9905 automatically align instructions to reduce branch penalties at the
9906 expense of some code density. The assembler attempts to widen density
9907 instructions to align branch targets and the instructions following call
9908 instructions. If there are not enough preceding safe density
9909 instructions to align a target, no widening will be performed. The
9910 default is @option{-mtarget-align}. These options do not affect the
9911 treatment of auto-aligned instructions like @code{LOOP}, which the
9912 assembler will always align, either by widening density instructions or
9913 by inserting no-op instructions.
9916 @itemx -mno-longcalls
9918 @opindex mno-longcalls
9919 When this option is enabled, GCC instructs the assembler to translate
9920 direct calls to indirect calls unless it can determine that the target
9921 of a direct call is in the range allowed by the call instruction. This
9922 translation typically occurs for calls to functions in other source
9923 files. Specifically, the assembler translates a direct @code{CALL}
9924 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
9925 The default is @option{-mno-longcalls}. This option should be used in
9926 programs where the call target can potentially be out of range. This
9927 option is implemented in the assembler, not the compiler, so the
9928 assembly code generated by GCC will still show direct call
9929 instructions---look at the disassembled object code to see the actual
9930 instructions. Note that the assembler will use an indirect call for
9931 every cross-file call, not just those that really will be out of range.
9934 @node Code Gen Options
9935 @section Options for Code Generation Conventions
9936 @cindex code generation conventions
9937 @cindex options, code generation
9938 @cindex run-time options
9940 These machine-independent options control the interface conventions
9941 used in code generation.
9943 Most of them have both positive and negative forms; the negative form
9944 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9945 one of the forms is listed---the one which is not the default. You
9946 can figure out the other form by either removing @samp{no-} or adding
9951 @opindex fexceptions
9952 Enable exception handling. Generates extra code needed to propagate
9953 exceptions. For some targets, this implies GCC will generate frame
9954 unwind information for all functions, which can produce significant data
9955 size overhead, although it does not affect execution. If you do not
9956 specify this option, GCC will enable it by default for languages like
9957 C++ which normally require exception handling, and disable it for
9958 languages like C that do not normally require it. However, you may need
9959 to enable this option when compiling C code that needs to interoperate
9960 properly with exception handlers written in C++. You may also wish to
9961 disable this option if you are compiling older C++ programs that don't
9962 use exception handling.
9964 @item -fnon-call-exceptions
9965 @opindex fnon-call-exceptions
9966 Generate code that allows trapping instructions to throw exceptions.
9967 Note that this requires platform-specific runtime support that does
9968 not exist everywhere. Moreover, it only allows @emph{trapping}
9969 instructions to throw exceptions, i.e.@: memory references or floating
9970 point instructions. It does not allow exceptions to be thrown from
9971 arbitrary signal handlers such as @code{SIGALRM}.
9973 @item -funwind-tables
9974 @opindex funwind-tables
9975 Similar to @option{-fexceptions}, except that it will just generate any needed
9976 static data, but will not affect the generated code in any other way.
9977 You will normally not enable this option; instead, a language processor
9978 that needs this handling would enable it on your behalf.
9980 @item -fasynchronous-unwind-tables
9981 @opindex funwind-tables
9982 Generate unwind table in dwarf2 format, if supported by target machine. The
9983 table is exact at each instruction boundary, so it can be used for stack
9984 unwinding from asynchronous events (such as debugger or garbage collector).
9986 @item -fpcc-struct-return
9987 @opindex fpcc-struct-return
9988 Return ``short'' @code{struct} and @code{union} values in memory like
9989 longer ones, rather than in registers. This convention is less
9990 efficient, but it has the advantage of allowing intercallability between
9991 GCC-compiled files and files compiled with other compilers, particularly
9992 the Portable C Compiler (pcc).
9994 The precise convention for returning structures in memory depends
9995 on the target configuration macros.
9997 Short structures and unions are those whose size and alignment match
9998 that of some integer type.
10000 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
10001 switch is not binary compatible with code compiled with the
10002 @option{-freg-struct-return} switch.
10003 Use it to conform to a non-default application binary interface.
10005 @item -freg-struct-return
10006 @opindex freg-struct-return
10007 Return @code{struct} and @code{union} values in registers when possible.
10008 This is more efficient for small structures than
10009 @option{-fpcc-struct-return}.
10011 If you specify neither @option{-fpcc-struct-return} nor
10012 @option{-freg-struct-return}, GCC defaults to whichever convention is
10013 standard for the target. If there is no standard convention, GCC
10014 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
10015 the principal compiler. In those cases, we can choose the standard, and
10016 we chose the more efficient register return alternative.
10018 @strong{Warning:} code compiled with the @option{-freg-struct-return}
10019 switch is not binary compatible with code compiled with the
10020 @option{-fpcc-struct-return} switch.
10021 Use it to conform to a non-default application binary interface.
10023 @item -fshort-enums
10024 @opindex fshort-enums
10025 Allocate to an @code{enum} type only as many bytes as it needs for the
10026 declared range of possible values. Specifically, the @code{enum} type
10027 will be equivalent to the smallest integer type which has enough room.
10029 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
10030 code that is not binary compatible with code generated without that switch.
10031 Use it to conform to a non-default application binary interface.
10033 @item -fshort-double
10034 @opindex fshort-double
10035 Use the same size for @code{double} as for @code{float}.
10037 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
10038 code that is not binary compatible with code generated without that switch.
10039 Use it to conform to a non-default application binary interface.
10041 @item -fshort-wchar
10042 @opindex fshort-wchar
10043 Override the underlying type for @samp{wchar_t} to be @samp{short
10044 unsigned int} instead of the default for the target. This option is
10045 useful for building programs to run under WINE@.
10047 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
10048 code that is not binary compatible with code generated without that switch.
10049 Use it to conform to a non-default application binary interface.
10051 @item -fshared-data
10052 @opindex fshared-data
10053 Requests that the data and non-@code{const} variables of this
10054 compilation be shared data rather than private data. The distinction
10055 makes sense only on certain operating systems, where shared data is
10056 shared between processes running the same program, while private data
10057 exists in one copy per process.
10060 @opindex fno-common
10061 In C, allocate even uninitialized global variables in the data section of the
10062 object file, rather than generating them as common blocks. This has the
10063 effect that if the same variable is declared (without @code{extern}) in
10064 two different compilations, you will get an error when you link them.
10065 The only reason this might be useful is if you wish to verify that the
10066 program will work on other systems which always work this way.
10070 Ignore the @samp{#ident} directive.
10072 @item -fno-gnu-linker
10073 @opindex fno-gnu-linker
10074 Do not output global initializations (such as C++ constructors and
10075 destructors) in the form used by the GNU linker (on systems where the GNU
10076 linker is the standard method of handling them). Use this option when
10077 you want to use a non-GNU linker, which also requires using the
10078 @command{collect2} program to make sure the system linker includes
10079 constructors and destructors. (@command{collect2} is included in the GCC
10080 distribution.) For systems which @emph{must} use @command{collect2}, the
10081 compiler driver @command{gcc} is configured to do this automatically.
10083 @item -finhibit-size-directive
10084 @opindex finhibit-size-directive
10085 Don't output a @code{.size} assembler directive, or anything else that
10086 would cause trouble if the function is split in the middle, and the
10087 two halves are placed at locations far apart in memory. This option is
10088 used when compiling @file{crtstuff.c}; you should not need to use it
10091 @item -fverbose-asm
10092 @opindex fverbose-asm
10093 Put extra commentary information in the generated assembly code to
10094 make it more readable. This option is generally only of use to those
10095 who actually need to read the generated assembly code (perhaps while
10096 debugging the compiler itself).
10098 @option{-fno-verbose-asm}, the default, causes the
10099 extra information to be omitted and is useful when comparing two assembler
10104 Consider all memory references through pointers to be volatile.
10106 @item -fvolatile-global
10107 @opindex fvolatile-global
10108 Consider all memory references to extern and global data items to
10109 be volatile. GCC does not consider static data items to be volatile
10110 because of this switch.
10112 @item -fvolatile-static
10113 @opindex fvolatile-static
10114 Consider all memory references to static data to be volatile.
10118 @cindex global offset table
10120 Generate position-independent code (PIC) suitable for use in a shared
10121 library, if supported for the target machine. Such code accesses all
10122 constant addresses through a global offset table (GOT)@. The dynamic
10123 loader resolves the GOT entries when the program starts (the dynamic
10124 loader is not part of GCC; it is part of the operating system). If
10125 the GOT size for the linked executable exceeds a machine-specific
10126 maximum size, you get an error message from the linker indicating that
10127 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
10128 instead. (These maximums are 16k on the m88k, 8k on the SPARC, and 32k
10129 on the m68k and RS/6000. The 386 has no such limit.)
10131 Position-independent code requires special support, and therefore works
10132 only on certain machines. For the 386, GCC supports PIC for System V
10133 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
10134 position-independent.
10138 If supported for the target machine, emit position-independent code,
10139 suitable for dynamic linking and avoiding any limit on the size of the
10140 global offset table. This option makes a difference on the m68k, m88k,
10143 Position-independent code requires special support, and therefore works
10144 only on certain machines.
10146 @item -ffixed-@var{reg}
10148 Treat the register named @var{reg} as a fixed register; generated code
10149 should never refer to it (except perhaps as a stack pointer, frame
10150 pointer or in some other fixed role).
10152 @var{reg} must be the name of a register. The register names accepted
10153 are machine-specific and are defined in the @code{REGISTER_NAMES}
10154 macro in the machine description macro file.
10156 This flag does not have a negative form, because it specifies a
10159 @item -fcall-used-@var{reg}
10160 @opindex fcall-used
10161 Treat the register named @var{reg} as an allocable register that is
10162 clobbered by function calls. It may be allocated for temporaries or
10163 variables that do not live across a call. Functions compiled this way
10164 will not save and restore the register @var{reg}.
10166 It is an error to used this flag with the frame pointer or stack pointer.
10167 Use of this flag for other registers that have fixed pervasive roles in
10168 the machine's execution model will produce disastrous results.
10170 This flag does not have a negative form, because it specifies a
10173 @item -fcall-saved-@var{reg}
10174 @opindex fcall-saved
10175 Treat the register named @var{reg} as an allocable register saved by
10176 functions. It may be allocated even for temporaries or variables that
10177 live across a call. Functions compiled this way will save and restore
10178 the register @var{reg} if they use it.
10180 It is an error to used this flag with the frame pointer or stack pointer.
10181 Use of this flag for other registers that have fixed pervasive roles in
10182 the machine's execution model will produce disastrous results.
10184 A different sort of disaster will result from the use of this flag for
10185 a register in which function values may be returned.
10187 This flag does not have a negative form, because it specifies a
10190 @item -fpack-struct
10191 @opindex fpack-struct
10192 Pack all structure members together without holes.
10194 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
10195 code that is not binary compatible with code generated without that switch.
10196 Additionally, it makes the code suboptimial.
10197 Use it to conform to a non-default application binary interface.
10199 @item -finstrument-functions
10200 @opindex finstrument-functions
10201 Generate instrumentation calls for entry and exit to functions. Just
10202 after function entry and just before function exit, the following
10203 profiling functions will be called with the address of the current
10204 function and its call site. (On some platforms,
10205 @code{__builtin_return_address} does not work beyond the current
10206 function, so the call site information may not be available to the
10207 profiling functions otherwise.)
10210 void __cyg_profile_func_enter (void *this_fn,
10212 void __cyg_profile_func_exit (void *this_fn,
10216 The first argument is the address of the start of the current function,
10217 which may be looked up exactly in the symbol table.
10219 This instrumentation is also done for functions expanded inline in other
10220 functions. The profiling calls will indicate where, conceptually, the
10221 inline function is entered and exited. This means that addressable
10222 versions of such functions must be available. If all your uses of a
10223 function are expanded inline, this may mean an additional expansion of
10224 code size. If you use @samp{extern inline} in your C code, an
10225 addressable version of such functions must be provided. (This is
10226 normally the case anyways, but if you get lucky and the optimizer always
10227 expands the functions inline, you might have gotten away without
10228 providing static copies.)
10230 A function may be given the attribute @code{no_instrument_function}, in
10231 which case this instrumentation will not be done. This can be used, for
10232 example, for the profiling functions listed above, high-priority
10233 interrupt routines, and any functions from which the profiling functions
10234 cannot safely be called (perhaps signal handlers, if the profiling
10235 routines generate output or allocate memory).
10237 @item -fstack-check
10238 @opindex fstack-check
10239 Generate code to verify that you do not go beyond the boundary of the
10240 stack. You should specify this flag if you are running in an
10241 environment with multiple threads, but only rarely need to specify it in
10242 a single-threaded environment since stack overflow is automatically
10243 detected on nearly all systems if there is only one stack.
10245 Note that this switch does not actually cause checking to be done; the
10246 operating system must do that. The switch causes generation of code
10247 to ensure that the operating system sees the stack being extended.
10249 @item -fstack-limit-register=@var{reg}
10250 @itemx -fstack-limit-symbol=@var{sym}
10251 @itemx -fno-stack-limit
10252 @opindex fstack-limit-register
10253 @opindex fstack-limit-symbol
10254 @opindex fno-stack-limit
10255 Generate code to ensure that the stack does not grow beyond a certain value,
10256 either the value of a register or the address of a symbol. If the stack
10257 would grow beyond the value, a signal is raised. For most targets,
10258 the signal is raised before the stack overruns the boundary, so
10259 it is possible to catch the signal without taking special precautions.
10261 For instance, if the stack starts at absolute address @samp{0x80000000}
10262 and grows downwards, you can use the flags
10263 @option{-fstack-limit-symbol=__stack_limit} and
10264 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
10265 of 128KB@. Note that this may only work with the GNU linker.
10267 @cindex aliasing of parameters
10268 @cindex parameters, aliased
10269 @item -fargument-alias
10270 @itemx -fargument-noalias
10271 @itemx -fargument-noalias-global
10272 @opindex fargument-alias
10273 @opindex fargument-noalias
10274 @opindex fargument-noalias-global
10275 Specify the possible relationships among parameters and between
10276 parameters and global data.
10278 @option{-fargument-alias} specifies that arguments (parameters) may
10279 alias each other and may alias global storage.@*
10280 @option{-fargument-noalias} specifies that arguments do not alias
10281 each other, but may alias global storage.@*
10282 @option{-fargument-noalias-global} specifies that arguments do not
10283 alias each other and do not alias global storage.
10285 Each language will automatically use whatever option is required by
10286 the language standard. You should not need to use these options yourself.
10288 @item -fleading-underscore
10289 @opindex fleading-underscore
10290 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
10291 change the way C symbols are represented in the object file. One use
10292 is to help link with legacy assembly code.
10294 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
10295 generate code that is not binary compatible with code generated without that
10296 switch. Use it to conform to a non-default application binary interface.
10297 Not all targets provide complete support for this switch.
10299 @item -ftls-model=@var{model}
10300 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
10301 The @var{model} argument should be one of @code{global-dynamic},
10302 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
10304 The default without @option{-fpic} is @code{initial-exec}; with
10305 @option{-fpic} the default is @code{global-dynamic}.
10310 @node Environment Variables
10311 @section Environment Variables Affecting GCC
10312 @cindex environment variables
10314 @c man begin ENVIRONMENT
10316 This section describes several environment variables that affect how GCC
10317 operates. Some of them work by specifying directories or prefixes to use
10318 when searching for various kinds of files. Some are used to specify other
10319 aspects of the compilation environment.
10321 Note that you can also specify places to search using options such as
10322 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
10323 take precedence over places specified using environment variables, which
10324 in turn take precedence over those specified by the configuration of GCC@.
10325 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
10326 GNU Compiler Collection (GCC) Internals}.
10331 @c @itemx LC_COLLATE
10333 @c @itemx LC_MONETARY
10334 @c @itemx LC_NUMERIC
10339 @c @findex LC_COLLATE
10340 @findex LC_MESSAGES
10341 @c @findex LC_MONETARY
10342 @c @findex LC_NUMERIC
10346 These environment variables control the way that GCC uses
10347 localization information that allow GCC to work with different
10348 national conventions. GCC inspects the locale categories
10349 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
10350 so. These locale categories can be set to any value supported by your
10351 installation. A typical value is @samp{en_UK} for English in the United
10354 The @env{LC_CTYPE} environment variable specifies character
10355 classification. GCC uses it to determine the character boundaries in
10356 a string; this is needed for some multibyte encodings that contain quote
10357 and escape characters that would otherwise be interpreted as a string
10360 The @env{LC_MESSAGES} environment variable specifies the language to
10361 use in diagnostic messages.
10363 If the @env{LC_ALL} environment variable is set, it overrides the value
10364 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10365 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10366 environment variable. If none of these variables are set, GCC
10367 defaults to traditional C English behavior.
10371 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10372 files. GCC uses temporary files to hold the output of one stage of
10373 compilation which is to be used as input to the next stage: for example,
10374 the output of the preprocessor, which is the input to the compiler
10377 @item GCC_EXEC_PREFIX
10378 @findex GCC_EXEC_PREFIX
10379 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10380 names of the subprograms executed by the compiler. No slash is added
10381 when this prefix is combined with the name of a subprogram, but you can
10382 specify a prefix that ends with a slash if you wish.
10384 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10385 an appropriate prefix to use based on the pathname it was invoked with.
10387 If GCC cannot find the subprogram using the specified prefix, it
10388 tries looking in the usual places for the subprogram.
10390 The default value of @env{GCC_EXEC_PREFIX} is
10391 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10392 of @code{prefix} when you ran the @file{configure} script.
10394 Other prefixes specified with @option{-B} take precedence over this prefix.
10396 This prefix is also used for finding files such as @file{crt0.o} that are
10399 In addition, the prefix is used in an unusual way in finding the
10400 directories to search for header files. For each of the standard
10401 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10402 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10403 replacing that beginning with the specified prefix to produce an
10404 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10405 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10406 These alternate directories are searched first; the standard directories
10409 @item COMPILER_PATH
10410 @findex COMPILER_PATH
10411 The value of @env{COMPILER_PATH} is a colon-separated list of
10412 directories, much like @env{PATH}. GCC tries the directories thus
10413 specified when searching for subprograms, if it can't find the
10414 subprograms using @env{GCC_EXEC_PREFIX}.
10417 @findex LIBRARY_PATH
10418 The value of @env{LIBRARY_PATH} is a colon-separated list of
10419 directories, much like @env{PATH}. When configured as a native compiler,
10420 GCC tries the directories thus specified when searching for special
10421 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10422 using GCC also uses these directories when searching for ordinary
10423 libraries for the @option{-l} option (but directories specified with
10424 @option{-L} come first).
10428 @cindex locale definition
10429 This variable is used to pass locale information to the compiler. One way in
10430 which this information is used is to determine the character set to be used
10431 when character literals, string literals and comments are parsed in C and C++.
10432 When the compiler is configured to allow multibyte characters,
10433 the following values for @env{LANG} are recognized:
10437 Recognize JIS characters.
10439 Recognize SJIS characters.
10441 Recognize EUCJP characters.
10444 If @env{LANG} is not defined, or if it has some other value, then the
10445 compiler will use mblen and mbtowc as defined by the default locale to
10446 recognize and translate multibyte characters.
10450 Some additional environments variables affect the behavior of the
10453 @include cppenv.texi
10457 @node Running Protoize
10458 @section Running Protoize
10460 The program @code{protoize} is an optional part of GCC@. You can use
10461 it to add prototypes to a program, thus converting the program to ISO
10462 C in one respect. The companion program @code{unprotoize} does the
10463 reverse: it removes argument types from any prototypes that are found.
10465 When you run these programs, you must specify a set of source files as
10466 command line arguments. The conversion programs start out by compiling
10467 these files to see what functions they define. The information gathered
10468 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10470 After scanning comes actual conversion. The specified files are all
10471 eligible to be converted; any files they include (whether sources or
10472 just headers) are eligible as well.
10474 But not all the eligible files are converted. By default,
10475 @code{protoize} and @code{unprotoize} convert only source and header
10476 files in the current directory. You can specify additional directories
10477 whose files should be converted with the @option{-d @var{directory}}
10478 option. You can also specify particular files to exclude with the
10479 @option{-x @var{file}} option. A file is converted if it is eligible, its
10480 directory name matches one of the specified directory names, and its
10481 name within the directory has not been excluded.
10483 Basic conversion with @code{protoize} consists of rewriting most
10484 function definitions and function declarations to specify the types of
10485 the arguments. The only ones not rewritten are those for varargs
10488 @code{protoize} optionally inserts prototype declarations at the
10489 beginning of the source file, to make them available for any calls that
10490 precede the function's definition. Or it can insert prototype
10491 declarations with block scope in the blocks where undeclared functions
10494 Basic conversion with @code{unprotoize} consists of rewriting most
10495 function declarations to remove any argument types, and rewriting
10496 function definitions to the old-style pre-ISO form.
10498 Both conversion programs print a warning for any function declaration or
10499 definition that they can't convert. You can suppress these warnings
10502 The output from @code{protoize} or @code{unprotoize} replaces the
10503 original source file. The original file is renamed to a name ending
10504 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10505 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10506 for DOS) file already exists, then the source file is simply discarded.
10508 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10509 scan the program and collect information about the functions it uses.
10510 So neither of these programs will work until GCC is installed.
10512 Here is a table of the options you can use with @code{protoize} and
10513 @code{unprotoize}. Each option works with both programs unless
10517 @item -B @var{directory}
10518 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10519 usual directory (normally @file{/usr/local/lib}). This file contains
10520 prototype information about standard system functions. This option
10521 applies only to @code{protoize}.
10523 @item -c @var{compilation-options}
10524 Use @var{compilation-options} as the options when running @code{gcc} to
10525 produce the @samp{.X} files. The special option @option{-aux-info} is
10526 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10528 Note that the compilation options must be given as a single argument to
10529 @code{protoize} or @code{unprotoize}. If you want to specify several
10530 @code{gcc} options, you must quote the entire set of compilation options
10531 to make them a single word in the shell.
10533 There are certain @code{gcc} arguments that you cannot use, because they
10534 would produce the wrong kind of output. These include @option{-g},
10535 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10536 the @var{compilation-options}, they are ignored.
10539 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10540 systems) instead of @samp{.c}. This is convenient if you are converting
10541 a C program to C++. This option applies only to @code{protoize}.
10544 Add explicit global declarations. This means inserting explicit
10545 declarations at the beginning of each source file for each function
10546 that is called in the file and was not declared. These declarations
10547 precede the first function definition that contains a call to an
10548 undeclared function. This option applies only to @code{protoize}.
10550 @item -i @var{string}
10551 Indent old-style parameter declarations with the string @var{string}.
10552 This option applies only to @code{protoize}.
10554 @code{unprotoize} converts prototyped function definitions to old-style
10555 function definitions, where the arguments are declared between the
10556 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10557 uses five spaces as the indentation. If you want to indent with just
10558 one space instead, use @option{-i " "}.
10561 Keep the @samp{.X} files. Normally, they are deleted after conversion
10565 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10566 a prototype declaration for each function in each block which calls the
10567 function without any declaration. This option applies only to
10571 Make no real changes. This mode just prints information about the conversions
10572 that would have been done without @option{-n}.
10575 Make no @samp{.save} files. The original files are simply deleted.
10576 Use this option with caution.
10578 @item -p @var{program}
10579 Use the program @var{program} as the compiler. Normally, the name
10580 @file{gcc} is used.
10583 Work quietly. Most warnings are suppressed.
10586 Print the version number, just like @option{-v} for @code{gcc}.
10589 If you need special compiler options to compile one of your program's
10590 source files, then you should generate that file's @samp{.X} file
10591 specially, by running @code{gcc} on that source file with the
10592 appropriate options and the option @option{-aux-info}. Then run
10593 @code{protoize} on the entire set of files. @code{protoize} will use
10594 the existing @samp{.X} file because it is newer than the source file.
10598 gcc -Dfoo=bar file1.c -aux-info file1.X
10603 You need to include the special files along with the rest in the
10604 @code{protoize} command, even though their @samp{.X} files already
10605 exist, because otherwise they won't get converted.
10607 @xref{Protoize Caveats}, for more information on how to use
10608 @code{protoize} successfully.