1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002 Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
8 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
9 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
11 Permission is granted to copy, distribute and/or modify this document
12 under the terms of the GNU Free Documentation License, Version 1.1 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU project C and C++ compiler
33 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
39 [@option{-o} @var{outfile}] @var{infile}@dots{}
41 Only the most useful options are listed here; see below for the
42 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
45 gpl(7), gfdl(7), fsf-funding(7),
46 cpp(1), gcov(1), g77(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
47 and the Info entries for @file{gcc}, @file{cpp}, @file{g77}, @file{as},
48 @file{ld}, @file{binutils} and @file{gdb}.
51 For instructions on reporting bugs, see
52 @w{@uref{http://gcc.gnu.org/bugs.html}}. Use of the @command{gccbug}
53 script to report bugs is recommended.
56 See the Info entry for @command{gcc}, or
57 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
58 for contributors to GCC@.
63 @chapter GCC Command Options
64 @cindex GCC command options
65 @cindex command options
66 @cindex options, GCC command
68 @c man begin DESCRIPTION
70 When you invoke GCC, it normally does preprocessing, compilation,
71 assembly and linking. The ``overall options'' allow you to stop this
72 process at an intermediate stage. For example, the @option{-c} option
73 says not to run the linker. Then the output consists of object files
74 output by the assembler.
76 Other options are passed on to one stage of processing. Some options
77 control the preprocessor and others the compiler itself. Yet other
78 options control the assembler and linker; most of these are not
79 documented here, since you rarely need to use any of them.
81 @cindex C compilation options
82 Most of the command line options that you can use with GCC are useful
83 for C programs; when an option is only useful with another language
84 (usually C++), the explanation says so explicitly. If the description
85 for a particular option does not mention a source language, you can use
86 that option with all supported languages.
88 @cindex C++ compilation options
89 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
90 options for compiling C++ programs.
92 @cindex grouping options
93 @cindex options, grouping
94 The @command{gcc} program accepts options and file names as operands. Many
95 options have multi-letter names; therefore multiple single-letter options
96 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
99 @cindex order of options
100 @cindex options, order
101 You can mix options and other arguments. For the most part, the order
102 you use doesn't matter. Order does matter when you use several options
103 of the same kind; for example, if you specify @option{-L} more than once,
104 the directories are searched in the order specified.
106 Many options have long names starting with @samp{-f} or with
107 @samp{-W}---for example, @option{-fforce-mem},
108 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
109 these have both positive and negative forms; the negative form of
110 @option{-ffoo} would be @option{-fno-foo}. This manual documents
111 only one of these two forms, whichever one is not the default.
115 @xref{Option Index}, for an index to GCC's options.
118 * Option Summary:: Brief list of all options, without explanations.
119 * Overall Options:: Controlling the kind of output:
120 an executable, object files, assembler files,
121 or preprocessed source.
122 * Invoking G++:: Compiling C++ programs.
123 * C Dialect Options:: Controlling the variant of C language compiled.
124 * C++ Dialect Options:: Variations on C++.
125 * Objective-C Dialect Options:: Variations on Objective-C.
126 * Language Independent Options:: Controlling how diagnostics should be
128 * Warning Options:: How picky should the compiler be?
129 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
130 * Optimize Options:: How much optimization?
131 * Preprocessor Options:: Controlling header files and macro definitions.
132 Also, getting dependency information for Make.
133 * Assembler Options:: Passing options to the assembler.
134 * Link Options:: Specifying libraries and so on.
135 * Directory Options:: Where to find header files and libraries.
136 Where to find the compiler executable files.
137 * Spec Files:: How to pass switches to sub-processes.
138 * Target Options:: Running a cross-compiler, or an old version of GCC.
139 * Submodel Options:: Specifying minor hardware or convention variations,
140 such as 68010 vs 68020.
141 * Code Gen Options:: Specifying conventions for function calls, data layout
143 * Environment Variables:: Env vars that affect GCC.
144 * Running Protoize:: Automatically adding or removing function prototypes.
150 @section Option Summary
152 Here is a summary of all the options, grouped by type. Explanations are
153 in the following sections.
156 @item Overall Options
157 @xref{Overall Options,,Options Controlling the Kind of Output}.
159 -c -S -E -o @var{file} -pipe -pass-exit-codes -x @var{language} @gol
160 -v -### --target-help --help}
162 @item C Language Options
163 @xref{C Dialect Options,,Options Controlling C Dialect}.
165 -ansi -std=@var{standard} -aux-info @var{filename} @gol
166 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
167 -fhosted -ffreestanding @gol
168 -trigraphs -traditional -traditional-cpp @gol
169 -fallow-single-precision -fcond-mismatch @gol
170 -fsigned-bitfields -fsigned-char @gol
171 -funsigned-bitfields -funsigned-char @gol
172 -fwritable-strings -fshort-wchar}
174 @item C++ Language Options
175 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
177 -fno-access-control -fcheck-new -fconserve-space @gol
178 -fno-const-strings -fdollars-in-identifiers @gol
179 -fno-elide-constructors @gol
180 -fno-enforce-eh-specs -fexternal-templates @gol
181 -falt-external-templates @gol
182 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
183 -fno-implicit-templates @gol
184 -fno-implicit-inline-templates @gol
185 -fno-implement-inlines -fms-extensions @gol
186 -fno-nonansi-builtins -fno-operator-names @gol
187 -fno-optional-diags -fpermissive @gol
188 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
189 -fuse-cxa-atexit -fvtable-gc -fno-weak -nostdinc++ @gol
190 -fno-default-inline -Wctor-dtor-privacy @gol
191 -Wnon-virtual-dtor -Wreorder @gol
192 -Weffc++ -Wno-deprecated @gol
193 -Wno-non-template-friend -Wold-style-cast @gol
194 -Woverloaded-virtual -Wno-pmf-conversions @gol
195 -Wsign-promo -Wsynth}
197 @item Objective-C Language Options
198 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
200 -fconstant-string-class=@var{class-name} @gol
201 -fgnu-runtime -fnext-runtime -gen-decls @gol
202 -Wno-protocol -Wselector}
204 @item Language Independent Options
205 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
207 -fmessage-length=@var{n} @gol
208 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
210 @item Warning Options
211 @xref{Warning Options,,Options to Request or Suppress Warnings}.
213 -fsyntax-only -pedantic -pedantic-errors @gol
214 -w -W -Wall -Waggregate-return @gol
215 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
216 -Wconversion -Wno-deprecated-declarations @gol
217 -Wdisabled-optimization -Wno-div-by-zero -Werror @gol
218 -Wfloat-equal -Wformat -Wformat=2 @gol
219 -Wformat-nonliteral -Wformat-security @gol
220 -Wimplicit -Wimplicit-int @gol
221 -Wimplicit-function-declaration @gol
222 -Werror-implicit-function-declaration @gol
223 -Wimport -Winline -Wno-endif-labels @gol
224 -Wlarger-than-@var{len} -Wlong-long @gol
225 -Wmain -Wmissing-braces -Wmissing-declarations @gol
226 -Wmissing-format-attribute -Wmissing-noreturn @gol
227 -Wno-multichar -Wno-format-extra-args -Wno-format-y2k @gol
228 -Wno-import -Wpacked -Wpadded @gol
229 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
230 -Wreturn-type -Wsequence-point -Wshadow @gol
231 -Wsign-compare -Wswitch -Wswitch-default -Wswitch-enum @gol
232 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
233 -Wunknown-pragmas -Wunreachable-code @gol
234 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
235 -Wunused-value -Wunused-variable -Wwrite-strings}
237 @item C-only Warning Options
239 -Wbad-function-cast -Wmissing-prototypes -Wnested-externs @gol
240 -Wstrict-prototypes -Wtraditional}
242 @item Debugging Options
243 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
245 -d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
246 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
247 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
248 -fdump-tree-original@r{[}-@var{n}@r{]} -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
249 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
251 -fprofile-arcs -ftest-coverage -ftime-report @gol
252 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
253 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
254 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
255 -print-multi-directory -print-multi-lib @gol
256 -print-prog-name=@var{program} -print-search-dirs -Q @gol
259 @item Optimization Options
260 @xref{Optimize Options,,Options that Control Optimization}.
262 -falign-functions=@var{n} -falign-jumps=@var{n} @gol
263 -falign-labels=@var{n} -falign-loops=@var{n} @gol
264 -fbranch-probabilities -fcaller-saves -fcprop-registers @gol
265 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
266 -fdelayed-branch -fdelete-null-pointer-checks @gol
267 -fexpensive-optimizations -ffast-math -ffloat-store @gol
268 -fforce-addr -fforce-mem -ffunction-sections @gol
269 -fgcse -fgcse-lm -fgcse-sm -floop-optimize -fcrossjumping @gol
270 -fif-conversion -fif-conversion2 @gol
271 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
272 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
273 -fmove-all-movables -fno-default-inline -fno-defer-pop @gol
274 -fno-function-cse -fno-guess-branch-probability @gol
275 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
276 -funsafe-math-optimizations -fno-trapping-math @gol
277 -fno-zero-initialized-in-bss @gol
278 -fomit-frame-pointer -foptimize-register-move @gol
279 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
280 -freduce-all-givs -fregmove -frename-registers @gol
281 -frerun-cse-after-loop -frerun-loop-opt @gol
282 -fschedule-insns -fschedule-insns2 @gol
283 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
284 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv @gol
285 -funroll-all-loops -funroll-loops @gol
286 --param @var{name}=@var{value}
287 -O -O0 -O1 -O2 -O3 -Os}
289 @item Preprocessor Options
290 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
292 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
293 -C -dD -dI -dM -dN @gol
294 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
295 -idirafter @var{dir} @gol
296 -include @var{file} -imacros @var{file} @gol
297 -iprefix @var{file} -iwithprefix @var{dir} @gol
298 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
299 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
300 -trigraphs -undef -U@var{macro} -Wp,@var{option}}
302 @item Assembler Option
303 @xref{Assembler Options,,Passing Options to the Assembler}.
308 @xref{Link Options,,Options for Linking}.
310 @var{object-file-name} -l@var{library} @gol
311 -nostartfiles -nodefaultlibs -nostdlib @gol
312 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
313 -Wl,@var{option} -Xlinker @var{option} @gol
316 @item Directory Options
317 @xref{Directory Options,,Options for Directory Search}.
319 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
322 @c I wrote this xref this way to avoid overfull hbox. -- rms
323 @xref{Target Options}.
327 @item Machine Dependent Options
328 @xref{Submodel Options,,Hardware Models and Configurations}.
330 @emph{M680x0 Options}
332 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
333 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
334 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
335 -malign-int -mstrict-align}
337 @emph{M68hc1x Options}
339 -m6811 -m6812 -m68hc11 -m68hc12 @gol
340 -mauto-incdec -mshort -msoft-reg-count=@var{count}}
348 -mcpu=@var{cpu-type} @gol
349 -mtune=@var{cpu-type} @gol
350 -mcmodel=@var{code-model} @gol
352 -mapp-regs -mbroken-saverestore -mcypress @gol
353 -mfaster-structs -mflat @gol
354 -mfpu -mhard-float -mhard-quad-float @gol
355 -mimpure-text -mlive-g0 -mno-app-regs @gol
356 -mno-faster-structs -mno-flat -mno-fpu @gol
357 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
358 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
359 -msupersparc -munaligned-doubles -mv8}
361 @emph{Convex Options}
363 -mc1 -mc2 -mc32 -mc34 -mc38 @gol
364 -margcount -mnoargcount @gol
365 -mlong32 -mlong64 @gol
366 -mvolatile-cache -mvolatile-nocache}
368 @emph{AMD29K Options}
370 -m29000 -m29050 -mbw -mnbw -mdw -mndw @gol
371 -mlarge -mnormal -msmall @gol
372 -mkernel-registers -mno-reuse-arg-regs @gol
373 -mno-stack-check -mno-storem-bug @gol
374 -mreuse-arg-regs -msoft-float -mstack-check @gol
375 -mstorem-bug -muser-registers}
379 -mapcs-frame -mno-apcs-frame @gol
380 -mapcs-26 -mapcs-32 @gol
381 -mapcs-stack-check -mno-apcs-stack-check @gol
382 -mapcs-float -mno-apcs-float @gol
383 -mapcs-reentrant -mno-apcs-reentrant @gol
384 -msched-prolog -mno-sched-prolog @gol
385 -mlittle-endian -mbig-endian -mwords-little-endian @gol
386 -malignment-traps -mno-alignment-traps @gol
387 -msoft-float -mhard-float -mfpe @gol
388 -mthumb-interwork -mno-thumb-interwork @gol
389 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
390 -mstructure-size-boundary=@var{n} @gol
391 -mbsd -mxopen -mno-symrename @gol
392 -mabort-on-noreturn @gol
393 -mlong-calls -mno-long-calls @gol
394 -msingle-pic-base -mno-single-pic-base @gol
395 -mpic-register=@var{reg} @gol
396 -mnop-fun-dllimport @gol
397 -mpoke-function-name @gol
399 -mtpcs-frame -mtpcs-leaf-frame @gol
400 -mcaller-super-interworking -mcallee-super-interworking }
402 @emph{MN10200 Options}
406 @emph{MN10300 Options}
408 -mmult-bug -mno-mult-bug @gol
409 -mam33 -mno-am33 @gol
412 @emph{M32R/D Options}
414 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
419 -m88000 -m88100 -m88110 -mbig-pic @gol
420 -mcheck-zero-division -mhandle-large-shift @gol
421 -midentify-revision -mno-check-zero-division @gol
422 -mno-ocs-debug-info -mno-ocs-frame-position @gol
423 -mno-optimize-arg-area -mno-serialize-volatile @gol
424 -mno-underscores -mocs-debug-info @gol
425 -mocs-frame-position -moptimize-arg-area @gol
426 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
427 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
428 -mversion-03.00 -mwarn-passed-structs}
430 @emph{RS/6000 and PowerPC Options}
432 -mcpu=@var{cpu-type} @gol
433 -mtune=@var{cpu-type} @gol
434 -mpower -mno-power -mpower2 -mno-power2 @gol
435 -mpowerpc -mpowerpc64 -mno-powerpc @gol
436 -maltivec -mno-altivec @gol
437 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
438 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
439 -mnew-mnemonics -mold-mnemonics @gol
440 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
441 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
442 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
443 -mstring -mno-string -mupdate -mno-update @gol
444 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
445 -mstrict-align -mno-strict-align -mrelocatable @gol
446 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
447 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
448 -mcall-aix -mcall-sysv -mcall-netbsd @gol
449 -maix-struct-return -msvr4-struct-return
450 -mabi=altivec -mabi=no-altivec @gol
451 -mprototype -mno-prototype @gol
452 -msim -mmvme -mads -myellowknife -memb -msdata @gol
453 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
457 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
458 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
459 -mminimum-fp-blocks -mnohc-struct-return}
463 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
464 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
465 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
466 -mgas -mgp32 -mgp64 @gol
467 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
468 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
469 -mmips-as -mmips-tfile -mno-abicalls @gol
470 -mno-embedded-data -mno-uninit-const-in-rodata @gol
471 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
472 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
473 -mrnames -msoft-float @gol
474 -m4650 -msingle-float -mmad @gol
475 -mstats -EL -EB -G @var{num} -nocpp @gol
476 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
477 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func}
479 @emph{i386 and x86-64 Options}
481 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
482 -masm=@var{dialect} -mno-fancy-math-387 @gol
483 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
484 -mno-wide-multiply -mrtd -malign-double @gol
485 -mpreferred-stack-boundary=@var{num} @gol
486 -mmmx -msse -msse2 -msse-math -m3dnow @gol
487 -mthreads -mno-align-stringops -minline-all-stringops @gol
488 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
489 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
495 -march=@var{architecture-type} @gol
496 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
497 -mfast-indirect-calls -mgas -mjump-in-delay @gol
498 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
499 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
500 -mno-jump-in-delay -mno-long-load-store @gol
501 -mno-portable-runtime -mno-soft-float @gol
502 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
503 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
504 -mschedule=@var{cpu-type} -mspace-regs}
506 @emph{Intel 960 Options}
508 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
509 -mcode-align -mcomplex-addr -mleaf-procedures @gol
510 -mic-compat -mic2.0-compat -mic3.0-compat @gol
511 -mintel-asm -mno-clean-linkage -mno-code-align @gol
512 -mno-complex-addr -mno-leaf-procedures @gol
513 -mno-old-align -mno-strict-align -mno-tail-call @gol
514 -mnumerics -mold-align -msoft-float -mstrict-align @gol
517 @emph{DEC Alpha Options}
519 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
520 -mieee -mieee-with-inexact -mieee-conformant @gol
521 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
522 -mtrap-precision=@var{mode} -mbuild-constants @gol
523 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
524 -mbwx -mmax -mfix -mcix @gol
525 -mfloat-vax -mfloat-ieee @gol
526 -mexplicit-relocs -msmall-data -mlarge-data @gol
527 -mmemory-latency=@var{time}}
529 @emph{DEC Alpha/VMS Options}
533 @emph{Clipper Options}
537 @emph{H8/300 Options}
539 -mrelax -mh -ms -mint32 -malign-300}
543 -m1 -m2 -m3 -m3e @gol
544 -m4-nofpu -m4-single-only -m4-single -m4 @gol
545 -m5-64media -m5-64media-nofpu @gol
546 -m5-32media -m5-32media-nofpu @gol
547 -m5-compact -m5-compact-nofpu @gol
548 -mb -ml -mdalign -mrelax @gol
549 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
550 -mieee -misize -mpadstruct -mspace @gol
551 -mprefergot -musermode}
553 @emph{System V Options}
555 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
560 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
561 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
563 @emph{TMS320C3x/C4x Options}
565 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
566 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
567 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
568 -mparallel-insns -mparallel-mpy -mpreserve-float}
572 -mlong-calls -mno-long-calls -mep -mno-ep @gol
573 -mprolog-function -mno-prolog-function -mspace @gol
574 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
579 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
580 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
581 -mregparam -mnoregparam -msb -mnosb @gol
582 -mbitfield -mnobitfield -mhimem -mnohimem}
586 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
587 -mcall-prologues -mno-tablejump -mtiny-stack}
591 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
592 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
593 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
594 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
595 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
599 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
600 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
601 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
606 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
607 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
608 -mconstant-gp -mauto-pic -minline-divide-min-latency @gol
609 -minline-divide-max-throughput -mno-dwarf2-asm @gol
610 -mfixed-range=@var{register-range}}
614 -mextmem -mextmemory -monchip -mno-asm-optimize -masm-optimize @gol
615 -mbranch-cost=@var{n} -mcond-exec=@var{n}}
617 @emph{S/390 and zSeries Options}
619 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
620 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
621 -m64 -m31 -mdebug -mno-debug}
625 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
626 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
627 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
628 -mstack-align -mdata-align -mconst-align @gol
629 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
630 -melf -maout -melinux -mlinux -sim -sim2}
632 @emph{PDP-11 Options}
634 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
635 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
636 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
637 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
638 -mbranch-expensive -mbranch-cheap @gol
639 -msplit -mno-split -munix-asm -mdec-asm}
641 @emph{Xstormy16 Options}
645 @emph{Xtensa Options}
647 -mbig-endian -mlittle-endian @gol
648 -mdensity -mno-density @gol
649 -mmac16 -mno-mac16 @gol
650 -mmul16 -mno-mul16 @gol
651 -mmul32 -mno-mul32 @gol
653 -mminmax -mno-minmax @gol
654 -msext -mno-sext @gol
655 -mbooleans -mno-booleans @gol
656 -mhard-float -msoft-float @gol
657 -mfused-madd -mno-fused-madd @gol
658 -mserialize-volatile -mno-serialize-volatile @gol
659 -mtext-section-literals -mno-text-section-literals @gol
660 -mtarget-align -mno-target-align @gol
661 -mlongcalls -mno-longcalls}
663 @item Code Generation Options
664 @xref{Code Gen Options,,Options for Code Generation Conventions}.
666 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
667 -ffixed-@var{reg} -fexceptions @gol
668 -fnon-call-exceptions -funwind-tables @gol
669 -fasynchronous-unwind-tables @gol
670 -finhibit-size-directive -finstrument-functions @gol
671 -fno-common -fno-ident -fno-gnu-linker @gol
672 -fpcc-struct-return -fpic -fPIC @gol
673 -freg-struct-return -fshared-data -fshort-enums @gol
674 -fshort-double -fvolatile @gol
675 -fvolatile-global -fvolatile-static @gol
676 -fverbose-asm -fpack-struct -fstack-check @gol
677 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
678 -fargument-alias -fargument-noalias @gol
679 -fargument-noalias-global -fleading-underscore}
683 * Overall Options:: Controlling the kind of output:
684 an executable, object files, assembler files,
685 or preprocessed source.
686 * C Dialect Options:: Controlling the variant of C language compiled.
687 * C++ Dialect Options:: Variations on C++.
688 * Objective-C Dialect Options:: Variations on Objective-C.
689 * Language Independent Options:: Controlling how diagnostics should be
691 * Warning Options:: How picky should the compiler be?
692 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
693 * Optimize Options:: How much optimization?
694 * Preprocessor Options:: Controlling header files and macro definitions.
695 Also, getting dependency information for Make.
696 * Assembler Options:: Passing options to the assembler.
697 * Link Options:: Specifying libraries and so on.
698 * Directory Options:: Where to find header files and libraries.
699 Where to find the compiler executable files.
700 * Spec Files:: How to pass switches to sub-processes.
701 * Target Options:: Running a cross-compiler, or an old version of GCC.
704 @node Overall Options
705 @section Options Controlling the Kind of Output
707 Compilation can involve up to four stages: preprocessing, compilation
708 proper, assembly and linking, always in that order. The first three
709 stages apply to an individual source file, and end by producing an
710 object file; linking combines all the object files (those newly
711 compiled, and those specified as input) into an executable file.
713 @cindex file name suffix
714 For any given input file, the file name suffix determines what kind of
719 C source code which must be preprocessed.
722 C source code which should not be preprocessed.
725 C++ source code which should not be preprocessed.
728 Objective-C source code. Note that you must link with the library
729 @file{libobjc.a} to make an Objective-C program work.
732 Objective-C source code which should not be preprocessed.
735 C header file (not to be compiled or linked).
739 @itemx @var{file}.cxx
740 @itemx @var{file}.cpp
741 @itemx @var{file}.c++
743 C++ source code which must be preprocessed. Note that in @samp{.cxx},
744 the last two letters must both be literally @samp{x}. Likewise,
745 @samp{.C} refers to a literal capital C@.
748 @itemx @var{file}.for
749 @itemx @var{file}.FOR
750 Fortran source code which should not be preprocessed.
753 @itemx @var{file}.fpp
754 @itemx @var{file}.FPP
755 Fortran source code which must be preprocessed (with the traditional
759 Fortran source code which must be preprocessed with a RATFOR
760 preprocessor (not included with GCC)@.
762 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
763 Using and Porting GNU Fortran}, for more details of the handling of
766 @c FIXME: Descriptions of Java file types.
773 Ada source code file which contains a library unit declaration (a
774 declaration of a package, subprogram, or generic, or a generic
775 instantiation), or a library unit renaming declaration (a package,
776 generic, or subprogram renaming declaration). Such files are also
779 @itemx @var{file}.adb
780 Ada source code file containing a library unit body (a subprogram or
781 package body). Such files are also called @dfn{bodies}.
783 @c GCC also knows about some suffixes for languages not yet included:
792 Assembler code which must be preprocessed.
795 An object file to be fed straight into linking.
796 Any file name with no recognized suffix is treated this way.
800 You can specify the input language explicitly with the @option{-x} option:
803 @item -x @var{language}
804 Specify explicitly the @var{language} for the following input files
805 (rather than letting the compiler choose a default based on the file
806 name suffix). This option applies to all following input files until
807 the next @option{-x} option. Possible values for @var{language} are:
809 c c-header cpp-output
811 objective-c objc-cpp-output
812 assembler assembler-with-cpp
814 f77 f77-cpp-input ratfor
819 Turn off any specification of a language, so that subsequent files are
820 handled according to their file name suffixes (as they are if @option{-x}
821 has not been used at all).
823 @item -pass-exit-codes
824 @opindex pass-exit-codes
825 Normally the @command{gcc} program will exit with the code of 1 if any
826 phase of the compiler returns a non-success return code. If you specify
827 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
828 numerically highest error produced by any phase that returned an error
832 If you only want some of the stages of compilation, you can use
833 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
834 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
835 @command{gcc} is to stop. Note that some combinations (for example,
836 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
841 Compile or assemble the source files, but do not link. The linking
842 stage simply is not done. The ultimate output is in the form of an
843 object file for each source file.
845 By default, the object file name for a source file is made by replacing
846 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
848 Unrecognized input files, not requiring compilation or assembly, are
853 Stop after the stage of compilation proper; do not assemble. The output
854 is in the form of an assembler code file for each non-assembler input
857 By default, the assembler file name for a source file is made by
858 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
860 Input files that don't require compilation are ignored.
864 Stop after the preprocessing stage; do not run the compiler proper. The
865 output is in the form of preprocessed source code, which is sent to the
868 Input files which don't require preprocessing are ignored.
870 @cindex output file option
873 Place output in file @var{file}. This applies regardless to whatever
874 sort of output is being produced, whether it be an executable file,
875 an object file, an assembler file or preprocessed C code.
877 Since only one output file can be specified, it does not make sense to
878 use @option{-o} when compiling more than one input file, unless you are
879 producing an executable file as output.
881 If @option{-o} is not specified, the default is to put an executable file
882 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
883 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
884 all preprocessed C source on standard output.
888 Print (on standard error output) the commands executed to run the stages
889 of compilation. Also print the version number of the compiler driver
890 program and of the preprocessor and the compiler proper.
894 Like @option{-v} except the commands are not executed and all command
895 arguments are quoted. This is useful for shell scripts to capture the
896 driver-generated command lines.
900 Use pipes rather than temporary files for communication between the
901 various stages of compilation. This fails to work on some systems where
902 the assembler is unable to read from a pipe; but the GNU assembler has
907 Print (on the standard output) a description of the command line options
908 understood by @command{gcc}. If the @option{-v} option is also specified
909 then @option{--help} will also be passed on to the various processes
910 invoked by @command{gcc}, so that they can display the command line options
911 they accept. If the @option{-W} option is also specified then command
912 line options which have no documentation associated with them will also
917 Print (on the standard output) a description of target specific command
918 line options for each tool.
922 @section Compiling C++ Programs
924 @cindex suffixes for C++ source
925 @cindex C++ source file suffixes
926 C++ source files conventionally use one of the suffixes @samp{.C},
927 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
928 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
929 files with these names and compiles them as C++ programs even if you
930 call the compiler the same way as for compiling C programs (usually with
931 the name @command{gcc}).
935 However, C++ programs often require class libraries as well as a
936 compiler that understands the C++ language---and under some
937 circumstances, you might want to compile programs from standard input,
938 or otherwise without a suffix that flags them as C++ programs.
939 @command{g++} is a program that calls GCC with the default language
940 set to C++, and automatically specifies linking against the C++
941 library. On many systems, @command{g++} is also
942 installed with the name @command{c++}.
944 @cindex invoking @command{g++}
945 When you compile C++ programs, you may specify many of the same
946 command-line options that you use for compiling programs in any
947 language; or command-line options meaningful for C and related
948 languages; or options that are meaningful only for C++ programs.
949 @xref{C Dialect Options,,Options Controlling C Dialect}, for
950 explanations of options for languages related to C@.
951 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
952 explanations of options that are meaningful only for C++ programs.
954 @node C Dialect Options
955 @section Options Controlling C Dialect
956 @cindex dialect options
957 @cindex language dialect options
958 @cindex options, dialect
960 The following options control the dialect of C (or languages derived
961 from C, such as C++ and Objective-C) that the compiler accepts:
968 In C mode, support all ISO C89 programs. In C++ mode,
969 remove GNU extensions that conflict with ISO C++.
971 This turns off certain features of GCC that are incompatible with ISO
972 C89 (when compiling C code), or of standard C++ (when compiling C++ code),
973 such as the @code{asm} and @code{typeof} keywords, and
974 predefined macros such as @code{unix} and @code{vax} that identify the
975 type of system you are using. It also enables the undesirable and
976 rarely used ISO trigraph feature. For the C compiler,
977 it disables recognition of C++ style @samp{//} comments as well as
978 the @code{inline} keyword.
980 The alternate keywords @code{__asm__}, @code{__extension__},
981 @code{__inline__} and @code{__typeof__} continue to work despite
982 @option{-ansi}. You would not want to use them in an ISO C program, of
983 course, but it is useful to put them in header files that might be included
984 in compilations done with @option{-ansi}. Alternate predefined macros
985 such as @code{__unix__} and @code{__vax__} are also available, with or
986 without @option{-ansi}.
988 The @option{-ansi} option does not cause non-ISO programs to be
989 rejected gratuitously. For that, @option{-pedantic} is required in
990 addition to @option{-ansi}. @xref{Warning Options}.
992 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
993 option is used. Some header files may notice this macro and refrain
994 from declaring certain functions or defining certain macros that the
995 ISO standard doesn't call for; this is to avoid interfering with any
996 programs that might use these names for other things.
998 Functions which would normally be built in but do not have semantics
999 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1000 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1001 built-in functions provided by GCC}, for details of the functions
1006 Determine the language standard. This option is currently only
1007 supported when compiling C@. A value for this option must be provided;
1013 ISO C89 (same as @option{-ansi}).
1015 @item iso9899:199409
1016 ISO C89 as modified in amendment 1.
1022 ISO C99. Note that this standard is not yet fully supported; see
1023 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1024 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1027 Default, ISO C89 plus GNU extensions (including some C99 features).
1031 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1032 this will become the default. The name @samp{gnu9x} is deprecated.
1036 Even when this option is not specified, you can still use some of the
1037 features of newer standards in so far as they do not conflict with
1038 previous C standards. For example, you may use @code{__restrict__} even
1039 when @option{-std=c99} is not specified.
1041 The @option{-std} options specifying some version of ISO C have the same
1042 effects as @option{-ansi}, except that features that were not in ISO C89
1043 but are in the specified version (for example, @samp{//} comments and
1044 the @code{inline} keyword in ISO C99) are not disabled.
1046 @xref{Standards,,Language Standards Supported by GCC}, for details of
1047 these standard versions.
1049 @item -aux-info @var{filename}
1051 Output to the given filename prototyped declarations for all functions
1052 declared and/or defined in a translation unit, including those in header
1053 files. This option is silently ignored in any language other than C@.
1055 Besides declarations, the file indicates, in comments, the origin of
1056 each declaration (source file and line), whether the declaration was
1057 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1058 @samp{O} for old, respectively, in the first character after the line
1059 number and the colon), and whether it came from a declaration or a
1060 definition (@samp{C} or @samp{F}, respectively, in the following
1061 character). In the case of function definitions, a K&R-style list of
1062 arguments followed by their declarations is also provided, inside
1063 comments, after the declaration.
1067 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1068 keyword, so that code can use these words as identifiers. You can use
1069 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1070 instead. @option{-ansi} implies @option{-fno-asm}.
1072 In C++, this switch only affects the @code{typeof} keyword, since
1073 @code{asm} and @code{inline} are standard keywords. You may want to
1074 use the @option{-fno-gnu-keywords} flag instead, which has the same
1075 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1076 switch only affects the @code{asm} and @code{typeof} keywords, since
1077 @code{inline} is a standard keyword in ISO C99.
1080 @itemx -fno-builtin-@var{function} @r{(C and Objective-C only)}
1081 @opindex fno-builtin
1082 @cindex built-in functions
1083 Don't recognize built-in functions that do not begin with
1084 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1085 functions provided by GCC}, for details of the functions affected,
1086 including those which are not built-in functions when @option{-ansi} or
1087 @option{-std} options for strict ISO C conformance are used because they
1088 do not have an ISO standard meaning.
1090 GCC normally generates special code to handle certain built-in functions
1091 more efficiently; for instance, calls to @code{alloca} may become single
1092 instructions that adjust the stack directly, and calls to @code{memcpy}
1093 may become inline copy loops. The resulting code is often both smaller
1094 and faster, but since the function calls no longer appear as such, you
1095 cannot set a breakpoint on those calls, nor can you change the behavior
1096 of the functions by linking with a different library.
1098 In C++, @option{-fno-builtin} is always in effect. The @option{-fbuiltin}
1099 option has no effect. Therefore, in C++, the only way to get the
1100 optimization benefits of built-in functions is to call the function
1101 using the @samp{__builtin_} prefix. The GNU C++ Standard Library uses
1102 built-in functions to implement many functions (like
1103 @code{std::strchr}), so that you automatically get efficient code.
1105 With the @option{-fno-builtin-@var{function}} option, not available
1106 when compiling C++, only the built-in function @var{function} is
1107 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1108 function is named this is not built-in in this version of GCC, this
1109 option is ignored. There is no corresponding
1110 @option{-fbuiltin-@var{function}} option; if you wish to enable
1111 built-in functions selectively when using @option{-fno-builtin} or
1112 @option{-ffreestanding}, you may define macros such as:
1115 #define abs(n) __builtin_abs ((n))
1116 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1121 @cindex hosted environment
1123 Assert that compilation takes place in a hosted environment. This implies
1124 @option{-fbuiltin}. A hosted environment is one in which the
1125 entire standard library is available, and in which @code{main} has a return
1126 type of @code{int}. Examples are nearly everything except a kernel.
1127 This is equivalent to @option{-fno-freestanding}.
1129 @item -ffreestanding
1130 @opindex ffreestanding
1131 @cindex hosted environment
1133 Assert that compilation takes place in a freestanding environment. This
1134 implies @option{-fno-builtin}. A freestanding environment
1135 is one in which the standard library may not exist, and program startup may
1136 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1137 This is equivalent to @option{-fno-hosted}.
1139 @xref{Standards,,Language Standards Supported by GCC}, for details of
1140 freestanding and hosted environments.
1144 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1145 options for strict ISO C conformance) implies @option{-trigraphs}.
1147 @cindex traditional C language
1148 @cindex C language, traditional
1150 @itemx -traditional-cpp
1151 @opindex traditional-cpp
1152 @opindex traditional
1153 Formerly, these options caused GCC to attempt to emulate a pre-standard
1154 C compiler. They are now only supported with the @option{-E} switch.
1155 The preprocessor continues to support a pre-standard mode. See the GNU
1156 CPP manual for details.
1158 @item -fcond-mismatch
1159 @opindex fcond-mismatch
1160 Allow conditional expressions with mismatched types in the second and
1161 third arguments. The value of such an expression is void. This option
1162 is not supported for C++.
1164 @item -funsigned-char
1165 @opindex funsigned-char
1166 Let the type @code{char} be unsigned, like @code{unsigned char}.
1168 Each kind of machine has a default for what @code{char} should
1169 be. It is either like @code{unsigned char} by default or like
1170 @code{signed char} by default.
1172 Ideally, a portable program should always use @code{signed char} or
1173 @code{unsigned char} when it depends on the signedness of an object.
1174 But many programs have been written to use plain @code{char} and
1175 expect it to be signed, or expect it to be unsigned, depending on the
1176 machines they were written for. This option, and its inverse, let you
1177 make such a program work with the opposite default.
1179 The type @code{char} is always a distinct type from each of
1180 @code{signed char} or @code{unsigned char}, even though its behavior
1181 is always just like one of those two.
1184 @opindex fsigned-char
1185 Let the type @code{char} be signed, like @code{signed char}.
1187 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1188 the negative form of @option{-funsigned-char}. Likewise, the option
1189 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1191 @item -fsigned-bitfields
1192 @itemx -funsigned-bitfields
1193 @itemx -fno-signed-bitfields
1194 @itemx -fno-unsigned-bitfields
1195 @opindex fsigned-bitfields
1196 @opindex funsigned-bitfields
1197 @opindex fno-signed-bitfields
1198 @opindex fno-unsigned-bitfields
1199 These options control whether a bit-field is signed or unsigned, when the
1200 declaration does not use either @code{signed} or @code{unsigned}. By
1201 default, such a bit-field is signed, because this is consistent: the
1202 basic integer types such as @code{int} are signed types.
1204 @item -fwritable-strings
1205 @opindex fwritable-strings
1206 Store string constants in the writable data segment and don't uniquize
1207 them. This is for compatibility with old programs which assume they can
1208 write into string constants.
1210 Writing into string constants is a very bad idea; ``constants'' should
1214 @opindex fshort-wchar
1215 Override the underlying type for @samp{wchar_t} to be @samp{short
1216 unsigned int} instead of the default for the target. This option is
1217 useful for building programs to run under WINE@.
1220 @node C++ Dialect Options
1221 @section Options Controlling C++ Dialect
1223 @cindex compiler options, C++
1224 @cindex C++ options, command line
1225 @cindex options, C++
1226 This section describes the command-line options that are only meaningful
1227 for C++ programs; but you can also use most of the GNU compiler options
1228 regardless of what language your program is in. For example, you
1229 might compile a file @code{firstClass.C} like this:
1232 g++ -g -frepo -O -c firstClass.C
1236 In this example, only @option{-frepo} is an option meant
1237 only for C++ programs; you can use the other options with any
1238 language supported by GCC@.
1240 Here is a list of options that are @emph{only} for compiling C++ programs:
1243 @item -fno-access-control
1244 @opindex fno-access-control
1245 Turn off all access checking. This switch is mainly useful for working
1246 around bugs in the access control code.
1250 Check that the pointer returned by @code{operator new} is non-null
1251 before attempting to modify the storage allocated. The current Working
1252 Paper requires that @code{operator new} never return a null pointer, so
1253 this check is normally unnecessary.
1255 An alternative to using this option is to specify that your
1256 @code{operator new} does not throw any exceptions; if you declare it
1257 @samp{throw()}, G++ will check the return value. See also @samp{new
1260 @item -fconserve-space
1261 @opindex fconserve-space
1262 Put uninitialized or runtime-initialized global variables into the
1263 common segment, as C does. This saves space in the executable at the
1264 cost of not diagnosing duplicate definitions. If you compile with this
1265 flag and your program mysteriously crashes after @code{main()} has
1266 completed, you may have an object that is being destroyed twice because
1267 two definitions were merged.
1269 This option is no longer useful on most targets, now that support has
1270 been added for putting variables into BSS without making them common.
1272 @item -fno-const-strings
1273 @opindex fno-const-strings
1274 Give string constants type @code{char *} instead of type @code{const
1275 char *}. By default, G++ uses type @code{const char *} as required by
1276 the standard. Even if you use @option{-fno-const-strings}, you cannot
1277 actually modify the value of a string constant, unless you also use
1278 @option{-fwritable-strings}.
1280 This option might be removed in a future release of G++. For maximum
1281 portability, you should structure your code so that it works with
1282 string constants that have type @code{const char *}.
1284 @item -fdollars-in-identifiers
1285 @opindex fdollars-in-identifiers
1286 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1287 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1288 @samp{$} by default on most target systems, but there are a few exceptions.)
1289 Traditional C allowed the character @samp{$} to form part of
1290 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1292 @item -fno-elide-constructors
1293 @opindex fno-elide-constructors
1294 The C++ standard allows an implementation to omit creating a temporary
1295 which is only used to initialize another object of the same type.
1296 Specifying this option disables that optimization, and forces G++ to
1297 call the copy constructor in all cases.
1299 @item -fno-enforce-eh-specs
1300 @opindex fno-enforce-eh-specs
1301 Don't check for violation of exception specifications at runtime. This
1302 option violates the C++ standard, but may be useful for reducing code
1303 size in production builds, much like defining @samp{NDEBUG}. The compiler
1304 will still optimize based on the exception specifications.
1306 @item -fexternal-templates
1307 @opindex fexternal-templates
1309 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1310 template instantiation; template instances are emitted or not according
1311 to the location of the template definition. @xref{Template
1312 Instantiation}, for more information.
1314 This option is deprecated.
1316 @item -falt-external-templates
1317 @opindex falt-external-templates
1318 Similar to @option{-fexternal-templates}, but template instances are
1319 emitted or not according to the place where they are first instantiated.
1320 @xref{Template Instantiation}, for more information.
1322 This option is deprecated.
1325 @itemx -fno-for-scope
1327 @opindex fno-for-scope
1328 If @option{-ffor-scope} is specified, the scope of variables declared in
1329 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1330 as specified by the C++ standard.
1331 If @option{-fno-for-scope} is specified, the scope of variables declared in
1332 a @i{for-init-statement} extends to the end of the enclosing scope,
1333 as was the case in old versions of G++, and other (traditional)
1334 implementations of C++.
1336 The default if neither flag is given to follow the standard,
1337 but to allow and give a warning for old-style code that would
1338 otherwise be invalid, or have different behavior.
1340 @item -fno-gnu-keywords
1341 @opindex fno-gnu-keywords
1342 Do not recognize @code{typeof} as a keyword, so that code can use this
1343 word as an identifier. You can use the keyword @code{__typeof__} instead.
1344 @option{-ansi} implies @option{-fno-gnu-keywords}.
1346 @item -fno-implicit-templates
1347 @opindex fno-implicit-templates
1348 Never emit code for non-inline templates which are instantiated
1349 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1350 @xref{Template Instantiation}, for more information.
1352 @item -fno-implicit-inline-templates
1353 @opindex fno-implicit-inline-templates
1354 Don't emit code for implicit instantiations of inline templates, either.
1355 The default is to handle inlines differently so that compiles with and
1356 without optimization will need the same set of explicit instantiations.
1358 @item -fno-implement-inlines
1359 @opindex fno-implement-inlines
1360 To save space, do not emit out-of-line copies of inline functions
1361 controlled by @samp{#pragma implementation}. This will cause linker
1362 errors if these functions are not inlined everywhere they are called.
1364 @item -fms-extensions
1365 @opindex fms-extensions
1366 Disable pedantic warnings about constructs used in MFC, such as implicit
1367 int and getting a pointer to member function via non-standard syntax.
1369 @item -fno-nonansi-builtins
1370 @opindex fno-nonansi-builtins
1371 Disable built-in declarations of functions that are not mandated by
1372 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1373 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1375 @item -fno-operator-names
1376 @opindex fno-operator-names
1377 Do not treat the operator name keywords @code{and}, @code{bitand},
1378 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1379 synonyms as keywords.
1381 @item -fno-optional-diags
1382 @opindex fno-optional-diags
1383 Disable diagnostics that the standard says a compiler does not need to
1384 issue. Currently, the only such diagnostic issued by G++ is the one for
1385 a name having multiple meanings within a class.
1388 @opindex fpermissive
1389 Downgrade messages about nonconformant code from errors to warnings. By
1390 default, G++ effectively sets @option{-pedantic-errors} without
1391 @option{-pedantic}; this option reverses that. This behavior and this
1392 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1396 Enable automatic template instantiation at link time. This option also
1397 implies @option{-fno-implicit-templates}. @xref{Template
1398 Instantiation}, for more information.
1402 Disable generation of information about every class with virtual
1403 functions for use by the C++ runtime type identification features
1404 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1405 of the language, you can save some space by using this flag. Note that
1406 exception handling uses the same information, but it will generate it as
1411 Emit statistics about front-end processing at the end of the compilation.
1412 This information is generally only useful to the G++ development team.
1414 @item -ftemplate-depth-@var{n}
1415 @opindex ftemplate-depth
1416 Set the maximum instantiation depth for template classes to @var{n}.
1417 A limit on the template instantiation depth is needed to detect
1418 endless recursions during template class instantiation. ANSI/ISO C++
1419 conforming programs must not rely on a maximum depth greater than 17.
1421 @item -fuse-cxa-atexit
1422 @opindex fuse-cxa-atexit
1423 Register destructors for objects with static storage duration with the
1424 @code{__cxa_atexit} function rather than the @code{atexit} function.
1425 This option is required for fully standards-compliant handling of static
1426 destructors, but will only work if your C library supports
1427 @code{__cxa_atexit}.
1431 Emit special relocations for vtables and virtual function references
1432 so that the linker can identify unused virtual functions and zero out
1433 vtable slots that refer to them. This is most useful with
1434 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1435 also discard the functions themselves.
1437 This optimization requires GNU as and GNU ld. Not all systems support
1438 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1442 Do not use weak symbol support, even if it is provided by the linker.
1443 By default, G++ will use weak symbols if they are available. This
1444 option exists only for testing, and should not be used by end-users;
1445 it will result in inferior code and has no benefits. This option may
1446 be removed in a future release of G++.
1450 Do not search for header files in the standard directories specific to
1451 C++, but do still search the other standard directories. (This option
1452 is used when building the C++ library.)
1455 In addition, these optimization, warning, and code generation options
1456 have meanings only for C++ programs:
1459 @item -fno-default-inline
1460 @opindex fno-default-inline
1461 Do not assume @samp{inline} for functions defined inside a class scope.
1462 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1463 functions will have linkage like inline functions; they just won't be
1466 @item -Wctor-dtor-privacy @r{(C++ only)}
1467 @opindex Wctor-dtor-privacy
1468 Warn when a class seems unusable, because all the constructors or
1469 destructors in a class are private and the class has no friends or
1470 public static member functions.
1472 @item -Wnon-virtual-dtor @r{(C++ only)}
1473 @opindex Wnon-virtual-dtor
1474 Warn when a class declares a non-virtual destructor that should probably
1475 be virtual, because it looks like the class will be used polymorphically.
1477 @item -Wreorder @r{(C++ only)}
1479 @cindex reordering, warning
1480 @cindex warning for reordering of member initializers
1481 Warn when the order of member initializers given in the code does not
1482 match the order in which they must be executed. For instance:
1488 A(): j (0), i (1) @{ @}
1492 Here the compiler will warn that the member initializers for @samp{i}
1493 and @samp{j} will be rearranged to match the declaration order of the
1497 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1500 @item -Weffc++ @r{(C++ only)}
1502 Warn about violations of the following style guidelines from Scott Meyers'
1503 @cite{Effective C++} book:
1507 Item 11: Define a copy constructor and an assignment operator for classes
1508 with dynamically allocated memory.
1511 Item 12: Prefer initialization to assignment in constructors.
1514 Item 14: Make destructors virtual in base classes.
1517 Item 15: Have @code{operator=} return a reference to @code{*this}.
1520 Item 23: Don't try to return a reference when you must return an object.
1524 and about violations of the following style guidelines from Scott Meyers'
1525 @cite{More Effective C++} book:
1529 Item 6: Distinguish between prefix and postfix forms of increment and
1530 decrement operators.
1533 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1537 If you use this option, you should be aware that the standard library
1538 headers do not obey all of these guidelines; you can use @samp{grep -v}
1539 to filter out those warnings.
1541 @item -Wno-deprecated @r{(C++ only)}
1542 @opindex Wno-deprecated
1543 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1545 @item -Wno-non-template-friend @r{(C++ only)}
1546 @opindex Wno-non-template-friend
1547 Disable warnings when non-templatized friend functions are declared
1548 within a template. With the advent of explicit template specification
1549 support in G++, if the name of the friend is an unqualified-id (i.e.,
1550 @samp{friend foo(int)}), the C++ language specification demands that the
1551 friend declare or define an ordinary, nontemplate function. (Section
1552 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1553 could be interpreted as a particular specialization of a templatized
1554 function. Because this non-conforming behavior is no longer the default
1555 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1556 check existing code for potential trouble spots, and is on by default.
1557 This new compiler behavior can be turned off with
1558 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1559 but disables the helpful warning.
1561 @item -Wold-style-cast @r{(C++ only)}
1562 @opindex Wold-style-cast
1563 Warn if an old-style (C-style) cast to a non-void type is used within
1564 a C++ program. The new-style casts (@samp{static_cast},
1565 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1566 unintended effects, and much easier to grep for.
1568 @item -Woverloaded-virtual @r{(C++ only)}
1569 @opindex Woverloaded-virtual
1570 @cindex overloaded virtual fn, warning
1571 @cindex warning for overloaded virtual fn
1572 Warn when a function declaration hides virtual functions from a
1573 base class. For example, in:
1580 struct B: public A @{
1585 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1593 will fail to compile.
1595 @item -Wno-pmf-conversions @r{(C++ only)}
1596 @opindex Wno-pmf-conversions
1597 Disable the diagnostic for converting a bound pointer to member function
1600 @item -Wsign-promo @r{(C++ only)}
1601 @opindex Wsign-promo
1602 Warn when overload resolution chooses a promotion from unsigned or
1603 enumeral type to a signed type over a conversion to an unsigned type of
1604 the same size. Previous versions of G++ would try to preserve
1605 unsignedness, but the standard mandates the current behavior.
1607 @item -Wsynth @r{(C++ only)}
1609 @cindex warning for synthesized methods
1610 @cindex synthesized methods, warning
1611 Warn when G++'s synthesis behavior does not match that of cfront. For
1617 A& operator = (int);
1627 In this example, G++ will synthesize a default @samp{A& operator =
1628 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1631 @node Objective-C Dialect Options
1632 @section Options Controlling Objective-C Dialect
1634 @cindex compiler options, Objective-C
1635 @cindex Objective-C options, command line
1636 @cindex options, Objective-C
1637 This section describes the command-line options that are only meaningful
1638 for Objective-C programs; but you can also use most of the GNU compiler
1639 options regardless of what language your program is in. For example,
1640 you might compile a file @code{some_class.m} like this:
1643 gcc -g -fgnu-runtime -O -c some_class.m
1647 In this example, only @option{-fgnu-runtime} is an option meant only for
1648 Objective-C programs; you can use the other options with any language
1651 Here is a list of options that are @emph{only} for compiling Objective-C
1655 @item -fconstant-string-class=@var{class-name}
1656 @opindex fconstant-string-class
1657 Use @var{class-name} as the name of the class to instantiate for each
1658 literal string specified with the syntax @code{@@"@dots{}"}. The default
1659 class name is @code{NXConstantString}.
1662 @opindex fgnu-runtime
1663 Generate object code compatible with the standard GNU Objective-C
1664 runtime. This is the default for most types of systems.
1666 @item -fnext-runtime
1667 @opindex fnext-runtime
1668 Generate output compatible with the NeXT runtime. This is the default
1669 for NeXT-based systems, including Darwin and Mac OS X@.
1673 Dump interface declarations for all classes seen in the source file to a
1674 file named @file{@var{sourcename}.decl}.
1677 @opindex Wno-protocol
1678 Do not warn if methods required by a protocol are not implemented
1679 in the class adopting it.
1683 Warn if a selector has multiple methods of different types defined.
1685 @c not documented because only avail via -Wp
1686 @c @item -print-objc-runtime-info
1690 @node Language Independent Options
1691 @section Options to Control Diagnostic Messages Formatting
1692 @cindex options to control diagnostics formatting
1693 @cindex diagnostic messages
1694 @cindex message formatting
1696 Traditionally, diagnostic messages have been formatted irrespective of
1697 the output device's aspect (e.g.@: its width, @dots{}). The options described
1698 below can be used to control the diagnostic messages formatting
1699 algorithm, e.g.@: how many characters per line, how often source location
1700 information should be reported. Right now, only the C++ front end can
1701 honor these options. However it is expected, in the near future, that
1702 the remaining front ends would be able to digest them correctly.
1705 @item -fmessage-length=@var{n}
1706 @opindex fmessage-length
1707 Try to format error messages so that they fit on lines of about @var{n}
1708 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1709 the front ends supported by GCC@. If @var{n} is zero, then no
1710 line-wrapping will be done; each error message will appear on a single
1713 @opindex fdiagnostics-show-location
1714 @item -fdiagnostics-show-location=once
1715 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1716 reporter to emit @emph{once} source location information; that is, in
1717 case the message is too long to fit on a single physical line and has to
1718 be wrapped, the source location won't be emitted (as prefix) again,
1719 over and over, in subsequent continuation lines. This is the default
1722 @item -fdiagnostics-show-location=every-line
1723 Only meaningful in line-wrapping mode. Instructs the diagnostic
1724 messages reporter to emit the same source location information (as
1725 prefix) for physical lines that result from the process of breaking
1726 a message which is too long to fit on a single line.
1730 @node Warning Options
1731 @section Options to Request or Suppress Warnings
1732 @cindex options to control warnings
1733 @cindex warning messages
1734 @cindex messages, warning
1735 @cindex suppressing warnings
1737 Warnings are diagnostic messages that report constructions which
1738 are not inherently erroneous but which are risky or suggest there
1739 may have been an error.
1741 You can request many specific warnings with options beginning @samp{-W},
1742 for example @option{-Wimplicit} to request warnings on implicit
1743 declarations. Each of these specific warning options also has a
1744 negative form beginning @samp{-Wno-} to turn off warnings;
1745 for example, @option{-Wno-implicit}. This manual lists only one of the
1746 two forms, whichever is not the default.
1748 The following options control the amount and kinds of warnings produced
1749 by GCC; for further, language-specific options also refer to
1750 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
1753 @cindex syntax checking
1755 @opindex fsyntax-only
1756 Check the code for syntax errors, but don't do anything beyond that.
1760 Issue all the warnings demanded by strict ISO C and ISO C++;
1761 reject all programs that use forbidden extensions, and some other
1762 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1763 version of the ISO C standard specified by any @option{-std} option used.
1765 Valid ISO C and ISO C++ programs should compile properly with or without
1766 this option (though a rare few will require @option{-ansi} or a
1767 @option{-std} option specifying the required version of ISO C)@. However,
1768 without this option, certain GNU extensions and traditional C and C++
1769 features are supported as well. With this option, they are rejected.
1771 @option{-pedantic} does not cause warning messages for use of the
1772 alternate keywords whose names begin and end with @samp{__}. Pedantic
1773 warnings are also disabled in the expression that follows
1774 @code{__extension__}. However, only system header files should use
1775 these escape routes; application programs should avoid them.
1776 @xref{Alternate Keywords}.
1778 Some users try to use @option{-pedantic} to check programs for strict ISO
1779 C conformance. They soon find that it does not do quite what they want:
1780 it finds some non-ISO practices, but not all---only those for which
1781 ISO C @emph{requires} a diagnostic, and some others for which
1782 diagnostics have been added.
1784 A feature to report any failure to conform to ISO C might be useful in
1785 some instances, but would require considerable additional work and would
1786 be quite different from @option{-pedantic}. We don't have plans to
1787 support such a feature in the near future.
1789 Where the standard specified with @option{-std} represents a GNU
1790 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1791 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1792 extended dialect is based. Warnings from @option{-pedantic} are given
1793 where they are required by the base standard. (It would not make sense
1794 for such warnings to be given only for features not in the specified GNU
1795 C dialect, since by definition the GNU dialects of C include all
1796 features the compiler supports with the given option, and there would be
1797 nothing to warn about.)
1799 @item -pedantic-errors
1800 @opindex pedantic-errors
1801 Like @option{-pedantic}, except that errors are produced rather than
1806 Inhibit all warning messages.
1810 Inhibit warning messages about the use of @samp{#import}.
1812 @item -Wchar-subscripts
1813 @opindex Wchar-subscripts
1814 Warn if an array subscript has type @code{char}. This is a common cause
1815 of error, as programmers often forget that this type is signed on some
1820 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1821 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1825 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1826 the arguments supplied have types appropriate to the format string
1827 specified, and that the conversions specified in the format string make
1828 sense. This includes standard functions, and others specified by format
1829 attributes (@pxref{Function Attributes}), in the @code{printf},
1830 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1831 not in the C standard) families.
1833 The formats are checked against the format features supported by GNU
1834 libc version 2.2. These include all ISO C89 and C99 features, as well
1835 as features from the Single Unix Specification and some BSD and GNU
1836 extensions. Other library implementations may not support all these
1837 features; GCC does not support warning about features that go beyond a
1838 particular library's limitations. However, if @option{-pedantic} is used
1839 with @option{-Wformat}, warnings will be given about format features not
1840 in the selected standard version (but not for @code{strfmon} formats,
1841 since those are not in any version of the C standard). @xref{C Dialect
1842 Options,,Options Controlling C Dialect}.
1844 @option{-Wformat} is included in @option{-Wall}. For more control over some
1845 aspects of format checking, the options @option{-Wno-format-y2k},
1846 @option{-Wno-format-extra-args}, @option{-Wformat-nonliteral},
1847 @option{-Wformat-security} and @option{-Wformat=2} are available, but are
1848 not included in @option{-Wall}.
1850 @item -Wno-format-y2k
1851 @opindex Wno-format-y2k
1852 If @option{-Wformat} is specified, do not warn about @code{strftime}
1853 formats which may yield only a two-digit year.
1855 @item -Wno-format-extra-args
1856 @opindex Wno-format-extra-args
1857 If @option{-Wformat} is specified, do not warn about excess arguments to a
1858 @code{printf} or @code{scanf} format function. The C standard specifies
1859 that such arguments are ignored.
1861 Where the unused arguments lie between used arguments that are
1862 specified with @samp{$} operand number specifications, normally
1863 warnings are still given, since the implementation could not know what
1864 type to pass to @code{va_arg} to skip the unused arguments. However,
1865 in the case of @code{scanf} formats, this option will suppress the
1866 warning if the unused arguments are all pointers, since the Single
1867 Unix Specification says that such unused arguments are allowed.
1869 @item -Wformat-nonliteral
1870 @opindex Wformat-nonliteral
1871 If @option{-Wformat} is specified, also warn if the format string is not a
1872 string literal and so cannot be checked, unless the format function
1873 takes its format arguments as a @code{va_list}.
1875 @item -Wformat-security
1876 @opindex Wformat-security
1877 If @option{-Wformat} is specified, also warn about uses of format
1878 functions that represent possible security problems. At present, this
1879 warns about calls to @code{printf} and @code{scanf} functions where the
1880 format string is not a string literal and there are no format arguments,
1881 as in @code{printf (foo);}. This may be a security hole if the format
1882 string came from untrusted input and contains @samp{%n}. (This is
1883 currently a subset of what @option{-Wformat-nonliteral} warns about, but
1884 in future warnings may be added to @option{-Wformat-security} that are not
1885 included in @option{-Wformat-nonliteral}.)
1889 Enable @option{-Wformat} plus format checks not included in
1890 @option{-Wformat}. Currently equivalent to @samp{-Wformat
1891 -Wformat-nonliteral -Wformat-security}.
1893 @item -Wimplicit-int
1894 @opindex Wimplicit-int
1895 Warn when a declaration does not specify a type.
1897 @item -Wimplicit-function-declaration
1898 @itemx -Werror-implicit-function-declaration
1899 @opindex Wimplicit-function-declaration
1900 @opindex Werror-implicit-function-declaration
1901 Give a warning (or error) whenever a function is used before being
1906 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
1910 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1911 function with external linkage, returning int, taking either zero
1912 arguments, two, or three arguments of appropriate types.
1914 @item -Wmissing-braces
1915 @opindex Wmissing-braces
1916 Warn if an aggregate or union initializer is not fully bracketed. In
1917 the following example, the initializer for @samp{a} is not fully
1918 bracketed, but that for @samp{b} is fully bracketed.
1921 int a[2][2] = @{ 0, 1, 2, 3 @};
1922 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
1926 @opindex Wparentheses
1927 Warn if parentheses are omitted in certain contexts, such
1928 as when there is an assignment in a context where a truth value
1929 is expected, or when operators are nested whose precedence people
1930 often get confused about.
1932 Also warn about constructions where there may be confusion to which
1933 @code{if} statement an @code{else} branch belongs. Here is an example of
1948 In C, every @code{else} branch belongs to the innermost possible @code{if}
1949 statement, which in this example is @code{if (b)}. This is often not
1950 what the programmer expected, as illustrated in the above example by
1951 indentation the programmer chose. When there is the potential for this
1952 confusion, GCC will issue a warning when this flag is specified.
1953 To eliminate the warning, add explicit braces around the innermost
1954 @code{if} statement so there is no way the @code{else} could belong to
1955 the enclosing @code{if}. The resulting code would look like this:
1971 @item -Wsequence-point
1972 @opindex Wsequence-point
1973 Warn about code that may have undefined semantics because of violations
1974 of sequence point rules in the C standard.
1976 The C standard defines the order in which expressions in a C program are
1977 evaluated in terms of @dfn{sequence points}, which represent a partial
1978 ordering between the execution of parts of the program: those executed
1979 before the sequence point, and those executed after it. These occur
1980 after the evaluation of a full expression (one which is not part of a
1981 larger expression), after the evaluation of the first operand of a
1982 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
1983 function is called (but after the evaluation of its arguments and the
1984 expression denoting the called function), and in certain other places.
1985 Other than as expressed by the sequence point rules, the order of
1986 evaluation of subexpressions of an expression is not specified. All
1987 these rules describe only a partial order rather than a total order,
1988 since, for example, if two functions are called within one expression
1989 with no sequence point between them, the order in which the functions
1990 are called is not specified. However, the standards committee have
1991 ruled that function calls do not overlap.
1993 It is not specified when between sequence points modifications to the
1994 values of objects take effect. Programs whose behavior depends on this
1995 have undefined behavior; the C standard specifies that ``Between the
1996 previous and next sequence point an object shall have its stored value
1997 modified at most once by the evaluation of an expression. Furthermore,
1998 the prior value shall be read only to determine the value to be
1999 stored.''. If a program breaks these rules, the results on any
2000 particular implementation are entirely unpredictable.
2002 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2003 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2004 diagnosed by this option, and it may give an occasional false positive
2005 result, but in general it has been found fairly effective at detecting
2006 this sort of problem in programs.
2008 The present implementation of this option only works for C programs. A
2009 future implementation may also work for C++ programs.
2011 The C standard is worded confusingly, therefore there is some debate
2012 over the precise meaning of the sequence point rules in subtle cases.
2013 Links to discussions of the problem, including proposed formal
2014 definitions, may be found on our readings page, at
2015 @w{@uref{http://gcc.gnu.org/readings.html}}.
2018 @opindex Wreturn-type
2019 Warn whenever a function is defined with a return-type that defaults to
2020 @code{int}. Also warn about any @code{return} statement with no
2021 return-value in a function whose return-type is not @code{void}.
2023 For C++, a function without return type always produces a diagnostic
2024 message, even when @option{-Wno-return-type} is specified. The only
2025 exceptions are @samp{main} and functions defined in system headers.
2029 Warn whenever a @code{switch} statement has an index of enumeral type
2030 and lacks a @code{case} for one or more of the named codes of that
2031 enumeration. (The presence of a @code{default} label prevents this
2032 warning.) @code{case} labels outside the enumeration range also
2033 provoke warnings when this option is used.
2035 @item -Wswitch-default
2036 @opindex Wswitch-switch
2037 Warn whenever a @code{switch} statement does not have a @code{default}
2041 @opindex Wswitch-enum
2042 Warn whenever a @code{switch} statement has an index of enumeral type
2043 and lacks a @code{case} for one or more of the named codes of that
2044 enumeration. @code{case} labels outside the enumeration range also
2045 provoke warnings when this option is used.
2049 Warn if any trigraphs are encountered that might change the meaning of
2050 the program (trigraphs within comments are not warned about).
2052 @item -Wunused-function
2053 @opindex Wunused-function
2054 Warn whenever a static function is declared but not defined or a
2055 non\-inline static function is unused.
2057 @item -Wunused-label
2058 @opindex Wunused-label
2059 Warn whenever a label is declared but not used.
2061 To suppress this warning use the @samp{unused} attribute
2062 (@pxref{Variable Attributes}).
2064 @item -Wunused-parameter
2065 @opindex Wunused-parameter
2066 Warn whenever a function parameter is unused aside from its declaration.
2068 To suppress this warning use the @samp{unused} attribute
2069 (@pxref{Variable Attributes}).
2071 @item -Wunused-variable
2072 @opindex Wunused-variable
2073 Warn whenever a local variable or non-constant static variable is unused
2074 aside from its declaration
2076 To suppress this warning use the @samp{unused} attribute
2077 (@pxref{Variable Attributes}).
2079 @item -Wunused-value
2080 @opindex Wunused-value
2081 Warn whenever a statement computes a result that is explicitly not used.
2083 To suppress this warning cast the expression to @samp{void}.
2087 All all the above @option{-Wunused} options combined.
2089 In order to get a warning about an unused function parameter, you must
2090 either specify @samp{-W -Wunused} or separately specify
2091 @option{-Wunused-parameter}.
2093 @item -Wuninitialized
2094 @opindex Wuninitialized
2095 Warn if an automatic variable is used without first being initialized or
2096 if a variable may be clobbered by a @code{setjmp} call.
2098 These warnings are possible only in optimizing compilation,
2099 because they require data flow information that is computed only
2100 when optimizing. If you don't specify @option{-O}, you simply won't
2103 These warnings occur only for variables that are candidates for
2104 register allocation. Therefore, they do not occur for a variable that
2105 is declared @code{volatile}, or whose address is taken, or whose size
2106 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2107 structures, unions or arrays, even when they are in registers.
2109 Note that there may be no warning about a variable that is used only
2110 to compute a value that itself is never used, because such
2111 computations may be deleted by data flow analysis before the warnings
2114 These warnings are made optional because GCC is not smart
2115 enough to see all the reasons why the code might be correct
2116 despite appearing to have an error. Here is one example of how
2137 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2138 always initialized, but GCC doesn't know this. Here is
2139 another common case:
2144 if (change_y) save_y = y, y = new_y;
2146 if (change_y) y = save_y;
2151 This has no bug because @code{save_y} is used only if it is set.
2153 @cindex @code{longjmp} warnings
2154 This option also warns when a non-volatile automatic variable might be
2155 changed by a call to @code{longjmp}. These warnings as well are possible
2156 only in optimizing compilation.
2158 The compiler sees only the calls to @code{setjmp}. It cannot know
2159 where @code{longjmp} will be called; in fact, a signal handler could
2160 call it at any point in the code. As a result, you may get a warning
2161 even when there is in fact no problem because @code{longjmp} cannot
2162 in fact be called at the place which would cause a problem.
2164 Some spurious warnings can be avoided if you declare all the functions
2165 you use that never return as @code{noreturn}. @xref{Function
2168 @item -Wreorder @r{(C++ only)}
2170 @cindex reordering, warning
2171 @cindex warning for reordering of member initializers
2172 Warn when the order of member initializers given in the code does not
2173 match the order in which they must be executed. For instance:
2175 @item -Wunknown-pragmas
2176 @opindex Wunknown-pragmas
2177 @cindex warning for unknown pragmas
2178 @cindex unknown pragmas, warning
2179 @cindex pragmas, warning of unknown
2180 Warn when a #pragma directive is encountered which is not understood by
2181 GCC@. If this command line option is used, warnings will even be issued
2182 for unknown pragmas in system header files. This is not the case if
2183 the warnings were only enabled by the @option{-Wall} command line option.
2187 All of the above @samp{-W} options combined. This enables all the
2188 warnings about constructions that some users consider questionable, and
2189 that are easy to avoid (or modify to prevent the warning), even in
2190 conjunction with macros.
2193 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2194 Some of them warn about constructions that users generally do not
2195 consider questionable, but which occasionally you might wish to check
2196 for; others warn about constructions that are necessary or hard to avoid
2197 in some cases, and there is no simple way to modify the code to suppress
2203 Print extra warning messages for these events:
2207 A function can return either with or without a value. (Falling
2208 off the end of the function body is considered returning without
2209 a value.) For example, this function would evoke such a
2223 An expression-statement or the left-hand side of a comma expression
2224 contains no side effects.
2225 To suppress the warning, cast the unused expression to void.
2226 For example, an expression such as @samp{x[i,j]} will cause a warning,
2227 but @samp{x[(void)i,j]} will not.
2230 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2233 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2234 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2235 that of ordinary mathematical notation.
2238 Storage-class specifiers like @code{static} are not the first things in
2239 a declaration. According to the C Standard, this usage is obsolescent.
2242 The return type of a function has a type qualifier such as @code{const}.
2243 Such a type qualifier has no effect, since the value returned by a
2244 function is not an lvalue. (But don't warn about the GNU extension of
2245 @code{volatile void} return types. That extension will be warned about
2246 if @option{-pedantic} is specified.)
2249 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2253 A comparison between signed and unsigned values could produce an
2254 incorrect result when the signed value is converted to unsigned.
2255 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2258 An aggregate has a partly bracketed initializer.
2259 For example, the following code would evoke such a warning,
2260 because braces are missing around the initializer for @code{x.h}:
2263 struct s @{ int f, g; @};
2264 struct t @{ struct s h; int i; @};
2265 struct t x = @{ 1, 2, 3 @};
2269 An aggregate has an initializer which does not initialize all members.
2270 For example, the following code would cause such a warning, because
2271 @code{x.h} would be implicitly initialized to zero:
2274 struct s @{ int f, g, h; @};
2275 struct s x = @{ 3, 4 @};
2279 @item -Wno-div-by-zero
2280 @opindex Wno-div-by-zero
2281 @opindex Wdiv-by-zero
2282 Do not warn about compile-time integer division by zero. Floating point
2283 division by zero is not warned about, as it can be a legitimate way of
2284 obtaining infinities and NaNs.
2286 @item -Wsystem-headers
2287 @opindex Wsystem-headers
2288 @cindex warnings from system headers
2289 @cindex system headers, warnings from
2290 Print warning messages for constructs found in system header files.
2291 Warnings from system headers are normally suppressed, on the assumption
2292 that they usually do not indicate real problems and would only make the
2293 compiler output harder to read. Using this command line option tells
2294 GCC to emit warnings from system headers as if they occurred in user
2295 code. However, note that using @option{-Wall} in conjunction with this
2296 option will @emph{not} warn about unknown pragmas in system
2297 headers---for that, @option{-Wunknown-pragmas} must also be used.
2300 @opindex Wfloat-equal
2301 Warn if floating point values are used in equality comparisons.
2303 The idea behind this is that sometimes it is convenient (for the
2304 programmer) to consider floating-point values as approximations to
2305 infinitely precise real numbers. If you are doing this, then you need
2306 to compute (by analysing the code, or in some other way) the maximum or
2307 likely maximum error that the computation introduces, and allow for it
2308 when performing comparisons (and when producing output, but that's a
2309 different problem). In particular, instead of testing for equality, you
2310 would check to see whether the two values have ranges that overlap; and
2311 this is done with the relational operators, so equality comparisons are
2314 @item -Wtraditional @r{(C only)}
2315 @opindex Wtraditional
2316 Warn about certain constructs that behave differently in traditional and
2317 ISO C@. Also warn about ISO C constructs that have no traditional C
2318 equivalent, and/or problematic constructs which should be avoided.
2322 Macro parameters that appear within string literals in the macro body.
2323 In traditional C macro replacement takes place within string literals,
2324 but does not in ISO C@.
2327 In traditional C, some preprocessor directives did not exist.
2328 Traditional preprocessors would only consider a line to be a directive
2329 if the @samp{#} appeared in column 1 on the line. Therefore
2330 @option{-Wtraditional} warns about directives that traditional C
2331 understands but would ignore because the @samp{#} does not appear as the
2332 first character on the line. It also suggests you hide directives like
2333 @samp{#pragma} not understood by traditional C by indenting them. Some
2334 traditional implementations would not recognize @samp{#elif}, so it
2335 suggests avoiding it altogether.
2338 A function-like macro that appears without arguments.
2341 The unary plus operator.
2344 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2345 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2346 constants.) Note, these suffixes appear in macros defined in the system
2347 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2348 Use of these macros in user code might normally lead to spurious
2349 warnings, however gcc's integrated preprocessor has enough context to
2350 avoid warning in these cases.
2353 A function declared external in one block and then used after the end of
2357 A @code{switch} statement has an operand of type @code{long}.
2360 A non-@code{static} function declaration follows a @code{static} one.
2361 This construct is not accepted by some traditional C compilers.
2364 The ISO type of an integer constant has a different width or
2365 signedness from its traditional type. This warning is only issued if
2366 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2367 typically represent bit patterns, are not warned about.
2370 Usage of ISO string concatenation is detected.
2373 Initialization of automatic aggregates.
2376 Identifier conflicts with labels. Traditional C lacks a separate
2377 namespace for labels.
2380 Initialization of unions. If the initializer is zero, the warning is
2381 omitted. This is done under the assumption that the zero initializer in
2382 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2383 initializer warnings and relies on default initialization to zero in the
2387 Conversions by prototypes between fixed/floating point values and vice
2388 versa. The absence of these prototypes when compiling with traditional
2389 C would cause serious problems. This is a subset of the possible
2390 conversion warnings, for the full set use @option{-Wconversion}.
2395 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2397 @item -Wendif-labels
2398 @opindex Wendif-labels
2399 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2403 Warn whenever a local variable shadows another local variable, parameter or
2404 global variable or whenever a built-in function is shadowed.
2406 @item -Wlarger-than-@var{len}
2407 @opindex Wlarger-than
2408 Warn whenever an object of larger than @var{len} bytes is defined.
2410 @item -Wpointer-arith
2411 @opindex Wpointer-arith
2412 Warn about anything that depends on the ``size of'' a function type or
2413 of @code{void}. GNU C assigns these types a size of 1, for
2414 convenience in calculations with @code{void *} pointers and pointers
2417 @item -Wbad-function-cast @r{(C only)}
2418 @opindex Wbad-function-cast
2419 Warn whenever a function call is cast to a non-matching type.
2420 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2424 Warn whenever a pointer is cast so as to remove a type qualifier from
2425 the target type. For example, warn if a @code{const char *} is cast
2426 to an ordinary @code{char *}.
2429 @opindex Wcast-align
2430 Warn whenever a pointer is cast such that the required alignment of the
2431 target is increased. For example, warn if a @code{char *} is cast to
2432 an @code{int *} on machines where integers can only be accessed at
2433 two- or four-byte boundaries.
2435 @item -Wwrite-strings
2436 @opindex Wwrite-strings
2437 When compiling C, give string constants the type @code{const
2438 char[@var{length}]} so that
2439 copying the address of one into a non-@code{const} @code{char *}
2440 pointer will get a warning; when compiling C++, warn about the
2441 deprecated conversion from string constants to @code{char *}.
2442 These warnings will help you find at
2443 compile time code that can try to write into a string constant, but
2444 only if you have been very careful about using @code{const} in
2445 declarations and prototypes. Otherwise, it will just be a nuisance;
2446 this is why we did not make @option{-Wall} request these warnings.
2449 @opindex Wconversion
2450 Warn if a prototype causes a type conversion that is different from what
2451 would happen to the same argument in the absence of a prototype. This
2452 includes conversions of fixed point to floating and vice versa, and
2453 conversions changing the width or signedness of a fixed point argument
2454 except when the same as the default promotion.
2456 Also, warn if a negative integer constant expression is implicitly
2457 converted to an unsigned type. For example, warn about the assignment
2458 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2459 casts like @code{(unsigned) -1}.
2461 @item -Wsign-compare
2462 @opindex Wsign-compare
2463 @cindex warning for comparison of signed and unsigned values
2464 @cindex comparison of signed and unsigned values, warning
2465 @cindex signed and unsigned values, comparison warning
2466 Warn when a comparison between signed and unsigned values could produce
2467 an incorrect result when the signed value is converted to unsigned.
2468 This warning is also enabled by @option{-W}; to get the other warnings
2469 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2471 @item -Waggregate-return
2472 @opindex Waggregate-return
2473 Warn if any functions that return structures or unions are defined or
2474 called. (In languages where you can return an array, this also elicits
2477 @item -Wstrict-prototypes @r{(C only)}
2478 @opindex Wstrict-prototypes
2479 Warn if a function is declared or defined without specifying the
2480 argument types. (An old-style function definition is permitted without
2481 a warning if preceded by a declaration which specifies the argument
2484 @item -Wmissing-prototypes @r{(C only)}
2485 @opindex Wmissing-prototypes
2486 Warn if a global function is defined without a previous prototype
2487 declaration. This warning is issued even if the definition itself
2488 provides a prototype. The aim is to detect global functions that fail
2489 to be declared in header files.
2491 @item -Wmissing-declarations
2492 @opindex Wmissing-declarations
2493 Warn if a global function is defined without a previous declaration.
2494 Do so even if the definition itself provides a prototype.
2495 Use this option to detect global functions that are not declared in
2498 @item -Wmissing-noreturn
2499 @opindex Wmissing-noreturn
2500 Warn about functions which might be candidates for attribute @code{noreturn}.
2501 Note these are only possible candidates, not absolute ones. Care should
2502 be taken to manually verify functions actually do not ever return before
2503 adding the @code{noreturn} attribute, otherwise subtle code generation
2504 bugs could be introduced. You will not get a warning for @code{main} in
2505 hosted C environments.
2507 @item -Wmissing-format-attribute
2508 @opindex Wmissing-format-attribute
2510 If @option{-Wformat} is enabled, also warn about functions which might be
2511 candidates for @code{format} attributes. Note these are only possible
2512 candidates, not absolute ones. GCC will guess that @code{format}
2513 attributes might be appropriate for any function that calls a function
2514 like @code{vprintf} or @code{vscanf}, but this might not always be the
2515 case, and some functions for which @code{format} attributes are
2516 appropriate may not be detected. This option has no effect unless
2517 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2519 @item -Wno-multichar
2520 @opindex Wno-multichar
2522 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2523 Usually they indicate a typo in the user's code, as they have
2524 implementation-defined values, and should not be used in portable code.
2526 @item -Wno-deprecated-declarations
2527 @opindex Wno-deprecated-declarations
2528 Do not warn about uses of functions, variables, and types marked as
2529 deprecated by using the @code{deprecated} attribute.
2530 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2531 @pxref{Type Attributes}.)
2535 Warn if a structure is given the packed attribute, but the packed
2536 attribute has no effect on the layout or size of the structure.
2537 Such structures may be mis-aligned for little benefit. For
2538 instance, in this code, the variable @code{f.x} in @code{struct bar}
2539 will be misaligned even though @code{struct bar} does not itself
2540 have the packed attribute:
2547 @} __attribute__((packed));
2557 Warn if padding is included in a structure, either to align an element
2558 of the structure or to align the whole structure. Sometimes when this
2559 happens it is possible to rearrange the fields of the structure to
2560 reduce the padding and so make the structure smaller.
2562 @item -Wredundant-decls
2563 @opindex Wredundant-decls
2564 Warn if anything is declared more than once in the same scope, even in
2565 cases where multiple declaration is valid and changes nothing.
2567 @item -Wnested-externs @r{(C only)}
2568 @opindex Wnested-externs
2569 Warn if an @code{extern} declaration is encountered within a function.
2571 @item -Wunreachable-code
2572 @opindex Wunreachable-code
2573 Warn if the compiler detects that code will never be executed.
2575 This option is intended to warn when the compiler detects that at
2576 least a whole line of source code will never be executed, because
2577 some condition is never satisfied or because it is after a
2578 procedure that never returns.
2580 It is possible for this option to produce a warning even though there
2581 are circumstances under which part of the affected line can be executed,
2582 so care should be taken when removing apparently-unreachable code.
2584 For instance, when a function is inlined, a warning may mean that the
2585 line is unreachable in only one inlined copy of the function.
2587 This option is not made part of @option{-Wall} because in a debugging
2588 version of a program there is often substantial code which checks
2589 correct functioning of the program and is, hopefully, unreachable
2590 because the program does work. Another common use of unreachable
2591 code is to provide behavior which is selectable at compile-time.
2595 Warn if a function can not be inlined and it was declared as inline.
2599 @opindex Wno-long-long
2600 Warn if @samp{long long} type is used. This is default. To inhibit
2601 the warning messages, use @option{-Wno-long-long}. Flags
2602 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2603 only when @option{-pedantic} flag is used.
2605 @item -Wdisabled-optimization
2606 @opindex Wdisabled-optimization
2607 Warn if a requested optimization pass is disabled. This warning does
2608 not generally indicate that there is anything wrong with your code; it
2609 merely indicates that GCC's optimizers were unable to handle the code
2610 effectively. Often, the problem is that your code is too big or too
2611 complex; GCC will refuse to optimize programs when the optimization
2612 itself is likely to take inordinate amounts of time.
2616 Make all warnings into errors.
2619 @node Debugging Options
2620 @section Options for Debugging Your Program or GCC
2621 @cindex options, debugging
2622 @cindex debugging information options
2624 GCC has various special options that are used for debugging
2625 either your program or GCC:
2630 Produce debugging information in the operating system's native format
2631 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2634 On most systems that use stabs format, @option{-g} enables use of extra
2635 debugging information that only GDB can use; this extra information
2636 makes debugging work better in GDB but will probably make other debuggers
2638 refuse to read the program. If you want to control for certain whether
2639 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2640 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2641 or @option{-gvms} (see below).
2643 Unlike most other C compilers, GCC allows you to use @option{-g} with
2644 @option{-O}. The shortcuts taken by optimized code may occasionally
2645 produce surprising results: some variables you declared may not exist
2646 at all; flow of control may briefly move where you did not expect it;
2647 some statements may not be executed because they compute constant
2648 results or their values were already at hand; some statements may
2649 execute in different places because they were moved out of loops.
2651 Nevertheless it proves possible to debug optimized output. This makes
2652 it reasonable to use the optimizer for programs that might have bugs.
2654 The following options are useful when GCC is generated with the
2655 capability for more than one debugging format.
2659 Produce debugging information for use by GDB@. This means to use the
2660 most expressive format available (DWARF 2, stabs, or the native format
2661 if neither of those are supported), including GDB extensions if at all
2666 Produce debugging information in stabs format (if that is supported),
2667 without GDB extensions. This is the format used by DBX on most BSD
2668 systems. On MIPS, Alpha and System V Release 4 systems this option
2669 produces stabs debugging output which is not understood by DBX or SDB@.
2670 On System V Release 4 systems this option requires the GNU assembler.
2674 Produce debugging information in stabs format (if that is supported),
2675 using GNU extensions understood only by the GNU debugger (GDB)@. The
2676 use of these extensions is likely to make other debuggers crash or
2677 refuse to read the program.
2681 Produce debugging information in COFF format (if that is supported).
2682 This is the format used by SDB on most System V systems prior to
2687 Produce debugging information in XCOFF format (if that is supported).
2688 This is the format used by the DBX debugger on IBM RS/6000 systems.
2692 Produce debugging information in XCOFF format (if that is supported),
2693 using GNU extensions understood only by the GNU debugger (GDB)@. The
2694 use of these extensions is likely to make other debuggers crash or
2695 refuse to read the program, and may cause assemblers other than the GNU
2696 assembler (GAS) to fail with an error.
2700 Produce debugging information in DWARF version 1 format (if that is
2701 supported). This is the format used by SDB on most System V Release 4
2706 Produce debugging information in DWARF version 1 format (if that is
2707 supported), using GNU extensions understood only by the GNU debugger
2708 (GDB)@. The use of these extensions is likely to make other debuggers
2709 crash or refuse to read the program.
2713 Produce debugging information in DWARF version 2 format (if that is
2714 supported). This is the format used by DBX on IRIX 6.
2718 Produce debugging information in VMS debug format (if that is
2719 supported). This is the format used by DEBUG on VMS systems.
2722 @itemx -ggdb@var{level}
2723 @itemx -gstabs@var{level}
2724 @itemx -gcoff@var{level}
2725 @itemx -gxcoff@var{level}
2726 @itemx -gvms@var{level}
2727 Request debugging information and also use @var{level} to specify how
2728 much information. The default level is 2.
2730 Level 1 produces minimal information, enough for making backtraces in
2731 parts of the program that you don't plan to debug. This includes
2732 descriptions of functions and external variables, but no information
2733 about local variables and no line numbers.
2735 Level 3 includes extra information, such as all the macro definitions
2736 present in the program. Some debuggers support macro expansion when
2737 you use @option{-g3}.
2739 Note that in order to avoid confusion between DWARF1 debug level 2,
2740 and DWARF2, neither @option{-gdwarf} nor @option{-gdwarf-2} accept
2741 a concatenated debug level. Instead use an additional @option{-g@var{level}}
2742 option to change the debug level for DWARF1 or DWARF2.
2747 Generate extra code to write profile information suitable for the
2748 analysis program @code{prof}. You must use this option when compiling
2749 the source files you want data about, and you must also use it when
2752 @cindex @code{gprof}
2755 Generate extra code to write profile information suitable for the
2756 analysis program @code{gprof}. You must use this option when compiling
2757 the source files you want data about, and you must also use it when
2763 Generate extra code to write profile information for basic blocks, which will
2764 record the number of times each basic block is executed, the basic block start
2765 address, and the function name containing the basic block. If @option{-g} is
2766 used, the line number and filename of the start of the basic block will also be
2767 recorded. If not overridden by the machine description, the default action is
2768 to append to the text file @file{bb.out}.
2770 This data could be analyzed by a program like @code{tcov}. Note,
2771 however, that the format of the data is not what @code{tcov} expects.
2772 Eventually GNU @code{gprof} should be extended to process this data.
2776 Makes the compiler print out each function name as it is compiled, and
2777 print some statistics about each pass when it finishes.
2780 @opindex ftime-report
2781 Makes the compiler print some statistics about the time consumed by each
2782 pass when it finishes.
2785 @opindex fmem-report
2786 Makes the compiler print some statistics about permanent memory
2787 allocation when it finishes.
2789 @item -fprofile-arcs
2790 @opindex fprofile-arcs
2791 Instrument @dfn{arcs} during compilation to generate coverage data
2792 or for profile-directed block ordering. During execution the program
2793 records how many times each branch is executed and how many times it is
2794 taken. When the compiled program exits it saves this data to a file
2795 called @file{@var{sourcename}.da} for each source file.
2797 For profile-directed block ordering, compile the program with
2798 @option{-fprofile-arcs} plus optimization and code generation options,
2799 generate the arc profile information by running the program on a
2800 selected workload, and then compile the program again with the same
2801 optimization and code generation options plus
2802 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2803 Control Optimization}).
2805 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
2806 when it is used with the @option{-ftest-coverage} option. GCC
2807 supports two methods of determining code coverage: the options that
2808 support @code{gcov}, and options @option{-a} and @option{-ax}, which
2809 write information to text files. The options that support @code{gcov}
2810 do not need to instrument every arc in the program, so a program compiled
2811 with them runs faster than a program compiled with @option{-a}, which
2812 adds instrumentation code to every basic block in the program. The
2813 tradeoff: since @code{gcov} does not have execution counts for all
2814 branches, it must start with the execution counts for the instrumented
2815 branches, and then iterate over the program flow graph until the entire
2816 graph has been solved. Hence, @code{gcov} runs a little more slowly than
2817 a program which uses information from @option{-a} and @option{-ax}.
2819 With @option{-fprofile-arcs}, for each function of your program GCC
2820 creates a program flow graph, then finds a spanning tree for the graph.
2821 Only arcs that are not on the spanning tree have to be instrumented: the
2822 compiler adds code to count the number of times that these arcs are
2823 executed. When an arc is the only exit or only entrance to a block, the
2824 instrumentation code can be added to the block; otherwise, a new basic
2825 block must be created to hold the instrumentation code.
2827 This option makes it possible to estimate branch probabilities and to
2828 calculate basic block execution counts. In general, basic block
2829 execution counts as provided by @option{-a} do not give enough
2830 information to estimate all branch probabilities.
2833 @item -ftest-coverage
2834 @opindex ftest-coverage
2835 Create data files for the @code{gcov} code-coverage utility
2836 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2837 The data file names begin with the name of your source file:
2840 @item @var{sourcename}.bb
2841 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2842 associate basic block execution counts with line numbers.
2844 @item @var{sourcename}.bbg
2845 A list of all arcs in the program flow graph. This allows @code{gcov}
2846 to reconstruct the program flow graph, so that it can compute all basic
2847 block and arc execution counts from the information in the
2848 @code{@var{sourcename}.da} file.
2851 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
2852 option adds instrumentation to the program, which then writes
2853 execution counts to another data file:
2856 @item @var{sourcename}.da
2857 Runtime arc execution counts, used in conjunction with the arc
2858 information in the file @code{@var{sourcename}.bbg}.
2861 Coverage data will map better to the source files if
2862 @option{-ftest-coverage} is used without optimization.
2864 @item -d@var{letters}
2866 Says to make debugging dumps during compilation at times specified by
2867 @var{letters}. This is used for debugging the compiler. The file names
2868 for most of the dumps are made by appending a pass number and a word to
2869 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2870 Here are the possible letters for use in @var{letters}, and their meanings:
2875 Annotate the assembler output with miscellaneous debugging information.
2878 Dump after computing branch probabilities, to @file{@var{file}.15.bp}.
2881 Dump after block reordering, to @file{@var{file}.30.bbro}.
2884 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
2887 Dump after the first if conversion, to the file @file{@var{file}.18.ce}.
2890 Dump after delayed branch scheduling, to @file{@var{file}.32.dbr}.
2893 Dump all macro definitions, at the end of preprocessing, in addition to
2897 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
2898 @file{@var{file}.07.ussa}.
2901 Dump after the second if conversion, to @file{@var{file}.27.ce2}.
2904 Dump after life analysis, to @file{@var{file}.16.life}.
2907 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.10.addressof}.
2910 Dump after global register allocation, to @file{@var{file}.22.greg}.
2913 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
2916 Dump after reg-to-stack conversion, to @file{@var{file}.29.stack}.
2919 Dump after post-reload optimizations, to @file{@var{file}.23.postreload}.
2922 Dump after GCSE, to @file{@var{file}.11.gcse}.
2925 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2928 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
2931 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
2934 Dump after local register allocation, to @file{@var{file}.21.lreg}.
2937 Dump after loop optimization, to @file{@var{file}.12.loop}.
2940 Dump after performing the machine dependent reorganisation pass, to
2941 @file{@var{file}.31.mach}.
2944 Dump after register renumbering, to @file{@var{file}.26.rnreg}.
2947 Dump after the register move pass, to @file{@var{file}.19.regmove}.
2950 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2953 Dump after the second scheduling pass, to @file{@var{file}.28.sched2}.
2956 Dump after CSE (including the jump optimization that sometimes follows
2957 CSE), to @file{@var{file}.09.cse}.
2960 Dump after the first scheduling pass, to @file{@var{file}.20.sched}.
2963 Dump after the second CSE pass (including the jump optimization that
2964 sometimes follows CSE), to @file{@var{file}.13.cse2}.
2967 Dump after null pointer ellimination pass ti @file{@var{file}.08.null}.
2970 Dump after the second flow pass, to @file{@var{file}.24.flow2}.
2973 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
2976 Dump after the peephole pass, to @file{@var{file}.25.peephole2}.
2979 Produce all the dumps listed above.
2982 Print statistics on memory usage, at the end of the run, to
2986 Annotate the assembler output with a comment indicating which
2987 pattern and alternative was used. The length of each instruction is
2991 Dump the RTL in the assembler output as a comment before each instruction.
2992 Also turns on @option{-dp} annotation.
2995 For each of the other indicated dump files (except for
2996 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
2997 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3000 Just generate RTL for a function instead of compiling it. Usually used
3004 Dump debugging information during parsing, to standard error.
3007 @item -fdump-unnumbered
3008 @opindex fdump-unnumbered
3009 When doing debugging dumps (see @option{-d} option above), suppress instruction
3010 numbers and line number note output. This makes it more feasible to
3011 use diff on debugging dumps for compiler invocations with different
3012 options, in particular with and without @option{-g}.
3014 @item -fdump-translation-unit @r{(C and C++ only)}
3015 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3016 @opindex fdump-translation-unit
3017 Dump a representation of the tree structure for the entire translation
3018 unit to a file. The file name is made by appending @file{.tu} to the
3019 source file name. If the @samp{-@var{options}} form is used, @var{options}
3020 controls the details of the dump as described for the
3021 @option{-fdump-tree} options.
3023 @item -fdump-class-hierarchy @r{(C++ only)}
3024 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3025 @opindex fdump-class-hierarchy
3026 Dump a representation of each class's hierarchy and virtual function
3027 table layout to a file. The file name is made by appending @file{.class}
3028 to the source file name. If the @samp{-@var{options}} form is used,
3029 @var{options} controls the details of the dump as described for the
3030 @option{-fdump-tree} options.
3032 @item -fdump-tree-@var{switch} @r{(C++ only)}
3033 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3035 Control the dumping at various stages of processing the intermediate
3036 language tree to a file. The file name is generated by appending a switch
3037 specific suffix to the source file name. If the @samp{-@var{options}}
3038 form is used, @var{options} is a list of @samp{-} separated options that
3039 control the details of the dump. Not all options are applicable to all
3040 dumps, those which are not meaningful will be ignored. The following
3041 options are available
3045 Print the address of each node. Usually this is not meaningful as it
3046 changes according to the environment and source file. Its primary use
3047 is for tying up a dump file with a debug environment.
3049 Inhibit dumping of members of a scope or body of a function merely
3050 because that scope has been reached. Only dump such items when they
3051 are directly reachable by some other path.
3053 Turn on all options.
3056 The following tree dumps are possible:
3059 Dump before any tree based optimization, to @file{@var{file}.original}.
3061 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3063 Dump after function inlining, to @file{@var{file}.inlined}.
3068 Store the usual ``temporary'' intermediate files permanently; place them
3069 in the current directory and name them based on the source file. Thus,
3070 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3071 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3072 preprocessed @file{foo.i} output file even though the compiler now
3073 normally uses an integrated preprocessor.
3077 Report the CPU time taken by each subprocess in the compilation
3078 sequence. For C source files, this is the compiler proper and assembler
3079 (plus the linker if linking is done). The output looks like this:
3086 The first number on each line is the ``user time,'' that is time spent
3087 executing the program itself. The second number is ``system time,''
3088 time spent executing operating system routines on behalf of the program.
3089 Both numbers are in seconds.
3091 @item -print-file-name=@var{library}
3092 @opindex print-file-name
3093 Print the full absolute name of the library file @var{library} that
3094 would be used when linking---and don't do anything else. With this
3095 option, GCC does not compile or link anything; it just prints the
3098 @item -print-multi-directory
3099 @opindex print-multi-directory
3100 Print the directory name corresponding to the multilib selected by any
3101 other switches present in the command line. This directory is supposed
3102 to exist in @env{GCC_EXEC_PREFIX}.
3104 @item -print-multi-lib
3105 @opindex print-multi-lib
3106 Print the mapping from multilib directory names to compiler switches
3107 that enable them. The directory name is separated from the switches by
3108 @samp{;}, and each switch starts with an @samp{@@} instead of the
3109 @samp{-}, without spaces between multiple switches. This is supposed to
3110 ease shell-processing.
3112 @item -print-prog-name=@var{program}
3113 @opindex print-prog-name
3114 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3116 @item -print-libgcc-file-name
3117 @opindex print-libgcc-file-name
3118 Same as @option{-print-file-name=libgcc.a}.
3120 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3121 but you do want to link with @file{libgcc.a}. You can do
3124 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3127 @item -print-search-dirs
3128 @opindex print-search-dirs
3129 Print the name of the configured installation directory and a list of
3130 program and library directories gcc will search---and don't do anything else.
3132 This is useful when gcc prints the error message
3133 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3134 To resolve this you either need to put @file{cpp0} and the other compiler
3135 components where gcc expects to find them, or you can set the environment
3136 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3137 Don't forget the trailing '/'.
3138 @xref{Environment Variables}.
3141 @opindex dumpmachine
3142 Print the compiler's target machine (for example,
3143 @samp{i686-pc-linux-gnu})---and don't do anything else.
3146 @opindex dumpversion
3147 Print the compiler version (for example, @samp{3.0})---and don't do
3152 Print the compiler's built-in specs---and don't do anything else. (This
3153 is used when GCC itself is being built.) @xref{Spec Files}.
3156 @node Optimize Options
3157 @section Options That Control Optimization
3158 @cindex optimize options
3159 @cindex options, optimization
3161 These options control various sorts of optimizations:
3168 Optimize. Optimizing compilation takes somewhat more time, and a lot
3169 more memory for a large function.
3171 Without @option{-O}, the compiler's goal is to reduce the cost of
3172 compilation and to make debugging produce the expected results.
3173 Statements are independent: if you stop the program with a breakpoint
3174 between statements, you can then assign a new value to any variable or
3175 change the program counter to any other statement in the function and
3176 get exactly the results you would expect from the source code.
3178 With @option{-O}, the compiler tries to reduce code size and execution
3179 time, without performing any optimizations that take a great deal of
3184 Optimize even more. GCC performs nearly all supported optimizations
3185 that do not involve a space-speed tradeoff. The compiler does not
3186 perform loop unrolling or function inlining when you specify @option{-O2}.
3187 As compared to @option{-O}, this option increases both compilation time
3188 and the performance of the generated code.
3190 @option{-O2} turns on all optional optimizations except for loop
3191 unrolling, function inlining, and register renaming. It also turns on
3192 the @option{-fforce-mem} and @option{-fstrict-aliasing} option on all
3193 machines and frame pointer elimination on machines where doing so does
3194 not interfere with debugging.
3196 Please note the warning under @option{-fgcse} about
3197 invoking @option{-O2} on programs that use computed gotos.
3201 Optimize yet more. @option{-O3} turns on all optimizations specified by
3202 @option{-O2} and also turns on the @option{-finline-functions} and
3203 @option{-frename-registers} options.
3211 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3212 do not typically increase code size. It also performs further
3213 optimizations designed to reduce code size.
3215 If you use multiple @option{-O} options, with or without level numbers,
3216 the last such option is the one that is effective.
3219 Options of the form @option{-f@var{flag}} specify machine-independent
3220 flags. Most flags have both positive and negative forms; the negative
3221 form of @option{-ffoo} would be @option{-fno-foo}. In the table below,
3222 only one of the forms is listed---the one which is not the default.
3223 You can figure out the other form by either removing @samp{no-} or
3228 @opindex ffloat-store
3229 Do not store floating point variables in registers, and inhibit other
3230 options that might change whether a floating point value is taken from a
3233 @cindex floating point precision
3234 This option prevents undesirable excess precision on machines such as
3235 the 68000 where the floating registers (of the 68881) keep more
3236 precision than a @code{double} is supposed to have. Similarly for the
3237 x86 architecture. For most programs, the excess precision does only
3238 good, but a few programs rely on the precise definition of IEEE floating
3239 point. Use @option{-ffloat-store} for such programs, after modifying
3240 them to store all pertinent intermediate computations into variables.
3242 @item -fno-default-inline
3243 @opindex fno-default-inline
3244 Do not make member functions inline by default merely because they are
3245 defined inside the class scope (C++ only). Otherwise, when you specify
3246 @w{@option{-O}}, member functions defined inside class scope are compiled
3247 inline by default; i.e., you don't need to add @samp{inline} in front of
3248 the member function name.
3250 @item -fno-defer-pop
3251 @opindex fno-defer-pop
3252 Always pop the arguments to each function call as soon as that function
3253 returns. For machines which must pop arguments after a function call,
3254 the compiler normally lets arguments accumulate on the stack for several
3255 function calls and pops them all at once.
3259 Force memory operands to be copied into registers before doing
3260 arithmetic on them. This produces better code by making all memory
3261 references potential common subexpressions. When they are not common
3262 subexpressions, instruction combination should eliminate the separate
3263 register-load. The @option{-O2} option turns on this option.
3266 @opindex fforce-addr
3267 Force memory address constants to be copied into registers before
3268 doing arithmetic on them. This may produce better code just as
3269 @option{-fforce-mem} may.
3271 @item -fomit-frame-pointer
3272 @opindex fomit-frame-pointer
3273 Don't keep the frame pointer in a register for functions that
3274 don't need one. This avoids the instructions to save, set up and
3275 restore frame pointers; it also makes an extra register available
3276 in many functions. @strong{It also makes debugging impossible on
3279 On some machines, such as the VAX, this flag has no effect, because
3280 the standard calling sequence automatically handles the frame pointer
3281 and nothing is saved by pretending it doesn't exist. The
3282 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3283 whether a target machine supports this flag. @xref{Registers,,Register
3284 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3286 @item -foptimize-sibling-calls
3287 @opindex foptimize-sibling-calls
3288 Optimize sibling and tail recursive calls.
3292 This option generates traps for signed overflow on addition, subtraction,
3293 multiplication operations.
3297 Don't pay attention to the @code{inline} keyword. Normally this option
3298 is used to keep the compiler from expanding any functions inline.
3299 Note that if you are not optimizing, no functions can be expanded inline.
3301 @item -finline-functions
3302 @opindex finline-functions
3303 Integrate all simple functions into their callers. The compiler
3304 heuristically decides which functions are simple enough to be worth
3305 integrating in this way.
3307 If all calls to a given function are integrated, and the function is
3308 declared @code{static}, then the function is normally not output as
3309 assembler code in its own right.
3311 @item -finline-limit=@var{n}
3312 @opindex finline-limit
3313 By default, gcc limits the size of functions that can be inlined. This flag
3314 allows the control of this limit for functions that are explicitly marked as
3315 inline (ie marked with the inline keyword or defined within the class
3316 definition in c++). @var{n} is the size of functions that can be inlined in
3317 number of pseudo instructions (not counting parameter handling). The default
3318 value of @var{n} is 600.
3319 Increasing this value can result in more inlined code at
3320 the cost of compilation time and memory consumption. Decreasing usually makes
3321 the compilation faster and less code will be inlined (which presumably
3322 means slower programs). This option is particularly useful for programs that
3323 use inlining heavily such as those based on recursive templates with C++.
3325 @emph{Note:} pseudo instruction represents, in this particular context, an
3326 abstract measurement of function's size. In no way, it represents a count
3327 of assembly instructions and as such its exact meaning might change from one
3328 release to an another.
3330 @item -fkeep-inline-functions
3331 @opindex fkeep-inline-functions
3332 Even if all calls to a given function are integrated, and the function
3333 is declared @code{static}, nevertheless output a separate run-time
3334 callable version of the function. This switch does not affect
3335 @code{extern inline} functions.
3337 @item -fkeep-static-consts
3338 @opindex fkeep-static-consts
3339 Emit variables declared @code{static const} when optimization isn't turned
3340 on, even if the variables aren't referenced.
3342 GCC enables this option by default. If you want to force the compiler to
3343 check if the variable was referenced, regardless of whether or not
3344 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3346 @item -fmerge-constants
3347 Attempt to merge identical constants (string constants and floating point
3348 constants) accross compilation units.
3350 This option is default for optimized compilation if assembler and linker
3351 support it. Use @option{-fno-merge-constants} to inhibit this behavior.
3353 @item -fmerge-all-constants
3354 Attempt to merge identical constants and identical variables.
3356 This option implies @option{-fmerge-constants}. In addition to
3357 @option{-fmerge-constants} this considers e.g. even constant initialized
3358 arrays or initialized constant variables with integral or floating point
3359 types. Languages like C or C++ require each non-automatic variable to
3360 have distinct location, so using this option will result in non-conforming
3363 @item -fno-function-cse
3364 @opindex fno-function-cse
3365 Do not put function addresses in registers; make each instruction that
3366 calls a constant function contain the function's address explicitly.
3368 This option results in less efficient code, but some strange hacks
3369 that alter the assembler output may be confused by the optimizations
3370 performed when this option is not used.
3374 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, and @*
3375 @option{-fno-trapping-math}.
3377 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3379 This option should never be turned on by any @option{-O} option since
3380 it can result in incorrect output for programs which depend on
3381 an exact implementation of IEEE or ISO rules/specifications for
3384 @item -fno-math-errno
3385 @opindex fno-math-errno
3386 Do not set ERRNO after calling math functions that are executed
3387 with a single instruction, e.g., sqrt. A program that relies on
3388 IEEE exceptions for math error handling may want to use this flag
3389 for speed while maintaining IEEE arithmetic compatibility.
3391 This option should never be turned on by any @option{-O} option since
3392 it can result in incorrect output for programs which depend on
3393 an exact implementation of IEEE or ISO rules/specifications for
3396 The default is @option{-fmath-errno}.
3398 @item -funsafe-math-optimizations
3399 @opindex funsafe-math-optimizations
3400 Allow optimizations for floating-point arithmetic that (a) assume
3401 that arguments and results are valid and (b) may violate IEEE or
3402 ANSI standards. When used at link-time, it may include libraries
3403 or startup files that change the default FPU control word or other
3404 similar optimizations.
3406 This option should never be turned on by any @option{-O} option since
3407 it can result in incorrect output for programs which depend on
3408 an exact implementation of IEEE or ISO rules/specifications for
3411 The default is @option{-fno-unsafe-math-optimizations}.
3413 @item -fno-trapping-math
3414 @opindex fno-trapping-math
3415 Compile code assuming that floating-point operations cannot generate
3416 user-visible traps. Setting this option may allow faster code
3417 if one relies on ``non-stop'' IEEE arithmetic, for example.
3419 This option should never be turned on by any @option{-O} option since
3420 it can result in incorrect output for programs which depend on
3421 an exact implementation of IEEE or ISO rules/specifications for
3424 The default is @option{-ftrapping-math}.
3426 @item -fno-zero-initialized-in-bss
3427 @opindex fno-zero-initialized-in-bss
3428 If the target supports a BSS section, GCC by default puts variables that
3429 are initialized to zero into BSS@. This can save space in the resulting
3432 This option turns off this behavior because some programs explicitly
3433 rely on variables going to the data section. E.g., so that the
3434 resulting executable can find the beginning of that section and/or make
3435 assumptions based on that.
3437 The default is @option{-fzero-initialized-in-bss}.
3440 The following options control specific optimizations. The @option{-O2}
3441 option turns on all of these optimizations except @option{-funroll-loops}
3442 and @option{-funroll-all-loops}. On most machines, the @option{-O} option
3443 turns on the @option{-fthread-jumps} and @option{-fdelayed-branch} options,
3444 but specific machines may handle it differently.
3446 You can use the following flags in the rare cases when ``fine-tuning''
3447 of optimizations to be performed is desired.
3449 Not all of the optimizations performed by GCC have @option{-f} options
3453 @item -fstrength-reduce
3454 @opindex fstrength-reduce
3455 Perform the optimizations of loop strength reduction and
3456 elimination of iteration variables.
3458 @item -fthread-jumps
3459 @opindex fthread-jumps
3460 Perform optimizations where we check to see if a jump branches to a
3461 location where another comparison subsumed by the first is found. If
3462 so, the first branch is redirected to either the destination of the
3463 second branch or a point immediately following it, depending on whether
3464 the condition is known to be true or false.
3466 @item -fcse-follow-jumps
3467 @opindex fcse-follow-jumps
3468 In common subexpression elimination, scan through jump instructions
3469 when the target of the jump is not reached by any other path. For
3470 example, when CSE encounters an @code{if} statement with an
3471 @code{else} clause, CSE will follow the jump when the condition
3474 @item -fcse-skip-blocks
3475 @opindex fcse-skip-blocks
3476 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3477 follow jumps which conditionally skip over blocks. When CSE
3478 encounters a simple @code{if} statement with no else clause,
3479 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3480 body of the @code{if}.
3482 @item -frerun-cse-after-loop
3483 @opindex frerun-cse-after-loop
3484 Re-run common subexpression elimination after loop optimizations has been
3487 @item -frerun-loop-opt
3488 @opindex frerun-loop-opt
3489 Run the loop optimizer twice.
3493 Perform a global common subexpression elimination pass.
3494 This pass also performs global constant and copy propagation.
3496 @emph{Note:} When compiling a program using computed gotos, a GCC
3497 extension, you may get better runtime performance if you disable
3498 the global common subexpression elmination pass by adding
3499 @option{-fno-gcse} to the command line.
3503 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3504 attempt to move loads which are only killed by stores into themselves. This
3505 allows a loop containing a load/store sequence to be changed to a load outside
3506 the loop, and a copy/store within the loop.
3510 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3511 subexpression elimination. This pass will attempt to move stores out of loops.
3512 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3513 can be changed to a load before the loop and a store after the loop.
3515 @item -floop-optimize
3516 @opindex floop-optimize
3517 Perform loop optimizations: move constant expressions out of loops, simplify
3518 exit test conditions and optionally do strength-reduction and loop unrolling as
3521 @item -fcrossjumping
3522 @opindex crossjumping
3523 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
3524 resulting code may or may not perform better than without cross-jumping.
3526 @item -fif-conversion
3527 @opindex if-conversion
3528 Attempt to transform conditional jumps into branch-less equivalents. This
3529 include use of conditional moves, min, max, set flags and abs instructions, and
3530 some tricks doable by standard arithmetics. The use of conditional execution
3531 on chips where it is available is controlled by @code{if-conversion2}.
3533 @item -fif-conversion2
3534 @opindex if-conversion2
3535 Use conditional execution (where available) to transform conditional jumps into
3536 branch-less equivalents.
3538 @item -fdelete-null-pointer-checks
3539 @opindex fdelete-null-pointer-checks
3540 Use global dataflow analysis to identify and eliminate useless checks
3541 for null pointers. The compiler assumes that dereferencing a null
3542 pointer would have halted the program. If a pointer is checked after
3543 it has already been dereferenced, it cannot be null.
3545 In some environments, this assumption is not true, and programs can
3546 safely dereference null pointers. Use
3547 @option{-fno-delete-null-pointer-checks} to disable this optimization
3548 for programs which depend on that behavior.
3550 @item -fexpensive-optimizations
3551 @opindex fexpensive-optimizations
3552 Perform a number of minor optimizations that are relatively expensive.
3554 @item -foptimize-register-move
3556 @opindex foptimize-register-move
3558 Attempt to reassign register numbers in move instructions and as
3559 operands of other simple instructions in order to maximize the amount of
3560 register tying. This is especially helpful on machines with two-operand
3561 instructions. GCC enables this optimization by default with @option{-O2}
3564 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3567 @item -fdelayed-branch
3568 @opindex fdelayed-branch
3569 If supported for the target machine, attempt to reorder instructions
3570 to exploit instruction slots available after delayed branch
3573 @item -fschedule-insns
3574 @opindex fschedule-insns
3575 If supported for the target machine, attempt to reorder instructions to
3576 eliminate execution stalls due to required data being unavailable. This
3577 helps machines that have slow floating point or memory load instructions
3578 by allowing other instructions to be issued until the result of the load
3579 or floating point instruction is required.
3581 @item -fschedule-insns2
3582 @opindex fschedule-insns2
3583 Similar to @option{-fschedule-insns}, but requests an additional pass of
3584 instruction scheduling after register allocation has been done. This is
3585 especially useful on machines with a relatively small number of
3586 registers and where memory load instructions take more than one cycle.
3588 @item -ffunction-sections
3589 @itemx -fdata-sections
3590 @opindex ffunction-sections
3591 @opindex fdata-sections
3592 Place each function or data item into its own section in the output
3593 file if the target supports arbitrary sections. The name of the
3594 function or the name of the data item determines the section's name
3597 Use these options on systems where the linker can perform optimizations
3598 to improve locality of reference in the instruction space. HPPA
3599 processors running HP-UX and Sparc processors running Solaris 2 have
3600 linkers with such optimizations. Other systems using the ELF object format
3601 as well as AIX may have these optimizations in the future.
3603 Only use these options when there are significant benefits from doing
3604 so. When you specify these options, the assembler and linker will
3605 create larger object and executable files and will also be slower.
3606 You will not be able to use @code{gprof} on all systems if you
3607 specify this option and you may have problems with debugging if
3608 you specify both this option and @option{-g}.
3610 @item -fcaller-saves
3611 @opindex fcaller-saves
3612 Enable values to be allocated in registers that will be clobbered by
3613 function calls, by emitting extra instructions to save and restore the
3614 registers around such calls. Such allocation is done only when it
3615 seems to result in better code than would otherwise be produced.
3617 This option is always enabled by default on certain machines, usually
3618 those which have no call-preserved registers to use instead.
3620 For all machines, optimization level 2 and higher enables this flag by
3623 @item -funroll-loops
3624 @opindex funroll-loops
3625 Unroll loops whose number of iterations can be determined at compile
3626 time or upon entry to the loop. @option{-funroll-loops} implies both
3627 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3628 option makes code larger, and may or may not make it run faster.
3630 @item -funroll-all-loops
3631 @opindex funroll-all-loops
3632 Unroll all loops, even if their number of iterations is uncertain when
3633 the loop is entered. This usually makes programs run more slowly.
3634 @option{-funroll-all-loops} implies the same options as
3635 @option{-funroll-loops},
3637 @item -fprefetch-loop-arrays
3638 @opindex fprefetch-loop-arrays
3639 If supported by the target machine, generate instructions to prefetch
3640 memory to improve the performance of loops that access large arrays.
3642 @item -fmove-all-movables
3643 @opindex fmove-all-movables
3644 Forces all invariant computations in loops to be moved
3647 @item -freduce-all-givs
3648 @opindex freduce-all-givs
3649 Forces all general-induction variables in loops to be
3652 @emph{Note:} When compiling programs written in Fortran,
3653 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3654 by default when you use the optimizer.
3656 These options may generate better or worse code; results are highly
3657 dependent on the structure of loops within the source code.
3659 These two options are intended to be removed someday, once
3660 they have helped determine the efficacy of various
3661 approaches to improving loop optimizations.
3663 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3664 know how use of these options affects
3665 the performance of your production code.
3666 We're very interested in code that runs @emph{slower}
3667 when these options are @emph{enabled}.
3670 @itemx -fno-peephole2
3671 @opindex fno-peephole
3672 @opindex fno-peephole2
3673 Disable any machine-specific peephole optimizations. The difference
3674 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
3675 are implemented in the compiler; some targets use one, some use the
3676 other, a few use both.
3678 @item -fbranch-probabilities
3679 @opindex fbranch-probabilities
3680 After running a program compiled with @option{-fprofile-arcs}
3681 (@pxref{Debugging Options,, Options for Debugging Your Program or
3682 @command{gcc}}), you can compile it a second time using
3683 @option{-fbranch-probabilities}, to improve optimizations based on
3684 the number of times each branch was taken. When the program
3685 compiled with @option{-fprofile-arcs} exits it saves arc execution
3686 counts to a file called @file{@var{sourcename}.da} for each source
3687 file The information in this data file is very dependent on the
3688 structure of the generated code, so you must use the same source code
3689 and the same optimization options for both compilations.
3691 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3692 note on the first instruction of each basic block, and a
3693 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3694 These can be used to improve optimization. Currently, they are only
3695 used in one place: in @file{reorg.c}, instead of guessing which path a
3696 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3697 exactly determine which path is taken more often.
3699 @item -fno-guess-branch-probability
3700 @opindex fno-guess-branch-probability
3701 Do not guess branch probabilities using a randomized model.
3703 Sometimes gcc will opt to use a randomized model to guess branch
3704 probabilities, when none are available from either profiling feedback
3705 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
3706 different runs of the compiler on the same program may produce different
3709 In a hard real-time system, people don't want different runs of the
3710 compiler to produce code that has different behavior; minimizing
3711 non-determinism is of paramount import. This switch allows users to
3712 reduce non-determinism, possibly at the expense of inferior
3715 @item -fstrict-aliasing
3716 @opindex fstrict-aliasing
3717 Allows the compiler to assume the strictest aliasing rules applicable to
3718 the language being compiled. For C (and C++), this activates
3719 optimizations based on the type of expressions. In particular, an
3720 object of one type is assumed never to reside at the same address as an
3721 object of a different type, unless the types are almost the same. For
3722 example, an @code{unsigned int} can alias an @code{int}, but not a
3723 @code{void*} or a @code{double}. A character type may alias any other
3726 Pay special attention to code like this:
3739 The practice of reading from a different union member than the one most
3740 recently written to (called ``type-punning'') is common. Even with
3741 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
3742 is accessed through the union type. So, the code above will work as
3743 expected. However, this code might not:
3754 Every language that wishes to perform language-specific alias analysis
3755 should define a function that computes, given an @code{tree}
3756 node, an alias set for the node. Nodes in different alias sets are not
3757 allowed to alias. For an example, see the C front-end function
3758 @code{c_get_alias_set}.
3760 For all machines, optimization level 2 and higher enables this flag by
3763 @item -falign-functions
3764 @itemx -falign-functions=@var{n}
3765 @opindex falign-functions
3766 Align the start of functions to the next power-of-two greater than
3767 @var{n}, skipping up to @var{n} bytes. For instance,
3768 @option{-falign-functions=32} aligns functions to the next 32-byte
3769 boundary, but @option{-falign-functions=24} would align to the next
3770 32-byte boundary only if this can be done by skipping 23 bytes or less.
3772 @option{-fno-align-functions} and @option{-falign-functions=1} are
3773 equivalent and mean that functions will not be aligned.
3775 Some assemblers only support this flag when @var{n} is a power of two;
3776 in that case, it is rounded up.
3778 If @var{n} is not specified, use a machine-dependent default.
3780 @item -falign-labels
3781 @itemx -falign-labels=@var{n}
3782 @opindex falign-labels
3783 Align all branch targets to a power-of-two boundary, skipping up to
3784 @var{n} bytes like @option{-falign-functions}. This option can easily
3785 make code slower, because it must insert dummy operations for when the
3786 branch target is reached in the usual flow of the code.
3788 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
3789 are greater than this value, then their values are used instead.
3791 If @var{n} is not specified, use a machine-dependent default which is
3792 very likely to be @samp{1}, meaning no alignment.
3795 @itemx -falign-loops=@var{n}
3796 @opindex falign-loops
3797 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3798 like @option{-falign-functions}. The hope is that the loop will be
3799 executed many times, which will make up for any execution of the dummy
3802 If @var{n} is not specified, use a machine-dependent default.
3805 @itemx -falign-jumps=@var{n}
3806 @opindex falign-jumps
3807 Align branch targets to a power-of-two boundary, for branch targets
3808 where the targets can only be reached by jumping, skipping up to @var{n}
3809 bytes like @option{-falign-functions}. In this case, no dummy operations
3812 If @var{n} is not specified, use a machine-dependent default.
3816 Perform optimizations in static single assignment form. Each function's
3817 flow graph is translated into SSA form, optimizations are performed, and
3818 the flow graph is translated back from SSA form. Users should not
3819 specify this option, since it is not yet ready for production use.
3823 Perform Sparse Conditional Constant Propagation in SSA form. Requires
3824 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
3828 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
3829 Like @option{-fssa}, this is an experimental feature.
3831 @item -fsingle-precision-constant
3832 @opindex fsingle-precision-constant
3833 Treat floating point constant as single precision constant instead of
3834 implicitly converting it to double precision constant.
3836 @item -frename-registers
3837 @opindex frename-registers
3838 Attempt to avoid false dependencies in scheduled code by making use
3839 of registers left over after register allocation. This optimization
3840 will most benefit processors with lots of registers. It can, however,
3841 make debugging impossible, since variables will no longer stay in
3842 a ``home register''.
3844 @item -fno-cprop-registers
3845 @opindex fno-cprop-registers
3846 After register allocation and post-register allocation instruction splitting,
3847 we perform a copy-propagation pass to try to reduce scheduling dependencies
3848 and occasionally eliminate the copy.
3850 @item --param @var{name}=@var{value}
3852 In some places, GCC uses various constants to control the amount of
3853 optimization that is done. For example, GCC will not inline functions
3854 that contain more that a certain number of instructions. You can
3855 control some of these constants on the command-line using the
3856 @option{--param} option.
3858 In each case, the @var{value} is an integer. The allowable choices for
3859 @var{name} are given in the following table:
3862 @item max-delay-slot-insn-search
3863 The maximum number of instructions to consider when looking for an
3864 instruction to fill a delay slot. If more than this arbitrary number of
3865 instructions is searched, the time savings from filling the delay slot
3866 will be minimal so stop searching. Increasing values mean more
3867 aggressive optimization, making the compile time increase with probably
3868 small improvement in executable run time.
3870 @item max-delay-slot-live-search
3871 When trying to fill delay slots, the maximum number of instructions to
3872 consider when searching for a block with valid live register
3873 information. Increasing this arbitrarily chosen value means more
3874 aggressive optimization, increasing the compile time. This parameter
3875 should be removed when the delay slot code is rewritten to maintain the
3878 @item max-gcse-memory
3879 The approximate maximum amount of memory that will be allocated in
3880 order to perform the global common subexpression elimination
3881 optimization. If more memory than specified is required, the
3882 optimization will not be done.
3884 @item max-gcse-passes
3885 The maximum number of passes of GCSE to run.
3887 @item max-pending-list-length
3888 The maximum number of pending dependencies scheduling will allow
3889 before flushing the current state and starting over. Large functions
3890 with few branches or calls can create excessively large lists which
3891 needlessly consume memory and resources.
3893 @item max-inline-insns
3894 If an function contains more than this many instructions, it
3895 will not be inlined. This option is precisely equivalent to
3896 @option{-finline-limit}.
3898 @item max-unrolled-insns
3899 The maximum number of instructions that a loop should have if that loop
3900 is unrolled, and if the loop is unrolled, it determines how many times
3901 the loop code is unrolled.
3906 @node Preprocessor Options
3907 @section Options Controlling the Preprocessor
3908 @cindex preprocessor options
3909 @cindex options, preprocessor
3911 These options control the C preprocessor, which is run on each C source
3912 file before actual compilation.
3914 If you use the @option{-E} option, nothing is done except preprocessing.
3915 Some of these options make sense only together with @option{-E} because
3916 they cause the preprocessor output to be unsuitable for actual
3920 You can use @option{-Wp,@var{option}} to bypass the compiler driver
3921 and pass @var{option} directly through to the preprocessor. If
3922 @var{option} contains commas, it is split into multiple options at the
3923 commas. However, many options are modified, translated or interpreted
3924 by the compiler driver before being passed to the preprocessor, and
3925 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
3926 interface is undocumented and subject to change, so whenever possible
3927 you should avoid using @option{-Wp} and let the driver handle the
3930 @include cppopts.texi
3932 @node Assembler Options
3933 @section Passing Options to the Assembler
3935 @c prevent bad page break with this line
3936 You can pass options to the assembler.
3939 @item -Wa,@var{option}
3941 Pass @var{option} as an option to the assembler. If @var{option}
3942 contains commas, it is split into multiple options at the commas.
3946 @section Options for Linking
3947 @cindex link options
3948 @cindex options, linking
3950 These options come into play when the compiler links object files into
3951 an executable output file. They are meaningless if the compiler is
3952 not doing a link step.
3956 @item @var{object-file-name}
3957 A file name that does not end in a special recognized suffix is
3958 considered to name an object file or library. (Object files are
3959 distinguished from libraries by the linker according to the file
3960 contents.) If linking is done, these object files are used as input
3969 If any of these options is used, then the linker is not run, and
3970 object file names should not be used as arguments. @xref{Overall
3974 @item -l@var{library}
3975 @itemx -l @var{library}
3977 Search the library named @var{library} when linking. (The second
3978 alternative with the library as a separate argument is only for
3979 POSIX compliance and is not recommended.)
3981 It makes a difference where in the command you write this option; the
3982 linker searches and processes libraries and object files in the order they
3983 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
3984 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
3985 to functions in @samp{z}, those functions may not be loaded.
3987 The linker searches a standard list of directories for the library,
3988 which is actually a file named @file{lib@var{library}.a}. The linker
3989 then uses this file as if it had been specified precisely by name.
3991 The directories searched include several standard system directories
3992 plus any that you specify with @option{-L}.
3994 Normally the files found this way are library files---archive files
3995 whose members are object files. The linker handles an archive file by
3996 scanning through it for members which define symbols that have so far
3997 been referenced but not defined. But if the file that is found is an
3998 ordinary object file, it is linked in the usual fashion. The only
3999 difference between using an @option{-l} option and specifying a file name
4000 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4001 and searches several directories.
4005 You need this special case of the @option{-l} option in order to
4006 link an Objective-C program.
4009 @opindex nostartfiles
4010 Do not use the standard system startup files when linking.
4011 The standard system libraries are used normally, unless @option{-nostdlib}
4012 or @option{-nodefaultlibs} is used.
4014 @item -nodefaultlibs
4015 @opindex nodefaultlibs
4016 Do not use the standard system libraries when linking.
4017 Only the libraries you specify will be passed to the linker.
4018 The standard startup files are used normally, unless @option{-nostartfiles}
4019 is used. The compiler may generate calls to memcmp, memset, and memcpy
4020 for System V (and ISO C) environments or to bcopy and bzero for
4021 BSD environments. These entries are usually resolved by entries in
4022 libc. These entry points should be supplied through some other
4023 mechanism when this option is specified.
4027 Do not use the standard system startup files or libraries when linking.
4028 No startup files and only the libraries you specify will be passed to
4029 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4030 for System V (and ISO C) environments or to bcopy and bzero for
4031 BSD environments. These entries are usually resolved by entries in
4032 libc. These entry points should be supplied through some other
4033 mechanism when this option is specified.
4035 @cindex @option{-lgcc}, use with @option{-nostdlib}
4036 @cindex @option{-nostdlib} and unresolved references
4037 @cindex unresolved references and @option{-nostdlib}
4038 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4039 @cindex @option{-nodefaultlibs} and unresolved references
4040 @cindex unresolved references and @option{-nodefaultlibs}
4041 One of the standard libraries bypassed by @option{-nostdlib} and
4042 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4043 that GCC uses to overcome shortcomings of particular machines, or special
4044 needs for some languages.
4045 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4046 Collection (GCC) Internals},
4047 for more discussion of @file{libgcc.a}.)
4048 In most cases, you need @file{libgcc.a} even when you want to avoid
4049 other standard libraries. In other words, when you specify @option{-nostdlib}
4050 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4051 This ensures that you have no unresolved references to internal GCC
4052 library subroutines. (For example, @samp{__main}, used to ensure C++
4053 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4054 GNU Compiler Collection (GCC) Internals}.)
4058 Remove all symbol table and relocation information from the executable.
4062 On systems that support dynamic linking, this prevents linking with the shared
4063 libraries. On other systems, this option has no effect.
4067 Produce a shared object which can then be linked with other objects to
4068 form an executable. Not all systems support this option. For predictable
4069 results, you must also specify the same set of options that were used to
4070 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4071 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4072 needs to build supplementary stub code for constructors to work. On
4073 multi-libbed systems, @samp{gcc -shared} must select the correct support
4074 libraries to link against. Failing to supply the correct flags may lead
4075 to subtle defects. Supplying them in cases where they are not necessary
4078 @item -shared-libgcc
4079 @itemx -static-libgcc
4080 @opindex shared-libgcc
4081 @opindex static-libgcc
4082 On systems that provide @file{libgcc} as a shared library, these options
4083 force the use of either the shared or static version respectively.
4084 If no shared version of @file{libgcc} was built when the compiler was
4085 configured, these options have no effect.
4087 There are several situations in which an application should use the
4088 shared @file{libgcc} instead of the static version. The most common
4089 of these is when the application wishes to throw and catch exceptions
4090 across different shared libraries. In that case, each of the libraries
4091 as well as the application itself should use the shared @file{libgcc}.
4093 Therefore, the G++ and GCJ drivers automatically add
4094 @option{-shared-libgcc} whenever you build a shared library or a main
4095 executable, because C++ and Java programs typically use exceptions, so
4096 this is the right thing to do.
4098 If, instead, you use the GCC driver to create shared libraries, you may
4099 find that they will not always be linked with the shared @file{libgcc}.
4100 If GCC finds, at its configuration time, that you have a GNU linker that
4101 does not support option @option{--eh-frame-hdr}, it will link the shared
4102 version of @file{libgcc} into shared libraries by default. Otherwise,
4103 it will take advantage of the linker and optimize away the linking with
4104 the shared version of @file{libgcc}, linking with the static version of
4105 libgcc by default. This allows exceptions to propagate through such
4106 shared libraries, without incurring relocation costs at library load
4109 However, if a library or main executable is supposed to throw or catch
4110 exceptions, you must link it using the G++ or GCJ driver, as appropriate
4111 for the languages used in the program, or using the option
4112 @option{-shared-libgcc}, such that it is linked with the shared
4117 Bind references to global symbols when building a shared object. Warn
4118 about any unresolved references (unless overridden by the link editor
4119 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4122 @item -Xlinker @var{option}
4124 Pass @var{option} as an option to the linker. You can use this to
4125 supply system-specific linker options which GCC does not know how to
4128 If you want to pass an option that takes an argument, you must use
4129 @option{-Xlinker} twice, once for the option and once for the argument.
4130 For example, to pass @option{-assert definitions}, you must write
4131 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4132 @option{-Xlinker "-assert definitions"}, because this passes the entire
4133 string as a single argument, which is not what the linker expects.
4135 @item -Wl,@var{option}
4137 Pass @var{option} as an option to the linker. If @var{option} contains
4138 commas, it is split into multiple options at the commas.
4140 @item -u @var{symbol}
4142 Pretend the symbol @var{symbol} is undefined, to force linking of
4143 library modules to define it. You can use @option{-u} multiple times with
4144 different symbols to force loading of additional library modules.
4147 @node Directory Options
4148 @section Options for Directory Search
4149 @cindex directory options
4150 @cindex options, directory search
4153 These options specify directories to search for header files, for
4154 libraries and for parts of the compiler:
4159 Add the directory @var{dir} to the head of the list of directories to be
4160 searched for header files. This can be used to override a system header
4161 file, substituting your own version, since these directories are
4162 searched before the system header file directories. However, you should
4163 not use this option to add directories that contain vendor-supplied
4164 system header files (use @option{-isystem} for that). If you use more than
4165 one @option{-I} option, the directories are scanned in left-to-right
4166 order; the standard system directories come after.
4168 If a standard system include directory, or a directory specified with
4169 @option{-isystem}, is also specified with @option{-I}, it will be
4170 searched only in the position requested by @option{-I}. Also, it will
4171 not be considered a system include directory. If that directory really
4172 does contain system headers, there is a good chance that they will
4173 break. For instance, if GCC's installation procedure edited the headers
4174 in @file{/usr/include} to fix bugs, @samp{-I/usr/include} will cause the
4175 original, buggy headers to be found instead of the corrected ones. GCC
4176 will issue a warning when a system include directory is hidden in this
4181 Any directories you specify with @option{-I} options before the @option{-I-}
4182 option are searched only for the case of @samp{#include "@var{file}"};
4183 they are not searched for @samp{#include <@var{file}>}.
4185 If additional directories are specified with @option{-I} options after
4186 the @option{-I-}, these directories are searched for all @samp{#include}
4187 directives. (Ordinarily @emph{all} @option{-I} directories are used
4190 In addition, the @option{-I-} option inhibits the use of the current
4191 directory (where the current input file came from) as the first search
4192 directory for @samp{#include "@var{file}"}. There is no way to
4193 override this effect of @option{-I-}. With @option{-I.} you can specify
4194 searching the directory which was current when the compiler was
4195 invoked. That is not exactly the same as what the preprocessor does
4196 by default, but it is often satisfactory.
4198 @option{-I-} does not inhibit the use of the standard system directories
4199 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4204 Add directory @var{dir} to the list of directories to be searched
4207 @item -B@var{prefix}
4209 This option specifies where to find the executables, libraries,
4210 include files, and data files of the compiler itself.
4212 The compiler driver program runs one or more of the subprograms
4213 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4214 @var{prefix} as a prefix for each program it tries to run, both with and
4215 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4217 For each subprogram to be run, the compiler driver first tries the
4218 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4219 was not specified, the driver tries two standard prefixes, which are
4220 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4221 those results in a file name that is found, the unmodified program
4222 name is searched for using the directories specified in your
4223 @env{PATH} environment variable.
4225 The compiler will check to see if the path provided by the @option{-B}
4226 refers to a directory, and if necessary it will add a directory
4227 separator character at the end of the path.
4229 @option{-B} prefixes that effectively specify directory names also apply
4230 to libraries in the linker, because the compiler translates these
4231 options into @option{-L} options for the linker. They also apply to
4232 includes files in the preprocessor, because the compiler translates these
4233 options into @option{-isystem} options for the preprocessor. In this case,
4234 the compiler appends @samp{include} to the prefix.
4236 The run-time support file @file{libgcc.a} can also be searched for using
4237 the @option{-B} prefix, if needed. If it is not found there, the two
4238 standard prefixes above are tried, and that is all. The file is left
4239 out of the link if it is not found by those means.
4241 Another way to specify a prefix much like the @option{-B} prefix is to use
4242 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4245 As a special kludge, if the path provided by @option{-B} is
4246 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4247 9, then it will be replaced by @file{[dir/]include}. This is to help
4248 with boot-strapping the compiler.
4250 @item -specs=@var{file}
4252 Process @var{file} after the compiler reads in the standard @file{specs}
4253 file, in order to override the defaults that the @file{gcc} driver
4254 program uses when determining what switches to pass to @file{cc1},
4255 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4256 @option{-specs=@var{file}} can be specified on the command line, and they
4257 are processed in order, from left to right.
4263 @section Specifying subprocesses and the switches to pass to them
4265 @command{gcc} is a driver program. It performs its job by invoking a
4266 sequence of other programs to do the work of compiling, assembling and
4267 linking. GCC interprets its command-line parameters and uses these to
4268 deduce which programs it should invoke, and which command-line options
4269 it ought to place on their command lines. This behavior is controlled
4270 by @dfn{spec strings}. In most cases there is one spec string for each
4271 program that GCC can invoke, but a few programs have multiple spec
4272 strings to control their behavior. The spec strings built into GCC can
4273 be overridden by using the @option{-specs=} command-line switch to specify
4276 @dfn{Spec files} are plaintext files that are used to construct spec
4277 strings. They consist of a sequence of directives separated by blank
4278 lines. The type of directive is determined by the first non-whitespace
4279 character on the line and it can be one of the following:
4282 @item %@var{command}
4283 Issues a @var{command} to the spec file processor. The commands that can
4287 @item %include <@var{file}>
4289 Search for @var{file} and insert its text at the current point in the
4292 @item %include_noerr <@var{file}>
4293 @cindex %include_noerr
4294 Just like @samp{%include}, but do not generate an error message if the include
4295 file cannot be found.
4297 @item %rename @var{old_name} @var{new_name}
4299 Rename the spec string @var{old_name} to @var{new_name}.
4303 @item *[@var{spec_name}]:
4304 This tells the compiler to create, override or delete the named spec
4305 string. All lines after this directive up to the next directive or
4306 blank line are considered to be the text for the spec string. If this
4307 results in an empty string then the spec will be deleted. (Or, if the
4308 spec did not exist, then nothing will happened.) Otherwise, if the spec
4309 does not currently exist a new spec will be created. If the spec does
4310 exist then its contents will be overridden by the text of this
4311 directive, unless the first character of that text is the @samp{+}
4312 character, in which case the text will be appended to the spec.
4314 @item [@var{suffix}]:
4315 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4316 and up to the next directive or blank line are considered to make up the
4317 spec string for the indicated suffix. When the compiler encounters an
4318 input file with the named suffix, it will processes the spec string in
4319 order to work out how to compile that file. For example:
4326 This says that any input file whose name ends in @samp{.ZZ} should be
4327 passed to the program @samp{z-compile}, which should be invoked with the
4328 command-line switch @option{-input} and with the result of performing the
4329 @samp{%i} substitution. (See below.)
4331 As an alternative to providing a spec string, the text that follows a
4332 suffix directive can be one of the following:
4335 @item @@@var{language}
4336 This says that the suffix is an alias for a known @var{language}. This is
4337 similar to using the @option{-x} command-line switch to GCC to specify a
4338 language explicitly. For example:
4345 Says that .ZZ files are, in fact, C++ source files.
4348 This causes an error messages saying:
4351 @var{name} compiler not installed on this system.
4355 GCC already has an extensive list of suffixes built into it.
4356 This directive will add an entry to the end of the list of suffixes, but
4357 since the list is searched from the end backwards, it is effectively
4358 possible to override earlier entries using this technique.
4362 GCC has the following spec strings built into it. Spec files can
4363 override these strings or create their own. Note that individual
4364 targets can also add their own spec strings to this list.
4367 asm Options to pass to the assembler
4368 asm_final Options to pass to the assembler post-processor
4369 cpp Options to pass to the C preprocessor
4370 cc1 Options to pass to the C compiler
4371 cc1plus Options to pass to the C++ compiler
4372 endfile Object files to include at the end of the link
4373 link Options to pass to the linker
4374 lib Libraries to include on the command line to the linker
4375 libgcc Decides which GCC support library to pass to the linker
4376 linker Sets the name of the linker
4377 predefines Defines to be passed to the C preprocessor
4378 signed_char Defines to pass to CPP to say whether @code{char} is signed
4380 startfile Object files to include at the start of the link
4383 Here is a small example of a spec file:
4389 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4392 This example renames the spec called @samp{lib} to @samp{old_lib} and
4393 then overrides the previous definition of @samp{lib} with a new one.
4394 The new definition adds in some extra command-line options before
4395 including the text of the old definition.
4397 @dfn{Spec strings} are a list of command-line options to be passed to their
4398 corresponding program. In addition, the spec strings can contain
4399 @samp{%}-prefixed sequences to substitute variable text or to
4400 conditionally insert text into the command line. Using these constructs
4401 it is possible to generate quite complex command lines.
4403 Here is a table of all defined @samp{%}-sequences for spec
4404 strings. Note that spaces are not generated automatically around the
4405 results of expanding these sequences. Therefore you can concatenate them
4406 together or combine them with constant text in a single argument.
4410 Substitute one @samp{%} into the program name or argument.
4413 Substitute the name of the input file being processed.
4416 Substitute the basename of the input file being processed.
4417 This is the substring up to (and not including) the last period
4418 and not including the directory.
4421 This is the same as @samp{%b}, but include the file suffix (text after
4425 Marks the argument containing or following the @samp{%d} as a
4426 temporary file name, so that that file will be deleted if GCC exits
4427 successfully. Unlike @samp{%g}, this contributes no text to the
4430 @item %g@var{suffix}
4431 Substitute a file name that has suffix @var{suffix} and is chosen
4432 once per compilation, and mark the argument in the same way as
4433 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4434 name is now chosen in a way that is hard to predict even when previously
4435 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4436 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4437 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4438 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4439 was simply substituted with a file name chosen once per compilation,
4440 without regard to any appended suffix (which was therefore treated
4441 just like ordinary text), making such attacks more likely to succeed.
4443 @item %u@var{suffix}
4444 Like @samp{%g}, but generates a new temporary file name even if
4445 @samp{%u@var{suffix}} was already seen.
4447 @item %U@var{suffix}
4448 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4449 new one if there is no such last file name. In the absence of any
4450 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4451 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4452 would involve the generation of two distinct file names, one
4453 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4454 simply substituted with a file name chosen for the previous @samp{%u},
4455 without regard to any appended suffix.
4457 @item %j@var{SUFFIX}
4458 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4459 writable, and if save-temps is off; otherwise, substitute the name
4460 of a temporary file, just like @samp{%u}. This temporary file is not
4461 meant for communication between processes, but rather as a junk
4464 @item %.@var{SUFFIX}
4465 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4466 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4467 terminated by the next space or %.
4470 Marks the argument containing or following the @samp{%w} as the
4471 designated output file of this compilation. This puts the argument
4472 into the sequence of arguments that @samp{%o} will substitute later.
4475 Substitutes the names of all the output files, with spaces
4476 automatically placed around them. You should write spaces
4477 around the @samp{%o} as well or the results are undefined.
4478 @samp{%o} is for use in the specs for running the linker.
4479 Input files whose names have no recognized suffix are not compiled
4480 at all, but they are included among the output files, so they will
4484 Substitutes the suffix for object files. Note that this is
4485 handled specially when it immediately follows @samp{%g, %u, or %U},
4486 because of the need for those to form complete file names. The
4487 handling is such that @samp{%O} is treated exactly as if it had already
4488 been substituted, except that @samp{%g, %u, and %U} do not currently
4489 support additional @var{suffix} characters following @samp{%O} as they would
4490 following, for example, @samp{.o}.
4493 Substitutes the standard macro predefinitions for the
4494 current target machine. Use this when running @code{cpp}.
4497 Like @samp{%p}, but puts @samp{__} before and after the name of each
4498 predefined macro, except for macros that start with @samp{__} or with
4499 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4503 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4506 Current argument is the name of a library or startup file of some sort.
4507 Search for that file in a standard list of directories and substitute
4508 the full name found.
4511 Print @var{str} as an error message. @var{str} is terminated by a newline.
4512 Use this when inconsistent options are detected.
4515 Output @samp{-} if the input for the current command is coming from a pipe.
4518 Substitute the contents of spec string @var{name} at this point.
4521 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4523 @item %x@{@var{option}@}
4524 Accumulate an option for @samp{%X}.
4527 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4531 Output the accumulated assembler options specified by @option{-Wa}.
4534 Output the accumulated preprocessor options specified by @option{-Wp}.
4537 Substitute the major version number of GCC@.
4538 (For version 2.9.5, this is 2.)
4541 Substitute the minor version number of GCC@.
4542 (For version 2.9.5, this is 9.)
4545 Substitute the patch level number of GCC@.
4546 (For version 2.9.5, this is 5.)
4549 Process the @code{asm} spec. This is used to compute the
4550 switches to be passed to the assembler.
4553 Process the @code{asm_final} spec. This is a spec string for
4554 passing switches to an assembler post-processor, if such a program is
4558 Process the @code{link} spec. This is the spec for computing the
4559 command line passed to the linker. Typically it will make use of the
4560 @samp{%L %G %S %D and %E} sequences.
4563 Dump out a @option{-L} option for each directory that GCC believes might
4564 contain startup files. If the target supports multilibs then the
4565 current multilib directory will be prepended to each of these paths.
4568 Output the multilib directory with directory separators replaced with
4569 @samp{_}. If multilib directories are not set, or the multilib directory is
4570 @file{.} then this option emits nothing.
4573 Process the @code{lib} spec. This is a spec string for deciding which
4574 libraries should be included on the command line to the linker.
4577 Process the @code{libgcc} spec. This is a spec string for deciding
4578 which GCC support library should be included on the command line to the linker.
4581 Process the @code{startfile} spec. This is a spec for deciding which
4582 object files should be the first ones passed to the linker. Typically
4583 this might be a file named @file{crt0.o}.
4586 Process the @code{endfile} spec. This is a spec string that specifies
4587 the last object files that will be passed to the linker.
4590 Process the @code{cpp} spec. This is used to construct the arguments
4591 to be passed to the C preprocessor.
4594 Process the @code{signed_char} spec. This is intended to be used
4595 to tell cpp whether a char is signed. It typically has the definition:
4597 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4601 Process the @code{cc1} spec. This is used to construct the options to be
4602 passed to the actual C compiler (@samp{cc1}).
4605 Process the @code{cc1plus} spec. This is used to construct the options to be
4606 passed to the actual C++ compiler (@samp{cc1plus}).
4609 Substitute the variable part of a matched option. See below.
4610 Note that each comma in the substituted string is replaced by
4614 Substitutes the @code{-S} switch, if that switch was given to GCC@.
4615 If that switch was not specified, this substitutes nothing. Note that
4616 the leading dash is omitted when specifying this option, and it is
4617 automatically inserted if the substitution is performed. Thus the spec
4618 string @samp{%@{foo@}} would match the command-line option @option{-foo}
4619 and would output the command line option @option{-foo}.
4621 @item %W@{@code{S}@}
4622 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4625 @item %@{@code{S}*@}
4626 Substitutes all the switches specified to GCC whose names start
4627 with @code{-S}, but which also take an argument. This is used for
4628 switches like @option{-o}, @option{-D}, @option{-I}, etc.
4629 GCC considers @option{-o foo} as being
4630 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4631 text, including the space. Thus two arguments would be generated.
4633 @item %@{^@code{S}*@}
4634 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4635 argument. Thus %@{^o*@} would only generate one argument, not two.
4637 @item %@{@code{S}*&@code{T}*@}
4638 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
4639 (the order of @code{S} and @code{T} in the spec is not significant).
4640 There can be any number of ampersand-separated variables; for each the
4641 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
4643 @item %@{<@code{S}@}
4644 Remove all occurrences of @code{-S} from the command line. Note---this
4645 command is position dependent. @samp{%} commands in the spec string
4646 before this option will see @code{-S}, @samp{%} commands in the spec
4647 string after this option will not.
4649 @item %@{@code{S}*:@code{X}@}
4650 Substitutes @code{X} if one or more switches whose names start with
4651 @code{-S} are specified to GCC@. Note that the tail part of the
4652 @code{-S} option (i.e.@: the part matched by the @samp{*}) will be substituted
4653 for each occurrence of @samp{%*} within @code{X}.
4655 @item %@{@code{S}:@code{X}@}
4656 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC@.
4658 @item %@{!@code{S}:@code{X}@}
4659 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC@.
4661 @item %@{|@code{S}:@code{X}@}
4662 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4664 @item %@{|!@code{S}:@code{X}@}
4665 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4667 @item %@{.@code{S}:@code{X}@}
4668 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4670 @item %@{!.@code{S}:@code{X}@}
4671 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4673 @item %@{@code{S}|@code{P}:@code{X}@}
4674 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. This may be
4675 combined with @samp{!} and @samp{.} sequences as well, although they
4676 have a stronger binding than the @samp{|}. For example a spec string
4680 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4683 will output the following command-line options from the following input
4684 command-line options:
4689 -d fred.c -foo -baz -boggle
4690 -d jim.d -bar -baz -boggle
4695 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4696 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4697 or spaces, or even newlines. They are processed as usual, as described
4700 The @option{-O}, @option{-f}, @option{-m}, and @option{-W}
4701 switches are handled specifically in these
4702 constructs. If another value of @option{-O} or the negated form of a @option{-f}, @option{-m}, or
4703 @option{-W} switch is found later in the command line, the earlier switch
4704 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4705 letter, which passes all matching options.
4707 The character @samp{|} at the beginning of the predicate text is used to indicate
4708 that a command should be piped to the following command, but only if @option{-pipe}
4711 It is built into GCC which switches take arguments and which do not.
4712 (You might think it would be useful to generalize this to allow each
4713 compiler's spec to say which switches take arguments. But this cannot
4714 be done in a consistent fashion. GCC cannot even decide which input
4715 files have been specified without knowing which switches take arguments,
4716 and it must know which input files to compile in order to tell which
4719 GCC also knows implicitly that arguments starting in @option{-l} are to be
4720 treated as compiler output files, and passed to the linker in their
4721 proper position among the other output files.
4723 @c man begin OPTIONS
4725 @node Target Options
4726 @section Specifying Target Machine and Compiler Version
4727 @cindex target options
4728 @cindex cross compiling
4729 @cindex specifying machine version
4730 @cindex specifying compiler version and target machine
4731 @cindex compiler version, specifying
4732 @cindex target machine, specifying
4734 By default, GCC compiles code for the same type of machine that you
4735 are using. However, it can also be installed as a cross-compiler, to
4736 compile for some other type of machine. In fact, several different
4737 configurations of GCC, for different target machines, can be
4738 installed side by side. Then you specify which one to use with the
4741 In addition, older and newer versions of GCC can be installed side
4742 by side. One of them (probably the newest) will be the default, but
4743 you may sometimes wish to use another.
4746 @item -b @var{machine}
4748 The argument @var{machine} specifies the target machine for compilation.
4749 This is useful when you have installed GCC as a cross-compiler.
4751 The value to use for @var{machine} is the same as was specified as the
4752 machine type when configuring GCC as a cross-compiler. For
4753 example, if a cross-compiler was configured with @samp{configure
4754 i386v}, meaning to compile for an 80386 running System V, then you
4755 would specify @option{-b i386v} to run that cross compiler.
4757 When you do not specify @option{-b}, it normally means to compile for
4758 the same type of machine that you are using.
4761 The @option{-b} option actually works by controlling part of
4762 the file name used for the executable files and libraries used for
4763 compilation. A given version of GCC, for a given target machine, is
4764 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.
4766 Thus, sites can customize the effect of @option{-b} either by
4767 changing the names of these directories or adding alternate names (or
4768 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
4769 file @file{80386} is a link to the file @file{i386v}, then @option{-b
4770 80386} becomes an alias for @option{-b i386v}.
4772 In one respect, @option{-b} does not completely change
4773 to a different compiler: the top-level driver program @command{gcc}
4774 that you originally invoked continues to run and invoke the other
4775 executables (preprocessor, compiler per se, assembler and linker)
4776 that do the real work. However, since no real work is done in the
4777 driver program, it usually does not matter that the driver program
4778 in use is not the one for the specified target.
4780 The only way that the driver program depends on the target machine is
4781 in the parsing and handling of special machine-specific options.
4782 However, this is controlled by a file which is found, along with the
4783 other executables, in the directory for the specified version and
4784 target machine. As a result, a single installed driver program adapts
4785 to any specified target machine, and sufficiently similar compiler
4788 The driver program executable does control one significant thing,
4789 however: the default version and target machine. Therefore, you can
4790 install different instances of the driver program, compiled for
4791 different targets or versions, under different names.
4793 For example, if the driver for version 2.0 is installed as @command{ogcc}
4794 and that for version 2.1 is installed as @command{gcc}, then the command
4795 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
4798 @node Submodel Options
4799 @section Hardware Models and Configurations
4800 @cindex submodel options
4801 @cindex specifying hardware config
4802 @cindex hardware models and configurations, specifying
4803 @cindex machine dependent options
4805 Earlier we discussed the standard option @option{-b} which chooses among
4806 different installed compilers for completely different target
4807 machines, such as VAX vs.@: 68000 vs.@: 80386.
4809 In addition, each of these target machine types can have its own
4810 special options, starting with @samp{-m}, to choose among various
4811 hardware models or configurations---for example, 68010 vs 68020,
4812 floating coprocessor or none. A single installed version of the
4813 compiler can compile for any model or configuration, according to the
4816 Some configurations of the compiler also support additional special
4817 options, usually for compatibility with other compilers on the same
4820 These options are defined by the macro @code{TARGET_SWITCHES} in the
4821 machine description. The default for the options is also defined by
4822 that macro, which enables you to change the defaults.
4836 * RS/6000 and PowerPC Options::
4839 * i386 and x86-64 Options::
4841 * Intel 960 Options::
4842 * DEC Alpha Options::
4843 * DEC Alpha/VMS Options::
4847 * System V Options::
4848 * TMS320C3x/C4x Options::
4856 * S/390 and zSeries Options::
4860 * Xstormy16 Options::
4864 @node M680x0 Options
4865 @subsection M680x0 Options
4866 @cindex M680x0 options
4868 These are the @samp{-m} options defined for the 68000 series. The default
4869 values for these options depends on which style of 68000 was selected when
4870 the compiler was configured; the defaults for the most common choices are
4878 Generate output for a 68000. This is the default
4879 when the compiler is configured for 68000-based systems.
4881 Use this option for microcontrollers with a 68000 or EC000 core,
4882 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
4888 Generate output for a 68020. This is the default
4889 when the compiler is configured for 68020-based systems.
4893 Generate output containing 68881 instructions for floating point.
4894 This is the default for most 68020 systems unless @option{--nfp} was
4895 specified when the compiler was configured.
4899 Generate output for a 68030. This is the default when the compiler is
4900 configured for 68030-based systems.
4904 Generate output for a 68040. This is the default when the compiler is
4905 configured for 68040-based systems.
4907 This option inhibits the use of 68881/68882 instructions that have to be
4908 emulated by software on the 68040. Use this option if your 68040 does not
4909 have code to emulate those instructions.
4913 Generate output for a 68060. This is the default when the compiler is
4914 configured for 68060-based systems.
4916 This option inhibits the use of 68020 and 68881/68882 instructions that
4917 have to be emulated by software on the 68060. Use this option if your 68060
4918 does not have code to emulate those instructions.
4922 Generate output for a CPU32. This is the default
4923 when the compiler is configured for CPU32-based systems.
4925 Use this option for microcontrollers with a
4926 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
4927 68336, 68340, 68341, 68349 and 68360.
4931 Generate output for a 520X ``coldfire'' family cpu. This is the default
4932 when the compiler is configured for 520X-based systems.
4934 Use this option for microcontroller with a 5200 core, including
4935 the MCF5202, MCF5203, MCF5204 and MCF5202.
4940 Generate output for a 68040, without using any of the new instructions.
4941 This results in code which can run relatively efficiently on either a
4942 68020/68881 or a 68030 or a 68040. The generated code does use the
4943 68881 instructions that are emulated on the 68040.
4947 Generate output for a 68060, without using any of the new instructions.
4948 This results in code which can run relatively efficiently on either a
4949 68020/68881 or a 68030 or a 68040. The generated code does use the
4950 68881 instructions that are emulated on the 68060.
4954 Generate output containing Sun FPA instructions for floating point.
4957 @opindex msoft-float
4958 Generate output containing library calls for floating point.
4959 @strong{Warning:} the requisite libraries are not available for all m68k
4960 targets. Normally the facilities of the machine's usual C compiler are
4961 used, but this can't be done directly in cross-compilation. You must
4962 make your own arrangements to provide suitable library functions for
4963 cross-compilation. The embedded targets @samp{m68k-*-aout} and
4964 @samp{m68k-*-coff} do provide software floating point support.
4968 Consider type @code{int} to be 16 bits wide, like @code{short int}.
4971 @opindex mnobitfield
4972 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
4973 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
4977 Do use the bit-field instructions. The @option{-m68020} option implies
4978 @option{-mbitfield}. This is the default if you use a configuration
4979 designed for a 68020.
4983 Use a different function-calling convention, in which functions
4984 that take a fixed number of arguments return with the @code{rtd}
4985 instruction, which pops their arguments while returning. This
4986 saves one instruction in the caller since there is no need to pop
4987 the arguments there.
4989 This calling convention is incompatible with the one normally
4990 used on Unix, so you cannot use it if you need to call libraries
4991 compiled with the Unix compiler.
4993 Also, you must provide function prototypes for all functions that
4994 take variable numbers of arguments (including @code{printf});
4995 otherwise incorrect code will be generated for calls to those
4998 In addition, seriously incorrect code will result if you call a
4999 function with too many arguments. (Normally, extra arguments are
5000 harmlessly ignored.)
5002 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5003 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5006 @itemx -mno-align-int
5008 @opindex mno-align-int
5009 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5010 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5011 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5012 Aligning variables on 32-bit boundaries produces code that runs somewhat
5013 faster on processors with 32-bit busses at the expense of more memory.
5015 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5016 align structures containing the above types differently than
5017 most published application binary interface specifications for the m68k.
5021 Use the pc-relative addressing mode of the 68000 directly, instead of
5022 using a global offset table. At present, this option implies @option{-fpic},
5023 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5024 not presently supported with @option{-mpcrel}, though this could be supported for
5025 68020 and higher processors.
5027 @item -mno-strict-align
5028 @itemx -mstrict-align
5029 @opindex mno-strict-align
5030 @opindex mstrict-align
5031 Do not (do) assume that unaligned memory references will be handled by
5036 @node M68hc1x Options
5037 @subsection M68hc1x Options
5038 @cindex M68hc1x options
5040 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5041 microcontrollers. The default values for these options depends on
5042 which style of microcontroller was selected when the compiler was configured;
5043 the defaults for the most common choices are given below.
5050 Generate output for a 68HC11. This is the default
5051 when the compiler is configured for 68HC11-based systems.
5057 Generate output for a 68HC12. This is the default
5058 when the compiler is configured for 68HC12-based systems.
5061 @opindex mauto-incdec
5062 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5067 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5069 @item -msoft-reg-count=@var{count}
5070 @opindex msoft-reg-count
5071 Specify the number of pseudo-soft registers which are used for the
5072 code generation. The maximum number is 32. Using more pseudo-soft
5073 register may or may not result in better code depending on the program.
5074 The default is 4 for 68HC11 and 2 for 68HC12.
5079 @subsection VAX Options
5082 These @samp{-m} options are defined for the VAX:
5087 Do not output certain jump instructions (@code{aobleq} and so on)
5088 that the Unix assembler for the VAX cannot handle across long
5093 Do output those jump instructions, on the assumption that you
5094 will assemble with the GNU assembler.
5098 Output code for g-format floating point numbers instead of d-format.
5102 @subsection SPARC Options
5103 @cindex SPARC options
5105 These @samp{-m} switches are supported on the SPARC:
5110 @opindex mno-app-regs
5112 Specify @option{-mapp-regs} to generate output using the global registers
5113 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5116 To be fully SVR4 ABI compliant at the cost of some performance loss,
5117 specify @option{-mno-app-regs}. You should compile libraries and system
5118 software with this option.
5123 @opindex mhard-float
5124 Generate output containing floating point instructions. This is the
5130 @opindex msoft-float
5131 Generate output containing library calls for floating point.
5132 @strong{Warning:} the requisite libraries are not available for all SPARC
5133 targets. Normally the facilities of the machine's usual C compiler are
5134 used, but this cannot be done directly in cross-compilation. You must make
5135 your own arrangements to provide suitable library functions for
5136 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5137 @samp{sparclite-*-*} do provide software floating point support.
5139 @option{-msoft-float} changes the calling convention in the output file;
5140 therefore, it is only useful if you compile @emph{all} of a program with
5141 this option. In particular, you need to compile @file{libgcc.a}, the
5142 library that comes with GCC, with @option{-msoft-float} in order for
5145 @item -mhard-quad-float
5146 @opindex mhard-quad-float
5147 Generate output containing quad-word (long double) floating point
5150 @item -msoft-quad-float
5151 @opindex msoft-quad-float
5152 Generate output containing library calls for quad-word (long double)
5153 floating point instructions. The functions called are those specified
5154 in the SPARC ABI@. This is the default.
5156 As of this writing, there are no sparc implementations that have hardware
5157 support for the quad-word floating point instructions. They all invoke
5158 a trap handler for one of these instructions, and then the trap handler
5159 emulates the effect of the instruction. Because of the trap handler overhead,
5160 this is much slower than calling the ABI library routines. Thus the
5161 @option{-msoft-quad-float} option is the default.
5167 With @option{-mflat}, the compiler does not generate save/restore instructions
5168 and will use a ``flat'' or single register window calling convention.
5169 This model uses %i7 as the frame pointer and is compatible with the normal
5170 register window model. Code from either may be intermixed.
5171 The local registers and the input registers (0--5) are still treated as
5172 ``call saved'' registers and will be saved on the stack as necessary.
5174 With @option{-mno-flat} (the default), the compiler emits save/restore
5175 instructions (except for leaf functions) and is the normal mode of operation.
5177 @item -mno-unaligned-doubles
5178 @itemx -munaligned-doubles
5179 @opindex mno-unaligned-doubles
5180 @opindex munaligned-doubles
5181 Assume that doubles have 8 byte alignment. This is the default.
5183 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5184 alignment only if they are contained in another type, or if they have an
5185 absolute address. Otherwise, it assumes they have 4 byte alignment.
5186 Specifying this option avoids some rare compatibility problems with code
5187 generated by other compilers. It is not the default because it results
5188 in a performance loss, especially for floating point code.
5190 @item -mno-faster-structs
5191 @itemx -mfaster-structs
5192 @opindex mno-faster-structs
5193 @opindex mfaster-structs
5194 With @option{-mfaster-structs}, the compiler assumes that structures
5195 should have 8 byte alignment. This enables the use of pairs of
5196 @code{ldd} and @code{std} instructions for copies in structure
5197 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5198 However, the use of this changed alignment directly violates the Sparc
5199 ABI@. Thus, it's intended only for use on targets where the developer
5200 acknowledges that their resulting code will not be directly in line with
5201 the rules of the ABI@.
5207 These two options select variations on the SPARC architecture.
5209 By default (unless specifically configured for the Fujitsu SPARClite),
5210 GCC generates code for the v7 variant of the SPARC architecture.
5212 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5213 code is that the compiler emits the integer multiply and integer
5214 divide instructions which exist in SPARC v8 but not in SPARC v7.
5216 @option{-msparclite} will give you SPARClite code. This adds the integer
5217 multiply, integer divide step and scan (@code{ffs}) instructions which
5218 exist in SPARClite but not in SPARC v7.
5220 These options are deprecated and will be deleted in a future GCC release.
5221 They have been replaced with @option{-mcpu=xxx}.
5226 @opindex msupersparc
5227 These two options select the processor for which the code is optimized.
5229 With @option{-mcypress} (the default), the compiler optimizes code for the
5230 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
5231 This is also appropriate for the older SparcStation 1, 2, IPX etc.
5233 With @option{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
5234 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
5235 of the full SPARC v8 instruction set.
5237 These options are deprecated and will be deleted in a future GCC release.
5238 They have been replaced with @option{-mcpu=xxx}.
5240 @item -mcpu=@var{cpu_type}
5242 Set the instruction set, register set, and instruction scheduling parameters
5243 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5244 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5245 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5246 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
5248 Default instruction scheduling parameters are used for values that select
5249 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5250 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5252 Here is a list of each supported architecture and their supported
5257 v8: supersparc, hypersparc
5258 sparclite: f930, f934, sparclite86x
5263 @item -mtune=@var{cpu_type}
5265 Set the instruction scheduling parameters for machine type
5266 @var{cpu_type}, but do not set the instruction set or register set that the
5267 option @option{-mcpu=@var{cpu_type}} would.
5269 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5270 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5271 that select a particular cpu implementation. Those are @samp{cypress},
5272 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5273 @samp{sparclite86x}, @samp{tsc701}, and @samp{ultrasparc}.
5277 These @samp{-m} switches are supported in addition to the above
5278 on the SPARCLET processor.
5281 @item -mlittle-endian
5282 @opindex mlittle-endian
5283 Generate code for a processor running in little-endian mode.
5287 Treat register @code{%g0} as a normal register.
5288 GCC will continue to clobber it as necessary but will not assume
5289 it always reads as 0.
5291 @item -mbroken-saverestore
5292 @opindex mbroken-saverestore
5293 Generate code that does not use non-trivial forms of the @code{save} and
5294 @code{restore} instructions. Early versions of the SPARCLET processor do
5295 not correctly handle @code{save} and @code{restore} instructions used with
5296 arguments. They correctly handle them used without arguments. A @code{save}
5297 instruction used without arguments increments the current window pointer
5298 but does not allocate a new stack frame. It is assumed that the window
5299 overflow trap handler will properly handle this case as will interrupt
5303 These @samp{-m} switches are supported in addition to the above
5304 on SPARC V9 processors in 64-bit environments.
5307 @item -mlittle-endian
5308 @opindex mlittle-endian
5309 Generate code for a processor running in little-endian mode.
5315 Generate code for a 32-bit or 64-bit environment.
5316 The 32-bit environment sets int, long and pointer to 32 bits.
5317 The 64-bit environment sets int to 32 bits and long and pointer
5320 @item -mcmodel=medlow
5321 @opindex mcmodel=medlow
5322 Generate code for the Medium/Low code model: the program must be linked
5323 in the low 32 bits of the address space. Pointers are 64 bits.
5324 Programs can be statically or dynamically linked.
5326 @item -mcmodel=medmid
5327 @opindex mcmodel=medmid
5328 Generate code for the Medium/Middle code model: the program must be linked
5329 in the low 44 bits of the address space, the text segment must be less than
5330 2G bytes, and data segment must be within 2G of the text segment.
5331 Pointers are 64 bits.
5333 @item -mcmodel=medany
5334 @opindex mcmodel=medany
5335 Generate code for the Medium/Anywhere code model: the program may be linked
5336 anywhere in the address space, the text segment must be less than
5337 2G bytes, and data segment must be within 2G of the text segment.
5338 Pointers are 64 bits.
5340 @item -mcmodel=embmedany
5341 @opindex mcmodel=embmedany
5342 Generate code for the Medium/Anywhere code model for embedded systems:
5343 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5344 (determined at link time). Register %g4 points to the base of the
5345 data segment. Pointers are still 64 bits.
5346 Programs are statically linked, PIC is not supported.
5349 @itemx -mno-stack-bias
5350 @opindex mstack-bias
5351 @opindex mno-stack-bias
5352 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5353 frame pointer if present, are offset by @minus{}2047 which must be added back
5354 when making stack frame references.
5355 Otherwise, assume no such offset is present.
5358 @node Convex Options
5359 @subsection Convex Options
5360 @cindex Convex options
5362 These @samp{-m} options are defined for Convex:
5367 Generate output for C1. The code will run on any Convex machine.
5368 The preprocessor symbol @code{__convex__c1__} is defined.
5372 Generate output for C2. Uses instructions not available on C1.
5373 Scheduling and other optimizations are chosen for max performance on C2.
5374 The preprocessor symbol @code{__convex_c2__} is defined.
5378 Generate output for C32xx. Uses instructions not available on C1.
5379 Scheduling and other optimizations are chosen for max performance on C32.
5380 The preprocessor symbol @code{__convex_c32__} is defined.
5384 Generate output for C34xx. Uses instructions not available on C1.
5385 Scheduling and other optimizations are chosen for max performance on C34.
5386 The preprocessor symbol @code{__convex_c34__} is defined.
5390 Generate output for C38xx. Uses instructions not available on C1.
5391 Scheduling and other optimizations are chosen for max performance on C38.
5392 The preprocessor symbol @code{__convex_c38__} is defined.
5396 Generate code which puts an argument count in the word preceding each
5397 argument list. This is compatible with regular CC, and a few programs
5398 may need the argument count word. GDB and other source-level debuggers
5399 do not need it; this info is in the symbol table.
5402 @opindex mnoargcount
5403 Omit the argument count word. This is the default.
5405 @item -mvolatile-cache
5406 @opindex mvolatile-cache
5407 Allow volatile references to be cached. This is the default.
5409 @item -mvolatile-nocache
5410 @opindex mvolatile-nocache
5411 Volatile references bypass the data cache, going all the way to memory.
5412 This is only needed for multi-processor code that does not use standard
5413 synchronization instructions. Making non-volatile references to volatile
5414 locations will not necessarily work.
5418 Type long is 32 bits, the same as type int. This is the default.
5422 Type long is 64 bits, the same as type long long. This option is useless,
5423 because no library support exists for it.
5426 @node AMD29K Options
5427 @subsection AMD29K Options
5428 @cindex AMD29K options
5430 These @samp{-m} options are defined for the AMD Am29000:
5435 @cindex DW bit (29k)
5436 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
5437 halfword operations are directly supported by the hardware. This is the
5442 Generate code that assumes the @code{DW} bit is not set.
5446 @cindex byte writes (29k)
5447 Generate code that assumes the system supports byte and halfword write
5448 operations. This is the default.
5452 Generate code that assumes the systems does not support byte and
5453 halfword write operations. @option{-mnbw} implies @option{-mndw}.
5457 @cindex memory model (29k)
5458 Use a small memory model that assumes that all function addresses are
5459 either within a single 256 KB segment or at an absolute address of less
5460 than 256k. This allows the @code{call} instruction to be used instead
5461 of a @code{const}, @code{consth}, @code{calli} sequence.
5465 Use the normal memory model: Generate @code{call} instructions only when
5466 calling functions in the same file and @code{calli} instructions
5467 otherwise. This works if each file occupies less than 256 KB but allows
5468 the entire executable to be larger than 256 KB@. This is the default.
5472 Always use @code{calli} instructions. Specify this option if you expect
5473 a single file to compile into more than 256 KB of code.
5477 @cindex processor selection (29k)
5478 Generate code for the Am29050.
5482 Generate code for the Am29000. This is the default.
5484 @item -mkernel-registers
5485 @opindex mkernel-registers
5486 @cindex kernel and user registers (29k)
5487 Generate references to registers @code{gr64-gr95} instead of to
5488 registers @code{gr96-gr127}. This option can be used when compiling
5489 kernel code that wants a set of global registers disjoint from that used
5492 Note that when this option is used, register names in @samp{-f} flags
5493 must use the normal, user-mode, names.
5495 @item -muser-registers
5496 @opindex muser-registers
5497 Use the normal set of global registers, @code{gr96-gr127}. This is the
5501 @itemx -mno-stack-check
5502 @opindex mstack-check
5503 @opindex mno-stack-check
5504 @cindex stack checks (29k)
5505 Insert (or do not insert) a call to @code{__msp_check} after each stack
5506 adjustment. This is often used for kernel code.
5509 @itemx -mno-storem-bug
5510 @opindex mstorem-bug
5511 @opindex mno-storem-bug
5512 @cindex storem bug (29k)
5513 @option{-mstorem-bug} handles 29k processors which cannot handle the
5514 separation of a mtsrim insn and a storem instruction (most 29000 chips
5515 to date, but not the 29050).
5517 @item -mno-reuse-arg-regs
5518 @itemx -mreuse-arg-regs
5519 @opindex mno-reuse-arg-regs
5520 @opindex mreuse-arg-regs
5521 @option{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
5522 registers for copying out arguments. This helps detect calling a function
5523 with fewer arguments than it was declared with.
5525 @item -mno-impure-text
5526 @itemx -mimpure-text
5527 @opindex mno-impure-text
5528 @opindex mimpure-text
5529 @option{-mimpure-text}, used in addition to @option{-shared}, tells the compiler to
5530 not pass @option{-assert pure-text} to the linker when linking a shared object.
5533 @opindex msoft-float
5534 Generate output containing library calls for floating point.
5535 @strong{Warning:} the requisite libraries are not part of GCC@.
5536 Normally the facilities of the machine's usual C compiler are used, but
5537 this can't be done directly in cross-compilation. You must make your
5538 own arrangements to provide suitable library functions for
5543 Do not generate multm or multmu instructions. This is useful for some embedded
5544 systems which do not have trap handlers for these instructions.
5548 @subsection ARM Options
5551 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5556 @opindex mapcs-frame
5557 Generate a stack frame that is compliant with the ARM Procedure Call
5558 Standard for all functions, even if this is not strictly necessary for
5559 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5560 with this option will cause the stack frames not to be generated for
5561 leaf functions. The default is @option{-mno-apcs-frame}.
5565 This is a synonym for @option{-mapcs-frame}.
5569 Generate code for a processor running with a 26-bit program counter,
5570 and conforming to the function calling standards for the APCS 26-bit
5571 option. This option replaces the @option{-m2} and @option{-m3} options
5572 of previous releases of the compiler.
5576 Generate code for a processor running with a 32-bit program counter,
5577 and conforming to the function calling standards for the APCS 32-bit
5578 option. This option replaces the @option{-m6} option of previous releases
5582 @c not currently implemented
5583 @item -mapcs-stack-check
5584 @opindex mapcs-stack-check
5585 Generate code to check the amount of stack space available upon entry to
5586 every function (that actually uses some stack space). If there is
5587 insufficient space available then either the function
5588 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5589 called, depending upon the amount of stack space required. The run time
5590 system is required to provide these functions. The default is
5591 @option{-mno-apcs-stack-check}, since this produces smaller code.
5593 @c not currently implemented
5595 @opindex mapcs-float
5596 Pass floating point arguments using the float point registers. This is
5597 one of the variants of the APCS@. This option is recommended if the
5598 target hardware has a floating point unit or if a lot of floating point
5599 arithmetic is going to be performed by the code. The default is
5600 @option{-mno-apcs-float}, since integer only code is slightly increased in
5601 size if @option{-mapcs-float} is used.
5603 @c not currently implemented
5604 @item -mapcs-reentrant
5605 @opindex mapcs-reentrant
5606 Generate reentrant, position independent code. The default is
5607 @option{-mno-apcs-reentrant}.
5610 @item -mthumb-interwork
5611 @opindex mthumb-interwork
5612 Generate code which supports calling between the ARM and Thumb
5613 instruction sets. Without this option the two instruction sets cannot
5614 be reliably used inside one program. The default is
5615 @option{-mno-thumb-interwork}, since slightly larger code is generated
5616 when @option{-mthumb-interwork} is specified.
5618 @item -mno-sched-prolog
5619 @opindex mno-sched-prolog
5620 Prevent the reordering of instructions in the function prolog, or the
5621 merging of those instruction with the instructions in the function's
5622 body. This means that all functions will start with a recognizable set
5623 of instructions (or in fact one of a choice from a small set of
5624 different function prologues), and this information can be used to
5625 locate the start if functions inside an executable piece of code. The
5626 default is @option{-msched-prolog}.
5629 @opindex mhard-float
5630 Generate output containing floating point instructions. This is the
5634 @opindex msoft-float
5635 Generate output containing library calls for floating point.
5636 @strong{Warning:} the requisite libraries are not available for all ARM
5637 targets. Normally the facilities of the machine's usual C compiler are
5638 used, but this cannot be done directly in cross-compilation. You must make
5639 your own arrangements to provide suitable library functions for
5642 @option{-msoft-float} changes the calling convention in the output file;
5643 therefore, it is only useful if you compile @emph{all} of a program with
5644 this option. In particular, you need to compile @file{libgcc.a}, the
5645 library that comes with GCC, with @option{-msoft-float} in order for
5648 @item -mlittle-endian
5649 @opindex mlittle-endian
5650 Generate code for a processor running in little-endian mode. This is
5651 the default for all standard configurations.
5654 @opindex mbig-endian
5655 Generate code for a processor running in big-endian mode; the default is
5656 to compile code for a little-endian processor.
5658 @item -mwords-little-endian
5659 @opindex mwords-little-endian
5660 This option only applies when generating code for big-endian processors.
5661 Generate code for a little-endian word order but a big-endian byte
5662 order. That is, a byte order of the form @samp{32107654}. Note: this
5663 option should only be used if you require compatibility with code for
5664 big-endian ARM processors generated by versions of the compiler prior to
5667 @item -malignment-traps
5668 @opindex malignment-traps
5669 Generate code that will not trap if the MMU has alignment traps enabled.
5670 On ARM architectures prior to ARMv4, there were no instructions to
5671 access half-word objects stored in memory. However, when reading from
5672 memory a feature of the ARM architecture allows a word load to be used,
5673 even if the address is unaligned, and the processor core will rotate the
5674 data as it is being loaded. This option tells the compiler that such
5675 misaligned accesses will cause a MMU trap and that it should instead
5676 synthesise the access as a series of byte accesses. The compiler can
5677 still use word accesses to load half-word data if it knows that the
5678 address is aligned to a word boundary.
5680 This option is ignored when compiling for ARM architecture 4 or later,
5681 since these processors have instructions to directly access half-word
5684 @item -mno-alignment-traps
5685 @opindex mno-alignment-traps
5686 Generate code that assumes that the MMU will not trap unaligned
5687 accesses. This produces better code when the target instruction set
5688 does not have half-word memory operations (i.e.@: implementations prior to
5691 Note that you cannot use this option to access unaligned word objects,
5692 since the processor will only fetch one 32-bit aligned object from
5695 The default setting for most targets is @option{-mno-alignment-traps}, since
5696 this produces better code when there are no half-word memory
5697 instructions available.
5699 @item -mshort-load-bytes
5700 @itemx -mno-short-load-words
5701 @opindex mshort-load-bytes
5702 @opindex mno-short-load-words
5703 These are deprecated aliases for @option{-malignment-traps}.
5705 @item -mno-short-load-bytes
5706 @itemx -mshort-load-words
5707 @opindex mno-short-load-bytes
5708 @opindex mshort-load-words
5709 This are deprecated aliases for @option{-mno-alignment-traps}.
5713 This option only applies to RISC iX@. Emulate the native BSD-mode
5714 compiler. This is the default if @option{-ansi} is not specified.
5718 This option only applies to RISC iX@. Emulate the native X/Open-mode
5721 @item -mno-symrename
5722 @opindex mno-symrename
5723 This option only applies to RISC iX@. Do not run the assembler
5724 post-processor, @samp{symrename}, after code has been assembled.
5725 Normally it is necessary to modify some of the standard symbols in
5726 preparation for linking with the RISC iX C library; this option
5727 suppresses this pass. The post-processor is never run when the
5728 compiler is built for cross-compilation.
5730 @item -mcpu=@var{name}
5732 This specifies the name of the target ARM processor. GCC uses this name
5733 to determine what kind of instructions it can emit when generating
5734 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
5735 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
5736 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
5737 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
5738 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
5739 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
5740 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
5741 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
5742 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
5743 @samp{arm1020t}, @samp{xscale}.
5745 @itemx -mtune=@var{name}
5747 This option is very similar to the @option{-mcpu=} option, except that
5748 instead of specifying the actual target processor type, and hence
5749 restricting which instructions can be used, it specifies that GCC should
5750 tune the performance of the code as if the target were of the type
5751 specified in this option, but still choosing the instructions that it
5752 will generate based on the cpu specified by a @option{-mcpu=} option.
5753 For some ARM implementations better performance can be obtained by using
5756 @item -march=@var{name}
5758 This specifies the name of the target ARM architecture. GCC uses this
5759 name to determine what kind of instructions it can emit when generating
5760 assembly code. This option can be used in conjunction with or instead
5761 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
5762 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
5763 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
5765 @item -mfpe=@var{number}
5766 @itemx -mfp=@var{number}
5769 This specifies the version of the floating point emulation available on
5770 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
5771 for @option{-mfpe=}, for compatibility with older versions of GCC@.
5773 @item -mstructure-size-boundary=@var{n}
5774 @opindex mstructure-size-boundary
5775 The size of all structures and unions will be rounded up to a multiple
5776 of the number of bits set by this option. Permissible values are 8 and
5777 32. The default value varies for different toolchains. For the COFF
5778 targeted toolchain the default value is 8. Specifying the larger number
5779 can produce faster, more efficient code, but can also increase the size
5780 of the program. The two values are potentially incompatible. Code
5781 compiled with one value cannot necessarily expect to work with code or
5782 libraries compiled with the other value, if they exchange information
5783 using structures or unions.
5785 @item -mabort-on-noreturn
5786 @opindex mabort-on-noreturn
5787 Generate a call to the function @code{abort} at the end of a
5788 @code{noreturn} function. It will be executed if the function tries to
5792 @itemx -mno-long-calls
5793 @opindex mlong-calls
5794 @opindex mno-long-calls
5795 Tells the compiler to perform function calls by first loading the
5796 address of the function into a register and then performing a subroutine
5797 call on this register. This switch is needed if the target function
5798 will lie outside of the 64 megabyte addressing range of the offset based
5799 version of subroutine call instruction.
5801 Even if this switch is enabled, not all function calls will be turned
5802 into long calls. The heuristic is that static functions, functions
5803 which have the @samp{short-call} attribute, functions that are inside
5804 the scope of a @samp{#pragma no_long_calls} directive and functions whose
5805 definitions have already been compiled within the current compilation
5806 unit, will not be turned into long calls. The exception to this rule is
5807 that weak function definitions, functions with the @samp{long-call}
5808 attribute or the @samp{section} attribute, and functions that are within
5809 the scope of a @samp{#pragma long_calls} directive, will always be
5810 turned into long calls.
5812 This feature is not enabled by default. Specifying
5813 @option{-mno-long-calls} will restore the default behavior, as will
5814 placing the function calls within the scope of a @samp{#pragma
5815 long_calls_off} directive. Note these switches have no effect on how
5816 the compiler generates code to handle function calls via function
5819 @item -mnop-fun-dllimport
5820 @opindex mnop-fun-dllimport
5821 Disable support for the @code{dllimport} attribute.
5823 @item -msingle-pic-base
5824 @opindex msingle-pic-base
5825 Treat the register used for PIC addressing as read-only, rather than
5826 loading it in the prologue for each function. The run-time system is
5827 responsible for initializing this register with an appropriate value
5828 before execution begins.
5830 @item -mpic-register=@var{reg}
5831 @opindex mpic-register
5832 Specify the register to be used for PIC addressing. The default is R10
5833 unless stack-checking is enabled, when R9 is used.
5835 @item -mpoke-function-name
5836 @opindex mpoke-function-name
5837 Write the name of each function into the text section, directly
5838 preceding the function prologue. The generated code is similar to this:
5842 .ascii "arm_poke_function_name", 0
5845 .word 0xff000000 + (t1 - t0)
5846 arm_poke_function_name
5848 stmfd sp!, @{fp, ip, lr, pc@}
5852 When performing a stack backtrace, code can inspect the value of
5853 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
5854 location @code{pc - 12} and the top 8 bits are set, then we know that
5855 there is a function name embedded immediately preceding this location
5856 and has length @code{((pc[-3]) & 0xff000000)}.
5860 Generate code for the 16-bit Thumb instruction set. The default is to
5861 use the 32-bit ARM instruction set.
5864 @opindex mtpcs-frame
5865 Generate a stack frame that is compliant with the Thumb Procedure Call
5866 Standard for all non-leaf functions. (A leaf function is one that does
5867 not call any other functions.) The default is @option{-mno-tpcs-frame}.
5869 @item -mtpcs-leaf-frame
5870 @opindex mtpcs-leaf-frame
5871 Generate a stack frame that is compliant with the Thumb Procedure Call
5872 Standard for all leaf functions. (A leaf function is one that does
5873 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
5875 @item -mcallee-super-interworking
5876 @opindex mcallee-super-interworking
5877 Gives all externally visible functions in the file being compiled an ARM
5878 instruction set header which switches to Thumb mode before executing the
5879 rest of the function. This allows these functions to be called from
5880 non-interworking code.
5882 @item -mcaller-super-interworking
5883 @opindex mcaller-super-interworking
5884 Allows calls via function pointers (including virtual functions) to
5885 execute correctly regardless of whether the target code has been
5886 compiled for interworking or not. There is a small overhead in the cost
5887 of executing a function pointer if this option is enabled.
5891 @node MN10200 Options
5892 @subsection MN10200 Options
5893 @cindex MN10200 options
5894 These @option{-m} options are defined for Matsushita MN10200 architectures:
5899 Indicate to the linker that it should perform a relaxation optimization pass
5900 to shorten branches, calls and absolute memory addresses. This option only
5901 has an effect when used on the command line for the final link step.
5903 This option makes symbolic debugging impossible.
5906 @node MN10300 Options
5907 @subsection MN10300 Options
5908 @cindex MN10300 options
5909 These @option{-m} options are defined for Matsushita MN10300 architectures:
5914 Generate code to avoid bugs in the multiply instructions for the MN10300
5915 processors. This is the default.
5918 @opindex mno-mult-bug
5919 Do not generate code to avoid bugs in the multiply instructions for the
5924 Generate code which uses features specific to the AM33 processor.
5928 Do not generate code which uses features specific to the AM33 processor. This
5933 Do not link in the C run-time initialization object file.
5937 Indicate to the linker that it should perform a relaxation optimization pass
5938 to shorten branches, calls and absolute memory addresses. This option only
5939 has an effect when used on the command line for the final link step.
5941 This option makes symbolic debugging impossible.
5945 @node M32R/D Options
5946 @subsection M32R/D Options
5947 @cindex M32R/D options
5949 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
5954 Generate code for the M32R/X@.
5958 Generate code for the M32R@. This is the default.
5960 @item -mcode-model=small
5961 @opindex mcode-model=small
5962 Assume all objects live in the lower 16MB of memory (so that their addresses
5963 can be loaded with the @code{ld24} instruction), and assume all subroutines
5964 are reachable with the @code{bl} instruction.
5965 This is the default.
5967 The addressability of a particular object can be set with the
5968 @code{model} attribute.
5970 @item -mcode-model=medium
5971 @opindex mcode-model=medium
5972 Assume objects may be anywhere in the 32-bit address space (the compiler
5973 will generate @code{seth/add3} instructions to load their addresses), and
5974 assume all subroutines are reachable with the @code{bl} instruction.
5976 @item -mcode-model=large
5977 @opindex mcode-model=large
5978 Assume objects may be anywhere in the 32-bit address space (the compiler
5979 will generate @code{seth/add3} instructions to load their addresses), and
5980 assume subroutines may not be reachable with the @code{bl} instruction
5981 (the compiler will generate the much slower @code{seth/add3/jl}
5982 instruction sequence).
5985 @opindex msdata=none
5986 Disable use of the small data area. Variables will be put into
5987 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
5988 @code{section} attribute has been specified).
5989 This is the default.
5991 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
5992 Objects may be explicitly put in the small data area with the
5993 @code{section} attribute using one of these sections.
5996 @opindex msdata=sdata
5997 Put small global and static data in the small data area, but do not
5998 generate special code to reference them.
6002 Put small global and static data in the small data area, and generate
6003 special instructions to reference them.
6007 @cindex smaller data references
6008 Put global and static objects less than or equal to @var{num} bytes
6009 into the small data or bss sections instead of the normal data or bss
6010 sections. The default value of @var{num} is 8.
6011 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6012 for this option to have any effect.
6014 All modules should be compiled with the same @option{-G @var{num}} value.
6015 Compiling with different values of @var{num} may or may not work; if it
6016 doesn't the linker will give an error message---incorrect code will not be
6022 @subsection M88K Options
6023 @cindex M88k options
6025 These @samp{-m} options are defined for Motorola 88k architectures:
6030 Generate code that works well on both the m88100 and the
6035 Generate code that works best for the m88100, but that also
6040 Generate code that works best for the m88110, and may not run
6045 Obsolete option to be removed from the next revision.
6048 @item -midentify-revision
6049 @opindex midentify-revision
6050 @cindex identifying source, compiler (88k)
6051 Include an @code{ident} directive in the assembler output recording the
6052 source file name, compiler name and version, timestamp, and compilation
6055 @item -mno-underscores
6056 @opindex mno-underscores
6057 @cindex underscores, avoiding (88k)
6058 In assembler output, emit symbol names without adding an underscore
6059 character at the beginning of each name. The default is to use an
6060 underscore as prefix on each name.
6062 @item -mocs-debug-info
6063 @itemx -mno-ocs-debug-info
6064 @opindex mocs-debug-info
6065 @opindex mno-ocs-debug-info
6067 @cindex debugging, 88k OCS
6068 Include (or omit) additional debugging information (about registers used
6069 in each stack frame) as specified in the 88open Object Compatibility
6070 Standard, ``OCS''@. This extra information allows debugging of code that
6071 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
6072 Delta 88 SVr3.2 is to include this information; other 88k configurations
6073 omit this information by default.
6075 @item -mocs-frame-position
6076 @opindex mocs-frame-position
6077 @cindex register positions in frame (88k)
6078 When emitting COFF debugging information for automatic variables and
6079 parameters stored on the stack, use the offset from the canonical frame
6080 address, which is the stack pointer (register 31) on entry to the
6081 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
6082 @option{-mocs-frame-position}; other 88k configurations have the default
6083 @option{-mno-ocs-frame-position}.
6085 @item -mno-ocs-frame-position
6086 @opindex mno-ocs-frame-position
6087 @cindex register positions in frame (88k)
6088 When emitting COFF debugging information for automatic variables and
6089 parameters stored on the stack, use the offset from the frame pointer
6090 register (register 30). When this option is in effect, the frame
6091 pointer is not eliminated when debugging information is selected by the
6094 @item -moptimize-arg-area
6095 @opindex moptimize-arg-area
6096 @cindex arguments in frame (88k)
6097 Save space by reorganizing the stack frame. This option generates code
6098 that does not agree with the 88open specifications, but uses less
6101 @itemx -mno-optimize-arg-area
6102 @opindex mno-optimize-arg-area
6103 Do not reorganize the stack frame to save space. This is the default.
6104 The generated conforms to the specification, but uses more memory.
6106 @item -mshort-data-@var{num}
6107 @opindex mshort-data
6108 @cindex smaller data references (88k)
6109 @cindex r0-relative references (88k)
6110 Generate smaller data references by making them relative to @code{r0},
6111 which allows loading a value using a single instruction (rather than the
6112 usual two). You control which data references are affected by
6113 specifying @var{num} with this option. For example, if you specify
6114 @option{-mshort-data-512}, then the data references affected are those
6115 involving displacements of less than 512 bytes.
6116 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6119 @item -mserialize-volatile
6120 @opindex mserialize-volatile
6121 @itemx -mno-serialize-volatile
6122 @opindex mno-serialize-volatile
6123 @cindex sequential consistency on 88k
6124 Do, or don't, generate code to guarantee sequential consistency
6125 of volatile memory references. By default, consistency is
6128 The order of memory references made by the MC88110 processor does
6129 not always match the order of the instructions requesting those
6130 references. In particular, a load instruction may execute before
6131 a preceding store instruction. Such reordering violates
6132 sequential consistency of volatile memory references, when there
6133 are multiple processors. When consistency must be guaranteed,
6134 GCC generates special instructions, as needed, to force
6135 execution in the proper order.
6137 The MC88100 processor does not reorder memory references and so
6138 always provides sequential consistency. However, by default, GCC
6139 generates the special instructions to guarantee consistency
6140 even when you use @option{-m88100}, so that the code may be run on an
6141 MC88110 processor. If you intend to run your code only on the
6142 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6144 The extra code generated to guarantee consistency may affect the
6145 performance of your application. If you know that you can safely
6146 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6152 @cindex assembler syntax, 88k
6154 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6155 related to System V release 4 (SVr4). This controls the following:
6159 Which variant of the assembler syntax to emit.
6161 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6162 that is used on System V release 4.
6164 @option{-msvr4} makes GCC issue additional declaration directives used in
6168 @option{-msvr4} is the default for the m88k-motorola-sysv4 and
6169 m88k-dg-dgux m88k configurations. @option{-msvr3} is the default for all
6170 other m88k configurations.
6172 @item -mversion-03.00
6173 @opindex mversion-03.00
6174 This option is obsolete, and is ignored.
6175 @c ??? which asm syntax better for GAS? option there too?
6177 @item -mno-check-zero-division
6178 @itemx -mcheck-zero-division
6179 @opindex mno-check-zero-division
6180 @opindex mcheck-zero-division
6181 @cindex zero division on 88k
6182 Do, or don't, generate code to guarantee that integer division by
6183 zero will be detected. By default, detection is guaranteed.
6185 Some models of the MC88100 processor fail to trap upon integer
6186 division by zero under certain conditions. By default, when
6187 compiling code that might be run on such a processor, GCC
6188 generates code that explicitly checks for zero-valued divisors
6189 and traps with exception number 503 when one is detected. Use of
6190 @option{-mno-check-zero-division} suppresses such checking for code
6191 generated to run on an MC88100 processor.
6193 GCC assumes that the MC88110 processor correctly detects all instances
6194 of integer division by zero. When @option{-m88110} is specified, no
6195 explicit checks for zero-valued divisors are generated, and both
6196 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6199 @item -muse-div-instruction
6200 @opindex muse-div-instruction
6201 @cindex divide instruction, 88k
6202 Use the div instruction for signed integer division on the
6203 MC88100 processor. By default, the div instruction is not used.
6205 On the MC88100 processor the signed integer division instruction
6206 div) traps to the operating system on a negative operand. The
6207 operating system transparently completes the operation, but at a
6208 large cost in execution time. By default, when compiling code
6209 that might be run on an MC88100 processor, GCC emulates signed
6210 integer division using the unsigned integer division instruction
6211 divu), thereby avoiding the large penalty of a trap to the
6212 operating system. Such emulation has its own, smaller, execution
6213 cost in both time and space. To the extent that your code's
6214 important signed integer division operations are performed on two
6215 nonnegative operands, it may be desirable to use the div
6216 instruction directly.
6218 On the MC88110 processor the div instruction (also known as the
6219 divs instruction) processes negative operands without trapping to
6220 the operating system. When @option{-m88110} is specified,
6221 @option{-muse-div-instruction} is ignored, and the div instruction is used
6222 for signed integer division.
6224 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6225 particular, the behavior of such a division with and without
6226 @option{-muse-div-instruction} may differ.
6228 @item -mtrap-large-shift
6229 @itemx -mhandle-large-shift
6230 @opindex mtrap-large-shift
6231 @opindex mhandle-large-shift
6232 @cindex bit shift overflow (88k)
6233 @cindex large bit shifts (88k)
6234 Include code to detect bit-shifts of more than 31 bits; respectively,
6235 trap such shifts or emit code to handle them properly. By default GCC
6236 makes no special provision for large bit shifts.
6238 @item -mwarn-passed-structs
6239 @opindex mwarn-passed-structs
6240 @cindex structure passing (88k)
6241 Warn when a function passes a struct as an argument or result.
6242 Structure-passing conventions have changed during the evolution of the C
6243 language, and are often the source of portability problems. By default,
6244 GCC issues no such warning.
6247 @c break page here to avoid unsightly interparagraph stretch.
6251 @node RS/6000 and PowerPC Options
6252 @subsection IBM RS/6000 and PowerPC Options
6253 @cindex RS/6000 and PowerPC Options
6254 @cindex IBM RS/6000 and PowerPC Options
6256 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6264 @itemx -mpowerpc-gpopt
6265 @itemx -mno-powerpc-gpopt
6266 @itemx -mpowerpc-gfxopt
6267 @itemx -mno-powerpc-gfxopt
6269 @itemx -mno-powerpc64
6275 @opindex mno-powerpc
6276 @opindex mpowerpc-gpopt
6277 @opindex mno-powerpc-gpopt
6278 @opindex mpowerpc-gfxopt
6279 @opindex mno-powerpc-gfxopt
6281 @opindex mno-powerpc64
6282 GCC supports two related instruction set architectures for the
6283 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6284 instructions supported by the @samp{rios} chip set used in the original
6285 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6286 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6287 the IBM 4xx microprocessors.
6289 Neither architecture is a subset of the other. However there is a
6290 large common subset of instructions supported by both. An MQ
6291 register is included in processors supporting the POWER architecture.
6293 You use these options to specify which instructions are available on the
6294 processor you are using. The default value of these options is
6295 determined when configuring GCC@. Specifying the
6296 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6297 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6298 rather than the options listed above.
6300 The @option{-mpower} option allows GCC to generate instructions that
6301 are found only in the POWER architecture and to use the MQ register.
6302 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6303 to generate instructions that are present in the POWER2 architecture but
6304 not the original POWER architecture.
6306 The @option{-mpowerpc} option allows GCC to generate instructions that
6307 are found only in the 32-bit subset of the PowerPC architecture.
6308 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6309 GCC to use the optional PowerPC architecture instructions in the
6310 General Purpose group, including floating-point square root. Specifying
6311 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6312 use the optional PowerPC architecture instructions in the Graphics
6313 group, including floating-point select.
6315 The @option{-mpowerpc64} option allows GCC to generate the additional
6316 64-bit instructions that are found in the full PowerPC64 architecture
6317 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6318 @option{-mno-powerpc64}.
6320 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6321 will use only the instructions in the common subset of both
6322 architectures plus some special AIX common-mode calls, and will not use
6323 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6324 permits GCC to use any instruction from either architecture and to
6325 allow use of the MQ register; specify this for the Motorola MPC601.
6327 @item -mnew-mnemonics
6328 @itemx -mold-mnemonics
6329 @opindex mnew-mnemonics
6330 @opindex mold-mnemonics
6331 Select which mnemonics to use in the generated assembler code. With
6332 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6333 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6334 assembler mnemonics defined for the POWER architecture. Instructions
6335 defined in only one architecture have only one mnemonic; GCC uses that
6336 mnemonic irrespective of which of these options is specified.
6338 GCC defaults to the mnemonics appropriate for the architecture in
6339 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6340 value of these option. Unless you are building a cross-compiler, you
6341 should normally not specify either @option{-mnew-mnemonics} or
6342 @option{-mold-mnemonics}, but should instead accept the default.
6344 @item -mcpu=@var{cpu_type}
6346 Set architecture type, register usage, choice of mnemonics, and
6347 instruction scheduling parameters for machine type @var{cpu_type}.
6348 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6349 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6350 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6351 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6352 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6353 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6355 @option{-mcpu=common} selects a completely generic processor. Code
6356 generated under this option will run on any POWER or PowerPC processor.
6357 GCC will use only the instructions in the common subset of both
6358 architectures, and will not use the MQ register. GCC assumes a generic
6359 processor model for scheduling purposes.
6361 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6362 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6363 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6364 types, with an appropriate, generic processor model assumed for
6365 scheduling purposes.
6367 The other options specify a specific processor. Code generated under
6368 those options will run best on that processor, and may not run at all on
6371 The @option{-mcpu} options automatically enable or disable other
6372 @option{-m} options as follows:
6376 @option{-mno-power}, @option{-mno-powerc}
6383 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6398 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6401 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6406 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6409 @item -mtune=@var{cpu_type}
6411 Set the instruction scheduling parameters for machine type
6412 @var{cpu_type}, but do not set the architecture type, register usage, or
6413 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6414 values for @var{cpu_type} are used for @option{-mtune} as for
6415 @option{-mcpu}. If both are specified, the code generated will use the
6416 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6417 scheduling parameters set by @option{-mtune}.
6422 @opindex mno-altivec
6423 These switches enable or disable the use of built-in functions that
6424 allow access to the AltiVec instruction set. You may also need to set
6425 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6429 @itemx -mno-fp-in-toc
6430 @itemx -mno-sum-in-toc
6431 @itemx -mminimal-toc
6433 @opindex mno-fp-in-toc
6434 @opindex mno-sum-in-toc
6435 @opindex mminimal-toc
6436 Modify generation of the TOC (Table Of Contents), which is created for
6437 every executable file. The @option{-mfull-toc} option is selected by
6438 default. In that case, GCC will allocate at least one TOC entry for
6439 each unique non-automatic variable reference in your program. GCC
6440 will also place floating-point constants in the TOC@. However, only
6441 16,384 entries are available in the TOC@.
6443 If you receive a linker error message that saying you have overflowed
6444 the available TOC space, you can reduce the amount of TOC space used
6445 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6446 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6447 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6448 generate code to calculate the sum of an address and a constant at
6449 run-time instead of putting that sum into the TOC@. You may specify one
6450 or both of these options. Each causes GCC to produce very slightly
6451 slower and larger code at the expense of conserving TOC space.
6453 If you still run out of space in the TOC even when you specify both of
6454 these options, specify @option{-mminimal-toc} instead. This option causes
6455 GCC to make only one TOC entry for every file. When you specify this
6456 option, GCC will produce code that is slower and larger but which
6457 uses extremely little TOC space. You may wish to use this option
6458 only on files that contain less frequently executed code.
6464 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6465 @code{long} type, and the infrastructure needed to support them.
6466 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6467 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6468 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6473 @opindex mno-xl-call
6474 On AIX, pass floating-point arguments to prototyped functions beyond the
6475 register save area (RSA) on the stack in addition to argument FPRs. The
6476 AIX calling convention was extended but not initially documented to
6477 handle an obscure K&R C case of calling a function that takes the
6478 address of its arguments with fewer arguments than declared. AIX XL
6479 compilers access floating point arguments which do not fit in the
6480 RSA from the stack when a subroutine is compiled without
6481 optimization. Because always storing floating-point arguments on the
6482 stack is inefficient and rarely needed, this option is not enabled by
6483 default and only is necessary when calling subroutines compiled by AIX
6484 XL compilers without optimization.
6488 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6489 application written to use message passing with special startup code to
6490 enable the application to run. The system must have PE installed in the
6491 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6492 must be overridden with the @option{-specs=} option to specify the
6493 appropriate directory location. The Parallel Environment does not
6494 support threads, so the @option{-mpe} option and the @option{-pthread}
6495 option are incompatible.
6499 @opindex msoft-float
6500 @opindex mhard-float
6501 Generate code that does not use (uses) the floating-point register set.
6502 Software floating point emulation is provided if you use the
6503 @option{-msoft-float} option, and pass the option to GCC when linking.
6506 @itemx -mno-multiple
6508 @opindex mno-multiple
6509 Generate code that uses (does not use) the load multiple word
6510 instructions and the store multiple word instructions. These
6511 instructions are generated by default on POWER systems, and not
6512 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6513 endian PowerPC systems, since those instructions do not work when the
6514 processor is in little endian mode. The exceptions are PPC740 and
6515 PPC750 which permit the instructions usage in little endian mode.
6521 Generate code that uses (does not use) the load string instructions
6522 and the store string word instructions to save multiple registers and
6523 do small block moves. These instructions are generated by default on
6524 POWER systems, and not generated on PowerPC systems. Do not use
6525 @option{-mstring} on little endian PowerPC systems, since those
6526 instructions do not work when the processor is in little endian mode.
6527 The exceptions are PPC740 and PPC750 which permit the instructions
6528 usage in little endian mode.
6534 Generate code that uses (does not use) the load or store instructions
6535 that update the base register to the address of the calculated memory
6536 location. These instructions are generated by default. If you use
6537 @option{-mno-update}, there is a small window between the time that the
6538 stack pointer is updated and the address of the previous frame is
6539 stored, which means code that walks the stack frame across interrupts or
6540 signals may get corrupted data.
6543 @itemx -mno-fused-madd
6544 @opindex mfused-madd
6545 @opindex mno-fused-madd
6546 Generate code that uses (does not use) the floating point multiply and
6547 accumulate instructions. These instructions are generated by default if
6548 hardware floating is used.
6550 @item -mno-bit-align
6552 @opindex mno-bit-align
6554 On System V.4 and embedded PowerPC systems do not (do) force structures
6555 and unions that contain bit-fields to be aligned to the base type of the
6558 For example, by default a structure containing nothing but 8
6559 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6560 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6561 the structure would be aligned to a 1 byte boundary and be one byte in
6564 @item -mno-strict-align
6565 @itemx -mstrict-align
6566 @opindex mno-strict-align
6567 @opindex mstrict-align
6568 On System V.4 and embedded PowerPC systems do not (do) assume that
6569 unaligned memory references will be handled by the system.
6572 @itemx -mno-relocatable
6573 @opindex mrelocatable
6574 @opindex mno-relocatable
6575 On embedded PowerPC systems generate code that allows (does not allow)
6576 the program to be relocated to a different address at runtime. If you
6577 use @option{-mrelocatable} on any module, all objects linked together must
6578 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6580 @item -mrelocatable-lib
6581 @itemx -mno-relocatable-lib
6582 @opindex mrelocatable-lib
6583 @opindex mno-relocatable-lib
6584 On embedded PowerPC systems generate code that allows (does not allow)
6585 the program to be relocated to a different address at runtime. Modules
6586 compiled with @option{-mrelocatable-lib} can be linked with either modules
6587 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6588 with modules compiled with the @option{-mrelocatable} options.
6594 On System V.4 and embedded PowerPC systems do not (do) assume that
6595 register 2 contains a pointer to a global area pointing to the addresses
6596 used in the program.
6599 @itemx -mlittle-endian
6601 @opindex mlittle-endian
6602 On System V.4 and embedded PowerPC systems compile code for the
6603 processor in little endian mode. The @option{-mlittle-endian} option is
6604 the same as @option{-mlittle}.
6609 @opindex mbig-endian
6610 On System V.4 and embedded PowerPC systems compile code for the
6611 processor in big endian mode. The @option{-mbig-endian} option is
6612 the same as @option{-mbig}.
6616 On System V.4 and embedded PowerPC systems compile code using calling
6617 conventions that adheres to the March 1995 draft of the System V
6618 Application Binary Interface, PowerPC processor supplement. This is the
6619 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6621 @item -mcall-sysv-eabi
6622 @opindex mcall-sysv-eabi
6623 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6625 @item -mcall-sysv-noeabi
6626 @opindex mcall-sysv-noeabi
6627 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6631 On System V.4 and embedded PowerPC systems compile code using calling
6632 conventions that are similar to those used on AIX@. This is the
6633 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6635 @item -mcall-solaris
6636 @opindex mcall-solaris
6637 On System V.4 and embedded PowerPC systems compile code for the Solaris
6641 @opindex mcall-linux
6642 On System V.4 and embedded PowerPC systems compile code for the
6643 Linux-based GNU system.
6647 On System V.4 and embedded PowerPC systems compile code for the
6648 Hurd-based GNU system.
6651 @opindex mcall-netbsd
6652 On System V.4 and embedded PowerPC systems compile code for the
6653 NetBSD operating system.
6655 @item -maix-struct-return
6656 @opindex maix-struct-return
6657 Return all structures in memory (as specified by the AIX ABI)@.
6659 @item -msvr4-struct-return
6660 @opindex msvr4-struct-return
6661 Return structures smaller than 8 bytes in registers (as specified by the
6665 @opindex mabi=altivec
6666 Extend the current ABI with AltiVec ABI extensions. This does not
6667 change the default ABI, instead it adds the AltiVec ABI extensions to
6670 @item -mabi=no-altivec
6671 @opindex mabi=no-altivec
6672 Disable AltiVec ABI extensions for the current ABI.
6675 @itemx -mno-prototype
6677 @opindex mno-prototype
6678 On System V.4 and embedded PowerPC systems assume that all calls to
6679 variable argument functions are properly prototyped. Otherwise, the
6680 compiler must insert an instruction before every non prototyped call to
6681 set or clear bit 6 of the condition code register (@var{CR}) to
6682 indicate whether floating point values were passed in the floating point
6683 registers in case the function takes a variable arguments. With
6684 @option{-mprototype}, only calls to prototyped variable argument functions
6685 will set or clear the bit.
6689 On embedded PowerPC systems, assume that the startup module is called
6690 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6691 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6696 On embedded PowerPC systems, assume that the startup module is called
6697 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6702 On embedded PowerPC systems, assume that the startup module is called
6703 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6707 @opindex myellowknife
6708 On embedded PowerPC systems, assume that the startup module is called
6709 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6714 On System V.4 and embedded PowerPC systems, specify that you are
6715 compiling for a VxWorks system.
6719 Specify that you are compiling for the WindISS simulation environment.
6723 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
6724 header to indicate that @samp{eabi} extended relocations are used.
6730 On System V.4 and embedded PowerPC systems do (do not) adhere to the
6731 Embedded Applications Binary Interface (eabi) which is a set of
6732 modifications to the System V.4 specifications. Selecting @option{-meabi}
6733 means that the stack is aligned to an 8 byte boundary, a function
6734 @code{__eabi} is called to from @code{main} to set up the eabi
6735 environment, and the @option{-msdata} option can use both @code{r2} and
6736 @code{r13} to point to two separate small data areas. Selecting
6737 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
6738 do not call an initialization function from @code{main}, and the
6739 @option{-msdata} option will only use @code{r13} to point to a single
6740 small data area. The @option{-meabi} option is on by default if you
6741 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
6744 @opindex msdata=eabi
6745 On System V.4 and embedded PowerPC systems, put small initialized
6746 @code{const} global and static data in the @samp{.sdata2} section, which
6747 is pointed to by register @code{r2}. Put small initialized
6748 non-@code{const} global and static data in the @samp{.sdata} section,
6749 which is pointed to by register @code{r13}. Put small uninitialized
6750 global and static data in the @samp{.sbss} section, which is adjacent to
6751 the @samp{.sdata} section. The @option{-msdata=eabi} option is
6752 incompatible with the @option{-mrelocatable} option. The
6753 @option{-msdata=eabi} option also sets the @option{-memb} option.
6756 @opindex msdata=sysv
6757 On System V.4 and embedded PowerPC systems, put small global and static
6758 data in the @samp{.sdata} section, which is pointed to by register
6759 @code{r13}. Put small uninitialized global and static data in the
6760 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
6761 The @option{-msdata=sysv} option is incompatible with the
6762 @option{-mrelocatable} option.
6764 @item -msdata=default
6766 @opindex msdata=default
6768 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
6769 compile code the same as @option{-msdata=eabi}, otherwise compile code the
6770 same as @option{-msdata=sysv}.
6773 @opindex msdata-data
6774 On System V.4 and embedded PowerPC systems, put small global and static
6775 data in the @samp{.sdata} section. Put small uninitialized global and
6776 static data in the @samp{.sbss} section. Do not use register @code{r13}
6777 to address small data however. This is the default behavior unless
6778 other @option{-msdata} options are used.
6782 @opindex msdata=none
6784 On embedded PowerPC systems, put all initialized global and static data
6785 in the @samp{.data} section, and all uninitialized data in the
6786 @samp{.bss} section.
6790 @cindex smaller data references (PowerPC)
6791 @cindex .sdata/.sdata2 references (PowerPC)
6792 On embedded PowerPC systems, put global and static items less than or
6793 equal to @var{num} bytes into the small data or bss sections instead of
6794 the normal data or bss section. By default, @var{num} is 8. The
6795 @option{-G @var{num}} switch is also passed to the linker.
6796 All modules should be compiled with the same @option{-G @var{num}} value.
6799 @itemx -mno-regnames
6801 @opindex mno-regnames
6802 On System V.4 and embedded PowerPC systems do (do not) emit register
6803 names in the assembly language output using symbolic forms.
6806 @itemx -mno-longcall
6808 @opindex mno-longcall
6809 Default to making all function calls via pointers, so that functions
6810 which reside further than 64 megabytes (67,108,864 bytes) from the
6811 current location can be called. This setting can be overridden by the
6812 @code{shortcall} function attribute, or by @code{#pragma longcall(0)}.
6814 Some linkers are capable of detecting out-of-range calls and generating
6815 glue code on the fly. On these systems, long calls are unnecessary and
6816 generate slower code. As of this writing, the AIX linker can do this,
6817 as can the GNU linker for PowerPC/64. It is planned to add this feature
6818 to the GNU linker for 32-bit PowerPC systems as well.
6820 In the future, we may cause GCC to ignore all longcall specifications
6821 when the linker is known to generate glue.
6825 Adds support for multithreading with the @dfn{pthreads} library.
6826 This option sets flags for both the preprocessor and linker.
6831 @subsection IBM RT Options
6833 @cindex IBM RT options
6835 These @samp{-m} options are defined for the IBM RT PC:
6839 @opindex min-line-mul
6840 Use an in-line code sequence for integer multiplies. This is the
6843 @item -mcall-lib-mul
6844 @opindex mcall-lib-mul
6845 Call @code{lmul$$} for integer multiples.
6847 @item -mfull-fp-blocks
6848 @opindex mfull-fp-blocks
6849 Generate full-size floating point data blocks, including the minimum
6850 amount of scratch space recommended by IBM@. This is the default.
6852 @item -mminimum-fp-blocks
6853 @opindex mminimum-fp-blocks
6854 Do not include extra scratch space in floating point data blocks. This
6855 results in smaller code, but slower execution, since scratch space must
6856 be allocated dynamically.
6858 @cindex @file{varargs.h} and RT PC
6859 @cindex @file{stdarg.h} and RT PC
6860 @item -mfp-arg-in-fpregs
6861 @opindex mfp-arg-in-fpregs
6862 Use a calling sequence incompatible with the IBM calling convention in
6863 which floating point arguments are passed in floating point registers.
6864 Note that @code{varargs.h} and @code{stdarg.h} will not work with
6865 floating point operands if this option is specified.
6867 @item -mfp-arg-in-gregs
6868 @opindex mfp-arg-in-gregs
6869 Use the normal calling convention for floating point arguments. This is
6872 @item -mhc-struct-return
6873 @opindex mhc-struct-return
6874 Return structures of more than one word in memory, rather than in a
6875 register. This provides compatibility with the MetaWare HighC (hc)
6876 compiler. Use the option @option{-fpcc-struct-return} for compatibility
6877 with the Portable C Compiler (pcc).
6879 @item -mnohc-struct-return
6880 @opindex mnohc-struct-return
6881 Return some structures of more than one word in registers, when
6882 convenient. This is the default. For compatibility with the
6883 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
6884 option @option{-mhc-struct-return}.
6888 @subsection MIPS Options
6889 @cindex MIPS options
6891 These @samp{-m} options are defined for the MIPS family of computers:
6895 @item -march=@var{cpu-type}
6897 Assume the defaults for the machine type @var{cpu-type} when generating
6898 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
6899 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
6900 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
6901 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
6902 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
6903 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc.
6905 @item -mtune=@var{cpu-type}
6907 Assume the defaults for the machine type @var{cpu-type} when scheduling
6908 instructions. The choices for @var{cpu-type} are @samp{r2000}, @samp{r3000},
6909 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
6910 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
6911 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
6912 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
6913 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
6914 @var{cpu-type} will schedule things appropriately for that particular
6915 chip, the compiler will not generate any code that does not meet level 1
6916 of the MIPS ISA (instruction set architecture) without a @option{-mipsX}
6917 or @option{-mabi} switch being used.
6919 @item -mcpu=@var{cpu-type}
6921 This is identical to specifying both @option{-march} and @option{-mtune}.
6925 Issue instructions from level 1 of the MIPS ISA@. This is the default.
6926 @samp{r3000} is the default @var{cpu-type} at this ISA level.
6930 Issue instructions from level 2 of the MIPS ISA (branch likely, square
6931 root instructions). @samp{r6000} is the default @var{cpu-type} at this
6936 Issue instructions from level 3 of the MIPS ISA (64-bit instructions).
6937 @samp{r4000} is the default @var{cpu-type} at this ISA level.
6941 Issue instructions from level 4 of the MIPS ISA (conditional move,
6942 prefetch, enhanced FPU instructions). @samp{r8000} is the default
6943 @var{cpu-type} at this ISA level.
6947 Assume that 32 32-bit floating point registers are available. This is
6952 Assume that 32 64-bit floating point registers are available. This is
6953 the default when the @option{-mips3} option is used.
6956 @itemx -mno-fused-madd
6957 @opindex mfused-madd
6958 @opindex mno-fused-madd
6959 Generate code that uses (does not use) the floating point multiply and
6960 accumulate instructions, when they are available. These instructions
6961 are generated by default if they are available, but this may be
6962 undesirable if the extra precision causes problems or on certain chips
6963 in the mode where denormals are rounded to zero where denormals
6964 generated by multiply and accumulate instructions cause exceptions
6969 Assume that 32 32-bit general purpose registers are available. This is
6974 Assume that 32 64-bit general purpose registers are available. This is
6975 the default when the @option{-mips3} option is used.
6979 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
6980 explanation of the default, and the width of pointers.
6984 Force long types to be 64 bits wide. See @option{-mlong32} for an
6985 explanation of the default, and the width of pointers.
6989 Force long, int, and pointer types to be 32 bits wide.
6991 If none of @option{-mlong32}, @option{-mlong64}, or @option{-mint64} are set,
6992 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
6993 For @option{-mabi=32}, and @option{-mabi=n32}, ints and longs are 32 bits
6994 wide. For @option{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
6995 For @option{-mabi=eabi} and either @option{-mips1} or @option{-mips2}, ints
6996 and longs are 32 bits wide. For @option{-mabi=eabi} and higher ISAs, ints
6997 are 32 bits, and longs are 64 bits wide. The width of pointer types is
6998 the smaller of the width of longs or the width of general purpose
6999 registers (which in turn depends on the ISA)@.
7011 Generate code for the indicated ABI@. The default instruction level is
7012 @option{-mips1} for @samp{32}, @option{-mips3} for @samp{n32}, and
7013 @option{-mips4} otherwise. Conversely, with @option{-mips1} or
7014 @option{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
7019 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7020 add normal debug information. This is the default for all
7021 platforms except for the OSF/1 reference platform, using the OSF/rose
7022 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7023 switches are used, the @file{mips-tfile} program will encapsulate the
7024 stabs within MIPS ECOFF@.
7028 Generate code for the GNU assembler. This is the default on the OSF/1
7029 reference platform, using the OSF/rose object format. Also, this is
7030 the default if the configure option @option{--with-gnu-as} is used.
7032 @item -msplit-addresses
7033 @itemx -mno-split-addresses
7034 @opindex msplit-addresses
7035 @opindex mno-split-addresses
7036 Generate code to load the high and low parts of address constants separately.
7037 This allows GCC to optimize away redundant loads of the high order
7038 bits of addresses. This optimization requires GNU as and GNU ld.
7039 This optimization is enabled by default for some embedded targets where
7040 GNU as and GNU ld are standard.
7046 The @option{-mrnames} switch says to output code using the MIPS software
7047 names for the registers, instead of the hardware names (ie, @var{a0}
7048 instead of @var{$4}). The only known assembler that supports this option
7049 is the Algorithmics assembler.
7055 The @option{-mgpopt} switch says to write all of the data declarations
7056 before the instructions in the text section, this allows the MIPS
7057 assembler to generate one word memory references instead of using two
7058 words for short global or static data items. This is on by default if
7059 optimization is selected.
7065 For each non-inline function processed, the @option{-mstats} switch
7066 causes the compiler to emit one line to the standard error file to
7067 print statistics about the program (number of registers saved, stack
7074 The @option{-mmemcpy} switch makes all block moves call the appropriate
7075 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7076 generating inline code.
7079 @itemx -mno-mips-tfile
7080 @opindex mmips-tfile
7081 @opindex mno-mips-tfile
7082 The @option{-mno-mips-tfile} switch causes the compiler not
7083 postprocess the object file with the @file{mips-tfile} program,
7084 after the MIPS assembler has generated it to add debug support. If
7085 @file{mips-tfile} is not run, then no local variables will be
7086 available to the debugger. In addition, @file{stage2} and
7087 @file{stage3} objects will have the temporary file names passed to the
7088 assembler embedded in the object file, which means the objects will
7089 not compare the same. The @option{-mno-mips-tfile} switch should only
7090 be used when there are bugs in the @file{mips-tfile} program that
7091 prevents compilation.
7094 @opindex msoft-float
7095 Generate output containing library calls for floating point.
7096 @strong{Warning:} the requisite libraries are not part of GCC@.
7097 Normally the facilities of the machine's usual C compiler are used, but
7098 this can't be done directly in cross-compilation. You must make your
7099 own arrangements to provide suitable library functions for
7103 @opindex mhard-float
7104 Generate output containing floating point instructions. This is the
7105 default if you use the unmodified sources.
7108 @itemx -mno-abicalls
7110 @opindex mno-abicalls
7111 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7112 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7113 position independent code.
7116 @itemx -mno-long-calls
7117 @opindex mlong-calls
7118 @opindex mno-long-calls
7119 Do all calls with the @samp{JALR} instruction, which requires
7120 loading up a function's address into a register before the call.
7121 You need to use this switch, if you call outside of the current
7122 512 megabyte segment to functions that are not through pointers.
7125 @itemx -mno-half-pic
7127 @opindex mno-half-pic
7128 Put pointers to extern references into the data section and load them
7129 up, rather than put the references in the text section.
7131 @item -membedded-pic
7132 @itemx -mno-embedded-pic
7133 @opindex membedded-pic
7134 @opindex mno-embedded-pic
7135 Generate PIC code suitable for some embedded systems. All calls are
7136 made using PC relative address, and all data is addressed using the $gp
7137 register. No more than 65536 bytes of global data may be used. This
7138 requires GNU as and GNU ld which do most of the work. This currently
7139 only works on targets which use ECOFF; it does not work with ELF@.
7141 @item -membedded-data
7142 @itemx -mno-embedded-data
7143 @opindex membedded-data
7144 @opindex mno-embedded-data
7145 Allocate variables to the read-only data section first if possible, then
7146 next in the small data section if possible, otherwise in data. This gives
7147 slightly slower code than the default, but reduces the amount of RAM required
7148 when executing, and thus may be preferred for some embedded systems.
7150 @item -muninit-const-in-rodata
7151 @itemx -mno-uninit-const-in-rodata
7152 @opindex muninit-const-in-rodata
7153 @opindex mno-uninit-const-in-rodata
7154 When used together with @option{-membedded-data}, it will always store uninitialized
7155 const variables in the read-only data section.
7157 @item -msingle-float
7158 @itemx -mdouble-float
7159 @opindex msingle-float
7160 @opindex mdouble-float
7161 The @option{-msingle-float} switch tells gcc to assume that the floating
7162 point coprocessor only supports single precision operations, as on the
7163 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7164 double precision operations. This is the default.
7170 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7171 as on the @samp{r4650} chip.
7175 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7176 @option{-mcpu=r4650}.
7182 Enable 16-bit instructions.
7186 Use the entry and exit pseudo ops. This option can only be used with
7191 Compile code for the processor in little endian mode.
7192 The requisite libraries are assumed to exist.
7196 Compile code for the processor in big endian mode.
7197 The requisite libraries are assumed to exist.
7201 @cindex smaller data references (MIPS)
7202 @cindex gp-relative references (MIPS)
7203 Put global and static items less than or equal to @var{num} bytes into
7204 the small data or bss sections instead of the normal data or bss
7205 section. This allows the assembler to emit one word memory reference
7206 instructions based on the global pointer (@var{gp} or @var{$28}),
7207 instead of the normal two words used. By default, @var{num} is 8 when
7208 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7209 @option{-G @var{num}} switch is also passed to the assembler and linker.
7210 All modules should be compiled with the same @option{-G @var{num}}
7215 Tell the MIPS assembler to not run its preprocessor over user
7216 assembler files (with a @samp{.s} suffix) when assembling them.
7220 Pass an option to gas which will cause nops to be inserted if
7221 the read of the destination register of an mfhi or mflo instruction
7222 occurs in the following two instructions.
7226 Do not include the default crt0.
7228 @item -mflush-func=@var{func}
7229 @itemx -mno-flush-func
7230 @opindex mflush-func
7231 Specifies the function to call to flush the I and D caches, or to not
7232 call any such function. If called, the function must take the same
7233 arguments as the common @code{_flush_func()}, that is, the address of the
7234 memory range for which the cache is being flushed, the size of the
7235 memory range, and the number 3 (to flush both caches). The default
7236 depends on the target gcc was configured for, but commonly is either
7237 @samp{_flush_func} or @samp{__cpu_flush}.
7240 These options are defined by the macro
7241 @code{TARGET_SWITCHES} in the machine description. The default for the
7242 options is also defined by that macro, which enables you to change the
7245 @node i386 and x86-64 Options
7246 @subsection Intel 386 and AMD x86-64 Options
7247 @cindex i386 Options
7248 @cindex x86-64 Options
7249 @cindex Intel 386 Options
7250 @cindex AMD x86-64 Options
7252 These @samp{-m} options are defined for the i386 and x86-64 family of
7256 @item -mcpu=@var{cpu-type}
7258 Tune to @var{cpu-type} everything applicable about the generated code, except
7259 for the ABI and the set of available instructions. The choices for
7260 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7261 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7262 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7263 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp}
7264 and @samp{athlon-mp}.
7266 While picking a specific @var{cpu-type} will schedule things appropriately
7267 for that particular chip, the compiler will not generate any code that
7268 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7269 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7270 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7271 AMD chips as opposed to the Intel ones.
7273 @item -march=@var{cpu-type}
7275 Generate instructions for the machine type @var{cpu-type}. The choices
7276 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7277 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7286 @opindex mpentiumpro
7287 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7288 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7289 These synonyms are deprecated.
7291 @item -mfpmath=@var{unit}
7293 generate floating point arithmetics for selected unit @var{unit}. the choices
7298 Use the standard 387 floating point coprocessor present majority of chips and
7299 emulated otherwise. Code compiled with this option will run almost everywhere.
7300 The temporary results are computed in 80bit precesion instead of precision
7301 specified by the type resulting in slightly different results compared to most
7302 of other chips. See @option{-ffloat-store} for more detailed description.
7304 This is the default choice for i386 compiler.
7307 Use scalar floating point instructions present in the SSE instruction set.
7308 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7309 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7310 instruction set supports only single precision arithmetics, thus the double and
7311 extended precision arithmetics is still done using 387. Later version, present
7312 only in Pentium4 and the future AMD x86-64 chips supports double precision
7315 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7316 @option{-msse2} switches to enable SSE extensions and make this option
7317 effective. For x86-64 compiler, these extensions are enabled by default.
7319 The resulting code should be considerably faster in majority of cases and avoid
7320 the numerical instability problems of 387 code, but may break some existing
7321 code that expects temporaries to be 80bit.
7323 This is the default choice for x86-64 compiler.
7326 Attempt to utilize both instruction sets at once. This effectivly double the
7327 amount of available registers and on chips with separate execution units for
7328 387 and SSE the execution resources too. Use this option with care, as it is
7329 still experimental, because gcc register allocator does not model separate
7330 functional units well resulting in instable performance.
7333 @item -masm=@var{dialect}
7334 @opindex masm=@var{dialect}
7335 Output asm instructions using selected @var{dialect}. Supported choices are
7336 @samp{intel} or @samp{att} (the default one).
7341 @opindex mno-ieee-fp
7342 Control whether or not the compiler uses IEEE floating point
7343 comparisons. These handle correctly the case where the result of a
7344 comparison is unordered.
7347 @opindex msoft-float
7348 Generate output containing library calls for floating point.
7349 @strong{Warning:} the requisite libraries are not part of GCC@.
7350 Normally the facilities of the machine's usual C compiler are used, but
7351 this can't be done directly in cross-compilation. You must make your
7352 own arrangements to provide suitable library functions for
7355 On machines where a function returns floating point results in the 80387
7356 register stack, some floating point opcodes may be emitted even if
7357 @option{-msoft-float} is used.
7359 @item -mno-fp-ret-in-387
7360 @opindex mno-fp-ret-in-387
7361 Do not use the FPU registers for return values of functions.
7363 The usual calling convention has functions return values of types
7364 @code{float} and @code{double} in an FPU register, even if there
7365 is no FPU@. The idea is that the operating system should emulate
7368 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7369 in ordinary CPU registers instead.
7371 @item -mno-fancy-math-387
7372 @opindex mno-fancy-math-387
7373 Some 387 emulators do not support the @code{sin}, @code{cos} and
7374 @code{sqrt} instructions for the 387. Specify this option to avoid
7375 generating those instructions. This option is the default on FreeBSD,
7376 OpenBSD and NetBSD@. This option is overridden when @option{-march}
7377 indicates that the target cpu will always have an FPU and so the
7378 instruction will not need emulation. As of revision 2.6.1, these
7379 instructions are not generated unless you also use the
7380 @option{-funsafe-math-optimizations} switch.
7382 @item -malign-double
7383 @itemx -mno-align-double
7384 @opindex malign-double
7385 @opindex mno-align-double
7386 Control whether GCC aligns @code{double}, @code{long double}, and
7387 @code{long long} variables on a two word boundary or a one word
7388 boundary. Aligning @code{double} variables on a two word boundary will
7389 produce code that runs somewhat faster on a @samp{Pentium} at the
7390 expense of more memory.
7392 @item -m128bit-long-double
7393 @opindex m128bit-long-double
7394 Control the size of @code{long double} type. i386 application binary interface
7395 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7396 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7397 impossible to reach with 12 byte long doubles in the array accesses.
7399 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7400 structures and arrays containing @code{long double} will change their size as
7401 well as function calling convention for function taking @code{long double}
7404 @item -m96bit-long-double
7405 @opindex m96bit-long-double
7406 Set the size of @code{long double} to 96 bits as required by the i386
7407 application binary interface. This is the default.
7410 @itemx -mno-svr3-shlib
7411 @opindex msvr3-shlib
7412 @opindex mno-svr3-shlib
7413 Control whether GCC places uninitialized local variables into the
7414 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7415 into @code{bss}. These options are meaningful only on System V Release 3.
7419 Use a different function-calling convention, in which functions that
7420 take a fixed number of arguments return with the @code{ret} @var{num}
7421 instruction, which pops their arguments while returning. This saves one
7422 instruction in the caller since there is no need to pop the arguments
7425 You can specify that an individual function is called with this calling
7426 sequence with the function attribute @samp{stdcall}. You can also
7427 override the @option{-mrtd} option by using the function attribute
7428 @samp{cdecl}. @xref{Function Attributes}.
7430 @strong{Warning:} this calling convention is incompatible with the one
7431 normally used on Unix, so you cannot use it if you need to call
7432 libraries compiled with the Unix compiler.
7434 Also, you must provide function prototypes for all functions that
7435 take variable numbers of arguments (including @code{printf});
7436 otherwise incorrect code will be generated for calls to those
7439 In addition, seriously incorrect code will result if you call a
7440 function with too many arguments. (Normally, extra arguments are
7441 harmlessly ignored.)
7443 @item -mregparm=@var{num}
7445 Control how many registers are used to pass integer arguments. By
7446 default, no registers are used to pass arguments, and at most 3
7447 registers can be used. You can control this behavior for a specific
7448 function by using the function attribute @samp{regparm}.
7449 @xref{Function Attributes}.
7451 @strong{Warning:} if you use this switch, and
7452 @var{num} is nonzero, then you must build all modules with the same
7453 value, including any libraries. This includes the system libraries and
7456 @item -mpreferred-stack-boundary=@var{num}
7457 @opindex mpreferred-stack-boundary
7458 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7459 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7460 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7461 size (@option{-Os}), in which case the default is the minimum correct
7462 alignment (4 bytes for x86, and 8 bytes for x86-64).
7464 On Pentium and PentiumPro, @code{double} and @code{long double} values
7465 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7466 suffer significant run time performance penalties. On Pentium III, the
7467 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7468 penalties if it is not 16 byte aligned.
7470 To ensure proper alignment of this values on the stack, the stack boundary
7471 must be as aligned as that required by any value stored on the stack.
7472 Further, every function must be generated such that it keeps the stack
7473 aligned. Thus calling a function compiled with a higher preferred
7474 stack boundary from a function compiled with a lower preferred stack
7475 boundary will most likely misalign the stack. It is recommended that
7476 libraries that use callbacks always use the default setting.
7478 This extra alignment does consume extra stack space, and generally
7479 increases code size. Code that is sensitive to stack space usage, such
7480 as embedded systems and operating system kernels, may want to reduce the
7481 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7497 These switches enable or disable the use of built-in functions that allow
7498 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7500 @xref{X86 Built-in Functions}, for details of the functions enabled
7501 and disabled by these switches.
7504 @itemx -mno-push-args
7506 @opindex mno-push-args
7507 Use PUSH operations to store outgoing parameters. This method is shorter
7508 and usually equally fast as method using SUB/MOV operations and is enabled
7509 by default. In some cases disabling it may improve performance because of
7510 improved scheduling and reduced dependencies.
7512 @item -maccumulate-outgoing-args
7513 @opindex maccumulate-outgoing-args
7514 If enabled, the maximum amount of space required for outgoing arguments will be
7515 computed in the function prologue. This is faster on most modern CPUs
7516 because of reduced dependencies, improved scheduling and reduced stack usage
7517 when preferred stack boundary is not equal to 2. The drawback is a notable
7518 increase in code size. This switch implies @option{-mno-push-args}.
7522 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7523 on thread-safe exception handling must compile and link all code with the
7524 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7525 @option{-D_MT}; when linking, it links in a special thread helper library
7526 @option{-lmingwthrd} which cleans up per thread exception handling data.
7528 @item -mno-align-stringops
7529 @opindex mno-align-stringops
7530 Do not align destination of inlined string operations. This switch reduces
7531 code size and improves performance in case the destination is already aligned,
7532 but gcc don't know about it.
7534 @item -minline-all-stringops
7535 @opindex minline-all-stringops
7536 By default GCC inlines string operations only when destination is known to be
7537 aligned at least to 4 byte boundary. This enables more inlining, increase code
7538 size, but may improve performance of code that depends on fast memcpy, strlen
7539 and memset for short lengths.
7541 @item -momit-leaf-frame-pointer
7542 @opindex momit-leaf-frame-pointer
7543 Don't keep the frame pointer in a register for leaf functions. This
7544 avoids the instructions to save, set up and restore frame pointers and
7545 makes an extra register available in leaf functions. The option
7546 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7547 which might make debugging harder.
7550 These @samp{-m} switches are supported in addition to the above
7551 on AMD x86-64 processors in 64-bit environments.
7558 Generate code for a 32-bit or 64-bit environment.
7559 The 32-bit environment sets int, long and pointer to 32 bits and
7560 generates code that runs on any i386 system.
7561 The 64-bit environment sets int to 32 bits and long and pointer
7562 to 64 bits and generates code for AMD's x86-64 architecture.
7565 @opindex no-red-zone
7566 Do not use a so called red zone for x86-64 code. The red zone is mandated
7567 by the x86-64 ABI, it is a 128-byte area beyond the location of the
7568 stack pointer that will not be modified by signal or interrupt handlers
7569 and therefore can be used for temporary data without adjusting the stack
7570 pointer. The flag @option{-mno-red-zone} disables this red zone.
7574 @subsection HPPA Options
7575 @cindex HPPA Options
7577 These @samp{-m} options are defined for the HPPA family of computers:
7580 @item -march=@var{architecture-type}
7582 Generate code for the specified architecture. The choices for
7583 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7584 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7585 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7586 architecture option for your machine. Code compiled for lower numbered
7587 architectures will run on higher numbered architectures, but not the
7590 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7591 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7595 @itemx -mpa-risc-1-1
7596 @itemx -mpa-risc-2-0
7597 @opindex mpa-risc-1-0
7598 @opindex mpa-risc-1-1
7599 @opindex mpa-risc-2-0
7600 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7603 @opindex mbig-switch
7604 Generate code suitable for big switch tables. Use this option only if
7605 the assembler/linker complain about out of range branches within a switch
7608 @item -mjump-in-delay
7609 @opindex mjump-in-delay
7610 Fill delay slots of function calls with unconditional jump instructions
7611 by modifying the return pointer for the function call to be the target
7612 of the conditional jump.
7614 @item -mdisable-fpregs
7615 @opindex mdisable-fpregs
7616 Prevent floating point registers from being used in any manner. This is
7617 necessary for compiling kernels which perform lazy context switching of
7618 floating point registers. If you use this option and attempt to perform
7619 floating point operations, the compiler will abort.
7621 @item -mdisable-indexing
7622 @opindex mdisable-indexing
7623 Prevent the compiler from using indexing address modes. This avoids some
7624 rather obscure problems when compiling MIG generated code under MACH@.
7626 @item -mno-space-regs
7627 @opindex mno-space-regs
7628 Generate code that assumes the target has no space registers. This allows
7629 GCC to generate faster indirect calls and use unscaled index address modes.
7631 Such code is suitable for level 0 PA systems and kernels.
7633 @item -mfast-indirect-calls
7634 @opindex mfast-indirect-calls
7635 Generate code that assumes calls never cross space boundaries. This
7636 allows GCC to emit code which performs faster indirect calls.
7638 This option will not work in the presence of shared libraries or nested
7641 @item -mlong-load-store
7642 @opindex mlong-load-store
7643 Generate 3-instruction load and store sequences as sometimes required by
7644 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7647 @item -mportable-runtime
7648 @opindex mportable-runtime
7649 Use the portable calling conventions proposed by HP for ELF systems.
7653 Enable the use of assembler directives only GAS understands.
7655 @item -mschedule=@var{cpu-type}
7657 Schedule code according to the constraints for the machine type
7658 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7659 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
7660 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
7661 proper scheduling option for your machine. The default scheduling is
7665 @opindex mlinker-opt
7666 Enable the optimization pass in the HPUX linker. Note this makes symbolic
7667 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
7668 in which they give bogus error messages when linking some programs.
7671 @opindex msoft-float
7672 Generate output containing library calls for floating point.
7673 @strong{Warning:} the requisite libraries are not available for all HPPA
7674 targets. Normally the facilities of the machine's usual C compiler are
7675 used, but this cannot be done directly in cross-compilation. You must make
7676 your own arrangements to provide suitable library functions for
7677 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7678 does provide software floating point support.
7680 @option{-msoft-float} changes the calling convention in the output file;
7681 therefore, it is only useful if you compile @emph{all} of a program with
7682 this option. In particular, you need to compile @file{libgcc.a}, the
7683 library that comes with GCC, with @option{-msoft-float} in order for
7687 @node Intel 960 Options
7688 @subsection Intel 960 Options
7690 These @samp{-m} options are defined for the Intel 960 implementations:
7693 @item -m@var{cpu-type}
7701 Assume the defaults for the machine type @var{cpu-type} for some of
7702 the other options, including instruction scheduling, floating point
7703 support, and addressing modes. The choices for @var{cpu-type} are
7704 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
7705 @samp{sa}, and @samp{sb}.
7712 @opindex msoft-float
7713 The @option{-mnumerics} option indicates that the processor does support
7714 floating-point instructions. The @option{-msoft-float} option indicates
7715 that floating-point support should not be assumed.
7717 @item -mleaf-procedures
7718 @itemx -mno-leaf-procedures
7719 @opindex mleaf-procedures
7720 @opindex mno-leaf-procedures
7721 Do (or do not) attempt to alter leaf procedures to be callable with the
7722 @code{bal} instruction as well as @code{call}. This will result in more
7723 efficient code for explicit calls when the @code{bal} instruction can be
7724 substituted by the assembler or linker, but less efficient code in other
7725 cases, such as calls via function pointers, or using a linker that doesn't
7726 support this optimization.
7729 @itemx -mno-tail-call
7731 @opindex mno-tail-call
7732 Do (or do not) make additional attempts (beyond those of the
7733 machine-independent portions of the compiler) to optimize tail-recursive
7734 calls into branches. You may not want to do this because the detection of
7735 cases where this is not valid is not totally complete. The default is
7736 @option{-mno-tail-call}.
7738 @item -mcomplex-addr
7739 @itemx -mno-complex-addr
7740 @opindex mcomplex-addr
7741 @opindex mno-complex-addr
7742 Assume (or do not assume) that the use of a complex addressing mode is a
7743 win on this implementation of the i960. Complex addressing modes may not
7744 be worthwhile on the K-series, but they definitely are on the C-series.
7745 The default is currently @option{-mcomplex-addr} for all processors except
7749 @itemx -mno-code-align
7750 @opindex mcode-align
7751 @opindex mno-code-align
7752 Align code to 8-byte boundaries for faster fetching (or don't bother).
7753 Currently turned on by default for C-series implementations only.
7756 @item -mclean-linkage
7757 @itemx -mno-clean-linkage
7758 @opindex mclean-linkage
7759 @opindex mno-clean-linkage
7760 These options are not fully implemented.
7764 @itemx -mic2.0-compat
7765 @itemx -mic3.0-compat
7767 @opindex mic2.0-compat
7768 @opindex mic3.0-compat
7769 Enable compatibility with iC960 v2.0 or v3.0.
7773 @opindex masm-compat
7775 Enable compatibility with the iC960 assembler.
7777 @item -mstrict-align
7778 @itemx -mno-strict-align
7779 @opindex mstrict-align
7780 @opindex mno-strict-align
7781 Do not permit (do permit) unaligned accesses.
7785 Enable structure-alignment compatibility with Intel's gcc release version
7786 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
7788 @item -mlong-double-64
7789 @opindex mlong-double-64
7790 Implement type @samp{long double} as 64-bit floating point numbers.
7791 Without the option @samp{long double} is implemented by 80-bit
7792 floating point numbers. The only reason we have it because there is
7793 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
7794 is only useful for people using soft-float targets. Otherwise, we
7795 should recommend against use of it.
7799 @node DEC Alpha Options
7800 @subsection DEC Alpha Options
7802 These @samp{-m} options are defined for the DEC Alpha implementations:
7805 @item -mno-soft-float
7807 @opindex mno-soft-float
7808 @opindex msoft-float
7809 Use (do not use) the hardware floating-point instructions for
7810 floating-point operations. When @option{-msoft-float} is specified,
7811 functions in @file{libgcc.a} will be used to perform floating-point
7812 operations. Unless they are replaced by routines that emulate the
7813 floating-point operations, or compiled in such a way as to call such
7814 emulations routines, these routines will issue floating-point
7815 operations. If you are compiling for an Alpha without floating-point
7816 operations, you must ensure that the library is built so as not to call
7819 Note that Alpha implementations without floating-point operations are
7820 required to have floating-point registers.
7825 @opindex mno-fp-regs
7826 Generate code that uses (does not use) the floating-point register set.
7827 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
7828 register set is not used, floating point operands are passed in integer
7829 registers as if they were integers and floating-point results are passed
7830 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
7831 so any function with a floating-point argument or return value called by code
7832 compiled with @option{-mno-fp-regs} must also be compiled with that
7835 A typical use of this option is building a kernel that does not use,
7836 and hence need not save and restore, any floating-point registers.
7840 The Alpha architecture implements floating-point hardware optimized for
7841 maximum performance. It is mostly compliant with the IEEE floating
7842 point standard. However, for full compliance, software assistance is
7843 required. This option generates code fully IEEE compliant code
7844 @emph{except} that the @var{inexact-flag} is not maintained (see below).
7845 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
7846 defined during compilation. The resulting code is less efficient but is
7847 able to correctly support denormalized numbers and exceptional IEEE
7848 values such as not-a-number and plus/minus infinity. Other Alpha
7849 compilers call this option @option{-ieee_with_no_inexact}.
7851 @item -mieee-with-inexact
7852 @opindex mieee-with-inexact
7853 This is like @option{-mieee} except the generated code also maintains
7854 the IEEE @var{inexact-flag}. Turning on this option causes the
7855 generated code to implement fully-compliant IEEE math. In addition to
7856 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
7857 macro. On some Alpha implementations the resulting code may execute
7858 significantly slower than the code generated by default. Since there is
7859 very little code that depends on the @var{inexact-flag}, you should
7860 normally not specify this option. Other Alpha compilers call this
7861 option @option{-ieee_with_inexact}.
7863 @item -mfp-trap-mode=@var{trap-mode}
7864 @opindex mfp-trap-mode
7865 This option controls what floating-point related traps are enabled.
7866 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
7867 The trap mode can be set to one of four values:
7871 This is the default (normal) setting. The only traps that are enabled
7872 are the ones that cannot be disabled in software (e.g., division by zero
7876 In addition to the traps enabled by @samp{n}, underflow traps are enabled
7880 Like @samp{su}, but the instructions are marked to be safe for software
7881 completion (see Alpha architecture manual for details).
7884 Like @samp{su}, but inexact traps are enabled as well.
7887 @item -mfp-rounding-mode=@var{rounding-mode}
7888 @opindex mfp-rounding-mode
7889 Selects the IEEE rounding mode. Other Alpha compilers call this option
7890 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
7895 Normal IEEE rounding mode. Floating point numbers are rounded towards
7896 the nearest machine number or towards the even machine number in case
7900 Round towards minus infinity.
7903 Chopped rounding mode. Floating point numbers are rounded towards zero.
7906 Dynamic rounding mode. A field in the floating point control register
7907 (@var{fpcr}, see Alpha architecture reference manual) controls the
7908 rounding mode in effect. The C library initializes this register for
7909 rounding towards plus infinity. Thus, unless your program modifies the
7910 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
7913 @item -mtrap-precision=@var{trap-precision}
7914 @opindex mtrap-precision
7915 In the Alpha architecture, floating point traps are imprecise. This
7916 means without software assistance it is impossible to recover from a
7917 floating trap and program execution normally needs to be terminated.
7918 GCC can generate code that can assist operating system trap handlers
7919 in determining the exact location that caused a floating point trap.
7920 Depending on the requirements of an application, different levels of
7921 precisions can be selected:
7925 Program precision. This option is the default and means a trap handler
7926 can only identify which program caused a floating point exception.
7929 Function precision. The trap handler can determine the function that
7930 caused a floating point exception.
7933 Instruction precision. The trap handler can determine the exact
7934 instruction that caused a floating point exception.
7937 Other Alpha compilers provide the equivalent options called
7938 @option{-scope_safe} and @option{-resumption_safe}.
7940 @item -mieee-conformant
7941 @opindex mieee-conformant
7942 This option marks the generated code as IEEE conformant. You must not
7943 use this option unless you also specify @option{-mtrap-precision=i} and either
7944 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
7945 is to emit the line @samp{.eflag 48} in the function prologue of the
7946 generated assembly file. Under DEC Unix, this has the effect that
7947 IEEE-conformant math library routines will be linked in.
7949 @item -mbuild-constants
7950 @opindex mbuild-constants
7951 Normally GCC examines a 32- or 64-bit integer constant to
7952 see if it can construct it from smaller constants in two or three
7953 instructions. If it cannot, it will output the constant as a literal and
7954 generate code to load it from the data segment at runtime.
7956 Use this option to require GCC to construct @emph{all} integer constants
7957 using code, even if it takes more instructions (the maximum is six).
7959 You would typically use this option to build a shared library dynamic
7960 loader. Itself a shared library, it must relocate itself in memory
7961 before it can find the variables and constants in its own data segment.
7967 Select whether to generate code to be assembled by the vendor-supplied
7968 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
7986 Indicate whether GCC should generate code to use the optional BWX,
7987 CIX, FIX and MAX instruction sets. The default is to use the instruction
7988 sets supported by the CPU type specified via @option{-mcpu=} option or that
7989 of the CPU on which GCC was built if none was specified.
7994 @opindex mfloat-ieee
7995 Generate code that uses (does not use) VAX F and G floating point
7996 arithmetic instead of IEEE single and double precision.
7998 @item -mexplicit-relocs
7999 @itemx -mno-explicit-relocs
8000 @opindex mexplicit-relocs
8001 @opindex mno-explicit-relocs
8002 Older Alpha assemblers provided no way to generate symbol relocations
8003 except via assembler macros. Use of these macros does not allow
8004 optimial instruction scheduling. GNU binutils as of version 2.12
8005 supports a new syntax that allows the compiler to explicitly mark
8006 which relocations should apply to which instructions. This option
8007 is mostly useful for debugging, as GCC detects the capabilities of
8008 the assembler when it is built and sets the default accordingly.
8012 @opindex msmall-data
8013 @opindex mlarge-data
8014 When @option{-mexplicit-relocs} is in effect, static data is
8015 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8016 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8017 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8018 16-bit relocations off of the @code{$gp} register. This limits the
8019 size of the small data area to 64KB, but allows the variables to be
8020 directly accessed via a single instruction.
8022 The default is @option{-mlarge-data}. With this option the data area
8023 is limited to just below 2GB. Programs that require more than 2GB of
8024 data must use @code{malloc} or @code{mmap} to allocate the data in the
8025 heap instead of in the program's data segment.
8027 When generating code for shared libraries, @option{-fpic} implies
8028 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8030 @item -mcpu=@var{cpu_type}
8032 Set the instruction set and instruction scheduling parameters for
8033 machine type @var{cpu_type}. You can specify either the @samp{EV}
8034 style name or the corresponding chip number. GCC supports scheduling
8035 parameters for the EV4, EV5 and EV6 family of processors and will
8036 choose the default values for the instruction set from the processor
8037 you specify. If you do not specify a processor type, GCC will default
8038 to the processor on which the compiler was built.
8040 Supported values for @var{cpu_type} are
8046 Schedules as an EV4 and has no instruction set extensions.
8050 Schedules as an EV5 and has no instruction set extensions.
8054 Schedules as an EV5 and supports the BWX extension.
8059 Schedules as an EV5 and supports the BWX and MAX extensions.
8063 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8067 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8070 @item -mtune=@var{cpu_type}
8072 Set only the instruction scheduling parameters for machine type
8073 @var{cpu_type}. The instruction set is not changed.
8075 @item -mmemory-latency=@var{time}
8076 @opindex mmemory-latency
8077 Sets the latency the scheduler should assume for typical memory
8078 references as seen by the application. This number is highly
8079 dependent on the memory access patterns used by the application
8080 and the size of the external cache on the machine.
8082 Valid options for @var{time} are
8086 A decimal number representing clock cycles.
8092 The compiler contains estimates of the number of clock cycles for
8093 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8094 (also called Dcache, Scache, and Bcache), as well as to main memory.
8095 Note that L3 is only valid for EV5.
8100 @node DEC Alpha/VMS Options
8101 @subsection DEC Alpha/VMS Options
8103 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8106 @item -mvms-return-codes
8107 @opindex mvms-return-codes
8108 Return VMS condition codes from main. The default is to return POSIX
8109 style condition (e.g.@ error) codes.
8112 @node Clipper Options
8113 @subsection Clipper Options
8115 These @samp{-m} options are defined for the Clipper implementations:
8120 Produce code for a C300 Clipper processor. This is the default.
8124 Produce code for a C400 Clipper processor, i.e.@: use floating point
8128 @node H8/300 Options
8129 @subsection H8/300 Options
8131 These @samp{-m} options are defined for the H8/300 implementations:
8136 Shorten some address references at link time, when possible; uses the
8137 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8138 ld.info, Using ld}, for a fuller description.
8142 Generate code for the H8/300H@.
8146 Generate code for the H8/S@.
8150 Generate code for the H8/S2600. This switch must be used with @option{-ms}.
8154 Make @code{int} data 32 bits by default.
8158 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
8159 The default for the H8/300H and H8/S is to align longs and floats on 4
8161 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8162 This option has no effect on the H8/300.
8166 @subsection SH Options
8168 These @samp{-m} options are defined for the SH implementations:
8173 Generate code for the SH1.
8177 Generate code for the SH2.
8181 Generate code for the SH3.
8185 Generate code for the SH3e.
8189 Generate code for the SH4 without a floating-point unit.
8191 @item -m4-single-only
8192 @opindex m4-single-only
8193 Generate code for the SH4 with a floating-point unit that only
8194 supports single-precision arithmetic.
8198 Generate code for the SH4 assuming the floating-point unit is in
8199 single-precision mode by default.
8203 Generate code for the SH4.
8207 Compile code for the processor in big endian mode.
8211 Compile code for the processor in little endian mode.
8215 Align doubles at 64-bit boundaries. Note that this changes the calling
8216 conventions, and thus some functions from the standard C library will
8217 not work unless you recompile it first with @option{-mdalign}.
8221 Shorten some address references at link time, when possible; uses the
8222 linker option @option{-relax}.
8226 Use 32-bit offsets in @code{switch} tables. The default is to use
8231 Enable the use of the instruction @code{fmovd}.
8235 Comply with the calling conventions defined by Hitachi.
8239 Mark the @code{MAC} register as call-clobbered, even if
8240 @option{-mhitachi} is given.
8244 Increase IEEE-compliance of floating-point code.
8248 Dump instruction size and location in the assembly code.
8252 This option is deprecated. It pads structures to multiple of 4 bytes,
8253 which is incompatible with the SH ABI@.
8257 Optimize for space instead of speed. Implied by @option{-Os}.
8261 When generating position-independent code, emit function calls using
8262 the Global Offset Table instead of the Procedure Linkage Table.
8266 Generate a library function call to invalidate instruction cache
8267 entries, after fixing up a trampoline. This library function call
8268 doesn't assume it can write to the whole memory address space. This
8269 is the default when the target is @code{sh-*-linux*}.
8272 @node System V Options
8273 @subsection Options for System V
8275 These additional options are available on System V Release 4 for
8276 compatibility with other compilers on those systems:
8281 Create a shared object.
8282 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8286 Identify the versions of each tool used by the compiler, in a
8287 @code{.ident} assembler directive in the output.
8291 Refrain from adding @code{.ident} directives to the output file (this is
8294 @item -YP,@var{dirs}
8296 Search the directories @var{dirs}, and no others, for libraries
8297 specified with @option{-l}.
8301 Look in the directory @var{dir} to find the M4 preprocessor.
8302 The assembler uses this option.
8303 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8304 @c the generic assembler that comes with Solaris takes just -Ym.
8307 @node TMS320C3x/C4x Options
8308 @subsection TMS320C3x/C4x Options
8309 @cindex TMS320C3x/C4x Options
8311 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8315 @item -mcpu=@var{cpu_type}
8317 Set the instruction set, register set, and instruction scheduling
8318 parameters for machine type @var{cpu_type}. Supported values for
8319 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8320 @samp{c44}. The default is @samp{c40} to generate code for the
8325 @itemx -msmall-memory
8327 @opindex mbig-memory
8329 @opindex msmall-memory
8331 Generates code for the big or small memory model. The small memory
8332 model assumed that all data fits into one 64K word page. At run-time
8333 the data page (DP) register must be set to point to the 64K page
8334 containing the .bss and .data program sections. The big memory model is
8335 the default and requires reloading of the DP register for every direct
8342 Allow (disallow) allocation of general integer operands into the block
8349 Enable (disable) generation of code using decrement and branch,
8350 DBcond(D), instructions. This is enabled by default for the C4x. To be
8351 on the safe side, this is disabled for the C3x, since the maximum
8352 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
8353 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
8354 that it can utilise the decrement and branch instruction, but will give
8355 up if there is more than one memory reference in the loop. Thus a loop
8356 where the loop counter is decremented can generate slightly more
8357 efficient code, in cases where the RPTB instruction cannot be utilised.
8359 @item -mdp-isr-reload
8361 @opindex mdp-isr-reload
8363 Force the DP register to be saved on entry to an interrupt service
8364 routine (ISR), reloaded to point to the data section, and restored on
8365 exit from the ISR@. This should not be required unless someone has
8366 violated the small memory model by modifying the DP register, say within
8373 For the C3x use the 24-bit MPYI instruction for integer multiplies
8374 instead of a library call to guarantee 32-bit results. Note that if one
8375 of the operands is a constant, then the multiplication will be performed
8376 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8377 then squaring operations are performed inline instead of a library call.
8380 @itemx -mno-fast-fix
8382 @opindex mno-fast-fix
8383 The C3x/C4x FIX instruction to convert a floating point value to an
8384 integer value chooses the nearest integer less than or equal to the
8385 floating point value rather than to the nearest integer. Thus if the
8386 floating point number is negative, the result will be incorrectly
8387 truncated an additional code is necessary to detect and correct this
8388 case. This option can be used to disable generation of the additional
8389 code required to correct the result.
8395 Enable (disable) generation of repeat block sequences using the RPTB
8396 instruction for zero overhead looping. The RPTB construct is only used
8397 for innermost loops that do not call functions or jump across the loop
8398 boundaries. There is no advantage having nested RPTB loops due to the
8399 overhead required to save and restore the RC, RS, and RE registers.
8400 This is enabled by default with @option{-O2}.
8402 @item -mrpts=@var{count}
8406 Enable (disable) the use of the single instruction repeat instruction
8407 RPTS@. If a repeat block contains a single instruction, and the loop
8408 count can be guaranteed to be less than the value @var{count}, GCC will
8409 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8410 then a RPTS will be emitted even if the loop count cannot be determined
8411 at compile time. Note that the repeated instruction following RPTS does
8412 not have to be reloaded from memory each iteration, thus freeing up the
8413 CPU buses for operands. However, since interrupts are blocked by this
8414 instruction, it is disabled by default.
8416 @item -mloop-unsigned
8417 @itemx -mno-loop-unsigned
8418 @opindex mloop-unsigned
8419 @opindex mno-loop-unsigned
8420 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8421 is @math{2^{31} + 1} since these instructions test if the iteration count is
8422 negative to terminate the loop. If the iteration count is unsigned
8423 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
8424 exceeded. This switch allows an unsigned iteration count.
8428 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8429 with. This also enforces compatibility with the API employed by the TI
8430 C3x C compiler. For example, long doubles are passed as structures
8431 rather than in floating point registers.
8437 Generate code that uses registers (stack) for passing arguments to functions.
8438 By default, arguments are passed in registers where possible rather
8439 than by pushing arguments on to the stack.
8441 @item -mparallel-insns
8442 @itemx -mno-parallel-insns
8443 @opindex mparallel-insns
8444 @opindex mno-parallel-insns
8445 Allow the generation of parallel instructions. This is enabled by
8446 default with @option{-O2}.
8448 @item -mparallel-mpy
8449 @itemx -mno-parallel-mpy
8450 @opindex mparallel-mpy
8451 @opindex mno-parallel-mpy
8452 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
8453 provided @option{-mparallel-insns} is also specified. These instructions have
8454 tight register constraints which can pessimize the code generation
8460 @subsection V850 Options
8461 @cindex V850 Options
8463 These @samp{-m} options are defined for V850 implementations:
8467 @itemx -mno-long-calls
8468 @opindex mlong-calls
8469 @opindex mno-long-calls
8470 Treat all calls as being far away (near). If calls are assumed to be
8471 far away, the compiler will always load the functions address up into a
8472 register, and call indirect through the pointer.
8478 Do not optimize (do optimize) basic blocks that use the same index
8479 pointer 4 or more times to copy pointer into the @code{ep} register, and
8480 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
8481 option is on by default if you optimize.
8483 @item -mno-prolog-function
8484 @itemx -mprolog-function
8485 @opindex mno-prolog-function
8486 @opindex mprolog-function
8487 Do not use (do use) external functions to save and restore registers at
8488 the prolog and epilog of a function. The external functions are slower,
8489 but use less code space if more than one function saves the same number
8490 of registers. The @option{-mprolog-function} option is on by default if
8495 Try to make the code as small as possible. At present, this just turns
8496 on the @option{-mep} and @option{-mprolog-function} options.
8500 Put static or global variables whose size is @var{n} bytes or less into
8501 the tiny data area that register @code{ep} points to. The tiny data
8502 area can hold up to 256 bytes in total (128 bytes for byte references).
8506 Put static or global variables whose size is @var{n} bytes or less into
8507 the small data area that register @code{gp} points to. The small data
8508 area can hold up to 64 kilobytes.
8512 Put static or global variables whose size is @var{n} bytes or less into
8513 the first 32 kilobytes of memory.
8517 Specify that the target processor is the V850.
8520 @opindex mbig-switch
8521 Generate code suitable for big switch tables. Use this option only if
8522 the assembler/linker complain about out of range branches within a switch
8527 @subsection ARC Options
8530 These options are defined for ARC implementations:
8535 Compile code for little endian mode. This is the default.
8539 Compile code for big endian mode.
8542 @opindex mmangle-cpu
8543 Prepend the name of the cpu to all public symbol names.
8544 In multiple-processor systems, there are many ARC variants with different
8545 instruction and register set characteristics. This flag prevents code
8546 compiled for one cpu to be linked with code compiled for another.
8547 No facility exists for handling variants that are ``almost identical''.
8548 This is an all or nothing option.
8550 @item -mcpu=@var{cpu}
8552 Compile code for ARC variant @var{cpu}.
8553 Which variants are supported depend on the configuration.
8554 All variants support @option{-mcpu=base}, this is the default.
8556 @item -mtext=@var{text-section}
8557 @itemx -mdata=@var{data-section}
8558 @itemx -mrodata=@var{readonly-data-section}
8562 Put functions, data, and readonly data in @var{text-section},
8563 @var{data-section}, and @var{readonly-data-section} respectively
8564 by default. This can be overridden with the @code{section} attribute.
8565 @xref{Variable Attributes}.
8570 @subsection NS32K Options
8571 @cindex NS32K options
8573 These are the @samp{-m} options defined for the 32000 series. The default
8574 values for these options depends on which style of 32000 was selected when
8575 the compiler was configured; the defaults for the most common choices are
8583 Generate output for a 32032. This is the default
8584 when the compiler is configured for 32032 and 32016 based systems.
8590 Generate output for a 32332. This is the default
8591 when the compiler is configured for 32332-based systems.
8597 Generate output for a 32532. This is the default
8598 when the compiler is configured for 32532-based systems.
8602 Generate output containing 32081 instructions for floating point.
8603 This is the default for all systems.
8607 Generate output containing 32381 instructions for floating point. This
8608 also implies @option{-m32081}. The 32381 is only compatible with the 32332
8609 and 32532 cpus. This is the default for the pc532-netbsd configuration.
8613 Try and generate multiply-add floating point instructions @code{polyF}
8614 and @code{dotF}. This option is only available if the @option{-m32381}
8615 option is in effect. Using these instructions requires changes to
8616 register allocation which generally has a negative impact on
8617 performance. This option should only be enabled when compiling code
8618 particularly likely to make heavy use of multiply-add instructions.
8621 @opindex mnomulti-add
8622 Do not try and generate multiply-add floating point instructions
8623 @code{polyF} and @code{dotF}. This is the default on all platforms.
8626 @opindex msoft-float
8627 Generate output containing library calls for floating point.
8628 @strong{Warning:} the requisite libraries may not be available.
8631 @opindex mnobitfield
8632 Do not use the bit-field instructions. On some machines it is faster to
8633 use shifting and masking operations. This is the default for the pc532.
8637 Do use the bit-field instructions. This is the default for all platforms
8642 Use a different function-calling convention, in which functions
8643 that take a fixed number of arguments return pop their
8644 arguments on return with the @code{ret} instruction.
8646 This calling convention is incompatible with the one normally
8647 used on Unix, so you cannot use it if you need to call libraries
8648 compiled with the Unix compiler.
8650 Also, you must provide function prototypes for all functions that
8651 take variable numbers of arguments (including @code{printf});
8652 otherwise incorrect code will be generated for calls to those
8655 In addition, seriously incorrect code will result if you call a
8656 function with too many arguments. (Normally, extra arguments are
8657 harmlessly ignored.)
8659 This option takes its name from the 680x0 @code{rtd} instruction.
8664 Use a different function-calling convention where the first two arguments
8665 are passed in registers.
8667 This calling convention is incompatible with the one normally
8668 used on Unix, so you cannot use it if you need to call libraries
8669 compiled with the Unix compiler.
8672 @opindex mnoregparam
8673 Do not pass any arguments in registers. This is the default for all
8678 It is OK to use the sb as an index register which is always loaded with
8679 zero. This is the default for the pc532-netbsd target.
8683 The sb register is not available for use or has not been initialized to
8684 zero by the run time system. This is the default for all targets except
8685 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
8686 @option{-fpic} is set.
8690 Many ns32000 series addressing modes use displacements of up to 512MB@.
8691 If an address is above 512MB then displacements from zero can not be used.
8692 This option causes code to be generated which can be loaded above 512MB@.
8693 This may be useful for operating systems or ROM code.
8697 Assume code will be loaded in the first 512MB of virtual address space.
8698 This is the default for all platforms.
8704 @subsection AVR Options
8707 These options are defined for AVR implementations:
8710 @item -mmcu=@var{mcu}
8712 Specify ATMEL AVR instruction set or MCU type.
8714 Instruction set avr1 is for the minimal AVR core, not supported by the C
8715 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8716 attiny11, attiny12, attiny15, attiny28).
8718 Instruction set avr2 (default) is for the classic AVR core with up to
8719 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8720 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8721 at90c8534, at90s8535).
8723 Instruction set avr3 is for the classic AVR core with up to 128K program
8724 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8726 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8727 memory space (MCU types: atmega8, atmega83, atmega85).
8729 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8730 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8731 atmega64, atmega128, at43usb355, at94k).
8735 Output instruction sizes to the asm file.
8737 @item -minit-stack=@var{N}
8738 @opindex minit-stack
8739 Specify the initial stack address, which may be a symbol or numeric value,
8740 @samp{__stack} is the default.
8742 @item -mno-interrupts
8743 @opindex mno-interrupts
8744 Generated code is not compatible with hardware interrupts.
8745 Code size will be smaller.
8747 @item -mcall-prologues
8748 @opindex mcall-prologues
8749 Functions prologues/epilogues expanded as call to appropriate
8750 subroutines. Code size will be smaller.
8752 @item -mno-tablejump
8753 @opindex mno-tablejump
8754 Do not generate tablejump insns which sometimes increase code size.
8757 @opindex mtiny-stack
8758 Change only the low 8 bits of the stack pointer.
8762 @subsection MCore Options
8763 @cindex MCore options
8765 These are the @samp{-m} options defined for the Motorola M*Core
8775 @opindex mno-hardlit
8776 Inline constants into the code stream if it can be done in two
8777 instructions or less.
8785 Use the divide instruction. (Enabled by default).
8787 @item -mrelax-immediate
8788 @itemx -mrelax-immediate
8789 @itemx -mno-relax-immediate
8790 @opindex mrelax-immediate
8791 @opindex mrelax-immediate
8792 @opindex mno-relax-immediate
8793 Allow arbitrary sized immediates in bit operations.
8795 @item -mwide-bitfields
8796 @itemx -mwide-bitfields
8797 @itemx -mno-wide-bitfields
8798 @opindex mwide-bitfields
8799 @opindex mwide-bitfields
8800 @opindex mno-wide-bitfields
8801 Always treat bit-fields as int-sized.
8803 @item -m4byte-functions
8804 @itemx -m4byte-functions
8805 @itemx -mno-4byte-functions
8806 @opindex m4byte-functions
8807 @opindex m4byte-functions
8808 @opindex mno-4byte-functions
8809 Force all functions to be aligned to a four byte boundary.
8811 @item -mcallgraph-data
8812 @itemx -mcallgraph-data
8813 @itemx -mno-callgraph-data
8814 @opindex mcallgraph-data
8815 @opindex mcallgraph-data
8816 @opindex mno-callgraph-data
8817 Emit callgraph information.
8821 @itemx -mno-slow-bytes
8822 @opindex mslow-bytes
8823 @opindex mslow-bytes
8824 @opindex mno-slow-bytes
8825 Prefer word access when reading byte quantities.
8827 @item -mlittle-endian
8828 @itemx -mlittle-endian
8830 @opindex mlittle-endian
8831 @opindex mlittle-endian
8832 @opindex mbig-endian
8833 Generate code for a little endian target.
8841 Generate code for the 210 processor.
8845 @subsection IA-64 Options
8846 @cindex IA-64 Options
8848 These are the @samp{-m} options defined for the Intel IA-64 architecture.
8852 @opindex mbig-endian
8853 Generate code for a big endian target. This is the default for HPUX@.
8855 @item -mlittle-endian
8856 @opindex mlittle-endian
8857 Generate code for a little endian target. This is the default for AIX5
8864 Generate (or don't) code for the GNU assembler. This is the default.
8865 @c Also, this is the default if the configure option @option{--with-gnu-as}
8872 Generate (or don't) code for the GNU linker. This is the default.
8873 @c Also, this is the default if the configure option @option{--with-gnu-ld}
8878 Generate code that does not use a global pointer register. The result
8879 is not position independent code, and violates the IA-64 ABI@.
8881 @item -mvolatile-asm-stop
8882 @itemx -mno-volatile-asm-stop
8883 @opindex mvolatile-asm-stop
8884 @opindex mno-volatile-asm-stop
8885 Generate (or don't) a stop bit immediately before and after volatile asm
8890 Generate code that works around Itanium B step errata.
8892 @item -mregister-names
8893 @itemx -mno-register-names
8894 @opindex mregister-names
8895 @opindex mno-register-names
8896 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
8897 the stacked registers. This may make assembler output more readable.
8903 Disable (or enable) optimizations that use the small data section. This may
8904 be useful for working around optimizer bugs.
8907 @opindex mconstant-gp
8908 Generate code that uses a single constant global pointer value. This is
8909 useful when compiling kernel code.
8913 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
8914 This is useful when compiling firmware code.
8916 @item -minline-divide-min-latency
8917 @opindex minline-divide-min-latency
8918 Generate code for inline divides using the minimum latency algorithm.
8920 @item -minline-divide-max-throughput
8921 @opindex minline-divide-max-throughput
8922 Generate code for inline divides using the maximum throughput algorithm.
8924 @item -mno-dwarf2-asm
8926 @opindex mno-dwarf2-asm
8927 @opindex mdwarf2-asm
8928 Don't (or do) generate assembler code for the DWARF2 line number debugging
8929 info. This may be useful when not using the GNU assembler.
8931 @item -mfixed-range=@var{register-range}
8932 @opindex mfixed-range
8933 Generate code treating the given register range as fixed registers.
8934 A fixed register is one that the register allocator can not use. This is
8935 useful when compiling kernel code. A register range is specified as
8936 two registers separated by a dash. Multiple register ranges can be
8937 specified separated by a comma.
8941 @subsection D30V Options
8942 @cindex D30V Options
8944 These @samp{-m} options are defined for D30V implementations:
8949 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
8950 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
8951 memory, which starts at location @code{0x80000000}.
8955 Same as the @option{-mextmem} switch.
8959 Link the @samp{.text} section into onchip text memory, which starts at
8960 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
8961 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
8962 into onchip data memory, which starts at location @code{0x20000000}.
8964 @item -mno-asm-optimize
8965 @itemx -masm-optimize
8966 @opindex mno-asm-optimize
8967 @opindex masm-optimize
8968 Disable (enable) passing @option{-O} to the assembler when optimizing.
8969 The assembler uses the @option{-O} option to automatically parallelize
8970 adjacent short instructions where possible.
8972 @item -mbranch-cost=@var{n}
8973 @opindex mbranch-cost
8974 Increase the internal costs of branches to @var{n}. Higher costs means
8975 that the compiler will issue more instructions to avoid doing a branch.
8978 @item -mcond-exec=@var{n}
8980 Specify the maximum number of conditionally executed instructions that
8981 replace a branch. The default is 4.
8984 @node S/390 and zSeries Options
8985 @subsection S/390 and zSeries Options
8986 @cindex S/390 and zSeries Options
8988 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
8993 @opindex mhard-float
8994 @opindex msoft-float
8995 Use (do not use) the hardware floating-point instructions and registers
8996 for floating-point operations. When @option{-msoft-float} is specified,
8997 functions in @file{libgcc.a} will be used to perform floating-point
8998 operations. When @option{-mhard-float} is specified, the compiler
8999 generates IEEE floating-point instructions. This is the default.
9002 @itemx -mno-backchain
9004 @opindex mno-backchain
9005 Generate (or do not generate) code which maintains an explicit
9006 backchain within the stack frame that points to the caller's frame.
9007 This is currently needed to allow debugging. The default is to
9008 generate the backchain.
9011 @itemx -mno-small-exec
9012 @opindex msmall-exec
9013 @opindex mno-small-exec
9014 Generate (or do not generate) code using the @code{bras} instruction
9015 to do subroutine calls.
9016 This only works reliably if the total executable size does not
9017 exceed 64k. The default is to use the @code{basr} instruction instead,
9018 which does not have this limitation.
9024 When @option{-m31} is specified, generate code compliant to the
9025 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9026 code compliant to the Linux for zSeries ABI@. This allows GCC in
9027 particular to generate 64-bit instructions. For the @samp{s390}
9028 targets, the default is @option{-m31}, while the @samp{s390x}
9029 targets default to @option{-m64}.
9035 Generate (or do not generate) code using the @code{mvcle} instruction
9036 to perform block moves. When @option{-mno-mvcle} is specifed,
9037 use a @code{mvc} loop instead. This is the default.
9043 Print (or do not print) additional debug information when compiling.
9044 The default is to not print debug information.
9049 @subsection CRIS Options
9050 @cindex CRIS Options
9052 These options are defined specifically for the CRIS ports.
9055 @item -march=@var{architecture-type}
9056 @itemx -mcpu=@var{architecture-type}
9059 Generate code for the specified architecture. The choices for
9060 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9061 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9062 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9065 @item -mtune=@var{architecture-type}
9067 Tune to @var{architecture-type} everything applicable about the generated
9068 code, except for the ABI and the set of available instructions. The
9069 choices for @var{architecture-type} are the same as for
9070 @option{-march=@var{architecture-type}}.
9072 @item -mmax-stack-frame=@var{n}
9073 @opindex mmax-stack-frame
9074 Warn when the stack frame of a function exceeds @var{n} bytes.
9076 @item -melinux-stacksize=@var{n}
9077 @opindex melinux-stacksize
9078 Only available with the @samp{cris-axis-aout} target. Arranges for
9079 indications in the program to the kernel loader that the stack of the
9080 program should be set to @var{n} bytes.
9086 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9087 @option{-march=v3} and @option{-march=v8} respectively.
9091 Enable CRIS-specific verbose debug-related information in the assembly
9092 code. This option also has the effect to turn off the @samp{#NO_APP}
9093 formatted-code indicator to the assembler at the beginning of the
9098 Do not use condition-code results from previous instruction; always emit
9099 compare and test instructions before use of condition codes.
9101 @item -mno-side-effects
9102 @opindex mno-side-effects
9103 Do not emit instructions with side-effects in addressing modes other than
9107 @itemx -mno-stack-align
9109 @itemx -mno-data-align
9110 @itemx -mconst-align
9111 @itemx -mno-const-align
9112 @opindex mstack-align
9113 @opindex mno-stack-align
9114 @opindex mdata-align
9115 @opindex mno-data-align
9116 @opindex mconst-align
9117 @opindex mno-const-align
9118 These options (no-options) arranges (eliminate arrangements) for the
9119 stack-frame, individual data and constants to be aligned for the maximum
9120 single data access size for the chosen CPU model. The default is to
9121 arrange for 32-bit alignment. ABI details such as structure layout are
9122 not affected by these options.
9130 Similar to the stack- data- and const-align options above, these options
9131 arrange for stack-frame, writable data and constants to all be 32-bit,
9132 16-bit or 8-bit aligned. The default is 32-bit alignment.
9134 @item -mno-prologue-epilogue
9135 @itemx -mprologue-epilogue
9136 @opindex mno-prologue-epilogue
9137 @opindex mprologue-epilogue
9138 With @option{-mno-prologue-epilogue}, the normal function prologue and
9139 epilogue that sets up the stack-frame are omitted and no return
9140 instructions or return sequences are generated in the code. Use this
9141 option only together with visual inspection of the compiled code: no
9142 warnings or errors are generated when call-saved registers must be saved,
9143 or storage for local variable needs to be allocated.
9149 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9150 instruction sequences that load addresses for functions from the PLT part
9151 of the GOT rather than (traditional on other architectures) calls to the
9152 PLT. The default is @option{-mgotplt}.
9156 Legacy no-op option only recognized with the cris-axis-aout target.
9160 Legacy no-op option only recognized with the cris-axis-elf and
9161 cris-axis-linux-gnu targets.
9165 Only recognized with the cris-axis-aout target, where it selects a
9166 GNU/linux-like multilib, include files and instruction set for
9171 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9175 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9176 to link with input-output functions from a simulator library. Code,
9177 initialized data and zero-initialized data are allocated consecutively.
9181 Like @option{-sim}, but pass linker options to locate initialized data at
9182 0x40000000 and zero-initialized data at 0x80000000.
9186 @subsection MMIX Options
9187 @cindex MMIX Options
9189 These options are defined for the MMIX:
9193 @itemx -mno-libfuncs
9195 @opindex mno-libfuncs
9196 Specify that intrinsic library functions are being compiled, passing all
9197 values in registers, no matter the size.
9202 @opindex mno-epsilon
9203 Generate floating-point comparison instructions that compare with respect
9204 to the @code{rE} epsilon register.
9206 @item -mabi=mmixware
9208 @opindex mabi-mmixware
9210 Generate code that passes function parameters and return values that (in
9211 the called function) are seen as registers @code{$0} and up, as opposed to
9212 the GNU ABI which uses global registers @code{$231} and up.
9215 @itemx -mno-zero-extend
9216 @opindex mzero-extend
9217 @opindex mno-zero-extend
9218 When reading data from memory in sizes shorter than 64 bits, use (do not
9219 use) zero-extending load instructions by default, rather than
9220 sign-extending ones.
9223 @itemx -mno-knuthdiv
9225 @opindex mno-knuthdiv
9226 Make the result of a division yielding a remainder have the same sign as
9227 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9228 remainder follows the sign of the dividend. Both methods are
9229 arithmetically valid, the latter being almost exclusively used.
9231 @item -mtoplevel-symbols
9232 @itemx -mno-toplevel-symbols
9233 @opindex mtoplevel-symbols
9234 @opindex mno-toplevel-symbols
9235 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9236 code can be used with the @code{PREFIX} assembly directive.
9240 Generate an executable in the ELF format, rather than the default
9241 @samp{mmo} format used by the @command{mmix} simulator.
9243 @item -mbranch-predict
9244 @itemx -mno-branch-predict
9245 @opindex mbranch-predict
9246 @opindex mno-branch-predict
9247 Use (do not use) the probable-branch instructions, when static branch
9248 prediction indicates a probable branch.
9250 @item -mbase-addresses
9251 @itemx -mno-base-addresses
9252 @opindex mbase-addresses
9253 @opindex mno-base-addresses
9254 Generate (do not generate) code that uses @emph{base addresses}. Using a
9255 base address automatically generates a request (handled by the assembler
9256 and the linker) for a constant to be set up in a global register. The
9257 register is used for one or more base address requests within the range 0
9258 to 255 from the value held in the register. The generally leads to short
9259 and fast code, but the number of different data items that can be
9260 addressed is limited. This means that a program that uses lots of static
9261 data may require @option{-mno-base-addresses}.
9264 @node PDP-11 Options
9265 @subsection PDP-11 Options
9266 @cindex PDP-11 Options
9268 These options are defined for the PDP-11:
9273 Use hardware FPP floating point. This is the default. (FIS floating
9274 point on the PDP-11/40 is not supported.)
9277 @opindex msoft-float
9278 Do not use hardware floating point.
9282 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9286 Return floating-point results in memory. This is the default.
9290 Generate code for a PDP-11/40.
9294 Generate code for a PDP-11/45. This is the default.
9298 Generate code for a PDP-11/10.
9300 @item -mbcopy-builtin
9301 @opindex bcopy-builtin
9302 Use inline @code{movstrhi} patterns for copying memory. This is the
9307 Do not use inline @code{movstrhi} patterns for copying memory.
9313 Use 16-bit @code{int}. This is the default.
9319 Use 32-bit @code{int}.
9324 @opindex mno-float32
9325 Use 64-bit @code{float}. This is the default.
9330 @opindex mno-float64
9331 Use 32-bit @code{float}.
9335 Use @code{abshi2} pattern. This is the default.
9339 Do not use @code{abshi2} pattern.
9341 @item -mbranch-expensive
9342 @opindex mbranch-expensive
9343 Pretend that branches are expensive. This is for experimenting with
9344 code generation only.
9346 @item -mbranch-cheap
9347 @opindex mbranch-cheap
9348 Do not pretend that branches are expensive. This is the default.
9352 Generate code for a system with split I&D.
9356 Generate code for a system without split I&D. This is the default.
9360 Use Unix assembler syntax. This is the default when configured for
9365 Use DEC assembler syntax. This is the default when configured for any
9366 PDP-11 target other than @samp{pdp11-*-bsd}.
9369 @node Xstormy16 Options
9370 @subsection Xstormy16 Options
9371 @cindex Xstormy16 Options
9373 These options are defined for Xstormy16:
9378 Choose startup files and linker script suitable for the simulator.
9381 @node Xtensa Options
9382 @subsection Xtensa Options
9383 @cindex Xtensa Options
9385 The Xtensa architecture is designed to support many different
9386 configurations. The compiler's default options can be set to match a
9387 particular Xtensa configuration by copying a configuration file into the
9388 GCC sources when building GCC@. The options below may be used to
9389 override the default options.
9393 @itemx -mlittle-endian
9394 @opindex mbig-endian
9395 @opindex mlittle-endian
9396 Specify big-endian or little-endian byte ordering for the target Xtensa
9402 @opindex mno-density
9403 Enable or disable use of the optional Xtensa code density instructions.
9409 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
9410 will generate MAC16 instructions from standard C code, with the
9411 limitation that it will use neither the MR register file nor any
9412 instruction that operates on the MR registers. When this option is
9413 disabled, GCC will translate 16-bit multiply/accumulate operations to a
9414 combination of core instructions and library calls, depending on whether
9415 any other multiplier options are enabled.
9421 Enable or disable use of the 16-bit integer multiplier option. When
9422 enabled, the compiler will generate 16-bit multiply instructions for
9423 multiplications of 16 bits or smaller in standard C code. When this
9424 option is disabled, the compiler will either use 32-bit multiply or
9425 MAC16 instructions if they are available or generate library calls to
9426 perform the multiply operations using shifts and adds.
9432 Enable or disable use of the 32-bit integer multiplier option. When
9433 enabled, the compiler will generate 32-bit multiply instructions for
9434 multiplications of 32 bits or smaller in standard C code. When this
9435 option is disabled, the compiler will generate library calls to perform
9436 the multiply operations using either shifts and adds or 16-bit multiply
9437 instructions if they are available.
9443 Enable or disable use of the optional normalization shift amount
9444 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
9450 Enable or disable use of the optional minimum and maximum value
9457 Enable or disable use of the optional sign extend (@code{SEXT})
9461 @itemx -mno-booleans
9463 @opindex mno-booleans
9464 Enable or disable support for the boolean register file used by Xtensa
9465 coprocessors. This is not typically useful by itself but may be
9466 required for other options that make use of the boolean registers (e.g.,
9467 the floating-point option).
9471 @opindex mhard-float
9472 @opindex msoft-float
9473 Enable or disable use of the floating-point option. When enabled, GCC
9474 generates floating-point instructions for 32-bit @code{float}
9475 operations. When this option is disabled, GCC generates library calls
9476 to emulate 32-bit floating-point operations using integer instructions.
9477 Regardless of this option, 64-bit @code{double} operations are always
9478 emulated with calls to library functions.
9481 @itemx -mno-fused-madd
9482 @opindex mfused-madd
9483 @opindex mno-fused-madd
9484 Enable or disable use of fused multiply/add and multiply/subtract
9485 instructions in the floating-point option. This has no effect if the
9486 floating-point option is not also enabled. Disabling fused multiply/add
9487 and multiply/subtract instructions forces the compiler to use separate
9488 instructions for the multiply and add/subtract operations. This may be
9489 desirable in some cases where strict IEEE 754-compliant results are
9490 required: the fused multiply add/subtract instructions do not round the
9491 intermediate result, thereby producing results with @emph{more} bits of
9492 precision than specified by the IEEE standard. Disabling fused multiply
9493 add/subtract instructions also ensures that the program output is not
9494 sensitive to the compiler's ability to combine multiply and add/subtract
9497 @item -mserialize-volatile
9498 @itemx -mno-serialize-volatile
9499 @opindex mserialize-volatile
9500 @opindex mno-serialize-volatile
9501 When this option is enabled, GCC inserts @code{MEMW} instructions before
9502 @code{volatile} memory references to guarantee sequential consistency.
9503 The default is @option{-mserialize-volatile}. Use
9504 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
9506 @item -mtext-section-literals
9507 @itemx -mno-text-section-literals
9508 @opindex mtext-section-literals
9509 @opindex mno-text-section-literals
9510 Control the treatment of literal pools. The default is
9511 @option{-mno-text-section-literals}, which places literals in a separate
9512 section in the output file. This allows the literal pool to be placed
9513 in a data RAM/ROM, and it also allows the linker to combine literal
9514 pools from separate object files to remove redundant literals and
9515 improve code size. With @option{-mtext-section-literals}, the literals
9516 are interspersed in the text section in order to keep them as close as
9517 possible to their references. This may be necessary for large assembly
9520 @item -mtarget-align
9521 @itemx -mno-target-align
9522 @opindex mtarget-align
9523 @opindex mno-target-align
9524 When this option is enabled, GCC instructs the assembler to
9525 automatically align instructions to reduce branch penalties at the
9526 expense of some code density. The assembler attempts to widen density
9527 instructions to align branch targets and the instructions following call
9528 instructions. If there are not enough preceding safe density
9529 instructions to align a target, no widening will be performed. The
9530 default is @option{-mtarget-align}. These options do not affect the
9531 treatment of auto-aligned instructions like @code{LOOP}, which the
9532 assembler will always align, either by widening density instructions or
9533 by inserting no-op instructions.
9536 @itemx -mno-longcalls
9538 @opindex mno-longcalls
9539 When this option is enabled, GCC instructs the assembler to translate
9540 direct calls to indirect calls unless it can determine that the target
9541 of a direct call is in the range allowed by the call instruction. This
9542 translation typically occurs for calls to functions in other source
9543 files. Specifically, the assembler translates a direct @code{CALL}
9544 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
9545 The default is @option{-mno-longcalls}. This option should be used in
9546 programs where the call target can potentially be out of range. This
9547 option is implemented in the assembler, not the compiler, so the
9548 assembly code generated by GCC will still show direct call
9549 instructions---look at the disassembled object code to see the actual
9550 instructions. Note that the assembler will use an indirect call for
9551 every cross-file call, not just those that really will be out of range.
9554 @node Code Gen Options
9555 @section Options for Code Generation Conventions
9556 @cindex code generation conventions
9557 @cindex options, code generation
9558 @cindex run-time options
9560 These machine-independent options control the interface conventions
9561 used in code generation.
9563 Most of them have both positive and negative forms; the negative form
9564 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
9565 one of the forms is listed---the one which is not the default. You
9566 can figure out the other form by either removing @samp{no-} or adding
9571 @opindex fexceptions
9572 Enable exception handling. Generates extra code needed to propagate
9573 exceptions. For some targets, this implies GCC will generate frame
9574 unwind information for all functions, which can produce significant data
9575 size overhead, although it does not affect execution. If you do not
9576 specify this option, GCC will enable it by default for languages like
9577 C++ which normally require exception handling, and disable it for
9578 languages like C that do not normally require it. However, you may need
9579 to enable this option when compiling C code that needs to interoperate
9580 properly with exception handlers written in C++. You may also wish to
9581 disable this option if you are compiling older C++ programs that don't
9582 use exception handling.
9584 @item -fnon-call-exceptions
9585 @opindex fnon-call-exceptions
9586 Generate code that allows trapping instructions to throw exceptions.
9587 Note that this requires platform-specific runtime support that does
9588 not exist everywhere. Moreover, it only allows @emph{trapping}
9589 instructions to throw exceptions, i.e.@: memory references or floating
9590 point instructions. It does not allow exceptions to be thrown from
9591 arbitrary signal handlers such as @code{SIGALRM}.
9593 @item -funwind-tables
9594 @opindex funwind-tables
9595 Similar to @option{-fexceptions}, except that it will just generate any needed
9596 static data, but will not affect the generated code in any other way.
9597 You will normally not enable this option; instead, a language processor
9598 that needs this handling would enable it on your behalf.
9600 @item -fasynchronous-unwind-tables
9601 @opindex funwind-tables
9602 Generate unwind table in dwarf2 format, if supported by target machine. The
9603 table is exact at each instruction boundary, so it can be used for stack
9604 unwinding from asynchronous events (such as debugger or garbage collector).
9606 @item -fpcc-struct-return
9607 @opindex fpcc-struct-return
9608 Return ``short'' @code{struct} and @code{union} values in memory like
9609 longer ones, rather than in registers. This convention is less
9610 efficient, but it has the advantage of allowing intercallability between
9611 GCC-compiled files and files compiled with other compilers.
9613 The precise convention for returning structures in memory depends
9614 on the target configuration macros.
9616 Short structures and unions are those whose size and alignment match
9617 that of some integer type.
9619 @item -freg-struct-return
9620 @opindex freg-struct-return
9621 Return @code{struct} and @code{union} values in registers when possible.
9622 This is more efficient for small structures than
9623 @option{-fpcc-struct-return}.
9625 If you specify neither @option{-fpcc-struct-return} nor
9626 @option{-freg-struct-return}, GCC defaults to whichever convention is
9627 standard for the target. If there is no standard convention, GCC
9628 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
9629 the principal compiler. In those cases, we can choose the standard, and
9630 we chose the more efficient register return alternative.
9633 @opindex fshort-enums
9634 Allocate to an @code{enum} type only as many bytes as it needs for the
9635 declared range of possible values. Specifically, the @code{enum} type
9636 will be equivalent to the smallest integer type which has enough room.
9638 @item -fshort-double
9639 @opindex fshort-double
9640 Use the same size for @code{double} as for @code{float}.
9643 @opindex fshared-data
9644 Requests that the data and non-@code{const} variables of this
9645 compilation be shared data rather than private data. The distinction
9646 makes sense only on certain operating systems, where shared data is
9647 shared between processes running the same program, while private data
9648 exists in one copy per process.
9652 In C, allocate even uninitialized global variables in the data section of the
9653 object file, rather than generating them as common blocks. This has the
9654 effect that if the same variable is declared (without @code{extern}) in
9655 two different compilations, you will get an error when you link them.
9656 The only reason this might be useful is if you wish to verify that the
9657 program will work on other systems which always work this way.
9661 Ignore the @samp{#ident} directive.
9663 @item -fno-gnu-linker
9664 @opindex fno-gnu-linker
9665 Do not output global initializations (such as C++ constructors and
9666 destructors) in the form used by the GNU linker (on systems where the GNU
9667 linker is the standard method of handling them). Use this option when
9668 you want to use a non-GNU linker, which also requires using the
9669 @command{collect2} program to make sure the system linker includes
9670 constructors and destructors. (@command{collect2} is included in the GCC
9671 distribution.) For systems which @emph{must} use @command{collect2}, the
9672 compiler driver @command{gcc} is configured to do this automatically.
9674 @item -finhibit-size-directive
9675 @opindex finhibit-size-directive
9676 Don't output a @code{.size} assembler directive, or anything else that
9677 would cause trouble if the function is split in the middle, and the
9678 two halves are placed at locations far apart in memory. This option is
9679 used when compiling @file{crtstuff.c}; you should not need to use it
9683 @opindex fverbose-asm
9684 Put extra commentary information in the generated assembly code to
9685 make it more readable. This option is generally only of use to those
9686 who actually need to read the generated assembly code (perhaps while
9687 debugging the compiler itself).
9689 @option{-fno-verbose-asm}, the default, causes the
9690 extra information to be omitted and is useful when comparing two assembler
9695 Consider all memory references through pointers to be volatile.
9697 @item -fvolatile-global
9698 @opindex fvolatile-global
9699 Consider all memory references to extern and global data items to
9700 be volatile. GCC does not consider static data items to be volatile
9701 because of this switch.
9703 @item -fvolatile-static
9704 @opindex fvolatile-static
9705 Consider all memory references to static data to be volatile.
9709 @cindex global offset table
9711 Generate position-independent code (PIC) suitable for use in a shared
9712 library, if supported for the target machine. Such code accesses all
9713 constant addresses through a global offset table (GOT)@. The dynamic
9714 loader resolves the GOT entries when the program starts (the dynamic
9715 loader is not part of GCC; it is part of the operating system). If
9716 the GOT size for the linked executable exceeds a machine-specific
9717 maximum size, you get an error message from the linker indicating that
9718 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
9719 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
9720 on the m68k and RS/6000. The 386 has no such limit.)
9722 Position-independent code requires special support, and therefore works
9723 only on certain machines. For the 386, GCC supports PIC for System V
9724 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
9725 position-independent.
9729 If supported for the target machine, emit position-independent code,
9730 suitable for dynamic linking and avoiding any limit on the size of the
9731 global offset table. This option makes a difference on the m68k, m88k,
9734 Position-independent code requires special support, and therefore works
9735 only on certain machines.
9737 @item -ffixed-@var{reg}
9739 Treat the register named @var{reg} as a fixed register; generated code
9740 should never refer to it (except perhaps as a stack pointer, frame
9741 pointer or in some other fixed role).
9743 @var{reg} must be the name of a register. The register names accepted
9744 are machine-specific and are defined in the @code{REGISTER_NAMES}
9745 macro in the machine description macro file.
9747 This flag does not have a negative form, because it specifies a
9750 @item -fcall-used-@var{reg}
9752 Treat the register named @var{reg} as an allocable register that is
9753 clobbered by function calls. It may be allocated for temporaries or
9754 variables that do not live across a call. Functions compiled this way
9755 will not save and restore the register @var{reg}.
9757 It is an error to used this flag with the frame pointer or stack pointer.
9758 Use of this flag for other registers that have fixed pervasive roles in
9759 the machine's execution model will produce disastrous results.
9761 This flag does not have a negative form, because it specifies a
9764 @item -fcall-saved-@var{reg}
9765 @opindex fcall-saved
9766 Treat the register named @var{reg} as an allocable register saved by
9767 functions. It may be allocated even for temporaries or variables that
9768 live across a call. Functions compiled this way will save and restore
9769 the register @var{reg} if they use it.
9771 It is an error to used this flag with the frame pointer or stack pointer.
9772 Use of this flag for other registers that have fixed pervasive roles in
9773 the machine's execution model will produce disastrous results.
9775 A different sort of disaster will result from the use of this flag for
9776 a register in which function values may be returned.
9778 This flag does not have a negative form, because it specifies a
9782 @opindex fpack-struct
9783 Pack all structure members together without holes. Usually you would
9784 not want to use this option, since it makes the code suboptimal, and
9785 the offsets of structure members won't agree with system libraries.
9787 @item -finstrument-functions
9788 @opindex finstrument-functions
9789 Generate instrumentation calls for entry and exit to functions. Just
9790 after function entry and just before function exit, the following
9791 profiling functions will be called with the address of the current
9792 function and its call site. (On some platforms,
9793 @code{__builtin_return_address} does not work beyond the current
9794 function, so the call site information may not be available to the
9795 profiling functions otherwise.)
9798 void __cyg_profile_func_enter (void *this_fn,
9800 void __cyg_profile_func_exit (void *this_fn,
9804 The first argument is the address of the start of the current function,
9805 which may be looked up exactly in the symbol table.
9807 This instrumentation is also done for functions expanded inline in other
9808 functions. The profiling calls will indicate where, conceptually, the
9809 inline function is entered and exited. This means that addressable
9810 versions of such functions must be available. If all your uses of a
9811 function are expanded inline, this may mean an additional expansion of
9812 code size. If you use @samp{extern inline} in your C code, an
9813 addressable version of such functions must be provided. (This is
9814 normally the case anyways, but if you get lucky and the optimizer always
9815 expands the functions inline, you might have gotten away without
9816 providing static copies.)
9818 A function may be given the attribute @code{no_instrument_function}, in
9819 which case this instrumentation will not be done. This can be used, for
9820 example, for the profiling functions listed above, high-priority
9821 interrupt routines, and any functions from which the profiling functions
9822 cannot safely be called (perhaps signal handlers, if the profiling
9823 routines generate output or allocate memory).
9826 @opindex fstack-check
9827 Generate code to verify that you do not go beyond the boundary of the
9828 stack. You should specify this flag if you are running in an
9829 environment with multiple threads, but only rarely need to specify it in
9830 a single-threaded environment since stack overflow is automatically
9831 detected on nearly all systems if there is only one stack.
9833 Note that this switch does not actually cause checking to be done; the
9834 operating system must do that. The switch causes generation of code
9835 to ensure that the operating system sees the stack being extended.
9837 @item -fstack-limit-register=@var{reg}
9838 @itemx -fstack-limit-symbol=@var{sym}
9839 @itemx -fno-stack-limit
9840 @opindex fstack-limit-register
9841 @opindex fstack-limit-symbol
9842 @opindex fno-stack-limit
9843 Generate code to ensure that the stack does not grow beyond a certain value,
9844 either the value of a register or the address of a symbol. If the stack
9845 would grow beyond the value, a signal is raised. For most targets,
9846 the signal is raised before the stack overruns the boundary, so
9847 it is possible to catch the signal without taking special precautions.
9849 For instance, if the stack starts at absolute address @samp{0x80000000}
9850 and grows downwards, you can use the flags
9851 @option{-fstack-limit-symbol=__stack_limit} and
9852 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
9853 of 128KB@. Note that this may only work with the GNU linker.
9855 @cindex aliasing of parameters
9856 @cindex parameters, aliased
9857 @item -fargument-alias
9858 @itemx -fargument-noalias
9859 @itemx -fargument-noalias-global
9860 @opindex fargument-alias
9861 @opindex fargument-noalias
9862 @opindex fargument-noalias-global
9863 Specify the possible relationships among parameters and between
9864 parameters and global data.
9866 @option{-fargument-alias} specifies that arguments (parameters) may
9867 alias each other and may alias global storage.@*
9868 @option{-fargument-noalias} specifies that arguments do not alias
9869 each other, but may alias global storage.@*
9870 @option{-fargument-noalias-global} specifies that arguments do not
9871 alias each other and do not alias global storage.
9873 Each language will automatically use whatever option is required by
9874 the language standard. You should not need to use these options yourself.
9876 @item -fleading-underscore
9877 @opindex fleading-underscore
9878 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
9879 change the way C symbols are represented in the object file. One use
9880 is to help link with legacy assembly code.
9882 Be warned that you should know what you are doing when invoking this
9883 option, and that not all targets provide complete support for it.
9888 @node Environment Variables
9889 @section Environment Variables Affecting GCC
9890 @cindex environment variables
9892 @c man begin ENVIRONMENT
9894 This section describes several environment variables that affect how GCC
9895 operates. Some of them work by specifying directories or prefixes to use
9896 when searching for various kinds of files. Some are used to specify other
9897 aspects of the compilation environment.
9899 Note that you can also specify places to search using options such as
9900 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
9901 take precedence over places specified using environment variables, which
9902 in turn take precedence over those specified by the configuration of GCC@.
9903 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
9904 GNU Compiler Collection (GCC) Internals}.
9909 @c @itemx LC_COLLATE
9911 @c @itemx LC_MONETARY
9912 @c @itemx LC_NUMERIC
9917 @c @findex LC_COLLATE
9919 @c @findex LC_MONETARY
9920 @c @findex LC_NUMERIC
9924 These environment variables control the way that GCC uses
9925 localization information that allow GCC to work with different
9926 national conventions. GCC inspects the locale categories
9927 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
9928 so. These locale categories can be set to any value supported by your
9929 installation. A typical value is @samp{en_UK} for English in the United
9932 The @env{LC_CTYPE} environment variable specifies character
9933 classification. GCC uses it to determine the character boundaries in
9934 a string; this is needed for some multibyte encodings that contain quote
9935 and escape characters that would otherwise be interpreted as a string
9938 The @env{LC_MESSAGES} environment variable specifies the language to
9939 use in diagnostic messages.
9941 If the @env{LC_ALL} environment variable is set, it overrides the value
9942 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
9943 and @env{LC_MESSAGES} default to the value of the @env{LANG}
9944 environment variable. If none of these variables are set, GCC
9945 defaults to traditional C English behavior.
9949 If @env{TMPDIR} is set, it specifies the directory to use for temporary
9950 files. GCC uses temporary files to hold the output of one stage of
9951 compilation which is to be used as input to the next stage: for example,
9952 the output of the preprocessor, which is the input to the compiler
9955 @item GCC_EXEC_PREFIX
9956 @findex GCC_EXEC_PREFIX
9957 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
9958 names of the subprograms executed by the compiler. No slash is added
9959 when this prefix is combined with the name of a subprogram, but you can
9960 specify a prefix that ends with a slash if you wish.
9962 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
9963 an appropriate prefix to use based on the pathname it was invoked with.
9965 If GCC cannot find the subprogram using the specified prefix, it
9966 tries looking in the usual places for the subprogram.
9968 The default value of @env{GCC_EXEC_PREFIX} is
9969 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
9970 of @code{prefix} when you ran the @file{configure} script.
9972 Other prefixes specified with @option{-B} take precedence over this prefix.
9974 This prefix is also used for finding files such as @file{crt0.o} that are
9977 In addition, the prefix is used in an unusual way in finding the
9978 directories to search for header files. For each of the standard
9979 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
9980 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
9981 replacing that beginning with the specified prefix to produce an
9982 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
9983 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
9984 These alternate directories are searched first; the standard directories
9988 @findex COMPILER_PATH
9989 The value of @env{COMPILER_PATH} is a colon-separated list of
9990 directories, much like @env{PATH}. GCC tries the directories thus
9991 specified when searching for subprograms, if it can't find the
9992 subprograms using @env{GCC_EXEC_PREFIX}.
9995 @findex LIBRARY_PATH
9996 The value of @env{LIBRARY_PATH} is a colon-separated list of
9997 directories, much like @env{PATH}. When configured as a native compiler,
9998 GCC tries the directories thus specified when searching for special
9999 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10000 using GCC also uses these directories when searching for ordinary
10001 libraries for the @option{-l} option (but directories specified with
10002 @option{-L} come first).
10006 @cindex locale definition
10007 This variable is used to pass locale information to the compiler. One way in
10008 which this information is used is to determine the character set to be used
10009 when character literals, string literals and comments are parsed in C and C++.
10010 When the compiler is configured to allow multibyte characters,
10011 the following values for @env{LANG} are recognized:
10015 Recognize JIS characters.
10017 Recognize SJIS characters.
10019 Recognize EUCJP characters.
10022 If @env{LANG} is not defined, or if it has some other value, then the
10023 compiler will use mblen and mbtowc as defined by the default locale to
10024 recognize and translate multibyte characters.
10028 Some additional environments variables affect the behavior of the
10031 @include cppenv.texi
10035 @node Running Protoize
10036 @section Running Protoize
10038 The program @code{protoize} is an optional part of GCC@. You can use
10039 it to add prototypes to a program, thus converting the program to ISO
10040 C in one respect. The companion program @code{unprotoize} does the
10041 reverse: it removes argument types from any prototypes that are found.
10043 When you run these programs, you must specify a set of source files as
10044 command line arguments. The conversion programs start out by compiling
10045 these files to see what functions they define. The information gathered
10046 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
10048 After scanning comes actual conversion. The specified files are all
10049 eligible to be converted; any files they include (whether sources or
10050 just headers) are eligible as well.
10052 But not all the eligible files are converted. By default,
10053 @code{protoize} and @code{unprotoize} convert only source and header
10054 files in the current directory. You can specify additional directories
10055 whose files should be converted with the @option{-d @var{directory}}
10056 option. You can also specify particular files to exclude with the
10057 @option{-x @var{file}} option. A file is converted if it is eligible, its
10058 directory name matches one of the specified directory names, and its
10059 name within the directory has not been excluded.
10061 Basic conversion with @code{protoize} consists of rewriting most
10062 function definitions and function declarations to specify the types of
10063 the arguments. The only ones not rewritten are those for varargs
10066 @code{protoize} optionally inserts prototype declarations at the
10067 beginning of the source file, to make them available for any calls that
10068 precede the function's definition. Or it can insert prototype
10069 declarations with block scope in the blocks where undeclared functions
10072 Basic conversion with @code{unprotoize} consists of rewriting most
10073 function declarations to remove any argument types, and rewriting
10074 function definitions to the old-style pre-ISO form.
10076 Both conversion programs print a warning for any function declaration or
10077 definition that they can't convert. You can suppress these warnings
10080 The output from @code{protoize} or @code{unprotoize} replaces the
10081 original source file. The original file is renamed to a name ending
10082 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
10083 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
10084 for DOS) file already exists, then the source file is simply discarded.
10086 @code{protoize} and @code{unprotoize} both depend on GCC itself to
10087 scan the program and collect information about the functions it uses.
10088 So neither of these programs will work until GCC is installed.
10090 Here is a table of the options you can use with @code{protoize} and
10091 @code{unprotoize}. Each option works with both programs unless
10095 @item -B @var{directory}
10096 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
10097 usual directory (normally @file{/usr/local/lib}). This file contains
10098 prototype information about standard system functions. This option
10099 applies only to @code{protoize}.
10101 @item -c @var{compilation-options}
10102 Use @var{compilation-options} as the options when running @code{gcc} to
10103 produce the @samp{.X} files. The special option @option{-aux-info} is
10104 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
10106 Note that the compilation options must be given as a single argument to
10107 @code{protoize} or @code{unprotoize}. If you want to specify several
10108 @code{gcc} options, you must quote the entire set of compilation options
10109 to make them a single word in the shell.
10111 There are certain @code{gcc} arguments that you cannot use, because they
10112 would produce the wrong kind of output. These include @option{-g},
10113 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
10114 the @var{compilation-options}, they are ignored.
10117 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
10118 systems) instead of @samp{.c}. This is convenient if you are converting
10119 a C program to C++. This option applies only to @code{protoize}.
10122 Add explicit global declarations. This means inserting explicit
10123 declarations at the beginning of each source file for each function
10124 that is called in the file and was not declared. These declarations
10125 precede the first function definition that contains a call to an
10126 undeclared function. This option applies only to @code{protoize}.
10128 @item -i @var{string}
10129 Indent old-style parameter declarations with the string @var{string}.
10130 This option applies only to @code{protoize}.
10132 @code{unprotoize} converts prototyped function definitions to old-style
10133 function definitions, where the arguments are declared between the
10134 argument list and the initial @samp{@{}. By default, @code{unprotoize}
10135 uses five spaces as the indentation. If you want to indent with just
10136 one space instead, use @option{-i " "}.
10139 Keep the @samp{.X} files. Normally, they are deleted after conversion
10143 Add explicit local declarations. @code{protoize} with @option{-l} inserts
10144 a prototype declaration for each function in each block which calls the
10145 function without any declaration. This option applies only to
10149 Make no real changes. This mode just prints information about the conversions
10150 that would have been done without @option{-n}.
10153 Make no @samp{.save} files. The original files are simply deleted.
10154 Use this option with caution.
10156 @item -p @var{program}
10157 Use the program @var{program} as the compiler. Normally, the name
10158 @file{gcc} is used.
10161 Work quietly. Most warnings are suppressed.
10164 Print the version number, just like @option{-v} for @code{gcc}.
10167 If you need special compiler options to compile one of your program's
10168 source files, then you should generate that file's @samp{.X} file
10169 specially, by running @code{gcc} on that source file with the
10170 appropriate options and the option @option{-aux-info}. Then run
10171 @code{protoize} on the entire set of files. @code{protoize} will use
10172 the existing @samp{.X} file because it is newer than the source file.
10176 gcc -Dfoo=bar file1.c -aux-info file1.X
10181 You need to include the special files along with the rest in the
10182 @code{protoize} command, even though their @samp{.X} files already
10183 exist, because otherwise they won't get converted.
10185 @xref{Protoize Caveats}, for more information on how to use
10186 @code{protoize} successfully.