1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003 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, 2003 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 1.2 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
69 When you invoke GCC, it normally does preprocessing, compilation,
70 assembly and linking. The ``overall options'' allow you to stop this
71 process at an intermediate stage. For example, the @option{-c} option
72 says not to run the linker. Then the output consists of object files
73 output by the assembler.
75 Other options are passed on to one stage of processing. Some options
76 control the preprocessor and others the compiler itself. Yet other
77 options control the assembler and linker; most of these are not
78 documented here, since you rarely need to use any of them.
80 @cindex C compilation options
81 Most of the command line options that you can use with GCC are useful
82 for C programs; when an option is only useful with another language
83 (usually C++), the explanation says so explicitly. If the description
84 for a particular option does not mention a source language, you can use
85 that option with all supported languages.
87 @cindex C++ compilation options
88 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
89 options for compiling C++ programs.
91 @cindex grouping options
92 @cindex options, grouping
93 The @command{gcc} program accepts options and file names as operands. Many
94 options have multi-letter names; therefore multiple single-letter options
95 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
98 @cindex order of options
99 @cindex options, order
100 You can mix options and other arguments. For the most part, the order
101 you use doesn't matter. Order does matter when you use several options
102 of the same kind; for example, if you specify @option{-L} more than once,
103 the directories are searched in the order specified.
105 Many options have long names starting with @samp{-f} or with
106 @samp{-W}---for example, @option{-fforce-mem},
107 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
108 these have both positive and negative forms; the negative form of
109 @option{-ffoo} would be @option{-fno-foo}. This manual documents
110 only one of these two forms, whichever one is not the default.
114 @xref{Option Index}, for an index to GCC's options.
117 * Option Summary:: Brief list of all options, without explanations.
118 * Overall Options:: Controlling the kind of output:
119 an executable, object files, assembler files,
120 or preprocessed source.
121 * Invoking G++:: Compiling C++ programs.
122 * C Dialect Options:: Controlling the variant of C language compiled.
123 * C++ Dialect Options:: Variations on C++.
124 * Objective-C Dialect Options:: Variations on Objective-C.
125 * Language Independent Options:: Controlling how diagnostics should be
127 * Warning Options:: How picky should the compiler be?
128 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
129 * Optimize Options:: How much optimization?
130 * Preprocessor Options:: Controlling header files and macro definitions.
131 Also, getting dependency information for Make.
132 * Assembler Options:: Passing options to the assembler.
133 * Link Options:: Specifying libraries and so on.
134 * Directory Options:: Where to find header files and libraries.
135 Where to find the compiler executable files.
136 * Spec Files:: How to pass switches to sub-processes.
137 * Target Options:: Running a cross-compiler, or an old version of GCC.
138 * Submodel Options:: Specifying minor hardware or convention variations,
139 such as 68010 vs 68020.
140 * Code Gen Options:: Specifying conventions for function calls, data layout
142 * Environment Variables:: Env vars that affect GCC.
143 * Precompiled Headers:: Compiling a header once, and using it many times.
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}.
158 @gccoptlist{-c -S -E -o @var{file} -pipe -pass-exit-codes @gol
159 -x @var{language} -v -### --help --target-help --version}
161 @item C Language Options
162 @xref{C Dialect Options,,Options Controlling C Dialect}.
163 @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol
164 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
165 -fhosted -ffreestanding -fms-extensions @gol
166 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
167 -fallow-single-precision -fcond-mismatch @gol
168 -fsigned-bitfields -fsigned-char @gol
169 -funsigned-bitfields -funsigned-char @gol
172 @item C++ Language Options
173 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
174 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
175 -fconserve-space -fno-const-strings @gol
176 -fno-elide-constructors @gol
177 -fno-enforce-eh-specs -fexternal-templates @gol
178 -falt-external-templates @gol
179 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
180 -fno-implicit-templates @gol
181 -fno-implicit-inline-templates @gol
182 -fno-implement-inlines -fms-extensions @gol
183 -fno-nonansi-builtins -fno-operator-names @gol
184 -fno-optional-diags -fpermissive @gol
185 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
186 -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
187 -fno-default-inline -Wabi -Wctor-dtor-privacy @gol
188 -Wnon-virtual-dtor -Wreorder @gol
189 -Weffc++ -Wno-deprecated @gol
190 -Wno-non-template-friend -Wold-style-cast @gol
191 -Woverloaded-virtual -Wno-pmf-conversions @gol
192 -Wsign-promo -Wsynth}
194 @item Objective-C Language Options
195 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
197 -fconstant-string-class=@var{class-name} @gol
198 -fgnu-runtime -fnext-runtime @gol
199 -fno-nil-receivers @gol
200 -fobjc-exceptions @gol
201 -freplace-objc-classes @gol
204 -Wno-protocol -Wselector -Wundeclared-selector}
206 @item Language Independent Options
207 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
208 @gccoptlist{-fmessage-length=@var{n} @gol
209 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
211 @item Warning Options
212 @xref{Warning Options,,Options to Request or Suppress Warnings}.
213 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
214 -w -Wextra -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 -Wendif-labels @gol
218 -Werror -Werror-implicit-function-declaration @gol
219 -Wfloat-equal -Wformat -Wformat=2 @gol
220 -Wno-format-extra-args -Wformat-nonliteral @gol
221 -Wformat-security -Wno-format-y2k @gol
222 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
223 -Wimport -Wno-import -Winit-self -Winline @gol
224 -Wno-invalid-offsetof -Winvalid-pch @gol
225 -Wlarger-than-@var{len} -Wlong-long @gol
226 -Wmain -Wmissing-braces @gol
227 -Wmissing-format-attribute -Wmissing-noreturn @gol
228 -Wno-multichar -Wnonnull -Wpacked -Wpadded @gol
229 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
230 -Wreturn-type -Wsequence-point -Wshadow @gol
231 -Wsign-compare -Wstrict-aliasing @gol
232 -Wswitch -Wswitch-default -Wswitch-enum @gol
233 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
234 -Wunknown-pragmas -Wunreachable-code @gol
235 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
236 -Wunused-value -Wunused-variable -Wwrite-strings}
238 @item C-only Warning Options
239 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
240 -Wmissing-prototypes -Wnested-externs -Wold-style-definition @gol
241 -Wstrict-prototypes -Wtraditional @gol
242 -Wdeclaration-after-statement}
244 @item Debugging Options
245 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
246 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
247 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
248 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
249 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
250 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
251 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
252 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
253 -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs @gol
254 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
255 -ftest-coverage -ftime-report @gol
256 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
257 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
258 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
259 -print-multi-directory -print-multi-lib @gol
260 -print-prog-name=@var{program} -print-search-dirs -Q @gol
263 @item Optimization Options
264 @xref{Optimize Options,,Options that Control Optimization}.
265 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
266 -falign-labels=@var{n} -falign-loops=@var{n} @gol
267 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
268 -fbranch-target-load-optimize2 -fcaller-saves -fcprop-registers @gol
269 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
270 -fdelayed-branch -fdelete-null-pointer-checks @gol
271 -fexpensive-optimizations -ffast-math -ffloat-store @gol
272 -fforce-addr -fforce-mem -ffunction-sections @gol
273 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -floop-optimize @gol
274 -fcrossjumping -fif-conversion -fif-conversion2 @gol
275 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
276 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
277 -fmove-all-movables -fnew-ra -fno-branch-count-reg @gol
278 -fno-default-inline -fno-defer-pop @gol
279 -fno-function-cse -fno-guess-branch-probability @gol
280 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
281 -funsafe-math-optimizations -ffinite-math-only @gol
282 -fno-trapping-math -fno-zero-initialized-in-bss @gol
283 -fomit-frame-pointer -foptimize-register-move @gol
284 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
285 -freduce-all-givs -fregmove -frename-registers @gol
286 -freorder-blocks -freorder-functions @gol
287 -frerun-cse-after-loop -frerun-loop-opt @gol
288 -frounding-math -fschedule-insns -fschedule-insns2 @gol
289 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
290 -fsched-spec-load-dangerous @gol
291 -fsched-stalled-insns=@var{n} -sched-stalled-insns-dep=@var{n} @gol
292 -fsched2-use-superblocks @gol
293 -fsched2-use-traces -fsignaling-nans @gol
294 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
295 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
296 -funroll-all-loops -funroll-loops -fpeel-loops @gol
297 -funswitch-loops -fold-unroll-loops -fold-unroll-all-loops @gol
298 --param @var{name}=@var{value}
299 -O -O0 -O1 -O2 -O3 -Os}
301 @item Preprocessor Options
302 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
303 @gccoptlist{-A@var{question}=@var{answer} @gol
304 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
305 -C -dD -dI -dM -dN @gol
306 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
307 -idirafter @var{dir} @gol
308 -include @var{file} -imacros @var{file} @gol
309 -iprefix @var{file} -iwithprefix @var{dir} @gol
310 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
311 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
312 -P -fworking-directory -remap @gol
313 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
314 -Xpreprocessor @var{option}}
316 @item Assembler Option
317 @xref{Assembler Options,,Passing Options to the Assembler}.
318 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
321 @xref{Link Options,,Options for Linking}.
322 @gccoptlist{@var{object-file-name} -l@var{library} @gol
323 -nostartfiles -nodefaultlibs -nostdlib -pie @gol
324 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
325 -Wl,@var{option} -Xlinker @var{option} @gol
328 @item Directory Options
329 @xref{Directory Options,,Options for Directory Search}.
330 @gccoptlist{-B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
333 @c I wrote this xref this way to avoid overfull hbox. -- rms
334 @xref{Target Options}.
335 @gccoptlist{-V @var{version} -b @var{machine}}
337 @item Machine Dependent Options
338 @xref{Submodel Options,,Hardware Models and Configurations}.
340 @emph{M680x0 Options}
341 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
342 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
343 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
344 -malign-int -mstrict-align}
346 @emph{M68hc1x Options}
347 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
348 -mauto-incdec -minmax -mlong-calls -mshort @gol
349 -msoft-reg-count=@var{count}}
352 @gccoptlist{-mg -mgnu -munix}
355 @gccoptlist{-mcpu=@var{cpu-type} @gol
356 -mtune=@var{cpu-type} @gol
357 -mcmodel=@var{code-model} @gol
359 -mapp-regs -mbroken-saverestore -mcypress @gol
360 -mfaster-structs -mflat @gol
361 -mfpu -mhard-float -mhard-quad-float @gol
362 -mimpure-text -mlittle-endian -mlive-g0 -mno-app-regs @gol
363 -mno-faster-structs -mno-flat -mno-fpu @gol
364 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
365 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
366 -msupersparc -munaligned-doubles -mv8}
369 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
370 -mapcs-26 -mapcs-32 @gol
371 -mapcs-stack-check -mno-apcs-stack-check @gol
372 -mapcs-float -mno-apcs-float @gol
373 -mapcs-reentrant -mno-apcs-reentrant @gol
374 -msched-prolog -mno-sched-prolog @gol
375 -mlittle-endian -mbig-endian -mwords-little-endian @gol
376 -malignment-traps -mno-alignment-traps @gol
377 -msoft-float -mhard-float -mfpe @gol
378 -mthumb-interwork -mno-thumb-interwork @gol
379 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
380 -mstructure-size-boundary=@var{n} @gol
381 -mabort-on-noreturn @gol
382 -mlong-calls -mno-long-calls @gol
383 -msingle-pic-base -mno-single-pic-base @gol
384 -mpic-register=@var{reg} @gol
385 -mnop-fun-dllimport @gol
386 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
387 -mpoke-function-name @gol
389 -mtpcs-frame -mtpcs-leaf-frame @gol
390 -mcaller-super-interworking -mcallee-super-interworking}
392 @emph{MN10200 Options}
395 @emph{MN10300 Options}
396 @gccoptlist{-mmult-bug -mno-mult-bug @gol
397 -mam33 -mno-am33 @gol
398 -mam33-2 -mno-am33-2 @gol
401 @emph{M32R/D Options}
402 @gccoptlist{-m32rx -m32r -mcode-model=@var{model-type} @gol
403 -msdata=@var{sdata-type} -G @var{num}}
406 @gccoptlist{-m88000 -m88100 -m88110 -mbig-pic @gol
407 -mcheck-zero-division -mhandle-large-shift @gol
408 -midentify-revision -mno-check-zero-division @gol
409 -mno-ocs-debug-info -mno-ocs-frame-position @gol
410 -mno-optimize-arg-area -mno-serialize-volatile @gol
411 -mno-underscores -mocs-debug-info @gol
412 -mocs-frame-position -moptimize-arg-area @gol
413 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
414 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
415 -mversion-03.00 -mwarn-passed-structs}
417 @emph{RS/6000 and PowerPC Options}
418 @gccoptlist{-mcpu=@var{cpu-type} @gol
419 -mtune=@var{cpu-type} @gol
420 -mpower -mno-power -mpower2 -mno-power2 @gol
421 -mpowerpc -mpowerpc64 -mno-powerpc @gol
422 -maltivec -mno-altivec @gol
423 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
424 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
425 -mnew-mnemonics -mold-mnemonics @gol
426 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
427 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
428 -malign-power -malign-natural @gol
429 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
430 -mstring -mno-string -mupdate -mno-update @gol
431 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
432 -mstrict-align -mno-strict-align -mrelocatable @gol
433 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
434 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
435 -mdynamic-no-pic @gol
436 -mprioritize-restricted-insns=@var{priority} @gol
437 -msched-costly-dep=@var{dependence_type} @gol
438 -minsert-sched-nops=@var{scheme} @gol
439 -mcall-sysv -mcall-netbsd @gol
440 -maix-struct-return -msvr4-struct-return @gol
441 -mabi=altivec -mabi=no-altivec @gol
442 -mabi=spe -mabi=no-spe @gol
443 -misel=yes -misel=no @gol
444 -mspe=yes -mspe=no @gol
445 -mfloat-gprs=yes -mfloat-gprs=no @gol
446 -mprototype -mno-prototype @gol
447 -msim -mmvme -mads -myellowknife -memb -msdata @gol
448 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
450 @emph{Darwin Options}
451 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
452 -arch_only -bind_at_load -bundle -bundle_loader @gol
453 -client_name -compatibility_version -current_version @gol
454 -dependency-file -dylib_file -dylinker_install_name @gol
455 -dynamic -dynamiclib -exported_symbols_list @gol
456 -filelist -flat_namespace -force_cpusubtype_ALL @gol
457 -force_flat_namespace -headerpad_max_install_names @gol
458 -image_base -init -install_name -keep_private_externs @gol
459 -multi_module -multiply_defined -multiply_defined_unused @gol
460 -noall_load -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
461 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
462 -private_bundle -read_only_relocs -sectalign @gol
463 -sectobjectsymbols -whyload -seg1addr @gol
464 -sectcreate -sectobjectsymbols -sectorder @gol
465 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
466 -segprot -segs_read_only_addr -segs_read_write_addr @gol
467 -single_module -static -sub_library -sub_umbrella @gol
468 -twolevel_namespace -umbrella -undefined @gol
469 -unexported_symbols_list -weak_reference_mismatches @gol
473 @gccoptlist{-mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
474 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
475 -mminimum-fp-blocks -mnohc-struct-return}
478 @gccoptlist{-mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
479 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
480 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
481 -mgas -mgp32 -mgp64 -mhard-float -mint64 -mips1 @gol
482 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
483 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
484 -mmips-as -mmips-tfile -mno-abicalls -mxgot @gol
485 -mno-embedded-data -mno-uninit-const-in-rodata @gol
486 -mno-embedded-pic -mno-long-calls @gol
487 -mno-memcpy -mno-mips-tfile -mno-rnames @gol
488 -mrnames -msoft-float @gol
489 -m4650 -msingle-float -mmad @gol
490 -EL -EB -G @var{num} -nocpp @gol
491 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi -mabi-fake-default @gol
492 -mfix7000 -mfix-sb1 -mno-fix-sb1 @gol
493 -mno-crt0 -mflush-func=@var{func} -mno-flush-func @gol
494 -mbranch-likely -mno-branch-likely}
496 @emph{i386 and x86-64 Options}
497 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
498 -mfpmath=@var{unit} @gol
499 -masm=@var{dialect} -mno-fancy-math-387 @gol
500 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
501 -mno-wide-multiply -mrtd -malign-double @gol
502 -mpreferred-stack-boundary=@var{num} @gol
503 -mmmx -msse -msse2 -mpni -m3dnow @gol
504 -mthreads -mno-align-stringops -minline-all-stringops @gol
505 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
506 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
507 -mno-red-zone -mno-tls-direct-seg-refs @gol
508 -mcmodel=@var{code-model} @gol
512 @gccoptlist{-march=@var{architecture-type} @gol
513 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
514 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
515 -mjump-in-delay -mlinker-opt -mlong-calls @gol
516 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
517 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
518 -mno-jump-in-delay -mno-long-load-store @gol
519 -mno-portable-runtime -mno-soft-float @gol
520 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
521 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
522 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
523 -nolibdld -static -threads}
525 @emph{Intel 960 Options}
526 @gccoptlist{-m@var{cpu-type} -masm-compat -mclean-linkage @gol
527 -mcode-align -mcomplex-addr -mleaf-procedures @gol
528 -mic-compat -mic2.0-compat -mic3.0-compat @gol
529 -mintel-asm -mno-clean-linkage -mno-code-align @gol
530 -mno-complex-addr -mno-leaf-procedures @gol
531 -mno-old-align -mno-strict-align -mno-tail-call @gol
532 -mnumerics -mold-align -msoft-float -mstrict-align @gol
535 @emph{DEC Alpha Options}
536 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
537 -mieee -mieee-with-inexact -mieee-conformant @gol
538 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
539 -mtrap-precision=@var{mode} -mbuild-constants @gol
540 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
541 -mbwx -mmax -mfix -mcix @gol
542 -mfloat-vax -mfloat-ieee @gol
543 -mexplicit-relocs -msmall-data -mlarge-data @gol
544 -msmall-text -mlarge-text @gol
545 -mmemory-latency=@var{time}}
547 @emph{DEC Alpha/VMS Options}
548 @gccoptlist{-mvms-return-codes}
550 @emph{H8/300 Options}
551 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
554 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
555 -m4-nofpu -m4-single-only -m4-single -m4 @gol
556 -m5-64media -m5-64media-nofpu @gol
557 -m5-32media -m5-32media-nofpu @gol
558 -m5-compact -m5-compact-nofpu @gol
559 -mb -ml -mdalign -mrelax @gol
560 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
561 -mieee -misize -mpadstruct -mspace @gol
562 -mprefergot -musermode}
564 @emph{System V Options}
565 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
568 @gccoptlist{-EB -EL @gol
569 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
570 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
572 @emph{TMS320C3x/C4x Options}
573 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
574 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
575 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
576 -mparallel-insns -mparallel-mpy -mpreserve-float}
579 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
580 -mprolog-function -mno-prolog-function -mspace @gol
581 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
582 -mapp-regs -mno-app-regs @gol
583 -mdisable-callt -mno-disable-callt @gol
589 @gccoptlist{-m32032 -m32332 -m32532 -m32081 -m32381 @gol
590 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
591 -mregparam -mnoregparam -msb -mnosb @gol
592 -mbitfield -mnobitfield -mhimem -mnohimem}
595 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
596 -mcall-prologues -mno-tablejump -mtiny-stack}
599 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
600 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
601 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
602 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
603 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
606 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
607 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
608 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
609 -mno-base-addresses -msingle-exit -mno-single-exit}
612 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
613 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
614 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
615 -minline-float-divide-max-throughput @gol
616 -minline-int-divide-min-latency @gol
617 -minline-int-divide-max-throughput -mno-dwarf2-asm @gol
618 -mfixed-range=@var{register-range}}
621 @gccoptlist{-mextmem -mextmemory -monchip -mno-asm-optimize @gol
622 -masm-optimize -mbranch-cost=@var{n} -mcond-exec=@var{n}}
624 @emph{S/390 and zSeries Options}
625 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
626 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
627 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
628 -m64 -m31 -mdebug -mno-debug -mesa -mzarch -mfused-madd -mno-fused-madd}
631 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
632 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
633 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
634 -mstack-align -mdata-align -mconst-align @gol
635 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
636 -melf -maout -melinux -mlinux -sim -sim2}
638 @emph{PDP-11 Options}
639 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
640 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
641 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
642 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
643 -mbranch-expensive -mbranch-cheap @gol
644 -msplit -mno-split -munix-asm -mdec-asm}
646 @emph{Xstormy16 Options}
649 @emph{Xtensa Options}
650 @gccoptlist{-mconst16 -mno-const16 @gol
651 -mfused-madd -mno-fused-madd @gol
652 -mtext-section-literals -mno-text-section-literals @gol
653 -mtarget-align -mno-target-align @gol
654 -mlongcalls -mno-longcalls}
657 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
658 -mhard-float -msoft-float @gol
659 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
660 -mdouble -mno-double @gol
661 -mmedia -mno-media -mmuladd -mno-muladd @gol
662 -mlibrary-pic -macc-4 -macc-8 @gol
663 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
664 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
665 -mvliw-branch -mno-vliw-branch @gol
666 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
667 -mno-nested-cond-exec -mtomcat-stats @gol
670 @item Code Generation Options
671 @xref{Code Gen Options,,Options for Code Generation Conventions}.
672 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
673 -ffixed-@var{reg} -fexceptions @gol
674 -fnon-call-exceptions -funwind-tables @gol
675 -fasynchronous-unwind-tables @gol
676 -finhibit-size-directive -finstrument-functions @gol
677 -fno-common -fno-ident -fno-gnu-linker @gol
678 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
679 -freg-struct-return -fshared-data -fshort-enums @gol
680 -fshort-double -fshort-wchar @gol
681 -fverbose-asm -fpack-struct -fstack-check @gol
682 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
683 -fargument-alias -fargument-noalias @gol
684 -fargument-noalias-global -fleading-underscore @gol
685 -ftls-model=@var{model} @gol
686 -ftrapv -fwrapv -fbounds-check}
690 * Overall Options:: Controlling the kind of output:
691 an executable, object files, assembler files,
692 or preprocessed source.
693 * C Dialect Options:: Controlling the variant of C language compiled.
694 * C++ Dialect Options:: Variations on C++.
695 * Objective-C Dialect Options:: Variations on Objective-C.
696 * Language Independent Options:: Controlling how diagnostics should be
698 * Warning Options:: How picky should the compiler be?
699 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
700 * Optimize Options:: How much optimization?
701 * Preprocessor Options:: Controlling header files and macro definitions.
702 Also, getting dependency information for Make.
703 * Assembler Options:: Passing options to the assembler.
704 * Link Options:: Specifying libraries and so on.
705 * Directory Options:: Where to find header files and libraries.
706 Where to find the compiler executable files.
707 * Spec Files:: How to pass switches to sub-processes.
708 * Target Options:: Running a cross-compiler, or an old version of GCC.
711 @node Overall Options
712 @section Options Controlling the Kind of Output
714 Compilation can involve up to four stages: preprocessing, compilation
715 proper, assembly and linking, always in that order. GCC is capable of
716 preprocessing and compiling several files either into several
717 assembler input files, or into one assembler input file; then each
718 assembler input file produces an object file, and linking combines all
719 the object files (those newly compiled, and those specified as input)
720 into an executable file.
722 @cindex file name suffix
723 For any given input file, the file name suffix determines what kind of
728 C source code which must be preprocessed.
731 C source code which should not be preprocessed.
734 C++ source code which should not be preprocessed.
737 Objective-C source code. Note that you must link with the library
738 @file{libobjc.a} to make an Objective-C program work.
741 Objective-C source code which should not be preprocessed.
744 C or C++ header file to be turned into a precompiled header.
748 @itemx @var{file}.cxx
749 @itemx @var{file}.cpp
750 @itemx @var{file}.CPP
751 @itemx @var{file}.c++
753 C++ source code which must be preprocessed. Note that in @samp{.cxx},
754 the last two letters must both be literally @samp{x}. Likewise,
755 @samp{.C} refers to a literal capital C@.
759 C++ header file to be turned into a precompiled header.
762 @itemx @var{file}.for
763 @itemx @var{file}.FOR
764 Fortran source code which should not be preprocessed.
767 @itemx @var{file}.fpp
768 @itemx @var{file}.FPP
769 Fortran source code which must be preprocessed (with the traditional
773 Fortran source code which must be preprocessed with a RATFOR
774 preprocessor (not included with GCC)@.
776 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
777 Using and Porting GNU Fortran}, for more details of the handling of
780 @c FIXME: Descriptions of Java file types.
787 Ada source code file which contains a library unit declaration (a
788 declaration of a package, subprogram, or generic, or a generic
789 instantiation), or a library unit renaming declaration (a package,
790 generic, or subprogram renaming declaration). Such files are also
793 @itemx @var{file}.adb
794 Ada source code file containing a library unit body (a subprogram or
795 package body). Such files are also called @dfn{bodies}.
797 @c GCC also knows about some suffixes for languages not yet included:
806 Assembler code which must be preprocessed.
809 An object file to be fed straight into linking.
810 Any file name with no recognized suffix is treated this way.
814 You can specify the input language explicitly with the @option{-x} option:
817 @item -x @var{language}
818 Specify explicitly the @var{language} for the following input files
819 (rather than letting the compiler choose a default based on the file
820 name suffix). This option applies to all following input files until
821 the next @option{-x} option. Possible values for @var{language} are:
823 c c-header cpp-output
824 c++ c++-header c++-cpp-output
825 objective-c objective-c-header objc-cpp-output
826 assembler assembler-with-cpp
828 f77 f77-cpp-input ratfor
834 Turn off any specification of a language, so that subsequent files are
835 handled according to their file name suffixes (as they are if @option{-x}
836 has not been used at all).
838 @item -pass-exit-codes
839 @opindex pass-exit-codes
840 Normally the @command{gcc} program will exit with the code of 1 if any
841 phase of the compiler returns a non-success return code. If you specify
842 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
843 numerically highest error produced by any phase that returned an error
847 If you only want some of the stages of compilation, you can use
848 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
849 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
850 @command{gcc} is to stop. Note that some combinations (for example,
851 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
856 Compile or assemble the source files, but do not link. The linking
857 stage simply is not done. The ultimate output is in the form of an
858 object file for each source file.
860 By default, the object file name for a source file is made by replacing
861 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
863 Unrecognized input files, not requiring compilation or assembly, are
868 Stop after the stage of compilation proper; do not assemble. The output
869 is in the form of an assembler code file for each non-assembler input
872 By default, the assembler file name for a source file is made by
873 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
875 Input files that don't require compilation are ignored.
879 Stop after the preprocessing stage; do not run the compiler proper. The
880 output is in the form of preprocessed source code, which is sent to the
883 Input files which don't require preprocessing are ignored.
885 @cindex output file option
888 Place output in file @var{file}. This applies regardless to whatever
889 sort of output is being produced, whether it be an executable file,
890 an object file, an assembler file or preprocessed C code.
892 If you specify @option{-o} when compiling more than one input file, or
893 you are producing an executable file as output, all the source files
894 on the command line will be compiled at once.
896 If @option{-o} is not specified, the default is to put an executable file
897 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
898 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
899 all preprocessed C source on standard output.
903 Print (on standard error output) the commands executed to run the stages
904 of compilation. Also print the version number of the compiler driver
905 program and of the preprocessor and the compiler proper.
909 Like @option{-v} except the commands are not executed and all command
910 arguments are quoted. This is useful for shell scripts to capture the
911 driver-generated command lines.
915 Use pipes rather than temporary files for communication between the
916 various stages of compilation. This fails to work on some systems where
917 the assembler is unable to read from a pipe; but the GNU assembler has
922 Print (on the standard output) a description of the command line options
923 understood by @command{gcc}. If the @option{-v} option is also specified
924 then @option{--help} will also be passed on to the various processes
925 invoked by @command{gcc}, so that they can display the command line options
926 they accept. If the @option{-Wextra} option is also specified then command
927 line options which have no documentation associated with them will also
932 Print (on the standard output) a description of target specific command
933 line options for each tool.
937 Display the version number and copyrights of the invoked GCC.
941 @section Compiling C++ Programs
943 @cindex suffixes for C++ source
944 @cindex C++ source file suffixes
945 C++ source files conventionally use one of the suffixes @samp{.C},
946 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
947 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
948 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
949 files with these names and compiles them as C++ programs even if you
950 call the compiler the same way as for compiling C programs (usually
951 with the name @command{gcc}).
955 However, C++ programs often require class libraries as well as a
956 compiler that understands the C++ language---and under some
957 circumstances, you might want to compile programs or header files from
958 standard input, or otherwise without a suffix that flags them as C++
959 programs. You might also like to precompile a C header file with a
960 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
961 program that calls GCC with the default language set to C++, and
962 automatically specifies linking against the C++ library. On many
963 systems, @command{g++} is also installed with the name @command{c++}.
965 @cindex invoking @command{g++}
966 When you compile C++ programs, you may specify many of the same
967 command-line options that you use for compiling programs in any
968 language; or command-line options meaningful for C and related
969 languages; or options that are meaningful only for C++ programs.
970 @xref{C Dialect Options,,Options Controlling C Dialect}, for
971 explanations of options for languages related to C@.
972 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
973 explanations of options that are meaningful only for C++ programs.
975 @node C Dialect Options
976 @section Options Controlling C Dialect
977 @cindex dialect options
978 @cindex language dialect options
979 @cindex options, dialect
981 The following options control the dialect of C (or languages derived
982 from C, such as C++ and Objective-C) that the compiler accepts:
989 In C mode, support all ISO C90 programs. In C++ mode,
990 remove GNU extensions that conflict with ISO C++.
992 This turns off certain features of GCC that are incompatible with ISO
993 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
994 such as the @code{asm} and @code{typeof} keywords, and
995 predefined macros such as @code{unix} and @code{vax} that identify the
996 type of system you are using. It also enables the undesirable and
997 rarely used ISO trigraph feature. For the C compiler,
998 it disables recognition of C++ style @samp{//} comments as well as
999 the @code{inline} keyword.
1001 The alternate keywords @code{__asm__}, @code{__extension__},
1002 @code{__inline__} and @code{__typeof__} continue to work despite
1003 @option{-ansi}. You would not want to use them in an ISO C program, of
1004 course, but it is useful to put them in header files that might be included
1005 in compilations done with @option{-ansi}. Alternate predefined macros
1006 such as @code{__unix__} and @code{__vax__} are also available, with or
1007 without @option{-ansi}.
1009 The @option{-ansi} option does not cause non-ISO programs to be
1010 rejected gratuitously. For that, @option{-pedantic} is required in
1011 addition to @option{-ansi}. @xref{Warning Options}.
1013 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1014 option is used. Some header files may notice this macro and refrain
1015 from declaring certain functions or defining certain macros that the
1016 ISO standard doesn't call for; this is to avoid interfering with any
1017 programs that might use these names for other things.
1019 Functions which would normally be built in but do not have semantics
1020 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1021 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1022 built-in functions provided by GCC}, for details of the functions
1027 Determine the language standard. This option is currently only
1028 supported when compiling C or C++. A value for this option must be
1029 provided; possible values are
1034 ISO C90 (same as @option{-ansi}).
1036 @item iso9899:199409
1037 ISO C90 as modified in amendment 1.
1043 ISO C99. Note that this standard is not yet fully supported; see
1044 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1045 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1048 Default, ISO C90 plus GNU extensions (including some C99 features).
1052 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1053 this will become the default. The name @samp{gnu9x} is deprecated.
1056 The 1998 ISO C++ standard plus amendments.
1059 The same as @option{-std=c++98} plus GNU extensions. This is the
1060 default for C++ code.
1063 Even when this option is not specified, you can still use some of the
1064 features of newer standards in so far as they do not conflict with
1065 previous C standards. For example, you may use @code{__restrict__} even
1066 when @option{-std=c99} is not specified.
1068 The @option{-std} options specifying some version of ISO C have the same
1069 effects as @option{-ansi}, except that features that were not in ISO C90
1070 but are in the specified version (for example, @samp{//} comments and
1071 the @code{inline} keyword in ISO C99) are not disabled.
1073 @xref{Standards,,Language Standards Supported by GCC}, for details of
1074 these standard versions.
1076 @item -aux-info @var{filename}
1078 Output to the given filename prototyped declarations for all functions
1079 declared and/or defined in a translation unit, including those in header
1080 files. This option is silently ignored in any language other than C@.
1082 Besides declarations, the file indicates, in comments, the origin of
1083 each declaration (source file and line), whether the declaration was
1084 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1085 @samp{O} for old, respectively, in the first character after the line
1086 number and the colon), and whether it came from a declaration or a
1087 definition (@samp{C} or @samp{F}, respectively, in the following
1088 character). In the case of function definitions, a K&R-style list of
1089 arguments followed by their declarations is also provided, inside
1090 comments, after the declaration.
1094 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1095 keyword, so that code can use these words as identifiers. You can use
1096 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1097 instead. @option{-ansi} implies @option{-fno-asm}.
1099 In C++, this switch only affects the @code{typeof} keyword, since
1100 @code{asm} and @code{inline} are standard keywords. You may want to
1101 use the @option{-fno-gnu-keywords} flag instead, which has the same
1102 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1103 switch only affects the @code{asm} and @code{typeof} keywords, since
1104 @code{inline} is a standard keyword in ISO C99.
1107 @itemx -fno-builtin-@var{function}
1108 @opindex fno-builtin
1109 @cindex built-in functions
1110 Don't recognize built-in functions that do not begin with
1111 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1112 functions provided by GCC}, for details of the functions affected,
1113 including those which are not built-in functions when @option{-ansi} or
1114 @option{-std} options for strict ISO C conformance are used because they
1115 do not have an ISO standard meaning.
1117 GCC normally generates special code to handle certain built-in functions
1118 more efficiently; for instance, calls to @code{alloca} may become single
1119 instructions that adjust the stack directly, and calls to @code{memcpy}
1120 may become inline copy loops. The resulting code is often both smaller
1121 and faster, but since the function calls no longer appear as such, you
1122 cannot set a breakpoint on those calls, nor can you change the behavior
1123 of the functions by linking with a different library.
1125 With the @option{-fno-builtin-@var{function}} option
1126 only the built-in function @var{function} is
1127 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1128 function is named this is not built-in in this version of GCC, this
1129 option is ignored. There is no corresponding
1130 @option{-fbuiltin-@var{function}} option; if you wish to enable
1131 built-in functions selectively when using @option{-fno-builtin} or
1132 @option{-ffreestanding}, you may define macros such as:
1135 #define abs(n) __builtin_abs ((n))
1136 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1141 @cindex hosted environment
1143 Assert that compilation takes place in a hosted environment. This implies
1144 @option{-fbuiltin}. A hosted environment is one in which the
1145 entire standard library is available, and in which @code{main} has a return
1146 type of @code{int}. Examples are nearly everything except a kernel.
1147 This is equivalent to @option{-fno-freestanding}.
1149 @item -ffreestanding
1150 @opindex ffreestanding
1151 @cindex hosted environment
1153 Assert that compilation takes place in a freestanding environment. This
1154 implies @option{-fno-builtin}. A freestanding environment
1155 is one in which the standard library may not exist, and program startup may
1156 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1157 This is equivalent to @option{-fno-hosted}.
1159 @xref{Standards,,Language Standards Supported by GCC}, for details of
1160 freestanding and hosted environments.
1162 @item -fms-extensions
1163 @opindex fms-extensions
1164 Accept some non-standard constructs used in Microsoft header files.
1168 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1169 options for strict ISO C conformance) implies @option{-trigraphs}.
1171 @item -no-integrated-cpp
1172 @opindex no-integrated-cpp
1173 Performs a compilation in two passes: preprocessing and compiling. This
1174 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1175 @option{-B} option. The user supplied compilation step can then add in
1176 an additional preprocessing step after normal preprocessing but before
1177 compiling. The default is to use the integrated cpp (internal cpp)
1179 The semantics of this option will change if "cc1", "cc1plus", and
1180 "cc1obj" are merged.
1182 @cindex traditional C language
1183 @cindex C language, traditional
1185 @itemx -traditional-cpp
1186 @opindex traditional-cpp
1187 @opindex traditional
1188 Formerly, these options caused GCC to attempt to emulate a pre-standard
1189 C compiler. They are now only supported with the @option{-E} switch.
1190 The preprocessor continues to support a pre-standard mode. See the GNU
1191 CPP manual for details.
1193 @item -fcond-mismatch
1194 @opindex fcond-mismatch
1195 Allow conditional expressions with mismatched types in the second and
1196 third arguments. The value of such an expression is void. This option
1197 is not supported for C++.
1199 @item -funsigned-char
1200 @opindex funsigned-char
1201 Let the type @code{char} be unsigned, like @code{unsigned char}.
1203 Each kind of machine has a default for what @code{char} should
1204 be. It is either like @code{unsigned char} by default or like
1205 @code{signed char} by default.
1207 Ideally, a portable program should always use @code{signed char} or
1208 @code{unsigned char} when it depends on the signedness of an object.
1209 But many programs have been written to use plain @code{char} and
1210 expect it to be signed, or expect it to be unsigned, depending on the
1211 machines they were written for. This option, and its inverse, let you
1212 make such a program work with the opposite default.
1214 The type @code{char} is always a distinct type from each of
1215 @code{signed char} or @code{unsigned char}, even though its behavior
1216 is always just like one of those two.
1219 @opindex fsigned-char
1220 Let the type @code{char} be signed, like @code{signed char}.
1222 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1223 the negative form of @option{-funsigned-char}. Likewise, the option
1224 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1226 @item -fsigned-bitfields
1227 @itemx -funsigned-bitfields
1228 @itemx -fno-signed-bitfields
1229 @itemx -fno-unsigned-bitfields
1230 @opindex fsigned-bitfields
1231 @opindex funsigned-bitfields
1232 @opindex fno-signed-bitfields
1233 @opindex fno-unsigned-bitfields
1234 These options control whether a bit-field is signed or unsigned, when the
1235 declaration does not use either @code{signed} or @code{unsigned}. By
1236 default, such a bit-field is signed, because this is consistent: the
1237 basic integer types such as @code{int} are signed types.
1239 @item -fwritable-strings
1240 @opindex fwritable-strings
1241 Store string constants in the writable data segment and don't uniquize
1242 them. This is for compatibility with old programs which assume they can
1243 write into string constants.
1245 Writing into string constants is a very bad idea; ``constants'' should
1249 @node C++ Dialect Options
1250 @section Options Controlling C++ Dialect
1252 @cindex compiler options, C++
1253 @cindex C++ options, command line
1254 @cindex options, C++
1255 This section describes the command-line options that are only meaningful
1256 for C++ programs; but you can also use most of the GNU compiler options
1257 regardless of what language your program is in. For example, you
1258 might compile a file @code{firstClass.C} like this:
1261 g++ -g -frepo -O -c firstClass.C
1265 In this example, only @option{-frepo} is an option meant
1266 only for C++ programs; you can use the other options with any
1267 language supported by GCC@.
1269 Here is a list of options that are @emph{only} for compiling C++ programs:
1273 @item -fabi-version=@var{n}
1274 @opindex fabi-version
1275 Use version @var{n} of the C++ ABI. Version 1 is the version of the C++
1276 ABI that first appeared in G++ 3.2. Version 0 will always be the
1277 version that conforms most closely to the C++ ABI specification.
1278 Therefore, the ABI obtained using version 0 will change as ABI bugs are
1281 The default is version 1.
1283 @item -fno-access-control
1284 @opindex fno-access-control
1285 Turn off all access checking. This switch is mainly useful for working
1286 around bugs in the access control code.
1290 Check that the pointer returned by @code{operator new} is non-null
1291 before attempting to modify the storage allocated. This check is
1292 normally unnecessary because the C++ standard specifies that
1293 @code{operator new} will only return @code{0} if it is declared
1294 @samp{throw()}, in which case the compiler will always check the
1295 return value even without this option. In all other cases, when
1296 @code{operator new} has a non-empty exception specification, memory
1297 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1298 @samp{new (nothrow)}.
1300 @item -fconserve-space
1301 @opindex fconserve-space
1302 Put uninitialized or runtime-initialized global variables into the
1303 common segment, as C does. This saves space in the executable at the
1304 cost of not diagnosing duplicate definitions. If you compile with this
1305 flag and your program mysteriously crashes after @code{main()} has
1306 completed, you may have an object that is being destroyed twice because
1307 two definitions were merged.
1309 This option is no longer useful on most targets, now that support has
1310 been added for putting variables into BSS without making them common.
1312 @item -fno-const-strings
1313 @opindex fno-const-strings
1314 Give string constants type @code{char *} instead of type @code{const
1315 char *}. By default, G++ uses type @code{const char *} as required by
1316 the standard. Even if you use @option{-fno-const-strings}, you cannot
1317 actually modify the value of a string constant, unless you also use
1318 @option{-fwritable-strings}.
1320 This option might be removed in a future release of G++. For maximum
1321 portability, you should structure your code so that it works with
1322 string constants that have type @code{const char *}.
1324 @item -fno-elide-constructors
1325 @opindex fno-elide-constructors
1326 The C++ standard allows an implementation to omit creating a temporary
1327 which is only used to initialize another object of the same type.
1328 Specifying this option disables that optimization, and forces G++ to
1329 call the copy constructor in all cases.
1331 @item -fno-enforce-eh-specs
1332 @opindex fno-enforce-eh-specs
1333 Don't check for violation of exception specifications at runtime. This
1334 option violates the C++ standard, but may be useful for reducing code
1335 size in production builds, much like defining @samp{NDEBUG}. The compiler
1336 will still optimize based on the exception specifications.
1338 @item -fexternal-templates
1339 @opindex fexternal-templates
1341 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1342 template instantiation; template instances are emitted or not according
1343 to the location of the template definition. @xref{Template
1344 Instantiation}, for more information.
1346 This option is deprecated.
1348 @item -falt-external-templates
1349 @opindex falt-external-templates
1350 Similar to @option{-fexternal-templates}, but template instances are
1351 emitted or not according to the place where they are first instantiated.
1352 @xref{Template Instantiation}, for more information.
1354 This option is deprecated.
1357 @itemx -fno-for-scope
1359 @opindex fno-for-scope
1360 If @option{-ffor-scope} is specified, the scope of variables declared in
1361 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1362 as specified by the C++ standard.
1363 If @option{-fno-for-scope} is specified, the scope of variables declared in
1364 a @i{for-init-statement} extends to the end of the enclosing scope,
1365 as was the case in old versions of G++, and other (traditional)
1366 implementations of C++.
1368 The default if neither flag is given to follow the standard,
1369 but to allow and give a warning for old-style code that would
1370 otherwise be invalid, or have different behavior.
1372 @item -fno-gnu-keywords
1373 @opindex fno-gnu-keywords
1374 Do not recognize @code{typeof} as a keyword, so that code can use this
1375 word as an identifier. You can use the keyword @code{__typeof__} instead.
1376 @option{-ansi} implies @option{-fno-gnu-keywords}.
1378 @item -fno-implicit-templates
1379 @opindex fno-implicit-templates
1380 Never emit code for non-inline templates which are instantiated
1381 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1382 @xref{Template Instantiation}, for more information.
1384 @item -fno-implicit-inline-templates
1385 @opindex fno-implicit-inline-templates
1386 Don't emit code for implicit instantiations of inline templates, either.
1387 The default is to handle inlines differently so that compiles with and
1388 without optimization will need the same set of explicit instantiations.
1390 @item -fno-implement-inlines
1391 @opindex fno-implement-inlines
1392 To save space, do not emit out-of-line copies of inline functions
1393 controlled by @samp{#pragma implementation}. This will cause linker
1394 errors if these functions are not inlined everywhere they are called.
1396 @item -fms-extensions
1397 @opindex fms-extensions
1398 Disable pedantic warnings about constructs used in MFC, such as implicit
1399 int and getting a pointer to member function via non-standard syntax.
1401 @item -fno-nonansi-builtins
1402 @opindex fno-nonansi-builtins
1403 Disable built-in declarations of functions that are not mandated by
1404 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1405 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1407 @item -fno-operator-names
1408 @opindex fno-operator-names
1409 Do not treat the operator name keywords @code{and}, @code{bitand},
1410 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1411 synonyms as keywords.
1413 @item -fno-optional-diags
1414 @opindex fno-optional-diags
1415 Disable diagnostics that the standard says a compiler does not need to
1416 issue. Currently, the only such diagnostic issued by G++ is the one for
1417 a name having multiple meanings within a class.
1420 @opindex fpermissive
1421 Downgrade some diagnostics about nonconformant code from errors to
1422 warnings. Thus, using @option{-fpermissive} will allow some
1423 nonconforming code to compile.
1427 Enable automatic template instantiation at link time. This option also
1428 implies @option{-fno-implicit-templates}. @xref{Template
1429 Instantiation}, for more information.
1433 Disable generation of information about every class with virtual
1434 functions for use by the C++ runtime type identification features
1435 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1436 of the language, you can save some space by using this flag. Note that
1437 exception handling uses the same information, but it will generate it as
1442 Emit statistics about front-end processing at the end of the compilation.
1443 This information is generally only useful to the G++ development team.
1445 @item -ftemplate-depth-@var{n}
1446 @opindex ftemplate-depth
1447 Set the maximum instantiation depth for template classes to @var{n}.
1448 A limit on the template instantiation depth is needed to detect
1449 endless recursions during template class instantiation. ANSI/ISO C++
1450 conforming programs must not rely on a maximum depth greater than 17.
1452 @item -fuse-cxa-atexit
1453 @opindex fuse-cxa-atexit
1454 Register destructors for objects with static storage duration with the
1455 @code{__cxa_atexit} function rather than the @code{atexit} function.
1456 This option is required for fully standards-compliant handling of static
1457 destructors, but will only work if your C library supports
1458 @code{__cxa_atexit}.
1462 Do not use weak symbol support, even if it is provided by the linker.
1463 By default, G++ will use weak symbols if they are available. This
1464 option exists only for testing, and should not be used by end-users;
1465 it will result in inferior code and has no benefits. This option may
1466 be removed in a future release of G++.
1470 Do not search for header files in the standard directories specific to
1471 C++, but do still search the other standard directories. (This option
1472 is used when building the C++ library.)
1475 In addition, these optimization, warning, and code generation options
1476 have meanings only for C++ programs:
1479 @item -fno-default-inline
1480 @opindex fno-default-inline
1481 Do not assume @samp{inline} for functions defined inside a class scope.
1482 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1483 functions will have linkage like inline functions; they just won't be
1486 @item -Wabi @r{(C++ only)}
1488 Warn when G++ generates code that is probably not compatible with the
1489 vendor-neutral C++ ABI. Although an effort has been made to warn about
1490 all such cases, there are probably some cases that are not warned about,
1491 even though G++ is generating incompatible code. There may also be
1492 cases where warnings are emitted even though the code that is generated
1495 You should rewrite your code to avoid these warnings if you are
1496 concerned about the fact that code generated by G++ may not be binary
1497 compatible with code generated by other compilers.
1499 The known incompatibilities at this point include:
1504 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1505 pack data into the same byte as a base class. For example:
1508 struct A @{ virtual void f(); int f1 : 1; @};
1509 struct B : public A @{ int f2 : 1; @};
1513 In this case, G++ will place @code{B::f2} into the same byte
1514 as@code{A::f1}; other compilers will not. You can avoid this problem
1515 by explicitly padding @code{A} so that its size is a multiple of the
1516 byte size on your platform; that will cause G++ and other compilers to
1517 layout @code{B} identically.
1520 Incorrect handling of tail-padding for virtual bases. G++ does not use
1521 tail padding when laying out virtual bases. For example:
1524 struct A @{ virtual void f(); char c1; @};
1525 struct B @{ B(); char c2; @};
1526 struct C : public A, public virtual B @{@};
1530 In this case, G++ will not place @code{B} into the tail-padding for
1531 @code{A}; other compilers will. You can avoid this problem by
1532 explicitly padding @code{A} so that its size is a multiple of its
1533 alignment (ignoring virtual base classes); that will cause G++ and other
1534 compilers to layout @code{C} identically.
1537 Incorrect handling of bit-fields with declared widths greater than that
1538 of their underlying types, when the bit-fields appear in a union. For
1542 union U @{ int i : 4096; @};
1546 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1547 union too small by the number of bits in an @code{int}.
1550 Empty classes can be placed at incorrect offsets. For example:
1560 struct C : public B, public A @{@};
1564 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1565 it should be placed at offset zero. G++ mistakenly believes that the
1566 @code{A} data member of @code{B} is already at offset zero.
1569 Names of template functions whose types involve @code{typename} or
1570 template template parameters can be mangled incorrectly.
1573 template <typename Q>
1574 void f(typename Q::X) @{@}
1576 template <template <typename> class Q>
1577 void f(typename Q<int>::X) @{@}
1581 Instantiations of these templates may be mangled incorrectly.
1585 @item -Wctor-dtor-privacy @r{(C++ only)}
1586 @opindex Wctor-dtor-privacy
1587 Warn when a class seems unusable because all the constructors or
1588 destructors in that class are private, and it has neither friends nor
1589 public static member functions.
1591 @item -Wnon-virtual-dtor @r{(C++ only)}
1592 @opindex Wnon-virtual-dtor
1593 Warn when a class appears to be polymorphic, thereby requiring a virtual
1594 destructor, yet it declares a non-virtual one.
1595 This warning is enabled by @option{-Wall}.
1597 @item -Wreorder @r{(C++ only)}
1599 @cindex reordering, warning
1600 @cindex warning for reordering of member initializers
1601 Warn when the order of member initializers given in the code does not
1602 match the order in which they must be executed. For instance:
1608 A(): j (0), i (1) @{ @}
1612 The compiler will rearrange the member initializers for @samp{i}
1613 and @samp{j} to match the declaration order of the members, emitting
1614 a warning to that effect. This warning is enabled by @option{-Wall}.
1617 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1620 @item -Weffc++ @r{(C++ only)}
1622 Warn about violations of the following style guidelines from Scott Meyers'
1623 @cite{Effective C++} book:
1627 Item 11: Define a copy constructor and an assignment operator for classes
1628 with dynamically allocated memory.
1631 Item 12: Prefer initialization to assignment in constructors.
1634 Item 14: Make destructors virtual in base classes.
1637 Item 15: Have @code{operator=} return a reference to @code{*this}.
1640 Item 23: Don't try to return a reference when you must return an object.
1644 Also warn about violations of the following style guidelines from
1645 Scott Meyers' @cite{More Effective C++} book:
1649 Item 6: Distinguish between prefix and postfix forms of increment and
1650 decrement operators.
1653 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1657 When selecting this option, be aware that the standard library
1658 headers do not obey all of these guidelines; use @samp{grep -v}
1659 to filter out those warnings.
1661 @item -Wno-deprecated @r{(C++ only)}
1662 @opindex Wno-deprecated
1663 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1665 @item -Wno-non-template-friend @r{(C++ only)}
1666 @opindex Wno-non-template-friend
1667 Disable warnings when non-templatized friend functions are declared
1668 within a template. Since the advent of explicit template specification
1669 support in G++, if the name of the friend is an unqualified-id (i.e.,
1670 @samp{friend foo(int)}), the C++ language specification demands that the
1671 friend declare or define an ordinary, nontemplate function. (Section
1672 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1673 could be interpreted as a particular specialization of a templatized
1674 function. Because this non-conforming behavior is no longer the default
1675 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1676 check existing code for potential trouble spots and is on by default.
1677 This new compiler behavior can be turned off with
1678 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1679 but disables the helpful warning.
1681 @item -Wold-style-cast @r{(C++ only)}
1682 @opindex Wold-style-cast
1683 Warn if an old-style (C-style) cast to a non-void type is used within
1684 a C++ program. The new-style casts (@samp{static_cast},
1685 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1686 unintended effects and much easier to search for.
1688 @item -Woverloaded-virtual @r{(C++ only)}
1689 @opindex Woverloaded-virtual
1690 @cindex overloaded virtual fn, warning
1691 @cindex warning for overloaded virtual fn
1692 Warn when a function declaration hides virtual functions from a
1693 base class. For example, in:
1700 struct B: public A @{
1705 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1713 will fail to compile.
1715 @item -Wno-pmf-conversions @r{(C++ only)}
1716 @opindex Wno-pmf-conversions
1717 Disable the diagnostic for converting a bound pointer to member function
1720 @item -Wsign-promo @r{(C++ only)}
1721 @opindex Wsign-promo
1722 Warn when overload resolution chooses a promotion from unsigned or
1723 enumeral type to a signed type, over a conversion to an unsigned type of
1724 the same size. Previous versions of G++ would try to preserve
1725 unsignedness, but the standard mandates the current behavior.
1727 @item -Wsynth @r{(C++ only)}
1729 @cindex warning for synthesized methods
1730 @cindex synthesized methods, warning
1731 Warn when G++'s synthesis behavior does not match that of cfront. For
1737 A& operator = (int);
1747 In this example, G++ will synthesize a default @samp{A& operator =
1748 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1751 @node Objective-C Dialect Options
1752 @section Options Controlling Objective-C Dialect
1754 @cindex compiler options, Objective-C
1755 @cindex Objective-C options, command line
1756 @cindex options, Objective-C
1757 (NOTE: This manual does not describe the Objective-C language itself. See
1758 @w{@uref{http://gcc.gnu.org/readings.html}} for references.)
1760 This section describes the command-line options that are only meaningful
1761 for Objective-C programs, but you can also use most of the GNU compiler
1762 options regardless of what language your program is in. For example,
1763 you might compile a file @code{some_class.m} like this:
1766 gcc -g -fgnu-runtime -O -c some_class.m
1770 In this example, @option{-fgnu-runtime} is an option meant only for
1771 Objective-C programs; you can use the other options with any language
1774 Here is a list of options that are @emph{only} for compiling Objective-C
1778 @item -fconstant-string-class=@var{class-name}
1779 @opindex fconstant-string-class
1780 Use @var{class-name} as the name of the class to instantiate for each
1781 literal string specified with the syntax @code{@@"@dots{}"}. The default
1782 class name is @code{NXConstantString} if the GNU runtime is being used, and
1783 @code{NSConstantString} if the NeXT runtime is being used (see below). The
1784 @option{-fconstant-cfstrings} option, if also present, will override the
1785 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
1786 to be laid out as constant CoreFoundation strings.
1789 @opindex fgnu-runtime
1790 Generate object code compatible with the standard GNU Objective-C
1791 runtime. This is the default for most types of systems.
1793 @item -fnext-runtime
1794 @opindex fnext-runtime
1795 Generate output compatible with the NeXT runtime. This is the default
1796 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1797 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1800 @item -fno-nil-receivers
1801 @opindex -fno-nil-receivers
1802 Assume that all Objective-C message dispatches (e.g.,
1803 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
1804 is not @code{nil}. This allows for more efficient entry points in the runtime to be
1805 used. Currently, this option is only available in conjunction with
1806 the NeXT runtime on Mac OS X 10.3 and later.
1808 @item -fobjc-exceptions
1809 @opindex -fobjc-exceptions
1810 Enable syntactic support for structured exception handling in Objective-C,
1811 similar to what is offered by C++ and Java. Currently, this option is only
1812 available in conjunction with the NeXT runtime on Mac OS X 10.3 and later.
1820 @@catch (AnObjCClass *exc) @{
1827 @@catch (AnotherClass *exc) @{
1830 @@catch (id allOthers) @{
1840 The @code{@@throw} statement may appear anywhere in an Objective-C or
1841 Objective-C++ program; when used inside of a @code{@@catch} block, the
1842 @code{@@throw} may appear without an argument (as shown above), in which case
1843 the object caught by the @code{@@catch} will be rethrown.
1845 Note that only (pointers to) Objective-C objects may be thrown and
1846 caught using this scheme. When an object is thrown, it will be caught
1847 by the nearest @code{@@catch} clause capable of handling objects of that type,
1848 analogously to how @code{catch} blocks work in C++ and Java. A
1849 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
1850 any and all Objective-C exceptions not caught by previous @code{@@catch}
1853 The @code{@@finally} clause, if present, will be executed upon exit from the
1854 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
1855 regardless of whether any exceptions are thrown, caught or rethrown
1856 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
1857 of the @code{finally} clause in Java.
1859 There are several caveats to using the new exception mechanism:
1863 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
1864 idioms provided by the @code{NSException} class, the new
1865 exceptions can only be used on Mac OS X 10.3 (Panther) and later
1866 systems, due to additional functionality needed in the (NeXT) Objective-C
1870 As mentioned above, the new exceptions do not support handling
1871 types other than Objective-C objects. Furthermore, when used from
1872 Objective-C++, the Objective-C exception model does not interoperate with C++
1873 exceptions at this time. This means you cannot @code{@@throw} an exception
1874 from Objective-C and @code{catch} it in C++, or vice versa
1875 (i.e., @code{throw @dots{} @@catch}).
1878 The @option{-fobjc-exceptions} switch also enables the use of synchronization
1879 blocks for thread-safe execution:
1882 @@synchronized (ObjCClass *guard) @{
1887 Upon entering the @code{@@synchronized} block, a thread of execution shall
1888 first check whether a lock has been placed on the corresponding @code{guard}
1889 object by another thread. If it has, the current thread shall wait until
1890 the other thread relinquishes its lock. Once @code{guard} becomes available,
1891 the current thread will place its own lock on it, execute the code contained in
1892 the @code{@@synchronized} block, and finally relinquish the lock (thereby
1893 making @code{guard} available to other threads).
1895 Unlike Java, Objective-C does not allow for entire methods to be marked
1896 @code{@@synchronized}. Note that throwing exceptions out of
1897 @code{@@synchronized} blocks is allowed, and will cause the guarding object
1898 to be unlocked properly.
1900 @item -freplace-objc-classes
1901 @opindex -freplace-objc-classes
1902 Emit a special marker instructing @command{ld(1)} not to statically link in
1903 the resulting object file, and allow @command{dyld(1)} to load it in at
1904 run time instead. This is used in conjunction with the Fix-and-Continue
1905 debugging mode, where the object file in question may be recompiled and
1906 dynamically reloaded in the course of program execution, without the need
1907 to restart the program itself. Currently, Fix-and-Continue functionality
1908 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
1912 @opindex -fzero-link
1913 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
1914 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
1915 compile time) with static class references that get initialized at load time,
1916 which improves run-time performance. Specifying the @option{-fzero-link} flag
1917 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
1918 to be retained. This is useful in Zero-Link debugging mode, since it allows
1919 for individual class implementations to be modified during program execution.
1923 Dump interface declarations for all classes seen in the source file to a
1924 file named @file{@var{sourcename}.decl}.
1927 @opindex Wno-protocol
1928 If a class is declared to implement a protocol, a warning is issued for
1929 every method in the protocol that is not implemented by the class. The
1930 default behavior is to issue a warning for every method not explicitly
1931 implemented in the class, even if a method implementation is inherited
1932 from the superclass. If you use the @code{-Wno-protocol} option, then
1933 methods inherited from the superclass are considered to be implemented,
1934 and no warning is issued for them.
1938 Warn if multiple methods of different types for the same selector are
1939 found during compilation. The check is performed on the list of methods
1940 in the final stage of compilation. Additionally, a check is performed
1941 for each selector appearing in a @code{@@selector(@dots{})}
1942 expression, and a corresponding method for that selector has been found
1943 during compilation. Because these checks scan the method table only at
1944 the end of compilation, these warnings are not produced if the final
1945 stage of compilation is not reached, for example because an error is
1946 found during compilation, or because the @code{-fsyntax-only} option is
1949 @item -Wundeclared-selector
1950 @opindex Wundeclared-selector
1951 Warn if a @code{@@selector(@dots{})} expression referring to an
1952 undeclared selector is found. A selector is considered undeclared if no
1953 method with that name has been declared before the
1954 @code{@@selector(@dots{})} expression, either explicitly in an
1955 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
1956 an @code{@@implementation} section. This option always performs its
1957 checks as soon as a @code{@@selector(@dots{})} expression is found,
1958 while @code{-Wselector} only performs its checks in the final stage of
1959 compilation. This also enforces the coding style convention
1960 that methods and selectors must be declared before being used.
1962 @c not documented because only avail via -Wp
1963 @c @item -print-objc-runtime-info
1967 @node Language Independent Options
1968 @section Options to Control Diagnostic Messages Formatting
1969 @cindex options to control diagnostics formatting
1970 @cindex diagnostic messages
1971 @cindex message formatting
1973 Traditionally, diagnostic messages have been formatted irrespective of
1974 the output device's aspect (e.g.@: its width, @dots{}). The options described
1975 below can be used to control the diagnostic messages formatting
1976 algorithm, e.g.@: how many characters per line, how often source location
1977 information should be reported. Right now, only the C++ front end can
1978 honor these options. However it is expected, in the near future, that
1979 the remaining front ends would be able to digest them correctly.
1982 @item -fmessage-length=@var{n}
1983 @opindex fmessage-length
1984 Try to format error messages so that they fit on lines of about @var{n}
1985 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1986 the front ends supported by GCC@. If @var{n} is zero, then no
1987 line-wrapping will be done; each error message will appear on a single
1990 @opindex fdiagnostics-show-location
1991 @item -fdiagnostics-show-location=once
1992 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1993 reporter to emit @emph{once} source location information; that is, in
1994 case the message is too long to fit on a single physical line and has to
1995 be wrapped, the source location won't be emitted (as prefix) again,
1996 over and over, in subsequent continuation lines. This is the default
1999 @item -fdiagnostics-show-location=every-line
2000 Only meaningful in line-wrapping mode. Instructs the diagnostic
2001 messages reporter to emit the same source location information (as
2002 prefix) for physical lines that result from the process of breaking
2003 a message which is too long to fit on a single line.
2007 @node Warning Options
2008 @section Options to Request or Suppress Warnings
2009 @cindex options to control warnings
2010 @cindex warning messages
2011 @cindex messages, warning
2012 @cindex suppressing warnings
2014 Warnings are diagnostic messages that report constructions which
2015 are not inherently erroneous but which are risky or suggest there
2016 may have been an error.
2018 You can request many specific warnings with options beginning @samp{-W},
2019 for example @option{-Wimplicit} to request warnings on implicit
2020 declarations. Each of these specific warning options also has a
2021 negative form beginning @samp{-Wno-} to turn off warnings;
2022 for example, @option{-Wno-implicit}. This manual lists only one of the
2023 two forms, whichever is not the default.
2025 The following options control the amount and kinds of warnings produced
2026 by GCC; for further, language-specific options also refer to
2027 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
2030 @cindex syntax checking
2032 @opindex fsyntax-only
2033 Check the code for syntax errors, but don't do anything beyond that.
2037 Issue all the warnings demanded by strict ISO C and ISO C++;
2038 reject all programs that use forbidden extensions, and some other
2039 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2040 version of the ISO C standard specified by any @option{-std} option used.
2042 Valid ISO C and ISO C++ programs should compile properly with or without
2043 this option (though a rare few will require @option{-ansi} or a
2044 @option{-std} option specifying the required version of ISO C)@. However,
2045 without this option, certain GNU extensions and traditional C and C++
2046 features are supported as well. With this option, they are rejected.
2048 @option{-pedantic} does not cause warning messages for use of the
2049 alternate keywords whose names begin and end with @samp{__}. Pedantic
2050 warnings are also disabled in the expression that follows
2051 @code{__extension__}. However, only system header files should use
2052 these escape routes; application programs should avoid them.
2053 @xref{Alternate Keywords}.
2055 Some users try to use @option{-pedantic} to check programs for strict ISO
2056 C conformance. They soon find that it does not do quite what they want:
2057 it finds some non-ISO practices, but not all---only those for which
2058 ISO C @emph{requires} a diagnostic, and some others for which
2059 diagnostics have been added.
2061 A feature to report any failure to conform to ISO C might be useful in
2062 some instances, but would require considerable additional work and would
2063 be quite different from @option{-pedantic}. We don't have plans to
2064 support such a feature in the near future.
2066 Where the standard specified with @option{-std} represents a GNU
2067 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2068 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2069 extended dialect is based. Warnings from @option{-pedantic} are given
2070 where they are required by the base standard. (It would not make sense
2071 for such warnings to be given only for features not in the specified GNU
2072 C dialect, since by definition the GNU dialects of C include all
2073 features the compiler supports with the given option, and there would be
2074 nothing to warn about.)
2076 @item -pedantic-errors
2077 @opindex pedantic-errors
2078 Like @option{-pedantic}, except that errors are produced rather than
2083 Inhibit all warning messages.
2087 Inhibit warning messages about the use of @samp{#import}.
2089 @item -Wchar-subscripts
2090 @opindex Wchar-subscripts
2091 Warn if an array subscript has type @code{char}. This is a common cause
2092 of error, as programmers often forget that this type is signed on some
2097 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2098 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2102 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2103 the arguments supplied have types appropriate to the format string
2104 specified, and that the conversions specified in the format string make
2105 sense. This includes standard functions, and others specified by format
2106 attributes (@pxref{Function Attributes}), in the @code{printf},
2107 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2108 not in the C standard) families.
2110 The formats are checked against the format features supported by GNU
2111 libc version 2.2. These include all ISO C90 and C99 features, as well
2112 as features from the Single Unix Specification and some BSD and GNU
2113 extensions. Other library implementations may not support all these
2114 features; GCC does not support warning about features that go beyond a
2115 particular library's limitations. However, if @option{-pedantic} is used
2116 with @option{-Wformat}, warnings will be given about format features not
2117 in the selected standard version (but not for @code{strfmon} formats,
2118 since those are not in any version of the C standard). @xref{C Dialect
2119 Options,,Options Controlling C Dialect}.
2121 Since @option{-Wformat} also checks for null format arguments for
2122 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2124 @option{-Wformat} is included in @option{-Wall}. For more control over some
2125 aspects of format checking, the options @option{-Wno-format-y2k},
2126 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2127 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2128 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2130 @item -Wno-format-y2k
2131 @opindex Wno-format-y2k
2132 If @option{-Wformat} is specified, do not warn about @code{strftime}
2133 formats which may yield only a two-digit year.
2135 @item -Wno-format-extra-args
2136 @opindex Wno-format-extra-args
2137 If @option{-Wformat} is specified, do not warn about excess arguments to a
2138 @code{printf} or @code{scanf} format function. The C standard specifies
2139 that such arguments are ignored.
2141 Where the unused arguments lie between used arguments that are
2142 specified with @samp{$} operand number specifications, normally
2143 warnings are still given, since the implementation could not know what
2144 type to pass to @code{va_arg} to skip the unused arguments. However,
2145 in the case of @code{scanf} formats, this option will suppress the
2146 warning if the unused arguments are all pointers, since the Single
2147 Unix Specification says that such unused arguments are allowed.
2149 @item -Wno-format-zero-length
2150 @opindex Wno-format-zero-length
2151 If @option{-Wformat} is specified, do not warn about zero-length formats.
2152 The C standard specifies that zero-length formats are allowed.
2154 @item -Wformat-nonliteral
2155 @opindex Wformat-nonliteral
2156 If @option{-Wformat} is specified, also warn if the format string is not a
2157 string literal and so cannot be checked, unless the format function
2158 takes its format arguments as a @code{va_list}.
2160 @item -Wformat-security
2161 @opindex Wformat-security
2162 If @option{-Wformat} is specified, also warn about uses of format
2163 functions that represent possible security problems. At present, this
2164 warns about calls to @code{printf} and @code{scanf} functions where the
2165 format string is not a string literal and there are no format arguments,
2166 as in @code{printf (foo);}. This may be a security hole if the format
2167 string came from untrusted input and contains @samp{%n}. (This is
2168 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2169 in future warnings may be added to @option{-Wformat-security} that are not
2170 included in @option{-Wformat-nonliteral}.)
2174 Enable @option{-Wformat} plus format checks not included in
2175 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2176 -Wformat-nonliteral -Wformat-security}.
2180 Warn about passing a null pointer for arguments marked as
2181 requiring a non-null value by the @code{nonnull} function attribute.
2183 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2184 can be disabled with the @option{-Wno-nonnull} option.
2186 @item -Winit-self @r{(C, C++, and Objective-C only)}
2188 Warn about uninitialized variables which are initialized with themselves.
2189 Note this option can only be used with the @option{-Wuninitialized} option,
2190 which in turn only works with @option{-O1} and above.
2192 For example, GCC will warn about @code{i} being uninitialized in the
2193 following snippet only when @option{-Winit-self} has been specified:
2204 @item -Wimplicit-int
2205 @opindex Wimplicit-int
2206 Warn when a declaration does not specify a type.
2208 @item -Wimplicit-function-declaration
2209 @itemx -Werror-implicit-function-declaration
2210 @opindex Wimplicit-function-declaration
2211 @opindex Werror-implicit-function-declaration
2212 Give a warning (or error) whenever a function is used before being
2217 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2221 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2222 function with external linkage, returning int, taking either zero
2223 arguments, two, or three arguments of appropriate types.
2225 @item -Wmissing-braces
2226 @opindex Wmissing-braces
2227 Warn if an aggregate or union initializer is not fully bracketed. In
2228 the following example, the initializer for @samp{a} is not fully
2229 bracketed, but that for @samp{b} is fully bracketed.
2232 int a[2][2] = @{ 0, 1, 2, 3 @};
2233 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2237 @opindex Wparentheses
2238 Warn if parentheses are omitted in certain contexts, such
2239 as when there is an assignment in a context where a truth value
2240 is expected, or when operators are nested whose precedence people
2241 often get confused about.
2243 Also warn about constructions where there may be confusion to which
2244 @code{if} statement an @code{else} branch belongs. Here is an example of
2259 In C, every @code{else} branch belongs to the innermost possible @code{if}
2260 statement, which in this example is @code{if (b)}. This is often not
2261 what the programmer expected, as illustrated in the above example by
2262 indentation the programmer chose. When there is the potential for this
2263 confusion, GCC will issue a warning when this flag is specified.
2264 To eliminate the warning, add explicit braces around the innermost
2265 @code{if} statement so there is no way the @code{else} could belong to
2266 the enclosing @code{if}. The resulting code would look like this:
2282 @item -Wsequence-point
2283 @opindex Wsequence-point
2284 Warn about code that may have undefined semantics because of violations
2285 of sequence point rules in the C standard.
2287 The C standard defines the order in which expressions in a C program are
2288 evaluated in terms of @dfn{sequence points}, which represent a partial
2289 ordering between the execution of parts of the program: those executed
2290 before the sequence point, and those executed after it. These occur
2291 after the evaluation of a full expression (one which is not part of a
2292 larger expression), after the evaluation of the first operand of a
2293 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2294 function is called (but after the evaluation of its arguments and the
2295 expression denoting the called function), and in certain other places.
2296 Other than as expressed by the sequence point rules, the order of
2297 evaluation of subexpressions of an expression is not specified. All
2298 these rules describe only a partial order rather than a total order,
2299 since, for example, if two functions are called within one expression
2300 with no sequence point between them, the order in which the functions
2301 are called is not specified. However, the standards committee have
2302 ruled that function calls do not overlap.
2304 It is not specified when between sequence points modifications to the
2305 values of objects take effect. Programs whose behavior depends on this
2306 have undefined behavior; the C standard specifies that ``Between the
2307 previous and next sequence point an object shall have its stored value
2308 modified at most once by the evaluation of an expression. Furthermore,
2309 the prior value shall be read only to determine the value to be
2310 stored.''. If a program breaks these rules, the results on any
2311 particular implementation are entirely unpredictable.
2313 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2314 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2315 diagnosed by this option, and it may give an occasional false positive
2316 result, but in general it has been found fairly effective at detecting
2317 this sort of problem in programs.
2319 The present implementation of this option only works for C programs. A
2320 future implementation may also work for C++ programs.
2322 The C standard is worded confusingly, therefore there is some debate
2323 over the precise meaning of the sequence point rules in subtle cases.
2324 Links to discussions of the problem, including proposed formal
2325 definitions, may be found on our readings page, at
2326 @w{@uref{http://gcc.gnu.org/readings.html}}.
2329 @opindex Wreturn-type
2330 Warn whenever a function is defined with a return-type that defaults to
2331 @code{int}. Also warn about any @code{return} statement with no
2332 return-value in a function whose return-type is not @code{void}.
2334 For C++, a function without return type always produces a diagnostic
2335 message, even when @option{-Wno-return-type} is specified. The only
2336 exceptions are @samp{main} and functions defined in system headers.
2340 Warn whenever a @code{switch} statement has an index of enumeral type
2341 and lacks a @code{case} for one or more of the named codes of that
2342 enumeration. (The presence of a @code{default} label prevents this
2343 warning.) @code{case} labels outside the enumeration range also
2344 provoke warnings when this option is used.
2346 @item -Wswitch-default
2347 @opindex Wswitch-switch
2348 Warn whenever a @code{switch} statement does not have a @code{default}
2352 @opindex Wswitch-enum
2353 Warn whenever a @code{switch} statement has an index of enumeral type
2354 and lacks a @code{case} for one or more of the named codes of that
2355 enumeration. @code{case} labels outside the enumeration range also
2356 provoke warnings when this option is used.
2360 Warn if any trigraphs are encountered that might change the meaning of
2361 the program (trigraphs within comments are not warned about).
2363 @item -Wunused-function
2364 @opindex Wunused-function
2365 Warn whenever a static function is declared but not defined or a
2366 non\-inline static function is unused.
2368 @item -Wunused-label
2369 @opindex Wunused-label
2370 Warn whenever a label is declared but not used.
2372 To suppress this warning use the @samp{unused} attribute
2373 (@pxref{Variable Attributes}).
2375 @item -Wunused-parameter
2376 @opindex Wunused-parameter
2377 Warn whenever a function parameter is unused aside from its declaration.
2379 To suppress this warning use the @samp{unused} attribute
2380 (@pxref{Variable Attributes}).
2382 @item -Wunused-variable
2383 @opindex Wunused-variable
2384 Warn whenever a local variable or non-constant static variable is unused
2385 aside from its declaration
2387 To suppress this warning use the @samp{unused} attribute
2388 (@pxref{Variable Attributes}).
2390 @item -Wunused-value
2391 @opindex Wunused-value
2392 Warn whenever a statement computes a result that is explicitly not used.
2394 To suppress this warning cast the expression to @samp{void}.
2398 All the above @option{-Wunused} options combined.
2400 In order to get a warning about an unused function parameter, you must
2401 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2402 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2404 @item -Wuninitialized
2405 @opindex Wuninitialized
2406 Warn if an automatic variable is used without first being initialized or
2407 if a variable may be clobbered by a @code{setjmp} call.
2409 These warnings are possible only in optimizing compilation,
2410 because they require data flow information that is computed only
2411 when optimizing. If you don't specify @option{-O}, you simply won't
2414 If you want to warn about code which uses the uninitialized value of the
2415 variable in its own initializer, use the @option{-Winit-self} option.
2417 These warnings occur only for variables that are candidates for
2418 register allocation. Therefore, they do not occur for a variable that
2419 is declared @code{volatile}, or whose address is taken, or whose size
2420 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2421 structures, unions or arrays, even when they are in registers.
2423 Note that there may be no warning about a variable that is used only
2424 to compute a value that itself is never used, because such
2425 computations may be deleted by data flow analysis before the warnings
2428 These warnings are made optional because GCC is not smart
2429 enough to see all the reasons why the code might be correct
2430 despite appearing to have an error. Here is one example of how
2451 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2452 always initialized, but GCC doesn't know this. Here is
2453 another common case:
2458 if (change_y) save_y = y, y = new_y;
2460 if (change_y) y = save_y;
2465 This has no bug because @code{save_y} is used only if it is set.
2467 @cindex @code{longjmp} warnings
2468 This option also warns when a non-volatile automatic variable might be
2469 changed by a call to @code{longjmp}. These warnings as well are possible
2470 only in optimizing compilation.
2472 The compiler sees only the calls to @code{setjmp}. It cannot know
2473 where @code{longjmp} will be called; in fact, a signal handler could
2474 call it at any point in the code. As a result, you may get a warning
2475 even when there is in fact no problem because @code{longjmp} cannot
2476 in fact be called at the place which would cause a problem.
2478 Some spurious warnings can be avoided if you declare all the functions
2479 you use that never return as @code{noreturn}. @xref{Function
2482 @item -Wunknown-pragmas
2483 @opindex Wunknown-pragmas
2484 @cindex warning for unknown pragmas
2485 @cindex unknown pragmas, warning
2486 @cindex pragmas, warning of unknown
2487 Warn when a #pragma directive is encountered which is not understood by
2488 GCC@. If this command line option is used, warnings will even be issued
2489 for unknown pragmas in system header files. This is not the case if
2490 the warnings were only enabled by the @option{-Wall} command line option.
2492 @item -Wstrict-aliasing
2493 @opindex Wstrict-aliasing
2494 This option is only active when @option{-fstrict-aliasing} is active.
2495 It warns about code which might break the strict aliasing rules that the
2496 compiler is using for optimization. The warning does not catch all
2497 cases, but does attempt to catch the more common pitfalls. It is
2498 included in @option{-Wall}.
2502 All of the above @samp{-W} options combined. This enables all the
2503 warnings about constructions that some users consider questionable, and
2504 that are easy to avoid (or modify to prevent the warning), even in
2505 conjunction with macros. This also enables some language-specific
2506 warnings described in @ref{C++ Dialect Options} and
2507 @ref{Objective-C Dialect Options}.
2510 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2511 Some of them warn about constructions that users generally do not
2512 consider questionable, but which occasionally you might wish to check
2513 for; others warn about constructions that are necessary or hard to avoid
2514 in some cases, and there is no simple way to modify the code to suppress
2521 (This option used to be called @option{-W}. The older name is still
2522 supported, but the newer name is more descriptive.) Print extra warning
2523 messages for these events:
2527 A function can return either with or without a value. (Falling
2528 off the end of the function body is considered returning without
2529 a value.) For example, this function would evoke such a
2543 An expression-statement or the left-hand side of a comma expression
2544 contains no side effects.
2545 To suppress the warning, cast the unused expression to void.
2546 For example, an expression such as @samp{x[i,j]} will cause a warning,
2547 but @samp{x[(void)i,j]} will not.
2550 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2553 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2554 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2555 that of ordinary mathematical notation.
2558 Storage-class specifiers like @code{static} are not the first things in
2559 a declaration. According to the C Standard, this usage is obsolescent.
2562 The return type of a function has a type qualifier such as @code{const}.
2563 Such a type qualifier has no effect, since the value returned by a
2564 function is not an lvalue. (But don't warn about the GNU extension of
2565 @code{volatile void} return types. That extension will be warned about
2566 if @option{-pedantic} is specified.)
2569 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2573 A comparison between signed and unsigned values could produce an
2574 incorrect result when the signed value is converted to unsigned.
2575 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2578 An aggregate has an initializer which does not initialize all members.
2579 For example, the following code would cause such a warning, because
2580 @code{x.h} would be implicitly initialized to zero:
2583 struct s @{ int f, g, h; @};
2584 struct s x = @{ 3, 4 @};
2588 A function parameter is declared without a type specifier in K&R-style
2596 An empty body occurs in an @samp{if} or @samp{else} statement.
2599 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2600 @samp{>}, or @samp{>=}.
2603 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2606 Any of several floating-point events that often indicate errors, such as
2607 overflow, underflow, loss of precision, etc.
2609 @item @r{(C++ only)}
2610 An enumerator and a non-enumerator both appear in a conditional expression.
2612 @item @r{(C++ only)}
2613 A non-static reference or non-static @samp{const} member appears in a
2614 class without constructors.
2616 @item @r{(C++ only)}
2617 Ambiguous virtual bases.
2619 @item @r{(C++ only)}
2620 Subscripting an array which has been declared @samp{register}.
2622 @item @r{(C++ only)}
2623 Taking the address of a variable which has been declared @samp{register}.
2625 @item @r{(C++ only)}
2626 A base class is not initialized in a derived class' copy constructor.
2629 @item -Wno-div-by-zero
2630 @opindex Wno-div-by-zero
2631 @opindex Wdiv-by-zero
2632 Do not warn about compile-time integer division by zero. Floating point
2633 division by zero is not warned about, as it can be a legitimate way of
2634 obtaining infinities and NaNs.
2636 @item -Wsystem-headers
2637 @opindex Wsystem-headers
2638 @cindex warnings from system headers
2639 @cindex system headers, warnings from
2640 Print warning messages for constructs found in system header files.
2641 Warnings from system headers are normally suppressed, on the assumption
2642 that they usually do not indicate real problems and would only make the
2643 compiler output harder to read. Using this command line option tells
2644 GCC to emit warnings from system headers as if they occurred in user
2645 code. However, note that using @option{-Wall} in conjunction with this
2646 option will @emph{not} warn about unknown pragmas in system
2647 headers---for that, @option{-Wunknown-pragmas} must also be used.
2650 @opindex Wfloat-equal
2651 Warn if floating point values are used in equality comparisons.
2653 The idea behind this is that sometimes it is convenient (for the
2654 programmer) to consider floating-point values as approximations to
2655 infinitely precise real numbers. If you are doing this, then you need
2656 to compute (by analyzing the code, or in some other way) the maximum or
2657 likely maximum error that the computation introduces, and allow for it
2658 when performing comparisons (and when producing output, but that's a
2659 different problem). In particular, instead of testing for equality, you
2660 would check to see whether the two values have ranges that overlap; and
2661 this is done with the relational operators, so equality comparisons are
2664 @item -Wtraditional @r{(C only)}
2665 @opindex Wtraditional
2666 Warn about certain constructs that behave differently in traditional and
2667 ISO C@. Also warn about ISO C constructs that have no traditional C
2668 equivalent, and/or problematic constructs which should be avoided.
2672 Macro parameters that appear within string literals in the macro body.
2673 In traditional C macro replacement takes place within string literals,
2674 but does not in ISO C@.
2677 In traditional C, some preprocessor directives did not exist.
2678 Traditional preprocessors would only consider a line to be a directive
2679 if the @samp{#} appeared in column 1 on the line. Therefore
2680 @option{-Wtraditional} warns about directives that traditional C
2681 understands but would ignore because the @samp{#} does not appear as the
2682 first character on the line. It also suggests you hide directives like
2683 @samp{#pragma} not understood by traditional C by indenting them. Some
2684 traditional implementations would not recognize @samp{#elif}, so it
2685 suggests avoiding it altogether.
2688 A function-like macro that appears without arguments.
2691 The unary plus operator.
2694 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2695 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2696 constants.) Note, these suffixes appear in macros defined in the system
2697 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2698 Use of these macros in user code might normally lead to spurious
2699 warnings, however gcc's integrated preprocessor has enough context to
2700 avoid warning in these cases.
2703 A function declared external in one block and then used after the end of
2707 A @code{switch} statement has an operand of type @code{long}.
2710 A non-@code{static} function declaration follows a @code{static} one.
2711 This construct is not accepted by some traditional C compilers.
2714 The ISO type of an integer constant has a different width or
2715 signedness from its traditional type. This warning is only issued if
2716 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2717 typically represent bit patterns, are not warned about.
2720 Usage of ISO string concatenation is detected.
2723 Initialization of automatic aggregates.
2726 Identifier conflicts with labels. Traditional C lacks a separate
2727 namespace for labels.
2730 Initialization of unions. If the initializer is zero, the warning is
2731 omitted. This is done under the assumption that the zero initializer in
2732 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2733 initializer warnings and relies on default initialization to zero in the
2737 Conversions by prototypes between fixed/floating point values and vice
2738 versa. The absence of these prototypes when compiling with traditional
2739 C would cause serious problems. This is a subset of the possible
2740 conversion warnings, for the full set use @option{-Wconversion}.
2743 Use of ISO C style function definitions. This warning intentionally is
2744 @emph{not} issued for prototype declarations or variadic functions
2745 because these ISO C features will appear in your code when using
2746 libiberty's traditional C compatibility macros, @code{PARAMS} and
2747 @code{VPARAMS}. This warning is also bypassed for nested functions
2748 because that feature is already a gcc extension and thus not relevant to
2749 traditional C compatibility.
2752 @item -Wdeclaration-after-statement @r{(C only)}
2753 @opindex Wdeclaration-after-statement
2754 Warn when a declaration is found after a statement in a block. This
2755 construct, known from C++, was introduced with ISO C99 and is by default
2756 allowed in GCC@. It is not supported by ISO C90 and was not supported by
2757 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
2761 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2763 @item -Wendif-labels
2764 @opindex Wendif-labels
2765 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2769 Warn whenever a local variable shadows another local variable, parameter or
2770 global variable or whenever a built-in function is shadowed.
2772 @item -Wlarger-than-@var{len}
2773 @opindex Wlarger-than
2774 Warn whenever an object of larger than @var{len} bytes is defined.
2776 @item -Wpointer-arith
2777 @opindex Wpointer-arith
2778 Warn about anything that depends on the ``size of'' a function type or
2779 of @code{void}. GNU C assigns these types a size of 1, for
2780 convenience in calculations with @code{void *} pointers and pointers
2783 @item -Wbad-function-cast @r{(C only)}
2784 @opindex Wbad-function-cast
2785 Warn whenever a function call is cast to a non-matching type.
2786 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2790 Warn whenever a pointer is cast so as to remove a type qualifier from
2791 the target type. For example, warn if a @code{const char *} is cast
2792 to an ordinary @code{char *}.
2795 @opindex Wcast-align
2796 Warn whenever a pointer is cast such that the required alignment of the
2797 target is increased. For example, warn if a @code{char *} is cast to
2798 an @code{int *} on machines where integers can only be accessed at
2799 two- or four-byte boundaries.
2801 @item -Wwrite-strings
2802 @opindex Wwrite-strings
2803 When compiling C, give string constants the type @code{const
2804 char[@var{length}]} so that
2805 copying the address of one into a non-@code{const} @code{char *}
2806 pointer will get a warning; when compiling C++, warn about the
2807 deprecated conversion from string constants to @code{char *}.
2808 These warnings will help you find at
2809 compile time code that can try to write into a string constant, but
2810 only if you have been very careful about using @code{const} in
2811 declarations and prototypes. Otherwise, it will just be a nuisance;
2812 this is why we did not make @option{-Wall} request these warnings.
2815 @opindex Wconversion
2816 Warn if a prototype causes a type conversion that is different from what
2817 would happen to the same argument in the absence of a prototype. This
2818 includes conversions of fixed point to floating and vice versa, and
2819 conversions changing the width or signedness of a fixed point argument
2820 except when the same as the default promotion.
2822 Also, warn if a negative integer constant expression is implicitly
2823 converted to an unsigned type. For example, warn about the assignment
2824 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2825 casts like @code{(unsigned) -1}.
2827 @item -Wsign-compare
2828 @opindex Wsign-compare
2829 @cindex warning for comparison of signed and unsigned values
2830 @cindex comparison of signed and unsigned values, warning
2831 @cindex signed and unsigned values, comparison warning
2832 Warn when a comparison between signed and unsigned values could produce
2833 an incorrect result when the signed value is converted to unsigned.
2834 This warning is also enabled by @option{-Wextra}; to get the other warnings
2835 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
2837 @item -Waggregate-return
2838 @opindex Waggregate-return
2839 Warn if any functions that return structures or unions are defined or
2840 called. (In languages where you can return an array, this also elicits
2843 @item -Wstrict-prototypes @r{(C only)}
2844 @opindex Wstrict-prototypes
2845 Warn if a function is declared or defined without specifying the
2846 argument types. (An old-style function definition is permitted without
2847 a warning if preceded by a declaration which specifies the argument
2850 @item -Wold-style-definition @r{(C only)}
2851 @opindex Wold-style-definition
2852 Warn if an old-style function definition is used. A warning is given
2853 even if there is a previous prototype.
2855 @item -Wmissing-prototypes @r{(C only)}
2856 @opindex Wmissing-prototypes
2857 Warn if a global function is defined without a previous prototype
2858 declaration. This warning is issued even if the definition itself
2859 provides a prototype. The aim is to detect global functions that fail
2860 to be declared in header files.
2862 @item -Wmissing-declarations @r{(C only)}
2863 @opindex Wmissing-declarations
2864 Warn if a global function is defined without a previous declaration.
2865 Do so even if the definition itself provides a prototype.
2866 Use this option to detect global functions that are not declared in
2869 @item -Wmissing-noreturn
2870 @opindex Wmissing-noreturn
2871 Warn about functions which might be candidates for attribute @code{noreturn}.
2872 Note these are only possible candidates, not absolute ones. Care should
2873 be taken to manually verify functions actually do not ever return before
2874 adding the @code{noreturn} attribute, otherwise subtle code generation
2875 bugs could be introduced. You will not get a warning for @code{main} in
2876 hosted C environments.
2878 @item -Wmissing-format-attribute
2879 @opindex Wmissing-format-attribute
2881 If @option{-Wformat} is enabled, also warn about functions which might be
2882 candidates for @code{format} attributes. Note these are only possible
2883 candidates, not absolute ones. GCC will guess that @code{format}
2884 attributes might be appropriate for any function that calls a function
2885 like @code{vprintf} or @code{vscanf}, but this might not always be the
2886 case, and some functions for which @code{format} attributes are
2887 appropriate may not be detected. This option has no effect unless
2888 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2890 @item -Wno-multichar
2891 @opindex Wno-multichar
2893 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2894 Usually they indicate a typo in the user's code, as they have
2895 implementation-defined values, and should not be used in portable code.
2897 @item -Wno-deprecated-declarations
2898 @opindex Wno-deprecated-declarations
2899 Do not warn about uses of functions, variables, and types marked as
2900 deprecated by using the @code{deprecated} attribute.
2901 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2902 @pxref{Type Attributes}.)
2906 Warn if a structure is given the packed attribute, but the packed
2907 attribute has no effect on the layout or size of the structure.
2908 Such structures may be mis-aligned for little benefit. For
2909 instance, in this code, the variable @code{f.x} in @code{struct bar}
2910 will be misaligned even though @code{struct bar} does not itself
2911 have the packed attribute:
2918 @} __attribute__((packed));
2928 Warn if padding is included in a structure, either to align an element
2929 of the structure or to align the whole structure. Sometimes when this
2930 happens it is possible to rearrange the fields of the structure to
2931 reduce the padding and so make the structure smaller.
2933 @item -Wredundant-decls
2934 @opindex Wredundant-decls
2935 Warn if anything is declared more than once in the same scope, even in
2936 cases where multiple declaration is valid and changes nothing.
2938 @item -Wnested-externs @r{(C only)}
2939 @opindex Wnested-externs
2940 Warn if an @code{extern} declaration is encountered within a function.
2942 @item -Wunreachable-code
2943 @opindex Wunreachable-code
2944 Warn if the compiler detects that code will never be executed.
2946 This option is intended to warn when the compiler detects that at
2947 least a whole line of source code will never be executed, because
2948 some condition is never satisfied or because it is after a
2949 procedure that never returns.
2951 It is possible for this option to produce a warning even though there
2952 are circumstances under which part of the affected line can be executed,
2953 so care should be taken when removing apparently-unreachable code.
2955 For instance, when a function is inlined, a warning may mean that the
2956 line is unreachable in only one inlined copy of the function.
2958 This option is not made part of @option{-Wall} because in a debugging
2959 version of a program there is often substantial code which checks
2960 correct functioning of the program and is, hopefully, unreachable
2961 because the program does work. Another common use of unreachable
2962 code is to provide behavior which is selectable at compile-time.
2966 Warn if a function can not be inlined and it was declared as inline.
2967 Even with this option, the compiler will not warn about failures to
2968 inline functions declared in system headers.
2970 The compiler uses a variety of heuristics to determine whether or not
2971 to inline a function. For example, the compiler takes into account
2972 the size of the function being inlined and the the amount of inlining
2973 that has already been done in the current function. Therefore,
2974 seemingly insignificant changes in the source program can cause the
2975 warnings produced by @option{-Winline} to appear or disappear.
2977 @item -Wno-invalid-offsetof @r{(C++ only)}
2978 @opindex Wno-invalid-offsetof
2979 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
2980 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
2981 to a non-POD type is undefined. In existing C++ implementations,
2982 however, @samp{offsetof} typically gives meaningful results even when
2983 applied to certain kinds of non-POD types. (Such as a simple
2984 @samp{struct} that fails to be a POD type only by virtue of having a
2985 constructor.) This flag is for users who are aware that they are
2986 writing nonportable code and who have deliberately chosen to ignore the
2989 The restrictions on @samp{offsetof} may be relaxed in a future version
2990 of the C++ standard.
2993 @opindex Winvalid-pch
2994 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
2995 the search path but can't be used.
2999 @opindex Wno-long-long
3000 Warn if @samp{long long} type is used. This is default. To inhibit
3001 the warning messages, use @option{-Wno-long-long}. Flags
3002 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3003 only when @option{-pedantic} flag is used.
3005 @item -Wdisabled-optimization
3006 @opindex Wdisabled-optimization
3007 Warn if a requested optimization pass is disabled. This warning does
3008 not generally indicate that there is anything wrong with your code; it
3009 merely indicates that GCC's optimizers were unable to handle the code
3010 effectively. Often, the problem is that your code is too big or too
3011 complex; GCC will refuse to optimize programs when the optimization
3012 itself is likely to take inordinate amounts of time.
3016 Make all warnings into errors.
3019 @node Debugging Options
3020 @section Options for Debugging Your Program or GCC
3021 @cindex options, debugging
3022 @cindex debugging information options
3024 GCC has various special options that are used for debugging
3025 either your program or GCC:
3030 Produce debugging information in the operating system's native format
3031 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
3034 On most systems that use stabs format, @option{-g} enables use of extra
3035 debugging information that only GDB can use; this extra information
3036 makes debugging work better in GDB but will probably make other debuggers
3038 refuse to read the program. If you want to control for certain whether
3039 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3040 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
3041 or @option{-gvms} (see below).
3043 Unlike most other C compilers, GCC allows you to use @option{-g} with
3044 @option{-O}. The shortcuts taken by optimized code may occasionally
3045 produce surprising results: some variables you declared may not exist
3046 at all; flow of control may briefly move where you did not expect it;
3047 some statements may not be executed because they compute constant
3048 results or their values were already at hand; some statements may
3049 execute in different places because they were moved out of loops.
3051 Nevertheless it proves possible to debug optimized output. This makes
3052 it reasonable to use the optimizer for programs that might have bugs.
3054 The following options are useful when GCC is generated with the
3055 capability for more than one debugging format.
3059 Produce debugging information for use by GDB@. This means to use the
3060 most expressive format available (DWARF 2, stabs, or the native format
3061 if neither of those are supported), including GDB extensions if at all
3066 Produce debugging information in stabs format (if that is supported),
3067 without GDB extensions. This is the format used by DBX on most BSD
3068 systems. On MIPS, Alpha and System V Release 4 systems this option
3069 produces stabs debugging output which is not understood by DBX or SDB@.
3070 On System V Release 4 systems this option requires the GNU assembler.
3072 @item -feliminate-unused-debug-symbols
3073 @opindex feliminate-unused-debug-symbols
3074 Produce debugging information in stabs format (if that is supported),
3075 for only symbols that are actually used.
3079 Produce debugging information in stabs format (if that is supported),
3080 using GNU extensions understood only by the GNU debugger (GDB)@. The
3081 use of these extensions is likely to make other debuggers crash or
3082 refuse to read the program.
3086 Produce debugging information in COFF format (if that is supported).
3087 This is the format used by SDB on most System V systems prior to
3092 Produce debugging information in XCOFF format (if that is supported).
3093 This is the format used by the DBX debugger on IBM RS/6000 systems.
3097 Produce debugging information in XCOFF format (if that is supported),
3098 using GNU extensions understood only by the GNU debugger (GDB)@. The
3099 use of these extensions is likely to make other debuggers crash or
3100 refuse to read the program, and may cause assemblers other than the GNU
3101 assembler (GAS) to fail with an error.
3105 Produce debugging information in DWARF version 1 format (if that is
3106 supported). This is the format used by SDB on most System V Release 4
3109 This option is deprecated.
3113 Produce debugging information in DWARF version 1 format (if that is
3114 supported), using GNU extensions understood only by the GNU debugger
3115 (GDB)@. The use of these extensions is likely to make other debuggers
3116 crash or refuse to read the program.
3118 This option is deprecated.
3122 Produce debugging information in DWARF version 2 format (if that is
3123 supported). This is the format used by DBX on IRIX 6.
3127 Produce debugging information in VMS debug format (if that is
3128 supported). This is the format used by DEBUG on VMS systems.
3131 @itemx -ggdb@var{level}
3132 @itemx -gstabs@var{level}
3133 @itemx -gcoff@var{level}
3134 @itemx -gxcoff@var{level}
3135 @itemx -gvms@var{level}
3136 Request debugging information and also use @var{level} to specify how
3137 much information. The default level is 2.
3139 Level 1 produces minimal information, enough for making backtraces in
3140 parts of the program that you don't plan to debug. This includes
3141 descriptions of functions and external variables, but no information
3142 about local variables and no line numbers.
3144 Level 3 includes extra information, such as all the macro definitions
3145 present in the program. Some debuggers support macro expansion when
3146 you use @option{-g3}.
3148 Note that in order to avoid confusion between DWARF1 debug level 2,
3149 and DWARF2, neither @option{-gdwarf} nor @option{-gdwarf-2} accept
3150 a concatenated debug level. Instead use an additional @option{-g@var{level}}
3151 option to change the debug level for DWARF1 or DWARF2.
3153 @item -feliminate-dwarf2-dups
3154 @opindex feliminate-dwarf2-dups
3155 Compress DWARF2 debugging information by eliminating duplicated
3156 information about each symbol. This option only makes sense when
3157 generating DWARF2 debugging information with @option{-gdwarf-2}.
3159 @cindex @command{prof}
3162 Generate extra code to write profile information suitable for the
3163 analysis program @command{prof}. You must use this option when compiling
3164 the source files you want data about, and you must also use it when
3167 @cindex @command{gprof}
3170 Generate extra code to write profile information suitable for the
3171 analysis program @command{gprof}. You must use this option when compiling
3172 the source files you want data about, and you must also use it when
3177 Makes the compiler print out each function name as it is compiled, and
3178 print some statistics about each pass when it finishes.
3181 @opindex ftime-report
3182 Makes the compiler print some statistics about the time consumed by each
3183 pass when it finishes.
3186 @opindex fmem-report
3187 Makes the compiler print some statistics about permanent memory
3188 allocation when it finishes.
3190 @item -fprofile-arcs
3191 @opindex fprofile-arcs
3192 Add code so that program flow @dfn{arcs} are instrumented. During
3193 execution the program records how many times each branch and call is
3194 executed and how many times it is taken or returns. When the compiled
3195 program exits it saves this data to a file called
3196 @file{@var{auxname}.gcda} for each source file. The data may be used for
3197 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3198 test coverage analysis (@option{-ftest-coverage}). Each object file's
3199 @var{auxname} is generated from the name of the output file, if
3200 explicitly specified and it is not the final executable, otherwise it is
3201 the basename of the source file. In both cases any suffix is removed
3202 (e.g. @file{foo.gcda} for input file @file{dir/foo.c}, or
3203 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3208 Compile the source files with @option{-fprofile-arcs} plus optimization
3209 and code generation options. For test coverage analysis, use the
3210 additional @option{-ftest-coverage} option. You do not need to profile
3211 every source file in a program.
3214 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3215 (the latter implies the former).
3218 Run the program on a representative workload to generate the arc profile
3219 information. This may be repeated any number of times. You can run
3220 concurrent instances of your program, and provided that the file system
3221 supports locking, the data files will be correctly updated. Also
3222 @code{fork} calls are detected and correctly handled (double counting
3226 For profile-directed optimizations, compile the source files again with
3227 the same optimization and code generation options plus
3228 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3229 Control Optimization}).
3232 For test coverage analysis, use @command{gcov} to produce human readable
3233 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3234 @command{gcov} documentation for further information.
3238 With @option{-fprofile-arcs}, for each function of your program GCC
3239 creates a program flow graph, then finds a spanning tree for the graph.
3240 Only arcs that are not on the spanning tree have to be instrumented: the
3241 compiler adds code to count the number of times that these arcs are
3242 executed. When an arc is the only exit or only entrance to a block, the
3243 instrumentation code can be added to the block; otherwise, a new basic
3244 block must be created to hold the instrumentation code.
3247 @item -ftest-coverage
3248 @opindex ftest-coverage
3249 Produce a notes file that the @command{gcov} code-coverage utility
3250 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3251 show program coverage. Each source file's note file is called
3252 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3253 above for a description of @var{auxname} and instructions on how to
3254 generate test coverage data. Coverage data will match the source files
3255 more closely, if you do not optimize.
3257 @item -d@var{letters}
3259 Says to make debugging dumps during compilation at times specified by
3260 @var{letters}. This is used for debugging the compiler. The file names
3261 for most of the dumps are made by appending a pass number and a word to
3262 the @var{dumpname}. @var{dumpname} is generated from the name of the
3263 output file, if explicitly specified and it is not an executable,
3264 otherwise it is the basename of the source file. In both cases any
3265 suffix is removed (e.g. @file{foo.00.rtl} or @file{foo.01.sibling}).
3266 Here are the possible letters for use in @var{letters}, and their
3272 Annotate the assembler output with miscellaneous debugging information.
3275 Dump after computing branch probabilities, to @file{@var{file}.16.bp}.
3278 Dump after block reordering, to @file{@var{file}.32.bbro}.
3281 Dump after instruction combination, to the file @file{@var{file}.22.combine}.
3284 Dump after the first if conversion, to the file @file{@var{file}.17.ce1}.
3285 Also dump after the second if conversion, to the file @file{@var{file}.23.ce2}.
3288 Dump after branch target load optimization, to to @file{@var{file}.34.btl}.
3289 Also dump after delayed branch scheduling, to @file{@var{file}.37.dbr}.
3292 Dump all macro definitions, at the end of preprocessing, in addition to
3296 Dump after SSA optimizations, to @file{@var{file}.05.ssa} and
3297 @file{@var{file}.010.ussa}.
3300 Dump after the second if conversion, to @file{@var{file}.33.ce3}.
3303 Dump after control and data flow analysis, to @file{@var{file}.15.cfg}.
3304 Also dump after life analysis, to @file{@var{file}.21.life}.
3307 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.11.addressof}.
3310 Dump after global register allocation, to @file{@var{file}.27.greg}.
3313 Dump after GCSE, to @file{@var{file}.12.gcse}.
3314 Also dump after jump bypassing and control flow optimizations, to
3315 @file{@var{file}.14.bypass}.
3318 Dump after finalization of EH handling code, to @file{@var{file}.03.eh}.
3321 Dump after sibling call optimizations, to @file{@var{file}.02.sibling}.
3324 Dump after the first jump optimization, to @file{@var{file}.04.jump}.
3327 Dump after conversion from registers to stack, to @file{@var{file}.36.stack}.
3330 Dump after local register allocation, to @file{@var{file}.26.lreg}.
3333 Dump after loop optimization passes, to @file{@var{file}.13.loop} and
3334 @file{@var{file}.19.loop2}.
3337 Dump after performing the machine dependent reorganization pass, to
3338 @file{@var{file}.37.mach}.
3341 Dump after register renumbering, to @file{@var{file}.31.rnreg}.
3344 Dump after the register move pass, to @file{@var{file}.24.regmove}.
3347 Dump after post-reload optimizations, to @file{@var{file}.28.postreload}.
3350 Dump after RTL generation, to @file{@var{file}.01.rtl}.
3353 Dump after the second scheduling pass, to @file{@var{file}.35.sched2}.
3356 Dump after CSE (including the jump optimization that sometimes follows
3357 CSE), to @file{@var{file}.019.cse}.
3360 Dump after the first scheduling pass, to @file{@var{file}.25.sched}.
3363 Dump after the second CSE pass (including the jump optimization that
3364 sometimes follows CSE), to @file{@var{file}.20.cse2}.
3367 Dump after running tracer, to @file{@var{file}.18.tracer}.
3370 Dump after null pointer elimination pass to @file{@var{file}.018.null}.
3373 Dump callgraph and unit-at-a-time optimization @file{@var{file}.00.unit}.
3376 Dump after the second flow pass, to @file{@var{file}.29.flow2}.
3379 Dump after SSA conditional constant propagation, to
3380 @file{@var{file}.06.ssaccp}.
3383 Dump after SSA dead code elimination, to @file{@var{file}.07.ssadce}.
3386 Dump after the peephole pass, to @file{@var{file}.30.peephole2}.
3389 Produce all the dumps listed above.
3392 Produce a core dump whenever an error occurs.
3395 Print statistics on memory usage, at the end of the run, to
3399 Annotate the assembler output with a comment indicating which
3400 pattern and alternative was used. The length of each instruction is
3404 Dump the RTL in the assembler output as a comment before each instruction.
3405 Also turns on @option{-dp} annotation.
3408 For each of the other indicated dump files (except for
3409 @file{@var{file}.01.rtl}), dump a representation of the control flow graph
3410 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3413 Just generate RTL for a function instead of compiling it. Usually used
3417 Dump debugging information during parsing, to standard error.
3420 @item -fdump-unnumbered
3421 @opindex fdump-unnumbered
3422 When doing debugging dumps (see @option{-d} option above), suppress instruction
3423 numbers and line number note output. This makes it more feasible to
3424 use diff on debugging dumps for compiler invocations with different
3425 options, in particular with and without @option{-g}.
3427 @item -fdump-translation-unit @r{(C and C++ only)}
3428 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3429 @opindex fdump-translation-unit
3430 Dump a representation of the tree structure for the entire translation
3431 unit to a file. The file name is made by appending @file{.tu} to the
3432 source file name. If the @samp{-@var{options}} form is used, @var{options}
3433 controls the details of the dump as described for the
3434 @option{-fdump-tree} options.
3436 @item -fdump-class-hierarchy @r{(C++ only)}
3437 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3438 @opindex fdump-class-hierarchy
3439 Dump a representation of each class's hierarchy and virtual function
3440 table layout to a file. The file name is made by appending @file{.class}
3441 to the source file name. If the @samp{-@var{options}} form is used,
3442 @var{options} controls the details of the dump as described for the
3443 @option{-fdump-tree} options.
3445 @item -fdump-tree-@var{switch} @r{(C++ only)}
3446 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3448 Control the dumping at various stages of processing the intermediate
3449 language tree to a file. The file name is generated by appending a switch
3450 specific suffix to the source file name. If the @samp{-@var{options}}
3451 form is used, @var{options} is a list of @samp{-} separated options that
3452 control the details of the dump. Not all options are applicable to all
3453 dumps, those which are not meaningful will be ignored. The following
3454 options are available
3458 Print the address of each node. Usually this is not meaningful as it
3459 changes according to the environment and source file. Its primary use
3460 is for tying up a dump file with a debug environment.
3462 Inhibit dumping of members of a scope or body of a function merely
3463 because that scope has been reached. Only dump such items when they
3464 are directly reachable by some other path.
3466 Turn on all options.
3469 The following tree dumps are possible:
3472 Dump before any tree based optimization, to @file{@var{file}.original}.
3474 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3476 Dump after function inlining, to @file{@var{file}.inlined}.
3479 @item -frandom-seed=@var{string}
3480 @opindex frandom-string
3481 This option provides a seed that GCC uses when it would otherwise use
3482 random numbers. It is used to generate certain symbol names
3483 that have to be different in every compiled file. It is also used to
3484 place unique stamps in coverage data files and the object files that
3485 produce them. You can use the @option{-frandom-seed} option to produce
3486 reproducibly identical object files.
3488 The @var{string} should be different for every file you compile.
3490 @item -fsched-verbose=@var{n}
3491 @opindex fsched-verbose
3492 On targets that use instruction scheduling, this option controls the
3493 amount of debugging output the scheduler prints. This information is
3494 written to standard error, unless @option{-dS} or @option{-dR} is
3495 specified, in which case it is output to the usual dump
3496 listing file, @file{.sched} or @file{.sched2} respectively. However
3497 for @var{n} greater than nine, the output is always printed to standard
3500 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
3501 same information as @option{-dRS}. For @var{n} greater than one, it
3502 also output basic block probabilities, detailed ready list information
3503 and unit/insn info. For @var{n} greater than two, it includes RTL
3504 at abort point, control-flow and regions info. And for @var{n} over
3505 four, @option{-fsched-verbose} also includes dependence info.
3509 Store the usual ``temporary'' intermediate files permanently; place them
3510 in the current directory and name them based on the source file. Thus,
3511 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3512 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3513 preprocessed @file{foo.i} output file even though the compiler now
3514 normally uses an integrated preprocessor.
3518 Report the CPU time taken by each subprocess in the compilation
3519 sequence. For C source files, this is the compiler proper and assembler
3520 (plus the linker if linking is done). The output looks like this:
3527 The first number on each line is the ``user time,'' that is time spent
3528 executing the program itself. The second number is ``system time,''
3529 time spent executing operating system routines on behalf of the program.
3530 Both numbers are in seconds.
3532 @item -print-file-name=@var{library}
3533 @opindex print-file-name
3534 Print the full absolute name of the library file @var{library} that
3535 would be used when linking---and don't do anything else. With this
3536 option, GCC does not compile or link anything; it just prints the
3539 @item -print-multi-directory
3540 @opindex print-multi-directory
3541 Print the directory name corresponding to the multilib selected by any
3542 other switches present in the command line. This directory is supposed
3543 to exist in @env{GCC_EXEC_PREFIX}.
3545 @item -print-multi-lib
3546 @opindex print-multi-lib
3547 Print the mapping from multilib directory names to compiler switches
3548 that enable them. The directory name is separated from the switches by
3549 @samp{;}, and each switch starts with an @samp{@@} instead of the
3550 @samp{-}, without spaces between multiple switches. This is supposed to
3551 ease shell-processing.
3553 @item -print-prog-name=@var{program}
3554 @opindex print-prog-name
3555 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3557 @item -print-libgcc-file-name
3558 @opindex print-libgcc-file-name
3559 Same as @option{-print-file-name=libgcc.a}.
3561 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3562 but you do want to link with @file{libgcc.a}. You can do
3565 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3568 @item -print-search-dirs
3569 @opindex print-search-dirs
3570 Print the name of the configured installation directory and a list of
3571 program and library directories gcc will search---and don't do anything else.
3573 This is useful when gcc prints the error message
3574 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3575 To resolve this you either need to put @file{cpp0} and the other compiler
3576 components where gcc expects to find them, or you can set the environment
3577 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3578 Don't forget the trailing '/'.
3579 @xref{Environment Variables}.
3582 @opindex dumpmachine
3583 Print the compiler's target machine (for example,
3584 @samp{i686-pc-linux-gnu})---and don't do anything else.
3587 @opindex dumpversion
3588 Print the compiler version (for example, @samp{3.0})---and don't do
3593 Print the compiler's built-in specs---and don't do anything else. (This
3594 is used when GCC itself is being built.) @xref{Spec Files}.
3596 @item -feliminate-unused-debug-types
3597 @opindex feliminate-unused-debug-types
3598 Normally, when producing DWARF2 output, GCC will emit debugging
3599 information for all types declared in a compilation
3600 unit, regardless of whether or not they are actually used
3601 in that compilation unit. Sometimes this is useful, such as
3602 if, in the debugger, you want to cast a value to a type that is
3603 not actually used in your program (but is declared). More often,
3604 however, this results in a significant amount of wasted space.
3605 With this option, GCC will avoid producing debug symbol output
3606 for types that are nowhere used in the source file being compiled.
3609 @node Optimize Options
3610 @section Options That Control Optimization
3611 @cindex optimize options
3612 @cindex options, optimization
3614 These options control various sorts of optimizations.
3616 Without any optimization option, the compiler's goal is to reduce the
3617 cost of compilation and to make debugging produce the expected
3618 results. Statements are independent: if you stop the program with a
3619 breakpoint between statements, you can then assign a new value to any
3620 variable or change the program counter to any other statement in the
3621 function and get exactly the results you would expect from the source
3624 Turning on optimization flags makes the compiler attempt to improve
3625 the performance and/or code size at the expense of compilation time
3626 and possibly the ability to debug the program.
3628 The compiler performs optimisation based on the knowledge it has of
3629 the program. Using the @option{-funit-at-a-time} flag will allow the
3630 compiler to consider information gained from later functions in the
3631 file when compiling a function. Compiling multiple files at once to a
3632 single output file (and using @option{-funit-at-a-time}) will allow
3633 the compiler to use information gained from all of the files when
3634 compiling each of them.
3636 Not all optimizations are controlled directly by a flag. Only
3637 optimizations that have a flag are listed.
3644 Optimize. Optimizing compilation takes somewhat more time, and a lot
3645 more memory for a large function.
3647 With @option{-O}, the compiler tries to reduce code size and execution
3648 time, without performing any optimizations that take a great deal of
3651 @option{-O} turns on the following optimization flags:
3652 @gccoptlist{-fdefer-pop @gol
3653 -fmerge-constants @gol
3655 -floop-optimize @gol
3657 -fif-conversion @gol
3658 -fif-conversion2 @gol
3659 -fdelayed-branch @gol
3660 -fguess-branch-probability @gol
3663 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
3664 where doing so does not interfere with debugging.
3668 Optimize even more. GCC performs nearly all supported optimizations
3669 that do not involve a space-speed tradeoff. The compiler does not
3670 perform loop unrolling or function inlining when you specify @option{-O2}.
3671 As compared to @option{-O}, this option increases both compilation time
3672 and the performance of the generated code.
3674 @option{-O2} turns on all optimization flags specified by @option{-O}. It
3675 also turns on the following optimization flags:
3676 @gccoptlist{-fforce-mem @gol
3677 -foptimize-sibling-calls @gol
3678 -fstrength-reduce @gol
3679 -fcse-follow-jumps -fcse-skip-blocks @gol
3680 -frerun-cse-after-loop -frerun-loop-opt @gol
3681 -fgcse -fgcse-lm -fgcse-sm -fgcse-las @gol
3682 -fdelete-null-pointer-checks @gol
3683 -fexpensive-optimizations @gol
3685 -fschedule-insns -fschedule-insns2 @gol
3686 -fsched-interblock -fsched-spec @gol
3689 -freorder-blocks -freorder-functions @gol
3690 -fstrict-aliasing @gol
3691 -funit-at-a-time -fweb @gol
3692 -falign-functions -falign-jumps @gol
3693 -falign-loops -falign-labels}
3695 Please note the warning under @option{-fgcse} about
3696 invoking @option{-O2} on programs that use computed gotos.
3700 Optimize yet more. @option{-O3} turns on all optimizations specified by
3701 @option{-O2} and also turns on the @option{-finline-functions},
3702 @option{-fweb} and @option{-frename-registers} options.
3706 Do not optimize. This is the default.
3710 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3711 do not typically increase code size. It also performs further
3712 optimizations designed to reduce code size.
3714 @option{-Os} disables the following optimization flags:
3715 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
3716 -falign-labels -freorder-blocks -fprefetch-loop-arrays}
3718 If you use multiple @option{-O} options, with or without level numbers,
3719 the last such option is the one that is effective.
3722 Options of the form @option{-f@var{flag}} specify machine-independent
3723 flags. Most flags have both positive and negative forms; the negative
3724 form of @option{-ffoo} would be @option{-fno-foo}. In the table
3725 below, only one of the forms is listed---the one you typically will
3726 use. You can figure out the other form by either removing @samp{no-}
3729 The following options control specific optimizations. They are either
3730 activated by @option{-O} options or are related to ones that are. You
3731 can use the following flags in the rare cases when ``fine-tuning'' of
3732 optimizations to be performed is desired.
3735 @item -fno-default-inline
3736 @opindex fno-default-inline
3737 Do not make member functions inline by default merely because they are
3738 defined inside the class scope (C++ only). Otherwise, when you specify
3739 @w{@option{-O}}, member functions defined inside class scope are compiled
3740 inline by default; i.e., you don't need to add @samp{inline} in front of
3741 the member function name.
3743 @item -fno-defer-pop
3744 @opindex fno-defer-pop
3745 Always pop the arguments to each function call as soon as that function
3746 returns. For machines which must pop arguments after a function call,
3747 the compiler normally lets arguments accumulate on the stack for several
3748 function calls and pops them all at once.
3750 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3754 Force memory operands to be copied into registers before doing
3755 arithmetic on them. This produces better code by making all memory
3756 references potential common subexpressions. When they are not common
3757 subexpressions, instruction combination should eliminate the separate
3760 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3763 @opindex fforce-addr
3764 Force memory address constants to be copied into registers before
3765 doing arithmetic on them. This may produce better code just as
3766 @option{-fforce-mem} may.
3768 @item -fomit-frame-pointer
3769 @opindex fomit-frame-pointer
3770 Don't keep the frame pointer in a register for functions that
3771 don't need one. This avoids the instructions to save, set up and
3772 restore frame pointers; it also makes an extra register available
3773 in many functions. @strong{It also makes debugging impossible on
3776 On some machines, such as the VAX, this flag has no effect, because
3777 the standard calling sequence automatically handles the frame pointer
3778 and nothing is saved by pretending it doesn't exist. The
3779 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3780 whether a target machine supports this flag. @xref{Registers,,Register
3781 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3783 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3785 @item -foptimize-sibling-calls
3786 @opindex foptimize-sibling-calls
3787 Optimize sibling and tail recursive calls.
3789 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3793 Don't pay attention to the @code{inline} keyword. Normally this option
3794 is used to keep the compiler from expanding any functions inline.
3795 Note that if you are not optimizing, no functions can be expanded inline.
3797 @item -finline-functions
3798 @opindex finline-functions
3799 Integrate all simple functions into their callers. The compiler
3800 heuristically decides which functions are simple enough to be worth
3801 integrating in this way.
3803 If all calls to a given function are integrated, and the function is
3804 declared @code{static}, then the function is normally not output as
3805 assembler code in its own right.
3807 Enabled at level @option{-O3}.
3809 @item -finline-limit=@var{n}
3810 @opindex finline-limit
3811 By default, gcc limits the size of functions that can be inlined. This flag
3812 allows the control of this limit for functions that are explicitly marked as
3813 inline (i.e., marked with the inline keyword or defined within the class
3814 definition in c++). @var{n} is the size of functions that can be inlined in
3815 number of pseudo instructions (not counting parameter handling). The default
3816 value of @var{n} is 600.
3817 Increasing this value can result in more inlined code at
3818 the cost of compilation time and memory consumption. Decreasing usually makes
3819 the compilation faster and less code will be inlined (which presumably
3820 means slower programs). This option is particularly useful for programs that
3821 use inlining heavily such as those based on recursive templates with C++.
3823 Inlining is actually controlled by a number of parameters, which may be
3824 specified individually by using @option{--param @var{name}=@var{value}}.
3825 The @option{-finline-limit=@var{n}} option sets some of these parameters
3829 @item max-inline-insns
3831 @item max-inline-insns-single
3832 is set to @var{n}/2.
3833 @item max-inline-insns-auto
3834 is set to @var{n}/2.
3835 @item min-inline-insns
3836 is set to 130 or @var{n}/4, whichever is smaller.
3837 @item max-inline-insns-rtl
3841 Using @option{-finline-limit=600} thus results in the default settings
3842 for these parameters. See below for a documentation of the individual
3843 parameters controlling inlining.
3845 @emph{Note:} pseudo instruction represents, in this particular context, an
3846 abstract measurement of function's size. In no way, it represents a count
3847 of assembly instructions and as such its exact meaning might change from one
3848 release to an another.
3850 @item -fkeep-inline-functions
3851 @opindex fkeep-inline-functions
3852 Even if all calls to a given function are integrated, and the function
3853 is declared @code{static}, nevertheless output a separate run-time
3854 callable version of the function. This switch does not affect
3855 @code{extern inline} functions.
3857 @item -fkeep-static-consts
3858 @opindex fkeep-static-consts
3859 Emit variables declared @code{static const} when optimization isn't turned
3860 on, even if the variables aren't referenced.
3862 GCC enables this option by default. If you want to force the compiler to
3863 check if the variable was referenced, regardless of whether or not
3864 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3866 @item -fmerge-constants
3867 Attempt to merge identical constants (string constants and floating point
3868 constants) across compilation units.
3870 This option is the default for optimized compilation if the assembler and
3871 linker support it. Use @option{-fno-merge-constants} to inhibit this
3874 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3876 @item -fmerge-all-constants
3877 Attempt to merge identical constants and identical variables.
3879 This option implies @option{-fmerge-constants}. In addition to
3880 @option{-fmerge-constants} this considers e.g. even constant initialized
3881 arrays or initialized constant variables with integral or floating point
3882 types. Languages like C or C++ require each non-automatic variable to
3883 have distinct location, so using this option will result in non-conforming
3888 Use a graph coloring register allocator. Currently this option is meant
3889 for testing, so we are interested to hear about miscompilations with
3892 @item -fno-branch-count-reg
3893 @opindex fno-branch-count-reg
3894 Do not use ``decrement and branch'' instructions on a count register,
3895 but instead generate a sequence of instructions that decrement a
3896 register, compare it against zero, then branch based upon the result.
3897 This option is only meaningful on architectures that support such
3898 instructions, which include x86, PowerPC, IA-64 and S/390.
3900 The default is @option{-fbranch-count-reg}, enabled when
3901 @option{-fstrength-reduce} is enabled.
3903 @item -fno-function-cse
3904 @opindex fno-function-cse
3905 Do not put function addresses in registers; make each instruction that
3906 calls a constant function contain the function's address explicitly.
3908 This option results in less efficient code, but some strange hacks
3909 that alter the assembler output may be confused by the optimizations
3910 performed when this option is not used.
3912 The default is @option{-ffunction-cse}
3914 @item -fno-zero-initialized-in-bss
3915 @opindex fno-zero-initialized-in-bss
3916 If the target supports a BSS section, GCC by default puts variables that
3917 are initialized to zero into BSS@. This can save space in the resulting
3920 This option turns off this behavior because some programs explicitly
3921 rely on variables going to the data section. E.g., so that the
3922 resulting executable can find the beginning of that section and/or make
3923 assumptions based on that.
3925 The default is @option{-fzero-initialized-in-bss}.
3927 @item -fstrength-reduce
3928 @opindex fstrength-reduce
3929 Perform the optimizations of loop strength reduction and
3930 elimination of iteration variables.
3932 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3934 @item -fthread-jumps
3935 @opindex fthread-jumps
3936 Perform optimizations where we check to see if a jump branches to a
3937 location where another comparison subsumed by the first is found. If
3938 so, the first branch is redirected to either the destination of the
3939 second branch or a point immediately following it, depending on whether
3940 the condition is known to be true or false.
3942 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3944 @item -fcse-follow-jumps
3945 @opindex fcse-follow-jumps
3946 In common subexpression elimination, scan through jump instructions
3947 when the target of the jump is not reached by any other path. For
3948 example, when CSE encounters an @code{if} statement with an
3949 @code{else} clause, CSE will follow the jump when the condition
3952 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3954 @item -fcse-skip-blocks
3955 @opindex fcse-skip-blocks
3956 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3957 follow jumps which conditionally skip over blocks. When CSE
3958 encounters a simple @code{if} statement with no else clause,
3959 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3960 body of the @code{if}.
3962 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3964 @item -frerun-cse-after-loop
3965 @opindex frerun-cse-after-loop
3966 Re-run common subexpression elimination after loop optimizations has been
3969 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3971 @item -frerun-loop-opt
3972 @opindex frerun-loop-opt
3973 Run the loop optimizer twice.
3975 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3979 Perform a global common subexpression elimination pass.
3980 This pass also performs global constant and copy propagation.
3982 @emph{Note:} When compiling a program using computed gotos, a GCC
3983 extension, you may get better runtime performance if you disable
3984 the global common subexpression elimination pass by adding
3985 @option{-fno-gcse} to the command line.
3987 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3991 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3992 attempt to move loads which are only killed by stores into themselves. This
3993 allows a loop containing a load/store sequence to be changed to a load outside
3994 the loop, and a copy/store within the loop.
3996 Enabled by default when gcse is enabled.
4000 When @option{-fgcse-sm} is enabled, a store motion pass is run after
4001 global common subexpression elimination. This pass will attempt to move
4002 stores out of loops. When used in conjunction with @option{-fgcse-lm},
4003 loops containing a load/store sequence can be changed to a load before
4004 the loop and a store after the loop.
4006 Enabled by default when gcse is enabled.
4010 When @option{-fgcse-las} is enabled, the global common subexpression
4011 elimination pass eliminates redundant loads that come after stores to the
4012 same memory location (both partial and full redundancies).
4014 Enabled by default when gcse is enabled.
4016 @item -floop-optimize
4017 @opindex floop-optimize
4018 Perform loop optimizations: move constant expressions out of loops, simplify
4019 exit test conditions and optionally do strength-reduction and loop unrolling as
4022 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4024 @item -fcrossjumping
4025 @opindex crossjumping
4026 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
4027 resulting code may or may not perform better than without cross-jumping.
4029 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4031 @item -fif-conversion
4032 @opindex if-conversion
4033 Attempt to transform conditional jumps into branch-less equivalents. This
4034 include use of conditional moves, min, max, set flags and abs instructions, and
4035 some tricks doable by standard arithmetics. The use of conditional execution
4036 on chips where it is available is controlled by @code{if-conversion2}.
4038 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4040 @item -fif-conversion2
4041 @opindex if-conversion2
4042 Use conditional execution (where available) to transform conditional jumps into
4043 branch-less equivalents.
4045 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4047 @item -fdelete-null-pointer-checks
4048 @opindex fdelete-null-pointer-checks
4049 Use global dataflow analysis to identify and eliminate useless checks
4050 for null pointers. The compiler assumes that dereferencing a null
4051 pointer would have halted the program. If a pointer is checked after
4052 it has already been dereferenced, it cannot be null.
4054 In some environments, this assumption is not true, and programs can
4055 safely dereference null pointers. Use
4056 @option{-fno-delete-null-pointer-checks} to disable this optimization
4057 for programs which depend on that behavior.
4059 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4061 @item -fexpensive-optimizations
4062 @opindex fexpensive-optimizations
4063 Perform a number of minor optimizations that are relatively expensive.
4065 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4067 @item -foptimize-register-move
4069 @opindex foptimize-register-move
4071 Attempt to reassign register numbers in move instructions and as
4072 operands of other simple instructions in order to maximize the amount of
4073 register tying. This is especially helpful on machines with two-operand
4076 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
4079 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4081 @item -fdelayed-branch
4082 @opindex fdelayed-branch
4083 If supported for the target machine, attempt to reorder instructions
4084 to exploit instruction slots available after delayed branch
4087 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4089 @item -fschedule-insns
4090 @opindex fschedule-insns
4091 If supported for the target machine, attempt to reorder instructions to
4092 eliminate execution stalls due to required data being unavailable. This
4093 helps machines that have slow floating point or memory load instructions
4094 by allowing other instructions to be issued until the result of the load
4095 or floating point instruction is required.
4097 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4099 @item -fschedule-insns2
4100 @opindex fschedule-insns2
4101 Similar to @option{-fschedule-insns}, but requests an additional pass of
4102 instruction scheduling after register allocation has been done. This is
4103 especially useful on machines with a relatively small number of
4104 registers and where memory load instructions take more than one cycle.
4106 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4108 @item -fno-sched-interblock
4109 @opindex fno-sched-interblock
4110 Don't schedule instructions across basic blocks. This is normally
4111 enabled by default when scheduling before register allocation, i.e.@:
4112 with @option{-fschedule-insns} or at @option{-O2} or higher.
4114 @item -fno-sched-spec
4115 @opindex fno-sched-spec
4116 Don't allow speculative motion of non-load instructions. This is normally
4117 enabled by default when scheduling before register allocation, i.e.@:
4118 with @option{-fschedule-insns} or at @option{-O2} or higher.
4120 @item -fsched-spec-load
4121 @opindex fsched-spec-load
4122 Allow speculative motion of some load instructions. This only makes
4123 sense when scheduling before register allocation, i.e.@: with
4124 @option{-fschedule-insns} or at @option{-O2} or higher.
4126 @item -fsched-spec-load-dangerous
4127 @opindex fsched-spec-load-dangerous
4128 Allow speculative motion of more load instructions. This only makes
4129 sense when scheduling before register allocation, i.e.@: with
4130 @option{-fschedule-insns} or at @option{-O2} or higher.
4132 @item -fsched-stalled-insns=@var{n}
4133 @opindex fsched-stalled-insns
4134 Define how many insns (if any) can be moved prematurely from the queue
4135 of stalled insns into the ready list, during the second scheduling pass.
4137 @item -fsched-stalled-insns-dep=@var{n}
4138 @opindex fsched-stalled-insns-dep
4139 Define how many insn groups (cycles) will be examined for a dependency
4140 on a stalled insn that is candidate for premature removal from the queue
4141 of stalled insns. Has an effect only during the second scheduling pass,
4142 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
4144 @item -fsched2-use-superblocks
4145 @opindex fsched2-use-superblocks
4146 When scheduling after register allocation, do use superblock scheduling
4147 algorithm. Superblock scheduling allows motion across basic block boundaries
4148 resulting on faster schedules. This option is experimental, as not all machine
4149 descriptions used by GCC model the CPU closely enough to avoid unreliable
4150 results from the algorithm.
4152 This only makes sense when scheduling after register allocation, i.e.@: with
4153 @option{-fschedule-insns2} or at @option{-O2} or higher.
4155 @item -fsched2-use-traces
4156 @opindex fsched2-use-traces
4157 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4158 allocation and additionally perform code duplication in order to increase the
4159 size of superblocks using tracer pass. See @option{-ftracer} for details on
4162 This mode should produce faster but significantly longer programs. Also
4163 without @code{-fbranch-probabilities} the traces constructed may not match the
4164 reality and hurt the performance. This only makes
4165 sense when scheduling after register allocation, i.e.@: with
4166 @option{-fschedule-insns2} or at @option{-O2} or higher.
4168 @item -fcaller-saves
4169 @opindex fcaller-saves
4170 Enable values to be allocated in registers that will be clobbered by
4171 function calls, by emitting extra instructions to save and restore the
4172 registers around such calls. Such allocation is done only when it
4173 seems to result in better code than would otherwise be produced.
4175 This option is always enabled by default on certain machines, usually
4176 those which have no call-preserved registers to use instead.
4178 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4180 @item -fmove-all-movables
4181 @opindex fmove-all-movables
4182 Forces all invariant computations in loops to be moved
4185 @item -freduce-all-givs
4186 @opindex freduce-all-givs
4187 Forces all general-induction variables in loops to be
4190 @emph{Note:} When compiling programs written in Fortran,
4191 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
4192 by default when you use the optimizer.
4194 These options may generate better or worse code; results are highly
4195 dependent on the structure of loops within the source code.
4197 These two options are intended to be removed someday, once
4198 they have helped determine the efficacy of various
4199 approaches to improving loop optimizations.
4201 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
4202 know how use of these options affects
4203 the performance of your production code.
4204 We're very interested in code that runs @emph{slower}
4205 when these options are @emph{enabled}.
4208 @itemx -fno-peephole2
4209 @opindex fno-peephole
4210 @opindex fno-peephole2
4211 Disable any machine-specific peephole optimizations. The difference
4212 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
4213 are implemented in the compiler; some targets use one, some use the
4214 other, a few use both.
4216 @option{-fpeephole} is enabled by default.
4217 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4219 @item -fno-guess-branch-probability
4220 @opindex fno-guess-branch-probability
4221 Do not guess branch probabilities using a randomized model.
4223 Sometimes gcc will opt to use a randomized model to guess branch
4224 probabilities, when none are available from either profiling feedback
4225 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
4226 different runs of the compiler on the same program may produce different
4229 In a hard real-time system, people don't want different runs of the
4230 compiler to produce code that has different behavior; minimizing
4231 non-determinism is of paramount import. This switch allows users to
4232 reduce non-determinism, possibly at the expense of inferior
4235 The default is @option{-fguess-branch-probability} at levels
4236 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4238 @item -freorder-blocks
4239 @opindex freorder-blocks
4240 Reorder basic blocks in the compiled function in order to reduce number of
4241 taken branches and improve code locality.
4243 Enabled at levels @option{-O2}, @option{-O3}.
4245 @item -freorder-functions
4246 @opindex freorder-functions
4247 Reorder basic blocks in the compiled function in order to reduce number of
4248 taken branches and improve code locality. This is implemented by using special
4249 subsections @code{text.hot} for most frequently executed functions and
4250 @code{text.unlikely} for unlikely executed functions. Reordering is done by
4251 the linker so object file format must support named sections and linker must
4252 place them in a reasonable way.
4254 Also profile feedback must be available in to make this option effective. See
4255 @option{-fprofile-arcs} for details.
4257 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4259 @item -fstrict-aliasing
4260 @opindex fstrict-aliasing
4261 Allows the compiler to assume the strictest aliasing rules applicable to
4262 the language being compiled. For C (and C++), this activates
4263 optimizations based on the type of expressions. In particular, an
4264 object of one type is assumed never to reside at the same address as an
4265 object of a different type, unless the types are almost the same. For
4266 example, an @code{unsigned int} can alias an @code{int}, but not a
4267 @code{void*} or a @code{double}. A character type may alias any other
4270 Pay special attention to code like this:
4283 The practice of reading from a different union member than the one most
4284 recently written to (called ``type-punning'') is common. Even with
4285 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
4286 is accessed through the union type. So, the code above will work as
4287 expected. However, this code might not:
4298 Every language that wishes to perform language-specific alias analysis
4299 should define a function that computes, given an @code{tree}
4300 node, an alias set for the node. Nodes in different alias sets are not
4301 allowed to alias. For an example, see the C front-end function
4302 @code{c_get_alias_set}.
4304 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4306 @item -falign-functions
4307 @itemx -falign-functions=@var{n}
4308 @opindex falign-functions
4309 Align the start of functions to the next power-of-two greater than
4310 @var{n}, skipping up to @var{n} bytes. For instance,
4311 @option{-falign-functions=32} aligns functions to the next 32-byte
4312 boundary, but @option{-falign-functions=24} would align to the next
4313 32-byte boundary only if this can be done by skipping 23 bytes or less.
4315 @option{-fno-align-functions} and @option{-falign-functions=1} are
4316 equivalent and mean that functions will not be aligned.
4318 Some assemblers only support this flag when @var{n} is a power of two;
4319 in that case, it is rounded up.
4321 If @var{n} is not specified or is zero, use a machine-dependent default.
4323 Enabled at levels @option{-O2}, @option{-O3}.
4325 @item -falign-labels
4326 @itemx -falign-labels=@var{n}
4327 @opindex falign-labels
4328 Align all branch targets to a power-of-two boundary, skipping up to
4329 @var{n} bytes like @option{-falign-functions}. This option can easily
4330 make code slower, because it must insert dummy operations for when the
4331 branch target is reached in the usual flow of the code.
4333 @option{-fno-align-labels} and @option{-falign-labels=1} are
4334 equivalent and mean that labels will not be aligned.
4336 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4337 are greater than this value, then their values are used instead.
4339 If @var{n} is not specified or is zero, use a machine-dependent default
4340 which is very likely to be @samp{1}, meaning no alignment.
4342 Enabled at levels @option{-O2}, @option{-O3}.
4345 @itemx -falign-loops=@var{n}
4346 @opindex falign-loops
4347 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4348 like @option{-falign-functions}. The hope is that the loop will be
4349 executed many times, which will make up for any execution of the dummy
4352 @option{-fno-align-loops} and @option{-falign-loops=1} are
4353 equivalent and mean that loops will not be aligned.
4355 If @var{n} is not specified or is zero, use a machine-dependent default.
4357 Enabled at levels @option{-O2}, @option{-O3}.
4360 @itemx -falign-jumps=@var{n}
4361 @opindex falign-jumps
4362 Align branch targets to a power-of-two boundary, for branch targets
4363 where the targets can only be reached by jumping, skipping up to @var{n}
4364 bytes like @option{-falign-functions}. In this case, no dummy operations
4367 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
4368 equivalent and mean that loops will not be aligned.
4370 If @var{n} is not specified or is zero, use a machine-dependent default.
4372 Enabled at levels @option{-O2}, @option{-O3}.
4374 @item -frename-registers
4375 @opindex frename-registers
4376 Attempt to avoid false dependencies in scheduled code by making use
4377 of registers left over after register allocation. This optimization
4378 will most benefit processors with lots of registers. It can, however,
4379 make debugging impossible, since variables will no longer stay in
4380 a ``home register''.
4384 Constructs webs as commonly used for register allocation purposes and assign
4385 each web individual pseudo register. This allows our register allocation pass
4386 to operate on pseudos directly, but also strengthens several other optimization
4387 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
4388 however, make debugging impossible, since variables will no longer stay in a
4391 Enabled at levels @option{-O3}.
4393 @item -fno-cprop-registers
4394 @opindex fno-cprop-registers
4395 After register allocation and post-register allocation instruction splitting,
4396 we perform a copy-propagation pass to try to reduce scheduling dependencies
4397 and occasionally eliminate the copy.
4399 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4403 The following options control compiler behavior regarding floating
4404 point arithmetic. These options trade off between speed and
4405 correctness. All must be specifically enabled.
4409 @opindex ffloat-store
4410 Do not store floating point variables in registers, and inhibit other
4411 options that might change whether a floating point value is taken from a
4414 @cindex floating point precision
4415 This option prevents undesirable excess precision on machines such as
4416 the 68000 where the floating registers (of the 68881) keep more
4417 precision than a @code{double} is supposed to have. Similarly for the
4418 x86 architecture. For most programs, the excess precision does only
4419 good, but a few programs rely on the precise definition of IEEE floating
4420 point. Use @option{-ffloat-store} for such programs, after modifying
4421 them to store all pertinent intermediate computations into variables.
4425 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
4426 @option{-fno-trapping-math}, @option{-ffinite-math-only},
4427 @option{-fno-rounding-math} and @option{-fno-signaling-nans}.
4429 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
4431 This option should never be turned on by any @option{-O} option since
4432 it can result in incorrect output for programs which depend on
4433 an exact implementation of IEEE or ISO rules/specifications for
4436 @item -fno-math-errno
4437 @opindex fno-math-errno
4438 Do not set ERRNO after calling math functions that are executed
4439 with a single instruction, e.g., sqrt. A program that relies on
4440 IEEE exceptions for math error handling may want to use this flag
4441 for speed while maintaining IEEE arithmetic compatibility.
4443 This option should never be turned on by any @option{-O} option since
4444 it can result in incorrect output for programs which depend on
4445 an exact implementation of IEEE or ISO rules/specifications for
4448 The default is @option{-fmath-errno}.
4450 @item -funsafe-math-optimizations
4451 @opindex funsafe-math-optimizations
4452 Allow optimizations for floating-point arithmetic that (a) assume
4453 that arguments and results are valid and (b) may violate IEEE or
4454 ANSI standards. When used at link-time, it may include libraries
4455 or startup files that change the default FPU control word or other
4456 similar optimizations.
4458 This option should never be turned on by any @option{-O} option since
4459 it can result in incorrect output for programs which depend on
4460 an exact implementation of IEEE or ISO rules/specifications for
4463 The default is @option{-fno-unsafe-math-optimizations}.
4465 @item -ffinite-math-only
4466 @opindex ffinite-math-only
4467 Allow optimizations for floating-point arithmetic that assume
4468 that arguments and results are not NaNs or +-Infs.
4470 This option should never be turned on by any @option{-O} option since
4471 it can result in incorrect output for programs which depend on
4472 an exact implementation of IEEE or ISO rules/specifications.
4474 The default is @option{-fno-finite-math-only}.
4476 @item -fno-trapping-math
4477 @opindex fno-trapping-math
4478 Compile code assuming that floating-point operations cannot generate
4479 user-visible traps. These traps include division by zero, overflow,
4480 underflow, inexact result and invalid operation. This option implies
4481 @option{-fno-signaling-nans}. Setting this option may allow faster
4482 code if one relies on ``non-stop'' IEEE arithmetic, for example.
4484 This option should never be turned on by any @option{-O} option since
4485 it can result in incorrect output for programs which depend on
4486 an exact implementation of IEEE or ISO rules/specifications for
4489 The default is @option{-ftrapping-math}.
4491 @item -frounding-math
4492 @opindex frounding-math
4493 Disable transformations and optimizations that assume default floating
4494 point rounding behavior. This is round-to-zero for all floating point
4495 to integer conversions, and round-to-nearest for all other arithmetic
4496 truncations. This option should be specified for programs that change
4497 the FP rounding mode dynamically, or that may be executed with a
4498 non-default rounding mode. This option disables constant folding of
4499 floating point expressions at compile-time (which may be affected by
4500 rounding mode) and arithmetic transformations that are unsafe in the
4501 presence of sign-dependent rounding modes.
4503 The default is @option{-fno-rounding-math}.
4505 This option is experimental and does not currently guarantee to
4506 disable all GCC optimizations that are affected by rounding mode.
4507 Future versions of gcc may provide finer control of this setting
4508 using C99's @code{FENV_ACCESS} pragma. This command line option
4509 will be used to specify the default state for @code{FENV_ACCESS}.
4511 @item -fsignaling-nans
4512 @opindex fsignaling-nans
4513 Compile code assuming that IEEE signaling NaNs may generate user-visible
4514 traps during floating-point operations. Setting this option disables
4515 optimizations that may change the number of exceptions visible with
4516 signaling NaNs. This option implies @option{-ftrapping-math}.
4518 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
4521 The default is @option{-fno-signaling-nans}.
4523 This option is experimental and does not currently guarantee to
4524 disable all GCC optimizations that affect signaling NaN behavior.
4526 @item -fsingle-precision-constant
4527 @opindex fsingle-precision-constant
4528 Treat floating point constant as single precision constant instead of
4529 implicitly converting it to double precision constant.
4534 The following options control optimizations that may improve
4535 performance, but are not enabled by any @option{-O} options. This
4536 section includes experimental options that may produce broken code.
4539 @item -fbranch-probabilities
4540 @opindex fbranch-probabilities
4541 After running a program compiled with @option{-fprofile-arcs}
4542 (@pxref{Debugging Options,, Options for Debugging Your Program or
4543 @command{gcc}}), you can compile it a second time using
4544 @option{-fbranch-probabilities}, to improve optimizations based on
4545 the number of times each branch was taken. When the program
4546 compiled with @option{-fprofile-arcs} exits it saves arc execution
4547 counts to a file called @file{@var{sourcename}.gcda} for each source
4548 file The information in this data file is very dependent on the
4549 structure of the generated code, so you must use the same source code
4550 and the same optimization options for both compilations.
4552 With @option{-fbranch-probabilities}, GCC puts a
4553 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
4554 These can be used to improve optimization. Currently, they are only
4555 used in one place: in @file{reorg.c}, instead of guessing which path a
4556 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
4557 exactly determine which path is taken more often.
4559 @item -fprofile-values
4560 @opindex fprofile-values
4561 If combined with @option{-fprofile-arcs}, it adds code so that some
4562 data about values of expressions in the program is gathered.
4564 With @option{-fbranch-probabilities}, it reads back the data gathered
4565 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
4566 notes to instructions for their later usage in optimizations.
4570 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
4571 a code to gather information about values of expressions.
4573 With @option{-fbranch-probabilities}, it reads back the data gathered
4574 and actually performs the optimizations based on them.
4575 Currently the optimizations include specialization of division operation
4576 using the knowledge about the value of the denominator.
4580 Use a graph coloring register allocator. Currently this option is meant
4581 for testing, so we are interested to hear about miscompilations with
4586 Perform tail duplication to enlarge superblock size. This transformation
4587 simplifies the control flow of the function allowing other optimizations to do
4590 @item -funit-at-a-time
4591 @opindex funit-at-a-time
4592 Parse the whole compilation unit before starting to produce code.
4593 This allows some extra optimizations to take place but consumes more
4596 @item -funroll-loops
4597 @opindex funroll-loops
4598 Unroll loops whose number of iterations can be determined at compile time or
4599 upon entry to the loop. @option{-funroll-loops} implies
4600 @option{-frerun-cse-after-loop}. It also turns on complete loop peeling
4601 (i.e. complete removal of loops with small constant number of iterations).
4602 This option makes code larger, and may or may not make it run faster.
4604 @item -funroll-all-loops
4605 @opindex funroll-all-loops
4606 Unroll all loops, even if their number of iterations is uncertain when
4607 the loop is entered. This usually makes programs run more slowly.
4608 @option{-funroll-all-loops} implies the same options as
4609 @option{-funroll-loops}.
4612 @opindex fpeel-loops
4613 Peels the loops for that there is enough information that they do not
4614 roll much (from profile feedback). It also turns on complete loop peeling
4615 (i.e. complete removal of loops with small constant number of iterations).
4617 @item -funswitch-loops
4618 @opindex funswitch-loops
4619 Move branches with loop invariant conditions out of the loop, with duplicates
4620 of the loop on both branches (modified according to result of the condition).
4622 @item -fold-unroll-loops
4623 @opindex fold-unroll-loops
4624 Unroll loops whose number of iterations can be determined at compile
4625 time or upon entry to the loop, using the old loop unroller whose loop
4626 recognition is based on notes from frontend. @option{-fold-unroll-loops} implies
4627 both @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
4628 option makes code larger, and may or may not make it run faster.
4630 @item -fold-unroll-all-loops
4631 @opindex fold-unroll-all-loops
4632 Unroll all loops, even if their number of iterations is uncertain when
4633 the loop is entered. This is done using the old loop unroller whose loop
4634 recognition is based on notes from frontend. This usually makes programs run more slowly.
4635 @option{-fold-unroll-all-loops} implies the same options as
4636 @option{-fold-unroll-loops}.
4638 @item -funswitch-loops
4639 @opindex funswitch-loops
4640 Move branches with loop invariant conditions out of the loop, with duplicates
4641 of the loop on both branches (modified according to result of the condition).
4643 @item -funswitch-loops
4644 @opindex funswitch-loops
4645 Move branches with loop invariant conditions out of the loop, with duplicates
4646 of the loop on both branches (modified according to result of the condition).
4648 @item -fprefetch-loop-arrays
4649 @opindex fprefetch-loop-arrays
4650 If supported by the target machine, generate instructions to prefetch
4651 memory to improve the performance of loops that access large arrays.
4653 Disabled at level @option{-Os}.
4655 @item -ffunction-sections
4656 @itemx -fdata-sections
4657 @opindex ffunction-sections
4658 @opindex fdata-sections
4659 Place each function or data item into its own section in the output
4660 file if the target supports arbitrary sections. The name of the
4661 function or the name of the data item determines the section's name
4664 Use these options on systems where the linker can perform optimizations
4665 to improve locality of reference in the instruction space. Most systems
4666 using the ELF object format and SPARC processors running Solaris 2 have
4667 linkers with such optimizations. AIX may have these optimizations in
4670 Only use these options when there are significant benefits from doing
4671 so. When you specify these options, the assembler and linker will
4672 create larger object and executable files and will also be slower.
4673 You will not be able to use @code{gprof} on all systems if you
4674 specify this option and you may have problems with debugging if
4675 you specify both this option and @option{-g}.
4679 Perform optimizations in static single assignment form. Each function's
4680 flow graph is translated into SSA form, optimizations are performed, and
4681 the flow graph is translated back from SSA form. Users should not
4682 specify this option, since it is not yet ready for production use.
4686 Perform Sparse Conditional Constant Propagation in SSA form. Requires
4687 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
4691 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
4692 Like @option{-fssa}, this is an experimental feature.
4694 @item -fbranch-target-load-optimize
4695 @opindex fbranch-target-load-optimize
4696 Perform branch target register load optimization before prologue / epilogue
4698 The use of target registers can typically be exposed only during reload,
4699 thus hoisting loads out of loops and doing inter-block scheduling needs
4700 a separate optimization pass.
4702 @item -fbranch-target-load-optimize2
4703 @opindex fbranch-target-load-optimize2
4704 Perform branch target register load optimization after prologue / epilogue
4710 @item --param @var{name}=@var{value}
4712 In some places, GCC uses various constants to control the amount of
4713 optimization that is done. For example, GCC will not inline functions
4714 that contain more that a certain number of instructions. You can
4715 control some of these constants on the command-line using the
4716 @option{--param} option.
4718 In each case, the @var{value} is an integer. The allowable choices for
4719 @var{name} are given in the following table:
4722 @item max-crossjump-edges
4723 The maximum number of incoming edges to consider for crossjumping.
4724 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
4725 the number of edges incoming to each block. Increasing values mean
4726 more aggressive optimization, making the compile time increase with
4727 probably small improvement in executable size.
4729 @item max-delay-slot-insn-search
4730 The maximum number of instructions to consider when looking for an
4731 instruction to fill a delay slot. If more than this arbitrary number of
4732 instructions is searched, the time savings from filling the delay slot
4733 will be minimal so stop searching. Increasing values mean more
4734 aggressive optimization, making the compile time increase with probably
4735 small improvement in executable run time.
4737 @item max-delay-slot-live-search
4738 When trying to fill delay slots, the maximum number of instructions to
4739 consider when searching for a block with valid live register
4740 information. Increasing this arbitrarily chosen value means more
4741 aggressive optimization, increasing the compile time. This parameter
4742 should be removed when the delay slot code is rewritten to maintain the
4745 @item max-gcse-memory
4746 The approximate maximum amount of memory that will be allocated in
4747 order to perform the global common subexpression elimination
4748 optimization. If more memory than specified is required, the
4749 optimization will not be done.
4751 @item max-gcse-passes
4752 The maximum number of passes of GCSE to run.
4754 @item max-pending-list-length
4755 The maximum number of pending dependencies scheduling will allow
4756 before flushing the current state and starting over. Large functions
4757 with few branches or calls can create excessively large lists which
4758 needlessly consume memory and resources.
4760 @item max-inline-insns-single
4761 Several parameters control the tree inliner used in gcc.
4762 This number sets the maximum number of instructions (counted in gcc's
4763 internal representation) in a single function that the tree inliner
4764 will consider for inlining. This only affects functions declared
4765 inline and methods implemented in a class declaration (C++).
4766 The default value is 500.
4768 @item max-inline-insns-auto
4769 When you use @option{-finline-functions} (included in @option{-O3}),
4770 a lot of functions that would otherwise not be considered for inlining
4771 by the compiler will be investigated. To those functions, a different
4772 (more restrictive) limit compared to functions declared inline can
4774 The default value is 150.
4776 @item max-inline-insns
4777 The tree inliner does decrease the allowable size for single functions
4778 to be inlined after we already inlined the number of instructions
4779 given here by repeated inlining. This number should be a factor of
4780 two or more larger than the single function limit.
4781 Higher numbers result in better runtime performance, but incur higher
4782 compile-time resource (CPU time, memory) requirements and result in
4783 larger binaries. Very high values are not advisable, as too large
4784 binaries may adversely affect runtime performance.
4785 The default value is 200.
4787 @item max-inline-slope
4788 After exceeding the maximum number of inlined instructions by repeated
4789 inlining, a linear function is used to decrease the allowable size
4790 for single functions. The slope of that function is the negative
4791 reciprocal of the number specified here.
4792 This parameter is ignored when @option{-funit-at-a-time} is used.
4793 The default value is 32.
4795 @item min-inline-insns
4796 The repeated inlining is throttled more and more by the linear function
4797 after exceeding the limit. To avoid too much throttling, a minimum for
4798 this function is specified here to allow repeated inlining for very small
4799 functions even when a lot of repeated inlining already has been done.
4800 This parameter is ignored when @option{-funit-at-a-time} is used.
4801 The default value is 10.
4803 @item large-function-insns
4804 The limit specifying really large functions. For functions greater than this
4805 limit inlining is constrained by @option{--param large-function-growth}.
4806 This parameter is useful primarily to avoid extreme compilation time caused by non-linear
4807 algorithms used by the backend.
4808 This parameter is ignored when @option{-funit-at-a-time} is not used.
4809 The default value is 30000.
4811 @item large-function-growth
4812 Specifies maximal growth of large functtion caused by inlining in percents.
4813 This parameter is ignored when @option{-funit-at-a-time} is not used.
4814 The default value is 200.
4816 @item inline-unit-growth
4817 Specifies maximal overall growth of the compilation unit caused by inlining.
4818 This parameter is ignored when @option{-funit-at-a-time} is not used.
4819 The default value is 150.
4821 @item max-inline-insns-rtl
4822 For languages that use the RTL inliner (this happens at a later stage
4823 than tree inlining), you can set the maximum allowable size (counted
4824 in RTL instructions) for the RTL inliner with this parameter.
4825 The default value is 600.
4828 @item max-unrolled-insns
4829 The maximum number of instructions that a loop should have if that loop
4830 is unrolled, and if the loop is unrolled, it determines how many times
4831 the loop code is unrolled.
4833 @item max-average-unrolled-insns
4834 The maximum number of instructions biased by probabilities of their execution
4835 that a loop should have if that loop is unrolled, and if the loop is unrolled,
4836 it determines how many times the loop code is unrolled.
4838 @item max-unroll-times
4839 The maximum number of unrollings of a single loop.
4841 @item max-peeled-insns
4842 The maximum number of instructions that a loop should have if that loop
4843 is peeled, and if the loop is peeled, it determines how many times
4844 the loop code is peeled.
4846 @item max-peel-times
4847 The maximum number of peelings of a single loop.
4849 @item max-completely-peeled-insns
4850 The maximum number of insns of a completely peeled loop.
4852 @item max-completely-peel-times
4853 The maximum number of iterations of a loop to be suitable for complete peeling.
4855 @item max-unswitch-insns
4856 The maximum number of insns of an unswitched loop.
4858 @item max-unswitch-level
4859 The maximum number of branches unswitched in a single loop.
4861 @item hot-bb-count-fraction
4862 Select fraction of the maximal count of repetitions of basic block in program
4863 given basic block needs to have to be considered hot.
4865 @item hot-bb-frequency-fraction
4866 Select fraction of the maximal frequency of executions of basic block in
4867 function given basic block needs to have to be considered hot
4869 @item tracer-dynamic-coverage
4870 @itemx tracer-dynamic-coverage-feedback
4872 This value is used to limit superblock formation once the given percentage of
4873 executed instructions is covered. This limits unnecessary code size
4876 The @option{tracer-dynamic-coverage-feedback} is used only when profile
4877 feedback is available. The real profiles (as opposed to statically estimated
4878 ones) are much less balanced allowing the threshold to be larger value.
4880 @item tracer-max-code-growth
4881 Stop tail duplication once code growth has reached given percentage. This is
4882 rather hokey argument, as most of the duplicates will be eliminated later in
4883 cross jumping, so it may be set to much higher values than is the desired code
4886 @item tracer-min-branch-ratio
4888 Stop reverse growth when the reverse probability of best edge is less than this
4889 threshold (in percent).
4891 @item tracer-min-branch-ratio
4892 @itemx tracer-min-branch-ratio-feedback
4894 Stop forward growth if the best edge do have probability lower than this
4897 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
4898 compilation for profile feedback and one for compilation without. The value
4899 for compilation with profile feedback needs to be more conservative (higher) in
4900 order to make tracer effective.
4902 @item max-cse-path-length
4904 Maximum number of basic blocks on path that cse considers.
4906 @item ggc-min-expand
4908 GCC uses a garbage collector to manage its own memory allocation. This
4909 parameter specifies the minimum percentage by which the garbage
4910 collector's heap should be allowed to expand between collections.
4911 Tuning this may improve compilation speed; it has no effect on code
4914 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
4915 RAM >= 1GB. If @code{getrlimit} is available, the notion of "RAM" is
4916 the smallest of actual RAM, RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If
4917 GCC is not able to calculate RAM on a particular platform, the lower
4918 bound of 30% is used. Setting this parameter and
4919 @option{ggc-min-heapsize} to zero causes a full collection to occur at
4920 every opportunity. This is extremely slow, but can be useful for
4923 @item ggc-min-heapsize
4925 Minimum size of the garbage collector's heap before it begins bothering
4926 to collect garbage. The first collection occurs after the heap expands
4927 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
4928 tuning this may improve compilation speed, and has no effect on code
4931 The default is RAM/8, with a lower bound of 4096 (four megabytes) and an
4932 upper bound of 131072 (128 megabytes). If @code{getrlimit} is
4933 available, the notion of "RAM" is the smallest of actual RAM,
4934 RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If GCC is not able to calculate
4935 RAM on a particular platform, the lower bound is used. Setting this
4936 parameter very large effectively disables garbage collection. Setting
4937 this parameter and @option{ggc-min-expand} to zero causes a full
4938 collection to occur at every opportunity.
4940 @item reorder-blocks-duplicate
4941 @itemx reorder-blocks-duplicate-feedback
4943 Used by basic block reordering pass to decide whether to use unconditional
4944 branch or duplicate the code on its destination. Code is duplicated when its
4945 estimated size is smaller than this value multiplied by the estimated size of
4946 unconditional jump in the hot spots of the program.
4948 The @option{reorder-block-duplicate-feedback} is used only when profile
4949 feedback is available and may be set to higher values than
4950 @option{reorder-block-duplicate} since information about the hot spots is more
4955 @node Preprocessor Options
4956 @section Options Controlling the Preprocessor
4957 @cindex preprocessor options
4958 @cindex options, preprocessor
4960 These options control the C preprocessor, which is run on each C source
4961 file before actual compilation.
4963 If you use the @option{-E} option, nothing is done except preprocessing.
4964 Some of these options make sense only together with @option{-E} because
4965 they cause the preprocessor output to be unsuitable for actual
4970 You can use @option{-Wp,@var{option}} to bypass the compiler driver
4971 and pass @var{option} directly through to the preprocessor. If
4972 @var{option} contains commas, it is split into multiple options at the
4973 commas. However, many options are modified, translated or interpreted
4974 by the compiler driver before being passed to the preprocessor, and
4975 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
4976 interface is undocumented and subject to change, so whenever possible
4977 you should avoid using @option{-Wp} and let the driver handle the
4980 @item -Xpreprocessor @var{option}
4981 @opindex preprocessor
4982 Pass @var{option} as an option to the preprocessor. You can use this to
4983 supply system-specific preprocessor options which GCC does not know how to
4986 If you want to pass an option that takes an argument, you must use
4987 @option{-Xpreprocessor} twice, once for the option and once for the argument.
4990 @include cppopts.texi
4992 @node Assembler Options
4993 @section Passing Options to the Assembler
4995 @c prevent bad page break with this line
4996 You can pass options to the assembler.
4999 @item -Wa,@var{option}
5001 Pass @var{option} as an option to the assembler. If @var{option}
5002 contains commas, it is split into multiple options at the commas.
5004 @item -Xassembler @var{option}
5006 Pass @var{option} as an option to the assembler. You can use this to
5007 supply system-specific assembler options which GCC does not know how to
5010 If you want to pass an option that takes an argument, you must use
5011 @option{-Xassembler} twice, once for the option and once for the argument.
5016 @section Options for Linking
5017 @cindex link options
5018 @cindex options, linking
5020 These options come into play when the compiler links object files into
5021 an executable output file. They are meaningless if the compiler is
5022 not doing a link step.
5026 @item @var{object-file-name}
5027 A file name that does not end in a special recognized suffix is
5028 considered to name an object file or library. (Object files are
5029 distinguished from libraries by the linker according to the file
5030 contents.) If linking is done, these object files are used as input
5039 If any of these options is used, then the linker is not run, and
5040 object file names should not be used as arguments. @xref{Overall
5044 @item -l@var{library}
5045 @itemx -l @var{library}
5047 Search the library named @var{library} when linking. (The second
5048 alternative with the library as a separate argument is only for
5049 POSIX compliance and is not recommended.)
5051 It makes a difference where in the command you write this option; the
5052 linker searches and processes libraries and object files in the order they
5053 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
5054 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
5055 to functions in @samp{z}, those functions may not be loaded.
5057 The linker searches a standard list of directories for the library,
5058 which is actually a file named @file{lib@var{library}.a}. The linker
5059 then uses this file as if it had been specified precisely by name.
5061 The directories searched include several standard system directories
5062 plus any that you specify with @option{-L}.
5064 Normally the files found this way are library files---archive files
5065 whose members are object files. The linker handles an archive file by
5066 scanning through it for members which define symbols that have so far
5067 been referenced but not defined. But if the file that is found is an
5068 ordinary object file, it is linked in the usual fashion. The only
5069 difference between using an @option{-l} option and specifying a file name
5070 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
5071 and searches several directories.
5075 You need this special case of the @option{-l} option in order to
5076 link an Objective-C program.
5079 @opindex nostartfiles
5080 Do not use the standard system startup files when linking.
5081 The standard system libraries are used normally, unless @option{-nostdlib}
5082 or @option{-nodefaultlibs} is used.
5084 @item -nodefaultlibs
5085 @opindex nodefaultlibs
5086 Do not use the standard system libraries when linking.
5087 Only the libraries you specify will be passed to the linker.
5088 The standard startup files are used normally, unless @option{-nostartfiles}
5089 is used. The compiler may generate calls to memcmp, memset, and memcpy
5090 for System V (and ISO C) environments or to bcopy and bzero for
5091 BSD environments. These entries are usually resolved by entries in
5092 libc. These entry points should be supplied through some other
5093 mechanism when this option is specified.
5097 Do not use the standard system startup files or libraries when linking.
5098 No startup files and only the libraries you specify will be passed to
5099 the linker. The compiler may generate calls to memcmp, memset, and memcpy
5100 for System V (and ISO C) environments or to bcopy and bzero for
5101 BSD environments. These entries are usually resolved by entries in
5102 libc. These entry points should be supplied through some other
5103 mechanism when this option is specified.
5105 @cindex @option{-lgcc}, use with @option{-nostdlib}
5106 @cindex @option{-nostdlib} and unresolved references
5107 @cindex unresolved references and @option{-nostdlib}
5108 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
5109 @cindex @option{-nodefaultlibs} and unresolved references
5110 @cindex unresolved references and @option{-nodefaultlibs}
5111 One of the standard libraries bypassed by @option{-nostdlib} and
5112 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
5113 that GCC uses to overcome shortcomings of particular machines, or special
5114 needs for some languages.
5115 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
5116 Collection (GCC) Internals},
5117 for more discussion of @file{libgcc.a}.)
5118 In most cases, you need @file{libgcc.a} even when you want to avoid
5119 other standard libraries. In other words, when you specify @option{-nostdlib}
5120 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
5121 This ensures that you have no unresolved references to internal GCC
5122 library subroutines. (For example, @samp{__main}, used to ensure C++
5123 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
5124 GNU Compiler Collection (GCC) Internals}.)
5128 Produce a position independent executable on targets which support it.
5129 For predictable results, you must also specify the same set of options
5130 that were used to generate code (@option{-fpie}, @option{-fPIE},
5131 or model suboptions) when you specify this option.
5135 Remove all symbol table and relocation information from the executable.
5139 On systems that support dynamic linking, this prevents linking with the shared
5140 libraries. On other systems, this option has no effect.
5144 Produce a shared object which can then be linked with other objects to
5145 form an executable. Not all systems support this option. For predictable
5146 results, you must also specify the same set of options that were used to
5147 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
5148 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
5149 needs to build supplementary stub code for constructors to work. On
5150 multi-libbed systems, @samp{gcc -shared} must select the correct support
5151 libraries to link against. Failing to supply the correct flags may lead
5152 to subtle defects. Supplying them in cases where they are not necessary
5155 @item -shared-libgcc
5156 @itemx -static-libgcc
5157 @opindex shared-libgcc
5158 @opindex static-libgcc
5159 On systems that provide @file{libgcc} as a shared library, these options
5160 force the use of either the shared or static version respectively.
5161 If no shared version of @file{libgcc} was built when the compiler was
5162 configured, these options have no effect.
5164 There are several situations in which an application should use the
5165 shared @file{libgcc} instead of the static version. The most common
5166 of these is when the application wishes to throw and catch exceptions
5167 across different shared libraries. In that case, each of the libraries
5168 as well as the application itself should use the shared @file{libgcc}.
5170 Therefore, the G++ and GCJ drivers automatically add
5171 @option{-shared-libgcc} whenever you build a shared library or a main
5172 executable, because C++ and Java programs typically use exceptions, so
5173 this is the right thing to do.
5175 If, instead, you use the GCC driver to create shared libraries, you may
5176 find that they will not always be linked with the shared @file{libgcc}.
5177 If GCC finds, at its configuration time, that you have a GNU linker that
5178 does not support option @option{--eh-frame-hdr}, it will link the shared
5179 version of @file{libgcc} into shared libraries by default. Otherwise,
5180 it will take advantage of the linker and optimize away the linking with
5181 the shared version of @file{libgcc}, linking with the static version of
5182 libgcc by default. This allows exceptions to propagate through such
5183 shared libraries, without incurring relocation costs at library load
5186 However, if a library or main executable is supposed to throw or catch
5187 exceptions, you must link it using the G++ or GCJ driver, as appropriate
5188 for the languages used in the program, or using the option
5189 @option{-shared-libgcc}, such that it is linked with the shared
5194 Bind references to global symbols when building a shared object. Warn
5195 about any unresolved references (unless overridden by the link editor
5196 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
5199 @item -Xlinker @var{option}
5201 Pass @var{option} as an option to the linker. You can use this to
5202 supply system-specific linker options which GCC does not know how to
5205 If you want to pass an option that takes an argument, you must use
5206 @option{-Xlinker} twice, once for the option and once for the argument.
5207 For example, to pass @option{-assert definitions}, you must write
5208 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
5209 @option{-Xlinker "-assert definitions"}, because this passes the entire
5210 string as a single argument, which is not what the linker expects.
5212 @item -Wl,@var{option}
5214 Pass @var{option} as an option to the linker. If @var{option} contains
5215 commas, it is split into multiple options at the commas.
5217 @item -u @var{symbol}
5219 Pretend the symbol @var{symbol} is undefined, to force linking of
5220 library modules to define it. You can use @option{-u} multiple times with
5221 different symbols to force loading of additional library modules.
5224 @node Directory Options
5225 @section Options for Directory Search
5226 @cindex directory options
5227 @cindex options, directory search
5230 These options specify directories to search for header files, for
5231 libraries and for parts of the compiler:
5236 Add the directory @var{dir} to the head of the list of directories to be
5237 searched for header files. This can be used to override a system header
5238 file, substituting your own version, since these directories are
5239 searched before the system header file directories. However, you should
5240 not use this option to add directories that contain vendor-supplied
5241 system header files (use @option{-isystem} for that). If you use more than
5242 one @option{-I} option, the directories are scanned in left-to-right
5243 order; the standard system directories come after.
5245 If a standard system include directory, or a directory specified with
5246 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
5247 option will be ignored. The directory will still be searched but as a
5248 system directory at its normal position in the system include chain.
5249 This is to ensure that GCC's procedure to fix buggy system headers and
5250 the ordering for the include_next directive are not inadvertently changed.
5251 If you really need to change the search order for system directories,
5252 use the @option{-nostdinc} and/or @option{-isystem} options.
5256 Any directories you specify with @option{-I} options before the @option{-I-}
5257 option are searched only for the case of @samp{#include "@var{file}"};
5258 they are not searched for @samp{#include <@var{file}>}.
5260 If additional directories are specified with @option{-I} options after
5261 the @option{-I-}, these directories are searched for all @samp{#include}
5262 directives. (Ordinarily @emph{all} @option{-I} directories are used
5265 In addition, the @option{-I-} option inhibits the use of the current
5266 directory (where the current input file came from) as the first search
5267 directory for @samp{#include "@var{file}"}. There is no way to
5268 override this effect of @option{-I-}. With @option{-I.} you can specify
5269 searching the directory which was current when the compiler was
5270 invoked. That is not exactly the same as what the preprocessor does
5271 by default, but it is often satisfactory.
5273 @option{-I-} does not inhibit the use of the standard system directories
5274 for header files. Thus, @option{-I-} and @option{-nostdinc} are
5279 Add directory @var{dir} to the list of directories to be searched
5282 @item -B@var{prefix}
5284 This option specifies where to find the executables, libraries,
5285 include files, and data files of the compiler itself.
5287 The compiler driver program runs one or more of the subprograms
5288 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
5289 @var{prefix} as a prefix for each program it tries to run, both with and
5290 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
5292 For each subprogram to be run, the compiler driver first tries the
5293 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
5294 was not specified, the driver tries two standard prefixes, which are
5295 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
5296 those results in a file name that is found, the unmodified program
5297 name is searched for using the directories specified in your
5298 @env{PATH} environment variable.
5300 The compiler will check to see if the path provided by the @option{-B}
5301 refers to a directory, and if necessary it will add a directory
5302 separator character at the end of the path.
5304 @option{-B} prefixes that effectively specify directory names also apply
5305 to libraries in the linker, because the compiler translates these
5306 options into @option{-L} options for the linker. They also apply to
5307 includes files in the preprocessor, because the compiler translates these
5308 options into @option{-isystem} options for the preprocessor. In this case,
5309 the compiler appends @samp{include} to the prefix.
5311 The run-time support file @file{libgcc.a} can also be searched for using
5312 the @option{-B} prefix, if needed. If it is not found there, the two
5313 standard prefixes above are tried, and that is all. The file is left
5314 out of the link if it is not found by those means.
5316 Another way to specify a prefix much like the @option{-B} prefix is to use
5317 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
5320 As a special kludge, if the path provided by @option{-B} is
5321 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
5322 9, then it will be replaced by @file{[dir/]include}. This is to help
5323 with boot-strapping the compiler.
5325 @item -specs=@var{file}
5327 Process @var{file} after the compiler reads in the standard @file{specs}
5328 file, in order to override the defaults that the @file{gcc} driver
5329 program uses when determining what switches to pass to @file{cc1},
5330 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
5331 @option{-specs=@var{file}} can be specified on the command line, and they
5332 are processed in order, from left to right.
5338 @section Specifying subprocesses and the switches to pass to them
5341 @command{gcc} is a driver program. It performs its job by invoking a
5342 sequence of other programs to do the work of compiling, assembling and
5343 linking. GCC interprets its command-line parameters and uses these to
5344 deduce which programs it should invoke, and which command-line options
5345 it ought to place on their command lines. This behavior is controlled
5346 by @dfn{spec strings}. In most cases there is one spec string for each
5347 program that GCC can invoke, but a few programs have multiple spec
5348 strings to control their behavior. The spec strings built into GCC can
5349 be overridden by using the @option{-specs=} command-line switch to specify
5352 @dfn{Spec files} are plaintext files that are used to construct spec
5353 strings. They consist of a sequence of directives separated by blank
5354 lines. The type of directive is determined by the first non-whitespace
5355 character on the line and it can be one of the following:
5358 @item %@var{command}
5359 Issues a @var{command} to the spec file processor. The commands that can
5363 @item %include <@var{file}>
5365 Search for @var{file} and insert its text at the current point in the
5368 @item %include_noerr <@var{file}>
5369 @cindex %include_noerr
5370 Just like @samp{%include}, but do not generate an error message if the include
5371 file cannot be found.
5373 @item %rename @var{old_name} @var{new_name}
5375 Rename the spec string @var{old_name} to @var{new_name}.
5379 @item *[@var{spec_name}]:
5380 This tells the compiler to create, override or delete the named spec
5381 string. All lines after this directive up to the next directive or
5382 blank line are considered to be the text for the spec string. If this
5383 results in an empty string then the spec will be deleted. (Or, if the
5384 spec did not exist, then nothing will happened.) Otherwise, if the spec
5385 does not currently exist a new spec will be created. If the spec does
5386 exist then its contents will be overridden by the text of this
5387 directive, unless the first character of that text is the @samp{+}
5388 character, in which case the text will be appended to the spec.
5390 @item [@var{suffix}]:
5391 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
5392 and up to the next directive or blank line are considered to make up the
5393 spec string for the indicated suffix. When the compiler encounters an
5394 input file with the named suffix, it will processes the spec string in
5395 order to work out how to compile that file. For example:
5402 This says that any input file whose name ends in @samp{.ZZ} should be
5403 passed to the program @samp{z-compile}, which should be invoked with the
5404 command-line switch @option{-input} and with the result of performing the
5405 @samp{%i} substitution. (See below.)
5407 As an alternative to providing a spec string, the text that follows a
5408 suffix directive can be one of the following:
5411 @item @@@var{language}
5412 This says that the suffix is an alias for a known @var{language}. This is
5413 similar to using the @option{-x} command-line switch to GCC to specify a
5414 language explicitly. For example:
5421 Says that .ZZ files are, in fact, C++ source files.
5424 This causes an error messages saying:
5427 @var{name} compiler not installed on this system.
5431 GCC already has an extensive list of suffixes built into it.
5432 This directive will add an entry to the end of the list of suffixes, but
5433 since the list is searched from the end backwards, it is effectively
5434 possible to override earlier entries using this technique.
5438 GCC has the following spec strings built into it. Spec files can
5439 override these strings or create their own. Note that individual
5440 targets can also add their own spec strings to this list.
5443 asm Options to pass to the assembler
5444 asm_final Options to pass to the assembler post-processor
5445 cpp Options to pass to the C preprocessor
5446 cc1 Options to pass to the C compiler
5447 cc1plus Options to pass to the C++ compiler
5448 endfile Object files to include at the end of the link
5449 link Options to pass to the linker
5450 lib Libraries to include on the command line to the linker
5451 libgcc Decides which GCC support library to pass to the linker
5452 linker Sets the name of the linker
5453 predefines Defines to be passed to the C preprocessor
5454 signed_char Defines to pass to CPP to say whether @code{char} is signed
5456 startfile Object files to include at the start of the link
5459 Here is a small example of a spec file:
5465 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
5468 This example renames the spec called @samp{lib} to @samp{old_lib} and
5469 then overrides the previous definition of @samp{lib} with a new one.
5470 The new definition adds in some extra command-line options before
5471 including the text of the old definition.
5473 @dfn{Spec strings} are a list of command-line options to be passed to their
5474 corresponding program. In addition, the spec strings can contain
5475 @samp{%}-prefixed sequences to substitute variable text or to
5476 conditionally insert text into the command line. Using these constructs
5477 it is possible to generate quite complex command lines.
5479 Here is a table of all defined @samp{%}-sequences for spec
5480 strings. Note that spaces are not generated automatically around the
5481 results of expanding these sequences. Therefore you can concatenate them
5482 together or combine them with constant text in a single argument.
5486 Substitute one @samp{%} into the program name or argument.
5489 Substitute the name of the input file being processed.
5492 Substitute the basename of the input file being processed.
5493 This is the substring up to (and not including) the last period
5494 and not including the directory.
5497 This is the same as @samp{%b}, but include the file suffix (text after
5501 Marks the argument containing or following the @samp{%d} as a
5502 temporary file name, so that that file will be deleted if GCC exits
5503 successfully. Unlike @samp{%g}, this contributes no text to the
5506 @item %g@var{suffix}
5507 Substitute a file name that has suffix @var{suffix} and is chosen
5508 once per compilation, and mark the argument in the same way as
5509 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
5510 name is now chosen in a way that is hard to predict even when previously
5511 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
5512 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
5513 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
5514 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
5515 was simply substituted with a file name chosen once per compilation,
5516 without regard to any appended suffix (which was therefore treated
5517 just like ordinary text), making such attacks more likely to succeed.
5519 @item %u@var{suffix}
5520 Like @samp{%g}, but generates a new temporary file name even if
5521 @samp{%u@var{suffix}} was already seen.
5523 @item %U@var{suffix}
5524 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
5525 new one if there is no such last file name. In the absence of any
5526 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
5527 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
5528 would involve the generation of two distinct file names, one
5529 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
5530 simply substituted with a file name chosen for the previous @samp{%u},
5531 without regard to any appended suffix.
5533 @item %j@var{suffix}
5534 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
5535 writable, and if save-temps is off; otherwise, substitute the name
5536 of a temporary file, just like @samp{%u}. This temporary file is not
5537 meant for communication between processes, but rather as a junk
5540 @item %|@var{suffix}
5541 @itemx %m@var{suffix}
5542 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
5543 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
5544 all. These are the two most common ways to instruct a program that it
5545 should read from standard input or write to standard output. If you
5546 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
5547 construct: see for example @file{f/lang-specs.h}.
5549 @item %.@var{SUFFIX}
5550 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
5551 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
5552 terminated by the next space or %.
5555 Marks the argument containing or following the @samp{%w} as the
5556 designated output file of this compilation. This puts the argument
5557 into the sequence of arguments that @samp{%o} will substitute later.
5560 Substitutes the names of all the output files, with spaces
5561 automatically placed around them. You should write spaces
5562 around the @samp{%o} as well or the results are undefined.
5563 @samp{%o} is for use in the specs for running the linker.
5564 Input files whose names have no recognized suffix are not compiled
5565 at all, but they are included among the output files, so they will
5569 Substitutes the suffix for object files. Note that this is
5570 handled specially when it immediately follows @samp{%g, %u, or %U},
5571 because of the need for those to form complete file names. The
5572 handling is such that @samp{%O} is treated exactly as if it had already
5573 been substituted, except that @samp{%g, %u, and %U} do not currently
5574 support additional @var{suffix} characters following @samp{%O} as they would
5575 following, for example, @samp{.o}.
5578 Substitutes the standard macro predefinitions for the
5579 current target machine. Use this when running @code{cpp}.
5582 Like @samp{%p}, but puts @samp{__} before and after the name of each
5583 predefined macro, except for macros that start with @samp{__} or with
5584 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
5588 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
5589 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), and
5590 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
5594 Current argument is the name of a library or startup file of some sort.
5595 Search for that file in a standard list of directories and substitute
5596 the full name found.
5599 Print @var{str} as an error message. @var{str} is terminated by a newline.
5600 Use this when inconsistent options are detected.
5603 Substitute the contents of spec string @var{name} at this point.
5606 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
5608 @item %x@{@var{option}@}
5609 Accumulate an option for @samp{%X}.
5612 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
5616 Output the accumulated assembler options specified by @option{-Wa}.
5619 Output the accumulated preprocessor options specified by @option{-Wp}.
5622 Process the @code{asm} spec. This is used to compute the
5623 switches to be passed to the assembler.
5626 Process the @code{asm_final} spec. This is a spec string for
5627 passing switches to an assembler post-processor, if such a program is
5631 Process the @code{link} spec. This is the spec for computing the
5632 command line passed to the linker. Typically it will make use of the
5633 @samp{%L %G %S %D and %E} sequences.
5636 Dump out a @option{-L} option for each directory that GCC believes might
5637 contain startup files. If the target supports multilibs then the
5638 current multilib directory will be prepended to each of these paths.
5641 Output the multilib directory with directory separators replaced with
5642 @samp{_}. If multilib directories are not set, or the multilib directory is
5643 @file{.} then this option emits nothing.
5646 Process the @code{lib} spec. This is a spec string for deciding which
5647 libraries should be included on the command line to the linker.
5650 Process the @code{libgcc} spec. This is a spec string for deciding
5651 which GCC support library should be included on the command line to the linker.
5654 Process the @code{startfile} spec. This is a spec for deciding which
5655 object files should be the first ones passed to the linker. Typically
5656 this might be a file named @file{crt0.o}.
5659 Process the @code{endfile} spec. This is a spec string that specifies
5660 the last object files that will be passed to the linker.
5663 Process the @code{cpp} spec. This is used to construct the arguments
5664 to be passed to the C preprocessor.
5667 Process the @code{signed_char} spec. This is intended to be used
5668 to tell cpp whether a char is signed. It typically has the definition:
5670 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
5674 Process the @code{cc1} spec. This is used to construct the options to be
5675 passed to the actual C compiler (@samp{cc1}).
5678 Process the @code{cc1plus} spec. This is used to construct the options to be
5679 passed to the actual C++ compiler (@samp{cc1plus}).
5682 Substitute the variable part of a matched option. See below.
5683 Note that each comma in the substituted string is replaced by
5687 Remove all occurrences of @code{-S} from the command line. Note---this
5688 command is position dependent. @samp{%} commands in the spec string
5689 before this one will see @code{-S}, @samp{%} commands in the spec string
5690 after this one will not.
5692 @item %:@var{function}(@var{args})
5693 Call the named function @var{function}, passing it @var{args}.
5694 @var{args} is first processed as a nested spec string, then split
5695 into an argument vector in the usual fashion. The function returns
5696 a string which is processed as if it had appeared literally as part
5697 of the current spec.
5699 The following built-in spec functions are provided:
5702 @item @code{if-exists}
5703 The @code{if-exists} spec function takes one argument, an absolute
5704 pathname to a file. If the file exists, @code{if-exists} returns the
5705 pathname. Here is a small example of its usage:
5709 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
5712 @item @code{if-exists-else}
5713 The @code{if-exists-else} spec function is similar to the @code{if-exists}
5714 spec function, except that it takes two arguments. The first argument is
5715 an absolute pathname to a file. If the file exists, @code{if-exists-else}
5716 returns the pathname. If it does not exist, it returns the second argument.
5717 This way, @code{if-exists-else} can be used to select one file or another,
5718 based on the existence of the first. Here is a small example of its usage:
5722 crt0%O%s %:if-exists(crti%O%s) \
5723 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
5728 Substitutes the @code{-S} switch, if that switch was given to GCC@.
5729 If that switch was not specified, this substitutes nothing. Note that
5730 the leading dash is omitted when specifying this option, and it is
5731 automatically inserted if the substitution is performed. Thus the spec
5732 string @samp{%@{foo@}} would match the command-line option @option{-foo}
5733 and would output the command line option @option{-foo}.
5735 @item %W@{@code{S}@}
5736 Like %@{@code{S}@} but mark last argument supplied within as a file to be
5739 @item %@{@code{S}*@}
5740 Substitutes all the switches specified to GCC whose names start
5741 with @code{-S}, but which also take an argument. This is used for
5742 switches like @option{-o}, @option{-D}, @option{-I}, etc.
5743 GCC considers @option{-o foo} as being
5744 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
5745 text, including the space. Thus two arguments would be generated.
5747 @item %@{@code{S}*&@code{T}*@}
5748 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
5749 (the order of @code{S} and @code{T} in the spec is not significant).
5750 There can be any number of ampersand-separated variables; for each the
5751 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
5753 @item %@{@code{S}:@code{X}@}
5754 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
5756 @item %@{!@code{S}:@code{X}@}
5757 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
5759 @item %@{@code{S}*:@code{X}@}
5760 Substitutes @code{X} if one or more switches whose names start with
5761 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
5762 once, no matter how many such switches appeared. However, if @code{%*}
5763 appears somewhere in @code{X}, then @code{X} will be substituted once
5764 for each matching switch, with the @code{%*} replaced by the part of
5765 that switch that matched the @code{*}.
5767 @item %@{.@code{S}:@code{X}@}
5768 Substitutes @code{X}, if processing a file with suffix @code{S}.
5770 @item %@{!.@code{S}:@code{X}@}
5771 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
5773 @item %@{@code{S}|@code{P}:@code{X}@}
5774 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
5775 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
5776 although they have a stronger binding than the @samp{|}. If @code{%*}
5777 appears in @code{X}, all of the alternatives must be starred, and only
5778 the first matching alternative is substituted.
5780 For example, a spec string like this:
5783 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
5786 will output the following command-line options from the following input
5787 command-line options:
5792 -d fred.c -foo -baz -boggle
5793 -d jim.d -bar -baz -boggle
5796 @item %@{S:X; T:Y; :D@}
5798 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
5799 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
5800 be as many clauses as you need. This may be combined with @code{.},
5801 @code{!}, @code{|}, and @code{*} as needed.
5806 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
5807 construct may contain other nested @samp{%} constructs or spaces, or
5808 even newlines. They are processed as usual, as described above.
5809 Trailing white space in @code{X} is ignored. White space may also
5810 appear anywhere on the left side of the colon in these constructs,
5811 except between @code{.} or @code{*} and the corresponding word.
5813 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
5814 handled specifically in these constructs. If another value of
5815 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
5816 @option{-W} switch is found later in the command line, the earlier
5817 switch value is ignored, except with @{@code{S}*@} where @code{S} is
5818 just one letter, which passes all matching options.
5820 The character @samp{|} at the beginning of the predicate text is used to
5821 indicate that a command should be piped to the following command, but
5822 only if @option{-pipe} is specified.
5824 It is built into GCC which switches take arguments and which do not.
5825 (You might think it would be useful to generalize this to allow each
5826 compiler's spec to say which switches take arguments. But this cannot
5827 be done in a consistent fashion. GCC cannot even decide which input
5828 files have been specified without knowing which switches take arguments,
5829 and it must know which input files to compile in order to tell which
5832 GCC also knows implicitly that arguments starting in @option{-l} are to be
5833 treated as compiler output files, and passed to the linker in their
5834 proper position among the other output files.
5836 @c man begin OPTIONS
5838 @node Target Options
5839 @section Specifying Target Machine and Compiler Version
5840 @cindex target options
5841 @cindex cross compiling
5842 @cindex specifying machine version
5843 @cindex specifying compiler version and target machine
5844 @cindex compiler version, specifying
5845 @cindex target machine, specifying
5847 The usual way to run GCC is to run the executable called @file{gcc}, or
5848 @file{<machine>-gcc} when cross-compiling, or
5849 @file{<machine>-gcc-<version>} to run a version other than the one that
5850 was installed last. Sometimes this is inconvenient, so GCC provides
5851 options that will switch to another cross-compiler or version.
5854 @item -b @var{machine}
5856 The argument @var{machine} specifies the target machine for compilation.
5858 The value to use for @var{machine} is the same as was specified as the
5859 machine type when configuring GCC as a cross-compiler. For
5860 example, if a cross-compiler was configured with @samp{configure
5861 i386v}, meaning to compile for an 80386 running System V, then you
5862 would specify @option{-b i386v} to run that cross compiler.
5864 @item -V @var{version}
5866 The argument @var{version} specifies which version of GCC to run.
5867 This is useful when multiple versions are installed. For example,
5868 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5871 The @option{-V} and @option{-b} options work by running the
5872 @file{<machine>-gcc-<version>} executable, so there's no real reason to
5873 use them if you can just run that directly.
5875 @node Submodel Options
5876 @section Hardware Models and Configurations
5877 @cindex submodel options
5878 @cindex specifying hardware config
5879 @cindex hardware models and configurations, specifying
5880 @cindex machine dependent options
5882 Earlier we discussed the standard option @option{-b} which chooses among
5883 different installed compilers for completely different target
5884 machines, such as VAX vs.@: 68000 vs.@: 80386.
5886 In addition, each of these target machine types can have its own
5887 special options, starting with @samp{-m}, to choose among various
5888 hardware models or configurations---for example, 68010 vs 68020,
5889 floating coprocessor or none. A single installed version of the
5890 compiler can compile for any model or configuration, according to the
5893 Some configurations of the compiler also support additional special
5894 options, usually for compatibility with other compilers on the same
5897 These options are defined by the macro @code{TARGET_SWITCHES} in the
5898 machine description. The default for the options is also defined by
5899 that macro, which enables you to change the defaults.
5911 * RS/6000 and PowerPC Options::
5915 * i386 and x86-64 Options::
5917 * Intel 960 Options::
5918 * DEC Alpha Options::
5919 * DEC Alpha/VMS Options::
5922 * System V Options::
5923 * TMS320C3x/C4x Options::
5931 * S/390 and zSeries Options::
5935 * Xstormy16 Options::
5940 @node M680x0 Options
5941 @subsection M680x0 Options
5942 @cindex M680x0 options
5944 These are the @samp{-m} options defined for the 68000 series. The default
5945 values for these options depends on which style of 68000 was selected when
5946 the compiler was configured; the defaults for the most common choices are
5954 Generate output for a 68000. This is the default
5955 when the compiler is configured for 68000-based systems.
5957 Use this option for microcontrollers with a 68000 or EC000 core,
5958 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5964 Generate output for a 68020. This is the default
5965 when the compiler is configured for 68020-based systems.
5969 Generate output containing 68881 instructions for floating point.
5970 This is the default for most 68020 systems unless @option{--nfp} was
5971 specified when the compiler was configured.
5975 Generate output for a 68030. This is the default when the compiler is
5976 configured for 68030-based systems.
5980 Generate output for a 68040. This is the default when the compiler is
5981 configured for 68040-based systems.
5983 This option inhibits the use of 68881/68882 instructions that have to be
5984 emulated by software on the 68040. Use this option if your 68040 does not
5985 have code to emulate those instructions.
5989 Generate output for a 68060. This is the default when the compiler is
5990 configured for 68060-based systems.
5992 This option inhibits the use of 68020 and 68881/68882 instructions that
5993 have to be emulated by software on the 68060. Use this option if your 68060
5994 does not have code to emulate those instructions.
5998 Generate output for a CPU32. This is the default
5999 when the compiler is configured for CPU32-based systems.
6001 Use this option for microcontrollers with a
6002 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
6003 68336, 68340, 68341, 68349 and 68360.
6007 Generate output for a 520X ``coldfire'' family cpu. This is the default
6008 when the compiler is configured for 520X-based systems.
6010 Use this option for microcontroller with a 5200 core, including
6011 the MCF5202, MCF5203, MCF5204 and MCF5202.
6016 Generate output for a 68040, without using any of the new instructions.
6017 This results in code which can run relatively efficiently on either a
6018 68020/68881 or a 68030 or a 68040. The generated code does use the
6019 68881 instructions that are emulated on the 68040.
6023 Generate output for a 68060, without using any of the new instructions.
6024 This results in code which can run relatively efficiently on either a
6025 68020/68881 or a 68030 or a 68040. The generated code does use the
6026 68881 instructions that are emulated on the 68060.
6029 @opindex msoft-float
6030 Generate output containing library calls for floating point.
6031 @strong{Warning:} the requisite libraries are not available for all m68k
6032 targets. Normally the facilities of the machine's usual C compiler are
6033 used, but this can't be done directly in cross-compilation. You must
6034 make your own arrangements to provide suitable library functions for
6035 cross-compilation. The embedded targets @samp{m68k-*-aout} and
6036 @samp{m68k-*-coff} do provide software floating point support.
6040 Consider type @code{int} to be 16 bits wide, like @code{short int}.
6043 @opindex mnobitfield
6044 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
6045 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
6049 Do use the bit-field instructions. The @option{-m68020} option implies
6050 @option{-mbitfield}. This is the default if you use a configuration
6051 designed for a 68020.
6055 Use a different function-calling convention, in which functions
6056 that take a fixed number of arguments return with the @code{rtd}
6057 instruction, which pops their arguments while returning. This
6058 saves one instruction in the caller since there is no need to pop
6059 the arguments there.
6061 This calling convention is incompatible with the one normally
6062 used on Unix, so you cannot use it if you need to call libraries
6063 compiled with the Unix compiler.
6065 Also, you must provide function prototypes for all functions that
6066 take variable numbers of arguments (including @code{printf});
6067 otherwise incorrect code will be generated for calls to those
6070 In addition, seriously incorrect code will result if you call a
6071 function with too many arguments. (Normally, extra arguments are
6072 harmlessly ignored.)
6074 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
6075 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
6078 @itemx -mno-align-int
6080 @opindex mno-align-int
6081 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
6082 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
6083 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
6084 Aligning variables on 32-bit boundaries produces code that runs somewhat
6085 faster on processors with 32-bit busses at the expense of more memory.
6087 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
6088 align structures containing the above types differently than
6089 most published application binary interface specifications for the m68k.
6093 Use the pc-relative addressing mode of the 68000 directly, instead of
6094 using a global offset table. At present, this option implies @option{-fpic},
6095 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
6096 not presently supported with @option{-mpcrel}, though this could be supported for
6097 68020 and higher processors.
6099 @item -mno-strict-align
6100 @itemx -mstrict-align
6101 @opindex mno-strict-align
6102 @opindex mstrict-align
6103 Do not (do) assume that unaligned memory references will be handled by
6108 @node M68hc1x Options
6109 @subsection M68hc1x Options
6110 @cindex M68hc1x options
6112 These are the @samp{-m} options defined for the 68hc11 and 68hc12
6113 microcontrollers. The default values for these options depends on
6114 which style of microcontroller was selected when the compiler was configured;
6115 the defaults for the most common choices are given below.
6122 Generate output for a 68HC11. This is the default
6123 when the compiler is configured for 68HC11-based systems.
6129 Generate output for a 68HC12. This is the default
6130 when the compiler is configured for 68HC12-based systems.
6136 Generate output for a 68HCS12.
6139 @opindex mauto-incdec
6140 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
6147 Enable the use of 68HC12 min and max instructions.
6150 @itemx -mno-long-calls
6151 @opindex mlong-calls
6152 @opindex mno-long-calls
6153 Treat all calls as being far away (near). If calls are assumed to be
6154 far away, the compiler will use the @code{call} instruction to
6155 call a function and the @code{rtc} instruction for returning.
6159 Consider type @code{int} to be 16 bits wide, like @code{short int}.
6161 @item -msoft-reg-count=@var{count}
6162 @opindex msoft-reg-count
6163 Specify the number of pseudo-soft registers which are used for the
6164 code generation. The maximum number is 32. Using more pseudo-soft
6165 register may or may not result in better code depending on the program.
6166 The default is 4 for 68HC11 and 2 for 68HC12.
6171 @subsection VAX Options
6174 These @samp{-m} options are defined for the VAX:
6179 Do not output certain jump instructions (@code{aobleq} and so on)
6180 that the Unix assembler for the VAX cannot handle across long
6185 Do output those jump instructions, on the assumption that you
6186 will assemble with the GNU assembler.
6190 Output code for g-format floating point numbers instead of d-format.
6194 @subsection SPARC Options
6195 @cindex SPARC options
6197 These @samp{-m} switches are supported on the SPARC:
6202 @opindex mno-app-regs
6204 Specify @option{-mapp-regs} to generate output using the global registers
6205 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
6208 To be fully SVR4 ABI compliant at the cost of some performance loss,
6209 specify @option{-mno-app-regs}. You should compile libraries and system
6210 software with this option.
6215 @opindex mhard-float
6216 Generate output containing floating point instructions. This is the
6222 @opindex msoft-float
6223 Generate output containing library calls for floating point.
6224 @strong{Warning:} the requisite libraries are not available for all SPARC
6225 targets. Normally the facilities of the machine's usual C compiler are
6226 used, but this cannot be done directly in cross-compilation. You must make
6227 your own arrangements to provide suitable library functions for
6228 cross-compilation. The embedded targets @samp{sparc-*-aout} and
6229 @samp{sparclite-*-*} do provide software floating point support.
6231 @option{-msoft-float} changes the calling convention in the output file;
6232 therefore, it is only useful if you compile @emph{all} of a program with
6233 this option. In particular, you need to compile @file{libgcc.a}, the
6234 library that comes with GCC, with @option{-msoft-float} in order for
6237 @item -mhard-quad-float
6238 @opindex mhard-quad-float
6239 Generate output containing quad-word (long double) floating point
6242 @item -msoft-quad-float
6243 @opindex msoft-quad-float
6244 Generate output containing library calls for quad-word (long double)
6245 floating point instructions. The functions called are those specified
6246 in the SPARC ABI@. This is the default.
6248 As of this writing, there are no sparc implementations that have hardware
6249 support for the quad-word floating point instructions. They all invoke
6250 a trap handler for one of these instructions, and then the trap handler
6251 emulates the effect of the instruction. Because of the trap handler overhead,
6252 this is much slower than calling the ABI library routines. Thus the
6253 @option{-msoft-quad-float} option is the default.
6259 With @option{-mflat}, the compiler does not generate save/restore instructions
6260 and will use a ``flat'' or single register window calling convention.
6261 This model uses %i7 as the frame pointer and is compatible with the normal
6262 register window model. Code from either may be intermixed.
6263 The local registers and the input registers (0--5) are still treated as
6264 ``call saved'' registers and will be saved on the stack as necessary.
6266 With @option{-mno-flat} (the default), the compiler emits save/restore
6267 instructions (except for leaf functions) and is the normal mode of operation.
6269 @item -mno-unaligned-doubles
6270 @itemx -munaligned-doubles
6271 @opindex mno-unaligned-doubles
6272 @opindex munaligned-doubles
6273 Assume that doubles have 8 byte alignment. This is the default.
6275 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
6276 alignment only if they are contained in another type, or if they have an
6277 absolute address. Otherwise, it assumes they have 4 byte alignment.
6278 Specifying this option avoids some rare compatibility problems with code
6279 generated by other compilers. It is not the default because it results
6280 in a performance loss, especially for floating point code.
6282 @item -mno-faster-structs
6283 @itemx -mfaster-structs
6284 @opindex mno-faster-structs
6285 @opindex mfaster-structs
6286 With @option{-mfaster-structs}, the compiler assumes that structures
6287 should have 8 byte alignment. This enables the use of pairs of
6288 @code{ldd} and @code{std} instructions for copies in structure
6289 assignment, in place of twice as many @code{ld} and @code{st} pairs.
6290 However, the use of this changed alignment directly violates the SPARC
6291 ABI@. Thus, it's intended only for use on targets where the developer
6292 acknowledges that their resulting code will not be directly in line with
6293 the rules of the ABI@.
6296 @opindex mimpure-text
6297 @option{-mimpure-text}, used in addition to @option{-shared}, tells
6298 the compiler to not pass @option{-z text} to the linker when linking a
6299 shared object. Using this option, you can link position-dependent
6300 code into a shared object.
6302 @option{-mimpure-text} suppresses the ``relocations remain against
6303 allocatable but non-writable sections'' linker error message.
6304 However, the necessary relocations will trigger copy-on-write, and the
6305 shared object is not actually shared across processes. Instead of
6306 using @option{-mimpure-text}, you should compile all source code with
6307 @option{-fpic} or @option{-fPIC}.
6309 This option is only available on SunOS and Solaris.
6315 These two options select variations on the SPARC architecture.
6317 By default (unless specifically configured for the Fujitsu SPARClite),
6318 GCC generates code for the v7 variant of the SPARC architecture.
6320 @option{-mv8} will give you SPARC v8 code. The only difference from v7
6321 code is that the compiler emits the integer multiply and integer
6322 divide instructions which exist in SPARC v8 but not in SPARC v7.
6324 @option{-msparclite} will give you SPARClite code. This adds the integer
6325 multiply, integer divide step and scan (@code{ffs}) instructions which
6326 exist in SPARClite but not in SPARC v7.
6328 These options are deprecated and will be deleted in a future GCC release.
6329 They have been replaced with @option{-mcpu=xxx}.
6334 @opindex msupersparc
6335 These two options select the processor for which the code is optimized.
6337 With @option{-mcypress} (the default), the compiler optimizes code for the
6338 Cypress CY7C602 chip, as used in the SPARCStation/SPARCServer 3xx series.
6339 This is also appropriate for the older SPARCStation 1, 2, IPX etc.
6341 With @option{-msupersparc} the compiler optimizes code for the SuperSPARC cpu, as
6342 used in the SPARCStation 10, 1000 and 2000 series. This flag also enables use
6343 of the full SPARC v8 instruction set.
6345 These options are deprecated and will be deleted in a future GCC release.
6346 They have been replaced with @option{-mcpu=xxx}.
6348 @item -mcpu=@var{cpu_type}
6350 Set the instruction set, register set, and instruction scheduling parameters
6351 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
6352 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
6353 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
6354 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
6357 Default instruction scheduling parameters are used for values that select
6358 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
6359 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
6361 Here is a list of each supported architecture and their supported
6366 v8: supersparc, hypersparc
6367 sparclite: f930, f934, sparclite86x
6369 v9: ultrasparc, ultrasparc3
6372 @item -mtune=@var{cpu_type}
6374 Set the instruction scheduling parameters for machine type
6375 @var{cpu_type}, but do not set the instruction set or register set that the
6376 option @option{-mcpu=@var{cpu_type}} would.
6378 The same values for @option{-mcpu=@var{cpu_type}} can be used for
6379 @option{-mtune=@var{cpu_type}}, but the only useful values are those
6380 that select a particular cpu implementation. Those are @samp{cypress},
6381 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
6382 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
6387 These @samp{-m} switches are supported in addition to the above
6388 on the SPARCLET processor.
6391 @item -mlittle-endian
6392 @opindex mlittle-endian
6393 Generate code for a processor running in little-endian mode.
6397 Treat register @code{%g0} as a normal register.
6398 GCC will continue to clobber it as necessary but will not assume
6399 it always reads as 0.
6401 @item -mbroken-saverestore
6402 @opindex mbroken-saverestore
6403 Generate code that does not use non-trivial forms of the @code{save} and
6404 @code{restore} instructions. Early versions of the SPARCLET processor do
6405 not correctly handle @code{save} and @code{restore} instructions used with
6406 arguments. They correctly handle them used without arguments. A @code{save}
6407 instruction used without arguments increments the current window pointer
6408 but does not allocate a new stack frame. It is assumed that the window
6409 overflow trap handler will properly handle this case as will interrupt
6413 These @samp{-m} switches are supported in addition to the above
6414 on SPARC V9 processors in 64-bit environments.
6417 @item -mlittle-endian
6418 @opindex mlittle-endian
6419 Generate code for a processor running in little-endian mode. It is only
6420 available for a few configurations and most notably not on Solaris.
6426 Generate code for a 32-bit or 64-bit environment.
6427 The 32-bit environment sets int, long and pointer to 32 bits.
6428 The 64-bit environment sets int to 32 bits and long and pointer
6431 @item -mcmodel=medlow
6432 @opindex mcmodel=medlow
6433 Generate code for the Medium/Low code model: the program must be linked
6434 in the low 32 bits of the address space. Pointers are 64 bits.
6435 Programs can be statically or dynamically linked.
6437 @item -mcmodel=medmid
6438 @opindex mcmodel=medmid
6439 Generate code for the Medium/Middle code model: the program must be linked
6440 in the low 44 bits of the address space, the text segment must be less than
6441 2G bytes, and data segment must be within 2G of the text segment.
6442 Pointers are 64 bits.
6444 @item -mcmodel=medany
6445 @opindex mcmodel=medany
6446 Generate code for the Medium/Anywhere code model: the program may be linked
6447 anywhere in the address space, the text segment must be less than
6448 2G bytes, and data segment must be within 2G of the text segment.
6449 Pointers are 64 bits.
6451 @item -mcmodel=embmedany
6452 @opindex mcmodel=embmedany
6453 Generate code for the Medium/Anywhere code model for embedded systems:
6454 assume a 32-bit text and a 32-bit data segment, both starting anywhere
6455 (determined at link time). Register %g4 points to the base of the
6456 data segment. Pointers are still 64 bits.
6457 Programs are statically linked, PIC is not supported.
6460 @itemx -mno-stack-bias
6461 @opindex mstack-bias
6462 @opindex mno-stack-bias
6463 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
6464 frame pointer if present, are offset by @minus{}2047 which must be added back
6465 when making stack frame references.
6466 Otherwise, assume no such offset is present.
6470 @subsection ARM Options
6473 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
6478 @opindex mapcs-frame
6479 Generate a stack frame that is compliant with the ARM Procedure Call
6480 Standard for all functions, even if this is not strictly necessary for
6481 correct execution of the code. Specifying @option{-fomit-frame-pointer}
6482 with this option will cause the stack frames not to be generated for
6483 leaf functions. The default is @option{-mno-apcs-frame}.
6487 This is a synonym for @option{-mapcs-frame}.
6491 Generate code for a processor running with a 26-bit program counter,
6492 and conforming to the function calling standards for the APCS 26-bit
6493 option. This option replaces the @option{-m2} and @option{-m3} options
6494 of previous releases of the compiler.
6498 Generate code for a processor running with a 32-bit program counter,
6499 and conforming to the function calling standards for the APCS 32-bit
6500 option. This option replaces the @option{-m6} option of previous releases
6504 @c not currently implemented
6505 @item -mapcs-stack-check
6506 @opindex mapcs-stack-check
6507 Generate code to check the amount of stack space available upon entry to
6508 every function (that actually uses some stack space). If there is
6509 insufficient space available then either the function
6510 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
6511 called, depending upon the amount of stack space required. The run time
6512 system is required to provide these functions. The default is
6513 @option{-mno-apcs-stack-check}, since this produces smaller code.
6515 @c not currently implemented
6517 @opindex mapcs-float
6518 Pass floating point arguments using the float point registers. This is
6519 one of the variants of the APCS@. This option is recommended if the
6520 target hardware has a floating point unit or if a lot of floating point
6521 arithmetic is going to be performed by the code. The default is
6522 @option{-mno-apcs-float}, since integer only code is slightly increased in
6523 size if @option{-mapcs-float} is used.
6525 @c not currently implemented
6526 @item -mapcs-reentrant
6527 @opindex mapcs-reentrant
6528 Generate reentrant, position independent code. The default is
6529 @option{-mno-apcs-reentrant}.
6532 @item -mthumb-interwork
6533 @opindex mthumb-interwork
6534 Generate code which supports calling between the ARM and Thumb
6535 instruction sets. Without this option the two instruction sets cannot
6536 be reliably used inside one program. The default is
6537 @option{-mno-thumb-interwork}, since slightly larger code is generated
6538 when @option{-mthumb-interwork} is specified.
6540 @item -mno-sched-prolog
6541 @opindex mno-sched-prolog
6542 Prevent the reordering of instructions in the function prolog, or the
6543 merging of those instruction with the instructions in the function's
6544 body. This means that all functions will start with a recognizable set
6545 of instructions (or in fact one of a choice from a small set of
6546 different function prologues), and this information can be used to
6547 locate the start if functions inside an executable piece of code. The
6548 default is @option{-msched-prolog}.
6551 @opindex mhard-float
6552 Generate output containing floating point instructions. This is the
6556 @opindex msoft-float
6557 Generate output containing library calls for floating point.
6558 @strong{Warning:} the requisite libraries are not available for all ARM
6559 targets. Normally the facilities of the machine's usual C compiler are
6560 used, but this cannot be done directly in cross-compilation. You must make
6561 your own arrangements to provide suitable library functions for
6564 @option{-msoft-float} changes the calling convention in the output file;
6565 therefore, it is only useful if you compile @emph{all} of a program with
6566 this option. In particular, you need to compile @file{libgcc.a}, the
6567 library that comes with GCC, with @option{-msoft-float} in order for
6570 @item -mlittle-endian
6571 @opindex mlittle-endian
6572 Generate code for a processor running in little-endian mode. This is
6573 the default for all standard configurations.
6576 @opindex mbig-endian
6577 Generate code for a processor running in big-endian mode; the default is
6578 to compile code for a little-endian processor.
6580 @item -mwords-little-endian
6581 @opindex mwords-little-endian
6582 This option only applies when generating code for big-endian processors.
6583 Generate code for a little-endian word order but a big-endian byte
6584 order. That is, a byte order of the form @samp{32107654}. Note: this
6585 option should only be used if you require compatibility with code for
6586 big-endian ARM processors generated by versions of the compiler prior to
6589 @item -malignment-traps
6590 @opindex malignment-traps
6591 Generate code that will not trap if the MMU has alignment traps enabled.
6592 On ARM architectures prior to ARMv4, there were no instructions to
6593 access half-word objects stored in memory. However, when reading from
6594 memory a feature of the ARM architecture allows a word load to be used,
6595 even if the address is unaligned, and the processor core will rotate the
6596 data as it is being loaded. This option tells the compiler that such
6597 misaligned accesses will cause a MMU trap and that it should instead
6598 synthesize the access as a series of byte accesses. The compiler can
6599 still use word accesses to load half-word data if it knows that the
6600 address is aligned to a word boundary.
6602 This option is ignored when compiling for ARM architecture 4 or later,
6603 since these processors have instructions to directly access half-word
6606 @item -mno-alignment-traps
6607 @opindex mno-alignment-traps
6608 Generate code that assumes that the MMU will not trap unaligned
6609 accesses. This produces better code when the target instruction set
6610 does not have half-word memory operations (i.e.@: implementations prior to
6613 Note that you cannot use this option to access unaligned word objects,
6614 since the processor will only fetch one 32-bit aligned object from
6617 The default setting for most targets is @option{-mno-alignment-traps}, since
6618 this produces better code when there are no half-word memory
6619 instructions available.
6621 @item -mshort-load-bytes
6622 @itemx -mno-short-load-words
6623 @opindex mshort-load-bytes
6624 @opindex mno-short-load-words
6625 These are deprecated aliases for @option{-malignment-traps}.
6627 @item -mno-short-load-bytes
6628 @itemx -mshort-load-words
6629 @opindex mno-short-load-bytes
6630 @opindex mshort-load-words
6631 This are deprecated aliases for @option{-mno-alignment-traps}.
6633 @item -mcpu=@var{name}
6635 This specifies the name of the target ARM processor. GCC uses this name
6636 to determine what kind of instructions it can emit when generating
6637 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6638 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6639 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6640 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6641 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6642 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6643 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6644 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6645 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
6646 @samp{arm1020t}, @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
6648 @itemx -mtune=@var{name}
6650 This option is very similar to the @option{-mcpu=} option, except that
6651 instead of specifying the actual target processor type, and hence
6652 restricting which instructions can be used, it specifies that GCC should
6653 tune the performance of the code as if the target were of the type
6654 specified in this option, but still choosing the instructions that it
6655 will generate based on the cpu specified by a @option{-mcpu=} option.
6656 For some ARM implementations better performance can be obtained by using
6659 @item -march=@var{name}
6661 This specifies the name of the target ARM architecture. GCC uses this
6662 name to determine what kind of instructions it can emit when generating
6663 assembly code. This option can be used in conjunction with or instead
6664 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6665 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6666 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{iwmmxt}, @samp{ep9312}.
6668 @item -mfpe=@var{number}
6669 @itemx -mfp=@var{number}
6672 This specifies the version of the floating point emulation available on
6673 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6674 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6676 @item -mstructure-size-boundary=@var{n}
6677 @opindex mstructure-size-boundary
6678 The size of all structures and unions will be rounded up to a multiple
6679 of the number of bits set by this option. Permissible values are 8 and
6680 32. The default value varies for different toolchains. For the COFF
6681 targeted toolchain the default value is 8. Specifying the larger number
6682 can produce faster, more efficient code, but can also increase the size
6683 of the program. The two values are potentially incompatible. Code
6684 compiled with one value cannot necessarily expect to work with code or
6685 libraries compiled with the other value, if they exchange information
6686 using structures or unions.
6688 @item -mabort-on-noreturn
6689 @opindex mabort-on-noreturn
6690 Generate a call to the function @code{abort} at the end of a
6691 @code{noreturn} function. It will be executed if the function tries to
6695 @itemx -mno-long-calls
6696 @opindex mlong-calls
6697 @opindex mno-long-calls
6698 Tells the compiler to perform function calls by first loading the
6699 address of the function into a register and then performing a subroutine
6700 call on this register. This switch is needed if the target function
6701 will lie outside of the 64 megabyte addressing range of the offset based
6702 version of subroutine call instruction.
6704 Even if this switch is enabled, not all function calls will be turned
6705 into long calls. The heuristic is that static functions, functions
6706 which have the @samp{short-call} attribute, functions that are inside
6707 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6708 definitions have already been compiled within the current compilation
6709 unit, will not be turned into long calls. The exception to this rule is
6710 that weak function definitions, functions with the @samp{long-call}
6711 attribute or the @samp{section} attribute, and functions that are within
6712 the scope of a @samp{#pragma long_calls} directive, will always be
6713 turned into long calls.
6715 This feature is not enabled by default. Specifying
6716 @option{-mno-long-calls} will restore the default behavior, as will
6717 placing the function calls within the scope of a @samp{#pragma
6718 long_calls_off} directive. Note these switches have no effect on how
6719 the compiler generates code to handle function calls via function
6722 @item -mnop-fun-dllimport
6723 @opindex mnop-fun-dllimport
6724 Disable support for the @code{dllimport} attribute.
6726 @item -msingle-pic-base
6727 @opindex msingle-pic-base
6728 Treat the register used for PIC addressing as read-only, rather than
6729 loading it in the prologue for each function. The run-time system is
6730 responsible for initializing this register with an appropriate value
6731 before execution begins.
6733 @item -mpic-register=@var{reg}
6734 @opindex mpic-register
6735 Specify the register to be used for PIC addressing. The default is R10
6736 unless stack-checking is enabled, when R9 is used.
6738 @item -mcirrus-fix-invalid-insns
6739 @opindex mcirrus-fix-invalid-insns
6740 @opindex mno-cirrus-fix-invalid-insns
6741 Insert NOPs into the instruction stream to in order to work around
6742 problems with invalid Maverick instruction combinations. This option
6743 is only valid if the @option{-mcpu=ep9312} option has been used to
6744 enable generation of instructions for the Cirrus Maverick floating
6745 point co-processor. This option is not enabled by default, since the
6746 problem is only present in older Maverick implementations. The default
6747 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
6750 @item -mpoke-function-name
6751 @opindex mpoke-function-name
6752 Write the name of each function into the text section, directly
6753 preceding the function prologue. The generated code is similar to this:
6757 .ascii "arm_poke_function_name", 0
6760 .word 0xff000000 + (t1 - t0)
6761 arm_poke_function_name
6763 stmfd sp!, @{fp, ip, lr, pc@}
6767 When performing a stack backtrace, code can inspect the value of
6768 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6769 location @code{pc - 12} and the top 8 bits are set, then we know that
6770 there is a function name embedded immediately preceding this location
6771 and has length @code{((pc[-3]) & 0xff000000)}.
6775 Generate code for the 16-bit Thumb instruction set. The default is to
6776 use the 32-bit ARM instruction set.
6779 @opindex mtpcs-frame
6780 Generate a stack frame that is compliant with the Thumb Procedure Call
6781 Standard for all non-leaf functions. (A leaf function is one that does
6782 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6784 @item -mtpcs-leaf-frame
6785 @opindex mtpcs-leaf-frame
6786 Generate a stack frame that is compliant with the Thumb Procedure Call
6787 Standard for all leaf functions. (A leaf function is one that does
6788 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6790 @item -mcallee-super-interworking
6791 @opindex mcallee-super-interworking
6792 Gives all externally visible functions in the file being compiled an ARM
6793 instruction set header which switches to Thumb mode before executing the
6794 rest of the function. This allows these functions to be called from
6795 non-interworking code.
6797 @item -mcaller-super-interworking
6798 @opindex mcaller-super-interworking
6799 Allows calls via function pointers (including virtual functions) to
6800 execute correctly regardless of whether the target code has been
6801 compiled for interworking or not. There is a small overhead in the cost
6802 of executing a function pointer if this option is enabled.
6806 @node MN10200 Options
6807 @subsection MN10200 Options
6808 @cindex MN10200 options
6810 These @option{-m} options are defined for Matsushita MN10200 architectures:
6815 Indicate to the linker that it should perform a relaxation optimization pass
6816 to shorten branches, calls and absolute memory addresses. This option only
6817 has an effect when used on the command line for the final link step.
6819 This option makes symbolic debugging impossible.
6822 @node MN10300 Options
6823 @subsection MN10300 Options
6824 @cindex MN10300 options
6826 These @option{-m} options are defined for Matsushita MN10300 architectures:
6831 Generate code to avoid bugs in the multiply instructions for the MN10300
6832 processors. This is the default.
6835 @opindex mno-mult-bug
6836 Do not generate code to avoid bugs in the multiply instructions for the
6841 Generate code which uses features specific to the AM33 processor.
6845 Do not generate code which uses features specific to the AM33 processor. This
6850 Do not link in the C run-time initialization object file.
6854 Indicate to the linker that it should perform a relaxation optimization pass
6855 to shorten branches, calls and absolute memory addresses. This option only
6856 has an effect when used on the command line for the final link step.
6858 This option makes symbolic debugging impossible.
6862 @node M32R/D Options
6863 @subsection M32R/D Options
6864 @cindex M32R/D options
6866 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6871 Generate code for the M32R/X@.
6875 Generate code for the M32R@. This is the default.
6877 @item -mcode-model=small
6878 @opindex mcode-model=small
6879 Assume all objects live in the lower 16MB of memory (so that their addresses
6880 can be loaded with the @code{ld24} instruction), and assume all subroutines
6881 are reachable with the @code{bl} instruction.
6882 This is the default.
6884 The addressability of a particular object can be set with the
6885 @code{model} attribute.
6887 @item -mcode-model=medium
6888 @opindex mcode-model=medium
6889 Assume objects may be anywhere in the 32-bit address space (the compiler
6890 will generate @code{seth/add3} instructions to load their addresses), and
6891 assume all subroutines are reachable with the @code{bl} instruction.
6893 @item -mcode-model=large
6894 @opindex mcode-model=large
6895 Assume objects may be anywhere in the 32-bit address space (the compiler
6896 will generate @code{seth/add3} instructions to load their addresses), and
6897 assume subroutines may not be reachable with the @code{bl} instruction
6898 (the compiler will generate the much slower @code{seth/add3/jl}
6899 instruction sequence).
6902 @opindex msdata=none
6903 Disable use of the small data area. Variables will be put into
6904 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6905 @code{section} attribute has been specified).
6906 This is the default.
6908 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6909 Objects may be explicitly put in the small data area with the
6910 @code{section} attribute using one of these sections.
6913 @opindex msdata=sdata
6914 Put small global and static data in the small data area, but do not
6915 generate special code to reference them.
6919 Put small global and static data in the small data area, and generate
6920 special instructions to reference them.
6924 @cindex smaller data references
6925 Put global and static objects less than or equal to @var{num} bytes
6926 into the small data or bss sections instead of the normal data or bss
6927 sections. The default value of @var{num} is 8.
6928 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6929 for this option to have any effect.
6931 All modules should be compiled with the same @option{-G @var{num}} value.
6932 Compiling with different values of @var{num} may or may not work; if it
6933 doesn't the linker will give an error message---incorrect code will not be
6939 @subsection M88K Options
6940 @cindex M88k options
6942 These @samp{-m} options are defined for Motorola 88k architectures:
6947 Generate code that works well on both the m88100 and the
6952 Generate code that works best for the m88100, but that also
6957 Generate code that works best for the m88110, and may not run
6962 Obsolete option to be removed from the next revision.
6965 @item -midentify-revision
6966 @opindex midentify-revision
6967 @cindex identifying source, compiler (88k)
6968 Include an @code{ident} directive in the assembler output recording the
6969 source file name, compiler name and version, timestamp, and compilation
6972 @item -mno-underscores
6973 @opindex mno-underscores
6974 @cindex underscores, avoiding (88k)
6975 In assembler output, emit symbol names without adding an underscore
6976 character at the beginning of each name. The default is to use an
6977 underscore as prefix on each name.
6979 @item -mocs-debug-info
6980 @itemx -mno-ocs-debug-info
6981 @opindex mocs-debug-info
6982 @opindex mno-ocs-debug-info
6984 @cindex debugging, 88k OCS
6985 Include (or omit) additional debugging information (about registers used
6986 in each stack frame) as specified in the 88open Object Compatibility
6987 Standard, ``OCS''@. This extra information allows debugging of code that
6988 has had the frame pointer eliminated. The default for SVr4 and Delta 88
6989 SVr3.2 is to include this information; other 88k configurations omit this
6990 information by default.
6992 @item -mocs-frame-position
6993 @opindex mocs-frame-position
6994 @cindex register positions in frame (88k)
6995 When emitting COFF debugging information for automatic variables and
6996 parameters stored on the stack, use the offset from the canonical frame
6997 address, which is the stack pointer (register 31) on entry to the
6998 function. The SVr4 and Delta88 SVr3.2, and BCS configurations use
6999 @option{-mocs-frame-position}; other 88k configurations have the default
7000 @option{-mno-ocs-frame-position}.
7002 @item -mno-ocs-frame-position
7003 @opindex mno-ocs-frame-position
7004 @cindex register positions in frame (88k)
7005 When emitting COFF debugging information for automatic variables and
7006 parameters stored on the stack, use the offset from the frame pointer
7007 register (register 30). When this option is in effect, the frame
7008 pointer is not eliminated when debugging information is selected by the
7011 @item -moptimize-arg-area
7012 @opindex moptimize-arg-area
7013 @cindex arguments in frame (88k)
7014 Save space by reorganizing the stack frame. This option generates code
7015 that does not agree with the 88open specifications, but uses less
7018 @itemx -mno-optimize-arg-area
7019 @opindex mno-optimize-arg-area
7020 Do not reorganize the stack frame to save space. This is the default.
7021 The generated conforms to the specification, but uses more memory.
7023 @item -mshort-data-@var{num}
7024 @opindex mshort-data
7025 @cindex smaller data references (88k)
7026 @cindex r0-relative references (88k)
7027 Generate smaller data references by making them relative to @code{r0},
7028 which allows loading a value using a single instruction (rather than the
7029 usual two). You control which data references are affected by
7030 specifying @var{num} with this option. For example, if you specify
7031 @option{-mshort-data-512}, then the data references affected are those
7032 involving displacements of less than 512 bytes.
7033 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
7036 @item -mserialize-volatile
7037 @opindex mserialize-volatile
7038 @itemx -mno-serialize-volatile
7039 @opindex mno-serialize-volatile
7040 @cindex sequential consistency on 88k
7041 Do, or don't, generate code to guarantee sequential consistency
7042 of volatile memory references. By default, consistency is
7045 The order of memory references made by the MC88110 processor does
7046 not always match the order of the instructions requesting those
7047 references. In particular, a load instruction may execute before
7048 a preceding store instruction. Such reordering violates
7049 sequential consistency of volatile memory references, when there
7050 are multiple processors. When consistency must be guaranteed,
7051 GCC generates special instructions, as needed, to force
7052 execution in the proper order.
7054 The MC88100 processor does not reorder memory references and so
7055 always provides sequential consistency. However, by default, GCC
7056 generates the special instructions to guarantee consistency
7057 even when you use @option{-m88100}, so that the code may be run on an
7058 MC88110 processor. If you intend to run your code only on the
7059 MC88100 processor, you may use @option{-mno-serialize-volatile}.
7061 The extra code generated to guarantee consistency may affect the
7062 performance of your application. If you know that you can safely
7063 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
7069 @cindex assembler syntax, 88k
7071 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
7072 related to System V release 4 (SVr4). This controls the following:
7076 Which variant of the assembler syntax to emit.
7078 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
7079 that is used on System V release 4.
7081 @option{-msvr4} makes GCC issue additional declaration directives used in
7085 @option{-msvr4} is the default for the m88k-motorola-sysv4 configuration.
7086 @option{-msvr3} is the default for all other m88k configurations.
7088 @item -mversion-03.00
7089 @opindex mversion-03.00
7090 This option is obsolete, and is ignored.
7091 @c ??? which asm syntax better for GAS? option there too?
7093 @item -mno-check-zero-division
7094 @itemx -mcheck-zero-division
7095 @opindex mno-check-zero-division
7096 @opindex mcheck-zero-division
7097 @cindex zero division on 88k
7098 Do, or don't, generate code to guarantee that integer division by
7099 zero will be detected. By default, detection is guaranteed.
7101 Some models of the MC88100 processor fail to trap upon integer
7102 division by zero under certain conditions. By default, when
7103 compiling code that might be run on such a processor, GCC
7104 generates code that explicitly checks for zero-valued divisors
7105 and traps with exception number 503 when one is detected. Use of
7106 @option{-mno-check-zero-division} suppresses such checking for code
7107 generated to run on an MC88100 processor.
7109 GCC assumes that the MC88110 processor correctly detects all instances
7110 of integer division by zero. When @option{-m88110} is specified, no
7111 explicit checks for zero-valued divisors are generated, and both
7112 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
7115 @item -muse-div-instruction
7116 @opindex muse-div-instruction
7117 @cindex divide instruction, 88k
7118 Use the div instruction for signed integer division on the
7119 MC88100 processor. By default, the div instruction is not used.
7121 On the MC88100 processor the signed integer division instruction
7122 div) traps to the operating system on a negative operand. The
7123 operating system transparently completes the operation, but at a
7124 large cost in execution time. By default, when compiling code
7125 that might be run on an MC88100 processor, GCC emulates signed
7126 integer division using the unsigned integer division instruction
7127 divu), thereby avoiding the large penalty of a trap to the
7128 operating system. Such emulation has its own, smaller, execution
7129 cost in both time and space. To the extent that your code's
7130 important signed integer division operations are performed on two
7131 nonnegative operands, it may be desirable to use the div
7132 instruction directly.
7134 On the MC88110 processor the div instruction (also known as the
7135 divs instruction) processes negative operands without trapping to
7136 the operating system. When @option{-m88110} is specified,
7137 @option{-muse-div-instruction} is ignored, and the div instruction is used
7138 for signed integer division.
7140 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
7141 particular, the behavior of such a division with and without
7142 @option{-muse-div-instruction} may differ.
7144 @item -mtrap-large-shift
7145 @itemx -mhandle-large-shift
7146 @opindex mtrap-large-shift
7147 @opindex mhandle-large-shift
7148 @cindex bit shift overflow (88k)
7149 @cindex large bit shifts (88k)
7150 Include code to detect bit-shifts of more than 31 bits; respectively,
7151 trap such shifts or emit code to handle them properly. By default GCC
7152 makes no special provision for large bit shifts.
7154 @item -mwarn-passed-structs
7155 @opindex mwarn-passed-structs
7156 @cindex structure passing (88k)
7157 Warn when a function passes a struct as an argument or result.
7158 Structure-passing conventions have changed during the evolution of the C
7159 language, and are often the source of portability problems. By default,
7160 GCC issues no such warning.
7163 @c break page here to avoid unsightly interparagraph stretch.
7167 @node RS/6000 and PowerPC Options
7168 @subsection IBM RS/6000 and PowerPC Options
7169 @cindex RS/6000 and PowerPC Options
7170 @cindex IBM RS/6000 and PowerPC Options
7172 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
7180 @itemx -mpowerpc-gpopt
7181 @itemx -mno-powerpc-gpopt
7182 @itemx -mpowerpc-gfxopt
7183 @itemx -mno-powerpc-gfxopt
7185 @itemx -mno-powerpc64
7191 @opindex mno-powerpc
7192 @opindex mpowerpc-gpopt
7193 @opindex mno-powerpc-gpopt
7194 @opindex mpowerpc-gfxopt
7195 @opindex mno-powerpc-gfxopt
7197 @opindex mno-powerpc64
7198 GCC supports two related instruction set architectures for the
7199 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
7200 instructions supported by the @samp{rios} chip set used in the original
7201 RS/6000 systems and the @dfn{PowerPC} instruction set is the
7202 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
7203 the IBM 4xx microprocessors.
7205 Neither architecture is a subset of the other. However there is a
7206 large common subset of instructions supported by both. An MQ
7207 register is included in processors supporting the POWER architecture.
7209 You use these options to specify which instructions are available on the
7210 processor you are using. The default value of these options is
7211 determined when configuring GCC@. Specifying the
7212 @option{-mcpu=@var{cpu_type}} overrides the specification of these
7213 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
7214 rather than the options listed above.
7216 The @option{-mpower} option allows GCC to generate instructions that
7217 are found only in the POWER architecture and to use the MQ register.
7218 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
7219 to generate instructions that are present in the POWER2 architecture but
7220 not the original POWER architecture.
7222 The @option{-mpowerpc} option allows GCC to generate instructions that
7223 are found only in the 32-bit subset of the PowerPC architecture.
7224 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
7225 GCC to use the optional PowerPC architecture instructions in the
7226 General Purpose group, including floating-point square root. Specifying
7227 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
7228 use the optional PowerPC architecture instructions in the Graphics
7229 group, including floating-point select.
7231 The @option{-mpowerpc64} option allows GCC to generate the additional
7232 64-bit instructions that are found in the full PowerPC64 architecture
7233 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
7234 @option{-mno-powerpc64}.
7236 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
7237 will use only the instructions in the common subset of both
7238 architectures plus some special AIX common-mode calls, and will not use
7239 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
7240 permits GCC to use any instruction from either architecture and to
7241 allow use of the MQ register; specify this for the Motorola MPC601.
7243 @item -mnew-mnemonics
7244 @itemx -mold-mnemonics
7245 @opindex mnew-mnemonics
7246 @opindex mold-mnemonics
7247 Select which mnemonics to use in the generated assembler code. With
7248 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
7249 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
7250 assembler mnemonics defined for the POWER architecture. Instructions
7251 defined in only one architecture have only one mnemonic; GCC uses that
7252 mnemonic irrespective of which of these options is specified.
7254 GCC defaults to the mnemonics appropriate for the architecture in
7255 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
7256 value of these option. Unless you are building a cross-compiler, you
7257 should normally not specify either @option{-mnew-mnemonics} or
7258 @option{-mold-mnemonics}, but should instead accept the default.
7260 @item -mcpu=@var{cpu_type}
7262 Set architecture type, register usage, choice of mnemonics, and
7263 instruction scheduling parameters for machine type @var{cpu_type}.
7264 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
7265 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
7266 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
7267 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
7268 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
7269 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
7271 @option{-mcpu=common} selects a completely generic processor. Code
7272 generated under this option will run on any POWER or PowerPC processor.
7273 GCC will use only the instructions in the common subset of both
7274 architectures, and will not use the MQ register. GCC assumes a generic
7275 processor model for scheduling purposes.
7277 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
7278 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
7279 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
7280 types, with an appropriate, generic processor model assumed for
7281 scheduling purposes.
7283 The other options specify a specific processor. Code generated under
7284 those options will run best on that processor, and may not run at all on
7287 The @option{-mcpu} options automatically enable or disable other
7288 @option{-m} options as follows:
7292 @option{-mno-power}, @option{-mno-powerpc}
7299 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
7314 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
7317 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
7322 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
7325 @item -mtune=@var{cpu_type}
7327 Set the instruction scheduling parameters for machine type
7328 @var{cpu_type}, but do not set the architecture type, register usage, or
7329 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
7330 values for @var{cpu_type} are used for @option{-mtune} as for
7331 @option{-mcpu}. If both are specified, the code generated will use the
7332 architecture, registers, and mnemonics set by @option{-mcpu}, but the
7333 scheduling parameters set by @option{-mtune}.
7338 @opindex mno-altivec
7339 These switches enable or disable the use of built-in functions that
7340 allow access to the AltiVec instruction set. You may also need to set
7341 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
7346 Extend the current ABI with SPE ABI extensions. This does not change
7347 the default ABI, instead it adds the SPE ABI extensions to the current
7351 @opindex mabi=no-spe
7352 Disable Booke SPE ABI extensions for the current ABI.
7354 @item -misel=@var{yes/no}
7357 This switch enables or disables the generation of ISEL instructions.
7359 @item -mspe=@var{yes/no}
7362 This switch enables or disables the generation of SPE simd
7365 @item -mfloat-gprs=@var{yes/no}
7367 @opindex mfloat-gprs
7368 This switch enables or disables the generation of floating point
7369 operations on the general purpose registers for architectures that
7370 support it. This option is currently only available on the MPC8540.
7373 @itemx -mno-fp-in-toc
7374 @itemx -mno-sum-in-toc
7375 @itemx -mminimal-toc
7377 @opindex mno-fp-in-toc
7378 @opindex mno-sum-in-toc
7379 @opindex mminimal-toc
7380 Modify generation of the TOC (Table Of Contents), which is created for
7381 every executable file. The @option{-mfull-toc} option is selected by
7382 default. In that case, GCC will allocate at least one TOC entry for
7383 each unique non-automatic variable reference in your program. GCC
7384 will also place floating-point constants in the TOC@. However, only
7385 16,384 entries are available in the TOC@.
7387 If you receive a linker error message that saying you have overflowed
7388 the available TOC space, you can reduce the amount of TOC space used
7389 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
7390 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
7391 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
7392 generate code to calculate the sum of an address and a constant at
7393 run-time instead of putting that sum into the TOC@. You may specify one
7394 or both of these options. Each causes GCC to produce very slightly
7395 slower and larger code at the expense of conserving TOC space.
7397 If you still run out of space in the TOC even when you specify both of
7398 these options, specify @option{-mminimal-toc} instead. This option causes
7399 GCC to make only one TOC entry for every file. When you specify this
7400 option, GCC will produce code that is slower and larger but which
7401 uses extremely little TOC space. You may wish to use this option
7402 only on files that contain less frequently executed code.
7408 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
7409 @code{long} type, and the infrastructure needed to support them.
7410 Specifying @option{-maix64} implies @option{-mpowerpc64} and
7411 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
7412 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
7417 @opindex mno-xl-call
7418 On AIX, pass floating-point arguments to prototyped functions beyond the
7419 register save area (RSA) on the stack in addition to argument FPRs. The
7420 AIX calling convention was extended but not initially documented to
7421 handle an obscure K&R C case of calling a function that takes the
7422 address of its arguments with fewer arguments than declared. AIX XL
7423 compilers access floating point arguments which do not fit in the
7424 RSA from the stack when a subroutine is compiled without
7425 optimization. Because always storing floating-point arguments on the
7426 stack is inefficient and rarely needed, this option is not enabled by
7427 default and only is necessary when calling subroutines compiled by AIX
7428 XL compilers without optimization.
7432 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
7433 application written to use message passing with special startup code to
7434 enable the application to run. The system must have PE installed in the
7435 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
7436 must be overridden with the @option{-specs=} option to specify the
7437 appropriate directory location. The Parallel Environment does not
7438 support threads, so the @option{-mpe} option and the @option{-pthread}
7439 option are incompatible.
7441 @item -malign-natural
7442 @itemx -malign-power
7443 @opindex malign-natural
7444 @opindex malign-power
7445 On AIX, Darwin, and 64-bit PowerPC Linux, the option
7446 @option{-malign-natural} overrides the ABI-defined alignment of larger
7447 types, such as floating-point doubles, on their natural size-based boundary.
7448 The option @option{-malign-power} instructs GCC to follow the ABI-specified
7449 alignment rules. GCC defaults to the standard alignment defined in the ABI.
7453 @opindex msoft-float
7454 @opindex mhard-float
7455 Generate code that does not use (uses) the floating-point register set.
7456 Software floating point emulation is provided if you use the
7457 @option{-msoft-float} option, and pass the option to GCC when linking.
7460 @itemx -mno-multiple
7462 @opindex mno-multiple
7463 Generate code that uses (does not use) the load multiple word
7464 instructions and the store multiple word instructions. These
7465 instructions are generated by default on POWER systems, and not
7466 generated on PowerPC systems. Do not use @option{-mmultiple} on little
7467 endian PowerPC systems, since those instructions do not work when the
7468 processor is in little endian mode. The exceptions are PPC740 and
7469 PPC750 which permit the instructions usage in little endian mode.
7475 Generate code that uses (does not use) the load string instructions
7476 and the store string word instructions to save multiple registers and
7477 do small block moves. These instructions are generated by default on
7478 POWER systems, and not generated on PowerPC systems. Do not use
7479 @option{-mstring} on little endian PowerPC systems, since those
7480 instructions do not work when the processor is in little endian mode.
7481 The exceptions are PPC740 and PPC750 which permit the instructions
7482 usage in little endian mode.
7488 Generate code that uses (does not use) the load or store instructions
7489 that update the base register to the address of the calculated memory
7490 location. These instructions are generated by default. If you use
7491 @option{-mno-update}, there is a small window between the time that the
7492 stack pointer is updated and the address of the previous frame is
7493 stored, which means code that walks the stack frame across interrupts or
7494 signals may get corrupted data.
7497 @itemx -mno-fused-madd
7498 @opindex mfused-madd
7499 @opindex mno-fused-madd
7500 Generate code that uses (does not use) the floating point multiply and
7501 accumulate instructions. These instructions are generated by default if
7502 hardware floating is used.
7504 @item -mno-bit-align
7506 @opindex mno-bit-align
7508 On System V.4 and embedded PowerPC systems do not (do) force structures
7509 and unions that contain bit-fields to be aligned to the base type of the
7512 For example, by default a structure containing nothing but 8
7513 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
7514 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
7515 the structure would be aligned to a 1 byte boundary and be one byte in
7518 @item -mno-strict-align
7519 @itemx -mstrict-align
7520 @opindex mno-strict-align
7521 @opindex mstrict-align
7522 On System V.4 and embedded PowerPC systems do not (do) assume that
7523 unaligned memory references will be handled by the system.
7526 @itemx -mno-relocatable
7527 @opindex mrelocatable
7528 @opindex mno-relocatable
7529 On embedded PowerPC systems generate code that allows (does not allow)
7530 the program to be relocated to a different address at runtime. If you
7531 use @option{-mrelocatable} on any module, all objects linked together must
7532 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
7534 @item -mrelocatable-lib
7535 @itemx -mno-relocatable-lib
7536 @opindex mrelocatable-lib
7537 @opindex mno-relocatable-lib
7538 On embedded PowerPC systems generate code that allows (does not allow)
7539 the program to be relocated to a different address at runtime. Modules
7540 compiled with @option{-mrelocatable-lib} can be linked with either modules
7541 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
7542 with modules compiled with the @option{-mrelocatable} options.
7548 On System V.4 and embedded PowerPC systems do not (do) assume that
7549 register 2 contains a pointer to a global area pointing to the addresses
7550 used in the program.
7553 @itemx -mlittle-endian
7555 @opindex mlittle-endian
7556 On System V.4 and embedded PowerPC systems compile code for the
7557 processor in little endian mode. The @option{-mlittle-endian} option is
7558 the same as @option{-mlittle}.
7563 @opindex mbig-endian
7564 On System V.4 and embedded PowerPC systems compile code for the
7565 processor in big endian mode. The @option{-mbig-endian} option is
7566 the same as @option{-mbig}.
7568 @item -mdynamic-no-pic
7569 @opindex mdynamic-no-pic
7570 On Darwin and Mac OS X systems, compile code so that it is not
7571 relocatable, but that its external references are relocatable. The
7572 resulting code is suitable for applications, but not shared
7575 @item -mprioritize-restricted-insns=@var{priority}
7576 @opindex mprioritize-restricted-insns
7577 This option controls the priority that is assigned to
7578 dispatch-slot restricted instructions during the second scheduling
7579 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
7580 @var{no/highest/second-highest} priority to dispatch slot restricted
7583 @item -msched-costly-dep=@var{dependence_type}
7584 @opindex msched-costly-dep
7585 This option controls which dependences are considered costly
7586 by the target during instruction scheduling. The argument
7587 @var{dependence_type} takes one of the following values:
7588 @var{no}: no dependence is costly,
7589 @var{all}: all dependences are costly,
7590 @var{true_store_to_load}: a true dependence from store to load is costly,
7591 @var{store_to_load}: any dependence from store to load is costly,
7592 @var{number}: any dependence which latency >= @var{number} is costly.
7594 @item -minsert-sched-nops=@var{scheme}
7595 @opindex minsert-sched-nops
7596 This option controls which nop insertion scheme will be used during
7597 the second scheduling pass. The argument @var{scheme} takes one of the
7599 @var{no}: Don't insert nops.
7600 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
7601 according to the scheduler's grouping.
7602 @var{regroup_exact}: Insert nops to force costly dependent insns into
7603 separate groups. Insert exactly as many nops as needed to force an insn
7604 to a new group, according to the estimatied processor grouping.
7605 @var{number}: Insert nops to force costly dependent insns into
7606 separate groups. Insert @var{number} nops to force an insn to a new group.
7610 On System V.4 and embedded PowerPC systems compile code using calling
7611 conventions that adheres to the March 1995 draft of the System V
7612 Application Binary Interface, PowerPC processor supplement. This is the
7613 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
7615 @item -mcall-sysv-eabi
7616 @opindex mcall-sysv-eabi
7617 Specify both @option{-mcall-sysv} and @option{-meabi} options.
7619 @item -mcall-sysv-noeabi
7620 @opindex mcall-sysv-noeabi
7621 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
7623 @item -mcall-solaris
7624 @opindex mcall-solaris
7625 On System V.4 and embedded PowerPC systems compile code for the Solaris
7629 @opindex mcall-linux
7630 On System V.4 and embedded PowerPC systems compile code for the
7631 Linux-based GNU system.
7635 On System V.4 and embedded PowerPC systems compile code for the
7636 Hurd-based GNU system.
7639 @opindex mcall-netbsd
7640 On System V.4 and embedded PowerPC systems compile code for the
7641 NetBSD operating system.
7643 @item -maix-struct-return
7644 @opindex maix-struct-return
7645 Return all structures in memory (as specified by the AIX ABI)@.
7647 @item -msvr4-struct-return
7648 @opindex msvr4-struct-return
7649 Return structures smaller than 8 bytes in registers (as specified by the
7653 @opindex mabi=altivec
7654 Extend the current ABI with AltiVec ABI extensions. This does not
7655 change the default ABI, instead it adds the AltiVec ABI extensions to
7658 @item -mabi=no-altivec
7659 @opindex mabi=no-altivec
7660 Disable AltiVec ABI extensions for the current ABI.
7663 @itemx -mno-prototype
7665 @opindex mno-prototype
7666 On System V.4 and embedded PowerPC systems assume that all calls to
7667 variable argument functions are properly prototyped. Otherwise, the
7668 compiler must insert an instruction before every non prototyped call to
7669 set or clear bit 6 of the condition code register (@var{CR}) to
7670 indicate whether floating point values were passed in the floating point
7671 registers in case the function takes a variable arguments. With
7672 @option{-mprototype}, only calls to prototyped variable argument functions
7673 will set or clear the bit.
7677 On embedded PowerPC systems, assume that the startup module is called
7678 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
7679 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
7684 On embedded PowerPC systems, assume that the startup module is called
7685 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
7690 On embedded PowerPC systems, assume that the startup module is called
7691 @file{crt0.o} and the standard C libraries are @file{libads.a} and
7695 @opindex myellowknife
7696 On embedded PowerPC systems, assume that the startup module is called
7697 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
7702 On System V.4 and embedded PowerPC systems, specify that you are
7703 compiling for a VxWorks system.
7707 Specify that you are compiling for the WindISS simulation environment.
7711 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7712 header to indicate that @samp{eabi} extended relocations are used.
7718 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7719 Embedded Applications Binary Interface (eabi) which is a set of
7720 modifications to the System V.4 specifications. Selecting @option{-meabi}
7721 means that the stack is aligned to an 8 byte boundary, a function
7722 @code{__eabi} is called to from @code{main} to set up the eabi
7723 environment, and the @option{-msdata} option can use both @code{r2} and
7724 @code{r13} to point to two separate small data areas. Selecting
7725 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7726 do not call an initialization function from @code{main}, and the
7727 @option{-msdata} option will only use @code{r13} to point to a single
7728 small data area. The @option{-meabi} option is on by default if you
7729 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7732 @opindex msdata=eabi
7733 On System V.4 and embedded PowerPC systems, put small initialized
7734 @code{const} global and static data in the @samp{.sdata2} section, which
7735 is pointed to by register @code{r2}. Put small initialized
7736 non-@code{const} global and static data in the @samp{.sdata} section,
7737 which is pointed to by register @code{r13}. Put small uninitialized
7738 global and static data in the @samp{.sbss} section, which is adjacent to
7739 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7740 incompatible with the @option{-mrelocatable} option. The
7741 @option{-msdata=eabi} option also sets the @option{-memb} option.
7744 @opindex msdata=sysv
7745 On System V.4 and embedded PowerPC systems, put small global and static
7746 data in the @samp{.sdata} section, which is pointed to by register
7747 @code{r13}. Put small uninitialized global and static data in the
7748 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7749 The @option{-msdata=sysv} option is incompatible with the
7750 @option{-mrelocatable} option.
7752 @item -msdata=default
7754 @opindex msdata=default
7756 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7757 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7758 same as @option{-msdata=sysv}.
7761 @opindex msdata-data
7762 On System V.4 and embedded PowerPC systems, put small global and static
7763 data in the @samp{.sdata} section. Put small uninitialized global and
7764 static data in the @samp{.sbss} section. Do not use register @code{r13}
7765 to address small data however. This is the default behavior unless
7766 other @option{-msdata} options are used.
7770 @opindex msdata=none
7772 On embedded PowerPC systems, put all initialized global and static data
7773 in the @samp{.data} section, and all uninitialized data in the
7774 @samp{.bss} section.
7778 @cindex smaller data references (PowerPC)
7779 @cindex .sdata/.sdata2 references (PowerPC)
7780 On embedded PowerPC systems, put global and static items less than or
7781 equal to @var{num} bytes into the small data or bss sections instead of
7782 the normal data or bss section. By default, @var{num} is 8. The
7783 @option{-G @var{num}} switch is also passed to the linker.
7784 All modules should be compiled with the same @option{-G @var{num}} value.
7787 @itemx -mno-regnames
7789 @opindex mno-regnames
7790 On System V.4 and embedded PowerPC systems do (do not) emit register
7791 names in the assembly language output using symbolic forms.
7794 @itemx -mno-longcall
7796 @opindex mno-longcall
7797 Default to making all function calls via pointers, so that functions
7798 which reside further than 64 megabytes (67,108,864 bytes) from the
7799 current location can be called. This setting can be overridden by the
7800 @code{shortcall} function attribute, or by @code{#pragma longcall(0)}.
7802 Some linkers are capable of detecting out-of-range calls and generating
7803 glue code on the fly. On these systems, long calls are unnecessary and
7804 generate slower code. As of this writing, the AIX linker can do this,
7805 as can the GNU linker for PowerPC/64. It is planned to add this feature
7806 to the GNU linker for 32-bit PowerPC systems as well.
7808 In the future, we may cause GCC to ignore all longcall specifications
7809 when the linker is known to generate glue.
7813 Adds support for multithreading with the @dfn{pthreads} library.
7814 This option sets flags for both the preprocessor and linker.
7818 @node Darwin Options
7819 @subsection Darwin Options
7820 @cindex Darwin options
7822 These options are defined for all architectures running the Darwin operating
7823 system. They are useful for compatibility with other Mac OS compilers.
7828 Loads all members of static archive libraries.
7829 See man ld(1) for more information.
7831 @item -arch_errors_fatal
7832 @opindex arch_errors_fatal
7833 Cause the errors having to do with files that have the wrong architecture
7837 @opindex bind_at_load
7838 Causes the output file to be marked such that the dynamic linker will
7839 bind all undefined references when the file is loaded or launched.
7843 Produce a Mach-o bundle format file.
7844 See man ld(1) for more information.
7846 @item -bundle_loader @var{executable}
7847 @opindex bundle_loader
7848 This specifies the @var{executable} that will be loading the build
7849 output file being linked. See man ld(1) for more information.
7851 @item -allowable_client @var{client_name}
7855 @item -compatibility_version
7856 @item -current_version
7857 @item -dependency-file
7859 @item -dylinker_install_name
7862 @item -exported_symbols_list
7864 @item -flat_namespace
7865 @item -force_cpusubtype_ALL
7866 @item -force_flat_namespace
7867 @item -headerpad_max_install_names
7871 @item -keep_private_externs
7873 @item -multiply_defined
7874 @item -multiply_defined_unused
7876 @item -nofixprebinding
7879 @item -noseglinkedit
7880 @item -pagezero_size
7882 @item -prebind_all_twolevel_modules
7883 @item -private_bundle
7884 @item -read_only_relocs
7886 @item -sectobjectsymbols
7890 @item -sectobjectsymbols
7892 @item -seg_addr_table
7893 @item -seg_addr_table_filename
7896 @item -segs_read_only_addr
7897 @item -segs_read_write_addr
7898 @item -single_module
7902 @item -twolevel_namespace
7905 @item -unexported_symbols_list
7906 @item -weak_reference_mismatches
7909 @opindex allowable_client
7911 @opindex client_name
7912 @opindex compatibility_version
7913 @opindex current_version
7914 @opindex dependency-file
7916 @opindex dylinker_install_name
7919 @opindex exported_symbols_list
7921 @opindex flat_namespace
7922 @opindex force_cpusubtype_ALL
7923 @opindex force_flat_namespace
7924 @opindex headerpad_max_install_names
7927 @opindex install_name
7928 @opindex keep_private_externs
7929 @opindex multi_module
7930 @opindex multiply_defined
7931 @opindex multiply_defined_unused
7933 @opindex nofixprebinding
7934 @opindex nomultidefs
7936 @opindex noseglinkedit
7937 @opindex pagezero_size
7939 @opindex prebind_all_twolevel_modules
7940 @opindex private_bundle
7941 @opindex read_only_relocs
7943 @opindex sectobjectsymbols
7947 @opindex sectobjectsymbols
7949 @opindex seg_addr_table
7950 @opindex seg_addr_table_filename
7951 @opindex seglinkedit
7953 @opindex segs_read_only_addr
7954 @opindex segs_read_write_addr
7955 @opindex single_module
7957 @opindex sub_library
7958 @opindex sub_umbrella
7959 @opindex twolevel_namespace
7962 @opindex unexported_symbols_list
7963 @opindex weak_reference_mismatches
7964 @opindex whatsloaded
7966 This options are available for Darwin linker. Darwin linker man page
7967 describes them in detail.
7972 @subsection IBM RT Options
7974 @cindex IBM RT options
7976 These @samp{-m} options are defined for the IBM RT PC:
7980 @opindex min-line-mul
7981 Use an in-line code sequence for integer multiplies. This is the
7984 @item -mcall-lib-mul
7985 @opindex mcall-lib-mul
7986 Call @code{lmul$$} for integer multiples.
7988 @item -mfull-fp-blocks
7989 @opindex mfull-fp-blocks
7990 Generate full-size floating point data blocks, including the minimum
7991 amount of scratch space recommended by IBM@. This is the default.
7993 @item -mminimum-fp-blocks
7994 @opindex mminimum-fp-blocks
7995 Do not include extra scratch space in floating point data blocks. This
7996 results in smaller code, but slower execution, since scratch space must
7997 be allocated dynamically.
7999 @cindex @file{stdarg.h} and RT PC
8000 @item -mfp-arg-in-fpregs
8001 @opindex mfp-arg-in-fpregs
8002 Use a calling sequence incompatible with the IBM calling convention in
8003 which floating point arguments are passed in floating point registers.
8004 Note that @code{stdarg.h} will not work with floating point operands
8005 if this option is specified.
8007 @item -mfp-arg-in-gregs
8008 @opindex mfp-arg-in-gregs
8009 Use the normal calling convention for floating point arguments. This is
8012 @item -mhc-struct-return
8013 @opindex mhc-struct-return
8014 Return structures of more than one word in memory, rather than in a
8015 register. This provides compatibility with the MetaWare HighC (hc)
8016 compiler. Use the option @option{-fpcc-struct-return} for compatibility
8017 with the Portable C Compiler (pcc).
8019 @item -mnohc-struct-return
8020 @opindex mnohc-struct-return
8021 Return some structures of more than one word in registers, when
8022 convenient. This is the default. For compatibility with the
8023 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
8024 option @option{-mhc-struct-return}.
8028 @subsection MIPS Options
8029 @cindex MIPS options
8031 These @samp{-m} options are defined for the MIPS family of computers:
8035 @item -march=@var{arch}
8037 Generate code that will run on @var{arch}, which can be the name of a
8038 generic MIPS ISA, or the name of a particular processor.
8040 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
8041 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
8042 The processor names are:
8043 @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
8045 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
8046 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000}, @samp{rm7000},
8050 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4300},
8051 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
8052 The special value @samp{from-abi} selects the
8053 most compatible architecture for the selected ABI (that is,
8054 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
8056 In processor names, a final @samp{000} can be abbreviated as @samp{k}
8057 (for example, @samp{-march=r2k}). Prefixes are optional, and
8058 @samp{vr} may be written @samp{r}.
8060 GCC defines two macros based on the value of this option. The first
8061 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
8062 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
8063 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
8064 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
8065 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
8067 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
8068 above. In other words, it will have the full prefix and will not
8069 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
8070 the macro names the resolved architecture (either @samp{"mips1"} or
8071 @samp{"mips3"}). It names the default architecture when no
8072 @option{-march} option is given.
8074 @item -mtune=@var{arch}
8076 Optimize for @var{arch}. Among other things, this option controls
8077 the way instructions are scheduled, and the perceived cost of arithmetic
8078 operations. The list of @var{arch} values is the same as for
8081 When this option is not used, GCC will optimize for the processor
8082 specified by @option{-march}. By using @option{-march} and
8083 @option{-mtune} together, it is possible to generate code that will
8084 run on a family of processors, but optimize the code for one
8085 particular member of that family.
8087 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
8088 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
8089 @samp{-march} ones described above.
8093 Equivalent to @samp{-march=mips1}.
8097 Equivalent to @samp{-march=mips2}.
8101 Equivalent to @samp{-march=mips3}.
8105 Equivalent to @samp{-march=mips4}.
8109 Equivalent to @samp{-march=mips32}.
8113 Equivalent to @samp{-march=mips32r2}.
8117 Equivalent to @samp{-march=mips64}.
8120 @itemx -mno-fused-madd
8121 @opindex mfused-madd
8122 @opindex mno-fused-madd
8123 Generate code that uses (does not use) the floating point multiply and
8124 accumulate instructions, when they are available. These instructions
8125 are generated by default if they are available, but this may be
8126 undesirable if the extra precision causes problems or on certain chips
8127 in the mode where denormals are rounded to zero where denormals
8128 generated by multiply and accumulate instructions cause exceptions
8133 Assume that floating point registers are 32 bits wide.
8137 Assume that floating point registers are 64 bits wide.
8141 Assume that general purpose registers are 32 bits wide.
8145 Assume that general purpose registers are 64 bits wide.
8149 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
8150 explanation of the default, and the width of pointers.
8154 Force long types to be 64 bits wide. See @option{-mlong32} for an
8155 explanation of the default, and the width of pointers.
8159 Force long, int, and pointer types to be 32 bits wide.
8161 The default size of ints, longs and pointers depends on the ABI@. All
8162 the supported ABIs use 32-bit ints. The n64 ABI uses 64-bit longs, as
8163 does the 64-bit Cygnus EABI; the others use 32-bit longs. Pointers
8164 are the same size as longs, or the same size as integer registers,
8165 whichever is smaller.
8179 Generate code for the given ABI@.
8181 Note that there are two embedded ABIs: @option{-mabi=eabi}
8182 selects the one defined by Cygnus while @option{-meabi=meabi}
8183 selects the one defined by MIPS@. Both these ABIs have
8184 32-bit and 64-bit variants. Normally, GCC will generate
8185 64-bit code when you select a 64-bit architecture, but you
8186 can use @option{-mgp32} to get 32-bit code instead.
8188 @item -mabi-fake-default
8189 @opindex mabi-fake-default
8190 You don't want to know what this option does. No, really. I mean
8191 it. Move on to the next option.
8193 What? You're still here? Oh, well@enddots{} Ok, here's the deal. GCC
8194 wants the default set of options to get the root of the multilib tree,
8195 and the shared library SONAMEs without any multilib-indicating
8196 suffixes. This is not convenience for @samp{mips64-linux-gnu}, since
8197 we want to default to the N32 ABI, while still being binary-compatible
8198 with @samp{mips-linux-gnu} if you stick to the O32 ABI@. Being
8199 binary-compatible means shared libraries should have the same SONAMEs,
8200 and libraries should live in the same location. Having O32 libraries
8201 in a sub-directory named say @file{o32} is not acceptable.
8203 So we trick GCC into believing that O32 is the default ABI, except
8204 that we override the default with some internal command-line
8205 processing magic. Problem is, if we stopped at that, and you then
8206 created a multilib-aware package that used the output of @command{gcc
8207 -print-multi-lib} to decide which multilibs to build, and how, and
8208 you'd find yourself in an awkward situation when you found out that
8209 some of the options listed ended up mapping to the same multilib, and
8210 none of your libraries was actually built for the multilib that
8211 @option{-print-multi-lib} claims to be the default. So we added this
8212 option that disables the default switcher, falling back to GCC's
8213 original notion of the default library. Confused yet?
8215 For short: don't ever use this option, unless you find it in the list
8216 of additional options to be used when building for multilibs, in the
8217 output of @option{gcc -print-multi-lib}.
8221 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
8222 add normal debug information. This is the default for all
8223 platforms except for the OSF/1 reference platform, using the OSF/rose
8224 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
8225 switches are used, the @file{mips-tfile} program will encapsulate the
8226 stabs within MIPS ECOFF@.
8230 Generate code for the GNU assembler. This is the default on the OSF/1
8231 reference platform, using the OSF/rose object format. Also, this is
8232 the default if the configure option @option{--with-gnu-as} is used.
8234 @item -msplit-addresses
8235 @itemx -mno-split-addresses
8236 @opindex msplit-addresses
8237 @opindex mno-split-addresses
8238 Generate code to load the high and low parts of address constants separately.
8239 This allows GCC to optimize away redundant loads of the high order
8240 bits of addresses. This optimization requires GNU as and GNU ld.
8241 This optimization is enabled by default for some embedded targets where
8242 GNU as and GNU ld are standard.
8248 The @option{-mrnames} switch says to output code using the MIPS software
8249 names for the registers, instead of the hardware names (ie, @var{a0}
8250 instead of @var{$4}). The only known assembler that supports this option
8251 is the Algorithmics assembler.
8257 The @option{-mmemcpy} switch makes all block moves call the appropriate
8258 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
8259 generating inline code.
8262 @itemx -mno-mips-tfile
8263 @opindex mmips-tfile
8264 @opindex mno-mips-tfile
8265 The @option{-mno-mips-tfile} switch causes the compiler not
8266 postprocess the object file with the @file{mips-tfile} program,
8267 after the MIPS assembler has generated it to add debug support. If
8268 @file{mips-tfile} is not run, then no local variables will be
8269 available to the debugger. In addition, @file{stage2} and
8270 @file{stage3} objects will have the temporary file names passed to the
8271 assembler embedded in the object file, which means the objects will
8272 not compare the same. The @option{-mno-mips-tfile} switch should only
8273 be used when there are bugs in the @file{mips-tfile} program that
8274 prevents compilation.
8277 @opindex msoft-float
8278 Generate output containing library calls for floating point.
8279 @strong{Warning:} the requisite libraries are not part of GCC@.
8280 Normally the facilities of the machine's usual C compiler are used, but
8281 this can't be done directly in cross-compilation. You must make your
8282 own arrangements to provide suitable library functions for
8286 @opindex mhard-float
8287 Generate output containing floating point instructions. This is the
8288 default if you use the unmodified sources.
8291 @itemx -mno-abicalls
8293 @opindex mno-abicalls
8294 Emit (or do not emit) the pseudo operations @samp{.abicalls},
8295 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
8296 position independent code.
8302 Lift (or do not lift) the usual restrictions on the size of the global
8305 GCC normally uses a single instruction to load values from the GOT.
8306 While this is relatively efficient, it will only work if the GOT
8307 is smaller than about 64k. Anything larger will cause the linker
8308 to report an error such as:
8310 @cindex relocation truncated to fit (MIPS)
8312 relocation truncated to fit: R_MIPS_GOT16 foobar
8315 If this happens, you should recompile your code with @option{-mxgot}.
8316 It should then work with very large GOTs, although it will also be
8317 less efficient, since it will take three instructions to fetch the
8318 value of a global symbol.
8320 Note that some linkers can create multiple GOTs. If you have such a
8321 linker, you should only need to use @option{-mxgot} when a single object
8322 file accesses more than 64k's worth of GOT entries. Very few do.
8324 These options have no effect unless GCC is generating position
8328 @itemx -mno-long-calls
8329 @opindex mlong-calls
8330 @opindex mno-long-calls
8331 Do all calls with the @samp{JALR} instruction, which requires
8332 loading up a function's address into a register before the call.
8333 You need to use this switch, if you call outside of the current
8334 512 megabyte segment to functions that are not through pointers.
8336 @item -membedded-pic
8337 @itemx -mno-embedded-pic
8338 @opindex membedded-pic
8339 @opindex mno-embedded-pic
8340 Generate PIC code suitable for some embedded systems. All calls are
8341 made using PC relative address, and all data is addressed using the $gp
8342 register. No more than 65536 bytes of global data may be used. This
8343 requires GNU as and GNU ld which do most of the work. This currently
8344 only works on targets which use ECOFF; it does not work with ELF@.
8346 @item -membedded-data
8347 @itemx -mno-embedded-data
8348 @opindex membedded-data
8349 @opindex mno-embedded-data
8350 Allocate variables to the read-only data section first if possible, then
8351 next in the small data section if possible, otherwise in data. This gives
8352 slightly slower code than the default, but reduces the amount of RAM required
8353 when executing, and thus may be preferred for some embedded systems.
8355 @item -muninit-const-in-rodata
8356 @itemx -mno-uninit-const-in-rodata
8357 @opindex muninit-const-in-rodata
8358 @opindex mno-uninit-const-in-rodata
8359 When used together with @option{-membedded-data}, it will always store uninitialized
8360 const variables in the read-only data section.
8362 @item -msingle-float
8363 @itemx -mdouble-float
8364 @opindex msingle-float
8365 @opindex mdouble-float
8366 The @option{-msingle-float} switch tells gcc to assume that the floating
8367 point coprocessor only supports single precision operations, as on the
8368 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
8369 double precision operations. This is the default.
8375 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
8376 as on the @samp{r4650} chip.
8380 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
8381 @option{-mcpu=r4650}.
8387 Enable 16-bit instructions.
8391 Compile code for the processor in little endian mode.
8392 The requisite libraries are assumed to exist.
8396 Compile code for the processor in big endian mode.
8397 The requisite libraries are assumed to exist.
8401 @cindex smaller data references (MIPS)
8402 @cindex gp-relative references (MIPS)
8403 Put global and static items less than or equal to @var{num} bytes into
8404 the small data or bss sections instead of the normal data or bss
8405 section. This allows the assembler to emit one word memory reference
8406 instructions based on the global pointer (@var{gp} or @var{$28}),
8407 instead of the normal two words used. By default, @var{num} is 8 when
8408 the MIPS assembler is used, and 0 when the GNU assembler is used. The
8409 @option{-G @var{num}} switch is also passed to the assembler and linker.
8410 All modules should be compiled with the same @option{-G @var{num}}
8415 Tell the MIPS assembler to not run its preprocessor over user
8416 assembler files (with a @samp{.s} suffix) when assembling them.
8420 Pass an option to gas which will cause nops to be inserted if
8421 the read of the destination register of an mfhi or mflo instruction
8422 occurs in the following two instructions.
8427 Work around certain SB-1 CPU core errata.
8428 (This flag currently works around the SB-1 revision 2
8429 ``F1'' and ``F2'' floating point errata.)
8433 Do not include the default crt0.
8435 @item -mflush-func=@var{func}
8436 @itemx -mno-flush-func
8437 @opindex mflush-func
8438 Specifies the function to call to flush the I and D caches, or to not
8439 call any such function. If called, the function must take the same
8440 arguments as the common @code{_flush_func()}, that is, the address of the
8441 memory range for which the cache is being flushed, the size of the
8442 memory range, and the number 3 (to flush both caches). The default
8443 depends on the target gcc was configured for, but commonly is either
8444 @samp{_flush_func} or @samp{__cpu_flush}.
8446 @item -mbranch-likely
8447 @itemx -mno-branch-likely
8448 @opindex mbranch-likely
8449 @opindex mno-branch-likely
8450 Enable or disable use of Branch Likely instructions, regardless of the
8451 default for the selected architecture. By default, Branch Likely
8452 instructions may be generated if they are supported by the selected
8453 architecture. An exception is for the MIPS32 and MIPS64 architectures
8454 and processors which implement those architectures; for those, Branch
8455 Likely instructions will not be generated by default because the MIPS32
8456 and MIPS64 architectures specifically deprecate their use.
8459 @node i386 and x86-64 Options
8460 @subsection Intel 386 and AMD x86-64 Options
8461 @cindex i386 Options
8462 @cindex x86-64 Options
8463 @cindex Intel 386 Options
8464 @cindex AMD x86-64 Options
8466 These @samp{-m} options are defined for the i386 and x86-64 family of
8470 @item -mtune=@var{cpu-type}
8472 Tune to @var{cpu-type} everything applicable about the generated code, except
8473 for the ABI and the set of available instructions. The choices for
8474 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
8475 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
8476 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
8477 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp},
8478 @samp{athlon-mp}, @samp{winchip-c6}, @samp{winchip2}, @samp{k8}, @samp{c3}
8481 While picking a specific @var{cpu-type} will schedule things appropriately
8482 for that particular chip, the compiler will not generate any code that
8483 does not run on the i386 without the @option{-march=@var{cpu-type}} option
8484 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
8485 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
8486 AMD chips as opposed to the Intel ones.
8488 @item -march=@var{cpu-type}
8490 Generate instructions for the machine type @var{cpu-type}. The choices
8491 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
8492 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
8494 @item -mcpu=@var{cpu-type}
8496 A deprecated synonym for @option{-mtune}.
8505 @opindex mpentiumpro
8506 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
8507 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
8508 These synonyms are deprecated.
8510 @item -mfpmath=@var{unit}
8512 generate floating point arithmetics for selected unit @var{unit}. the choices
8517 Use the standard 387 floating point coprocessor present majority of chips and
8518 emulated otherwise. Code compiled with this option will run almost everywhere.
8519 The temporary results are computed in 80bit precision instead of precision
8520 specified by the type resulting in slightly different results compared to most
8521 of other chips. See @option{-ffloat-store} for more detailed description.
8523 This is the default choice for i386 compiler.
8526 Use scalar floating point instructions present in the SSE instruction set.
8527 This instruction set is supported by Pentium3 and newer chips, in the AMD line
8528 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
8529 instruction set supports only single precision arithmetics, thus the double and
8530 extended precision arithmetics is still done using 387. Later version, present
8531 only in Pentium4 and the future AMD x86-64 chips supports double precision
8534 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
8535 @option{-msse2} switches to enable SSE extensions and make this option
8536 effective. For x86-64 compiler, these extensions are enabled by default.
8538 The resulting code should be considerably faster in majority of cases and avoid
8539 the numerical instability problems of 387 code, but may break some existing
8540 code that expects temporaries to be 80bit.
8542 This is the default choice for x86-64 compiler.
8545 Use all SSE extensions enabled by @option{-msse2} as well as the new
8546 SSE extensions in Prescott New Instructions. @option{-mpni} also
8547 enables 2 builtin functions, @code{__builtin_ia32_monitor} and
8548 @code{__builtin_ia32_mwait}, for new instructions @code{monitor} and
8552 Attempt to utilize both instruction sets at once. This effectively double the
8553 amount of available registers and on chips with separate execution units for
8554 387 and SSE the execution resources too. Use this option with care, as it is
8555 still experimental, because gcc register allocator does not model separate
8556 functional units well resulting in instable performance.
8559 @item -masm=@var{dialect}
8560 @opindex masm=@var{dialect}
8561 Output asm instructions using selected @var{dialect}. Supported choices are
8562 @samp{intel} or @samp{att} (the default one).
8567 @opindex mno-ieee-fp
8568 Control whether or not the compiler uses IEEE floating point
8569 comparisons. These handle correctly the case where the result of a
8570 comparison is unordered.
8573 @opindex msoft-float
8574 Generate output containing library calls for floating point.
8575 @strong{Warning:} the requisite libraries are not part of GCC@.
8576 Normally the facilities of the machine's usual C compiler are used, but
8577 this can't be done directly in cross-compilation. You must make your
8578 own arrangements to provide suitable library functions for
8581 On machines where a function returns floating point results in the 80387
8582 register stack, some floating point opcodes may be emitted even if
8583 @option{-msoft-float} is used.
8585 @item -mno-fp-ret-in-387
8586 @opindex mno-fp-ret-in-387
8587 Do not use the FPU registers for return values of functions.
8589 The usual calling convention has functions return values of types
8590 @code{float} and @code{double} in an FPU register, even if there
8591 is no FPU@. The idea is that the operating system should emulate
8594 The option @option{-mno-fp-ret-in-387} causes such values to be returned
8595 in ordinary CPU registers instead.
8597 @item -mno-fancy-math-387
8598 @opindex mno-fancy-math-387
8599 Some 387 emulators do not support the @code{sin}, @code{cos} and
8600 @code{sqrt} instructions for the 387. Specify this option to avoid
8601 generating those instructions. This option is the default on FreeBSD,
8602 OpenBSD and NetBSD@. This option is overridden when @option{-march}
8603 indicates that the target cpu will always have an FPU and so the
8604 instruction will not need emulation. As of revision 2.6.1, these
8605 instructions are not generated unless you also use the
8606 @option{-funsafe-math-optimizations} switch.
8608 @item -malign-double
8609 @itemx -mno-align-double
8610 @opindex malign-double
8611 @opindex mno-align-double
8612 Control whether GCC aligns @code{double}, @code{long double}, and
8613 @code{long long} variables on a two word boundary or a one word
8614 boundary. Aligning @code{double} variables on a two word boundary will
8615 produce code that runs somewhat faster on a @samp{Pentium} at the
8616 expense of more memory.
8618 @strong{Warning:} if you use the @option{-malign-double} switch,
8619 structures containing the above types will be aligned differently than
8620 the published application binary interface specifications for the 386
8621 and will not be binary compatible with structures in code compiled
8622 without that switch.
8624 @item -m96bit-long-double
8625 @item -m128bit-long-double
8626 @opindex m96bit-long-double
8627 @opindex m128bit-long-double
8628 These switches control the size of @code{long double} type. The i386
8629 application binary interface specifies the size to be 96 bits,
8630 so @option{-m96bit-long-double} is the default in 32 bit mode.
8632 Modern architectures (Pentium and newer) would prefer @code{long double}
8633 to be aligned to an 8 or 16 byte boundary. In arrays or structures
8634 conforming to the ABI, this would not be possible. So specifying a
8635 @option{-m128bit-long-double} will align @code{long double}
8636 to a 16 byte boundary by padding the @code{long double} with an additional
8639 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
8640 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
8642 Notice that neither of these options enable any extra precision over the x87
8643 standard of 80 bits for a @code{long double}.
8645 @strong{Warning:} if you override the default value for your target ABI, the
8646 structures and arrays containing @code{long double} will change their size as
8647 well as function calling convention for function taking @code{long double}
8648 will be modified. Hence they will not be binary compatible with arrays or
8649 structures in code compiled without that switch.
8653 @itemx -mno-svr3-shlib
8654 @opindex msvr3-shlib
8655 @opindex mno-svr3-shlib
8656 Control whether GCC places uninitialized local variables into the
8657 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
8658 into @code{bss}. These options are meaningful only on System V Release 3.
8662 Use a different function-calling convention, in which functions that
8663 take a fixed number of arguments return with the @code{ret} @var{num}
8664 instruction, which pops their arguments while returning. This saves one
8665 instruction in the caller since there is no need to pop the arguments
8668 You can specify that an individual function is called with this calling
8669 sequence with the function attribute @samp{stdcall}. You can also
8670 override the @option{-mrtd} option by using the function attribute
8671 @samp{cdecl}. @xref{Function Attributes}.
8673 @strong{Warning:} this calling convention is incompatible with the one
8674 normally used on Unix, so you cannot use it if you need to call
8675 libraries compiled with the Unix compiler.
8677 Also, you must provide function prototypes for all functions that
8678 take variable numbers of arguments (including @code{printf});
8679 otherwise incorrect code will be generated for calls to those
8682 In addition, seriously incorrect code will result if you call a
8683 function with too many arguments. (Normally, extra arguments are
8684 harmlessly ignored.)
8686 @item -mregparm=@var{num}
8688 Control how many registers are used to pass integer arguments. By
8689 default, no registers are used to pass arguments, and at most 3
8690 registers can be used. You can control this behavior for a specific
8691 function by using the function attribute @samp{regparm}.
8692 @xref{Function Attributes}.
8694 @strong{Warning:} if you use this switch, and
8695 @var{num} is nonzero, then you must build all modules with the same
8696 value, including any libraries. This includes the system libraries and
8699 @item -mpreferred-stack-boundary=@var{num}
8700 @opindex mpreferred-stack-boundary
8701 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
8702 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
8703 the default is 4 (16 bytes or 128 bits), except when optimizing for code
8704 size (@option{-Os}), in which case the default is the minimum correct
8705 alignment (4 bytes for x86, and 8 bytes for x86-64).
8707 On Pentium and PentiumPro, @code{double} and @code{long double} values
8708 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
8709 suffer significant run time performance penalties. On Pentium III, the
8710 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
8711 penalties if it is not 16 byte aligned.
8713 To ensure proper alignment of this values on the stack, the stack boundary
8714 must be as aligned as that required by any value stored on the stack.
8715 Further, every function must be generated such that it keeps the stack
8716 aligned. Thus calling a function compiled with a higher preferred
8717 stack boundary from a function compiled with a lower preferred stack
8718 boundary will most likely misalign the stack. It is recommended that
8719 libraries that use callbacks always use the default setting.
8721 This extra alignment does consume extra stack space, and generally
8722 increases code size. Code that is sensitive to stack space usage, such
8723 as embedded systems and operating system kernels, may want to reduce the
8724 preferred alignment to @option{-mpreferred-stack-boundary=2}.
8742 These switches enable or disable the use of built-in functions that allow
8743 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
8745 @xref{X86 Built-in Functions}, for details of the functions enabled
8746 and disabled by these switches.
8748 To have SSE/SSE2 instructions generated automatically from floating-point
8749 code, see @option{-mfpmath=sse}.
8752 @itemx -mno-push-args
8754 @opindex mno-push-args
8755 Use PUSH operations to store outgoing parameters. This method is shorter
8756 and usually equally fast as method using SUB/MOV operations and is enabled
8757 by default. In some cases disabling it may improve performance because of
8758 improved scheduling and reduced dependencies.
8760 @item -maccumulate-outgoing-args
8761 @opindex maccumulate-outgoing-args
8762 If enabled, the maximum amount of space required for outgoing arguments will be
8763 computed in the function prologue. This is faster on most modern CPUs
8764 because of reduced dependencies, improved scheduling and reduced stack usage
8765 when preferred stack boundary is not equal to 2. The drawback is a notable
8766 increase in code size. This switch implies @option{-mno-push-args}.
8770 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
8771 on thread-safe exception handling must compile and link all code with the
8772 @option{-mthreads} option. When compiling, @option{-mthreads} defines
8773 @option{-D_MT}; when linking, it links in a special thread helper library
8774 @option{-lmingwthrd} which cleans up per thread exception handling data.
8776 @item -mno-align-stringops
8777 @opindex mno-align-stringops
8778 Do not align destination of inlined string operations. This switch reduces
8779 code size and improves performance in case the destination is already aligned,
8780 but gcc don't know about it.
8782 @item -minline-all-stringops
8783 @opindex minline-all-stringops
8784 By default GCC inlines string operations only when destination is known to be
8785 aligned at least to 4 byte boundary. This enables more inlining, increase code
8786 size, but may improve performance of code that depends on fast memcpy, strlen
8787 and memset for short lengths.
8789 @item -momit-leaf-frame-pointer
8790 @opindex momit-leaf-frame-pointer
8791 Don't keep the frame pointer in a register for leaf functions. This
8792 avoids the instructions to save, set up and restore frame pointers and
8793 makes an extra register available in leaf functions. The option
8794 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8795 which might make debugging harder.
8797 @item -mtls-direct-seg-refs
8798 @itemx -mno-tls-direct-seg-refs
8799 @opindex mtls-direct-seg-refs
8800 Controls whether TLS variables may be accessed with offsets from the
8801 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
8802 or whether the thread base pointer must be added. Whether or not this
8803 is legal depends on the operating system, and whether it maps the
8804 segment to cover the entire TLS area.
8806 For systems that use GNU libc, the default is on.
8809 These @samp{-m} switches are supported in addition to the above
8810 on AMD x86-64 processors in 64-bit environments.
8817 Generate code for a 32-bit or 64-bit environment.
8818 The 32-bit environment sets int, long and pointer to 32 bits and
8819 generates code that runs on any i386 system.
8820 The 64-bit environment sets int to 32 bits and long and pointer
8821 to 64 bits and generates code for AMD's x86-64 architecture.
8824 @opindex no-red-zone
8825 Do not use a so called red zone for x86-64 code. The red zone is mandated
8826 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8827 stack pointer that will not be modified by signal or interrupt handlers
8828 and therefore can be used for temporary data without adjusting the stack
8829 pointer. The flag @option{-mno-red-zone} disables this red zone.
8831 @item -mcmodel=small
8832 @opindex mcmodel=small
8833 Generate code for the small code model: the program and its symbols must
8834 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
8835 Programs can be statically or dynamically linked. This is the default
8838 @item -mcmodel=kernel
8839 @opindex mcmodel=kernel
8840 Generate code for the kernel code model. The kernel runs in the
8841 negative 2 GB of the address space.
8842 This model has to be used for Linux kernel code.
8844 @item -mcmodel=medium
8845 @opindex mcmodel=medium
8846 Generate code for the medium model: The program is linked in the lower 2
8847 GB of the address space but symbols can be located anywhere in the
8848 address space. Programs can be statically or dynamically linked, but
8849 building of shared libraries are not supported with the medium model.
8851 @item -mcmodel=large
8852 @opindex mcmodel=large
8853 Generate code for the large model: This model makes no assumptions
8854 about addresses and sizes of sections. Currently GCC does not implement
8859 @subsection HPPA Options
8860 @cindex HPPA Options
8862 These @samp{-m} options are defined for the HPPA family of computers:
8865 @item -march=@var{architecture-type}
8867 Generate code for the specified architecture. The choices for
8868 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8869 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8870 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8871 architecture option for your machine. Code compiled for lower numbered
8872 architectures will run on higher numbered architectures, but not the
8875 PA 2.0 support currently requires gas snapshot 19990413 or later. The
8876 next release of binutils (current is 2.9.1) will probably contain PA 2.0
8880 @itemx -mpa-risc-1-1
8881 @itemx -mpa-risc-2-0
8882 @opindex mpa-risc-1-0
8883 @opindex mpa-risc-1-1
8884 @opindex mpa-risc-2-0
8885 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8888 @opindex mbig-switch
8889 Generate code suitable for big switch tables. Use this option only if
8890 the assembler/linker complain about out of range branches within a switch
8893 @item -mjump-in-delay
8894 @opindex mjump-in-delay
8895 Fill delay slots of function calls with unconditional jump instructions
8896 by modifying the return pointer for the function call to be the target
8897 of the conditional jump.
8899 @item -mdisable-fpregs
8900 @opindex mdisable-fpregs
8901 Prevent floating point registers from being used in any manner. This is
8902 necessary for compiling kernels which perform lazy context switching of
8903 floating point registers. If you use this option and attempt to perform
8904 floating point operations, the compiler will abort.
8906 @item -mdisable-indexing
8907 @opindex mdisable-indexing
8908 Prevent the compiler from using indexing address modes. This avoids some
8909 rather obscure problems when compiling MIG generated code under MACH@.
8911 @item -mno-space-regs
8912 @opindex mno-space-regs
8913 Generate code that assumes the target has no space registers. This allows
8914 GCC to generate faster indirect calls and use unscaled index address modes.
8916 Such code is suitable for level 0 PA systems and kernels.
8918 @item -mfast-indirect-calls
8919 @opindex mfast-indirect-calls
8920 Generate code that assumes calls never cross space boundaries. This
8921 allows GCC to emit code which performs faster indirect calls.
8923 This option will not work in the presence of shared libraries or nested
8926 @item -mlong-load-store
8927 @opindex mlong-load-store
8928 Generate 3-instruction load and store sequences as sometimes required by
8929 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8932 @item -mportable-runtime
8933 @opindex mportable-runtime
8934 Use the portable calling conventions proposed by HP for ELF systems.
8938 Enable the use of assembler directives only GAS understands.
8940 @item -mschedule=@var{cpu-type}
8942 Schedule code according to the constraints for the machine type
8943 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8944 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8945 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8946 proper scheduling option for your machine. The default scheduling is
8950 @opindex mlinker-opt
8951 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8952 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8953 linkers in which they give bogus error messages when linking some programs.
8956 @opindex msoft-float
8957 Generate output containing library calls for floating point.
8958 @strong{Warning:} the requisite libraries are not available for all HPPA
8959 targets. Normally the facilities of the machine's usual C compiler are
8960 used, but this cannot be done directly in cross-compilation. You must make
8961 your own arrangements to provide suitable library functions for
8962 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8963 does provide software floating point support.
8965 @option{-msoft-float} changes the calling convention in the output file;
8966 therefore, it is only useful if you compile @emph{all} of a program with
8967 this option. In particular, you need to compile @file{libgcc.a}, the
8968 library that comes with GCC, with @option{-msoft-float} in order for
8973 Generate the predefine, @code{_SIO}, for server IO. The default is
8974 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
8975 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO. These
8976 options are available under HP-UX and HI-UX.
8980 Use GNU ld specific options. This passes @option{-shared} to ld when
8981 building a shared library. It is the default when GCC is configured,
8982 explicitly or implicitly, with the GNU linker. This option does not
8983 have any affect on which ld is called, it only changes what parameters
8984 are passed to that ld. The ld that is called is determined by the
8985 @option{--with-ld} configure option, gcc's program search path, and
8986 finally by the user's @env{PATH}. The linker used by GCC can be printed
8987 using @samp{which `gcc -print-prog-name=ld`}.
8991 Use HP ld specific options. This passes @option{-b} to ld when building
8992 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8993 links. It is the default when GCC is configured, explicitly or
8994 implicitly, with the HP linker. This option does not have any affect on
8995 which ld is called, it only changes what parameters are passed to that
8996 ld. The ld that is called is determined by the @option{--with-ld}
8997 configure option, gcc's program search path, and finally by the user's
8998 @env{PATH}. The linker used by GCC can be printed using @samp{which
8999 `gcc -print-prog-name=ld`}.
9002 @opindex mno-long-calls
9003 Generate code that uses long call sequences. This ensures that a call
9004 is always able to reach linker generated stubs. The default is to generate
9005 long calls only when the distance from the call site to the beginning
9006 of the function or translation unit, as the case may be, exceeds a
9007 predefined limit set by the branch type being used. The limits for
9008 normal calls are 7,600,000 and 240,000 bytes, respectively for the
9009 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
9012 Distances are measured from the beginning of functions when using the
9013 @option{-ffunction-sections} option, or when using the @option{-mgas}
9014 and @option{-mno-portable-runtime} options together under HP-UX with
9017 It is normally not desirable to use this option as it will degrade
9018 performance. However, it may be useful in large applications,
9019 particularly when partial linking is used to build the application.
9021 The types of long calls used depends on the capabilities of the
9022 assembler and linker, and the type of code being generated. The
9023 impact on systems that support long absolute calls, and long pic
9024 symbol-difference or pc-relative calls should be relatively small.
9025 However, an indirect call is used on 32-bit ELF systems in pic code
9026 and it is quite long.
9030 Suppress the generation of link options to search libdld.sl when the
9031 @option{-static} option is specified on HP-UX 10 and later.
9035 The HP-UX implementation of setlocale in libc has a dependency on
9036 libdld.sl. There isn't an archive version of libdld.sl. Thus,
9037 when the @option{-static} option is specified, special link options
9038 are needed to resolve this dependency.
9040 On HP-UX 10 and later, the GCC driver adds the necessary options to
9041 link with libdld.sl when the @option{-static} option is specified.
9042 This causes the resulting binary to be dynamic. On the 64-bit port,
9043 the linkers generate dynamic binaries by default in any case. The
9044 @option{-nolibdld} option can be used to prevent the GCC driver from
9045 adding these link options.
9049 Add support for multithreading with the @dfn{dce thread} library
9050 under HP-UX. This option sets flags for both the preprocessor and
9054 @node Intel 960 Options
9055 @subsection Intel 960 Options
9057 These @samp{-m} options are defined for the Intel 960 implementations:
9060 @item -m@var{cpu-type}
9068 Assume the defaults for the machine type @var{cpu-type} for some of
9069 the other options, including instruction scheduling, floating point
9070 support, and addressing modes. The choices for @var{cpu-type} are
9071 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
9072 @samp{sa}, and @samp{sb}.
9079 @opindex msoft-float
9080 The @option{-mnumerics} option indicates that the processor does support
9081 floating-point instructions. The @option{-msoft-float} option indicates
9082 that floating-point support should not be assumed.
9084 @item -mleaf-procedures
9085 @itemx -mno-leaf-procedures
9086 @opindex mleaf-procedures
9087 @opindex mno-leaf-procedures
9088 Do (or do not) attempt to alter leaf procedures to be callable with the
9089 @code{bal} instruction as well as @code{call}. This will result in more
9090 efficient code for explicit calls when the @code{bal} instruction can be
9091 substituted by the assembler or linker, but less efficient code in other
9092 cases, such as calls via function pointers, or using a linker that doesn't
9093 support this optimization.
9096 @itemx -mno-tail-call
9098 @opindex mno-tail-call
9099 Do (or do not) make additional attempts (beyond those of the
9100 machine-independent portions of the compiler) to optimize tail-recursive
9101 calls into branches. You may not want to do this because the detection of
9102 cases where this is not valid is not totally complete. The default is
9103 @option{-mno-tail-call}.
9105 @item -mcomplex-addr
9106 @itemx -mno-complex-addr
9107 @opindex mcomplex-addr
9108 @opindex mno-complex-addr
9109 Assume (or do not assume) that the use of a complex addressing mode is a
9110 win on this implementation of the i960. Complex addressing modes may not
9111 be worthwhile on the K-series, but they definitely are on the C-series.
9112 The default is currently @option{-mcomplex-addr} for all processors except
9116 @itemx -mno-code-align
9117 @opindex mcode-align
9118 @opindex mno-code-align
9119 Align code to 8-byte boundaries for faster fetching (or don't bother).
9120 Currently turned on by default for C-series implementations only.
9123 @item -mclean-linkage
9124 @itemx -mno-clean-linkage
9125 @opindex mclean-linkage
9126 @opindex mno-clean-linkage
9127 These options are not fully implemented.
9131 @itemx -mic2.0-compat
9132 @itemx -mic3.0-compat
9134 @opindex mic2.0-compat
9135 @opindex mic3.0-compat
9136 Enable compatibility with iC960 v2.0 or v3.0.
9140 @opindex masm-compat
9142 Enable compatibility with the iC960 assembler.
9144 @item -mstrict-align
9145 @itemx -mno-strict-align
9146 @opindex mstrict-align
9147 @opindex mno-strict-align
9148 Do not permit (do permit) unaligned accesses.
9152 Enable structure-alignment compatibility with Intel's gcc release version
9153 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
9155 @item -mlong-double-64
9156 @opindex mlong-double-64
9157 Implement type @samp{long double} as 64-bit floating point numbers.
9158 Without the option @samp{long double} is implemented by 80-bit
9159 floating point numbers. The only reason we have it because there is
9160 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
9161 is only useful for people using soft-float targets. Otherwise, we
9162 should recommend against use of it.
9166 @node DEC Alpha Options
9167 @subsection DEC Alpha Options
9169 These @samp{-m} options are defined for the DEC Alpha implementations:
9172 @item -mno-soft-float
9174 @opindex mno-soft-float
9175 @opindex msoft-float
9176 Use (do not use) the hardware floating-point instructions for
9177 floating-point operations. When @option{-msoft-float} is specified,
9178 functions in @file{libgcc.a} will be used to perform floating-point
9179 operations. Unless they are replaced by routines that emulate the
9180 floating-point operations, or compiled in such a way as to call such
9181 emulations routines, these routines will issue floating-point
9182 operations. If you are compiling for an Alpha without floating-point
9183 operations, you must ensure that the library is built so as not to call
9186 Note that Alpha implementations without floating-point operations are
9187 required to have floating-point registers.
9192 @opindex mno-fp-regs
9193 Generate code that uses (does not use) the floating-point register set.
9194 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
9195 register set is not used, floating point operands are passed in integer
9196 registers as if they were integers and floating-point results are passed
9197 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
9198 so any function with a floating-point argument or return value called by code
9199 compiled with @option{-mno-fp-regs} must also be compiled with that
9202 A typical use of this option is building a kernel that does not use,
9203 and hence need not save and restore, any floating-point registers.
9207 The Alpha architecture implements floating-point hardware optimized for
9208 maximum performance. It is mostly compliant with the IEEE floating
9209 point standard. However, for full compliance, software assistance is
9210 required. This option generates code fully IEEE compliant code
9211 @emph{except} that the @var{inexact-flag} is not maintained (see below).
9212 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
9213 defined during compilation. The resulting code is less efficient but is
9214 able to correctly support denormalized numbers and exceptional IEEE
9215 values such as not-a-number and plus/minus infinity. Other Alpha
9216 compilers call this option @option{-ieee_with_no_inexact}.
9218 @item -mieee-with-inexact
9219 @opindex mieee-with-inexact
9220 This is like @option{-mieee} except the generated code also maintains
9221 the IEEE @var{inexact-flag}. Turning on this option causes the
9222 generated code to implement fully-compliant IEEE math. In addition to
9223 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
9224 macro. On some Alpha implementations the resulting code may execute
9225 significantly slower than the code generated by default. Since there is
9226 very little code that depends on the @var{inexact-flag}, you should
9227 normally not specify this option. Other Alpha compilers call this
9228 option @option{-ieee_with_inexact}.
9230 @item -mfp-trap-mode=@var{trap-mode}
9231 @opindex mfp-trap-mode
9232 This option controls what floating-point related traps are enabled.
9233 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9234 The trap mode can be set to one of four values:
9238 This is the default (normal) setting. The only traps that are enabled
9239 are the ones that cannot be disabled in software (e.g., division by zero
9243 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9247 Like @samp{su}, but the instructions are marked to be safe for software
9248 completion (see Alpha architecture manual for details).
9251 Like @samp{su}, but inexact traps are enabled as well.
9254 @item -mfp-rounding-mode=@var{rounding-mode}
9255 @opindex mfp-rounding-mode
9256 Selects the IEEE rounding mode. Other Alpha compilers call this option
9257 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9262 Normal IEEE rounding mode. Floating point numbers are rounded towards
9263 the nearest machine number or towards the even machine number in case
9267 Round towards minus infinity.
9270 Chopped rounding mode. Floating point numbers are rounded towards zero.
9273 Dynamic rounding mode. A field in the floating point control register
9274 (@var{fpcr}, see Alpha architecture reference manual) controls the
9275 rounding mode in effect. The C library initializes this register for
9276 rounding towards plus infinity. Thus, unless your program modifies the
9277 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9280 @item -mtrap-precision=@var{trap-precision}
9281 @opindex mtrap-precision
9282 In the Alpha architecture, floating point traps are imprecise. This
9283 means without software assistance it is impossible to recover from a
9284 floating trap and program execution normally needs to be terminated.
9285 GCC can generate code that can assist operating system trap handlers
9286 in determining the exact location that caused a floating point trap.
9287 Depending on the requirements of an application, different levels of
9288 precisions can be selected:
9292 Program precision. This option is the default and means a trap handler
9293 can only identify which program caused a floating point exception.
9296 Function precision. The trap handler can determine the function that
9297 caused a floating point exception.
9300 Instruction precision. The trap handler can determine the exact
9301 instruction that caused a floating point exception.
9304 Other Alpha compilers provide the equivalent options called
9305 @option{-scope_safe} and @option{-resumption_safe}.
9307 @item -mieee-conformant
9308 @opindex mieee-conformant
9309 This option marks the generated code as IEEE conformant. You must not
9310 use this option unless you also specify @option{-mtrap-precision=i} and either
9311 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9312 is to emit the line @samp{.eflag 48} in the function prologue of the
9313 generated assembly file. Under DEC Unix, this has the effect that
9314 IEEE-conformant math library routines will be linked in.
9316 @item -mbuild-constants
9317 @opindex mbuild-constants
9318 Normally GCC examines a 32- or 64-bit integer constant to
9319 see if it can construct it from smaller constants in two or three
9320 instructions. If it cannot, it will output the constant as a literal and
9321 generate code to load it from the data segment at runtime.
9323 Use this option to require GCC to construct @emph{all} integer constants
9324 using code, even if it takes more instructions (the maximum is six).
9326 You would typically use this option to build a shared library dynamic
9327 loader. Itself a shared library, it must relocate itself in memory
9328 before it can find the variables and constants in its own data segment.
9334 Select whether to generate code to be assembled by the vendor-supplied
9335 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9353 Indicate whether GCC should generate code to use the optional BWX,
9354 CIX, FIX and MAX instruction sets. The default is to use the instruction
9355 sets supported by the CPU type specified via @option{-mcpu=} option or that
9356 of the CPU on which GCC was built if none was specified.
9361 @opindex mfloat-ieee
9362 Generate code that uses (does not use) VAX F and G floating point
9363 arithmetic instead of IEEE single and double precision.
9365 @item -mexplicit-relocs
9366 @itemx -mno-explicit-relocs
9367 @opindex mexplicit-relocs
9368 @opindex mno-explicit-relocs
9369 Older Alpha assemblers provided no way to generate symbol relocations
9370 except via assembler macros. Use of these macros does not allow
9371 optimal instruction scheduling. GNU binutils as of version 2.12
9372 supports a new syntax that allows the compiler to explicitly mark
9373 which relocations should apply to which instructions. This option
9374 is mostly useful for debugging, as GCC detects the capabilities of
9375 the assembler when it is built and sets the default accordingly.
9379 @opindex msmall-data
9380 @opindex mlarge-data
9381 When @option{-mexplicit-relocs} is in effect, static data is
9382 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9383 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9384 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9385 16-bit relocations off of the @code{$gp} register. This limits the
9386 size of the small data area to 64KB, but allows the variables to be
9387 directly accessed via a single instruction.
9389 The default is @option{-mlarge-data}. With this option the data area
9390 is limited to just below 2GB. Programs that require more than 2GB of
9391 data must use @code{malloc} or @code{mmap} to allocate the data in the
9392 heap instead of in the program's data segment.
9394 When generating code for shared libraries, @option{-fpic} implies
9395 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9399 @opindex msmall-text
9400 @opindex mlarge-text
9401 When @option{-msmall-text} is used, the compiler assumes that the
9402 code of the entire program (or shared library) fits in 4MB, and is
9403 thus reachable with a branch instruction. When @option{-msmall-data}
9404 is used, the compiler can assume that all local symbols share the
9405 same @code{$gp} value, and thus reduce the number of instructions
9406 required for a function call from 4 to 1.
9408 The default is @option{-mlarge-text}.
9410 @item -mcpu=@var{cpu_type}
9412 Set the instruction set and instruction scheduling parameters for
9413 machine type @var{cpu_type}. You can specify either the @samp{EV}
9414 style name or the corresponding chip number. GCC supports scheduling
9415 parameters for the EV4, EV5 and EV6 family of processors and will
9416 choose the default values for the instruction set from the processor
9417 you specify. If you do not specify a processor type, GCC will default
9418 to the processor on which the compiler was built.
9420 Supported values for @var{cpu_type} are
9426 Schedules as an EV4 and has no instruction set extensions.
9430 Schedules as an EV5 and has no instruction set extensions.
9434 Schedules as an EV5 and supports the BWX extension.
9439 Schedules as an EV5 and supports the BWX and MAX extensions.
9443 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9447 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9450 @item -mtune=@var{cpu_type}
9452 Set only the instruction scheduling parameters for machine type
9453 @var{cpu_type}. The instruction set is not changed.
9455 @item -mmemory-latency=@var{time}
9456 @opindex mmemory-latency
9457 Sets the latency the scheduler should assume for typical memory
9458 references as seen by the application. This number is highly
9459 dependent on the memory access patterns used by the application
9460 and the size of the external cache on the machine.
9462 Valid options for @var{time} are
9466 A decimal number representing clock cycles.
9472 The compiler contains estimates of the number of clock cycles for
9473 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9474 (also called Dcache, Scache, and Bcache), as well as to main memory.
9475 Note that L3 is only valid for EV5.
9480 @node DEC Alpha/VMS Options
9481 @subsection DEC Alpha/VMS Options
9483 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9486 @item -mvms-return-codes
9487 @opindex mvms-return-codes
9488 Return VMS condition codes from main. The default is to return POSIX
9489 style condition (e.g.@ error) codes.
9492 @node H8/300 Options
9493 @subsection H8/300 Options
9495 These @samp{-m} options are defined for the H8/300 implementations:
9500 Shorten some address references at link time, when possible; uses the
9501 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9502 ld.info, Using ld}, for a fuller description.
9506 Generate code for the H8/300H@.
9510 Generate code for the H8S@.
9514 Generate code for the H8S and H8/300H in the normal mode. This switch
9515 must be used either with -mh or -ms.
9519 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9523 Make @code{int} data 32 bits by default.
9527 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9528 The default for the H8/300H and H8S is to align longs and floats on 4
9530 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9531 This option has no effect on the H8/300.
9535 @subsection SH Options
9537 These @samp{-m} options are defined for the SH implementations:
9542 Generate code for the SH1.
9546 Generate code for the SH2.
9549 Generate code for the SH2e.
9553 Generate code for the SH3.
9557 Generate code for the SH3e.
9561 Generate code for the SH4 without a floating-point unit.
9563 @item -m4-single-only
9564 @opindex m4-single-only
9565 Generate code for the SH4 with a floating-point unit that only
9566 supports single-precision arithmetic.
9570 Generate code for the SH4 assuming the floating-point unit is in
9571 single-precision mode by default.
9575 Generate code for the SH4.
9579 Compile code for the processor in big endian mode.
9583 Compile code for the processor in little endian mode.
9587 Align doubles at 64-bit boundaries. Note that this changes the calling
9588 conventions, and thus some functions from the standard C library will
9589 not work unless you recompile it first with @option{-mdalign}.
9593 Shorten some address references at link time, when possible; uses the
9594 linker option @option{-relax}.
9598 Use 32-bit offsets in @code{switch} tables. The default is to use
9603 Enable the use of the instruction @code{fmovd}.
9607 Comply with the calling conventions defined by Renesas.
9611 Mark the @code{MAC} register as call-clobbered, even if
9612 @option{-mhitachi} is given.
9616 Increase IEEE-compliance of floating-point code.
9620 Dump instruction size and location in the assembly code.
9624 This option is deprecated. It pads structures to multiple of 4 bytes,
9625 which is incompatible with the SH ABI@.
9629 Optimize for space instead of speed. Implied by @option{-Os}.
9633 When generating position-independent code, emit function calls using
9634 the Global Offset Table instead of the Procedure Linkage Table.
9638 Generate a library function call to invalidate instruction cache
9639 entries, after fixing up a trampoline. This library function call
9640 doesn't assume it can write to the whole memory address space. This
9641 is the default when the target is @code{sh-*-linux*}.
9644 @node System V Options
9645 @subsection Options for System V
9647 These additional options are available on System V Release 4 for
9648 compatibility with other compilers on those systems:
9653 Create a shared object.
9654 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
9658 Identify the versions of each tool used by the compiler, in a
9659 @code{.ident} assembler directive in the output.
9663 Refrain from adding @code{.ident} directives to the output file (this is
9666 @item -YP,@var{dirs}
9668 Search the directories @var{dirs}, and no others, for libraries
9669 specified with @option{-l}.
9673 Look in the directory @var{dir} to find the M4 preprocessor.
9674 The assembler uses this option.
9675 @c This is supposed to go with a -Yd for predefined M4 macro files, but
9676 @c the generic assembler that comes with Solaris takes just -Ym.
9679 @node TMS320C3x/C4x Options
9680 @subsection TMS320C3x/C4x Options
9681 @cindex TMS320C3x/C4x Options
9683 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
9687 @item -mcpu=@var{cpu_type}
9689 Set the instruction set, register set, and instruction scheduling
9690 parameters for machine type @var{cpu_type}. Supported values for
9691 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
9692 @samp{c44}. The default is @samp{c40} to generate code for the
9697 @itemx -msmall-memory
9699 @opindex mbig-memory
9701 @opindex msmall-memory
9703 Generates code for the big or small memory model. The small memory
9704 model assumed that all data fits into one 64K word page. At run-time
9705 the data page (DP) register must be set to point to the 64K page
9706 containing the .bss and .data program sections. The big memory model is
9707 the default and requires reloading of the DP register for every direct
9714 Allow (disallow) allocation of general integer operands into the block
9721 Enable (disable) generation of code using decrement and branch,
9722 DBcond(D), instructions. This is enabled by default for the C4x. To be
9723 on the safe side, this is disabled for the C3x, since the maximum
9724 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
9725 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
9726 that it can utilize the decrement and branch instruction, but will give
9727 up if there is more than one memory reference in the loop. Thus a loop
9728 where the loop counter is decremented can generate slightly more
9729 efficient code, in cases where the RPTB instruction cannot be utilized.
9731 @item -mdp-isr-reload
9733 @opindex mdp-isr-reload
9735 Force the DP register to be saved on entry to an interrupt service
9736 routine (ISR), reloaded to point to the data section, and restored on
9737 exit from the ISR@. This should not be required unless someone has
9738 violated the small memory model by modifying the DP register, say within
9745 For the C3x use the 24-bit MPYI instruction for integer multiplies
9746 instead of a library call to guarantee 32-bit results. Note that if one
9747 of the operands is a constant, then the multiplication will be performed
9748 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
9749 then squaring operations are performed inline instead of a library call.
9752 @itemx -mno-fast-fix
9754 @opindex mno-fast-fix
9755 The C3x/C4x FIX instruction to convert a floating point value to an
9756 integer value chooses the nearest integer less than or equal to the
9757 floating point value rather than to the nearest integer. Thus if the
9758 floating point number is negative, the result will be incorrectly
9759 truncated an additional code is necessary to detect and correct this
9760 case. This option can be used to disable generation of the additional
9761 code required to correct the result.
9767 Enable (disable) generation of repeat block sequences using the RPTB
9768 instruction for zero overhead looping. The RPTB construct is only used
9769 for innermost loops that do not call functions or jump across the loop
9770 boundaries. There is no advantage having nested RPTB loops due to the
9771 overhead required to save and restore the RC, RS, and RE registers.
9772 This is enabled by default with @option{-O2}.
9774 @item -mrpts=@var{count}
9778 Enable (disable) the use of the single instruction repeat instruction
9779 RPTS@. If a repeat block contains a single instruction, and the loop
9780 count can be guaranteed to be less than the value @var{count}, GCC will
9781 emit a RPTS instruction instead of a RPTB@. If no value is specified,
9782 then a RPTS will be emitted even if the loop count cannot be determined
9783 at compile time. Note that the repeated instruction following RPTS does
9784 not have to be reloaded from memory each iteration, thus freeing up the
9785 CPU buses for operands. However, since interrupts are blocked by this
9786 instruction, it is disabled by default.
9788 @item -mloop-unsigned
9789 @itemx -mno-loop-unsigned
9790 @opindex mloop-unsigned
9791 @opindex mno-loop-unsigned
9792 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
9793 is @math{2^{31} + 1} since these instructions test if the iteration count is
9794 negative to terminate the loop. If the iteration count is unsigned
9795 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
9796 exceeded. This switch allows an unsigned iteration count.
9800 Try to emit an assembler syntax that the TI assembler (asm30) is happy
9801 with. This also enforces compatibility with the API employed by the TI
9802 C3x C compiler. For example, long doubles are passed as structures
9803 rather than in floating point registers.
9809 Generate code that uses registers (stack) for passing arguments to functions.
9810 By default, arguments are passed in registers where possible rather
9811 than by pushing arguments on to the stack.
9813 @item -mparallel-insns
9814 @itemx -mno-parallel-insns
9815 @opindex mparallel-insns
9816 @opindex mno-parallel-insns
9817 Allow the generation of parallel instructions. This is enabled by
9818 default with @option{-O2}.
9820 @item -mparallel-mpy
9821 @itemx -mno-parallel-mpy
9822 @opindex mparallel-mpy
9823 @opindex mno-parallel-mpy
9824 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
9825 provided @option{-mparallel-insns} is also specified. These instructions have
9826 tight register constraints which can pessimize the code generation
9832 @subsection V850 Options
9833 @cindex V850 Options
9835 These @samp{-m} options are defined for V850 implementations:
9839 @itemx -mno-long-calls
9840 @opindex mlong-calls
9841 @opindex mno-long-calls
9842 Treat all calls as being far away (near). If calls are assumed to be
9843 far away, the compiler will always load the functions address up into a
9844 register, and call indirect through the pointer.
9850 Do not optimize (do optimize) basic blocks that use the same index
9851 pointer 4 or more times to copy pointer into the @code{ep} register, and
9852 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
9853 option is on by default if you optimize.
9855 @item -mno-prolog-function
9856 @itemx -mprolog-function
9857 @opindex mno-prolog-function
9858 @opindex mprolog-function
9859 Do not use (do use) external functions to save and restore registers
9860 at the prologue and epilogue of a function. The external functions
9861 are slower, but use less code space if more than one function saves
9862 the same number of registers. The @option{-mprolog-function} option
9863 is on by default if you optimize.
9867 Try to make the code as small as possible. At present, this just turns
9868 on the @option{-mep} and @option{-mprolog-function} options.
9872 Put static or global variables whose size is @var{n} bytes or less into
9873 the tiny data area that register @code{ep} points to. The tiny data
9874 area can hold up to 256 bytes in total (128 bytes for byte references).
9878 Put static or global variables whose size is @var{n} bytes or less into
9879 the small data area that register @code{gp} points to. The small data
9880 area can hold up to 64 kilobytes.
9884 Put static or global variables whose size is @var{n} bytes or less into
9885 the first 32 kilobytes of memory.
9889 Specify that the target processor is the V850.
9892 @opindex mbig-switch
9893 Generate code suitable for big switch tables. Use this option only if
9894 the assembler/linker complain about out of range branches within a switch
9899 This option will cause r2 and r5 to be used in the code generated by
9900 the compiler. This setting is the default.
9903 @opindex mno-app-regs
9904 This option will cause r2 and r5 to be treated as fixed registers.
9908 Specify that the target processor is the V850E1. The preprocessor
9909 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
9910 this option is used.
9914 Specify that the target processor is the V850E. The preprocessor
9915 constant @samp{__v850e__} will be defined if this option is used.
9917 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
9918 are defined then a default target processor will be chosen and the
9919 relevant @samp{__v850*__} preprocessor constant will be defined.
9921 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
9922 defined, regardless of which processor variant is the target.
9924 @item -mdisable-callt
9925 @opindex mdisable-callt
9926 This option will suppress generation of the CALLT instruction for the
9927 v850e and v850e1 flavors of the v850 architecture. The default is
9928 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
9933 @subsection ARC Options
9936 These options are defined for ARC implementations:
9941 Compile code for little endian mode. This is the default.
9945 Compile code for big endian mode.
9948 @opindex mmangle-cpu
9949 Prepend the name of the cpu to all public symbol names.
9950 In multiple-processor systems, there are many ARC variants with different
9951 instruction and register set characteristics. This flag prevents code
9952 compiled for one cpu to be linked with code compiled for another.
9953 No facility exists for handling variants that are ``almost identical''.
9954 This is an all or nothing option.
9956 @item -mcpu=@var{cpu}
9958 Compile code for ARC variant @var{cpu}.
9959 Which variants are supported depend on the configuration.
9960 All variants support @option{-mcpu=base}, this is the default.
9962 @item -mtext=@var{text-section}
9963 @itemx -mdata=@var{data-section}
9964 @itemx -mrodata=@var{readonly-data-section}
9968 Put functions, data, and readonly data in @var{text-section},
9969 @var{data-section}, and @var{readonly-data-section} respectively
9970 by default. This can be overridden with the @code{section} attribute.
9971 @xref{Variable Attributes}.
9976 @subsection NS32K Options
9977 @cindex NS32K options
9979 These are the @samp{-m} options defined for the 32000 series. The default
9980 values for these options depends on which style of 32000 was selected when
9981 the compiler was configured; the defaults for the most common choices are
9989 Generate output for a 32032. This is the default
9990 when the compiler is configured for 32032 and 32016 based systems.
9996 Generate output for a 32332. This is the default
9997 when the compiler is configured for 32332-based systems.
10003 Generate output for a 32532. This is the default
10004 when the compiler is configured for 32532-based systems.
10008 Generate output containing 32081 instructions for floating point.
10009 This is the default for all systems.
10013 Generate output containing 32381 instructions for floating point. This
10014 also implies @option{-m32081}. The 32381 is only compatible with the 32332
10015 and 32532 cpus. This is the default for the pc532-netbsd configuration.
10018 @opindex mmulti-add
10019 Try and generate multiply-add floating point instructions @code{polyF}
10020 and @code{dotF}. This option is only available if the @option{-m32381}
10021 option is in effect. Using these instructions requires changes to
10022 register allocation which generally has a negative impact on
10023 performance. This option should only be enabled when compiling code
10024 particularly likely to make heavy use of multiply-add instructions.
10026 @item -mnomulti-add
10027 @opindex mnomulti-add
10028 Do not try and generate multiply-add floating point instructions
10029 @code{polyF} and @code{dotF}. This is the default on all platforms.
10032 @opindex msoft-float
10033 Generate output containing library calls for floating point.
10034 @strong{Warning:} the requisite libraries may not be available.
10036 @item -mieee-compare
10037 @itemx -mno-ieee-compare
10038 @opindex mieee-compare
10039 @opindex mno-ieee-compare
10040 Control whether or not the compiler uses IEEE floating point
10041 comparisons. These handle correctly the case where the result of a
10042 comparison is unordered.
10043 @strong{Warning:} the requisite kernel support may not be available.
10046 @opindex mnobitfield
10047 Do not use the bit-field instructions. On some machines it is faster to
10048 use shifting and masking operations. This is the default for the pc532.
10052 Do use the bit-field instructions. This is the default for all platforms
10057 Use a different function-calling convention, in which functions
10058 that take a fixed number of arguments return pop their
10059 arguments on return with the @code{ret} instruction.
10061 This calling convention is incompatible with the one normally
10062 used on Unix, so you cannot use it if you need to call libraries
10063 compiled with the Unix compiler.
10065 Also, you must provide function prototypes for all functions that
10066 take variable numbers of arguments (including @code{printf});
10067 otherwise incorrect code will be generated for calls to those
10070 In addition, seriously incorrect code will result if you call a
10071 function with too many arguments. (Normally, extra arguments are
10072 harmlessly ignored.)
10074 This option takes its name from the 680x0 @code{rtd} instruction.
10079 Use a different function-calling convention where the first two arguments
10080 are passed in registers.
10082 This calling convention is incompatible with the one normally
10083 used on Unix, so you cannot use it if you need to call libraries
10084 compiled with the Unix compiler.
10087 @opindex mnoregparam
10088 Do not pass any arguments in registers. This is the default for all
10093 It is OK to use the sb as an index register which is always loaded with
10094 zero. This is the default for the pc532-netbsd target.
10098 The sb register is not available for use or has not been initialized to
10099 zero by the run time system. This is the default for all targets except
10100 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
10101 @option{-fpic} is set.
10105 Many ns32000 series addressing modes use displacements of up to 512MB@.
10106 If an address is above 512MB then displacements from zero can not be used.
10107 This option causes code to be generated which can be loaded above 512MB@.
10108 This may be useful for operating systems or ROM code.
10112 Assume code will be loaded in the first 512MB of virtual address space.
10113 This is the default for all platforms.
10119 @subsection AVR Options
10120 @cindex AVR Options
10122 These options are defined for AVR implementations:
10125 @item -mmcu=@var{mcu}
10127 Specify ATMEL AVR instruction set or MCU type.
10129 Instruction set avr1 is for the minimal AVR core, not supported by the C
10130 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
10131 attiny11, attiny12, attiny15, attiny28).
10133 Instruction set avr2 (default) is for the classic AVR core with up to
10134 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
10135 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
10136 at90c8534, at90s8535).
10138 Instruction set avr3 is for the classic AVR core with up to 128K program
10139 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
10141 Instruction set avr4 is for the enhanced AVR core with up to 8K program
10142 memory space (MCU types: atmega8, atmega83, atmega85).
10144 Instruction set avr5 is for the enhanced AVR core with up to 128K program
10145 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
10146 atmega64, atmega128, at43usb355, at94k).
10150 Output instruction sizes to the asm file.
10152 @item -minit-stack=@var{N}
10153 @opindex minit-stack
10154 Specify the initial stack address, which may be a symbol or numeric value,
10155 @samp{__stack} is the default.
10157 @item -mno-interrupts
10158 @opindex mno-interrupts
10159 Generated code is not compatible with hardware interrupts.
10160 Code size will be smaller.
10162 @item -mcall-prologues
10163 @opindex mcall-prologues
10164 Functions prologues/epilogues expanded as call to appropriate
10165 subroutines. Code size will be smaller.
10167 @item -mno-tablejump
10168 @opindex mno-tablejump
10169 Do not generate tablejump insns which sometimes increase code size.
10172 @opindex mtiny-stack
10173 Change only the low 8 bits of the stack pointer.
10176 @node MCore Options
10177 @subsection MCore Options
10178 @cindex MCore options
10180 These are the @samp{-m} options defined for the Motorola M*Core
10186 @itemx -mno-hardlit
10188 @opindex mno-hardlit
10189 Inline constants into the code stream if it can be done in two
10190 instructions or less.
10196 Use the divide instruction. (Enabled by default).
10198 @item -mrelax-immediate
10199 @itemx -mno-relax-immediate
10200 @opindex mrelax-immediate
10201 @opindex mno-relax-immediate
10202 Allow arbitrary sized immediates in bit operations.
10204 @item -mwide-bitfields
10205 @itemx -mno-wide-bitfields
10206 @opindex mwide-bitfields
10207 @opindex mno-wide-bitfields
10208 Always treat bit-fields as int-sized.
10210 @item -m4byte-functions
10211 @itemx -mno-4byte-functions
10212 @opindex m4byte-functions
10213 @opindex mno-4byte-functions
10214 Force all functions to be aligned to a four byte boundary.
10216 @item -mcallgraph-data
10217 @itemx -mno-callgraph-data
10218 @opindex mcallgraph-data
10219 @opindex mno-callgraph-data
10220 Emit callgraph information.
10223 @itemx -mno-slow-bytes
10224 @opindex mslow-bytes
10225 @opindex mno-slow-bytes
10226 Prefer word access when reading byte quantities.
10228 @item -mlittle-endian
10229 @itemx -mbig-endian
10230 @opindex mlittle-endian
10231 @opindex mbig-endian
10232 Generate code for a little endian target.
10238 Generate code for the 210 processor.
10241 @node IA-64 Options
10242 @subsection IA-64 Options
10243 @cindex IA-64 Options
10245 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10249 @opindex mbig-endian
10250 Generate code for a big endian target. This is the default for HP-UX@.
10252 @item -mlittle-endian
10253 @opindex mlittle-endian
10254 Generate code for a little endian target. This is the default for AIX5
10260 @opindex mno-gnu-as
10261 Generate (or don't) code for the GNU assembler. This is the default.
10262 @c Also, this is the default if the configure option @option{--with-gnu-as}
10268 @opindex mno-gnu-ld
10269 Generate (or don't) code for the GNU linker. This is the default.
10270 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10275 Generate code that does not use a global pointer register. The result
10276 is not position independent code, and violates the IA-64 ABI@.
10278 @item -mvolatile-asm-stop
10279 @itemx -mno-volatile-asm-stop
10280 @opindex mvolatile-asm-stop
10281 @opindex mno-volatile-asm-stop
10282 Generate (or don't) a stop bit immediately before and after volatile asm
10287 Generate code that works around Itanium B step errata.
10289 @item -mregister-names
10290 @itemx -mno-register-names
10291 @opindex mregister-names
10292 @opindex mno-register-names
10293 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
10294 the stacked registers. This may make assembler output more readable.
10300 Disable (or enable) optimizations that use the small data section. This may
10301 be useful for working around optimizer bugs.
10303 @item -mconstant-gp
10304 @opindex mconstant-gp
10305 Generate code that uses a single constant global pointer value. This is
10306 useful when compiling kernel code.
10310 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
10311 This is useful when compiling firmware code.
10313 @item -minline-float-divide-min-latency
10314 @opindex minline-float-divide-min-latency
10315 Generate code for inline divides of floating point values
10316 using the minimum latency algorithm.
10318 @item -minline-float-divide-max-throughput
10319 @opindex minline-float-divide-max-throughput
10320 Generate code for inline divides of floating point values
10321 using the maximum throughput algorithm.
10323 @item -minline-int-divide-min-latency
10324 @opindex minline-int-divide-min-latency
10325 Generate code for inline divides of integer values
10326 using the minimum latency algorithm.
10328 @item -minline-int-divide-max-throughput
10329 @opindex minline-int-divide-max-throughput
10330 Generate code for inline divides of integer values
10331 using the maximum throughput algorithm.
10333 @item -mno-dwarf2-asm
10334 @itemx -mdwarf2-asm
10335 @opindex mno-dwarf2-asm
10336 @opindex mdwarf2-asm
10337 Don't (or do) generate assembler code for the DWARF2 line number debugging
10338 info. This may be useful when not using the GNU assembler.
10340 @item -mfixed-range=@var{register-range}
10341 @opindex mfixed-range
10342 Generate code treating the given register range as fixed registers.
10343 A fixed register is one that the register allocator can not use. This is
10344 useful when compiling kernel code. A register range is specified as
10345 two registers separated by a dash. Multiple register ranges can be
10346 specified separated by a comma.
10348 @item -mearly-stop-bits
10349 @itemx -mno-early-stop-bits
10350 @opindex mearly-stop-bits
10351 @opindex mno-early-stop-bits
10352 Allow stop bits to be placed earlier than immediately preceding the
10353 instruction that triggered the stop bit. This can improve instruction
10354 scheduling, but does not always do so.
10358 @subsection D30V Options
10359 @cindex D30V Options
10361 These @samp{-m} options are defined for D30V implementations:
10366 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
10367 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
10368 memory, which starts at location @code{0x80000000}.
10371 @opindex mextmemory
10372 Same as the @option{-mextmem} switch.
10376 Link the @samp{.text} section into onchip text memory, which starts at
10377 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
10378 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
10379 into onchip data memory, which starts at location @code{0x20000000}.
10381 @item -mno-asm-optimize
10382 @itemx -masm-optimize
10383 @opindex mno-asm-optimize
10384 @opindex masm-optimize
10385 Disable (enable) passing @option{-O} to the assembler when optimizing.
10386 The assembler uses the @option{-O} option to automatically parallelize
10387 adjacent short instructions where possible.
10389 @item -mbranch-cost=@var{n}
10390 @opindex mbranch-cost
10391 Increase the internal costs of branches to @var{n}. Higher costs means
10392 that the compiler will issue more instructions to avoid doing a branch.
10395 @item -mcond-exec=@var{n}
10396 @opindex mcond-exec
10397 Specify the maximum number of conditionally executed instructions that
10398 replace a branch. The default is 4.
10401 @node S/390 and zSeries Options
10402 @subsection S/390 and zSeries Options
10403 @cindex S/390 and zSeries Options
10405 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
10409 @itemx -msoft-float
10410 @opindex mhard-float
10411 @opindex msoft-float
10412 Use (do not use) the hardware floating-point instructions and registers
10413 for floating-point operations. When @option{-msoft-float} is specified,
10414 functions in @file{libgcc.a} will be used to perform floating-point
10415 operations. When @option{-mhard-float} is specified, the compiler
10416 generates IEEE floating-point instructions. This is the default.
10419 @itemx -mno-backchain
10420 @opindex mbackchain
10421 @opindex mno-backchain
10422 Generate (or do not generate) code which maintains an explicit
10423 backchain within the stack frame that points to the caller's frame.
10424 This is currently needed to allow debugging. The default is to
10425 generate the backchain.
10428 @itemx -mno-small-exec
10429 @opindex msmall-exec
10430 @opindex mno-small-exec
10431 Generate (or do not generate) code using the @code{bras} instruction
10432 to do subroutine calls.
10433 This only works reliably if the total executable size does not
10434 exceed 64k. The default is to use the @code{basr} instruction instead,
10435 which does not have this limitation.
10441 When @option{-m31} is specified, generate code compliant to the
10442 Linux for S/390 ABI@. When @option{-m64} is specified, generate
10443 code compliant to the Linux for zSeries ABI@. This allows GCC in
10444 particular to generate 64-bit instructions. For the @samp{s390}
10445 targets, the default is @option{-m31}, while the @samp{s390x}
10446 targets default to @option{-m64}.
10452 When @option{-mzarch} is specified, generate code using the
10453 instructions available on z/Architecture.
10454 When @option{-mesa} is specified, generate code using the
10455 instructions available on ESA/390. Note that @option{-mesa} is
10456 not possible with @option{-m64}.
10457 When generating code compliant to the Linux for S/390 ABI,
10458 the default is @option{-mesa}. When generating code compliant
10459 to the Linux for zSeries ABI, the default is @option{-mzarch}.
10465 Generate (or do not generate) code using the @code{mvcle} instruction
10466 to perform block moves. When @option{-mno-mvcle} is specified,
10467 use a @code{mvc} loop instead. This is the default.
10473 Print (or do not print) additional debug information when compiling.
10474 The default is to not print debug information.
10476 @item -march=@var{cpu-type}
10478 Generate code that will run on @var{cpu-type}, which is the name of a system
10479 representing a certain processor type. Possible values for
10480 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
10481 When generating code using the instructions available on z/Architecture,
10482 the default is @option{-march=z900}. Otherwise, the default is
10483 @option{-march=g5}.
10485 @item -mtune=@var{cpu-type}
10487 Tune to @var{cpu-type} everything applicable about the generated code,
10488 except for the ABI and the set of available instructions.
10489 The list of @var{cpu-type} values is the same as for @option{-march}.
10490 The default is the value used for @option{-march}.
10493 @itemx -mno-fused-madd
10494 @opindex mfused-madd
10495 @opindex mno-fused-madd
10496 Generate code that uses (does not use) the floating point multiply and
10497 accumulate instructions. These instructions are generated by default if
10498 hardware floating point is used.
10502 @subsection CRIS Options
10503 @cindex CRIS Options
10505 These options are defined specifically for the CRIS ports.
10508 @item -march=@var{architecture-type}
10509 @itemx -mcpu=@var{architecture-type}
10512 Generate code for the specified architecture. The choices for
10513 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10514 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
10515 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10518 @item -mtune=@var{architecture-type}
10520 Tune to @var{architecture-type} everything applicable about the generated
10521 code, except for the ABI and the set of available instructions. The
10522 choices for @var{architecture-type} are the same as for
10523 @option{-march=@var{architecture-type}}.
10525 @item -mmax-stack-frame=@var{n}
10526 @opindex mmax-stack-frame
10527 Warn when the stack frame of a function exceeds @var{n} bytes.
10529 @item -melinux-stacksize=@var{n}
10530 @opindex melinux-stacksize
10531 Only available with the @samp{cris-axis-aout} target. Arranges for
10532 indications in the program to the kernel loader that the stack of the
10533 program should be set to @var{n} bytes.
10539 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10540 @option{-march=v3} and @option{-march=v8} respectively.
10544 Enable CRIS-specific verbose debug-related information in the assembly
10545 code. This option also has the effect to turn off the @samp{#NO_APP}
10546 formatted-code indicator to the assembler at the beginning of the
10551 Do not use condition-code results from previous instruction; always emit
10552 compare and test instructions before use of condition codes.
10554 @item -mno-side-effects
10555 @opindex mno-side-effects
10556 Do not emit instructions with side-effects in addressing modes other than
10559 @item -mstack-align
10560 @itemx -mno-stack-align
10561 @itemx -mdata-align
10562 @itemx -mno-data-align
10563 @itemx -mconst-align
10564 @itemx -mno-const-align
10565 @opindex mstack-align
10566 @opindex mno-stack-align
10567 @opindex mdata-align
10568 @opindex mno-data-align
10569 @opindex mconst-align
10570 @opindex mno-const-align
10571 These options (no-options) arranges (eliminate arrangements) for the
10572 stack-frame, individual data and constants to be aligned for the maximum
10573 single data access size for the chosen CPU model. The default is to
10574 arrange for 32-bit alignment. ABI details such as structure layout are
10575 not affected by these options.
10583 Similar to the stack- data- and const-align options above, these options
10584 arrange for stack-frame, writable data and constants to all be 32-bit,
10585 16-bit or 8-bit aligned. The default is 32-bit alignment.
10587 @item -mno-prologue-epilogue
10588 @itemx -mprologue-epilogue
10589 @opindex mno-prologue-epilogue
10590 @opindex mprologue-epilogue
10591 With @option{-mno-prologue-epilogue}, the normal function prologue and
10592 epilogue that sets up the stack-frame are omitted and no return
10593 instructions or return sequences are generated in the code. Use this
10594 option only together with visual inspection of the compiled code: no
10595 warnings or errors are generated when call-saved registers must be saved,
10596 or storage for local variable needs to be allocated.
10600 @opindex mno-gotplt
10602 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10603 instruction sequences that load addresses for functions from the PLT part
10604 of the GOT rather than (traditional on other architectures) calls to the
10605 PLT. The default is @option{-mgotplt}.
10609 Legacy no-op option only recognized with the cris-axis-aout target.
10613 Legacy no-op option only recognized with the cris-axis-elf and
10614 cris-axis-linux-gnu targets.
10618 Only recognized with the cris-axis-aout target, where it selects a
10619 GNU/linux-like multilib, include files and instruction set for
10620 @option{-march=v8}.
10624 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10628 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
10629 to link with input-output functions from a simulator library. Code,
10630 initialized data and zero-initialized data are allocated consecutively.
10634 Like @option{-sim}, but pass linker options to locate initialized data at
10635 0x40000000 and zero-initialized data at 0x80000000.
10639 @subsection MMIX Options
10640 @cindex MMIX Options
10642 These options are defined for the MMIX:
10646 @itemx -mno-libfuncs
10648 @opindex mno-libfuncs
10649 Specify that intrinsic library functions are being compiled, passing all
10650 values in registers, no matter the size.
10653 @itemx -mno-epsilon
10655 @opindex mno-epsilon
10656 Generate floating-point comparison instructions that compare with respect
10657 to the @code{rE} epsilon register.
10659 @item -mabi=mmixware
10661 @opindex mabi-mmixware
10663 Generate code that passes function parameters and return values that (in
10664 the called function) are seen as registers @code{$0} and up, as opposed to
10665 the GNU ABI which uses global registers @code{$231} and up.
10667 @item -mzero-extend
10668 @itemx -mno-zero-extend
10669 @opindex mzero-extend
10670 @opindex mno-zero-extend
10671 When reading data from memory in sizes shorter than 64 bits, use (do not
10672 use) zero-extending load instructions by default, rather than
10673 sign-extending ones.
10676 @itemx -mno-knuthdiv
10678 @opindex mno-knuthdiv
10679 Make the result of a division yielding a remainder have the same sign as
10680 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
10681 remainder follows the sign of the dividend. Both methods are
10682 arithmetically valid, the latter being almost exclusively used.
10684 @item -mtoplevel-symbols
10685 @itemx -mno-toplevel-symbols
10686 @opindex mtoplevel-symbols
10687 @opindex mno-toplevel-symbols
10688 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
10689 code can be used with the @code{PREFIX} assembly directive.
10693 Generate an executable in the ELF format, rather than the default
10694 @samp{mmo} format used by the @command{mmix} simulator.
10696 @item -mbranch-predict
10697 @itemx -mno-branch-predict
10698 @opindex mbranch-predict
10699 @opindex mno-branch-predict
10700 Use (do not use) the probable-branch instructions, when static branch
10701 prediction indicates a probable branch.
10703 @item -mbase-addresses
10704 @itemx -mno-base-addresses
10705 @opindex mbase-addresses
10706 @opindex mno-base-addresses
10707 Generate (do not generate) code that uses @emph{base addresses}. Using a
10708 base address automatically generates a request (handled by the assembler
10709 and the linker) for a constant to be set up in a global register. The
10710 register is used for one or more base address requests within the range 0
10711 to 255 from the value held in the register. The generally leads to short
10712 and fast code, but the number of different data items that can be
10713 addressed is limited. This means that a program that uses lots of static
10714 data may require @option{-mno-base-addresses}.
10716 @item -msingle-exit
10717 @itemx -mno-single-exit
10718 @opindex msingle-exit
10719 @opindex mno-single-exit
10720 Force (do not force) generated code to have a single exit point in each
10724 @node PDP-11 Options
10725 @subsection PDP-11 Options
10726 @cindex PDP-11 Options
10728 These options are defined for the PDP-11:
10733 Use hardware FPP floating point. This is the default. (FIS floating
10734 point on the PDP-11/40 is not supported.)
10737 @opindex msoft-float
10738 Do not use hardware floating point.
10742 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
10746 Return floating-point results in memory. This is the default.
10750 Generate code for a PDP-11/40.
10754 Generate code for a PDP-11/45. This is the default.
10758 Generate code for a PDP-11/10.
10760 @item -mbcopy-builtin
10761 @opindex bcopy-builtin
10762 Use inline @code{movstrhi} patterns for copying memory. This is the
10767 Do not use inline @code{movstrhi} patterns for copying memory.
10773 Use 16-bit @code{int}. This is the default.
10779 Use 32-bit @code{int}.
10782 @itemx -mno-float32
10784 @opindex mno-float32
10785 Use 64-bit @code{float}. This is the default.
10790 @opindex mno-float64
10791 Use 32-bit @code{float}.
10795 Use @code{abshi2} pattern. This is the default.
10799 Do not use @code{abshi2} pattern.
10801 @item -mbranch-expensive
10802 @opindex mbranch-expensive
10803 Pretend that branches are expensive. This is for experimenting with
10804 code generation only.
10806 @item -mbranch-cheap
10807 @opindex mbranch-cheap
10808 Do not pretend that branches are expensive. This is the default.
10812 Generate code for a system with split I&D.
10816 Generate code for a system without split I&D. This is the default.
10820 Use Unix assembler syntax. This is the default when configured for
10821 @samp{pdp11-*-bsd}.
10825 Use DEC assembler syntax. This is the default when configured for any
10826 PDP-11 target other than @samp{pdp11-*-bsd}.
10829 @node Xstormy16 Options
10830 @subsection Xstormy16 Options
10831 @cindex Xstormy16 Options
10833 These options are defined for Xstormy16:
10838 Choose startup files and linker script suitable for the simulator.
10842 @subsection FRV Options
10843 @cindex FRV Options
10849 Only use the first 32 general purpose registers.
10854 Use all 64 general purpose registers.
10859 Use only the first 32 floating point registers.
10864 Use all 64 floating point registers
10867 @opindex mhard-float
10869 Use hardware instructions for floating point operations.
10872 @opindex msoft-float
10874 Use library routines for floating point operations.
10879 Dynamically allocate condition code registers.
10884 Do not try to dynamically allocate condition code registers, only
10885 use @code{icc0} and @code{fcc0}.
10890 Change ABI to use double word insns.
10895 Do not use double word instructions.
10900 Use floating point double instructions.
10903 @opindex mno-double
10905 Do not use floating point double instructions.
10910 Use media instructions.
10915 Do not use media instructions.
10920 Use multiply and add/subtract instructions.
10923 @opindex mno-muladd
10925 Do not use multiply and add/subtract instructions.
10927 @item -mlibrary-pic
10928 @opindex mlibrary-pic
10930 Enable PIC support for building libraries
10935 Use only the first four media accumulator registers.
10940 Use all eight media accumulator registers.
10945 Pack VLIW instructions.
10950 Do not pack VLIW instructions.
10953 @opindex mno-eflags
10955 Do not mark ABI switches in e_flags.
10958 @opindex mcond-move
10960 Enable the use of conditional-move instructions (default).
10962 This switch is mainly for debugging the compiler and will likely be removed
10963 in a future version.
10965 @item -mno-cond-move
10966 @opindex mno-cond-move
10968 Disable the use of conditional-move instructions.
10970 This switch is mainly for debugging the compiler and will likely be removed
10971 in a future version.
10976 Enable the use of conditional set instructions (default).
10978 This switch is mainly for debugging the compiler and will likely be removed
10979 in a future version.
10984 Disable the use of conditional set instructions.
10986 This switch is mainly for debugging the compiler and will likely be removed
10987 in a future version.
10990 @opindex mcond-exec
10992 Enable the use of conditional execution (default).
10994 This switch is mainly for debugging the compiler and will likely be removed
10995 in a future version.
10997 @item -mno-cond-exec
10998 @opindex mno-cond-exec
11000 Disable the use of conditional execution.
11002 This switch is mainly for debugging the compiler and will likely be removed
11003 in a future version.
11005 @item -mvliw-branch
11006 @opindex mvliw-branch
11008 Run a pass to pack branches into VLIW instructions (default).
11010 This switch is mainly for debugging the compiler and will likely be removed
11011 in a future version.
11013 @item -mno-vliw-branch
11014 @opindex mno-vliw-branch
11016 Do not run a pass to pack branches into VLIW instructions.
11018 This switch is mainly for debugging the compiler and will likely be removed
11019 in a future version.
11021 @item -mmulti-cond-exec
11022 @opindex mmulti-cond-exec
11024 Enable optimization of @code{&&} and @code{||} in conditional execution
11027 This switch is mainly for debugging the compiler and will likely be removed
11028 in a future version.
11030 @item -mno-multi-cond-exec
11031 @opindex mno-multi-cond-exec
11033 Disable optimization of @code{&&} and @code{||} in conditional execution.
11035 This switch is mainly for debugging the compiler and will likely be removed
11036 in a future version.
11038 @item -mnested-cond-exec
11039 @opindex mnested-cond-exec
11041 Enable nested conditional execution optimizations (default).
11043 This switch is mainly for debugging the compiler and will likely be removed
11044 in a future version.
11046 @item -mno-nested-cond-exec
11047 @opindex mno-nested-cond-exec
11049 Disable nested conditional execution optimizations.
11051 This switch is mainly for debugging the compiler and will likely be removed
11052 in a future version.
11054 @item -mtomcat-stats
11055 @opindex mtomcat-stats
11057 Cause gas to print out tomcat statistics.
11059 @item -mcpu=@var{cpu}
11062 Select the processor type for which to generate code. Possible values are
11063 @samp{simple}, @samp{tomcat}, @samp{fr500}, @samp{fr400}, @samp{fr300},
11068 @node Xtensa Options
11069 @subsection Xtensa Options
11070 @cindex Xtensa Options
11072 These options are supported for Xtensa targets:
11076 @itemx -mno-const16
11078 @opindex mno-const16
11079 Enable or disable use of @code{CONST16} instructions for loading
11080 constant values. The @code{CONST16} instruction is currently not a
11081 standard option from Tensilica. When enabled, @code{CONST16}
11082 instructions are always used in place of the standard @code{L32R}
11083 instructions. The use of @code{CONST16} is enabled by default only if
11084 the @code{L32R} instruction is not available.
11087 @itemx -mno-fused-madd
11088 @opindex mfused-madd
11089 @opindex mno-fused-madd
11090 Enable or disable use of fused multiply/add and multiply/subtract
11091 instructions in the floating-point option. This has no effect if the
11092 floating-point option is not also enabled. Disabling fused multiply/add
11093 and multiply/subtract instructions forces the compiler to use separate
11094 instructions for the multiply and add/subtract operations. This may be
11095 desirable in some cases where strict IEEE 754-compliant results are
11096 required: the fused multiply add/subtract instructions do not round the
11097 intermediate result, thereby producing results with @emph{more} bits of
11098 precision than specified by the IEEE standard. Disabling fused multiply
11099 add/subtract instructions also ensures that the program output is not
11100 sensitive to the compiler's ability to combine multiply and add/subtract
11103 @item -mtext-section-literals
11104 @itemx -mno-text-section-literals
11105 @opindex mtext-section-literals
11106 @opindex mno-text-section-literals
11107 Control the treatment of literal pools. The default is
11108 @option{-mno-text-section-literals}, which places literals in a separate
11109 section in the output file. This allows the literal pool to be placed
11110 in a data RAM/ROM, and it also allows the linker to combine literal
11111 pools from separate object files to remove redundant literals and
11112 improve code size. With @option{-mtext-section-literals}, the literals
11113 are interspersed in the text section in order to keep them as close as
11114 possible to their references. This may be necessary for large assembly
11117 @item -mtarget-align
11118 @itemx -mno-target-align
11119 @opindex mtarget-align
11120 @opindex mno-target-align
11121 When this option is enabled, GCC instructs the assembler to
11122 automatically align instructions to reduce branch penalties at the
11123 expense of some code density. The assembler attempts to widen density
11124 instructions to align branch targets and the instructions following call
11125 instructions. If there are not enough preceding safe density
11126 instructions to align a target, no widening will be performed. The
11127 default is @option{-mtarget-align}. These options do not affect the
11128 treatment of auto-aligned instructions like @code{LOOP}, which the
11129 assembler will always align, either by widening density instructions or
11130 by inserting no-op instructions.
11133 @itemx -mno-longcalls
11134 @opindex mlongcalls
11135 @opindex mno-longcalls
11136 When this option is enabled, GCC instructs the assembler to translate
11137 direct calls to indirect calls unless it can determine that the target
11138 of a direct call is in the range allowed by the call instruction. This
11139 translation typically occurs for calls to functions in other source
11140 files. Specifically, the assembler translates a direct @code{CALL}
11141 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
11142 The default is @option{-mno-longcalls}. This option should be used in
11143 programs where the call target can potentially be out of range. This
11144 option is implemented in the assembler, not the compiler, so the
11145 assembly code generated by GCC will still show direct call
11146 instructions---look at the disassembled object code to see the actual
11147 instructions. Note that the assembler will use an indirect call for
11148 every cross-file call, not just those that really will be out of range.
11151 @node Code Gen Options
11152 @section Options for Code Generation Conventions
11153 @cindex code generation conventions
11154 @cindex options, code generation
11155 @cindex run-time options
11157 These machine-independent options control the interface conventions
11158 used in code generation.
11160 Most of them have both positive and negative forms; the negative form
11161 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
11162 one of the forms is listed---the one which is not the default. You
11163 can figure out the other form by either removing @samp{no-} or adding
11167 @item -fbounds-check
11168 @opindex fbounds-check
11169 For front-ends that support it, generate additional code to check that
11170 indices used to access arrays are within the declared range. This is
11171 currently only supported by the Java and Fortran 77 front-ends, where
11172 this option defaults to true and false respectively.
11176 This option generates traps for signed overflow on addition, subtraction,
11177 multiplication operations.
11181 This option instructs the compiler to assume that signed arithmetic
11182 overflow of addition, subtraction and multiplication wraps around
11183 using twos-complement representation. This flag enables some optimizations
11184 and disables other. This option is enabled by default for the Java
11185 front-end, as required by the Java language specification.
11188 @opindex fexceptions
11189 Enable exception handling. Generates extra code needed to propagate
11190 exceptions. For some targets, this implies GCC will generate frame
11191 unwind information for all functions, which can produce significant data
11192 size overhead, although it does not affect execution. If you do not
11193 specify this option, GCC will enable it by default for languages like
11194 C++ which normally require exception handling, and disable it for
11195 languages like C that do not normally require it. However, you may need
11196 to enable this option when compiling C code that needs to interoperate
11197 properly with exception handlers written in C++. You may also wish to
11198 disable this option if you are compiling older C++ programs that don't
11199 use exception handling.
11201 @item -fnon-call-exceptions
11202 @opindex fnon-call-exceptions
11203 Generate code that allows trapping instructions to throw exceptions.
11204 Note that this requires platform-specific runtime support that does
11205 not exist everywhere. Moreover, it only allows @emph{trapping}
11206 instructions to throw exceptions, i.e.@: memory references or floating
11207 point instructions. It does not allow exceptions to be thrown from
11208 arbitrary signal handlers such as @code{SIGALRM}.
11210 @item -funwind-tables
11211 @opindex funwind-tables
11212 Similar to @option{-fexceptions}, except that it will just generate any needed
11213 static data, but will not affect the generated code in any other way.
11214 You will normally not enable this option; instead, a language processor
11215 that needs this handling would enable it on your behalf.
11217 @item -fasynchronous-unwind-tables
11218 @opindex funwind-tables
11219 Generate unwind table in dwarf2 format, if supported by target machine. The
11220 table is exact at each instruction boundary, so it can be used for stack
11221 unwinding from asynchronous events (such as debugger or garbage collector).
11223 @item -fpcc-struct-return
11224 @opindex fpcc-struct-return
11225 Return ``short'' @code{struct} and @code{union} values in memory like
11226 longer ones, rather than in registers. This convention is less
11227 efficient, but it has the advantage of allowing intercallability between
11228 GCC-compiled files and files compiled with other compilers, particularly
11229 the Portable C Compiler (pcc).
11231 The precise convention for returning structures in memory depends
11232 on the target configuration macros.
11234 Short structures and unions are those whose size and alignment match
11235 that of some integer type.
11237 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
11238 switch is not binary compatible with code compiled with the
11239 @option{-freg-struct-return} switch.
11240 Use it to conform to a non-default application binary interface.
11242 @item -freg-struct-return
11243 @opindex freg-struct-return
11244 Return @code{struct} and @code{union} values in registers when possible.
11245 This is more efficient for small structures than
11246 @option{-fpcc-struct-return}.
11248 If you specify neither @option{-fpcc-struct-return} nor
11249 @option{-freg-struct-return}, GCC defaults to whichever convention is
11250 standard for the target. If there is no standard convention, GCC
11251 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
11252 the principal compiler. In those cases, we can choose the standard, and
11253 we chose the more efficient register return alternative.
11255 @strong{Warning:} code compiled with the @option{-freg-struct-return}
11256 switch is not binary compatible with code compiled with the
11257 @option{-fpcc-struct-return} switch.
11258 Use it to conform to a non-default application binary interface.
11260 @item -fshort-enums
11261 @opindex fshort-enums
11262 Allocate to an @code{enum} type only as many bytes as it needs for the
11263 declared range of possible values. Specifically, the @code{enum} type
11264 will be equivalent to the smallest integer type which has enough room.
11266 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
11267 code that is not binary compatible with code generated without that switch.
11268 Use it to conform to a non-default application binary interface.
11270 @item -fshort-double
11271 @opindex fshort-double
11272 Use the same size for @code{double} as for @code{float}.
11274 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
11275 code that is not binary compatible with code generated without that switch.
11276 Use it to conform to a non-default application binary interface.
11278 @item -fshort-wchar
11279 @opindex fshort-wchar
11280 Override the underlying type for @samp{wchar_t} to be @samp{short
11281 unsigned int} instead of the default for the target. This option is
11282 useful for building programs to run under WINE@.
11284 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
11285 code that is not binary compatible with code generated without that switch.
11286 Use it to conform to a non-default application binary interface.
11288 @item -fshared-data
11289 @opindex fshared-data
11290 Requests that the data and non-@code{const} variables of this
11291 compilation be shared data rather than private data. The distinction
11292 makes sense only on certain operating systems, where shared data is
11293 shared between processes running the same program, while private data
11294 exists in one copy per process.
11297 @opindex fno-common
11298 In C, allocate even uninitialized global variables in the data section of the
11299 object file, rather than generating them as common blocks. This has the
11300 effect that if the same variable is declared (without @code{extern}) in
11301 two different compilations, you will get an error when you link them.
11302 The only reason this might be useful is if you wish to verify that the
11303 program will work on other systems which always work this way.
11307 Ignore the @samp{#ident} directive.
11309 @item -fno-gnu-linker
11310 @opindex fno-gnu-linker
11311 Do not output global initializations (such as C++ constructors and
11312 destructors) in the form used by the GNU linker (on systems where the GNU
11313 linker is the standard method of handling them). Use this option when
11314 you want to use a non-GNU linker, which also requires using the
11315 @command{collect2} program to make sure the system linker includes
11316 constructors and destructors. (@command{collect2} is included in the GCC
11317 distribution.) For systems which @emph{must} use @command{collect2}, the
11318 compiler driver @command{gcc} is configured to do this automatically.
11320 @item -finhibit-size-directive
11321 @opindex finhibit-size-directive
11322 Don't output a @code{.size} assembler directive, or anything else that
11323 would cause trouble if the function is split in the middle, and the
11324 two halves are placed at locations far apart in memory. This option is
11325 used when compiling @file{crtstuff.c}; you should not need to use it
11328 @item -fverbose-asm
11329 @opindex fverbose-asm
11330 Put extra commentary information in the generated assembly code to
11331 make it more readable. This option is generally only of use to those
11332 who actually need to read the generated assembly code (perhaps while
11333 debugging the compiler itself).
11335 @option{-fno-verbose-asm}, the default, causes the
11336 extra information to be omitted and is useful when comparing two assembler
11341 @cindex global offset table
11343 Generate position-independent code (PIC) suitable for use in a shared
11344 library, if supported for the target machine. Such code accesses all
11345 constant addresses through a global offset table (GOT)@. The dynamic
11346 loader resolves the GOT entries when the program starts (the dynamic
11347 loader is not part of GCC; it is part of the operating system). If
11348 the GOT size for the linked executable exceeds a machine-specific
11349 maximum size, you get an error message from the linker indicating that
11350 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
11351 instead. (These maximums are 16k on the m88k, 8k on the SPARC, and 32k
11352 on the m68k and RS/6000. The 386 has no such limit.)
11354 Position-independent code requires special support, and therefore works
11355 only on certain machines. For the 386, GCC supports PIC for System V
11356 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
11357 position-independent.
11361 If supported for the target machine, emit position-independent code,
11362 suitable for dynamic linking and avoiding any limit on the size of the
11363 global offset table. This option makes a difference on the m68k, m88k,
11366 Position-independent code requires special support, and therefore works
11367 only on certain machines.
11373 These options are similar to @option{-fpic} and @option{-fPIC}, but
11374 generated position independent code can be only linked into executables.
11375 Usually these options are used when @option{-pie} GCC option will be
11376 used during linking.
11378 @item -ffixed-@var{reg}
11380 Treat the register named @var{reg} as a fixed register; generated code
11381 should never refer to it (except perhaps as a stack pointer, frame
11382 pointer or in some other fixed role).
11384 @var{reg} must be the name of a register. The register names accepted
11385 are machine-specific and are defined in the @code{REGISTER_NAMES}
11386 macro in the machine description macro file.
11388 This flag does not have a negative form, because it specifies a
11391 @item -fcall-used-@var{reg}
11392 @opindex fcall-used
11393 Treat the register named @var{reg} as an allocable register that is
11394 clobbered by function calls. It may be allocated for temporaries or
11395 variables that do not live across a call. Functions compiled this way
11396 will not save and restore the register @var{reg}.
11398 It is an error to used this flag with the frame pointer or stack pointer.
11399 Use of this flag for other registers that have fixed pervasive roles in
11400 the machine's execution model will produce disastrous results.
11402 This flag does not have a negative form, because it specifies a
11405 @item -fcall-saved-@var{reg}
11406 @opindex fcall-saved
11407 Treat the register named @var{reg} as an allocable register saved by
11408 functions. It may be allocated even for temporaries or variables that
11409 live across a call. Functions compiled this way will save and restore
11410 the register @var{reg} if they use it.
11412 It is an error to used this flag with the frame pointer or stack pointer.
11413 Use of this flag for other registers that have fixed pervasive roles in
11414 the machine's execution model will produce disastrous results.
11416 A different sort of disaster will result from the use of this flag for
11417 a register in which function values may be returned.
11419 This flag does not have a negative form, because it specifies a
11422 @item -fpack-struct
11423 @opindex fpack-struct
11424 Pack all structure members together without holes.
11426 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
11427 code that is not binary compatible with code generated without that switch.
11428 Additionally, it makes the code suboptimal.
11429 Use it to conform to a non-default application binary interface.
11431 @item -finstrument-functions
11432 @opindex finstrument-functions
11433 Generate instrumentation calls for entry and exit to functions. Just
11434 after function entry and just before function exit, the following
11435 profiling functions will be called with the address of the current
11436 function and its call site. (On some platforms,
11437 @code{__builtin_return_address} does not work beyond the current
11438 function, so the call site information may not be available to the
11439 profiling functions otherwise.)
11442 void __cyg_profile_func_enter (void *this_fn,
11444 void __cyg_profile_func_exit (void *this_fn,
11448 The first argument is the address of the start of the current function,
11449 which may be looked up exactly in the symbol table.
11451 This instrumentation is also done for functions expanded inline in other
11452 functions. The profiling calls will indicate where, conceptually, the
11453 inline function is entered and exited. This means that addressable
11454 versions of such functions must be available. If all your uses of a
11455 function are expanded inline, this may mean an additional expansion of
11456 code size. If you use @samp{extern inline} in your C code, an
11457 addressable version of such functions must be provided. (This is
11458 normally the case anyways, but if you get lucky and the optimizer always
11459 expands the functions inline, you might have gotten away without
11460 providing static copies.)
11462 A function may be given the attribute @code{no_instrument_function}, in
11463 which case this instrumentation will not be done. This can be used, for
11464 example, for the profiling functions listed above, high-priority
11465 interrupt routines, and any functions from which the profiling functions
11466 cannot safely be called (perhaps signal handlers, if the profiling
11467 routines generate output or allocate memory).
11469 @item -fstack-check
11470 @opindex fstack-check
11471 Generate code to verify that you do not go beyond the boundary of the
11472 stack. You should specify this flag if you are running in an
11473 environment with multiple threads, but only rarely need to specify it in
11474 a single-threaded environment since stack overflow is automatically
11475 detected on nearly all systems if there is only one stack.
11477 Note that this switch does not actually cause checking to be done; the
11478 operating system must do that. The switch causes generation of code
11479 to ensure that the operating system sees the stack being extended.
11481 @item -fstack-limit-register=@var{reg}
11482 @itemx -fstack-limit-symbol=@var{sym}
11483 @itemx -fno-stack-limit
11484 @opindex fstack-limit-register
11485 @opindex fstack-limit-symbol
11486 @opindex fno-stack-limit
11487 Generate code to ensure that the stack does not grow beyond a certain value,
11488 either the value of a register or the address of a symbol. If the stack
11489 would grow beyond the value, a signal is raised. For most targets,
11490 the signal is raised before the stack overruns the boundary, so
11491 it is possible to catch the signal without taking special precautions.
11493 For instance, if the stack starts at absolute address @samp{0x80000000}
11494 and grows downwards, you can use the flags
11495 @option{-fstack-limit-symbol=__stack_limit} and
11496 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
11497 of 128KB@. Note that this may only work with the GNU linker.
11499 @cindex aliasing of parameters
11500 @cindex parameters, aliased
11501 @item -fargument-alias
11502 @itemx -fargument-noalias
11503 @itemx -fargument-noalias-global
11504 @opindex fargument-alias
11505 @opindex fargument-noalias
11506 @opindex fargument-noalias-global
11507 Specify the possible relationships among parameters and between
11508 parameters and global data.
11510 @option{-fargument-alias} specifies that arguments (parameters) may
11511 alias each other and may alias global storage.@*
11512 @option{-fargument-noalias} specifies that arguments do not alias
11513 each other, but may alias global storage.@*
11514 @option{-fargument-noalias-global} specifies that arguments do not
11515 alias each other and do not alias global storage.
11517 Each language will automatically use whatever option is required by
11518 the language standard. You should not need to use these options yourself.
11520 @item -fleading-underscore
11521 @opindex fleading-underscore
11522 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
11523 change the way C symbols are represented in the object file. One use
11524 is to help link with legacy assembly code.
11526 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
11527 generate code that is not binary compatible with code generated without that
11528 switch. Use it to conform to a non-default application binary interface.
11529 Not all targets provide complete support for this switch.
11531 @item -ftls-model=@var{model}
11532 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
11533 The @var{model} argument should be one of @code{global-dynamic},
11534 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
11536 The default without @option{-fpic} is @code{initial-exec}; with
11537 @option{-fpic} the default is @code{global-dynamic}.
11542 @node Environment Variables
11543 @section Environment Variables Affecting GCC
11544 @cindex environment variables
11546 @c man begin ENVIRONMENT
11547 This section describes several environment variables that affect how GCC
11548 operates. Some of them work by specifying directories or prefixes to use
11549 when searching for various kinds of files. Some are used to specify other
11550 aspects of the compilation environment.
11552 Note that you can also specify places to search using options such as
11553 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
11554 take precedence over places specified using environment variables, which
11555 in turn take precedence over those specified by the configuration of GCC@.
11556 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
11557 GNU Compiler Collection (GCC) Internals}.
11562 @c @itemx LC_COLLATE
11564 @c @itemx LC_MONETARY
11565 @c @itemx LC_NUMERIC
11570 @c @findex LC_COLLATE
11571 @findex LC_MESSAGES
11572 @c @findex LC_MONETARY
11573 @c @findex LC_NUMERIC
11577 These environment variables control the way that GCC uses
11578 localization information that allow GCC to work with different
11579 national conventions. GCC inspects the locale categories
11580 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
11581 so. These locale categories can be set to any value supported by your
11582 installation. A typical value is @samp{en_UK} for English in the United
11585 The @env{LC_CTYPE} environment variable specifies character
11586 classification. GCC uses it to determine the character boundaries in
11587 a string; this is needed for some multibyte encodings that contain quote
11588 and escape characters that would otherwise be interpreted as a string
11591 The @env{LC_MESSAGES} environment variable specifies the language to
11592 use in diagnostic messages.
11594 If the @env{LC_ALL} environment variable is set, it overrides the value
11595 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
11596 and @env{LC_MESSAGES} default to the value of the @env{LANG}
11597 environment variable. If none of these variables are set, GCC
11598 defaults to traditional C English behavior.
11602 If @env{TMPDIR} is set, it specifies the directory to use for temporary
11603 files. GCC uses temporary files to hold the output of one stage of
11604 compilation which is to be used as input to the next stage: for example,
11605 the output of the preprocessor, which is the input to the compiler
11608 @item GCC_EXEC_PREFIX
11609 @findex GCC_EXEC_PREFIX
11610 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
11611 names of the subprograms executed by the compiler. No slash is added
11612 when this prefix is combined with the name of a subprogram, but you can
11613 specify a prefix that ends with a slash if you wish.
11615 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
11616 an appropriate prefix to use based on the pathname it was invoked with.
11618 If GCC cannot find the subprogram using the specified prefix, it
11619 tries looking in the usual places for the subprogram.
11621 The default value of @env{GCC_EXEC_PREFIX} is
11622 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
11623 of @code{prefix} when you ran the @file{configure} script.
11625 Other prefixes specified with @option{-B} take precedence over this prefix.
11627 This prefix is also used for finding files such as @file{crt0.o} that are
11630 In addition, the prefix is used in an unusual way in finding the
11631 directories to search for header files. For each of the standard
11632 directories whose name normally begins with @samp{/usr/local/lib/gcc}
11633 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
11634 replacing that beginning with the specified prefix to produce an
11635 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
11636 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
11637 These alternate directories are searched first; the standard directories
11640 @item COMPILER_PATH
11641 @findex COMPILER_PATH
11642 The value of @env{COMPILER_PATH} is a colon-separated list of
11643 directories, much like @env{PATH}. GCC tries the directories thus
11644 specified when searching for subprograms, if it can't find the
11645 subprograms using @env{GCC_EXEC_PREFIX}.
11648 @findex LIBRARY_PATH
11649 The value of @env{LIBRARY_PATH} is a colon-separated list of
11650 directories, much like @env{PATH}. When configured as a native compiler,
11651 GCC tries the directories thus specified when searching for special
11652 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
11653 using GCC also uses these directories when searching for ordinary
11654 libraries for the @option{-l} option (but directories specified with
11655 @option{-L} come first).
11659 @cindex locale definition
11660 This variable is used to pass locale information to the compiler. One way in
11661 which this information is used is to determine the character set to be used
11662 when character literals, string literals and comments are parsed in C and C++.
11663 When the compiler is configured to allow multibyte characters,
11664 the following values for @env{LANG} are recognized:
11668 Recognize JIS characters.
11670 Recognize SJIS characters.
11672 Recognize EUCJP characters.
11675 If @env{LANG} is not defined, or if it has some other value, then the
11676 compiler will use mblen and mbtowc as defined by the default locale to
11677 recognize and translate multibyte characters.
11681 Some additional environments variables affect the behavior of the
11684 @include cppenv.texi
11688 @node Precompiled Headers
11689 @section Using Precompiled Headers
11690 @cindex precompiled headers
11691 @cindex speed of compilation
11693 Often large projects have many header files that are included in every
11694 source file. The time the compiler takes to process these header files
11695 over and over again can account for nearly all of the time required to
11696 build the project. To make builds faster, GCC allows users to
11697 `precompile' a header file; then, if builds can use the precompiled
11698 header file they will be much faster.
11700 To create a precompiled header file, simply compile it as you would any
11701 other file, if necessary using the @option{-x} option to make the driver
11702 treat it as a C or C++ header file. You will probably want to use a
11703 tool like @command{make} to keep the precompiled header up-to-date when
11704 the headers it contains change.
11706 A precompiled header file will be searched for when @code{#include} is
11707 seen in the compilation. As it searches for the included file
11708 (@pxref{Search Path,,Search Path,cpp.info,The C Preprocessor}) the
11709 compiler looks for a precompiled header in each directory just before it
11710 looks for the include file in that directory. The name searched for is
11711 the name specified in the @code{#include} with @samp{.gch} appended. If
11712 the precompiled header file can't be used, it is ignored.
11714 For instance, if you have @code{#include "all.h"}, and you have
11715 @file{all.h.gch} in the same directory as @file{all.h}, then the
11716 precompiled header file will be used if possible, and the original
11717 header will be used otherwise.
11719 Alternatively, you might decide to put the precompiled header file in a
11720 directory and use @option{-I} to ensure that directory is searched
11721 before (or instead of) the directory containing the original header.
11722 Then, if you want to check that the precompiled header file is always
11723 used, you can put a file of the same name as the original header in this
11724 directory containing an @code{#error} command.
11726 This also works with @option{-include}. So yet another way to use
11727 precompiled headers, good for projects not designed with precompiled
11728 header files in mind, is to simply take most of the header files used by
11729 a project, include them from another header file, precompile that header
11730 file, and @option{-include} the precompiled header. If the header files
11731 have guards against multiple inclusion, they will be skipped because
11732 they've already been included (in the precompiled header).
11734 If you need to precompile the same header file for different
11735 languages, targets, or compiler options, you can instead make a
11736 @emph{directory} named like @file{all.h.gch}, and put each precompiled
11737 header in the directory. (It doesn't matter what you call the files
11738 in the directory, every precompiled header in the directory will be
11739 considered.) The first precompiled header encountered in the
11740 directory that is valid for this compilation will be used; they're
11741 searched in no particular order.
11743 There are many other possibilities, limited only by your imagination,
11744 good sense, and the constraints of your build system.
11746 A precompiled header file can be used only when these conditions apply:
11750 Only one precompiled header can be used in a particular compilation.
11752 A precompiled header can't be used once the first C token is seen. You
11753 can have preprocessor directives before a precompiled header; you can
11754 even include a precompiled header from inside another header, so long as
11755 there are no C tokens before the @code{#include}.
11757 The precompiled header file must be produced for the same language as
11758 the current compilation. You can't use a C precompiled header for a C++
11761 The precompiled header file must be produced by the same compiler
11762 version and configuration as the current compilation is using.
11763 The easiest way to guarantee this is to use the same compiler binary
11764 for creating and using precompiled headers.
11766 Any macros defined before the precompiled header (including with
11767 @option{-D}) must either be defined in the same way as when the
11768 precompiled header was generated, or must not affect the precompiled
11769 header, which usually means that the they don't appear in the
11770 precompiled header at all.
11772 Certain command-line options must be defined in the same way as when the
11773 precompiled header was generated. At present, it's not clear which
11774 options are safe to change and which are not; the safest choice is to
11775 use exactly the same options when generating and using the precompiled
11779 For all of these but the last, the compiler will automatically ignore
11780 the precompiled header if the conditions aren't met. For the last item,
11781 some option changes will cause the precompiled header to be rejected,
11782 but not all incompatible option combinations have yet been found. If
11783 you find a new incompatible combination, please consider filing a bug
11784 report, see @ref{Bugs}.
11786 @node Running Protoize
11787 @section Running Protoize
11789 The program @code{protoize} is an optional part of GCC@. You can use
11790 it to add prototypes to a program, thus converting the program to ISO
11791 C in one respect. The companion program @code{unprotoize} does the
11792 reverse: it removes argument types from any prototypes that are found.
11794 When you run these programs, you must specify a set of source files as
11795 command line arguments. The conversion programs start out by compiling
11796 these files to see what functions they define. The information gathered
11797 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
11799 After scanning comes actual conversion. The specified files are all
11800 eligible to be converted; any files they include (whether sources or
11801 just headers) are eligible as well.
11803 But not all the eligible files are converted. By default,
11804 @code{protoize} and @code{unprotoize} convert only source and header
11805 files in the current directory. You can specify additional directories
11806 whose files should be converted with the @option{-d @var{directory}}
11807 option. You can also specify particular files to exclude with the
11808 @option{-x @var{file}} option. A file is converted if it is eligible, its
11809 directory name matches one of the specified directory names, and its
11810 name within the directory has not been excluded.
11812 Basic conversion with @code{protoize} consists of rewriting most
11813 function definitions and function declarations to specify the types of
11814 the arguments. The only ones not rewritten are those for varargs
11817 @code{protoize} optionally inserts prototype declarations at the
11818 beginning of the source file, to make them available for any calls that
11819 precede the function's definition. Or it can insert prototype
11820 declarations with block scope in the blocks where undeclared functions
11823 Basic conversion with @code{unprotoize} consists of rewriting most
11824 function declarations to remove any argument types, and rewriting
11825 function definitions to the old-style pre-ISO form.
11827 Both conversion programs print a warning for any function declaration or
11828 definition that they can't convert. You can suppress these warnings
11831 The output from @code{protoize} or @code{unprotoize} replaces the
11832 original source file. The original file is renamed to a name ending
11833 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
11834 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
11835 for DOS) file already exists, then the source file is simply discarded.
11837 @code{protoize} and @code{unprotoize} both depend on GCC itself to
11838 scan the program and collect information about the functions it uses.
11839 So neither of these programs will work until GCC is installed.
11841 Here is a table of the options you can use with @code{protoize} and
11842 @code{unprotoize}. Each option works with both programs unless
11846 @item -B @var{directory}
11847 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
11848 usual directory (normally @file{/usr/local/lib}). This file contains
11849 prototype information about standard system functions. This option
11850 applies only to @code{protoize}.
11852 @item -c @var{compilation-options}
11853 Use @var{compilation-options} as the options when running @command{gcc} to
11854 produce the @samp{.X} files. The special option @option{-aux-info} is
11855 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
11857 Note that the compilation options must be given as a single argument to
11858 @code{protoize} or @code{unprotoize}. If you want to specify several
11859 @command{gcc} options, you must quote the entire set of compilation options
11860 to make them a single word in the shell.
11862 There are certain @command{gcc} arguments that you cannot use, because they
11863 would produce the wrong kind of output. These include @option{-g},
11864 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
11865 the @var{compilation-options}, they are ignored.
11868 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
11869 systems) instead of @samp{.c}. This is convenient if you are converting
11870 a C program to C++. This option applies only to @code{protoize}.
11873 Add explicit global declarations. This means inserting explicit
11874 declarations at the beginning of each source file for each function
11875 that is called in the file and was not declared. These declarations
11876 precede the first function definition that contains a call to an
11877 undeclared function. This option applies only to @code{protoize}.
11879 @item -i @var{string}
11880 Indent old-style parameter declarations with the string @var{string}.
11881 This option applies only to @code{protoize}.
11883 @code{unprotoize} converts prototyped function definitions to old-style
11884 function definitions, where the arguments are declared between the
11885 argument list and the initial @samp{@{}. By default, @code{unprotoize}
11886 uses five spaces as the indentation. If you want to indent with just
11887 one space instead, use @option{-i " "}.
11890 Keep the @samp{.X} files. Normally, they are deleted after conversion
11894 Add explicit local declarations. @code{protoize} with @option{-l} inserts
11895 a prototype declaration for each function in each block which calls the
11896 function without any declaration. This option applies only to
11900 Make no real changes. This mode just prints information about the conversions
11901 that would have been done without @option{-n}.
11904 Make no @samp{.save} files. The original files are simply deleted.
11905 Use this option with caution.
11907 @item -p @var{program}
11908 Use the program @var{program} as the compiler. Normally, the name
11909 @file{gcc} is used.
11912 Work quietly. Most warnings are suppressed.
11915 Print the version number, just like @option{-v} for @command{gcc}.
11918 If you need special compiler options to compile one of your program's
11919 source files, then you should generate that file's @samp{.X} file
11920 specially, by running @command{gcc} on that source file with the
11921 appropriate options and the option @option{-aux-info}. Then run
11922 @code{protoize} on the entire set of files. @code{protoize} will use
11923 the existing @samp{.X} file because it is newer than the source file.
11927 gcc -Dfoo=bar file1.c -aux-info file1.X
11932 You need to include the special files along with the rest in the
11933 @code{protoize} command, even though their @samp{.X} files already
11934 exist, because otherwise they won't get converted.
11936 @xref{Protoize Caveats}, for more information on how to use
11937 @code{protoize} successfully.