1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004 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, 2004 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}
171 @item C++ Language Options
172 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
173 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
174 -fconserve-space -fno-const-strings @gol
175 -fno-elide-constructors @gol
176 -fno-enforce-eh-specs @gol
177 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
178 -fno-implicit-templates @gol
179 -fno-implicit-inline-templates @gol
180 -fno-implement-inlines -fms-extensions @gol
181 -fno-nonansi-builtins -fno-operator-names @gol
182 -fno-optional-diags -fpermissive @gol
183 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
184 -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
185 -fno-default-inline -Wabi -Wctor-dtor-privacy @gol
186 -Wnon-virtual-dtor -Wreorder @gol
187 -Weffc++ -Wno-deprecated @gol
188 -Wno-non-template-friend -Wold-style-cast @gol
189 -Woverloaded-virtual -Wno-pmf-conversions @gol
190 -Wsign-promo -Wsynth}
192 @item Objective-C Language Options
193 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
195 -fconstant-string-class=@var{class-name} @gol
196 -fgnu-runtime -fnext-runtime @gol
197 -fno-nil-receivers @gol
198 -fobjc-exceptions @gol
199 -freplace-objc-classes @gol
202 -Wno-protocol -Wselector -Wundeclared-selector}
204 @item Language Independent Options
205 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
206 @gccoptlist{-fmessage-length=@var{n} @gol
207 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
209 @item Warning Options
210 @xref{Warning Options,,Options to Request or Suppress Warnings}.
211 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
212 -w -Wextra -Wall -Waggregate-return @gol
213 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
214 -Wconversion -Wno-deprecated-declarations @gol
215 -Wdisabled-optimization -Wno-div-by-zero -Wendif-labels @gol
216 -Werror -Werror-implicit-function-declaration @gol
217 -Wfloat-equal -Wformat -Wformat=2 @gol
218 -Wno-format-extra-args -Wformat-nonliteral @gol
219 -Wformat-security -Wformat-y2k @gol
220 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
221 -Wimport -Wno-import -Winit-self -Winline @gol
222 -Wno-invalid-offsetof -Winvalid-pch @gol
223 -Wlarger-than-@var{len} -Wlong-long @gol
224 -Wmain -Wmissing-braces @gol
225 -Wmissing-format-attribute -Wmissing-noreturn @gol
226 -Wno-multichar -Wnonnull -Wpacked -Wpadded @gol
227 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
228 -Wreturn-type -Wsequence-point -Wshadow @gol
229 -Wsign-compare -Wstrict-aliasing -Wstrict-aliasing=2 @gol
230 -Wswitch -Wswitch-default -Wswitch-enum @gol
231 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
232 -Wunknown-pragmas -Wunreachable-code @gol
233 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
234 -Wunused-value -Wunused-variable -Wwrite-strings @gol
237 @item C-only Warning Options
238 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
239 -Wmissing-prototypes -Wnested-externs -Wold-style-definition @gol
240 -Wstrict-prototypes -Wtraditional @gol
241 -Wdeclaration-after-statement}
243 @item Debugging Options
244 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
245 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
246 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
247 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
248 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
249 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
250 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
251 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
252 -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs @gol
253 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
254 -ftest-coverage -ftime-report -fvar-tracking @gol
255 -g -g@var{level} -gcoff -gdwarf-2 @gol
256 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
257 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
258 -print-multi-directory -print-multi-lib @gol
259 -print-prog-name=@var{program} -print-search-dirs -Q @gol
262 @item Optimization Options
263 @xref{Optimize Options,,Options that Control Optimization}.
264 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
265 -falign-labels=@var{n} -falign-loops=@var{n} @gol
266 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
267 -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol
268 -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 -fprofile-generate -fprofile-use @gol
286 -freduce-all-givs -fregmove -frename-registers @gol
287 -freorder-blocks -freorder-functions @gol
288 -frerun-cse-after-loop -frerun-loop-opt @gol
289 -frounding-math -fschedule-insns -fschedule-insns2 @gol
290 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
291 -fsched-spec-load-dangerous @gol
292 -fsched-stalled-insns=@var{n} -sched-stalled-insns-dep=@var{n} @gol
293 -fsched2-use-superblocks @gol
294 -fsched2-use-traces -fsignaling-nans @gol
295 -fsingle-precision-constant @gol
296 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
297 -funroll-all-loops -funroll-loops -fpeel-loops @gol
298 -funswitch-loops -fold-unroll-loops -fold-unroll-all-loops @gol
299 --param @var{name}=@var{value}
300 -O -O0 -O1 -O2 -O3 -Os}
302 @item Preprocessor Options
303 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
304 @gccoptlist{-A@var{question}=@var{answer} @gol
305 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
306 -C -dD -dI -dM -dN @gol
307 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
308 -idirafter @var{dir} @gol
309 -include @var{file} -imacros @var{file} @gol
310 -iprefix @var{file} -iwithprefix @var{dir} @gol
311 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
312 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
313 -P -fworking-directory -remap @gol
314 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
315 -Xpreprocessor @var{option}}
317 @item Assembler Option
318 @xref{Assembler Options,,Passing Options to the Assembler}.
319 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
322 @xref{Link Options,,Options for Linking}.
323 @gccoptlist{@var{object-file-name} -l@var{library} @gol
324 -nostartfiles -nodefaultlibs -nostdlib -pie @gol
325 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
326 -Wl,@var{option} -Xlinker @var{option} @gol
329 @item Directory Options
330 @xref{Directory Options,,Options for Directory Search}.
331 @gccoptlist{-B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
334 @c I wrote this xref this way to avoid overfull hbox. -- rms
335 @xref{Target Options}.
336 @gccoptlist{-V @var{version} -b @var{machine}}
338 @item Machine Dependent Options
339 @xref{Submodel Options,,Hardware Models and Configurations}.
341 @emph{M680x0 Options}
342 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
343 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
344 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
345 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
346 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
348 @emph{M68hc1x Options}
349 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
350 -mauto-incdec -minmax -mlong-calls -mshort @gol
351 -msoft-reg-count=@var{count}}
354 @gccoptlist{-mg -mgnu -munix}
357 @gccoptlist{-mcpu=@var{cpu-type} @gol
358 -mtune=@var{cpu-type} @gol
359 -mcmodel=@var{code-model} @gol
360 -m32 -m64 -mapp-regs -mno-app-regs @gol
361 -mfaster-structs -mno-faster-structs @gol
362 -mfpu -mno-fpu -mhard-float -msoft-float @gol
363 -mhard-quad-float -msoft-quad-float @gol
364 -mimpure-text -mno-impure-text -mlittle-endian @gol
365 -mstack-bias -mno-stack-bias @gol
366 -munaligned-doubles -mno-unaligned-doubles @gol
367 -mv8plus -mno-v8plus -mvis -mno-vis}
370 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
371 -mabi=@var{name} @gol
372 -mapcs-26 -mapcs-32 @gol
373 -mapcs-stack-check -mno-apcs-stack-check @gol
374 -mapcs-float -mno-apcs-float @gol
375 -mapcs-reentrant -mno-apcs-reentrant @gol
376 -msched-prolog -mno-sched-prolog @gol
377 -mlittle-endian -mbig-endian -mwords-little-endian @gol
378 -malignment-traps -mno-alignment-traps @gol
379 -mfloat-abi=@var{name} soft-float -mhard-float -mfpe @gol
380 -mthumb-interwork -mno-thumb-interwork @gol
381 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
382 -mstructure-size-boundary=@var{n} @gol
383 -mabort-on-noreturn @gol
384 -mlong-calls -mno-long-calls @gol
385 -msingle-pic-base -mno-single-pic-base @gol
386 -mpic-register=@var{reg} @gol
387 -mnop-fun-dllimport @gol
388 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
389 -mpoke-function-name @gol
391 -mtpcs-frame -mtpcs-leaf-frame @gol
392 -mcaller-super-interworking -mcallee-super-interworking}
394 @emph{MN10300 Options}
395 @gccoptlist{-mmult-bug -mno-mult-bug @gol
396 -mam33 -mno-am33 @gol
397 -mam33-2 -mno-am33-2 @gol
400 @emph{M32R/D Options}
401 @gccoptlist{-m32r2 -m32rx -m32r @gol
403 -malign-loops -mno-align-loops @gol
404 -missue-rate=@var{number} @gol
405 -mbranch-cost=@var{number} @gol
406 -mmodel=@var{code-size-model-type} @gol
407 -msdata=@var{sdata-type} @gol
408 -mno-flush-func -mflush-func=@var{name} @gol
409 -mno-flush-trap -mflush-trap=@var{number} @gol
412 @emph{RS/6000 and PowerPC Options}
413 @gccoptlist{-mcpu=@var{cpu-type} @gol
414 -mtune=@var{cpu-type} @gol
415 -mpower -mno-power -mpower2 -mno-power2 @gol
416 -mpowerpc -mpowerpc64 -mno-powerpc @gol
417 -maltivec -mno-altivec @gol
418 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
419 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
420 -mnew-mnemonics -mold-mnemonics @gol
421 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
422 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
423 -malign-power -malign-natural @gol
424 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
425 -mstring -mno-string -mupdate -mno-update @gol
426 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
427 -mstrict-align -mno-strict-align -mrelocatable @gol
428 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
429 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
430 -mdynamic-no-pic @gol
431 -mprioritize-restricted-insns=@var{priority} @gol
432 -msched-costly-dep=@var{dependence_type} @gol
433 -minsert-sched-nops=@var{scheme} @gol
434 -mcall-sysv -mcall-netbsd @gol
435 -maix-struct-return -msvr4-struct-return @gol
436 -mabi=altivec -mabi=no-altivec @gol
437 -mabi=spe -mabi=no-spe @gol
438 -misel=yes -misel=no @gol
439 -mspe=yes -mspe=no @gol
440 -mfloat-gprs=yes -mfloat-gprs=no @gol
441 -mprototype -mno-prototype @gol
442 -msim -mmvme -mads -myellowknife -memb -msdata @gol
443 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
445 @emph{Darwin Options}
446 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
447 -arch_only -bind_at_load -bundle -bundle_loader @gol
448 -client_name -compatibility_version -current_version @gol
449 -dependency-file -dylib_file -dylinker_install_name @gol
450 -dynamic -dynamiclib -exported_symbols_list @gol
451 -filelist -flat_namespace -force_cpusubtype_ALL @gol
452 -force_flat_namespace -headerpad_max_install_names @gol
453 -image_base -init -install_name -keep_private_externs @gol
454 -multi_module -multiply_defined -multiply_defined_unused @gol
455 -noall_load -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
456 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
457 -private_bundle -read_only_relocs -sectalign @gol
458 -sectobjectsymbols -whyload -seg1addr @gol
459 -sectcreate -sectobjectsymbols -sectorder @gol
460 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
461 -segprot -segs_read_only_addr -segs_read_write_addr @gol
462 -single_module -static -sub_library -sub_umbrella @gol
463 -twolevel_namespace -umbrella -undefined @gol
464 -unexported_symbols_list -weak_reference_mismatches @gol
468 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
469 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
470 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
471 -mxgot -mno-xgot -membedded-pic -mno-embedded-pic @gol
472 -mgp32 -mgp64 -mfp32 -mfp64 -mhard-float -msoft-float @gol
473 -msingle-float -mdouble-float -mint64 -mlong64 -mlong32 @gol
474 -G@var{num} -membedded-data -mno-embedded-data @gol
475 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
476 -msplit-addresses -mno-split-addresses @gol
477 -mexplicit-relocs -mno-explicit-relocs @gol
478 -mrnames -mno-rnames @gol
479 -mcheck-zero-division -mno-check-zero-division @gol
480 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
481 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
482 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
483 -mfix-vr4122-bugs -mno-fix-vr4122-bugs -mfix-sb1 -mno-fix-sb1 @gol
484 -mflush-func=@var{func} -mno-flush-func @gol
485 -mbranch-likely -mno-branch-likely}
487 @emph{i386 and x86-64 Options}
488 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
489 -mfpmath=@var{unit} @gol
490 -masm=@var{dialect} -mno-fancy-math-387 @gol
491 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
492 -mno-wide-multiply -mrtd -malign-double @gol
493 -mpreferred-stack-boundary=@var{num} @gol
494 -mmmx -msse -msse2 -msse3 -m3dnow @gol
495 -mthreads -mno-align-stringops -minline-all-stringops @gol
496 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
497 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
498 -mno-red-zone -mno-tls-direct-seg-refs @gol
499 -mcmodel=@var{code-model} @gol
503 @gccoptlist{-march=@var{architecture-type} @gol
504 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
505 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
506 -mjump-in-delay -mlinker-opt -mlong-calls @gol
507 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
508 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
509 -mno-jump-in-delay -mno-long-load-store @gol
510 -mno-portable-runtime -mno-soft-float @gol
511 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
512 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
513 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
514 -nolibdld -static -threads}
516 @emph{DEC Alpha Options}
517 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
518 -mieee -mieee-with-inexact -mieee-conformant @gol
519 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
520 -mtrap-precision=@var{mode} -mbuild-constants @gol
521 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
522 -mbwx -mmax -mfix -mcix @gol
523 -mfloat-vax -mfloat-ieee @gol
524 -mexplicit-relocs -msmall-data -mlarge-data @gol
525 -msmall-text -mlarge-text @gol
526 -mmemory-latency=@var{time}}
528 @emph{DEC Alpha/VMS Options}
529 @gccoptlist{-mvms-return-codes}
531 @emph{H8/300 Options}
532 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
535 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
536 -m4-nofpu -m4-single-only -m4-single -m4 @gol
537 -m5-64media -m5-64media-nofpu @gol
538 -m5-32media -m5-32media-nofpu @gol
539 -m5-compact -m5-compact-nofpu @gol
540 -mb -ml -mdalign -mrelax @gol
541 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
542 -mieee -misize -mpadstruct -mspace @gol
543 -mprefergot -musermode}
545 @emph{System V Options}
546 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
549 @gccoptlist{-EB -EL @gol
550 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
551 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
553 @emph{TMS320C3x/C4x Options}
554 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
555 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
556 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
557 -mparallel-insns -mparallel-mpy -mpreserve-float}
560 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
561 -mprolog-function -mno-prolog-function -mspace @gol
562 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
563 -mapp-regs -mno-app-regs @gol
564 -mdisable-callt -mno-disable-callt @gol
570 @gccoptlist{-m32032 -m32332 -m32532 -m32081 -m32381 @gol
571 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
572 -mregparam -mnoregparam -msb -mnosb @gol
573 -mbitfield -mnobitfield -mhimem -mnohimem}
576 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
577 -mcall-prologues -mno-tablejump -mtiny-stack}
580 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
581 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
582 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
583 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
584 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
587 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
588 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
589 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
590 -mno-base-addresses -msingle-exit -mno-single-exit}
593 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
594 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
595 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
596 -minline-float-divide-max-throughput @gol
597 -minline-int-divide-min-latency @gol
598 -minline-int-divide-max-throughput -mno-dwarf2-asm @gol
599 -mfixed-range=@var{register-range}}
601 @emph{S/390 and zSeries Options}
602 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
603 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
604 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
605 -m64 -m31 -mdebug -mno-debug -mesa -mzarch -mfused-madd -mno-fused-madd}
608 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
609 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
610 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
611 -mstack-align -mdata-align -mconst-align @gol
612 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
613 -melf -maout -melinux -mlinux -sim -sim2 @gol
614 -mmul-bug-workaround -mno-mul-bug-workaround}
616 @emph{PDP-11 Options}
617 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
618 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
619 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
620 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
621 -mbranch-expensive -mbranch-cheap @gol
622 -msplit -mno-split -munix-asm -mdec-asm}
624 @emph{Xstormy16 Options}
627 @emph{Xtensa Options}
628 @gccoptlist{-mconst16 -mno-const16 @gol
629 -mfused-madd -mno-fused-madd @gol
630 -mtext-section-literals -mno-text-section-literals @gol
631 -mtarget-align -mno-target-align @gol
632 -mlongcalls -mno-longcalls}
635 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
636 -mhard-float -msoft-float @gol
637 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
638 -mdouble -mno-double @gol
639 -mmedia -mno-media -mmuladd -mno-muladd @gol
640 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic -mlinked-fp @gol
641 -mlibrary-pic -macc-4 -macc-8 @gol
642 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
643 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
644 -mvliw-branch -mno-vliw-branch @gol
645 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
646 -mno-nested-cond-exec -mtomcat-stats @gol
649 @item Code Generation Options
650 @xref{Code Gen Options,,Options for Code Generation Conventions}.
651 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
652 -ffixed-@var{reg} -fexceptions @gol
653 -fnon-call-exceptions -funwind-tables @gol
654 -fasynchronous-unwind-tables @gol
655 -finhibit-size-directive -finstrument-functions @gol
656 -fno-common -fno-ident @gol
657 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
658 -freg-struct-return -fshared-data -fshort-enums @gol
659 -fshort-double -fshort-wchar @gol
660 -fverbose-asm -fpack-struct -fstack-check @gol
661 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
662 -fargument-alias -fargument-noalias @gol
663 -fargument-noalias-global -fleading-underscore @gol
664 -ftls-model=@var{model} @gol
665 -ftrapv -fwrapv -fbounds-check}
669 * Overall Options:: Controlling the kind of output:
670 an executable, object files, assembler files,
671 or preprocessed source.
672 * C Dialect Options:: Controlling the variant of C language compiled.
673 * C++ Dialect Options:: Variations on C++.
674 * Objective-C Dialect Options:: Variations on Objective-C.
675 * Language Independent Options:: Controlling how diagnostics should be
677 * Warning Options:: How picky should the compiler be?
678 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
679 * Optimize Options:: How much optimization?
680 * Preprocessor Options:: Controlling header files and macro definitions.
681 Also, getting dependency information for Make.
682 * Assembler Options:: Passing options to the assembler.
683 * Link Options:: Specifying libraries and so on.
684 * Directory Options:: Where to find header files and libraries.
685 Where to find the compiler executable files.
686 * Spec Files:: How to pass switches to sub-processes.
687 * Target Options:: Running a cross-compiler, or an old version of GCC.
690 @node Overall Options
691 @section Options Controlling the Kind of Output
693 Compilation can involve up to four stages: preprocessing, compilation
694 proper, assembly and linking, always in that order. GCC is capable of
695 preprocessing and compiling several files either into several
696 assembler input files, or into one assembler input file; then each
697 assembler input file produces an object file, and linking combines all
698 the object files (those newly compiled, and those specified as input)
699 into an executable file.
701 @cindex file name suffix
702 For any given input file, the file name suffix determines what kind of
707 C source code which must be preprocessed.
710 C source code which should not be preprocessed.
713 C++ source code which should not be preprocessed.
716 Objective-C source code. Note that you must link with the library
717 @file{libobjc.a} to make an Objective-C program work.
720 Objective-C source code which should not be preprocessed.
723 C or C++ header file to be turned into a precompiled header.
727 @itemx @var{file}.cxx
728 @itemx @var{file}.cpp
729 @itemx @var{file}.CPP
730 @itemx @var{file}.c++
732 C++ source code which must be preprocessed. Note that in @samp{.cxx},
733 the last two letters must both be literally @samp{x}. Likewise,
734 @samp{.C} refers to a literal capital C@.
738 C++ header file to be turned into a precompiled header.
741 @itemx @var{file}.for
742 @itemx @var{file}.FOR
743 Fortran source code which should not be preprocessed.
746 @itemx @var{file}.fpp
747 @itemx @var{file}.FPP
748 Fortran source code which must be preprocessed (with the traditional
752 Fortran source code which must be preprocessed with a RATFOR
753 preprocessor (not included with GCC)@.
755 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
756 Using and Porting GNU Fortran}, for more details of the handling of
759 @c FIXME: Descriptions of Java file types.
766 Ada source code file which contains a library unit declaration (a
767 declaration of a package, subprogram, or generic, or a generic
768 instantiation), or a library unit renaming declaration (a package,
769 generic, or subprogram renaming declaration). Such files are also
772 @itemx @var{file}.adb
773 Ada source code file containing a library unit body (a subprogram or
774 package body). Such files are also called @dfn{bodies}.
776 @c GCC also knows about some suffixes for languages not yet included:
785 Assembler code which must be preprocessed.
788 An object file to be fed straight into linking.
789 Any file name with no recognized suffix is treated this way.
793 You can specify the input language explicitly with the @option{-x} option:
796 @item -x @var{language}
797 Specify explicitly the @var{language} for the following input files
798 (rather than letting the compiler choose a default based on the file
799 name suffix). This option applies to all following input files until
800 the next @option{-x} option. Possible values for @var{language} are:
802 c c-header cpp-output
803 c++ c++-header c++-cpp-output
804 objective-c objective-c-header objc-cpp-output
805 assembler assembler-with-cpp
807 f77 f77-cpp-input ratfor
813 Turn off any specification of a language, so that subsequent files are
814 handled according to their file name suffixes (as they are if @option{-x}
815 has not been used at all).
817 @item -pass-exit-codes
818 @opindex pass-exit-codes
819 Normally the @command{gcc} program will exit with the code of 1 if any
820 phase of the compiler returns a non-success return code. If you specify
821 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
822 numerically highest error produced by any phase that returned an error
826 If you only want some of the stages of compilation, you can use
827 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
828 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
829 @command{gcc} is to stop. Note that some combinations (for example,
830 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
835 Compile or assemble the source files, but do not link. The linking
836 stage simply is not done. The ultimate output is in the form of an
837 object file for each source file.
839 By default, the object file name for a source file is made by replacing
840 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
842 Unrecognized input files, not requiring compilation or assembly, are
847 Stop after the stage of compilation proper; do not assemble. The output
848 is in the form of an assembler code file for each non-assembler input
851 By default, the assembler file name for a source file is made by
852 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
854 Input files that don't require compilation are ignored.
858 Stop after the preprocessing stage; do not run the compiler proper. The
859 output is in the form of preprocessed source code, which is sent to the
862 Input files which don't require preprocessing are ignored.
864 @cindex output file option
867 Place output in file @var{file}. This applies regardless to whatever
868 sort of output is being produced, whether it be an executable file,
869 an object file, an assembler file or preprocessed C code.
871 If you specify @option{-o} when compiling more than one input file, or
872 you are producing an executable file as output, all the source files
873 on the command line will be compiled at once.
875 If @option{-o} is not specified, the default is to put an executable file
876 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
877 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
878 all preprocessed C source on standard output.
882 Print (on standard error output) the commands executed to run the stages
883 of compilation. Also print the version number of the compiler driver
884 program and of the preprocessor and the compiler proper.
888 Like @option{-v} except the commands are not executed and all command
889 arguments are quoted. This is useful for shell scripts to capture the
890 driver-generated command lines.
894 Use pipes rather than temporary files for communication between the
895 various stages of compilation. This fails to work on some systems where
896 the assembler is unable to read from a pipe; but the GNU assembler has
901 Print (on the standard output) a description of the command line options
902 understood by @command{gcc}. If the @option{-v} option is also specified
903 then @option{--help} will also be passed on to the various processes
904 invoked by @command{gcc}, so that they can display the command line options
905 they accept. If the @option{-Wextra} option is also specified then command
906 line options which have no documentation associated with them will also
911 Print (on the standard output) a description of target specific command
912 line options for each tool.
916 Display the version number and copyrights of the invoked GCC.
920 @section Compiling C++ Programs
922 @cindex suffixes for C++ source
923 @cindex C++ source file suffixes
924 C++ source files conventionally use one of the suffixes @samp{.C},
925 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
926 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
927 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
928 files with these names and compiles them as C++ programs even if you
929 call the compiler the same way as for compiling C programs (usually
930 with the name @command{gcc}).
934 However, C++ programs often require class libraries as well as a
935 compiler that understands the C++ language---and under some
936 circumstances, you might want to compile programs or header files from
937 standard input, or otherwise without a suffix that flags them as C++
938 programs. You might also like to precompile a C header file with a
939 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
940 program that calls GCC with the default language set to C++, and
941 automatically specifies linking against the C++ library. On many
942 systems, @command{g++} is also installed with the name @command{c++}.
944 @cindex invoking @command{g++}
945 When you compile C++ programs, you may specify many of the same
946 command-line options that you use for compiling programs in any
947 language; or command-line options meaningful for C and related
948 languages; or options that are meaningful only for C++ programs.
949 @xref{C Dialect Options,,Options Controlling C Dialect}, for
950 explanations of options for languages related to C@.
951 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
952 explanations of options that are meaningful only for C++ programs.
954 @node C Dialect Options
955 @section Options Controlling C Dialect
956 @cindex dialect options
957 @cindex language dialect options
958 @cindex options, dialect
960 The following options control the dialect of C (or languages derived
961 from C, such as C++ and Objective-C) that the compiler accepts:
968 In C mode, support all ISO C90 programs. In C++ mode,
969 remove GNU extensions that conflict with ISO C++.
971 This turns off certain features of GCC that are incompatible with ISO
972 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
973 such as the @code{asm} and @code{typeof} keywords, and
974 predefined macros such as @code{unix} and @code{vax} that identify the
975 type of system you are using. It also enables the undesirable and
976 rarely used ISO trigraph feature. For the C compiler,
977 it disables recognition of C++ style @samp{//} comments as well as
978 the @code{inline} keyword.
980 The alternate keywords @code{__asm__}, @code{__extension__},
981 @code{__inline__} and @code{__typeof__} continue to work despite
982 @option{-ansi}. You would not want to use them in an ISO C program, of
983 course, but it is useful to put them in header files that might be included
984 in compilations done with @option{-ansi}. Alternate predefined macros
985 such as @code{__unix__} and @code{__vax__} are also available, with or
986 without @option{-ansi}.
988 The @option{-ansi} option does not cause non-ISO programs to be
989 rejected gratuitously. For that, @option{-pedantic} is required in
990 addition to @option{-ansi}. @xref{Warning Options}.
992 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
993 option is used. Some header files may notice this macro and refrain
994 from declaring certain functions or defining certain macros that the
995 ISO standard doesn't call for; this is to avoid interfering with any
996 programs that might use these names for other things.
998 Functions which would normally be built in but do not have semantics
999 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1000 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1001 built-in functions provided by GCC}, for details of the functions
1006 Determine the language standard. This option is currently only
1007 supported when compiling C or C++. A value for this option must be
1008 provided; possible values are
1013 ISO C90 (same as @option{-ansi}).
1015 @item iso9899:199409
1016 ISO C90 as modified in amendment 1.
1022 ISO C99. Note that this standard is not yet fully supported; see
1023 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1024 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1027 Default, ISO C90 plus GNU extensions (including some C99 features).
1031 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1032 this will become the default. The name @samp{gnu9x} is deprecated.
1035 The 1998 ISO C++ standard plus amendments.
1038 The same as @option{-std=c++98} plus GNU extensions. This is the
1039 default for C++ code.
1042 Even when this option is not specified, you can still use some of the
1043 features of newer standards in so far as they do not conflict with
1044 previous C standards. For example, you may use @code{__restrict__} even
1045 when @option{-std=c99} is not specified.
1047 The @option{-std} options specifying some version of ISO C have the same
1048 effects as @option{-ansi}, except that features that were not in ISO C90
1049 but are in the specified version (for example, @samp{//} comments and
1050 the @code{inline} keyword in ISO C99) are not disabled.
1052 @xref{Standards,,Language Standards Supported by GCC}, for details of
1053 these standard versions.
1055 @item -aux-info @var{filename}
1057 Output to the given filename prototyped declarations for all functions
1058 declared and/or defined in a translation unit, including those in header
1059 files. This option is silently ignored in any language other than C@.
1061 Besides declarations, the file indicates, in comments, the origin of
1062 each declaration (source file and line), whether the declaration was
1063 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1064 @samp{O} for old, respectively, in the first character after the line
1065 number and the colon), and whether it came from a declaration or a
1066 definition (@samp{C} or @samp{F}, respectively, in the following
1067 character). In the case of function definitions, a K&R-style list of
1068 arguments followed by their declarations is also provided, inside
1069 comments, after the declaration.
1073 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1074 keyword, so that code can use these words as identifiers. You can use
1075 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1076 instead. @option{-ansi} implies @option{-fno-asm}.
1078 In C++, this switch only affects the @code{typeof} keyword, since
1079 @code{asm} and @code{inline} are standard keywords. You may want to
1080 use the @option{-fno-gnu-keywords} flag instead, which has the same
1081 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1082 switch only affects the @code{asm} and @code{typeof} keywords, since
1083 @code{inline} is a standard keyword in ISO C99.
1086 @itemx -fno-builtin-@var{function}
1087 @opindex fno-builtin
1088 @cindex built-in functions
1089 Don't recognize built-in functions that do not begin with
1090 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1091 functions provided by GCC}, for details of the functions affected,
1092 including those which are not built-in functions when @option{-ansi} or
1093 @option{-std} options for strict ISO C conformance are used because they
1094 do not have an ISO standard meaning.
1096 GCC normally generates special code to handle certain built-in functions
1097 more efficiently; for instance, calls to @code{alloca} may become single
1098 instructions that adjust the stack directly, and calls to @code{memcpy}
1099 may become inline copy loops. The resulting code is often both smaller
1100 and faster, but since the function calls no longer appear as such, you
1101 cannot set a breakpoint on those calls, nor can you change the behavior
1102 of the functions by linking with a different library.
1104 With the @option{-fno-builtin-@var{function}} option
1105 only the built-in function @var{function} is
1106 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1107 function is named this is not built-in in this version of GCC, this
1108 option is ignored. There is no corresponding
1109 @option{-fbuiltin-@var{function}} option; if you wish to enable
1110 built-in functions selectively when using @option{-fno-builtin} or
1111 @option{-ffreestanding}, you may define macros such as:
1114 #define abs(n) __builtin_abs ((n))
1115 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1120 @cindex hosted environment
1122 Assert that compilation takes place in a hosted environment. This implies
1123 @option{-fbuiltin}. A hosted environment is one in which the
1124 entire standard library is available, and in which @code{main} has a return
1125 type of @code{int}. Examples are nearly everything except a kernel.
1126 This is equivalent to @option{-fno-freestanding}.
1128 @item -ffreestanding
1129 @opindex ffreestanding
1130 @cindex hosted environment
1132 Assert that compilation takes place in a freestanding environment. This
1133 implies @option{-fno-builtin}. A freestanding environment
1134 is one in which the standard library may not exist, and program startup may
1135 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1136 This is equivalent to @option{-fno-hosted}.
1138 @xref{Standards,,Language Standards Supported by GCC}, for details of
1139 freestanding and hosted environments.
1141 @item -fms-extensions
1142 @opindex fms-extensions
1143 Accept some non-standard constructs used in Microsoft header files.
1147 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1148 options for strict ISO C conformance) implies @option{-trigraphs}.
1150 @item -no-integrated-cpp
1151 @opindex no-integrated-cpp
1152 Performs a compilation in two passes: preprocessing and compiling. This
1153 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1154 @option{-B} option. The user supplied compilation step can then add in
1155 an additional preprocessing step after normal preprocessing but before
1156 compiling. The default is to use the integrated cpp (internal cpp)
1158 The semantics of this option will change if "cc1", "cc1plus", and
1159 "cc1obj" are merged.
1161 @cindex traditional C language
1162 @cindex C language, traditional
1164 @itemx -traditional-cpp
1165 @opindex traditional-cpp
1166 @opindex traditional
1167 Formerly, these options caused GCC to attempt to emulate a pre-standard
1168 C compiler. They are now only supported with the @option{-E} switch.
1169 The preprocessor continues to support a pre-standard mode. See the GNU
1170 CPP manual for details.
1172 @item -fcond-mismatch
1173 @opindex fcond-mismatch
1174 Allow conditional expressions with mismatched types in the second and
1175 third arguments. The value of such an expression is void. This option
1176 is not supported for C++.
1178 @item -funsigned-char
1179 @opindex funsigned-char
1180 Let the type @code{char} be unsigned, like @code{unsigned char}.
1182 Each kind of machine has a default for what @code{char} should
1183 be. It is either like @code{unsigned char} by default or like
1184 @code{signed char} by default.
1186 Ideally, a portable program should always use @code{signed char} or
1187 @code{unsigned char} when it depends on the signedness of an object.
1188 But many programs have been written to use plain @code{char} and
1189 expect it to be signed, or expect it to be unsigned, depending on the
1190 machines they were written for. This option, and its inverse, let you
1191 make such a program work with the opposite default.
1193 The type @code{char} is always a distinct type from each of
1194 @code{signed char} or @code{unsigned char}, even though its behavior
1195 is always just like one of those two.
1198 @opindex fsigned-char
1199 Let the type @code{char} be signed, like @code{signed char}.
1201 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1202 the negative form of @option{-funsigned-char}. Likewise, the option
1203 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1205 @item -fsigned-bitfields
1206 @itemx -funsigned-bitfields
1207 @itemx -fno-signed-bitfields
1208 @itemx -fno-unsigned-bitfields
1209 @opindex fsigned-bitfields
1210 @opindex funsigned-bitfields
1211 @opindex fno-signed-bitfields
1212 @opindex fno-unsigned-bitfields
1213 These options control whether a bit-field is signed or unsigned, when the
1214 declaration does not use either @code{signed} or @code{unsigned}. By
1215 default, such a bit-field is signed, because this is consistent: the
1216 basic integer types such as @code{int} are signed types.
1219 @node C++ Dialect Options
1220 @section Options Controlling C++ Dialect
1222 @cindex compiler options, C++
1223 @cindex C++ options, command line
1224 @cindex options, C++
1225 This section describes the command-line options that are only meaningful
1226 for C++ programs; but you can also use most of the GNU compiler options
1227 regardless of what language your program is in. For example, you
1228 might compile a file @code{firstClass.C} like this:
1231 g++ -g -frepo -O -c firstClass.C
1235 In this example, only @option{-frepo} is an option meant
1236 only for C++ programs; you can use the other options with any
1237 language supported by GCC@.
1239 Here is a list of options that are @emph{only} for compiling C++ programs:
1243 @item -fabi-version=@var{n}
1244 @opindex fabi-version
1245 Use version @var{n} of the C++ ABI. Version 2 is the version of the
1246 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1247 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1248 the version that conforms most closely to the C++ ABI specification.
1249 Therefore, the ABI obtained using version 0 will change as ABI bugs
1252 The default is version 2.
1254 @item -fno-access-control
1255 @opindex fno-access-control
1256 Turn off all access checking. This switch is mainly useful for working
1257 around bugs in the access control code.
1261 Check that the pointer returned by @code{operator new} is non-null
1262 before attempting to modify the storage allocated. This check is
1263 normally unnecessary because the C++ standard specifies that
1264 @code{operator new} will only return @code{0} if it is declared
1265 @samp{throw()}, in which case the compiler will always check the
1266 return value even without this option. In all other cases, when
1267 @code{operator new} has a non-empty exception specification, memory
1268 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1269 @samp{new (nothrow)}.
1271 @item -fconserve-space
1272 @opindex fconserve-space
1273 Put uninitialized or runtime-initialized global variables into the
1274 common segment, as C does. This saves space in the executable at the
1275 cost of not diagnosing duplicate definitions. If you compile with this
1276 flag and your program mysteriously crashes after @code{main()} has
1277 completed, you may have an object that is being destroyed twice because
1278 two definitions were merged.
1280 This option is no longer useful on most targets, now that support has
1281 been added for putting variables into BSS without making them common.
1283 @item -fno-const-strings
1284 @opindex fno-const-strings
1285 Give string constants type @code{char *} instead of type @code{const
1286 char *}. By default, G++ uses type @code{const char *} as required by
1287 the standard. Even if you use @option{-fno-const-strings}, you cannot
1288 actually modify the value of a string constant.
1290 This option might be removed in a future release of G++. For maximum
1291 portability, you should structure your code so that it works with
1292 string constants that have type @code{const char *}.
1294 @item -fno-elide-constructors
1295 @opindex fno-elide-constructors
1296 The C++ standard allows an implementation to omit creating a temporary
1297 which is only used to initialize another object of the same type.
1298 Specifying this option disables that optimization, and forces G++ to
1299 call the copy constructor in all cases.
1301 @item -fno-enforce-eh-specs
1302 @opindex fno-enforce-eh-specs
1303 Don't check for violation of exception specifications at runtime. This
1304 option violates the C++ standard, but may be useful for reducing code
1305 size in production builds, much like defining @samp{NDEBUG}. The compiler
1306 will still optimize based on the exception specifications.
1309 @itemx -fno-for-scope
1311 @opindex fno-for-scope
1312 If @option{-ffor-scope} is specified, the scope of variables declared in
1313 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1314 as specified by the C++ standard.
1315 If @option{-fno-for-scope} is specified, the scope of variables declared in
1316 a @i{for-init-statement} extends to the end of the enclosing scope,
1317 as was the case in old versions of G++, and other (traditional)
1318 implementations of C++.
1320 The default if neither flag is given to follow the standard,
1321 but to allow and give a warning for old-style code that would
1322 otherwise be invalid, or have different behavior.
1324 @item -fno-gnu-keywords
1325 @opindex fno-gnu-keywords
1326 Do not recognize @code{typeof} as a keyword, so that code can use this
1327 word as an identifier. You can use the keyword @code{__typeof__} instead.
1328 @option{-ansi} implies @option{-fno-gnu-keywords}.
1330 @item -fno-implicit-templates
1331 @opindex fno-implicit-templates
1332 Never emit code for non-inline templates which are instantiated
1333 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1334 @xref{Template Instantiation}, for more information.
1336 @item -fno-implicit-inline-templates
1337 @opindex fno-implicit-inline-templates
1338 Don't emit code for implicit instantiations of inline templates, either.
1339 The default is to handle inlines differently so that compiles with and
1340 without optimization will need the same set of explicit instantiations.
1342 @item -fno-implement-inlines
1343 @opindex fno-implement-inlines
1344 To save space, do not emit out-of-line copies of inline functions
1345 controlled by @samp{#pragma implementation}. This will cause linker
1346 errors if these functions are not inlined everywhere they are called.
1348 @item -fms-extensions
1349 @opindex fms-extensions
1350 Disable pedantic warnings about constructs used in MFC, such as implicit
1351 int and getting a pointer to member function via non-standard syntax.
1353 @item -fno-nonansi-builtins
1354 @opindex fno-nonansi-builtins
1355 Disable built-in declarations of functions that are not mandated by
1356 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1357 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1359 @item -fno-operator-names
1360 @opindex fno-operator-names
1361 Do not treat the operator name keywords @code{and}, @code{bitand},
1362 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1363 synonyms as keywords.
1365 @item -fno-optional-diags
1366 @opindex fno-optional-diags
1367 Disable diagnostics that the standard says a compiler does not need to
1368 issue. Currently, the only such diagnostic issued by G++ is the one for
1369 a name having multiple meanings within a class.
1372 @opindex fpermissive
1373 Downgrade some diagnostics about nonconformant code from errors to
1374 warnings. Thus, using @option{-fpermissive} will allow some
1375 nonconforming code to compile.
1379 Enable automatic template instantiation at link time. This option also
1380 implies @option{-fno-implicit-templates}. @xref{Template
1381 Instantiation}, for more information.
1385 Disable generation of information about every class with virtual
1386 functions for use by the C++ runtime type identification features
1387 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1388 of the language, you can save some space by using this flag. Note that
1389 exception handling uses the same information, but it will generate it as
1394 Emit statistics about front-end processing at the end of the compilation.
1395 This information is generally only useful to the G++ development team.
1397 @item -ftemplate-depth-@var{n}
1398 @opindex ftemplate-depth
1399 Set the maximum instantiation depth for template classes to @var{n}.
1400 A limit on the template instantiation depth is needed to detect
1401 endless recursions during template class instantiation. ANSI/ISO C++
1402 conforming programs must not rely on a maximum depth greater than 17.
1404 @item -fuse-cxa-atexit
1405 @opindex fuse-cxa-atexit
1406 Register destructors for objects with static storage duration with the
1407 @code{__cxa_atexit} function rather than the @code{atexit} function.
1408 This option is required for fully standards-compliant handling of static
1409 destructors, but will only work if your C library supports
1410 @code{__cxa_atexit}.
1414 Do not use weak symbol support, even if it is provided by the linker.
1415 By default, G++ will use weak symbols if they are available. This
1416 option exists only for testing, and should not be used by end-users;
1417 it will result in inferior code and has no benefits. This option may
1418 be removed in a future release of G++.
1422 Do not search for header files in the standard directories specific to
1423 C++, but do still search the other standard directories. (This option
1424 is used when building the C++ library.)
1427 In addition, these optimization, warning, and code generation options
1428 have meanings only for C++ programs:
1431 @item -fno-default-inline
1432 @opindex fno-default-inline
1433 Do not assume @samp{inline} for functions defined inside a class scope.
1434 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1435 functions will have linkage like inline functions; they just won't be
1438 @item -Wabi @r{(C++ only)}
1440 Warn when G++ generates code that is probably not compatible with the
1441 vendor-neutral C++ ABI. Although an effort has been made to warn about
1442 all such cases, there are probably some cases that are not warned about,
1443 even though G++ is generating incompatible code. There may also be
1444 cases where warnings are emitted even though the code that is generated
1447 You should rewrite your code to avoid these warnings if you are
1448 concerned about the fact that code generated by G++ may not be binary
1449 compatible with code generated by other compilers.
1451 The known incompatibilities at this point include:
1456 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1457 pack data into the same byte as a base class. For example:
1460 struct A @{ virtual void f(); int f1 : 1; @};
1461 struct B : public A @{ int f2 : 1; @};
1465 In this case, G++ will place @code{B::f2} into the same byte
1466 as@code{A::f1}; other compilers will not. You can avoid this problem
1467 by explicitly padding @code{A} so that its size is a multiple of the
1468 byte size on your platform; that will cause G++ and other compilers to
1469 layout @code{B} identically.
1472 Incorrect handling of tail-padding for virtual bases. G++ does not use
1473 tail padding when laying out virtual bases. For example:
1476 struct A @{ virtual void f(); char c1; @};
1477 struct B @{ B(); char c2; @};
1478 struct C : public A, public virtual B @{@};
1482 In this case, G++ will not place @code{B} into the tail-padding for
1483 @code{A}; other compilers will. You can avoid this problem by
1484 explicitly padding @code{A} so that its size is a multiple of its
1485 alignment (ignoring virtual base classes); that will cause G++ and other
1486 compilers to layout @code{C} identically.
1489 Incorrect handling of bit-fields with declared widths greater than that
1490 of their underlying types, when the bit-fields appear in a union. For
1494 union U @{ int i : 4096; @};
1498 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1499 union too small by the number of bits in an @code{int}.
1502 Empty classes can be placed at incorrect offsets. For example:
1512 struct C : public B, public A @{@};
1516 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1517 it should be placed at offset zero. G++ mistakenly believes that the
1518 @code{A} data member of @code{B} is already at offset zero.
1521 Names of template functions whose types involve @code{typename} or
1522 template template parameters can be mangled incorrectly.
1525 template <typename Q>
1526 void f(typename Q::X) @{@}
1528 template <template <typename> class Q>
1529 void f(typename Q<int>::X) @{@}
1533 Instantiations of these templates may be mangled incorrectly.
1537 @item -Wctor-dtor-privacy @r{(C++ only)}
1538 @opindex Wctor-dtor-privacy
1539 Warn when a class seems unusable because all the constructors or
1540 destructors in that class are private, and it has neither friends nor
1541 public static member functions.
1543 @item -Wnon-virtual-dtor @r{(C++ only)}
1544 @opindex Wnon-virtual-dtor
1545 Warn when a class appears to be polymorphic, thereby requiring a virtual
1546 destructor, yet it declares a non-virtual one.
1547 This warning is enabled by @option{-Wall}.
1549 @item -Wreorder @r{(C++ only)}
1551 @cindex reordering, warning
1552 @cindex warning for reordering of member initializers
1553 Warn when the order of member initializers given in the code does not
1554 match the order in which they must be executed. For instance:
1560 A(): j (0), i (1) @{ @}
1564 The compiler will rearrange the member initializers for @samp{i}
1565 and @samp{j} to match the declaration order of the members, emitting
1566 a warning to that effect. This warning is enabled by @option{-Wall}.
1569 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1572 @item -Weffc++ @r{(C++ only)}
1574 Warn about violations of the following style guidelines from Scott Meyers'
1575 @cite{Effective C++} book:
1579 Item 11: Define a copy constructor and an assignment operator for classes
1580 with dynamically allocated memory.
1583 Item 12: Prefer initialization to assignment in constructors.
1586 Item 14: Make destructors virtual in base classes.
1589 Item 15: Have @code{operator=} return a reference to @code{*this}.
1592 Item 23: Don't try to return a reference when you must return an object.
1596 Also warn about violations of the following style guidelines from
1597 Scott Meyers' @cite{More Effective C++} book:
1601 Item 6: Distinguish between prefix and postfix forms of increment and
1602 decrement operators.
1605 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1609 When selecting this option, be aware that the standard library
1610 headers do not obey all of these guidelines; use @samp{grep -v}
1611 to filter out those warnings.
1613 @item -Wno-deprecated @r{(C++ only)}
1614 @opindex Wno-deprecated
1615 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1617 @item -Wno-non-template-friend @r{(C++ only)}
1618 @opindex Wno-non-template-friend
1619 Disable warnings when non-templatized friend functions are declared
1620 within a template. Since the advent of explicit template specification
1621 support in G++, if the name of the friend is an unqualified-id (i.e.,
1622 @samp{friend foo(int)}), the C++ language specification demands that the
1623 friend declare or define an ordinary, nontemplate function. (Section
1624 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1625 could be interpreted as a particular specialization of a templatized
1626 function. Because this non-conforming behavior is no longer the default
1627 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1628 check existing code for potential trouble spots and is on by default.
1629 This new compiler behavior can be turned off with
1630 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1631 but disables the helpful warning.
1633 @item -Wold-style-cast @r{(C++ only)}
1634 @opindex Wold-style-cast
1635 Warn if an old-style (C-style) cast to a non-void type is used within
1636 a C++ program. The new-style casts (@samp{static_cast},
1637 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1638 unintended effects and much easier to search for.
1640 @item -Woverloaded-virtual @r{(C++ only)}
1641 @opindex Woverloaded-virtual
1642 @cindex overloaded virtual fn, warning
1643 @cindex warning for overloaded virtual fn
1644 Warn when a function declaration hides virtual functions from a
1645 base class. For example, in:
1652 struct B: public A @{
1657 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1665 will fail to compile.
1667 @item -Wno-pmf-conversions @r{(C++ only)}
1668 @opindex Wno-pmf-conversions
1669 Disable the diagnostic for converting a bound pointer to member function
1672 @item -Wsign-promo @r{(C++ only)}
1673 @opindex Wsign-promo
1674 Warn when overload resolution chooses a promotion from unsigned or
1675 enumeral type to a signed type, over a conversion to an unsigned type of
1676 the same size. Previous versions of G++ would try to preserve
1677 unsignedness, but the standard mandates the current behavior.
1679 @item -Wsynth @r{(C++ only)}
1681 @cindex warning for synthesized methods
1682 @cindex synthesized methods, warning
1683 Warn when G++'s synthesis behavior does not match that of cfront. For
1689 A& operator = (int);
1699 In this example, G++ will synthesize a default @samp{A& operator =
1700 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1703 @node Objective-C Dialect Options
1704 @section Options Controlling Objective-C Dialect
1706 @cindex compiler options, Objective-C
1707 @cindex Objective-C options, command line
1708 @cindex options, Objective-C
1709 (NOTE: This manual does not describe the Objective-C language itself. See
1710 @w{@uref{http://gcc.gnu.org/readings.html}} for references.)
1712 This section describes the command-line options that are only meaningful
1713 for Objective-C programs, but you can also use most of the GNU compiler
1714 options regardless of what language your program is in. For example,
1715 you might compile a file @code{some_class.m} like this:
1718 gcc -g -fgnu-runtime -O -c some_class.m
1722 In this example, @option{-fgnu-runtime} is an option meant only for
1723 Objective-C programs; you can use the other options with any language
1726 Here is a list of options that are @emph{only} for compiling Objective-C
1730 @item -fconstant-string-class=@var{class-name}
1731 @opindex fconstant-string-class
1732 Use @var{class-name} as the name of the class to instantiate for each
1733 literal string specified with the syntax @code{@@"@dots{}"}. The default
1734 class name is @code{NXConstantString} if the GNU runtime is being used, and
1735 @code{NSConstantString} if the NeXT runtime is being used (see below). The
1736 @option{-fconstant-cfstrings} option, if also present, will override the
1737 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
1738 to be laid out as constant CoreFoundation strings.
1741 @opindex fgnu-runtime
1742 Generate object code compatible with the standard GNU Objective-C
1743 runtime. This is the default for most types of systems.
1745 @item -fnext-runtime
1746 @opindex fnext-runtime
1747 Generate output compatible with the NeXT runtime. This is the default
1748 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1749 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1752 @item -fno-nil-receivers
1753 @opindex -fno-nil-receivers
1754 Assume that all Objective-C message dispatches (e.g.,
1755 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
1756 is not @code{nil}. This allows for more efficient entry points in the runtime to be
1757 used. Currently, this option is only available in conjunction with
1758 the NeXT runtime on Mac OS X 10.3 and later.
1760 @item -fobjc-exceptions
1761 @opindex -fobjc-exceptions
1762 Enable syntactic support for structured exception handling in Objective-C,
1763 similar to what is offered by C++ and Java. Currently, this option is only
1764 available in conjunction with the NeXT runtime on Mac OS X 10.3 and later.
1772 @@catch (AnObjCClass *exc) @{
1779 @@catch (AnotherClass *exc) @{
1782 @@catch (id allOthers) @{
1792 The @code{@@throw} statement may appear anywhere in an Objective-C or
1793 Objective-C++ program; when used inside of a @code{@@catch} block, the
1794 @code{@@throw} may appear without an argument (as shown above), in which case
1795 the object caught by the @code{@@catch} will be rethrown.
1797 Note that only (pointers to) Objective-C objects may be thrown and
1798 caught using this scheme. When an object is thrown, it will be caught
1799 by the nearest @code{@@catch} clause capable of handling objects of that type,
1800 analogously to how @code{catch} blocks work in C++ and Java. A
1801 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
1802 any and all Objective-C exceptions not caught by previous @code{@@catch}
1805 The @code{@@finally} clause, if present, will be executed upon exit from the
1806 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
1807 regardless of whether any exceptions are thrown, caught or rethrown
1808 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
1809 of the @code{finally} clause in Java.
1811 There are several caveats to using the new exception mechanism:
1815 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
1816 idioms provided by the @code{NSException} class, the new
1817 exceptions can only be used on Mac OS X 10.3 (Panther) and later
1818 systems, due to additional functionality needed in the (NeXT) Objective-C
1822 As mentioned above, the new exceptions do not support handling
1823 types other than Objective-C objects. Furthermore, when used from
1824 Objective-C++, the Objective-C exception model does not interoperate with C++
1825 exceptions at this time. This means you cannot @code{@@throw} an exception
1826 from Objective-C and @code{catch} it in C++, or vice versa
1827 (i.e., @code{throw @dots{} @@catch}).
1830 The @option{-fobjc-exceptions} switch also enables the use of synchronization
1831 blocks for thread-safe execution:
1834 @@synchronized (ObjCClass *guard) @{
1839 Upon entering the @code{@@synchronized} block, a thread of execution shall
1840 first check whether a lock has been placed on the corresponding @code{guard}
1841 object by another thread. If it has, the current thread shall wait until
1842 the other thread relinquishes its lock. Once @code{guard} becomes available,
1843 the current thread will place its own lock on it, execute the code contained in
1844 the @code{@@synchronized} block, and finally relinquish the lock (thereby
1845 making @code{guard} available to other threads).
1847 Unlike Java, Objective-C does not allow for entire methods to be marked
1848 @code{@@synchronized}. Note that throwing exceptions out of
1849 @code{@@synchronized} blocks is allowed, and will cause the guarding object
1850 to be unlocked properly.
1852 @item -freplace-objc-classes
1853 @opindex -freplace-objc-classes
1854 Emit a special marker instructing @command{ld(1)} not to statically link in
1855 the resulting object file, and allow @command{dyld(1)} to load it in at
1856 run time instead. This is used in conjunction with the Fix-and-Continue
1857 debugging mode, where the object file in question may be recompiled and
1858 dynamically reloaded in the course of program execution, without the need
1859 to restart the program itself. Currently, Fix-and-Continue functionality
1860 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
1864 @opindex -fzero-link
1865 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
1866 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
1867 compile time) with static class references that get initialized at load time,
1868 which improves run-time performance. Specifying the @option{-fzero-link} flag
1869 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
1870 to be retained. This is useful in Zero-Link debugging mode, since it allows
1871 for individual class implementations to be modified during program execution.
1875 Dump interface declarations for all classes seen in the source file to a
1876 file named @file{@var{sourcename}.decl}.
1879 @opindex Wno-protocol
1880 If a class is declared to implement a protocol, a warning is issued for
1881 every method in the protocol that is not implemented by the class. The
1882 default behavior is to issue a warning for every method not explicitly
1883 implemented in the class, even if a method implementation is inherited
1884 from the superclass. If you use the @code{-Wno-protocol} option, then
1885 methods inherited from the superclass are considered to be implemented,
1886 and no warning is issued for them.
1890 Warn if multiple methods of different types for the same selector are
1891 found during compilation. The check is performed on the list of methods
1892 in the final stage of compilation. Additionally, a check is performed
1893 for each selector appearing in a @code{@@selector(@dots{})}
1894 expression, and a corresponding method for that selector has been found
1895 during compilation. Because these checks scan the method table only at
1896 the end of compilation, these warnings are not produced if the final
1897 stage of compilation is not reached, for example because an error is
1898 found during compilation, or because the @code{-fsyntax-only} option is
1901 @item -Wundeclared-selector
1902 @opindex Wundeclared-selector
1903 Warn if a @code{@@selector(@dots{})} expression referring to an
1904 undeclared selector is found. A selector is considered undeclared if no
1905 method with that name has been declared before the
1906 @code{@@selector(@dots{})} expression, either explicitly in an
1907 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
1908 an @code{@@implementation} section. This option always performs its
1909 checks as soon as a @code{@@selector(@dots{})} expression is found,
1910 while @code{-Wselector} only performs its checks in the final stage of
1911 compilation. This also enforces the coding style convention
1912 that methods and selectors must be declared before being used.
1914 @item -print-objc-runtime-info
1915 @opindex -print-objc-runtime-info
1916 Generate C header describing the largest structure that is passed by
1921 @node Language Independent Options
1922 @section Options to Control Diagnostic Messages Formatting
1923 @cindex options to control diagnostics formatting
1924 @cindex diagnostic messages
1925 @cindex message formatting
1927 Traditionally, diagnostic messages have been formatted irrespective of
1928 the output device's aspect (e.g.@: its width, @dots{}). The options described
1929 below can be used to control the diagnostic messages formatting
1930 algorithm, e.g.@: how many characters per line, how often source location
1931 information should be reported. Right now, only the C++ front end can
1932 honor these options. However it is expected, in the near future, that
1933 the remaining front ends would be able to digest them correctly.
1936 @item -fmessage-length=@var{n}
1937 @opindex fmessage-length
1938 Try to format error messages so that they fit on lines of about @var{n}
1939 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1940 the front ends supported by GCC@. If @var{n} is zero, then no
1941 line-wrapping will be done; each error message will appear on a single
1944 @opindex fdiagnostics-show-location
1945 @item -fdiagnostics-show-location=once
1946 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1947 reporter to emit @emph{once} source location information; that is, in
1948 case the message is too long to fit on a single physical line and has to
1949 be wrapped, the source location won't be emitted (as prefix) again,
1950 over and over, in subsequent continuation lines. This is the default
1953 @item -fdiagnostics-show-location=every-line
1954 Only meaningful in line-wrapping mode. Instructs the diagnostic
1955 messages reporter to emit the same source location information (as
1956 prefix) for physical lines that result from the process of breaking
1957 a message which is too long to fit on a single line.
1961 @node Warning Options
1962 @section Options to Request or Suppress Warnings
1963 @cindex options to control warnings
1964 @cindex warning messages
1965 @cindex messages, warning
1966 @cindex suppressing warnings
1968 Warnings are diagnostic messages that report constructions which
1969 are not inherently erroneous but which are risky or suggest there
1970 may have been an error.
1972 You can request many specific warnings with options beginning @samp{-W},
1973 for example @option{-Wimplicit} to request warnings on implicit
1974 declarations. Each of these specific warning options also has a
1975 negative form beginning @samp{-Wno-} to turn off warnings;
1976 for example, @option{-Wno-implicit}. This manual lists only one of the
1977 two forms, whichever is not the default.
1979 The following options control the amount and kinds of warnings produced
1980 by GCC; for further, language-specific options also refer to
1981 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
1984 @cindex syntax checking
1986 @opindex fsyntax-only
1987 Check the code for syntax errors, but don't do anything beyond that.
1991 Issue all the warnings demanded by strict ISO C and ISO C++;
1992 reject all programs that use forbidden extensions, and some other
1993 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1994 version of the ISO C standard specified by any @option{-std} option used.
1996 Valid ISO C and ISO C++ programs should compile properly with or without
1997 this option (though a rare few will require @option{-ansi} or a
1998 @option{-std} option specifying the required version of ISO C)@. However,
1999 without this option, certain GNU extensions and traditional C and C++
2000 features are supported as well. With this option, they are rejected.
2002 @option{-pedantic} does not cause warning messages for use of the
2003 alternate keywords whose names begin and end with @samp{__}. Pedantic
2004 warnings are also disabled in the expression that follows
2005 @code{__extension__}. However, only system header files should use
2006 these escape routes; application programs should avoid them.
2007 @xref{Alternate Keywords}.
2009 Some users try to use @option{-pedantic} to check programs for strict ISO
2010 C conformance. They soon find that it does not do quite what they want:
2011 it finds some non-ISO practices, but not all---only those for which
2012 ISO C @emph{requires} a diagnostic, and some others for which
2013 diagnostics have been added.
2015 A feature to report any failure to conform to ISO C might be useful in
2016 some instances, but would require considerable additional work and would
2017 be quite different from @option{-pedantic}. We don't have plans to
2018 support such a feature in the near future.
2020 Where the standard specified with @option{-std} represents a GNU
2021 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2022 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2023 extended dialect is based. Warnings from @option{-pedantic} are given
2024 where they are required by the base standard. (It would not make sense
2025 for such warnings to be given only for features not in the specified GNU
2026 C dialect, since by definition the GNU dialects of C include all
2027 features the compiler supports with the given option, and there would be
2028 nothing to warn about.)
2030 @item -pedantic-errors
2031 @opindex pedantic-errors
2032 Like @option{-pedantic}, except that errors are produced rather than
2037 Inhibit all warning messages.
2041 Inhibit warning messages about the use of @samp{#import}.
2043 @item -Wchar-subscripts
2044 @opindex Wchar-subscripts
2045 Warn if an array subscript has type @code{char}. This is a common cause
2046 of error, as programmers often forget that this type is signed on some
2051 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2052 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2056 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2057 the arguments supplied have types appropriate to the format string
2058 specified, and that the conversions specified in the format string make
2059 sense. This includes standard functions, and others specified by format
2060 attributes (@pxref{Function Attributes}), in the @code{printf},
2061 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2062 not in the C standard) families.
2064 The formats are checked against the format features supported by GNU
2065 libc version 2.2. These include all ISO C90 and C99 features, as well
2066 as features from the Single Unix Specification and some BSD and GNU
2067 extensions. Other library implementations may not support all these
2068 features; GCC does not support warning about features that go beyond a
2069 particular library's limitations. However, if @option{-pedantic} is used
2070 with @option{-Wformat}, warnings will be given about format features not
2071 in the selected standard version (but not for @code{strfmon} formats,
2072 since those are not in any version of the C standard). @xref{C Dialect
2073 Options,,Options Controlling C Dialect}.
2075 Since @option{-Wformat} also checks for null format arguments for
2076 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2078 @option{-Wformat} is included in @option{-Wall}. For more control over some
2079 aspects of format checking, the options @option{-Wformat-y2k},
2080 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2081 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2082 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2085 @opindex Wformat-y2k
2086 If @option{-Wformat} is specified, also warn about @code{strftime}
2087 formats which may yield only a two-digit year.
2089 @item -Wno-format-extra-args
2090 @opindex Wno-format-extra-args
2091 If @option{-Wformat} is specified, do not warn about excess arguments to a
2092 @code{printf} or @code{scanf} format function. The C standard specifies
2093 that such arguments are ignored.
2095 Where the unused arguments lie between used arguments that are
2096 specified with @samp{$} operand number specifications, normally
2097 warnings are still given, since the implementation could not know what
2098 type to pass to @code{va_arg} to skip the unused arguments. However,
2099 in the case of @code{scanf} formats, this option will suppress the
2100 warning if the unused arguments are all pointers, since the Single
2101 Unix Specification says that such unused arguments are allowed.
2103 @item -Wno-format-zero-length
2104 @opindex Wno-format-zero-length
2105 If @option{-Wformat} is specified, do not warn about zero-length formats.
2106 The C standard specifies that zero-length formats are allowed.
2108 @item -Wformat-nonliteral
2109 @opindex Wformat-nonliteral
2110 If @option{-Wformat} is specified, also warn if the format string is not a
2111 string literal and so cannot be checked, unless the format function
2112 takes its format arguments as a @code{va_list}.
2114 @item -Wformat-security
2115 @opindex Wformat-security
2116 If @option{-Wformat} is specified, also warn about uses of format
2117 functions that represent possible security problems. At present, this
2118 warns about calls to @code{printf} and @code{scanf} functions where the
2119 format string is not a string literal and there are no format arguments,
2120 as in @code{printf (foo);}. This may be a security hole if the format
2121 string came from untrusted input and contains @samp{%n}. (This is
2122 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2123 in future warnings may be added to @option{-Wformat-security} that are not
2124 included in @option{-Wformat-nonliteral}.)
2128 Enable @option{-Wformat} plus format checks not included in
2129 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2130 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2134 Warn about passing a null pointer for arguments marked as
2135 requiring a non-null value by the @code{nonnull} function attribute.
2137 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2138 can be disabled with the @option{-Wno-nonnull} option.
2140 @item -Winit-self @r{(C, C++, and Objective-C only)}
2142 Warn about uninitialized variables which are initialized with themselves.
2143 Note this option can only be used with the @option{-Wuninitialized} option,
2144 which in turn only works with @option{-O1} and above.
2146 For example, GCC will warn about @code{i} being uninitialized in the
2147 following snippet only when @option{-Winit-self} has been specified:
2158 @item -Wimplicit-int
2159 @opindex Wimplicit-int
2160 Warn when a declaration does not specify a type.
2162 @item -Wimplicit-function-declaration
2163 @itemx -Werror-implicit-function-declaration
2164 @opindex Wimplicit-function-declaration
2165 @opindex Werror-implicit-function-declaration
2166 Give a warning (or error) whenever a function is used before being
2171 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2175 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2176 function with external linkage, returning int, taking either zero
2177 arguments, two, or three arguments of appropriate types.
2179 @item -Wmissing-braces
2180 @opindex Wmissing-braces
2181 Warn if an aggregate or union initializer is not fully bracketed. In
2182 the following example, the initializer for @samp{a} is not fully
2183 bracketed, but that for @samp{b} is fully bracketed.
2186 int a[2][2] = @{ 0, 1, 2, 3 @};
2187 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2191 @opindex Wparentheses
2192 Warn if parentheses are omitted in certain contexts, such
2193 as when there is an assignment in a context where a truth value
2194 is expected, or when operators are nested whose precedence people
2195 often get confused about.
2197 Also warn about constructions where there may be confusion to which
2198 @code{if} statement an @code{else} branch belongs. Here is an example of
2213 In C, every @code{else} branch belongs to the innermost possible @code{if}
2214 statement, which in this example is @code{if (b)}. This is often not
2215 what the programmer expected, as illustrated in the above example by
2216 indentation the programmer chose. When there is the potential for this
2217 confusion, GCC will issue a warning when this flag is specified.
2218 To eliminate the warning, add explicit braces around the innermost
2219 @code{if} statement so there is no way the @code{else} could belong to
2220 the enclosing @code{if}. The resulting code would look like this:
2236 @item -Wsequence-point
2237 @opindex Wsequence-point
2238 Warn about code that may have undefined semantics because of violations
2239 of sequence point rules in the C standard.
2241 The C standard defines the order in which expressions in a C program are
2242 evaluated in terms of @dfn{sequence points}, which represent a partial
2243 ordering between the execution of parts of the program: those executed
2244 before the sequence point, and those executed after it. These occur
2245 after the evaluation of a full expression (one which is not part of a
2246 larger expression), after the evaluation of the first operand of a
2247 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2248 function is called (but after the evaluation of its arguments and the
2249 expression denoting the called function), and in certain other places.
2250 Other than as expressed by the sequence point rules, the order of
2251 evaluation of subexpressions of an expression is not specified. All
2252 these rules describe only a partial order rather than a total order,
2253 since, for example, if two functions are called within one expression
2254 with no sequence point between them, the order in which the functions
2255 are called is not specified. However, the standards committee have
2256 ruled that function calls do not overlap.
2258 It is not specified when between sequence points modifications to the
2259 values of objects take effect. Programs whose behavior depends on this
2260 have undefined behavior; the C standard specifies that ``Between the
2261 previous and next sequence point an object shall have its stored value
2262 modified at most once by the evaluation of an expression. Furthermore,
2263 the prior value shall be read only to determine the value to be
2264 stored.''. If a program breaks these rules, the results on any
2265 particular implementation are entirely unpredictable.
2267 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2268 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2269 diagnosed by this option, and it may give an occasional false positive
2270 result, but in general it has been found fairly effective at detecting
2271 this sort of problem in programs.
2273 The present implementation of this option only works for C programs. A
2274 future implementation may also work for C++ programs.
2276 The C standard is worded confusingly, therefore there is some debate
2277 over the precise meaning of the sequence point rules in subtle cases.
2278 Links to discussions of the problem, including proposed formal
2279 definitions, may be found on our readings page, at
2280 @w{@uref{http://gcc.gnu.org/readings.html}}.
2283 @opindex Wreturn-type
2284 Warn whenever a function is defined with a return-type that defaults to
2285 @code{int}. Also warn about any @code{return} statement with no
2286 return-value in a function whose return-type is not @code{void}.
2288 For C++, a function without return type always produces a diagnostic
2289 message, even when @option{-Wno-return-type} is specified. The only
2290 exceptions are @samp{main} and functions defined in system headers.
2294 Warn whenever a @code{switch} statement has an index of enumeral type
2295 and lacks a @code{case} for one or more of the named codes of that
2296 enumeration. (The presence of a @code{default} label prevents this
2297 warning.) @code{case} labels outside the enumeration range also
2298 provoke warnings when this option is used.
2300 @item -Wswitch-default
2301 @opindex Wswitch-switch
2302 Warn whenever a @code{switch} statement does not have a @code{default}
2306 @opindex Wswitch-enum
2307 Warn whenever a @code{switch} statement has an index of enumeral type
2308 and lacks a @code{case} for one or more of the named codes of that
2309 enumeration. @code{case} labels outside the enumeration range also
2310 provoke warnings when this option is used.
2314 Warn if any trigraphs are encountered that might change the meaning of
2315 the program (trigraphs within comments are not warned about).
2317 @item -Wunused-function
2318 @opindex Wunused-function
2319 Warn whenever a static function is declared but not defined or a
2320 non\-inline static function is unused.
2322 @item -Wunused-label
2323 @opindex Wunused-label
2324 Warn whenever a label is declared but not used.
2326 To suppress this warning use the @samp{unused} attribute
2327 (@pxref{Variable Attributes}).
2329 @item -Wunused-parameter
2330 @opindex Wunused-parameter
2331 Warn whenever a function parameter is unused aside from its declaration.
2333 To suppress this warning use the @samp{unused} attribute
2334 (@pxref{Variable Attributes}).
2336 @item -Wunused-variable
2337 @opindex Wunused-variable
2338 Warn whenever a local variable or non-constant static variable is unused
2339 aside from its declaration
2341 To suppress this warning use the @samp{unused} attribute
2342 (@pxref{Variable Attributes}).
2344 @item -Wunused-value
2345 @opindex Wunused-value
2346 Warn whenever a statement computes a result that is explicitly not used.
2348 To suppress this warning cast the expression to @samp{void}.
2352 All the above @option{-Wunused} options combined.
2354 In order to get a warning about an unused function parameter, you must
2355 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2356 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2358 @item -Wuninitialized
2359 @opindex Wuninitialized
2360 Warn if an automatic variable is used without first being initialized or
2361 if a variable may be clobbered by a @code{setjmp} call.
2363 These warnings are possible only in optimizing compilation,
2364 because they require data flow information that is computed only
2365 when optimizing. If you don't specify @option{-O}, you simply won't
2368 If you want to warn about code which uses the uninitialized value of the
2369 variable in its own initializer, use the @option{-Winit-self} option.
2371 These warnings occur only for variables that are candidates for
2372 register allocation. Therefore, they do not occur for a variable that
2373 is declared @code{volatile}, or whose address is taken, or whose size
2374 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2375 structures, unions or arrays, even when they are in registers.
2377 Note that there may be no warning about a variable that is used only
2378 to compute a value that itself is never used, because such
2379 computations may be deleted by data flow analysis before the warnings
2382 These warnings are made optional because GCC is not smart
2383 enough to see all the reasons why the code might be correct
2384 despite appearing to have an error. Here is one example of how
2405 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2406 always initialized, but GCC doesn't know this. Here is
2407 another common case:
2412 if (change_y) save_y = y, y = new_y;
2414 if (change_y) y = save_y;
2419 This has no bug because @code{save_y} is used only if it is set.
2421 @cindex @code{longjmp} warnings
2422 This option also warns when a non-volatile automatic variable might be
2423 changed by a call to @code{longjmp}. These warnings as well are possible
2424 only in optimizing compilation.
2426 The compiler sees only the calls to @code{setjmp}. It cannot know
2427 where @code{longjmp} will be called; in fact, a signal handler could
2428 call it at any point in the code. As a result, you may get a warning
2429 even when there is in fact no problem because @code{longjmp} cannot
2430 in fact be called at the place which would cause a problem.
2432 Some spurious warnings can be avoided if you declare all the functions
2433 you use that never return as @code{noreturn}. @xref{Function
2436 @item -Wunknown-pragmas
2437 @opindex Wunknown-pragmas
2438 @cindex warning for unknown pragmas
2439 @cindex unknown pragmas, warning
2440 @cindex pragmas, warning of unknown
2441 Warn when a #pragma directive is encountered which is not understood by
2442 GCC@. If this command line option is used, warnings will even be issued
2443 for unknown pragmas in system header files. This is not the case if
2444 the warnings were only enabled by the @option{-Wall} command line option.
2446 @item -Wstrict-aliasing
2447 @opindex Wstrict-aliasing
2448 This option is only active when @option{-fstrict-aliasing} is active.
2449 It warns about code which might break the strict aliasing rules that the
2450 compiler is using for optimization. The warning does not catch all
2451 cases, but does attempt to catch the more common pitfalls. It is
2452 included in @option{-Wall}.
2454 @item -Wstrict-aliasing=2
2455 @opindex Wstrict-aliasing=2
2456 This option is only active when @option{-fstrict-aliasing} is active.
2457 It warns about all code which might break the strict aliasing rules that the
2458 compiler is using for optimization. This warning catches all cases, but
2459 it will also give a warning for some ambiguous cases that are safe.
2463 All of the above @samp{-W} options combined. This enables all the
2464 warnings about constructions that some users consider questionable, and
2465 that are easy to avoid (or modify to prevent the warning), even in
2466 conjunction with macros. This also enables some language-specific
2467 warnings described in @ref{C++ Dialect Options} and
2468 @ref{Objective-C Dialect Options}.
2471 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2472 Some of them warn about constructions that users generally do not
2473 consider questionable, but which occasionally you might wish to check
2474 for; others warn about constructions that are necessary or hard to avoid
2475 in some cases, and there is no simple way to modify the code to suppress
2482 (This option used to be called @option{-W}. The older name is still
2483 supported, but the newer name is more descriptive.) Print extra warning
2484 messages for these events:
2488 A function can return either with or without a value. (Falling
2489 off the end of the function body is considered returning without
2490 a value.) For example, this function would evoke such a
2504 An expression-statement or the left-hand side of a comma expression
2505 contains no side effects.
2506 To suppress the warning, cast the unused expression to void.
2507 For example, an expression such as @samp{x[i,j]} will cause a warning,
2508 but @samp{x[(void)i,j]} will not.
2511 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2514 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2515 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2516 that of ordinary mathematical notation.
2519 Storage-class specifiers like @code{static} are not the first things in
2520 a declaration. According to the C Standard, this usage is obsolescent.
2523 The return type of a function has a type qualifier such as @code{const}.
2524 Such a type qualifier has no effect, since the value returned by a
2525 function is not an lvalue. (But don't warn about the GNU extension of
2526 @code{volatile void} return types. That extension will be warned about
2527 if @option{-pedantic} is specified.)
2530 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2534 A comparison between signed and unsigned values could produce an
2535 incorrect result when the signed value is converted to unsigned.
2536 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2539 An aggregate has an initializer which does not initialize all members.
2540 For example, the following code would cause such a warning, because
2541 @code{x.h} would be implicitly initialized to zero:
2544 struct s @{ int f, g, h; @};
2545 struct s x = @{ 3, 4 @};
2549 A function parameter is declared without a type specifier in K&R-style
2557 An empty body occurs in an @samp{if} or @samp{else} statement.
2560 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2561 @samp{>}, or @samp{>=}.
2564 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2567 Any of several floating-point events that often indicate errors, such as
2568 overflow, underflow, loss of precision, etc.
2570 @item @r{(C++ only)}
2571 An enumerator and a non-enumerator both appear in a conditional expression.
2573 @item @r{(C++ only)}
2574 A non-static reference or non-static @samp{const} member appears in a
2575 class without constructors.
2577 @item @r{(C++ only)}
2578 Ambiguous virtual bases.
2580 @item @r{(C++ only)}
2581 Subscripting an array which has been declared @samp{register}.
2583 @item @r{(C++ only)}
2584 Taking the address of a variable which has been declared @samp{register}.
2586 @item @r{(C++ only)}
2587 A base class is not initialized in a derived class' copy constructor.
2590 @item -Wno-div-by-zero
2591 @opindex Wno-div-by-zero
2592 @opindex Wdiv-by-zero
2593 Do not warn about compile-time integer division by zero. Floating point
2594 division by zero is not warned about, as it can be a legitimate way of
2595 obtaining infinities and NaNs.
2597 @item -Wsystem-headers
2598 @opindex Wsystem-headers
2599 @cindex warnings from system headers
2600 @cindex system headers, warnings from
2601 Print warning messages for constructs found in system header files.
2602 Warnings from system headers are normally suppressed, on the assumption
2603 that they usually do not indicate real problems and would only make the
2604 compiler output harder to read. Using this command line option tells
2605 GCC to emit warnings from system headers as if they occurred in user
2606 code. However, note that using @option{-Wall} in conjunction with this
2607 option will @emph{not} warn about unknown pragmas in system
2608 headers---for that, @option{-Wunknown-pragmas} must also be used.
2611 @opindex Wfloat-equal
2612 Warn if floating point values are used in equality comparisons.
2614 The idea behind this is that sometimes it is convenient (for the
2615 programmer) to consider floating-point values as approximations to
2616 infinitely precise real numbers. If you are doing this, then you need
2617 to compute (by analyzing the code, or in some other way) the maximum or
2618 likely maximum error that the computation introduces, and allow for it
2619 when performing comparisons (and when producing output, but that's a
2620 different problem). In particular, instead of testing for equality, you
2621 would check to see whether the two values have ranges that overlap; and
2622 this is done with the relational operators, so equality comparisons are
2625 @item -Wtraditional @r{(C only)}
2626 @opindex Wtraditional
2627 Warn about certain constructs that behave differently in traditional and
2628 ISO C@. Also warn about ISO C constructs that have no traditional C
2629 equivalent, and/or problematic constructs which should be avoided.
2633 Macro parameters that appear within string literals in the macro body.
2634 In traditional C macro replacement takes place within string literals,
2635 but does not in ISO C@.
2638 In traditional C, some preprocessor directives did not exist.
2639 Traditional preprocessors would only consider a line to be a directive
2640 if the @samp{#} appeared in column 1 on the line. Therefore
2641 @option{-Wtraditional} warns about directives that traditional C
2642 understands but would ignore because the @samp{#} does not appear as the
2643 first character on the line. It also suggests you hide directives like
2644 @samp{#pragma} not understood by traditional C by indenting them. Some
2645 traditional implementations would not recognize @samp{#elif}, so it
2646 suggests avoiding it altogether.
2649 A function-like macro that appears without arguments.
2652 The unary plus operator.
2655 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2656 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2657 constants.) Note, these suffixes appear in macros defined in the system
2658 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2659 Use of these macros in user code might normally lead to spurious
2660 warnings, however GCC's integrated preprocessor has enough context to
2661 avoid warning in these cases.
2664 A function declared external in one block and then used after the end of
2668 A @code{switch} statement has an operand of type @code{long}.
2671 A non-@code{static} function declaration follows a @code{static} one.
2672 This construct is not accepted by some traditional C compilers.
2675 The ISO type of an integer constant has a different width or
2676 signedness from its traditional type. This warning is only issued if
2677 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2678 typically represent bit patterns, are not warned about.
2681 Usage of ISO string concatenation is detected.
2684 Initialization of automatic aggregates.
2687 Identifier conflicts with labels. Traditional C lacks a separate
2688 namespace for labels.
2691 Initialization of unions. If the initializer is zero, the warning is
2692 omitted. This is done under the assumption that the zero initializer in
2693 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2694 initializer warnings and relies on default initialization to zero in the
2698 Conversions by prototypes between fixed/floating point values and vice
2699 versa. The absence of these prototypes when compiling with traditional
2700 C would cause serious problems. This is a subset of the possible
2701 conversion warnings, for the full set use @option{-Wconversion}.
2704 Use of ISO C style function definitions. This warning intentionally is
2705 @emph{not} issued for prototype declarations or variadic functions
2706 because these ISO C features will appear in your code when using
2707 libiberty's traditional C compatibility macros, @code{PARAMS} and
2708 @code{VPARAMS}. This warning is also bypassed for nested functions
2709 because that feature is already a GCC extension and thus not relevant to
2710 traditional C compatibility.
2713 @item -Wdeclaration-after-statement @r{(C only)}
2714 @opindex Wdeclaration-after-statement
2715 Warn when a declaration is found after a statement in a block. This
2716 construct, known from C++, was introduced with ISO C99 and is by default
2717 allowed in GCC@. It is not supported by ISO C90 and was not supported by
2718 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
2722 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2724 @item -Wendif-labels
2725 @opindex Wendif-labels
2726 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2730 Warn whenever a local variable shadows another local variable, parameter or
2731 global variable or whenever a built-in function is shadowed.
2733 @item -Wlarger-than-@var{len}
2734 @opindex Wlarger-than
2735 Warn whenever an object of larger than @var{len} bytes is defined.
2737 @item -Wpointer-arith
2738 @opindex Wpointer-arith
2739 Warn about anything that depends on the ``size of'' a function type or
2740 of @code{void}. GNU C assigns these types a size of 1, for
2741 convenience in calculations with @code{void *} pointers and pointers
2744 @item -Wbad-function-cast @r{(C only)}
2745 @opindex Wbad-function-cast
2746 Warn whenever a function call is cast to a non-matching type.
2747 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2751 Warn whenever a pointer is cast so as to remove a type qualifier from
2752 the target type. For example, warn if a @code{const char *} is cast
2753 to an ordinary @code{char *}.
2756 @opindex Wcast-align
2757 Warn whenever a pointer is cast such that the required alignment of the
2758 target is increased. For example, warn if a @code{char *} is cast to
2759 an @code{int *} on machines where integers can only be accessed at
2760 two- or four-byte boundaries.
2762 @item -Wwrite-strings
2763 @opindex Wwrite-strings
2764 When compiling C, give string constants the type @code{const
2765 char[@var{length}]} so that
2766 copying the address of one into a non-@code{const} @code{char *}
2767 pointer will get a warning; when compiling C++, warn about the
2768 deprecated conversion from string constants to @code{char *}.
2769 These warnings will help you find at
2770 compile time code that can try to write into a string constant, but
2771 only if you have been very careful about using @code{const} in
2772 declarations and prototypes. Otherwise, it will just be a nuisance;
2773 this is why we did not make @option{-Wall} request these warnings.
2776 @opindex Wconversion
2777 Warn if a prototype causes a type conversion that is different from what
2778 would happen to the same argument in the absence of a prototype. This
2779 includes conversions of fixed point to floating and vice versa, and
2780 conversions changing the width or signedness of a fixed point argument
2781 except when the same as the default promotion.
2783 Also, warn if a negative integer constant expression is implicitly
2784 converted to an unsigned type. For example, warn about the assignment
2785 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2786 casts like @code{(unsigned) -1}.
2788 @item -Wsign-compare
2789 @opindex Wsign-compare
2790 @cindex warning for comparison of signed and unsigned values
2791 @cindex comparison of signed and unsigned values, warning
2792 @cindex signed and unsigned values, comparison warning
2793 Warn when a comparison between signed and unsigned values could produce
2794 an incorrect result when the signed value is converted to unsigned.
2795 This warning is also enabled by @option{-Wextra}; to get the other warnings
2796 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
2798 @item -Waggregate-return
2799 @opindex Waggregate-return
2800 Warn if any functions that return structures or unions are defined or
2801 called. (In languages where you can return an array, this also elicits
2804 @item -Wstrict-prototypes @r{(C only)}
2805 @opindex Wstrict-prototypes
2806 Warn if a function is declared or defined without specifying the
2807 argument types. (An old-style function definition is permitted without
2808 a warning if preceded by a declaration which specifies the argument
2811 @item -Wold-style-definition @r{(C only)}
2812 @opindex Wold-style-definition
2813 Warn if an old-style function definition is used. A warning is given
2814 even if there is a previous prototype.
2816 @item -Wmissing-prototypes @r{(C only)}
2817 @opindex Wmissing-prototypes
2818 Warn if a global function is defined without a previous prototype
2819 declaration. This warning is issued even if the definition itself
2820 provides a prototype. The aim is to detect global functions that fail
2821 to be declared in header files.
2823 @item -Wmissing-declarations @r{(C only)}
2824 @opindex Wmissing-declarations
2825 Warn if a global function is defined without a previous declaration.
2826 Do so even if the definition itself provides a prototype.
2827 Use this option to detect global functions that are not declared in
2830 @item -Wmissing-noreturn
2831 @opindex Wmissing-noreturn
2832 Warn about functions which might be candidates for attribute @code{noreturn}.
2833 Note these are only possible candidates, not absolute ones. Care should
2834 be taken to manually verify functions actually do not ever return before
2835 adding the @code{noreturn} attribute, otherwise subtle code generation
2836 bugs could be introduced. You will not get a warning for @code{main} in
2837 hosted C environments.
2839 @item -Wmissing-format-attribute
2840 @opindex Wmissing-format-attribute
2842 If @option{-Wformat} is enabled, also warn about functions which might be
2843 candidates for @code{format} attributes. Note these are only possible
2844 candidates, not absolute ones. GCC will guess that @code{format}
2845 attributes might be appropriate for any function that calls a function
2846 like @code{vprintf} or @code{vscanf}, but this might not always be the
2847 case, and some functions for which @code{format} attributes are
2848 appropriate may not be detected. This option has no effect unless
2849 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2851 @item -Wno-multichar
2852 @opindex Wno-multichar
2854 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2855 Usually they indicate a typo in the user's code, as they have
2856 implementation-defined values, and should not be used in portable code.
2858 @item -Wno-deprecated-declarations
2859 @opindex Wno-deprecated-declarations
2860 Do not warn about uses of functions, variables, and types marked as
2861 deprecated by using the @code{deprecated} attribute.
2862 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2863 @pxref{Type Attributes}.)
2867 Warn if a structure is given the packed attribute, but the packed
2868 attribute has no effect on the layout or size of the structure.
2869 Such structures may be mis-aligned for little benefit. For
2870 instance, in this code, the variable @code{f.x} in @code{struct bar}
2871 will be misaligned even though @code{struct bar} does not itself
2872 have the packed attribute:
2879 @} __attribute__((packed));
2889 Warn if padding is included in a structure, either to align an element
2890 of the structure or to align the whole structure. Sometimes when this
2891 happens it is possible to rearrange the fields of the structure to
2892 reduce the padding and so make the structure smaller.
2894 @item -Wredundant-decls
2895 @opindex Wredundant-decls
2896 Warn if anything is declared more than once in the same scope, even in
2897 cases where multiple declaration is valid and changes nothing.
2899 @item -Wnested-externs @r{(C only)}
2900 @opindex Wnested-externs
2901 Warn if an @code{extern} declaration is encountered within a function.
2903 @item -Wunreachable-code
2904 @opindex Wunreachable-code
2905 Warn if the compiler detects that code will never be executed.
2907 This option is intended to warn when the compiler detects that at
2908 least a whole line of source code will never be executed, because
2909 some condition is never satisfied or because it is after a
2910 procedure that never returns.
2912 It is possible for this option to produce a warning even though there
2913 are circumstances under which part of the affected line can be executed,
2914 so care should be taken when removing apparently-unreachable code.
2916 For instance, when a function is inlined, a warning may mean that the
2917 line is unreachable in only one inlined copy of the function.
2919 This option is not made part of @option{-Wall} because in a debugging
2920 version of a program there is often substantial code which checks
2921 correct functioning of the program and is, hopefully, unreachable
2922 because the program does work. Another common use of unreachable
2923 code is to provide behavior which is selectable at compile-time.
2927 Warn if a function can not be inlined and it was declared as inline.
2928 Even with this option, the compiler will not warn about failures to
2929 inline functions declared in system headers.
2931 The compiler uses a variety of heuristics to determine whether or not
2932 to inline a function. For example, the compiler takes into account
2933 the size of the function being inlined and the the amount of inlining
2934 that has already been done in the current function. Therefore,
2935 seemingly insignificant changes in the source program can cause the
2936 warnings produced by @option{-Winline} to appear or disappear.
2938 @item -Wno-invalid-offsetof @r{(C++ only)}
2939 @opindex Wno-invalid-offsetof
2940 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
2941 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
2942 to a non-POD type is undefined. In existing C++ implementations,
2943 however, @samp{offsetof} typically gives meaningful results even when
2944 applied to certain kinds of non-POD types. (Such as a simple
2945 @samp{struct} that fails to be a POD type only by virtue of having a
2946 constructor.) This flag is for users who are aware that they are
2947 writing nonportable code and who have deliberately chosen to ignore the
2950 The restrictions on @samp{offsetof} may be relaxed in a future version
2951 of the C++ standard.
2954 @opindex Winvalid-pch
2955 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
2956 the search path but can't be used.
2960 @opindex Wno-long-long
2961 Warn if @samp{long long} type is used. This is default. To inhibit
2962 the warning messages, use @option{-Wno-long-long}. Flags
2963 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2964 only when @option{-pedantic} flag is used.
2966 @item -Wvariadic-macros
2967 @opindex Wvariadic-macros
2968 @opindex Wno-variadic-macros
2969 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
2970 alternate syntax when in pedantic ISO C99 mode. This is default.
2971 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
2973 @item -Wdisabled-optimization
2974 @opindex Wdisabled-optimization
2975 Warn if a requested optimization pass is disabled. This warning does
2976 not generally indicate that there is anything wrong with your code; it
2977 merely indicates that GCC's optimizers were unable to handle the code
2978 effectively. Often, the problem is that your code is too big or too
2979 complex; GCC will refuse to optimize programs when the optimization
2980 itself is likely to take inordinate amounts of time.
2984 Make all warnings into errors.
2987 @node Debugging Options
2988 @section Options for Debugging Your Program or GCC
2989 @cindex options, debugging
2990 @cindex debugging information options
2992 GCC has various special options that are used for debugging
2993 either your program or GCC:
2998 Produce debugging information in the operating system's native format
2999 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
3002 On most systems that use stabs format, @option{-g} enables use of extra
3003 debugging information that only GDB can use; this extra information
3004 makes debugging work better in GDB but will probably make other debuggers
3006 refuse to read the program. If you want to control for certain whether
3007 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3008 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3010 Unlike most other C compilers, GCC allows you to use @option{-g} with
3011 @option{-O}. The shortcuts taken by optimized code may occasionally
3012 produce surprising results: some variables you declared may not exist
3013 at all; flow of control may briefly move where you did not expect it;
3014 some statements may not be executed because they compute constant
3015 results or their values were already at hand; some statements may
3016 execute in different places because they were moved out of loops.
3018 Nevertheless it proves possible to debug optimized output. This makes
3019 it reasonable to use the optimizer for programs that might have bugs.
3021 The following options are useful when GCC is generated with the
3022 capability for more than one debugging format.
3026 Produce debugging information for use by GDB@. This means to use the
3027 most expressive format available (DWARF 2, stabs, or the native format
3028 if neither of those are supported), including GDB extensions if at all
3033 Produce debugging information in stabs format (if that is supported),
3034 without GDB extensions. This is the format used by DBX on most BSD
3035 systems. On MIPS, Alpha and System V Release 4 systems this option
3036 produces stabs debugging output which is not understood by DBX or SDB@.
3037 On System V Release 4 systems this option requires the GNU assembler.
3039 @item -feliminate-unused-debug-symbols
3040 @opindex feliminate-unused-debug-symbols
3041 Produce debugging information in stabs format (if that is supported),
3042 for only symbols that are actually used.
3046 Produce debugging information in stabs format (if that is supported),
3047 using GNU extensions understood only by the GNU debugger (GDB)@. The
3048 use of these extensions is likely to make other debuggers crash or
3049 refuse to read the program.
3053 Produce debugging information in COFF format (if that is supported).
3054 This is the format used by SDB on most System V systems prior to
3059 Produce debugging information in XCOFF format (if that is supported).
3060 This is the format used by the DBX debugger on IBM RS/6000 systems.
3064 Produce debugging information in XCOFF format (if that is supported),
3065 using GNU extensions understood only by the GNU debugger (GDB)@. The
3066 use of these extensions is likely to make other debuggers crash or
3067 refuse to read the program, and may cause assemblers other than the GNU
3068 assembler (GAS) to fail with an error.
3072 Produce debugging information in DWARF version 2 format (if that is
3073 supported). This is the format used by DBX on IRIX 6.
3077 Produce debugging information in VMS debug format (if that is
3078 supported). This is the format used by DEBUG on VMS systems.
3081 @itemx -ggdb@var{level}
3082 @itemx -gstabs@var{level}
3083 @itemx -gcoff@var{level}
3084 @itemx -gxcoff@var{level}
3085 @itemx -gvms@var{level}
3086 Request debugging information and also use @var{level} to specify how
3087 much information. The default level is 2.
3089 Level 1 produces minimal information, enough for making backtraces in
3090 parts of the program that you don't plan to debug. This includes
3091 descriptions of functions and external variables, but no information
3092 about local variables and no line numbers.
3094 Level 3 includes extra information, such as all the macro definitions
3095 present in the program. Some debuggers support macro expansion when
3096 you use @option{-g3}.
3098 Note that in order to avoid confusion between DWARF1 debug level 2,
3099 and DWARF2 @option{-gdwarf-2} does not accept a concatenated debug
3100 level. Instead use an additional @option{-g@var{level}} option to
3101 change the debug level for DWARF2.
3103 @item -feliminate-dwarf2-dups
3104 @opindex feliminate-dwarf2-dups
3105 Compress DWARF2 debugging information by eliminating duplicated
3106 information about each symbol. This option only makes sense when
3107 generating DWARF2 debugging information with @option{-gdwarf-2}.
3109 @cindex @command{prof}
3112 Generate extra code to write profile information suitable for the
3113 analysis program @command{prof}. You must use this option when compiling
3114 the source files you want data about, and you must also use it when
3117 @cindex @command{gprof}
3120 Generate extra code to write profile information suitable for the
3121 analysis program @command{gprof}. You must use this option when compiling
3122 the source files you want data about, and you must also use it when
3127 Makes the compiler print out each function name as it is compiled, and
3128 print some statistics about each pass when it finishes.
3131 @opindex ftime-report
3132 Makes the compiler print some statistics about the time consumed by each
3133 pass when it finishes.
3136 @opindex fmem-report
3137 Makes the compiler print some statistics about permanent memory
3138 allocation when it finishes.
3140 @item -fprofile-arcs
3141 @opindex fprofile-arcs
3142 Add code so that program flow @dfn{arcs} are instrumented. During
3143 execution the program records how many times each branch and call is
3144 executed and how many times it is taken or returns. When the compiled
3145 program exits it saves this data to a file called
3146 @file{@var{auxname}.gcda} for each source file. The data may be used for
3147 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3148 test coverage analysis (@option{-ftest-coverage}). Each object file's
3149 @var{auxname} is generated from the name of the output file, if
3150 explicitly specified and it is not the final executable, otherwise it is
3151 the basename of the source file. In both cases any suffix is removed
3152 (e.g. @file{foo.gcda} for input file @file{dir/foo.c}, or
3153 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3158 Compile the source files with @option{-fprofile-arcs} plus optimization
3159 and code generation options. For test coverage analysis, use the
3160 additional @option{-ftest-coverage} option. You do not need to profile
3161 every source file in a program.
3164 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3165 (the latter implies the former).
3168 Run the program on a representative workload to generate the arc profile
3169 information. This may be repeated any number of times. You can run
3170 concurrent instances of your program, and provided that the file system
3171 supports locking, the data files will be correctly updated. Also
3172 @code{fork} calls are detected and correctly handled (double counting
3176 For profile-directed optimizations, compile the source files again with
3177 the same optimization and code generation options plus
3178 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3179 Control Optimization}).
3182 For test coverage analysis, use @command{gcov} to produce human readable
3183 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3184 @command{gcov} documentation for further information.
3188 With @option{-fprofile-arcs}, for each function of your program GCC
3189 creates a program flow graph, then finds a spanning tree for the graph.
3190 Only arcs that are not on the spanning tree have to be instrumented: the
3191 compiler adds code to count the number of times that these arcs are
3192 executed. When an arc is the only exit or only entrance to a block, the
3193 instrumentation code can be added to the block; otherwise, a new basic
3194 block must be created to hold the instrumentation code.
3197 @item -ftest-coverage
3198 @opindex ftest-coverage
3199 Produce a notes file that the @command{gcov} code-coverage utility
3200 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3201 show program coverage. Each source file's note file is called
3202 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3203 above for a description of @var{auxname} and instructions on how to
3204 generate test coverage data. Coverage data will match the source files
3205 more closely, if you do not optimize.
3207 @item -d@var{letters}
3209 Says to make debugging dumps during compilation at times specified by
3210 @var{letters}. This is used for debugging the compiler. The file names
3211 for most of the dumps are made by appending a pass number and a word to
3212 the @var{dumpname}. @var{dumpname} is generated from the name of the
3213 output file, if explicitly specified and it is not an executable,
3214 otherwise it is the basename of the source file. In both cases any
3215 suffix is removed (e.g. @file{foo.01.rtl} or @file{foo.02.sibling}).
3216 Here are the possible letters for use in @var{letters}, and their
3222 Annotate the assembler output with miscellaneous debugging information.
3225 Dump after computing branch probabilities, to @file{@var{file}.12.bp}.
3228 Dump after block reordering, to @file{@var{file}.31.bbro}.
3231 Dump after instruction combination, to the file @file{@var{file}.20.combine}.
3234 Dump after the first if conversion, to the file @file{@var{file}.14.ce1}.
3235 Also dump after the second if conversion, to the file @file{@var{file}.21.ce2}.
3238 Dump after branch target load optimization, to to @file{@var{file}.32.btl}.
3239 Also dump after delayed branch scheduling, to @file{@var{file}.36.dbr}.
3242 Dump all macro definitions, at the end of preprocessing, in addition to
3246 Dump after the third if conversion, to @file{@var{file}.30.ce3}.
3249 Dump after control and data flow analysis, to @file{@var{file}.11.cfg}.
3250 Also dump after life analysis, to @file{@var{file}.19.life}.
3253 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.07.addressof}.
3256 Dump after global register allocation, to @file{@var{file}.25.greg}.
3259 Dump after GCSE, to @file{@var{file}.08.gcse}.
3260 Also dump after jump bypassing and control flow optimizations, to
3261 @file{@var{file}.10.bypass}.
3264 Dump after finalization of EH handling code, to @file{@var{file}.03.eh}.
3267 Dump after sibling call optimizations, to @file{@var{file}.02.sibling}.
3270 Dump after the first jump optimization, to @file{@var{file}.04.jump}.
3273 Dump after conversion from registers to stack, to @file{@var{file}.34.stack}.
3276 Dump after local register allocation, to @file{@var{file}.24.lreg}.
3279 Dump after loop optimization passes, to @file{@var{file}.09.loop} and
3280 @file{@var{file}.16.loop2}.
3283 Dump after performing the machine dependent reorganization pass, to
3284 @file{@var{file}.35.mach}.
3287 Dump after register renumbering, to @file{@var{file}.29.rnreg}.
3290 Dump after the register move pass, to @file{@var{file}.22.regmove}.
3293 Dump after post-reload optimizations, to @file{@var{file}.26.postreload}.
3296 Dump after RTL generation, to @file{@var{file}.01.rtl}.
3299 Dump after the second scheduling pass, to @file{@var{file}.33.sched2}.
3302 Dump after CSE (including the jump optimization that sometimes follows
3303 CSE), to @file{@var{file}.06.cse}.
3306 Dump after the first scheduling pass, to @file{@var{file}.23.sched}.
3309 Dump after the second CSE pass (including the jump optimization that
3310 sometimes follows CSE), to @file{@var{file}.18.cse2}.
3313 Dump after running tracer, to @file{@var{file}.15.tracer}.
3316 Dump after null pointer elimination pass to @file{@var{file}.05.null}.
3319 Dump callgraph and unit-at-a-time optimization @file{@var{file}.00.unit}.
3322 Dump after the value profile transformations, to @file{@var{file}.13.vpt}.
3323 Also dump after variable tracking, to @file{@var{file}.35.vartrack}.
3326 Dump after the second flow pass, to @file{@var{file}.27.flow2}.
3329 Dump after the peephole pass, to @file{@var{file}.28.peephole2}.
3332 Dump after constructing the web, to @file{@var{file}.17.web}.
3335 Produce all the dumps listed above.
3338 Produce a core dump whenever an error occurs.
3341 Print statistics on memory usage, at the end of the run, to
3345 Annotate the assembler output with a comment indicating which
3346 pattern and alternative was used. The length of each instruction is
3350 Dump the RTL in the assembler output as a comment before each instruction.
3351 Also turns on @option{-dp} annotation.
3354 For each of the other indicated dump files (except for
3355 @file{@var{file}.01.rtl}), dump a representation of the control flow graph
3356 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3359 Just generate RTL for a function instead of compiling it. Usually used
3363 Dump debugging information during parsing, to standard error.
3366 @item -fdump-unnumbered
3367 @opindex fdump-unnumbered
3368 When doing debugging dumps (see @option{-d} option above), suppress instruction
3369 numbers and line number note output. This makes it more feasible to
3370 use diff on debugging dumps for compiler invocations with different
3371 options, in particular with and without @option{-g}.
3373 @item -fdump-translation-unit @r{(C and C++ only)}
3374 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3375 @opindex fdump-translation-unit
3376 Dump a representation of the tree structure for the entire translation
3377 unit to a file. The file name is made by appending @file{.tu} to the
3378 source file name. If the @samp{-@var{options}} form is used, @var{options}
3379 controls the details of the dump as described for the
3380 @option{-fdump-tree} options.
3382 @item -fdump-class-hierarchy @r{(C++ only)}
3383 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3384 @opindex fdump-class-hierarchy
3385 Dump a representation of each class's hierarchy and virtual function
3386 table layout to a file. The file name is made by appending @file{.class}
3387 to the source file name. If the @samp{-@var{options}} form is used,
3388 @var{options} controls the details of the dump as described for the
3389 @option{-fdump-tree} options.
3391 @item -fdump-tree-@var{switch} @r{(C++ only)}
3392 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3394 Control the dumping at various stages of processing the intermediate
3395 language tree to a file. The file name is generated by appending a switch
3396 specific suffix to the source file name. If the @samp{-@var{options}}
3397 form is used, @var{options} is a list of @samp{-} separated options that
3398 control the details of the dump. Not all options are applicable to all
3399 dumps, those which are not meaningful will be ignored. The following
3400 options are available
3404 Print the address of each node. Usually this is not meaningful as it
3405 changes according to the environment and source file. Its primary use
3406 is for tying up a dump file with a debug environment.
3408 Inhibit dumping of members of a scope or body of a function merely
3409 because that scope has been reached. Only dump such items when they
3410 are directly reachable by some other path.
3412 Turn on all options.
3415 The following tree dumps are possible:
3418 Dump before any tree based optimization, to @file{@var{file}.original}.
3420 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3422 Dump after function inlining, to @file{@var{file}.inlined}.
3425 @item -frandom-seed=@var{string}
3426 @opindex frandom-string
3427 This option provides a seed that GCC uses when it would otherwise use
3428 random numbers. It is used to generate certain symbol names
3429 that have to be different in every compiled file. It is also used to
3430 place unique stamps in coverage data files and the object files that
3431 produce them. You can use the @option{-frandom-seed} option to produce
3432 reproducibly identical object files.
3434 The @var{string} should be different for every file you compile.
3436 @item -fsched-verbose=@var{n}
3437 @opindex fsched-verbose
3438 On targets that use instruction scheduling, this option controls the
3439 amount of debugging output the scheduler prints. This information is
3440 written to standard error, unless @option{-dS} or @option{-dR} is
3441 specified, in which case it is output to the usual dump
3442 listing file, @file{.sched} or @file{.sched2} respectively. However
3443 for @var{n} greater than nine, the output is always printed to standard
3446 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
3447 same information as @option{-dRS}. For @var{n} greater than one, it
3448 also output basic block probabilities, detailed ready list information
3449 and unit/insn info. For @var{n} greater than two, it includes RTL
3450 at abort point, control-flow and regions info. And for @var{n} over
3451 four, @option{-fsched-verbose} also includes dependence info.
3455 Store the usual ``temporary'' intermediate files permanently; place them
3456 in the current directory and name them based on the source file. Thus,
3457 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3458 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3459 preprocessed @file{foo.i} output file even though the compiler now
3460 normally uses an integrated preprocessor.
3464 Report the CPU time taken by each subprocess in the compilation
3465 sequence. For C source files, this is the compiler proper and assembler
3466 (plus the linker if linking is done). The output looks like this:
3473 The first number on each line is the ``user time,'' that is time spent
3474 executing the program itself. The second number is ``system time,''
3475 time spent executing operating system routines on behalf of the program.
3476 Both numbers are in seconds.
3478 @item -fvar-tracking
3479 @opindex fvar-tracking
3480 Run variable tracking pass. It computes where variables are stored at each
3481 position in code. Better debugging information is then generated
3482 (if the debugging information format supports this information).
3484 It is enabled by default when compiling with optimization (@option{-Os},
3485 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
3486 the debug info format supports it.
3488 @item -print-file-name=@var{library}
3489 @opindex print-file-name
3490 Print the full absolute name of the library file @var{library} that
3491 would be used when linking---and don't do anything else. With this
3492 option, GCC does not compile or link anything; it just prints the
3495 @item -print-multi-directory
3496 @opindex print-multi-directory
3497 Print the directory name corresponding to the multilib selected by any
3498 other switches present in the command line. This directory is supposed
3499 to exist in @env{GCC_EXEC_PREFIX}.
3501 @item -print-multi-lib
3502 @opindex print-multi-lib
3503 Print the mapping from multilib directory names to compiler switches
3504 that enable them. The directory name is separated from the switches by
3505 @samp{;}, and each switch starts with an @samp{@@} instead of the
3506 @samp{-}, without spaces between multiple switches. This is supposed to
3507 ease shell-processing.
3509 @item -print-prog-name=@var{program}
3510 @opindex print-prog-name
3511 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3513 @item -print-libgcc-file-name
3514 @opindex print-libgcc-file-name
3515 Same as @option{-print-file-name=libgcc.a}.
3517 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3518 but you do want to link with @file{libgcc.a}. You can do
3521 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3524 @item -print-search-dirs
3525 @opindex print-search-dirs
3526 Print the name of the configured installation directory and a list of
3527 program and library directories @command{gcc} will search---and don't do anything else.
3529 This is useful when @command{gcc} prints the error message
3530 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3531 To resolve this you either need to put @file{cpp0} and the other compiler
3532 components where @command{gcc} expects to find them, or you can set the environment
3533 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3534 Don't forget the trailing '/'.
3535 @xref{Environment Variables}.
3538 @opindex dumpmachine
3539 Print the compiler's target machine (for example,
3540 @samp{i686-pc-linux-gnu})---and don't do anything else.
3543 @opindex dumpversion
3544 Print the compiler version (for example, @samp{3.0})---and don't do
3549 Print the compiler's built-in specs---and don't do anything else. (This
3550 is used when GCC itself is being built.) @xref{Spec Files}.
3552 @item -feliminate-unused-debug-types
3553 @opindex feliminate-unused-debug-types
3554 Normally, when producing DWARF2 output, GCC will emit debugging
3555 information for all types declared in a compilation
3556 unit, regardless of whether or not they are actually used
3557 in that compilation unit. Sometimes this is useful, such as
3558 if, in the debugger, you want to cast a value to a type that is
3559 not actually used in your program (but is declared). More often,
3560 however, this results in a significant amount of wasted space.
3561 With this option, GCC will avoid producing debug symbol output
3562 for types that are nowhere used in the source file being compiled.
3565 @node Optimize Options
3566 @section Options That Control Optimization
3567 @cindex optimize options
3568 @cindex options, optimization
3570 These options control various sorts of optimizations.
3572 Without any optimization option, the compiler's goal is to reduce the
3573 cost of compilation and to make debugging produce the expected
3574 results. Statements are independent: if you stop the program with a
3575 breakpoint between statements, you can then assign a new value to any
3576 variable or change the program counter to any other statement in the
3577 function and get exactly the results you would expect from the source
3580 Turning on optimization flags makes the compiler attempt to improve
3581 the performance and/or code size at the expense of compilation time
3582 and possibly the ability to debug the program.
3584 The compiler performs optimization based on the knowledge it has of
3585 the program. Using the @option{-funit-at-a-time} flag will allow the
3586 compiler to consider information gained from later functions in the
3587 file when compiling a function. Compiling multiple files at once to a
3588 single output file (and using @option{-funit-at-a-time}) will allow
3589 the compiler to use information gained from all of the files when
3590 compiling each of them.
3592 Not all optimizations are controlled directly by a flag. Only
3593 optimizations that have a flag are listed.
3600 Optimize. Optimizing compilation takes somewhat more time, and a lot
3601 more memory for a large function.
3603 With @option{-O}, the compiler tries to reduce code size and execution
3604 time, without performing any optimizations that take a great deal of
3607 @option{-O} turns on the following optimization flags:
3608 @gccoptlist{-fdefer-pop @gol
3609 -fmerge-constants @gol
3611 -floop-optimize @gol
3612 -fif-conversion @gol
3613 -fif-conversion2 @gol
3614 -fdelayed-branch @gol
3615 -fguess-branch-probability @gol
3618 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
3619 where doing so does not interfere with debugging.
3623 Optimize even more. GCC performs nearly all supported optimizations
3624 that do not involve a space-speed tradeoff. The compiler does not
3625 perform loop unrolling or function inlining when you specify @option{-O2}.
3626 As compared to @option{-O}, this option increases both compilation time
3627 and the performance of the generated code.
3629 @option{-O2} turns on all optimization flags specified by @option{-O}. It
3630 also turns on the following optimization flags:
3631 @gccoptlist{-fforce-mem @gol
3632 -foptimize-sibling-calls @gol
3633 -fstrength-reduce @gol
3634 -fcse-follow-jumps -fcse-skip-blocks @gol
3635 -frerun-cse-after-loop -frerun-loop-opt @gol
3636 -fgcse -fgcse-lm -fgcse-sm -fgcse-las @gol
3637 -fdelete-null-pointer-checks @gol
3638 -fexpensive-optimizations @gol
3640 -fschedule-insns -fschedule-insns2 @gol
3641 -fsched-interblock -fsched-spec @gol
3644 -freorder-blocks -freorder-functions @gol
3645 -fstrict-aliasing @gol
3646 -funit-at-a-time @gol
3647 -falign-functions -falign-jumps @gol
3648 -falign-loops -falign-labels @gol
3651 Please note the warning under @option{-fgcse} about
3652 invoking @option{-O2} on programs that use computed gotos.
3656 Optimize yet more. @option{-O3} turns on all optimizations specified by
3657 @option{-O2} and also turns on the @option{-finline-functions},
3658 @option{-fweb} and @option{-frename-registers} options.
3662 Do not optimize. This is the default.
3666 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3667 do not typically increase code size. It also performs further
3668 optimizations designed to reduce code size.
3670 @option{-Os} disables the following optimization flags:
3671 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
3672 -falign-labels -freorder-blocks -fprefetch-loop-arrays}
3674 If you use multiple @option{-O} options, with or without level numbers,
3675 the last such option is the one that is effective.
3678 Options of the form @option{-f@var{flag}} specify machine-independent
3679 flags. Most flags have both positive and negative forms; the negative
3680 form of @option{-ffoo} would be @option{-fno-foo}. In the table
3681 below, only one of the forms is listed---the one you typically will
3682 use. You can figure out the other form by either removing @samp{no-}
3685 The following options control specific optimizations. They are either
3686 activated by @option{-O} options or are related to ones that are. You
3687 can use the following flags in the rare cases when ``fine-tuning'' of
3688 optimizations to be performed is desired.
3691 @item -fno-default-inline
3692 @opindex fno-default-inline
3693 Do not make member functions inline by default merely because they are
3694 defined inside the class scope (C++ only). Otherwise, when you specify
3695 @w{@option{-O}}, member functions defined inside class scope are compiled
3696 inline by default; i.e., you don't need to add @samp{inline} in front of
3697 the member function name.
3699 @item -fno-defer-pop
3700 @opindex fno-defer-pop
3701 Always pop the arguments to each function call as soon as that function
3702 returns. For machines which must pop arguments after a function call,
3703 the compiler normally lets arguments accumulate on the stack for several
3704 function calls and pops them all at once.
3706 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3710 Force memory operands to be copied into registers before doing
3711 arithmetic on them. This produces better code by making all memory
3712 references potential common subexpressions. When they are not common
3713 subexpressions, instruction combination should eliminate the separate
3716 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3719 @opindex fforce-addr
3720 Force memory address constants to be copied into registers before
3721 doing arithmetic on them. This may produce better code just as
3722 @option{-fforce-mem} may.
3724 @item -fomit-frame-pointer
3725 @opindex fomit-frame-pointer
3726 Don't keep the frame pointer in a register for functions that
3727 don't need one. This avoids the instructions to save, set up and
3728 restore frame pointers; it also makes an extra register available
3729 in many functions. @strong{It also makes debugging impossible on
3732 On some machines, such as the VAX, this flag has no effect, because
3733 the standard calling sequence automatically handles the frame pointer
3734 and nothing is saved by pretending it doesn't exist. The
3735 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3736 whether a target machine supports this flag. @xref{Registers,,Register
3737 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3739 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3741 @item -foptimize-sibling-calls
3742 @opindex foptimize-sibling-calls
3743 Optimize sibling and tail recursive calls.
3745 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3749 Don't pay attention to the @code{inline} keyword. Normally this option
3750 is used to keep the compiler from expanding any functions inline.
3751 Note that if you are not optimizing, no functions can be expanded inline.
3753 @item -finline-functions
3754 @opindex finline-functions
3755 Integrate all simple functions into their callers. The compiler
3756 heuristically decides which functions are simple enough to be worth
3757 integrating in this way.
3759 If all calls to a given function are integrated, and the function is
3760 declared @code{static}, then the function is normally not output as
3761 assembler code in its own right.
3763 Enabled at level @option{-O3}.
3765 @item -finline-limit=@var{n}
3766 @opindex finline-limit
3767 By default, GCC limits the size of functions that can be inlined. This flag
3768 allows the control of this limit for functions that are explicitly marked as
3769 inline (i.e., marked with the inline keyword or defined within the class
3770 definition in c++). @var{n} is the size of functions that can be inlined in
3771 number of pseudo instructions (not counting parameter handling). The default
3772 value of @var{n} is 600.
3773 Increasing this value can result in more inlined code at
3774 the cost of compilation time and memory consumption. Decreasing usually makes
3775 the compilation faster and less code will be inlined (which presumably
3776 means slower programs). This option is particularly useful for programs that
3777 use inlining heavily such as those based on recursive templates with C++.
3779 Inlining is actually controlled by a number of parameters, which may be
3780 specified individually by using @option{--param @var{name}=@var{value}}.
3781 The @option{-finline-limit=@var{n}} option sets some of these parameters
3785 @item max-inline-insns-single
3786 is set to @var{n}/2.
3787 @item max-inline-insns-auto
3788 is set to @var{n}/2.
3789 @item min-inline-insns
3790 is set to 130 or @var{n}/4, whichever is smaller.
3791 @item max-inline-insns-rtl
3795 See below for a documentation of the individual
3796 parameters controlling inlining.
3798 @emph{Note:} pseudo instruction represents, in this particular context, an
3799 abstract measurement of function's size. In no way, it represents a count
3800 of assembly instructions and as such its exact meaning might change from one
3801 release to an another.
3803 @item -fkeep-inline-functions
3804 @opindex fkeep-inline-functions
3805 Even if all calls to a given function are integrated, and the function
3806 is declared @code{static}, nevertheless output a separate run-time
3807 callable version of the function. This switch does not affect
3808 @code{extern inline} functions.
3810 @item -fkeep-static-consts
3811 @opindex fkeep-static-consts
3812 Emit variables declared @code{static const} when optimization isn't turned
3813 on, even if the variables aren't referenced.
3815 GCC enables this option by default. If you want to force the compiler to
3816 check if the variable was referenced, regardless of whether or not
3817 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3819 @item -fmerge-constants
3820 Attempt to merge identical constants (string constants and floating point
3821 constants) across compilation units.
3823 This option is the default for optimized compilation if the assembler and
3824 linker support it. Use @option{-fno-merge-constants} to inhibit this
3827 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3829 @item -fmerge-all-constants
3830 Attempt to merge identical constants and identical variables.
3832 This option implies @option{-fmerge-constants}. In addition to
3833 @option{-fmerge-constants} this considers e.g. even constant initialized
3834 arrays or initialized constant variables with integral or floating point
3835 types. Languages like C or C++ require each non-automatic variable to
3836 have distinct location, so using this option will result in non-conforming
3841 Use a graph coloring register allocator. Currently this option is meant
3842 only for testing. Users should not specify this option, since it is not
3843 yet ready for production use.
3845 @item -fno-branch-count-reg
3846 @opindex fno-branch-count-reg
3847 Do not use ``decrement and branch'' instructions on a count register,
3848 but instead generate a sequence of instructions that decrement a
3849 register, compare it against zero, then branch based upon the result.
3850 This option is only meaningful on architectures that support such
3851 instructions, which include x86, PowerPC, IA-64 and S/390.
3853 The default is @option{-fbranch-count-reg}, enabled when
3854 @option{-fstrength-reduce} is enabled.
3856 @item -fno-function-cse
3857 @opindex fno-function-cse
3858 Do not put function addresses in registers; make each instruction that
3859 calls a constant function contain the function's address explicitly.
3861 This option results in less efficient code, but some strange hacks
3862 that alter the assembler output may be confused by the optimizations
3863 performed when this option is not used.
3865 The default is @option{-ffunction-cse}
3867 @item -fno-zero-initialized-in-bss
3868 @opindex fno-zero-initialized-in-bss
3869 If the target supports a BSS section, GCC by default puts variables that
3870 are initialized to zero into BSS@. This can save space in the resulting
3873 This option turns off this behavior because some programs explicitly
3874 rely on variables going to the data section. E.g., so that the
3875 resulting executable can find the beginning of that section and/or make
3876 assumptions based on that.
3878 The default is @option{-fzero-initialized-in-bss}.
3880 @item -fstrength-reduce
3881 @opindex fstrength-reduce
3882 Perform the optimizations of loop strength reduction and
3883 elimination of iteration variables.
3885 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3887 @item -fthread-jumps
3888 @opindex fthread-jumps
3889 Perform optimizations where we check to see if a jump branches to a
3890 location where another comparison subsumed by the first is found. If
3891 so, the first branch is redirected to either the destination of the
3892 second branch or a point immediately following it, depending on whether
3893 the condition is known to be true or false.
3895 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3897 @item -fcse-follow-jumps
3898 @opindex fcse-follow-jumps
3899 In common subexpression elimination, scan through jump instructions
3900 when the target of the jump is not reached by any other path. For
3901 example, when CSE encounters an @code{if} statement with an
3902 @code{else} clause, CSE will follow the jump when the condition
3905 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3907 @item -fcse-skip-blocks
3908 @opindex fcse-skip-blocks
3909 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3910 follow jumps which conditionally skip over blocks. When CSE
3911 encounters a simple @code{if} statement with no else clause,
3912 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3913 body of the @code{if}.
3915 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3917 @item -frerun-cse-after-loop
3918 @opindex frerun-cse-after-loop
3919 Re-run common subexpression elimination after loop optimizations has been
3922 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3924 @item -frerun-loop-opt
3925 @opindex frerun-loop-opt
3926 Run the loop optimizer twice.
3928 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3932 Perform a global common subexpression elimination pass.
3933 This pass also performs global constant and copy propagation.
3935 @emph{Note:} When compiling a program using computed gotos, a GCC
3936 extension, you may get better runtime performance if you disable
3937 the global common subexpression elimination pass by adding
3938 @option{-fno-gcse} to the command line.
3940 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3944 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3945 attempt to move loads which are only killed by stores into themselves. This
3946 allows a loop containing a load/store sequence to be changed to a load outside
3947 the loop, and a copy/store within the loop.
3949 Enabled by default when gcse is enabled.
3953 When @option{-fgcse-sm} is enabled, a store motion pass is run after
3954 global common subexpression elimination. This pass will attempt to move
3955 stores out of loops. When used in conjunction with @option{-fgcse-lm},
3956 loops containing a load/store sequence can be changed to a load before
3957 the loop and a store after the loop.
3959 Enabled by default when gcse is enabled.
3963 When @option{-fgcse-las} is enabled, the global common subexpression
3964 elimination pass eliminates redundant loads that come after stores to the
3965 same memory location (both partial and full redundancies).
3967 Enabled by default when gcse is enabled.
3969 @item -floop-optimize
3970 @opindex floop-optimize
3971 Perform loop optimizations: move constant expressions out of loops, simplify
3972 exit test conditions and optionally do strength-reduction and loop unrolling as
3975 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3977 @item -fcrossjumping
3978 @opindex crossjumping
3979 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
3980 resulting code may or may not perform better than without cross-jumping.
3982 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3984 @item -fif-conversion
3985 @opindex if-conversion
3986 Attempt to transform conditional jumps into branch-less equivalents. This
3987 include use of conditional moves, min, max, set flags and abs instructions, and
3988 some tricks doable by standard arithmetics. The use of conditional execution
3989 on chips where it is available is controlled by @code{if-conversion2}.
3991 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3993 @item -fif-conversion2
3994 @opindex if-conversion2
3995 Use conditional execution (where available) to transform conditional jumps into
3996 branch-less equivalents.
3998 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4000 @item -fdelete-null-pointer-checks
4001 @opindex fdelete-null-pointer-checks
4002 Use global dataflow analysis to identify and eliminate useless checks
4003 for null pointers. The compiler assumes that dereferencing a null
4004 pointer would have halted the program. If a pointer is checked after
4005 it has already been dereferenced, it cannot be null.
4007 In some environments, this assumption is not true, and programs can
4008 safely dereference null pointers. Use
4009 @option{-fno-delete-null-pointer-checks} to disable this optimization
4010 for programs which depend on that behavior.
4012 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4014 @item -fexpensive-optimizations
4015 @opindex fexpensive-optimizations
4016 Perform a number of minor optimizations that are relatively expensive.
4018 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4020 @item -foptimize-register-move
4022 @opindex foptimize-register-move
4024 Attempt to reassign register numbers in move instructions and as
4025 operands of other simple instructions in order to maximize the amount of
4026 register tying. This is especially helpful on machines with two-operand
4029 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
4032 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4034 @item -fdelayed-branch
4035 @opindex fdelayed-branch
4036 If supported for the target machine, attempt to reorder instructions
4037 to exploit instruction slots available after delayed branch
4040 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4042 @item -fschedule-insns
4043 @opindex fschedule-insns
4044 If supported for the target machine, attempt to reorder instructions to
4045 eliminate execution stalls due to required data being unavailable. This
4046 helps machines that have slow floating point or memory load instructions
4047 by allowing other instructions to be issued until the result of the load
4048 or floating point instruction is required.
4050 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4052 @item -fschedule-insns2
4053 @opindex fschedule-insns2
4054 Similar to @option{-fschedule-insns}, but requests an additional pass of
4055 instruction scheduling after register allocation has been done. This is
4056 especially useful on machines with a relatively small number of
4057 registers and where memory load instructions take more than one cycle.
4059 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4061 @item -fno-sched-interblock
4062 @opindex fno-sched-interblock
4063 Don't schedule instructions across basic blocks. This is normally
4064 enabled by default when scheduling before register allocation, i.e.@:
4065 with @option{-fschedule-insns} or at @option{-O2} or higher.
4067 @item -fno-sched-spec
4068 @opindex fno-sched-spec
4069 Don't allow speculative motion of non-load instructions. This is normally
4070 enabled by default when scheduling before register allocation, i.e.@:
4071 with @option{-fschedule-insns} or at @option{-O2} or higher.
4073 @item -fsched-spec-load
4074 @opindex fsched-spec-load
4075 Allow speculative motion of some load instructions. This only makes
4076 sense when scheduling before register allocation, i.e.@: with
4077 @option{-fschedule-insns} or at @option{-O2} or higher.
4079 @item -fsched-spec-load-dangerous
4080 @opindex fsched-spec-load-dangerous
4081 Allow speculative motion of more load instructions. This only makes
4082 sense when scheduling before register allocation, i.e.@: with
4083 @option{-fschedule-insns} or at @option{-O2} or higher.
4085 @item -fsched-stalled-insns=@var{n}
4086 @opindex fsched-stalled-insns
4087 Define how many insns (if any) can be moved prematurely from the queue
4088 of stalled insns into the ready list, during the second scheduling pass.
4090 @item -fsched-stalled-insns-dep=@var{n}
4091 @opindex fsched-stalled-insns-dep
4092 Define how many insn groups (cycles) will be examined for a dependency
4093 on a stalled insn that is candidate for premature removal from the queue
4094 of stalled insns. Has an effect only during the second scheduling pass,
4095 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
4097 @item -fsched2-use-superblocks
4098 @opindex fsched2-use-superblocks
4099 When scheduling after register allocation, do use superblock scheduling
4100 algorithm. Superblock scheduling allows motion across basic block boundaries
4101 resulting on faster schedules. This option is experimental, as not all machine
4102 descriptions used by GCC model the CPU closely enough to avoid unreliable
4103 results from the algorithm.
4105 This only makes sense when scheduling after register allocation, i.e.@: with
4106 @option{-fschedule-insns2} or at @option{-O2} or higher.
4108 @item -fsched2-use-traces
4109 @opindex fsched2-use-traces
4110 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4111 allocation and additionally perform code duplication in order to increase the
4112 size of superblocks using tracer pass. See @option{-ftracer} for details on
4115 This mode should produce faster but significantly longer programs. Also
4116 without @code{-fbranch-probabilities} the traces constructed may not match the
4117 reality and hurt the performance. This only makes
4118 sense when scheduling after register allocation, i.e.@: with
4119 @option{-fschedule-insns2} or at @option{-O2} or higher.
4121 @item -fcaller-saves
4122 @opindex fcaller-saves
4123 Enable values to be allocated in registers that will be clobbered by
4124 function calls, by emitting extra instructions to save and restore the
4125 registers around such calls. Such allocation is done only when it
4126 seems to result in better code than would otherwise be produced.
4128 This option is always enabled by default on certain machines, usually
4129 those which have no call-preserved registers to use instead.
4131 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4133 @item -fmove-all-movables
4134 @opindex fmove-all-movables
4135 Forces all invariant computations in loops to be moved
4138 @item -freduce-all-givs
4139 @opindex freduce-all-givs
4140 Forces all general-induction variables in loops to be
4143 @emph{Note:} When compiling programs written in Fortran,
4144 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
4145 by default when you use the optimizer.
4147 These options may generate better or worse code; results are highly
4148 dependent on the structure of loops within the source code.
4150 These two options are intended to be removed someday, once
4151 they have helped determine the efficacy of various
4152 approaches to improving loop optimizations.
4154 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
4155 know how use of these options affects
4156 the performance of your production code.
4157 We're very interested in code that runs @emph{slower}
4158 when these options are @emph{enabled}.
4161 @itemx -fno-peephole2
4162 @opindex fno-peephole
4163 @opindex fno-peephole2
4164 Disable any machine-specific peephole optimizations. The difference
4165 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
4166 are implemented in the compiler; some targets use one, some use the
4167 other, a few use both.
4169 @option{-fpeephole} is enabled by default.
4170 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4172 @item -fno-guess-branch-probability
4173 @opindex fno-guess-branch-probability
4174 Do not guess branch probabilities using a randomized model.
4176 Sometimes GCC will opt to use a randomized model to guess branch
4177 probabilities, when none are available from either profiling feedback
4178 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
4179 different runs of the compiler on the same program may produce different
4182 In a hard real-time system, people don't want different runs of the
4183 compiler to produce code that has different behavior; minimizing
4184 non-determinism is of paramount import. This switch allows users to
4185 reduce non-determinism, possibly at the expense of inferior
4188 The default is @option{-fguess-branch-probability} at levels
4189 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4191 @item -freorder-blocks
4192 @opindex freorder-blocks
4193 Reorder basic blocks in the compiled function in order to reduce number of
4194 taken branches and improve code locality.
4196 Enabled at levels @option{-O2}, @option{-O3}.
4198 @item -freorder-functions
4199 @opindex freorder-functions
4200 Reorder basic blocks in the compiled function in order to reduce number of
4201 taken branches and improve code locality. This is implemented by using special
4202 subsections @code{text.hot} for most frequently executed functions and
4203 @code{text.unlikely} for unlikely executed functions. Reordering is done by
4204 the linker so object file format must support named sections and linker must
4205 place them in a reasonable way.
4207 Also profile feedback must be available in to make this option effective. See
4208 @option{-fprofile-arcs} for details.
4210 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4212 @item -fstrict-aliasing
4213 @opindex fstrict-aliasing
4214 Allows the compiler to assume the strictest aliasing rules applicable to
4215 the language being compiled. For C (and C++), this activates
4216 optimizations based on the type of expressions. In particular, an
4217 object of one type is assumed never to reside at the same address as an
4218 object of a different type, unless the types are almost the same. For
4219 example, an @code{unsigned int} can alias an @code{int}, but not a
4220 @code{void*} or a @code{double}. A character type may alias any other
4223 Pay special attention to code like this:
4236 The practice of reading from a different union member than the one most
4237 recently written to (called ``type-punning'') is common. Even with
4238 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
4239 is accessed through the union type. So, the code above will work as
4240 expected. However, this code might not:
4251 Every language that wishes to perform language-specific alias analysis
4252 should define a function that computes, given an @code{tree}
4253 node, an alias set for the node. Nodes in different alias sets are not
4254 allowed to alias. For an example, see the C front-end function
4255 @code{c_get_alias_set}.
4257 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4259 @item -falign-functions
4260 @itemx -falign-functions=@var{n}
4261 @opindex falign-functions
4262 Align the start of functions to the next power-of-two greater than
4263 @var{n}, skipping up to @var{n} bytes. For instance,
4264 @option{-falign-functions=32} aligns functions to the next 32-byte
4265 boundary, but @option{-falign-functions=24} would align to the next
4266 32-byte boundary only if this can be done by skipping 23 bytes or less.
4268 @option{-fno-align-functions} and @option{-falign-functions=1} are
4269 equivalent and mean that functions will not be aligned.
4271 Some assemblers only support this flag when @var{n} is a power of two;
4272 in that case, it is rounded up.
4274 If @var{n} is not specified or is zero, use a machine-dependent default.
4276 Enabled at levels @option{-O2}, @option{-O3}.
4278 @item -falign-labels
4279 @itemx -falign-labels=@var{n}
4280 @opindex falign-labels
4281 Align all branch targets to a power-of-two boundary, skipping up to
4282 @var{n} bytes like @option{-falign-functions}. This option can easily
4283 make code slower, because it must insert dummy operations for when the
4284 branch target is reached in the usual flow of the code.
4286 @option{-fno-align-labels} and @option{-falign-labels=1} are
4287 equivalent and mean that labels will not be aligned.
4289 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4290 are greater than this value, then their values are used instead.
4292 If @var{n} is not specified or is zero, use a machine-dependent default
4293 which is very likely to be @samp{1}, meaning no alignment.
4295 Enabled at levels @option{-O2}, @option{-O3}.
4298 @itemx -falign-loops=@var{n}
4299 @opindex falign-loops
4300 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4301 like @option{-falign-functions}. The hope is that the loop will be
4302 executed many times, which will make up for any execution of the dummy
4305 @option{-fno-align-loops} and @option{-falign-loops=1} are
4306 equivalent and mean that loops will not be aligned.
4308 If @var{n} is not specified or is zero, use a machine-dependent default.
4310 Enabled at levels @option{-O2}, @option{-O3}.
4313 @itemx -falign-jumps=@var{n}
4314 @opindex falign-jumps
4315 Align branch targets to a power-of-two boundary, for branch targets
4316 where the targets can only be reached by jumping, skipping up to @var{n}
4317 bytes like @option{-falign-functions}. In this case, no dummy operations
4320 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
4321 equivalent and mean that loops will not be aligned.
4323 If @var{n} is not specified or is zero, use a machine-dependent default.
4325 Enabled at levels @option{-O2}, @option{-O3}.
4327 @item -frename-registers
4328 @opindex frename-registers
4329 Attempt to avoid false dependencies in scheduled code by making use
4330 of registers left over after register allocation. This optimization
4331 will most benefit processors with lots of registers. It can, however,
4332 make debugging impossible, since variables will no longer stay in
4333 a ``home register''.
4335 Enabled at levels @option{-O3}.
4339 Constructs webs as commonly used for register allocation purposes and assign
4340 each web individual pseudo register. This allows our register allocation pass
4341 to operate on pseudos directly, but also strengthens several other optimization
4342 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
4343 however, make debugging impossible, since variables will no longer stay in a
4346 Enabled at levels @option{-O3}.
4348 @item -fno-cprop-registers
4349 @opindex fno-cprop-registers
4350 After register allocation and post-register allocation instruction splitting,
4351 we perform a copy-propagation pass to try to reduce scheduling dependencies
4352 and occasionally eliminate the copy.
4354 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4356 @item -fprofile-generate
4357 @opindex fprofile-generate
4358 Enable options usually used for instrumenting application to produce profile usefull
4359 for later recompilation profile feedback based optimization.
4361 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}
4364 @opindex fprofile-use
4365 Enable profile feedback directed optimizations, and optimizations
4366 generally profitable only with profile feedback available.
4368 The following options are enabled: @code{-fbranch-probabilities},
4369 @code{-fvpt}, @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}.
4373 The following options control compiler behavior regarding floating
4374 point arithmetic. These options trade off between speed and
4375 correctness. All must be specifically enabled.
4379 @opindex ffloat-store
4380 Do not store floating point variables in registers, and inhibit other
4381 options that might change whether a floating point value is taken from a
4384 @cindex floating point precision
4385 This option prevents undesirable excess precision on machines such as
4386 the 68000 where the floating registers (of the 68881) keep more
4387 precision than a @code{double} is supposed to have. Similarly for the
4388 x86 architecture. For most programs, the excess precision does only
4389 good, but a few programs rely on the precise definition of IEEE floating
4390 point. Use @option{-ffloat-store} for such programs, after modifying
4391 them to store all pertinent intermediate computations into variables.
4395 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
4396 @option{-fno-trapping-math}, @option{-ffinite-math-only},
4397 @option{-fno-rounding-math} and @option{-fno-signaling-nans}.
4399 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
4401 This option should never be turned on by any @option{-O} option since
4402 it can result in incorrect output for programs which depend on
4403 an exact implementation of IEEE or ISO rules/specifications for
4406 @item -fno-math-errno
4407 @opindex fno-math-errno
4408 Do not set ERRNO after calling math functions that are executed
4409 with a single instruction, e.g., sqrt. A program that relies on
4410 IEEE exceptions for math error handling may want to use this flag
4411 for speed while maintaining IEEE arithmetic compatibility.
4413 This option should never be turned on by any @option{-O} option since
4414 it can result in incorrect output for programs which depend on
4415 an exact implementation of IEEE or ISO rules/specifications for
4418 The default is @option{-fmath-errno}.
4420 @item -funsafe-math-optimizations
4421 @opindex funsafe-math-optimizations
4422 Allow optimizations for floating-point arithmetic that (a) assume
4423 that arguments and results are valid and (b) may violate IEEE or
4424 ANSI standards. When used at link-time, it may include libraries
4425 or startup files that change the default FPU control word or other
4426 similar optimizations.
4428 This option should never be turned on by any @option{-O} option since
4429 it can result in incorrect output for programs which depend on
4430 an exact implementation of IEEE or ISO rules/specifications for
4433 The default is @option{-fno-unsafe-math-optimizations}.
4435 @item -ffinite-math-only
4436 @opindex ffinite-math-only
4437 Allow optimizations for floating-point arithmetic that assume
4438 that arguments and results are not NaNs or +-Infs.
4440 This option should never be turned on by any @option{-O} option since
4441 it can result in incorrect output for programs which depend on
4442 an exact implementation of IEEE or ISO rules/specifications.
4444 The default is @option{-fno-finite-math-only}.
4446 @item -fno-trapping-math
4447 @opindex fno-trapping-math
4448 Compile code assuming that floating-point operations cannot generate
4449 user-visible traps. These traps include division by zero, overflow,
4450 underflow, inexact result and invalid operation. This option implies
4451 @option{-fno-signaling-nans}. Setting this option may allow faster
4452 code if one relies on ``non-stop'' IEEE arithmetic, for example.
4454 This option should never be turned on by any @option{-O} option since
4455 it can result in incorrect output for programs which depend on
4456 an exact implementation of IEEE or ISO rules/specifications for
4459 The default is @option{-ftrapping-math}.
4461 @item -frounding-math
4462 @opindex frounding-math
4463 Disable transformations and optimizations that assume default floating
4464 point rounding behavior. This is round-to-zero for all floating point
4465 to integer conversions, and round-to-nearest for all other arithmetic
4466 truncations. This option should be specified for programs that change
4467 the FP rounding mode dynamically, or that may be executed with a
4468 non-default rounding mode. This option disables constant folding of
4469 floating point expressions at compile-time (which may be affected by
4470 rounding mode) and arithmetic transformations that are unsafe in the
4471 presence of sign-dependent rounding modes.
4473 The default is @option{-fno-rounding-math}.
4475 This option is experimental and does not currently guarantee to
4476 disable all GCC optimizations that are affected by rounding mode.
4477 Future versions of GCC may provide finer control of this setting
4478 using C99's @code{FENV_ACCESS} pragma. This command line option
4479 will be used to specify the default state for @code{FENV_ACCESS}.
4481 @item -fsignaling-nans
4482 @opindex fsignaling-nans
4483 Compile code assuming that IEEE signaling NaNs may generate user-visible
4484 traps during floating-point operations. Setting this option disables
4485 optimizations that may change the number of exceptions visible with
4486 signaling NaNs. This option implies @option{-ftrapping-math}.
4488 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
4491 The default is @option{-fno-signaling-nans}.
4493 This option is experimental and does not currently guarantee to
4494 disable all GCC optimizations that affect signaling NaN behavior.
4496 @item -fsingle-precision-constant
4497 @opindex fsingle-precision-constant
4498 Treat floating point constant as single precision constant instead of
4499 implicitly converting it to double precision constant.
4504 The following options control optimizations that may improve
4505 performance, but are not enabled by any @option{-O} options. This
4506 section includes experimental options that may produce broken code.
4509 @item -fbranch-probabilities
4510 @opindex fbranch-probabilities
4511 After running a program compiled with @option{-fprofile-arcs}
4512 (@pxref{Debugging Options,, Options for Debugging Your Program or
4513 @command{gcc}}), you can compile it a second time using
4514 @option{-fbranch-probabilities}, to improve optimizations based on
4515 the number of times each branch was taken. When the program
4516 compiled with @option{-fprofile-arcs} exits it saves arc execution
4517 counts to a file called @file{@var{sourcename}.gcda} for each source
4518 file The information in this data file is very dependent on the
4519 structure of the generated code, so you must use the same source code
4520 and the same optimization options for both compilations.
4522 With @option{-fbranch-probabilities}, GCC puts a
4523 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
4524 These can be used to improve optimization. Currently, they are only
4525 used in one place: in @file{reorg.c}, instead of guessing which path a
4526 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
4527 exactly determine which path is taken more often.
4529 @item -fprofile-values
4530 @opindex fprofile-values
4531 If combined with @option{-fprofile-arcs}, it adds code so that some
4532 data about values of expressions in the program is gathered.
4534 With @option{-fbranch-probabilities}, it reads back the data gathered
4535 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
4536 notes to instructions for their later usage in optimizations.
4538 Enabled with @option{-profile-generate} and @option{-profile-use}.
4542 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
4543 a code to gather information about values of expressions.
4545 With @option{-fbranch-probabilities}, it reads back the data gathered
4546 and actually performs the optimizations based on them.
4547 Currently the optimizations include specialization of division operation
4548 using the knowledge about the value of the denominator.
4550 Enabled with @option{-profile-generate} and @option{-profile-use}.
4554 Use a graph coloring register allocator. Currently this option is meant
4555 for testing, so we are interested to hear about miscompilations with
4560 Perform tail duplication to enlarge superblock size. This transformation
4561 simplifies the control flow of the function allowing other optimizations to do
4564 Enabled with @option{-profile-use}.
4566 @item -funit-at-a-time
4567 @opindex funit-at-a-time
4568 Parse the whole compilation unit before starting to produce code.
4569 This allows some extra optimizations to take place but consumes more
4572 Enabled at levels @option{-O2}, @option{-O3}.
4574 @item -funroll-loops
4575 @opindex funroll-loops
4576 Unroll loops whose number of iterations can be determined at compile time or
4577 upon entry to the loop. @option{-funroll-loops} implies
4578 @option{-frerun-cse-after-loop}. It also turns on complete loop peeling
4579 (i.e. complete removal of loops with small constant number of iterations).
4580 This option makes code larger, and may or may not make it run faster.
4582 Enabled with @option{-profile-use}.
4584 @item -funroll-all-loops
4585 @opindex funroll-all-loops
4586 Unroll all loops, even if their number of iterations is uncertain when
4587 the loop is entered. This usually makes programs run more slowly.
4588 @option{-funroll-all-loops} implies the same options as
4589 @option{-funroll-loops}.
4592 @opindex fpeel-loops
4593 Peels the loops for that there is enough information that they do not
4594 roll much (from profile feedback). It also turns on complete loop peeling
4595 (i.e. complete removal of loops with small constant number of iterations).
4597 Enabled with @option{-profile-use}.
4599 @item -funswitch-loops
4600 @opindex funswitch-loops
4601 Move branches with loop invariant conditions out of the loop, with duplicates
4602 of the loop on both branches (modified according to result of the condition).
4604 @item -fold-unroll-loops
4605 @opindex fold-unroll-loops
4606 Unroll loops whose number of iterations can be determined at compile
4607 time or upon entry to the loop, using the old loop unroller whose loop
4608 recognition is based on notes from frontend. @option{-fold-unroll-loops} implies
4609 both @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
4610 option makes code larger, and may or may not make it run faster.
4612 @item -fold-unroll-all-loops
4613 @opindex fold-unroll-all-loops
4614 Unroll all loops, even if their number of iterations is uncertain when
4615 the loop is entered. This is done using the old loop unroller whose loop
4616 recognition is based on notes from frontend. This usually makes programs run more slowly.
4617 @option{-fold-unroll-all-loops} implies the same options as
4618 @option{-fold-unroll-loops}.
4620 @item -fprefetch-loop-arrays
4621 @opindex fprefetch-loop-arrays
4622 If supported by the target machine, generate instructions to prefetch
4623 memory to improve the performance of loops that access large arrays.
4625 Disabled at level @option{-Os}.
4627 @item -ffunction-sections
4628 @itemx -fdata-sections
4629 @opindex ffunction-sections
4630 @opindex fdata-sections
4631 Place each function or data item into its own section in the output
4632 file if the target supports arbitrary sections. The name of the
4633 function or the name of the data item determines the section's name
4636 Use these options on systems where the linker can perform optimizations
4637 to improve locality of reference in the instruction space. Most systems
4638 using the ELF object format and SPARC processors running Solaris 2 have
4639 linkers with such optimizations. AIX may have these optimizations in
4642 Only use these options when there are significant benefits from doing
4643 so. When you specify these options, the assembler and linker will
4644 create larger object and executable files and will also be slower.
4645 You will not be able to use @code{gprof} on all systems if you
4646 specify this option and you may have problems with debugging if
4647 you specify both this option and @option{-g}.
4649 @item -fbranch-target-load-optimize
4650 @opindex fbranch-target-load-optimize
4651 Perform branch target register load optimization before prologue / epilogue
4653 The use of target registers can typically be exposed only during reload,
4654 thus hoisting loads out of loops and doing inter-block scheduling needs
4655 a separate optimization pass.
4657 @item -fbranch-target-load-optimize2
4658 @opindex fbranch-target-load-optimize2
4659 Perform branch target register load optimization after prologue / epilogue
4662 @item -fbtr-bb-exclusive
4663 @opindex fbtr-bb-exclusive
4664 When performing branch target register load optimization, don't reuse
4665 branch target registers in within any basic block.
4667 @item --param @var{name}=@var{value}
4669 In some places, GCC uses various constants to control the amount of
4670 optimization that is done. For example, GCC will not inline functions
4671 that contain more that a certain number of instructions. You can
4672 control some of these constants on the command-line using the
4673 @option{--param} option.
4675 The names of specific parameters, and the meaning of the values, are
4676 tied to the internals of the compiler, and are subject to change
4677 without notice in future releases.
4679 In each case, the @var{value} is an integer. The allowable choices for
4680 @var{name} are given in the following table:
4683 @item max-crossjump-edges
4684 The maximum number of incoming edges to consider for crossjumping.
4685 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
4686 the number of edges incoming to each block. Increasing values mean
4687 more aggressive optimization, making the compile time increase with
4688 probably small improvement in executable size.
4690 @item max-delay-slot-insn-search
4691 The maximum number of instructions to consider when looking for an
4692 instruction to fill a delay slot. If more than this arbitrary number of
4693 instructions is searched, the time savings from filling the delay slot
4694 will be minimal so stop searching. Increasing values mean more
4695 aggressive optimization, making the compile time increase with probably
4696 small improvement in executable run time.
4698 @item max-delay-slot-live-search
4699 When trying to fill delay slots, the maximum number of instructions to
4700 consider when searching for a block with valid live register
4701 information. Increasing this arbitrarily chosen value means more
4702 aggressive optimization, increasing the compile time. This parameter
4703 should be removed when the delay slot code is rewritten to maintain the
4706 @item max-gcse-memory
4707 The approximate maximum amount of memory that will be allocated in
4708 order to perform the global common subexpression elimination
4709 optimization. If more memory than specified is required, the
4710 optimization will not be done.
4712 @item max-gcse-passes
4713 The maximum number of passes of GCSE to run. The default is 1.
4715 @item max-pending-list-length
4716 The maximum number of pending dependencies scheduling will allow
4717 before flushing the current state and starting over. Large functions
4718 with few branches or calls can create excessively large lists which
4719 needlessly consume memory and resources.
4721 @item max-inline-insns-single
4722 Several parameters control the tree inliner used in gcc.
4723 This number sets the maximum number of instructions (counted in GCC's
4724 internal representation) in a single function that the tree inliner
4725 will consider for inlining. This only affects functions declared
4726 inline and methods implemented in a class declaration (C++).
4727 The default value is 500.
4729 @item max-inline-insns-auto
4730 When you use @option{-finline-functions} (included in @option{-O3}),
4731 a lot of functions that would otherwise not be considered for inlining
4732 by the compiler will be investigated. To those functions, a different
4733 (more restrictive) limit compared to functions declared inline can
4735 The default value is 120.
4737 @item large-function-insns
4738 The limit specifying really large functions. For functions greater than this
4739 limit inlining is constrained by @option{--param large-function-growth}.
4740 This parameter is useful primarily to avoid extreme compilation time caused by non-linear
4741 algorithms used by the backend.
4742 This parameter is ignored when @option{-funit-at-a-time} is not used.
4743 The default value is 3000.
4745 @item large-function-growth
4746 Specifies maximal growth of large function caused by inlining in percents.
4747 This parameter is ignored when @option{-funit-at-a-time} is not used.
4748 The default value is 200.
4750 @item inline-unit-growth
4751 Specifies maximal overall growth of the compilation unit caused by inlining.
4752 This parameter is ignored when @option{-funit-at-a-time} is not used.
4753 The default value is 150.
4755 @item max-inline-insns-rtl
4756 For languages that use the RTL inliner (this happens at a later stage
4757 than tree inlining), you can set the maximum allowable size (counted
4758 in RTL instructions) for the RTL inliner with this parameter.
4759 The default value is 600.
4761 @item max-unrolled-insns
4762 The maximum number of instructions that a loop should have if that loop
4763 is unrolled, and if the loop is unrolled, it determines how many times
4764 the loop code is unrolled.
4766 @item max-average-unrolled-insns
4767 The maximum number of instructions biased by probabilities of their execution
4768 that a loop should have if that loop is unrolled, and if the loop is unrolled,
4769 it determines how many times the loop code is unrolled.
4771 @item max-unroll-times
4772 The maximum number of unrollings of a single loop.
4774 @item max-peeled-insns
4775 The maximum number of instructions that a loop should have if that loop
4776 is peeled, and if the loop is peeled, it determines how many times
4777 the loop code is peeled.
4779 @item max-peel-times
4780 The maximum number of peelings of a single loop.
4782 @item max-completely-peeled-insns
4783 The maximum number of insns of a completely peeled loop.
4785 @item max-completely-peel-times
4786 The maximum number of iterations of a loop to be suitable for complete peeling.
4788 @item max-unswitch-insns
4789 The maximum number of insns of an unswitched loop.
4791 @item max-unswitch-level
4792 The maximum number of branches unswitched in a single loop.
4794 @item hot-bb-count-fraction
4795 Select fraction of the maximal count of repetitions of basic block in program
4796 given basic block needs to have to be considered hot.
4798 @item hot-bb-frequency-fraction
4799 Select fraction of the maximal frequency of executions of basic block in
4800 function given basic block needs to have to be considered hot
4802 @item tracer-dynamic-coverage
4803 @itemx tracer-dynamic-coverage-feedback
4805 This value is used to limit superblock formation once the given percentage of
4806 executed instructions is covered. This limits unnecessary code size
4809 The @option{tracer-dynamic-coverage-feedback} is used only when profile
4810 feedback is available. The real profiles (as opposed to statically estimated
4811 ones) are much less balanced allowing the threshold to be larger value.
4813 @item tracer-max-code-growth
4814 Stop tail duplication once code growth has reached given percentage. This is
4815 rather hokey argument, as most of the duplicates will be eliminated later in
4816 cross jumping, so it may be set to much higher values than is the desired code
4819 @item tracer-min-branch-ratio
4821 Stop reverse growth when the reverse probability of best edge is less than this
4822 threshold (in percent).
4824 @item tracer-min-branch-ratio
4825 @itemx tracer-min-branch-ratio-feedback
4827 Stop forward growth if the best edge do have probability lower than this
4830 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
4831 compilation for profile feedback and one for compilation without. The value
4832 for compilation with profile feedback needs to be more conservative (higher) in
4833 order to make tracer effective.
4835 @item max-cse-path-length
4837 Maximum number of basic blocks on path that cse considers. The default is 10.
4839 @item ggc-min-expand
4841 GCC uses a garbage collector to manage its own memory allocation. This
4842 parameter specifies the minimum percentage by which the garbage
4843 collector's heap should be allowed to expand between collections.
4844 Tuning this may improve compilation speed; it has no effect on code
4847 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
4848 RAM >= 1GB. If @code{getrlimit} is available, the notion of "RAM" is
4849 the smallest of actual RAM, RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If
4850 GCC is not able to calculate RAM on a particular platform, the lower
4851 bound of 30% is used. Setting this parameter and
4852 @option{ggc-min-heapsize} to zero causes a full collection to occur at
4853 every opportunity. This is extremely slow, but can be useful for
4856 @item ggc-min-heapsize
4858 Minimum size of the garbage collector's heap before it begins bothering
4859 to collect garbage. The first collection occurs after the heap expands
4860 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
4861 tuning this may improve compilation speed, and has no effect on code
4864 The default is RAM/8, with a lower bound of 4096 (four megabytes) and an
4865 upper bound of 131072 (128 megabytes). If @code{getrlimit} is
4866 available, the notion of "RAM" is the smallest of actual RAM,
4867 RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If GCC is not able to calculate
4868 RAM on a particular platform, the lower bound is used. Setting this
4869 parameter very large effectively disables garbage collection. Setting
4870 this parameter and @option{ggc-min-expand} to zero causes a full
4871 collection to occur at every opportunity.
4873 @item max-reload-search-insns
4874 The maximum number of instruction reload should look backward for equivalent
4875 register. Increasing values mean more aggressive optimization, making the
4876 compile time increase with probably slightly better performance. The default
4879 @item max-cselib-memory-location
4880 The maximum number of memory locations cselib should take into acount.
4881 Increasing values mean more aggressive optimization, making the compile time
4882 increase with probably slightly better performance. The default value is 500.
4884 @item reorder-blocks-duplicate
4885 @itemx reorder-blocks-duplicate-feedback
4887 Used by basic block reordering pass to decide whether to use unconditional
4888 branch or duplicate the code on its destination. Code is duplicated when its
4889 estimated size is smaller than this value multiplied by the estimated size of
4890 unconditional jump in the hot spots of the program.
4892 The @option{reorder-block-duplicate-feedback} is used only when profile
4893 feedback is available and may be set to higher values than
4894 @option{reorder-block-duplicate} since information about the hot spots is more
4897 @item max-sched-region-blocks
4898 The maximum number of blocks in a region to be considered for
4899 interblock scheduling. The default value is 10.
4901 @item max-sched-region-insns
4902 The maximum number of insns in a region to be considered for
4903 interblock scheduling. The default value is 100.
4907 @node Preprocessor Options
4908 @section Options Controlling the Preprocessor
4909 @cindex preprocessor options
4910 @cindex options, preprocessor
4912 These options control the C preprocessor, which is run on each C source
4913 file before actual compilation.
4915 If you use the @option{-E} option, nothing is done except preprocessing.
4916 Some of these options make sense only together with @option{-E} because
4917 they cause the preprocessor output to be unsuitable for actual
4922 You can use @option{-Wp,@var{option}} to bypass the compiler driver
4923 and pass @var{option} directly through to the preprocessor. If
4924 @var{option} contains commas, it is split into multiple options at the
4925 commas. However, many options are modified, translated or interpreted
4926 by the compiler driver before being passed to the preprocessor, and
4927 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
4928 interface is undocumented and subject to change, so whenever possible
4929 you should avoid using @option{-Wp} and let the driver handle the
4932 @item -Xpreprocessor @var{option}
4933 @opindex preprocessor
4934 Pass @var{option} as an option to the preprocessor. You can use this to
4935 supply system-specific preprocessor options which GCC does not know how to
4938 If you want to pass an option that takes an argument, you must use
4939 @option{-Xpreprocessor} twice, once for the option and once for the argument.
4942 @include cppopts.texi
4944 @node Assembler Options
4945 @section Passing Options to the Assembler
4947 @c prevent bad page break with this line
4948 You can pass options to the assembler.
4951 @item -Wa,@var{option}
4953 Pass @var{option} as an option to the assembler. If @var{option}
4954 contains commas, it is split into multiple options at the commas.
4956 @item -Xassembler @var{option}
4958 Pass @var{option} as an option to the assembler. You can use this to
4959 supply system-specific assembler options which GCC does not know how to
4962 If you want to pass an option that takes an argument, you must use
4963 @option{-Xassembler} twice, once for the option and once for the argument.
4968 @section Options for Linking
4969 @cindex link options
4970 @cindex options, linking
4972 These options come into play when the compiler links object files into
4973 an executable output file. They are meaningless if the compiler is
4974 not doing a link step.
4978 @item @var{object-file-name}
4979 A file name that does not end in a special recognized suffix is
4980 considered to name an object file or library. (Object files are
4981 distinguished from libraries by the linker according to the file
4982 contents.) If linking is done, these object files are used as input
4991 If any of these options is used, then the linker is not run, and
4992 object file names should not be used as arguments. @xref{Overall
4996 @item -l@var{library}
4997 @itemx -l @var{library}
4999 Search the library named @var{library} when linking. (The second
5000 alternative with the library as a separate argument is only for
5001 POSIX compliance and is not recommended.)
5003 It makes a difference where in the command you write this option; the
5004 linker searches and processes libraries and object files in the order they
5005 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
5006 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
5007 to functions in @samp{z}, those functions may not be loaded.
5009 The linker searches a standard list of directories for the library,
5010 which is actually a file named @file{lib@var{library}.a}. The linker
5011 then uses this file as if it had been specified precisely by name.
5013 The directories searched include several standard system directories
5014 plus any that you specify with @option{-L}.
5016 Normally the files found this way are library files---archive files
5017 whose members are object files. The linker handles an archive file by
5018 scanning through it for members which define symbols that have so far
5019 been referenced but not defined. But if the file that is found is an
5020 ordinary object file, it is linked in the usual fashion. The only
5021 difference between using an @option{-l} option and specifying a file name
5022 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
5023 and searches several directories.
5027 You need this special case of the @option{-l} option in order to
5028 link an Objective-C program.
5031 @opindex nostartfiles
5032 Do not use the standard system startup files when linking.
5033 The standard system libraries are used normally, unless @option{-nostdlib}
5034 or @option{-nodefaultlibs} is used.
5036 @item -nodefaultlibs
5037 @opindex nodefaultlibs
5038 Do not use the standard system libraries when linking.
5039 Only the libraries you specify will be passed to the linker.
5040 The standard startup files are used normally, unless @option{-nostartfiles}
5041 is used. The compiler may generate calls to memcmp, memset, and memcpy
5042 for System V (and ISO C) environments or to bcopy and bzero for
5043 BSD environments. These entries are usually resolved by entries in
5044 libc. These entry points should be supplied through some other
5045 mechanism when this option is specified.
5049 Do not use the standard system startup files or libraries when linking.
5050 No startup files and only the libraries you specify will be passed to
5051 the linker. The compiler may generate calls to memcmp, memset, and memcpy
5052 for System V (and ISO C) environments or to bcopy and bzero for
5053 BSD environments. These entries are usually resolved by entries in
5054 libc. These entry points should be supplied through some other
5055 mechanism when this option is specified.
5057 @cindex @option{-lgcc}, use with @option{-nostdlib}
5058 @cindex @option{-nostdlib} and unresolved references
5059 @cindex unresolved references and @option{-nostdlib}
5060 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
5061 @cindex @option{-nodefaultlibs} and unresolved references
5062 @cindex unresolved references and @option{-nodefaultlibs}
5063 One of the standard libraries bypassed by @option{-nostdlib} and
5064 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
5065 that GCC uses to overcome shortcomings of particular machines, or special
5066 needs for some languages.
5067 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
5068 Collection (GCC) Internals},
5069 for more discussion of @file{libgcc.a}.)
5070 In most cases, you need @file{libgcc.a} even when you want to avoid
5071 other standard libraries. In other words, when you specify @option{-nostdlib}
5072 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
5073 This ensures that you have no unresolved references to internal GCC
5074 library subroutines. (For example, @samp{__main}, used to ensure C++
5075 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
5076 GNU Compiler Collection (GCC) Internals}.)
5080 Produce a position independent executable on targets which support it.
5081 For predictable results, you must also specify the same set of options
5082 that were used to generate code (@option{-fpie}, @option{-fPIE},
5083 or model suboptions) when you specify this option.
5087 Remove all symbol table and relocation information from the executable.
5091 On systems that support dynamic linking, this prevents linking with the shared
5092 libraries. On other systems, this option has no effect.
5096 Produce a shared object which can then be linked with other objects to
5097 form an executable. Not all systems support this option. For predictable
5098 results, you must also specify the same set of options that were used to
5099 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
5100 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
5101 needs to build supplementary stub code for constructors to work. On
5102 multi-libbed systems, @samp{gcc -shared} must select the correct support
5103 libraries to link against. Failing to supply the correct flags may lead
5104 to subtle defects. Supplying them in cases where they are not necessary
5107 @item -shared-libgcc
5108 @itemx -static-libgcc
5109 @opindex shared-libgcc
5110 @opindex static-libgcc
5111 On systems that provide @file{libgcc} as a shared library, these options
5112 force the use of either the shared or static version respectively.
5113 If no shared version of @file{libgcc} was built when the compiler was
5114 configured, these options have no effect.
5116 There are several situations in which an application should use the
5117 shared @file{libgcc} instead of the static version. The most common
5118 of these is when the application wishes to throw and catch exceptions
5119 across different shared libraries. In that case, each of the libraries
5120 as well as the application itself should use the shared @file{libgcc}.
5122 Therefore, the G++ and GCJ drivers automatically add
5123 @option{-shared-libgcc} whenever you build a shared library or a main
5124 executable, because C++ and Java programs typically use exceptions, so
5125 this is the right thing to do.
5127 If, instead, you use the GCC driver to create shared libraries, you may
5128 find that they will not always be linked with the shared @file{libgcc}.
5129 If GCC finds, at its configuration time, that you have a GNU linker that
5130 does not support option @option{--eh-frame-hdr}, it will link the shared
5131 version of @file{libgcc} into shared libraries by default. Otherwise,
5132 it will take advantage of the linker and optimize away the linking with
5133 the shared version of @file{libgcc}, linking with the static version of
5134 libgcc by default. This allows exceptions to propagate through such
5135 shared libraries, without incurring relocation costs at library load
5138 However, if a library or main executable is supposed to throw or catch
5139 exceptions, you must link it using the G++ or GCJ driver, as appropriate
5140 for the languages used in the program, or using the option
5141 @option{-shared-libgcc}, such that it is linked with the shared
5146 Bind references to global symbols when building a shared object. Warn
5147 about any unresolved references (unless overridden by the link editor
5148 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
5151 @item -Xlinker @var{option}
5153 Pass @var{option} as an option to the linker. You can use this to
5154 supply system-specific linker options which GCC does not know how to
5157 If you want to pass an option that takes an argument, you must use
5158 @option{-Xlinker} twice, once for the option and once for the argument.
5159 For example, to pass @option{-assert definitions}, you must write
5160 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
5161 @option{-Xlinker "-assert definitions"}, because this passes the entire
5162 string as a single argument, which is not what the linker expects.
5164 @item -Wl,@var{option}
5166 Pass @var{option} as an option to the linker. If @var{option} contains
5167 commas, it is split into multiple options at the commas.
5169 @item -u @var{symbol}
5171 Pretend the symbol @var{symbol} is undefined, to force linking of
5172 library modules to define it. You can use @option{-u} multiple times with
5173 different symbols to force loading of additional library modules.
5176 @node Directory Options
5177 @section Options for Directory Search
5178 @cindex directory options
5179 @cindex options, directory search
5182 These options specify directories to search for header files, for
5183 libraries and for parts of the compiler:
5188 Add the directory @var{dir} to the head of the list of directories to be
5189 searched for header files. This can be used to override a system header
5190 file, substituting your own version, since these directories are
5191 searched before the system header file directories. However, you should
5192 not use this option to add directories that contain vendor-supplied
5193 system header files (use @option{-isystem} for that). If you use more than
5194 one @option{-I} option, the directories are scanned in left-to-right
5195 order; the standard system directories come after.
5197 If a standard system include directory, or a directory specified with
5198 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
5199 option will be ignored. The directory will still be searched but as a
5200 system directory at its normal position in the system include chain.
5201 This is to ensure that GCC's procedure to fix buggy system headers and
5202 the ordering for the include_next directive are not inadvertently changed.
5203 If you really need to change the search order for system directories,
5204 use the @option{-nostdinc} and/or @option{-isystem} options.
5208 Any directories you specify with @option{-I} options before the @option{-I-}
5209 option are searched only for the case of @samp{#include "@var{file}"};
5210 they are not searched for @samp{#include <@var{file}>}.
5212 If additional directories are specified with @option{-I} options after
5213 the @option{-I-}, these directories are searched for all @samp{#include}
5214 directives. (Ordinarily @emph{all} @option{-I} directories are used
5217 In addition, the @option{-I-} option inhibits the use of the current
5218 directory (where the current input file came from) as the first search
5219 directory for @samp{#include "@var{file}"}. There is no way to
5220 override this effect of @option{-I-}. With @option{-I.} you can specify
5221 searching the directory which was current when the compiler was
5222 invoked. That is not exactly the same as what the preprocessor does
5223 by default, but it is often satisfactory.
5225 @option{-I-} does not inhibit the use of the standard system directories
5226 for header files. Thus, @option{-I-} and @option{-nostdinc} are
5231 Add directory @var{dir} to the list of directories to be searched
5234 @item -B@var{prefix}
5236 This option specifies where to find the executables, libraries,
5237 include files, and data files of the compiler itself.
5239 The compiler driver program runs one or more of the subprograms
5240 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
5241 @var{prefix} as a prefix for each program it tries to run, both with and
5242 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
5244 For each subprogram to be run, the compiler driver first tries the
5245 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
5246 was not specified, the driver tries two standard prefixes, which are
5247 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
5248 those results in a file name that is found, the unmodified program
5249 name is searched for using the directories specified in your
5250 @env{PATH} environment variable.
5252 The compiler will check to see if the path provided by the @option{-B}
5253 refers to a directory, and if necessary it will add a directory
5254 separator character at the end of the path.
5256 @option{-B} prefixes that effectively specify directory names also apply
5257 to libraries in the linker, because the compiler translates these
5258 options into @option{-L} options for the linker. They also apply to
5259 includes files in the preprocessor, because the compiler translates these
5260 options into @option{-isystem} options for the preprocessor. In this case,
5261 the compiler appends @samp{include} to the prefix.
5263 The run-time support file @file{libgcc.a} can also be searched for using
5264 the @option{-B} prefix, if needed. If it is not found there, the two
5265 standard prefixes above are tried, and that is all. The file is left
5266 out of the link if it is not found by those means.
5268 Another way to specify a prefix much like the @option{-B} prefix is to use
5269 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
5272 As a special kludge, if the path provided by @option{-B} is
5273 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
5274 9, then it will be replaced by @file{[dir/]include}. This is to help
5275 with boot-strapping the compiler.
5277 @item -specs=@var{file}
5279 Process @var{file} after the compiler reads in the standard @file{specs}
5280 file, in order to override the defaults that the @file{gcc} driver
5281 program uses when determining what switches to pass to @file{cc1},
5282 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
5283 @option{-specs=@var{file}} can be specified on the command line, and they
5284 are processed in order, from left to right.
5290 @section Specifying subprocesses and the switches to pass to them
5293 @command{gcc} is a driver program. It performs its job by invoking a
5294 sequence of other programs to do the work of compiling, assembling and
5295 linking. GCC interprets its command-line parameters and uses these to
5296 deduce which programs it should invoke, and which command-line options
5297 it ought to place on their command lines. This behavior is controlled
5298 by @dfn{spec strings}. In most cases there is one spec string for each
5299 program that GCC can invoke, but a few programs have multiple spec
5300 strings to control their behavior. The spec strings built into GCC can
5301 be overridden by using the @option{-specs=} command-line switch to specify
5304 @dfn{Spec files} are plaintext files that are used to construct spec
5305 strings. They consist of a sequence of directives separated by blank
5306 lines. The type of directive is determined by the first non-whitespace
5307 character on the line and it can be one of the following:
5310 @item %@var{command}
5311 Issues a @var{command} to the spec file processor. The commands that can
5315 @item %include <@var{file}>
5317 Search for @var{file} and insert its text at the current point in the
5320 @item %include_noerr <@var{file}>
5321 @cindex %include_noerr
5322 Just like @samp{%include}, but do not generate an error message if the include
5323 file cannot be found.
5325 @item %rename @var{old_name} @var{new_name}
5327 Rename the spec string @var{old_name} to @var{new_name}.
5331 @item *[@var{spec_name}]:
5332 This tells the compiler to create, override or delete the named spec
5333 string. All lines after this directive up to the next directive or
5334 blank line are considered to be the text for the spec string. If this
5335 results in an empty string then the spec will be deleted. (Or, if the
5336 spec did not exist, then nothing will happened.) Otherwise, if the spec
5337 does not currently exist a new spec will be created. If the spec does
5338 exist then its contents will be overridden by the text of this
5339 directive, unless the first character of that text is the @samp{+}
5340 character, in which case the text will be appended to the spec.
5342 @item [@var{suffix}]:
5343 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
5344 and up to the next directive or blank line are considered to make up the
5345 spec string for the indicated suffix. When the compiler encounters an
5346 input file with the named suffix, it will processes the spec string in
5347 order to work out how to compile that file. For example:
5354 This says that any input file whose name ends in @samp{.ZZ} should be
5355 passed to the program @samp{z-compile}, which should be invoked with the
5356 command-line switch @option{-input} and with the result of performing the
5357 @samp{%i} substitution. (See below.)
5359 As an alternative to providing a spec string, the text that follows a
5360 suffix directive can be one of the following:
5363 @item @@@var{language}
5364 This says that the suffix is an alias for a known @var{language}. This is
5365 similar to using the @option{-x} command-line switch to GCC to specify a
5366 language explicitly. For example:
5373 Says that .ZZ files are, in fact, C++ source files.
5376 This causes an error messages saying:
5379 @var{name} compiler not installed on this system.
5383 GCC already has an extensive list of suffixes built into it.
5384 This directive will add an entry to the end of the list of suffixes, but
5385 since the list is searched from the end backwards, it is effectively
5386 possible to override earlier entries using this technique.
5390 GCC has the following spec strings built into it. Spec files can
5391 override these strings or create their own. Note that individual
5392 targets can also add their own spec strings to this list.
5395 asm Options to pass to the assembler
5396 asm_final Options to pass to the assembler post-processor
5397 cpp Options to pass to the C preprocessor
5398 cc1 Options to pass to the C compiler
5399 cc1plus Options to pass to the C++ compiler
5400 endfile Object files to include at the end of the link
5401 link Options to pass to the linker
5402 lib Libraries to include on the command line to the linker
5403 libgcc Decides which GCC support library to pass to the linker
5404 linker Sets the name of the linker
5405 predefines Defines to be passed to the C preprocessor
5406 signed_char Defines to pass to CPP to say whether @code{char} is signed
5408 startfile Object files to include at the start of the link
5411 Here is a small example of a spec file:
5417 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
5420 This example renames the spec called @samp{lib} to @samp{old_lib} and
5421 then overrides the previous definition of @samp{lib} with a new one.
5422 The new definition adds in some extra command-line options before
5423 including the text of the old definition.
5425 @dfn{Spec strings} are a list of command-line options to be passed to their
5426 corresponding program. In addition, the spec strings can contain
5427 @samp{%}-prefixed sequences to substitute variable text or to
5428 conditionally insert text into the command line. Using these constructs
5429 it is possible to generate quite complex command lines.
5431 Here is a table of all defined @samp{%}-sequences for spec
5432 strings. Note that spaces are not generated automatically around the
5433 results of expanding these sequences. Therefore you can concatenate them
5434 together or combine them with constant text in a single argument.
5438 Substitute one @samp{%} into the program name or argument.
5441 Substitute the name of the input file being processed.
5444 Substitute the basename of the input file being processed.
5445 This is the substring up to (and not including) the last period
5446 and not including the directory.
5449 This is the same as @samp{%b}, but include the file suffix (text after
5453 Marks the argument containing or following the @samp{%d} as a
5454 temporary file name, so that that file will be deleted if GCC exits
5455 successfully. Unlike @samp{%g}, this contributes no text to the
5458 @item %g@var{suffix}
5459 Substitute a file name that has suffix @var{suffix} and is chosen
5460 once per compilation, and mark the argument in the same way as
5461 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
5462 name is now chosen in a way that is hard to predict even when previously
5463 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
5464 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
5465 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
5466 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
5467 was simply substituted with a file name chosen once per compilation,
5468 without regard to any appended suffix (which was therefore treated
5469 just like ordinary text), making such attacks more likely to succeed.
5471 @item %u@var{suffix}
5472 Like @samp{%g}, but generates a new temporary file name even if
5473 @samp{%u@var{suffix}} was already seen.
5475 @item %U@var{suffix}
5476 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
5477 new one if there is no such last file name. In the absence of any
5478 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
5479 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
5480 would involve the generation of two distinct file names, one
5481 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
5482 simply substituted with a file name chosen for the previous @samp{%u},
5483 without regard to any appended suffix.
5485 @item %j@var{suffix}
5486 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
5487 writable, and if save-temps is off; otherwise, substitute the name
5488 of a temporary file, just like @samp{%u}. This temporary file is not
5489 meant for communication between processes, but rather as a junk
5492 @item %|@var{suffix}
5493 @itemx %m@var{suffix}
5494 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
5495 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
5496 all. These are the two most common ways to instruct a program that it
5497 should read from standard input or write to standard output. If you
5498 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
5499 construct: see for example @file{f/lang-specs.h}.
5501 @item %.@var{SUFFIX}
5502 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
5503 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
5504 terminated by the next space or %.
5507 Marks the argument containing or following the @samp{%w} as the
5508 designated output file of this compilation. This puts the argument
5509 into the sequence of arguments that @samp{%o} will substitute later.
5512 Substitutes the names of all the output files, with spaces
5513 automatically placed around them. You should write spaces
5514 around the @samp{%o} as well or the results are undefined.
5515 @samp{%o} is for use in the specs for running the linker.
5516 Input files whose names have no recognized suffix are not compiled
5517 at all, but they are included among the output files, so they will
5521 Substitutes the suffix for object files. Note that this is
5522 handled specially when it immediately follows @samp{%g, %u, or %U},
5523 because of the need for those to form complete file names. The
5524 handling is such that @samp{%O} is treated exactly as if it had already
5525 been substituted, except that @samp{%g, %u, and %U} do not currently
5526 support additional @var{suffix} characters following @samp{%O} as they would
5527 following, for example, @samp{.o}.
5530 Substitutes the standard macro predefinitions for the
5531 current target machine. Use this when running @code{cpp}.
5534 Like @samp{%p}, but puts @samp{__} before and after the name of each
5535 predefined macro, except for macros that start with @samp{__} or with
5536 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
5540 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
5541 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), and
5542 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
5546 Current argument is the name of a library or startup file of some sort.
5547 Search for that file in a standard list of directories and substitute
5548 the full name found.
5551 Print @var{str} as an error message. @var{str} is terminated by a newline.
5552 Use this when inconsistent options are detected.
5555 Substitute the contents of spec string @var{name} at this point.
5558 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
5560 @item %x@{@var{option}@}
5561 Accumulate an option for @samp{%X}.
5564 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
5568 Output the accumulated assembler options specified by @option{-Wa}.
5571 Output the accumulated preprocessor options specified by @option{-Wp}.
5574 Process the @code{asm} spec. This is used to compute the
5575 switches to be passed to the assembler.
5578 Process the @code{asm_final} spec. This is a spec string for
5579 passing switches to an assembler post-processor, if such a program is
5583 Process the @code{link} spec. This is the spec for computing the
5584 command line passed to the linker. Typically it will make use of the
5585 @samp{%L %G %S %D and %E} sequences.
5588 Dump out a @option{-L} option for each directory that GCC believes might
5589 contain startup files. If the target supports multilibs then the
5590 current multilib directory will be prepended to each of these paths.
5593 Output the multilib directory with directory separators replaced with
5594 @samp{_}. If multilib directories are not set, or the multilib directory is
5595 @file{.} then this option emits nothing.
5598 Process the @code{lib} spec. This is a spec string for deciding which
5599 libraries should be included on the command line to the linker.
5602 Process the @code{libgcc} spec. This is a spec string for deciding
5603 which GCC support library should be included on the command line to the linker.
5606 Process the @code{startfile} spec. This is a spec for deciding which
5607 object files should be the first ones passed to the linker. Typically
5608 this might be a file named @file{crt0.o}.
5611 Process the @code{endfile} spec. This is a spec string that specifies
5612 the last object files that will be passed to the linker.
5615 Process the @code{cpp} spec. This is used to construct the arguments
5616 to be passed to the C preprocessor.
5619 Process the @code{signed_char} spec. This is intended to be used
5620 to tell cpp whether a char is signed. It typically has the definition:
5622 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
5626 Process the @code{cc1} spec. This is used to construct the options to be
5627 passed to the actual C compiler (@samp{cc1}).
5630 Process the @code{cc1plus} spec. This is used to construct the options to be
5631 passed to the actual C++ compiler (@samp{cc1plus}).
5634 Substitute the variable part of a matched option. See below.
5635 Note that each comma in the substituted string is replaced by
5639 Remove all occurrences of @code{-S} from the command line. Note---this
5640 command is position dependent. @samp{%} commands in the spec string
5641 before this one will see @code{-S}, @samp{%} commands in the spec string
5642 after this one will not.
5644 @item %:@var{function}(@var{args})
5645 Call the named function @var{function}, passing it @var{args}.
5646 @var{args} is first processed as a nested spec string, then split
5647 into an argument vector in the usual fashion. The function returns
5648 a string which is processed as if it had appeared literally as part
5649 of the current spec.
5651 The following built-in spec functions are provided:
5654 @item @code{if-exists}
5655 The @code{if-exists} spec function takes one argument, an absolute
5656 pathname to a file. If the file exists, @code{if-exists} returns the
5657 pathname. Here is a small example of its usage:
5661 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
5664 @item @code{if-exists-else}
5665 The @code{if-exists-else} spec function is similar to the @code{if-exists}
5666 spec function, except that it takes two arguments. The first argument is
5667 an absolute pathname to a file. If the file exists, @code{if-exists-else}
5668 returns the pathname. If it does not exist, it returns the second argument.
5669 This way, @code{if-exists-else} can be used to select one file or another,
5670 based on the existence of the first. Here is a small example of its usage:
5674 crt0%O%s %:if-exists(crti%O%s) \
5675 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
5680 Substitutes the @code{-S} switch, if that switch was given to GCC@.
5681 If that switch was not specified, this substitutes nothing. Note that
5682 the leading dash is omitted when specifying this option, and it is
5683 automatically inserted if the substitution is performed. Thus the spec
5684 string @samp{%@{foo@}} would match the command-line option @option{-foo}
5685 and would output the command line option @option{-foo}.
5687 @item %W@{@code{S}@}
5688 Like %@{@code{S}@} but mark last argument supplied within as a file to be
5691 @item %@{@code{S}*@}
5692 Substitutes all the switches specified to GCC whose names start
5693 with @code{-S}, but which also take an argument. This is used for
5694 switches like @option{-o}, @option{-D}, @option{-I}, etc.
5695 GCC considers @option{-o foo} as being
5696 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
5697 text, including the space. Thus two arguments would be generated.
5699 @item %@{@code{S}*&@code{T}*@}
5700 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
5701 (the order of @code{S} and @code{T} in the spec is not significant).
5702 There can be any number of ampersand-separated variables; for each the
5703 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
5705 @item %@{@code{S}:@code{X}@}
5706 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
5708 @item %@{!@code{S}:@code{X}@}
5709 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
5711 @item %@{@code{S}*:@code{X}@}
5712 Substitutes @code{X} if one or more switches whose names start with
5713 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
5714 once, no matter how many such switches appeared. However, if @code{%*}
5715 appears somewhere in @code{X}, then @code{X} will be substituted once
5716 for each matching switch, with the @code{%*} replaced by the part of
5717 that switch that matched the @code{*}.
5719 @item %@{.@code{S}:@code{X}@}
5720 Substitutes @code{X}, if processing a file with suffix @code{S}.
5722 @item %@{!.@code{S}:@code{X}@}
5723 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
5725 @item %@{@code{S}|@code{P}:@code{X}@}
5726 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
5727 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
5728 although they have a stronger binding than the @samp{|}. If @code{%*}
5729 appears in @code{X}, all of the alternatives must be starred, and only
5730 the first matching alternative is substituted.
5732 For example, a spec string like this:
5735 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
5738 will output the following command-line options from the following input
5739 command-line options:
5744 -d fred.c -foo -baz -boggle
5745 -d jim.d -bar -baz -boggle
5748 @item %@{S:X; T:Y; :D@}
5750 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
5751 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
5752 be as many clauses as you need. This may be combined with @code{.},
5753 @code{!}, @code{|}, and @code{*} as needed.
5758 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
5759 construct may contain other nested @samp{%} constructs or spaces, or
5760 even newlines. They are processed as usual, as described above.
5761 Trailing white space in @code{X} is ignored. White space may also
5762 appear anywhere on the left side of the colon in these constructs,
5763 except between @code{.} or @code{*} and the corresponding word.
5765 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
5766 handled specifically in these constructs. If another value of
5767 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
5768 @option{-W} switch is found later in the command line, the earlier
5769 switch value is ignored, except with @{@code{S}*@} where @code{S} is
5770 just one letter, which passes all matching options.
5772 The character @samp{|} at the beginning of the predicate text is used to
5773 indicate that a command should be piped to the following command, but
5774 only if @option{-pipe} is specified.
5776 It is built into GCC which switches take arguments and which do not.
5777 (You might think it would be useful to generalize this to allow each
5778 compiler's spec to say which switches take arguments. But this cannot
5779 be done in a consistent fashion. GCC cannot even decide which input
5780 files have been specified without knowing which switches take arguments,
5781 and it must know which input files to compile in order to tell which
5784 GCC also knows implicitly that arguments starting in @option{-l} are to be
5785 treated as compiler output files, and passed to the linker in their
5786 proper position among the other output files.
5788 @c man begin OPTIONS
5790 @node Target Options
5791 @section Specifying Target Machine and Compiler Version
5792 @cindex target options
5793 @cindex cross compiling
5794 @cindex specifying machine version
5795 @cindex specifying compiler version and target machine
5796 @cindex compiler version, specifying
5797 @cindex target machine, specifying
5799 The usual way to run GCC is to run the executable called @file{gcc}, or
5800 @file{<machine>-gcc} when cross-compiling, or
5801 @file{<machine>-gcc-<version>} to run a version other than the one that
5802 was installed last. Sometimes this is inconvenient, so GCC provides
5803 options that will switch to another cross-compiler or version.
5806 @item -b @var{machine}
5808 The argument @var{machine} specifies the target machine for compilation.
5810 The value to use for @var{machine} is the same as was specified as the
5811 machine type when configuring GCC as a cross-compiler. For
5812 example, if a cross-compiler was configured with @samp{configure
5813 i386v}, meaning to compile for an 80386 running System V, then you
5814 would specify @option{-b i386v} to run that cross compiler.
5816 @item -V @var{version}
5818 The argument @var{version} specifies which version of GCC to run.
5819 This is useful when multiple versions are installed. For example,
5820 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5823 The @option{-V} and @option{-b} options work by running the
5824 @file{<machine>-gcc-<version>} executable, so there's no real reason to
5825 use them if you can just run that directly.
5827 @node Submodel Options
5828 @section Hardware Models and Configurations
5829 @cindex submodel options
5830 @cindex specifying hardware config
5831 @cindex hardware models and configurations, specifying
5832 @cindex machine dependent options
5834 Earlier we discussed the standard option @option{-b} which chooses among
5835 different installed compilers for completely different target
5836 machines, such as VAX vs.@: 68000 vs.@: 80386.
5838 In addition, each of these target machine types can have its own
5839 special options, starting with @samp{-m}, to choose among various
5840 hardware models or configurations---for example, 68010 vs 68020,
5841 floating coprocessor or none. A single installed version of the
5842 compiler can compile for any model or configuration, according to the
5845 Some configurations of the compiler also support additional special
5846 options, usually for compatibility with other compilers on the same
5849 These options are defined by the macro @code{TARGET_SWITCHES} in the
5850 machine description. The default for the options is also defined by
5851 that macro, which enables you to change the defaults.
5861 * RS/6000 and PowerPC Options::
5864 * i386 and x86-64 Options::
5866 * DEC Alpha Options::
5867 * DEC Alpha/VMS Options::
5870 * System V Options::
5871 * TMS320C3x/C4x Options::
5878 * S/390 and zSeries Options::
5882 * Xstormy16 Options::
5887 @node M680x0 Options
5888 @subsection M680x0 Options
5889 @cindex M680x0 options
5891 These are the @samp{-m} options defined for the 68000 series. The default
5892 values for these options depends on which style of 68000 was selected when
5893 the compiler was configured; the defaults for the most common choices are
5901 Generate output for a 68000. This is the default
5902 when the compiler is configured for 68000-based systems.
5904 Use this option for microcontrollers with a 68000 or EC000 core,
5905 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5911 Generate output for a 68020. This is the default
5912 when the compiler is configured for 68020-based systems.
5916 Generate output containing 68881 instructions for floating point.
5917 This is the default for most 68020 systems unless @option{--nfp} was
5918 specified when the compiler was configured.
5922 Generate output for a 68030. This is the default when the compiler is
5923 configured for 68030-based systems.
5927 Generate output for a 68040. This is the default when the compiler is
5928 configured for 68040-based systems.
5930 This option inhibits the use of 68881/68882 instructions that have to be
5931 emulated by software on the 68040. Use this option if your 68040 does not
5932 have code to emulate those instructions.
5936 Generate output for a 68060. This is the default when the compiler is
5937 configured for 68060-based systems.
5939 This option inhibits the use of 68020 and 68881/68882 instructions that
5940 have to be emulated by software on the 68060. Use this option if your 68060
5941 does not have code to emulate those instructions.
5945 Generate output for a CPU32. This is the default
5946 when the compiler is configured for CPU32-based systems.
5948 Use this option for microcontrollers with a
5949 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5950 68336, 68340, 68341, 68349 and 68360.
5954 Generate output for a 520X ``coldfire'' family cpu. This is the default
5955 when the compiler is configured for 520X-based systems.
5957 Use this option for microcontroller with a 5200 core, including
5958 the MCF5202, MCF5203, MCF5204 and MCF5202.
5963 Generate output for a 68040, without using any of the new instructions.
5964 This results in code which can run relatively efficiently on either a
5965 68020/68881 or a 68030 or a 68040. The generated code does use the
5966 68881 instructions that are emulated on the 68040.
5970 Generate output for a 68060, without using any of the new instructions.
5971 This results in code which can run relatively efficiently on either a
5972 68020/68881 or a 68030 or a 68040. The generated code does use the
5973 68881 instructions that are emulated on the 68060.
5976 @opindex msoft-float
5977 Generate output containing library calls for floating point.
5978 @strong{Warning:} the requisite libraries are not available for all m68k
5979 targets. Normally the facilities of the machine's usual C compiler are
5980 used, but this can't be done directly in cross-compilation. You must
5981 make your own arrangements to provide suitable library functions for
5982 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5983 @samp{m68k-*-coff} do provide software floating point support.
5987 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5990 @opindex mnobitfield
5991 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5992 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5996 Do use the bit-field instructions. The @option{-m68020} option implies
5997 @option{-mbitfield}. This is the default if you use a configuration
5998 designed for a 68020.
6002 Use a different function-calling convention, in which functions
6003 that take a fixed number of arguments return with the @code{rtd}
6004 instruction, which pops their arguments while returning. This
6005 saves one instruction in the caller since there is no need to pop
6006 the arguments there.
6008 This calling convention is incompatible with the one normally
6009 used on Unix, so you cannot use it if you need to call libraries
6010 compiled with the Unix compiler.
6012 Also, you must provide function prototypes for all functions that
6013 take variable numbers of arguments (including @code{printf});
6014 otherwise incorrect code will be generated for calls to those
6017 In addition, seriously incorrect code will result if you call a
6018 function with too many arguments. (Normally, extra arguments are
6019 harmlessly ignored.)
6021 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
6022 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
6025 @itemx -mno-align-int
6027 @opindex mno-align-int
6028 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
6029 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
6030 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
6031 Aligning variables on 32-bit boundaries produces code that runs somewhat
6032 faster on processors with 32-bit busses at the expense of more memory.
6034 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
6035 align structures containing the above types differently than
6036 most published application binary interface specifications for the m68k.
6040 Use the pc-relative addressing mode of the 68000 directly, instead of
6041 using a global offset table. At present, this option implies @option{-fpic},
6042 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
6043 not presently supported with @option{-mpcrel}, though this could be supported for
6044 68020 and higher processors.
6046 @item -mno-strict-align
6047 @itemx -mstrict-align
6048 @opindex mno-strict-align
6049 @opindex mstrict-align
6050 Do not (do) assume that unaligned memory references will be handled by
6054 Generate code that allows the data segment to be located in a different
6055 area of memory from the text segment. This allows for execute in place in
6056 an environment without virtual memory management. This option implies -fPIC.
6059 Generate code that assumes that the data segment follows the text segment.
6060 This is the default.
6062 @item -mid-shared-library
6063 Generate code that supports shared libraries via the library ID method.
6064 This allows for execute in place and shared libraries in an environment
6065 without virtual memory management. This option implies -fPIC.
6067 @item -mno-id-shared-library
6068 Generate code that doesn't assume ID based shared libraries are being used.
6069 This is the default.
6071 @item -mshared-library-id=n
6072 Specified the identification number of the ID based shared library being
6073 compiled. Specifying a value of 0 will generate more compact code, specifying
6074 other values will force the allocation of that number to the current
6075 library but is no more space or time efficient than omitting this option.
6079 @node M68hc1x Options
6080 @subsection M68hc1x Options
6081 @cindex M68hc1x options
6083 These are the @samp{-m} options defined for the 68hc11 and 68hc12
6084 microcontrollers. The default values for these options depends on
6085 which style of microcontroller was selected when the compiler was configured;
6086 the defaults for the most common choices are given below.
6093 Generate output for a 68HC11. This is the default
6094 when the compiler is configured for 68HC11-based systems.
6100 Generate output for a 68HC12. This is the default
6101 when the compiler is configured for 68HC12-based systems.
6107 Generate output for a 68HCS12.
6110 @opindex mauto-incdec
6111 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
6118 Enable the use of 68HC12 min and max instructions.
6121 @itemx -mno-long-calls
6122 @opindex mlong-calls
6123 @opindex mno-long-calls
6124 Treat all calls as being far away (near). If calls are assumed to be
6125 far away, the compiler will use the @code{call} instruction to
6126 call a function and the @code{rtc} instruction for returning.
6130 Consider type @code{int} to be 16 bits wide, like @code{short int}.
6132 @item -msoft-reg-count=@var{count}
6133 @opindex msoft-reg-count
6134 Specify the number of pseudo-soft registers which are used for the
6135 code generation. The maximum number is 32. Using more pseudo-soft
6136 register may or may not result in better code depending on the program.
6137 The default is 4 for 68HC11 and 2 for 68HC12.
6142 @subsection VAX Options
6145 These @samp{-m} options are defined for the VAX:
6150 Do not output certain jump instructions (@code{aobleq} and so on)
6151 that the Unix assembler for the VAX cannot handle across long
6156 Do output those jump instructions, on the assumption that you
6157 will assemble with the GNU assembler.
6161 Output code for g-format floating point numbers instead of d-format.
6165 @subsection SPARC Options
6166 @cindex SPARC options
6168 These @samp{-m} options are supported on the SPARC:
6173 @opindex mno-app-regs
6175 Specify @option{-mapp-regs} to generate output using the global registers
6176 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
6179 To be fully SVR4 ABI compliant at the cost of some performance loss,
6180 specify @option{-mno-app-regs}. You should compile libraries and system
6181 software with this option.
6186 @opindex mhard-float
6187 Generate output containing floating point instructions. This is the
6193 @opindex msoft-float
6194 Generate output containing library calls for floating point.
6195 @strong{Warning:} the requisite libraries are not available for all SPARC
6196 targets. Normally the facilities of the machine's usual C compiler are
6197 used, but this cannot be done directly in cross-compilation. You must make
6198 your own arrangements to provide suitable library functions for
6199 cross-compilation. The embedded targets @samp{sparc-*-aout} and
6200 @samp{sparclite-*-*} do provide software floating point support.
6202 @option{-msoft-float} changes the calling convention in the output file;
6203 therefore, it is only useful if you compile @emph{all} of a program with
6204 this option. In particular, you need to compile @file{libgcc.a}, the
6205 library that comes with GCC, with @option{-msoft-float} in order for
6208 @item -mhard-quad-float
6209 @opindex mhard-quad-float
6210 Generate output containing quad-word (long double) floating point
6213 @item -msoft-quad-float
6214 @opindex msoft-quad-float
6215 Generate output containing library calls for quad-word (long double)
6216 floating point instructions. The functions called are those specified
6217 in the SPARC ABI@. This is the default.
6219 As of this writing, there are no SPARC implementations that have hardware
6220 support for the quad-word floating point instructions. They all invoke
6221 a trap handler for one of these instructions, and then the trap handler
6222 emulates the effect of the instruction. Because of the trap handler overhead,
6223 this is much slower than calling the ABI library routines. Thus the
6224 @option{-msoft-quad-float} option is the default.
6226 @item -mno-unaligned-doubles
6227 @itemx -munaligned-doubles
6228 @opindex mno-unaligned-doubles
6229 @opindex munaligned-doubles
6230 Assume that doubles have 8 byte alignment. This is the default.
6232 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
6233 alignment only if they are contained in another type, or if they have an
6234 absolute address. Otherwise, it assumes they have 4 byte alignment.
6235 Specifying this option avoids some rare compatibility problems with code
6236 generated by other compilers. It is not the default because it results
6237 in a performance loss, especially for floating point code.
6239 @item -mno-faster-structs
6240 @itemx -mfaster-structs
6241 @opindex mno-faster-structs
6242 @opindex mfaster-structs
6243 With @option{-mfaster-structs}, the compiler assumes that structures
6244 should have 8 byte alignment. This enables the use of pairs of
6245 @code{ldd} and @code{std} instructions for copies in structure
6246 assignment, in place of twice as many @code{ld} and @code{st} pairs.
6247 However, the use of this changed alignment directly violates the SPARC
6248 ABI@. Thus, it's intended only for use on targets where the developer
6249 acknowledges that their resulting code will not be directly in line with
6250 the rules of the ABI@.
6253 @opindex mimpure-text
6254 @option{-mimpure-text}, used in addition to @option{-shared}, tells
6255 the compiler to not pass @option{-z text} to the linker when linking a
6256 shared object. Using this option, you can link position-dependent
6257 code into a shared object.
6259 @option{-mimpure-text} suppresses the ``relocations remain against
6260 allocatable but non-writable sections'' linker error message.
6261 However, the necessary relocations will trigger copy-on-write, and the
6262 shared object is not actually shared across processes. Instead of
6263 using @option{-mimpure-text}, you should compile all source code with
6264 @option{-fpic} or @option{-fPIC}.
6266 This option is only available on SunOS and Solaris.
6268 @item -mcpu=@var{cpu_type}
6270 Set the instruction set, register set, and instruction scheduling parameters
6271 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
6272 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
6273 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
6274 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
6277 Default instruction scheduling parameters are used for values that select
6278 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
6279 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
6281 Here is a list of each supported architecture and their supported
6286 v8: supersparc, hypersparc
6287 sparclite: f930, f934, sparclite86x
6289 v9: ultrasparc, ultrasparc3
6292 By default (unless configured otherwise), GCC generates code for the V7
6293 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
6294 additionally optimizes it for the Cypress CY7C602 chip, as used in the
6295 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
6296 SPARCStation 1, 2, IPX etc.
6298 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
6299 architecture. The only difference from V7 code is that the compiler emits
6300 the integer multiply and integer divide instructions which exist in SPARC-V8
6301 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
6302 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
6305 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
6306 the SPARC architecture. This adds the integer multiply, integer divide step
6307 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
6308 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
6309 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU. With
6310 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
6311 MB86934 chip, which is the more recent SPARClite with FPU.
6313 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
6314 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
6315 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
6316 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
6317 optimizes it for the TEMIC SPARClet chip.
6319 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
6320 architecture. This adds 64-bit integer and floating-point move instructions,
6321 3 additional floating-point condition code registers and conditional move
6322 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
6323 optimizes it for the Sun UltraSPARC I/II chips. With
6324 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
6325 Sun UltraSPARC III chip.
6327 @item -mtune=@var{cpu_type}
6329 Set the instruction scheduling parameters for machine type
6330 @var{cpu_type}, but do not set the instruction set or register set that the
6331 option @option{-mcpu=@var{cpu_type}} would.
6333 The same values for @option{-mcpu=@var{cpu_type}} can be used for
6334 @option{-mtune=@var{cpu_type}}, but the only useful values are those
6335 that select a particular cpu implementation. Those are @samp{cypress},
6336 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
6337 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
6343 @opindex -mno-v8plus
6344 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI. The
6345 difference from the V8 ABI is that the global and out registers are
6346 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
6347 mode for all SPARC-V9 processors.
6353 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
6354 Visual Instruction Set extensions. The default is @option{-mno-vis}.
6357 These @samp{-m} options are supported in addition to the above
6358 on SPARC-V9 processors in 64-bit environments:
6361 @item -mlittle-endian
6362 @opindex mlittle-endian
6363 Generate code for a processor running in little-endian mode. It is only
6364 available for a few configurations and most notably not on Solaris.
6370 Generate code for a 32-bit or 64-bit environment.
6371 The 32-bit environment sets int, long and pointer to 32 bits.
6372 The 64-bit environment sets int to 32 bits and long and pointer
6375 @item -mcmodel=medlow
6376 @opindex mcmodel=medlow
6377 Generate code for the Medium/Low code model: 64-bit addresses, programs
6378 must be linked in the low 32 bits of memory. Programs can be statically
6379 or dynamically linked.
6381 @item -mcmodel=medmid
6382 @opindex mcmodel=medmid
6383 Generate code for the Medium/Middle code model: 64-bit addresses, programs
6384 must be linked in the low 44 bits of memory, the text and data segments must
6385 be less than 2GB in size and the data segment must be located within 2GB of
6388 @item -mcmodel=medany
6389 @opindex mcmodel=medany
6390 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
6391 may be linked anywhere in memory, the text and data segments must be less
6392 than 2GB in size and the data segment must be located within 2GB of the
6395 @item -mcmodel=embmedany
6396 @opindex mcmodel=embmedany
6397 Generate code for the Medium/Anywhere code model for embedded systems:
6398 64-bit addresses, the text and data segments must be less than 2GB in
6399 size, both starting anywhere in memory (determined at link time). The
6400 global register %g4 points to the base of the data segment. Programs
6401 are statically linked and PIC is not supported.
6404 @itemx -mno-stack-bias
6405 @opindex mstack-bias
6406 @opindex mno-stack-bias
6407 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
6408 frame pointer if present, are offset by @minus{}2047 which must be added back
6409 when making stack frame references. This is the default in 64-bit mode.
6410 Otherwise, assume no such offset is present.
6414 @subsection ARM Options
6417 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
6421 @item -mabi=@var{name}
6423 Generate code for the specified ABI. Permissible values are: @samp{apcs-gnu},
6424 @samp{atpcs}, @samp{aapcs} and @samp{iwmmxt}.
6427 @opindex mapcs-frame
6428 Generate a stack frame that is compliant with the ARM Procedure Call
6429 Standard for all functions, even if this is not strictly necessary for
6430 correct execution of the code. Specifying @option{-fomit-frame-pointer}
6431 with this option will cause the stack frames not to be generated for
6432 leaf functions. The default is @option{-mno-apcs-frame}.
6436 This is a synonym for @option{-mapcs-frame}.
6440 Generate code for a processor running with a 26-bit program counter,
6441 and conforming to the function calling standards for the APCS 26-bit
6442 option. This option replaces the @option{-m2} and @option{-m3} options
6443 of previous releases of the compiler.
6447 Generate code for a processor running with a 32-bit program counter,
6448 and conforming to the function calling standards for the APCS 32-bit
6449 option. This option replaces the @option{-m6} option of previous releases
6453 @c not currently implemented
6454 @item -mapcs-stack-check
6455 @opindex mapcs-stack-check
6456 Generate code to check the amount of stack space available upon entry to
6457 every function (that actually uses some stack space). If there is
6458 insufficient space available then either the function
6459 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
6460 called, depending upon the amount of stack space required. The run time
6461 system is required to provide these functions. The default is
6462 @option{-mno-apcs-stack-check}, since this produces smaller code.
6464 @c not currently implemented
6466 @opindex mapcs-float
6467 Pass floating point arguments using the float point registers. This is
6468 one of the variants of the APCS@. This option is recommended if the
6469 target hardware has a floating point unit or if a lot of floating point
6470 arithmetic is going to be performed by the code. The default is
6471 @option{-mno-apcs-float}, since integer only code is slightly increased in
6472 size if @option{-mapcs-float} is used.
6474 @c not currently implemented
6475 @item -mapcs-reentrant
6476 @opindex mapcs-reentrant
6477 Generate reentrant, position independent code. The default is
6478 @option{-mno-apcs-reentrant}.
6481 @item -mthumb-interwork
6482 @opindex mthumb-interwork
6483 Generate code which supports calling between the ARM and Thumb
6484 instruction sets. Without this option the two instruction sets cannot
6485 be reliably used inside one program. The default is
6486 @option{-mno-thumb-interwork}, since slightly larger code is generated
6487 when @option{-mthumb-interwork} is specified.
6489 @item -mno-sched-prolog
6490 @opindex mno-sched-prolog
6491 Prevent the reordering of instructions in the function prolog, or the
6492 merging of those instruction with the instructions in the function's
6493 body. This means that all functions will start with a recognizable set
6494 of instructions (or in fact one of a choice from a small set of
6495 different function prologues), and this information can be used to
6496 locate the start if functions inside an executable piece of code. The
6497 default is @option{-msched-prolog}.
6500 @opindex mhard-float
6501 Generate output containing floating point instructions. This is the
6505 @opindex msoft-float
6506 Generate output containing library calls for floating point.
6507 @strong{Warning:} the requisite libraries are not available for all ARM
6508 targets. Normally the facilities of the machine's usual C compiler are
6509 used, but this cannot be done directly in cross-compilation. You must make
6510 your own arrangements to provide suitable library functions for
6513 @option{-msoft-float} changes the calling convention in the output file;
6514 therefore, it is only useful if you compile @emph{all} of a program with
6515 this option. In particular, you need to compile @file{libgcc.a}, the
6516 library that comes with GCC, with @option{-msoft-float} in order for
6519 @item -mfloat-abi=@var{name}
6521 Specifies which ABI to use for floating point values. Permissible values
6522 are: @samp{soft}, @samp{softfp} and @samp{hard}.
6524 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
6525 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
6526 of floating point instructions, but still uses the soft-float calling
6529 @item -mlittle-endian
6530 @opindex mlittle-endian
6531 Generate code for a processor running in little-endian mode. This is
6532 the default for all standard configurations.
6535 @opindex mbig-endian
6536 Generate code for a processor running in big-endian mode; the default is
6537 to compile code for a little-endian processor.
6539 @item -mwords-little-endian
6540 @opindex mwords-little-endian
6541 This option only applies when generating code for big-endian processors.
6542 Generate code for a little-endian word order but a big-endian byte
6543 order. That is, a byte order of the form @samp{32107654}. Note: this
6544 option should only be used if you require compatibility with code for
6545 big-endian ARM processors generated by versions of the compiler prior to
6548 @item -malignment-traps
6549 @opindex malignment-traps
6550 Generate code that will not trap if the MMU has alignment traps enabled.
6551 On ARM architectures prior to ARMv4, there were no instructions to
6552 access half-word objects stored in memory. However, when reading from
6553 memory a feature of the ARM architecture allows a word load to be used,
6554 even if the address is unaligned, and the processor core will rotate the
6555 data as it is being loaded. This option tells the compiler that such
6556 misaligned accesses will cause a MMU trap and that it should instead
6557 synthesize the access as a series of byte accesses. The compiler can
6558 still use word accesses to load half-word data if it knows that the
6559 address is aligned to a word boundary.
6561 This option is ignored when compiling for ARM architecture 4 or later,
6562 since these processors have instructions to directly access half-word
6565 @item -mno-alignment-traps
6566 @opindex mno-alignment-traps
6567 Generate code that assumes that the MMU will not trap unaligned
6568 accesses. This produces better code when the target instruction set
6569 does not have half-word memory operations (i.e.@: implementations prior to
6572 Note that you cannot use this option to access unaligned word objects,
6573 since the processor will only fetch one 32-bit aligned object from
6576 The default setting for most targets is @option{-mno-alignment-traps}, since
6577 this produces better code when there are no half-word memory
6578 instructions available.
6580 @item -mshort-load-bytes
6581 @itemx -mno-short-load-words
6582 @opindex mshort-load-bytes
6583 @opindex mno-short-load-words
6584 These are deprecated aliases for @option{-malignment-traps}.
6586 @item -mno-short-load-bytes
6587 @itemx -mshort-load-words
6588 @opindex mno-short-load-bytes
6589 @opindex mshort-load-words
6590 This are deprecated aliases for @option{-mno-alignment-traps}.
6592 @item -mcpu=@var{name}
6594 This specifies the name of the target ARM processor. GCC uses this name
6595 to determine what kind of instructions it can emit when generating
6596 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6597 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6598 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6599 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6600 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6601 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6602 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6603 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6604 @samp{arm920t}, @samp{arm926ejs}, @samp{arm940t}, @samp{arm9tdmi},
6605 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ejs},
6606 @samp{arm1136js}, @samp{arm1136jfs} ,@samp{xscale}, @samp{iwmmxt},
6609 @itemx -mtune=@var{name}
6611 This option is very similar to the @option{-mcpu=} option, except that
6612 instead of specifying the actual target processor type, and hence
6613 restricting which instructions can be used, it specifies that GCC should
6614 tune the performance of the code as if the target were of the type
6615 specified in this option, but still choosing the instructions that it
6616 will generate based on the cpu specified by a @option{-mcpu=} option.
6617 For some ARM implementations better performance can be obtained by using
6620 @item -march=@var{name}
6622 This specifies the name of the target ARM architecture. GCC uses this
6623 name to determine what kind of instructions it can emit when generating
6624 assembly code. This option can be used in conjunction with or instead
6625 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6626 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6627 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
6628 @samp{iwmmxt}, @samp{ep9312}.
6630 @item -mfpu=@var{name}
6631 @itemx -mfpe=@var{number}
6632 @itemx -mfp=@var{number}
6636 This specifies what floating point hardware (or hardware emulation) is
6637 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
6638 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
6639 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
6640 with older versions of GCC@.
6642 If @option{-msoft-float} is specified this specifies the format of
6643 floating point values.
6645 @item -mstructure-size-boundary=@var{n}
6646 @opindex mstructure-size-boundary
6647 The size of all structures and unions will be rounded up to a multiple
6648 of the number of bits set by this option. Permissible values are 8, 32
6649 and 64. The default value varies for different toolchains. For the COFF
6650 targeted toolchain the default value is 8. A value of 64 is only allowed
6651 if the underlying ABI supports it.
6653 Specifying the larger number can produce faster, more efficient code, but
6654 can also increase the size of the program. Different values are potentially
6655 incompatible. Code compiled with one value cannot necessarily expect to
6656 work with code or libraries compiled with another value, if they exchange
6657 information using structures or unions.
6659 @item -mabort-on-noreturn
6660 @opindex mabort-on-noreturn
6661 Generate a call to the function @code{abort} at the end of a
6662 @code{noreturn} function. It will be executed if the function tries to
6666 @itemx -mno-long-calls
6667 @opindex mlong-calls
6668 @opindex mno-long-calls
6669 Tells the compiler to perform function calls by first loading the
6670 address of the function into a register and then performing a subroutine
6671 call on this register. This switch is needed if the target function
6672 will lie outside of the 64 megabyte addressing range of the offset based
6673 version of subroutine call instruction.
6675 Even if this switch is enabled, not all function calls will be turned
6676 into long calls. The heuristic is that static functions, functions
6677 which have the @samp{short-call} attribute, functions that are inside
6678 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6679 definitions have already been compiled within the current compilation
6680 unit, will not be turned into long calls. The exception to this rule is
6681 that weak function definitions, functions with the @samp{long-call}
6682 attribute or the @samp{section} attribute, and functions that are within
6683 the scope of a @samp{#pragma long_calls} directive, will always be
6684 turned into long calls.
6686 This feature is not enabled by default. Specifying
6687 @option{-mno-long-calls} will restore the default behavior, as will
6688 placing the function calls within the scope of a @samp{#pragma
6689 long_calls_off} directive. Note these switches have no effect on how
6690 the compiler generates code to handle function calls via function
6693 @item -mnop-fun-dllimport
6694 @opindex mnop-fun-dllimport
6695 Disable support for the @code{dllimport} attribute.
6697 @item -msingle-pic-base
6698 @opindex msingle-pic-base
6699 Treat the register used for PIC addressing as read-only, rather than
6700 loading it in the prologue for each function. The run-time system is
6701 responsible for initializing this register with an appropriate value
6702 before execution begins.
6704 @item -mpic-register=@var{reg}
6705 @opindex mpic-register
6706 Specify the register to be used for PIC addressing. The default is R10
6707 unless stack-checking is enabled, when R9 is used.
6709 @item -mcirrus-fix-invalid-insns
6710 @opindex mcirrus-fix-invalid-insns
6711 @opindex mno-cirrus-fix-invalid-insns
6712 Insert NOPs into the instruction stream to in order to work around
6713 problems with invalid Maverick instruction combinations. This option
6714 is only valid if the @option{-mcpu=ep9312} option has been used to
6715 enable generation of instructions for the Cirrus Maverick floating
6716 point co-processor. This option is not enabled by default, since the
6717 problem is only present in older Maverick implementations. The default
6718 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
6721 @item -mpoke-function-name
6722 @opindex mpoke-function-name
6723 Write the name of each function into the text section, directly
6724 preceding the function prologue. The generated code is similar to this:
6728 .ascii "arm_poke_function_name", 0
6731 .word 0xff000000 + (t1 - t0)
6732 arm_poke_function_name
6734 stmfd sp!, @{fp, ip, lr, pc@}
6738 When performing a stack backtrace, code can inspect the value of
6739 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6740 location @code{pc - 12} and the top 8 bits are set, then we know that
6741 there is a function name embedded immediately preceding this location
6742 and has length @code{((pc[-3]) & 0xff000000)}.
6746 Generate code for the 16-bit Thumb instruction set. The default is to
6747 use the 32-bit ARM instruction set.
6750 @opindex mtpcs-frame
6751 Generate a stack frame that is compliant with the Thumb Procedure Call
6752 Standard for all non-leaf functions. (A leaf function is one that does
6753 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6755 @item -mtpcs-leaf-frame
6756 @opindex mtpcs-leaf-frame
6757 Generate a stack frame that is compliant with the Thumb Procedure Call
6758 Standard for all leaf functions. (A leaf function is one that does
6759 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6761 @item -mcallee-super-interworking
6762 @opindex mcallee-super-interworking
6763 Gives all externally visible functions in the file being compiled an ARM
6764 instruction set header which switches to Thumb mode before executing the
6765 rest of the function. This allows these functions to be called from
6766 non-interworking code.
6768 @item -mcaller-super-interworking
6769 @opindex mcaller-super-interworking
6770 Allows calls via function pointers (including virtual functions) to
6771 execute correctly regardless of whether the target code has been
6772 compiled for interworking or not. There is a small overhead in the cost
6773 of executing a function pointer if this option is enabled.
6777 @node MN10300 Options
6778 @subsection MN10300 Options
6779 @cindex MN10300 options
6781 These @option{-m} options are defined for Matsushita MN10300 architectures:
6786 Generate code to avoid bugs in the multiply instructions for the MN10300
6787 processors. This is the default.
6790 @opindex mno-mult-bug
6791 Do not generate code to avoid bugs in the multiply instructions for the
6796 Generate code which uses features specific to the AM33 processor.
6800 Do not generate code which uses features specific to the AM33 processor. This
6805 Do not link in the C run-time initialization object file.
6809 Indicate to the linker that it should perform a relaxation optimization pass
6810 to shorten branches, calls and absolute memory addresses. This option only
6811 has an effect when used on the command line for the final link step.
6813 This option makes symbolic debugging impossible.
6817 @node M32R/D Options
6818 @subsection M32R/D Options
6819 @cindex M32R/D options
6821 These @option{-m} options are defined for Renesas M32R/D architectures:
6826 Generate code for the M32R/2@.
6830 Generate code for the M32R/X@.
6834 Generate code for the M32R@. This is the default.
6837 @opindex mmodel=small
6838 Assume all objects live in the lower 16MB of memory (so that their addresses
6839 can be loaded with the @code{ld24} instruction), and assume all subroutines
6840 are reachable with the @code{bl} instruction.
6841 This is the default.
6843 The addressability of a particular object can be set with the
6844 @code{model} attribute.
6846 @item -mmodel=medium
6847 @opindex mmodel=medium
6848 Assume objects may be anywhere in the 32-bit address space (the compiler
6849 will generate @code{seth/add3} instructions to load their addresses), and
6850 assume all subroutines are reachable with the @code{bl} instruction.
6853 @opindex mmodel=large
6854 Assume objects may be anywhere in the 32-bit address space (the compiler
6855 will generate @code{seth/add3} instructions to load their addresses), and
6856 assume subroutines may not be reachable with the @code{bl} instruction
6857 (the compiler will generate the much slower @code{seth/add3/jl}
6858 instruction sequence).
6861 @opindex msdata=none
6862 Disable use of the small data area. Variables will be put into
6863 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6864 @code{section} attribute has been specified).
6865 This is the default.
6867 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6868 Objects may be explicitly put in the small data area with the
6869 @code{section} attribute using one of these sections.
6872 @opindex msdata=sdata
6873 Put small global and static data in the small data area, but do not
6874 generate special code to reference them.
6878 Put small global and static data in the small data area, and generate
6879 special instructions to reference them.
6883 @cindex smaller data references
6884 Put global and static objects less than or equal to @var{num} bytes
6885 into the small data or bss sections instead of the normal data or bss
6886 sections. The default value of @var{num} is 8.
6887 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6888 for this option to have any effect.
6890 All modules should be compiled with the same @option{-G @var{num}} value.
6891 Compiling with different values of @var{num} may or may not work; if it
6892 doesn't the linker will give an error message---incorrect code will not be
6897 Makes the M32R specific code in the compiler display some statistics
6898 that might help in debugging programs.
6901 @opindex malign-loops
6902 Align all loops to a 32-byte boundary.
6904 @item -mno-align-loops
6905 @opindex mno-align-loops
6906 Do not enforce a 32-byte alignment for loops. This is the default.
6908 @item -missue-rate=@var{number}
6909 @opindex missue-rate=@var{number}
6910 Issue @var{number} instructions per cycle. @var{number} can only be 1
6913 @item -mbranch-cost=@var{number}
6914 @opindex mbranch-cost=@var{number}
6915 @var{number} can only be 1 or 2. If it is 1 then branches will be
6916 preferred over conditional code, if it is 2, then the opposite will
6919 @item -mflush-trap=@var{number}
6920 @opindex mflush-trap=@var{number}
6921 Specifies the trap number to use to flush the cache. The default is
6922 12. Valid numbers are between 0 and 15 inclusive.
6924 @item -mno-flush-trap
6925 @opindex mno-flush-trap
6926 Specifies that the cache cannot be flushed by using a trap.
6928 @item -mflush-func=@var{name}
6929 @opindex mflush-func=@var{name}
6930 Specifies the name of the operating system function to call to flush
6931 the cache. The default is @emph{_flush_cache}, but a function call
6932 will only be used if a trap is not available.
6934 @item -mno-flush-func
6935 @opindex mno-flush-func
6936 Indicates that there is no OS function for flushing the cache.
6940 @node RS/6000 and PowerPC Options
6941 @subsection IBM RS/6000 and PowerPC Options
6942 @cindex RS/6000 and PowerPC Options
6943 @cindex IBM RS/6000 and PowerPC Options
6945 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6953 @itemx -mpowerpc-gpopt
6954 @itemx -mno-powerpc-gpopt
6955 @itemx -mpowerpc-gfxopt
6956 @itemx -mno-powerpc-gfxopt
6958 @itemx -mno-powerpc64
6964 @opindex mno-powerpc
6965 @opindex mpowerpc-gpopt
6966 @opindex mno-powerpc-gpopt
6967 @opindex mpowerpc-gfxopt
6968 @opindex mno-powerpc-gfxopt
6970 @opindex mno-powerpc64
6971 GCC supports two related instruction set architectures for the
6972 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6973 instructions supported by the @samp{rios} chip set used in the original
6974 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6975 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6976 the IBM 4xx microprocessors.
6978 Neither architecture is a subset of the other. However there is a
6979 large common subset of instructions supported by both. An MQ
6980 register is included in processors supporting the POWER architecture.
6982 You use these options to specify which instructions are available on the
6983 processor you are using. The default value of these options is
6984 determined when configuring GCC@. Specifying the
6985 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6986 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6987 rather than the options listed above.
6989 The @option{-mpower} option allows GCC to generate instructions that
6990 are found only in the POWER architecture and to use the MQ register.
6991 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6992 to generate instructions that are present in the POWER2 architecture but
6993 not the original POWER architecture.
6995 The @option{-mpowerpc} option allows GCC to generate instructions that
6996 are found only in the 32-bit subset of the PowerPC architecture.
6997 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6998 GCC to use the optional PowerPC architecture instructions in the
6999 General Purpose group, including floating-point square root. Specifying
7000 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
7001 use the optional PowerPC architecture instructions in the Graphics
7002 group, including floating-point select.
7004 The @option{-mpowerpc64} option allows GCC to generate the additional
7005 64-bit instructions that are found in the full PowerPC64 architecture
7006 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
7007 @option{-mno-powerpc64}.
7009 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
7010 will use only the instructions in the common subset of both
7011 architectures plus some special AIX common-mode calls, and will not use
7012 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
7013 permits GCC to use any instruction from either architecture and to
7014 allow use of the MQ register; specify this for the Motorola MPC601.
7016 @item -mnew-mnemonics
7017 @itemx -mold-mnemonics
7018 @opindex mnew-mnemonics
7019 @opindex mold-mnemonics
7020 Select which mnemonics to use in the generated assembler code. With
7021 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
7022 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
7023 assembler mnemonics defined for the POWER architecture. Instructions
7024 defined in only one architecture have only one mnemonic; GCC uses that
7025 mnemonic irrespective of which of these options is specified.
7027 GCC defaults to the mnemonics appropriate for the architecture in
7028 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
7029 value of these option. Unless you are building a cross-compiler, you
7030 should normally not specify either @option{-mnew-mnemonics} or
7031 @option{-mold-mnemonics}, but should instead accept the default.
7033 @item -mcpu=@var{cpu_type}
7035 Set architecture type, register usage, choice of mnemonics, and
7036 instruction scheduling parameters for machine type @var{cpu_type}.
7037 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
7038 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
7039 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
7040 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
7041 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
7042 @samp{860}, @samp{970}, @samp{common}, @samp{ec603e}, @samp{G3},
7043 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
7044 @samp{power4}, @samp{powerpc}, @samp{powerpc64}, @samp{rios},
7045 @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64a}.
7047 @option{-mcpu=common} selects a completely generic processor. Code
7048 generated under this option will run on any POWER or PowerPC processor.
7049 GCC will use only the instructions in the common subset of both
7050 architectures, and will not use the MQ register. GCC assumes a generic
7051 processor model for scheduling purposes.
7053 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
7054 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
7055 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
7056 types, with an appropriate, generic processor model assumed for
7057 scheduling purposes.
7059 The other options specify a specific processor. Code generated under
7060 those options will run best on that processor, and may not run at all on
7063 The @option{-mcpu} options automatically enable or disable the
7064 following options: @option{-maltivec}, @option{-mhard-float},
7065 @option{-mmfcrf}, @option{-mmultiple}, @option{-mnew-mnemonics},
7066 @option{-mpower}, @option{-mpower2}, @option{-mpowerpc64},
7067 @option{-mpowerpc-gpopt}, @option{-mpowerpc-gfxopt},
7068 @option{-mstring}. The particular options set for any particular CPU
7069 will vary between compiler versions, depending on what setting seems
7070 to produce optimal code for that CPU; it doesn't necessarily reflect
7071 the actual hardware's capabilities. If you wish to set an individual
7072 option to a particular value, you may specify it after the
7073 @option{-mcpu} option, like @samp{-mcpu=970 -mno-altivec}.
7075 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
7076 not enabled or disabled by the @option{-mcpu} option at present, since
7077 AIX does not have full support for these options. You may still
7078 enable or disable them individually if you're sure it'll work in your
7081 @item -mtune=@var{cpu_type}
7083 Set the instruction scheduling parameters for machine type
7084 @var{cpu_type}, but do not set the architecture type, register usage, or
7085 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
7086 values for @var{cpu_type} are used for @option{-mtune} as for
7087 @option{-mcpu}. If both are specified, the code generated will use the
7088 architecture, registers, and mnemonics set by @option{-mcpu}, but the
7089 scheduling parameters set by @option{-mtune}.
7094 @opindex mno-altivec
7095 These switches enable or disable the use of built-in functions that
7096 allow access to the AltiVec instruction set. You may also need to set
7097 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
7102 Extend the current ABI with SPE ABI extensions. This does not change
7103 the default ABI, instead it adds the SPE ABI extensions to the current
7107 @opindex mabi=no-spe
7108 Disable Booke SPE ABI extensions for the current ABI.
7110 @item -misel=@var{yes/no}
7113 This switch enables or disables the generation of ISEL instructions.
7115 @item -mspe=@var{yes/no}
7118 This switch enables or disables the generation of SPE simd
7121 @item -mfloat-gprs=@var{yes/no}
7123 @opindex mfloat-gprs
7124 This switch enables or disables the generation of floating point
7125 operations on the general purpose registers for architectures that
7126 support it. This option is currently only available on the MPC8540.
7129 @itemx -mno-fp-in-toc
7130 @itemx -mno-sum-in-toc
7131 @itemx -mminimal-toc
7133 @opindex mno-fp-in-toc
7134 @opindex mno-sum-in-toc
7135 @opindex mminimal-toc
7136 Modify generation of the TOC (Table Of Contents), which is created for
7137 every executable file. The @option{-mfull-toc} option is selected by
7138 default. In that case, GCC will allocate at least one TOC entry for
7139 each unique non-automatic variable reference in your program. GCC
7140 will also place floating-point constants in the TOC@. However, only
7141 16,384 entries are available in the TOC@.
7143 If you receive a linker error message that saying you have overflowed
7144 the available TOC space, you can reduce the amount of TOC space used
7145 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
7146 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
7147 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
7148 generate code to calculate the sum of an address and a constant at
7149 run-time instead of putting that sum into the TOC@. You may specify one
7150 or both of these options. Each causes GCC to produce very slightly
7151 slower and larger code at the expense of conserving TOC space.
7153 If you still run out of space in the TOC even when you specify both of
7154 these options, specify @option{-mminimal-toc} instead. This option causes
7155 GCC to make only one TOC entry for every file. When you specify this
7156 option, GCC will produce code that is slower and larger but which
7157 uses extremely little TOC space. You may wish to use this option
7158 only on files that contain less frequently executed code.
7164 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
7165 @code{long} type, and the infrastructure needed to support them.
7166 Specifying @option{-maix64} implies @option{-mpowerpc64} and
7167 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
7168 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
7173 @opindex mno-xl-call
7174 On AIX, pass floating-point arguments to prototyped functions beyond the
7175 register save area (RSA) on the stack in addition to argument FPRs. The
7176 AIX calling convention was extended but not initially documented to
7177 handle an obscure K&R C case of calling a function that takes the
7178 address of its arguments with fewer arguments than declared. AIX XL
7179 compilers access floating point arguments which do not fit in the
7180 RSA from the stack when a subroutine is compiled without
7181 optimization. Because always storing floating-point arguments on the
7182 stack is inefficient and rarely needed, this option is not enabled by
7183 default and only is necessary when calling subroutines compiled by AIX
7184 XL compilers without optimization.
7188 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
7189 application written to use message passing with special startup code to
7190 enable the application to run. The system must have PE installed in the
7191 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
7192 must be overridden with the @option{-specs=} option to specify the
7193 appropriate directory location. The Parallel Environment does not
7194 support threads, so the @option{-mpe} option and the @option{-pthread}
7195 option are incompatible.
7197 @item -malign-natural
7198 @itemx -malign-power
7199 @opindex malign-natural
7200 @opindex malign-power
7201 On AIX, Darwin, and 64-bit PowerPC GNU/Linux, the option
7202 @option{-malign-natural} overrides the ABI-defined alignment of larger
7203 types, such as floating-point doubles, on their natural size-based boundary.
7204 The option @option{-malign-power} instructs GCC to follow the ABI-specified
7205 alignment rules. GCC defaults to the standard alignment defined in the ABI.
7209 @opindex msoft-float
7210 @opindex mhard-float
7211 Generate code that does not use (uses) the floating-point register set.
7212 Software floating point emulation is provided if you use the
7213 @option{-msoft-float} option, and pass the option to GCC when linking.
7216 @itemx -mno-multiple
7218 @opindex mno-multiple
7219 Generate code that uses (does not use) the load multiple word
7220 instructions and the store multiple word instructions. These
7221 instructions are generated by default on POWER systems, and not
7222 generated on PowerPC systems. Do not use @option{-mmultiple} on little
7223 endian PowerPC systems, since those instructions do not work when the
7224 processor is in little endian mode. The exceptions are PPC740 and
7225 PPC750 which permit the instructions usage in little endian mode.
7231 Generate code that uses (does not use) the load string instructions
7232 and the store string word instructions to save multiple registers and
7233 do small block moves. These instructions are generated by default on
7234 POWER systems, and not generated on PowerPC systems. Do not use
7235 @option{-mstring} on little endian PowerPC systems, since those
7236 instructions do not work when the processor is in little endian mode.
7237 The exceptions are PPC740 and PPC750 which permit the instructions
7238 usage in little endian mode.
7244 Generate code that uses (does not use) the load or store instructions
7245 that update the base register to the address of the calculated memory
7246 location. These instructions are generated by default. If you use
7247 @option{-mno-update}, there is a small window between the time that the
7248 stack pointer is updated and the address of the previous frame is
7249 stored, which means code that walks the stack frame across interrupts or
7250 signals may get corrupted data.
7253 @itemx -mno-fused-madd
7254 @opindex mfused-madd
7255 @opindex mno-fused-madd
7256 Generate code that uses (does not use) the floating point multiply and
7257 accumulate instructions. These instructions are generated by default if
7258 hardware floating is used.
7260 @item -mno-bit-align
7262 @opindex mno-bit-align
7264 On System V.4 and embedded PowerPC systems do not (do) force structures
7265 and unions that contain bit-fields to be aligned to the base type of the
7268 For example, by default a structure containing nothing but 8
7269 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
7270 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
7271 the structure would be aligned to a 1 byte boundary and be one byte in
7274 @item -mno-strict-align
7275 @itemx -mstrict-align
7276 @opindex mno-strict-align
7277 @opindex mstrict-align
7278 On System V.4 and embedded PowerPC systems do not (do) assume that
7279 unaligned memory references will be handled by the system.
7282 @itemx -mno-relocatable
7283 @opindex mrelocatable
7284 @opindex mno-relocatable
7285 On embedded PowerPC systems generate code that allows (does not allow)
7286 the program to be relocated to a different address at runtime. If you
7287 use @option{-mrelocatable} on any module, all objects linked together must
7288 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
7290 @item -mrelocatable-lib
7291 @itemx -mno-relocatable-lib
7292 @opindex mrelocatable-lib
7293 @opindex mno-relocatable-lib
7294 On embedded PowerPC systems generate code that allows (does not allow)
7295 the program to be relocated to a different address at runtime. Modules
7296 compiled with @option{-mrelocatable-lib} can be linked with either modules
7297 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
7298 with modules compiled with the @option{-mrelocatable} options.
7304 On System V.4 and embedded PowerPC systems do not (do) assume that
7305 register 2 contains a pointer to a global area pointing to the addresses
7306 used in the program.
7309 @itemx -mlittle-endian
7311 @opindex mlittle-endian
7312 On System V.4 and embedded PowerPC systems compile code for the
7313 processor in little endian mode. The @option{-mlittle-endian} option is
7314 the same as @option{-mlittle}.
7319 @opindex mbig-endian
7320 On System V.4 and embedded PowerPC systems compile code for the
7321 processor in big endian mode. The @option{-mbig-endian} option is
7322 the same as @option{-mbig}.
7324 @item -mdynamic-no-pic
7325 @opindex mdynamic-no-pic
7326 On Darwin and Mac OS X systems, compile code so that it is not
7327 relocatable, but that its external references are relocatable. The
7328 resulting code is suitable for applications, but not shared
7331 @item -mprioritize-restricted-insns=@var{priority}
7332 @opindex mprioritize-restricted-insns
7333 This option controls the priority that is assigned to
7334 dispatch-slot restricted instructions during the second scheduling
7335 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
7336 @var{no/highest/second-highest} priority to dispatch slot restricted
7339 @item -msched-costly-dep=@var{dependence_type}
7340 @opindex msched-costly-dep
7341 This option controls which dependences are considered costly
7342 by the target during instruction scheduling. The argument
7343 @var{dependence_type} takes one of the following values:
7344 @var{no}: no dependence is costly,
7345 @var{all}: all dependences are costly,
7346 @var{true_store_to_load}: a true dependence from store to load is costly,
7347 @var{store_to_load}: any dependence from store to load is costly,
7348 @var{number}: any dependence which latency >= @var{number} is costly.
7350 @item -minsert-sched-nops=@var{scheme}
7351 @opindex minsert-sched-nops
7352 This option controls which nop insertion scheme will be used during
7353 the second scheduling pass. The argument @var{scheme} takes one of the
7355 @var{no}: Don't insert nops.
7356 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
7357 according to the scheduler's grouping.
7358 @var{regroup_exact}: Insert nops to force costly dependent insns into
7359 separate groups. Insert exactly as many nops as needed to force an insn
7360 to a new group, according to the estimated processor grouping.
7361 @var{number}: Insert nops to force costly dependent insns into
7362 separate groups. Insert @var{number} nops to force an insn to a new group.
7366 On System V.4 and embedded PowerPC systems compile code using calling
7367 conventions that adheres to the March 1995 draft of the System V
7368 Application Binary Interface, PowerPC processor supplement. This is the
7369 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
7371 @item -mcall-sysv-eabi
7372 @opindex mcall-sysv-eabi
7373 Specify both @option{-mcall-sysv} and @option{-meabi} options.
7375 @item -mcall-sysv-noeabi
7376 @opindex mcall-sysv-noeabi
7377 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
7379 @item -mcall-solaris
7380 @opindex mcall-solaris
7381 On System V.4 and embedded PowerPC systems compile code for the Solaris
7385 @opindex mcall-linux
7386 On System V.4 and embedded PowerPC systems compile code for the
7387 Linux-based GNU system.
7391 On System V.4 and embedded PowerPC systems compile code for the
7392 Hurd-based GNU system.
7395 @opindex mcall-netbsd
7396 On System V.4 and embedded PowerPC systems compile code for the
7397 NetBSD operating system.
7399 @item -maix-struct-return
7400 @opindex maix-struct-return
7401 Return all structures in memory (as specified by the AIX ABI)@.
7403 @item -msvr4-struct-return
7404 @opindex msvr4-struct-return
7405 Return structures smaller than 8 bytes in registers (as specified by the
7409 @opindex mabi=altivec
7410 Extend the current ABI with AltiVec ABI extensions. This does not
7411 change the default ABI, instead it adds the AltiVec ABI extensions to
7414 @item -mabi=no-altivec
7415 @opindex mabi=no-altivec
7416 Disable AltiVec ABI extensions for the current ABI.
7419 @itemx -mno-prototype
7421 @opindex mno-prototype
7422 On System V.4 and embedded PowerPC systems assume that all calls to
7423 variable argument functions are properly prototyped. Otherwise, the
7424 compiler must insert an instruction before every non prototyped call to
7425 set or clear bit 6 of the condition code register (@var{CR}) to
7426 indicate whether floating point values were passed in the floating point
7427 registers in case the function takes a variable arguments. With
7428 @option{-mprototype}, only calls to prototyped variable argument functions
7429 will set or clear the bit.
7433 On embedded PowerPC systems, assume that the startup module is called
7434 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
7435 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
7440 On embedded PowerPC systems, assume that the startup module is called
7441 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
7446 On embedded PowerPC systems, assume that the startup module is called
7447 @file{crt0.o} and the standard C libraries are @file{libads.a} and
7451 @opindex myellowknife
7452 On embedded PowerPC systems, assume that the startup module is called
7453 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
7458 On System V.4 and embedded PowerPC systems, specify that you are
7459 compiling for a VxWorks system.
7463 Specify that you are compiling for the WindISS simulation environment.
7467 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7468 header to indicate that @samp{eabi} extended relocations are used.
7474 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7475 Embedded Applications Binary Interface (eabi) which is a set of
7476 modifications to the System V.4 specifications. Selecting @option{-meabi}
7477 means that the stack is aligned to an 8 byte boundary, a function
7478 @code{__eabi} is called to from @code{main} to set up the eabi
7479 environment, and the @option{-msdata} option can use both @code{r2} and
7480 @code{r13} to point to two separate small data areas. Selecting
7481 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7482 do not call an initialization function from @code{main}, and the
7483 @option{-msdata} option will only use @code{r13} to point to a single
7484 small data area. The @option{-meabi} option is on by default if you
7485 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7488 @opindex msdata=eabi
7489 On System V.4 and embedded PowerPC systems, put small initialized
7490 @code{const} global and static data in the @samp{.sdata2} section, which
7491 is pointed to by register @code{r2}. Put small initialized
7492 non-@code{const} global and static data in the @samp{.sdata} section,
7493 which is pointed to by register @code{r13}. Put small uninitialized
7494 global and static data in the @samp{.sbss} section, which is adjacent to
7495 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7496 incompatible with the @option{-mrelocatable} option. The
7497 @option{-msdata=eabi} option also sets the @option{-memb} option.
7500 @opindex msdata=sysv
7501 On System V.4 and embedded PowerPC systems, put small global and static
7502 data in the @samp{.sdata} section, which is pointed to by register
7503 @code{r13}. Put small uninitialized global and static data in the
7504 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7505 The @option{-msdata=sysv} option is incompatible with the
7506 @option{-mrelocatable} option.
7508 @item -msdata=default
7510 @opindex msdata=default
7512 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7513 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7514 same as @option{-msdata=sysv}.
7517 @opindex msdata-data
7518 On System V.4 and embedded PowerPC systems, put small global and static
7519 data in the @samp{.sdata} section. Put small uninitialized global and
7520 static data in the @samp{.sbss} section. Do not use register @code{r13}
7521 to address small data however. This is the default behavior unless
7522 other @option{-msdata} options are used.
7526 @opindex msdata=none
7528 On embedded PowerPC systems, put all initialized global and static data
7529 in the @samp{.data} section, and all uninitialized data in the
7530 @samp{.bss} section.
7534 @cindex smaller data references (PowerPC)
7535 @cindex .sdata/.sdata2 references (PowerPC)
7536 On embedded PowerPC systems, put global and static items less than or
7537 equal to @var{num} bytes into the small data or bss sections instead of
7538 the normal data or bss section. By default, @var{num} is 8. The
7539 @option{-G @var{num}} switch is also passed to the linker.
7540 All modules should be compiled with the same @option{-G @var{num}} value.
7543 @itemx -mno-regnames
7545 @opindex mno-regnames
7546 On System V.4 and embedded PowerPC systems do (do not) emit register
7547 names in the assembly language output using symbolic forms.
7550 @itemx -mno-longcall
7552 @opindex mno-longcall
7553 Default to making all function calls indirectly, using a register, so
7554 that functions which reside further than 32 megabytes (33,554,432
7555 bytes) from the current location can be called. This setting can be
7556 overridden by the @code{shortcall} function attribute, or by
7557 @code{#pragma longcall(0)}.
7559 Some linkers are capable of detecting out-of-range calls and generating
7560 glue code on the fly. On these systems, long calls are unnecessary and
7561 generate slower code. As of this writing, the AIX linker can do this,
7562 as can the GNU linker for PowerPC/64. It is planned to add this feature
7563 to the GNU linker for 32-bit PowerPC systems as well.
7565 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
7566 callee, L42'', plus a ``branch island'' (glue code). The two target
7567 addresses represent the callee and the ``branch island.'' The
7568 Darwin/PPC linker will prefer the first address and generate a ``bl
7569 callee'' if the PPC ``bl'' instruction will reach the callee directly;
7570 otherwise, the linker will generate ``bl L42'' to call the ``branch
7571 island.'' The ``branch island'' is appended to the body of the
7572 calling function; it computes the full 32-bit address of the callee
7575 On Mach-O (Darwin) systems, this option directs the compiler emit to
7576 the glue for every direct call, and the Darwin linker decides whether
7577 to use or discard it.
7579 In the future, we may cause GCC to ignore all longcall specifications
7580 when the linker is known to generate glue.
7584 Adds support for multithreading with the @dfn{pthreads} library.
7585 This option sets flags for both the preprocessor and linker.
7589 @node Darwin Options
7590 @subsection Darwin Options
7591 @cindex Darwin options
7593 These options are defined for all architectures running the Darwin operating
7594 system. They are useful for compatibility with other Mac OS compilers.
7599 Loads all members of static archive libraries.
7600 See man ld(1) for more information.
7602 @item -arch_errors_fatal
7603 @opindex arch_errors_fatal
7604 Cause the errors having to do with files that have the wrong architecture
7608 @opindex bind_at_load
7609 Causes the output file to be marked such that the dynamic linker will
7610 bind all undefined references when the file is loaded or launched.
7614 Produce a Mach-o bundle format file.
7615 See man ld(1) for more information.
7617 @item -bundle_loader @var{executable}
7618 @opindex bundle_loader
7619 This specifies the @var{executable} that will be loading the build
7620 output file being linked. See man ld(1) for more information.
7622 @item -allowable_client @var{client_name}
7626 @itemx -compatibility_version
7627 @itemx -current_version
7628 @itemx -dependency-file
7630 @itemx -dylinker_install_name
7633 @itemx -exported_symbols_list
7635 @itemx -flat_namespace
7636 @itemx -force_cpusubtype_ALL
7637 @itemx -force_flat_namespace
7638 @itemx -headerpad_max_install_names
7641 @itemx -install_name
7642 @itemx -keep_private_externs
7643 @itemx -multi_module
7644 @itemx -multiply_defined
7645 @itemx -multiply_defined_unused
7647 @itemx -nofixprebinding
7650 @itemx -noseglinkedit
7651 @itemx -pagezero_size
7653 @itemx -prebind_all_twolevel_modules
7654 @itemx -private_bundle
7655 @itemx -read_only_relocs
7657 @itemx -sectobjectsymbols
7661 @itemx -sectobjectsymbols
7663 @itemx -seg_addr_table
7664 @itemx -seg_addr_table_filename
7667 @itemx -segs_read_only_addr
7668 @itemx -segs_read_write_addr
7669 @itemx -single_module
7672 @itemx -sub_umbrella
7673 @itemx -twolevel_namespace
7676 @itemx -unexported_symbols_list
7677 @itemx -weak_reference_mismatches
7680 @opindex allowable_client
7682 @opindex client_name
7683 @opindex compatibility_version
7684 @opindex current_version
7685 @opindex dependency-file
7687 @opindex dylinker_install_name
7690 @opindex exported_symbols_list
7692 @opindex flat_namespace
7693 @opindex force_cpusubtype_ALL
7694 @opindex force_flat_namespace
7695 @opindex headerpad_max_install_names
7698 @opindex install_name
7699 @opindex keep_private_externs
7700 @opindex multi_module
7701 @opindex multiply_defined
7702 @opindex multiply_defined_unused
7704 @opindex nofixprebinding
7705 @opindex nomultidefs
7707 @opindex noseglinkedit
7708 @opindex pagezero_size
7710 @opindex prebind_all_twolevel_modules
7711 @opindex private_bundle
7712 @opindex read_only_relocs
7714 @opindex sectobjectsymbols
7718 @opindex sectobjectsymbols
7720 @opindex seg_addr_table
7721 @opindex seg_addr_table_filename
7722 @opindex seglinkedit
7724 @opindex segs_read_only_addr
7725 @opindex segs_read_write_addr
7726 @opindex single_module
7728 @opindex sub_library
7729 @opindex sub_umbrella
7730 @opindex twolevel_namespace
7733 @opindex unexported_symbols_list
7734 @opindex weak_reference_mismatches
7735 @opindex whatsloaded
7737 These options are available for Darwin linker. Darwin linker man page
7738 describes them in detail.
7743 @subsection MIPS Options
7744 @cindex MIPS options
7750 Generate big-endian code.
7754 Generate little-endian code. This is the default for @samp{mips*el-*-*}
7757 @item -march=@var{arch}
7759 Generate code that will run on @var{arch}, which can be the name of a
7760 generic MIPS ISA, or the name of a particular processor.
7762 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
7763 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
7764 The processor names are:
7765 @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
7767 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
7768 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000}, @samp{rm7000},
7772 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
7773 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
7774 The special value @samp{from-abi} selects the
7775 most compatible architecture for the selected ABI (that is,
7776 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
7778 In processor names, a final @samp{000} can be abbreviated as @samp{k}
7779 (for example, @samp{-march=r2k}). Prefixes are optional, and
7780 @samp{vr} may be written @samp{r}.
7782 GCC defines two macros based on the value of this option. The first
7783 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
7784 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
7785 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
7786 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
7787 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
7789 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
7790 above. In other words, it will have the full prefix and will not
7791 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
7792 the macro names the resolved architecture (either @samp{"mips1"} or
7793 @samp{"mips3"}). It names the default architecture when no
7794 @option{-march} option is given.
7796 @item -mtune=@var{arch}
7798 Optimize for @var{arch}. Among other things, this option controls
7799 the way instructions are scheduled, and the perceived cost of arithmetic
7800 operations. The list of @var{arch} values is the same as for
7803 When this option is not used, GCC will optimize for the processor
7804 specified by @option{-march}. By using @option{-march} and
7805 @option{-mtune} together, it is possible to generate code that will
7806 run on a family of processors, but optimize the code for one
7807 particular member of that family.
7809 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
7810 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
7811 @samp{-march} ones described above.
7815 Equivalent to @samp{-march=mips1}.
7819 Equivalent to @samp{-march=mips2}.
7823 Equivalent to @samp{-march=mips3}.
7827 Equivalent to @samp{-march=mips4}.
7831 Equivalent to @samp{-march=mips32}.
7835 Equivalent to @samp{-march=mips32r2}.
7839 Equivalent to @samp{-march=mips64}.
7845 Use (do not use) the MIPS16 ISA.
7857 Generate code for the given ABI@.
7859 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
7860 generates 64-bit code when you select a 64-bit architecture, but you
7861 can use @option{-mgp32} to get 32-bit code instead.
7864 @itemx -mno-abicalls
7866 @opindex mno-abicalls
7867 Generate (do not generate) SVR4-style position-independent code.
7868 @option{-mabicalls} is the default for SVR4-based systems.
7874 Lift (do not lift) the usual restrictions on the size of the global
7877 GCC normally uses a single instruction to load values from the GOT.
7878 While this is relatively efficient, it will only work if the GOT
7879 is smaller than about 64k. Anything larger will cause the linker
7880 to report an error such as:
7882 @cindex relocation truncated to fit (MIPS)
7884 relocation truncated to fit: R_MIPS_GOT16 foobar
7887 If this happens, you should recompile your code with @option{-mxgot}.
7888 It should then work with very large GOTs, although it will also be
7889 less efficient, since it will take three instructions to fetch the
7890 value of a global symbol.
7892 Note that some linkers can create multiple GOTs. If you have such a
7893 linker, you should only need to use @option{-mxgot} when a single object
7894 file accesses more than 64k's worth of GOT entries. Very few do.
7896 These options have no effect unless GCC is generating position
7899 @item -membedded-pic
7900 @itemx -mno-embedded-pic
7901 @opindex membedded-pic
7902 @opindex mno-embedded-pic
7903 Generate (do not generate) position-independent code suitable for some
7904 embedded systems. All calls are made using PC relative addresses, and
7905 all data is addressed using the $gp register. No more than 65536
7906 bytes of global data may be used. This requires GNU as and GNU ld,
7907 which do most of the work.
7911 Assume that general-purpose registers are 32 bits wide.
7915 Assume that general-purpose registers are 64 bits wide.
7919 Assume that floating-point registers are 32 bits wide.
7923 Assume that floating-point registers are 64 bits wide.
7926 @opindex mhard-float
7927 Use floating-point coprocessor instructions.
7930 @opindex msoft-float
7931 Do not use floating-point coprocessor instructions. Implement
7932 floating-point calculations using library calls instead.
7934 @item -msingle-float
7935 @opindex msingle-float
7936 Assume that the floating-point coprocessor only supports single-precision
7939 @itemx -mdouble-float
7940 @opindex mdouble-float
7941 Assume that the floating-point coprocessor supports double-precision
7942 operations. This is the default.
7946 Force @code{int} and @code{long} types to be 64 bits wide. See
7947 @option{-mlong32} for an explanation of the default and the way
7948 that the pointer size is determined.
7952 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
7953 an explanation of the default and the way that the pointer size is
7958 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
7960 The default size of @code{int}s, @code{long}s and pointers depends on
7961 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
7962 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
7963 32-bit @code{long}s. Pointers are the same size as @code{long}s,
7964 or the same size as integer registers, whichever is smaller.
7968 @cindex smaller data references (MIPS)
7969 @cindex gp-relative references (MIPS)
7970 Put global and static items less than or equal to @var{num} bytes into
7971 the small data or bss section instead of the normal data or bss section.
7972 This allows the data to be accessed using a single instruction.
7974 All modules should be compiled with the same @option{-G @var{num}}
7977 @item -membedded-data
7978 @itemx -mno-embedded-data
7979 @opindex membedded-data
7980 @opindex mno-embedded-data
7981 Allocate variables to the read-only data section first if possible, then
7982 next in the small data section if possible, otherwise in data. This gives
7983 slightly slower code than the default, but reduces the amount of RAM required
7984 when executing, and thus may be preferred for some embedded systems.
7986 @item -muninit-const-in-rodata
7987 @itemx -mno-uninit-const-in-rodata
7988 @opindex muninit-const-in-rodata
7989 @opindex mno-uninit-const-in-rodata
7990 Put uninitialized @code{const} variables in the read-only data section.
7991 This option is only meaningful in conjunction with @option{-membedded-data}.
7993 @item -msplit-addresses
7994 @itemx -mno-split-addresses
7995 @opindex msplit-addresses
7996 @opindex mno-split-addresses
7997 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
7998 relocation operators. This option has been superceded by
7999 @option{-mexplicit-relocs} but is retained for backwards compatibility.
8001 @item -mexplicit-relocs
8002 @itemx -mno-explicit-relocs
8003 @opindex mexplicit-relocs
8004 @opindex mno-explicit-relocs
8005 Use (do not use) assembler relocation operators when dealing with symbolic
8006 addresses. The alternative, selected by @option{-mno-explicit-relocs},
8007 is to use assembler macros instead.
8009 @option{-mexplicit-relocs} is usually the default if GCC was configured
8010 to use an assembler that supports relocation operators. However, the
8011 combination of @option{-mabicalls} and @option{-fno-unit-at-a-time}
8012 implies @option{-mno-explicit-relocs} unless explicitly overridden.
8013 This is because, when generating abicalls, the choice of relocation
8014 depends on whether a symbol is local or global. In some rare cases,
8015 GCC will not be able to decide this until the whole compilation unit
8022 Generate (do not generate) code that refers to registers using their
8023 software names. The default is @option{-mno-rnames}, which tells GCC
8024 to use hardware names like @samp{$4} instead of software names like
8025 @samp{a0}. The only assembler known to support @option{-rnames} is
8026 the Algorithmics assembler.
8028 @item -mcheck-zero-division
8029 @itemx -mno-check-zero-division
8030 @opindex mcheck-zero-division
8031 @opindex mno-check-zero-division
8032 Trap (do not trap) on integer division by zero. The default is
8033 @option{-mcheck-zero-division}.
8039 Force (do not force) the use of @code{memcpy()} for non-trivial block
8040 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
8041 most constant-sized copies.
8044 @itemx -mno-long-calls
8045 @opindex mlong-calls
8046 @opindex mno-long-calls
8047 Disable (do not disable) use of the @code{jal} instruction. Calling
8048 functions using @code{jal} is more efficient but requires the caller
8049 and callee to be in the same 256 megabyte segment.
8051 This option has no effect on abicalls code. The default is
8052 @option{-mno-long-calls}.
8058 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
8059 instructions, as provided by the R4650 ISA.
8062 @itemx -mno-fused-madd
8063 @opindex mfused-madd
8064 @opindex mno-fused-madd
8065 Enable (disable) use of the floating point multiply-accumulate
8066 instructions, when they are available. The default is
8067 @option{-mfused-madd}.
8069 When multiply-accumulate instructions are used, the intermediate
8070 product is calculated to infinite precision and is not subject to
8071 the FCSR Flush to Zero bit. This may be undesirable in some
8076 Tell the MIPS assembler to not run its preprocessor over user
8077 assembler files (with a @samp{.s} suffix) when assembling them.
8080 @itemx -mno-fix-r4000
8082 @opindex mno-fix-r4000
8083 Work around certain R4000 CPU errata:
8086 A double-word or a variable shift may give an incorrect result if executed
8087 immediately after starting an integer division.
8089 A double-word or a variable shift may give an incorrect result if executed
8090 while an integer multiplication is in progress.
8092 An integer division may give an incorrect result if started in a delay slot
8093 of a taken branch or a jump.
8097 @itemx -mno-fix-r4400
8099 @opindex mno-fix-r4400
8100 Work around certain R4400 CPU errata:
8103 A double-word or a variable shift may give an incorrect result if executed
8104 immediately after starting an integer division.
8107 @item -mfix-vr4122-bugs
8108 @itemx -mno-fix-vr4122-bugs
8109 @opindex mfix-vr4122-bugs
8110 Work around certain VR4122 errata:
8113 @code{dmultu} does not always produce the correct result.
8115 @code{div} and @code{ddiv} do not always produce the correct result if one
8116 of the operands is negative.
8118 The workarounds for the division errata rely on special functions in
8119 @file{libgcc.a}. At present, these functions are only provided by
8120 the @code{mips64vr*-elf} configurations.
8122 Other VR4122 errata require a nop to be inserted between certain pairs of
8123 instructions. These errata are handled by the assembler, not by GCC itself.
8128 Work around certain SB-1 CPU core errata.
8129 (This flag currently works around the SB-1 revision 2
8130 ``F1'' and ``F2'' floating point errata.)
8132 @item -mflush-func=@var{func}
8133 @itemx -mno-flush-func
8134 @opindex mflush-func
8135 Specifies the function to call to flush the I and D caches, or to not
8136 call any such function. If called, the function must take the same
8137 arguments as the common @code{_flush_func()}, that is, the address of the
8138 memory range for which the cache is being flushed, the size of the
8139 memory range, and the number 3 (to flush both caches). The default
8140 depends on the target GCC was configured for, but commonly is either
8141 @samp{_flush_func} or @samp{__cpu_flush}.
8143 @item -mbranch-likely
8144 @itemx -mno-branch-likely
8145 @opindex mbranch-likely
8146 @opindex mno-branch-likely
8147 Enable or disable use of Branch Likely instructions, regardless of the
8148 default for the selected architecture. By default, Branch Likely
8149 instructions may be generated if they are supported by the selected
8150 architecture. An exception is for the MIPS32 and MIPS64 architectures
8151 and processors which implement those architectures; for those, Branch
8152 Likely instructions will not be generated by default because the MIPS32
8153 and MIPS64 architectures specifically deprecate their use.
8156 @node i386 and x86-64 Options
8157 @subsection Intel 386 and AMD x86-64 Options
8158 @cindex i386 Options
8159 @cindex x86-64 Options
8160 @cindex Intel 386 Options
8161 @cindex AMD x86-64 Options
8163 These @samp{-m} options are defined for the i386 and x86-64 family of
8167 @item -mtune=@var{cpu-type}
8169 Tune to @var{cpu-type} everything applicable about the generated code, except
8170 for the ABI and the set of available instructions. The choices for
8174 Original Intel's i386 CPU.
8176 Intel's i486 CPU. (No scheduling is implemented for this chip.)
8178 Intel Pentium CPU with no MMX support.
8180 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
8181 @item i686, pentiumpro
8182 Intel PentiumPro CPU.
8184 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
8185 @item pentium3, pentium3m
8186 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
8189 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
8190 support. Used by Centrino notebooks.
8191 @item pentium4, pentium4m
8192 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
8194 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
8197 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
8198 SSE2 and SSE3 instruction set support.
8200 AMD K6 CPU with MMX instruction set support.
8202 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
8203 @item athlon, athlon-tbird
8204 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
8206 @item athlon-4, athlon-xp, athlon-mp
8207 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
8208 instruction set support.
8209 @item k8, opteron, athlon64, athlon-fx
8210 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
8211 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
8213 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
8216 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
8217 instruction set support.
8219 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
8220 implemented for this chip.)
8222 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
8223 implemented for this chip.)
8226 While picking a specific @var{cpu-type} will schedule things appropriately
8227 for that particular chip, the compiler will not generate any code that
8228 does not run on the i386 without the @option{-march=@var{cpu-type}} option
8231 @item -march=@var{cpu-type}
8233 Generate instructions for the machine type @var{cpu-type}. The choices
8234 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
8235 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
8237 @item -mcpu=@var{cpu-type}
8239 A deprecated synonym for @option{-mtune}.
8248 @opindex mpentiumpro
8249 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
8250 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
8251 These synonyms are deprecated.
8253 @item -mfpmath=@var{unit}
8255 Generate floating point arithmetics for selected unit @var{unit}. The choices
8260 Use the standard 387 floating point coprocessor present majority of chips and
8261 emulated otherwise. Code compiled with this option will run almost everywhere.
8262 The temporary results are computed in 80bit precision instead of precision
8263 specified by the type resulting in slightly different results compared to most
8264 of other chips. See @option{-ffloat-store} for more detailed description.
8266 This is the default choice for i386 compiler.
8269 Use scalar floating point instructions present in the SSE instruction set.
8270 This instruction set is supported by Pentium3 and newer chips, in the AMD line
8271 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
8272 instruction set supports only single precision arithmetics, thus the double and
8273 extended precision arithmetics is still done using 387. Later version, present
8274 only in Pentium4 and the future AMD x86-64 chips supports double precision
8277 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
8278 @option{-msse2} switches to enable SSE extensions and make this option
8279 effective. For x86-64 compiler, these extensions are enabled by default.
8281 The resulting code should be considerably faster in the majority of cases and avoid
8282 the numerical instability problems of 387 code, but may break some existing
8283 code that expects temporaries to be 80bit.
8285 This is the default choice for the x86-64 compiler.
8288 Attempt to utilize both instruction sets at once. This effectively double the
8289 amount of available registers and on chips with separate execution units for
8290 387 and SSE the execution resources too. Use this option with care, as it is
8291 still experimental, because the GCC register allocator does not model separate
8292 functional units well resulting in instable performance.
8295 @item -masm=@var{dialect}
8296 @opindex masm=@var{dialect}
8297 Output asm instructions using selected @var{dialect}. Supported choices are
8298 @samp{intel} or @samp{att} (the default one).
8303 @opindex mno-ieee-fp
8304 Control whether or not the compiler uses IEEE floating point
8305 comparisons. These handle correctly the case where the result of a
8306 comparison is unordered.
8309 @opindex msoft-float
8310 Generate output containing library calls for floating point.
8311 @strong{Warning:} the requisite libraries are not part of GCC@.
8312 Normally the facilities of the machine's usual C compiler are used, but
8313 this can't be done directly in cross-compilation. You must make your
8314 own arrangements to provide suitable library functions for
8317 On machines where a function returns floating point results in the 80387
8318 register stack, some floating point opcodes may be emitted even if
8319 @option{-msoft-float} is used.
8321 @item -mno-fp-ret-in-387
8322 @opindex mno-fp-ret-in-387
8323 Do not use the FPU registers for return values of functions.
8325 The usual calling convention has functions return values of types
8326 @code{float} and @code{double} in an FPU register, even if there
8327 is no FPU@. The idea is that the operating system should emulate
8330 The option @option{-mno-fp-ret-in-387} causes such values to be returned
8331 in ordinary CPU registers instead.
8333 @item -mno-fancy-math-387
8334 @opindex mno-fancy-math-387
8335 Some 387 emulators do not support the @code{sin}, @code{cos} and
8336 @code{sqrt} instructions for the 387. Specify this option to avoid
8337 generating those instructions. This option is the default on FreeBSD,
8338 OpenBSD and NetBSD@. This option is overridden when @option{-march}
8339 indicates that the target cpu will always have an FPU and so the
8340 instruction will not need emulation. As of revision 2.6.1, these
8341 instructions are not generated unless you also use the
8342 @option{-funsafe-math-optimizations} switch.
8344 @item -malign-double
8345 @itemx -mno-align-double
8346 @opindex malign-double
8347 @opindex mno-align-double
8348 Control whether GCC aligns @code{double}, @code{long double}, and
8349 @code{long long} variables on a two word boundary or a one word
8350 boundary. Aligning @code{double} variables on a two word boundary will
8351 produce code that runs somewhat faster on a @samp{Pentium} at the
8352 expense of more memory.
8354 @strong{Warning:} if you use the @option{-malign-double} switch,
8355 structures containing the above types will be aligned differently than
8356 the published application binary interface specifications for the 386
8357 and will not be binary compatible with structures in code compiled
8358 without that switch.
8360 @item -m96bit-long-double
8361 @itemx -m128bit-long-double
8362 @opindex m96bit-long-double
8363 @opindex m128bit-long-double
8364 These switches control the size of @code{long double} type. The i386
8365 application binary interface specifies the size to be 96 bits,
8366 so @option{-m96bit-long-double} is the default in 32 bit mode.
8368 Modern architectures (Pentium and newer) would prefer @code{long double}
8369 to be aligned to an 8 or 16 byte boundary. In arrays or structures
8370 conforming to the ABI, this would not be possible. So specifying a
8371 @option{-m128bit-long-double} will align @code{long double}
8372 to a 16 byte boundary by padding the @code{long double} with an additional
8375 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
8376 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
8378 Notice that neither of these options enable any extra precision over the x87
8379 standard of 80 bits for a @code{long double}.
8381 @strong{Warning:} if you override the default value for your target ABI, the
8382 structures and arrays containing @code{long double} variables will change
8383 their size as well as function calling convention for function taking
8384 @code{long double} will be modified. Hence they will not be binary
8385 compatible with arrays or structures in code compiled without that switch.
8389 @itemx -mno-svr3-shlib
8390 @opindex msvr3-shlib
8391 @opindex mno-svr3-shlib
8392 Control whether GCC places uninitialized local variables into the
8393 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
8394 into @code{bss}. These options are meaningful only on System V Release 3.
8398 Use a different function-calling convention, in which functions that
8399 take a fixed number of arguments return with the @code{ret} @var{num}
8400 instruction, which pops their arguments while returning. This saves one
8401 instruction in the caller since there is no need to pop the arguments
8404 You can specify that an individual function is called with this calling
8405 sequence with the function attribute @samp{stdcall}. You can also
8406 override the @option{-mrtd} option by using the function attribute
8407 @samp{cdecl}. @xref{Function Attributes}.
8409 @strong{Warning:} this calling convention is incompatible with the one
8410 normally used on Unix, so you cannot use it if you need to call
8411 libraries compiled with the Unix compiler.
8413 Also, you must provide function prototypes for all functions that
8414 take variable numbers of arguments (including @code{printf});
8415 otherwise incorrect code will be generated for calls to those
8418 In addition, seriously incorrect code will result if you call a
8419 function with too many arguments. (Normally, extra arguments are
8420 harmlessly ignored.)
8422 @item -mregparm=@var{num}
8424 Control how many registers are used to pass integer arguments. By
8425 default, no registers are used to pass arguments, and at most 3
8426 registers can be used. You can control this behavior for a specific
8427 function by using the function attribute @samp{regparm}.
8428 @xref{Function Attributes}.
8430 @strong{Warning:} if you use this switch, and
8431 @var{num} is nonzero, then you must build all modules with the same
8432 value, including any libraries. This includes the system libraries and
8435 @item -mpreferred-stack-boundary=@var{num}
8436 @opindex mpreferred-stack-boundary
8437 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
8438 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
8439 the default is 4 (16 bytes or 128 bits), except when optimizing for code
8440 size (@option{-Os}), in which case the default is the minimum correct
8441 alignment (4 bytes for x86, and 8 bytes for x86-64).
8443 On Pentium and PentiumPro, @code{double} and @code{long double} values
8444 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
8445 suffer significant run time performance penalties. On Pentium III, the
8446 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
8447 penalties if it is not 16 byte aligned.
8449 To ensure proper alignment of this values on the stack, the stack boundary
8450 must be as aligned as that required by any value stored on the stack.
8451 Further, every function must be generated such that it keeps the stack
8452 aligned. Thus calling a function compiled with a higher preferred
8453 stack boundary from a function compiled with a lower preferred stack
8454 boundary will most likely misalign the stack. It is recommended that
8455 libraries that use callbacks always use the default setting.
8457 This extra alignment does consume extra stack space, and generally
8458 increases code size. Code that is sensitive to stack space usage, such
8459 as embedded systems and operating system kernels, may want to reduce the
8460 preferred alignment to @option{-mpreferred-stack-boundary=2}.
8478 These switches enable or disable the use of built-in functions that allow
8479 direct access to the MMX, SSE, SSE2, SSE3 and 3Dnow extensions of the
8482 @xref{X86 Built-in Functions}, for details of the functions enabled
8483 and disabled by these switches.
8485 To have SSE/SSE2 instructions generated automatically from floating-point
8486 code, see @option{-mfpmath=sse}.
8489 @itemx -mno-push-args
8491 @opindex mno-push-args
8492 Use PUSH operations to store outgoing parameters. This method is shorter
8493 and usually equally fast as method using SUB/MOV operations and is enabled
8494 by default. In some cases disabling it may improve performance because of
8495 improved scheduling and reduced dependencies.
8497 @item -maccumulate-outgoing-args
8498 @opindex maccumulate-outgoing-args
8499 If enabled, the maximum amount of space required for outgoing arguments will be
8500 computed in the function prologue. This is faster on most modern CPUs
8501 because of reduced dependencies, improved scheduling and reduced stack usage
8502 when preferred stack boundary is not equal to 2. The drawback is a notable
8503 increase in code size. This switch implies @option{-mno-push-args}.
8507 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
8508 on thread-safe exception handling must compile and link all code with the
8509 @option{-mthreads} option. When compiling, @option{-mthreads} defines
8510 @option{-D_MT}; when linking, it links in a special thread helper library
8511 @option{-lmingwthrd} which cleans up per thread exception handling data.
8513 @item -mno-align-stringops
8514 @opindex mno-align-stringops
8515 Do not align destination of inlined string operations. This switch reduces
8516 code size and improves performance in case the destination is already aligned,
8517 but GCC doesn't know about it.
8519 @item -minline-all-stringops
8520 @opindex minline-all-stringops
8521 By default GCC inlines string operations only when destination is known to be
8522 aligned at least to 4 byte boundary. This enables more inlining, increase code
8523 size, but may improve performance of code that depends on fast memcpy, strlen
8524 and memset for short lengths.
8526 @item -momit-leaf-frame-pointer
8527 @opindex momit-leaf-frame-pointer
8528 Don't keep the frame pointer in a register for leaf functions. This
8529 avoids the instructions to save, set up and restore frame pointers and
8530 makes an extra register available in leaf functions. The option
8531 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8532 which might make debugging harder.
8534 @item -mtls-direct-seg-refs
8535 @itemx -mno-tls-direct-seg-refs
8536 @opindex mtls-direct-seg-refs
8537 Controls whether TLS variables may be accessed with offsets from the
8538 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
8539 or whether the thread base pointer must be added. Whether or not this
8540 is legal depends on the operating system, and whether it maps the
8541 segment to cover the entire TLS area.
8543 For systems that use GNU libc, the default is on.
8546 These @samp{-m} switches are supported in addition to the above
8547 on AMD x86-64 processors in 64-bit environments.
8554 Generate code for a 32-bit or 64-bit environment.
8555 The 32-bit environment sets int, long and pointer to 32 bits and
8556 generates code that runs on any i386 system.
8557 The 64-bit environment sets int to 32 bits and long and pointer
8558 to 64 bits and generates code for AMD's x86-64 architecture.
8561 @opindex no-red-zone
8562 Do not use a so called red zone for x86-64 code. The red zone is mandated
8563 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8564 stack pointer that will not be modified by signal or interrupt handlers
8565 and therefore can be used for temporary data without adjusting the stack
8566 pointer. The flag @option{-mno-red-zone} disables this red zone.
8568 @item -mcmodel=small
8569 @opindex mcmodel=small
8570 Generate code for the small code model: the program and its symbols must
8571 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
8572 Programs can be statically or dynamically linked. This is the default
8575 @item -mcmodel=kernel
8576 @opindex mcmodel=kernel
8577 Generate code for the kernel code model. The kernel runs in the
8578 negative 2 GB of the address space.
8579 This model has to be used for Linux kernel code.
8581 @item -mcmodel=medium
8582 @opindex mcmodel=medium
8583 Generate code for the medium model: The program is linked in the lower 2
8584 GB of the address space but symbols can be located anywhere in the
8585 address space. Programs can be statically or dynamically linked, but
8586 building of shared libraries are not supported with the medium model.
8588 @item -mcmodel=large
8589 @opindex mcmodel=large
8590 Generate code for the large model: This model makes no assumptions
8591 about addresses and sizes of sections. Currently GCC does not implement
8596 @subsection HPPA Options
8597 @cindex HPPA Options
8599 These @samp{-m} options are defined for the HPPA family of computers:
8602 @item -march=@var{architecture-type}
8604 Generate code for the specified architecture. The choices for
8605 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8606 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8607 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8608 architecture option for your machine. Code compiled for lower numbered
8609 architectures will run on higher numbered architectures, but not the
8612 PA 2.0 support currently requires gas snapshot 19990413 or later. The
8613 next release of binutils (current is 2.9.1) will probably contain PA 2.0
8617 @itemx -mpa-risc-1-1
8618 @itemx -mpa-risc-2-0
8619 @opindex mpa-risc-1-0
8620 @opindex mpa-risc-1-1
8621 @opindex mpa-risc-2-0
8622 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8625 @opindex mbig-switch
8626 Generate code suitable for big switch tables. Use this option only if
8627 the assembler/linker complain about out of range branches within a switch
8630 @item -mjump-in-delay
8631 @opindex mjump-in-delay
8632 Fill delay slots of function calls with unconditional jump instructions
8633 by modifying the return pointer for the function call to be the target
8634 of the conditional jump.
8636 @item -mdisable-fpregs
8637 @opindex mdisable-fpregs
8638 Prevent floating point registers from being used in any manner. This is
8639 necessary for compiling kernels which perform lazy context switching of
8640 floating point registers. If you use this option and attempt to perform
8641 floating point operations, the compiler will abort.
8643 @item -mdisable-indexing
8644 @opindex mdisable-indexing
8645 Prevent the compiler from using indexing address modes. This avoids some
8646 rather obscure problems when compiling MIG generated code under MACH@.
8648 @item -mno-space-regs
8649 @opindex mno-space-regs
8650 Generate code that assumes the target has no space registers. This allows
8651 GCC to generate faster indirect calls and use unscaled index address modes.
8653 Such code is suitable for level 0 PA systems and kernels.
8655 @item -mfast-indirect-calls
8656 @opindex mfast-indirect-calls
8657 Generate code that assumes calls never cross space boundaries. This
8658 allows GCC to emit code which performs faster indirect calls.
8660 This option will not work in the presence of shared libraries or nested
8663 @item -mlong-load-store
8664 @opindex mlong-load-store
8665 Generate 3-instruction load and store sequences as sometimes required by
8666 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8669 @item -mportable-runtime
8670 @opindex mportable-runtime
8671 Use the portable calling conventions proposed by HP for ELF systems.
8675 Enable the use of assembler directives only GAS understands.
8677 @item -mschedule=@var{cpu-type}
8679 Schedule code according to the constraints for the machine type
8680 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8681 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8682 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8683 proper scheduling option for your machine. The default scheduling is
8687 @opindex mlinker-opt
8688 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8689 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8690 linkers in which they give bogus error messages when linking some programs.
8693 @opindex msoft-float
8694 Generate output containing library calls for floating point.
8695 @strong{Warning:} the requisite libraries are not available for all HPPA
8696 targets. Normally the facilities of the machine's usual C compiler are
8697 used, but this cannot be done directly in cross-compilation. You must make
8698 your own arrangements to provide suitable library functions for
8699 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8700 does provide software floating point support.
8702 @option{-msoft-float} changes the calling convention in the output file;
8703 therefore, it is only useful if you compile @emph{all} of a program with
8704 this option. In particular, you need to compile @file{libgcc.a}, the
8705 library that comes with GCC, with @option{-msoft-float} in order for
8710 Generate the predefine, @code{_SIO}, for server IO. The default is
8711 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
8712 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO. These
8713 options are available under HP-UX and HI-UX.
8717 Use GNU ld specific options. This passes @option{-shared} to ld when
8718 building a shared library. It is the default when GCC is configured,
8719 explicitly or implicitly, with the GNU linker. This option does not
8720 have any affect on which ld is called, it only changes what parameters
8721 are passed to that ld. The ld that is called is determined by the
8722 @option{--with-ld} configure option, GCC's program search path, and
8723 finally by the user's @env{PATH}. The linker used by GCC can be printed
8724 using @samp{which `gcc -print-prog-name=ld`}.
8728 Use HP ld specific options. This passes @option{-b} to ld when building
8729 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8730 links. It is the default when GCC is configured, explicitly or
8731 implicitly, with the HP linker. This option does not have any affect on
8732 which ld is called, it only changes what parameters are passed to that
8733 ld. The ld that is called is determined by the @option{--with-ld}
8734 configure option, GCC's program search path, and finally by the user's
8735 @env{PATH}. The linker used by GCC can be printed using @samp{which
8736 `gcc -print-prog-name=ld`}.
8741 Select the FDPIC ABI, that uses function descriptors to represent
8742 pointers to functions. Without any PIC/PIE-related options, it
8743 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
8744 assumes GOT entries and small data are within a 12-bit range from the
8745 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
8746 are computed with 32 bits.
8749 @opindex minline-plt
8751 Enable inlining of PLT entries in function calls to functions that are
8752 not known to bind locally. It has no effect without @option{-mfdpic}.
8753 It's enabled by default if optimizing for speed and compiling for
8754 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
8755 optimization option such as @option{-O3} or above is present in the
8761 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
8762 that is known to be in read-only sections. It's enabled by default,
8763 except for @option{-fpic} or @option{-fpie}: even though it may help
8764 make the global offset table smaller, it trades 1 instruction for 4.
8765 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
8766 one of which may be shared by multiple symbols, and it avoids the need
8767 for a GOT entry for the referenced symbol, so it's more likely to be a
8768 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
8770 @item -multilib-library-pic
8771 @opindex multilib-library-pic
8773 Link with the (library, not FD) pic libraries. It's implied by
8774 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
8775 @option{-fpic} without @option{-mfdpic}. You should never have to use
8781 Follow the EABI requirement of always creating a frame pointer whenever
8782 a stack frame is allocated. This option is enabled by default and can
8783 be disabled with @option{-mno-linked-fp}.
8786 @opindex mno-long-calls
8787 Generate code that uses long call sequences. This ensures that a call
8788 is always able to reach linker generated stubs. The default is to generate
8789 long calls only when the distance from the call site to the beginning
8790 of the function or translation unit, as the case may be, exceeds a
8791 predefined limit set by the branch type being used. The limits for
8792 normal calls are 7,600,000 and 240,000 bytes, respectively for the
8793 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
8796 Distances are measured from the beginning of functions when using the
8797 @option{-ffunction-sections} option, or when using the @option{-mgas}
8798 and @option{-mno-portable-runtime} options together under HP-UX with
8801 It is normally not desirable to use this option as it will degrade
8802 performance. However, it may be useful in large applications,
8803 particularly when partial linking is used to build the application.
8805 The types of long calls used depends on the capabilities of the
8806 assembler and linker, and the type of code being generated. The
8807 impact on systems that support long absolute calls, and long pic
8808 symbol-difference or pc-relative calls should be relatively small.
8809 However, an indirect call is used on 32-bit ELF systems in pic code
8810 and it is quite long.
8814 Suppress the generation of link options to search libdld.sl when the
8815 @option{-static} option is specified on HP-UX 10 and later.
8819 The HP-UX implementation of setlocale in libc has a dependency on
8820 libdld.sl. There isn't an archive version of libdld.sl. Thus,
8821 when the @option{-static} option is specified, special link options
8822 are needed to resolve this dependency.
8824 On HP-UX 10 and later, the GCC driver adds the necessary options to
8825 link with libdld.sl when the @option{-static} option is specified.
8826 This causes the resulting binary to be dynamic. On the 64-bit port,
8827 the linkers generate dynamic binaries by default in any case. The
8828 @option{-nolibdld} option can be used to prevent the GCC driver from
8829 adding these link options.
8833 Add support for multithreading with the @dfn{dce thread} library
8834 under HP-UX. This option sets flags for both the preprocessor and
8838 @node DEC Alpha Options
8839 @subsection DEC Alpha Options
8841 These @samp{-m} options are defined for the DEC Alpha implementations:
8844 @item -mno-soft-float
8846 @opindex mno-soft-float
8847 @opindex msoft-float
8848 Use (do not use) the hardware floating-point instructions for
8849 floating-point operations. When @option{-msoft-float} is specified,
8850 functions in @file{libgcc.a} will be used to perform floating-point
8851 operations. Unless they are replaced by routines that emulate the
8852 floating-point operations, or compiled in such a way as to call such
8853 emulations routines, these routines will issue floating-point
8854 operations. If you are compiling for an Alpha without floating-point
8855 operations, you must ensure that the library is built so as not to call
8858 Note that Alpha implementations without floating-point operations are
8859 required to have floating-point registers.
8864 @opindex mno-fp-regs
8865 Generate code that uses (does not use) the floating-point register set.
8866 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8867 register set is not used, floating point operands are passed in integer
8868 registers as if they were integers and floating-point results are passed
8869 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8870 so any function with a floating-point argument or return value called by code
8871 compiled with @option{-mno-fp-regs} must also be compiled with that
8874 A typical use of this option is building a kernel that does not use,
8875 and hence need not save and restore, any floating-point registers.
8879 The Alpha architecture implements floating-point hardware optimized for
8880 maximum performance. It is mostly compliant with the IEEE floating
8881 point standard. However, for full compliance, software assistance is
8882 required. This option generates code fully IEEE compliant code
8883 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8884 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8885 defined during compilation. The resulting code is less efficient but is
8886 able to correctly support denormalized numbers and exceptional IEEE
8887 values such as not-a-number and plus/minus infinity. Other Alpha
8888 compilers call this option @option{-ieee_with_no_inexact}.
8890 @item -mieee-with-inexact
8891 @opindex mieee-with-inexact
8892 This is like @option{-mieee} except the generated code also maintains
8893 the IEEE @var{inexact-flag}. Turning on this option causes the
8894 generated code to implement fully-compliant IEEE math. In addition to
8895 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8896 macro. On some Alpha implementations the resulting code may execute
8897 significantly slower than the code generated by default. Since there is
8898 very little code that depends on the @var{inexact-flag}, you should
8899 normally not specify this option. Other Alpha compilers call this
8900 option @option{-ieee_with_inexact}.
8902 @item -mfp-trap-mode=@var{trap-mode}
8903 @opindex mfp-trap-mode
8904 This option controls what floating-point related traps are enabled.
8905 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8906 The trap mode can be set to one of four values:
8910 This is the default (normal) setting. The only traps that are enabled
8911 are the ones that cannot be disabled in software (e.g., division by zero
8915 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8919 Like @samp{su}, but the instructions are marked to be safe for software
8920 completion (see Alpha architecture manual for details).
8923 Like @samp{su}, but inexact traps are enabled as well.
8926 @item -mfp-rounding-mode=@var{rounding-mode}
8927 @opindex mfp-rounding-mode
8928 Selects the IEEE rounding mode. Other Alpha compilers call this option
8929 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8934 Normal IEEE rounding mode. Floating point numbers are rounded towards
8935 the nearest machine number or towards the even machine number in case
8939 Round towards minus infinity.
8942 Chopped rounding mode. Floating point numbers are rounded towards zero.
8945 Dynamic rounding mode. A field in the floating point control register
8946 (@var{fpcr}, see Alpha architecture reference manual) controls the
8947 rounding mode in effect. The C library initializes this register for
8948 rounding towards plus infinity. Thus, unless your program modifies the
8949 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8952 @item -mtrap-precision=@var{trap-precision}
8953 @opindex mtrap-precision
8954 In the Alpha architecture, floating point traps are imprecise. This
8955 means without software assistance it is impossible to recover from a
8956 floating trap and program execution normally needs to be terminated.
8957 GCC can generate code that can assist operating system trap handlers
8958 in determining the exact location that caused a floating point trap.
8959 Depending on the requirements of an application, different levels of
8960 precisions can be selected:
8964 Program precision. This option is the default and means a trap handler
8965 can only identify which program caused a floating point exception.
8968 Function precision. The trap handler can determine the function that
8969 caused a floating point exception.
8972 Instruction precision. The trap handler can determine the exact
8973 instruction that caused a floating point exception.
8976 Other Alpha compilers provide the equivalent options called
8977 @option{-scope_safe} and @option{-resumption_safe}.
8979 @item -mieee-conformant
8980 @opindex mieee-conformant
8981 This option marks the generated code as IEEE conformant. You must not
8982 use this option unless you also specify @option{-mtrap-precision=i} and either
8983 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8984 is to emit the line @samp{.eflag 48} in the function prologue of the
8985 generated assembly file. Under DEC Unix, this has the effect that
8986 IEEE-conformant math library routines will be linked in.
8988 @item -mbuild-constants
8989 @opindex mbuild-constants
8990 Normally GCC examines a 32- or 64-bit integer constant to
8991 see if it can construct it from smaller constants in two or three
8992 instructions. If it cannot, it will output the constant as a literal and
8993 generate code to load it from the data segment at runtime.
8995 Use this option to require GCC to construct @emph{all} integer constants
8996 using code, even if it takes more instructions (the maximum is six).
8998 You would typically use this option to build a shared library dynamic
8999 loader. Itself a shared library, it must relocate itself in memory
9000 before it can find the variables and constants in its own data segment.
9006 Select whether to generate code to be assembled by the vendor-supplied
9007 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9025 Indicate whether GCC should generate code to use the optional BWX,
9026 CIX, FIX and MAX instruction sets. The default is to use the instruction
9027 sets supported by the CPU type specified via @option{-mcpu=} option or that
9028 of the CPU on which GCC was built if none was specified.
9033 @opindex mfloat-ieee
9034 Generate code that uses (does not use) VAX F and G floating point
9035 arithmetic instead of IEEE single and double precision.
9037 @item -mexplicit-relocs
9038 @itemx -mno-explicit-relocs
9039 @opindex mexplicit-relocs
9040 @opindex mno-explicit-relocs
9041 Older Alpha assemblers provided no way to generate symbol relocations
9042 except via assembler macros. Use of these macros does not allow
9043 optimal instruction scheduling. GNU binutils as of version 2.12
9044 supports a new syntax that allows the compiler to explicitly mark
9045 which relocations should apply to which instructions. This option
9046 is mostly useful for debugging, as GCC detects the capabilities of
9047 the assembler when it is built and sets the default accordingly.
9051 @opindex msmall-data
9052 @opindex mlarge-data
9053 When @option{-mexplicit-relocs} is in effect, static data is
9054 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9055 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9056 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9057 16-bit relocations off of the @code{$gp} register. This limits the
9058 size of the small data area to 64KB, but allows the variables to be
9059 directly accessed via a single instruction.
9061 The default is @option{-mlarge-data}. With this option the data area
9062 is limited to just below 2GB. Programs that require more than 2GB of
9063 data must use @code{malloc} or @code{mmap} to allocate the data in the
9064 heap instead of in the program's data segment.
9066 When generating code for shared libraries, @option{-fpic} implies
9067 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9071 @opindex msmall-text
9072 @opindex mlarge-text
9073 When @option{-msmall-text} is used, the compiler assumes that the
9074 code of the entire program (or shared library) fits in 4MB, and is
9075 thus reachable with a branch instruction. When @option{-msmall-data}
9076 is used, the compiler can assume that all local symbols share the
9077 same @code{$gp} value, and thus reduce the number of instructions
9078 required for a function call from 4 to 1.
9080 The default is @option{-mlarge-text}.
9082 @item -mcpu=@var{cpu_type}
9084 Set the instruction set and instruction scheduling parameters for
9085 machine type @var{cpu_type}. You can specify either the @samp{EV}
9086 style name or the corresponding chip number. GCC supports scheduling
9087 parameters for the EV4, EV5 and EV6 family of processors and will
9088 choose the default values for the instruction set from the processor
9089 you specify. If you do not specify a processor type, GCC will default
9090 to the processor on which the compiler was built.
9092 Supported values for @var{cpu_type} are
9098 Schedules as an EV4 and has no instruction set extensions.
9102 Schedules as an EV5 and has no instruction set extensions.
9106 Schedules as an EV5 and supports the BWX extension.
9111 Schedules as an EV5 and supports the BWX and MAX extensions.
9115 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9119 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9122 @item -mtune=@var{cpu_type}
9124 Set only the instruction scheduling parameters for machine type
9125 @var{cpu_type}. The instruction set is not changed.
9127 @item -mmemory-latency=@var{time}
9128 @opindex mmemory-latency
9129 Sets the latency the scheduler should assume for typical memory
9130 references as seen by the application. This number is highly
9131 dependent on the memory access patterns used by the application
9132 and the size of the external cache on the machine.
9134 Valid options for @var{time} are
9138 A decimal number representing clock cycles.
9144 The compiler contains estimates of the number of clock cycles for
9145 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9146 (also called Dcache, Scache, and Bcache), as well as to main memory.
9147 Note that L3 is only valid for EV5.
9152 @node DEC Alpha/VMS Options
9153 @subsection DEC Alpha/VMS Options
9155 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9158 @item -mvms-return-codes
9159 @opindex mvms-return-codes
9160 Return VMS condition codes from main. The default is to return POSIX
9161 style condition (e.g.@ error) codes.
9164 @node H8/300 Options
9165 @subsection H8/300 Options
9167 These @samp{-m} options are defined for the H8/300 implementations:
9172 Shorten some address references at link time, when possible; uses the
9173 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9174 ld, Using ld}, for a fuller description.
9178 Generate code for the H8/300H@.
9182 Generate code for the H8S@.
9186 Generate code for the H8S and H8/300H in the normal mode. This switch
9187 must be used either with -mh or -ms.
9191 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9195 Make @code{int} data 32 bits by default.
9199 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9200 The default for the H8/300H and H8S is to align longs and floats on 4
9202 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9203 This option has no effect on the H8/300.
9207 @subsection SH Options
9209 These @samp{-m} options are defined for the SH implementations:
9214 Generate code for the SH1.
9218 Generate code for the SH2.
9221 Generate code for the SH2e.
9225 Generate code for the SH3.
9229 Generate code for the SH3e.
9233 Generate code for the SH4 without a floating-point unit.
9235 @item -m4-single-only
9236 @opindex m4-single-only
9237 Generate code for the SH4 with a floating-point unit that only
9238 supports single-precision arithmetic.
9242 Generate code for the SH4 assuming the floating-point unit is in
9243 single-precision mode by default.
9247 Generate code for the SH4.
9251 Compile code for the processor in big endian mode.
9255 Compile code for the processor in little endian mode.
9259 Align doubles at 64-bit boundaries. Note that this changes the calling
9260 conventions, and thus some functions from the standard C library will
9261 not work unless you recompile it first with @option{-mdalign}.
9265 Shorten some address references at link time, when possible; uses the
9266 linker option @option{-relax}.
9270 Use 32-bit offsets in @code{switch} tables. The default is to use
9275 Enable the use of the instruction @code{fmovd}.
9279 Comply with the calling conventions defined by Renesas.
9283 Mark the @code{MAC} register as call-clobbered, even if
9284 @option{-mhitachi} is given.
9288 Increase IEEE-compliance of floating-point code.
9292 Dump instruction size and location in the assembly code.
9296 This option is deprecated. It pads structures to multiple of 4 bytes,
9297 which is incompatible with the SH ABI@.
9301 Optimize for space instead of speed. Implied by @option{-Os}.
9305 When generating position-independent code, emit function calls using
9306 the Global Offset Table instead of the Procedure Linkage Table.
9310 Generate a library function call to invalidate instruction cache
9311 entries, after fixing up a trampoline. This library function call
9312 doesn't assume it can write to the whole memory address space. This
9313 is the default when the target is @code{sh-*-linux*}.
9316 @node System V Options
9317 @subsection Options for System V
9319 These additional options are available on System V Release 4 for
9320 compatibility with other compilers on those systems:
9325 Create a shared object.
9326 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
9330 Identify the versions of each tool used by the compiler, in a
9331 @code{.ident} assembler directive in the output.
9335 Refrain from adding @code{.ident} directives to the output file (this is
9338 @item -YP,@var{dirs}
9340 Search the directories @var{dirs}, and no others, for libraries
9341 specified with @option{-l}.
9345 Look in the directory @var{dir} to find the M4 preprocessor.
9346 The assembler uses this option.
9347 @c This is supposed to go with a -Yd for predefined M4 macro files, but
9348 @c the generic assembler that comes with Solaris takes just -Ym.
9351 @node TMS320C3x/C4x Options
9352 @subsection TMS320C3x/C4x Options
9353 @cindex TMS320C3x/C4x Options
9355 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
9359 @item -mcpu=@var{cpu_type}
9361 Set the instruction set, register set, and instruction scheduling
9362 parameters for machine type @var{cpu_type}. Supported values for
9363 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
9364 @samp{c44}. The default is @samp{c40} to generate code for the
9369 @itemx -msmall-memory
9371 @opindex mbig-memory
9373 @opindex msmall-memory
9375 Generates code for the big or small memory model. The small memory
9376 model assumed that all data fits into one 64K word page. At run-time
9377 the data page (DP) register must be set to point to the 64K page
9378 containing the .bss and .data program sections. The big memory model is
9379 the default and requires reloading of the DP register for every direct
9386 Allow (disallow) allocation of general integer operands into the block
9393 Enable (disable) generation of code using decrement and branch,
9394 DBcond(D), instructions. This is enabled by default for the C4x. To be
9395 on the safe side, this is disabled for the C3x, since the maximum
9396 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
9397 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
9398 that it can utilize the decrement and branch instruction, but will give
9399 up if there is more than one memory reference in the loop. Thus a loop
9400 where the loop counter is decremented can generate slightly more
9401 efficient code, in cases where the RPTB instruction cannot be utilized.
9403 @item -mdp-isr-reload
9405 @opindex mdp-isr-reload
9407 Force the DP register to be saved on entry to an interrupt service
9408 routine (ISR), reloaded to point to the data section, and restored on
9409 exit from the ISR@. This should not be required unless someone has
9410 violated the small memory model by modifying the DP register, say within
9417 For the C3x use the 24-bit MPYI instruction for integer multiplies
9418 instead of a library call to guarantee 32-bit results. Note that if one
9419 of the operands is a constant, then the multiplication will be performed
9420 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
9421 then squaring operations are performed inline instead of a library call.
9424 @itemx -mno-fast-fix
9426 @opindex mno-fast-fix
9427 The C3x/C4x FIX instruction to convert a floating point value to an
9428 integer value chooses the nearest integer less than or equal to the
9429 floating point value rather than to the nearest integer. Thus if the
9430 floating point number is negative, the result will be incorrectly
9431 truncated an additional code is necessary to detect and correct this
9432 case. This option can be used to disable generation of the additional
9433 code required to correct the result.
9439 Enable (disable) generation of repeat block sequences using the RPTB
9440 instruction for zero overhead looping. The RPTB construct is only used
9441 for innermost loops that do not call functions or jump across the loop
9442 boundaries. There is no advantage having nested RPTB loops due to the
9443 overhead required to save and restore the RC, RS, and RE registers.
9444 This is enabled by default with @option{-O2}.
9446 @item -mrpts=@var{count}
9450 Enable (disable) the use of the single instruction repeat instruction
9451 RPTS@. If a repeat block contains a single instruction, and the loop
9452 count can be guaranteed to be less than the value @var{count}, GCC will
9453 emit a RPTS instruction instead of a RPTB@. If no value is specified,
9454 then a RPTS will be emitted even if the loop count cannot be determined
9455 at compile time. Note that the repeated instruction following RPTS does
9456 not have to be reloaded from memory each iteration, thus freeing up the
9457 CPU buses for operands. However, since interrupts are blocked by this
9458 instruction, it is disabled by default.
9460 @item -mloop-unsigned
9461 @itemx -mno-loop-unsigned
9462 @opindex mloop-unsigned
9463 @opindex mno-loop-unsigned
9464 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
9465 is @math{2^{31} + 1} since these instructions test if the iteration count is
9466 negative to terminate the loop. If the iteration count is unsigned
9467 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
9468 exceeded. This switch allows an unsigned iteration count.
9472 Try to emit an assembler syntax that the TI assembler (asm30) is happy
9473 with. This also enforces compatibility with the API employed by the TI
9474 C3x C compiler. For example, long doubles are passed as structures
9475 rather than in floating point registers.
9481 Generate code that uses registers (stack) for passing arguments to functions.
9482 By default, arguments are passed in registers where possible rather
9483 than by pushing arguments on to the stack.
9485 @item -mparallel-insns
9486 @itemx -mno-parallel-insns
9487 @opindex mparallel-insns
9488 @opindex mno-parallel-insns
9489 Allow the generation of parallel instructions. This is enabled by
9490 default with @option{-O2}.
9492 @item -mparallel-mpy
9493 @itemx -mno-parallel-mpy
9494 @opindex mparallel-mpy
9495 @opindex mno-parallel-mpy
9496 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
9497 provided @option{-mparallel-insns} is also specified. These instructions have
9498 tight register constraints which can pessimize the code generation
9504 @subsection V850 Options
9505 @cindex V850 Options
9507 These @samp{-m} options are defined for V850 implementations:
9511 @itemx -mno-long-calls
9512 @opindex mlong-calls
9513 @opindex mno-long-calls
9514 Treat all calls as being far away (near). If calls are assumed to be
9515 far away, the compiler will always load the functions address up into a
9516 register, and call indirect through the pointer.
9522 Do not optimize (do optimize) basic blocks that use the same index
9523 pointer 4 or more times to copy pointer into the @code{ep} register, and
9524 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
9525 option is on by default if you optimize.
9527 @item -mno-prolog-function
9528 @itemx -mprolog-function
9529 @opindex mno-prolog-function
9530 @opindex mprolog-function
9531 Do not use (do use) external functions to save and restore registers
9532 at the prologue and epilogue of a function. The external functions
9533 are slower, but use less code space if more than one function saves
9534 the same number of registers. The @option{-mprolog-function} option
9535 is on by default if you optimize.
9539 Try to make the code as small as possible. At present, this just turns
9540 on the @option{-mep} and @option{-mprolog-function} options.
9544 Put static or global variables whose size is @var{n} bytes or less into
9545 the tiny data area that register @code{ep} points to. The tiny data
9546 area can hold up to 256 bytes in total (128 bytes for byte references).
9550 Put static or global variables whose size is @var{n} bytes or less into
9551 the small data area that register @code{gp} points to. The small data
9552 area can hold up to 64 kilobytes.
9556 Put static or global variables whose size is @var{n} bytes or less into
9557 the first 32 kilobytes of memory.
9561 Specify that the target processor is the V850.
9564 @opindex mbig-switch
9565 Generate code suitable for big switch tables. Use this option only if
9566 the assembler/linker complain about out of range branches within a switch
9571 This option will cause r2 and r5 to be used in the code generated by
9572 the compiler. This setting is the default.
9575 @opindex mno-app-regs
9576 This option will cause r2 and r5 to be treated as fixed registers.
9580 Specify that the target processor is the V850E1. The preprocessor
9581 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
9582 this option is used.
9586 Specify that the target processor is the V850E. The preprocessor
9587 constant @samp{__v850e__} will be defined if this option is used.
9589 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
9590 are defined then a default target processor will be chosen and the
9591 relevant @samp{__v850*__} preprocessor constant will be defined.
9593 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
9594 defined, regardless of which processor variant is the target.
9596 @item -mdisable-callt
9597 @opindex mdisable-callt
9598 This option will suppress generation of the CALLT instruction for the
9599 v850e and v850e1 flavors of the v850 architecture. The default is
9600 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
9605 @subsection ARC Options
9608 These options are defined for ARC implementations:
9613 Compile code for little endian mode. This is the default.
9617 Compile code for big endian mode.
9620 @opindex mmangle-cpu
9621 Prepend the name of the cpu to all public symbol names.
9622 In multiple-processor systems, there are many ARC variants with different
9623 instruction and register set characteristics. This flag prevents code
9624 compiled for one cpu to be linked with code compiled for another.
9625 No facility exists for handling variants that are ``almost identical''.
9626 This is an all or nothing option.
9628 @item -mcpu=@var{cpu}
9630 Compile code for ARC variant @var{cpu}.
9631 Which variants are supported depend on the configuration.
9632 All variants support @option{-mcpu=base}, this is the default.
9634 @item -mtext=@var{text-section}
9635 @itemx -mdata=@var{data-section}
9636 @itemx -mrodata=@var{readonly-data-section}
9640 Put functions, data, and readonly data in @var{text-section},
9641 @var{data-section}, and @var{readonly-data-section} respectively
9642 by default. This can be overridden with the @code{section} attribute.
9643 @xref{Variable Attributes}.
9648 @subsection NS32K Options
9649 @cindex NS32K options
9651 These are the @samp{-m} options defined for the 32000 series. The default
9652 values for these options depends on which style of 32000 was selected when
9653 the compiler was configured; the defaults for the most common choices are
9661 Generate output for a 32032. This is the default
9662 when the compiler is configured for 32032 and 32016 based systems.
9668 Generate output for a 32332. This is the default
9669 when the compiler is configured for 32332-based systems.
9675 Generate output for a 32532. This is the default
9676 when the compiler is configured for 32532-based systems.
9680 Generate output containing 32081 instructions for floating point.
9681 This is the default for all systems.
9685 Generate output containing 32381 instructions for floating point. This
9686 also implies @option{-m32081}. The 32381 is only compatible with the 32332
9687 and 32532 cpus. This is the default for the pc532-netbsd configuration.
9691 Try and generate multiply-add floating point instructions @code{polyF}
9692 and @code{dotF}. This option is only available if the @option{-m32381}
9693 option is in effect. Using these instructions requires changes to
9694 register allocation which generally has a negative impact on
9695 performance. This option should only be enabled when compiling code
9696 particularly likely to make heavy use of multiply-add instructions.
9699 @opindex mnomulti-add
9700 Do not try and generate multiply-add floating point instructions
9701 @code{polyF} and @code{dotF}. This is the default on all platforms.
9704 @opindex msoft-float
9705 Generate output containing library calls for floating point.
9706 @strong{Warning:} the requisite libraries may not be available.
9708 @item -mieee-compare
9709 @itemx -mno-ieee-compare
9710 @opindex mieee-compare
9711 @opindex mno-ieee-compare
9712 Control whether or not the compiler uses IEEE floating point
9713 comparisons. These handle correctly the case where the result of a
9714 comparison is unordered.
9715 @strong{Warning:} the requisite kernel support may not be available.
9718 @opindex mnobitfield
9719 Do not use the bit-field instructions. On some machines it is faster to
9720 use shifting and masking operations. This is the default for the pc532.
9724 Do use the bit-field instructions. This is the default for all platforms
9729 Use a different function-calling convention, in which functions
9730 that take a fixed number of arguments return pop their
9731 arguments on return with the @code{ret} instruction.
9733 This calling convention is incompatible with the one normally
9734 used on Unix, so you cannot use it if you need to call libraries
9735 compiled with the Unix compiler.
9737 Also, you must provide function prototypes for all functions that
9738 take variable numbers of arguments (including @code{printf});
9739 otherwise incorrect code will be generated for calls to those
9742 In addition, seriously incorrect code will result if you call a
9743 function with too many arguments. (Normally, extra arguments are
9744 harmlessly ignored.)
9746 This option takes its name from the 680x0 @code{rtd} instruction.
9751 Use a different function-calling convention where the first two arguments
9752 are passed in registers.
9754 This calling convention is incompatible with the one normally
9755 used on Unix, so you cannot use it if you need to call libraries
9756 compiled with the Unix compiler.
9759 @opindex mnoregparam
9760 Do not pass any arguments in registers. This is the default for all
9765 It is OK to use the sb as an index register which is always loaded with
9766 zero. This is the default for the pc532-netbsd target.
9770 The sb register is not available for use or has not been initialized to
9771 zero by the run time system. This is the default for all targets except
9772 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
9773 @option{-fpic} is set.
9777 Many ns32000 series addressing modes use displacements of up to 512MB@.
9778 If an address is above 512MB then displacements from zero can not be used.
9779 This option causes code to be generated which can be loaded above 512MB@.
9780 This may be useful for operating systems or ROM code.
9784 Assume code will be loaded in the first 512MB of virtual address space.
9785 This is the default for all platforms.
9791 @subsection AVR Options
9794 These options are defined for AVR implementations:
9797 @item -mmcu=@var{mcu}
9799 Specify ATMEL AVR instruction set or MCU type.
9801 Instruction set avr1 is for the minimal AVR core, not supported by the C
9802 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
9803 attiny11, attiny12, attiny15, attiny28).
9805 Instruction set avr2 (default) is for the classic AVR core with up to
9806 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
9807 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
9808 at90c8534, at90s8535).
9810 Instruction set avr3 is for the classic AVR core with up to 128K program
9811 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
9813 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9814 memory space (MCU types: atmega8, atmega83, atmega85).
9816 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9817 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9818 atmega64, atmega128, at43usb355, at94k).
9822 Output instruction sizes to the asm file.
9824 @item -minit-stack=@var{N}
9825 @opindex minit-stack
9826 Specify the initial stack address, which may be a symbol or numeric value,
9827 @samp{__stack} is the default.
9829 @item -mno-interrupts
9830 @opindex mno-interrupts
9831 Generated code is not compatible with hardware interrupts.
9832 Code size will be smaller.
9834 @item -mcall-prologues
9835 @opindex mcall-prologues
9836 Functions prologues/epilogues expanded as call to appropriate
9837 subroutines. Code size will be smaller.
9839 @item -mno-tablejump
9840 @opindex mno-tablejump
9841 Do not generate tablejump insns which sometimes increase code size.
9844 @opindex mtiny-stack
9845 Change only the low 8 bits of the stack pointer.
9849 @subsection MCore Options
9850 @cindex MCore options
9852 These are the @samp{-m} options defined for the Motorola M*Core
9860 @opindex mno-hardlit
9861 Inline constants into the code stream if it can be done in two
9862 instructions or less.
9868 Use the divide instruction. (Enabled by default).
9870 @item -mrelax-immediate
9871 @itemx -mno-relax-immediate
9872 @opindex mrelax-immediate
9873 @opindex mno-relax-immediate
9874 Allow arbitrary sized immediates in bit operations.
9876 @item -mwide-bitfields
9877 @itemx -mno-wide-bitfields
9878 @opindex mwide-bitfields
9879 @opindex mno-wide-bitfields
9880 Always treat bit-fields as int-sized.
9882 @item -m4byte-functions
9883 @itemx -mno-4byte-functions
9884 @opindex m4byte-functions
9885 @opindex mno-4byte-functions
9886 Force all functions to be aligned to a four byte boundary.
9888 @item -mcallgraph-data
9889 @itemx -mno-callgraph-data
9890 @opindex mcallgraph-data
9891 @opindex mno-callgraph-data
9892 Emit callgraph information.
9895 @itemx -mno-slow-bytes
9896 @opindex mslow-bytes
9897 @opindex mno-slow-bytes
9898 Prefer word access when reading byte quantities.
9900 @item -mlittle-endian
9902 @opindex mlittle-endian
9903 @opindex mbig-endian
9904 Generate code for a little endian target.
9910 Generate code for the 210 processor.
9914 @subsection IA-64 Options
9915 @cindex IA-64 Options
9917 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9921 @opindex mbig-endian
9922 Generate code for a big endian target. This is the default for HP-UX@.
9924 @item -mlittle-endian
9925 @opindex mlittle-endian
9926 Generate code for a little endian target. This is the default for AIX5
9933 Generate (or don't) code for the GNU assembler. This is the default.
9934 @c Also, this is the default if the configure option @option{--with-gnu-as}
9941 Generate (or don't) code for the GNU linker. This is the default.
9942 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9947 Generate code that does not use a global pointer register. The result
9948 is not position independent code, and violates the IA-64 ABI@.
9950 @item -mvolatile-asm-stop
9951 @itemx -mno-volatile-asm-stop
9952 @opindex mvolatile-asm-stop
9953 @opindex mno-volatile-asm-stop
9954 Generate (or don't) a stop bit immediately before and after volatile asm
9959 Generate code that works around Itanium B step errata.
9961 @item -mregister-names
9962 @itemx -mno-register-names
9963 @opindex mregister-names
9964 @opindex mno-register-names
9965 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9966 the stacked registers. This may make assembler output more readable.
9972 Disable (or enable) optimizations that use the small data section. This may
9973 be useful for working around optimizer bugs.
9976 @opindex mconstant-gp
9977 Generate code that uses a single constant global pointer value. This is
9978 useful when compiling kernel code.
9982 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9983 This is useful when compiling firmware code.
9985 @item -minline-float-divide-min-latency
9986 @opindex minline-float-divide-min-latency
9987 Generate code for inline divides of floating point values
9988 using the minimum latency algorithm.
9990 @item -minline-float-divide-max-throughput
9991 @opindex minline-float-divide-max-throughput
9992 Generate code for inline divides of floating point values
9993 using the maximum throughput algorithm.
9995 @item -minline-int-divide-min-latency
9996 @opindex minline-int-divide-min-latency
9997 Generate code for inline divides of integer values
9998 using the minimum latency algorithm.
10000 @item -minline-int-divide-max-throughput
10001 @opindex minline-int-divide-max-throughput
10002 Generate code for inline divides of integer values
10003 using the maximum throughput algorithm.
10005 @item -mno-dwarf2-asm
10006 @itemx -mdwarf2-asm
10007 @opindex mno-dwarf2-asm
10008 @opindex mdwarf2-asm
10009 Don't (or do) generate assembler code for the DWARF2 line number debugging
10010 info. This may be useful when not using the GNU assembler.
10012 @item -mfixed-range=@var{register-range}
10013 @opindex mfixed-range
10014 Generate code treating the given register range as fixed registers.
10015 A fixed register is one that the register allocator can not use. This is
10016 useful when compiling kernel code. A register range is specified as
10017 two registers separated by a dash. Multiple register ranges can be
10018 specified separated by a comma.
10020 @item -mearly-stop-bits
10021 @itemx -mno-early-stop-bits
10022 @opindex mearly-stop-bits
10023 @opindex mno-early-stop-bits
10024 Allow stop bits to be placed earlier than immediately preceding the
10025 instruction that triggered the stop bit. This can improve instruction
10026 scheduling, but does not always do so.
10029 @node S/390 and zSeries Options
10030 @subsection S/390 and zSeries Options
10031 @cindex S/390 and zSeries Options
10033 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
10037 @itemx -msoft-float
10038 @opindex mhard-float
10039 @opindex msoft-float
10040 Use (do not use) the hardware floating-point instructions and registers
10041 for floating-point operations. When @option{-msoft-float} is specified,
10042 functions in @file{libgcc.a} will be used to perform floating-point
10043 operations. When @option{-mhard-float} is specified, the compiler
10044 generates IEEE floating-point instructions. This is the default.
10047 @itemx -mno-backchain
10048 @opindex mbackchain
10049 @opindex mno-backchain
10050 Generate (or do not generate) code which maintains an explicit
10051 backchain within the stack frame that points to the caller's frame.
10052 This may be needed to allow debugging using tools that do not understand
10053 DWARF-2 call frame information. The default is not to generate the
10057 @itemx -mno-small-exec
10058 @opindex msmall-exec
10059 @opindex mno-small-exec
10060 Generate (or do not generate) code using the @code{bras} instruction
10061 to do subroutine calls.
10062 This only works reliably if the total executable size does not
10063 exceed 64k. The default is to use the @code{basr} instruction instead,
10064 which does not have this limitation.
10070 When @option{-m31} is specified, generate code compliant to the
10071 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
10072 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
10073 particular to generate 64-bit instructions. For the @samp{s390}
10074 targets, the default is @option{-m31}, while the @samp{s390x}
10075 targets default to @option{-m64}.
10081 When @option{-mzarch} is specified, generate code using the
10082 instructions available on z/Architecture.
10083 When @option{-mesa} is specified, generate code using the
10084 instructions available on ESA/390. Note that @option{-mesa} is
10085 not possible with @option{-m64}.
10086 When generating code compliant to the GNU/Linux for S/390 ABI,
10087 the default is @option{-mesa}. When generating code compliant
10088 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
10094 Generate (or do not generate) code using the @code{mvcle} instruction
10095 to perform block moves. When @option{-mno-mvcle} is specified,
10096 use a @code{mvc} loop instead. This is the default.
10102 Print (or do not print) additional debug information when compiling.
10103 The default is to not print debug information.
10105 @item -march=@var{cpu-type}
10107 Generate code that will run on @var{cpu-type}, which is the name of a system
10108 representing a certain processor type. Possible values for
10109 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
10110 When generating code using the instructions available on z/Architecture,
10111 the default is @option{-march=z900}. Otherwise, the default is
10112 @option{-march=g5}.
10114 @item -mtune=@var{cpu-type}
10116 Tune to @var{cpu-type} everything applicable about the generated code,
10117 except for the ABI and the set of available instructions.
10118 The list of @var{cpu-type} values is the same as for @option{-march}.
10119 The default is the value used for @option{-march}.
10122 @itemx -mno-fused-madd
10123 @opindex mfused-madd
10124 @opindex mno-fused-madd
10125 Generate code that uses (does not use) the floating point multiply and
10126 accumulate instructions. These instructions are generated by default if
10127 hardware floating point is used.
10131 @subsection CRIS Options
10132 @cindex CRIS Options
10134 These options are defined specifically for the CRIS ports.
10137 @item -march=@var{architecture-type}
10138 @itemx -mcpu=@var{architecture-type}
10141 Generate code for the specified architecture. The choices for
10142 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10143 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
10144 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10147 @item -mtune=@var{architecture-type}
10149 Tune to @var{architecture-type} everything applicable about the generated
10150 code, except for the ABI and the set of available instructions. The
10151 choices for @var{architecture-type} are the same as for
10152 @option{-march=@var{architecture-type}}.
10154 @item -mmax-stack-frame=@var{n}
10155 @opindex mmax-stack-frame
10156 Warn when the stack frame of a function exceeds @var{n} bytes.
10158 @item -melinux-stacksize=@var{n}
10159 @opindex melinux-stacksize
10160 Only available with the @samp{cris-axis-aout} target. Arranges for
10161 indications in the program to the kernel loader that the stack of the
10162 program should be set to @var{n} bytes.
10168 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10169 @option{-march=v3} and @option{-march=v8} respectively.
10171 @item -mmul-bug-workaround
10172 @itemx -mno-mul-bug-workaround
10173 @opindex mmul-bug-workaround
10174 @opindex mno-mul-bug-workaround
10175 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
10176 models where it applies. This option is active by default.
10180 Enable CRIS-specific verbose debug-related information in the assembly
10181 code. This option also has the effect to turn off the @samp{#NO_APP}
10182 formatted-code indicator to the assembler at the beginning of the
10187 Do not use condition-code results from previous instruction; always emit
10188 compare and test instructions before use of condition codes.
10190 @item -mno-side-effects
10191 @opindex mno-side-effects
10192 Do not emit instructions with side-effects in addressing modes other than
10195 @item -mstack-align
10196 @itemx -mno-stack-align
10197 @itemx -mdata-align
10198 @itemx -mno-data-align
10199 @itemx -mconst-align
10200 @itemx -mno-const-align
10201 @opindex mstack-align
10202 @opindex mno-stack-align
10203 @opindex mdata-align
10204 @opindex mno-data-align
10205 @opindex mconst-align
10206 @opindex mno-const-align
10207 These options (no-options) arranges (eliminate arrangements) for the
10208 stack-frame, individual data and constants to be aligned for the maximum
10209 single data access size for the chosen CPU model. The default is to
10210 arrange for 32-bit alignment. ABI details such as structure layout are
10211 not affected by these options.
10219 Similar to the stack- data- and const-align options above, these options
10220 arrange for stack-frame, writable data and constants to all be 32-bit,
10221 16-bit or 8-bit aligned. The default is 32-bit alignment.
10223 @item -mno-prologue-epilogue
10224 @itemx -mprologue-epilogue
10225 @opindex mno-prologue-epilogue
10226 @opindex mprologue-epilogue
10227 With @option{-mno-prologue-epilogue}, the normal function prologue and
10228 epilogue that sets up the stack-frame are omitted and no return
10229 instructions or return sequences are generated in the code. Use this
10230 option only together with visual inspection of the compiled code: no
10231 warnings or errors are generated when call-saved registers must be saved,
10232 or storage for local variable needs to be allocated.
10236 @opindex mno-gotplt
10238 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10239 instruction sequences that load addresses for functions from the PLT part
10240 of the GOT rather than (traditional on other architectures) calls to the
10241 PLT. The default is @option{-mgotplt}.
10245 Legacy no-op option only recognized with the cris-axis-aout target.
10249 Legacy no-op option only recognized with the cris-axis-elf and
10250 cris-axis-linux-gnu targets.
10254 Only recognized with the cris-axis-aout target, where it selects a
10255 GNU/linux-like multilib, include files and instruction set for
10256 @option{-march=v8}.
10260 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10264 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
10265 to link with input-output functions from a simulator library. Code,
10266 initialized data and zero-initialized data are allocated consecutively.
10270 Like @option{-sim}, but pass linker options to locate initialized data at
10271 0x40000000 and zero-initialized data at 0x80000000.
10275 @subsection MMIX Options
10276 @cindex MMIX Options
10278 These options are defined for the MMIX:
10282 @itemx -mno-libfuncs
10284 @opindex mno-libfuncs
10285 Specify that intrinsic library functions are being compiled, passing all
10286 values in registers, no matter the size.
10289 @itemx -mno-epsilon
10291 @opindex mno-epsilon
10292 Generate floating-point comparison instructions that compare with respect
10293 to the @code{rE} epsilon register.
10295 @item -mabi=mmixware
10297 @opindex mabi-mmixware
10299 Generate code that passes function parameters and return values that (in
10300 the called function) are seen as registers @code{$0} and up, as opposed to
10301 the GNU ABI which uses global registers @code{$231} and up.
10303 @item -mzero-extend
10304 @itemx -mno-zero-extend
10305 @opindex mzero-extend
10306 @opindex mno-zero-extend
10307 When reading data from memory in sizes shorter than 64 bits, use (do not
10308 use) zero-extending load instructions by default, rather than
10309 sign-extending ones.
10312 @itemx -mno-knuthdiv
10314 @opindex mno-knuthdiv
10315 Make the result of a division yielding a remainder have the same sign as
10316 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
10317 remainder follows the sign of the dividend. Both methods are
10318 arithmetically valid, the latter being almost exclusively used.
10320 @item -mtoplevel-symbols
10321 @itemx -mno-toplevel-symbols
10322 @opindex mtoplevel-symbols
10323 @opindex mno-toplevel-symbols
10324 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
10325 code can be used with the @code{PREFIX} assembly directive.
10329 Generate an executable in the ELF format, rather than the default
10330 @samp{mmo} format used by the @command{mmix} simulator.
10332 @item -mbranch-predict
10333 @itemx -mno-branch-predict
10334 @opindex mbranch-predict
10335 @opindex mno-branch-predict
10336 Use (do not use) the probable-branch instructions, when static branch
10337 prediction indicates a probable branch.
10339 @item -mbase-addresses
10340 @itemx -mno-base-addresses
10341 @opindex mbase-addresses
10342 @opindex mno-base-addresses
10343 Generate (do not generate) code that uses @emph{base addresses}. Using a
10344 base address automatically generates a request (handled by the assembler
10345 and the linker) for a constant to be set up in a global register. The
10346 register is used for one or more base address requests within the range 0
10347 to 255 from the value held in the register. The generally leads to short
10348 and fast code, but the number of different data items that can be
10349 addressed is limited. This means that a program that uses lots of static
10350 data may require @option{-mno-base-addresses}.
10352 @item -msingle-exit
10353 @itemx -mno-single-exit
10354 @opindex msingle-exit
10355 @opindex mno-single-exit
10356 Force (do not force) generated code to have a single exit point in each
10360 @node PDP-11 Options
10361 @subsection PDP-11 Options
10362 @cindex PDP-11 Options
10364 These options are defined for the PDP-11:
10369 Use hardware FPP floating point. This is the default. (FIS floating
10370 point on the PDP-11/40 is not supported.)
10373 @opindex msoft-float
10374 Do not use hardware floating point.
10378 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
10382 Return floating-point results in memory. This is the default.
10386 Generate code for a PDP-11/40.
10390 Generate code for a PDP-11/45. This is the default.
10394 Generate code for a PDP-11/10.
10396 @item -mbcopy-builtin
10397 @opindex bcopy-builtin
10398 Use inline @code{movstrhi} patterns for copying memory. This is the
10403 Do not use inline @code{movstrhi} patterns for copying memory.
10409 Use 16-bit @code{int}. This is the default.
10415 Use 32-bit @code{int}.
10418 @itemx -mno-float32
10420 @opindex mno-float32
10421 Use 64-bit @code{float}. This is the default.
10424 @itemx -mno-float64
10426 @opindex mno-float64
10427 Use 32-bit @code{float}.
10431 Use @code{abshi2} pattern. This is the default.
10435 Do not use @code{abshi2} pattern.
10437 @item -mbranch-expensive
10438 @opindex mbranch-expensive
10439 Pretend that branches are expensive. This is for experimenting with
10440 code generation only.
10442 @item -mbranch-cheap
10443 @opindex mbranch-cheap
10444 Do not pretend that branches are expensive. This is the default.
10448 Generate code for a system with split I&D.
10452 Generate code for a system without split I&D. This is the default.
10456 Use Unix assembler syntax. This is the default when configured for
10457 @samp{pdp11-*-bsd}.
10461 Use DEC assembler syntax. This is the default when configured for any
10462 PDP-11 target other than @samp{pdp11-*-bsd}.
10465 @node Xstormy16 Options
10466 @subsection Xstormy16 Options
10467 @cindex Xstormy16 Options
10469 These options are defined for Xstormy16:
10474 Choose startup files and linker script suitable for the simulator.
10478 @subsection FRV Options
10479 @cindex FRV Options
10485 Only use the first 32 general purpose registers.
10490 Use all 64 general purpose registers.
10495 Use only the first 32 floating point registers.
10500 Use all 64 floating point registers
10503 @opindex mhard-float
10505 Use hardware instructions for floating point operations.
10508 @opindex msoft-float
10510 Use library routines for floating point operations.
10515 Dynamically allocate condition code registers.
10520 Do not try to dynamically allocate condition code registers, only
10521 use @code{icc0} and @code{fcc0}.
10526 Change ABI to use double word insns.
10531 Do not use double word instructions.
10536 Use floating point double instructions.
10539 @opindex mno-double
10541 Do not use floating point double instructions.
10546 Use media instructions.
10551 Do not use media instructions.
10556 Use multiply and add/subtract instructions.
10559 @opindex mno-muladd
10561 Do not use multiply and add/subtract instructions.
10563 @item -mlibrary-pic
10564 @opindex mlibrary-pic
10566 Generate position-independent EABI code.
10571 Use only the first four media accumulator registers.
10576 Use all eight media accumulator registers.
10581 Pack VLIW instructions.
10586 Do not pack VLIW instructions.
10589 @opindex mno-eflags
10591 Do not mark ABI switches in e_flags.
10594 @opindex mcond-move
10596 Enable the use of conditional-move instructions (default).
10598 This switch is mainly for debugging the compiler and will likely be removed
10599 in a future version.
10601 @item -mno-cond-move
10602 @opindex mno-cond-move
10604 Disable the use of conditional-move instructions.
10606 This switch is mainly for debugging the compiler and will likely be removed
10607 in a future version.
10612 Enable the use of conditional set instructions (default).
10614 This switch is mainly for debugging the compiler and will likely be removed
10615 in a future version.
10620 Disable the use of conditional set instructions.
10622 This switch is mainly for debugging the compiler and will likely be removed
10623 in a future version.
10626 @opindex mcond-exec
10628 Enable the use of conditional execution (default).
10630 This switch is mainly for debugging the compiler and will likely be removed
10631 in a future version.
10633 @item -mno-cond-exec
10634 @opindex mno-cond-exec
10636 Disable the use of conditional execution.
10638 This switch is mainly for debugging the compiler and will likely be removed
10639 in a future version.
10641 @item -mvliw-branch
10642 @opindex mvliw-branch
10644 Run a pass to pack branches into VLIW instructions (default).
10646 This switch is mainly for debugging the compiler and will likely be removed
10647 in a future version.
10649 @item -mno-vliw-branch
10650 @opindex mno-vliw-branch
10652 Do not run a pass to pack branches into VLIW instructions.
10654 This switch is mainly for debugging the compiler and will likely be removed
10655 in a future version.
10657 @item -mmulti-cond-exec
10658 @opindex mmulti-cond-exec
10660 Enable optimization of @code{&&} and @code{||} in conditional execution
10663 This switch is mainly for debugging the compiler and will likely be removed
10664 in a future version.
10666 @item -mno-multi-cond-exec
10667 @opindex mno-multi-cond-exec
10669 Disable optimization of @code{&&} and @code{||} in conditional execution.
10671 This switch is mainly for debugging the compiler and will likely be removed
10672 in a future version.
10674 @item -mnested-cond-exec
10675 @opindex mnested-cond-exec
10677 Enable nested conditional execution optimizations (default).
10679 This switch is mainly for debugging the compiler and will likely be removed
10680 in a future version.
10682 @item -mno-nested-cond-exec
10683 @opindex mno-nested-cond-exec
10685 Disable nested conditional execution optimizations.
10687 This switch is mainly for debugging the compiler and will likely be removed
10688 in a future version.
10690 @item -mtomcat-stats
10691 @opindex mtomcat-stats
10693 Cause gas to print out tomcat statistics.
10695 @item -mcpu=@var{cpu}
10698 Select the processor type for which to generate code. Possible values are
10699 @samp{simple}, @samp{tomcat}, @samp{fr500}, @samp{fr400}, @samp{fr300},
10704 @node Xtensa Options
10705 @subsection Xtensa Options
10706 @cindex Xtensa Options
10708 These options are supported for Xtensa targets:
10712 @itemx -mno-const16
10714 @opindex mno-const16
10715 Enable or disable use of @code{CONST16} instructions for loading
10716 constant values. The @code{CONST16} instruction is currently not a
10717 standard option from Tensilica. When enabled, @code{CONST16}
10718 instructions are always used in place of the standard @code{L32R}
10719 instructions. The use of @code{CONST16} is enabled by default only if
10720 the @code{L32R} instruction is not available.
10723 @itemx -mno-fused-madd
10724 @opindex mfused-madd
10725 @opindex mno-fused-madd
10726 Enable or disable use of fused multiply/add and multiply/subtract
10727 instructions in the floating-point option. This has no effect if the
10728 floating-point option is not also enabled. Disabling fused multiply/add
10729 and multiply/subtract instructions forces the compiler to use separate
10730 instructions for the multiply and add/subtract operations. This may be
10731 desirable in some cases where strict IEEE 754-compliant results are
10732 required: the fused multiply add/subtract instructions do not round the
10733 intermediate result, thereby producing results with @emph{more} bits of
10734 precision than specified by the IEEE standard. Disabling fused multiply
10735 add/subtract instructions also ensures that the program output is not
10736 sensitive to the compiler's ability to combine multiply and add/subtract
10739 @item -mtext-section-literals
10740 @itemx -mno-text-section-literals
10741 @opindex mtext-section-literals
10742 @opindex mno-text-section-literals
10743 Control the treatment of literal pools. The default is
10744 @option{-mno-text-section-literals}, which places literals in a separate
10745 section in the output file. This allows the literal pool to be placed
10746 in a data RAM/ROM, and it also allows the linker to combine literal
10747 pools from separate object files to remove redundant literals and
10748 improve code size. With @option{-mtext-section-literals}, the literals
10749 are interspersed in the text section in order to keep them as close as
10750 possible to their references. This may be necessary for large assembly
10753 @item -mtarget-align
10754 @itemx -mno-target-align
10755 @opindex mtarget-align
10756 @opindex mno-target-align
10757 When this option is enabled, GCC instructs the assembler to
10758 automatically align instructions to reduce branch penalties at the
10759 expense of some code density. The assembler attempts to widen density
10760 instructions to align branch targets and the instructions following call
10761 instructions. If there are not enough preceding safe density
10762 instructions to align a target, no widening will be performed. The
10763 default is @option{-mtarget-align}. These options do not affect the
10764 treatment of auto-aligned instructions like @code{LOOP}, which the
10765 assembler will always align, either by widening density instructions or
10766 by inserting no-op instructions.
10769 @itemx -mno-longcalls
10770 @opindex mlongcalls
10771 @opindex mno-longcalls
10772 When this option is enabled, GCC instructs the assembler to translate
10773 direct calls to indirect calls unless it can determine that the target
10774 of a direct call is in the range allowed by the call instruction. This
10775 translation typically occurs for calls to functions in other source
10776 files. Specifically, the assembler translates a direct @code{CALL}
10777 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
10778 The default is @option{-mno-longcalls}. This option should be used in
10779 programs where the call target can potentially be out of range. This
10780 option is implemented in the assembler, not the compiler, so the
10781 assembly code generated by GCC will still show direct call
10782 instructions---look at the disassembled object code to see the actual
10783 instructions. Note that the assembler will use an indirect call for
10784 every cross-file call, not just those that really will be out of range.
10787 @node Code Gen Options
10788 @section Options for Code Generation Conventions
10789 @cindex code generation conventions
10790 @cindex options, code generation
10791 @cindex run-time options
10793 These machine-independent options control the interface conventions
10794 used in code generation.
10796 Most of them have both positive and negative forms; the negative form
10797 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
10798 one of the forms is listed---the one which is not the default. You
10799 can figure out the other form by either removing @samp{no-} or adding
10803 @item -fbounds-check
10804 @opindex fbounds-check
10805 For front-ends that support it, generate additional code to check that
10806 indices used to access arrays are within the declared range. This is
10807 currently only supported by the Java and Fortran 77 front-ends, where
10808 this option defaults to true and false respectively.
10812 This option generates traps for signed overflow on addition, subtraction,
10813 multiplication operations.
10817 This option instructs the compiler to assume that signed arithmetic
10818 overflow of addition, subtraction and multiplication wraps around
10819 using twos-complement representation. This flag enables some optimizations
10820 and disables other. This option is enabled by default for the Java
10821 front-end, as required by the Java language specification.
10824 @opindex fexceptions
10825 Enable exception handling. Generates extra code needed to propagate
10826 exceptions. For some targets, this implies GCC will generate frame
10827 unwind information for all functions, which can produce significant data
10828 size overhead, although it does not affect execution. If you do not
10829 specify this option, GCC will enable it by default for languages like
10830 C++ which normally require exception handling, and disable it for
10831 languages like C that do not normally require it. However, you may need
10832 to enable this option when compiling C code that needs to interoperate
10833 properly with exception handlers written in C++. You may also wish to
10834 disable this option if you are compiling older C++ programs that don't
10835 use exception handling.
10837 @item -fnon-call-exceptions
10838 @opindex fnon-call-exceptions
10839 Generate code that allows trapping instructions to throw exceptions.
10840 Note that this requires platform-specific runtime support that does
10841 not exist everywhere. Moreover, it only allows @emph{trapping}
10842 instructions to throw exceptions, i.e.@: memory references or floating
10843 point instructions. It does not allow exceptions to be thrown from
10844 arbitrary signal handlers such as @code{SIGALRM}.
10846 @item -funwind-tables
10847 @opindex funwind-tables
10848 Similar to @option{-fexceptions}, except that it will just generate any needed
10849 static data, but will not affect the generated code in any other way.
10850 You will normally not enable this option; instead, a language processor
10851 that needs this handling would enable it on your behalf.
10853 @item -fasynchronous-unwind-tables
10854 @opindex funwind-tables
10855 Generate unwind table in dwarf2 format, if supported by target machine. The
10856 table is exact at each instruction boundary, so it can be used for stack
10857 unwinding from asynchronous events (such as debugger or garbage collector).
10859 @item -fpcc-struct-return
10860 @opindex fpcc-struct-return
10861 Return ``short'' @code{struct} and @code{union} values in memory like
10862 longer ones, rather than in registers. This convention is less
10863 efficient, but it has the advantage of allowing intercallability between
10864 GCC-compiled files and files compiled with other compilers, particularly
10865 the Portable C Compiler (pcc).
10867 The precise convention for returning structures in memory depends
10868 on the target configuration macros.
10870 Short structures and unions are those whose size and alignment match
10871 that of some integer type.
10873 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
10874 switch is not binary compatible with code compiled with the
10875 @option{-freg-struct-return} switch.
10876 Use it to conform to a non-default application binary interface.
10878 @item -freg-struct-return
10879 @opindex freg-struct-return
10880 Return @code{struct} and @code{union} values in registers when possible.
10881 This is more efficient for small structures than
10882 @option{-fpcc-struct-return}.
10884 If you specify neither @option{-fpcc-struct-return} nor
10885 @option{-freg-struct-return}, GCC defaults to whichever convention is
10886 standard for the target. If there is no standard convention, GCC
10887 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
10888 the principal compiler. In those cases, we can choose the standard, and
10889 we chose the more efficient register return alternative.
10891 @strong{Warning:} code compiled with the @option{-freg-struct-return}
10892 switch is not binary compatible with code compiled with the
10893 @option{-fpcc-struct-return} switch.
10894 Use it to conform to a non-default application binary interface.
10896 @item -fshort-enums
10897 @opindex fshort-enums
10898 Allocate to an @code{enum} type only as many bytes as it needs for the
10899 declared range of possible values. Specifically, the @code{enum} type
10900 will be equivalent to the smallest integer type which has enough room.
10902 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
10903 code that is not binary compatible with code generated without that switch.
10904 Use it to conform to a non-default application binary interface.
10906 @item -fshort-double
10907 @opindex fshort-double
10908 Use the same size for @code{double} as for @code{float}.
10910 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
10911 code that is not binary compatible with code generated without that switch.
10912 Use it to conform to a non-default application binary interface.
10914 @item -fshort-wchar
10915 @opindex fshort-wchar
10916 Override the underlying type for @samp{wchar_t} to be @samp{short
10917 unsigned int} instead of the default for the target. This option is
10918 useful for building programs to run under WINE@.
10920 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
10921 code that is not binary compatible with code generated without that switch.
10922 Use it to conform to a non-default application binary interface.
10924 @item -fshared-data
10925 @opindex fshared-data
10926 Requests that the data and non-@code{const} variables of this
10927 compilation be shared data rather than private data. The distinction
10928 makes sense only on certain operating systems, where shared data is
10929 shared between processes running the same program, while private data
10930 exists in one copy per process.
10933 @opindex fno-common
10934 In C, allocate even uninitialized global variables in the data section of the
10935 object file, rather than generating them as common blocks. This has the
10936 effect that if the same variable is declared (without @code{extern}) in
10937 two different compilations, you will get an error when you link them.
10938 The only reason this might be useful is if you wish to verify that the
10939 program will work on other systems which always work this way.
10943 Ignore the @samp{#ident} directive.
10945 @item -finhibit-size-directive
10946 @opindex finhibit-size-directive
10947 Don't output a @code{.size} assembler directive, or anything else that
10948 would cause trouble if the function is split in the middle, and the
10949 two halves are placed at locations far apart in memory. This option is
10950 used when compiling @file{crtstuff.c}; you should not need to use it
10953 @item -fverbose-asm
10954 @opindex fverbose-asm
10955 Put extra commentary information in the generated assembly code to
10956 make it more readable. This option is generally only of use to those
10957 who actually need to read the generated assembly code (perhaps while
10958 debugging the compiler itself).
10960 @option{-fno-verbose-asm}, the default, causes the
10961 extra information to be omitted and is useful when comparing two assembler
10966 @cindex global offset table
10968 Generate position-independent code (PIC) suitable for use in a shared
10969 library, if supported for the target machine. Such code accesses all
10970 constant addresses through a global offset table (GOT)@. The dynamic
10971 loader resolves the GOT entries when the program starts (the dynamic
10972 loader is not part of GCC; it is part of the operating system). If
10973 the GOT size for the linked executable exceeds a machine-specific
10974 maximum size, you get an error message from the linker indicating that
10975 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
10976 instead. (These maximums are 8k on the SPARC and 32k
10977 on the m68k and RS/6000. The 386 has no such limit.)
10979 Position-independent code requires special support, and therefore works
10980 only on certain machines. For the 386, GCC supports PIC for System V
10981 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
10982 position-independent.
10986 If supported for the target machine, emit position-independent code,
10987 suitable for dynamic linking and avoiding any limit on the size of the
10988 global offset table. This option makes a difference on the m68k
10991 Position-independent code requires special support, and therefore works
10992 only on certain machines.
10998 These options are similar to @option{-fpic} and @option{-fPIC}, but
10999 generated position independent code can be only linked into executables.
11000 Usually these options are used when @option{-pie} GCC option will be
11001 used during linking.
11003 @item -ffixed-@var{reg}
11005 Treat the register named @var{reg} as a fixed register; generated code
11006 should never refer to it (except perhaps as a stack pointer, frame
11007 pointer or in some other fixed role).
11009 @var{reg} must be the name of a register. The register names accepted
11010 are machine-specific and are defined in the @code{REGISTER_NAMES}
11011 macro in the machine description macro file.
11013 This flag does not have a negative form, because it specifies a
11016 @item -fcall-used-@var{reg}
11017 @opindex fcall-used
11018 Treat the register named @var{reg} as an allocable register that is
11019 clobbered by function calls. It may be allocated for temporaries or
11020 variables that do not live across a call. Functions compiled this way
11021 will not save and restore the register @var{reg}.
11023 It is an error to used this flag with the frame pointer or stack pointer.
11024 Use of this flag for other registers that have fixed pervasive roles in
11025 the machine's execution model will produce disastrous results.
11027 This flag does not have a negative form, because it specifies a
11030 @item -fcall-saved-@var{reg}
11031 @opindex fcall-saved
11032 Treat the register named @var{reg} as an allocable register saved by
11033 functions. It may be allocated even for temporaries or variables that
11034 live across a call. Functions compiled this way will save and restore
11035 the register @var{reg} if they use it.
11037 It is an error to used this flag with the frame pointer or stack pointer.
11038 Use of this flag for other registers that have fixed pervasive roles in
11039 the machine's execution model will produce disastrous results.
11041 A different sort of disaster will result from the use of this flag for
11042 a register in which function values may be returned.
11044 This flag does not have a negative form, because it specifies a
11047 @item -fpack-struct
11048 @opindex fpack-struct
11049 Pack all structure members together without holes.
11051 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
11052 code that is not binary compatible with code generated without that switch.
11053 Additionally, it makes the code suboptimal.
11054 Use it to conform to a non-default application binary interface.
11056 @item -finstrument-functions
11057 @opindex finstrument-functions
11058 Generate instrumentation calls for entry and exit to functions. Just
11059 after function entry and just before function exit, the following
11060 profiling functions will be called with the address of the current
11061 function and its call site. (On some platforms,
11062 @code{__builtin_return_address} does not work beyond the current
11063 function, so the call site information may not be available to the
11064 profiling functions otherwise.)
11067 void __cyg_profile_func_enter (void *this_fn,
11069 void __cyg_profile_func_exit (void *this_fn,
11073 The first argument is the address of the start of the current function,
11074 which may be looked up exactly in the symbol table.
11076 This instrumentation is also done for functions expanded inline in other
11077 functions. The profiling calls will indicate where, conceptually, the
11078 inline function is entered and exited. This means that addressable
11079 versions of such functions must be available. If all your uses of a
11080 function are expanded inline, this may mean an additional expansion of
11081 code size. If you use @samp{extern inline} in your C code, an
11082 addressable version of such functions must be provided. (This is
11083 normally the case anyways, but if you get lucky and the optimizer always
11084 expands the functions inline, you might have gotten away without
11085 providing static copies.)
11087 A function may be given the attribute @code{no_instrument_function}, in
11088 which case this instrumentation will not be done. This can be used, for
11089 example, for the profiling functions listed above, high-priority
11090 interrupt routines, and any functions from which the profiling functions
11091 cannot safely be called (perhaps signal handlers, if the profiling
11092 routines generate output or allocate memory).
11094 @item -fstack-check
11095 @opindex fstack-check
11096 Generate code to verify that you do not go beyond the boundary of the
11097 stack. You should specify this flag if you are running in an
11098 environment with multiple threads, but only rarely need to specify it in
11099 a single-threaded environment since stack overflow is automatically
11100 detected on nearly all systems if there is only one stack.
11102 Note that this switch does not actually cause checking to be done; the
11103 operating system must do that. The switch causes generation of code
11104 to ensure that the operating system sees the stack being extended.
11106 @item -fstack-limit-register=@var{reg}
11107 @itemx -fstack-limit-symbol=@var{sym}
11108 @itemx -fno-stack-limit
11109 @opindex fstack-limit-register
11110 @opindex fstack-limit-symbol
11111 @opindex fno-stack-limit
11112 Generate code to ensure that the stack does not grow beyond a certain value,
11113 either the value of a register or the address of a symbol. If the stack
11114 would grow beyond the value, a signal is raised. For most targets,
11115 the signal is raised before the stack overruns the boundary, so
11116 it is possible to catch the signal without taking special precautions.
11118 For instance, if the stack starts at absolute address @samp{0x80000000}
11119 and grows downwards, you can use the flags
11120 @option{-fstack-limit-symbol=__stack_limit} and
11121 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
11122 of 128KB@. Note that this may only work with the GNU linker.
11124 @cindex aliasing of parameters
11125 @cindex parameters, aliased
11126 @item -fargument-alias
11127 @itemx -fargument-noalias
11128 @itemx -fargument-noalias-global
11129 @opindex fargument-alias
11130 @opindex fargument-noalias
11131 @opindex fargument-noalias-global
11132 Specify the possible relationships among parameters and between
11133 parameters and global data.
11135 @option{-fargument-alias} specifies that arguments (parameters) may
11136 alias each other and may alias global storage.@*
11137 @option{-fargument-noalias} specifies that arguments do not alias
11138 each other, but may alias global storage.@*
11139 @option{-fargument-noalias-global} specifies that arguments do not
11140 alias each other and do not alias global storage.
11142 Each language will automatically use whatever option is required by
11143 the language standard. You should not need to use these options yourself.
11145 @item -fleading-underscore
11146 @opindex fleading-underscore
11147 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
11148 change the way C symbols are represented in the object file. One use
11149 is to help link with legacy assembly code.
11151 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
11152 generate code that is not binary compatible with code generated without that
11153 switch. Use it to conform to a non-default application binary interface.
11154 Not all targets provide complete support for this switch.
11156 @item -ftls-model=@var{model}
11157 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
11158 The @var{model} argument should be one of @code{global-dynamic},
11159 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
11161 The default without @option{-fpic} is @code{initial-exec}; with
11162 @option{-fpic} the default is @code{global-dynamic}.
11167 @node Environment Variables
11168 @section Environment Variables Affecting GCC
11169 @cindex environment variables
11171 @c man begin ENVIRONMENT
11172 This section describes several environment variables that affect how GCC
11173 operates. Some of them work by specifying directories or prefixes to use
11174 when searching for various kinds of files. Some are used to specify other
11175 aspects of the compilation environment.
11177 Note that you can also specify places to search using options such as
11178 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
11179 take precedence over places specified using environment variables, which
11180 in turn take precedence over those specified by the configuration of GCC@.
11181 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
11182 GNU Compiler Collection (GCC) Internals}.
11187 @c @itemx LC_COLLATE
11189 @c @itemx LC_MONETARY
11190 @c @itemx LC_NUMERIC
11195 @c @findex LC_COLLATE
11196 @findex LC_MESSAGES
11197 @c @findex LC_MONETARY
11198 @c @findex LC_NUMERIC
11202 These environment variables control the way that GCC uses
11203 localization information that allow GCC to work with different
11204 national conventions. GCC inspects the locale categories
11205 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
11206 so. These locale categories can be set to any value supported by your
11207 installation. A typical value is @samp{en_UK} for English in the United
11210 The @env{LC_CTYPE} environment variable specifies character
11211 classification. GCC uses it to determine the character boundaries in
11212 a string; this is needed for some multibyte encodings that contain quote
11213 and escape characters that would otherwise be interpreted as a string
11216 The @env{LC_MESSAGES} environment variable specifies the language to
11217 use in diagnostic messages.
11219 If the @env{LC_ALL} environment variable is set, it overrides the value
11220 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
11221 and @env{LC_MESSAGES} default to the value of the @env{LANG}
11222 environment variable. If none of these variables are set, GCC
11223 defaults to traditional C English behavior.
11227 If @env{TMPDIR} is set, it specifies the directory to use for temporary
11228 files. GCC uses temporary files to hold the output of one stage of
11229 compilation which is to be used as input to the next stage: for example,
11230 the output of the preprocessor, which is the input to the compiler
11233 @item GCC_EXEC_PREFIX
11234 @findex GCC_EXEC_PREFIX
11235 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
11236 names of the subprograms executed by the compiler. No slash is added
11237 when this prefix is combined with the name of a subprogram, but you can
11238 specify a prefix that ends with a slash if you wish.
11240 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
11241 an appropriate prefix to use based on the pathname it was invoked with.
11243 If GCC cannot find the subprogram using the specified prefix, it
11244 tries looking in the usual places for the subprogram.
11246 The default value of @env{GCC_EXEC_PREFIX} is
11247 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
11248 of @code{prefix} when you ran the @file{configure} script.
11250 Other prefixes specified with @option{-B} take precedence over this prefix.
11252 This prefix is also used for finding files such as @file{crt0.o} that are
11255 In addition, the prefix is used in an unusual way in finding the
11256 directories to search for header files. For each of the standard
11257 directories whose name normally begins with @samp{/usr/local/lib/gcc}
11258 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
11259 replacing that beginning with the specified prefix to produce an
11260 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
11261 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
11262 These alternate directories are searched first; the standard directories
11265 @item COMPILER_PATH
11266 @findex COMPILER_PATH
11267 The value of @env{COMPILER_PATH} is a colon-separated list of
11268 directories, much like @env{PATH}. GCC tries the directories thus
11269 specified when searching for subprograms, if it can't find the
11270 subprograms using @env{GCC_EXEC_PREFIX}.
11273 @findex LIBRARY_PATH
11274 The value of @env{LIBRARY_PATH} is a colon-separated list of
11275 directories, much like @env{PATH}. When configured as a native compiler,
11276 GCC tries the directories thus specified when searching for special
11277 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
11278 using GCC also uses these directories when searching for ordinary
11279 libraries for the @option{-l} option (but directories specified with
11280 @option{-L} come first).
11284 @cindex locale definition
11285 This variable is used to pass locale information to the compiler. One way in
11286 which this information is used is to determine the character set to be used
11287 when character literals, string literals and comments are parsed in C and C++.
11288 When the compiler is configured to allow multibyte characters,
11289 the following values for @env{LANG} are recognized:
11293 Recognize JIS characters.
11295 Recognize SJIS characters.
11297 Recognize EUCJP characters.
11300 If @env{LANG} is not defined, or if it has some other value, then the
11301 compiler will use mblen and mbtowc as defined by the default locale to
11302 recognize and translate multibyte characters.
11306 Some additional environments variables affect the behavior of the
11309 @include cppenv.texi
11313 @node Precompiled Headers
11314 @section Using Precompiled Headers
11315 @cindex precompiled headers
11316 @cindex speed of compilation
11318 Often large projects have many header files that are included in every
11319 source file. The time the compiler takes to process these header files
11320 over and over again can account for nearly all of the time required to
11321 build the project. To make builds faster, GCC allows users to
11322 `precompile' a header file; then, if builds can use the precompiled
11323 header file they will be much faster.
11325 To create a precompiled header file, simply compile it as you would any
11326 other file, if necessary using the @option{-x} option to make the driver
11327 treat it as a C or C++ header file. You will probably want to use a
11328 tool like @command{make} to keep the precompiled header up-to-date when
11329 the headers it contains change.
11331 A precompiled header file will be searched for when @code{#include} is
11332 seen in the compilation. As it searches for the included file
11333 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
11334 compiler looks for a precompiled header in each directory just before it
11335 looks for the include file in that directory. The name searched for is
11336 the name specified in the @code{#include} with @samp{.gch} appended. If
11337 the precompiled header file can't be used, it is ignored.
11339 For instance, if you have @code{#include "all.h"}, and you have
11340 @file{all.h.gch} in the same directory as @file{all.h}, then the
11341 precompiled header file will be used if possible, and the original
11342 header will be used otherwise.
11344 Alternatively, you might decide to put the precompiled header file in a
11345 directory and use @option{-I} to ensure that directory is searched
11346 before (or instead of) the directory containing the original header.
11347 Then, if you want to check that the precompiled header file is always
11348 used, you can put a file of the same name as the original header in this
11349 directory containing an @code{#error} command.
11351 This also works with @option{-include}. So yet another way to use
11352 precompiled headers, good for projects not designed with precompiled
11353 header files in mind, is to simply take most of the header files used by
11354 a project, include them from another header file, precompile that header
11355 file, and @option{-include} the precompiled header. If the header files
11356 have guards against multiple inclusion, they will be skipped because
11357 they've already been included (in the precompiled header).
11359 If you need to precompile the same header file for different
11360 languages, targets, or compiler options, you can instead make a
11361 @emph{directory} named like @file{all.h.gch}, and put each precompiled
11362 header in the directory. (It doesn't matter what you call the files
11363 in the directory, every precompiled header in the directory will be
11364 considered.) The first precompiled header encountered in the
11365 directory that is valid for this compilation will be used; they're
11366 searched in no particular order.
11368 There are many other possibilities, limited only by your imagination,
11369 good sense, and the constraints of your build system.
11371 A precompiled header file can be used only when these conditions apply:
11375 Only one precompiled header can be used in a particular compilation.
11377 A precompiled header can't be used once the first C token is seen. You
11378 can have preprocessor directives before a precompiled header; you can
11379 even include a precompiled header from inside another header, so long as
11380 there are no C tokens before the @code{#include}.
11382 The precompiled header file must be produced for the same language as
11383 the current compilation. You can't use a C precompiled header for a C++
11386 The precompiled header file must be produced by the same compiler
11387 version and configuration as the current compilation is using.
11388 The easiest way to guarantee this is to use the same compiler binary
11389 for creating and using precompiled headers.
11391 Any macros defined before the precompiled header (including with
11392 @option{-D}) must either be defined in the same way as when the
11393 precompiled header was generated, or must not affect the precompiled
11394 header, which usually means that the they don't appear in the
11395 precompiled header at all.
11397 Certain command-line options must be defined in the same way as when the
11398 precompiled header was generated. At present, it's not clear which
11399 options are safe to change and which are not; the safest choice is to
11400 use exactly the same options when generating and using the precompiled
11404 For all of these but the last, the compiler will automatically ignore
11405 the precompiled header if the conditions aren't met. For the last item,
11406 some option changes will cause the precompiled header to be rejected,
11407 but not all incompatible option combinations have yet been found. If
11408 you find a new incompatible combination, please consider filing a bug
11409 report, see @ref{Bugs}.
11411 @node Running Protoize
11412 @section Running Protoize
11414 The program @code{protoize} is an optional part of GCC@. You can use
11415 it to add prototypes to a program, thus converting the program to ISO
11416 C in one respect. The companion program @code{unprotoize} does the
11417 reverse: it removes argument types from any prototypes that are found.
11419 When you run these programs, you must specify a set of source files as
11420 command line arguments. The conversion programs start out by compiling
11421 these files to see what functions they define. The information gathered
11422 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
11424 After scanning comes actual conversion. The specified files are all
11425 eligible to be converted; any files they include (whether sources or
11426 just headers) are eligible as well.
11428 But not all the eligible files are converted. By default,
11429 @code{protoize} and @code{unprotoize} convert only source and header
11430 files in the current directory. You can specify additional directories
11431 whose files should be converted with the @option{-d @var{directory}}
11432 option. You can also specify particular files to exclude with the
11433 @option{-x @var{file}} option. A file is converted if it is eligible, its
11434 directory name matches one of the specified directory names, and its
11435 name within the directory has not been excluded.
11437 Basic conversion with @code{protoize} consists of rewriting most
11438 function definitions and function declarations to specify the types of
11439 the arguments. The only ones not rewritten are those for varargs
11442 @code{protoize} optionally inserts prototype declarations at the
11443 beginning of the source file, to make them available for any calls that
11444 precede the function's definition. Or it can insert prototype
11445 declarations with block scope in the blocks where undeclared functions
11448 Basic conversion with @code{unprotoize} consists of rewriting most
11449 function declarations to remove any argument types, and rewriting
11450 function definitions to the old-style pre-ISO form.
11452 Both conversion programs print a warning for any function declaration or
11453 definition that they can't convert. You can suppress these warnings
11456 The output from @code{protoize} or @code{unprotoize} replaces the
11457 original source file. The original file is renamed to a name ending
11458 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
11459 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
11460 for DOS) file already exists, then the source file is simply discarded.
11462 @code{protoize} and @code{unprotoize} both depend on GCC itself to
11463 scan the program and collect information about the functions it uses.
11464 So neither of these programs will work until GCC is installed.
11466 Here is a table of the options you can use with @code{protoize} and
11467 @code{unprotoize}. Each option works with both programs unless
11471 @item -B @var{directory}
11472 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
11473 usual directory (normally @file{/usr/local/lib}). This file contains
11474 prototype information about standard system functions. This option
11475 applies only to @code{protoize}.
11477 @item -c @var{compilation-options}
11478 Use @var{compilation-options} as the options when running @command{gcc} to
11479 produce the @samp{.X} files. The special option @option{-aux-info} is
11480 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
11482 Note that the compilation options must be given as a single argument to
11483 @code{protoize} or @code{unprotoize}. If you want to specify several
11484 @command{gcc} options, you must quote the entire set of compilation options
11485 to make them a single word in the shell.
11487 There are certain @command{gcc} arguments that you cannot use, because they
11488 would produce the wrong kind of output. These include @option{-g},
11489 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
11490 the @var{compilation-options}, they are ignored.
11493 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
11494 systems) instead of @samp{.c}. This is convenient if you are converting
11495 a C program to C++. This option applies only to @code{protoize}.
11498 Add explicit global declarations. This means inserting explicit
11499 declarations at the beginning of each source file for each function
11500 that is called in the file and was not declared. These declarations
11501 precede the first function definition that contains a call to an
11502 undeclared function. This option applies only to @code{protoize}.
11504 @item -i @var{string}
11505 Indent old-style parameter declarations with the string @var{string}.
11506 This option applies only to @code{protoize}.
11508 @code{unprotoize} converts prototyped function definitions to old-style
11509 function definitions, where the arguments are declared between the
11510 argument list and the initial @samp{@{}. By default, @code{unprotoize}
11511 uses five spaces as the indentation. If you want to indent with just
11512 one space instead, use @option{-i " "}.
11515 Keep the @samp{.X} files. Normally, they are deleted after conversion
11519 Add explicit local declarations. @code{protoize} with @option{-l} inserts
11520 a prototype declaration for each function in each block which calls the
11521 function without any declaration. This option applies only to
11525 Make no real changes. This mode just prints information about the conversions
11526 that would have been done without @option{-n}.
11529 Make no @samp{.save} files. The original files are simply deleted.
11530 Use this option with caution.
11532 @item -p @var{program}
11533 Use the program @var{program} as the compiler. Normally, the name
11534 @file{gcc} is used.
11537 Work quietly. Most warnings are suppressed.
11540 Print the version number, just like @option{-v} for @command{gcc}.
11543 If you need special compiler options to compile one of your program's
11544 source files, then you should generate that file's @samp{.X} file
11545 specially, by running @command{gcc} on that source file with the
11546 appropriate options and the option @option{-aux-info}. Then run
11547 @code{protoize} on the entire set of files. @code{protoize} will use
11548 the existing @samp{.X} file because it is newer than the source file.
11552 gcc -Dfoo=bar file1.c -aux-info file1.X
11557 You need to include the special files along with the rest in the
11558 @code{protoize} command, even though their @samp{.X} files already
11559 exist, because otherwise they won't get converted.
11561 @xref{Protoize Caveats}, for more information on how to use
11562 @code{protoize} successfully.