1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003 Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
8 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
9 1998, 1999, 2000, 2001, 2002, 2003 Free Software Foundation, Inc.
11 Permission is granted to copy, distribute and/or modify this document
12 under the terms of the GNU Free Documentation License, Version 1.2 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU project C and C++ compiler
33 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
39 [@option{-o} @var{outfile}] @var{infile}@dots{}
41 Only the most useful options are listed here; see below for the
42 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
45 gpl(7), gfdl(7), fsf-funding(7),
46 cpp(1), gcov(1), g77(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
47 and the Info entries for @file{gcc}, @file{cpp}, @file{g77}, @file{as},
48 @file{ld}, @file{binutils} and @file{gdb}.
51 For instructions on reporting bugs, see
52 @w{@uref{http://gcc.gnu.org/bugs.html}}. Use of the @command{gccbug}
53 script to report bugs is recommended.
56 See the Info entry for @command{gcc}, or
57 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
58 for contributors to GCC@.
63 @chapter GCC Command Options
64 @cindex GCC command options
65 @cindex command options
66 @cindex options, GCC command
68 @c man begin DESCRIPTION
69 When you invoke GCC, it normally does preprocessing, compilation,
70 assembly and linking. The ``overall options'' allow you to stop this
71 process at an intermediate stage. For example, the @option{-c} option
72 says not to run the linker. Then the output consists of object files
73 output by the assembler.
75 Other options are passed on to one stage of processing. Some options
76 control the preprocessor and others the compiler itself. Yet other
77 options control the assembler and linker; most of these are not
78 documented here, since you rarely need to use any of them.
80 @cindex C compilation options
81 Most of the command line options that you can use with GCC are useful
82 for C programs; when an option is only useful with another language
83 (usually C++), the explanation says so explicitly. If the description
84 for a particular option does not mention a source language, you can use
85 that option with all supported languages.
87 @cindex C++ compilation options
88 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
89 options for compiling C++ programs.
91 @cindex grouping options
92 @cindex options, grouping
93 The @command{gcc} program accepts options and file names as operands. Many
94 options have multi-letter names; therefore multiple single-letter options
95 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
98 @cindex order of options
99 @cindex options, order
100 You can mix options and other arguments. For the most part, the order
101 you use doesn't matter. Order does matter when you use several options
102 of the same kind; for example, if you specify @option{-L} more than once,
103 the directories are searched in the order specified.
105 Many options have long names starting with @samp{-f} or with
106 @samp{-W}---for example, @option{-fforce-mem},
107 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
108 these have both positive and negative forms; the negative form of
109 @option{-ffoo} would be @option{-fno-foo}. This manual documents
110 only one of these two forms, whichever one is not the default.
114 @xref{Option Index}, for an index to GCC's options.
117 * Option Summary:: Brief list of all options, without explanations.
118 * Overall Options:: Controlling the kind of output:
119 an executable, object files, assembler files,
120 or preprocessed source.
121 * Invoking G++:: Compiling C++ programs.
122 * C Dialect Options:: Controlling the variant of C language compiled.
123 * C++ Dialect Options:: Variations on C++.
124 * Objective-C Dialect Options:: Variations on Objective-C.
125 * Language Independent Options:: Controlling how diagnostics should be
127 * Warning Options:: How picky should the compiler be?
128 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
129 * Optimize Options:: How much optimization?
130 * Preprocessor Options:: Controlling header files and macro definitions.
131 Also, getting dependency information for Make.
132 * Assembler Options:: Passing options to the assembler.
133 * Link Options:: Specifying libraries and so on.
134 * Directory Options:: Where to find header files and libraries.
135 Where to find the compiler executable files.
136 * Spec Files:: How to pass switches to sub-processes.
137 * Target Options:: Running a cross-compiler, or an old version of GCC.
138 * Submodel Options:: Specifying minor hardware or convention variations,
139 such as 68010 vs 68020.
140 * Code Gen Options:: Specifying conventions for function calls, data layout
142 * Environment Variables:: Env vars that affect GCC.
143 * Precompiled Headers:: Compiling a header once, and using it many times.
144 * Running Protoize:: Automatically adding or removing function prototypes.
150 @section Option Summary
152 Here is a summary of all the options, grouped by type. Explanations are
153 in the following sections.
156 @item Overall Options
157 @xref{Overall Options,,Options Controlling the Kind of Output}.
158 @gccoptlist{-c -S -E -o @var{file} -pipe -pass-exit-codes @gol
159 -x @var{language} -v -### --help --target-help --version}
161 @item C Language Options
162 @xref{C Dialect Options,,Options Controlling C Dialect}.
163 @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol
164 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
165 -fhosted -ffreestanding -fms-extensions @gol
166 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
167 -fallow-single-precision -fcond-mismatch @gol
168 -fsigned-bitfields -fsigned-char @gol
169 -funsigned-bitfields -funsigned-char @gol
172 @item C++ Language Options
173 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
174 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
175 -fconserve-space -fno-const-strings @gol
176 -fno-elide-constructors @gol
177 -fno-enforce-eh-specs -fexternal-templates @gol
178 -falt-external-templates @gol
179 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
180 -fno-implicit-templates @gol
181 -fno-implicit-inline-templates @gol
182 -fno-implement-inlines -fms-extensions @gol
183 -fno-nonansi-builtins -fno-operator-names @gol
184 -fno-optional-diags -fpermissive @gol
185 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
186 -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
187 -fno-default-inline -Wabi -Wctor-dtor-privacy @gol
188 -Wnon-virtual-dtor -Wreorder @gol
189 -Weffc++ -Wno-deprecated @gol
190 -Wno-non-template-friend -Wold-style-cast @gol
191 -Woverloaded-virtual -Wno-pmf-conversions @gol
192 -Wsign-promo -Wsynth}
194 @item Objective-C Language Options
195 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
197 -fconstant-string-class=@var{class-name} @gol
198 -fgnu-runtime -fnext-runtime @gol
199 -fno-nil-receivers @gol
200 -fobjc-exceptions @gol
201 -freplace-objc-classes @gol
204 -Wno-protocol -Wselector -Wundeclared-selector}
206 @item Language Independent Options
207 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
208 @gccoptlist{-fmessage-length=@var{n} @gol
209 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
211 @item Warning Options
212 @xref{Warning Options,,Options to Request or Suppress Warnings}.
213 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
214 -w -Wextra -Wall -Waggregate-return @gol
215 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
216 -Wconversion -Wno-deprecated-declarations @gol
217 -Wdisabled-optimization -Wno-div-by-zero -Wendif-labels @gol
218 -Werror -Werror-implicit-function-declaration @gol
219 -Wfloat-equal -Wformat -Wformat=2 @gol
220 -Wno-format-extra-args -Wformat-nonliteral @gol
221 -Wformat-security -Wno-format-y2k @gol
222 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
223 -Wimport -Wno-import -Winit-self -Winline @gol
224 -Wno-invalid-offsetof -Winvalid-pch @gol
225 -Wlarger-than-@var{len} -Wlong-long @gol
226 -Wmain -Wmissing-braces @gol
227 -Wmissing-format-attribute -Wmissing-noreturn @gol
228 -Wno-multichar -Wnonnull -Wpacked -Wpadded @gol
229 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
230 -Wreturn-type -Wsequence-point -Wshadow @gol
231 -Wsign-compare -Wstrict-aliasing @gol
232 -Wswitch -Wswitch-default -Wswitch-enum @gol
233 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
234 -Wunknown-pragmas -Wunreachable-code @gol
235 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
236 -Wunused-value -Wunused-variable -Wwrite-strings}
238 @item C-only Warning Options
239 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
240 -Wmissing-prototypes -Wnested-externs -Wold-style-definition @gol
241 -Wstrict-prototypes -Wtraditional @gol
242 -Wdeclaration-after-statement}
244 @item Debugging Options
245 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
246 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
247 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
248 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
249 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
250 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
251 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
252 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
253 -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs @gol
254 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
255 -ftest-coverage -ftime-report @gol
256 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
257 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
258 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
259 -print-multi-directory -print-multi-lib @gol
260 -print-prog-name=@var{program} -print-search-dirs -Q @gol
263 @item Optimization Options
264 @xref{Optimize Options,,Options that Control Optimization}.
265 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
266 -falign-labels=@var{n} -falign-loops=@var{n} @gol
267 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
268 -fbranch-target-load-optimize2 -fcaller-saves -fcprop-registers @gol
269 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
270 -fdelayed-branch -fdelete-null-pointer-checks @gol
271 -fexpensive-optimizations -ffast-math -ffloat-store @gol
272 -fforce-addr -fforce-mem -ffunction-sections @gol
273 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -floop-optimize @gol
274 -fcrossjumping -fif-conversion -fif-conversion2 @gol
275 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
276 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
277 -fmove-all-movables -fnew-ra -fno-branch-count-reg @gol
278 -fno-default-inline -fno-defer-pop @gol
279 -fno-function-cse -fno-guess-branch-probability @gol
280 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
281 -funsafe-math-optimizations -ffinite-math-only @gol
282 -fno-trapping-math -fno-zero-initialized-in-bss @gol
283 -fomit-frame-pointer -foptimize-register-move @gol
284 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
285 -freduce-all-givs -fregmove -frename-registers @gol
286 -freorder-blocks -freorder-functions @gol
287 -frerun-cse-after-loop -frerun-loop-opt @gol
288 -frounding-math -fschedule-insns -fschedule-insns2 @gol
289 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
290 -fsched-spec-load-dangerous @gol
291 -fsched-stalled-insns=@var{n} -sched-stalled-insns-dep=@var{n} @gol
292 -fsched2-use-superblocks @gol
293 -fsched2-use-traces -fsignaling-nans @gol
294 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
295 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
296 -funroll-all-loops -funroll-loops -fpeel-loops @gol
297 -funswitch-loops -fold-unroll-loops -fold-unroll-all-loops @gol
298 --param @var{name}=@var{value}
299 -O -O0 -O1 -O2 -O3 -Os}
301 @item Preprocessor Options
302 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
303 @gccoptlist{-A@var{question}=@var{answer} @gol
304 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
305 -C -dD -dI -dM -dN @gol
306 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
307 -idirafter @var{dir} @gol
308 -include @var{file} -imacros @var{file} @gol
309 -iprefix @var{file} -iwithprefix @var{dir} @gol
310 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
311 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
312 -P -fworking-directory -remap @gol
313 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
314 -Xpreprocessor @var{option}}
316 @item Assembler Option
317 @xref{Assembler Options,,Passing Options to the Assembler}.
318 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
321 @xref{Link Options,,Options for Linking}.
322 @gccoptlist{@var{object-file-name} -l@var{library} @gol
323 -nostartfiles -nodefaultlibs -nostdlib -pie @gol
324 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
325 -Wl,@var{option} -Xlinker @var{option} @gol
328 @item Directory Options
329 @xref{Directory Options,,Options for Directory Search}.
330 @gccoptlist{-B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
333 @c I wrote this xref this way to avoid overfull hbox. -- rms
334 @xref{Target Options}.
335 @gccoptlist{-V @var{version} -b @var{machine}}
337 @item Machine Dependent Options
338 @xref{Submodel Options,,Hardware Models and Configurations}.
340 @emph{M680x0 Options}
341 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
342 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
343 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
344 -malign-int -mstrict-align}
346 @emph{M68hc1x Options}
347 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
348 -mauto-incdec -minmax -mlong-calls -mshort @gol
349 -msoft-reg-count=@var{count}}
352 @gccoptlist{-mg -mgnu -munix}
355 @gccoptlist{-mcpu=@var{cpu-type} @gol
356 -mtune=@var{cpu-type} @gol
357 -mcmodel=@var{code-model} @gol
359 -mapp-regs -mbroken-saverestore -mcypress @gol
360 -mfaster-structs -mflat @gol
361 -mfpu -mhard-float -mhard-quad-float @gol
362 -mimpure-text -mlittle-endian -mlive-g0 -mno-app-regs @gol
363 -mno-faster-structs -mno-flat -mno-fpu @gol
364 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
365 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
366 -msupersparc -munaligned-doubles -mv8}
369 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
370 -mapcs-26 -mapcs-32 @gol
371 -mapcs-stack-check -mno-apcs-stack-check @gol
372 -mapcs-float -mno-apcs-float @gol
373 -mapcs-reentrant -mno-apcs-reentrant @gol
374 -msched-prolog -mno-sched-prolog @gol
375 -mlittle-endian -mbig-endian -mwords-little-endian @gol
376 -malignment-traps -mno-alignment-traps @gol
377 -msoft-float -mhard-float -mfpe @gol
378 -mthumb-interwork -mno-thumb-interwork @gol
379 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
380 -mstructure-size-boundary=@var{n} @gol
381 -mabort-on-noreturn @gol
382 -mlong-calls -mno-long-calls @gol
383 -msingle-pic-base -mno-single-pic-base @gol
384 -mpic-register=@var{reg} @gol
385 -mnop-fun-dllimport @gol
386 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
387 -mpoke-function-name @gol
389 -mtpcs-frame -mtpcs-leaf-frame @gol
390 -mcaller-super-interworking -mcallee-super-interworking}
392 @emph{MN10200 Options}
395 @emph{MN10300 Options}
396 @gccoptlist{-mmult-bug -mno-mult-bug @gol
397 -mam33 -mno-am33 @gol
398 -mam33-2 -mno-am33-2 @gol
401 @emph{M32R/D Options}
402 @gccoptlist{-m32rx -m32r -mcode-model=@var{model-type} @gol
403 -msdata=@var{sdata-type} -G @var{num}}
406 @gccoptlist{-m88000 -m88100 -m88110 -mbig-pic @gol
407 -mcheck-zero-division -mhandle-large-shift @gol
408 -midentify-revision -mno-check-zero-division @gol
409 -mno-ocs-debug-info -mno-ocs-frame-position @gol
410 -mno-optimize-arg-area -mno-serialize-volatile @gol
411 -mno-underscores -mocs-debug-info @gol
412 -mocs-frame-position -moptimize-arg-area @gol
413 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
414 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
415 -mversion-03.00 -mwarn-passed-structs}
417 @emph{RS/6000 and PowerPC Options}
418 @gccoptlist{-mcpu=@var{cpu-type} @gol
419 -mtune=@var{cpu-type} @gol
420 -mpower -mno-power -mpower2 -mno-power2 @gol
421 -mpowerpc -mpowerpc64 -mno-powerpc @gol
422 -maltivec -mno-altivec @gol
423 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
424 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
425 -mnew-mnemonics -mold-mnemonics @gol
426 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
427 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
428 -malign-power -malign-natural @gol
429 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
430 -mstring -mno-string -mupdate -mno-update @gol
431 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
432 -mstrict-align -mno-strict-align -mrelocatable @gol
433 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
434 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
435 -mdynamic-no-pic @gol
436 -mprioritize-restricted-insns=@var{priority} @gol
437 -msched-costly-dep=@var{dependence_type} @gol
438 -minsert-sched-nops=@var{scheme} @gol
439 -mcall-sysv -mcall-netbsd @gol
440 -maix-struct-return -msvr4-struct-return @gol
441 -mabi=altivec -mabi=no-altivec @gol
442 -mabi=spe -mabi=no-spe @gol
443 -misel=yes -misel=no @gol
444 -mspe=yes -mspe=no @gol
445 -mfloat-gprs=yes -mfloat-gprs=no @gol
446 -mprototype -mno-prototype @gol
447 -msim -mmvme -mads -myellowknife -memb -msdata @gol
448 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
450 @emph{Darwin Options}
451 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
452 -arch_only -bind_at_load -bundle -bundle_loader @gol
453 -client_name -compatibility_version -current_version @gol
454 -dependency-file -dylib_file -dylinker_install_name @gol
455 -dynamic -dynamiclib -exported_symbols_list @gol
456 -filelist -flat_namespace -force_cpusubtype_ALL @gol
457 -force_flat_namespace -headerpad_max_install_names @gol
458 -image_base -init -install_name -keep_private_externs @gol
459 -multi_module -multiply_defined -multiply_defined_unused @gol
460 -noall_load -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
461 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
462 -private_bundle -read_only_relocs -sectalign @gol
463 -sectobjectsymbols -whyload -seg1addr @gol
464 -sectcreate -sectobjectsymbols -sectorder @gol
465 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
466 -segprot -segs_read_only_addr -segs_read_write_addr @gol
467 -single_module -static -sub_library -sub_umbrella @gol
468 -twolevel_namespace -umbrella -undefined @gol
469 -unexported_symbols_list -weak_reference_mismatches @gol
473 @gccoptlist{-mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
474 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
475 -mminimum-fp-blocks -mnohc-struct-return}
478 @gccoptlist{-mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
479 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
480 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
481 -mgas -mgp32 -mgp64 -mhard-float -mint64 -mips1 @gol
482 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
483 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
484 -mmips-as -mmips-tfile -mno-abicalls -mxgot @gol
485 -mno-embedded-data -mno-uninit-const-in-rodata @gol
486 -mno-embedded-pic -mno-long-calls @gol
487 -mno-memcpy -mno-mips-tfile -mno-rnames @gol
488 -mrnames -msoft-float @gol
489 -m4650 -msingle-float -mmad @gol
490 -EL -EB -G @var{num} -nocpp @gol
491 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi -mabi-fake-default @gol
492 -mfix7000 -mfix-sb1 -mno-fix-sb1 @gol
493 -mno-crt0 -mflush-func=@var{func} -mno-flush-func @gol
494 -mbranch-likely -mno-branch-likely}
496 @emph{i386 and x86-64 Options}
497 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
498 -mfpmath=@var{unit} @gol
499 -masm=@var{dialect} -mno-fancy-math-387 @gol
500 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
501 -mno-wide-multiply -mrtd -malign-double @gol
502 -mpreferred-stack-boundary=@var{num} @gol
503 -mmmx -msse -msse2 -mpni -m3dnow @gol
504 -mthreads -mno-align-stringops -minline-all-stringops @gol
505 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
506 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
507 -mno-red-zone -mno-tls-direct-seg-refs @gol
508 -mcmodel=@var{code-model} @gol
512 @gccoptlist{-march=@var{architecture-type} @gol
513 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
514 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
515 -mjump-in-delay -mlinker-opt -mlong-calls @gol
516 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
517 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
518 -mno-jump-in-delay -mno-long-load-store @gol
519 -mno-portable-runtime -mno-soft-float @gol
520 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
521 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
522 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
523 -nolibdld -static -threads}
525 @emph{Intel 960 Options}
526 @gccoptlist{-m@var{cpu-type} -masm-compat -mclean-linkage @gol
527 -mcode-align -mcomplex-addr -mleaf-procedures @gol
528 -mic-compat -mic2.0-compat -mic3.0-compat @gol
529 -mintel-asm -mno-clean-linkage -mno-code-align @gol
530 -mno-complex-addr -mno-leaf-procedures @gol
531 -mno-old-align -mno-strict-align -mno-tail-call @gol
532 -mnumerics -mold-align -msoft-float -mstrict-align @gol
535 @emph{DEC Alpha Options}
536 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
537 -mieee -mieee-with-inexact -mieee-conformant @gol
538 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
539 -mtrap-precision=@var{mode} -mbuild-constants @gol
540 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
541 -mbwx -mmax -mfix -mcix @gol
542 -mfloat-vax -mfloat-ieee @gol
543 -mexplicit-relocs -msmall-data -mlarge-data @gol
544 -msmall-text -mlarge-text @gol
545 -mmemory-latency=@var{time}}
547 @emph{DEC Alpha/VMS Options}
548 @gccoptlist{-mvms-return-codes}
550 @emph{H8/300 Options}
551 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
554 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
555 -m4-nofpu -m4-single-only -m4-single -m4 @gol
556 -m5-64media -m5-64media-nofpu @gol
557 -m5-32media -m5-32media-nofpu @gol
558 -m5-compact -m5-compact-nofpu @gol
559 -mb -ml -mdalign -mrelax @gol
560 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
561 -mieee -misize -mpadstruct -mspace @gol
562 -mprefergot -musermode}
564 @emph{System V Options}
565 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
568 @gccoptlist{-EB -EL @gol
569 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
570 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
572 @emph{TMS320C3x/C4x Options}
573 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
574 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
575 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
576 -mparallel-insns -mparallel-mpy -mpreserve-float}
579 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
580 -mprolog-function -mno-prolog-function -mspace @gol
581 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
582 -mapp-regs -mno-app-regs @gol
583 -mdisable-callt -mno-disable-callt @gol
589 @gccoptlist{-m32032 -m32332 -m32532 -m32081 -m32381 @gol
590 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
591 -mregparam -mnoregparam -msb -mnosb @gol
592 -mbitfield -mnobitfield -mhimem -mnohimem}
595 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
596 -mcall-prologues -mno-tablejump -mtiny-stack}
599 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
600 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
601 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
602 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
603 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
606 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
607 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
608 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
609 -mno-base-addresses -msingle-exit -mno-single-exit}
612 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
613 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
614 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
615 -minline-float-divide-max-throughput @gol
616 -minline-int-divide-min-latency @gol
617 -minline-int-divide-max-throughput -mno-dwarf2-asm @gol
618 -mfixed-range=@var{register-range}}
621 @gccoptlist{-mextmem -mextmemory -monchip -mno-asm-optimize @gol
622 -masm-optimize -mbranch-cost=@var{n} -mcond-exec=@var{n}}
624 @emph{S/390 and zSeries Options}
625 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
626 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
627 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
628 -m64 -m31 -mdebug -mno-debug -mesa -mzarch -mfused-madd -mno-fused-madd}
631 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
632 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
633 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
634 -mstack-align -mdata-align -mconst-align @gol
635 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
636 -melf -maout -melinux -mlinux -sim -sim2}
638 @emph{PDP-11 Options}
639 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
640 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
641 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
642 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
643 -mbranch-expensive -mbranch-cheap @gol
644 -msplit -mno-split -munix-asm -mdec-asm}
646 @emph{Xstormy16 Options}
649 @emph{Xtensa Options}
650 @gccoptlist{-mconst16 -mno-const16 @gol
651 -mfused-madd -mno-fused-madd @gol
652 -mtext-section-literals -mno-text-section-literals @gol
653 -mtarget-align -mno-target-align @gol
654 -mlongcalls -mno-longcalls}
657 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
658 -mhard-float -msoft-float @gol
659 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
660 -mdouble -mno-double @gol
661 -mmedia -mno-media -mmuladd -mno-muladd @gol
662 -mlibrary-pic -macc-4 -macc-8 @gol
663 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
664 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
665 -mvliw-branch -mno-vliw-branch @gol
666 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
667 -mno-nested-cond-exec -mtomcat-stats @gol
670 @item Code Generation Options
671 @xref{Code Gen Options,,Options for Code Generation Conventions}.
672 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
673 -ffixed-@var{reg} -fexceptions @gol
674 -fnon-call-exceptions -funwind-tables @gol
675 -fasynchronous-unwind-tables @gol
676 -finhibit-size-directive -finstrument-functions @gol
677 -fno-common -fno-ident -fno-gnu-linker @gol
678 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
679 -freg-struct-return -fshared-data -fshort-enums @gol
680 -fshort-double -fshort-wchar @gol
681 -fverbose-asm -fpack-struct -fstack-check @gol
682 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
683 -fargument-alias -fargument-noalias @gol
684 -fargument-noalias-global -fleading-underscore @gol
685 -ftls-model=@var{model} @gol
686 -ftrapv -fwrapv -fbounds-check}
690 * Overall Options:: Controlling the kind of output:
691 an executable, object files, assembler files,
692 or preprocessed source.
693 * C Dialect Options:: Controlling the variant of C language compiled.
694 * C++ Dialect Options:: Variations on C++.
695 * Objective-C Dialect Options:: Variations on Objective-C.
696 * Language Independent Options:: Controlling how diagnostics should be
698 * Warning Options:: How picky should the compiler be?
699 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
700 * Optimize Options:: How much optimization?
701 * Preprocessor Options:: Controlling header files and macro definitions.
702 Also, getting dependency information for Make.
703 * Assembler Options:: Passing options to the assembler.
704 * Link Options:: Specifying libraries and so on.
705 * Directory Options:: Where to find header files and libraries.
706 Where to find the compiler executable files.
707 * Spec Files:: How to pass switches to sub-processes.
708 * Target Options:: Running a cross-compiler, or an old version of GCC.
711 @node Overall Options
712 @section Options Controlling the Kind of Output
714 Compilation can involve up to four stages: preprocessing, compilation
715 proper, assembly and linking, always in that order. GCC is capable of
716 preprocessing and compiling several files either into several
717 assembler input files, or into one assembler input file; then each
718 assembler input file produces an object file, and linking combines all
719 the object files (those newly compiled, and those specified as input)
720 into an executable file.
722 @cindex file name suffix
723 For any given input file, the file name suffix determines what kind of
728 C source code which must be preprocessed.
731 C source code which should not be preprocessed.
734 C++ source code which should not be preprocessed.
737 Objective-C source code. Note that you must link with the library
738 @file{libobjc.a} to make an Objective-C program work.
741 Objective-C source code which should not be preprocessed.
744 C or C++ header file to be turned into a precompiled header.
748 @itemx @var{file}.cxx
749 @itemx @var{file}.cpp
750 @itemx @var{file}.CPP
751 @itemx @var{file}.c++
753 C++ source code which must be preprocessed. Note that in @samp{.cxx},
754 the last two letters must both be literally @samp{x}. Likewise,
755 @samp{.C} refers to a literal capital C@.
759 C++ header file to be turned into a precompiled header.
762 @itemx @var{file}.for
763 @itemx @var{file}.FOR
764 Fortran source code which should not be preprocessed.
767 @itemx @var{file}.fpp
768 @itemx @var{file}.FPP
769 Fortran source code which must be preprocessed (with the traditional
773 Fortran source code which must be preprocessed with a RATFOR
774 preprocessor (not included with GCC)@.
776 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
777 Using and Porting GNU Fortran}, for more details of the handling of
780 @c FIXME: Descriptions of Java file types.
787 Ada source code file which contains a library unit declaration (a
788 declaration of a package, subprogram, or generic, or a generic
789 instantiation), or a library unit renaming declaration (a package,
790 generic, or subprogram renaming declaration). Such files are also
793 @itemx @var{file}.adb
794 Ada source code file containing a library unit body (a subprogram or
795 package body). Such files are also called @dfn{bodies}.
797 @c GCC also knows about some suffixes for languages not yet included:
806 Assembler code which must be preprocessed.
809 An object file to be fed straight into linking.
810 Any file name with no recognized suffix is treated this way.
814 You can specify the input language explicitly with the @option{-x} option:
817 @item -x @var{language}
818 Specify explicitly the @var{language} for the following input files
819 (rather than letting the compiler choose a default based on the file
820 name suffix). This option applies to all following input files until
821 the next @option{-x} option. Possible values for @var{language} are:
823 c c-header cpp-output
824 c++ c++-header c++-cpp-output
825 objective-c objective-c-header objc-cpp-output
826 assembler assembler-with-cpp
828 f77 f77-cpp-input ratfor
834 Turn off any specification of a language, so that subsequent files are
835 handled according to their file name suffixes (as they are if @option{-x}
836 has not been used at all).
838 @item -pass-exit-codes
839 @opindex pass-exit-codes
840 Normally the @command{gcc} program will exit with the code of 1 if any
841 phase of the compiler returns a non-success return code. If you specify
842 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
843 numerically highest error produced by any phase that returned an error
847 If you only want some of the stages of compilation, you can use
848 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
849 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
850 @command{gcc} is to stop. Note that some combinations (for example,
851 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
856 Compile or assemble the source files, but do not link. The linking
857 stage simply is not done. The ultimate output is in the form of an
858 object file for each source file.
860 By default, the object file name for a source file is made by replacing
861 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
863 Unrecognized input files, not requiring compilation or assembly, are
868 Stop after the stage of compilation proper; do not assemble. The output
869 is in the form of an assembler code file for each non-assembler input
872 By default, the assembler file name for a source file is made by
873 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
875 Input files that don't require compilation are ignored.
879 Stop after the preprocessing stage; do not run the compiler proper. The
880 output is in the form of preprocessed source code, which is sent to the
883 Input files which don't require preprocessing are ignored.
885 @cindex output file option
888 Place output in file @var{file}. This applies regardless to whatever
889 sort of output is being produced, whether it be an executable file,
890 an object file, an assembler file or preprocessed C code.
892 If you specify @option{-o} when compiling more than one input file, or
893 you are producing an executable file as output, all the source files
894 on the command line will be compiled at once.
896 If @option{-o} is not specified, the default is to put an executable file
897 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
898 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
899 all preprocessed C source on standard output.
903 Print (on standard error output) the commands executed to run the stages
904 of compilation. Also print the version number of the compiler driver
905 program and of the preprocessor and the compiler proper.
909 Like @option{-v} except the commands are not executed and all command
910 arguments are quoted. This is useful for shell scripts to capture the
911 driver-generated command lines.
915 Use pipes rather than temporary files for communication between the
916 various stages of compilation. This fails to work on some systems where
917 the assembler is unable to read from a pipe; but the GNU assembler has
922 Print (on the standard output) a description of the command line options
923 understood by @command{gcc}. If the @option{-v} option is also specified
924 then @option{--help} will also be passed on to the various processes
925 invoked by @command{gcc}, so that they can display the command line options
926 they accept. If the @option{-Wextra} option is also specified then command
927 line options which have no documentation associated with them will also
932 Print (on the standard output) a description of target specific command
933 line options for each tool.
937 Display the version number and copyrights of the invoked GCC.
941 @section Compiling C++ Programs
943 @cindex suffixes for C++ source
944 @cindex C++ source file suffixes
945 C++ source files conventionally use one of the suffixes @samp{.C},
946 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
947 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
948 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
949 files with these names and compiles them as C++ programs even if you
950 call the compiler the same way as for compiling C programs (usually
951 with the name @command{gcc}).
955 However, C++ programs often require class libraries as well as a
956 compiler that understands the C++ language---and under some
957 circumstances, you might want to compile programs or header files from
958 standard input, or otherwise without a suffix that flags them as C++
959 programs. You might also like to precompile a C header file with a
960 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
961 program that calls GCC with the default language set to C++, and
962 automatically specifies linking against the C++ library. On many
963 systems, @command{g++} is also installed with the name @command{c++}.
965 @cindex invoking @command{g++}
966 When you compile C++ programs, you may specify many of the same
967 command-line options that you use for compiling programs in any
968 language; or command-line options meaningful for C and related
969 languages; or options that are meaningful only for C++ programs.
970 @xref{C Dialect Options,,Options Controlling C Dialect}, for
971 explanations of options for languages related to C@.
972 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
973 explanations of options that are meaningful only for C++ programs.
975 @node C Dialect Options
976 @section Options Controlling C Dialect
977 @cindex dialect options
978 @cindex language dialect options
979 @cindex options, dialect
981 The following options control the dialect of C (or languages derived
982 from C, such as C++ and Objective-C) that the compiler accepts:
989 In C mode, support all ISO C90 programs. In C++ mode,
990 remove GNU extensions that conflict with ISO C++.
992 This turns off certain features of GCC that are incompatible with ISO
993 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
994 such as the @code{asm} and @code{typeof} keywords, and
995 predefined macros such as @code{unix} and @code{vax} that identify the
996 type of system you are using. It also enables the undesirable and
997 rarely used ISO trigraph feature. For the C compiler,
998 it disables recognition of C++ style @samp{//} comments as well as
999 the @code{inline} keyword.
1001 The alternate keywords @code{__asm__}, @code{__extension__},
1002 @code{__inline__} and @code{__typeof__} continue to work despite
1003 @option{-ansi}. You would not want to use them in an ISO C program, of
1004 course, but it is useful to put them in header files that might be included
1005 in compilations done with @option{-ansi}. Alternate predefined macros
1006 such as @code{__unix__} and @code{__vax__} are also available, with or
1007 without @option{-ansi}.
1009 The @option{-ansi} option does not cause non-ISO programs to be
1010 rejected gratuitously. For that, @option{-pedantic} is required in
1011 addition to @option{-ansi}. @xref{Warning Options}.
1013 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1014 option is used. Some header files may notice this macro and refrain
1015 from declaring certain functions or defining certain macros that the
1016 ISO standard doesn't call for; this is to avoid interfering with any
1017 programs that might use these names for other things.
1019 Functions which would normally be built in but do not have semantics
1020 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1021 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1022 built-in functions provided by GCC}, for details of the functions
1027 Determine the language standard. This option is currently only
1028 supported when compiling C or C++. A value for this option must be
1029 provided; possible values are
1034 ISO C90 (same as @option{-ansi}).
1036 @item iso9899:199409
1037 ISO C90 as modified in amendment 1.
1043 ISO C99. Note that this standard is not yet fully supported; see
1044 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1045 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1048 Default, ISO C90 plus GNU extensions (including some C99 features).
1052 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1053 this will become the default. The name @samp{gnu9x} is deprecated.
1056 The 1998 ISO C++ standard plus amendments.
1059 The same as @option{-std=c++98} plus GNU extensions. This is the
1060 default for C++ code.
1063 Even when this option is not specified, you can still use some of the
1064 features of newer standards in so far as they do not conflict with
1065 previous C standards. For example, you may use @code{__restrict__} even
1066 when @option{-std=c99} is not specified.
1068 The @option{-std} options specifying some version of ISO C have the same
1069 effects as @option{-ansi}, except that features that were not in ISO C90
1070 but are in the specified version (for example, @samp{//} comments and
1071 the @code{inline} keyword in ISO C99) are not disabled.
1073 @xref{Standards,,Language Standards Supported by GCC}, for details of
1074 these standard versions.
1076 @item -aux-info @var{filename}
1078 Output to the given filename prototyped declarations for all functions
1079 declared and/or defined in a translation unit, including those in header
1080 files. This option is silently ignored in any language other than C@.
1082 Besides declarations, the file indicates, in comments, the origin of
1083 each declaration (source file and line), whether the declaration was
1084 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1085 @samp{O} for old, respectively, in the first character after the line
1086 number and the colon), and whether it came from a declaration or a
1087 definition (@samp{C} or @samp{F}, respectively, in the following
1088 character). In the case of function definitions, a K&R-style list of
1089 arguments followed by their declarations is also provided, inside
1090 comments, after the declaration.
1094 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1095 keyword, so that code can use these words as identifiers. You can use
1096 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1097 instead. @option{-ansi} implies @option{-fno-asm}.
1099 In C++, this switch only affects the @code{typeof} keyword, since
1100 @code{asm} and @code{inline} are standard keywords. You may want to
1101 use the @option{-fno-gnu-keywords} flag instead, which has the same
1102 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1103 switch only affects the @code{asm} and @code{typeof} keywords, since
1104 @code{inline} is a standard keyword in ISO C99.
1107 @itemx -fno-builtin-@var{function}
1108 @opindex fno-builtin
1109 @cindex built-in functions
1110 Don't recognize built-in functions that do not begin with
1111 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1112 functions provided by GCC}, for details of the functions affected,
1113 including those which are not built-in functions when @option{-ansi} or
1114 @option{-std} options for strict ISO C conformance are used because they
1115 do not have an ISO standard meaning.
1117 GCC normally generates special code to handle certain built-in functions
1118 more efficiently; for instance, calls to @code{alloca} may become single
1119 instructions that adjust the stack directly, and calls to @code{memcpy}
1120 may become inline copy loops. The resulting code is often both smaller
1121 and faster, but since the function calls no longer appear as such, you
1122 cannot set a breakpoint on those calls, nor can you change the behavior
1123 of the functions by linking with a different library.
1125 With the @option{-fno-builtin-@var{function}} option
1126 only the built-in function @var{function} is
1127 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1128 function is named this is not built-in in this version of GCC, this
1129 option is ignored. There is no corresponding
1130 @option{-fbuiltin-@var{function}} option; if you wish to enable
1131 built-in functions selectively when using @option{-fno-builtin} or
1132 @option{-ffreestanding}, you may define macros such as:
1135 #define abs(n) __builtin_abs ((n))
1136 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1141 @cindex hosted environment
1143 Assert that compilation takes place in a hosted environment. This implies
1144 @option{-fbuiltin}. A hosted environment is one in which the
1145 entire standard library is available, and in which @code{main} has a return
1146 type of @code{int}. Examples are nearly everything except a kernel.
1147 This is equivalent to @option{-fno-freestanding}.
1149 @item -ffreestanding
1150 @opindex ffreestanding
1151 @cindex hosted environment
1153 Assert that compilation takes place in a freestanding environment. This
1154 implies @option{-fno-builtin}. A freestanding environment
1155 is one in which the standard library may not exist, and program startup may
1156 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1157 This is equivalent to @option{-fno-hosted}.
1159 @xref{Standards,,Language Standards Supported by GCC}, for details of
1160 freestanding and hosted environments.
1162 @item -fms-extensions
1163 @opindex fms-extensions
1164 Accept some non-standard constructs used in Microsoft header files.
1168 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1169 options for strict ISO C conformance) implies @option{-trigraphs}.
1171 @item -no-integrated-cpp
1172 @opindex no-integrated-cpp
1173 Performs a compilation in two passes: preprocessing and compiling. This
1174 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1175 @option{-B} option. The user supplied compilation step can then add in
1176 an additional preprocessing step after normal preprocessing but before
1177 compiling. The default is to use the integrated cpp (internal cpp)
1179 The semantics of this option will change if "cc1", "cc1plus", and
1180 "cc1obj" are merged.
1182 @cindex traditional C language
1183 @cindex C language, traditional
1185 @itemx -traditional-cpp
1186 @opindex traditional-cpp
1187 @opindex traditional
1188 Formerly, these options caused GCC to attempt to emulate a pre-standard
1189 C compiler. They are now only supported with the @option{-E} switch.
1190 The preprocessor continues to support a pre-standard mode. See the GNU
1191 CPP manual for details.
1193 @item -fcond-mismatch
1194 @opindex fcond-mismatch
1195 Allow conditional expressions with mismatched types in the second and
1196 third arguments. The value of such an expression is void. This option
1197 is not supported for C++.
1199 @item -funsigned-char
1200 @opindex funsigned-char
1201 Let the type @code{char} be unsigned, like @code{unsigned char}.
1203 Each kind of machine has a default for what @code{char} should
1204 be. It is either like @code{unsigned char} by default or like
1205 @code{signed char} by default.
1207 Ideally, a portable program should always use @code{signed char} or
1208 @code{unsigned char} when it depends on the signedness of an object.
1209 But many programs have been written to use plain @code{char} and
1210 expect it to be signed, or expect it to be unsigned, depending on the
1211 machines they were written for. This option, and its inverse, let you
1212 make such a program work with the opposite default.
1214 The type @code{char} is always a distinct type from each of
1215 @code{signed char} or @code{unsigned char}, even though its behavior
1216 is always just like one of those two.
1219 @opindex fsigned-char
1220 Let the type @code{char} be signed, like @code{signed char}.
1222 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1223 the negative form of @option{-funsigned-char}. Likewise, the option
1224 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1226 @item -fsigned-bitfields
1227 @itemx -funsigned-bitfields
1228 @itemx -fno-signed-bitfields
1229 @itemx -fno-unsigned-bitfields
1230 @opindex fsigned-bitfields
1231 @opindex funsigned-bitfields
1232 @opindex fno-signed-bitfields
1233 @opindex fno-unsigned-bitfields
1234 These options control whether a bit-field is signed or unsigned, when the
1235 declaration does not use either @code{signed} or @code{unsigned}. By
1236 default, such a bit-field is signed, because this is consistent: the
1237 basic integer types such as @code{int} are signed types.
1239 @item -fwritable-strings
1240 @opindex fwritable-strings
1241 Store string constants in the writable data segment and don't uniquize
1242 them. This is for compatibility with old programs which assume they can
1243 write into string constants.
1245 Writing into string constants is a very bad idea; ``constants'' should
1249 @node C++ Dialect Options
1250 @section Options Controlling C++ Dialect
1252 @cindex compiler options, C++
1253 @cindex C++ options, command line
1254 @cindex options, C++
1255 This section describes the command-line options that are only meaningful
1256 for C++ programs; but you can also use most of the GNU compiler options
1257 regardless of what language your program is in. For example, you
1258 might compile a file @code{firstClass.C} like this:
1261 g++ -g -frepo -O -c firstClass.C
1265 In this example, only @option{-frepo} is an option meant
1266 only for C++ programs; you can use the other options with any
1267 language supported by GCC@.
1269 Here is a list of options that are @emph{only} for compiling C++ programs:
1273 @item -fabi-version=@var{n}
1274 @opindex fabi-version
1275 Use version @var{n} of the C++ ABI. Version 1 is the version of the C++
1276 ABI that first appeared in G++ 3.2. Version 0 will always be the
1277 version that conforms most closely to the C++ ABI specification.
1278 Therefore, the ABI obtained using version 0 will change as ABI bugs are
1281 The default is version 1.
1283 @item -fno-access-control
1284 @opindex fno-access-control
1285 Turn off all access checking. This switch is mainly useful for working
1286 around bugs in the access control code.
1290 Check that the pointer returned by @code{operator new} is non-null
1291 before attempting to modify the storage allocated. This check is
1292 normally unnecessary because the C++ standard specifies that
1293 @code{operator new} will only return @code{0} if it is declared
1294 @samp{throw()}, in which case the compiler will always check the
1295 return value even without this option. In all other cases, when
1296 @code{operator new} has a non-empty exception specification, memory
1297 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1298 @samp{new (nothrow)}.
1300 @item -fconserve-space
1301 @opindex fconserve-space
1302 Put uninitialized or runtime-initialized global variables into the
1303 common segment, as C does. This saves space in the executable at the
1304 cost of not diagnosing duplicate definitions. If you compile with this
1305 flag and your program mysteriously crashes after @code{main()} has
1306 completed, you may have an object that is being destroyed twice because
1307 two definitions were merged.
1309 This option is no longer useful on most targets, now that support has
1310 been added for putting variables into BSS without making them common.
1312 @item -fno-const-strings
1313 @opindex fno-const-strings
1314 Give string constants type @code{char *} instead of type @code{const
1315 char *}. By default, G++ uses type @code{const char *} as required by
1316 the standard. Even if you use @option{-fno-const-strings}, you cannot
1317 actually modify the value of a string constant, unless you also use
1318 @option{-fwritable-strings}.
1320 This option might be removed in a future release of G++. For maximum
1321 portability, you should structure your code so that it works with
1322 string constants that have type @code{const char *}.
1324 @item -fno-elide-constructors
1325 @opindex fno-elide-constructors
1326 The C++ standard allows an implementation to omit creating a temporary
1327 which is only used to initialize another object of the same type.
1328 Specifying this option disables that optimization, and forces G++ to
1329 call the copy constructor in all cases.
1331 @item -fno-enforce-eh-specs
1332 @opindex fno-enforce-eh-specs
1333 Don't check for violation of exception specifications at runtime. This
1334 option violates the C++ standard, but may be useful for reducing code
1335 size in production builds, much like defining @samp{NDEBUG}. The compiler
1336 will still optimize based on the exception specifications.
1338 @item -fexternal-templates
1339 @opindex fexternal-templates
1341 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1342 template instantiation; template instances are emitted or not according
1343 to the location of the template definition. @xref{Template
1344 Instantiation}, for more information.
1346 This option is deprecated.
1348 @item -falt-external-templates
1349 @opindex falt-external-templates
1350 Similar to @option{-fexternal-templates}, but template instances are
1351 emitted or not according to the place where they are first instantiated.
1352 @xref{Template Instantiation}, for more information.
1354 This option is deprecated.
1357 @itemx -fno-for-scope
1359 @opindex fno-for-scope
1360 If @option{-ffor-scope} is specified, the scope of variables declared in
1361 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1362 as specified by the C++ standard.
1363 If @option{-fno-for-scope} is specified, the scope of variables declared in
1364 a @i{for-init-statement} extends to the end of the enclosing scope,
1365 as was the case in old versions of G++, and other (traditional)
1366 implementations of C++.
1368 The default if neither flag is given to follow the standard,
1369 but to allow and give a warning for old-style code that would
1370 otherwise be invalid, or have different behavior.
1372 @item -fno-gnu-keywords
1373 @opindex fno-gnu-keywords
1374 Do not recognize @code{typeof} as a keyword, so that code can use this
1375 word as an identifier. You can use the keyword @code{__typeof__} instead.
1376 @option{-ansi} implies @option{-fno-gnu-keywords}.
1378 @item -fno-implicit-templates
1379 @opindex fno-implicit-templates
1380 Never emit code for non-inline templates which are instantiated
1381 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1382 @xref{Template Instantiation}, for more information.
1384 @item -fno-implicit-inline-templates
1385 @opindex fno-implicit-inline-templates
1386 Don't emit code for implicit instantiations of inline templates, either.
1387 The default is to handle inlines differently so that compiles with and
1388 without optimization will need the same set of explicit instantiations.
1390 @item -fno-implement-inlines
1391 @opindex fno-implement-inlines
1392 To save space, do not emit out-of-line copies of inline functions
1393 controlled by @samp{#pragma implementation}. This will cause linker
1394 errors if these functions are not inlined everywhere they are called.
1396 @item -fms-extensions
1397 @opindex fms-extensions
1398 Disable pedantic warnings about constructs used in MFC, such as implicit
1399 int and getting a pointer to member function via non-standard syntax.
1401 @item -fno-nonansi-builtins
1402 @opindex fno-nonansi-builtins
1403 Disable built-in declarations of functions that are not mandated by
1404 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1405 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1407 @item -fno-operator-names
1408 @opindex fno-operator-names
1409 Do not treat the operator name keywords @code{and}, @code{bitand},
1410 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1411 synonyms as keywords.
1413 @item -fno-optional-diags
1414 @opindex fno-optional-diags
1415 Disable diagnostics that the standard says a compiler does not need to
1416 issue. Currently, the only such diagnostic issued by G++ is the one for
1417 a name having multiple meanings within a class.
1420 @opindex fpermissive
1421 Downgrade some diagnostics about nonconformant code from errors to
1422 warnings. Thus, using @option{-fpermissive} will allow some
1423 nonconforming code to compile.
1427 Enable automatic template instantiation at link time. This option also
1428 implies @option{-fno-implicit-templates}. @xref{Template
1429 Instantiation}, for more information.
1433 Disable generation of information about every class with virtual
1434 functions for use by the C++ runtime type identification features
1435 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1436 of the language, you can save some space by using this flag. Note that
1437 exception handling uses the same information, but it will generate it as
1442 Emit statistics about front-end processing at the end of the compilation.
1443 This information is generally only useful to the G++ development team.
1445 @item -ftemplate-depth-@var{n}
1446 @opindex ftemplate-depth
1447 Set the maximum instantiation depth for template classes to @var{n}.
1448 A limit on the template instantiation depth is needed to detect
1449 endless recursions during template class instantiation. ANSI/ISO C++
1450 conforming programs must not rely on a maximum depth greater than 17.
1452 @item -fuse-cxa-atexit
1453 @opindex fuse-cxa-atexit
1454 Register destructors for objects with static storage duration with the
1455 @code{__cxa_atexit} function rather than the @code{atexit} function.
1456 This option is required for fully standards-compliant handling of static
1457 destructors, but will only work if your C library supports
1458 @code{__cxa_atexit}.
1462 Do not use weak symbol support, even if it is provided by the linker.
1463 By default, G++ will use weak symbols if they are available. This
1464 option exists only for testing, and should not be used by end-users;
1465 it will result in inferior code and has no benefits. This option may
1466 be removed in a future release of G++.
1470 Do not search for header files in the standard directories specific to
1471 C++, but do still search the other standard directories. (This option
1472 is used when building the C++ library.)
1475 In addition, these optimization, warning, and code generation options
1476 have meanings only for C++ programs:
1479 @item -fno-default-inline
1480 @opindex fno-default-inline
1481 Do not assume @samp{inline} for functions defined inside a class scope.
1482 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1483 functions will have linkage like inline functions; they just won't be
1486 @item -Wabi @r{(C++ only)}
1488 Warn when G++ generates code that is probably not compatible with the
1489 vendor-neutral C++ ABI. Although an effort has been made to warn about
1490 all such cases, there are probably some cases that are not warned about,
1491 even though G++ is generating incompatible code. There may also be
1492 cases where warnings are emitted even though the code that is generated
1495 You should rewrite your code to avoid these warnings if you are
1496 concerned about the fact that code generated by G++ may not be binary
1497 compatible with code generated by other compilers.
1499 The known incompatibilities at this point include:
1504 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1505 pack data into the same byte as a base class. For example:
1508 struct A @{ virtual void f(); int f1 : 1; @};
1509 struct B : public A @{ int f2 : 1; @};
1513 In this case, G++ will place @code{B::f2} into the same byte
1514 as@code{A::f1}; other compilers will not. You can avoid this problem
1515 by explicitly padding @code{A} so that its size is a multiple of the
1516 byte size on your platform; that will cause G++ and other compilers to
1517 layout @code{B} identically.
1520 Incorrect handling of tail-padding for virtual bases. G++ does not use
1521 tail padding when laying out virtual bases. For example:
1524 struct A @{ virtual void f(); char c1; @};
1525 struct B @{ B(); char c2; @};
1526 struct C : public A, public virtual B @{@};
1530 In this case, G++ will not place @code{B} into the tail-padding for
1531 @code{A}; other compilers will. You can avoid this problem by
1532 explicitly padding @code{A} so that its size is a multiple of its
1533 alignment (ignoring virtual base classes); that will cause G++ and other
1534 compilers to layout @code{C} identically.
1537 Incorrect handling of bit-fields with declared widths greater than that
1538 of their underlying types, when the bit-fields appear in a union. For
1542 union U @{ int i : 4096; @};
1546 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1547 union too small by the number of bits in an @code{int}.
1550 Empty classes can be placed at incorrect offsets. For example:
1560 struct C : public B, public A @{@};
1564 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1565 it should be placed at offset zero. G++ mistakenly believes that the
1566 @code{A} data member of @code{B} is already at offset zero.
1569 Names of template functions whose types involve @code{typename} or
1570 template template parameters can be mangled incorrectly.
1573 template <typename Q>
1574 void f(typename Q::X) @{@}
1576 template <template <typename> class Q>
1577 void f(typename Q<int>::X) @{@}
1581 Instantiations of these templates may be mangled incorrectly.
1585 @item -Wctor-dtor-privacy @r{(C++ only)}
1586 @opindex Wctor-dtor-privacy
1587 Warn when a class seems unusable because all the constructors or
1588 destructors in that class are private, and it has neither friends nor
1589 public static member functions.
1591 @item -Wnon-virtual-dtor @r{(C++ only)}
1592 @opindex Wnon-virtual-dtor
1593 Warn when a class appears to be polymorphic, thereby requiring a virtual
1594 destructor, yet it declares a non-virtual one.
1595 This warning is enabled by @option{-Wall}.
1597 @item -Wreorder @r{(C++ only)}
1599 @cindex reordering, warning
1600 @cindex warning for reordering of member initializers
1601 Warn when the order of member initializers given in the code does not
1602 match the order in which they must be executed. For instance:
1608 A(): j (0), i (1) @{ @}
1612 The compiler will rearrange the member initializers for @samp{i}
1613 and @samp{j} to match the declaration order of the members, emitting
1614 a warning to that effect. This warning is enabled by @option{-Wall}.
1617 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1620 @item -Weffc++ @r{(C++ only)}
1622 Warn about violations of the following style guidelines from Scott Meyers'
1623 @cite{Effective C++} book:
1627 Item 11: Define a copy constructor and an assignment operator for classes
1628 with dynamically allocated memory.
1631 Item 12: Prefer initialization to assignment in constructors.
1634 Item 14: Make destructors virtual in base classes.
1637 Item 15: Have @code{operator=} return a reference to @code{*this}.
1640 Item 23: Don't try to return a reference when you must return an object.
1644 Also warn about violations of the following style guidelines from
1645 Scott Meyers' @cite{More Effective C++} book:
1649 Item 6: Distinguish between prefix and postfix forms of increment and
1650 decrement operators.
1653 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1657 When selecting this option, be aware that the standard library
1658 headers do not obey all of these guidelines; use @samp{grep -v}
1659 to filter out those warnings.
1661 @item -Wno-deprecated @r{(C++ only)}
1662 @opindex Wno-deprecated
1663 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1665 @item -Wno-non-template-friend @r{(C++ only)}
1666 @opindex Wno-non-template-friend
1667 Disable warnings when non-templatized friend functions are declared
1668 within a template. Since the advent of explicit template specification
1669 support in G++, if the name of the friend is an unqualified-id (i.e.,
1670 @samp{friend foo(int)}), the C++ language specification demands that the
1671 friend declare or define an ordinary, nontemplate function. (Section
1672 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1673 could be interpreted as a particular specialization of a templatized
1674 function. Because this non-conforming behavior is no longer the default
1675 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1676 check existing code for potential trouble spots and is on by default.
1677 This new compiler behavior can be turned off with
1678 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1679 but disables the helpful warning.
1681 @item -Wold-style-cast @r{(C++ only)}
1682 @opindex Wold-style-cast
1683 Warn if an old-style (C-style) cast to a non-void type is used within
1684 a C++ program. The new-style casts (@samp{static_cast},
1685 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1686 unintended effects and much easier to search for.
1688 @item -Woverloaded-virtual @r{(C++ only)}
1689 @opindex Woverloaded-virtual
1690 @cindex overloaded virtual fn, warning
1691 @cindex warning for overloaded virtual fn
1692 Warn when a function declaration hides virtual functions from a
1693 base class. For example, in:
1700 struct B: public A @{
1705 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1713 will fail to compile.
1715 @item -Wno-pmf-conversions @r{(C++ only)}
1716 @opindex Wno-pmf-conversions
1717 Disable the diagnostic for converting a bound pointer to member function
1720 @item -Wsign-promo @r{(C++ only)}
1721 @opindex Wsign-promo
1722 Warn when overload resolution chooses a promotion from unsigned or
1723 enumeral type to a signed type, over a conversion to an unsigned type of
1724 the same size. Previous versions of G++ would try to preserve
1725 unsignedness, but the standard mandates the current behavior.
1727 @item -Wsynth @r{(C++ only)}
1729 @cindex warning for synthesized methods
1730 @cindex synthesized methods, warning
1731 Warn when G++'s synthesis behavior does not match that of cfront. For
1737 A& operator = (int);
1747 In this example, G++ will synthesize a default @samp{A& operator =
1748 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1751 @node Objective-C Dialect Options
1752 @section Options Controlling Objective-C Dialect
1754 @cindex compiler options, Objective-C
1755 @cindex Objective-C options, command line
1756 @cindex options, Objective-C
1757 (NOTE: This manual does not describe the Objective-C language itself. See
1758 @w{@uref{http://gcc.gnu.org/readings.html}} for references.)
1760 This section describes the command-line options that are only meaningful
1761 for Objective-C programs, but you can also use most of the GNU compiler
1762 options regardless of what language your program is in. For example,
1763 you might compile a file @code{some_class.m} like this:
1766 gcc -g -fgnu-runtime -O -c some_class.m
1770 In this example, @option{-fgnu-runtime} is an option meant only for
1771 Objective-C programs; you can use the other options with any language
1774 Here is a list of options that are @emph{only} for compiling Objective-C
1778 @item -fconstant-string-class=@var{class-name}
1779 @opindex fconstant-string-class
1780 Use @var{class-name} as the name of the class to instantiate for each
1781 literal string specified with the syntax @code{@@"@dots{}"}. The default
1782 class name is @code{NXConstantString} if the GNU runtime is being used, and
1783 @code{NSConstantString} if the NeXT runtime is being used (see below). The
1784 @option{-fconstant-cfstrings} option, if also present, will override the
1785 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
1786 to be laid out as constant CoreFoundation strings.
1789 @opindex fgnu-runtime
1790 Generate object code compatible with the standard GNU Objective-C
1791 runtime. This is the default for most types of systems.
1793 @item -fnext-runtime
1794 @opindex fnext-runtime
1795 Generate output compatible with the NeXT runtime. This is the default
1796 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1797 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1800 @item -fno-nil-receivers
1801 @opindex -fno-nil-receivers
1802 Assume that all Objective-C message dispatches (e.g.,
1803 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
1804 is not @code{nil}. This allows for more efficient entry points in the runtime to be
1805 used. Currently, this option is only available in conjunction with
1806 the NeXT runtime on Mac OS X 10.3 and later.
1808 @item -fobjc-exceptions
1809 @opindex -fobjc-exceptions
1810 Enable syntactic support for structured exception handling in Objective-C,
1811 similar to what is offered by C++ and Java. Currently, this option is only
1812 available in conjunction with the NeXT runtime on Mac OS X 10.3 and later.
1820 @@catch (AnObjCClass *exc) @{
1827 @@catch (AnotherClass *exc) @{
1830 @@catch (id allOthers) @{
1840 The @code{@@throw} statement may appear anywhere in an Objective-C or
1841 Objective-C++ program; when used inside of a @code{@@catch} block, the
1842 @code{@@throw} may appear without an argument (as shown above), in which case
1843 the object caught by the @code{@@catch} will be rethrown.
1845 Note that only (pointers to) Objective-C objects may be thrown and
1846 caught using this scheme. When an object is thrown, it will be caught
1847 by the nearest @code{@@catch} clause capable of handling objects of that type,
1848 analogously to how @code{catch} blocks work in C++ and Java. A
1849 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
1850 any and all Objective-C exceptions not caught by previous @code{@@catch}
1853 The @code{@@finally} clause, if present, will be executed upon exit from the
1854 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
1855 regardless of whether any exceptions are thrown, caught or rethrown
1856 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
1857 of the @code{finally} clause in Java.
1859 There are several caveats to using the new exception mechanism:
1863 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
1864 idioms provided by the @code{NSException} class, the new
1865 exceptions can only be used on Mac OS X 10.3 (Panther) and later
1866 systems, due to additional functionality needed in the (NeXT) Objective-C
1870 As mentioned above, the new exceptions do not support handling
1871 types other than Objective-C objects. Furthermore, when used from
1872 Objective-C++, the Objective-C exception model does not interoperate with C++
1873 exceptions at this time. This means you cannot @code{@@throw} an exception
1874 from Objective-C and @code{catch} it in C++, or vice versa
1875 (i.e., @code{throw @dots{} @@catch}).
1878 The @option{-fobjc-exceptions} switch also enables the use of synchronization
1879 blocks for thread-safe execution:
1882 @@synchronized (ObjCClass *guard) @{
1887 Upon entering the @code{@@synchronized} block, a thread of execution shall
1888 first check whether a lock has been placed on the corresponding @code{guard}
1889 object by another thread. If it has, the current thread shall wait until
1890 the other thread relinquishes its lock. Once @code{guard} becomes available,
1891 the current thread will place its own lock on it, execute the code contained in
1892 the @code{@@synchronized} block, and finally relinquish the lock (thereby
1893 making @code{guard} available to other threads).
1895 Unlike Java, Objective-C does not allow for entire methods to be marked
1896 @code{@@synchronized}. Note that throwing exceptions out of
1897 @code{@@synchronized} blocks is allowed, and will cause the guarding object
1898 to be unlocked properly.
1900 @item -freplace-objc-classes
1901 @opindex -freplace-objc-classes
1902 Emit a special marker instructing @command{ld(1)} not to statically link in
1903 the resulting object file, and allow @command{dyld(1)} to load it in at
1904 run time instead. This is used in conjunction with the Fix-and-Continue
1905 debugging mode, where the object file in question may be recompiled and
1906 dynamically reloaded in the course of program execution, without the need
1907 to restart the program itself. Currently, Fix-and-Continue functionality
1908 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
1912 @opindex -fzero-link
1913 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
1914 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
1915 compile time) with static class references that get initialized at load time,
1916 which improves run-time performance. Specifying the @option{-fzero-link} flag
1917 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
1918 to be retained. This is useful in Zero-Link debugging mode, since it allows
1919 for individual class implementations to be modified during program execution.
1923 Dump interface declarations for all classes seen in the source file to a
1924 file named @file{@var{sourcename}.decl}.
1927 @opindex Wno-protocol
1928 If a class is declared to implement a protocol, a warning is issued for
1929 every method in the protocol that is not implemented by the class. The
1930 default behavior is to issue a warning for every method not explicitly
1931 implemented in the class, even if a method implementation is inherited
1932 from the superclass. If you use the @code{-Wno-protocol} option, then
1933 methods inherited from the superclass are considered to be implemented,
1934 and no warning is issued for them.
1938 Warn if multiple methods of different types for the same selector are
1939 found during compilation. The check is performed on the list of methods
1940 in the final stage of compilation. Additionally, a check is performed
1941 for each selector appearing in a @code{@@selector(@dots{})}
1942 expression, and a corresponding method for that selector has been found
1943 during compilation. Because these checks scan the method table only at
1944 the end of compilation, these warnings are not produced if the final
1945 stage of compilation is not reached, for example because an error is
1946 found during compilation, or because the @code{-fsyntax-only} option is
1949 @item -Wundeclared-selector
1950 @opindex Wundeclared-selector
1951 Warn if a @code{@@selector(@dots{})} expression referring to an
1952 undeclared selector is found. A selector is considered undeclared if no
1953 method with that name has been declared before the
1954 @code{@@selector(@dots{})} expression, either explicitly in an
1955 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
1956 an @code{@@implementation} section. This option always performs its
1957 checks as soon as a @code{@@selector(@dots{})} expression is found,
1958 while @code{-Wselector} only performs its checks in the final stage of
1959 compilation. This also enforces the coding style convention
1960 that methods and selectors must be declared before being used.
1962 @item -print-objc-runtime-info
1963 @opindex -print-objc-runtime-info
1964 Generate C header describing the largest structure that is passed by
1969 @node Language Independent Options
1970 @section Options to Control Diagnostic Messages Formatting
1971 @cindex options to control diagnostics formatting
1972 @cindex diagnostic messages
1973 @cindex message formatting
1975 Traditionally, diagnostic messages have been formatted irrespective of
1976 the output device's aspect (e.g.@: its width, @dots{}). The options described
1977 below can be used to control the diagnostic messages formatting
1978 algorithm, e.g.@: how many characters per line, how often source location
1979 information should be reported. Right now, only the C++ front end can
1980 honor these options. However it is expected, in the near future, that
1981 the remaining front ends would be able to digest them correctly.
1984 @item -fmessage-length=@var{n}
1985 @opindex fmessage-length
1986 Try to format error messages so that they fit on lines of about @var{n}
1987 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1988 the front ends supported by GCC@. If @var{n} is zero, then no
1989 line-wrapping will be done; each error message will appear on a single
1992 @opindex fdiagnostics-show-location
1993 @item -fdiagnostics-show-location=once
1994 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1995 reporter to emit @emph{once} source location information; that is, in
1996 case the message is too long to fit on a single physical line and has to
1997 be wrapped, the source location won't be emitted (as prefix) again,
1998 over and over, in subsequent continuation lines. This is the default
2001 @item -fdiagnostics-show-location=every-line
2002 Only meaningful in line-wrapping mode. Instructs the diagnostic
2003 messages reporter to emit the same source location information (as
2004 prefix) for physical lines that result from the process of breaking
2005 a message which is too long to fit on a single line.
2009 @node Warning Options
2010 @section Options to Request or Suppress Warnings
2011 @cindex options to control warnings
2012 @cindex warning messages
2013 @cindex messages, warning
2014 @cindex suppressing warnings
2016 Warnings are diagnostic messages that report constructions which
2017 are not inherently erroneous but which are risky or suggest there
2018 may have been an error.
2020 You can request many specific warnings with options beginning @samp{-W},
2021 for example @option{-Wimplicit} to request warnings on implicit
2022 declarations. Each of these specific warning options also has a
2023 negative form beginning @samp{-Wno-} to turn off warnings;
2024 for example, @option{-Wno-implicit}. This manual lists only one of the
2025 two forms, whichever is not the default.
2027 The following options control the amount and kinds of warnings produced
2028 by GCC; for further, language-specific options also refer to
2029 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
2032 @cindex syntax checking
2034 @opindex fsyntax-only
2035 Check the code for syntax errors, but don't do anything beyond that.
2039 Issue all the warnings demanded by strict ISO C and ISO C++;
2040 reject all programs that use forbidden extensions, and some other
2041 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2042 version of the ISO C standard specified by any @option{-std} option used.
2044 Valid ISO C and ISO C++ programs should compile properly with or without
2045 this option (though a rare few will require @option{-ansi} or a
2046 @option{-std} option specifying the required version of ISO C)@. However,
2047 without this option, certain GNU extensions and traditional C and C++
2048 features are supported as well. With this option, they are rejected.
2050 @option{-pedantic} does not cause warning messages for use of the
2051 alternate keywords whose names begin and end with @samp{__}. Pedantic
2052 warnings are also disabled in the expression that follows
2053 @code{__extension__}. However, only system header files should use
2054 these escape routes; application programs should avoid them.
2055 @xref{Alternate Keywords}.
2057 Some users try to use @option{-pedantic} to check programs for strict ISO
2058 C conformance. They soon find that it does not do quite what they want:
2059 it finds some non-ISO practices, but not all---only those for which
2060 ISO C @emph{requires} a diagnostic, and some others for which
2061 diagnostics have been added.
2063 A feature to report any failure to conform to ISO C might be useful in
2064 some instances, but would require considerable additional work and would
2065 be quite different from @option{-pedantic}. We don't have plans to
2066 support such a feature in the near future.
2068 Where the standard specified with @option{-std} represents a GNU
2069 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2070 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2071 extended dialect is based. Warnings from @option{-pedantic} are given
2072 where they are required by the base standard. (It would not make sense
2073 for such warnings to be given only for features not in the specified GNU
2074 C dialect, since by definition the GNU dialects of C include all
2075 features the compiler supports with the given option, and there would be
2076 nothing to warn about.)
2078 @item -pedantic-errors
2079 @opindex pedantic-errors
2080 Like @option{-pedantic}, except that errors are produced rather than
2085 Inhibit all warning messages.
2089 Inhibit warning messages about the use of @samp{#import}.
2091 @item -Wchar-subscripts
2092 @opindex Wchar-subscripts
2093 Warn if an array subscript has type @code{char}. This is a common cause
2094 of error, as programmers often forget that this type is signed on some
2099 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2100 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2104 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2105 the arguments supplied have types appropriate to the format string
2106 specified, and that the conversions specified in the format string make
2107 sense. This includes standard functions, and others specified by format
2108 attributes (@pxref{Function Attributes}), in the @code{printf},
2109 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2110 not in the C standard) families.
2112 The formats are checked against the format features supported by GNU
2113 libc version 2.2. These include all ISO C90 and C99 features, as well
2114 as features from the Single Unix Specification and some BSD and GNU
2115 extensions. Other library implementations may not support all these
2116 features; GCC does not support warning about features that go beyond a
2117 particular library's limitations. However, if @option{-pedantic} is used
2118 with @option{-Wformat}, warnings will be given about format features not
2119 in the selected standard version (but not for @code{strfmon} formats,
2120 since those are not in any version of the C standard). @xref{C Dialect
2121 Options,,Options Controlling C Dialect}.
2123 Since @option{-Wformat} also checks for null format arguments for
2124 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2126 @option{-Wformat} is included in @option{-Wall}. For more control over some
2127 aspects of format checking, the options @option{-Wno-format-y2k},
2128 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2129 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2130 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2132 @item -Wno-format-y2k
2133 @opindex Wno-format-y2k
2134 If @option{-Wformat} is specified, do not warn about @code{strftime}
2135 formats which may yield only a two-digit year.
2137 @item -Wno-format-extra-args
2138 @opindex Wno-format-extra-args
2139 If @option{-Wformat} is specified, do not warn about excess arguments to a
2140 @code{printf} or @code{scanf} format function. The C standard specifies
2141 that such arguments are ignored.
2143 Where the unused arguments lie between used arguments that are
2144 specified with @samp{$} operand number specifications, normally
2145 warnings are still given, since the implementation could not know what
2146 type to pass to @code{va_arg} to skip the unused arguments. However,
2147 in the case of @code{scanf} formats, this option will suppress the
2148 warning if the unused arguments are all pointers, since the Single
2149 Unix Specification says that such unused arguments are allowed.
2151 @item -Wno-format-zero-length
2152 @opindex Wno-format-zero-length
2153 If @option{-Wformat} is specified, do not warn about zero-length formats.
2154 The C standard specifies that zero-length formats are allowed.
2156 @item -Wformat-nonliteral
2157 @opindex Wformat-nonliteral
2158 If @option{-Wformat} is specified, also warn if the format string is not a
2159 string literal and so cannot be checked, unless the format function
2160 takes its format arguments as a @code{va_list}.
2162 @item -Wformat-security
2163 @opindex Wformat-security
2164 If @option{-Wformat} is specified, also warn about uses of format
2165 functions that represent possible security problems. At present, this
2166 warns about calls to @code{printf} and @code{scanf} functions where the
2167 format string is not a string literal and there are no format arguments,
2168 as in @code{printf (foo);}. This may be a security hole if the format
2169 string came from untrusted input and contains @samp{%n}. (This is
2170 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2171 in future warnings may be added to @option{-Wformat-security} that are not
2172 included in @option{-Wformat-nonliteral}.)
2176 Enable @option{-Wformat} plus format checks not included in
2177 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2178 -Wformat-nonliteral -Wformat-security}.
2182 Warn about passing a null pointer for arguments marked as
2183 requiring a non-null value by the @code{nonnull} function attribute.
2185 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2186 can be disabled with the @option{-Wno-nonnull} option.
2188 @item -Winit-self @r{(C, C++, and Objective-C only)}
2190 Warn about uninitialized variables which are initialized with themselves.
2191 Note this option can only be used with the @option{-Wuninitialized} option,
2192 which in turn only works with @option{-O1} and above.
2194 For example, GCC will warn about @code{i} being uninitialized in the
2195 following snippet only when @option{-Winit-self} has been specified:
2206 @item -Wimplicit-int
2207 @opindex Wimplicit-int
2208 Warn when a declaration does not specify a type.
2210 @item -Wimplicit-function-declaration
2211 @itemx -Werror-implicit-function-declaration
2212 @opindex Wimplicit-function-declaration
2213 @opindex Werror-implicit-function-declaration
2214 Give a warning (or error) whenever a function is used before being
2219 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2223 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2224 function with external linkage, returning int, taking either zero
2225 arguments, two, or three arguments of appropriate types.
2227 @item -Wmissing-braces
2228 @opindex Wmissing-braces
2229 Warn if an aggregate or union initializer is not fully bracketed. In
2230 the following example, the initializer for @samp{a} is not fully
2231 bracketed, but that for @samp{b} is fully bracketed.
2234 int a[2][2] = @{ 0, 1, 2, 3 @};
2235 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2239 @opindex Wparentheses
2240 Warn if parentheses are omitted in certain contexts, such
2241 as when there is an assignment in a context where a truth value
2242 is expected, or when operators are nested whose precedence people
2243 often get confused about.
2245 Also warn about constructions where there may be confusion to which
2246 @code{if} statement an @code{else} branch belongs. Here is an example of
2261 In C, every @code{else} branch belongs to the innermost possible @code{if}
2262 statement, which in this example is @code{if (b)}. This is often not
2263 what the programmer expected, as illustrated in the above example by
2264 indentation the programmer chose. When there is the potential for this
2265 confusion, GCC will issue a warning when this flag is specified.
2266 To eliminate the warning, add explicit braces around the innermost
2267 @code{if} statement so there is no way the @code{else} could belong to
2268 the enclosing @code{if}. The resulting code would look like this:
2284 @item -Wsequence-point
2285 @opindex Wsequence-point
2286 Warn about code that may have undefined semantics because of violations
2287 of sequence point rules in the C standard.
2289 The C standard defines the order in which expressions in a C program are
2290 evaluated in terms of @dfn{sequence points}, which represent a partial
2291 ordering between the execution of parts of the program: those executed
2292 before the sequence point, and those executed after it. These occur
2293 after the evaluation of a full expression (one which is not part of a
2294 larger expression), after the evaluation of the first operand of a
2295 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2296 function is called (but after the evaluation of its arguments and the
2297 expression denoting the called function), and in certain other places.
2298 Other than as expressed by the sequence point rules, the order of
2299 evaluation of subexpressions of an expression is not specified. All
2300 these rules describe only a partial order rather than a total order,
2301 since, for example, if two functions are called within one expression
2302 with no sequence point between them, the order in which the functions
2303 are called is not specified. However, the standards committee have
2304 ruled that function calls do not overlap.
2306 It is not specified when between sequence points modifications to the
2307 values of objects take effect. Programs whose behavior depends on this
2308 have undefined behavior; the C standard specifies that ``Between the
2309 previous and next sequence point an object shall have its stored value
2310 modified at most once by the evaluation of an expression. Furthermore,
2311 the prior value shall be read only to determine the value to be
2312 stored.''. If a program breaks these rules, the results on any
2313 particular implementation are entirely unpredictable.
2315 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2316 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2317 diagnosed by this option, and it may give an occasional false positive
2318 result, but in general it has been found fairly effective at detecting
2319 this sort of problem in programs.
2321 The present implementation of this option only works for C programs. A
2322 future implementation may also work for C++ programs.
2324 The C standard is worded confusingly, therefore there is some debate
2325 over the precise meaning of the sequence point rules in subtle cases.
2326 Links to discussions of the problem, including proposed formal
2327 definitions, may be found on our readings page, at
2328 @w{@uref{http://gcc.gnu.org/readings.html}}.
2331 @opindex Wreturn-type
2332 Warn whenever a function is defined with a return-type that defaults to
2333 @code{int}. Also warn about any @code{return} statement with no
2334 return-value in a function whose return-type is not @code{void}.
2336 For C++, a function without return type always produces a diagnostic
2337 message, even when @option{-Wno-return-type} is specified. The only
2338 exceptions are @samp{main} and functions defined in system headers.
2342 Warn whenever a @code{switch} statement has an index of enumeral type
2343 and lacks a @code{case} for one or more of the named codes of that
2344 enumeration. (The presence of a @code{default} label prevents this
2345 warning.) @code{case} labels outside the enumeration range also
2346 provoke warnings when this option is used.
2348 @item -Wswitch-default
2349 @opindex Wswitch-switch
2350 Warn whenever a @code{switch} statement does not have a @code{default}
2354 @opindex Wswitch-enum
2355 Warn whenever a @code{switch} statement has an index of enumeral type
2356 and lacks a @code{case} for one or more of the named codes of that
2357 enumeration. @code{case} labels outside the enumeration range also
2358 provoke warnings when this option is used.
2362 Warn if any trigraphs are encountered that might change the meaning of
2363 the program (trigraphs within comments are not warned about).
2365 @item -Wunused-function
2366 @opindex Wunused-function
2367 Warn whenever a static function is declared but not defined or a
2368 non\-inline static function is unused.
2370 @item -Wunused-label
2371 @opindex Wunused-label
2372 Warn whenever a label is declared but not used.
2374 To suppress this warning use the @samp{unused} attribute
2375 (@pxref{Variable Attributes}).
2377 @item -Wunused-parameter
2378 @opindex Wunused-parameter
2379 Warn whenever a function parameter is unused aside from its declaration.
2381 To suppress this warning use the @samp{unused} attribute
2382 (@pxref{Variable Attributes}).
2384 @item -Wunused-variable
2385 @opindex Wunused-variable
2386 Warn whenever a local variable or non-constant static variable is unused
2387 aside from its declaration
2389 To suppress this warning use the @samp{unused} attribute
2390 (@pxref{Variable Attributes}).
2392 @item -Wunused-value
2393 @opindex Wunused-value
2394 Warn whenever a statement computes a result that is explicitly not used.
2396 To suppress this warning cast the expression to @samp{void}.
2400 All the above @option{-Wunused} options combined.
2402 In order to get a warning about an unused function parameter, you must
2403 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2404 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2406 @item -Wuninitialized
2407 @opindex Wuninitialized
2408 Warn if an automatic variable is used without first being initialized or
2409 if a variable may be clobbered by a @code{setjmp} call.
2411 These warnings are possible only in optimizing compilation,
2412 because they require data flow information that is computed only
2413 when optimizing. If you don't specify @option{-O}, you simply won't
2416 If you want to warn about code which uses the uninitialized value of the
2417 variable in its own initializer, use the @option{-Winit-self} option.
2419 These warnings occur only for variables that are candidates for
2420 register allocation. Therefore, they do not occur for a variable that
2421 is declared @code{volatile}, or whose address is taken, or whose size
2422 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2423 structures, unions or arrays, even when they are in registers.
2425 Note that there may be no warning about a variable that is used only
2426 to compute a value that itself is never used, because such
2427 computations may be deleted by data flow analysis before the warnings
2430 These warnings are made optional because GCC is not smart
2431 enough to see all the reasons why the code might be correct
2432 despite appearing to have an error. Here is one example of how
2453 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2454 always initialized, but GCC doesn't know this. Here is
2455 another common case:
2460 if (change_y) save_y = y, y = new_y;
2462 if (change_y) y = save_y;
2467 This has no bug because @code{save_y} is used only if it is set.
2469 @cindex @code{longjmp} warnings
2470 This option also warns when a non-volatile automatic variable might be
2471 changed by a call to @code{longjmp}. These warnings as well are possible
2472 only in optimizing compilation.
2474 The compiler sees only the calls to @code{setjmp}. It cannot know
2475 where @code{longjmp} will be called; in fact, a signal handler could
2476 call it at any point in the code. As a result, you may get a warning
2477 even when there is in fact no problem because @code{longjmp} cannot
2478 in fact be called at the place which would cause a problem.
2480 Some spurious warnings can be avoided if you declare all the functions
2481 you use that never return as @code{noreturn}. @xref{Function
2484 @item -Wunknown-pragmas
2485 @opindex Wunknown-pragmas
2486 @cindex warning for unknown pragmas
2487 @cindex unknown pragmas, warning
2488 @cindex pragmas, warning of unknown
2489 Warn when a #pragma directive is encountered which is not understood by
2490 GCC@. If this command line option is used, warnings will even be issued
2491 for unknown pragmas in system header files. This is not the case if
2492 the warnings were only enabled by the @option{-Wall} command line option.
2494 @item -Wstrict-aliasing
2495 @opindex Wstrict-aliasing
2496 This option is only active when @option{-fstrict-aliasing} is active.
2497 It warns about code which might break the strict aliasing rules that the
2498 compiler is using for optimization. The warning does not catch all
2499 cases, but does attempt to catch the more common pitfalls. It is
2500 included in @option{-Wall}.
2504 All of the above @samp{-W} options combined. This enables all the
2505 warnings about constructions that some users consider questionable, and
2506 that are easy to avoid (or modify to prevent the warning), even in
2507 conjunction with macros. This also enables some language-specific
2508 warnings described in @ref{C++ Dialect Options} and
2509 @ref{Objective-C Dialect Options}.
2512 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2513 Some of them warn about constructions that users generally do not
2514 consider questionable, but which occasionally you might wish to check
2515 for; others warn about constructions that are necessary or hard to avoid
2516 in some cases, and there is no simple way to modify the code to suppress
2523 (This option used to be called @option{-W}. The older name is still
2524 supported, but the newer name is more descriptive.) Print extra warning
2525 messages for these events:
2529 A function can return either with or without a value. (Falling
2530 off the end of the function body is considered returning without
2531 a value.) For example, this function would evoke such a
2545 An expression-statement or the left-hand side of a comma expression
2546 contains no side effects.
2547 To suppress the warning, cast the unused expression to void.
2548 For example, an expression such as @samp{x[i,j]} will cause a warning,
2549 but @samp{x[(void)i,j]} will not.
2552 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2555 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2556 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2557 that of ordinary mathematical notation.
2560 Storage-class specifiers like @code{static} are not the first things in
2561 a declaration. According to the C Standard, this usage is obsolescent.
2564 The return type of a function has a type qualifier such as @code{const}.
2565 Such a type qualifier has no effect, since the value returned by a
2566 function is not an lvalue. (But don't warn about the GNU extension of
2567 @code{volatile void} return types. That extension will be warned about
2568 if @option{-pedantic} is specified.)
2571 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2575 A comparison between signed and unsigned values could produce an
2576 incorrect result when the signed value is converted to unsigned.
2577 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2580 An aggregate has an initializer which does not initialize all members.
2581 For example, the following code would cause such a warning, because
2582 @code{x.h} would be implicitly initialized to zero:
2585 struct s @{ int f, g, h; @};
2586 struct s x = @{ 3, 4 @};
2590 A function parameter is declared without a type specifier in K&R-style
2598 An empty body occurs in an @samp{if} or @samp{else} statement.
2601 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2602 @samp{>}, or @samp{>=}.
2605 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2608 Any of several floating-point events that often indicate errors, such as
2609 overflow, underflow, loss of precision, etc.
2611 @item @r{(C++ only)}
2612 An enumerator and a non-enumerator both appear in a conditional expression.
2614 @item @r{(C++ only)}
2615 A non-static reference or non-static @samp{const} member appears in a
2616 class without constructors.
2618 @item @r{(C++ only)}
2619 Ambiguous virtual bases.
2621 @item @r{(C++ only)}
2622 Subscripting an array which has been declared @samp{register}.
2624 @item @r{(C++ only)}
2625 Taking the address of a variable which has been declared @samp{register}.
2627 @item @r{(C++ only)}
2628 A base class is not initialized in a derived class' copy constructor.
2631 @item -Wno-div-by-zero
2632 @opindex Wno-div-by-zero
2633 @opindex Wdiv-by-zero
2634 Do not warn about compile-time integer division by zero. Floating point
2635 division by zero is not warned about, as it can be a legitimate way of
2636 obtaining infinities and NaNs.
2638 @item -Wsystem-headers
2639 @opindex Wsystem-headers
2640 @cindex warnings from system headers
2641 @cindex system headers, warnings from
2642 Print warning messages for constructs found in system header files.
2643 Warnings from system headers are normally suppressed, on the assumption
2644 that they usually do not indicate real problems and would only make the
2645 compiler output harder to read. Using this command line option tells
2646 GCC to emit warnings from system headers as if they occurred in user
2647 code. However, note that using @option{-Wall} in conjunction with this
2648 option will @emph{not} warn about unknown pragmas in system
2649 headers---for that, @option{-Wunknown-pragmas} must also be used.
2652 @opindex Wfloat-equal
2653 Warn if floating point values are used in equality comparisons.
2655 The idea behind this is that sometimes it is convenient (for the
2656 programmer) to consider floating-point values as approximations to
2657 infinitely precise real numbers. If you are doing this, then you need
2658 to compute (by analyzing the code, or in some other way) the maximum or
2659 likely maximum error that the computation introduces, and allow for it
2660 when performing comparisons (and when producing output, but that's a
2661 different problem). In particular, instead of testing for equality, you
2662 would check to see whether the two values have ranges that overlap; and
2663 this is done with the relational operators, so equality comparisons are
2666 @item -Wtraditional @r{(C only)}
2667 @opindex Wtraditional
2668 Warn about certain constructs that behave differently in traditional and
2669 ISO C@. Also warn about ISO C constructs that have no traditional C
2670 equivalent, and/or problematic constructs which should be avoided.
2674 Macro parameters that appear within string literals in the macro body.
2675 In traditional C macro replacement takes place within string literals,
2676 but does not in ISO C@.
2679 In traditional C, some preprocessor directives did not exist.
2680 Traditional preprocessors would only consider a line to be a directive
2681 if the @samp{#} appeared in column 1 on the line. Therefore
2682 @option{-Wtraditional} warns about directives that traditional C
2683 understands but would ignore because the @samp{#} does not appear as the
2684 first character on the line. It also suggests you hide directives like
2685 @samp{#pragma} not understood by traditional C by indenting them. Some
2686 traditional implementations would not recognize @samp{#elif}, so it
2687 suggests avoiding it altogether.
2690 A function-like macro that appears without arguments.
2693 The unary plus operator.
2696 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2697 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2698 constants.) Note, these suffixes appear in macros defined in the system
2699 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2700 Use of these macros in user code might normally lead to spurious
2701 warnings, however gcc's integrated preprocessor has enough context to
2702 avoid warning in these cases.
2705 A function declared external in one block and then used after the end of
2709 A @code{switch} statement has an operand of type @code{long}.
2712 A non-@code{static} function declaration follows a @code{static} one.
2713 This construct is not accepted by some traditional C compilers.
2716 The ISO type of an integer constant has a different width or
2717 signedness from its traditional type. This warning is only issued if
2718 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2719 typically represent bit patterns, are not warned about.
2722 Usage of ISO string concatenation is detected.
2725 Initialization of automatic aggregates.
2728 Identifier conflicts with labels. Traditional C lacks a separate
2729 namespace for labels.
2732 Initialization of unions. If the initializer is zero, the warning is
2733 omitted. This is done under the assumption that the zero initializer in
2734 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2735 initializer warnings and relies on default initialization to zero in the
2739 Conversions by prototypes between fixed/floating point values and vice
2740 versa. The absence of these prototypes when compiling with traditional
2741 C would cause serious problems. This is a subset of the possible
2742 conversion warnings, for the full set use @option{-Wconversion}.
2745 Use of ISO C style function definitions. This warning intentionally is
2746 @emph{not} issued for prototype declarations or variadic functions
2747 because these ISO C features will appear in your code when using
2748 libiberty's traditional C compatibility macros, @code{PARAMS} and
2749 @code{VPARAMS}. This warning is also bypassed for nested functions
2750 because that feature is already a gcc extension and thus not relevant to
2751 traditional C compatibility.
2754 @item -Wdeclaration-after-statement @r{(C only)}
2755 @opindex Wdeclaration-after-statement
2756 Warn when a declaration is found after a statement in a block. This
2757 construct, known from C++, was introduced with ISO C99 and is by default
2758 allowed in GCC@. It is not supported by ISO C90 and was not supported by
2759 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
2763 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2765 @item -Wendif-labels
2766 @opindex Wendif-labels
2767 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2771 Warn whenever a local variable shadows another local variable, parameter or
2772 global variable or whenever a built-in function is shadowed.
2774 @item -Wlarger-than-@var{len}
2775 @opindex Wlarger-than
2776 Warn whenever an object of larger than @var{len} bytes is defined.
2778 @item -Wpointer-arith
2779 @opindex Wpointer-arith
2780 Warn about anything that depends on the ``size of'' a function type or
2781 of @code{void}. GNU C assigns these types a size of 1, for
2782 convenience in calculations with @code{void *} pointers and pointers
2785 @item -Wbad-function-cast @r{(C only)}
2786 @opindex Wbad-function-cast
2787 Warn whenever a function call is cast to a non-matching type.
2788 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2792 Warn whenever a pointer is cast so as to remove a type qualifier from
2793 the target type. For example, warn if a @code{const char *} is cast
2794 to an ordinary @code{char *}.
2797 @opindex Wcast-align
2798 Warn whenever a pointer is cast such that the required alignment of the
2799 target is increased. For example, warn if a @code{char *} is cast to
2800 an @code{int *} on machines where integers can only be accessed at
2801 two- or four-byte boundaries.
2803 @item -Wwrite-strings
2804 @opindex Wwrite-strings
2805 When compiling C, give string constants the type @code{const
2806 char[@var{length}]} so that
2807 copying the address of one into a non-@code{const} @code{char *}
2808 pointer will get a warning; when compiling C++, warn about the
2809 deprecated conversion from string constants to @code{char *}.
2810 These warnings will help you find at
2811 compile time code that can try to write into a string constant, but
2812 only if you have been very careful about using @code{const} in
2813 declarations and prototypes. Otherwise, it will just be a nuisance;
2814 this is why we did not make @option{-Wall} request these warnings.
2817 @opindex Wconversion
2818 Warn if a prototype causes a type conversion that is different from what
2819 would happen to the same argument in the absence of a prototype. This
2820 includes conversions of fixed point to floating and vice versa, and
2821 conversions changing the width or signedness of a fixed point argument
2822 except when the same as the default promotion.
2824 Also, warn if a negative integer constant expression is implicitly
2825 converted to an unsigned type. For example, warn about the assignment
2826 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2827 casts like @code{(unsigned) -1}.
2829 @item -Wsign-compare
2830 @opindex Wsign-compare
2831 @cindex warning for comparison of signed and unsigned values
2832 @cindex comparison of signed and unsigned values, warning
2833 @cindex signed and unsigned values, comparison warning
2834 Warn when a comparison between signed and unsigned values could produce
2835 an incorrect result when the signed value is converted to unsigned.
2836 This warning is also enabled by @option{-Wextra}; to get the other warnings
2837 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
2839 @item -Waggregate-return
2840 @opindex Waggregate-return
2841 Warn if any functions that return structures or unions are defined or
2842 called. (In languages where you can return an array, this also elicits
2845 @item -Wstrict-prototypes @r{(C only)}
2846 @opindex Wstrict-prototypes
2847 Warn if a function is declared or defined without specifying the
2848 argument types. (An old-style function definition is permitted without
2849 a warning if preceded by a declaration which specifies the argument
2852 @item -Wold-style-definition @r{(C only)}
2853 @opindex Wold-style-definition
2854 Warn if an old-style function definition is used. A warning is given
2855 even if there is a previous prototype.
2857 @item -Wmissing-prototypes @r{(C only)}
2858 @opindex Wmissing-prototypes
2859 Warn if a global function is defined without a previous prototype
2860 declaration. This warning is issued even if the definition itself
2861 provides a prototype. The aim is to detect global functions that fail
2862 to be declared in header files.
2864 @item -Wmissing-declarations @r{(C only)}
2865 @opindex Wmissing-declarations
2866 Warn if a global function is defined without a previous declaration.
2867 Do so even if the definition itself provides a prototype.
2868 Use this option to detect global functions that are not declared in
2871 @item -Wmissing-noreturn
2872 @opindex Wmissing-noreturn
2873 Warn about functions which might be candidates for attribute @code{noreturn}.
2874 Note these are only possible candidates, not absolute ones. Care should
2875 be taken to manually verify functions actually do not ever return before
2876 adding the @code{noreturn} attribute, otherwise subtle code generation
2877 bugs could be introduced. You will not get a warning for @code{main} in
2878 hosted C environments.
2880 @item -Wmissing-format-attribute
2881 @opindex Wmissing-format-attribute
2883 If @option{-Wformat} is enabled, also warn about functions which might be
2884 candidates for @code{format} attributes. Note these are only possible
2885 candidates, not absolute ones. GCC will guess that @code{format}
2886 attributes might be appropriate for any function that calls a function
2887 like @code{vprintf} or @code{vscanf}, but this might not always be the
2888 case, and some functions for which @code{format} attributes are
2889 appropriate may not be detected. This option has no effect unless
2890 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2892 @item -Wno-multichar
2893 @opindex Wno-multichar
2895 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2896 Usually they indicate a typo in the user's code, as they have
2897 implementation-defined values, and should not be used in portable code.
2899 @item -Wno-deprecated-declarations
2900 @opindex Wno-deprecated-declarations
2901 Do not warn about uses of functions, variables, and types marked as
2902 deprecated by using the @code{deprecated} attribute.
2903 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2904 @pxref{Type Attributes}.)
2908 Warn if a structure is given the packed attribute, but the packed
2909 attribute has no effect on the layout or size of the structure.
2910 Such structures may be mis-aligned for little benefit. For
2911 instance, in this code, the variable @code{f.x} in @code{struct bar}
2912 will be misaligned even though @code{struct bar} does not itself
2913 have the packed attribute:
2920 @} __attribute__((packed));
2930 Warn if padding is included in a structure, either to align an element
2931 of the structure or to align the whole structure. Sometimes when this
2932 happens it is possible to rearrange the fields of the structure to
2933 reduce the padding and so make the structure smaller.
2935 @item -Wredundant-decls
2936 @opindex Wredundant-decls
2937 Warn if anything is declared more than once in the same scope, even in
2938 cases where multiple declaration is valid and changes nothing.
2940 @item -Wnested-externs @r{(C only)}
2941 @opindex Wnested-externs
2942 Warn if an @code{extern} declaration is encountered within a function.
2944 @item -Wunreachable-code
2945 @opindex Wunreachable-code
2946 Warn if the compiler detects that code will never be executed.
2948 This option is intended to warn when the compiler detects that at
2949 least a whole line of source code will never be executed, because
2950 some condition is never satisfied or because it is after a
2951 procedure that never returns.
2953 It is possible for this option to produce a warning even though there
2954 are circumstances under which part of the affected line can be executed,
2955 so care should be taken when removing apparently-unreachable code.
2957 For instance, when a function is inlined, a warning may mean that the
2958 line is unreachable in only one inlined copy of the function.
2960 This option is not made part of @option{-Wall} because in a debugging
2961 version of a program there is often substantial code which checks
2962 correct functioning of the program and is, hopefully, unreachable
2963 because the program does work. Another common use of unreachable
2964 code is to provide behavior which is selectable at compile-time.
2968 Warn if a function can not be inlined and it was declared as inline.
2969 Even with this option, the compiler will not warn about failures to
2970 inline functions declared in system headers.
2972 The compiler uses a variety of heuristics to determine whether or not
2973 to inline a function. For example, the compiler takes into account
2974 the size of the function being inlined and the the amount of inlining
2975 that has already been done in the current function. Therefore,
2976 seemingly insignificant changes in the source program can cause the
2977 warnings produced by @option{-Winline} to appear or disappear.
2979 @item -Wno-invalid-offsetof @r{(C++ only)}
2980 @opindex Wno-invalid-offsetof
2981 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
2982 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
2983 to a non-POD type is undefined. In existing C++ implementations,
2984 however, @samp{offsetof} typically gives meaningful results even when
2985 applied to certain kinds of non-POD types. (Such as a simple
2986 @samp{struct} that fails to be a POD type only by virtue of having a
2987 constructor.) This flag is for users who are aware that they are
2988 writing nonportable code and who have deliberately chosen to ignore the
2991 The restrictions on @samp{offsetof} may be relaxed in a future version
2992 of the C++ standard.
2995 @opindex Winvalid-pch
2996 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
2997 the search path but can't be used.
3001 @opindex Wno-long-long
3002 Warn if @samp{long long} type is used. This is default. To inhibit
3003 the warning messages, use @option{-Wno-long-long}. Flags
3004 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3005 only when @option{-pedantic} flag is used.
3007 @item -Wdisabled-optimization
3008 @opindex Wdisabled-optimization
3009 Warn if a requested optimization pass is disabled. This warning does
3010 not generally indicate that there is anything wrong with your code; it
3011 merely indicates that GCC's optimizers were unable to handle the code
3012 effectively. Often, the problem is that your code is too big or too
3013 complex; GCC will refuse to optimize programs when the optimization
3014 itself is likely to take inordinate amounts of time.
3018 Make all warnings into errors.
3021 @node Debugging Options
3022 @section Options for Debugging Your Program or GCC
3023 @cindex options, debugging
3024 @cindex debugging information options
3026 GCC has various special options that are used for debugging
3027 either your program or GCC:
3032 Produce debugging information in the operating system's native format
3033 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
3036 On most systems that use stabs format, @option{-g} enables use of extra
3037 debugging information that only GDB can use; this extra information
3038 makes debugging work better in GDB but will probably make other debuggers
3040 refuse to read the program. If you want to control for certain whether
3041 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3042 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
3043 or @option{-gvms} (see below).
3045 Unlike most other C compilers, GCC allows you to use @option{-g} with
3046 @option{-O}. The shortcuts taken by optimized code may occasionally
3047 produce surprising results: some variables you declared may not exist
3048 at all; flow of control may briefly move where you did not expect it;
3049 some statements may not be executed because they compute constant
3050 results or their values were already at hand; some statements may
3051 execute in different places because they were moved out of loops.
3053 Nevertheless it proves possible to debug optimized output. This makes
3054 it reasonable to use the optimizer for programs that might have bugs.
3056 The following options are useful when GCC is generated with the
3057 capability for more than one debugging format.
3061 Produce debugging information for use by GDB@. This means to use the
3062 most expressive format available (DWARF 2, stabs, or the native format
3063 if neither of those are supported), including GDB extensions if at all
3068 Produce debugging information in stabs format (if that is supported),
3069 without GDB extensions. This is the format used by DBX on most BSD
3070 systems. On MIPS, Alpha and System V Release 4 systems this option
3071 produces stabs debugging output which is not understood by DBX or SDB@.
3072 On System V Release 4 systems this option requires the GNU assembler.
3074 @item -feliminate-unused-debug-symbols
3075 @opindex feliminate-unused-debug-symbols
3076 Produce debugging information in stabs format (if that is supported),
3077 for only symbols that are actually used.
3081 Produce debugging information in stabs format (if that is supported),
3082 using GNU extensions understood only by the GNU debugger (GDB)@. The
3083 use of these extensions is likely to make other debuggers crash or
3084 refuse to read the program.
3088 Produce debugging information in COFF format (if that is supported).
3089 This is the format used by SDB on most System V systems prior to
3094 Produce debugging information in XCOFF format (if that is supported).
3095 This is the format used by the DBX debugger on IBM RS/6000 systems.
3099 Produce debugging information in XCOFF format (if that is supported),
3100 using GNU extensions understood only by the GNU debugger (GDB)@. The
3101 use of these extensions is likely to make other debuggers crash or
3102 refuse to read the program, and may cause assemblers other than the GNU
3103 assembler (GAS) to fail with an error.
3107 Produce debugging information in DWARF version 1 format (if that is
3108 supported). This is the format used by SDB on most System V Release 4
3111 This option is deprecated.
3115 Produce debugging information in DWARF version 1 format (if that is
3116 supported), using GNU extensions understood only by the GNU debugger
3117 (GDB)@. The use of these extensions is likely to make other debuggers
3118 crash or refuse to read the program.
3120 This option is deprecated.
3124 Produce debugging information in DWARF version 2 format (if that is
3125 supported). This is the format used by DBX on IRIX 6.
3129 Produce debugging information in VMS debug format (if that is
3130 supported). This is the format used by DEBUG on VMS systems.
3133 @itemx -ggdb@var{level}
3134 @itemx -gstabs@var{level}
3135 @itemx -gcoff@var{level}
3136 @itemx -gxcoff@var{level}
3137 @itemx -gvms@var{level}
3138 Request debugging information and also use @var{level} to specify how
3139 much information. The default level is 2.
3141 Level 1 produces minimal information, enough for making backtraces in
3142 parts of the program that you don't plan to debug. This includes
3143 descriptions of functions and external variables, but no information
3144 about local variables and no line numbers.
3146 Level 3 includes extra information, such as all the macro definitions
3147 present in the program. Some debuggers support macro expansion when
3148 you use @option{-g3}.
3150 Note that in order to avoid confusion between DWARF1 debug level 2,
3151 and DWARF2, neither @option{-gdwarf} nor @option{-gdwarf-2} accept
3152 a concatenated debug level. Instead use an additional @option{-g@var{level}}
3153 option to change the debug level for DWARF1 or DWARF2.
3155 @item -feliminate-dwarf2-dups
3156 @opindex feliminate-dwarf2-dups
3157 Compress DWARF2 debugging information by eliminating duplicated
3158 information about each symbol. This option only makes sense when
3159 generating DWARF2 debugging information with @option{-gdwarf-2}.
3161 @cindex @command{prof}
3164 Generate extra code to write profile information suitable for the
3165 analysis program @command{prof}. You must use this option when compiling
3166 the source files you want data about, and you must also use it when
3169 @cindex @command{gprof}
3172 Generate extra code to write profile information suitable for the
3173 analysis program @command{gprof}. You must use this option when compiling
3174 the source files you want data about, and you must also use it when
3179 Makes the compiler print out each function name as it is compiled, and
3180 print some statistics about each pass when it finishes.
3183 @opindex ftime-report
3184 Makes the compiler print some statistics about the time consumed by each
3185 pass when it finishes.
3188 @opindex fmem-report
3189 Makes the compiler print some statistics about permanent memory
3190 allocation when it finishes.
3192 @item -fprofile-arcs
3193 @opindex fprofile-arcs
3194 Add code so that program flow @dfn{arcs} are instrumented. During
3195 execution the program records how many times each branch and call is
3196 executed and how many times it is taken or returns. When the compiled
3197 program exits it saves this data to a file called
3198 @file{@var{auxname}.gcda} for each source file. The data may be used for
3199 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3200 test coverage analysis (@option{-ftest-coverage}). Each object file's
3201 @var{auxname} is generated from the name of the output file, if
3202 explicitly specified and it is not the final executable, otherwise it is
3203 the basename of the source file. In both cases any suffix is removed
3204 (e.g. @file{foo.gcda} for input file @file{dir/foo.c}, or
3205 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3210 Compile the source files with @option{-fprofile-arcs} plus optimization
3211 and code generation options. For test coverage analysis, use the
3212 additional @option{-ftest-coverage} option. You do not need to profile
3213 every source file in a program.
3216 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3217 (the latter implies the former).
3220 Run the program on a representative workload to generate the arc profile
3221 information. This may be repeated any number of times. You can run
3222 concurrent instances of your program, and provided that the file system
3223 supports locking, the data files will be correctly updated. Also
3224 @code{fork} calls are detected and correctly handled (double counting
3228 For profile-directed optimizations, compile the source files again with
3229 the same optimization and code generation options plus
3230 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3231 Control Optimization}).
3234 For test coverage analysis, use @command{gcov} to produce human readable
3235 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3236 @command{gcov} documentation for further information.
3240 With @option{-fprofile-arcs}, for each function of your program GCC
3241 creates a program flow graph, then finds a spanning tree for the graph.
3242 Only arcs that are not on the spanning tree have to be instrumented: the
3243 compiler adds code to count the number of times that these arcs are
3244 executed. When an arc is the only exit or only entrance to a block, the
3245 instrumentation code can be added to the block; otherwise, a new basic
3246 block must be created to hold the instrumentation code.
3249 @item -ftest-coverage
3250 @opindex ftest-coverage
3251 Produce a notes file that the @command{gcov} code-coverage utility
3252 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3253 show program coverage. Each source file's note file is called
3254 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3255 above for a description of @var{auxname} and instructions on how to
3256 generate test coverage data. Coverage data will match the source files
3257 more closely, if you do not optimize.
3259 @item -d@var{letters}
3261 Says to make debugging dumps during compilation at times specified by
3262 @var{letters}. This is used for debugging the compiler. The file names
3263 for most of the dumps are made by appending a pass number and a word to
3264 the @var{dumpname}. @var{dumpname} is generated from the name of the
3265 output file, if explicitly specified and it is not an executable,
3266 otherwise it is the basename of the source file. In both cases any
3267 suffix is removed (e.g. @file{foo.00.rtl} or @file{foo.01.sibling}).
3268 Here are the possible letters for use in @var{letters}, and their
3274 Annotate the assembler output with miscellaneous debugging information.
3277 Dump after computing branch probabilities, to @file{@var{file}.16.bp}.
3280 Dump after block reordering, to @file{@var{file}.32.bbro}.
3283 Dump after instruction combination, to the file @file{@var{file}.22.combine}.
3286 Dump after the first if conversion, to the file @file{@var{file}.17.ce1}.
3287 Also dump after the second if conversion, to the file @file{@var{file}.23.ce2}.
3290 Dump after branch target load optimization, to to @file{@var{file}.34.btl}.
3291 Also dump after delayed branch scheduling, to @file{@var{file}.37.dbr}.
3294 Dump all macro definitions, at the end of preprocessing, in addition to
3298 Dump after SSA optimizations, to @file{@var{file}.05.ssa} and
3299 @file{@var{file}.010.ussa}.
3302 Dump after the second if conversion, to @file{@var{file}.33.ce3}.
3305 Dump after control and data flow analysis, to @file{@var{file}.15.cfg}.
3306 Also dump after life analysis, to @file{@var{file}.21.life}.
3309 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.11.addressof}.
3312 Dump after global register allocation, to @file{@var{file}.27.greg}.
3315 Dump after GCSE, to @file{@var{file}.12.gcse}.
3316 Also dump after jump bypassing and control flow optimizations, to
3317 @file{@var{file}.14.bypass}.
3320 Dump after finalization of EH handling code, to @file{@var{file}.03.eh}.
3323 Dump after sibling call optimizations, to @file{@var{file}.02.sibling}.
3326 Dump after the first jump optimization, to @file{@var{file}.04.jump}.
3329 Dump after conversion from registers to stack, to @file{@var{file}.36.stack}.
3332 Dump after local register allocation, to @file{@var{file}.26.lreg}.
3335 Dump after loop optimization passes, to @file{@var{file}.13.loop} and
3336 @file{@var{file}.19.loop2}.
3339 Dump after performing the machine dependent reorganization pass, to
3340 @file{@var{file}.37.mach}.
3343 Dump after register renumbering, to @file{@var{file}.31.rnreg}.
3346 Dump after the register move pass, to @file{@var{file}.24.regmove}.
3349 Dump after post-reload optimizations, to @file{@var{file}.28.postreload}.
3352 Dump after RTL generation, to @file{@var{file}.01.rtl}.
3355 Dump after the second scheduling pass, to @file{@var{file}.35.sched2}.
3358 Dump after CSE (including the jump optimization that sometimes follows
3359 CSE), to @file{@var{file}.019.cse}.
3362 Dump after the first scheduling pass, to @file{@var{file}.25.sched}.
3365 Dump after the second CSE pass (including the jump optimization that
3366 sometimes follows CSE), to @file{@var{file}.20.cse2}.
3369 Dump after running tracer, to @file{@var{file}.18.tracer}.
3372 Dump after null pointer elimination pass to @file{@var{file}.018.null}.
3375 Dump callgraph and unit-at-a-time optimization @file{@var{file}.00.unit}.
3378 Dump after the second flow pass, to @file{@var{file}.29.flow2}.
3381 Dump after SSA conditional constant propagation, to
3382 @file{@var{file}.06.ssaccp}.
3385 Dump after SSA dead code elimination, to @file{@var{file}.07.ssadce}.
3388 Dump after the peephole pass, to @file{@var{file}.30.peephole2}.
3391 Produce all the dumps listed above.
3394 Produce a core dump whenever an error occurs.
3397 Print statistics on memory usage, at the end of the run, to
3401 Annotate the assembler output with a comment indicating which
3402 pattern and alternative was used. The length of each instruction is
3406 Dump the RTL in the assembler output as a comment before each instruction.
3407 Also turns on @option{-dp} annotation.
3410 For each of the other indicated dump files (except for
3411 @file{@var{file}.01.rtl}), dump a representation of the control flow graph
3412 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3415 Just generate RTL for a function instead of compiling it. Usually used
3419 Dump debugging information during parsing, to standard error.
3422 @item -fdump-unnumbered
3423 @opindex fdump-unnumbered
3424 When doing debugging dumps (see @option{-d} option above), suppress instruction
3425 numbers and line number note output. This makes it more feasible to
3426 use diff on debugging dumps for compiler invocations with different
3427 options, in particular with and without @option{-g}.
3429 @item -fdump-translation-unit @r{(C and C++ only)}
3430 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3431 @opindex fdump-translation-unit
3432 Dump a representation of the tree structure for the entire translation
3433 unit to a file. The file name is made by appending @file{.tu} to the
3434 source file name. If the @samp{-@var{options}} form is used, @var{options}
3435 controls the details of the dump as described for the
3436 @option{-fdump-tree} options.
3438 @item -fdump-class-hierarchy @r{(C++ only)}
3439 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3440 @opindex fdump-class-hierarchy
3441 Dump a representation of each class's hierarchy and virtual function
3442 table layout to a file. The file name is made by appending @file{.class}
3443 to the source file name. If the @samp{-@var{options}} form is used,
3444 @var{options} controls the details of the dump as described for the
3445 @option{-fdump-tree} options.
3447 @item -fdump-tree-@var{switch} @r{(C++ only)}
3448 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3450 Control the dumping at various stages of processing the intermediate
3451 language tree to a file. The file name is generated by appending a switch
3452 specific suffix to the source file name. If the @samp{-@var{options}}
3453 form is used, @var{options} is a list of @samp{-} separated options that
3454 control the details of the dump. Not all options are applicable to all
3455 dumps, those which are not meaningful will be ignored. The following
3456 options are available
3460 Print the address of each node. Usually this is not meaningful as it
3461 changes according to the environment and source file. Its primary use
3462 is for tying up a dump file with a debug environment.
3464 Inhibit dumping of members of a scope or body of a function merely
3465 because that scope has been reached. Only dump such items when they
3466 are directly reachable by some other path.
3468 Turn on all options.
3471 The following tree dumps are possible:
3474 Dump before any tree based optimization, to @file{@var{file}.original}.
3476 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3478 Dump after function inlining, to @file{@var{file}.inlined}.
3481 @item -frandom-seed=@var{string}
3482 @opindex frandom-string
3483 This option provides a seed that GCC uses when it would otherwise use
3484 random numbers. It is used to generate certain symbol names
3485 that have to be different in every compiled file. It is also used to
3486 place unique stamps in coverage data files and the object files that
3487 produce them. You can use the @option{-frandom-seed} option to produce
3488 reproducibly identical object files.
3490 The @var{string} should be different for every file you compile.
3492 @item -fsched-verbose=@var{n}
3493 @opindex fsched-verbose
3494 On targets that use instruction scheduling, this option controls the
3495 amount of debugging output the scheduler prints. This information is
3496 written to standard error, unless @option{-dS} or @option{-dR} is
3497 specified, in which case it is output to the usual dump
3498 listing file, @file{.sched} or @file{.sched2} respectively. However
3499 for @var{n} greater than nine, the output is always printed to standard
3502 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
3503 same information as @option{-dRS}. For @var{n} greater than one, it
3504 also output basic block probabilities, detailed ready list information
3505 and unit/insn info. For @var{n} greater than two, it includes RTL
3506 at abort point, control-flow and regions info. And for @var{n} over
3507 four, @option{-fsched-verbose} also includes dependence info.
3511 Store the usual ``temporary'' intermediate files permanently; place them
3512 in the current directory and name them based on the source file. Thus,
3513 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3514 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3515 preprocessed @file{foo.i} output file even though the compiler now
3516 normally uses an integrated preprocessor.
3520 Report the CPU time taken by each subprocess in the compilation
3521 sequence. For C source files, this is the compiler proper and assembler
3522 (plus the linker if linking is done). The output looks like this:
3529 The first number on each line is the ``user time,'' that is time spent
3530 executing the program itself. The second number is ``system time,''
3531 time spent executing operating system routines on behalf of the program.
3532 Both numbers are in seconds.
3534 @item -print-file-name=@var{library}
3535 @opindex print-file-name
3536 Print the full absolute name of the library file @var{library} that
3537 would be used when linking---and don't do anything else. With this
3538 option, GCC does not compile or link anything; it just prints the
3541 @item -print-multi-directory
3542 @opindex print-multi-directory
3543 Print the directory name corresponding to the multilib selected by any
3544 other switches present in the command line. This directory is supposed
3545 to exist in @env{GCC_EXEC_PREFIX}.
3547 @item -print-multi-lib
3548 @opindex print-multi-lib
3549 Print the mapping from multilib directory names to compiler switches
3550 that enable them. The directory name is separated from the switches by
3551 @samp{;}, and each switch starts with an @samp{@@} instead of the
3552 @samp{-}, without spaces between multiple switches. This is supposed to
3553 ease shell-processing.
3555 @item -print-prog-name=@var{program}
3556 @opindex print-prog-name
3557 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3559 @item -print-libgcc-file-name
3560 @opindex print-libgcc-file-name
3561 Same as @option{-print-file-name=libgcc.a}.
3563 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3564 but you do want to link with @file{libgcc.a}. You can do
3567 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3570 @item -print-search-dirs
3571 @opindex print-search-dirs
3572 Print the name of the configured installation directory and a list of
3573 program and library directories gcc will search---and don't do anything else.
3575 This is useful when gcc prints the error message
3576 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3577 To resolve this you either need to put @file{cpp0} and the other compiler
3578 components where gcc expects to find them, or you can set the environment
3579 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3580 Don't forget the trailing '/'.
3581 @xref{Environment Variables}.
3584 @opindex dumpmachine
3585 Print the compiler's target machine (for example,
3586 @samp{i686-pc-linux-gnu})---and don't do anything else.
3589 @opindex dumpversion
3590 Print the compiler version (for example, @samp{3.0})---and don't do
3595 Print the compiler's built-in specs---and don't do anything else. (This
3596 is used when GCC itself is being built.) @xref{Spec Files}.
3598 @item -feliminate-unused-debug-types
3599 @opindex feliminate-unused-debug-types
3600 Normally, when producing DWARF2 output, GCC will emit debugging
3601 information for all types declared in a compilation
3602 unit, regardless of whether or not they are actually used
3603 in that compilation unit. Sometimes this is useful, such as
3604 if, in the debugger, you want to cast a value to a type that is
3605 not actually used in your program (but is declared). More often,
3606 however, this results in a significant amount of wasted space.
3607 With this option, GCC will avoid producing debug symbol output
3608 for types that are nowhere used in the source file being compiled.
3611 @node Optimize Options
3612 @section Options That Control Optimization
3613 @cindex optimize options
3614 @cindex options, optimization
3616 These options control various sorts of optimizations.
3618 Without any optimization option, the compiler's goal is to reduce the
3619 cost of compilation and to make debugging produce the expected
3620 results. Statements are independent: if you stop the program with a
3621 breakpoint between statements, you can then assign a new value to any
3622 variable or change the program counter to any other statement in the
3623 function and get exactly the results you would expect from the source
3626 Turning on optimization flags makes the compiler attempt to improve
3627 the performance and/or code size at the expense of compilation time
3628 and possibly the ability to debug the program.
3630 The compiler performs optimisation based on the knowledge it has of
3631 the program. Using the @option{-funit-at-a-time} flag will allow the
3632 compiler to consider information gained from later functions in the
3633 file when compiling a function. Compiling multiple files at once to a
3634 single output file (and using @option{-funit-at-a-time}) will allow
3635 the compiler to use information gained from all of the files when
3636 compiling each of them.
3638 Not all optimizations are controlled directly by a flag. Only
3639 optimizations that have a flag are listed.
3646 Optimize. Optimizing compilation takes somewhat more time, and a lot
3647 more memory for a large function.
3649 With @option{-O}, the compiler tries to reduce code size and execution
3650 time, without performing any optimizations that take a great deal of
3653 @option{-O} turns on the following optimization flags:
3654 @gccoptlist{-fdefer-pop @gol
3655 -fmerge-constants @gol
3657 -floop-optimize @gol
3659 -fif-conversion @gol
3660 -fif-conversion2 @gol
3661 -fdelayed-branch @gol
3662 -fguess-branch-probability @gol
3665 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
3666 where doing so does not interfere with debugging.
3670 Optimize even more. GCC performs nearly all supported optimizations
3671 that do not involve a space-speed tradeoff. The compiler does not
3672 perform loop unrolling or function inlining when you specify @option{-O2}.
3673 As compared to @option{-O}, this option increases both compilation time
3674 and the performance of the generated code.
3676 @option{-O2} turns on all optimization flags specified by @option{-O}. It
3677 also turns on the following optimization flags:
3678 @gccoptlist{-fforce-mem @gol
3679 -foptimize-sibling-calls @gol
3680 -fstrength-reduce @gol
3681 -fcse-follow-jumps -fcse-skip-blocks @gol
3682 -frerun-cse-after-loop -frerun-loop-opt @gol
3683 -fgcse -fgcse-lm -fgcse-sm -fgcse-las @gol
3684 -fdelete-null-pointer-checks @gol
3685 -fexpensive-optimizations @gol
3687 -fschedule-insns -fschedule-insns2 @gol
3688 -fsched-interblock -fsched-spec @gol
3691 -freorder-blocks -freorder-functions @gol
3692 -fstrict-aliasing @gol
3693 -funit-at-a-time -fweb @gol
3694 -falign-functions -falign-jumps @gol
3695 -falign-loops -falign-labels}
3697 Please note the warning under @option{-fgcse} about
3698 invoking @option{-O2} on programs that use computed gotos.
3702 Optimize yet more. @option{-O3} turns on all optimizations specified by
3703 @option{-O2} and also turns on the @option{-finline-functions},
3704 @option{-fweb} and @option{-frename-registers} options.
3708 Do not optimize. This is the default.
3712 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3713 do not typically increase code size. It also performs further
3714 optimizations designed to reduce code size.
3716 @option{-Os} disables the following optimization flags:
3717 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
3718 -falign-labels -freorder-blocks -fprefetch-loop-arrays}
3720 If you use multiple @option{-O} options, with or without level numbers,
3721 the last such option is the one that is effective.
3724 Options of the form @option{-f@var{flag}} specify machine-independent
3725 flags. Most flags have both positive and negative forms; the negative
3726 form of @option{-ffoo} would be @option{-fno-foo}. In the table
3727 below, only one of the forms is listed---the one you typically will
3728 use. You can figure out the other form by either removing @samp{no-}
3731 The following options control specific optimizations. They are either
3732 activated by @option{-O} options or are related to ones that are. You
3733 can use the following flags in the rare cases when ``fine-tuning'' of
3734 optimizations to be performed is desired.
3737 @item -fno-default-inline
3738 @opindex fno-default-inline
3739 Do not make member functions inline by default merely because they are
3740 defined inside the class scope (C++ only). Otherwise, when you specify
3741 @w{@option{-O}}, member functions defined inside class scope are compiled
3742 inline by default; i.e., you don't need to add @samp{inline} in front of
3743 the member function name.
3745 @item -fno-defer-pop
3746 @opindex fno-defer-pop
3747 Always pop the arguments to each function call as soon as that function
3748 returns. For machines which must pop arguments after a function call,
3749 the compiler normally lets arguments accumulate on the stack for several
3750 function calls and pops them all at once.
3752 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3756 Force memory operands to be copied into registers before doing
3757 arithmetic on them. This produces better code by making all memory
3758 references potential common subexpressions. When they are not common
3759 subexpressions, instruction combination should eliminate the separate
3762 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3765 @opindex fforce-addr
3766 Force memory address constants to be copied into registers before
3767 doing arithmetic on them. This may produce better code just as
3768 @option{-fforce-mem} may.
3770 @item -fomit-frame-pointer
3771 @opindex fomit-frame-pointer
3772 Don't keep the frame pointer in a register for functions that
3773 don't need one. This avoids the instructions to save, set up and
3774 restore frame pointers; it also makes an extra register available
3775 in many functions. @strong{It also makes debugging impossible on
3778 On some machines, such as the VAX, this flag has no effect, because
3779 the standard calling sequence automatically handles the frame pointer
3780 and nothing is saved by pretending it doesn't exist. The
3781 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3782 whether a target machine supports this flag. @xref{Registers,,Register
3783 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3785 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3787 @item -foptimize-sibling-calls
3788 @opindex foptimize-sibling-calls
3789 Optimize sibling and tail recursive calls.
3791 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3795 Don't pay attention to the @code{inline} keyword. Normally this option
3796 is used to keep the compiler from expanding any functions inline.
3797 Note that if you are not optimizing, no functions can be expanded inline.
3799 @item -finline-functions
3800 @opindex finline-functions
3801 Integrate all simple functions into their callers. The compiler
3802 heuristically decides which functions are simple enough to be worth
3803 integrating in this way.
3805 If all calls to a given function are integrated, and the function is
3806 declared @code{static}, then the function is normally not output as
3807 assembler code in its own right.
3809 Enabled at level @option{-O3}.
3811 @item -finline-limit=@var{n}
3812 @opindex finline-limit
3813 By default, gcc limits the size of functions that can be inlined. This flag
3814 allows the control of this limit for functions that are explicitly marked as
3815 inline (i.e., marked with the inline keyword or defined within the class
3816 definition in c++). @var{n} is the size of functions that can be inlined in
3817 number of pseudo instructions (not counting parameter handling). The default
3818 value of @var{n} is 600.
3819 Increasing this value can result in more inlined code at
3820 the cost of compilation time and memory consumption. Decreasing usually makes
3821 the compilation faster and less code will be inlined (which presumably
3822 means slower programs). This option is particularly useful for programs that
3823 use inlining heavily such as those based on recursive templates with C++.
3825 Inlining is actually controlled by a number of parameters, which may be
3826 specified individually by using @option{--param @var{name}=@var{value}}.
3827 The @option{-finline-limit=@var{n}} option sets some of these parameters
3831 @item max-inline-insns
3833 @item max-inline-insns-single
3834 is set to @var{n}/2.
3835 @item max-inline-insns-auto
3836 is set to @var{n}/2.
3837 @item min-inline-insns
3838 is set to 130 or @var{n}/4, whichever is smaller.
3839 @item max-inline-insns-rtl
3843 Using @option{-finline-limit=600} thus results in the default settings
3844 for these parameters. See below for a documentation of the individual
3845 parameters controlling inlining.
3847 @emph{Note:} pseudo instruction represents, in this particular context, an
3848 abstract measurement of function's size. In no way, it represents a count
3849 of assembly instructions and as such its exact meaning might change from one
3850 release to an another.
3852 @item -fkeep-inline-functions
3853 @opindex fkeep-inline-functions
3854 Even if all calls to a given function are integrated, and the function
3855 is declared @code{static}, nevertheless output a separate run-time
3856 callable version of the function. This switch does not affect
3857 @code{extern inline} functions.
3859 @item -fkeep-static-consts
3860 @opindex fkeep-static-consts
3861 Emit variables declared @code{static const} when optimization isn't turned
3862 on, even if the variables aren't referenced.
3864 GCC enables this option by default. If you want to force the compiler to
3865 check if the variable was referenced, regardless of whether or not
3866 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3868 @item -fmerge-constants
3869 Attempt to merge identical constants (string constants and floating point
3870 constants) across compilation units.
3872 This option is the default for optimized compilation if the assembler and
3873 linker support it. Use @option{-fno-merge-constants} to inhibit this
3876 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3878 @item -fmerge-all-constants
3879 Attempt to merge identical constants and identical variables.
3881 This option implies @option{-fmerge-constants}. In addition to
3882 @option{-fmerge-constants} this considers e.g. even constant initialized
3883 arrays or initialized constant variables with integral or floating point
3884 types. Languages like C or C++ require each non-automatic variable to
3885 have distinct location, so using this option will result in non-conforming
3890 Use a graph coloring register allocator. Currently this option is meant
3891 for testing, so we are interested to hear about miscompilations with
3894 @item -fno-branch-count-reg
3895 @opindex fno-branch-count-reg
3896 Do not use ``decrement and branch'' instructions on a count register,
3897 but instead generate a sequence of instructions that decrement a
3898 register, compare it against zero, then branch based upon the result.
3899 This option is only meaningful on architectures that support such
3900 instructions, which include x86, PowerPC, IA-64 and S/390.
3902 The default is @option{-fbranch-count-reg}, enabled when
3903 @option{-fstrength-reduce} is enabled.
3905 @item -fno-function-cse
3906 @opindex fno-function-cse
3907 Do not put function addresses in registers; make each instruction that
3908 calls a constant function contain the function's address explicitly.
3910 This option results in less efficient code, but some strange hacks
3911 that alter the assembler output may be confused by the optimizations
3912 performed when this option is not used.
3914 The default is @option{-ffunction-cse}
3916 @item -fno-zero-initialized-in-bss
3917 @opindex fno-zero-initialized-in-bss
3918 If the target supports a BSS section, GCC by default puts variables that
3919 are initialized to zero into BSS@. This can save space in the resulting
3922 This option turns off this behavior because some programs explicitly
3923 rely on variables going to the data section. E.g., so that the
3924 resulting executable can find the beginning of that section and/or make
3925 assumptions based on that.
3927 The default is @option{-fzero-initialized-in-bss}.
3929 @item -fstrength-reduce
3930 @opindex fstrength-reduce
3931 Perform the optimizations of loop strength reduction and
3932 elimination of iteration variables.
3934 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3936 @item -fthread-jumps
3937 @opindex fthread-jumps
3938 Perform optimizations where we check to see if a jump branches to a
3939 location where another comparison subsumed by the first is found. If
3940 so, the first branch is redirected to either the destination of the
3941 second branch or a point immediately following it, depending on whether
3942 the condition is known to be true or false.
3944 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3946 @item -fcse-follow-jumps
3947 @opindex fcse-follow-jumps
3948 In common subexpression elimination, scan through jump instructions
3949 when the target of the jump is not reached by any other path. For
3950 example, when CSE encounters an @code{if} statement with an
3951 @code{else} clause, CSE will follow the jump when the condition
3954 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3956 @item -fcse-skip-blocks
3957 @opindex fcse-skip-blocks
3958 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3959 follow jumps which conditionally skip over blocks. When CSE
3960 encounters a simple @code{if} statement with no else clause,
3961 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3962 body of the @code{if}.
3964 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3966 @item -frerun-cse-after-loop
3967 @opindex frerun-cse-after-loop
3968 Re-run common subexpression elimination after loop optimizations has been
3971 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3973 @item -frerun-loop-opt
3974 @opindex frerun-loop-opt
3975 Run the loop optimizer twice.
3977 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3981 Perform a global common subexpression elimination pass.
3982 This pass also performs global constant and copy propagation.
3984 @emph{Note:} When compiling a program using computed gotos, a GCC
3985 extension, you may get better runtime performance if you disable
3986 the global common subexpression elimination pass by adding
3987 @option{-fno-gcse} to the command line.
3989 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3993 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3994 attempt to move loads which are only killed by stores into themselves. This
3995 allows a loop containing a load/store sequence to be changed to a load outside
3996 the loop, and a copy/store within the loop.
3998 Enabled by default when gcse is enabled.
4002 When @option{-fgcse-sm} is enabled, a store motion pass is run after
4003 global common subexpression elimination. This pass will attempt to move
4004 stores out of loops. When used in conjunction with @option{-fgcse-lm},
4005 loops containing a load/store sequence can be changed to a load before
4006 the loop and a store after the loop.
4008 Enabled by default when gcse is enabled.
4012 When @option{-fgcse-las} is enabled, the global common subexpression
4013 elimination pass eliminates redundant loads that come after stores to the
4014 same memory location (both partial and full redundancies).
4016 Enabled by default when gcse is enabled.
4018 @item -floop-optimize
4019 @opindex floop-optimize
4020 Perform loop optimizations: move constant expressions out of loops, simplify
4021 exit test conditions and optionally do strength-reduction and loop unrolling as
4024 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4026 @item -fcrossjumping
4027 @opindex crossjumping
4028 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
4029 resulting code may or may not perform better than without cross-jumping.
4031 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4033 @item -fif-conversion
4034 @opindex if-conversion
4035 Attempt to transform conditional jumps into branch-less equivalents. This
4036 include use of conditional moves, min, max, set flags and abs instructions, and
4037 some tricks doable by standard arithmetics. The use of conditional execution
4038 on chips where it is available is controlled by @code{if-conversion2}.
4040 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4042 @item -fif-conversion2
4043 @opindex if-conversion2
4044 Use conditional execution (where available) to transform conditional jumps into
4045 branch-less equivalents.
4047 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4049 @item -fdelete-null-pointer-checks
4050 @opindex fdelete-null-pointer-checks
4051 Use global dataflow analysis to identify and eliminate useless checks
4052 for null pointers. The compiler assumes that dereferencing a null
4053 pointer would have halted the program. If a pointer is checked after
4054 it has already been dereferenced, it cannot be null.
4056 In some environments, this assumption is not true, and programs can
4057 safely dereference null pointers. Use
4058 @option{-fno-delete-null-pointer-checks} to disable this optimization
4059 for programs which depend on that behavior.
4061 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4063 @item -fexpensive-optimizations
4064 @opindex fexpensive-optimizations
4065 Perform a number of minor optimizations that are relatively expensive.
4067 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4069 @item -foptimize-register-move
4071 @opindex foptimize-register-move
4073 Attempt to reassign register numbers in move instructions and as
4074 operands of other simple instructions in order to maximize the amount of
4075 register tying. This is especially helpful on machines with two-operand
4078 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
4081 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4083 @item -fdelayed-branch
4084 @opindex fdelayed-branch
4085 If supported for the target machine, attempt to reorder instructions
4086 to exploit instruction slots available after delayed branch
4089 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4091 @item -fschedule-insns
4092 @opindex fschedule-insns
4093 If supported for the target machine, attempt to reorder instructions to
4094 eliminate execution stalls due to required data being unavailable. This
4095 helps machines that have slow floating point or memory load instructions
4096 by allowing other instructions to be issued until the result of the load
4097 or floating point instruction is required.
4099 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4101 @item -fschedule-insns2
4102 @opindex fschedule-insns2
4103 Similar to @option{-fschedule-insns}, but requests an additional pass of
4104 instruction scheduling after register allocation has been done. This is
4105 especially useful on machines with a relatively small number of
4106 registers and where memory load instructions take more than one cycle.
4108 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4110 @item -fno-sched-interblock
4111 @opindex fno-sched-interblock
4112 Don't schedule instructions across basic blocks. This is normally
4113 enabled by default when scheduling before register allocation, i.e.@:
4114 with @option{-fschedule-insns} or at @option{-O2} or higher.
4116 @item -fno-sched-spec
4117 @opindex fno-sched-spec
4118 Don't allow speculative motion of non-load instructions. This is normally
4119 enabled by default when scheduling before register allocation, i.e.@:
4120 with @option{-fschedule-insns} or at @option{-O2} or higher.
4122 @item -fsched-spec-load
4123 @opindex fsched-spec-load
4124 Allow speculative motion of some load instructions. This only makes
4125 sense when scheduling before register allocation, i.e.@: with
4126 @option{-fschedule-insns} or at @option{-O2} or higher.
4128 @item -fsched-spec-load-dangerous
4129 @opindex fsched-spec-load-dangerous
4130 Allow speculative motion of more load instructions. This only makes
4131 sense when scheduling before register allocation, i.e.@: with
4132 @option{-fschedule-insns} or at @option{-O2} or higher.
4134 @item -fsched-stalled-insns=@var{n}
4135 @opindex fsched-stalled-insns
4136 Define how many insns (if any) can be moved prematurely from the queue
4137 of stalled insns into the ready list, during the second scheduling pass.
4139 @item -fsched-stalled-insns-dep=@var{n}
4140 @opindex fsched-stalled-insns-dep
4141 Define how many insn groups (cycles) will be examined for a dependency
4142 on a stalled insn that is candidate for premature removal from the queue
4143 of stalled insns. Has an effect only during the second scheduling pass,
4144 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
4146 @item -fsched2-use-superblocks
4147 @opindex fsched2-use-superblocks
4148 When scheduling after register allocation, do use superblock scheduling
4149 algorithm. Superblock scheduling allows motion across basic block boundaries
4150 resulting on faster schedules. This option is experimental, as not all machine
4151 descriptions used by GCC model the CPU closely enough to avoid unreliable
4152 results from the algorithm.
4154 This only makes sense when scheduling after register allocation, i.e.@: with
4155 @option{-fschedule-insns2} or at @option{-O2} or higher.
4157 @item -fsched2-use-traces
4158 @opindex fsched2-use-traces
4159 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4160 allocation and additionally perform code duplication in order to increase the
4161 size of superblocks using tracer pass. See @option{-ftracer} for details on
4164 This mode should produce faster but significantly longer programs. Also
4165 without @code{-fbranch-probabilities} the traces constructed may not match the
4166 reality and hurt the performance. This only makes
4167 sense when scheduling after register allocation, i.e.@: with
4168 @option{-fschedule-insns2} or at @option{-O2} or higher.
4170 @item -fcaller-saves
4171 @opindex fcaller-saves
4172 Enable values to be allocated in registers that will be clobbered by
4173 function calls, by emitting extra instructions to save and restore the
4174 registers around such calls. Such allocation is done only when it
4175 seems to result in better code than would otherwise be produced.
4177 This option is always enabled by default on certain machines, usually
4178 those which have no call-preserved registers to use instead.
4180 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4182 @item -fmove-all-movables
4183 @opindex fmove-all-movables
4184 Forces all invariant computations in loops to be moved
4187 @item -freduce-all-givs
4188 @opindex freduce-all-givs
4189 Forces all general-induction variables in loops to be
4192 @emph{Note:} When compiling programs written in Fortran,
4193 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
4194 by default when you use the optimizer.
4196 These options may generate better or worse code; results are highly
4197 dependent on the structure of loops within the source code.
4199 These two options are intended to be removed someday, once
4200 they have helped determine the efficacy of various
4201 approaches to improving loop optimizations.
4203 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
4204 know how use of these options affects
4205 the performance of your production code.
4206 We're very interested in code that runs @emph{slower}
4207 when these options are @emph{enabled}.
4210 @itemx -fno-peephole2
4211 @opindex fno-peephole
4212 @opindex fno-peephole2
4213 Disable any machine-specific peephole optimizations. The difference
4214 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
4215 are implemented in the compiler; some targets use one, some use the
4216 other, a few use both.
4218 @option{-fpeephole} is enabled by default.
4219 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4221 @item -fno-guess-branch-probability
4222 @opindex fno-guess-branch-probability
4223 Do not guess branch probabilities using a randomized model.
4225 Sometimes gcc will opt to use a randomized model to guess branch
4226 probabilities, when none are available from either profiling feedback
4227 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
4228 different runs of the compiler on the same program may produce different
4231 In a hard real-time system, people don't want different runs of the
4232 compiler to produce code that has different behavior; minimizing
4233 non-determinism is of paramount import. This switch allows users to
4234 reduce non-determinism, possibly at the expense of inferior
4237 The default is @option{-fguess-branch-probability} at levels
4238 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4240 @item -freorder-blocks
4241 @opindex freorder-blocks
4242 Reorder basic blocks in the compiled function in order to reduce number of
4243 taken branches and improve code locality.
4245 Enabled at levels @option{-O2}, @option{-O3}.
4247 @item -freorder-functions
4248 @opindex freorder-functions
4249 Reorder basic blocks in the compiled function in order to reduce number of
4250 taken branches and improve code locality. This is implemented by using special
4251 subsections @code{text.hot} for most frequently executed functions and
4252 @code{text.unlikely} for unlikely executed functions. Reordering is done by
4253 the linker so object file format must support named sections and linker must
4254 place them in a reasonable way.
4256 Also profile feedback must be available in to make this option effective. See
4257 @option{-fprofile-arcs} for details.
4259 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4261 @item -fstrict-aliasing
4262 @opindex fstrict-aliasing
4263 Allows the compiler to assume the strictest aliasing rules applicable to
4264 the language being compiled. For C (and C++), this activates
4265 optimizations based on the type of expressions. In particular, an
4266 object of one type is assumed never to reside at the same address as an
4267 object of a different type, unless the types are almost the same. For
4268 example, an @code{unsigned int} can alias an @code{int}, but not a
4269 @code{void*} or a @code{double}. A character type may alias any other
4272 Pay special attention to code like this:
4285 The practice of reading from a different union member than the one most
4286 recently written to (called ``type-punning'') is common. Even with
4287 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
4288 is accessed through the union type. So, the code above will work as
4289 expected. However, this code might not:
4300 Every language that wishes to perform language-specific alias analysis
4301 should define a function that computes, given an @code{tree}
4302 node, an alias set for the node. Nodes in different alias sets are not
4303 allowed to alias. For an example, see the C front-end function
4304 @code{c_get_alias_set}.
4306 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4308 @item -falign-functions
4309 @itemx -falign-functions=@var{n}
4310 @opindex falign-functions
4311 Align the start of functions to the next power-of-two greater than
4312 @var{n}, skipping up to @var{n} bytes. For instance,
4313 @option{-falign-functions=32} aligns functions to the next 32-byte
4314 boundary, but @option{-falign-functions=24} would align to the next
4315 32-byte boundary only if this can be done by skipping 23 bytes or less.
4317 @option{-fno-align-functions} and @option{-falign-functions=1} are
4318 equivalent and mean that functions will not be aligned.
4320 Some assemblers only support this flag when @var{n} is a power of two;
4321 in that case, it is rounded up.
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 -falign-labels
4328 @itemx -falign-labels=@var{n}
4329 @opindex falign-labels
4330 Align all branch targets to a power-of-two boundary, skipping up to
4331 @var{n} bytes like @option{-falign-functions}. This option can easily
4332 make code slower, because it must insert dummy operations for when the
4333 branch target is reached in the usual flow of the code.
4335 @option{-fno-align-labels} and @option{-falign-labels=1} are
4336 equivalent and mean that labels will not be aligned.
4338 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4339 are greater than this value, then their values are used instead.
4341 If @var{n} is not specified or is zero, use a machine-dependent default
4342 which is very likely to be @samp{1}, meaning no alignment.
4344 Enabled at levels @option{-O2}, @option{-O3}.
4347 @itemx -falign-loops=@var{n}
4348 @opindex falign-loops
4349 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4350 like @option{-falign-functions}. The hope is that the loop will be
4351 executed many times, which will make up for any execution of the dummy
4354 @option{-fno-align-loops} and @option{-falign-loops=1} are
4355 equivalent and mean that loops will not be aligned.
4357 If @var{n} is not specified or is zero, use a machine-dependent default.
4359 Enabled at levels @option{-O2}, @option{-O3}.
4362 @itemx -falign-jumps=@var{n}
4363 @opindex falign-jumps
4364 Align branch targets to a power-of-two boundary, for branch targets
4365 where the targets can only be reached by jumping, skipping up to @var{n}
4366 bytes like @option{-falign-functions}. In this case, no dummy operations
4369 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
4370 equivalent and mean that loops will not be aligned.
4372 If @var{n} is not specified or is zero, use a machine-dependent default.
4374 Enabled at levels @option{-O2}, @option{-O3}.
4376 @item -frename-registers
4377 @opindex frename-registers
4378 Attempt to avoid false dependencies in scheduled code by making use
4379 of registers left over after register allocation. This optimization
4380 will most benefit processors with lots of registers. It can, however,
4381 make debugging impossible, since variables will no longer stay in
4382 a ``home register''.
4386 Constructs webs as commonly used for register allocation purposes and assign
4387 each web individual pseudo register. This allows our register allocation pass
4388 to operate on pseudos directly, but also strengthens several other optimization
4389 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
4390 however, make debugging impossible, since variables will no longer stay in a
4393 Enabled at levels @option{-O3}.
4395 @item -fno-cprop-registers
4396 @opindex fno-cprop-registers
4397 After register allocation and post-register allocation instruction splitting,
4398 we perform a copy-propagation pass to try to reduce scheduling dependencies
4399 and occasionally eliminate the copy.
4401 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4405 The following options control compiler behavior regarding floating
4406 point arithmetic. These options trade off between speed and
4407 correctness. All must be specifically enabled.
4411 @opindex ffloat-store
4412 Do not store floating point variables in registers, and inhibit other
4413 options that might change whether a floating point value is taken from a
4416 @cindex floating point precision
4417 This option prevents undesirable excess precision on machines such as
4418 the 68000 where the floating registers (of the 68881) keep more
4419 precision than a @code{double} is supposed to have. Similarly for the
4420 x86 architecture. For most programs, the excess precision does only
4421 good, but a few programs rely on the precise definition of IEEE floating
4422 point. Use @option{-ffloat-store} for such programs, after modifying
4423 them to store all pertinent intermediate computations into variables.
4427 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
4428 @option{-fno-trapping-math}, @option{-ffinite-math-only},
4429 @option{-fno-rounding-math} and @option{-fno-signaling-nans}.
4431 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
4433 This option should never be turned on by any @option{-O} option since
4434 it can result in incorrect output for programs which depend on
4435 an exact implementation of IEEE or ISO rules/specifications for
4438 @item -fno-math-errno
4439 @opindex fno-math-errno
4440 Do not set ERRNO after calling math functions that are executed
4441 with a single instruction, e.g., sqrt. A program that relies on
4442 IEEE exceptions for math error handling may want to use this flag
4443 for speed while maintaining IEEE arithmetic compatibility.
4445 This option should never be turned on by any @option{-O} option since
4446 it can result in incorrect output for programs which depend on
4447 an exact implementation of IEEE or ISO rules/specifications for
4450 The default is @option{-fmath-errno}.
4452 @item -funsafe-math-optimizations
4453 @opindex funsafe-math-optimizations
4454 Allow optimizations for floating-point arithmetic that (a) assume
4455 that arguments and results are valid and (b) may violate IEEE or
4456 ANSI standards. When used at link-time, it may include libraries
4457 or startup files that change the default FPU control word or other
4458 similar optimizations.
4460 This option should never be turned on by any @option{-O} option since
4461 it can result in incorrect output for programs which depend on
4462 an exact implementation of IEEE or ISO rules/specifications for
4465 The default is @option{-fno-unsafe-math-optimizations}.
4467 @item -ffinite-math-only
4468 @opindex ffinite-math-only
4469 Allow optimizations for floating-point arithmetic that assume
4470 that arguments and results are not NaNs or +-Infs.
4472 This option should never be turned on by any @option{-O} option since
4473 it can result in incorrect output for programs which depend on
4474 an exact implementation of IEEE or ISO rules/specifications.
4476 The default is @option{-fno-finite-math-only}.
4478 @item -fno-trapping-math
4479 @opindex fno-trapping-math
4480 Compile code assuming that floating-point operations cannot generate
4481 user-visible traps. These traps include division by zero, overflow,
4482 underflow, inexact result and invalid operation. This option implies
4483 @option{-fno-signaling-nans}. Setting this option may allow faster
4484 code if one relies on ``non-stop'' IEEE arithmetic, for example.
4486 This option should never be turned on by any @option{-O} option since
4487 it can result in incorrect output for programs which depend on
4488 an exact implementation of IEEE or ISO rules/specifications for
4491 The default is @option{-ftrapping-math}.
4493 @item -frounding-math
4494 @opindex frounding-math
4495 Disable transformations and optimizations that assume default floating
4496 point rounding behavior. This is round-to-zero for all floating point
4497 to integer conversions, and round-to-nearest for all other arithmetic
4498 truncations. This option should be specified for programs that change
4499 the FP rounding mode dynamically, or that may be executed with a
4500 non-default rounding mode. This option disables constant folding of
4501 floating point expressions at compile-time (which may be affected by
4502 rounding mode) and arithmetic transformations that are unsafe in the
4503 presence of sign-dependent rounding modes.
4505 The default is @option{-fno-rounding-math}.
4507 This option is experimental and does not currently guarantee to
4508 disable all GCC optimizations that are affected by rounding mode.
4509 Future versions of gcc may provide finer control of this setting
4510 using C99's @code{FENV_ACCESS} pragma. This command line option
4511 will be used to specify the default state for @code{FENV_ACCESS}.
4513 @item -fsignaling-nans
4514 @opindex fsignaling-nans
4515 Compile code assuming that IEEE signaling NaNs may generate user-visible
4516 traps during floating-point operations. Setting this option disables
4517 optimizations that may change the number of exceptions visible with
4518 signaling NaNs. This option implies @option{-ftrapping-math}.
4520 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
4523 The default is @option{-fno-signaling-nans}.
4525 This option is experimental and does not currently guarantee to
4526 disable all GCC optimizations that affect signaling NaN behavior.
4528 @item -fsingle-precision-constant
4529 @opindex fsingle-precision-constant
4530 Treat floating point constant as single precision constant instead of
4531 implicitly converting it to double precision constant.
4536 The following options control optimizations that may improve
4537 performance, but are not enabled by any @option{-O} options. This
4538 section includes experimental options that may produce broken code.
4541 @item -fbranch-probabilities
4542 @opindex fbranch-probabilities
4543 After running a program compiled with @option{-fprofile-arcs}
4544 (@pxref{Debugging Options,, Options for Debugging Your Program or
4545 @command{gcc}}), you can compile it a second time using
4546 @option{-fbranch-probabilities}, to improve optimizations based on
4547 the number of times each branch was taken. When the program
4548 compiled with @option{-fprofile-arcs} exits it saves arc execution
4549 counts to a file called @file{@var{sourcename}.gcda} for each source
4550 file The information in this data file is very dependent on the
4551 structure of the generated code, so you must use the same source code
4552 and the same optimization options for both compilations.
4554 With @option{-fbranch-probabilities}, GCC puts a
4555 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
4556 These can be used to improve optimization. Currently, they are only
4557 used in one place: in @file{reorg.c}, instead of guessing which path a
4558 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
4559 exactly determine which path is taken more often.
4561 @item -fprofile-values
4562 @opindex fprofile-values
4563 If combined with @option{-fprofile-arcs}, it adds code so that some
4564 data about values of expressions in the program is gathered.
4566 With @option{-fbranch-probabilities}, it reads back the data gathered
4567 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
4568 notes to instructions for their later usage in optimizations.
4572 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
4573 a code to gather information about values of expressions.
4575 With @option{-fbranch-probabilities}, it reads back the data gathered
4576 and actually performs the optimizations based on them.
4577 Currently the optimizations include specialization of division operation
4578 using the knowledge about the value of the denominator.
4582 Use a graph coloring register allocator. Currently this option is meant
4583 for testing, so we are interested to hear about miscompilations with
4588 Perform tail duplication to enlarge superblock size. This transformation
4589 simplifies the control flow of the function allowing other optimizations to do
4592 @item -funit-at-a-time
4593 @opindex funit-at-a-time
4594 Parse the whole compilation unit before starting to produce code.
4595 This allows some extra optimizations to take place but consumes more
4598 @item -funroll-loops
4599 @opindex funroll-loops
4600 Unroll loops whose number of iterations can be determined at compile time or
4601 upon entry to the loop. @option{-funroll-loops} implies
4602 @option{-frerun-cse-after-loop}. It also turns on complete loop peeling
4603 (i.e. complete removal of loops with small constant number of iterations).
4604 This option makes code larger, and may or may not make it run faster.
4606 @item -funroll-all-loops
4607 @opindex funroll-all-loops
4608 Unroll all loops, even if their number of iterations is uncertain when
4609 the loop is entered. This usually makes programs run more slowly.
4610 @option{-funroll-all-loops} implies the same options as
4611 @option{-funroll-loops}.
4614 @opindex fpeel-loops
4615 Peels the loops for that there is enough information that they do not
4616 roll much (from profile feedback). It also turns on complete loop peeling
4617 (i.e. complete removal of loops with small constant number of iterations).
4619 @item -funswitch-loops
4620 @opindex funswitch-loops
4621 Move branches with loop invariant conditions out of the loop, with duplicates
4622 of the loop on both branches (modified according to result of the condition).
4624 @item -fold-unroll-loops
4625 @opindex fold-unroll-loops
4626 Unroll loops whose number of iterations can be determined at compile
4627 time or upon entry to the loop, using the old loop unroller whose loop
4628 recognition is based on notes from frontend. @option{-fold-unroll-loops} implies
4629 both @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
4630 option makes code larger, and may or may not make it run faster.
4632 @item -fold-unroll-all-loops
4633 @opindex fold-unroll-all-loops
4634 Unroll all loops, even if their number of iterations is uncertain when
4635 the loop is entered. This is done using the old loop unroller whose loop
4636 recognition is based on notes from frontend. This usually makes programs run more slowly.
4637 @option{-fold-unroll-all-loops} implies the same options as
4638 @option{-fold-unroll-loops}.
4640 @item -funswitch-loops
4641 @opindex funswitch-loops
4642 Move branches with loop invariant conditions out of the loop, with duplicates
4643 of the loop on both branches (modified according to result of the condition).
4645 @item -funswitch-loops
4646 @opindex funswitch-loops
4647 Move branches with loop invariant conditions out of the loop, with duplicates
4648 of the loop on both branches (modified according to result of the condition).
4650 @item -fprefetch-loop-arrays
4651 @opindex fprefetch-loop-arrays
4652 If supported by the target machine, generate instructions to prefetch
4653 memory to improve the performance of loops that access large arrays.
4655 Disabled at level @option{-Os}.
4657 @item -ffunction-sections
4658 @itemx -fdata-sections
4659 @opindex ffunction-sections
4660 @opindex fdata-sections
4661 Place each function or data item into its own section in the output
4662 file if the target supports arbitrary sections. The name of the
4663 function or the name of the data item determines the section's name
4666 Use these options on systems where the linker can perform optimizations
4667 to improve locality of reference in the instruction space. Most systems
4668 using the ELF object format and SPARC processors running Solaris 2 have
4669 linkers with such optimizations. AIX may have these optimizations in
4672 Only use these options when there are significant benefits from doing
4673 so. When you specify these options, the assembler and linker will
4674 create larger object and executable files and will also be slower.
4675 You will not be able to use @code{gprof} on all systems if you
4676 specify this option and you may have problems with debugging if
4677 you specify both this option and @option{-g}.
4681 Perform optimizations in static single assignment form. Each function's
4682 flow graph is translated into SSA form, optimizations are performed, and
4683 the flow graph is translated back from SSA form. Users should not
4684 specify this option, since it is not yet ready for production use.
4688 Perform Sparse Conditional Constant Propagation in SSA form. Requires
4689 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
4693 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
4694 Like @option{-fssa}, this is an experimental feature.
4696 @item -fbranch-target-load-optimize
4697 @opindex fbranch-target-load-optimize
4698 Perform branch target register load optimization before prologue / epilogue
4700 The use of target registers can typically be exposed only during reload,
4701 thus hoisting loads out of loops and doing inter-block scheduling needs
4702 a separate optimization pass.
4704 @item -fbranch-target-load-optimize2
4705 @opindex fbranch-target-load-optimize2
4706 Perform branch target register load optimization after prologue / epilogue
4712 @item --param @var{name}=@var{value}
4714 In some places, GCC uses various constants to control the amount of
4715 optimization that is done. For example, GCC will not inline functions
4716 that contain more that a certain number of instructions. You can
4717 control some of these constants on the command-line using the
4718 @option{--param} option.
4720 In each case, the @var{value} is an integer. The allowable choices for
4721 @var{name} are given in the following table:
4724 @item max-crossjump-edges
4725 The maximum number of incoming edges to consider for crossjumping.
4726 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
4727 the number of edges incoming to each block. Increasing values mean
4728 more aggressive optimization, making the compile time increase with
4729 probably small improvement in executable size.
4731 @item max-delay-slot-insn-search
4732 The maximum number of instructions to consider when looking for an
4733 instruction to fill a delay slot. If more than this arbitrary number of
4734 instructions is searched, the time savings from filling the delay slot
4735 will be minimal so stop searching. Increasing values mean more
4736 aggressive optimization, making the compile time increase with probably
4737 small improvement in executable run time.
4739 @item max-delay-slot-live-search
4740 When trying to fill delay slots, the maximum number of instructions to
4741 consider when searching for a block with valid live register
4742 information. Increasing this arbitrarily chosen value means more
4743 aggressive optimization, increasing the compile time. This parameter
4744 should be removed when the delay slot code is rewritten to maintain the
4747 @item max-gcse-memory
4748 The approximate maximum amount of memory that will be allocated in
4749 order to perform the global common subexpression elimination
4750 optimization. If more memory than specified is required, the
4751 optimization will not be done.
4753 @item max-gcse-passes
4754 The maximum number of passes of GCSE to run.
4756 @item max-pending-list-length
4757 The maximum number of pending dependencies scheduling will allow
4758 before flushing the current state and starting over. Large functions
4759 with few branches or calls can create excessively large lists which
4760 needlessly consume memory and resources.
4762 @item max-inline-insns-single
4763 Several parameters control the tree inliner used in gcc.
4764 This number sets the maximum number of instructions (counted in gcc's
4765 internal representation) in a single function that the tree inliner
4766 will consider for inlining. This only affects functions declared
4767 inline and methods implemented in a class declaration (C++).
4768 The default value is 500.
4770 @item max-inline-insns-auto
4771 When you use @option{-finline-functions} (included in @option{-O3}),
4772 a lot of functions that would otherwise not be considered for inlining
4773 by the compiler will be investigated. To those functions, a different
4774 (more restrictive) limit compared to functions declared inline can
4776 The default value is 150.
4778 @item max-inline-insns
4779 The tree inliner does decrease the allowable size for single functions
4780 to be inlined after we already inlined the number of instructions
4781 given here by repeated inlining. This number should be a factor of
4782 two or more larger than the single function limit.
4783 Higher numbers result in better runtime performance, but incur higher
4784 compile-time resource (CPU time, memory) requirements and result in
4785 larger binaries. Very high values are not advisable, as too large
4786 binaries may adversely affect runtime performance.
4787 The default value is 200.
4789 @item large-function-insns
4790 The limit specifying really large functions. For functions greater than this
4791 limit inlining is constrained by @option{--param large-function-growth}.
4792 This parameter is useful primarily to avoid extreme compilation time caused by non-linear
4793 algorithms used by the backend.
4794 This parameter is ignored when @option{-funit-at-a-time} is not used.
4795 The default value is 30000.
4797 @item large-function-growth
4798 Specifies maximal growth of large functtion caused by inlining in percents.
4799 This parameter is ignored when @option{-funit-at-a-time} is not used.
4800 The default value is 200.
4802 @item inline-unit-growth
4803 Specifies maximal overall growth of the compilation unit caused by inlining.
4804 This parameter is ignored when @option{-funit-at-a-time} is not used.
4805 The default value is 150.
4807 @item max-inline-insns-rtl
4808 For languages that use the RTL inliner (this happens at a later stage
4809 than tree inlining), you can set the maximum allowable size (counted
4810 in RTL instructions) for the RTL inliner with this parameter.
4811 The default value is 600.
4814 @item max-unrolled-insns
4815 The maximum number of instructions that a loop should have if that loop
4816 is unrolled, and if the loop is unrolled, it determines how many times
4817 the loop code is unrolled.
4819 @item max-average-unrolled-insns
4820 The maximum number of instructions biased by probabilities of their execution
4821 that a loop should have if that loop is unrolled, and if the loop is unrolled,
4822 it determines how many times the loop code is unrolled.
4824 @item max-unroll-times
4825 The maximum number of unrollings of a single loop.
4827 @item max-peeled-insns
4828 The maximum number of instructions that a loop should have if that loop
4829 is peeled, and if the loop is peeled, it determines how many times
4830 the loop code is peeled.
4832 @item max-peel-times
4833 The maximum number of peelings of a single loop.
4835 @item max-completely-peeled-insns
4836 The maximum number of insns of a completely peeled loop.
4838 @item max-completely-peel-times
4839 The maximum number of iterations of a loop to be suitable for complete peeling.
4841 @item max-unswitch-insns
4842 The maximum number of insns of an unswitched loop.
4844 @item max-unswitch-level
4845 The maximum number of branches unswitched in a single loop.
4847 @item hot-bb-count-fraction
4848 Select fraction of the maximal count of repetitions of basic block in program
4849 given basic block needs to have to be considered hot.
4851 @item hot-bb-frequency-fraction
4852 Select fraction of the maximal frequency of executions of basic block in
4853 function given basic block needs to have to be considered hot
4855 @item tracer-dynamic-coverage
4856 @itemx tracer-dynamic-coverage-feedback
4858 This value is used to limit superblock formation once the given percentage of
4859 executed instructions is covered. This limits unnecessary code size
4862 The @option{tracer-dynamic-coverage-feedback} is used only when profile
4863 feedback is available. The real profiles (as opposed to statically estimated
4864 ones) are much less balanced allowing the threshold to be larger value.
4866 @item tracer-max-code-growth
4867 Stop tail duplication once code growth has reached given percentage. This is
4868 rather hokey argument, as most of the duplicates will be eliminated later in
4869 cross jumping, so it may be set to much higher values than is the desired code
4872 @item tracer-min-branch-ratio
4874 Stop reverse growth when the reverse probability of best edge is less than this
4875 threshold (in percent).
4877 @item tracer-min-branch-ratio
4878 @itemx tracer-min-branch-ratio-feedback
4880 Stop forward growth if the best edge do have probability lower than this
4883 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
4884 compilation for profile feedback and one for compilation without. The value
4885 for compilation with profile feedback needs to be more conservative (higher) in
4886 order to make tracer effective.
4888 @item max-cse-path-length
4890 Maximum number of basic blocks on path that cse considers.
4892 @item ggc-min-expand
4894 GCC uses a garbage collector to manage its own memory allocation. This
4895 parameter specifies the minimum percentage by which the garbage
4896 collector's heap should be allowed to expand between collections.
4897 Tuning this may improve compilation speed; it has no effect on code
4900 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
4901 RAM >= 1GB. If @code{getrlimit} is available, the notion of "RAM" is
4902 the smallest of actual RAM, RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If
4903 GCC is not able to calculate RAM on a particular platform, the lower
4904 bound of 30% is used. Setting this parameter and
4905 @option{ggc-min-heapsize} to zero causes a full collection to occur at
4906 every opportunity. This is extremely slow, but can be useful for
4909 @item ggc-min-heapsize
4911 Minimum size of the garbage collector's heap before it begins bothering
4912 to collect garbage. The first collection occurs after the heap expands
4913 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
4914 tuning this may improve compilation speed, and has no effect on code
4917 The default is RAM/8, with a lower bound of 4096 (four megabytes) and an
4918 upper bound of 131072 (128 megabytes). If @code{getrlimit} is
4919 available, the notion of "RAM" is the smallest of actual RAM,
4920 RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If GCC is not able to calculate
4921 RAM on a particular platform, the lower bound is used. Setting this
4922 parameter very large effectively disables garbage collection. Setting
4923 this parameter and @option{ggc-min-expand} to zero causes a full
4924 collection to occur at every opportunity.
4926 @item reorder-blocks-duplicate
4927 @itemx reorder-blocks-duplicate-feedback
4929 Used by basic block reordering pass to decide whether to use unconditional
4930 branch or duplicate the code on its destination. Code is duplicated when its
4931 estimated size is smaller than this value multiplied by the estimated size of
4932 unconditional jump in the hot spots of the program.
4934 The @option{reorder-block-duplicate-feedback} is used only when profile
4935 feedback is available and may be set to higher values than
4936 @option{reorder-block-duplicate} since information about the hot spots is more
4941 @node Preprocessor Options
4942 @section Options Controlling the Preprocessor
4943 @cindex preprocessor options
4944 @cindex options, preprocessor
4946 These options control the C preprocessor, which is run on each C source
4947 file before actual compilation.
4949 If you use the @option{-E} option, nothing is done except preprocessing.
4950 Some of these options make sense only together with @option{-E} because
4951 they cause the preprocessor output to be unsuitable for actual
4956 You can use @option{-Wp,@var{option}} to bypass the compiler driver
4957 and pass @var{option} directly through to the preprocessor. If
4958 @var{option} contains commas, it is split into multiple options at the
4959 commas. However, many options are modified, translated or interpreted
4960 by the compiler driver before being passed to the preprocessor, and
4961 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
4962 interface is undocumented and subject to change, so whenever possible
4963 you should avoid using @option{-Wp} and let the driver handle the
4966 @item -Xpreprocessor @var{option}
4967 @opindex preprocessor
4968 Pass @var{option} as an option to the preprocessor. You can use this to
4969 supply system-specific preprocessor options which GCC does not know how to
4972 If you want to pass an option that takes an argument, you must use
4973 @option{-Xpreprocessor} twice, once for the option and once for the argument.
4976 @include cppopts.texi
4978 @node Assembler Options
4979 @section Passing Options to the Assembler
4981 @c prevent bad page break with this line
4982 You can pass options to the assembler.
4985 @item -Wa,@var{option}
4987 Pass @var{option} as an option to the assembler. If @var{option}
4988 contains commas, it is split into multiple options at the commas.
4990 @item -Xassembler @var{option}
4992 Pass @var{option} as an option to the assembler. You can use this to
4993 supply system-specific assembler options which GCC does not know how to
4996 If you want to pass an option that takes an argument, you must use
4997 @option{-Xassembler} twice, once for the option and once for the argument.
5002 @section Options for Linking
5003 @cindex link options
5004 @cindex options, linking
5006 These options come into play when the compiler links object files into
5007 an executable output file. They are meaningless if the compiler is
5008 not doing a link step.
5012 @item @var{object-file-name}
5013 A file name that does not end in a special recognized suffix is
5014 considered to name an object file or library. (Object files are
5015 distinguished from libraries by the linker according to the file
5016 contents.) If linking is done, these object files are used as input
5025 If any of these options is used, then the linker is not run, and
5026 object file names should not be used as arguments. @xref{Overall
5030 @item -l@var{library}
5031 @itemx -l @var{library}
5033 Search the library named @var{library} when linking. (The second
5034 alternative with the library as a separate argument is only for
5035 POSIX compliance and is not recommended.)
5037 It makes a difference where in the command you write this option; the
5038 linker searches and processes libraries and object files in the order they
5039 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
5040 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
5041 to functions in @samp{z}, those functions may not be loaded.
5043 The linker searches a standard list of directories for the library,
5044 which is actually a file named @file{lib@var{library}.a}. The linker
5045 then uses this file as if it had been specified precisely by name.
5047 The directories searched include several standard system directories
5048 plus any that you specify with @option{-L}.
5050 Normally the files found this way are library files---archive files
5051 whose members are object files. The linker handles an archive file by
5052 scanning through it for members which define symbols that have so far
5053 been referenced but not defined. But if the file that is found is an
5054 ordinary object file, it is linked in the usual fashion. The only
5055 difference between using an @option{-l} option and specifying a file name
5056 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
5057 and searches several directories.
5061 You need this special case of the @option{-l} option in order to
5062 link an Objective-C program.
5065 @opindex nostartfiles
5066 Do not use the standard system startup files when linking.
5067 The standard system libraries are used normally, unless @option{-nostdlib}
5068 or @option{-nodefaultlibs} is used.
5070 @item -nodefaultlibs
5071 @opindex nodefaultlibs
5072 Do not use the standard system libraries when linking.
5073 Only the libraries you specify will be passed to the linker.
5074 The standard startup files are used normally, unless @option{-nostartfiles}
5075 is used. The compiler may generate calls to memcmp, memset, and memcpy
5076 for System V (and ISO C) environments or to bcopy and bzero for
5077 BSD environments. These entries are usually resolved by entries in
5078 libc. These entry points should be supplied through some other
5079 mechanism when this option is specified.
5083 Do not use the standard system startup files or libraries when linking.
5084 No startup files and only the libraries you specify will be passed to
5085 the linker. The compiler may generate calls to memcmp, memset, and memcpy
5086 for System V (and ISO C) environments or to bcopy and bzero for
5087 BSD environments. These entries are usually resolved by entries in
5088 libc. These entry points should be supplied through some other
5089 mechanism when this option is specified.
5091 @cindex @option{-lgcc}, use with @option{-nostdlib}
5092 @cindex @option{-nostdlib} and unresolved references
5093 @cindex unresolved references and @option{-nostdlib}
5094 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
5095 @cindex @option{-nodefaultlibs} and unresolved references
5096 @cindex unresolved references and @option{-nodefaultlibs}
5097 One of the standard libraries bypassed by @option{-nostdlib} and
5098 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
5099 that GCC uses to overcome shortcomings of particular machines, or special
5100 needs for some languages.
5101 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
5102 Collection (GCC) Internals},
5103 for more discussion of @file{libgcc.a}.)
5104 In most cases, you need @file{libgcc.a} even when you want to avoid
5105 other standard libraries. In other words, when you specify @option{-nostdlib}
5106 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
5107 This ensures that you have no unresolved references to internal GCC
5108 library subroutines. (For example, @samp{__main}, used to ensure C++
5109 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
5110 GNU Compiler Collection (GCC) Internals}.)
5114 Produce a position independent executable on targets which support it.
5115 For predictable results, you must also specify the same set of options
5116 that were used to generate code (@option{-fpie}, @option{-fPIE},
5117 or model suboptions) when you specify this option.
5121 Remove all symbol table and relocation information from the executable.
5125 On systems that support dynamic linking, this prevents linking with the shared
5126 libraries. On other systems, this option has no effect.
5130 Produce a shared object which can then be linked with other objects to
5131 form an executable. Not all systems support this option. For predictable
5132 results, you must also specify the same set of options that were used to
5133 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
5134 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
5135 needs to build supplementary stub code for constructors to work. On
5136 multi-libbed systems, @samp{gcc -shared} must select the correct support
5137 libraries to link against. Failing to supply the correct flags may lead
5138 to subtle defects. Supplying them in cases where they are not necessary
5141 @item -shared-libgcc
5142 @itemx -static-libgcc
5143 @opindex shared-libgcc
5144 @opindex static-libgcc
5145 On systems that provide @file{libgcc} as a shared library, these options
5146 force the use of either the shared or static version respectively.
5147 If no shared version of @file{libgcc} was built when the compiler was
5148 configured, these options have no effect.
5150 There are several situations in which an application should use the
5151 shared @file{libgcc} instead of the static version. The most common
5152 of these is when the application wishes to throw and catch exceptions
5153 across different shared libraries. In that case, each of the libraries
5154 as well as the application itself should use the shared @file{libgcc}.
5156 Therefore, the G++ and GCJ drivers automatically add
5157 @option{-shared-libgcc} whenever you build a shared library or a main
5158 executable, because C++ and Java programs typically use exceptions, so
5159 this is the right thing to do.
5161 If, instead, you use the GCC driver to create shared libraries, you may
5162 find that they will not always be linked with the shared @file{libgcc}.
5163 If GCC finds, at its configuration time, that you have a GNU linker that
5164 does not support option @option{--eh-frame-hdr}, it will link the shared
5165 version of @file{libgcc} into shared libraries by default. Otherwise,
5166 it will take advantage of the linker and optimize away the linking with
5167 the shared version of @file{libgcc}, linking with the static version of
5168 libgcc by default. This allows exceptions to propagate through such
5169 shared libraries, without incurring relocation costs at library load
5172 However, if a library or main executable is supposed to throw or catch
5173 exceptions, you must link it using the G++ or GCJ driver, as appropriate
5174 for the languages used in the program, or using the option
5175 @option{-shared-libgcc}, such that it is linked with the shared
5180 Bind references to global symbols when building a shared object. Warn
5181 about any unresolved references (unless overridden by the link editor
5182 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
5185 @item -Xlinker @var{option}
5187 Pass @var{option} as an option to the linker. You can use this to
5188 supply system-specific linker options which GCC does not know how to
5191 If you want to pass an option that takes an argument, you must use
5192 @option{-Xlinker} twice, once for the option and once for the argument.
5193 For example, to pass @option{-assert definitions}, you must write
5194 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
5195 @option{-Xlinker "-assert definitions"}, because this passes the entire
5196 string as a single argument, which is not what the linker expects.
5198 @item -Wl,@var{option}
5200 Pass @var{option} as an option to the linker. If @var{option} contains
5201 commas, it is split into multiple options at the commas.
5203 @item -u @var{symbol}
5205 Pretend the symbol @var{symbol} is undefined, to force linking of
5206 library modules to define it. You can use @option{-u} multiple times with
5207 different symbols to force loading of additional library modules.
5210 @node Directory Options
5211 @section Options for Directory Search
5212 @cindex directory options
5213 @cindex options, directory search
5216 These options specify directories to search for header files, for
5217 libraries and for parts of the compiler:
5222 Add the directory @var{dir} to the head of the list of directories to be
5223 searched for header files. This can be used to override a system header
5224 file, substituting your own version, since these directories are
5225 searched before the system header file directories. However, you should
5226 not use this option to add directories that contain vendor-supplied
5227 system header files (use @option{-isystem} for that). If you use more than
5228 one @option{-I} option, the directories are scanned in left-to-right
5229 order; the standard system directories come after.
5231 If a standard system include directory, or a directory specified with
5232 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
5233 option will be ignored. The directory will still be searched but as a
5234 system directory at its normal position in the system include chain.
5235 This is to ensure that GCC's procedure to fix buggy system headers and
5236 the ordering for the include_next directive are not inadvertently changed.
5237 If you really need to change the search order for system directories,
5238 use the @option{-nostdinc} and/or @option{-isystem} options.
5242 Any directories you specify with @option{-I} options before the @option{-I-}
5243 option are searched only for the case of @samp{#include "@var{file}"};
5244 they are not searched for @samp{#include <@var{file}>}.
5246 If additional directories are specified with @option{-I} options after
5247 the @option{-I-}, these directories are searched for all @samp{#include}
5248 directives. (Ordinarily @emph{all} @option{-I} directories are used
5251 In addition, the @option{-I-} option inhibits the use of the current
5252 directory (where the current input file came from) as the first search
5253 directory for @samp{#include "@var{file}"}. There is no way to
5254 override this effect of @option{-I-}. With @option{-I.} you can specify
5255 searching the directory which was current when the compiler was
5256 invoked. That is not exactly the same as what the preprocessor does
5257 by default, but it is often satisfactory.
5259 @option{-I-} does not inhibit the use of the standard system directories
5260 for header files. Thus, @option{-I-} and @option{-nostdinc} are
5265 Add directory @var{dir} to the list of directories to be searched
5268 @item -B@var{prefix}
5270 This option specifies where to find the executables, libraries,
5271 include files, and data files of the compiler itself.
5273 The compiler driver program runs one or more of the subprograms
5274 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
5275 @var{prefix} as a prefix for each program it tries to run, both with and
5276 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
5278 For each subprogram to be run, the compiler driver first tries the
5279 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
5280 was not specified, the driver tries two standard prefixes, which are
5281 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
5282 those results in a file name that is found, the unmodified program
5283 name is searched for using the directories specified in your
5284 @env{PATH} environment variable.
5286 The compiler will check to see if the path provided by the @option{-B}
5287 refers to a directory, and if necessary it will add a directory
5288 separator character at the end of the path.
5290 @option{-B} prefixes that effectively specify directory names also apply
5291 to libraries in the linker, because the compiler translates these
5292 options into @option{-L} options for the linker. They also apply to
5293 includes files in the preprocessor, because the compiler translates these
5294 options into @option{-isystem} options for the preprocessor. In this case,
5295 the compiler appends @samp{include} to the prefix.
5297 The run-time support file @file{libgcc.a} can also be searched for using
5298 the @option{-B} prefix, if needed. If it is not found there, the two
5299 standard prefixes above are tried, and that is all. The file is left
5300 out of the link if it is not found by those means.
5302 Another way to specify a prefix much like the @option{-B} prefix is to use
5303 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
5306 As a special kludge, if the path provided by @option{-B} is
5307 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
5308 9, then it will be replaced by @file{[dir/]include}. This is to help
5309 with boot-strapping the compiler.
5311 @item -specs=@var{file}
5313 Process @var{file} after the compiler reads in the standard @file{specs}
5314 file, in order to override the defaults that the @file{gcc} driver
5315 program uses when determining what switches to pass to @file{cc1},
5316 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
5317 @option{-specs=@var{file}} can be specified on the command line, and they
5318 are processed in order, from left to right.
5324 @section Specifying subprocesses and the switches to pass to them
5327 @command{gcc} is a driver program. It performs its job by invoking a
5328 sequence of other programs to do the work of compiling, assembling and
5329 linking. GCC interprets its command-line parameters and uses these to
5330 deduce which programs it should invoke, and which command-line options
5331 it ought to place on their command lines. This behavior is controlled
5332 by @dfn{spec strings}. In most cases there is one spec string for each
5333 program that GCC can invoke, but a few programs have multiple spec
5334 strings to control their behavior. The spec strings built into GCC can
5335 be overridden by using the @option{-specs=} command-line switch to specify
5338 @dfn{Spec files} are plaintext files that are used to construct spec
5339 strings. They consist of a sequence of directives separated by blank
5340 lines. The type of directive is determined by the first non-whitespace
5341 character on the line and it can be one of the following:
5344 @item %@var{command}
5345 Issues a @var{command} to the spec file processor. The commands that can
5349 @item %include <@var{file}>
5351 Search for @var{file} and insert its text at the current point in the
5354 @item %include_noerr <@var{file}>
5355 @cindex %include_noerr
5356 Just like @samp{%include}, but do not generate an error message if the include
5357 file cannot be found.
5359 @item %rename @var{old_name} @var{new_name}
5361 Rename the spec string @var{old_name} to @var{new_name}.
5365 @item *[@var{spec_name}]:
5366 This tells the compiler to create, override or delete the named spec
5367 string. All lines after this directive up to the next directive or
5368 blank line are considered to be the text for the spec string. If this
5369 results in an empty string then the spec will be deleted. (Or, if the
5370 spec did not exist, then nothing will happened.) Otherwise, if the spec
5371 does not currently exist a new spec will be created. If the spec does
5372 exist then its contents will be overridden by the text of this
5373 directive, unless the first character of that text is the @samp{+}
5374 character, in which case the text will be appended to the spec.
5376 @item [@var{suffix}]:
5377 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
5378 and up to the next directive or blank line are considered to make up the
5379 spec string for the indicated suffix. When the compiler encounters an
5380 input file with the named suffix, it will processes the spec string in
5381 order to work out how to compile that file. For example:
5388 This says that any input file whose name ends in @samp{.ZZ} should be
5389 passed to the program @samp{z-compile}, which should be invoked with the
5390 command-line switch @option{-input} and with the result of performing the
5391 @samp{%i} substitution. (See below.)
5393 As an alternative to providing a spec string, the text that follows a
5394 suffix directive can be one of the following:
5397 @item @@@var{language}
5398 This says that the suffix is an alias for a known @var{language}. This is
5399 similar to using the @option{-x} command-line switch to GCC to specify a
5400 language explicitly. For example:
5407 Says that .ZZ files are, in fact, C++ source files.
5410 This causes an error messages saying:
5413 @var{name} compiler not installed on this system.
5417 GCC already has an extensive list of suffixes built into it.
5418 This directive will add an entry to the end of the list of suffixes, but
5419 since the list is searched from the end backwards, it is effectively
5420 possible to override earlier entries using this technique.
5424 GCC has the following spec strings built into it. Spec files can
5425 override these strings or create their own. Note that individual
5426 targets can also add their own spec strings to this list.
5429 asm Options to pass to the assembler
5430 asm_final Options to pass to the assembler post-processor
5431 cpp Options to pass to the C preprocessor
5432 cc1 Options to pass to the C compiler
5433 cc1plus Options to pass to the C++ compiler
5434 endfile Object files to include at the end of the link
5435 link Options to pass to the linker
5436 lib Libraries to include on the command line to the linker
5437 libgcc Decides which GCC support library to pass to the linker
5438 linker Sets the name of the linker
5439 predefines Defines to be passed to the C preprocessor
5440 signed_char Defines to pass to CPP to say whether @code{char} is signed
5442 startfile Object files to include at the start of the link
5445 Here is a small example of a spec file:
5451 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
5454 This example renames the spec called @samp{lib} to @samp{old_lib} and
5455 then overrides the previous definition of @samp{lib} with a new one.
5456 The new definition adds in some extra command-line options before
5457 including the text of the old definition.
5459 @dfn{Spec strings} are a list of command-line options to be passed to their
5460 corresponding program. In addition, the spec strings can contain
5461 @samp{%}-prefixed sequences to substitute variable text or to
5462 conditionally insert text into the command line. Using these constructs
5463 it is possible to generate quite complex command lines.
5465 Here is a table of all defined @samp{%}-sequences for spec
5466 strings. Note that spaces are not generated automatically around the
5467 results of expanding these sequences. Therefore you can concatenate them
5468 together or combine them with constant text in a single argument.
5472 Substitute one @samp{%} into the program name or argument.
5475 Substitute the name of the input file being processed.
5478 Substitute the basename of the input file being processed.
5479 This is the substring up to (and not including) the last period
5480 and not including the directory.
5483 This is the same as @samp{%b}, but include the file suffix (text after
5487 Marks the argument containing or following the @samp{%d} as a
5488 temporary file name, so that that file will be deleted if GCC exits
5489 successfully. Unlike @samp{%g}, this contributes no text to the
5492 @item %g@var{suffix}
5493 Substitute a file name that has suffix @var{suffix} and is chosen
5494 once per compilation, and mark the argument in the same way as
5495 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
5496 name is now chosen in a way that is hard to predict even when previously
5497 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
5498 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
5499 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
5500 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
5501 was simply substituted with a file name chosen once per compilation,
5502 without regard to any appended suffix (which was therefore treated
5503 just like ordinary text), making such attacks more likely to succeed.
5505 @item %u@var{suffix}
5506 Like @samp{%g}, but generates a new temporary file name even if
5507 @samp{%u@var{suffix}} was already seen.
5509 @item %U@var{suffix}
5510 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
5511 new one if there is no such last file name. In the absence of any
5512 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
5513 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
5514 would involve the generation of two distinct file names, one
5515 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
5516 simply substituted with a file name chosen for the previous @samp{%u},
5517 without regard to any appended suffix.
5519 @item %j@var{suffix}
5520 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
5521 writable, and if save-temps is off; otherwise, substitute the name
5522 of a temporary file, just like @samp{%u}. This temporary file is not
5523 meant for communication between processes, but rather as a junk
5526 @item %|@var{suffix}
5527 @itemx %m@var{suffix}
5528 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
5529 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
5530 all. These are the two most common ways to instruct a program that it
5531 should read from standard input or write to standard output. If you
5532 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
5533 construct: see for example @file{f/lang-specs.h}.
5535 @item %.@var{SUFFIX}
5536 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
5537 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
5538 terminated by the next space or %.
5541 Marks the argument containing or following the @samp{%w} as the
5542 designated output file of this compilation. This puts the argument
5543 into the sequence of arguments that @samp{%o} will substitute later.
5546 Substitutes the names of all the output files, with spaces
5547 automatically placed around them. You should write spaces
5548 around the @samp{%o} as well or the results are undefined.
5549 @samp{%o} is for use in the specs for running the linker.
5550 Input files whose names have no recognized suffix are not compiled
5551 at all, but they are included among the output files, so they will
5555 Substitutes the suffix for object files. Note that this is
5556 handled specially when it immediately follows @samp{%g, %u, or %U},
5557 because of the need for those to form complete file names. The
5558 handling is such that @samp{%O} is treated exactly as if it had already
5559 been substituted, except that @samp{%g, %u, and %U} do not currently
5560 support additional @var{suffix} characters following @samp{%O} as they would
5561 following, for example, @samp{.o}.
5564 Substitutes the standard macro predefinitions for the
5565 current target machine. Use this when running @code{cpp}.
5568 Like @samp{%p}, but puts @samp{__} before and after the name of each
5569 predefined macro, except for macros that start with @samp{__} or with
5570 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
5574 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
5575 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), and
5576 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
5580 Current argument is the name of a library or startup file of some sort.
5581 Search for that file in a standard list of directories and substitute
5582 the full name found.
5585 Print @var{str} as an error message. @var{str} is terminated by a newline.
5586 Use this when inconsistent options are detected.
5589 Substitute the contents of spec string @var{name} at this point.
5592 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
5594 @item %x@{@var{option}@}
5595 Accumulate an option for @samp{%X}.
5598 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
5602 Output the accumulated assembler options specified by @option{-Wa}.
5605 Output the accumulated preprocessor options specified by @option{-Wp}.
5608 Process the @code{asm} spec. This is used to compute the
5609 switches to be passed to the assembler.
5612 Process the @code{asm_final} spec. This is a spec string for
5613 passing switches to an assembler post-processor, if such a program is
5617 Process the @code{link} spec. This is the spec for computing the
5618 command line passed to the linker. Typically it will make use of the
5619 @samp{%L %G %S %D and %E} sequences.
5622 Dump out a @option{-L} option for each directory that GCC believes might
5623 contain startup files. If the target supports multilibs then the
5624 current multilib directory will be prepended to each of these paths.
5627 Output the multilib directory with directory separators replaced with
5628 @samp{_}. If multilib directories are not set, or the multilib directory is
5629 @file{.} then this option emits nothing.
5632 Process the @code{lib} spec. This is a spec string for deciding which
5633 libraries should be included on the command line to the linker.
5636 Process the @code{libgcc} spec. This is a spec string for deciding
5637 which GCC support library should be included on the command line to the linker.
5640 Process the @code{startfile} spec. This is a spec for deciding which
5641 object files should be the first ones passed to the linker. Typically
5642 this might be a file named @file{crt0.o}.
5645 Process the @code{endfile} spec. This is a spec string that specifies
5646 the last object files that will be passed to the linker.
5649 Process the @code{cpp} spec. This is used to construct the arguments
5650 to be passed to the C preprocessor.
5653 Process the @code{signed_char} spec. This is intended to be used
5654 to tell cpp whether a char is signed. It typically has the definition:
5656 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
5660 Process the @code{cc1} spec. This is used to construct the options to be
5661 passed to the actual C compiler (@samp{cc1}).
5664 Process the @code{cc1plus} spec. This is used to construct the options to be
5665 passed to the actual C++ compiler (@samp{cc1plus}).
5668 Substitute the variable part of a matched option. See below.
5669 Note that each comma in the substituted string is replaced by
5673 Remove all occurrences of @code{-S} from the command line. Note---this
5674 command is position dependent. @samp{%} commands in the spec string
5675 before this one will see @code{-S}, @samp{%} commands in the spec string
5676 after this one will not.
5678 @item %:@var{function}(@var{args})
5679 Call the named function @var{function}, passing it @var{args}.
5680 @var{args} is first processed as a nested spec string, then split
5681 into an argument vector in the usual fashion. The function returns
5682 a string which is processed as if it had appeared literally as part
5683 of the current spec.
5685 The following built-in spec functions are provided:
5688 @item @code{if-exists}
5689 The @code{if-exists} spec function takes one argument, an absolute
5690 pathname to a file. If the file exists, @code{if-exists} returns the
5691 pathname. Here is a small example of its usage:
5695 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
5698 @item @code{if-exists-else}
5699 The @code{if-exists-else} spec function is similar to the @code{if-exists}
5700 spec function, except that it takes two arguments. The first argument is
5701 an absolute pathname to a file. If the file exists, @code{if-exists-else}
5702 returns the pathname. If it does not exist, it returns the second argument.
5703 This way, @code{if-exists-else} can be used to select one file or another,
5704 based on the existence of the first. Here is a small example of its usage:
5708 crt0%O%s %:if-exists(crti%O%s) \
5709 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
5714 Substitutes the @code{-S} switch, if that switch was given to GCC@.
5715 If that switch was not specified, this substitutes nothing. Note that
5716 the leading dash is omitted when specifying this option, and it is
5717 automatically inserted if the substitution is performed. Thus the spec
5718 string @samp{%@{foo@}} would match the command-line option @option{-foo}
5719 and would output the command line option @option{-foo}.
5721 @item %W@{@code{S}@}
5722 Like %@{@code{S}@} but mark last argument supplied within as a file to be
5725 @item %@{@code{S}*@}
5726 Substitutes all the switches specified to GCC whose names start
5727 with @code{-S}, but which also take an argument. This is used for
5728 switches like @option{-o}, @option{-D}, @option{-I}, etc.
5729 GCC considers @option{-o foo} as being
5730 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
5731 text, including the space. Thus two arguments would be generated.
5733 @item %@{@code{S}*&@code{T}*@}
5734 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
5735 (the order of @code{S} and @code{T} in the spec is not significant).
5736 There can be any number of ampersand-separated variables; for each the
5737 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
5739 @item %@{@code{S}:@code{X}@}
5740 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
5742 @item %@{!@code{S}:@code{X}@}
5743 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
5745 @item %@{@code{S}*:@code{X}@}
5746 Substitutes @code{X} if one or more switches whose names start with
5747 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
5748 once, no matter how many such switches appeared. However, if @code{%*}
5749 appears somewhere in @code{X}, then @code{X} will be substituted once
5750 for each matching switch, with the @code{%*} replaced by the part of
5751 that switch that matched the @code{*}.
5753 @item %@{.@code{S}:@code{X}@}
5754 Substitutes @code{X}, if processing a file with suffix @code{S}.
5756 @item %@{!.@code{S}:@code{X}@}
5757 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
5759 @item %@{@code{S}|@code{P}:@code{X}@}
5760 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
5761 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
5762 although they have a stronger binding than the @samp{|}. If @code{%*}
5763 appears in @code{X}, all of the alternatives must be starred, and only
5764 the first matching alternative is substituted.
5766 For example, a spec string like this:
5769 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
5772 will output the following command-line options from the following input
5773 command-line options:
5778 -d fred.c -foo -baz -boggle
5779 -d jim.d -bar -baz -boggle
5782 @item %@{S:X; T:Y; :D@}
5784 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
5785 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
5786 be as many clauses as you need. This may be combined with @code{.},
5787 @code{!}, @code{|}, and @code{*} as needed.
5792 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
5793 construct may contain other nested @samp{%} constructs or spaces, or
5794 even newlines. They are processed as usual, as described above.
5795 Trailing white space in @code{X} is ignored. White space may also
5796 appear anywhere on the left side of the colon in these constructs,
5797 except between @code{.} or @code{*} and the corresponding word.
5799 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
5800 handled specifically in these constructs. If another value of
5801 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
5802 @option{-W} switch is found later in the command line, the earlier
5803 switch value is ignored, except with @{@code{S}*@} where @code{S} is
5804 just one letter, which passes all matching options.
5806 The character @samp{|} at the beginning of the predicate text is used to
5807 indicate that a command should be piped to the following command, but
5808 only if @option{-pipe} is specified.
5810 It is built into GCC which switches take arguments and which do not.
5811 (You might think it would be useful to generalize this to allow each
5812 compiler's spec to say which switches take arguments. But this cannot
5813 be done in a consistent fashion. GCC cannot even decide which input
5814 files have been specified without knowing which switches take arguments,
5815 and it must know which input files to compile in order to tell which
5818 GCC also knows implicitly that arguments starting in @option{-l} are to be
5819 treated as compiler output files, and passed to the linker in their
5820 proper position among the other output files.
5822 @c man begin OPTIONS
5824 @node Target Options
5825 @section Specifying Target Machine and Compiler Version
5826 @cindex target options
5827 @cindex cross compiling
5828 @cindex specifying machine version
5829 @cindex specifying compiler version and target machine
5830 @cindex compiler version, specifying
5831 @cindex target machine, specifying
5833 The usual way to run GCC is to run the executable called @file{gcc}, or
5834 @file{<machine>-gcc} when cross-compiling, or
5835 @file{<machine>-gcc-<version>} to run a version other than the one that
5836 was installed last. Sometimes this is inconvenient, so GCC provides
5837 options that will switch to another cross-compiler or version.
5840 @item -b @var{machine}
5842 The argument @var{machine} specifies the target machine for compilation.
5844 The value to use for @var{machine} is the same as was specified as the
5845 machine type when configuring GCC as a cross-compiler. For
5846 example, if a cross-compiler was configured with @samp{configure
5847 i386v}, meaning to compile for an 80386 running System V, then you
5848 would specify @option{-b i386v} to run that cross compiler.
5850 @item -V @var{version}
5852 The argument @var{version} specifies which version of GCC to run.
5853 This is useful when multiple versions are installed. For example,
5854 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5857 The @option{-V} and @option{-b} options work by running the
5858 @file{<machine>-gcc-<version>} executable, so there's no real reason to
5859 use them if you can just run that directly.
5861 @node Submodel Options
5862 @section Hardware Models and Configurations
5863 @cindex submodel options
5864 @cindex specifying hardware config
5865 @cindex hardware models and configurations, specifying
5866 @cindex machine dependent options
5868 Earlier we discussed the standard option @option{-b} which chooses among
5869 different installed compilers for completely different target
5870 machines, such as VAX vs.@: 68000 vs.@: 80386.
5872 In addition, each of these target machine types can have its own
5873 special options, starting with @samp{-m}, to choose among various
5874 hardware models or configurations---for example, 68010 vs 68020,
5875 floating coprocessor or none. A single installed version of the
5876 compiler can compile for any model or configuration, according to the
5879 Some configurations of the compiler also support additional special
5880 options, usually for compatibility with other compilers on the same
5883 These options are defined by the macro @code{TARGET_SWITCHES} in the
5884 machine description. The default for the options is also defined by
5885 that macro, which enables you to change the defaults.
5897 * RS/6000 and PowerPC Options::
5901 * i386 and x86-64 Options::
5903 * Intel 960 Options::
5904 * DEC Alpha Options::
5905 * DEC Alpha/VMS Options::
5908 * System V Options::
5909 * TMS320C3x/C4x Options::
5917 * S/390 and zSeries Options::
5921 * Xstormy16 Options::
5926 @node M680x0 Options
5927 @subsection M680x0 Options
5928 @cindex M680x0 options
5930 These are the @samp{-m} options defined for the 68000 series. The default
5931 values for these options depends on which style of 68000 was selected when
5932 the compiler was configured; the defaults for the most common choices are
5940 Generate output for a 68000. This is the default
5941 when the compiler is configured for 68000-based systems.
5943 Use this option for microcontrollers with a 68000 or EC000 core,
5944 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5950 Generate output for a 68020. This is the default
5951 when the compiler is configured for 68020-based systems.
5955 Generate output containing 68881 instructions for floating point.
5956 This is the default for most 68020 systems unless @option{--nfp} was
5957 specified when the compiler was configured.
5961 Generate output for a 68030. This is the default when the compiler is
5962 configured for 68030-based systems.
5966 Generate output for a 68040. This is the default when the compiler is
5967 configured for 68040-based systems.
5969 This option inhibits the use of 68881/68882 instructions that have to be
5970 emulated by software on the 68040. Use this option if your 68040 does not
5971 have code to emulate those instructions.
5975 Generate output for a 68060. This is the default when the compiler is
5976 configured for 68060-based systems.
5978 This option inhibits the use of 68020 and 68881/68882 instructions that
5979 have to be emulated by software on the 68060. Use this option if your 68060
5980 does not have code to emulate those instructions.
5984 Generate output for a CPU32. This is the default
5985 when the compiler is configured for CPU32-based systems.
5987 Use this option for microcontrollers with a
5988 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5989 68336, 68340, 68341, 68349 and 68360.
5993 Generate output for a 520X ``coldfire'' family cpu. This is the default
5994 when the compiler is configured for 520X-based systems.
5996 Use this option for microcontroller with a 5200 core, including
5997 the MCF5202, MCF5203, MCF5204 and MCF5202.
6002 Generate output for a 68040, without using any of the new instructions.
6003 This results in code which can run relatively efficiently on either a
6004 68020/68881 or a 68030 or a 68040. The generated code does use the
6005 68881 instructions that are emulated on the 68040.
6009 Generate output for a 68060, without using any of the new instructions.
6010 This results in code which can run relatively efficiently on either a
6011 68020/68881 or a 68030 or a 68040. The generated code does use the
6012 68881 instructions that are emulated on the 68060.
6015 @opindex msoft-float
6016 Generate output containing library calls for floating point.
6017 @strong{Warning:} the requisite libraries are not available for all m68k
6018 targets. Normally the facilities of the machine's usual C compiler are
6019 used, but this can't be done directly in cross-compilation. You must
6020 make your own arrangements to provide suitable library functions for
6021 cross-compilation. The embedded targets @samp{m68k-*-aout} and
6022 @samp{m68k-*-coff} do provide software floating point support.
6026 Consider type @code{int} to be 16 bits wide, like @code{short int}.
6029 @opindex mnobitfield
6030 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
6031 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
6035 Do use the bit-field instructions. The @option{-m68020} option implies
6036 @option{-mbitfield}. This is the default if you use a configuration
6037 designed for a 68020.
6041 Use a different function-calling convention, in which functions
6042 that take a fixed number of arguments return with the @code{rtd}
6043 instruction, which pops their arguments while returning. This
6044 saves one instruction in the caller since there is no need to pop
6045 the arguments there.
6047 This calling convention is incompatible with the one normally
6048 used on Unix, so you cannot use it if you need to call libraries
6049 compiled with the Unix compiler.
6051 Also, you must provide function prototypes for all functions that
6052 take variable numbers of arguments (including @code{printf});
6053 otherwise incorrect code will be generated for calls to those
6056 In addition, seriously incorrect code will result if you call a
6057 function with too many arguments. (Normally, extra arguments are
6058 harmlessly ignored.)
6060 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
6061 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
6064 @itemx -mno-align-int
6066 @opindex mno-align-int
6067 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
6068 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
6069 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
6070 Aligning variables on 32-bit boundaries produces code that runs somewhat
6071 faster on processors with 32-bit busses at the expense of more memory.
6073 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
6074 align structures containing the above types differently than
6075 most published application binary interface specifications for the m68k.
6079 Use the pc-relative addressing mode of the 68000 directly, instead of
6080 using a global offset table. At present, this option implies @option{-fpic},
6081 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
6082 not presently supported with @option{-mpcrel}, though this could be supported for
6083 68020 and higher processors.
6085 @item -mno-strict-align
6086 @itemx -mstrict-align
6087 @opindex mno-strict-align
6088 @opindex mstrict-align
6089 Do not (do) assume that unaligned memory references will be handled by
6094 @node M68hc1x Options
6095 @subsection M68hc1x Options
6096 @cindex M68hc1x options
6098 These are the @samp{-m} options defined for the 68hc11 and 68hc12
6099 microcontrollers. The default values for these options depends on
6100 which style of microcontroller was selected when the compiler was configured;
6101 the defaults for the most common choices are given below.
6108 Generate output for a 68HC11. This is the default
6109 when the compiler is configured for 68HC11-based systems.
6115 Generate output for a 68HC12. This is the default
6116 when the compiler is configured for 68HC12-based systems.
6122 Generate output for a 68HCS12.
6125 @opindex mauto-incdec
6126 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
6133 Enable the use of 68HC12 min and max instructions.
6136 @itemx -mno-long-calls
6137 @opindex mlong-calls
6138 @opindex mno-long-calls
6139 Treat all calls as being far away (near). If calls are assumed to be
6140 far away, the compiler will use the @code{call} instruction to
6141 call a function and the @code{rtc} instruction for returning.
6145 Consider type @code{int} to be 16 bits wide, like @code{short int}.
6147 @item -msoft-reg-count=@var{count}
6148 @opindex msoft-reg-count
6149 Specify the number of pseudo-soft registers which are used for the
6150 code generation. The maximum number is 32. Using more pseudo-soft
6151 register may or may not result in better code depending on the program.
6152 The default is 4 for 68HC11 and 2 for 68HC12.
6157 @subsection VAX Options
6160 These @samp{-m} options are defined for the VAX:
6165 Do not output certain jump instructions (@code{aobleq} and so on)
6166 that the Unix assembler for the VAX cannot handle across long
6171 Do output those jump instructions, on the assumption that you
6172 will assemble with the GNU assembler.
6176 Output code for g-format floating point numbers instead of d-format.
6180 @subsection SPARC Options
6181 @cindex SPARC options
6183 These @samp{-m} switches are supported on the SPARC:
6188 @opindex mno-app-regs
6190 Specify @option{-mapp-regs} to generate output using the global registers
6191 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
6194 To be fully SVR4 ABI compliant at the cost of some performance loss,
6195 specify @option{-mno-app-regs}. You should compile libraries and system
6196 software with this option.
6201 @opindex mhard-float
6202 Generate output containing floating point instructions. This is the
6208 @opindex msoft-float
6209 Generate output containing library calls for floating point.
6210 @strong{Warning:} the requisite libraries are not available for all SPARC
6211 targets. Normally the facilities of the machine's usual C compiler are
6212 used, but this cannot be done directly in cross-compilation. You must make
6213 your own arrangements to provide suitable library functions for
6214 cross-compilation. The embedded targets @samp{sparc-*-aout} and
6215 @samp{sparclite-*-*} do provide software floating point support.
6217 @option{-msoft-float} changes the calling convention in the output file;
6218 therefore, it is only useful if you compile @emph{all} of a program with
6219 this option. In particular, you need to compile @file{libgcc.a}, the
6220 library that comes with GCC, with @option{-msoft-float} in order for
6223 @item -mhard-quad-float
6224 @opindex mhard-quad-float
6225 Generate output containing quad-word (long double) floating point
6228 @item -msoft-quad-float
6229 @opindex msoft-quad-float
6230 Generate output containing library calls for quad-word (long double)
6231 floating point instructions. The functions called are those specified
6232 in the SPARC ABI@. This is the default.
6234 As of this writing, there are no sparc implementations that have hardware
6235 support for the quad-word floating point instructions. They all invoke
6236 a trap handler for one of these instructions, and then the trap handler
6237 emulates the effect of the instruction. Because of the trap handler overhead,
6238 this is much slower than calling the ABI library routines. Thus the
6239 @option{-msoft-quad-float} option is the default.
6245 With @option{-mflat}, the compiler does not generate save/restore instructions
6246 and will use a ``flat'' or single register window calling convention.
6247 This model uses %i7 as the frame pointer and is compatible with the normal
6248 register window model. Code from either may be intermixed.
6249 The local registers and the input registers (0--5) are still treated as
6250 ``call saved'' registers and will be saved on the stack as necessary.
6252 With @option{-mno-flat} (the default), the compiler emits save/restore
6253 instructions (except for leaf functions) and is the normal mode of operation.
6255 @item -mno-unaligned-doubles
6256 @itemx -munaligned-doubles
6257 @opindex mno-unaligned-doubles
6258 @opindex munaligned-doubles
6259 Assume that doubles have 8 byte alignment. This is the default.
6261 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
6262 alignment only if they are contained in another type, or if they have an
6263 absolute address. Otherwise, it assumes they have 4 byte alignment.
6264 Specifying this option avoids some rare compatibility problems with code
6265 generated by other compilers. It is not the default because it results
6266 in a performance loss, especially for floating point code.
6268 @item -mno-faster-structs
6269 @itemx -mfaster-structs
6270 @opindex mno-faster-structs
6271 @opindex mfaster-structs
6272 With @option{-mfaster-structs}, the compiler assumes that structures
6273 should have 8 byte alignment. This enables the use of pairs of
6274 @code{ldd} and @code{std} instructions for copies in structure
6275 assignment, in place of twice as many @code{ld} and @code{st} pairs.
6276 However, the use of this changed alignment directly violates the SPARC
6277 ABI@. Thus, it's intended only for use on targets where the developer
6278 acknowledges that their resulting code will not be directly in line with
6279 the rules of the ABI@.
6282 @opindex mimpure-text
6283 @option{-mimpure-text}, used in addition to @option{-shared}, tells
6284 the compiler to not pass @option{-z text} to the linker when linking a
6285 shared object. Using this option, you can link position-dependent
6286 code into a shared object.
6288 @option{-mimpure-text} suppresses the ``relocations remain against
6289 allocatable but non-writable sections'' linker error message.
6290 However, the necessary relocations will trigger copy-on-write, and the
6291 shared object is not actually shared across processes. Instead of
6292 using @option{-mimpure-text}, you should compile all source code with
6293 @option{-fpic} or @option{-fPIC}.
6295 This option is only available on SunOS and Solaris.
6301 These two options select variations on the SPARC architecture.
6303 By default (unless specifically configured for the Fujitsu SPARClite),
6304 GCC generates code for the v7 variant of the SPARC architecture.
6306 @option{-mv8} will give you SPARC v8 code. The only difference from v7
6307 code is that the compiler emits the integer multiply and integer
6308 divide instructions which exist in SPARC v8 but not in SPARC v7.
6310 @option{-msparclite} will give you SPARClite code. This adds the integer
6311 multiply, integer divide step and scan (@code{ffs}) instructions which
6312 exist in SPARClite but not in SPARC v7.
6314 These options are deprecated and will be deleted in a future GCC release.
6315 They have been replaced with @option{-mcpu=xxx}.
6320 @opindex msupersparc
6321 These two options select the processor for which the code is optimized.
6323 With @option{-mcypress} (the default), the compiler optimizes code for the
6324 Cypress CY7C602 chip, as used in the SPARCStation/SPARCServer 3xx series.
6325 This is also appropriate for the older SPARCStation 1, 2, IPX etc.
6327 With @option{-msupersparc} the compiler optimizes code for the SuperSPARC cpu, as
6328 used in the SPARCStation 10, 1000 and 2000 series. This flag also enables use
6329 of the full SPARC v8 instruction set.
6331 These options are deprecated and will be deleted in a future GCC release.
6332 They have been replaced with @option{-mcpu=xxx}.
6334 @item -mcpu=@var{cpu_type}
6336 Set the instruction set, register set, and instruction scheduling parameters
6337 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
6338 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
6339 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
6340 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
6343 Default instruction scheduling parameters are used for values that select
6344 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
6345 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
6347 Here is a list of each supported architecture and their supported
6352 v8: supersparc, hypersparc
6353 sparclite: f930, f934, sparclite86x
6355 v9: ultrasparc, ultrasparc3
6358 @item -mtune=@var{cpu_type}
6360 Set the instruction scheduling parameters for machine type
6361 @var{cpu_type}, but do not set the instruction set or register set that the
6362 option @option{-mcpu=@var{cpu_type}} would.
6364 The same values for @option{-mcpu=@var{cpu_type}} can be used for
6365 @option{-mtune=@var{cpu_type}}, but the only useful values are those
6366 that select a particular cpu implementation. Those are @samp{cypress},
6367 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
6368 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
6373 These @samp{-m} switches are supported in addition to the above
6374 on the SPARCLET processor.
6377 @item -mlittle-endian
6378 @opindex mlittle-endian
6379 Generate code for a processor running in little-endian mode.
6383 Treat register @code{%g0} as a normal register.
6384 GCC will continue to clobber it as necessary but will not assume
6385 it always reads as 0.
6387 @item -mbroken-saverestore
6388 @opindex mbroken-saverestore
6389 Generate code that does not use non-trivial forms of the @code{save} and
6390 @code{restore} instructions. Early versions of the SPARCLET processor do
6391 not correctly handle @code{save} and @code{restore} instructions used with
6392 arguments. They correctly handle them used without arguments. A @code{save}
6393 instruction used without arguments increments the current window pointer
6394 but does not allocate a new stack frame. It is assumed that the window
6395 overflow trap handler will properly handle this case as will interrupt
6399 These @samp{-m} switches are supported in addition to the above
6400 on SPARC V9 processors in 64-bit environments.
6403 @item -mlittle-endian
6404 @opindex mlittle-endian
6405 Generate code for a processor running in little-endian mode. It is only
6406 available for a few configurations and most notably not on Solaris.
6412 Generate code for a 32-bit or 64-bit environment.
6413 The 32-bit environment sets int, long and pointer to 32 bits.
6414 The 64-bit environment sets int to 32 bits and long and pointer
6417 @item -mcmodel=medlow
6418 @opindex mcmodel=medlow
6419 Generate code for the Medium/Low code model: the program must be linked
6420 in the low 32 bits of the address space. Pointers are 64 bits.
6421 Programs can be statically or dynamically linked.
6423 @item -mcmodel=medmid
6424 @opindex mcmodel=medmid
6425 Generate code for the Medium/Middle code model: the program must be linked
6426 in the low 44 bits of the address space, the text segment must be less than
6427 2G bytes, and data segment must be within 2G of the text segment.
6428 Pointers are 64 bits.
6430 @item -mcmodel=medany
6431 @opindex mcmodel=medany
6432 Generate code for the Medium/Anywhere code model: the program may be linked
6433 anywhere in the address space, the text segment must be less than
6434 2G bytes, and data segment must be within 2G of the text segment.
6435 Pointers are 64 bits.
6437 @item -mcmodel=embmedany
6438 @opindex mcmodel=embmedany
6439 Generate code for the Medium/Anywhere code model for embedded systems:
6440 assume a 32-bit text and a 32-bit data segment, both starting anywhere
6441 (determined at link time). Register %g4 points to the base of the
6442 data segment. Pointers are still 64 bits.
6443 Programs are statically linked, PIC is not supported.
6446 @itemx -mno-stack-bias
6447 @opindex mstack-bias
6448 @opindex mno-stack-bias
6449 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
6450 frame pointer if present, are offset by @minus{}2047 which must be added back
6451 when making stack frame references.
6452 Otherwise, assume no such offset is present.
6456 @subsection ARM Options
6459 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
6464 @opindex mapcs-frame
6465 Generate a stack frame that is compliant with the ARM Procedure Call
6466 Standard for all functions, even if this is not strictly necessary for
6467 correct execution of the code. Specifying @option{-fomit-frame-pointer}
6468 with this option will cause the stack frames not to be generated for
6469 leaf functions. The default is @option{-mno-apcs-frame}.
6473 This is a synonym for @option{-mapcs-frame}.
6477 Generate code for a processor running with a 26-bit program counter,
6478 and conforming to the function calling standards for the APCS 26-bit
6479 option. This option replaces the @option{-m2} and @option{-m3} options
6480 of previous releases of the compiler.
6484 Generate code for a processor running with a 32-bit program counter,
6485 and conforming to the function calling standards for the APCS 32-bit
6486 option. This option replaces the @option{-m6} option of previous releases
6490 @c not currently implemented
6491 @item -mapcs-stack-check
6492 @opindex mapcs-stack-check
6493 Generate code to check the amount of stack space available upon entry to
6494 every function (that actually uses some stack space). If there is
6495 insufficient space available then either the function
6496 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
6497 called, depending upon the amount of stack space required. The run time
6498 system is required to provide these functions. The default is
6499 @option{-mno-apcs-stack-check}, since this produces smaller code.
6501 @c not currently implemented
6503 @opindex mapcs-float
6504 Pass floating point arguments using the float point registers. This is
6505 one of the variants of the APCS@. This option is recommended if the
6506 target hardware has a floating point unit or if a lot of floating point
6507 arithmetic is going to be performed by the code. The default is
6508 @option{-mno-apcs-float}, since integer only code is slightly increased in
6509 size if @option{-mapcs-float} is used.
6511 @c not currently implemented
6512 @item -mapcs-reentrant
6513 @opindex mapcs-reentrant
6514 Generate reentrant, position independent code. The default is
6515 @option{-mno-apcs-reentrant}.
6518 @item -mthumb-interwork
6519 @opindex mthumb-interwork
6520 Generate code which supports calling between the ARM and Thumb
6521 instruction sets. Without this option the two instruction sets cannot
6522 be reliably used inside one program. The default is
6523 @option{-mno-thumb-interwork}, since slightly larger code is generated
6524 when @option{-mthumb-interwork} is specified.
6526 @item -mno-sched-prolog
6527 @opindex mno-sched-prolog
6528 Prevent the reordering of instructions in the function prolog, or the
6529 merging of those instruction with the instructions in the function's
6530 body. This means that all functions will start with a recognizable set
6531 of instructions (or in fact one of a choice from a small set of
6532 different function prologues), and this information can be used to
6533 locate the start if functions inside an executable piece of code. The
6534 default is @option{-msched-prolog}.
6537 @opindex mhard-float
6538 Generate output containing floating point instructions. This is the
6542 @opindex msoft-float
6543 Generate output containing library calls for floating point.
6544 @strong{Warning:} the requisite libraries are not available for all ARM
6545 targets. Normally the facilities of the machine's usual C compiler are
6546 used, but this cannot be done directly in cross-compilation. You must make
6547 your own arrangements to provide suitable library functions for
6550 @option{-msoft-float} changes the calling convention in the output file;
6551 therefore, it is only useful if you compile @emph{all} of a program with
6552 this option. In particular, you need to compile @file{libgcc.a}, the
6553 library that comes with GCC, with @option{-msoft-float} in order for
6556 @item -mlittle-endian
6557 @opindex mlittle-endian
6558 Generate code for a processor running in little-endian mode. This is
6559 the default for all standard configurations.
6562 @opindex mbig-endian
6563 Generate code for a processor running in big-endian mode; the default is
6564 to compile code for a little-endian processor.
6566 @item -mwords-little-endian
6567 @opindex mwords-little-endian
6568 This option only applies when generating code for big-endian processors.
6569 Generate code for a little-endian word order but a big-endian byte
6570 order. That is, a byte order of the form @samp{32107654}. Note: this
6571 option should only be used if you require compatibility with code for
6572 big-endian ARM processors generated by versions of the compiler prior to
6575 @item -malignment-traps
6576 @opindex malignment-traps
6577 Generate code that will not trap if the MMU has alignment traps enabled.
6578 On ARM architectures prior to ARMv4, there were no instructions to
6579 access half-word objects stored in memory. However, when reading from
6580 memory a feature of the ARM architecture allows a word load to be used,
6581 even if the address is unaligned, and the processor core will rotate the
6582 data as it is being loaded. This option tells the compiler that such
6583 misaligned accesses will cause a MMU trap and that it should instead
6584 synthesize the access as a series of byte accesses. The compiler can
6585 still use word accesses to load half-word data if it knows that the
6586 address is aligned to a word boundary.
6588 This option is ignored when compiling for ARM architecture 4 or later,
6589 since these processors have instructions to directly access half-word
6592 @item -mno-alignment-traps
6593 @opindex mno-alignment-traps
6594 Generate code that assumes that the MMU will not trap unaligned
6595 accesses. This produces better code when the target instruction set
6596 does not have half-word memory operations (i.e.@: implementations prior to
6599 Note that you cannot use this option to access unaligned word objects,
6600 since the processor will only fetch one 32-bit aligned object from
6603 The default setting for most targets is @option{-mno-alignment-traps}, since
6604 this produces better code when there are no half-word memory
6605 instructions available.
6607 @item -mshort-load-bytes
6608 @itemx -mno-short-load-words
6609 @opindex mshort-load-bytes
6610 @opindex mno-short-load-words
6611 These are deprecated aliases for @option{-malignment-traps}.
6613 @item -mno-short-load-bytes
6614 @itemx -mshort-load-words
6615 @opindex mno-short-load-bytes
6616 @opindex mshort-load-words
6617 This are deprecated aliases for @option{-mno-alignment-traps}.
6619 @item -mcpu=@var{name}
6621 This specifies the name of the target ARM processor. GCC uses this name
6622 to determine what kind of instructions it can emit when generating
6623 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6624 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6625 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6626 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6627 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6628 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6629 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6630 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6631 @samp{arm920t}, @samp{arm926ejs}, @samp{arm940t}, @samp{arm9tdmi},
6632 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ejs},
6633 @samp{arm1136js}, @samp{arm1136jfs} ,@samp{xscale}, @samp{iwmmxt},
6636 @itemx -mtune=@var{name}
6638 This option is very similar to the @option{-mcpu=} option, except that
6639 instead of specifying the actual target processor type, and hence
6640 restricting which instructions can be used, it specifies that GCC should
6641 tune the performance of the code as if the target were of the type
6642 specified in this option, but still choosing the instructions that it
6643 will generate based on the cpu specified by a @option{-mcpu=} option.
6644 For some ARM implementations better performance can be obtained by using
6647 @item -march=@var{name}
6649 This specifies the name of the target ARM architecture. GCC uses this
6650 name to determine what kind of instructions it can emit when generating
6651 assembly code. This option can be used in conjunction with or instead
6652 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6653 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6654 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6j},
6655 @samp{iwmmxt}, @samp{ep9312}.
6657 @item -mfpe=@var{number}
6658 @itemx -mfp=@var{number}
6661 This specifies the version of the floating point emulation available on
6662 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6663 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6665 @item -mstructure-size-boundary=@var{n}
6666 @opindex mstructure-size-boundary
6667 The size of all structures and unions will be rounded up to a multiple
6668 of the number of bits set by this option. Permissible values are 8 and
6669 32. The default value varies for different toolchains. For the COFF
6670 targeted toolchain the default value is 8. Specifying the larger number
6671 can produce faster, more efficient code, but can also increase the size
6672 of the program. The two values are potentially incompatible. Code
6673 compiled with one value cannot necessarily expect to work with code or
6674 libraries compiled with the other value, if they exchange information
6675 using structures or unions.
6677 @item -mabort-on-noreturn
6678 @opindex mabort-on-noreturn
6679 Generate a call to the function @code{abort} at the end of a
6680 @code{noreturn} function. It will be executed if the function tries to
6684 @itemx -mno-long-calls
6685 @opindex mlong-calls
6686 @opindex mno-long-calls
6687 Tells the compiler to perform function calls by first loading the
6688 address of the function into a register and then performing a subroutine
6689 call on this register. This switch is needed if the target function
6690 will lie outside of the 64 megabyte addressing range of the offset based
6691 version of subroutine call instruction.
6693 Even if this switch is enabled, not all function calls will be turned
6694 into long calls. The heuristic is that static functions, functions
6695 which have the @samp{short-call} attribute, functions that are inside
6696 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6697 definitions have already been compiled within the current compilation
6698 unit, will not be turned into long calls. The exception to this rule is
6699 that weak function definitions, functions with the @samp{long-call}
6700 attribute or the @samp{section} attribute, and functions that are within
6701 the scope of a @samp{#pragma long_calls} directive, will always be
6702 turned into long calls.
6704 This feature is not enabled by default. Specifying
6705 @option{-mno-long-calls} will restore the default behavior, as will
6706 placing the function calls within the scope of a @samp{#pragma
6707 long_calls_off} directive. Note these switches have no effect on how
6708 the compiler generates code to handle function calls via function
6711 @item -mnop-fun-dllimport
6712 @opindex mnop-fun-dllimport
6713 Disable support for the @code{dllimport} attribute.
6715 @item -msingle-pic-base
6716 @opindex msingle-pic-base
6717 Treat the register used for PIC addressing as read-only, rather than
6718 loading it in the prologue for each function. The run-time system is
6719 responsible for initializing this register with an appropriate value
6720 before execution begins.
6722 @item -mpic-register=@var{reg}
6723 @opindex mpic-register
6724 Specify the register to be used for PIC addressing. The default is R10
6725 unless stack-checking is enabled, when R9 is used.
6727 @item -mcirrus-fix-invalid-insns
6728 @opindex mcirrus-fix-invalid-insns
6729 @opindex mno-cirrus-fix-invalid-insns
6730 Insert NOPs into the instruction stream to in order to work around
6731 problems with invalid Maverick instruction combinations. This option
6732 is only valid if the @option{-mcpu=ep9312} option has been used to
6733 enable generation of instructions for the Cirrus Maverick floating
6734 point co-processor. This option is not enabled by default, since the
6735 problem is only present in older Maverick implementations. The default
6736 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
6739 @item -mpoke-function-name
6740 @opindex mpoke-function-name
6741 Write the name of each function into the text section, directly
6742 preceding the function prologue. The generated code is similar to this:
6746 .ascii "arm_poke_function_name", 0
6749 .word 0xff000000 + (t1 - t0)
6750 arm_poke_function_name
6752 stmfd sp!, @{fp, ip, lr, pc@}
6756 When performing a stack backtrace, code can inspect the value of
6757 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6758 location @code{pc - 12} and the top 8 bits are set, then we know that
6759 there is a function name embedded immediately preceding this location
6760 and has length @code{((pc[-3]) & 0xff000000)}.
6764 Generate code for the 16-bit Thumb instruction set. The default is to
6765 use the 32-bit ARM instruction set.
6768 @opindex mtpcs-frame
6769 Generate a stack frame that is compliant with the Thumb Procedure Call
6770 Standard for all non-leaf functions. (A leaf function is one that does
6771 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6773 @item -mtpcs-leaf-frame
6774 @opindex mtpcs-leaf-frame
6775 Generate a stack frame that is compliant with the Thumb Procedure Call
6776 Standard for all leaf functions. (A leaf function is one that does
6777 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6779 @item -mcallee-super-interworking
6780 @opindex mcallee-super-interworking
6781 Gives all externally visible functions in the file being compiled an ARM
6782 instruction set header which switches to Thumb mode before executing the
6783 rest of the function. This allows these functions to be called from
6784 non-interworking code.
6786 @item -mcaller-super-interworking
6787 @opindex mcaller-super-interworking
6788 Allows calls via function pointers (including virtual functions) to
6789 execute correctly regardless of whether the target code has been
6790 compiled for interworking or not. There is a small overhead in the cost
6791 of executing a function pointer if this option is enabled.
6795 @node MN10200 Options
6796 @subsection MN10200 Options
6797 @cindex MN10200 options
6799 These @option{-m} options are defined for Matsushita MN10200 architectures:
6804 Indicate to the linker that it should perform a relaxation optimization pass
6805 to shorten branches, calls and absolute memory addresses. This option only
6806 has an effect when used on the command line for the final link step.
6808 This option makes symbolic debugging impossible.
6811 @node MN10300 Options
6812 @subsection MN10300 Options
6813 @cindex MN10300 options
6815 These @option{-m} options are defined for Matsushita MN10300 architectures:
6820 Generate code to avoid bugs in the multiply instructions for the MN10300
6821 processors. This is the default.
6824 @opindex mno-mult-bug
6825 Do not generate code to avoid bugs in the multiply instructions for the
6830 Generate code which uses features specific to the AM33 processor.
6834 Do not generate code which uses features specific to the AM33 processor. This
6839 Do not link in the C run-time initialization object file.
6843 Indicate to the linker that it should perform a relaxation optimization pass
6844 to shorten branches, calls and absolute memory addresses. This option only
6845 has an effect when used on the command line for the final link step.
6847 This option makes symbolic debugging impossible.
6851 @node M32R/D Options
6852 @subsection M32R/D Options
6853 @cindex M32R/D options
6855 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6860 Generate code for the M32R/X@.
6864 Generate code for the M32R@. This is the default.
6866 @item -mcode-model=small
6867 @opindex mcode-model=small
6868 Assume all objects live in the lower 16MB of memory (so that their addresses
6869 can be loaded with the @code{ld24} instruction), and assume all subroutines
6870 are reachable with the @code{bl} instruction.
6871 This is the default.
6873 The addressability of a particular object can be set with the
6874 @code{model} attribute.
6876 @item -mcode-model=medium
6877 @opindex mcode-model=medium
6878 Assume objects may be anywhere in the 32-bit address space (the compiler
6879 will generate @code{seth/add3} instructions to load their addresses), and
6880 assume all subroutines are reachable with the @code{bl} instruction.
6882 @item -mcode-model=large
6883 @opindex mcode-model=large
6884 Assume objects may be anywhere in the 32-bit address space (the compiler
6885 will generate @code{seth/add3} instructions to load their addresses), and
6886 assume subroutines may not be reachable with the @code{bl} instruction
6887 (the compiler will generate the much slower @code{seth/add3/jl}
6888 instruction sequence).
6891 @opindex msdata=none
6892 Disable use of the small data area. Variables will be put into
6893 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6894 @code{section} attribute has been specified).
6895 This is the default.
6897 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6898 Objects may be explicitly put in the small data area with the
6899 @code{section} attribute using one of these sections.
6902 @opindex msdata=sdata
6903 Put small global and static data in the small data area, but do not
6904 generate special code to reference them.
6908 Put small global and static data in the small data area, and generate
6909 special instructions to reference them.
6913 @cindex smaller data references
6914 Put global and static objects less than or equal to @var{num} bytes
6915 into the small data or bss sections instead of the normal data or bss
6916 sections. The default value of @var{num} is 8.
6917 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6918 for this option to have any effect.
6920 All modules should be compiled with the same @option{-G @var{num}} value.
6921 Compiling with different values of @var{num} may or may not work; if it
6922 doesn't the linker will give an error message---incorrect code will not be
6928 @subsection M88K Options
6929 @cindex M88k options
6931 These @samp{-m} options are defined for Motorola 88k architectures:
6936 Generate code that works well on both the m88100 and the
6941 Generate code that works best for the m88100, but that also
6946 Generate code that works best for the m88110, and may not run
6951 Obsolete option to be removed from the next revision.
6954 @item -midentify-revision
6955 @opindex midentify-revision
6956 @cindex identifying source, compiler (88k)
6957 Include an @code{ident} directive in the assembler output recording the
6958 source file name, compiler name and version, timestamp, and compilation
6961 @item -mno-underscores
6962 @opindex mno-underscores
6963 @cindex underscores, avoiding (88k)
6964 In assembler output, emit symbol names without adding an underscore
6965 character at the beginning of each name. The default is to use an
6966 underscore as prefix on each name.
6968 @item -mocs-debug-info
6969 @itemx -mno-ocs-debug-info
6970 @opindex mocs-debug-info
6971 @opindex mno-ocs-debug-info
6973 @cindex debugging, 88k OCS
6974 Include (or omit) additional debugging information (about registers used
6975 in each stack frame) as specified in the 88open Object Compatibility
6976 Standard, ``OCS''@. This extra information allows debugging of code that
6977 has had the frame pointer eliminated. The default for SVr4 and Delta 88
6978 SVr3.2 is to include this information; other 88k configurations omit this
6979 information by default.
6981 @item -mocs-frame-position
6982 @opindex mocs-frame-position
6983 @cindex register positions in frame (88k)
6984 When emitting COFF debugging information for automatic variables and
6985 parameters stored on the stack, use the offset from the canonical frame
6986 address, which is the stack pointer (register 31) on entry to the
6987 function. The SVr4 and Delta88 SVr3.2, and BCS configurations use
6988 @option{-mocs-frame-position}; other 88k configurations have the default
6989 @option{-mno-ocs-frame-position}.
6991 @item -mno-ocs-frame-position
6992 @opindex mno-ocs-frame-position
6993 @cindex register positions in frame (88k)
6994 When emitting COFF debugging information for automatic variables and
6995 parameters stored on the stack, use the offset from the frame pointer
6996 register (register 30). When this option is in effect, the frame
6997 pointer is not eliminated when debugging information is selected by the
7000 @item -moptimize-arg-area
7001 @opindex moptimize-arg-area
7002 @cindex arguments in frame (88k)
7003 Save space by reorganizing the stack frame. This option generates code
7004 that does not agree with the 88open specifications, but uses less
7007 @itemx -mno-optimize-arg-area
7008 @opindex mno-optimize-arg-area
7009 Do not reorganize the stack frame to save space. This is the default.
7010 The generated conforms to the specification, but uses more memory.
7012 @item -mshort-data-@var{num}
7013 @opindex mshort-data
7014 @cindex smaller data references (88k)
7015 @cindex r0-relative references (88k)
7016 Generate smaller data references by making them relative to @code{r0},
7017 which allows loading a value using a single instruction (rather than the
7018 usual two). You control which data references are affected by
7019 specifying @var{num} with this option. For example, if you specify
7020 @option{-mshort-data-512}, then the data references affected are those
7021 involving displacements of less than 512 bytes.
7022 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
7025 @item -mserialize-volatile
7026 @opindex mserialize-volatile
7027 @itemx -mno-serialize-volatile
7028 @opindex mno-serialize-volatile
7029 @cindex sequential consistency on 88k
7030 Do, or don't, generate code to guarantee sequential consistency
7031 of volatile memory references. By default, consistency is
7034 The order of memory references made by the MC88110 processor does
7035 not always match the order of the instructions requesting those
7036 references. In particular, a load instruction may execute before
7037 a preceding store instruction. Such reordering violates
7038 sequential consistency of volatile memory references, when there
7039 are multiple processors. When consistency must be guaranteed,
7040 GCC generates special instructions, as needed, to force
7041 execution in the proper order.
7043 The MC88100 processor does not reorder memory references and so
7044 always provides sequential consistency. However, by default, GCC
7045 generates the special instructions to guarantee consistency
7046 even when you use @option{-m88100}, so that the code may be run on an
7047 MC88110 processor. If you intend to run your code only on the
7048 MC88100 processor, you may use @option{-mno-serialize-volatile}.
7050 The extra code generated to guarantee consistency may affect the
7051 performance of your application. If you know that you can safely
7052 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
7058 @cindex assembler syntax, 88k
7060 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
7061 related to System V release 4 (SVr4). This controls the following:
7065 Which variant of the assembler syntax to emit.
7067 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
7068 that is used on System V release 4.
7070 @option{-msvr4} makes GCC issue additional declaration directives used in
7074 @option{-msvr4} is the default for the m88k-motorola-sysv4 configuration.
7075 @option{-msvr3} is the default for all other m88k configurations.
7077 @item -mversion-03.00
7078 @opindex mversion-03.00
7079 This option is obsolete, and is ignored.
7080 @c ??? which asm syntax better for GAS? option there too?
7082 @item -mno-check-zero-division
7083 @itemx -mcheck-zero-division
7084 @opindex mno-check-zero-division
7085 @opindex mcheck-zero-division
7086 @cindex zero division on 88k
7087 Do, or don't, generate code to guarantee that integer division by
7088 zero will be detected. By default, detection is guaranteed.
7090 Some models of the MC88100 processor fail to trap upon integer
7091 division by zero under certain conditions. By default, when
7092 compiling code that might be run on such a processor, GCC
7093 generates code that explicitly checks for zero-valued divisors
7094 and traps with exception number 503 when one is detected. Use of
7095 @option{-mno-check-zero-division} suppresses such checking for code
7096 generated to run on an MC88100 processor.
7098 GCC assumes that the MC88110 processor correctly detects all instances
7099 of integer division by zero. When @option{-m88110} is specified, no
7100 explicit checks for zero-valued divisors are generated, and both
7101 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
7104 @item -muse-div-instruction
7105 @opindex muse-div-instruction
7106 @cindex divide instruction, 88k
7107 Use the div instruction for signed integer division on the
7108 MC88100 processor. By default, the div instruction is not used.
7110 On the MC88100 processor the signed integer division instruction
7111 div) traps to the operating system on a negative operand. The
7112 operating system transparently completes the operation, but at a
7113 large cost in execution time. By default, when compiling code
7114 that might be run on an MC88100 processor, GCC emulates signed
7115 integer division using the unsigned integer division instruction
7116 divu), thereby avoiding the large penalty of a trap to the
7117 operating system. Such emulation has its own, smaller, execution
7118 cost in both time and space. To the extent that your code's
7119 important signed integer division operations are performed on two
7120 nonnegative operands, it may be desirable to use the div
7121 instruction directly.
7123 On the MC88110 processor the div instruction (also known as the
7124 divs instruction) processes negative operands without trapping to
7125 the operating system. When @option{-m88110} is specified,
7126 @option{-muse-div-instruction} is ignored, and the div instruction is used
7127 for signed integer division.
7129 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
7130 particular, the behavior of such a division with and without
7131 @option{-muse-div-instruction} may differ.
7133 @item -mtrap-large-shift
7134 @itemx -mhandle-large-shift
7135 @opindex mtrap-large-shift
7136 @opindex mhandle-large-shift
7137 @cindex bit shift overflow (88k)
7138 @cindex large bit shifts (88k)
7139 Include code to detect bit-shifts of more than 31 bits; respectively,
7140 trap such shifts or emit code to handle them properly. By default GCC
7141 makes no special provision for large bit shifts.
7143 @item -mwarn-passed-structs
7144 @opindex mwarn-passed-structs
7145 @cindex structure passing (88k)
7146 Warn when a function passes a struct as an argument or result.
7147 Structure-passing conventions have changed during the evolution of the C
7148 language, and are often the source of portability problems. By default,
7149 GCC issues no such warning.
7152 @c break page here to avoid unsightly interparagraph stretch.
7156 @node RS/6000 and PowerPC Options
7157 @subsection IBM RS/6000 and PowerPC Options
7158 @cindex RS/6000 and PowerPC Options
7159 @cindex IBM RS/6000 and PowerPC Options
7161 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
7169 @itemx -mpowerpc-gpopt
7170 @itemx -mno-powerpc-gpopt
7171 @itemx -mpowerpc-gfxopt
7172 @itemx -mno-powerpc-gfxopt
7174 @itemx -mno-powerpc64
7180 @opindex mno-powerpc
7181 @opindex mpowerpc-gpopt
7182 @opindex mno-powerpc-gpopt
7183 @opindex mpowerpc-gfxopt
7184 @opindex mno-powerpc-gfxopt
7186 @opindex mno-powerpc64
7187 GCC supports two related instruction set architectures for the
7188 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
7189 instructions supported by the @samp{rios} chip set used in the original
7190 RS/6000 systems and the @dfn{PowerPC} instruction set is the
7191 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
7192 the IBM 4xx microprocessors.
7194 Neither architecture is a subset of the other. However there is a
7195 large common subset of instructions supported by both. An MQ
7196 register is included in processors supporting the POWER architecture.
7198 You use these options to specify which instructions are available on the
7199 processor you are using. The default value of these options is
7200 determined when configuring GCC@. Specifying the
7201 @option{-mcpu=@var{cpu_type}} overrides the specification of these
7202 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
7203 rather than the options listed above.
7205 The @option{-mpower} option allows GCC to generate instructions that
7206 are found only in the POWER architecture and to use the MQ register.
7207 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
7208 to generate instructions that are present in the POWER2 architecture but
7209 not the original POWER architecture.
7211 The @option{-mpowerpc} option allows GCC to generate instructions that
7212 are found only in the 32-bit subset of the PowerPC architecture.
7213 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
7214 GCC to use the optional PowerPC architecture instructions in the
7215 General Purpose group, including floating-point square root. Specifying
7216 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
7217 use the optional PowerPC architecture instructions in the Graphics
7218 group, including floating-point select.
7220 The @option{-mpowerpc64} option allows GCC to generate the additional
7221 64-bit instructions that are found in the full PowerPC64 architecture
7222 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
7223 @option{-mno-powerpc64}.
7225 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
7226 will use only the instructions in the common subset of both
7227 architectures plus some special AIX common-mode calls, and will not use
7228 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
7229 permits GCC to use any instruction from either architecture and to
7230 allow use of the MQ register; specify this for the Motorola MPC601.
7232 @item -mnew-mnemonics
7233 @itemx -mold-mnemonics
7234 @opindex mnew-mnemonics
7235 @opindex mold-mnemonics
7236 Select which mnemonics to use in the generated assembler code. With
7237 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
7238 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
7239 assembler mnemonics defined for the POWER architecture. Instructions
7240 defined in only one architecture have only one mnemonic; GCC uses that
7241 mnemonic irrespective of which of these options is specified.
7243 GCC defaults to the mnemonics appropriate for the architecture in
7244 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
7245 value of these option. Unless you are building a cross-compiler, you
7246 should normally not specify either @option{-mnew-mnemonics} or
7247 @option{-mold-mnemonics}, but should instead accept the default.
7249 @item -mcpu=@var{cpu_type}
7251 Set architecture type, register usage, choice of mnemonics, and
7252 instruction scheduling parameters for machine type @var{cpu_type}.
7253 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
7254 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
7255 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
7256 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
7257 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
7258 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
7260 @option{-mcpu=common} selects a completely generic processor. Code
7261 generated under this option will run on any POWER or PowerPC processor.
7262 GCC will use only the instructions in the common subset of both
7263 architectures, and will not use the MQ register. GCC assumes a generic
7264 processor model for scheduling purposes.
7266 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
7267 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
7268 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
7269 types, with an appropriate, generic processor model assumed for
7270 scheduling purposes.
7272 The other options specify a specific processor. Code generated under
7273 those options will run best on that processor, and may not run at all on
7276 The @option{-mcpu} options automatically enable or disable other
7277 @option{-m} options as follows:
7281 @option{-mno-power}, @option{-mno-powerpc}
7288 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
7303 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
7306 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
7311 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
7314 @item -mtune=@var{cpu_type}
7316 Set the instruction scheduling parameters for machine type
7317 @var{cpu_type}, but do not set the architecture type, register usage, or
7318 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
7319 values for @var{cpu_type} are used for @option{-mtune} as for
7320 @option{-mcpu}. If both are specified, the code generated will use the
7321 architecture, registers, and mnemonics set by @option{-mcpu}, but the
7322 scheduling parameters set by @option{-mtune}.
7327 @opindex mno-altivec
7328 These switches enable or disable the use of built-in functions that
7329 allow access to the AltiVec instruction set. You may also need to set
7330 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
7335 Extend the current ABI with SPE ABI extensions. This does not change
7336 the default ABI, instead it adds the SPE ABI extensions to the current
7340 @opindex mabi=no-spe
7341 Disable Booke SPE ABI extensions for the current ABI.
7343 @item -misel=@var{yes/no}
7346 This switch enables or disables the generation of ISEL instructions.
7348 @item -mspe=@var{yes/no}
7351 This switch enables or disables the generation of SPE simd
7354 @item -mfloat-gprs=@var{yes/no}
7356 @opindex mfloat-gprs
7357 This switch enables or disables the generation of floating point
7358 operations on the general purpose registers for architectures that
7359 support it. This option is currently only available on the MPC8540.
7362 @itemx -mno-fp-in-toc
7363 @itemx -mno-sum-in-toc
7364 @itemx -mminimal-toc
7366 @opindex mno-fp-in-toc
7367 @opindex mno-sum-in-toc
7368 @opindex mminimal-toc
7369 Modify generation of the TOC (Table Of Contents), which is created for
7370 every executable file. The @option{-mfull-toc} option is selected by
7371 default. In that case, GCC will allocate at least one TOC entry for
7372 each unique non-automatic variable reference in your program. GCC
7373 will also place floating-point constants in the TOC@. However, only
7374 16,384 entries are available in the TOC@.
7376 If you receive a linker error message that saying you have overflowed
7377 the available TOC space, you can reduce the amount of TOC space used
7378 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
7379 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
7380 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
7381 generate code to calculate the sum of an address and a constant at
7382 run-time instead of putting that sum into the TOC@. You may specify one
7383 or both of these options. Each causes GCC to produce very slightly
7384 slower and larger code at the expense of conserving TOC space.
7386 If you still run out of space in the TOC even when you specify both of
7387 these options, specify @option{-mminimal-toc} instead. This option causes
7388 GCC to make only one TOC entry for every file. When you specify this
7389 option, GCC will produce code that is slower and larger but which
7390 uses extremely little TOC space. You may wish to use this option
7391 only on files that contain less frequently executed code.
7397 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
7398 @code{long} type, and the infrastructure needed to support them.
7399 Specifying @option{-maix64} implies @option{-mpowerpc64} and
7400 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
7401 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
7406 @opindex mno-xl-call
7407 On AIX, pass floating-point arguments to prototyped functions beyond the
7408 register save area (RSA) on the stack in addition to argument FPRs. The
7409 AIX calling convention was extended but not initially documented to
7410 handle an obscure K&R C case of calling a function that takes the
7411 address of its arguments with fewer arguments than declared. AIX XL
7412 compilers access floating point arguments which do not fit in the
7413 RSA from the stack when a subroutine is compiled without
7414 optimization. Because always storing floating-point arguments on the
7415 stack is inefficient and rarely needed, this option is not enabled by
7416 default and only is necessary when calling subroutines compiled by AIX
7417 XL compilers without optimization.
7421 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
7422 application written to use message passing with special startup code to
7423 enable the application to run. The system must have PE installed in the
7424 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
7425 must be overridden with the @option{-specs=} option to specify the
7426 appropriate directory location. The Parallel Environment does not
7427 support threads, so the @option{-mpe} option and the @option{-pthread}
7428 option are incompatible.
7430 @item -malign-natural
7431 @itemx -malign-power
7432 @opindex malign-natural
7433 @opindex malign-power
7434 On AIX, Darwin, and 64-bit PowerPC Linux, the option
7435 @option{-malign-natural} overrides the ABI-defined alignment of larger
7436 types, such as floating-point doubles, on their natural size-based boundary.
7437 The option @option{-malign-power} instructs GCC to follow the ABI-specified
7438 alignment rules. GCC defaults to the standard alignment defined in the ABI.
7442 @opindex msoft-float
7443 @opindex mhard-float
7444 Generate code that does not use (uses) the floating-point register set.
7445 Software floating point emulation is provided if you use the
7446 @option{-msoft-float} option, and pass the option to GCC when linking.
7449 @itemx -mno-multiple
7451 @opindex mno-multiple
7452 Generate code that uses (does not use) the load multiple word
7453 instructions and the store multiple word instructions. These
7454 instructions are generated by default on POWER systems, and not
7455 generated on PowerPC systems. Do not use @option{-mmultiple} on little
7456 endian PowerPC systems, since those instructions do not work when the
7457 processor is in little endian mode. The exceptions are PPC740 and
7458 PPC750 which permit the instructions usage in little endian mode.
7464 Generate code that uses (does not use) the load string instructions
7465 and the store string word instructions to save multiple registers and
7466 do small block moves. These instructions are generated by default on
7467 POWER systems, and not generated on PowerPC systems. Do not use
7468 @option{-mstring} on little endian PowerPC systems, since those
7469 instructions do not work when the processor is in little endian mode.
7470 The exceptions are PPC740 and PPC750 which permit the instructions
7471 usage in little endian mode.
7477 Generate code that uses (does not use) the load or store instructions
7478 that update the base register to the address of the calculated memory
7479 location. These instructions are generated by default. If you use
7480 @option{-mno-update}, there is a small window between the time that the
7481 stack pointer is updated and the address of the previous frame is
7482 stored, which means code that walks the stack frame across interrupts or
7483 signals may get corrupted data.
7486 @itemx -mno-fused-madd
7487 @opindex mfused-madd
7488 @opindex mno-fused-madd
7489 Generate code that uses (does not use) the floating point multiply and
7490 accumulate instructions. These instructions are generated by default if
7491 hardware floating is used.
7493 @item -mno-bit-align
7495 @opindex mno-bit-align
7497 On System V.4 and embedded PowerPC systems do not (do) force structures
7498 and unions that contain bit-fields to be aligned to the base type of the
7501 For example, by default a structure containing nothing but 8
7502 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
7503 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
7504 the structure would be aligned to a 1 byte boundary and be one byte in
7507 @item -mno-strict-align
7508 @itemx -mstrict-align
7509 @opindex mno-strict-align
7510 @opindex mstrict-align
7511 On System V.4 and embedded PowerPC systems do not (do) assume that
7512 unaligned memory references will be handled by the system.
7515 @itemx -mno-relocatable
7516 @opindex mrelocatable
7517 @opindex mno-relocatable
7518 On embedded PowerPC systems generate code that allows (does not allow)
7519 the program to be relocated to a different address at runtime. If you
7520 use @option{-mrelocatable} on any module, all objects linked together must
7521 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
7523 @item -mrelocatable-lib
7524 @itemx -mno-relocatable-lib
7525 @opindex mrelocatable-lib
7526 @opindex mno-relocatable-lib
7527 On embedded PowerPC systems generate code that allows (does not allow)
7528 the program to be relocated to a different address at runtime. Modules
7529 compiled with @option{-mrelocatable-lib} can be linked with either modules
7530 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
7531 with modules compiled with the @option{-mrelocatable} options.
7537 On System V.4 and embedded PowerPC systems do not (do) assume that
7538 register 2 contains a pointer to a global area pointing to the addresses
7539 used in the program.
7542 @itemx -mlittle-endian
7544 @opindex mlittle-endian
7545 On System V.4 and embedded PowerPC systems compile code for the
7546 processor in little endian mode. The @option{-mlittle-endian} option is
7547 the same as @option{-mlittle}.
7552 @opindex mbig-endian
7553 On System V.4 and embedded PowerPC systems compile code for the
7554 processor in big endian mode. The @option{-mbig-endian} option is
7555 the same as @option{-mbig}.
7557 @item -mdynamic-no-pic
7558 @opindex mdynamic-no-pic
7559 On Darwin and Mac OS X systems, compile code so that it is not
7560 relocatable, but that its external references are relocatable. The
7561 resulting code is suitable for applications, but not shared
7564 @item -mprioritize-restricted-insns=@var{priority}
7565 @opindex mprioritize-restricted-insns
7566 This option controls the priority that is assigned to
7567 dispatch-slot restricted instructions during the second scheduling
7568 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
7569 @var{no/highest/second-highest} priority to dispatch slot restricted
7572 @item -msched-costly-dep=@var{dependence_type}
7573 @opindex msched-costly-dep
7574 This option controls which dependences are considered costly
7575 by the target during instruction scheduling. The argument
7576 @var{dependence_type} takes one of the following values:
7577 @var{no}: no dependence is costly,
7578 @var{all}: all dependences are costly,
7579 @var{true_store_to_load}: a true dependence from store to load is costly,
7580 @var{store_to_load}: any dependence from store to load is costly,
7581 @var{number}: any dependence which latency >= @var{number} is costly.
7583 @item -minsert-sched-nops=@var{scheme}
7584 @opindex minsert-sched-nops
7585 This option controls which nop insertion scheme will be used during
7586 the second scheduling pass. The argument @var{scheme} takes one of the
7588 @var{no}: Don't insert nops.
7589 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
7590 according to the scheduler's grouping.
7591 @var{regroup_exact}: Insert nops to force costly dependent insns into
7592 separate groups. Insert exactly as many nops as needed to force an insn
7593 to a new group, according to the estimatied processor grouping.
7594 @var{number}: Insert nops to force costly dependent insns into
7595 separate groups. Insert @var{number} nops to force an insn to a new group.
7599 On System V.4 and embedded PowerPC systems compile code using calling
7600 conventions that adheres to the March 1995 draft of the System V
7601 Application Binary Interface, PowerPC processor supplement. This is the
7602 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
7604 @item -mcall-sysv-eabi
7605 @opindex mcall-sysv-eabi
7606 Specify both @option{-mcall-sysv} and @option{-meabi} options.
7608 @item -mcall-sysv-noeabi
7609 @opindex mcall-sysv-noeabi
7610 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
7612 @item -mcall-solaris
7613 @opindex mcall-solaris
7614 On System V.4 and embedded PowerPC systems compile code for the Solaris
7618 @opindex mcall-linux
7619 On System V.4 and embedded PowerPC systems compile code for the
7620 Linux-based GNU system.
7624 On System V.4 and embedded PowerPC systems compile code for the
7625 Hurd-based GNU system.
7628 @opindex mcall-netbsd
7629 On System V.4 and embedded PowerPC systems compile code for the
7630 NetBSD operating system.
7632 @item -maix-struct-return
7633 @opindex maix-struct-return
7634 Return all structures in memory (as specified by the AIX ABI)@.
7636 @item -msvr4-struct-return
7637 @opindex msvr4-struct-return
7638 Return structures smaller than 8 bytes in registers (as specified by the
7642 @opindex mabi=altivec
7643 Extend the current ABI with AltiVec ABI extensions. This does not
7644 change the default ABI, instead it adds the AltiVec ABI extensions to
7647 @item -mabi=no-altivec
7648 @opindex mabi=no-altivec
7649 Disable AltiVec ABI extensions for the current ABI.
7652 @itemx -mno-prototype
7654 @opindex mno-prototype
7655 On System V.4 and embedded PowerPC systems assume that all calls to
7656 variable argument functions are properly prototyped. Otherwise, the
7657 compiler must insert an instruction before every non prototyped call to
7658 set or clear bit 6 of the condition code register (@var{CR}) to
7659 indicate whether floating point values were passed in the floating point
7660 registers in case the function takes a variable arguments. With
7661 @option{-mprototype}, only calls to prototyped variable argument functions
7662 will set or clear the bit.
7666 On embedded PowerPC systems, assume that the startup module is called
7667 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
7668 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
7673 On embedded PowerPC systems, assume that the startup module is called
7674 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
7679 On embedded PowerPC systems, assume that the startup module is called
7680 @file{crt0.o} and the standard C libraries are @file{libads.a} and
7684 @opindex myellowknife
7685 On embedded PowerPC systems, assume that the startup module is called
7686 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
7691 On System V.4 and embedded PowerPC systems, specify that you are
7692 compiling for a VxWorks system.
7696 Specify that you are compiling for the WindISS simulation environment.
7700 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7701 header to indicate that @samp{eabi} extended relocations are used.
7707 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7708 Embedded Applications Binary Interface (eabi) which is a set of
7709 modifications to the System V.4 specifications. Selecting @option{-meabi}
7710 means that the stack is aligned to an 8 byte boundary, a function
7711 @code{__eabi} is called to from @code{main} to set up the eabi
7712 environment, and the @option{-msdata} option can use both @code{r2} and
7713 @code{r13} to point to two separate small data areas. Selecting
7714 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7715 do not call an initialization function from @code{main}, and the
7716 @option{-msdata} option will only use @code{r13} to point to a single
7717 small data area. The @option{-meabi} option is on by default if you
7718 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7721 @opindex msdata=eabi
7722 On System V.4 and embedded PowerPC systems, put small initialized
7723 @code{const} global and static data in the @samp{.sdata2} section, which
7724 is pointed to by register @code{r2}. Put small initialized
7725 non-@code{const} global and static data in the @samp{.sdata} section,
7726 which is pointed to by register @code{r13}. Put small uninitialized
7727 global and static data in the @samp{.sbss} section, which is adjacent to
7728 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7729 incompatible with the @option{-mrelocatable} option. The
7730 @option{-msdata=eabi} option also sets the @option{-memb} option.
7733 @opindex msdata=sysv
7734 On System V.4 and embedded PowerPC systems, put small global and static
7735 data in the @samp{.sdata} section, which is pointed to by register
7736 @code{r13}. Put small uninitialized global and static data in the
7737 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7738 The @option{-msdata=sysv} option is incompatible with the
7739 @option{-mrelocatable} option.
7741 @item -msdata=default
7743 @opindex msdata=default
7745 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7746 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7747 same as @option{-msdata=sysv}.
7750 @opindex msdata-data
7751 On System V.4 and embedded PowerPC systems, put small global and static
7752 data in the @samp{.sdata} section. Put small uninitialized global and
7753 static data in the @samp{.sbss} section. Do not use register @code{r13}
7754 to address small data however. This is the default behavior unless
7755 other @option{-msdata} options are used.
7759 @opindex msdata=none
7761 On embedded PowerPC systems, put all initialized global and static data
7762 in the @samp{.data} section, and all uninitialized data in the
7763 @samp{.bss} section.
7767 @cindex smaller data references (PowerPC)
7768 @cindex .sdata/.sdata2 references (PowerPC)
7769 On embedded PowerPC systems, put global and static items less than or
7770 equal to @var{num} bytes into the small data or bss sections instead of
7771 the normal data or bss section. By default, @var{num} is 8. The
7772 @option{-G @var{num}} switch is also passed to the linker.
7773 All modules should be compiled with the same @option{-G @var{num}} value.
7776 @itemx -mno-regnames
7778 @opindex mno-regnames
7779 On System V.4 and embedded PowerPC systems do (do not) emit register
7780 names in the assembly language output using symbolic forms.
7783 @itemx -mno-longcall
7785 @opindex mno-longcall
7786 Default to making all function calls via pointers, so that functions
7787 which reside further than 64 megabytes (67,108,864 bytes) from the
7788 current location can be called. This setting can be overridden by the
7789 @code{shortcall} function attribute, or by @code{#pragma longcall(0)}.
7791 Some linkers are capable of detecting out-of-range calls and generating
7792 glue code on the fly. On these systems, long calls are unnecessary and
7793 generate slower code. As of this writing, the AIX linker can do this,
7794 as can the GNU linker for PowerPC/64. It is planned to add this feature
7795 to the GNU linker for 32-bit PowerPC systems as well.
7797 In the future, we may cause GCC to ignore all longcall specifications
7798 when the linker is known to generate glue.
7802 Adds support for multithreading with the @dfn{pthreads} library.
7803 This option sets flags for both the preprocessor and linker.
7807 @node Darwin Options
7808 @subsection Darwin Options
7809 @cindex Darwin options
7811 These options are defined for all architectures running the Darwin operating
7812 system. They are useful for compatibility with other Mac OS compilers.
7817 Loads all members of static archive libraries.
7818 See man ld(1) for more information.
7820 @item -arch_errors_fatal
7821 @opindex arch_errors_fatal
7822 Cause the errors having to do with files that have the wrong architecture
7826 @opindex bind_at_load
7827 Causes the output file to be marked such that the dynamic linker will
7828 bind all undefined references when the file is loaded or launched.
7832 Produce a Mach-o bundle format file.
7833 See man ld(1) for more information.
7835 @item -bundle_loader @var{executable}
7836 @opindex bundle_loader
7837 This specifies the @var{executable} that will be loading the build
7838 output file being linked. See man ld(1) for more information.
7840 @item -allowable_client @var{client_name}
7844 @item -compatibility_version
7845 @item -current_version
7846 @item -dependency-file
7848 @item -dylinker_install_name
7851 @item -exported_symbols_list
7853 @item -flat_namespace
7854 @item -force_cpusubtype_ALL
7855 @item -force_flat_namespace
7856 @item -headerpad_max_install_names
7860 @item -keep_private_externs
7862 @item -multiply_defined
7863 @item -multiply_defined_unused
7865 @item -nofixprebinding
7868 @item -noseglinkedit
7869 @item -pagezero_size
7871 @item -prebind_all_twolevel_modules
7872 @item -private_bundle
7873 @item -read_only_relocs
7875 @item -sectobjectsymbols
7879 @item -sectobjectsymbols
7881 @item -seg_addr_table
7882 @item -seg_addr_table_filename
7885 @item -segs_read_only_addr
7886 @item -segs_read_write_addr
7887 @item -single_module
7891 @item -twolevel_namespace
7894 @item -unexported_symbols_list
7895 @item -weak_reference_mismatches
7898 @opindex allowable_client
7900 @opindex client_name
7901 @opindex compatibility_version
7902 @opindex current_version
7903 @opindex dependency-file
7905 @opindex dylinker_install_name
7908 @opindex exported_symbols_list
7910 @opindex flat_namespace
7911 @opindex force_cpusubtype_ALL
7912 @opindex force_flat_namespace
7913 @opindex headerpad_max_install_names
7916 @opindex install_name
7917 @opindex keep_private_externs
7918 @opindex multi_module
7919 @opindex multiply_defined
7920 @opindex multiply_defined_unused
7922 @opindex nofixprebinding
7923 @opindex nomultidefs
7925 @opindex noseglinkedit
7926 @opindex pagezero_size
7928 @opindex prebind_all_twolevel_modules
7929 @opindex private_bundle
7930 @opindex read_only_relocs
7932 @opindex sectobjectsymbols
7936 @opindex sectobjectsymbols
7938 @opindex seg_addr_table
7939 @opindex seg_addr_table_filename
7940 @opindex seglinkedit
7942 @opindex segs_read_only_addr
7943 @opindex segs_read_write_addr
7944 @opindex single_module
7946 @opindex sub_library
7947 @opindex sub_umbrella
7948 @opindex twolevel_namespace
7951 @opindex unexported_symbols_list
7952 @opindex weak_reference_mismatches
7953 @opindex whatsloaded
7955 This options are available for Darwin linker. Darwin linker man page
7956 describes them in detail.
7961 @subsection IBM RT Options
7963 @cindex IBM RT options
7965 These @samp{-m} options are defined for the IBM RT PC:
7969 @opindex min-line-mul
7970 Use an in-line code sequence for integer multiplies. This is the
7973 @item -mcall-lib-mul
7974 @opindex mcall-lib-mul
7975 Call @code{lmul$$} for integer multiples.
7977 @item -mfull-fp-blocks
7978 @opindex mfull-fp-blocks
7979 Generate full-size floating point data blocks, including the minimum
7980 amount of scratch space recommended by IBM@. This is the default.
7982 @item -mminimum-fp-blocks
7983 @opindex mminimum-fp-blocks
7984 Do not include extra scratch space in floating point data blocks. This
7985 results in smaller code, but slower execution, since scratch space must
7986 be allocated dynamically.
7988 @cindex @file{stdarg.h} and RT PC
7989 @item -mfp-arg-in-fpregs
7990 @opindex mfp-arg-in-fpregs
7991 Use a calling sequence incompatible with the IBM calling convention in
7992 which floating point arguments are passed in floating point registers.
7993 Note that @code{stdarg.h} will not work with floating point operands
7994 if this option is specified.
7996 @item -mfp-arg-in-gregs
7997 @opindex mfp-arg-in-gregs
7998 Use the normal calling convention for floating point arguments. This is
8001 @item -mhc-struct-return
8002 @opindex mhc-struct-return
8003 Return structures of more than one word in memory, rather than in a
8004 register. This provides compatibility with the MetaWare HighC (hc)
8005 compiler. Use the option @option{-fpcc-struct-return} for compatibility
8006 with the Portable C Compiler (pcc).
8008 @item -mnohc-struct-return
8009 @opindex mnohc-struct-return
8010 Return some structures of more than one word in registers, when
8011 convenient. This is the default. For compatibility with the
8012 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
8013 option @option{-mhc-struct-return}.
8017 @subsection MIPS Options
8018 @cindex MIPS options
8020 These @samp{-m} options are defined for the MIPS family of computers:
8024 @item -march=@var{arch}
8026 Generate code that will run on @var{arch}, which can be the name of a
8027 generic MIPS ISA, or the name of a particular processor.
8029 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
8030 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
8031 The processor names are:
8032 @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
8034 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
8035 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000}, @samp{rm7000},
8039 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4300},
8040 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
8041 The special value @samp{from-abi} selects the
8042 most compatible architecture for the selected ABI (that is,
8043 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
8045 In processor names, a final @samp{000} can be abbreviated as @samp{k}
8046 (for example, @samp{-march=r2k}). Prefixes are optional, and
8047 @samp{vr} may be written @samp{r}.
8049 GCC defines two macros based on the value of this option. The first
8050 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
8051 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
8052 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
8053 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
8054 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
8056 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
8057 above. In other words, it will have the full prefix and will not
8058 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
8059 the macro names the resolved architecture (either @samp{"mips1"} or
8060 @samp{"mips3"}). It names the default architecture when no
8061 @option{-march} option is given.
8063 @item -mtune=@var{arch}
8065 Optimize for @var{arch}. Among other things, this option controls
8066 the way instructions are scheduled, and the perceived cost of arithmetic
8067 operations. The list of @var{arch} values is the same as for
8070 When this option is not used, GCC will optimize for the processor
8071 specified by @option{-march}. By using @option{-march} and
8072 @option{-mtune} together, it is possible to generate code that will
8073 run on a family of processors, but optimize the code for one
8074 particular member of that family.
8076 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
8077 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
8078 @samp{-march} ones described above.
8082 Equivalent to @samp{-march=mips1}.
8086 Equivalent to @samp{-march=mips2}.
8090 Equivalent to @samp{-march=mips3}.
8094 Equivalent to @samp{-march=mips4}.
8098 Equivalent to @samp{-march=mips32}.
8102 Equivalent to @samp{-march=mips32r2}.
8106 Equivalent to @samp{-march=mips64}.
8109 @itemx -mno-fused-madd
8110 @opindex mfused-madd
8111 @opindex mno-fused-madd
8112 Generate code that uses (does not use) the floating point multiply and
8113 accumulate instructions, when they are available. These instructions
8114 are generated by default if they are available, but this may be
8115 undesirable if the extra precision causes problems or on certain chips
8116 in the mode where denormals are rounded to zero where denormals
8117 generated by multiply and accumulate instructions cause exceptions
8122 Assume that floating point registers are 32 bits wide.
8126 Assume that floating point registers are 64 bits wide.
8130 Assume that general purpose registers are 32 bits wide.
8134 Assume that general purpose registers are 64 bits wide.
8138 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
8139 explanation of the default, and the width of pointers.
8143 Force long types to be 64 bits wide. See @option{-mlong32} for an
8144 explanation of the default, and the width of pointers.
8148 Force long, int, and pointer types to be 32 bits wide.
8150 The default size of ints, longs and pointers depends on the ABI@. All
8151 the supported ABIs use 32-bit ints. The n64 ABI uses 64-bit longs, as
8152 does the 64-bit Cygnus EABI; the others use 32-bit longs. Pointers
8153 are the same size as longs, or the same size as integer registers,
8154 whichever is smaller.
8168 Generate code for the given ABI@.
8170 Note that there are two embedded ABIs: @option{-mabi=eabi}
8171 selects the one defined by Cygnus while @option{-meabi=meabi}
8172 selects the one defined by MIPS@. Both these ABIs have
8173 32-bit and 64-bit variants. Normally, GCC will generate
8174 64-bit code when you select a 64-bit architecture, but you
8175 can use @option{-mgp32} to get 32-bit code instead.
8177 @item -mabi-fake-default
8178 @opindex mabi-fake-default
8179 You don't want to know what this option does. No, really. I mean
8180 it. Move on to the next option.
8182 What? You're still here? Oh, well@enddots{} Ok, here's the deal. GCC
8183 wants the default set of options to get the root of the multilib tree,
8184 and the shared library SONAMEs without any multilib-indicating
8185 suffixes. This is not convenience for @samp{mips64-linux-gnu}, since
8186 we want to default to the N32 ABI, while still being binary-compatible
8187 with @samp{mips-linux-gnu} if you stick to the O32 ABI@. Being
8188 binary-compatible means shared libraries should have the same SONAMEs,
8189 and libraries should live in the same location. Having O32 libraries
8190 in a sub-directory named say @file{o32} is not acceptable.
8192 So we trick GCC into believing that O32 is the default ABI, except
8193 that we override the default with some internal command-line
8194 processing magic. Problem is, if we stopped at that, and you then
8195 created a multilib-aware package that used the output of @command{gcc
8196 -print-multi-lib} to decide which multilibs to build, and how, and
8197 you'd find yourself in an awkward situation when you found out that
8198 some of the options listed ended up mapping to the same multilib, and
8199 none of your libraries was actually built for the multilib that
8200 @option{-print-multi-lib} claims to be the default. So we added this
8201 option that disables the default switcher, falling back to GCC's
8202 original notion of the default library. Confused yet?
8204 For short: don't ever use this option, unless you find it in the list
8205 of additional options to be used when building for multilibs, in the
8206 output of @option{gcc -print-multi-lib}.
8210 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
8211 add normal debug information. This is the default for all
8212 platforms except for the OSF/1 reference platform, using the OSF/rose
8213 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
8214 switches are used, the @file{mips-tfile} program will encapsulate the
8215 stabs within MIPS ECOFF@.
8219 Generate code for the GNU assembler. This is the default on the OSF/1
8220 reference platform, using the OSF/rose object format. Also, this is
8221 the default if the configure option @option{--with-gnu-as} is used.
8223 @item -msplit-addresses
8224 @itemx -mno-split-addresses
8225 @opindex msplit-addresses
8226 @opindex mno-split-addresses
8227 Generate code to load the high and low parts of address constants separately.
8228 This allows GCC to optimize away redundant loads of the high order
8229 bits of addresses. This optimization requires GNU as and GNU ld.
8230 This optimization is enabled by default for some embedded targets where
8231 GNU as and GNU ld are standard.
8237 The @option{-mrnames} switch says to output code using the MIPS software
8238 names for the registers, instead of the hardware names (ie, @var{a0}
8239 instead of @var{$4}). The only known assembler that supports this option
8240 is the Algorithmics assembler.
8246 The @option{-mmemcpy} switch makes all block moves call the appropriate
8247 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
8248 generating inline code.
8251 @itemx -mno-mips-tfile
8252 @opindex mmips-tfile
8253 @opindex mno-mips-tfile
8254 The @option{-mno-mips-tfile} switch causes the compiler not
8255 postprocess the object file with the @file{mips-tfile} program,
8256 after the MIPS assembler has generated it to add debug support. If
8257 @file{mips-tfile} is not run, then no local variables will be
8258 available to the debugger. In addition, @file{stage2} and
8259 @file{stage3} objects will have the temporary file names passed to the
8260 assembler embedded in the object file, which means the objects will
8261 not compare the same. The @option{-mno-mips-tfile} switch should only
8262 be used when there are bugs in the @file{mips-tfile} program that
8263 prevents compilation.
8266 @opindex msoft-float
8267 Generate output containing library calls for floating point.
8268 @strong{Warning:} the requisite libraries are not part of GCC@.
8269 Normally the facilities of the machine's usual C compiler are used, but
8270 this can't be done directly in cross-compilation. You must make your
8271 own arrangements to provide suitable library functions for
8275 @opindex mhard-float
8276 Generate output containing floating point instructions. This is the
8277 default if you use the unmodified sources.
8280 @itemx -mno-abicalls
8282 @opindex mno-abicalls
8283 Emit (or do not emit) the pseudo operations @samp{.abicalls},
8284 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
8285 position independent code.
8291 Lift (or do not lift) the usual restrictions on the size of the global
8294 GCC normally uses a single instruction to load values from the GOT.
8295 While this is relatively efficient, it will only work if the GOT
8296 is smaller than about 64k. Anything larger will cause the linker
8297 to report an error such as:
8299 @cindex relocation truncated to fit (MIPS)
8301 relocation truncated to fit: R_MIPS_GOT16 foobar
8304 If this happens, you should recompile your code with @option{-mxgot}.
8305 It should then work with very large GOTs, although it will also be
8306 less efficient, since it will take three instructions to fetch the
8307 value of a global symbol.
8309 Note that some linkers can create multiple GOTs. If you have such a
8310 linker, you should only need to use @option{-mxgot} when a single object
8311 file accesses more than 64k's worth of GOT entries. Very few do.
8313 These options have no effect unless GCC is generating position
8317 @itemx -mno-long-calls
8318 @opindex mlong-calls
8319 @opindex mno-long-calls
8320 Do all calls with the @samp{JALR} instruction, which requires
8321 loading up a function's address into a register before the call.
8322 You need to use this switch, if you call outside of the current
8323 512 megabyte segment to functions that are not through pointers.
8325 @item -membedded-pic
8326 @itemx -mno-embedded-pic
8327 @opindex membedded-pic
8328 @opindex mno-embedded-pic
8329 Generate PIC code suitable for some embedded systems. All calls are
8330 made using PC relative address, and all data is addressed using the $gp
8331 register. No more than 65536 bytes of global data may be used. This
8332 requires GNU as and GNU ld which do most of the work. This currently
8333 only works on targets which use ECOFF; it does not work with ELF@.
8335 @item -membedded-data
8336 @itemx -mno-embedded-data
8337 @opindex membedded-data
8338 @opindex mno-embedded-data
8339 Allocate variables to the read-only data section first if possible, then
8340 next in the small data section if possible, otherwise in data. This gives
8341 slightly slower code than the default, but reduces the amount of RAM required
8342 when executing, and thus may be preferred for some embedded systems.
8344 @item -muninit-const-in-rodata
8345 @itemx -mno-uninit-const-in-rodata
8346 @opindex muninit-const-in-rodata
8347 @opindex mno-uninit-const-in-rodata
8348 When used together with @option{-membedded-data}, it will always store uninitialized
8349 const variables in the read-only data section.
8351 @item -msingle-float
8352 @itemx -mdouble-float
8353 @opindex msingle-float
8354 @opindex mdouble-float
8355 The @option{-msingle-float} switch tells gcc to assume that the floating
8356 point coprocessor only supports single precision operations, as on the
8357 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
8358 double precision operations. This is the default.
8364 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
8365 as on the @samp{r4650} chip.
8369 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
8370 @option{-mcpu=r4650}.
8376 Enable 16-bit instructions.
8380 Compile code for the processor in little endian mode.
8381 The requisite libraries are assumed to exist.
8385 Compile code for the processor in big endian mode.
8386 The requisite libraries are assumed to exist.
8390 @cindex smaller data references (MIPS)
8391 @cindex gp-relative references (MIPS)
8392 Put global and static items less than or equal to @var{num} bytes into
8393 the small data or bss sections instead of the normal data or bss
8394 section. This allows the assembler to emit one word memory reference
8395 instructions based on the global pointer (@var{gp} or @var{$28}),
8396 instead of the normal two words used. By default, @var{num} is 8 when
8397 the MIPS assembler is used, and 0 when the GNU assembler is used. The
8398 @option{-G @var{num}} switch is also passed to the assembler and linker.
8399 All modules should be compiled with the same @option{-G @var{num}}
8404 Tell the MIPS assembler to not run its preprocessor over user
8405 assembler files (with a @samp{.s} suffix) when assembling them.
8409 Pass an option to gas which will cause nops to be inserted if
8410 the read of the destination register of an mfhi or mflo instruction
8411 occurs in the following two instructions.
8416 Work around certain SB-1 CPU core errata.
8417 (This flag currently works around the SB-1 revision 2
8418 ``F1'' and ``F2'' floating point errata.)
8422 Do not include the default crt0.
8424 @item -mflush-func=@var{func}
8425 @itemx -mno-flush-func
8426 @opindex mflush-func
8427 Specifies the function to call to flush the I and D caches, or to not
8428 call any such function. If called, the function must take the same
8429 arguments as the common @code{_flush_func()}, that is, the address of the
8430 memory range for which the cache is being flushed, the size of the
8431 memory range, and the number 3 (to flush both caches). The default
8432 depends on the target gcc was configured for, but commonly is either
8433 @samp{_flush_func} or @samp{__cpu_flush}.
8435 @item -mbranch-likely
8436 @itemx -mno-branch-likely
8437 @opindex mbranch-likely
8438 @opindex mno-branch-likely
8439 Enable or disable use of Branch Likely instructions, regardless of the
8440 default for the selected architecture. By default, Branch Likely
8441 instructions may be generated if they are supported by the selected
8442 architecture. An exception is for the MIPS32 and MIPS64 architectures
8443 and processors which implement those architectures; for those, Branch
8444 Likely instructions will not be generated by default because the MIPS32
8445 and MIPS64 architectures specifically deprecate their use.
8448 @node i386 and x86-64 Options
8449 @subsection Intel 386 and AMD x86-64 Options
8450 @cindex i386 Options
8451 @cindex x86-64 Options
8452 @cindex Intel 386 Options
8453 @cindex AMD x86-64 Options
8455 These @samp{-m} options are defined for the i386 and x86-64 family of
8459 @item -mtune=@var{cpu-type}
8461 Tune to @var{cpu-type} everything applicable about the generated code, except
8462 for the ABI and the set of available instructions. The choices for
8463 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
8464 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
8465 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
8466 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp},
8467 @samp{athlon-mp}, @samp{winchip-c6}, @samp{winchip2}, @samp{k8}, @samp{c3}
8470 While picking a specific @var{cpu-type} will schedule things appropriately
8471 for that particular chip, the compiler will not generate any code that
8472 does not run on the i386 without the @option{-march=@var{cpu-type}} option
8473 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
8474 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
8475 AMD chips as opposed to the Intel ones.
8477 @item -march=@var{cpu-type}
8479 Generate instructions for the machine type @var{cpu-type}. The choices
8480 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
8481 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
8483 @item -mcpu=@var{cpu-type}
8485 A deprecated synonym for @option{-mtune}.
8494 @opindex mpentiumpro
8495 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
8496 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
8497 These synonyms are deprecated.
8499 @item -mfpmath=@var{unit}
8501 generate floating point arithmetics for selected unit @var{unit}. the choices
8506 Use the standard 387 floating point coprocessor present majority of chips and
8507 emulated otherwise. Code compiled with this option will run almost everywhere.
8508 The temporary results are computed in 80bit precision instead of precision
8509 specified by the type resulting in slightly different results compared to most
8510 of other chips. See @option{-ffloat-store} for more detailed description.
8512 This is the default choice for i386 compiler.
8515 Use scalar floating point instructions present in the SSE instruction set.
8516 This instruction set is supported by Pentium3 and newer chips, in the AMD line
8517 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
8518 instruction set supports only single precision arithmetics, thus the double and
8519 extended precision arithmetics is still done using 387. Later version, present
8520 only in Pentium4 and the future AMD x86-64 chips supports double precision
8523 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
8524 @option{-msse2} switches to enable SSE extensions and make this option
8525 effective. For x86-64 compiler, these extensions are enabled by default.
8527 The resulting code should be considerably faster in majority of cases and avoid
8528 the numerical instability problems of 387 code, but may break some existing
8529 code that expects temporaries to be 80bit.
8531 This is the default choice for x86-64 compiler.
8534 Use all SSE extensions enabled by @option{-msse2} as well as the new
8535 SSE extensions in Prescott New Instructions. @option{-mpni} also
8536 enables 2 builtin functions, @code{__builtin_ia32_monitor} and
8537 @code{__builtin_ia32_mwait}, for new instructions @code{monitor} and
8541 Attempt to utilize both instruction sets at once. This effectively double the
8542 amount of available registers and on chips with separate execution units for
8543 387 and SSE the execution resources too. Use this option with care, as it is
8544 still experimental, because gcc register allocator does not model separate
8545 functional units well resulting in instable performance.
8548 @item -masm=@var{dialect}
8549 @opindex masm=@var{dialect}
8550 Output asm instructions using selected @var{dialect}. Supported choices are
8551 @samp{intel} or @samp{att} (the default one).
8556 @opindex mno-ieee-fp
8557 Control whether or not the compiler uses IEEE floating point
8558 comparisons. These handle correctly the case where the result of a
8559 comparison is unordered.
8562 @opindex msoft-float
8563 Generate output containing library calls for floating point.
8564 @strong{Warning:} the requisite libraries are not part of GCC@.
8565 Normally the facilities of the machine's usual C compiler are used, but
8566 this can't be done directly in cross-compilation. You must make your
8567 own arrangements to provide suitable library functions for
8570 On machines where a function returns floating point results in the 80387
8571 register stack, some floating point opcodes may be emitted even if
8572 @option{-msoft-float} is used.
8574 @item -mno-fp-ret-in-387
8575 @opindex mno-fp-ret-in-387
8576 Do not use the FPU registers for return values of functions.
8578 The usual calling convention has functions return values of types
8579 @code{float} and @code{double} in an FPU register, even if there
8580 is no FPU@. The idea is that the operating system should emulate
8583 The option @option{-mno-fp-ret-in-387} causes such values to be returned
8584 in ordinary CPU registers instead.
8586 @item -mno-fancy-math-387
8587 @opindex mno-fancy-math-387
8588 Some 387 emulators do not support the @code{sin}, @code{cos} and
8589 @code{sqrt} instructions for the 387. Specify this option to avoid
8590 generating those instructions. This option is the default on FreeBSD,
8591 OpenBSD and NetBSD@. This option is overridden when @option{-march}
8592 indicates that the target cpu will always have an FPU and so the
8593 instruction will not need emulation. As of revision 2.6.1, these
8594 instructions are not generated unless you also use the
8595 @option{-funsafe-math-optimizations} switch.
8597 @item -malign-double
8598 @itemx -mno-align-double
8599 @opindex malign-double
8600 @opindex mno-align-double
8601 Control whether GCC aligns @code{double}, @code{long double}, and
8602 @code{long long} variables on a two word boundary or a one word
8603 boundary. Aligning @code{double} variables on a two word boundary will
8604 produce code that runs somewhat faster on a @samp{Pentium} at the
8605 expense of more memory.
8607 @strong{Warning:} if you use the @option{-malign-double} switch,
8608 structures containing the above types will be aligned differently than
8609 the published application binary interface specifications for the 386
8610 and will not be binary compatible with structures in code compiled
8611 without that switch.
8613 @item -m96bit-long-double
8614 @item -m128bit-long-double
8615 @opindex m96bit-long-double
8616 @opindex m128bit-long-double
8617 These switches control the size of @code{long double} type. The i386
8618 application binary interface specifies the size to be 96 bits,
8619 so @option{-m96bit-long-double} is the default in 32 bit mode.
8621 Modern architectures (Pentium and newer) would prefer @code{long double}
8622 to be aligned to an 8 or 16 byte boundary. In arrays or structures
8623 conforming to the ABI, this would not be possible. So specifying a
8624 @option{-m128bit-long-double} will align @code{long double}
8625 to a 16 byte boundary by padding the @code{long double} with an additional
8628 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
8629 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
8631 Notice that neither of these options enable any extra precision over the x87
8632 standard of 80 bits for a @code{long double}.
8634 @strong{Warning:} if you override the default value for your target ABI, the
8635 structures and arrays containing @code{long double} will change their size as
8636 well as function calling convention for function taking @code{long double}
8637 will be modified. Hence they will not be binary compatible with arrays or
8638 structures in code compiled without that switch.
8642 @itemx -mno-svr3-shlib
8643 @opindex msvr3-shlib
8644 @opindex mno-svr3-shlib
8645 Control whether GCC places uninitialized local variables into the
8646 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
8647 into @code{bss}. These options are meaningful only on System V Release 3.
8651 Use a different function-calling convention, in which functions that
8652 take a fixed number of arguments return with the @code{ret} @var{num}
8653 instruction, which pops their arguments while returning. This saves one
8654 instruction in the caller since there is no need to pop the arguments
8657 You can specify that an individual function is called with this calling
8658 sequence with the function attribute @samp{stdcall}. You can also
8659 override the @option{-mrtd} option by using the function attribute
8660 @samp{cdecl}. @xref{Function Attributes}.
8662 @strong{Warning:} this calling convention is incompatible with the one
8663 normally used on Unix, so you cannot use it if you need to call
8664 libraries compiled with the Unix compiler.
8666 Also, you must provide function prototypes for all functions that
8667 take variable numbers of arguments (including @code{printf});
8668 otherwise incorrect code will be generated for calls to those
8671 In addition, seriously incorrect code will result if you call a
8672 function with too many arguments. (Normally, extra arguments are
8673 harmlessly ignored.)
8675 @item -mregparm=@var{num}
8677 Control how many registers are used to pass integer arguments. By
8678 default, no registers are used to pass arguments, and at most 3
8679 registers can be used. You can control this behavior for a specific
8680 function by using the function attribute @samp{regparm}.
8681 @xref{Function Attributes}.
8683 @strong{Warning:} if you use this switch, and
8684 @var{num} is nonzero, then you must build all modules with the same
8685 value, including any libraries. This includes the system libraries and
8688 @item -mpreferred-stack-boundary=@var{num}
8689 @opindex mpreferred-stack-boundary
8690 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
8691 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
8692 the default is 4 (16 bytes or 128 bits), except when optimizing for code
8693 size (@option{-Os}), in which case the default is the minimum correct
8694 alignment (4 bytes for x86, and 8 bytes for x86-64).
8696 On Pentium and PentiumPro, @code{double} and @code{long double} values
8697 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
8698 suffer significant run time performance penalties. On Pentium III, the
8699 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
8700 penalties if it is not 16 byte aligned.
8702 To ensure proper alignment of this values on the stack, the stack boundary
8703 must be as aligned as that required by any value stored on the stack.
8704 Further, every function must be generated such that it keeps the stack
8705 aligned. Thus calling a function compiled with a higher preferred
8706 stack boundary from a function compiled with a lower preferred stack
8707 boundary will most likely misalign the stack. It is recommended that
8708 libraries that use callbacks always use the default setting.
8710 This extra alignment does consume extra stack space, and generally
8711 increases code size. Code that is sensitive to stack space usage, such
8712 as embedded systems and operating system kernels, may want to reduce the
8713 preferred alignment to @option{-mpreferred-stack-boundary=2}.
8731 These switches enable or disable the use of built-in functions that allow
8732 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
8734 @xref{X86 Built-in Functions}, for details of the functions enabled
8735 and disabled by these switches.
8737 To have SSE/SSE2 instructions generated automatically from floating-point
8738 code, see @option{-mfpmath=sse}.
8741 @itemx -mno-push-args
8743 @opindex mno-push-args
8744 Use PUSH operations to store outgoing parameters. This method is shorter
8745 and usually equally fast as method using SUB/MOV operations and is enabled
8746 by default. In some cases disabling it may improve performance because of
8747 improved scheduling and reduced dependencies.
8749 @item -maccumulate-outgoing-args
8750 @opindex maccumulate-outgoing-args
8751 If enabled, the maximum amount of space required for outgoing arguments will be
8752 computed in the function prologue. This is faster on most modern CPUs
8753 because of reduced dependencies, improved scheduling and reduced stack usage
8754 when preferred stack boundary is not equal to 2. The drawback is a notable
8755 increase in code size. This switch implies @option{-mno-push-args}.
8759 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
8760 on thread-safe exception handling must compile and link all code with the
8761 @option{-mthreads} option. When compiling, @option{-mthreads} defines
8762 @option{-D_MT}; when linking, it links in a special thread helper library
8763 @option{-lmingwthrd} which cleans up per thread exception handling data.
8765 @item -mno-align-stringops
8766 @opindex mno-align-stringops
8767 Do not align destination of inlined string operations. This switch reduces
8768 code size and improves performance in case the destination is already aligned,
8769 but gcc don't know about it.
8771 @item -minline-all-stringops
8772 @opindex minline-all-stringops
8773 By default GCC inlines string operations only when destination is known to be
8774 aligned at least to 4 byte boundary. This enables more inlining, increase code
8775 size, but may improve performance of code that depends on fast memcpy, strlen
8776 and memset for short lengths.
8778 @item -momit-leaf-frame-pointer
8779 @opindex momit-leaf-frame-pointer
8780 Don't keep the frame pointer in a register for leaf functions. This
8781 avoids the instructions to save, set up and restore frame pointers and
8782 makes an extra register available in leaf functions. The option
8783 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8784 which might make debugging harder.
8786 @item -mtls-direct-seg-refs
8787 @itemx -mno-tls-direct-seg-refs
8788 @opindex mtls-direct-seg-refs
8789 Controls whether TLS variables may be accessed with offsets from the
8790 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
8791 or whether the thread base pointer must be added. Whether or not this
8792 is legal depends on the operating system, and whether it maps the
8793 segment to cover the entire TLS area.
8795 For systems that use GNU libc, the default is on.
8798 These @samp{-m} switches are supported in addition to the above
8799 on AMD x86-64 processors in 64-bit environments.
8806 Generate code for a 32-bit or 64-bit environment.
8807 The 32-bit environment sets int, long and pointer to 32 bits and
8808 generates code that runs on any i386 system.
8809 The 64-bit environment sets int to 32 bits and long and pointer
8810 to 64 bits and generates code for AMD's x86-64 architecture.
8813 @opindex no-red-zone
8814 Do not use a so called red zone for x86-64 code. The red zone is mandated
8815 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8816 stack pointer that will not be modified by signal or interrupt handlers
8817 and therefore can be used for temporary data without adjusting the stack
8818 pointer. The flag @option{-mno-red-zone} disables this red zone.
8820 @item -mcmodel=small
8821 @opindex mcmodel=small
8822 Generate code for the small code model: the program and its symbols must
8823 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
8824 Programs can be statically or dynamically linked. This is the default
8827 @item -mcmodel=kernel
8828 @opindex mcmodel=kernel
8829 Generate code for the kernel code model. The kernel runs in the
8830 negative 2 GB of the address space.
8831 This model has to be used for Linux kernel code.
8833 @item -mcmodel=medium
8834 @opindex mcmodel=medium
8835 Generate code for the medium model: The program is linked in the lower 2
8836 GB of the address space but symbols can be located anywhere in the
8837 address space. Programs can be statically or dynamically linked, but
8838 building of shared libraries are not supported with the medium model.
8840 @item -mcmodel=large
8841 @opindex mcmodel=large
8842 Generate code for the large model: This model makes no assumptions
8843 about addresses and sizes of sections. Currently GCC does not implement
8848 @subsection HPPA Options
8849 @cindex HPPA Options
8851 These @samp{-m} options are defined for the HPPA family of computers:
8854 @item -march=@var{architecture-type}
8856 Generate code for the specified architecture. The choices for
8857 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8858 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8859 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8860 architecture option for your machine. Code compiled for lower numbered
8861 architectures will run on higher numbered architectures, but not the
8864 PA 2.0 support currently requires gas snapshot 19990413 or later. The
8865 next release of binutils (current is 2.9.1) will probably contain PA 2.0
8869 @itemx -mpa-risc-1-1
8870 @itemx -mpa-risc-2-0
8871 @opindex mpa-risc-1-0
8872 @opindex mpa-risc-1-1
8873 @opindex mpa-risc-2-0
8874 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8877 @opindex mbig-switch
8878 Generate code suitable for big switch tables. Use this option only if
8879 the assembler/linker complain about out of range branches within a switch
8882 @item -mjump-in-delay
8883 @opindex mjump-in-delay
8884 Fill delay slots of function calls with unconditional jump instructions
8885 by modifying the return pointer for the function call to be the target
8886 of the conditional jump.
8888 @item -mdisable-fpregs
8889 @opindex mdisable-fpregs
8890 Prevent floating point registers from being used in any manner. This is
8891 necessary for compiling kernels which perform lazy context switching of
8892 floating point registers. If you use this option and attempt to perform
8893 floating point operations, the compiler will abort.
8895 @item -mdisable-indexing
8896 @opindex mdisable-indexing
8897 Prevent the compiler from using indexing address modes. This avoids some
8898 rather obscure problems when compiling MIG generated code under MACH@.
8900 @item -mno-space-regs
8901 @opindex mno-space-regs
8902 Generate code that assumes the target has no space registers. This allows
8903 GCC to generate faster indirect calls and use unscaled index address modes.
8905 Such code is suitable for level 0 PA systems and kernels.
8907 @item -mfast-indirect-calls
8908 @opindex mfast-indirect-calls
8909 Generate code that assumes calls never cross space boundaries. This
8910 allows GCC to emit code which performs faster indirect calls.
8912 This option will not work in the presence of shared libraries or nested
8915 @item -mlong-load-store
8916 @opindex mlong-load-store
8917 Generate 3-instruction load and store sequences as sometimes required by
8918 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8921 @item -mportable-runtime
8922 @opindex mportable-runtime
8923 Use the portable calling conventions proposed by HP for ELF systems.
8927 Enable the use of assembler directives only GAS understands.
8929 @item -mschedule=@var{cpu-type}
8931 Schedule code according to the constraints for the machine type
8932 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8933 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8934 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8935 proper scheduling option for your machine. The default scheduling is
8939 @opindex mlinker-opt
8940 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8941 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8942 linkers in which they give bogus error messages when linking some programs.
8945 @opindex msoft-float
8946 Generate output containing library calls for floating point.
8947 @strong{Warning:} the requisite libraries are not available for all HPPA
8948 targets. Normally the facilities of the machine's usual C compiler are
8949 used, but this cannot be done directly in cross-compilation. You must make
8950 your own arrangements to provide suitable library functions for
8951 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8952 does provide software floating point support.
8954 @option{-msoft-float} changes the calling convention in the output file;
8955 therefore, it is only useful if you compile @emph{all} of a program with
8956 this option. In particular, you need to compile @file{libgcc.a}, the
8957 library that comes with GCC, with @option{-msoft-float} in order for
8962 Generate the predefine, @code{_SIO}, for server IO. The default is
8963 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
8964 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO. These
8965 options are available under HP-UX and HI-UX.
8969 Use GNU ld specific options. This passes @option{-shared} to ld when
8970 building a shared library. It is the default when GCC is configured,
8971 explicitly or implicitly, with the GNU linker. This option does not
8972 have any affect on which ld is called, it only changes what parameters
8973 are passed to that ld. The ld that is called is determined by the
8974 @option{--with-ld} configure option, gcc's program search path, and
8975 finally by the user's @env{PATH}. The linker used by GCC can be printed
8976 using @samp{which `gcc -print-prog-name=ld`}.
8980 Use HP ld specific options. This passes @option{-b} to ld when building
8981 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8982 links. It is the default when GCC is configured, explicitly or
8983 implicitly, with the HP linker. This option does not have any affect on
8984 which ld is called, it only changes what parameters are passed to that
8985 ld. The ld that is called is determined by the @option{--with-ld}
8986 configure option, gcc's program search path, and finally by the user's
8987 @env{PATH}. The linker used by GCC can be printed using @samp{which
8988 `gcc -print-prog-name=ld`}.
8991 @opindex mno-long-calls
8992 Generate code that uses long call sequences. This ensures that a call
8993 is always able to reach linker generated stubs. The default is to generate
8994 long calls only when the distance from the call site to the beginning
8995 of the function or translation unit, as the case may be, exceeds a
8996 predefined limit set by the branch type being used. The limits for
8997 normal calls are 7,600,000 and 240,000 bytes, respectively for the
8998 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
9001 Distances are measured from the beginning of functions when using the
9002 @option{-ffunction-sections} option, or when using the @option{-mgas}
9003 and @option{-mno-portable-runtime} options together under HP-UX with
9006 It is normally not desirable to use this option as it will degrade
9007 performance. However, it may be useful in large applications,
9008 particularly when partial linking is used to build the application.
9010 The types of long calls used depends on the capabilities of the
9011 assembler and linker, and the type of code being generated. The
9012 impact on systems that support long absolute calls, and long pic
9013 symbol-difference or pc-relative calls should be relatively small.
9014 However, an indirect call is used on 32-bit ELF systems in pic code
9015 and it is quite long.
9019 Suppress the generation of link options to search libdld.sl when the
9020 @option{-static} option is specified on HP-UX 10 and later.
9024 The HP-UX implementation of setlocale in libc has a dependency on
9025 libdld.sl. There isn't an archive version of libdld.sl. Thus,
9026 when the @option{-static} option is specified, special link options
9027 are needed to resolve this dependency.
9029 On HP-UX 10 and later, the GCC driver adds the necessary options to
9030 link with libdld.sl when the @option{-static} option is specified.
9031 This causes the resulting binary to be dynamic. On the 64-bit port,
9032 the linkers generate dynamic binaries by default in any case. The
9033 @option{-nolibdld} option can be used to prevent the GCC driver from
9034 adding these link options.
9038 Add support for multithreading with the @dfn{dce thread} library
9039 under HP-UX. This option sets flags for both the preprocessor and
9043 @node Intel 960 Options
9044 @subsection Intel 960 Options
9046 These @samp{-m} options are defined for the Intel 960 implementations:
9049 @item -m@var{cpu-type}
9057 Assume the defaults for the machine type @var{cpu-type} for some of
9058 the other options, including instruction scheduling, floating point
9059 support, and addressing modes. The choices for @var{cpu-type} are
9060 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
9061 @samp{sa}, and @samp{sb}.
9068 @opindex msoft-float
9069 The @option{-mnumerics} option indicates that the processor does support
9070 floating-point instructions. The @option{-msoft-float} option indicates
9071 that floating-point support should not be assumed.
9073 @item -mleaf-procedures
9074 @itemx -mno-leaf-procedures
9075 @opindex mleaf-procedures
9076 @opindex mno-leaf-procedures
9077 Do (or do not) attempt to alter leaf procedures to be callable with the
9078 @code{bal} instruction as well as @code{call}. This will result in more
9079 efficient code for explicit calls when the @code{bal} instruction can be
9080 substituted by the assembler or linker, but less efficient code in other
9081 cases, such as calls via function pointers, or using a linker that doesn't
9082 support this optimization.
9085 @itemx -mno-tail-call
9087 @opindex mno-tail-call
9088 Do (or do not) make additional attempts (beyond those of the
9089 machine-independent portions of the compiler) to optimize tail-recursive
9090 calls into branches. You may not want to do this because the detection of
9091 cases where this is not valid is not totally complete. The default is
9092 @option{-mno-tail-call}.
9094 @item -mcomplex-addr
9095 @itemx -mno-complex-addr
9096 @opindex mcomplex-addr
9097 @opindex mno-complex-addr
9098 Assume (or do not assume) that the use of a complex addressing mode is a
9099 win on this implementation of the i960. Complex addressing modes may not
9100 be worthwhile on the K-series, but they definitely are on the C-series.
9101 The default is currently @option{-mcomplex-addr} for all processors except
9105 @itemx -mno-code-align
9106 @opindex mcode-align
9107 @opindex mno-code-align
9108 Align code to 8-byte boundaries for faster fetching (or don't bother).
9109 Currently turned on by default for C-series implementations only.
9112 @item -mclean-linkage
9113 @itemx -mno-clean-linkage
9114 @opindex mclean-linkage
9115 @opindex mno-clean-linkage
9116 These options are not fully implemented.
9120 @itemx -mic2.0-compat
9121 @itemx -mic3.0-compat
9123 @opindex mic2.0-compat
9124 @opindex mic3.0-compat
9125 Enable compatibility with iC960 v2.0 or v3.0.
9129 @opindex masm-compat
9131 Enable compatibility with the iC960 assembler.
9133 @item -mstrict-align
9134 @itemx -mno-strict-align
9135 @opindex mstrict-align
9136 @opindex mno-strict-align
9137 Do not permit (do permit) unaligned accesses.
9141 Enable structure-alignment compatibility with Intel's gcc release version
9142 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
9144 @item -mlong-double-64
9145 @opindex mlong-double-64
9146 Implement type @samp{long double} as 64-bit floating point numbers.
9147 Without the option @samp{long double} is implemented by 80-bit
9148 floating point numbers. The only reason we have it because there is
9149 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
9150 is only useful for people using soft-float targets. Otherwise, we
9151 should recommend against use of it.
9155 @node DEC Alpha Options
9156 @subsection DEC Alpha Options
9158 These @samp{-m} options are defined for the DEC Alpha implementations:
9161 @item -mno-soft-float
9163 @opindex mno-soft-float
9164 @opindex msoft-float
9165 Use (do not use) the hardware floating-point instructions for
9166 floating-point operations. When @option{-msoft-float} is specified,
9167 functions in @file{libgcc.a} will be used to perform floating-point
9168 operations. Unless they are replaced by routines that emulate the
9169 floating-point operations, or compiled in such a way as to call such
9170 emulations routines, these routines will issue floating-point
9171 operations. If you are compiling for an Alpha without floating-point
9172 operations, you must ensure that the library is built so as not to call
9175 Note that Alpha implementations without floating-point operations are
9176 required to have floating-point registers.
9181 @opindex mno-fp-regs
9182 Generate code that uses (does not use) the floating-point register set.
9183 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
9184 register set is not used, floating point operands are passed in integer
9185 registers as if they were integers and floating-point results are passed
9186 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
9187 so any function with a floating-point argument or return value called by code
9188 compiled with @option{-mno-fp-regs} must also be compiled with that
9191 A typical use of this option is building a kernel that does not use,
9192 and hence need not save and restore, any floating-point registers.
9196 The Alpha architecture implements floating-point hardware optimized for
9197 maximum performance. It is mostly compliant with the IEEE floating
9198 point standard. However, for full compliance, software assistance is
9199 required. This option generates code fully IEEE compliant code
9200 @emph{except} that the @var{inexact-flag} is not maintained (see below).
9201 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
9202 defined during compilation. The resulting code is less efficient but is
9203 able to correctly support denormalized numbers and exceptional IEEE
9204 values such as not-a-number and plus/minus infinity. Other Alpha
9205 compilers call this option @option{-ieee_with_no_inexact}.
9207 @item -mieee-with-inexact
9208 @opindex mieee-with-inexact
9209 This is like @option{-mieee} except the generated code also maintains
9210 the IEEE @var{inexact-flag}. Turning on this option causes the
9211 generated code to implement fully-compliant IEEE math. In addition to
9212 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
9213 macro. On some Alpha implementations the resulting code may execute
9214 significantly slower than the code generated by default. Since there is
9215 very little code that depends on the @var{inexact-flag}, you should
9216 normally not specify this option. Other Alpha compilers call this
9217 option @option{-ieee_with_inexact}.
9219 @item -mfp-trap-mode=@var{trap-mode}
9220 @opindex mfp-trap-mode
9221 This option controls what floating-point related traps are enabled.
9222 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9223 The trap mode can be set to one of four values:
9227 This is the default (normal) setting. The only traps that are enabled
9228 are the ones that cannot be disabled in software (e.g., division by zero
9232 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9236 Like @samp{su}, but the instructions are marked to be safe for software
9237 completion (see Alpha architecture manual for details).
9240 Like @samp{su}, but inexact traps are enabled as well.
9243 @item -mfp-rounding-mode=@var{rounding-mode}
9244 @opindex mfp-rounding-mode
9245 Selects the IEEE rounding mode. Other Alpha compilers call this option
9246 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9251 Normal IEEE rounding mode. Floating point numbers are rounded towards
9252 the nearest machine number or towards the even machine number in case
9256 Round towards minus infinity.
9259 Chopped rounding mode. Floating point numbers are rounded towards zero.
9262 Dynamic rounding mode. A field in the floating point control register
9263 (@var{fpcr}, see Alpha architecture reference manual) controls the
9264 rounding mode in effect. The C library initializes this register for
9265 rounding towards plus infinity. Thus, unless your program modifies the
9266 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9269 @item -mtrap-precision=@var{trap-precision}
9270 @opindex mtrap-precision
9271 In the Alpha architecture, floating point traps are imprecise. This
9272 means without software assistance it is impossible to recover from a
9273 floating trap and program execution normally needs to be terminated.
9274 GCC can generate code that can assist operating system trap handlers
9275 in determining the exact location that caused a floating point trap.
9276 Depending on the requirements of an application, different levels of
9277 precisions can be selected:
9281 Program precision. This option is the default and means a trap handler
9282 can only identify which program caused a floating point exception.
9285 Function precision. The trap handler can determine the function that
9286 caused a floating point exception.
9289 Instruction precision. The trap handler can determine the exact
9290 instruction that caused a floating point exception.
9293 Other Alpha compilers provide the equivalent options called
9294 @option{-scope_safe} and @option{-resumption_safe}.
9296 @item -mieee-conformant
9297 @opindex mieee-conformant
9298 This option marks the generated code as IEEE conformant. You must not
9299 use this option unless you also specify @option{-mtrap-precision=i} and either
9300 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9301 is to emit the line @samp{.eflag 48} in the function prologue of the
9302 generated assembly file. Under DEC Unix, this has the effect that
9303 IEEE-conformant math library routines will be linked in.
9305 @item -mbuild-constants
9306 @opindex mbuild-constants
9307 Normally GCC examines a 32- or 64-bit integer constant to
9308 see if it can construct it from smaller constants in two or three
9309 instructions. If it cannot, it will output the constant as a literal and
9310 generate code to load it from the data segment at runtime.
9312 Use this option to require GCC to construct @emph{all} integer constants
9313 using code, even if it takes more instructions (the maximum is six).
9315 You would typically use this option to build a shared library dynamic
9316 loader. Itself a shared library, it must relocate itself in memory
9317 before it can find the variables and constants in its own data segment.
9323 Select whether to generate code to be assembled by the vendor-supplied
9324 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9342 Indicate whether GCC should generate code to use the optional BWX,
9343 CIX, FIX and MAX instruction sets. The default is to use the instruction
9344 sets supported by the CPU type specified via @option{-mcpu=} option or that
9345 of the CPU on which GCC was built if none was specified.
9350 @opindex mfloat-ieee
9351 Generate code that uses (does not use) VAX F and G floating point
9352 arithmetic instead of IEEE single and double precision.
9354 @item -mexplicit-relocs
9355 @itemx -mno-explicit-relocs
9356 @opindex mexplicit-relocs
9357 @opindex mno-explicit-relocs
9358 Older Alpha assemblers provided no way to generate symbol relocations
9359 except via assembler macros. Use of these macros does not allow
9360 optimal instruction scheduling. GNU binutils as of version 2.12
9361 supports a new syntax that allows the compiler to explicitly mark
9362 which relocations should apply to which instructions. This option
9363 is mostly useful for debugging, as GCC detects the capabilities of
9364 the assembler when it is built and sets the default accordingly.
9368 @opindex msmall-data
9369 @opindex mlarge-data
9370 When @option{-mexplicit-relocs} is in effect, static data is
9371 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9372 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9373 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9374 16-bit relocations off of the @code{$gp} register. This limits the
9375 size of the small data area to 64KB, but allows the variables to be
9376 directly accessed via a single instruction.
9378 The default is @option{-mlarge-data}. With this option the data area
9379 is limited to just below 2GB. Programs that require more than 2GB of
9380 data must use @code{malloc} or @code{mmap} to allocate the data in the
9381 heap instead of in the program's data segment.
9383 When generating code for shared libraries, @option{-fpic} implies
9384 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9388 @opindex msmall-text
9389 @opindex mlarge-text
9390 When @option{-msmall-text} is used, the compiler assumes that the
9391 code of the entire program (or shared library) fits in 4MB, and is
9392 thus reachable with a branch instruction. When @option{-msmall-data}
9393 is used, the compiler can assume that all local symbols share the
9394 same @code{$gp} value, and thus reduce the number of instructions
9395 required for a function call from 4 to 1.
9397 The default is @option{-mlarge-text}.
9399 @item -mcpu=@var{cpu_type}
9401 Set the instruction set and instruction scheduling parameters for
9402 machine type @var{cpu_type}. You can specify either the @samp{EV}
9403 style name or the corresponding chip number. GCC supports scheduling
9404 parameters for the EV4, EV5 and EV6 family of processors and will
9405 choose the default values for the instruction set from the processor
9406 you specify. If you do not specify a processor type, GCC will default
9407 to the processor on which the compiler was built.
9409 Supported values for @var{cpu_type} are
9415 Schedules as an EV4 and has no instruction set extensions.
9419 Schedules as an EV5 and has no instruction set extensions.
9423 Schedules as an EV5 and supports the BWX extension.
9428 Schedules as an EV5 and supports the BWX and MAX extensions.
9432 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9436 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9439 @item -mtune=@var{cpu_type}
9441 Set only the instruction scheduling parameters for machine type
9442 @var{cpu_type}. The instruction set is not changed.
9444 @item -mmemory-latency=@var{time}
9445 @opindex mmemory-latency
9446 Sets the latency the scheduler should assume for typical memory
9447 references as seen by the application. This number is highly
9448 dependent on the memory access patterns used by the application
9449 and the size of the external cache on the machine.
9451 Valid options for @var{time} are
9455 A decimal number representing clock cycles.
9461 The compiler contains estimates of the number of clock cycles for
9462 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9463 (also called Dcache, Scache, and Bcache), as well as to main memory.
9464 Note that L3 is only valid for EV5.
9469 @node DEC Alpha/VMS Options
9470 @subsection DEC Alpha/VMS Options
9472 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9475 @item -mvms-return-codes
9476 @opindex mvms-return-codes
9477 Return VMS condition codes from main. The default is to return POSIX
9478 style condition (e.g.@ error) codes.
9481 @node H8/300 Options
9482 @subsection H8/300 Options
9484 These @samp{-m} options are defined for the H8/300 implementations:
9489 Shorten some address references at link time, when possible; uses the
9490 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9491 ld.info, Using ld}, for a fuller description.
9495 Generate code for the H8/300H@.
9499 Generate code for the H8S@.
9503 Generate code for the H8S and H8/300H in the normal mode. This switch
9504 must be used either with -mh or -ms.
9508 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9512 Make @code{int} data 32 bits by default.
9516 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9517 The default for the H8/300H and H8S is to align longs and floats on 4
9519 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9520 This option has no effect on the H8/300.
9524 @subsection SH Options
9526 These @samp{-m} options are defined for the SH implementations:
9531 Generate code for the SH1.
9535 Generate code for the SH2.
9538 Generate code for the SH2e.
9542 Generate code for the SH3.
9546 Generate code for the SH3e.
9550 Generate code for the SH4 without a floating-point unit.
9552 @item -m4-single-only
9553 @opindex m4-single-only
9554 Generate code for the SH4 with a floating-point unit that only
9555 supports single-precision arithmetic.
9559 Generate code for the SH4 assuming the floating-point unit is in
9560 single-precision mode by default.
9564 Generate code for the SH4.
9568 Compile code for the processor in big endian mode.
9572 Compile code for the processor in little endian mode.
9576 Align doubles at 64-bit boundaries. Note that this changes the calling
9577 conventions, and thus some functions from the standard C library will
9578 not work unless you recompile it first with @option{-mdalign}.
9582 Shorten some address references at link time, when possible; uses the
9583 linker option @option{-relax}.
9587 Use 32-bit offsets in @code{switch} tables. The default is to use
9592 Enable the use of the instruction @code{fmovd}.
9596 Comply with the calling conventions defined by Renesas.
9600 Mark the @code{MAC} register as call-clobbered, even if
9601 @option{-mhitachi} is given.
9605 Increase IEEE-compliance of floating-point code.
9609 Dump instruction size and location in the assembly code.
9613 This option is deprecated. It pads structures to multiple of 4 bytes,
9614 which is incompatible with the SH ABI@.
9618 Optimize for space instead of speed. Implied by @option{-Os}.
9622 When generating position-independent code, emit function calls using
9623 the Global Offset Table instead of the Procedure Linkage Table.
9627 Generate a library function call to invalidate instruction cache
9628 entries, after fixing up a trampoline. This library function call
9629 doesn't assume it can write to the whole memory address space. This
9630 is the default when the target is @code{sh-*-linux*}.
9633 @node System V Options
9634 @subsection Options for System V
9636 These additional options are available on System V Release 4 for
9637 compatibility with other compilers on those systems:
9642 Create a shared object.
9643 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
9647 Identify the versions of each tool used by the compiler, in a
9648 @code{.ident} assembler directive in the output.
9652 Refrain from adding @code{.ident} directives to the output file (this is
9655 @item -YP,@var{dirs}
9657 Search the directories @var{dirs}, and no others, for libraries
9658 specified with @option{-l}.
9662 Look in the directory @var{dir} to find the M4 preprocessor.
9663 The assembler uses this option.
9664 @c This is supposed to go with a -Yd for predefined M4 macro files, but
9665 @c the generic assembler that comes with Solaris takes just -Ym.
9668 @node TMS320C3x/C4x Options
9669 @subsection TMS320C3x/C4x Options
9670 @cindex TMS320C3x/C4x Options
9672 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
9676 @item -mcpu=@var{cpu_type}
9678 Set the instruction set, register set, and instruction scheduling
9679 parameters for machine type @var{cpu_type}. Supported values for
9680 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
9681 @samp{c44}. The default is @samp{c40} to generate code for the
9686 @itemx -msmall-memory
9688 @opindex mbig-memory
9690 @opindex msmall-memory
9692 Generates code for the big or small memory model. The small memory
9693 model assumed that all data fits into one 64K word page. At run-time
9694 the data page (DP) register must be set to point to the 64K page
9695 containing the .bss and .data program sections. The big memory model is
9696 the default and requires reloading of the DP register for every direct
9703 Allow (disallow) allocation of general integer operands into the block
9710 Enable (disable) generation of code using decrement and branch,
9711 DBcond(D), instructions. This is enabled by default for the C4x. To be
9712 on the safe side, this is disabled for the C3x, since the maximum
9713 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
9714 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
9715 that it can utilize the decrement and branch instruction, but will give
9716 up if there is more than one memory reference in the loop. Thus a loop
9717 where the loop counter is decremented can generate slightly more
9718 efficient code, in cases where the RPTB instruction cannot be utilized.
9720 @item -mdp-isr-reload
9722 @opindex mdp-isr-reload
9724 Force the DP register to be saved on entry to an interrupt service
9725 routine (ISR), reloaded to point to the data section, and restored on
9726 exit from the ISR@. This should not be required unless someone has
9727 violated the small memory model by modifying the DP register, say within
9734 For the C3x use the 24-bit MPYI instruction for integer multiplies
9735 instead of a library call to guarantee 32-bit results. Note that if one
9736 of the operands is a constant, then the multiplication will be performed
9737 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
9738 then squaring operations are performed inline instead of a library call.
9741 @itemx -mno-fast-fix
9743 @opindex mno-fast-fix
9744 The C3x/C4x FIX instruction to convert a floating point value to an
9745 integer value chooses the nearest integer less than or equal to the
9746 floating point value rather than to the nearest integer. Thus if the
9747 floating point number is negative, the result will be incorrectly
9748 truncated an additional code is necessary to detect and correct this
9749 case. This option can be used to disable generation of the additional
9750 code required to correct the result.
9756 Enable (disable) generation of repeat block sequences using the RPTB
9757 instruction for zero overhead looping. The RPTB construct is only used
9758 for innermost loops that do not call functions or jump across the loop
9759 boundaries. There is no advantage having nested RPTB loops due to the
9760 overhead required to save and restore the RC, RS, and RE registers.
9761 This is enabled by default with @option{-O2}.
9763 @item -mrpts=@var{count}
9767 Enable (disable) the use of the single instruction repeat instruction
9768 RPTS@. If a repeat block contains a single instruction, and the loop
9769 count can be guaranteed to be less than the value @var{count}, GCC will
9770 emit a RPTS instruction instead of a RPTB@. If no value is specified,
9771 then a RPTS will be emitted even if the loop count cannot be determined
9772 at compile time. Note that the repeated instruction following RPTS does
9773 not have to be reloaded from memory each iteration, thus freeing up the
9774 CPU buses for operands. However, since interrupts are blocked by this
9775 instruction, it is disabled by default.
9777 @item -mloop-unsigned
9778 @itemx -mno-loop-unsigned
9779 @opindex mloop-unsigned
9780 @opindex mno-loop-unsigned
9781 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
9782 is @math{2^{31} + 1} since these instructions test if the iteration count is
9783 negative to terminate the loop. If the iteration count is unsigned
9784 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
9785 exceeded. This switch allows an unsigned iteration count.
9789 Try to emit an assembler syntax that the TI assembler (asm30) is happy
9790 with. This also enforces compatibility with the API employed by the TI
9791 C3x C compiler. For example, long doubles are passed as structures
9792 rather than in floating point registers.
9798 Generate code that uses registers (stack) for passing arguments to functions.
9799 By default, arguments are passed in registers where possible rather
9800 than by pushing arguments on to the stack.
9802 @item -mparallel-insns
9803 @itemx -mno-parallel-insns
9804 @opindex mparallel-insns
9805 @opindex mno-parallel-insns
9806 Allow the generation of parallel instructions. This is enabled by
9807 default with @option{-O2}.
9809 @item -mparallel-mpy
9810 @itemx -mno-parallel-mpy
9811 @opindex mparallel-mpy
9812 @opindex mno-parallel-mpy
9813 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
9814 provided @option{-mparallel-insns} is also specified. These instructions have
9815 tight register constraints which can pessimize the code generation
9821 @subsection V850 Options
9822 @cindex V850 Options
9824 These @samp{-m} options are defined for V850 implementations:
9828 @itemx -mno-long-calls
9829 @opindex mlong-calls
9830 @opindex mno-long-calls
9831 Treat all calls as being far away (near). If calls are assumed to be
9832 far away, the compiler will always load the functions address up into a
9833 register, and call indirect through the pointer.
9839 Do not optimize (do optimize) basic blocks that use the same index
9840 pointer 4 or more times to copy pointer into the @code{ep} register, and
9841 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
9842 option is on by default if you optimize.
9844 @item -mno-prolog-function
9845 @itemx -mprolog-function
9846 @opindex mno-prolog-function
9847 @opindex mprolog-function
9848 Do not use (do use) external functions to save and restore registers
9849 at the prologue and epilogue of a function. The external functions
9850 are slower, but use less code space if more than one function saves
9851 the same number of registers. The @option{-mprolog-function} option
9852 is on by default if you optimize.
9856 Try to make the code as small as possible. At present, this just turns
9857 on the @option{-mep} and @option{-mprolog-function} options.
9861 Put static or global variables whose size is @var{n} bytes or less into
9862 the tiny data area that register @code{ep} points to. The tiny data
9863 area can hold up to 256 bytes in total (128 bytes for byte references).
9867 Put static or global variables whose size is @var{n} bytes or less into
9868 the small data area that register @code{gp} points to. The small data
9869 area can hold up to 64 kilobytes.
9873 Put static or global variables whose size is @var{n} bytes or less into
9874 the first 32 kilobytes of memory.
9878 Specify that the target processor is the V850.
9881 @opindex mbig-switch
9882 Generate code suitable for big switch tables. Use this option only if
9883 the assembler/linker complain about out of range branches within a switch
9888 This option will cause r2 and r5 to be used in the code generated by
9889 the compiler. This setting is the default.
9892 @opindex mno-app-regs
9893 This option will cause r2 and r5 to be treated as fixed registers.
9897 Specify that the target processor is the V850E1. The preprocessor
9898 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
9899 this option is used.
9903 Specify that the target processor is the V850E. The preprocessor
9904 constant @samp{__v850e__} will be defined if this option is used.
9906 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
9907 are defined then a default target processor will be chosen and the
9908 relevant @samp{__v850*__} preprocessor constant will be defined.
9910 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
9911 defined, regardless of which processor variant is the target.
9913 @item -mdisable-callt
9914 @opindex mdisable-callt
9915 This option will suppress generation of the CALLT instruction for the
9916 v850e and v850e1 flavors of the v850 architecture. The default is
9917 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
9922 @subsection ARC Options
9925 These options are defined for ARC implementations:
9930 Compile code for little endian mode. This is the default.
9934 Compile code for big endian mode.
9937 @opindex mmangle-cpu
9938 Prepend the name of the cpu to all public symbol names.
9939 In multiple-processor systems, there are many ARC variants with different
9940 instruction and register set characteristics. This flag prevents code
9941 compiled for one cpu to be linked with code compiled for another.
9942 No facility exists for handling variants that are ``almost identical''.
9943 This is an all or nothing option.
9945 @item -mcpu=@var{cpu}
9947 Compile code for ARC variant @var{cpu}.
9948 Which variants are supported depend on the configuration.
9949 All variants support @option{-mcpu=base}, this is the default.
9951 @item -mtext=@var{text-section}
9952 @itemx -mdata=@var{data-section}
9953 @itemx -mrodata=@var{readonly-data-section}
9957 Put functions, data, and readonly data in @var{text-section},
9958 @var{data-section}, and @var{readonly-data-section} respectively
9959 by default. This can be overridden with the @code{section} attribute.
9960 @xref{Variable Attributes}.
9965 @subsection NS32K Options
9966 @cindex NS32K options
9968 These are the @samp{-m} options defined for the 32000 series. The default
9969 values for these options depends on which style of 32000 was selected when
9970 the compiler was configured; the defaults for the most common choices are
9978 Generate output for a 32032. This is the default
9979 when the compiler is configured for 32032 and 32016 based systems.
9985 Generate output for a 32332. This is the default
9986 when the compiler is configured for 32332-based systems.
9992 Generate output for a 32532. This is the default
9993 when the compiler is configured for 32532-based systems.
9997 Generate output containing 32081 instructions for floating point.
9998 This is the default for all systems.
10002 Generate output containing 32381 instructions for floating point. This
10003 also implies @option{-m32081}. The 32381 is only compatible with the 32332
10004 and 32532 cpus. This is the default for the pc532-netbsd configuration.
10007 @opindex mmulti-add
10008 Try and generate multiply-add floating point instructions @code{polyF}
10009 and @code{dotF}. This option is only available if the @option{-m32381}
10010 option is in effect. Using these instructions requires changes to
10011 register allocation which generally has a negative impact on
10012 performance. This option should only be enabled when compiling code
10013 particularly likely to make heavy use of multiply-add instructions.
10015 @item -mnomulti-add
10016 @opindex mnomulti-add
10017 Do not try and generate multiply-add floating point instructions
10018 @code{polyF} and @code{dotF}. This is the default on all platforms.
10021 @opindex msoft-float
10022 Generate output containing library calls for floating point.
10023 @strong{Warning:} the requisite libraries may not be available.
10025 @item -mieee-compare
10026 @itemx -mno-ieee-compare
10027 @opindex mieee-compare
10028 @opindex mno-ieee-compare
10029 Control whether or not the compiler uses IEEE floating point
10030 comparisons. These handle correctly the case where the result of a
10031 comparison is unordered.
10032 @strong{Warning:} the requisite kernel support may not be available.
10035 @opindex mnobitfield
10036 Do not use the bit-field instructions. On some machines it is faster to
10037 use shifting and masking operations. This is the default for the pc532.
10041 Do use the bit-field instructions. This is the default for all platforms
10046 Use a different function-calling convention, in which functions
10047 that take a fixed number of arguments return pop their
10048 arguments on return with the @code{ret} instruction.
10050 This calling convention is incompatible with the one normally
10051 used on Unix, so you cannot use it if you need to call libraries
10052 compiled with the Unix compiler.
10054 Also, you must provide function prototypes for all functions that
10055 take variable numbers of arguments (including @code{printf});
10056 otherwise incorrect code will be generated for calls to those
10059 In addition, seriously incorrect code will result if you call a
10060 function with too many arguments. (Normally, extra arguments are
10061 harmlessly ignored.)
10063 This option takes its name from the 680x0 @code{rtd} instruction.
10068 Use a different function-calling convention where the first two arguments
10069 are passed in registers.
10071 This calling convention is incompatible with the one normally
10072 used on Unix, so you cannot use it if you need to call libraries
10073 compiled with the Unix compiler.
10076 @opindex mnoregparam
10077 Do not pass any arguments in registers. This is the default for all
10082 It is OK to use the sb as an index register which is always loaded with
10083 zero. This is the default for the pc532-netbsd target.
10087 The sb register is not available for use or has not been initialized to
10088 zero by the run time system. This is the default for all targets except
10089 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
10090 @option{-fpic} is set.
10094 Many ns32000 series addressing modes use displacements of up to 512MB@.
10095 If an address is above 512MB then displacements from zero can not be used.
10096 This option causes code to be generated which can be loaded above 512MB@.
10097 This may be useful for operating systems or ROM code.
10101 Assume code will be loaded in the first 512MB of virtual address space.
10102 This is the default for all platforms.
10108 @subsection AVR Options
10109 @cindex AVR Options
10111 These options are defined for AVR implementations:
10114 @item -mmcu=@var{mcu}
10116 Specify ATMEL AVR instruction set or MCU type.
10118 Instruction set avr1 is for the minimal AVR core, not supported by the C
10119 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
10120 attiny11, attiny12, attiny15, attiny28).
10122 Instruction set avr2 (default) is for the classic AVR core with up to
10123 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
10124 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
10125 at90c8534, at90s8535).
10127 Instruction set avr3 is for the classic AVR core with up to 128K program
10128 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
10130 Instruction set avr4 is for the enhanced AVR core with up to 8K program
10131 memory space (MCU types: atmega8, atmega83, atmega85).
10133 Instruction set avr5 is for the enhanced AVR core with up to 128K program
10134 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
10135 atmega64, atmega128, at43usb355, at94k).
10139 Output instruction sizes to the asm file.
10141 @item -minit-stack=@var{N}
10142 @opindex minit-stack
10143 Specify the initial stack address, which may be a symbol or numeric value,
10144 @samp{__stack} is the default.
10146 @item -mno-interrupts
10147 @opindex mno-interrupts
10148 Generated code is not compatible with hardware interrupts.
10149 Code size will be smaller.
10151 @item -mcall-prologues
10152 @opindex mcall-prologues
10153 Functions prologues/epilogues expanded as call to appropriate
10154 subroutines. Code size will be smaller.
10156 @item -mno-tablejump
10157 @opindex mno-tablejump
10158 Do not generate tablejump insns which sometimes increase code size.
10161 @opindex mtiny-stack
10162 Change only the low 8 bits of the stack pointer.
10165 @node MCore Options
10166 @subsection MCore Options
10167 @cindex MCore options
10169 These are the @samp{-m} options defined for the Motorola M*Core
10175 @itemx -mno-hardlit
10177 @opindex mno-hardlit
10178 Inline constants into the code stream if it can be done in two
10179 instructions or less.
10185 Use the divide instruction. (Enabled by default).
10187 @item -mrelax-immediate
10188 @itemx -mno-relax-immediate
10189 @opindex mrelax-immediate
10190 @opindex mno-relax-immediate
10191 Allow arbitrary sized immediates in bit operations.
10193 @item -mwide-bitfields
10194 @itemx -mno-wide-bitfields
10195 @opindex mwide-bitfields
10196 @opindex mno-wide-bitfields
10197 Always treat bit-fields as int-sized.
10199 @item -m4byte-functions
10200 @itemx -mno-4byte-functions
10201 @opindex m4byte-functions
10202 @opindex mno-4byte-functions
10203 Force all functions to be aligned to a four byte boundary.
10205 @item -mcallgraph-data
10206 @itemx -mno-callgraph-data
10207 @opindex mcallgraph-data
10208 @opindex mno-callgraph-data
10209 Emit callgraph information.
10212 @itemx -mno-slow-bytes
10213 @opindex mslow-bytes
10214 @opindex mno-slow-bytes
10215 Prefer word access when reading byte quantities.
10217 @item -mlittle-endian
10218 @itemx -mbig-endian
10219 @opindex mlittle-endian
10220 @opindex mbig-endian
10221 Generate code for a little endian target.
10227 Generate code for the 210 processor.
10230 @node IA-64 Options
10231 @subsection IA-64 Options
10232 @cindex IA-64 Options
10234 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10238 @opindex mbig-endian
10239 Generate code for a big endian target. This is the default for HP-UX@.
10241 @item -mlittle-endian
10242 @opindex mlittle-endian
10243 Generate code for a little endian target. This is the default for AIX5
10249 @opindex mno-gnu-as
10250 Generate (or don't) code for the GNU assembler. This is the default.
10251 @c Also, this is the default if the configure option @option{--with-gnu-as}
10257 @opindex mno-gnu-ld
10258 Generate (or don't) code for the GNU linker. This is the default.
10259 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10264 Generate code that does not use a global pointer register. The result
10265 is not position independent code, and violates the IA-64 ABI@.
10267 @item -mvolatile-asm-stop
10268 @itemx -mno-volatile-asm-stop
10269 @opindex mvolatile-asm-stop
10270 @opindex mno-volatile-asm-stop
10271 Generate (or don't) a stop bit immediately before and after volatile asm
10276 Generate code that works around Itanium B step errata.
10278 @item -mregister-names
10279 @itemx -mno-register-names
10280 @opindex mregister-names
10281 @opindex mno-register-names
10282 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
10283 the stacked registers. This may make assembler output more readable.
10289 Disable (or enable) optimizations that use the small data section. This may
10290 be useful for working around optimizer bugs.
10292 @item -mconstant-gp
10293 @opindex mconstant-gp
10294 Generate code that uses a single constant global pointer value. This is
10295 useful when compiling kernel code.
10299 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
10300 This is useful when compiling firmware code.
10302 @item -minline-float-divide-min-latency
10303 @opindex minline-float-divide-min-latency
10304 Generate code for inline divides of floating point values
10305 using the minimum latency algorithm.
10307 @item -minline-float-divide-max-throughput
10308 @opindex minline-float-divide-max-throughput
10309 Generate code for inline divides of floating point values
10310 using the maximum throughput algorithm.
10312 @item -minline-int-divide-min-latency
10313 @opindex minline-int-divide-min-latency
10314 Generate code for inline divides of integer values
10315 using the minimum latency algorithm.
10317 @item -minline-int-divide-max-throughput
10318 @opindex minline-int-divide-max-throughput
10319 Generate code for inline divides of integer values
10320 using the maximum throughput algorithm.
10322 @item -mno-dwarf2-asm
10323 @itemx -mdwarf2-asm
10324 @opindex mno-dwarf2-asm
10325 @opindex mdwarf2-asm
10326 Don't (or do) generate assembler code for the DWARF2 line number debugging
10327 info. This may be useful when not using the GNU assembler.
10329 @item -mfixed-range=@var{register-range}
10330 @opindex mfixed-range
10331 Generate code treating the given register range as fixed registers.
10332 A fixed register is one that the register allocator can not use. This is
10333 useful when compiling kernel code. A register range is specified as
10334 two registers separated by a dash. Multiple register ranges can be
10335 specified separated by a comma.
10337 @item -mearly-stop-bits
10338 @itemx -mno-early-stop-bits
10339 @opindex mearly-stop-bits
10340 @opindex mno-early-stop-bits
10341 Allow stop bits to be placed earlier than immediately preceding the
10342 instruction that triggered the stop bit. This can improve instruction
10343 scheduling, but does not always do so.
10347 @subsection D30V Options
10348 @cindex D30V Options
10350 These @samp{-m} options are defined for D30V implementations:
10355 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
10356 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
10357 memory, which starts at location @code{0x80000000}.
10360 @opindex mextmemory
10361 Same as the @option{-mextmem} switch.
10365 Link the @samp{.text} section into onchip text memory, which starts at
10366 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
10367 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
10368 into onchip data memory, which starts at location @code{0x20000000}.
10370 @item -mno-asm-optimize
10371 @itemx -masm-optimize
10372 @opindex mno-asm-optimize
10373 @opindex masm-optimize
10374 Disable (enable) passing @option{-O} to the assembler when optimizing.
10375 The assembler uses the @option{-O} option to automatically parallelize
10376 adjacent short instructions where possible.
10378 @item -mbranch-cost=@var{n}
10379 @opindex mbranch-cost
10380 Increase the internal costs of branches to @var{n}. Higher costs means
10381 that the compiler will issue more instructions to avoid doing a branch.
10384 @item -mcond-exec=@var{n}
10385 @opindex mcond-exec
10386 Specify the maximum number of conditionally executed instructions that
10387 replace a branch. The default is 4.
10390 @node S/390 and zSeries Options
10391 @subsection S/390 and zSeries Options
10392 @cindex S/390 and zSeries Options
10394 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
10398 @itemx -msoft-float
10399 @opindex mhard-float
10400 @opindex msoft-float
10401 Use (do not use) the hardware floating-point instructions and registers
10402 for floating-point operations. When @option{-msoft-float} is specified,
10403 functions in @file{libgcc.a} will be used to perform floating-point
10404 operations. When @option{-mhard-float} is specified, the compiler
10405 generates IEEE floating-point instructions. This is the default.
10408 @itemx -mno-backchain
10409 @opindex mbackchain
10410 @opindex mno-backchain
10411 Generate (or do not generate) code which maintains an explicit
10412 backchain within the stack frame that points to the caller's frame.
10413 This is currently needed to allow debugging. The default is to
10414 generate the backchain.
10417 @itemx -mno-small-exec
10418 @opindex msmall-exec
10419 @opindex mno-small-exec
10420 Generate (or do not generate) code using the @code{bras} instruction
10421 to do subroutine calls.
10422 This only works reliably if the total executable size does not
10423 exceed 64k. The default is to use the @code{basr} instruction instead,
10424 which does not have this limitation.
10430 When @option{-m31} is specified, generate code compliant to the
10431 Linux for S/390 ABI@. When @option{-m64} is specified, generate
10432 code compliant to the Linux for zSeries ABI@. This allows GCC in
10433 particular to generate 64-bit instructions. For the @samp{s390}
10434 targets, the default is @option{-m31}, while the @samp{s390x}
10435 targets default to @option{-m64}.
10441 When @option{-mzarch} is specified, generate code using the
10442 instructions available on z/Architecture.
10443 When @option{-mesa} is specified, generate code using the
10444 instructions available on ESA/390. Note that @option{-mesa} is
10445 not possible with @option{-m64}.
10446 When generating code compliant to the Linux for S/390 ABI,
10447 the default is @option{-mesa}. When generating code compliant
10448 to the Linux for zSeries ABI, the default is @option{-mzarch}.
10454 Generate (or do not generate) code using the @code{mvcle} instruction
10455 to perform block moves. When @option{-mno-mvcle} is specified,
10456 use a @code{mvc} loop instead. This is the default.
10462 Print (or do not print) additional debug information when compiling.
10463 The default is to not print debug information.
10465 @item -march=@var{cpu-type}
10467 Generate code that will run on @var{cpu-type}, which is the name of a system
10468 representing a certain processor type. Possible values for
10469 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
10470 When generating code using the instructions available on z/Architecture,
10471 the default is @option{-march=z900}. Otherwise, the default is
10472 @option{-march=g5}.
10474 @item -mtune=@var{cpu-type}
10476 Tune to @var{cpu-type} everything applicable about the generated code,
10477 except for the ABI and the set of available instructions.
10478 The list of @var{cpu-type} values is the same as for @option{-march}.
10479 The default is the value used for @option{-march}.
10482 @itemx -mno-fused-madd
10483 @opindex mfused-madd
10484 @opindex mno-fused-madd
10485 Generate code that uses (does not use) the floating point multiply and
10486 accumulate instructions. These instructions are generated by default if
10487 hardware floating point is used.
10491 @subsection CRIS Options
10492 @cindex CRIS Options
10494 These options are defined specifically for the CRIS ports.
10497 @item -march=@var{architecture-type}
10498 @itemx -mcpu=@var{architecture-type}
10501 Generate code for the specified architecture. The choices for
10502 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10503 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
10504 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10507 @item -mtune=@var{architecture-type}
10509 Tune to @var{architecture-type} everything applicable about the generated
10510 code, except for the ABI and the set of available instructions. The
10511 choices for @var{architecture-type} are the same as for
10512 @option{-march=@var{architecture-type}}.
10514 @item -mmax-stack-frame=@var{n}
10515 @opindex mmax-stack-frame
10516 Warn when the stack frame of a function exceeds @var{n} bytes.
10518 @item -melinux-stacksize=@var{n}
10519 @opindex melinux-stacksize
10520 Only available with the @samp{cris-axis-aout} target. Arranges for
10521 indications in the program to the kernel loader that the stack of the
10522 program should be set to @var{n} bytes.
10528 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10529 @option{-march=v3} and @option{-march=v8} respectively.
10533 Enable CRIS-specific verbose debug-related information in the assembly
10534 code. This option also has the effect to turn off the @samp{#NO_APP}
10535 formatted-code indicator to the assembler at the beginning of the
10540 Do not use condition-code results from previous instruction; always emit
10541 compare and test instructions before use of condition codes.
10543 @item -mno-side-effects
10544 @opindex mno-side-effects
10545 Do not emit instructions with side-effects in addressing modes other than
10548 @item -mstack-align
10549 @itemx -mno-stack-align
10550 @itemx -mdata-align
10551 @itemx -mno-data-align
10552 @itemx -mconst-align
10553 @itemx -mno-const-align
10554 @opindex mstack-align
10555 @opindex mno-stack-align
10556 @opindex mdata-align
10557 @opindex mno-data-align
10558 @opindex mconst-align
10559 @opindex mno-const-align
10560 These options (no-options) arranges (eliminate arrangements) for the
10561 stack-frame, individual data and constants to be aligned for the maximum
10562 single data access size for the chosen CPU model. The default is to
10563 arrange for 32-bit alignment. ABI details such as structure layout are
10564 not affected by these options.
10572 Similar to the stack- data- and const-align options above, these options
10573 arrange for stack-frame, writable data and constants to all be 32-bit,
10574 16-bit or 8-bit aligned. The default is 32-bit alignment.
10576 @item -mno-prologue-epilogue
10577 @itemx -mprologue-epilogue
10578 @opindex mno-prologue-epilogue
10579 @opindex mprologue-epilogue
10580 With @option{-mno-prologue-epilogue}, the normal function prologue and
10581 epilogue that sets up the stack-frame are omitted and no return
10582 instructions or return sequences are generated in the code. Use this
10583 option only together with visual inspection of the compiled code: no
10584 warnings or errors are generated when call-saved registers must be saved,
10585 or storage for local variable needs to be allocated.
10589 @opindex mno-gotplt
10591 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10592 instruction sequences that load addresses for functions from the PLT part
10593 of the GOT rather than (traditional on other architectures) calls to the
10594 PLT. The default is @option{-mgotplt}.
10598 Legacy no-op option only recognized with the cris-axis-aout target.
10602 Legacy no-op option only recognized with the cris-axis-elf and
10603 cris-axis-linux-gnu targets.
10607 Only recognized with the cris-axis-aout target, where it selects a
10608 GNU/linux-like multilib, include files and instruction set for
10609 @option{-march=v8}.
10613 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10617 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
10618 to link with input-output functions from a simulator library. Code,
10619 initialized data and zero-initialized data are allocated consecutively.
10623 Like @option{-sim}, but pass linker options to locate initialized data at
10624 0x40000000 and zero-initialized data at 0x80000000.
10628 @subsection MMIX Options
10629 @cindex MMIX Options
10631 These options are defined for the MMIX:
10635 @itemx -mno-libfuncs
10637 @opindex mno-libfuncs
10638 Specify that intrinsic library functions are being compiled, passing all
10639 values in registers, no matter the size.
10642 @itemx -mno-epsilon
10644 @opindex mno-epsilon
10645 Generate floating-point comparison instructions that compare with respect
10646 to the @code{rE} epsilon register.
10648 @item -mabi=mmixware
10650 @opindex mabi-mmixware
10652 Generate code that passes function parameters and return values that (in
10653 the called function) are seen as registers @code{$0} and up, as opposed to
10654 the GNU ABI which uses global registers @code{$231} and up.
10656 @item -mzero-extend
10657 @itemx -mno-zero-extend
10658 @opindex mzero-extend
10659 @opindex mno-zero-extend
10660 When reading data from memory in sizes shorter than 64 bits, use (do not
10661 use) zero-extending load instructions by default, rather than
10662 sign-extending ones.
10665 @itemx -mno-knuthdiv
10667 @opindex mno-knuthdiv
10668 Make the result of a division yielding a remainder have the same sign as
10669 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
10670 remainder follows the sign of the dividend. Both methods are
10671 arithmetically valid, the latter being almost exclusively used.
10673 @item -mtoplevel-symbols
10674 @itemx -mno-toplevel-symbols
10675 @opindex mtoplevel-symbols
10676 @opindex mno-toplevel-symbols
10677 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
10678 code can be used with the @code{PREFIX} assembly directive.
10682 Generate an executable in the ELF format, rather than the default
10683 @samp{mmo} format used by the @command{mmix} simulator.
10685 @item -mbranch-predict
10686 @itemx -mno-branch-predict
10687 @opindex mbranch-predict
10688 @opindex mno-branch-predict
10689 Use (do not use) the probable-branch instructions, when static branch
10690 prediction indicates a probable branch.
10692 @item -mbase-addresses
10693 @itemx -mno-base-addresses
10694 @opindex mbase-addresses
10695 @opindex mno-base-addresses
10696 Generate (do not generate) code that uses @emph{base addresses}. Using a
10697 base address automatically generates a request (handled by the assembler
10698 and the linker) for a constant to be set up in a global register. The
10699 register is used for one or more base address requests within the range 0
10700 to 255 from the value held in the register. The generally leads to short
10701 and fast code, but the number of different data items that can be
10702 addressed is limited. This means that a program that uses lots of static
10703 data may require @option{-mno-base-addresses}.
10705 @item -msingle-exit
10706 @itemx -mno-single-exit
10707 @opindex msingle-exit
10708 @opindex mno-single-exit
10709 Force (do not force) generated code to have a single exit point in each
10713 @node PDP-11 Options
10714 @subsection PDP-11 Options
10715 @cindex PDP-11 Options
10717 These options are defined for the PDP-11:
10722 Use hardware FPP floating point. This is the default. (FIS floating
10723 point on the PDP-11/40 is not supported.)
10726 @opindex msoft-float
10727 Do not use hardware floating point.
10731 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
10735 Return floating-point results in memory. This is the default.
10739 Generate code for a PDP-11/40.
10743 Generate code for a PDP-11/45. This is the default.
10747 Generate code for a PDP-11/10.
10749 @item -mbcopy-builtin
10750 @opindex bcopy-builtin
10751 Use inline @code{movstrhi} patterns for copying memory. This is the
10756 Do not use inline @code{movstrhi} patterns for copying memory.
10762 Use 16-bit @code{int}. This is the default.
10768 Use 32-bit @code{int}.
10771 @itemx -mno-float32
10773 @opindex mno-float32
10774 Use 64-bit @code{float}. This is the default.
10779 @opindex mno-float64
10780 Use 32-bit @code{float}.
10784 Use @code{abshi2} pattern. This is the default.
10788 Do not use @code{abshi2} pattern.
10790 @item -mbranch-expensive
10791 @opindex mbranch-expensive
10792 Pretend that branches are expensive. This is for experimenting with
10793 code generation only.
10795 @item -mbranch-cheap
10796 @opindex mbranch-cheap
10797 Do not pretend that branches are expensive. This is the default.
10801 Generate code for a system with split I&D.
10805 Generate code for a system without split I&D. This is the default.
10809 Use Unix assembler syntax. This is the default when configured for
10810 @samp{pdp11-*-bsd}.
10814 Use DEC assembler syntax. This is the default when configured for any
10815 PDP-11 target other than @samp{pdp11-*-bsd}.
10818 @node Xstormy16 Options
10819 @subsection Xstormy16 Options
10820 @cindex Xstormy16 Options
10822 These options are defined for Xstormy16:
10827 Choose startup files and linker script suitable for the simulator.
10831 @subsection FRV Options
10832 @cindex FRV Options
10838 Only use the first 32 general purpose registers.
10843 Use all 64 general purpose registers.
10848 Use only the first 32 floating point registers.
10853 Use all 64 floating point registers
10856 @opindex mhard-float
10858 Use hardware instructions for floating point operations.
10861 @opindex msoft-float
10863 Use library routines for floating point operations.
10868 Dynamically allocate condition code registers.
10873 Do not try to dynamically allocate condition code registers, only
10874 use @code{icc0} and @code{fcc0}.
10879 Change ABI to use double word insns.
10884 Do not use double word instructions.
10889 Use floating point double instructions.
10892 @opindex mno-double
10894 Do not use floating point double instructions.
10899 Use media instructions.
10904 Do not use media instructions.
10909 Use multiply and add/subtract instructions.
10912 @opindex mno-muladd
10914 Do not use multiply and add/subtract instructions.
10916 @item -mlibrary-pic
10917 @opindex mlibrary-pic
10919 Enable PIC support for building libraries
10924 Use only the first four media accumulator registers.
10929 Use all eight media accumulator registers.
10934 Pack VLIW instructions.
10939 Do not pack VLIW instructions.
10942 @opindex mno-eflags
10944 Do not mark ABI switches in e_flags.
10947 @opindex mcond-move
10949 Enable the use of conditional-move instructions (default).
10951 This switch is mainly for debugging the compiler and will likely be removed
10952 in a future version.
10954 @item -mno-cond-move
10955 @opindex mno-cond-move
10957 Disable the use of conditional-move instructions.
10959 This switch is mainly for debugging the compiler and will likely be removed
10960 in a future version.
10965 Enable the use of conditional set instructions (default).
10967 This switch is mainly for debugging the compiler and will likely be removed
10968 in a future version.
10973 Disable the use of conditional set instructions.
10975 This switch is mainly for debugging the compiler and will likely be removed
10976 in a future version.
10979 @opindex mcond-exec
10981 Enable the use of conditional execution (default).
10983 This switch is mainly for debugging the compiler and will likely be removed
10984 in a future version.
10986 @item -mno-cond-exec
10987 @opindex mno-cond-exec
10989 Disable the use of conditional execution.
10991 This switch is mainly for debugging the compiler and will likely be removed
10992 in a future version.
10994 @item -mvliw-branch
10995 @opindex mvliw-branch
10997 Run a pass to pack branches into VLIW instructions (default).
10999 This switch is mainly for debugging the compiler and will likely be removed
11000 in a future version.
11002 @item -mno-vliw-branch
11003 @opindex mno-vliw-branch
11005 Do not run a pass to pack branches into VLIW instructions.
11007 This switch is mainly for debugging the compiler and will likely be removed
11008 in a future version.
11010 @item -mmulti-cond-exec
11011 @opindex mmulti-cond-exec
11013 Enable optimization of @code{&&} and @code{||} in conditional execution
11016 This switch is mainly for debugging the compiler and will likely be removed
11017 in a future version.
11019 @item -mno-multi-cond-exec
11020 @opindex mno-multi-cond-exec
11022 Disable optimization of @code{&&} and @code{||} in conditional execution.
11024 This switch is mainly for debugging the compiler and will likely be removed
11025 in a future version.
11027 @item -mnested-cond-exec
11028 @opindex mnested-cond-exec
11030 Enable nested conditional execution optimizations (default).
11032 This switch is mainly for debugging the compiler and will likely be removed
11033 in a future version.
11035 @item -mno-nested-cond-exec
11036 @opindex mno-nested-cond-exec
11038 Disable nested conditional execution optimizations.
11040 This switch is mainly for debugging the compiler and will likely be removed
11041 in a future version.
11043 @item -mtomcat-stats
11044 @opindex mtomcat-stats
11046 Cause gas to print out tomcat statistics.
11048 @item -mcpu=@var{cpu}
11051 Select the processor type for which to generate code. Possible values are
11052 @samp{simple}, @samp{tomcat}, @samp{fr500}, @samp{fr400}, @samp{fr300},
11057 @node Xtensa Options
11058 @subsection Xtensa Options
11059 @cindex Xtensa Options
11061 These options are supported for Xtensa targets:
11065 @itemx -mno-const16
11067 @opindex mno-const16
11068 Enable or disable use of @code{CONST16} instructions for loading
11069 constant values. The @code{CONST16} instruction is currently not a
11070 standard option from Tensilica. When enabled, @code{CONST16}
11071 instructions are always used in place of the standard @code{L32R}
11072 instructions. The use of @code{CONST16} is enabled by default only if
11073 the @code{L32R} instruction is not available.
11076 @itemx -mno-fused-madd
11077 @opindex mfused-madd
11078 @opindex mno-fused-madd
11079 Enable or disable use of fused multiply/add and multiply/subtract
11080 instructions in the floating-point option. This has no effect if the
11081 floating-point option is not also enabled. Disabling fused multiply/add
11082 and multiply/subtract instructions forces the compiler to use separate
11083 instructions for the multiply and add/subtract operations. This may be
11084 desirable in some cases where strict IEEE 754-compliant results are
11085 required: the fused multiply add/subtract instructions do not round the
11086 intermediate result, thereby producing results with @emph{more} bits of
11087 precision than specified by the IEEE standard. Disabling fused multiply
11088 add/subtract instructions also ensures that the program output is not
11089 sensitive to the compiler's ability to combine multiply and add/subtract
11092 @item -mtext-section-literals
11093 @itemx -mno-text-section-literals
11094 @opindex mtext-section-literals
11095 @opindex mno-text-section-literals
11096 Control the treatment of literal pools. The default is
11097 @option{-mno-text-section-literals}, which places literals in a separate
11098 section in the output file. This allows the literal pool to be placed
11099 in a data RAM/ROM, and it also allows the linker to combine literal
11100 pools from separate object files to remove redundant literals and
11101 improve code size. With @option{-mtext-section-literals}, the literals
11102 are interspersed in the text section in order to keep them as close as
11103 possible to their references. This may be necessary for large assembly
11106 @item -mtarget-align
11107 @itemx -mno-target-align
11108 @opindex mtarget-align
11109 @opindex mno-target-align
11110 When this option is enabled, GCC instructs the assembler to
11111 automatically align instructions to reduce branch penalties at the
11112 expense of some code density. The assembler attempts to widen density
11113 instructions to align branch targets and the instructions following call
11114 instructions. If there are not enough preceding safe density
11115 instructions to align a target, no widening will be performed. The
11116 default is @option{-mtarget-align}. These options do not affect the
11117 treatment of auto-aligned instructions like @code{LOOP}, which the
11118 assembler will always align, either by widening density instructions or
11119 by inserting no-op instructions.
11122 @itemx -mno-longcalls
11123 @opindex mlongcalls
11124 @opindex mno-longcalls
11125 When this option is enabled, GCC instructs the assembler to translate
11126 direct calls to indirect calls unless it can determine that the target
11127 of a direct call is in the range allowed by the call instruction. This
11128 translation typically occurs for calls to functions in other source
11129 files. Specifically, the assembler translates a direct @code{CALL}
11130 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
11131 The default is @option{-mno-longcalls}. This option should be used in
11132 programs where the call target can potentially be out of range. This
11133 option is implemented in the assembler, not the compiler, so the
11134 assembly code generated by GCC will still show direct call
11135 instructions---look at the disassembled object code to see the actual
11136 instructions. Note that the assembler will use an indirect call for
11137 every cross-file call, not just those that really will be out of range.
11140 @node Code Gen Options
11141 @section Options for Code Generation Conventions
11142 @cindex code generation conventions
11143 @cindex options, code generation
11144 @cindex run-time options
11146 These machine-independent options control the interface conventions
11147 used in code generation.
11149 Most of them have both positive and negative forms; the negative form
11150 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
11151 one of the forms is listed---the one which is not the default. You
11152 can figure out the other form by either removing @samp{no-} or adding
11156 @item -fbounds-check
11157 @opindex fbounds-check
11158 For front-ends that support it, generate additional code to check that
11159 indices used to access arrays are within the declared range. This is
11160 currently only supported by the Java and Fortran 77 front-ends, where
11161 this option defaults to true and false respectively.
11165 This option generates traps for signed overflow on addition, subtraction,
11166 multiplication operations.
11170 This option instructs the compiler to assume that signed arithmetic
11171 overflow of addition, subtraction and multiplication wraps around
11172 using twos-complement representation. This flag enables some optimizations
11173 and disables other. This option is enabled by default for the Java
11174 front-end, as required by the Java language specification.
11177 @opindex fexceptions
11178 Enable exception handling. Generates extra code needed to propagate
11179 exceptions. For some targets, this implies GCC will generate frame
11180 unwind information for all functions, which can produce significant data
11181 size overhead, although it does not affect execution. If you do not
11182 specify this option, GCC will enable it by default for languages like
11183 C++ which normally require exception handling, and disable it for
11184 languages like C that do not normally require it. However, you may need
11185 to enable this option when compiling C code that needs to interoperate
11186 properly with exception handlers written in C++. You may also wish to
11187 disable this option if you are compiling older C++ programs that don't
11188 use exception handling.
11190 @item -fnon-call-exceptions
11191 @opindex fnon-call-exceptions
11192 Generate code that allows trapping instructions to throw exceptions.
11193 Note that this requires platform-specific runtime support that does
11194 not exist everywhere. Moreover, it only allows @emph{trapping}
11195 instructions to throw exceptions, i.e.@: memory references or floating
11196 point instructions. It does not allow exceptions to be thrown from
11197 arbitrary signal handlers such as @code{SIGALRM}.
11199 @item -funwind-tables
11200 @opindex funwind-tables
11201 Similar to @option{-fexceptions}, except that it will just generate any needed
11202 static data, but will not affect the generated code in any other way.
11203 You will normally not enable this option; instead, a language processor
11204 that needs this handling would enable it on your behalf.
11206 @item -fasynchronous-unwind-tables
11207 @opindex funwind-tables
11208 Generate unwind table in dwarf2 format, if supported by target machine. The
11209 table is exact at each instruction boundary, so it can be used for stack
11210 unwinding from asynchronous events (such as debugger or garbage collector).
11212 @item -fpcc-struct-return
11213 @opindex fpcc-struct-return
11214 Return ``short'' @code{struct} and @code{union} values in memory like
11215 longer ones, rather than in registers. This convention is less
11216 efficient, but it has the advantage of allowing intercallability between
11217 GCC-compiled files and files compiled with other compilers, particularly
11218 the Portable C Compiler (pcc).
11220 The precise convention for returning structures in memory depends
11221 on the target configuration macros.
11223 Short structures and unions are those whose size and alignment match
11224 that of some integer type.
11226 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
11227 switch is not binary compatible with code compiled with the
11228 @option{-freg-struct-return} switch.
11229 Use it to conform to a non-default application binary interface.
11231 @item -freg-struct-return
11232 @opindex freg-struct-return
11233 Return @code{struct} and @code{union} values in registers when possible.
11234 This is more efficient for small structures than
11235 @option{-fpcc-struct-return}.
11237 If you specify neither @option{-fpcc-struct-return} nor
11238 @option{-freg-struct-return}, GCC defaults to whichever convention is
11239 standard for the target. If there is no standard convention, GCC
11240 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
11241 the principal compiler. In those cases, we can choose the standard, and
11242 we chose the more efficient register return alternative.
11244 @strong{Warning:} code compiled with the @option{-freg-struct-return}
11245 switch is not binary compatible with code compiled with the
11246 @option{-fpcc-struct-return} switch.
11247 Use it to conform to a non-default application binary interface.
11249 @item -fshort-enums
11250 @opindex fshort-enums
11251 Allocate to an @code{enum} type only as many bytes as it needs for the
11252 declared range of possible values. Specifically, the @code{enum} type
11253 will be equivalent to the smallest integer type which has enough room.
11255 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
11256 code that is not binary compatible with code generated without that switch.
11257 Use it to conform to a non-default application binary interface.
11259 @item -fshort-double
11260 @opindex fshort-double
11261 Use the same size for @code{double} as for @code{float}.
11263 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
11264 code that is not binary compatible with code generated without that switch.
11265 Use it to conform to a non-default application binary interface.
11267 @item -fshort-wchar
11268 @opindex fshort-wchar
11269 Override the underlying type for @samp{wchar_t} to be @samp{short
11270 unsigned int} instead of the default for the target. This option is
11271 useful for building programs to run under WINE@.
11273 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
11274 code that is not binary compatible with code generated without that switch.
11275 Use it to conform to a non-default application binary interface.
11277 @item -fshared-data
11278 @opindex fshared-data
11279 Requests that the data and non-@code{const} variables of this
11280 compilation be shared data rather than private data. The distinction
11281 makes sense only on certain operating systems, where shared data is
11282 shared between processes running the same program, while private data
11283 exists in one copy per process.
11286 @opindex fno-common
11287 In C, allocate even uninitialized global variables in the data section of the
11288 object file, rather than generating them as common blocks. This has the
11289 effect that if the same variable is declared (without @code{extern}) in
11290 two different compilations, you will get an error when you link them.
11291 The only reason this might be useful is if you wish to verify that the
11292 program will work on other systems which always work this way.
11296 Ignore the @samp{#ident} directive.
11298 @item -fno-gnu-linker
11299 @opindex fno-gnu-linker
11300 Do not output global initializations (such as C++ constructors and
11301 destructors) in the form used by the GNU linker (on systems where the GNU
11302 linker is the standard method of handling them). Use this option when
11303 you want to use a non-GNU linker, which also requires using the
11304 @command{collect2} program to make sure the system linker includes
11305 constructors and destructors. (@command{collect2} is included in the GCC
11306 distribution.) For systems which @emph{must} use @command{collect2}, the
11307 compiler driver @command{gcc} is configured to do this automatically.
11309 @item -finhibit-size-directive
11310 @opindex finhibit-size-directive
11311 Don't output a @code{.size} assembler directive, or anything else that
11312 would cause trouble if the function is split in the middle, and the
11313 two halves are placed at locations far apart in memory. This option is
11314 used when compiling @file{crtstuff.c}; you should not need to use it
11317 @item -fverbose-asm
11318 @opindex fverbose-asm
11319 Put extra commentary information in the generated assembly code to
11320 make it more readable. This option is generally only of use to those
11321 who actually need to read the generated assembly code (perhaps while
11322 debugging the compiler itself).
11324 @option{-fno-verbose-asm}, the default, causes the
11325 extra information to be omitted and is useful when comparing two assembler
11330 @cindex global offset table
11332 Generate position-independent code (PIC) suitable for use in a shared
11333 library, if supported for the target machine. Such code accesses all
11334 constant addresses through a global offset table (GOT)@. The dynamic
11335 loader resolves the GOT entries when the program starts (the dynamic
11336 loader is not part of GCC; it is part of the operating system). If
11337 the GOT size for the linked executable exceeds a machine-specific
11338 maximum size, you get an error message from the linker indicating that
11339 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
11340 instead. (These maximums are 16k on the m88k, 8k on the SPARC, and 32k
11341 on the m68k and RS/6000. The 386 has no such limit.)
11343 Position-independent code requires special support, and therefore works
11344 only on certain machines. For the 386, GCC supports PIC for System V
11345 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
11346 position-independent.
11350 If supported for the target machine, emit position-independent code,
11351 suitable for dynamic linking and avoiding any limit on the size of the
11352 global offset table. This option makes a difference on the m68k, m88k,
11355 Position-independent code requires special support, and therefore works
11356 only on certain machines.
11362 These options are similar to @option{-fpic} and @option{-fPIC}, but
11363 generated position independent code can be only linked into executables.
11364 Usually these options are used when @option{-pie} GCC option will be
11365 used during linking.
11367 @item -ffixed-@var{reg}
11369 Treat the register named @var{reg} as a fixed register; generated code
11370 should never refer to it (except perhaps as a stack pointer, frame
11371 pointer or in some other fixed role).
11373 @var{reg} must be the name of a register. The register names accepted
11374 are machine-specific and are defined in the @code{REGISTER_NAMES}
11375 macro in the machine description macro file.
11377 This flag does not have a negative form, because it specifies a
11380 @item -fcall-used-@var{reg}
11381 @opindex fcall-used
11382 Treat the register named @var{reg} as an allocable register that is
11383 clobbered by function calls. It may be allocated for temporaries or
11384 variables that do not live across a call. Functions compiled this way
11385 will not save and restore the register @var{reg}.
11387 It is an error to used this flag with the frame pointer or stack pointer.
11388 Use of this flag for other registers that have fixed pervasive roles in
11389 the machine's execution model will produce disastrous results.
11391 This flag does not have a negative form, because it specifies a
11394 @item -fcall-saved-@var{reg}
11395 @opindex fcall-saved
11396 Treat the register named @var{reg} as an allocable register saved by
11397 functions. It may be allocated even for temporaries or variables that
11398 live across a call. Functions compiled this way will save and restore
11399 the register @var{reg} if they use it.
11401 It is an error to used this flag with the frame pointer or stack pointer.
11402 Use of this flag for other registers that have fixed pervasive roles in
11403 the machine's execution model will produce disastrous results.
11405 A different sort of disaster will result from the use of this flag for
11406 a register in which function values may be returned.
11408 This flag does not have a negative form, because it specifies a
11411 @item -fpack-struct
11412 @opindex fpack-struct
11413 Pack all structure members together without holes.
11415 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
11416 code that is not binary compatible with code generated without that switch.
11417 Additionally, it makes the code suboptimal.
11418 Use it to conform to a non-default application binary interface.
11420 @item -finstrument-functions
11421 @opindex finstrument-functions
11422 Generate instrumentation calls for entry and exit to functions. Just
11423 after function entry and just before function exit, the following
11424 profiling functions will be called with the address of the current
11425 function and its call site. (On some platforms,
11426 @code{__builtin_return_address} does not work beyond the current
11427 function, so the call site information may not be available to the
11428 profiling functions otherwise.)
11431 void __cyg_profile_func_enter (void *this_fn,
11433 void __cyg_profile_func_exit (void *this_fn,
11437 The first argument is the address of the start of the current function,
11438 which may be looked up exactly in the symbol table.
11440 This instrumentation is also done for functions expanded inline in other
11441 functions. The profiling calls will indicate where, conceptually, the
11442 inline function is entered and exited. This means that addressable
11443 versions of such functions must be available. If all your uses of a
11444 function are expanded inline, this may mean an additional expansion of
11445 code size. If you use @samp{extern inline} in your C code, an
11446 addressable version of such functions must be provided. (This is
11447 normally the case anyways, but if you get lucky and the optimizer always
11448 expands the functions inline, you might have gotten away without
11449 providing static copies.)
11451 A function may be given the attribute @code{no_instrument_function}, in
11452 which case this instrumentation will not be done. This can be used, for
11453 example, for the profiling functions listed above, high-priority
11454 interrupt routines, and any functions from which the profiling functions
11455 cannot safely be called (perhaps signal handlers, if the profiling
11456 routines generate output or allocate memory).
11458 @item -fstack-check
11459 @opindex fstack-check
11460 Generate code to verify that you do not go beyond the boundary of the
11461 stack. You should specify this flag if you are running in an
11462 environment with multiple threads, but only rarely need to specify it in
11463 a single-threaded environment since stack overflow is automatically
11464 detected on nearly all systems if there is only one stack.
11466 Note that this switch does not actually cause checking to be done; the
11467 operating system must do that. The switch causes generation of code
11468 to ensure that the operating system sees the stack being extended.
11470 @item -fstack-limit-register=@var{reg}
11471 @itemx -fstack-limit-symbol=@var{sym}
11472 @itemx -fno-stack-limit
11473 @opindex fstack-limit-register
11474 @opindex fstack-limit-symbol
11475 @opindex fno-stack-limit
11476 Generate code to ensure that the stack does not grow beyond a certain value,
11477 either the value of a register or the address of a symbol. If the stack
11478 would grow beyond the value, a signal is raised. For most targets,
11479 the signal is raised before the stack overruns the boundary, so
11480 it is possible to catch the signal without taking special precautions.
11482 For instance, if the stack starts at absolute address @samp{0x80000000}
11483 and grows downwards, you can use the flags
11484 @option{-fstack-limit-symbol=__stack_limit} and
11485 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
11486 of 128KB@. Note that this may only work with the GNU linker.
11488 @cindex aliasing of parameters
11489 @cindex parameters, aliased
11490 @item -fargument-alias
11491 @itemx -fargument-noalias
11492 @itemx -fargument-noalias-global
11493 @opindex fargument-alias
11494 @opindex fargument-noalias
11495 @opindex fargument-noalias-global
11496 Specify the possible relationships among parameters and between
11497 parameters and global data.
11499 @option{-fargument-alias} specifies that arguments (parameters) may
11500 alias each other and may alias global storage.@*
11501 @option{-fargument-noalias} specifies that arguments do not alias
11502 each other, but may alias global storage.@*
11503 @option{-fargument-noalias-global} specifies that arguments do not
11504 alias each other and do not alias global storage.
11506 Each language will automatically use whatever option is required by
11507 the language standard. You should not need to use these options yourself.
11509 @item -fleading-underscore
11510 @opindex fleading-underscore
11511 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
11512 change the way C symbols are represented in the object file. One use
11513 is to help link with legacy assembly code.
11515 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
11516 generate code that is not binary compatible with code generated without that
11517 switch. Use it to conform to a non-default application binary interface.
11518 Not all targets provide complete support for this switch.
11520 @item -ftls-model=@var{model}
11521 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
11522 The @var{model} argument should be one of @code{global-dynamic},
11523 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
11525 The default without @option{-fpic} is @code{initial-exec}; with
11526 @option{-fpic} the default is @code{global-dynamic}.
11531 @node Environment Variables
11532 @section Environment Variables Affecting GCC
11533 @cindex environment variables
11535 @c man begin ENVIRONMENT
11536 This section describes several environment variables that affect how GCC
11537 operates. Some of them work by specifying directories or prefixes to use
11538 when searching for various kinds of files. Some are used to specify other
11539 aspects of the compilation environment.
11541 Note that you can also specify places to search using options such as
11542 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
11543 take precedence over places specified using environment variables, which
11544 in turn take precedence over those specified by the configuration of GCC@.
11545 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
11546 GNU Compiler Collection (GCC) Internals}.
11551 @c @itemx LC_COLLATE
11553 @c @itemx LC_MONETARY
11554 @c @itemx LC_NUMERIC
11559 @c @findex LC_COLLATE
11560 @findex LC_MESSAGES
11561 @c @findex LC_MONETARY
11562 @c @findex LC_NUMERIC
11566 These environment variables control the way that GCC uses
11567 localization information that allow GCC to work with different
11568 national conventions. GCC inspects the locale categories
11569 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
11570 so. These locale categories can be set to any value supported by your
11571 installation. A typical value is @samp{en_UK} for English in the United
11574 The @env{LC_CTYPE} environment variable specifies character
11575 classification. GCC uses it to determine the character boundaries in
11576 a string; this is needed for some multibyte encodings that contain quote
11577 and escape characters that would otherwise be interpreted as a string
11580 The @env{LC_MESSAGES} environment variable specifies the language to
11581 use in diagnostic messages.
11583 If the @env{LC_ALL} environment variable is set, it overrides the value
11584 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
11585 and @env{LC_MESSAGES} default to the value of the @env{LANG}
11586 environment variable. If none of these variables are set, GCC
11587 defaults to traditional C English behavior.
11591 If @env{TMPDIR} is set, it specifies the directory to use for temporary
11592 files. GCC uses temporary files to hold the output of one stage of
11593 compilation which is to be used as input to the next stage: for example,
11594 the output of the preprocessor, which is the input to the compiler
11597 @item GCC_EXEC_PREFIX
11598 @findex GCC_EXEC_PREFIX
11599 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
11600 names of the subprograms executed by the compiler. No slash is added
11601 when this prefix is combined with the name of a subprogram, but you can
11602 specify a prefix that ends with a slash if you wish.
11604 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
11605 an appropriate prefix to use based on the pathname it was invoked with.
11607 If GCC cannot find the subprogram using the specified prefix, it
11608 tries looking in the usual places for the subprogram.
11610 The default value of @env{GCC_EXEC_PREFIX} is
11611 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
11612 of @code{prefix} when you ran the @file{configure} script.
11614 Other prefixes specified with @option{-B} take precedence over this prefix.
11616 This prefix is also used for finding files such as @file{crt0.o} that are
11619 In addition, the prefix is used in an unusual way in finding the
11620 directories to search for header files. For each of the standard
11621 directories whose name normally begins with @samp{/usr/local/lib/gcc}
11622 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
11623 replacing that beginning with the specified prefix to produce an
11624 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
11625 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
11626 These alternate directories are searched first; the standard directories
11629 @item COMPILER_PATH
11630 @findex COMPILER_PATH
11631 The value of @env{COMPILER_PATH} is a colon-separated list of
11632 directories, much like @env{PATH}. GCC tries the directories thus
11633 specified when searching for subprograms, if it can't find the
11634 subprograms using @env{GCC_EXEC_PREFIX}.
11637 @findex LIBRARY_PATH
11638 The value of @env{LIBRARY_PATH} is a colon-separated list of
11639 directories, much like @env{PATH}. When configured as a native compiler,
11640 GCC tries the directories thus specified when searching for special
11641 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
11642 using GCC also uses these directories when searching for ordinary
11643 libraries for the @option{-l} option (but directories specified with
11644 @option{-L} come first).
11648 @cindex locale definition
11649 This variable is used to pass locale information to the compiler. One way in
11650 which this information is used is to determine the character set to be used
11651 when character literals, string literals and comments are parsed in C and C++.
11652 When the compiler is configured to allow multibyte characters,
11653 the following values for @env{LANG} are recognized:
11657 Recognize JIS characters.
11659 Recognize SJIS characters.
11661 Recognize EUCJP characters.
11664 If @env{LANG} is not defined, or if it has some other value, then the
11665 compiler will use mblen and mbtowc as defined by the default locale to
11666 recognize and translate multibyte characters.
11670 Some additional environments variables affect the behavior of the
11673 @include cppenv.texi
11677 @node Precompiled Headers
11678 @section Using Precompiled Headers
11679 @cindex precompiled headers
11680 @cindex speed of compilation
11682 Often large projects have many header files that are included in every
11683 source file. The time the compiler takes to process these header files
11684 over and over again can account for nearly all of the time required to
11685 build the project. To make builds faster, GCC allows users to
11686 `precompile' a header file; then, if builds can use the precompiled
11687 header file they will be much faster.
11689 To create a precompiled header file, simply compile it as you would any
11690 other file, if necessary using the @option{-x} option to make the driver
11691 treat it as a C or C++ header file. You will probably want to use a
11692 tool like @command{make} to keep the precompiled header up-to-date when
11693 the headers it contains change.
11695 A precompiled header file will be searched for when @code{#include} is
11696 seen in the compilation. As it searches for the included file
11697 (@pxref{Search Path,,Search Path,cpp.info,The C Preprocessor}) the
11698 compiler looks for a precompiled header in each directory just before it
11699 looks for the include file in that directory. The name searched for is
11700 the name specified in the @code{#include} with @samp{.gch} appended. If
11701 the precompiled header file can't be used, it is ignored.
11703 For instance, if you have @code{#include "all.h"}, and you have
11704 @file{all.h.gch} in the same directory as @file{all.h}, then the
11705 precompiled header file will be used if possible, and the original
11706 header will be used otherwise.
11708 Alternatively, you might decide to put the precompiled header file in a
11709 directory and use @option{-I} to ensure that directory is searched
11710 before (or instead of) the directory containing the original header.
11711 Then, if you want to check that the precompiled header file is always
11712 used, you can put a file of the same name as the original header in this
11713 directory containing an @code{#error} command.
11715 This also works with @option{-include}. So yet another way to use
11716 precompiled headers, good for projects not designed with precompiled
11717 header files in mind, is to simply take most of the header files used by
11718 a project, include them from another header file, precompile that header
11719 file, and @option{-include} the precompiled header. If the header files
11720 have guards against multiple inclusion, they will be skipped because
11721 they've already been included (in the precompiled header).
11723 If you need to precompile the same header file for different
11724 languages, targets, or compiler options, you can instead make a
11725 @emph{directory} named like @file{all.h.gch}, and put each precompiled
11726 header in the directory. (It doesn't matter what you call the files
11727 in the directory, every precompiled header in the directory will be
11728 considered.) The first precompiled header encountered in the
11729 directory that is valid for this compilation will be used; they're
11730 searched in no particular order.
11732 There are many other possibilities, limited only by your imagination,
11733 good sense, and the constraints of your build system.
11735 A precompiled header file can be used only when these conditions apply:
11739 Only one precompiled header can be used in a particular compilation.
11741 A precompiled header can't be used once the first C token is seen. You
11742 can have preprocessor directives before a precompiled header; you can
11743 even include a precompiled header from inside another header, so long as
11744 there are no C tokens before the @code{#include}.
11746 The precompiled header file must be produced for the same language as
11747 the current compilation. You can't use a C precompiled header for a C++
11750 The precompiled header file must be produced by the same compiler
11751 version and configuration as the current compilation is using.
11752 The easiest way to guarantee this is to use the same compiler binary
11753 for creating and using precompiled headers.
11755 Any macros defined before the precompiled header (including with
11756 @option{-D}) must either be defined in the same way as when the
11757 precompiled header was generated, or must not affect the precompiled
11758 header, which usually means that the they don't appear in the
11759 precompiled header at all.
11761 Certain command-line options must be defined in the same way as when the
11762 precompiled header was generated. At present, it's not clear which
11763 options are safe to change and which are not; the safest choice is to
11764 use exactly the same options when generating and using the precompiled
11768 For all of these but the last, the compiler will automatically ignore
11769 the precompiled header if the conditions aren't met. For the last item,
11770 some option changes will cause the precompiled header to be rejected,
11771 but not all incompatible option combinations have yet been found. If
11772 you find a new incompatible combination, please consider filing a bug
11773 report, see @ref{Bugs}.
11775 @node Running Protoize
11776 @section Running Protoize
11778 The program @code{protoize} is an optional part of GCC@. You can use
11779 it to add prototypes to a program, thus converting the program to ISO
11780 C in one respect. The companion program @code{unprotoize} does the
11781 reverse: it removes argument types from any prototypes that are found.
11783 When you run these programs, you must specify a set of source files as
11784 command line arguments. The conversion programs start out by compiling
11785 these files to see what functions they define. The information gathered
11786 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
11788 After scanning comes actual conversion. The specified files are all
11789 eligible to be converted; any files they include (whether sources or
11790 just headers) are eligible as well.
11792 But not all the eligible files are converted. By default,
11793 @code{protoize} and @code{unprotoize} convert only source and header
11794 files in the current directory. You can specify additional directories
11795 whose files should be converted with the @option{-d @var{directory}}
11796 option. You can also specify particular files to exclude with the
11797 @option{-x @var{file}} option. A file is converted if it is eligible, its
11798 directory name matches one of the specified directory names, and its
11799 name within the directory has not been excluded.
11801 Basic conversion with @code{protoize} consists of rewriting most
11802 function definitions and function declarations to specify the types of
11803 the arguments. The only ones not rewritten are those for varargs
11806 @code{protoize} optionally inserts prototype declarations at the
11807 beginning of the source file, to make them available for any calls that
11808 precede the function's definition. Or it can insert prototype
11809 declarations with block scope in the blocks where undeclared functions
11812 Basic conversion with @code{unprotoize} consists of rewriting most
11813 function declarations to remove any argument types, and rewriting
11814 function definitions to the old-style pre-ISO form.
11816 Both conversion programs print a warning for any function declaration or
11817 definition that they can't convert. You can suppress these warnings
11820 The output from @code{protoize} or @code{unprotoize} replaces the
11821 original source file. The original file is renamed to a name ending
11822 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
11823 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
11824 for DOS) file already exists, then the source file is simply discarded.
11826 @code{protoize} and @code{unprotoize} both depend on GCC itself to
11827 scan the program and collect information about the functions it uses.
11828 So neither of these programs will work until GCC is installed.
11830 Here is a table of the options you can use with @code{protoize} and
11831 @code{unprotoize}. Each option works with both programs unless
11835 @item -B @var{directory}
11836 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
11837 usual directory (normally @file{/usr/local/lib}). This file contains
11838 prototype information about standard system functions. This option
11839 applies only to @code{protoize}.
11841 @item -c @var{compilation-options}
11842 Use @var{compilation-options} as the options when running @command{gcc} to
11843 produce the @samp{.X} files. The special option @option{-aux-info} is
11844 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
11846 Note that the compilation options must be given as a single argument to
11847 @code{protoize} or @code{unprotoize}. If you want to specify several
11848 @command{gcc} options, you must quote the entire set of compilation options
11849 to make them a single word in the shell.
11851 There are certain @command{gcc} arguments that you cannot use, because they
11852 would produce the wrong kind of output. These include @option{-g},
11853 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
11854 the @var{compilation-options}, they are ignored.
11857 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
11858 systems) instead of @samp{.c}. This is convenient if you are converting
11859 a C program to C++. This option applies only to @code{protoize}.
11862 Add explicit global declarations. This means inserting explicit
11863 declarations at the beginning of each source file for each function
11864 that is called in the file and was not declared. These declarations
11865 precede the first function definition that contains a call to an
11866 undeclared function. This option applies only to @code{protoize}.
11868 @item -i @var{string}
11869 Indent old-style parameter declarations with the string @var{string}.
11870 This option applies only to @code{protoize}.
11872 @code{unprotoize} converts prototyped function definitions to old-style
11873 function definitions, where the arguments are declared between the
11874 argument list and the initial @samp{@{}. By default, @code{unprotoize}
11875 uses five spaces as the indentation. If you want to indent with just
11876 one space instead, use @option{-i " "}.
11879 Keep the @samp{.X} files. Normally, they are deleted after conversion
11883 Add explicit local declarations. @code{protoize} with @option{-l} inserts
11884 a prototype declaration for each function in each block which calls the
11885 function without any declaration. This option applies only to
11889 Make no real changes. This mode just prints information about the conversions
11890 that would have been done without @option{-n}.
11893 Make no @samp{.save} files. The original files are simply deleted.
11894 Use this option with caution.
11896 @item -p @var{program}
11897 Use the program @var{program} as the compiler. Normally, the name
11898 @file{gcc} is used.
11901 Work quietly. Most warnings are suppressed.
11904 Print the version number, just like @option{-v} for @command{gcc}.
11907 If you need special compiler options to compile one of your program's
11908 source files, then you should generate that file's @samp{.X} file
11909 specially, by running @command{gcc} on that source file with the
11910 appropriate options and the option @option{-aux-info}. Then run
11911 @code{protoize} on the entire set of files. @code{protoize} will use
11912 the existing @samp{.X} file because it is newer than the source file.
11916 gcc -Dfoo=bar file1.c -aux-info file1.X
11921 You need to include the special files along with the rest in the
11922 @code{protoize} command, even though their @samp{.X} files already
11923 exist, because otherwise they won't get converted.
11925 @xref{Protoize Caveats}, for more information on how to use
11926 @code{protoize} successfully.