1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
11 @c man begin COPYRIGHT
12 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
13 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
15 Permission is granted to copy, distribute and/or modify this document
16 under the terms of the GNU Free Documentation License, Version 1.2 or
17 any later version published by the Free Software Foundation; with the
18 Invariant Sections being ``GNU General Public License'' and ``Funding
19 Free Software'', the Front-Cover texts being (a) (see below), and with
20 the Back-Cover Texts being (b) (see below). A copy of the license is
21 included in the gfdl(7) man page.
23 (a) The FSF's Front-Cover Text is:
27 (b) The FSF's Back-Cover Text is:
29 You have freedom to copy and modify this GNU Manual, like GNU
30 software. Copies published by the Free Software Foundation raise
31 funds for GNU development.
33 @c Set file name and title for the man page.
35 @settitle GNU project C and C++ compiler
37 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
38 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
39 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
40 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
41 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
42 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
43 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
45 Only the most useful options are listed here; see below for the
46 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
49 gpl(7), gfdl(7), fsf-funding(7),
50 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
51 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
52 @file{ld}, @file{binutils} and @file{gdb}.
55 For instructions on reporting bugs, see
56 @w{@uref{http://gcc.gnu.org/bugs.html}}.
59 See the Info entry for @command{gcc}, or
60 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
61 for contributors to GCC@.
66 @chapter GCC Command Options
67 @cindex GCC command options
68 @cindex command options
69 @cindex options, GCC command
71 @c man begin DESCRIPTION
72 When you invoke GCC, it normally does preprocessing, compilation,
73 assembly and linking. The ``overall options'' allow you to stop this
74 process at an intermediate stage. For example, the @option{-c} option
75 says not to run the linker. Then the output consists of object files
76 output by the assembler.
78 Other options are passed on to one stage of processing. Some options
79 control the preprocessor and others the compiler itself. Yet other
80 options control the assembler and linker; most of these are not
81 documented here, since you rarely need to use any of them.
83 @cindex C compilation options
84 Most of the command line options that you can use with GCC are useful
85 for C programs; when an option is only useful with another language
86 (usually C++), the explanation says so explicitly. If the description
87 for a particular option does not mention a source language, you can use
88 that option with all supported languages.
90 @cindex C++ compilation options
91 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
92 options for compiling C++ programs.
94 @cindex grouping options
95 @cindex options, grouping
96 The @command{gcc} program accepts options and file names as operands. Many
97 options have multi-letter names; therefore multiple single-letter options
98 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
101 @cindex order of options
102 @cindex options, order
103 You can mix options and other arguments. For the most part, the order
104 you use doesn't matter. Order does matter when you use several options
105 of the same kind; for example, if you specify @option{-L} more than once,
106 the directories are searched in the order specified.
108 Many options have long names starting with @samp{-f} or with
109 @samp{-W}---for example,
110 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
111 these have both positive and negative forms; the negative form of
112 @option{-ffoo} would be @option{-fno-foo}. This manual documents
113 only one of these two forms, whichever one is not the default.
117 @xref{Option Index}, for an index to GCC's options.
120 * Option Summary:: Brief list of all options, without explanations.
121 * Overall Options:: Controlling the kind of output:
122 an executable, object files, assembler files,
123 or preprocessed source.
124 * Invoking G++:: Compiling C++ programs.
125 * C Dialect Options:: Controlling the variant of C language compiled.
126 * C++ Dialect Options:: Variations on C++.
127 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
129 * Language Independent Options:: Controlling how diagnostics should be
131 * Warning Options:: How picky should the compiler be?
132 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
133 * Optimize Options:: How much optimization?
134 * Preprocessor Options:: Controlling header files and macro definitions.
135 Also, getting dependency information for Make.
136 * Assembler Options:: Passing options to the assembler.
137 * Link Options:: Specifying libraries and so on.
138 * Directory Options:: Where to find header files and libraries.
139 Where to find the compiler executable files.
140 * Spec Files:: How to pass switches to sub-processes.
141 * Target Options:: Running a cross-compiler, or an old version of GCC.
142 * Submodel Options:: Specifying minor hardware or convention variations,
143 such as 68010 vs 68020.
144 * Code Gen Options:: Specifying conventions for function calls, data layout
146 * Environment Variables:: Env vars that affect GCC.
147 * Precompiled Headers:: Compiling a header once, and using it many times.
148 * Running Protoize:: Automatically adding or removing function prototypes.
154 @section Option Summary
156 Here is a summary of all the options, grouped by type. Explanations are
157 in the following sections.
160 @item Overall Options
161 @xref{Overall Options,,Options Controlling the Kind of Output}.
162 @gccoptlist{-c -S -E -o @var{file} -combine -pipe -pass-exit-codes @gol
163 -x @var{language} -v -### --help --target-help --version @@@var{file}}
165 @item C Language Options
166 @xref{C Dialect Options,,Options Controlling C Dialect}.
167 @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol
168 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
169 -fhosted -ffreestanding -fms-extensions @gol
170 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
171 -fallow-single-precision -fcond-mismatch @gol
172 -fsigned-bitfields -fsigned-char @gol
173 -funsigned-bitfields -funsigned-char}
175 @item C++ Language Options
176 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
177 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
178 -fconserve-space -ffriend-injection -fno-const-strings @gol
179 -fno-elide-constructors @gol
180 -fno-enforce-eh-specs @gol
181 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
182 -fno-implicit-templates @gol
183 -fno-implicit-inline-templates @gol
184 -fno-implement-inlines -fms-extensions @gol
185 -fno-nonansi-builtins -fno-operator-names @gol
186 -fno-optional-diags -fpermissive @gol
187 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
188 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
189 -fno-default-inline -fvisibility-inlines-hidden @gol
190 -Wabi -Wctor-dtor-privacy @gol
191 -Wnon-virtual-dtor -Wreorder @gol
192 -Weffc++ -Wno-deprecated -Wstrict-null-sentinel @gol
193 -Wno-non-template-friend -Wold-style-cast @gol
194 -Woverloaded-virtual -Wno-pmf-conversions @gol
197 @item Objective-C and Objective-C++ Language Options
198 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
199 Objective-C and Objective-C++ Dialects}.
201 -fconstant-string-class=@var{class-name} @gol
202 -fgnu-runtime -fnext-runtime @gol
203 -fno-nil-receivers @gol
204 -fobjc-call-cxx-cdtors @gol
205 -fobjc-direct-dispatch @gol
206 -fobjc-exceptions @gol
208 -freplace-objc-classes @gol
211 -Wassign-intercept @gol
212 -Wno-protocol -Wselector @gol
213 -Wstrict-selector-match @gol
214 -Wundeclared-selector}
216 @item Language Independent Options
217 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
218 @gccoptlist{-fmessage-length=@var{n} @gol
219 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}} @gol
220 -fdiagnostics-show-options
222 @item Warning Options
223 @xref{Warning Options,,Options to Request or Suppress Warnings}.
224 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
225 -w -Wextra -Wall -Waggregate-return -Walways-true -Wno-attributes @gol
226 -Wc++-compat -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
227 -Wconversion -Wno-deprecated-declarations @gol
228 -Wdisabled-optimization -Wno-div-by-zero -Wno-endif-labels @gol
229 -Werror -Werror-implicit-function-declaration @gol
230 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
231 -Wno-format-extra-args -Wformat-nonliteral @gol
232 -Wformat-security -Wformat-y2k @gol
233 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
234 -Wimport -Wno-import -Winit-self -Winline @gol
235 -Wno-int-to-pointer-cast @gol
236 -Wno-invalid-offsetof -Winvalid-pch @gol
237 -Wlarger-than-@var{len} -Wunsafe-loop-optimizations -Wlong-long @gol
238 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
239 -Wmissing-format-attribute -Wmissing-include-dirs @gol
240 -Wmissing-noreturn @gol
241 -Wno-multichar -Wnonnull -Wpacked -Wpadded @gol
242 -Wparentheses -Wpointer-arith -Wno-pointer-to-int-cast @gol
243 -Wredundant-decls @gol
244 -Wreturn-type -Wsequence-point -Wshadow @gol
245 -Wsign-compare -Wstack-protector @gol
246 -Wstrict-aliasing -Wstrict-aliasing=2 @gol
247 -Wstring-literal-comparison @gol
248 -Wswitch -Wswitch-default -Wswitch-enum @gol
249 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
250 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
251 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
252 -Wunused-value -Wunused-variable -Wvariadic-macros @gol
253 -Wvolatile-register-var -Wwrite-strings}
255 @item C-only Warning Options
256 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
257 -Wmissing-prototypes -Wnested-externs -Wold-style-definition @gol
258 -Wstrict-prototypes -Wtraditional @gol
259 -Wdeclaration-after-statement -Wno-pointer-sign}
261 @item Debugging Options
262 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
263 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
264 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
265 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
266 -fdump-ipa-all -fdump-ipa-cgraph @gol
268 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
269 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
270 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
271 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
273 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
274 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
275 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
276 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
277 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
278 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
279 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
280 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
281 -fdump-tree-nrv -fdump-tree-vect @gol
282 -fdump-tree-sink @gol
283 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
284 -fdump-tree-salias @gol
285 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
287 -ftree-vectorizer-verbose=@var{n} @gol
288 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
289 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
290 -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs @gol
291 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
292 -ftest-coverage -ftime-report -fvar-tracking @gol
293 -g -g@var{level} -gcoff -gdwarf-2 @gol
294 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
295 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
296 -print-multi-directory -print-multi-lib @gol
297 -print-prog-name=@var{program} -print-search-dirs -Q @gol
300 @item Optimization Options
301 @xref{Optimize Options,,Options that Control Optimization}.
302 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
303 -falign-labels=@var{n} -falign-loops=@var{n} @gol
304 -fbounds-check -fmudflap -fmudflapth -fmudflapir @gol
305 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
306 -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol
307 -fcaller-saves -fcprop-registers -fcse-follow-jumps @gol
308 -fcse-skip-blocks -fcx-limited-range -fdata-sections @gol
309 -fdelayed-branch -fdelete-null-pointer-checks -fearly-inlining @gol
310 -fexpensive-optimizations -ffast-math -ffloat-store @gol
311 -fforce-addr -ffunction-sections @gol
312 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -fgcse-after-reload @gol
313 -floop-optimize -fcrossjumping -fif-conversion -fif-conversion2 @gol
314 -finline-functions -finline-functions-called-once @gol
315 -finline-limit=@var{n} -fkeep-inline-functions @gol
316 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
317 -fmodulo-sched -fno-branch-count-reg @gol
318 -fno-default-inline -fno-defer-pop -floop-optimize2 -fmove-loop-invariants @gol
319 -fno-function-cse -fno-guess-branch-probability @gol
320 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
321 -funsafe-math-optimizations -funsafe-loop-optimizations -ffinite-math-only @gol
322 -fno-trapping-math -fno-zero-initialized-in-bss @gol
323 -fomit-frame-pointer -foptimize-register-move @gol
324 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
325 -fprofile-generate -fprofile-use @gol
326 -fregmove -frename-registers @gol
327 -freorder-blocks -freorder-blocks-and-partition -freorder-functions @gol
328 -frerun-cse-after-loop -frerun-loop-opt @gol
329 -frounding-math -fschedule-insns -fschedule-insns2 @gol
330 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
331 -fsched-spec-load-dangerous @gol
332 -fsched-stalled-insns=@var{n} -fsched-stalled-insns-dep=@var{n} @gol
333 -fsched2-use-superblocks @gol
334 -fsched2-use-traces -freschedule-modulo-scheduled-loops @gol
335 -fsignaling-nans -fsingle-precision-constant @gol
336 -fstack-protector -fstack-protector-all @gol
337 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
338 -funroll-all-loops -funroll-loops -fpeel-loops @gol
339 -fsplit-ivs-in-unroller -funswitch-loops @gol
340 -fvariable-expansion-in-unroller @gol
341 -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol
342 -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts @gol
343 -ftree-dominator-opts -ftree-dse -ftree-copyrename -ftree-sink @gol
344 -ftree-ch -ftree-sra -ftree-ter -ftree-lrs -ftree-fre -ftree-vectorize @gol
345 -ftree-vect-loop-version -ftree-salias -fweb @gol
346 -ftree-copy-prop -ftree-store-ccp -ftree-store-copy-prop -fwhole-program @gol
347 --param @var{name}=@var{value}
348 -O -O0 -O1 -O2 -O3 -Os}
350 @item Preprocessor Options
351 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
352 @gccoptlist{-A@var{question}=@var{answer} @gol
353 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
354 -C -dD -dI -dM -dN @gol
355 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
356 -idirafter @var{dir} @gol
357 -include @var{file} -imacros @var{file} @gol
358 -iprefix @var{file} -iwithprefix @var{dir} @gol
359 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
360 -isysroot @var{dir} @gol
361 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
362 -P -fworking-directory -remap @gol
363 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
364 -Xpreprocessor @var{option}}
366 @item Assembler Option
367 @xref{Assembler Options,,Passing Options to the Assembler}.
368 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
371 @xref{Link Options,,Options for Linking}.
372 @gccoptlist{@var{object-file-name} -l@var{library} @gol
373 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
374 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
375 -Wl,@var{option} -Xlinker @var{option} @gol
378 @item Directory Options
379 @xref{Directory Options,,Options for Directory Search}.
380 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
381 -specs=@var{file} -I- --sysroot=@var{dir}}
384 @c I wrote this xref this way to avoid overfull hbox. -- rms
385 @xref{Target Options}.
386 @gccoptlist{-V @var{version} -b @var{machine}}
388 @item Machine Dependent Options
389 @xref{Submodel Options,,Hardware Models and Configurations}.
390 @c This list is ordered alphanumerically by subsection name.
391 @c Try and put the significant identifier (CPU or system) first,
392 @c so users have a clue at guessing where the ones they want will be.
395 @gccoptlist{-EB -EL @gol
396 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
397 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
400 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
401 -mabi=@var{name} @gol
402 -mapcs-stack-check -mno-apcs-stack-check @gol
403 -mapcs-float -mno-apcs-float @gol
404 -mapcs-reentrant -mno-apcs-reentrant @gol
405 -msched-prolog -mno-sched-prolog @gol
406 -mlittle-endian -mbig-endian -mwords-little-endian @gol
407 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
408 -mthumb-interwork -mno-thumb-interwork @gol
409 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
410 -mstructure-size-boundary=@var{n} @gol
411 -mabort-on-noreturn @gol
412 -mlong-calls -mno-long-calls @gol
413 -msingle-pic-base -mno-single-pic-base @gol
414 -mpic-register=@var{reg} @gol
415 -mnop-fun-dllimport @gol
416 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
417 -mpoke-function-name @gol
419 -mtpcs-frame -mtpcs-leaf-frame @gol
420 -mcaller-super-interworking -mcallee-super-interworking @gol
424 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
425 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
427 @emph{Blackfin Options}
428 @gccoptlist{-momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
429 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
430 -mlow-64k -mno-low64k -mid-shared-library @gol
431 -mno-id-shared-library -mshared-library-id=@var{n} @gol
432 -mlong-calls -mno-long-calls}
435 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
436 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
437 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
438 -mstack-align -mdata-align -mconst-align @gol
439 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
440 -melf -maout -melinux -mlinux -sim -sim2 @gol
441 -mmul-bug-workaround -mno-mul-bug-workaround}
444 @gccoptlist{-mmac -mpush-args}
446 @emph{Darwin Options}
447 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
448 -arch_only -bind_at_load -bundle -bundle_loader @gol
449 -client_name -compatibility_version -current_version @gol
451 -dependency-file -dylib_file -dylinker_install_name @gol
452 -dynamic -dynamiclib -exported_symbols_list @gol
453 -filelist -flat_namespace -force_cpusubtype_ALL @gol
454 -force_flat_namespace -headerpad_max_install_names @gol
455 -image_base -init -install_name -keep_private_externs @gol
456 -multi_module -multiply_defined -multiply_defined_unused @gol
457 -noall_load -no_dead_strip_inits_and_terms @gol
458 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
459 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
460 -private_bundle -read_only_relocs -sectalign @gol
461 -sectobjectsymbols -whyload -seg1addr @gol
462 -sectcreate -sectobjectsymbols -sectorder @gol
463 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
464 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
465 -segprot -segs_read_only_addr -segs_read_write_addr @gol
466 -single_module -static -sub_library -sub_umbrella @gol
467 -twolevel_namespace -umbrella -undefined @gol
468 -unexported_symbols_list -weak_reference_mismatches @gol
469 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
472 @emph{DEC Alpha Options}
473 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
474 -mieee -mieee-with-inexact -mieee-conformant @gol
475 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
476 -mtrap-precision=@var{mode} -mbuild-constants @gol
477 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
478 -mbwx -mmax -mfix -mcix @gol
479 -mfloat-vax -mfloat-ieee @gol
480 -mexplicit-relocs -msmall-data -mlarge-data @gol
481 -msmall-text -mlarge-text @gol
482 -mmemory-latency=@var{time}}
484 @emph{DEC Alpha/VMS Options}
485 @gccoptlist{-mvms-return-codes}
488 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
489 -mhard-float -msoft-float @gol
490 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
491 -mdouble -mno-double @gol
492 -mmedia -mno-media -mmuladd -mno-muladd @gol
493 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
494 -mlinked-fp -mlong-calls -malign-labels @gol
495 -mlibrary-pic -macc-4 -macc-8 @gol
496 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
497 -moptimize-membar -mno-optimize-membar @gol
498 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
499 -mvliw-branch -mno-vliw-branch @gol
500 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
501 -mno-nested-cond-exec -mtomcat-stats @gol
505 @emph{H8/300 Options}
506 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
509 @gccoptlist{-march=@var{architecture-type} @gol
510 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
511 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
512 -mfixed-range=@var{register-range} @gol
513 -mjump-in-delay -mlinker-opt -mlong-calls @gol
514 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
515 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
516 -mno-jump-in-delay -mno-long-load-store @gol
517 -mno-portable-runtime -mno-soft-float @gol
518 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
519 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
520 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
521 -munix=@var{unix-std} -nolibdld -static -threads}
523 @emph{i386 and x86-64 Options}
524 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
525 -mfpmath=@var{unit} @gol
526 -masm=@var{dialect} -mno-fancy-math-387 @gol
527 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
528 -mno-wide-multiply -mrtd -malign-double @gol
529 -mpreferred-stack-boundary=@var{num} @gol
530 -mmmx -msse -msse2 -msse3 -m3dnow @gol
531 -mthreads -mno-align-stringops -minline-all-stringops @gol
532 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
533 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
534 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
535 -mcmodel=@var{code-model} @gol
536 -m32 -m64 -mlarge-data-threshold=@var{num}}
539 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
540 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
541 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
542 -minline-float-divide-max-throughput @gol
543 -minline-int-divide-min-latency @gol
544 -minline-int-divide-max-throughput @gol
545 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
546 -mno-dwarf2-asm -mearly-stop-bits @gol
547 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
548 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64}
550 @emph{M32R/D Options}
551 @gccoptlist{-m32r2 -m32rx -m32r @gol
553 -malign-loops -mno-align-loops @gol
554 -missue-rate=@var{number} @gol
555 -mbranch-cost=@var{number} @gol
556 -mmodel=@var{code-size-model-type} @gol
557 -msdata=@var{sdata-type} @gol
558 -mno-flush-func -mflush-func=@var{name} @gol
559 -mno-flush-trap -mflush-trap=@var{number} @gol
563 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
565 @emph{M680x0 Options}
566 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
567 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
568 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
569 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
570 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
572 @emph{M68hc1x Options}
573 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
574 -mauto-incdec -minmax -mlong-calls -mshort @gol
575 -msoft-reg-count=@var{count}}
578 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
579 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
580 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
581 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
582 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
585 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
586 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
587 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
588 -mxgot -mno-xgot -mgp32 -mgp64 -mfp32 -mfp64 @gol
589 -mhard-float -msoft-float -msingle-float -mdouble-float @gol
590 -mdsp -mpaired-single -mips3d @gol
591 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
592 -G@var{num} -membedded-data -mno-embedded-data @gol
593 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
594 -msplit-addresses -mno-split-addresses @gol
595 -mexplicit-relocs -mno-explicit-relocs @gol
596 -mcheck-zero-division -mno-check-zero-division @gol
597 -mdivide-traps -mdivide-breaks @gol
598 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
599 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
600 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
601 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 @gol
602 -mfix-sb1 -mno-fix-sb1 @gol
603 -mflush-func=@var{func} -mno-flush-func @gol
604 -mbranch-likely -mno-branch-likely @gol
605 -mfp-exceptions -mno-fp-exceptions @gol
606 -mvr4130-align -mno-vr4130-align}
609 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
610 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
611 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
612 -mno-base-addresses -msingle-exit -mno-single-exit}
614 @emph{MN10300 Options}
615 @gccoptlist{-mmult-bug -mno-mult-bug @gol
616 -mam33 -mno-am33 @gol
617 -mam33-2 -mno-am33-2 @gol
618 -mreturn-pointer-on-d0 @gol
622 @gccoptlist{-mno-crt0 -mbacc -msim @gol
623 -march=@var{cpu-type} }
625 @emph{PDP-11 Options}
626 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
627 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
628 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
629 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
630 -mbranch-expensive -mbranch-cheap @gol
631 -msplit -mno-split -munix-asm -mdec-asm}
633 @emph{PowerPC Options}
634 See RS/6000 and PowerPC Options.
636 @emph{RS/6000 and PowerPC Options}
637 @gccoptlist{-mcpu=@var{cpu-type} @gol
638 -mtune=@var{cpu-type} @gol
639 -mpower -mno-power -mpower2 -mno-power2 @gol
640 -mpowerpc -mpowerpc64 -mno-powerpc @gol
641 -maltivec -mno-altivec @gol
642 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
643 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
644 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
645 -mnew-mnemonics -mold-mnemonics @gol
646 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
647 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
648 -malign-power -malign-natural @gol
649 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
650 -mstring -mno-string -mupdate -mno-update @gol
651 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
652 -mstrict-align -mno-strict-align -mrelocatable @gol
653 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
654 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
655 -mdynamic-no-pic -maltivec -mswdiv @gol
656 -mprioritize-restricted-insns=@var{priority} @gol
657 -msched-costly-dep=@var{dependence_type} @gol
658 -minsert-sched-nops=@var{scheme} @gol
659 -mcall-sysv -mcall-netbsd @gol
660 -maix-struct-return -msvr4-struct-return @gol
661 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
662 -misel -mno-isel @gol
663 -misel=yes -misel=no @gol
665 -mspe=yes -mspe=no @gol
666 -mvrsave -mno-vrsave @gol
667 -mmulhw -mno-mulhw @gol
668 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
669 -mprototype -mno-prototype @gol
670 -msim -mmvme -mads -myellowknife -memb -msdata @gol
671 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
673 @emph{S/390 and zSeries Options}
674 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
675 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
676 -mpacked-stack -mno-packed-stack @gol
677 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
678 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
679 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
680 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
683 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
684 -m4-nofpu -m4-single-only -m4-single -m4 @gol
685 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
686 -m5-64media -m5-64media-nofpu @gol
687 -m5-32media -m5-32media-nofpu @gol
688 -m5-compact -m5-compact-nofpu @gol
689 -mb -ml -mdalign -mrelax @gol
690 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
691 -mieee -misize -mpadstruct -mspace @gol
692 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
693 -mdivsi3_libfunc=@var{name} @gol
694 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
698 @gccoptlist{-mcpu=@var{cpu-type} @gol
699 -mtune=@var{cpu-type} @gol
700 -mcmodel=@var{code-model} @gol
701 -m32 -m64 -mapp-regs -mno-app-regs @gol
702 -mfaster-structs -mno-faster-structs @gol
703 -mfpu -mno-fpu -mhard-float -msoft-float @gol
704 -mhard-quad-float -msoft-quad-float @gol
705 -mimpure-text -mno-impure-text -mlittle-endian @gol
706 -mstack-bias -mno-stack-bias @gol
707 -munaligned-doubles -mno-unaligned-doubles @gol
708 -mv8plus -mno-v8plus -mvis -mno-vis
711 @emph{System V Options}
712 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
714 @emph{TMS320C3x/C4x Options}
715 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
716 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
717 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
718 -mparallel-insns -mparallel-mpy -mpreserve-float}
721 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
722 -mprolog-function -mno-prolog-function -mspace @gol
723 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
724 -mapp-regs -mno-app-regs @gol
725 -mdisable-callt -mno-disable-callt @gol
731 @gccoptlist{-mg -mgnu -munix}
733 @emph{x86-64 Options}
734 See i386 and x86-64 Options.
736 @emph{Xstormy16 Options}
739 @emph{Xtensa Options}
740 @gccoptlist{-mconst16 -mno-const16 @gol
741 -mfused-madd -mno-fused-madd @gol
742 -mtext-section-literals -mno-text-section-literals @gol
743 -mtarget-align -mno-target-align @gol
744 -mlongcalls -mno-longcalls}
746 @emph{zSeries Options}
747 See S/390 and zSeries Options.
749 @item Code Generation Options
750 @xref{Code Gen Options,,Options for Code Generation Conventions}.
751 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
752 -ffixed-@var{reg} -fexceptions @gol
753 -fnon-call-exceptions -funwind-tables @gol
754 -fasynchronous-unwind-tables @gol
755 -finhibit-size-directive -finstrument-functions @gol
756 -fno-common -fno-ident @gol
757 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
758 -fno-jump-tables @gol
759 -freg-struct-return -fshared-data -fshort-enums @gol
760 -fshort-double -fshort-wchar @gol
761 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
762 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
763 -fargument-alias -fargument-noalias @gol
764 -fargument-noalias-global -fleading-underscore @gol
765 -ftls-model=@var{model} @gol
766 -ftrapv -fwrapv -fbounds-check @gol
771 * Overall Options:: Controlling the kind of output:
772 an executable, object files, assembler files,
773 or preprocessed source.
774 * C Dialect Options:: Controlling the variant of C language compiled.
775 * C++ Dialect Options:: Variations on C++.
776 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
778 * Language Independent Options:: Controlling how diagnostics should be
780 * Warning Options:: How picky should the compiler be?
781 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
782 * Optimize Options:: How much optimization?
783 * Preprocessor Options:: Controlling header files and macro definitions.
784 Also, getting dependency information for Make.
785 * Assembler Options:: Passing options to the assembler.
786 * Link Options:: Specifying libraries and so on.
787 * Directory Options:: Where to find header files and libraries.
788 Where to find the compiler executable files.
789 * Spec Files:: How to pass switches to sub-processes.
790 * Target Options:: Running a cross-compiler, or an old version of GCC.
793 @node Overall Options
794 @section Options Controlling the Kind of Output
796 Compilation can involve up to four stages: preprocessing, compilation
797 proper, assembly and linking, always in that order. GCC is capable of
798 preprocessing and compiling several files either into several
799 assembler input files, or into one assembler input file; then each
800 assembler input file produces an object file, and linking combines all
801 the object files (those newly compiled, and those specified as input)
802 into an executable file.
804 @cindex file name suffix
805 For any given input file, the file name suffix determines what kind of
810 C source code which must be preprocessed.
813 C source code which should not be preprocessed.
816 C++ source code which should not be preprocessed.
819 Objective-C source code. Note that you must link with the @file{libobjc}
820 library to make an Objective-C program work.
823 Objective-C source code which should not be preprocessed.
827 Objective-C++ source code. Note that you must link with the @file{libobjc}
828 library to make an Objective-C++ program work. Note that @samp{.M} refers
829 to a literal capital M@.
832 Objective-C++ source code which should not be preprocessed.
835 C, C++, Objective-C or Objective-C++ header file to be turned into a
840 @itemx @var{file}.cxx
841 @itemx @var{file}.cpp
842 @itemx @var{file}.CPP
843 @itemx @var{file}.c++
845 C++ source code which must be preprocessed. Note that in @samp{.cxx},
846 the last two letters must both be literally @samp{x}. Likewise,
847 @samp{.C} refers to a literal capital C@.
851 Objective-C++ source code which must be preprocessed.
854 Objective-C++ source code which should not be preprocessed.
858 C++ header file to be turned into a precompiled header.
861 @itemx @var{file}.for
862 @itemx @var{file}.FOR
863 Fixed form Fortran source code which should not be preprocessed.
866 @itemx @var{file}.fpp
867 @itemx @var{file}.FPP
868 Fixed form Fortran source code which must be preprocessed (with the traditional
872 @itemx @var{file}.f95
873 Free form Fortran source code which should not be preprocessed.
876 @itemx @var{file}.F95
877 Free form Fortran source code which must be preprocessed (with the
878 traditional preprocessor).
880 @c FIXME: Descriptions of Java file types.
887 Ada source code file which contains a library unit declaration (a
888 declaration of a package, subprogram, or generic, or a generic
889 instantiation), or a library unit renaming declaration (a package,
890 generic, or subprogram renaming declaration). Such files are also
893 @itemx @var{file}.adb
894 Ada source code file containing a library unit body (a subprogram or
895 package body). Such files are also called @dfn{bodies}.
897 @c GCC also knows about some suffixes for languages not yet included:
908 Assembler code which must be preprocessed.
911 An object file to be fed straight into linking.
912 Any file name with no recognized suffix is treated this way.
916 You can specify the input language explicitly with the @option{-x} option:
919 @item -x @var{language}
920 Specify explicitly the @var{language} for the following input files
921 (rather than letting the compiler choose a default based on the file
922 name suffix). This option applies to all following input files until
923 the next @option{-x} option. Possible values for @var{language} are:
925 c c-header c-cpp-output
926 c++ c++-header c++-cpp-output
927 objective-c objective-c-header objective-c-cpp-output
928 objective-c++ objective-c++-header objective-c++-cpp-output
929 assembler assembler-with-cpp
938 Turn off any specification of a language, so that subsequent files are
939 handled according to their file name suffixes (as they are if @option{-x}
940 has not been used at all).
942 @item -pass-exit-codes
943 @opindex pass-exit-codes
944 Normally the @command{gcc} program will exit with the code of 1 if any
945 phase of the compiler returns a non-success return code. If you specify
946 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
947 numerically highest error produced by any phase that returned an error
951 If you only want some of the stages of compilation, you can use
952 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
953 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
954 @command{gcc} is to stop. Note that some combinations (for example,
955 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
960 Compile or assemble the source files, but do not link. The linking
961 stage simply is not done. The ultimate output is in the form of an
962 object file for each source file.
964 By default, the object file name for a source file is made by replacing
965 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
967 Unrecognized input files, not requiring compilation or assembly, are
972 Stop after the stage of compilation proper; do not assemble. The output
973 is in the form of an assembler code file for each non-assembler input
976 By default, the assembler file name for a source file is made by
977 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
979 Input files that don't require compilation are ignored.
983 Stop after the preprocessing stage; do not run the compiler proper. The
984 output is in the form of preprocessed source code, which is sent to the
987 Input files which don't require preprocessing are ignored.
989 @cindex output file option
992 Place output in file @var{file}. This applies regardless to whatever
993 sort of output is being produced, whether it be an executable file,
994 an object file, an assembler file or preprocessed C code.
996 If @option{-o} is not specified, the default is to put an executable
997 file in @file{a.out}, the object file for
998 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
999 assembler file in @file{@var{source}.s}, a precompiled header file in
1000 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1005 Print (on standard error output) the commands executed to run the stages
1006 of compilation. Also print the version number of the compiler driver
1007 program and of the preprocessor and the compiler proper.
1011 Like @option{-v} except the commands are not executed and all command
1012 arguments are quoted. This is useful for shell scripts to capture the
1013 driver-generated command lines.
1017 Use pipes rather than temporary files for communication between the
1018 various stages of compilation. This fails to work on some systems where
1019 the assembler is unable to read from a pipe; but the GNU assembler has
1024 If you are compiling multiple source files, this option tells the driver
1025 to pass all the source files to the compiler at once (for those
1026 languages for which the compiler can handle this). This will allow
1027 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1028 language for which this is supported is C@. If you pass source files for
1029 multiple languages to the driver, using this option, the driver will invoke
1030 the compiler(s) that support IMA once each, passing each compiler all the
1031 source files appropriate for it. For those languages that do not support
1032 IMA this option will be ignored, and the compiler will be invoked once for
1033 each source file in that language. If you use this option in conjunction
1034 with @option{-save-temps}, the compiler will generate multiple
1036 (one for each source file), but only one (combined) @file{.o} or
1041 Print (on the standard output) a description of the command line options
1042 understood by @command{gcc}. If the @option{-v} option is also specified
1043 then @option{--help} will also be passed on to the various processes
1044 invoked by @command{gcc}, so that they can display the command line options
1045 they accept. If the @option{-Wextra} option is also specified then command
1046 line options which have no documentation associated with them will also
1050 @opindex target-help
1051 Print (on the standard output) a description of target specific command
1052 line options for each tool.
1056 Display the version number and copyrights of the invoked GCC@.
1058 @include @value{srcdir}/../libiberty/at-file.texi
1062 @section Compiling C++ Programs
1064 @cindex suffixes for C++ source
1065 @cindex C++ source file suffixes
1066 C++ source files conventionally use one of the suffixes @samp{.C},
1067 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1068 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
1069 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1070 files with these names and compiles them as C++ programs even if you
1071 call the compiler the same way as for compiling C programs (usually
1072 with the name @command{gcc}).
1076 However, C++ programs often require class libraries as well as a
1077 compiler that understands the C++ language---and under some
1078 circumstances, you might want to compile programs or header files from
1079 standard input, or otherwise without a suffix that flags them as C++
1080 programs. You might also like to precompile a C header file with a
1081 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
1082 program that calls GCC with the default language set to C++, and
1083 automatically specifies linking against the C++ library. On many
1084 systems, @command{g++} is also installed with the name @command{c++}.
1086 @cindex invoking @command{g++}
1087 When you compile C++ programs, you may specify many of the same
1088 command-line options that you use for compiling programs in any
1089 language; or command-line options meaningful for C and related
1090 languages; or options that are meaningful only for C++ programs.
1091 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1092 explanations of options for languages related to C@.
1093 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1094 explanations of options that are meaningful only for C++ programs.
1096 @node C Dialect Options
1097 @section Options Controlling C Dialect
1098 @cindex dialect options
1099 @cindex language dialect options
1100 @cindex options, dialect
1102 The following options control the dialect of C (or languages derived
1103 from C, such as C++, Objective-C and Objective-C++) that the compiler
1107 @cindex ANSI support
1111 In C mode, support all ISO C90 programs. In C++ mode,
1112 remove GNU extensions that conflict with ISO C++.
1114 This turns off certain features of GCC that are incompatible with ISO
1115 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1116 such as the @code{asm} and @code{typeof} keywords, and
1117 predefined macros such as @code{unix} and @code{vax} that identify the
1118 type of system you are using. It also enables the undesirable and
1119 rarely used ISO trigraph feature. For the C compiler,
1120 it disables recognition of C++ style @samp{//} comments as well as
1121 the @code{inline} keyword.
1123 The alternate keywords @code{__asm__}, @code{__extension__},
1124 @code{__inline__} and @code{__typeof__} continue to work despite
1125 @option{-ansi}. You would not want to use them in an ISO C program, of
1126 course, but it is useful to put them in header files that might be included
1127 in compilations done with @option{-ansi}. Alternate predefined macros
1128 such as @code{__unix__} and @code{__vax__} are also available, with or
1129 without @option{-ansi}.
1131 The @option{-ansi} option does not cause non-ISO programs to be
1132 rejected gratuitously. For that, @option{-pedantic} is required in
1133 addition to @option{-ansi}. @xref{Warning Options}.
1135 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1136 option is used. Some header files may notice this macro and refrain
1137 from declaring certain functions or defining certain macros that the
1138 ISO standard doesn't call for; this is to avoid interfering with any
1139 programs that might use these names for other things.
1141 Functions which would normally be built in but do not have semantics
1142 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1143 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1144 built-in functions provided by GCC}, for details of the functions
1149 Determine the language standard. This option is currently only
1150 supported when compiling C or C++. A value for this option must be
1151 provided; possible values are
1156 ISO C90 (same as @option{-ansi}).
1158 @item iso9899:199409
1159 ISO C90 as modified in amendment 1.
1165 ISO C99. Note that this standard is not yet fully supported; see
1166 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1167 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1170 Default, ISO C90 plus GNU extensions (including some C99 features).
1174 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1175 this will become the default. The name @samp{gnu9x} is deprecated.
1178 The 1998 ISO C++ standard plus amendments.
1181 The same as @option{-std=c++98} plus GNU extensions. This is the
1182 default for C++ code.
1185 Even when this option is not specified, you can still use some of the
1186 features of newer standards in so far as they do not conflict with
1187 previous C standards. For example, you may use @code{__restrict__} even
1188 when @option{-std=c99} is not specified.
1190 The @option{-std} options specifying some version of ISO C have the same
1191 effects as @option{-ansi}, except that features that were not in ISO C90
1192 but are in the specified version (for example, @samp{//} comments and
1193 the @code{inline} keyword in ISO C99) are not disabled.
1195 @xref{Standards,,Language Standards Supported by GCC}, for details of
1196 these standard versions.
1198 @item -aux-info @var{filename}
1200 Output to the given filename prototyped declarations for all functions
1201 declared and/or defined in a translation unit, including those in header
1202 files. This option is silently ignored in any language other than C@.
1204 Besides declarations, the file indicates, in comments, the origin of
1205 each declaration (source file and line), whether the declaration was
1206 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1207 @samp{O} for old, respectively, in the first character after the line
1208 number and the colon), and whether it came from a declaration or a
1209 definition (@samp{C} or @samp{F}, respectively, in the following
1210 character). In the case of function definitions, a K&R-style list of
1211 arguments followed by their declarations is also provided, inside
1212 comments, after the declaration.
1216 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1217 keyword, so that code can use these words as identifiers. You can use
1218 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1219 instead. @option{-ansi} implies @option{-fno-asm}.
1221 In C++, this switch only affects the @code{typeof} keyword, since
1222 @code{asm} and @code{inline} are standard keywords. You may want to
1223 use the @option{-fno-gnu-keywords} flag instead, which has the same
1224 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1225 switch only affects the @code{asm} and @code{typeof} keywords, since
1226 @code{inline} is a standard keyword in ISO C99.
1229 @itemx -fno-builtin-@var{function}
1230 @opindex fno-builtin
1231 @cindex built-in functions
1232 Don't recognize built-in functions that do not begin with
1233 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1234 functions provided by GCC}, for details of the functions affected,
1235 including those which are not built-in functions when @option{-ansi} or
1236 @option{-std} options for strict ISO C conformance are used because they
1237 do not have an ISO standard meaning.
1239 GCC normally generates special code to handle certain built-in functions
1240 more efficiently; for instance, calls to @code{alloca} may become single
1241 instructions that adjust the stack directly, and calls to @code{memcpy}
1242 may become inline copy loops. The resulting code is often both smaller
1243 and faster, but since the function calls no longer appear as such, you
1244 cannot set a breakpoint on those calls, nor can you change the behavior
1245 of the functions by linking with a different library. In addition,
1246 when a function is recognized as a built-in function, GCC may use
1247 information about that function to warn about problems with calls to
1248 that function, or to generate more efficient code, even if the
1249 resulting code still contains calls to that function. For example,
1250 warnings are given with @option{-Wformat} for bad calls to
1251 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1252 known not to modify global memory.
1254 With the @option{-fno-builtin-@var{function}} option
1255 only the built-in function @var{function} is
1256 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1257 function is named this is not built-in in this version of GCC, this
1258 option is ignored. There is no corresponding
1259 @option{-fbuiltin-@var{function}} option; if you wish to enable
1260 built-in functions selectively when using @option{-fno-builtin} or
1261 @option{-ffreestanding}, you may define macros such as:
1264 #define abs(n) __builtin_abs ((n))
1265 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1270 @cindex hosted environment
1272 Assert that compilation takes place in a hosted environment. This implies
1273 @option{-fbuiltin}. A hosted environment is one in which the
1274 entire standard library is available, and in which @code{main} has a return
1275 type of @code{int}. Examples are nearly everything except a kernel.
1276 This is equivalent to @option{-fno-freestanding}.
1278 @item -ffreestanding
1279 @opindex ffreestanding
1280 @cindex hosted environment
1282 Assert that compilation takes place in a freestanding environment. This
1283 implies @option{-fno-builtin}. A freestanding environment
1284 is one in which the standard library may not exist, and program startup may
1285 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1286 This is equivalent to @option{-fno-hosted}.
1288 @xref{Standards,,Language Standards Supported by GCC}, for details of
1289 freestanding and hosted environments.
1291 @item -fms-extensions
1292 @opindex fms-extensions
1293 Accept some non-standard constructs used in Microsoft header files.
1295 Some cases of unnamed fields in structures and unions are only
1296 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1297 fields within structs/unions}, for details.
1301 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1302 options for strict ISO C conformance) implies @option{-trigraphs}.
1304 @item -no-integrated-cpp
1305 @opindex no-integrated-cpp
1306 Performs a compilation in two passes: preprocessing and compiling. This
1307 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1308 @option{-B} option. The user supplied compilation step can then add in
1309 an additional preprocessing step after normal preprocessing but before
1310 compiling. The default is to use the integrated cpp (internal cpp)
1312 The semantics of this option will change if "cc1", "cc1plus", and
1313 "cc1obj" are merged.
1315 @cindex traditional C language
1316 @cindex C language, traditional
1318 @itemx -traditional-cpp
1319 @opindex traditional-cpp
1320 @opindex traditional
1321 Formerly, these options caused GCC to attempt to emulate a pre-standard
1322 C compiler. They are now only supported with the @option{-E} switch.
1323 The preprocessor continues to support a pre-standard mode. See the GNU
1324 CPP manual for details.
1326 @item -fcond-mismatch
1327 @opindex fcond-mismatch
1328 Allow conditional expressions with mismatched types in the second and
1329 third arguments. The value of such an expression is void. This option
1330 is not supported for C++.
1332 @item -funsigned-char
1333 @opindex funsigned-char
1334 Let the type @code{char} be unsigned, like @code{unsigned char}.
1336 Each kind of machine has a default for what @code{char} should
1337 be. It is either like @code{unsigned char} by default or like
1338 @code{signed char} by default.
1340 Ideally, a portable program should always use @code{signed char} or
1341 @code{unsigned char} when it depends on the signedness of an object.
1342 But many programs have been written to use plain @code{char} and
1343 expect it to be signed, or expect it to be unsigned, depending on the
1344 machines they were written for. This option, and its inverse, let you
1345 make such a program work with the opposite default.
1347 The type @code{char} is always a distinct type from each of
1348 @code{signed char} or @code{unsigned char}, even though its behavior
1349 is always just like one of those two.
1352 @opindex fsigned-char
1353 Let the type @code{char} be signed, like @code{signed char}.
1355 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1356 the negative form of @option{-funsigned-char}. Likewise, the option
1357 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1359 @item -fsigned-bitfields
1360 @itemx -funsigned-bitfields
1361 @itemx -fno-signed-bitfields
1362 @itemx -fno-unsigned-bitfields
1363 @opindex fsigned-bitfields
1364 @opindex funsigned-bitfields
1365 @opindex fno-signed-bitfields
1366 @opindex fno-unsigned-bitfields
1367 These options control whether a bit-field is signed or unsigned, when the
1368 declaration does not use either @code{signed} or @code{unsigned}. By
1369 default, such a bit-field is signed, because this is consistent: the
1370 basic integer types such as @code{int} are signed types.
1373 @node C++ Dialect Options
1374 @section Options Controlling C++ Dialect
1376 @cindex compiler options, C++
1377 @cindex C++ options, command line
1378 @cindex options, C++
1379 This section describes the command-line options that are only meaningful
1380 for C++ programs; but you can also use most of the GNU compiler options
1381 regardless of what language your program is in. For example, you
1382 might compile a file @code{firstClass.C} like this:
1385 g++ -g -frepo -O -c firstClass.C
1389 In this example, only @option{-frepo} is an option meant
1390 only for C++ programs; you can use the other options with any
1391 language supported by GCC@.
1393 Here is a list of options that are @emph{only} for compiling C++ programs:
1397 @item -fabi-version=@var{n}
1398 @opindex fabi-version
1399 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1400 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1401 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1402 the version that conforms most closely to the C++ ABI specification.
1403 Therefore, the ABI obtained using version 0 will change as ABI bugs
1406 The default is version 2.
1408 @item -fno-access-control
1409 @opindex fno-access-control
1410 Turn off all access checking. This switch is mainly useful for working
1411 around bugs in the access control code.
1415 Check that the pointer returned by @code{operator new} is non-null
1416 before attempting to modify the storage allocated. This check is
1417 normally unnecessary because the C++ standard specifies that
1418 @code{operator new} will only return @code{0} if it is declared
1419 @samp{throw()}, in which case the compiler will always check the
1420 return value even without this option. In all other cases, when
1421 @code{operator new} has a non-empty exception specification, memory
1422 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1423 @samp{new (nothrow)}.
1425 @item -fconserve-space
1426 @opindex fconserve-space
1427 Put uninitialized or runtime-initialized global variables into the
1428 common segment, as C does. This saves space in the executable at the
1429 cost of not diagnosing duplicate definitions. If you compile with this
1430 flag and your program mysteriously crashes after @code{main()} has
1431 completed, you may have an object that is being destroyed twice because
1432 two definitions were merged.
1434 This option is no longer useful on most targets, now that support has
1435 been added for putting variables into BSS without making them common.
1437 @item -ffriend-injection
1438 @opindex ffriend-injection
1439 Inject friend functions into the enclosing namespace, so that they are
1440 visible outside the scope of the class in which they are declared.
1441 Friend functions were documented to work this way in the old Annotated
1442 C++ Reference Manual, and versions of G++ before 4.1 always worked
1443 that way. However, in ISO C++ a friend function which is not declared
1444 in an enclosing scope can only be found using argument dependent
1445 lookup. This option causes friends to be injected as they were in
1448 This option is for compatibility, and may be removed in a future
1451 @item -fno-const-strings
1452 @opindex fno-const-strings
1453 Give string constants type @code{char *} instead of type @code{const
1454 char *}. By default, G++ uses type @code{const char *} as required by
1455 the standard. Even if you use @option{-fno-const-strings}, you cannot
1456 actually modify the value of a string constant.
1458 This option might be removed in a future release of G++. For maximum
1459 portability, you should structure your code so that it works with
1460 string constants that have type @code{const char *}.
1462 @item -fno-elide-constructors
1463 @opindex fno-elide-constructors
1464 The C++ standard allows an implementation to omit creating a temporary
1465 which is only used to initialize another object of the same type.
1466 Specifying this option disables that optimization, and forces G++ to
1467 call the copy constructor in all cases.
1469 @item -fno-enforce-eh-specs
1470 @opindex fno-enforce-eh-specs
1471 Don't generate code to check for violation of exception specifications
1472 at runtime. This option violates the C++ standard, but may be useful
1473 for reducing code size in production builds, much like defining
1474 @samp{NDEBUG}. This does not give user code permission to throw
1475 exceptions in violation of the exception specifications; the compiler
1476 will still optimize based on the specifications, so throwing an
1477 unexpected exception will result in undefined behavior.
1480 @itemx -fno-for-scope
1482 @opindex fno-for-scope
1483 If @option{-ffor-scope} is specified, the scope of variables declared in
1484 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1485 as specified by the C++ standard.
1486 If @option{-fno-for-scope} is specified, the scope of variables declared in
1487 a @i{for-init-statement} extends to the end of the enclosing scope,
1488 as was the case in old versions of G++, and other (traditional)
1489 implementations of C++.
1491 The default if neither flag is given to follow the standard,
1492 but to allow and give a warning for old-style code that would
1493 otherwise be invalid, or have different behavior.
1495 @item -fno-gnu-keywords
1496 @opindex fno-gnu-keywords
1497 Do not recognize @code{typeof} as a keyword, so that code can use this
1498 word as an identifier. You can use the keyword @code{__typeof__} instead.
1499 @option{-ansi} implies @option{-fno-gnu-keywords}.
1501 @item -fno-implicit-templates
1502 @opindex fno-implicit-templates
1503 Never emit code for non-inline templates which are instantiated
1504 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1505 @xref{Template Instantiation}, for more information.
1507 @item -fno-implicit-inline-templates
1508 @opindex fno-implicit-inline-templates
1509 Don't emit code for implicit instantiations of inline templates, either.
1510 The default is to handle inlines differently so that compiles with and
1511 without optimization will need the same set of explicit instantiations.
1513 @item -fno-implement-inlines
1514 @opindex fno-implement-inlines
1515 To save space, do not emit out-of-line copies of inline functions
1516 controlled by @samp{#pragma implementation}. This will cause linker
1517 errors if these functions are not inlined everywhere they are called.
1519 @item -fms-extensions
1520 @opindex fms-extensions
1521 Disable pedantic warnings about constructs used in MFC, such as implicit
1522 int and getting a pointer to member function via non-standard syntax.
1524 @item -fno-nonansi-builtins
1525 @opindex fno-nonansi-builtins
1526 Disable built-in declarations of functions that are not mandated by
1527 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1528 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1530 @item -fno-operator-names
1531 @opindex fno-operator-names
1532 Do not treat the operator name keywords @code{and}, @code{bitand},
1533 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1534 synonyms as keywords.
1536 @item -fno-optional-diags
1537 @opindex fno-optional-diags
1538 Disable diagnostics that the standard says a compiler does not need to
1539 issue. Currently, the only such diagnostic issued by G++ is the one for
1540 a name having multiple meanings within a class.
1543 @opindex fpermissive
1544 Downgrade some diagnostics about nonconformant code from errors to
1545 warnings. Thus, using @option{-fpermissive} will allow some
1546 nonconforming code to compile.
1550 Enable automatic template instantiation at link time. This option also
1551 implies @option{-fno-implicit-templates}. @xref{Template
1552 Instantiation}, for more information.
1556 Disable generation of information about every class with virtual
1557 functions for use by the C++ runtime type identification features
1558 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1559 of the language, you can save some space by using this flag. Note that
1560 exception handling uses the same information, but it will generate it as
1565 Emit statistics about front-end processing at the end of the compilation.
1566 This information is generally only useful to the G++ development team.
1568 @item -ftemplate-depth-@var{n}
1569 @opindex ftemplate-depth
1570 Set the maximum instantiation depth for template classes to @var{n}.
1571 A limit on the template instantiation depth is needed to detect
1572 endless recursions during template class instantiation. ANSI/ISO C++
1573 conforming programs must not rely on a maximum depth greater than 17.
1575 @item -fno-threadsafe-statics
1576 @opindex fno-threadsafe-statics
1577 Do not emit the extra code to use the routines specified in the C++
1578 ABI for thread-safe initialization of local statics. You can use this
1579 option to reduce code size slightly in code that doesn't need to be
1582 @item -fuse-cxa-atexit
1583 @opindex fuse-cxa-atexit
1584 Register destructors for objects with static storage duration with the
1585 @code{__cxa_atexit} function rather than the @code{atexit} function.
1586 This option is required for fully standards-compliant handling of static
1587 destructors, but will only work if your C library supports
1588 @code{__cxa_atexit}.
1590 @item -fvisibility-inlines-hidden
1591 @opindex fvisibility-inlines-hidden
1592 Causes all inlined methods to be marked with
1593 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1594 appear in the export table of a DSO and do not require a PLT indirection
1595 when used within the DSO@. Enabling this option can have a dramatic effect
1596 on load and link times of a DSO as it massively reduces the size of the
1597 dynamic export table when the library makes heavy use of templates. While
1598 it can cause bloating through duplication of code within each DSO where
1599 it is used, often the wastage is less than the considerable space occupied
1600 by a long symbol name in the export table which is typical when using
1601 templates and namespaces. For even more savings, combine with the
1602 @option{-fvisibility=hidden} switch.
1606 Do not use weak symbol support, even if it is provided by the linker.
1607 By default, G++ will use weak symbols if they are available. This
1608 option exists only for testing, and should not be used by end-users;
1609 it will result in inferior code and has no benefits. This option may
1610 be removed in a future release of G++.
1614 Do not search for header files in the standard directories specific to
1615 C++, but do still search the other standard directories. (This option
1616 is used when building the C++ library.)
1619 In addition, these optimization, warning, and code generation options
1620 have meanings only for C++ programs:
1623 @item -fno-default-inline
1624 @opindex fno-default-inline
1625 Do not assume @samp{inline} for functions defined inside a class scope.
1626 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1627 functions will have linkage like inline functions; they just won't be
1630 @item -Wabi @r{(C++ only)}
1632 Warn when G++ generates code that is probably not compatible with the
1633 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1634 all such cases, there are probably some cases that are not warned about,
1635 even though G++ is generating incompatible code. There may also be
1636 cases where warnings are emitted even though the code that is generated
1639 You should rewrite your code to avoid these warnings if you are
1640 concerned about the fact that code generated by G++ may not be binary
1641 compatible with code generated by other compilers.
1643 The known incompatibilities at this point include:
1648 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1649 pack data into the same byte as a base class. For example:
1652 struct A @{ virtual void f(); int f1 : 1; @};
1653 struct B : public A @{ int f2 : 1; @};
1657 In this case, G++ will place @code{B::f2} into the same byte
1658 as@code{A::f1}; other compilers will not. You can avoid this problem
1659 by explicitly padding @code{A} so that its size is a multiple of the
1660 byte size on your platform; that will cause G++ and other compilers to
1661 layout @code{B} identically.
1664 Incorrect handling of tail-padding for virtual bases. G++ does not use
1665 tail padding when laying out virtual bases. For example:
1668 struct A @{ virtual void f(); char c1; @};
1669 struct B @{ B(); char c2; @};
1670 struct C : public A, public virtual B @{@};
1674 In this case, G++ will not place @code{B} into the tail-padding for
1675 @code{A}; other compilers will. You can avoid this problem by
1676 explicitly padding @code{A} so that its size is a multiple of its
1677 alignment (ignoring virtual base classes); that will cause G++ and other
1678 compilers to layout @code{C} identically.
1681 Incorrect handling of bit-fields with declared widths greater than that
1682 of their underlying types, when the bit-fields appear in a union. For
1686 union U @{ int i : 4096; @};
1690 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1691 union too small by the number of bits in an @code{int}.
1694 Empty classes can be placed at incorrect offsets. For example:
1704 struct C : public B, public A @{@};
1708 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1709 it should be placed at offset zero. G++ mistakenly believes that the
1710 @code{A} data member of @code{B} is already at offset zero.
1713 Names of template functions whose types involve @code{typename} or
1714 template template parameters can be mangled incorrectly.
1717 template <typename Q>
1718 void f(typename Q::X) @{@}
1720 template <template <typename> class Q>
1721 void f(typename Q<int>::X) @{@}
1725 Instantiations of these templates may be mangled incorrectly.
1729 @item -Wctor-dtor-privacy @r{(C++ only)}
1730 @opindex Wctor-dtor-privacy
1731 Warn when a class seems unusable because all the constructors or
1732 destructors in that class are private, and it has neither friends nor
1733 public static member functions.
1735 @item -Wnon-virtual-dtor @r{(C++ only)}
1736 @opindex Wnon-virtual-dtor
1737 Warn when a class appears to be polymorphic, thereby requiring a virtual
1738 destructor, yet it declares a non-virtual one.
1739 This warning is enabled by @option{-Wall}.
1741 @item -Wreorder @r{(C++ only)}
1743 @cindex reordering, warning
1744 @cindex warning for reordering of member initializers
1745 Warn when the order of member initializers given in the code does not
1746 match the order in which they must be executed. For instance:
1752 A(): j (0), i (1) @{ @}
1756 The compiler will rearrange the member initializers for @samp{i}
1757 and @samp{j} to match the declaration order of the members, emitting
1758 a warning to that effect. This warning is enabled by @option{-Wall}.
1761 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1764 @item -Weffc++ @r{(C++ only)}
1766 Warn about violations of the following style guidelines from Scott Meyers'
1767 @cite{Effective C++} book:
1771 Item 11: Define a copy constructor and an assignment operator for classes
1772 with dynamically allocated memory.
1775 Item 12: Prefer initialization to assignment in constructors.
1778 Item 14: Make destructors virtual in base classes.
1781 Item 15: Have @code{operator=} return a reference to @code{*this}.
1784 Item 23: Don't try to return a reference when you must return an object.
1788 Also warn about violations of the following style guidelines from
1789 Scott Meyers' @cite{More Effective C++} book:
1793 Item 6: Distinguish between prefix and postfix forms of increment and
1794 decrement operators.
1797 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1801 When selecting this option, be aware that the standard library
1802 headers do not obey all of these guidelines; use @samp{grep -v}
1803 to filter out those warnings.
1805 @item -Wno-deprecated @r{(C++ only)}
1806 @opindex Wno-deprecated
1807 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1809 @item -Wstrict-null-sentinel @r{(C++ only)}
1810 @opindex Wstrict-null-sentinel
1811 Warn also about the use of an uncasted @code{NULL} as sentinel. When
1812 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
1813 to @code{__null}. Although it is a null pointer constant not a null pointer,
1814 it is guaranteed to of the same size as a pointer. But this use is
1815 not portable across different compilers.
1817 @item -Wno-non-template-friend @r{(C++ only)}
1818 @opindex Wno-non-template-friend
1819 Disable warnings when non-templatized friend functions are declared
1820 within a template. Since the advent of explicit template specification
1821 support in G++, if the name of the friend is an unqualified-id (i.e.,
1822 @samp{friend foo(int)}), the C++ language specification demands that the
1823 friend declare or define an ordinary, nontemplate function. (Section
1824 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1825 could be interpreted as a particular specialization of a templatized
1826 function. Because this non-conforming behavior is no longer the default
1827 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1828 check existing code for potential trouble spots and is on by default.
1829 This new compiler behavior can be turned off with
1830 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1831 but disables the helpful warning.
1833 @item -Wold-style-cast @r{(C++ only)}
1834 @opindex Wold-style-cast
1835 Warn if an old-style (C-style) cast to a non-void type is used within
1836 a C++ program. The new-style casts (@samp{dynamic_cast},
1837 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
1838 less vulnerable to unintended effects and much easier to search for.
1840 @item -Woverloaded-virtual @r{(C++ only)}
1841 @opindex Woverloaded-virtual
1842 @cindex overloaded virtual fn, warning
1843 @cindex warning for overloaded virtual fn
1844 Warn when a function declaration hides virtual functions from a
1845 base class. For example, in:
1852 struct B: public A @{
1857 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1865 will fail to compile.
1867 @item -Wno-pmf-conversions @r{(C++ only)}
1868 @opindex Wno-pmf-conversions
1869 Disable the diagnostic for converting a bound pointer to member function
1872 @item -Wsign-promo @r{(C++ only)}
1873 @opindex Wsign-promo
1874 Warn when overload resolution chooses a promotion from unsigned or
1875 enumerated type to a signed type, over a conversion to an unsigned type of
1876 the same size. Previous versions of G++ would try to preserve
1877 unsignedness, but the standard mandates the current behavior.
1882 A& operator = (int);
1892 In this example, G++ will synthesize a default @samp{A& operator =
1893 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1896 @node Objective-C and Objective-C++ Dialect Options
1897 @section Options Controlling Objective-C and Objective-C++ Dialects
1899 @cindex compiler options, Objective-C and Objective-C++
1900 @cindex Objective-C and Objective-C++ options, command line
1901 @cindex options, Objective-C and Objective-C++
1902 (NOTE: This manual does not describe the Objective-C and Objective-C++
1903 languages themselves. See @xref{Standards,,Language Standards
1904 Supported by GCC}, for references.)
1906 This section describes the command-line options that are only meaningful
1907 for Objective-C and Objective-C++ programs, but you can also use most of
1908 the language-independent GNU compiler options.
1909 For example, you might compile a file @code{some_class.m} like this:
1912 gcc -g -fgnu-runtime -O -c some_class.m
1916 In this example, @option{-fgnu-runtime} is an option meant only for
1917 Objective-C and Objective-C++ programs; you can use the other options with
1918 any language supported by GCC@.
1920 Note that since Objective-C is an extension of the C language, Objective-C
1921 compilations may also use options specific to the C front-end (e.g.,
1922 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
1923 C++-specific options (e.g., @option{-Wabi}).
1925 Here is a list of options that are @emph{only} for compiling Objective-C
1926 and Objective-C++ programs:
1929 @item -fconstant-string-class=@var{class-name}
1930 @opindex fconstant-string-class
1931 Use @var{class-name} as the name of the class to instantiate for each
1932 literal string specified with the syntax @code{@@"@dots{}"}. The default
1933 class name is @code{NXConstantString} if the GNU runtime is being used, and
1934 @code{NSConstantString} if the NeXT runtime is being used (see below). The
1935 @option{-fconstant-cfstrings} option, if also present, will override the
1936 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
1937 to be laid out as constant CoreFoundation strings.
1940 @opindex fgnu-runtime
1941 Generate object code compatible with the standard GNU Objective-C
1942 runtime. This is the default for most types of systems.
1944 @item -fnext-runtime
1945 @opindex fnext-runtime
1946 Generate output compatible with the NeXT runtime. This is the default
1947 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1948 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1951 @item -fno-nil-receivers
1952 @opindex fno-nil-receivers
1953 Assume that all Objective-C message dispatches (e.g.,
1954 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
1955 is not @code{nil}. This allows for more efficient entry points in the runtime
1956 to be used. Currently, this option is only available in conjunction with
1957 the NeXT runtime on Mac OS X 10.3 and later.
1959 @item -fobjc-call-cxx-cdtors
1960 @opindex fobjc-call-cxx-cdtors
1961 For each Objective-C class, check if any of its instance variables is a
1962 C++ object with a non-trivial default constructor. If so, synthesize a
1963 special @code{- (id) .cxx_construct} instance method that will run
1964 non-trivial default constructors on any such instance variables, in order,
1965 and then return @code{self}. Similarly, check if any instance variable
1966 is a C++ object with a non-trivial destructor, and if so, synthesize a
1967 special @code{- (void) .cxx_destruct} method that will run
1968 all such default destructors, in reverse order.
1970 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
1971 thusly generated will only operate on instance variables declared in the
1972 current Objective-C class, and not those inherited from superclasses. It
1973 is the responsibility of the Objective-C runtime to invoke all such methods
1974 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
1975 will be invoked by the runtime immediately after a new object
1976 instance is allocated; the @code{- (void) .cxx_destruct} methods will
1977 be invoked immediately before the runtime deallocates an object instance.
1979 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
1980 support for invoking the @code{- (id) .cxx_construct} and
1981 @code{- (void) .cxx_destruct} methods.
1983 @item -fobjc-direct-dispatch
1984 @opindex fobjc-direct-dispatch
1985 Allow fast jumps to the message dispatcher. On Darwin this is
1986 accomplished via the comm page.
1988 @item -fobjc-exceptions
1989 @opindex fobjc-exceptions
1990 Enable syntactic support for structured exception handling in Objective-C,
1991 similar to what is offered by C++ and Java. This option is
1992 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2001 @@catch (AnObjCClass *exc) @{
2008 @@catch (AnotherClass *exc) @{
2011 @@catch (id allOthers) @{
2021 The @code{@@throw} statement may appear anywhere in an Objective-C or
2022 Objective-C++ program; when used inside of a @code{@@catch} block, the
2023 @code{@@throw} may appear without an argument (as shown above), in which case
2024 the object caught by the @code{@@catch} will be rethrown.
2026 Note that only (pointers to) Objective-C objects may be thrown and
2027 caught using this scheme. When an object is thrown, it will be caught
2028 by the nearest @code{@@catch} clause capable of handling objects of that type,
2029 analogously to how @code{catch} blocks work in C++ and Java. A
2030 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2031 any and all Objective-C exceptions not caught by previous @code{@@catch}
2034 The @code{@@finally} clause, if present, will be executed upon exit from the
2035 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2036 regardless of whether any exceptions are thrown, caught or rethrown
2037 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2038 of the @code{finally} clause in Java.
2040 There are several caveats to using the new exception mechanism:
2044 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2045 idioms provided by the @code{NSException} class, the new
2046 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2047 systems, due to additional functionality needed in the (NeXT) Objective-C
2051 As mentioned above, the new exceptions do not support handling
2052 types other than Objective-C objects. Furthermore, when used from
2053 Objective-C++, the Objective-C exception model does not interoperate with C++
2054 exceptions at this time. This means you cannot @code{@@throw} an exception
2055 from Objective-C and @code{catch} it in C++, or vice versa
2056 (i.e., @code{throw @dots{} @@catch}).
2059 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2060 blocks for thread-safe execution:
2063 @@synchronized (ObjCClass *guard) @{
2068 Upon entering the @code{@@synchronized} block, a thread of execution shall
2069 first check whether a lock has been placed on the corresponding @code{guard}
2070 object by another thread. If it has, the current thread shall wait until
2071 the other thread relinquishes its lock. Once @code{guard} becomes available,
2072 the current thread will place its own lock on it, execute the code contained in
2073 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2074 making @code{guard} available to other threads).
2076 Unlike Java, Objective-C does not allow for entire methods to be marked
2077 @code{@@synchronized}. Note that throwing exceptions out of
2078 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2079 to be unlocked properly.
2083 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2085 @item -freplace-objc-classes
2086 @opindex freplace-objc-classes
2087 Emit a special marker instructing @command{ld(1)} not to statically link in
2088 the resulting object file, and allow @command{dyld(1)} to load it in at
2089 run time instead. This is used in conjunction with the Fix-and-Continue
2090 debugging mode, where the object file in question may be recompiled and
2091 dynamically reloaded in the course of program execution, without the need
2092 to restart the program itself. Currently, Fix-and-Continue functionality
2093 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2098 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2099 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2100 compile time) with static class references that get initialized at load time,
2101 which improves run-time performance. Specifying the @option{-fzero-link} flag
2102 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2103 to be retained. This is useful in Zero-Link debugging mode, since it allows
2104 for individual class implementations to be modified during program execution.
2108 Dump interface declarations for all classes seen in the source file to a
2109 file named @file{@var{sourcename}.decl}.
2111 @item -Wassign-intercept
2112 @opindex Wassign-intercept
2113 Warn whenever an Objective-C assignment is being intercepted by the
2117 @opindex Wno-protocol
2118 If a class is declared to implement a protocol, a warning is issued for
2119 every method in the protocol that is not implemented by the class. The
2120 default behavior is to issue a warning for every method not explicitly
2121 implemented in the class, even if a method implementation is inherited
2122 from the superclass. If you use the @option{-Wno-protocol} option, then
2123 methods inherited from the superclass are considered to be implemented,
2124 and no warning is issued for them.
2128 Warn if multiple methods of different types for the same selector are
2129 found during compilation. The check is performed on the list of methods
2130 in the final stage of compilation. Additionally, a check is performed
2131 for each selector appearing in a @code{@@selector(@dots{})}
2132 expression, and a corresponding method for that selector has been found
2133 during compilation. Because these checks scan the method table only at
2134 the end of compilation, these warnings are not produced if the final
2135 stage of compilation is not reached, for example because an error is
2136 found during compilation, or because the @option{-fsyntax-only} option is
2139 @item -Wstrict-selector-match
2140 @opindex Wstrict-selector-match
2141 Warn if multiple methods with differing argument and/or return types are
2142 found for a given selector when attempting to send a message using this
2143 selector to a receiver of type @code{id} or @code{Class}. When this flag
2144 is off (which is the default behavior), the compiler will omit such warnings
2145 if any differences found are confined to types which share the same size
2148 @item -Wundeclared-selector
2149 @opindex Wundeclared-selector
2150 Warn if a @code{@@selector(@dots{})} expression referring to an
2151 undeclared selector is found. A selector is considered undeclared if no
2152 method with that name has been declared before the
2153 @code{@@selector(@dots{})} expression, either explicitly in an
2154 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2155 an @code{@@implementation} section. This option always performs its
2156 checks as soon as a @code{@@selector(@dots{})} expression is found,
2157 while @option{-Wselector} only performs its checks in the final stage of
2158 compilation. This also enforces the coding style convention
2159 that methods and selectors must be declared before being used.
2161 @item -print-objc-runtime-info
2162 @opindex print-objc-runtime-info
2163 Generate C header describing the largest structure that is passed by
2168 @node Language Independent Options
2169 @section Options to Control Diagnostic Messages Formatting
2170 @cindex options to control diagnostics formatting
2171 @cindex diagnostic messages
2172 @cindex message formatting
2174 Traditionally, diagnostic messages have been formatted irrespective of
2175 the output device's aspect (e.g.@: its width, @dots{}). The options described
2176 below can be used to control the diagnostic messages formatting
2177 algorithm, e.g.@: how many characters per line, how often source location
2178 information should be reported. Right now, only the C++ front end can
2179 honor these options. However it is expected, in the near future, that
2180 the remaining front ends would be able to digest them correctly.
2183 @item -fmessage-length=@var{n}
2184 @opindex fmessage-length
2185 Try to format error messages so that they fit on lines of about @var{n}
2186 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2187 the front ends supported by GCC@. If @var{n} is zero, then no
2188 line-wrapping will be done; each error message will appear on a single
2191 @opindex fdiagnostics-show-location
2192 @item -fdiagnostics-show-location=once
2193 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2194 reporter to emit @emph{once} source location information; that is, in
2195 case the message is too long to fit on a single physical line and has to
2196 be wrapped, the source location won't be emitted (as prefix) again,
2197 over and over, in subsequent continuation lines. This is the default
2200 @item -fdiagnostics-show-location=every-line
2201 Only meaningful in line-wrapping mode. Instructs the diagnostic
2202 messages reporter to emit the same source location information (as
2203 prefix) for physical lines that result from the process of breaking
2204 a message which is too long to fit on a single line.
2206 @item -fdiagnostics-show-options
2207 @opindex fdiagnostics-show-options
2208 This option instructs the diagnostic machinery to add text to each
2209 diagnostic emitted, which indicates which command line option directly
2210 controls that diagnostic, when such an option is known to the
2211 diagnostic machinery.
2215 @node Warning Options
2216 @section Options to Request or Suppress Warnings
2217 @cindex options to control warnings
2218 @cindex warning messages
2219 @cindex messages, warning
2220 @cindex suppressing warnings
2222 Warnings are diagnostic messages that report constructions which
2223 are not inherently erroneous but which are risky or suggest there
2224 may have been an error.
2226 You can request many specific warnings with options beginning @samp{-W},
2227 for example @option{-Wimplicit} to request warnings on implicit
2228 declarations. Each of these specific warning options also has a
2229 negative form beginning @samp{-Wno-} to turn off warnings;
2230 for example, @option{-Wno-implicit}. This manual lists only one of the
2231 two forms, whichever is not the default.
2233 The following options control the amount and kinds of warnings produced
2234 by GCC; for further, language-specific options also refer to
2235 @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect
2239 @cindex syntax checking
2241 @opindex fsyntax-only
2242 Check the code for syntax errors, but don't do anything beyond that.
2246 Issue all the warnings demanded by strict ISO C and ISO C++;
2247 reject all programs that use forbidden extensions, and some other
2248 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2249 version of the ISO C standard specified by any @option{-std} option used.
2251 Valid ISO C and ISO C++ programs should compile properly with or without
2252 this option (though a rare few will require @option{-ansi} or a
2253 @option{-std} option specifying the required version of ISO C)@. However,
2254 without this option, certain GNU extensions and traditional C and C++
2255 features are supported as well. With this option, they are rejected.
2257 @option{-pedantic} does not cause warning messages for use of the
2258 alternate keywords whose names begin and end with @samp{__}. Pedantic
2259 warnings are also disabled in the expression that follows
2260 @code{__extension__}. However, only system header files should use
2261 these escape routes; application programs should avoid them.
2262 @xref{Alternate Keywords}.
2264 Some users try to use @option{-pedantic} to check programs for strict ISO
2265 C conformance. They soon find that it does not do quite what they want:
2266 it finds some non-ISO practices, but not all---only those for which
2267 ISO C @emph{requires} a diagnostic, and some others for which
2268 diagnostics have been added.
2270 A feature to report any failure to conform to ISO C might be useful in
2271 some instances, but would require considerable additional work and would
2272 be quite different from @option{-pedantic}. We don't have plans to
2273 support such a feature in the near future.
2275 Where the standard specified with @option{-std} represents a GNU
2276 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2277 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2278 extended dialect is based. Warnings from @option{-pedantic} are given
2279 where they are required by the base standard. (It would not make sense
2280 for such warnings to be given only for features not in the specified GNU
2281 C dialect, since by definition the GNU dialects of C include all
2282 features the compiler supports with the given option, and there would be
2283 nothing to warn about.)
2285 @item -pedantic-errors
2286 @opindex pedantic-errors
2287 Like @option{-pedantic}, except that errors are produced rather than
2292 Inhibit all warning messages.
2296 Inhibit warning messages about the use of @samp{#import}.
2298 @item -Wchar-subscripts
2299 @opindex Wchar-subscripts
2300 Warn if an array subscript has type @code{char}. This is a common cause
2301 of error, as programmers often forget that this type is signed on some
2303 This warning is enabled by @option{-Wall}.
2307 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2308 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2309 This warning is enabled by @option{-Wall}.
2311 @item -Wfatal-errors
2312 @opindex Wfatal-errors
2313 This option causes the compiler to abort compilation on the first error
2314 occurred rather than trying to keep going and printing further error
2319 @opindex ffreestanding
2320 @opindex fno-builtin
2321 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2322 the arguments supplied have types appropriate to the format string
2323 specified, and that the conversions specified in the format string make
2324 sense. This includes standard functions, and others specified by format
2325 attributes (@pxref{Function Attributes}), in the @code{printf},
2326 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2327 not in the C standard) families (or other target-specific families).
2328 Which functions are checked without format attributes having been
2329 specified depends on the standard version selected, and such checks of
2330 functions without the attribute specified are disabled by
2331 @option{-ffreestanding} or @option{-fno-builtin}.
2333 The formats are checked against the format features supported by GNU
2334 libc version 2.2. These include all ISO C90 and C99 features, as well
2335 as features from the Single Unix Specification and some BSD and GNU
2336 extensions. Other library implementations may not support all these
2337 features; GCC does not support warning about features that go beyond a
2338 particular library's limitations. However, if @option{-pedantic} is used
2339 with @option{-Wformat}, warnings will be given about format features not
2340 in the selected standard version (but not for @code{strfmon} formats,
2341 since those are not in any version of the C standard). @xref{C Dialect
2342 Options,,Options Controlling C Dialect}.
2344 Since @option{-Wformat} also checks for null format arguments for
2345 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2347 @option{-Wformat} is included in @option{-Wall}. For more control over some
2348 aspects of format checking, the options @option{-Wformat-y2k},
2349 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2350 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2351 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2354 @opindex Wformat-y2k
2355 If @option{-Wformat} is specified, also warn about @code{strftime}
2356 formats which may yield only a two-digit year.
2358 @item -Wno-format-extra-args
2359 @opindex Wno-format-extra-args
2360 If @option{-Wformat} is specified, do not warn about excess arguments to a
2361 @code{printf} or @code{scanf} format function. The C standard specifies
2362 that such arguments are ignored.
2364 Where the unused arguments lie between used arguments that are
2365 specified with @samp{$} operand number specifications, normally
2366 warnings are still given, since the implementation could not know what
2367 type to pass to @code{va_arg} to skip the unused arguments. However,
2368 in the case of @code{scanf} formats, this option will suppress the
2369 warning if the unused arguments are all pointers, since the Single
2370 Unix Specification says that such unused arguments are allowed.
2372 @item -Wno-format-zero-length
2373 @opindex Wno-format-zero-length
2374 If @option{-Wformat} is specified, do not warn about zero-length formats.
2375 The C standard specifies that zero-length formats are allowed.
2377 @item -Wformat-nonliteral
2378 @opindex Wformat-nonliteral
2379 If @option{-Wformat} is specified, also warn if the format string is not a
2380 string literal and so cannot be checked, unless the format function
2381 takes its format arguments as a @code{va_list}.
2383 @item -Wformat-security
2384 @opindex Wformat-security
2385 If @option{-Wformat} is specified, also warn about uses of format
2386 functions that represent possible security problems. At present, this
2387 warns about calls to @code{printf} and @code{scanf} functions where the
2388 format string is not a string literal and there are no format arguments,
2389 as in @code{printf (foo);}. This may be a security hole if the format
2390 string came from untrusted input and contains @samp{%n}. (This is
2391 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2392 in future warnings may be added to @option{-Wformat-security} that are not
2393 included in @option{-Wformat-nonliteral}.)
2397 Enable @option{-Wformat} plus format checks not included in
2398 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2399 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2403 Warn about passing a null pointer for arguments marked as
2404 requiring a non-null value by the @code{nonnull} function attribute.
2406 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2407 can be disabled with the @option{-Wno-nonnull} option.
2409 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2411 Warn about uninitialized variables which are initialized with themselves.
2412 Note this option can only be used with the @option{-Wuninitialized} option,
2413 which in turn only works with @option{-O1} and above.
2415 For example, GCC will warn about @code{i} being uninitialized in the
2416 following snippet only when @option{-Winit-self} has been specified:
2427 @item -Wimplicit-int
2428 @opindex Wimplicit-int
2429 Warn when a declaration does not specify a type.
2430 This warning is enabled by @option{-Wall}.
2432 @item -Wimplicit-function-declaration
2433 @itemx -Werror-implicit-function-declaration
2434 @opindex Wimplicit-function-declaration
2435 @opindex Werror-implicit-function-declaration
2436 Give a warning (or error) whenever a function is used before being
2437 declared. The form @option{-Wno-error-implicit-function-declaration}
2439 This warning is enabled by @option{-Wall} (as a warning, not an error).
2443 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2444 This warning is enabled by @option{-Wall}.
2448 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2449 function with external linkage, returning int, taking either zero
2450 arguments, two, or three arguments of appropriate types.
2451 This warning is enabled by @option{-Wall}.
2453 @item -Wmissing-braces
2454 @opindex Wmissing-braces
2455 Warn if an aggregate or union initializer is not fully bracketed. In
2456 the following example, the initializer for @samp{a} is not fully
2457 bracketed, but that for @samp{b} is fully bracketed.
2460 int a[2][2] = @{ 0, 1, 2, 3 @};
2461 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2464 This warning is enabled by @option{-Wall}.
2466 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2467 @opindex Wmissing-include-dirs
2468 Warn if a user-supplied include directory does not exist.
2471 @opindex Wparentheses
2472 Warn if parentheses are omitted in certain contexts, such
2473 as when there is an assignment in a context where a truth value
2474 is expected, or when operators are nested whose precedence people
2475 often get confused about. Only the warning for an assignment used as
2476 a truth value is supported when compiling C++; the other warnings are
2477 only supported when compiling C@.
2479 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2480 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2481 interpretation from that of ordinary mathematical notation.
2483 Also warn about constructions where there may be confusion to which
2484 @code{if} statement an @code{else} branch belongs. Here is an example of
2499 In C, every @code{else} branch belongs to the innermost possible @code{if}
2500 statement, which in this example is @code{if (b)}. This is often not
2501 what the programmer expected, as illustrated in the above example by
2502 indentation the programmer chose. When there is the potential for this
2503 confusion, GCC will issue a warning when this flag is specified.
2504 To eliminate the warning, add explicit braces around the innermost
2505 @code{if} statement so there is no way the @code{else} could belong to
2506 the enclosing @code{if}. The resulting code would look like this:
2522 This warning is enabled by @option{-Wall}.
2524 @item -Wsequence-point
2525 @opindex Wsequence-point
2526 Warn about code that may have undefined semantics because of violations
2527 of sequence point rules in the C standard.
2529 The C standard defines the order in which expressions in a C program are
2530 evaluated in terms of @dfn{sequence points}, which represent a partial
2531 ordering between the execution of parts of the program: those executed
2532 before the sequence point, and those executed after it. These occur
2533 after the evaluation of a full expression (one which is not part of a
2534 larger expression), after the evaluation of the first operand of a
2535 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2536 function is called (but after the evaluation of its arguments and the
2537 expression denoting the called function), and in certain other places.
2538 Other than as expressed by the sequence point rules, the order of
2539 evaluation of subexpressions of an expression is not specified. All
2540 these rules describe only a partial order rather than a total order,
2541 since, for example, if two functions are called within one expression
2542 with no sequence point between them, the order in which the functions
2543 are called is not specified. However, the standards committee have
2544 ruled that function calls do not overlap.
2546 It is not specified when between sequence points modifications to the
2547 values of objects take effect. Programs whose behavior depends on this
2548 have undefined behavior; the C standard specifies that ``Between the
2549 previous and next sequence point an object shall have its stored value
2550 modified at most once by the evaluation of an expression. Furthermore,
2551 the prior value shall be read only to determine the value to be
2552 stored.''. If a program breaks these rules, the results on any
2553 particular implementation are entirely unpredictable.
2555 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2556 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2557 diagnosed by this option, and it may give an occasional false positive
2558 result, but in general it has been found fairly effective at detecting
2559 this sort of problem in programs.
2561 The present implementation of this option only works for C programs. A
2562 future implementation may also work for C++ programs.
2564 The C standard is worded confusingly, therefore there is some debate
2565 over the precise meaning of the sequence point rules in subtle cases.
2566 Links to discussions of the problem, including proposed formal
2567 definitions, may be found on the GCC readings page, at
2568 @w{@uref{http://gcc.gnu.org/readings.html}}.
2570 This warning is enabled by @option{-Wall}.
2573 @opindex Wreturn-type
2574 Warn whenever a function is defined with a return-type that defaults to
2575 @code{int}. Also warn about any @code{return} statement with no
2576 return-value in a function whose return-type is not @code{void}.
2578 For C, also warn if the return type of a function has a type qualifier
2579 such as @code{const}. Such a type qualifier has no effect, since the
2580 value returned by a function is not an lvalue. ISO C prohibits
2581 qualified @code{void} return types on function definitions, so such
2582 return types always receive a warning even without this option.
2584 For C++, a function without return type always produces a diagnostic
2585 message, even when @option{-Wno-return-type} is specified. The only
2586 exceptions are @samp{main} and functions defined in system headers.
2588 This warning is enabled by @option{-Wall}.
2592 Warn whenever a @code{switch} statement has an index of enumerated type
2593 and lacks a @code{case} for one or more of the named codes of that
2594 enumeration. (The presence of a @code{default} label prevents this
2595 warning.) @code{case} labels outside the enumeration range also
2596 provoke warnings when this option is used.
2597 This warning is enabled by @option{-Wall}.
2599 @item -Wswitch-default
2600 @opindex Wswitch-switch
2601 Warn whenever a @code{switch} statement does not have a @code{default}
2605 @opindex Wswitch-enum
2606 Warn whenever a @code{switch} statement has an index of enumerated type
2607 and lacks a @code{case} for one or more of the named codes of that
2608 enumeration. @code{case} labels outside the enumeration range also
2609 provoke warnings when this option is used.
2613 Warn if any trigraphs are encountered that might change the meaning of
2614 the program (trigraphs within comments are not warned about).
2615 This warning is enabled by @option{-Wall}.
2617 @item -Wunused-function
2618 @opindex Wunused-function
2619 Warn whenever a static function is declared but not defined or a
2620 non-inline static function is unused.
2621 This warning is enabled by @option{-Wall}.
2623 @item -Wunused-label
2624 @opindex Wunused-label
2625 Warn whenever a label is declared but not used.
2626 This warning is enabled by @option{-Wall}.
2628 To suppress this warning use the @samp{unused} attribute
2629 (@pxref{Variable Attributes}).
2631 @item -Wunused-parameter
2632 @opindex Wunused-parameter
2633 Warn whenever a function parameter is unused aside from its declaration.
2635 To suppress this warning use the @samp{unused} attribute
2636 (@pxref{Variable Attributes}).
2638 @item -Wunused-variable
2639 @opindex Wunused-variable
2640 Warn whenever a local variable or non-constant static variable is unused
2641 aside from its declaration
2642 This warning is enabled by @option{-Wall}.
2644 To suppress this warning use the @samp{unused} attribute
2645 (@pxref{Variable Attributes}).
2647 @item -Wunused-value
2648 @opindex Wunused-value
2649 Warn whenever a statement computes a result that is explicitly not used.
2650 This warning is enabled by @option{-Wall}.
2652 To suppress this warning cast the expression to @samp{void}.
2656 All the above @option{-Wunused} options combined.
2658 In order to get a warning about an unused function parameter, you must
2659 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2660 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2662 @item -Wuninitialized
2663 @opindex Wuninitialized
2664 Warn if an automatic variable is used without first being initialized or
2665 if a variable may be clobbered by a @code{setjmp} call.
2667 These warnings are possible only in optimizing compilation,
2668 because they require data flow information that is computed only
2669 when optimizing. If you don't specify @option{-O}, you simply won't
2672 If you want to warn about code which uses the uninitialized value of the
2673 variable in its own initializer, use the @option{-Winit-self} option.
2675 These warnings occur for individual uninitialized or clobbered
2676 elements of structure, union or array variables as well as for
2677 variables which are uninitialized or clobbered as a whole. They do
2678 not occur for variables or elements declared @code{volatile}. Because
2679 these warnings depend on optimization, the exact variables or elements
2680 for which there are warnings will depend on the precise optimization
2681 options and version of GCC used.
2683 Note that there may be no warning about a variable that is used only
2684 to compute a value that itself is never used, because such
2685 computations may be deleted by data flow analysis before the warnings
2688 These warnings are made optional because GCC is not smart
2689 enough to see all the reasons why the code might be correct
2690 despite appearing to have an error. Here is one example of how
2711 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2712 always initialized, but GCC doesn't know this. Here is
2713 another common case:
2718 if (change_y) save_y = y, y = new_y;
2720 if (change_y) y = save_y;
2725 This has no bug because @code{save_y} is used only if it is set.
2727 @cindex @code{longjmp} warnings
2728 This option also warns when a non-volatile automatic variable might be
2729 changed by a call to @code{longjmp}. These warnings as well are possible
2730 only in optimizing compilation.
2732 The compiler sees only the calls to @code{setjmp}. It cannot know
2733 where @code{longjmp} will be called; in fact, a signal handler could
2734 call it at any point in the code. As a result, you may get a warning
2735 even when there is in fact no problem because @code{longjmp} cannot
2736 in fact be called at the place which would cause a problem.
2738 Some spurious warnings can be avoided if you declare all the functions
2739 you use that never return as @code{noreturn}. @xref{Function
2742 This warning is enabled by @option{-Wall}.
2744 @item -Wunknown-pragmas
2745 @opindex Wunknown-pragmas
2746 @cindex warning for unknown pragmas
2747 @cindex unknown pragmas, warning
2748 @cindex pragmas, warning of unknown
2749 Warn when a #pragma directive is encountered which is not understood by
2750 GCC@. If this command line option is used, warnings will even be issued
2751 for unknown pragmas in system header files. This is not the case if
2752 the warnings were only enabled by the @option{-Wall} command line option.
2755 @opindex Wno-pragmas
2757 Do not warn about misuses of pragmas, such as incorrect parameters,
2758 invalid syntax, or conflicts between pragmas. See also
2759 @samp{-Wunknown-pragmas}.
2761 @item -Wstrict-aliasing
2762 @opindex Wstrict-aliasing
2763 This option is only active when @option{-fstrict-aliasing} is active.
2764 It warns about code which might break the strict aliasing rules that the
2765 compiler is using for optimization. The warning does not catch all
2766 cases, but does attempt to catch the more common pitfalls. It is
2767 included in @option{-Wall}.
2769 @item -Wstrict-aliasing=2
2770 @opindex Wstrict-aliasing=2
2771 This option is only active when @option{-fstrict-aliasing} is active.
2772 It warns about code which might break the strict aliasing rules that the
2773 compiler is using for optimization. This warning catches more cases than
2774 @option{-Wstrict-aliasing}, but it will also give a warning for some ambiguous
2775 cases that are safe.
2779 All of the above @samp{-W} options combined. This enables all the
2780 warnings about constructions that some users consider questionable, and
2781 that are easy to avoid (or modify to prevent the warning), even in
2782 conjunction with macros. This also enables some language-specific
2783 warnings described in @ref{C++ Dialect Options} and
2784 @ref{Objective-C and Objective-C++ Dialect Options}.
2787 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2788 Some of them warn about constructions that users generally do not
2789 consider questionable, but which occasionally you might wish to check
2790 for; others warn about constructions that are necessary or hard to avoid
2791 in some cases, and there is no simple way to modify the code to suppress
2798 (This option used to be called @option{-W}. The older name is still
2799 supported, but the newer name is more descriptive.) Print extra warning
2800 messages for these events:
2804 A function can return either with or without a value. (Falling
2805 off the end of the function body is considered returning without
2806 a value.) For example, this function would evoke such a
2820 An expression-statement or the left-hand side of a comma expression
2821 contains no side effects.
2822 To suppress the warning, cast the unused expression to void.
2823 For example, an expression such as @samp{x[i,j]} will cause a warning,
2824 but @samp{x[(void)i,j]} will not.
2827 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2830 Storage-class specifiers like @code{static} are not the first things in
2831 a declaration. According to the C Standard, this usage is obsolescent.
2834 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2838 A comparison between signed and unsigned values could produce an
2839 incorrect result when the signed value is converted to unsigned.
2840 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2843 An aggregate has an initializer which does not initialize all members.
2844 This warning can be independently controlled by
2845 @option{-Wmissing-field-initializers}.
2848 A function parameter is declared without a type specifier in K&R-style
2856 An empty body occurs in an @samp{if} or @samp{else} statement.
2859 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2860 @samp{>}, or @samp{>=}.
2863 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2866 Any of several floating-point events that often indicate errors, such as
2867 overflow, underflow, loss of precision, etc.
2869 @item @r{(C++ only)}
2870 An enumerator and a non-enumerator both appear in a conditional expression.
2872 @item @r{(C++ only)}
2873 A non-static reference or non-static @samp{const} member appears in a
2874 class without constructors.
2876 @item @r{(C++ only)}
2877 Ambiguous virtual bases.
2879 @item @r{(C++ only)}
2880 Subscripting an array which has been declared @samp{register}.
2882 @item @r{(C++ only)}
2883 Taking the address of a variable which has been declared @samp{register}.
2885 @item @r{(C++ only)}
2886 A base class is not initialized in a derived class' copy constructor.
2889 @item -Wno-div-by-zero
2890 @opindex Wno-div-by-zero
2891 @opindex Wdiv-by-zero
2892 Do not warn about compile-time integer division by zero. Floating point
2893 division by zero is not warned about, as it can be a legitimate way of
2894 obtaining infinities and NaNs.
2896 @item -Wsystem-headers
2897 @opindex Wsystem-headers
2898 @cindex warnings from system headers
2899 @cindex system headers, warnings from
2900 Print warning messages for constructs found in system header files.
2901 Warnings from system headers are normally suppressed, on the assumption
2902 that they usually do not indicate real problems and would only make the
2903 compiler output harder to read. Using this command line option tells
2904 GCC to emit warnings from system headers as if they occurred in user
2905 code. However, note that using @option{-Wall} in conjunction with this
2906 option will @emph{not} warn about unknown pragmas in system
2907 headers---for that, @option{-Wunknown-pragmas} must also be used.
2910 @opindex Wfloat-equal
2911 Warn if floating point values are used in equality comparisons.
2913 The idea behind this is that sometimes it is convenient (for the
2914 programmer) to consider floating-point values as approximations to
2915 infinitely precise real numbers. If you are doing this, then you need
2916 to compute (by analyzing the code, or in some other way) the maximum or
2917 likely maximum error that the computation introduces, and allow for it
2918 when performing comparisons (and when producing output, but that's a
2919 different problem). In particular, instead of testing for equality, you
2920 would check to see whether the two values have ranges that overlap; and
2921 this is done with the relational operators, so equality comparisons are
2924 @item -Wtraditional @r{(C only)}
2925 @opindex Wtraditional
2926 Warn about certain constructs that behave differently in traditional and
2927 ISO C@. Also warn about ISO C constructs that have no traditional C
2928 equivalent, and/or problematic constructs which should be avoided.
2932 Macro parameters that appear within string literals in the macro body.
2933 In traditional C macro replacement takes place within string literals,
2934 but does not in ISO C@.
2937 In traditional C, some preprocessor directives did not exist.
2938 Traditional preprocessors would only consider a line to be a directive
2939 if the @samp{#} appeared in column 1 on the line. Therefore
2940 @option{-Wtraditional} warns about directives that traditional C
2941 understands but would ignore because the @samp{#} does not appear as the
2942 first character on the line. It also suggests you hide directives like
2943 @samp{#pragma} not understood by traditional C by indenting them. Some
2944 traditional implementations would not recognize @samp{#elif}, so it
2945 suggests avoiding it altogether.
2948 A function-like macro that appears without arguments.
2951 The unary plus operator.
2954 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2955 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2956 constants.) Note, these suffixes appear in macros defined in the system
2957 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2958 Use of these macros in user code might normally lead to spurious
2959 warnings, however GCC's integrated preprocessor has enough context to
2960 avoid warning in these cases.
2963 A function declared external in one block and then used after the end of
2967 A @code{switch} statement has an operand of type @code{long}.
2970 A non-@code{static} function declaration follows a @code{static} one.
2971 This construct is not accepted by some traditional C compilers.
2974 The ISO type of an integer constant has a different width or
2975 signedness from its traditional type. This warning is only issued if
2976 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2977 typically represent bit patterns, are not warned about.
2980 Usage of ISO string concatenation is detected.
2983 Initialization of automatic aggregates.
2986 Identifier conflicts with labels. Traditional C lacks a separate
2987 namespace for labels.
2990 Initialization of unions. If the initializer is zero, the warning is
2991 omitted. This is done under the assumption that the zero initializer in
2992 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2993 initializer warnings and relies on default initialization to zero in the
2997 Conversions by prototypes between fixed/floating point values and vice
2998 versa. The absence of these prototypes when compiling with traditional
2999 C would cause serious problems. This is a subset of the possible
3000 conversion warnings, for the full set use @option{-Wconversion}.
3003 Use of ISO C style function definitions. This warning intentionally is
3004 @emph{not} issued for prototype declarations or variadic functions
3005 because these ISO C features will appear in your code when using
3006 libiberty's traditional C compatibility macros, @code{PARAMS} and
3007 @code{VPARAMS}. This warning is also bypassed for nested functions
3008 because that feature is already a GCC extension and thus not relevant to
3009 traditional C compatibility.
3012 @item -Wdeclaration-after-statement @r{(C only)}
3013 @opindex Wdeclaration-after-statement
3014 Warn when a declaration is found after a statement in a block. This
3015 construct, known from C++, was introduced with ISO C99 and is by default
3016 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3017 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3021 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3023 @item -Wno-endif-labels
3024 @opindex Wno-endif-labels
3025 @opindex Wendif-labels
3026 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3030 Warn whenever a local variable shadows another local variable, parameter or
3031 global variable or whenever a built-in function is shadowed.
3033 @item -Wlarger-than-@var{len}
3034 @opindex Wlarger-than
3035 Warn whenever an object of larger than @var{len} bytes is defined.
3037 @item -Wunsafe-loop-optimizations
3038 @opindex Wunsafe-loop-optimizations
3039 Warn if the loop cannot be optimized because the compiler could not
3040 assume anything on the bounds of the loop indices. With
3041 @option{-funsafe-loop-optimizations} warn if the compiler made
3044 @item -Wpointer-arith
3045 @opindex Wpointer-arith
3046 Warn about anything that depends on the ``size of'' a function type or
3047 of @code{void}. GNU C assigns these types a size of 1, for
3048 convenience in calculations with @code{void *} pointers and pointers
3051 @item -Wbad-function-cast @r{(C only)}
3052 @opindex Wbad-function-cast
3053 Warn whenever a function call is cast to a non-matching type.
3054 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3057 Warn about ISO C constructs that are outside of the common subset of
3058 ISO C and ISO C++, e.g.@: request for implicit conversion from
3059 @code{void *} to a pointer to non-@code{void} type.
3063 Warn whenever a pointer is cast so as to remove a type qualifier from
3064 the target type. For example, warn if a @code{const char *} is cast
3065 to an ordinary @code{char *}.
3068 @opindex Wcast-align
3069 Warn whenever a pointer is cast such that the required alignment of the
3070 target is increased. For example, warn if a @code{char *} is cast to
3071 an @code{int *} on machines where integers can only be accessed at
3072 two- or four-byte boundaries.
3074 @item -Wwrite-strings
3075 @opindex Wwrite-strings
3076 When compiling C, give string constants the type @code{const
3077 char[@var{length}]} so that
3078 copying the address of one into a non-@code{const} @code{char *}
3079 pointer will get a warning; when compiling C++, warn about the
3080 deprecated conversion from string constants to @code{char *}.
3081 These warnings will help you find at
3082 compile time code that can try to write into a string constant, but
3083 only if you have been very careful about using @code{const} in
3084 declarations and prototypes. Otherwise, it will just be a nuisance;
3085 this is why we did not make @option{-Wall} request these warnings.
3088 @opindex Wconversion
3089 Warn if a prototype causes a type conversion that is different from what
3090 would happen to the same argument in the absence of a prototype. This
3091 includes conversions of fixed point to floating and vice versa, and
3092 conversions changing the width or signedness of a fixed point argument
3093 except when the same as the default promotion.
3095 Also, warn if a negative integer constant expression is implicitly
3096 converted to an unsigned type. For example, warn about the assignment
3097 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
3098 casts like @code{(unsigned) -1}.
3100 @item -Wsign-compare
3101 @opindex Wsign-compare
3102 @cindex warning for comparison of signed and unsigned values
3103 @cindex comparison of signed and unsigned values, warning
3104 @cindex signed and unsigned values, comparison warning
3105 Warn when a comparison between signed and unsigned values could produce
3106 an incorrect result when the signed value is converted to unsigned.
3107 This warning is also enabled by @option{-Wextra}; to get the other warnings
3108 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3110 @item -Waggregate-return
3111 @opindex Waggregate-return
3112 Warn if any functions that return structures or unions are defined or
3113 called. (In languages where you can return an array, this also elicits
3117 @opindex Walways-true
3118 Warn about comparisons which are always true such as testing if
3119 unsigned values are greater than or equal to zero. This warning is
3120 enabled by @option{-Wall}.
3122 @item -Wno-attributes
3123 @opindex Wno-attributes
3124 @opindex Wattributes
3125 Do not warn if an unexpected @code{__attribute__} is used, such as
3126 unrecognized attributes, function attributes applied to variables,
3127 etc. This will not stop errors for incorrect use of supported
3130 @item -Wstrict-prototypes @r{(C only)}
3131 @opindex Wstrict-prototypes
3132 Warn if a function is declared or defined without specifying the
3133 argument types. (An old-style function definition is permitted without
3134 a warning if preceded by a declaration which specifies the argument
3137 @item -Wold-style-definition @r{(C only)}
3138 @opindex Wold-style-definition
3139 Warn if an old-style function definition is used. A warning is given
3140 even if there is a previous prototype.
3142 @item -Wmissing-prototypes @r{(C only)}
3143 @opindex Wmissing-prototypes
3144 Warn if a global function is defined without a previous prototype
3145 declaration. This warning is issued even if the definition itself
3146 provides a prototype. The aim is to detect global functions that fail
3147 to be declared in header files.
3149 @item -Wmissing-declarations @r{(C only)}
3150 @opindex Wmissing-declarations
3151 Warn if a global function is defined without a previous declaration.
3152 Do so even if the definition itself provides a prototype.
3153 Use this option to detect global functions that are not declared in
3156 @item -Wmissing-field-initializers
3157 @opindex Wmissing-field-initializers
3160 Warn if a structure's initializer has some fields missing. For
3161 example, the following code would cause such a warning, because
3162 @code{x.h} is implicitly zero:
3165 struct s @{ int f, g, h; @};
3166 struct s x = @{ 3, 4 @};
3169 This option does not warn about designated initializers, so the following
3170 modification would not trigger a warning:
3173 struct s @{ int f, g, h; @};
3174 struct s x = @{ .f = 3, .g = 4 @};
3177 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3178 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3180 @item -Wmissing-noreturn
3181 @opindex Wmissing-noreturn
3182 Warn about functions which might be candidates for attribute @code{noreturn}.
3183 Note these are only possible candidates, not absolute ones. Care should
3184 be taken to manually verify functions actually do not ever return before
3185 adding the @code{noreturn} attribute, otherwise subtle code generation
3186 bugs could be introduced. You will not get a warning for @code{main} in
3187 hosted C environments.
3189 @item -Wmissing-format-attribute
3190 @opindex Wmissing-format-attribute
3192 Warn about function pointers which might be candidates for @code{format}
3193 attributes. Note these are only possible candidates, not absolute ones.
3194 GCC will guess that function pointers with @code{format} attributes that
3195 are used in assignment, initialization, parameter passing or return
3196 statements should have a corresponding @code{format} attribute in the
3197 resulting type. I.e.@: the left-hand side of the assignment or
3198 initialization, the type of the parameter variable, or the return type
3199 of the containing function respectively should also have a @code{format}
3200 attribute to avoid the warning.
3202 GCC will also warn about function definitions which might be
3203 candidates for @code{format} attributes. Again, these are only
3204 possible candidates. GCC will guess that @code{format} attributes
3205 might be appropriate for any function that calls a function like
3206 @code{vprintf} or @code{vscanf}, but this might not always be the
3207 case, and some functions for which @code{format} attributes are
3208 appropriate may not be detected.
3210 @item -Wno-multichar
3211 @opindex Wno-multichar
3213 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3214 Usually they indicate a typo in the user's code, as they have
3215 implementation-defined values, and should not be used in portable code.
3217 @item -Wnormalized=<none|id|nfc|nfkc>
3218 @opindex Wnormalized
3221 @cindex character set, input normalization
3222 In ISO C and ISO C++, two identifiers are different if they are
3223 different sequences of characters. However, sometimes when characters
3224 outside the basic ASCII character set are used, you can have two
3225 different character sequences that look the same. To avoid confusion,
3226 the ISO 10646 standard sets out some @dfn{normalization rules} which
3227 when applied ensure that two sequences that look the same are turned into
3228 the same sequence. GCC can warn you if you are using identifiers which
3229 have not been normalized; this option controls that warning.
3231 There are four levels of warning that GCC supports. The default is
3232 @option{-Wnormalized=nfc}, which warns about any identifier which is
3233 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3234 recommended form for most uses.
3236 Unfortunately, there are some characters which ISO C and ISO C++ allow
3237 in identifiers that when turned into NFC aren't allowable as
3238 identifiers. That is, there's no way to use these symbols in portable
3239 ISO C or C++ and have all your identifiers in NFC.
3240 @option{-Wnormalized=id} suppresses the warning for these characters.
3241 It is hoped that future versions of the standards involved will correct
3242 this, which is why this option is not the default.
3244 You can switch the warning off for all characters by writing
3245 @option{-Wnormalized=none}. You would only want to do this if you
3246 were using some other normalization scheme (like ``D''), because
3247 otherwise you can easily create bugs that are literally impossible to see.
3249 Some characters in ISO 10646 have distinct meanings but look identical
3250 in some fonts or display methodologies, especially once formatting has
3251 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3252 LETTER N'', will display just like a regular @code{n} which has been
3253 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3254 normalisation scheme to convert all these into a standard form as
3255 well, and GCC will warn if your code is not in NFKC if you use
3256 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3257 about every identifier that contains the letter O because it might be
3258 confused with the digit 0, and so is not the default, but may be
3259 useful as a local coding convention if the programming environment is
3260 unable to be fixed to display these characters distinctly.
3262 @item -Wno-deprecated-declarations
3263 @opindex Wno-deprecated-declarations
3264 Do not warn about uses of functions, variables, and types marked as
3265 deprecated by using the @code{deprecated} attribute.
3266 (@pxref{Function Attributes}, @pxref{Variable Attributes},
3267 @pxref{Type Attributes}.)
3271 Warn if a structure is given the packed attribute, but the packed
3272 attribute has no effect on the layout or size of the structure.
3273 Such structures may be mis-aligned for little benefit. For
3274 instance, in this code, the variable @code{f.x} in @code{struct bar}
3275 will be misaligned even though @code{struct bar} does not itself
3276 have the packed attribute:
3283 @} __attribute__((packed));
3293 Warn if padding is included in a structure, either to align an element
3294 of the structure or to align the whole structure. Sometimes when this
3295 happens it is possible to rearrange the fields of the structure to
3296 reduce the padding and so make the structure smaller.
3298 @item -Wredundant-decls
3299 @opindex Wredundant-decls
3300 Warn if anything is declared more than once in the same scope, even in
3301 cases where multiple declaration is valid and changes nothing.
3303 @item -Wnested-externs @r{(C only)}
3304 @opindex Wnested-externs
3305 Warn if an @code{extern} declaration is encountered within a function.
3307 @item -Wunreachable-code
3308 @opindex Wunreachable-code
3309 Warn if the compiler detects that code will never be executed.
3311 This option is intended to warn when the compiler detects that at
3312 least a whole line of source code will never be executed, because
3313 some condition is never satisfied or because it is after a
3314 procedure that never returns.
3316 It is possible for this option to produce a warning even though there
3317 are circumstances under which part of the affected line can be executed,
3318 so care should be taken when removing apparently-unreachable code.
3320 For instance, when a function is inlined, a warning may mean that the
3321 line is unreachable in only one inlined copy of the function.
3323 This option is not made part of @option{-Wall} because in a debugging
3324 version of a program there is often substantial code which checks
3325 correct functioning of the program and is, hopefully, unreachable
3326 because the program does work. Another common use of unreachable
3327 code is to provide behavior which is selectable at compile-time.
3331 Warn if a function can not be inlined and it was declared as inline.
3332 Even with this option, the compiler will not warn about failures to
3333 inline functions declared in system headers.
3335 The compiler uses a variety of heuristics to determine whether or not
3336 to inline a function. For example, the compiler takes into account
3337 the size of the function being inlined and the amount of inlining
3338 that has already been done in the current function. Therefore,
3339 seemingly insignificant changes in the source program can cause the
3340 warnings produced by @option{-Winline} to appear or disappear.
3342 @item -Wno-invalid-offsetof @r{(C++ only)}
3343 @opindex Wno-invalid-offsetof
3344 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3345 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3346 to a non-POD type is undefined. In existing C++ implementations,
3347 however, @samp{offsetof} typically gives meaningful results even when
3348 applied to certain kinds of non-POD types. (Such as a simple
3349 @samp{struct} that fails to be a POD type only by virtue of having a
3350 constructor.) This flag is for users who are aware that they are
3351 writing nonportable code and who have deliberately chosen to ignore the
3354 The restrictions on @samp{offsetof} may be relaxed in a future version
3355 of the C++ standard.
3357 @item -Wno-int-to-pointer-cast @r{(C only)}
3358 @opindex Wno-int-to-pointer-cast
3359 Suppress warnings from casts to pointer type of an integer of a
3362 @item -Wno-pointer-to-int-cast @r{(C only)}
3363 @opindex Wno-pointer-to-int-cast
3364 Suppress warnings from casts from a pointer to an integer type of a
3368 @opindex Winvalid-pch
3369 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3370 the search path but can't be used.
3374 @opindex Wno-long-long
3375 Warn if @samp{long long} type is used. This is default. To inhibit
3376 the warning messages, use @option{-Wno-long-long}. Flags
3377 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3378 only when @option{-pedantic} flag is used.
3380 @item -Wvariadic-macros
3381 @opindex Wvariadic-macros
3382 @opindex Wno-variadic-macros
3383 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3384 alternate syntax when in pedantic ISO C99 mode. This is default.
3385 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3387 @item -Wvolatile-register-var
3388 @opindex Wvolatile-register-var
3389 @opindex Wno-volatile-register-var
3390 Warn if a register variable is declared volatile. The volatile
3391 modifier does not inhibit all optimizations that may eliminate reads
3392 and/or writes to register variables.
3394 @item -Wdisabled-optimization
3395 @opindex Wdisabled-optimization
3396 Warn if a requested optimization pass is disabled. This warning does
3397 not generally indicate that there is anything wrong with your code; it
3398 merely indicates that GCC's optimizers were unable to handle the code
3399 effectively. Often, the problem is that your code is too big or too
3400 complex; GCC will refuse to optimize programs when the optimization
3401 itself is likely to take inordinate amounts of time.
3403 @item -Wno-pointer-sign
3404 @opindex Wno-pointer-sign
3405 Don't warn for pointer argument passing or assignment with different signedness.
3406 Only useful in the negative form since this warning is enabled by default.
3407 This option is only supported for C and Objective-C@.
3411 Make all warnings into errors.
3413 @item -Wstack-protector
3414 @opindex Wstack-protector
3415 This option is only active when @option{-fstack-protector} is active. It
3416 warns about functions that will not be protected against stack smashing.
3418 @item -Wstring-literal-comparison
3419 @opindex Wstring-literal-comparison
3420 Warn about suspicious comparisons to string literal constants. In C,
3421 direct comparisons against the memory address of a string literal, such
3422 as @code{if (x == "abc")}, typically indicate a programmer error, and
3423 even when intentional, result in unspecified behavior and are not portable.
3424 Usually these warnings alert that the programmer intended to use
3425 @code{strcmp}. This warning is enabled by @option{-Wall}.
3429 @node Debugging Options
3430 @section Options for Debugging Your Program or GCC
3431 @cindex options, debugging
3432 @cindex debugging information options
3434 GCC has various special options that are used for debugging
3435 either your program or GCC:
3440 Produce debugging information in the operating system's native format
3441 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
3444 On most systems that use stabs format, @option{-g} enables use of extra
3445 debugging information that only GDB can use; this extra information
3446 makes debugging work better in GDB but will probably make other debuggers
3448 refuse to read the program. If you want to control for certain whether
3449 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3450 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3452 GCC allows you to use @option{-g} with
3453 @option{-O}. The shortcuts taken by optimized code may occasionally
3454 produce surprising results: some variables you declared may not exist
3455 at all; flow of control may briefly move where you did not expect it;
3456 some statements may not be executed because they compute constant
3457 results or their values were already at hand; some statements may
3458 execute in different places because they were moved out of loops.
3460 Nevertheless it proves possible to debug optimized output. This makes
3461 it reasonable to use the optimizer for programs that might have bugs.
3463 The following options are useful when GCC is generated with the
3464 capability for more than one debugging format.
3468 Produce debugging information for use by GDB@. This means to use the
3469 most expressive format available (DWARF 2, stabs, or the native format
3470 if neither of those are supported), including GDB extensions if at all
3475 Produce debugging information in stabs format (if that is supported),
3476 without GDB extensions. This is the format used by DBX on most BSD
3477 systems. On MIPS, Alpha and System V Release 4 systems this option
3478 produces stabs debugging output which is not understood by DBX or SDB@.
3479 On System V Release 4 systems this option requires the GNU assembler.
3481 @item -feliminate-unused-debug-symbols
3482 @opindex feliminate-unused-debug-symbols
3483 Produce debugging information in stabs format (if that is supported),
3484 for only symbols that are actually used.
3488 Produce debugging information in stabs format (if that is supported),
3489 using GNU extensions understood only by the GNU debugger (GDB)@. The
3490 use of these extensions is likely to make other debuggers crash or
3491 refuse to read the program.
3495 Produce debugging information in COFF format (if that is supported).
3496 This is the format used by SDB on most System V systems prior to
3501 Produce debugging information in XCOFF format (if that is supported).
3502 This is the format used by the DBX debugger on IBM RS/6000 systems.
3506 Produce debugging information in XCOFF format (if that is supported),
3507 using GNU extensions understood only by the GNU debugger (GDB)@. The
3508 use of these extensions is likely to make other debuggers crash or
3509 refuse to read the program, and may cause assemblers other than the GNU
3510 assembler (GAS) to fail with an error.
3514 Produce debugging information in DWARF version 2 format (if that is
3515 supported). This is the format used by DBX on IRIX 6. With this
3516 option, GCC uses features of DWARF version 3 when they are useful;
3517 version 3 is upward compatible with version 2, but may still cause
3518 problems for older debuggers.
3522 Produce debugging information in VMS debug format (if that is
3523 supported). This is the format used by DEBUG on VMS systems.
3526 @itemx -ggdb@var{level}
3527 @itemx -gstabs@var{level}
3528 @itemx -gcoff@var{level}
3529 @itemx -gxcoff@var{level}
3530 @itemx -gvms@var{level}
3531 Request debugging information and also use @var{level} to specify how
3532 much information. The default level is 2.
3534 Level 1 produces minimal information, enough for making backtraces in
3535 parts of the program that you don't plan to debug. This includes
3536 descriptions of functions and external variables, but no information
3537 about local variables and no line numbers.
3539 Level 3 includes extra information, such as all the macro definitions
3540 present in the program. Some debuggers support macro expansion when
3541 you use @option{-g3}.
3543 @option{-gdwarf-2} does not accept a concatenated debug level, because
3544 GCC used to support an option @option{-gdwarf} that meant to generate
3545 debug information in version 1 of the DWARF format (which is very
3546 different from version 2), and it would have been too confusing. That
3547 debug format is long obsolete, but the option cannot be changed now.
3548 Instead use an additional @option{-g@var{level}} option to change the
3549 debug level for DWARF2.
3551 @item -feliminate-dwarf2-dups
3552 @opindex feliminate-dwarf2-dups
3553 Compress DWARF2 debugging information by eliminating duplicated
3554 information about each symbol. This option only makes sense when
3555 generating DWARF2 debugging information with @option{-gdwarf-2}.
3557 @cindex @command{prof}
3560 Generate extra code to write profile information suitable for the
3561 analysis program @command{prof}. You must use this option when compiling
3562 the source files you want data about, and you must also use it when
3565 @cindex @command{gprof}
3568 Generate extra code to write profile information suitable for the
3569 analysis program @command{gprof}. You must use this option when compiling
3570 the source files you want data about, and you must also use it when
3575 Makes the compiler print out each function name as it is compiled, and
3576 print some statistics about each pass when it finishes.
3579 @opindex ftime-report
3580 Makes the compiler print some statistics about the time consumed by each
3581 pass when it finishes.
3584 @opindex fmem-report
3585 Makes the compiler print some statistics about permanent memory
3586 allocation when it finishes.
3588 @item -fprofile-arcs
3589 @opindex fprofile-arcs
3590 Add code so that program flow @dfn{arcs} are instrumented. During
3591 execution the program records how many times each branch and call is
3592 executed and how many times it is taken or returns. When the compiled
3593 program exits it saves this data to a file called
3594 @file{@var{auxname}.gcda} for each source file. The data may be used for
3595 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3596 test coverage analysis (@option{-ftest-coverage}). Each object file's
3597 @var{auxname} is generated from the name of the output file, if
3598 explicitly specified and it is not the final executable, otherwise it is
3599 the basename of the source file. In both cases any suffix is removed
3600 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
3601 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3602 @xref{Cross-profiling}.
3604 @cindex @command{gcov}
3608 This option is used to compile and link code instrumented for coverage
3609 analysis. The option is a synonym for @option{-fprofile-arcs}
3610 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
3611 linking). See the documentation for those options for more details.
3616 Compile the source files with @option{-fprofile-arcs} plus optimization
3617 and code generation options. For test coverage analysis, use the
3618 additional @option{-ftest-coverage} option. You do not need to profile
3619 every source file in a program.
3622 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3623 (the latter implies the former).
3626 Run the program on a representative workload to generate the arc profile
3627 information. This may be repeated any number of times. You can run
3628 concurrent instances of your program, and provided that the file system
3629 supports locking, the data files will be correctly updated. Also
3630 @code{fork} calls are detected and correctly handled (double counting
3634 For profile-directed optimizations, compile the source files again with
3635 the same optimization and code generation options plus
3636 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3637 Control Optimization}).
3640 For test coverage analysis, use @command{gcov} to produce human readable
3641 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3642 @command{gcov} documentation for further information.
3646 With @option{-fprofile-arcs}, for each function of your program GCC
3647 creates a program flow graph, then finds a spanning tree for the graph.
3648 Only arcs that are not on the spanning tree have to be instrumented: the
3649 compiler adds code to count the number of times that these arcs are
3650 executed. When an arc is the only exit or only entrance to a block, the
3651 instrumentation code can be added to the block; otherwise, a new basic
3652 block must be created to hold the instrumentation code.
3655 @item -ftest-coverage
3656 @opindex ftest-coverage
3657 Produce a notes file that the @command{gcov} code-coverage utility
3658 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3659 show program coverage. Each source file's note file is called
3660 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3661 above for a description of @var{auxname} and instructions on how to
3662 generate test coverage data. Coverage data will match the source files
3663 more closely, if you do not optimize.
3665 @item -d@var{letters}
3666 @item -fdump-rtl-@var{pass}
3668 Says to make debugging dumps during compilation at times specified by
3669 @var{letters}. This is used for debugging the RTL-based passes of the
3670 compiler. The file names for most of the dumps are made by appending a
3671 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
3672 from the name of the output file, if explicitly specified and it is not
3673 an executable, otherwise it is the basename of the source file.
3675 Most debug dumps can be enabled either passing a letter to the @option{-d}
3676 option, or with a long @option{-fdump-rtl} switch; here are the possible
3677 letters for use in @var{letters} and @var{pass}, and their meanings:
3682 Annotate the assembler output with miscellaneous debugging information.
3685 @itemx -fdump-rtl-bp
3687 @opindex fdump-rtl-bp
3688 Dump after computing branch probabilities, to @file{@var{file}.09.bp}.
3691 @itemx -fdump-rtl-bbro
3693 @opindex fdump-rtl-bbro
3694 Dump after block reordering, to @file{@var{file}.30.bbro}.
3697 @itemx -fdump-rtl-combine
3699 @opindex fdump-rtl-combine
3700 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
3703 @itemx -fdump-rtl-ce1
3704 @itemx -fdump-rtl-ce2
3706 @opindex fdump-rtl-ce1
3707 @opindex fdump-rtl-ce2
3708 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
3709 first if conversion, to the file @file{@var{file}.11.ce1}. @option{-dC}
3710 and @option{-fdump-rtl-ce2} enable dumping after the second if
3711 conversion, to the file @file{@var{file}.18.ce2}.
3714 @itemx -fdump-rtl-btl
3715 @itemx -fdump-rtl-dbr
3717 @opindex fdump-rtl-btl
3718 @opindex fdump-rtl-dbr
3719 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
3720 target load optimization, to @file{@var{file}.31.btl}. @option{-dd}
3721 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
3722 scheduling, to @file{@var{file}.36.dbr}.
3726 Dump all macro definitions, at the end of preprocessing, in addition to
3730 @itemx -fdump-rtl-ce3
3732 @opindex fdump-rtl-ce3
3733 Dump after the third if conversion, to @file{@var{file}.28.ce3}.
3736 @itemx -fdump-rtl-cfg
3737 @itemx -fdump-rtl-life
3739 @opindex fdump-rtl-cfg
3740 @opindex fdump-rtl-life
3741 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
3742 and data flow analysis, to @file{@var{file}.08.cfg}. @option{-df}
3743 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
3744 to @file{@var{file}.16.life}.
3747 @itemx -fdump-rtl-greg
3749 @opindex fdump-rtl-greg
3750 Dump after global register allocation, to @file{@var{file}.23.greg}.
3753 @itemx -fdump-rtl-gcse
3754 @itemx -fdump-rtl-bypass
3756 @opindex fdump-rtl-gcse
3757 @opindex fdump-rtl-bypass
3758 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
3759 @file{@var{file}.05.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
3760 enable dumping after jump bypassing and control flow optimizations, to
3761 @file{@var{file}.07.bypass}.
3764 @itemx -fdump-rtl-eh
3766 @opindex fdump-rtl-eh
3767 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
3770 @itemx -fdump-rtl-sibling
3772 @opindex fdump-rtl-sibling
3773 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
3776 @itemx -fdump-rtl-jump
3778 @opindex fdump-rtl-jump
3779 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
3782 @itemx -fdump-rtl-stack
3784 @opindex fdump-rtl-stack
3785 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
3788 @itemx -fdump-rtl-lreg
3790 @opindex fdump-rtl-lreg
3791 Dump after local register allocation, to @file{@var{file}.22.lreg}.
3794 @itemx -fdump-rtl-loop
3795 @itemx -fdump-rtl-loop2
3797 @opindex fdump-rtl-loop
3798 @opindex fdump-rtl-loop2
3799 @option{-dL} and @option{-fdump-rtl-loop} enable dumping after the first
3800 loop optimization pass, to @file{@var{file}.06.loop}. @option{-dL} and
3801 @option{-fdump-rtl-loop2} enable dumping after the second pass, to
3802 @file{@var{file}.13.loop2}.
3805 @itemx -fdump-rtl-sms
3807 @opindex fdump-rtl-sms
3808 Dump after modulo scheduling, to @file{@var{file}.20.sms}.
3811 @itemx -fdump-rtl-mach
3813 @opindex fdump-rtl-mach
3814 Dump after performing the machine dependent reorganization pass, to
3815 @file{@var{file}.35.mach}.
3818 @itemx -fdump-rtl-rnreg
3820 @opindex fdump-rtl-rnreg
3821 Dump after register renumbering, to @file{@var{file}.29.rnreg}.
3824 @itemx -fdump-rtl-regmove
3826 @opindex fdump-rtl-regmove
3827 Dump after the register move pass, to @file{@var{file}.19.regmove}.
3830 @itemx -fdump-rtl-postreload
3832 @opindex fdump-rtl-postreload
3833 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
3836 @itemx -fdump-rtl-expand
3838 @opindex fdump-rtl-expand
3839 Dump after RTL generation, to @file{@var{file}.00.expand}.
3842 @itemx -fdump-rtl-sched2
3844 @opindex fdump-rtl-sched2
3845 Dump after the second scheduling pass, to @file{@var{file}.32.sched2}.
3848 @itemx -fdump-rtl-cse
3850 @opindex fdump-rtl-cse
3851 Dump after CSE (including the jump optimization that sometimes follows
3852 CSE), to @file{@var{file}.04.cse}.
3855 @itemx -fdump-rtl-sched
3857 @opindex fdump-rtl-sched
3858 Dump after the first scheduling pass, to @file{@var{file}.21.sched}.
3861 @itemx -fdump-rtl-cse2
3863 @opindex fdump-rtl-cse2
3864 Dump after the second CSE pass (including the jump optimization that
3865 sometimes follows CSE), to @file{@var{file}.15.cse2}.
3868 @itemx -fdump-rtl-tracer
3870 @opindex fdump-rtl-tracer
3871 Dump after running tracer, to @file{@var{file}.12.tracer}.
3874 @itemx -fdump-rtl-vpt
3875 @itemx -fdump-rtl-vartrack
3877 @opindex fdump-rtl-vpt
3878 @opindex fdump-rtl-vartrack
3879 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
3880 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
3881 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
3882 to @file{@var{file}.34.vartrack}.
3885 @itemx -fdump-rtl-flow2
3887 @opindex fdump-rtl-flow2
3888 Dump after the second flow pass, to @file{@var{file}.26.flow2}.
3891 @itemx -fdump-rtl-peephole2
3893 @opindex fdump-rtl-peephole2
3894 Dump after the peephole pass, to @file{@var{file}.27.peephole2}.
3897 @itemx -fdump-rtl-web
3899 @opindex fdump-rtl-web
3900 Dump after live range splitting, to @file{@var{file}.14.web}.
3903 @itemx -fdump-rtl-all
3905 @opindex fdump-rtl-all
3906 Produce all the dumps listed above.
3910 Produce a core dump whenever an error occurs.
3914 Print statistics on memory usage, at the end of the run, to
3919 Annotate the assembler output with a comment indicating which
3920 pattern and alternative was used. The length of each instruction is
3925 Dump the RTL in the assembler output as a comment before each instruction.
3926 Also turns on @option{-dp} annotation.
3930 For each of the other indicated dump files (either with @option{-d} or
3931 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
3932 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3936 Just generate RTL for a function instead of compiling it. Usually used
3937 with @samp{r} (@option{-fdump-rtl-expand}).
3941 Dump debugging information during parsing, to standard error.
3944 @item -fdump-unnumbered
3945 @opindex fdump-unnumbered
3946 When doing debugging dumps (see @option{-d} option above), suppress instruction
3947 numbers and line number note output. This makes it more feasible to
3948 use diff on debugging dumps for compiler invocations with different
3949 options, in particular with and without @option{-g}.
3951 @item -fdump-translation-unit @r{(C++ only)}
3952 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
3953 @opindex fdump-translation-unit
3954 Dump a representation of the tree structure for the entire translation
3955 unit to a file. The file name is made by appending @file{.tu} to the
3956 source file name. If the @samp{-@var{options}} form is used, @var{options}
3957 controls the details of the dump as described for the
3958 @option{-fdump-tree} options.
3960 @item -fdump-class-hierarchy @r{(C++ only)}
3961 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3962 @opindex fdump-class-hierarchy
3963 Dump a representation of each class's hierarchy and virtual function
3964 table layout to a file. The file name is made by appending @file{.class}
3965 to the source file name. If the @samp{-@var{options}} form is used,
3966 @var{options} controls the details of the dump as described for the
3967 @option{-fdump-tree} options.
3969 @item -fdump-ipa-@var{switch}
3971 Control the dumping at various stages of inter-procedural analysis
3972 language tree to a file. The file name is generated by appending a switch
3973 specific suffix to the source file name. The following dumps are possible:
3977 Enables all inter-procedural analysis dumps; currently the only produced
3978 dump is the @samp{cgraph} dump.
3981 Dumps information about call-graph optimization, unused function removal,
3982 and inlining decisions.
3985 @item -fdump-tree-@var{switch}
3986 @itemx -fdump-tree-@var{switch}-@var{options}
3988 Control the dumping at various stages of processing the intermediate
3989 language tree to a file. The file name is generated by appending a switch
3990 specific suffix to the source file name. If the @samp{-@var{options}}
3991 form is used, @var{options} is a list of @samp{-} separated options that
3992 control the details of the dump. Not all options are applicable to all
3993 dumps, those which are not meaningful will be ignored. The following
3994 options are available
3998 Print the address of each node. Usually this is not meaningful as it
3999 changes according to the environment and source file. Its primary use
4000 is for tying up a dump file with a debug environment.
4002 Inhibit dumping of members of a scope or body of a function merely
4003 because that scope has been reached. Only dump such items when they
4004 are directly reachable by some other path. When dumping pretty-printed
4005 trees, this option inhibits dumping the bodies of control structures.
4007 Print a raw representation of the tree. By default, trees are
4008 pretty-printed into a C-like representation.
4010 Enable more detailed dumps (not honored by every dump option).
4012 Enable dumping various statistics about the pass (not honored by every dump
4015 Enable showing basic block boundaries (disabled in raw dumps).
4017 Enable showing virtual operands for every statement.
4019 Enable showing line numbers for statements.
4021 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4023 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4026 The following tree dumps are possible:
4030 Dump before any tree based optimization, to @file{@var{file}.original}.
4033 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4036 Dump after function inlining, to @file{@var{file}.inlined}.
4039 @opindex fdump-tree-gimple
4040 Dump each function before and after the gimplification pass to a file. The
4041 file name is made by appending @file{.gimple} to the source file name.
4044 @opindex fdump-tree-cfg
4045 Dump the control flow graph of each function to a file. The file name is
4046 made by appending @file{.cfg} to the source file name.
4049 @opindex fdump-tree-vcg
4050 Dump the control flow graph of each function to a file in VCG format. The
4051 file name is made by appending @file{.vcg} to the source file name. Note
4052 that if the file contains more than one function, the generated file cannot
4053 be used directly by VCG@. You will need to cut and paste each function's
4054 graph into its own separate file first.
4057 @opindex fdump-tree-ch
4058 Dump each function after copying loop headers. The file name is made by
4059 appending @file{.ch} to the source file name.
4062 @opindex fdump-tree-ssa
4063 Dump SSA related information to a file. The file name is made by appending
4064 @file{.ssa} to the source file name.
4067 @opindex fdump-tree-salias
4068 Dump structure aliasing variable information to a file. This file name
4069 is made by appending @file{.salias} to the source file name.
4072 @opindex fdump-tree-alias
4073 Dump aliasing information for each function. The file name is made by
4074 appending @file{.alias} to the source file name.
4077 @opindex fdump-tree-ccp
4078 Dump each function after CCP@. The file name is made by appending
4079 @file{.ccp} to the source file name.
4082 @opindex fdump-tree-storeccp
4083 Dump each function after STORE-CCP. The file name is made by appending
4084 @file{.storeccp} to the source file name.
4087 @opindex fdump-tree-pre
4088 Dump trees after partial redundancy elimination. The file name is made
4089 by appending @file{.pre} to the source file name.
4092 @opindex fdump-tree-fre
4093 Dump trees after full redundancy elimination. The file name is made
4094 by appending @file{.fre} to the source file name.
4097 @opindex fdump-tree-copyprop
4098 Dump trees after copy propagation. The file name is made
4099 by appending @file{.copyprop} to the source file name.
4101 @item store_copyprop
4102 @opindex fdump-tree-store_copyprop
4103 Dump trees after store copy-propagation. The file name is made
4104 by appending @file{.store_copyprop} to the source file name.
4107 @opindex fdump-tree-dce
4108 Dump each function after dead code elimination. The file name is made by
4109 appending @file{.dce} to the source file name.
4112 @opindex fdump-tree-mudflap
4113 Dump each function after adding mudflap instrumentation. The file name is
4114 made by appending @file{.mudflap} to the source file name.
4117 @opindex fdump-tree-sra
4118 Dump each function after performing scalar replacement of aggregates. The
4119 file name is made by appending @file{.sra} to the source file name.
4122 @opindex fdump-tree-sink
4123 Dump each function after performing code sinking. The file name is made
4124 by appending @file{.sink} to the source file name.
4127 @opindex fdump-tree-dom
4128 Dump each function after applying dominator tree optimizations. The file
4129 name is made by appending @file{.dom} to the source file name.
4132 @opindex fdump-tree-dse
4133 Dump each function after applying dead store elimination. The file
4134 name is made by appending @file{.dse} to the source file name.
4137 @opindex fdump-tree-phiopt
4138 Dump each function after optimizing PHI nodes into straightline code. The file
4139 name is made by appending @file{.phiopt} to the source file name.
4142 @opindex fdump-tree-forwprop
4143 Dump each function after forward propagating single use variables. The file
4144 name is made by appending @file{.forwprop} to the source file name.
4147 @opindex fdump-tree-copyrename
4148 Dump each function after applying the copy rename optimization. The file
4149 name is made by appending @file{.copyrename} to the source file name.
4152 @opindex fdump-tree-nrv
4153 Dump each function after applying the named return value optimization on
4154 generic trees. The file name is made by appending @file{.nrv} to the source
4158 @opindex fdump-tree-vect
4159 Dump each function after applying vectorization of loops. The file name is
4160 made by appending @file{.vect} to the source file name.
4163 @opindex fdump-tree-vrp
4164 Dump each function after Value Range Propagation (VRP). The file name
4165 is made by appending @file{.vrp} to the source file name.
4168 @opindex fdump-tree-all
4169 Enable all the available tree dumps with the flags provided in this option.
4172 @item -ftree-vectorizer-verbose=@var{n}
4173 @opindex ftree-vectorizer-verbose
4174 This option controls the amount of debugging output the vectorizer prints.
4175 This information is written to standard error, unless @option{-fdump-tree-all}
4176 or @option{-fdump-tree-vect} is specified, in which case it is output to the
4177 usual dump listing file, @file{.vect}.
4179 @item -frandom-seed=@var{string}
4180 @opindex frandom-string
4181 This option provides a seed that GCC uses when it would otherwise use
4182 random numbers. It is used to generate certain symbol names
4183 that have to be different in every compiled file. It is also used to
4184 place unique stamps in coverage data files and the object files that
4185 produce them. You can use the @option{-frandom-seed} option to produce
4186 reproducibly identical object files.
4188 The @var{string} should be different for every file you compile.
4190 @item -fsched-verbose=@var{n}
4191 @opindex fsched-verbose
4192 On targets that use instruction scheduling, this option controls the
4193 amount of debugging output the scheduler prints. This information is
4194 written to standard error, unless @option{-dS} or @option{-dR} is
4195 specified, in which case it is output to the usual dump
4196 listing file, @file{.sched} or @file{.sched2} respectively. However
4197 for @var{n} greater than nine, the output is always printed to standard
4200 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4201 same information as @option{-dRS}. For @var{n} greater than one, it
4202 also output basic block probabilities, detailed ready list information
4203 and unit/insn info. For @var{n} greater than two, it includes RTL
4204 at abort point, control-flow and regions info. And for @var{n} over
4205 four, @option{-fsched-verbose} also includes dependence info.
4209 Store the usual ``temporary'' intermediate files permanently; place them
4210 in the current directory and name them based on the source file. Thus,
4211 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4212 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4213 preprocessed @file{foo.i} output file even though the compiler now
4214 normally uses an integrated preprocessor.
4216 When used in combination with the @option{-x} command line option,
4217 @option{-save-temps} is sensible enough to avoid over writing an
4218 input source file with the same extension as an intermediate file.
4219 The corresponding intermediate file may be obtained by renaming the
4220 source file before using @option{-save-temps}.
4224 Report the CPU time taken by each subprocess in the compilation
4225 sequence. For C source files, this is the compiler proper and assembler
4226 (plus the linker if linking is done). The output looks like this:
4233 The first number on each line is the ``user time'', that is time spent
4234 executing the program itself. The second number is ``system time'',
4235 time spent executing operating system routines on behalf of the program.
4236 Both numbers are in seconds.
4238 @item -fvar-tracking
4239 @opindex fvar-tracking
4240 Run variable tracking pass. It computes where variables are stored at each
4241 position in code. Better debugging information is then generated
4242 (if the debugging information format supports this information).
4244 It is enabled by default when compiling with optimization (@option{-Os},
4245 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
4246 the debug info format supports it.
4248 @item -print-file-name=@var{library}
4249 @opindex print-file-name
4250 Print the full absolute name of the library file @var{library} that
4251 would be used when linking---and don't do anything else. With this
4252 option, GCC does not compile or link anything; it just prints the
4255 @item -print-multi-directory
4256 @opindex print-multi-directory
4257 Print the directory name corresponding to the multilib selected by any
4258 other switches present in the command line. This directory is supposed
4259 to exist in @env{GCC_EXEC_PREFIX}.
4261 @item -print-multi-lib
4262 @opindex print-multi-lib
4263 Print the mapping from multilib directory names to compiler switches
4264 that enable them. The directory name is separated from the switches by
4265 @samp{;}, and each switch starts with an @samp{@@} instead of the
4266 @samp{-}, without spaces between multiple switches. This is supposed to
4267 ease shell-processing.
4269 @item -print-prog-name=@var{program}
4270 @opindex print-prog-name
4271 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
4273 @item -print-libgcc-file-name
4274 @opindex print-libgcc-file-name
4275 Same as @option{-print-file-name=libgcc.a}.
4277 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
4278 but you do want to link with @file{libgcc.a}. You can do
4281 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
4284 @item -print-search-dirs
4285 @opindex print-search-dirs
4286 Print the name of the configured installation directory and a list of
4287 program and library directories @command{gcc} will search---and don't do anything else.
4289 This is useful when @command{gcc} prints the error message
4290 @samp{installation problem, cannot exec cpp0: No such file or directory}.
4291 To resolve this you either need to put @file{cpp0} and the other compiler
4292 components where @command{gcc} expects to find them, or you can set the environment
4293 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
4294 Don't forget the trailing @samp{/}.
4295 @xref{Environment Variables}.
4298 @opindex dumpmachine
4299 Print the compiler's target machine (for example,
4300 @samp{i686-pc-linux-gnu})---and don't do anything else.
4303 @opindex dumpversion
4304 Print the compiler version (for example, @samp{3.0})---and don't do
4309 Print the compiler's built-in specs---and don't do anything else. (This
4310 is used when GCC itself is being built.) @xref{Spec Files}.
4312 @item -feliminate-unused-debug-types
4313 @opindex feliminate-unused-debug-types
4314 Normally, when producing DWARF2 output, GCC will emit debugging
4315 information for all types declared in a compilation
4316 unit, regardless of whether or not they are actually used
4317 in that compilation unit. Sometimes this is useful, such as
4318 if, in the debugger, you want to cast a value to a type that is
4319 not actually used in your program (but is declared). More often,
4320 however, this results in a significant amount of wasted space.
4321 With this option, GCC will avoid producing debug symbol output
4322 for types that are nowhere used in the source file being compiled.
4325 @node Optimize Options
4326 @section Options That Control Optimization
4327 @cindex optimize options
4328 @cindex options, optimization
4330 These options control various sorts of optimizations.
4332 Without any optimization option, the compiler's goal is to reduce the
4333 cost of compilation and to make debugging produce the expected
4334 results. Statements are independent: if you stop the program with a
4335 breakpoint between statements, you can then assign a new value to any
4336 variable or change the program counter to any other statement in the
4337 function and get exactly the results you would expect from the source
4340 Turning on optimization flags makes the compiler attempt to improve
4341 the performance and/or code size at the expense of compilation time
4342 and possibly the ability to debug the program.
4344 The compiler performs optimization based on the knowledge it has of
4345 the program. Optimization levels @option{-O2} and above, in
4346 particular, enable @emph{unit-at-a-time} mode, which allows the
4347 compiler to consider information gained from later functions in
4348 the file when compiling a function. Compiling multiple files at
4349 once to a single output file in @emph{unit-at-a-time} mode allows
4350 the compiler to use information gained from all of the files when
4351 compiling each of them.
4353 Not all optimizations are controlled directly by a flag. Only
4354 optimizations that have a flag are listed.
4361 Optimize. Optimizing compilation takes somewhat more time, and a lot
4362 more memory for a large function.
4364 With @option{-O}, the compiler tries to reduce code size and execution
4365 time, without performing any optimizations that take a great deal of
4368 @option{-O} turns on the following optimization flags:
4369 @gccoptlist{-fdefer-pop @gol
4370 -fdelayed-branch @gol
4371 -fguess-branch-probability @gol
4372 -fcprop-registers @gol
4373 -floop-optimize @gol
4374 -fif-conversion @gol
4375 -fif-conversion2 @gol
4378 -ftree-dominator-opts @gol
4383 -ftree-copyrename @gol
4388 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
4389 where doing so does not interfere with debugging.
4393 Optimize even more. GCC performs nearly all supported optimizations
4394 that do not involve a space-speed tradeoff. The compiler does not
4395 perform loop unrolling or function inlining when you specify @option{-O2}.
4396 As compared to @option{-O}, this option increases both compilation time
4397 and the performance of the generated code.
4399 @option{-O2} turns on all optimization flags specified by @option{-O}. It
4400 also turns on the following optimization flags:
4401 @gccoptlist{-fthread-jumps @gol
4403 -foptimize-sibling-calls @gol
4404 -fcse-follow-jumps -fcse-skip-blocks @gol
4405 -fgcse -fgcse-lm @gol
4406 -fexpensive-optimizations @gol
4407 -fstrength-reduce @gol
4408 -frerun-cse-after-loop -frerun-loop-opt @gol
4411 -fschedule-insns -fschedule-insns2 @gol
4412 -fsched-interblock -fsched-spec @gol
4414 -fstrict-aliasing @gol
4415 -fdelete-null-pointer-checks @gol
4416 -freorder-blocks -freorder-functions @gol
4417 -funit-at-a-time @gol
4418 -falign-functions -falign-jumps @gol
4419 -falign-loops -falign-labels @gol
4423 Please note the warning under @option{-fgcse} about
4424 invoking @option{-O2} on programs that use computed gotos.
4428 Optimize yet more. @option{-O3} turns on all optimizations specified by
4429 @option{-O2} and also turns on the @option{-finline-functions},
4430 @option{-funswitch-loops} and @option{-fgcse-after-reload} options.
4434 Do not optimize. This is the default.
4438 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
4439 do not typically increase code size. It also performs further
4440 optimizations designed to reduce code size.
4442 @option{-Os} disables the following optimization flags:
4443 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
4444 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
4445 -fprefetch-loop-arrays -ftree-vect-loop-version}
4447 If you use multiple @option{-O} options, with or without level numbers,
4448 the last such option is the one that is effective.
4451 Options of the form @option{-f@var{flag}} specify machine-independent
4452 flags. Most flags have both positive and negative forms; the negative
4453 form of @option{-ffoo} would be @option{-fno-foo}. In the table
4454 below, only one of the forms is listed---the one you typically will
4455 use. You can figure out the other form by either removing @samp{no-}
4458 The following options control specific optimizations. They are either
4459 activated by @option{-O} options or are related to ones that are. You
4460 can use the following flags in the rare cases when ``fine-tuning'' of
4461 optimizations to be performed is desired.
4464 @item -fno-default-inline
4465 @opindex fno-default-inline
4466 Do not make member functions inline by default merely because they are
4467 defined inside the class scope (C++ only). Otherwise, when you specify
4468 @w{@option{-O}}, member functions defined inside class scope are compiled
4469 inline by default; i.e., you don't need to add @samp{inline} in front of
4470 the member function name.
4472 @item -fno-defer-pop
4473 @opindex fno-defer-pop
4474 Always pop the arguments to each function call as soon as that function
4475 returns. For machines which must pop arguments after a function call,
4476 the compiler normally lets arguments accumulate on the stack for several
4477 function calls and pops them all at once.
4479 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4483 Force memory operands to be copied into registers before doing
4484 arithmetic on them. This produces better code by making all memory
4485 references potential common subexpressions. When they are not common
4486 subexpressions, instruction combination should eliminate the separate
4487 register-load. This option is now a nop and will be removed in 4.2.
4490 @opindex fforce-addr
4491 Force memory address constants to be copied into registers before
4492 doing arithmetic on them.
4494 @item -fomit-frame-pointer
4495 @opindex fomit-frame-pointer
4496 Don't keep the frame pointer in a register for functions that
4497 don't need one. This avoids the instructions to save, set up and
4498 restore frame pointers; it also makes an extra register available
4499 in many functions. @strong{It also makes debugging impossible on
4502 On some machines, such as the VAX, this flag has no effect, because
4503 the standard calling sequence automatically handles the frame pointer
4504 and nothing is saved by pretending it doesn't exist. The
4505 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
4506 whether a target machine supports this flag. @xref{Registers,,Register
4507 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
4509 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4511 @item -foptimize-sibling-calls
4512 @opindex foptimize-sibling-calls
4513 Optimize sibling and tail recursive calls.
4515 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4519 Don't pay attention to the @code{inline} keyword. Normally this option
4520 is used to keep the compiler from expanding any functions inline.
4521 Note that if you are not optimizing, no functions can be expanded inline.
4523 @item -finline-functions
4524 @opindex finline-functions
4525 Integrate all simple functions into their callers. The compiler
4526 heuristically decides which functions are simple enough to be worth
4527 integrating in this way.
4529 If all calls to a given function are integrated, and the function is
4530 declared @code{static}, then the function is normally not output as
4531 assembler code in its own right.
4533 Enabled at level @option{-O3}.
4535 @item -finline-functions-called-once
4536 @opindex finline-functions-called-once
4537 Consider all @code{static} functions called once for inlining into their
4538 caller even if they are not marked @code{inline}. If a call to a given
4539 function is integrated, then the function is not output as assembler code
4542 Enabled if @option{-funit-at-a-time} is enabled.
4544 @item -fearly-inlining
4545 @opindex fearly-inlining
4546 Inline functions marked by @code{always_inline} and functions whose body seems
4547 smaller than the function call overhead early before doing
4548 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
4549 makes profiling significantly cheaper and usually inlining faster on programs
4550 having large chains of nested wrapper functions.
4554 @item -finline-limit=@var{n}
4555 @opindex finline-limit
4556 By default, GCC limits the size of functions that can be inlined. This flag
4557 allows the control of this limit for functions that are explicitly marked as
4558 inline (i.e., marked with the inline keyword or defined within the class
4559 definition in c++). @var{n} is the size of functions that can be inlined in
4560 number of pseudo instructions (not counting parameter handling). The default
4561 value of @var{n} is 600.
4562 Increasing this value can result in more inlined code at
4563 the cost of compilation time and memory consumption. Decreasing usually makes
4564 the compilation faster and less code will be inlined (which presumably
4565 means slower programs). This option is particularly useful for programs that
4566 use inlining heavily such as those based on recursive templates with C++.
4568 Inlining is actually controlled by a number of parameters, which may be
4569 specified individually by using @option{--param @var{name}=@var{value}}.
4570 The @option{-finline-limit=@var{n}} option sets some of these parameters
4574 @item max-inline-insns-single
4575 is set to @var{n}/2.
4576 @item max-inline-insns-auto
4577 is set to @var{n}/2.
4578 @item min-inline-insns
4579 is set to 130 or @var{n}/4, whichever is smaller.
4580 @item max-inline-insns-rtl
4584 See below for a documentation of the individual
4585 parameters controlling inlining.
4587 @emph{Note:} pseudo instruction represents, in this particular context, an
4588 abstract measurement of function's size. In no way does it represent a count
4589 of assembly instructions and as such its exact meaning might change from one
4590 release to an another.
4592 @item -fkeep-inline-functions
4593 @opindex fkeep-inline-functions
4594 In C, emit @code{static} functions that are declared @code{inline}
4595 into the object file, even if the function has been inlined into all
4596 of its callers. This switch does not affect functions using the
4597 @code{extern inline} extension in GNU C@. In C++, emit any and all
4598 inline functions into the object file.
4600 @item -fkeep-static-consts
4601 @opindex fkeep-static-consts
4602 Emit variables declared @code{static const} when optimization isn't turned
4603 on, even if the variables aren't referenced.
4605 GCC enables this option by default. If you want to force the compiler to
4606 check if the variable was referenced, regardless of whether or not
4607 optimization is turned on, use the @option{-fno-keep-static-consts} option.
4609 @item -fmerge-constants
4610 Attempt to merge identical constants (string constants and floating point
4611 constants) across compilation units.
4613 This option is the default for optimized compilation if the assembler and
4614 linker support it. Use @option{-fno-merge-constants} to inhibit this
4617 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4619 @item -fmerge-all-constants
4620 Attempt to merge identical constants and identical variables.
4622 This option implies @option{-fmerge-constants}. In addition to
4623 @option{-fmerge-constants} this considers e.g.@: even constant initialized
4624 arrays or initialized constant variables with integral or floating point
4625 types. Languages like C or C++ require each non-automatic variable to
4626 have distinct location, so using this option will result in non-conforming
4629 @item -fmodulo-sched
4630 @opindex fmodulo-sched
4631 Perform swing modulo scheduling immediately before the first scheduling
4632 pass. This pass looks at innermost loops and reorders their
4633 instructions by overlapping different iterations.
4635 @item -fno-branch-count-reg
4636 @opindex fno-branch-count-reg
4637 Do not use ``decrement and branch'' instructions on a count register,
4638 but instead generate a sequence of instructions that decrement a
4639 register, compare it against zero, then branch based upon the result.
4640 This option is only meaningful on architectures that support such
4641 instructions, which include x86, PowerPC, IA-64 and S/390.
4643 The default is @option{-fbranch-count-reg}, enabled when
4644 @option{-fstrength-reduce} is enabled.
4646 @item -fno-function-cse
4647 @opindex fno-function-cse
4648 Do not put function addresses in registers; make each instruction that
4649 calls a constant function contain the function's address explicitly.
4651 This option results in less efficient code, but some strange hacks
4652 that alter the assembler output may be confused by the optimizations
4653 performed when this option is not used.
4655 The default is @option{-ffunction-cse}
4657 @item -fno-zero-initialized-in-bss
4658 @opindex fno-zero-initialized-in-bss
4659 If the target supports a BSS section, GCC by default puts variables that
4660 are initialized to zero into BSS@. This can save space in the resulting
4663 This option turns off this behavior because some programs explicitly
4664 rely on variables going to the data section. E.g., so that the
4665 resulting executable can find the beginning of that section and/or make
4666 assumptions based on that.
4668 The default is @option{-fzero-initialized-in-bss}.
4670 @item -fbounds-check
4671 @opindex fbounds-check
4672 For front-ends that support it, generate additional code to check that
4673 indices used to access arrays are within the declared range. This is
4674 currently only supported by the Java and Fortran front-ends, where
4675 this option defaults to true and false respectively.
4677 @item -fmudflap -fmudflapth -fmudflapir
4681 @cindex bounds checking
4683 For front-ends that support it (C and C++), instrument all risky
4684 pointer/array dereferencing operations, some standard library
4685 string/heap functions, and some other associated constructs with
4686 range/validity tests. Modules so instrumented should be immune to
4687 buffer overflows, invalid heap use, and some other classes of C/C++
4688 programming errors. The instrumentation relies on a separate runtime
4689 library (@file{libmudflap}), which will be linked into a program if
4690 @option{-fmudflap} is given at link time. Run-time behavior of the
4691 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
4692 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
4695 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
4696 link if your program is multi-threaded. Use @option{-fmudflapir}, in
4697 addition to @option{-fmudflap} or @option{-fmudflapth}, if
4698 instrumentation should ignore pointer reads. This produces less
4699 instrumentation (and therefore faster execution) and still provides
4700 some protection against outright memory corrupting writes, but allows
4701 erroneously read data to propagate within a program.
4703 @item -fstrength-reduce
4704 @opindex fstrength-reduce
4705 Perform the optimizations of loop strength reduction and
4706 elimination of iteration variables.
4708 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4710 @item -fthread-jumps
4711 @opindex fthread-jumps
4712 Perform optimizations where we check to see if a jump branches to a
4713 location where another comparison subsumed by the first is found. If
4714 so, the first branch is redirected to either the destination of the
4715 second branch or a point immediately following it, depending on whether
4716 the condition is known to be true or false.
4718 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4720 @item -fcse-follow-jumps
4721 @opindex fcse-follow-jumps
4722 In common subexpression elimination, scan through jump instructions
4723 when the target of the jump is not reached by any other path. For
4724 example, when CSE encounters an @code{if} statement with an
4725 @code{else} clause, CSE will follow the jump when the condition
4728 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4730 @item -fcse-skip-blocks
4731 @opindex fcse-skip-blocks
4732 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
4733 follow jumps which conditionally skip over blocks. When CSE
4734 encounters a simple @code{if} statement with no else clause,
4735 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
4736 body of the @code{if}.
4738 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4740 @item -frerun-cse-after-loop
4741 @opindex frerun-cse-after-loop
4742 Re-run common subexpression elimination after loop optimizations has been
4745 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4747 @item -frerun-loop-opt
4748 @opindex frerun-loop-opt
4749 Run the loop optimizer twice.
4751 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4755 Perform a global common subexpression elimination pass.
4756 This pass also performs global constant and copy propagation.
4758 @emph{Note:} When compiling a program using computed gotos, a GCC
4759 extension, you may get better runtime performance if you disable
4760 the global common subexpression elimination pass by adding
4761 @option{-fno-gcse} to the command line.
4763 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4767 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
4768 attempt to move loads which are only killed by stores into themselves. This
4769 allows a loop containing a load/store sequence to be changed to a load outside
4770 the loop, and a copy/store within the loop.
4772 Enabled by default when gcse is enabled.
4776 When @option{-fgcse-sm} is enabled, a store motion pass is run after
4777 global common subexpression elimination. This pass will attempt to move
4778 stores out of loops. When used in conjunction with @option{-fgcse-lm},
4779 loops containing a load/store sequence can be changed to a load before
4780 the loop and a store after the loop.
4782 Not enabled at any optimization level.
4786 When @option{-fgcse-las} is enabled, the global common subexpression
4787 elimination pass eliminates redundant loads that come after stores to the
4788 same memory location (both partial and full redundancies).
4790 Not enabled at any optimization level.
4792 @item -fgcse-after-reload
4793 @opindex fgcse-after-reload
4794 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
4795 pass is performed after reload. The purpose of this pass is to cleanup
4798 @item -floop-optimize
4799 @opindex floop-optimize
4800 Perform loop optimizations: move constant expressions out of loops, simplify
4801 exit test conditions and optionally do strength-reduction as well.
4803 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4805 @item -floop-optimize2
4806 @opindex floop-optimize2
4807 Perform loop optimizations using the new loop optimizer. The optimizations
4808 (loop unrolling, peeling and unswitching, loop invariant motion) are enabled
4811 @item -funsafe-loop-optimizations
4812 @opindex funsafe-loop-optimizations
4813 If given, the loop optimizer will assume that loop indices do not
4814 overflow, and that the loops with nontrivial exit condition are not
4815 infinite. This enables a wider range of loop optimizations even if
4816 the loop optimizer itself cannot prove that these assumptions are valid.
4817 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
4818 if it finds this kind of loop.
4820 @item -fcrossjumping
4821 @opindex crossjumping
4822 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
4823 resulting code may or may not perform better than without cross-jumping.
4825 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4827 @item -fif-conversion
4828 @opindex if-conversion
4829 Attempt to transform conditional jumps into branch-less equivalents. This
4830 include use of conditional moves, min, max, set flags and abs instructions, and
4831 some tricks doable by standard arithmetics. The use of conditional execution
4832 on chips where it is available is controlled by @code{if-conversion2}.
4834 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4836 @item -fif-conversion2
4837 @opindex if-conversion2
4838 Use conditional execution (where available) to transform conditional jumps into
4839 branch-less equivalents.
4841 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4843 @item -fdelete-null-pointer-checks
4844 @opindex fdelete-null-pointer-checks
4845 Use global dataflow analysis to identify and eliminate useless checks
4846 for null pointers. The compiler assumes that dereferencing a null
4847 pointer would have halted the program. If a pointer is checked after
4848 it has already been dereferenced, it cannot be null.
4850 In some environments, this assumption is not true, and programs can
4851 safely dereference null pointers. Use
4852 @option{-fno-delete-null-pointer-checks} to disable this optimization
4853 for programs which depend on that behavior.
4855 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4857 @item -fexpensive-optimizations
4858 @opindex fexpensive-optimizations
4859 Perform a number of minor optimizations that are relatively expensive.
4861 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4863 @item -foptimize-register-move
4865 @opindex foptimize-register-move
4867 Attempt to reassign register numbers in move instructions and as
4868 operands of other simple instructions in order to maximize the amount of
4869 register tying. This is especially helpful on machines with two-operand
4872 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
4875 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4877 @item -fdelayed-branch
4878 @opindex fdelayed-branch
4879 If supported for the target machine, attempt to reorder instructions
4880 to exploit instruction slots available after delayed branch
4883 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4885 @item -fschedule-insns
4886 @opindex fschedule-insns
4887 If supported for the target machine, attempt to reorder instructions to
4888 eliminate execution stalls due to required data being unavailable. This
4889 helps machines that have slow floating point or memory load instructions
4890 by allowing other instructions to be issued until the result of the load
4891 or floating point instruction is required.
4893 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4895 @item -fschedule-insns2
4896 @opindex fschedule-insns2
4897 Similar to @option{-fschedule-insns}, but requests an additional pass of
4898 instruction scheduling after register allocation has been done. This is
4899 especially useful on machines with a relatively small number of
4900 registers and where memory load instructions take more than one cycle.
4902 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4904 @item -fno-sched-interblock
4905 @opindex fno-sched-interblock
4906 Don't schedule instructions across basic blocks. This is normally
4907 enabled by default when scheduling before register allocation, i.e.@:
4908 with @option{-fschedule-insns} or at @option{-O2} or higher.
4910 @item -fno-sched-spec
4911 @opindex fno-sched-spec
4912 Don't allow speculative motion of non-load instructions. This is normally
4913 enabled by default when scheduling before register allocation, i.e.@:
4914 with @option{-fschedule-insns} or at @option{-O2} or higher.
4916 @item -fsched-spec-load
4917 @opindex fsched-spec-load
4918 Allow speculative motion of some load instructions. This only makes
4919 sense when scheduling before register allocation, i.e.@: with
4920 @option{-fschedule-insns} or at @option{-O2} or higher.
4922 @item -fsched-spec-load-dangerous
4923 @opindex fsched-spec-load-dangerous
4924 Allow speculative motion of more load instructions. This only makes
4925 sense when scheduling before register allocation, i.e.@: with
4926 @option{-fschedule-insns} or at @option{-O2} or higher.
4928 @item -fsched-stalled-insns=@var{n}
4929 @opindex fsched-stalled-insns
4930 Define how many insns (if any) can be moved prematurely from the queue
4931 of stalled insns into the ready list, during the second scheduling pass.
4933 @item -fsched-stalled-insns-dep=@var{n}
4934 @opindex fsched-stalled-insns-dep
4935 Define how many insn groups (cycles) will be examined for a dependency
4936 on a stalled insn that is candidate for premature removal from the queue
4937 of stalled insns. Has an effect only during the second scheduling pass,
4938 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
4940 @item -fsched2-use-superblocks
4941 @opindex fsched2-use-superblocks
4942 When scheduling after register allocation, do use superblock scheduling
4943 algorithm. Superblock scheduling allows motion across basic block boundaries
4944 resulting on faster schedules. This option is experimental, as not all machine
4945 descriptions used by GCC model the CPU closely enough to avoid unreliable
4946 results from the algorithm.
4948 This only makes sense when scheduling after register allocation, i.e.@: with
4949 @option{-fschedule-insns2} or at @option{-O2} or higher.
4951 @item -fsched2-use-traces
4952 @opindex fsched2-use-traces
4953 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4954 allocation and additionally perform code duplication in order to increase the
4955 size of superblocks using tracer pass. See @option{-ftracer} for details on
4958 This mode should produce faster but significantly longer programs. Also
4959 without @option{-fbranch-probabilities} the traces constructed may not
4960 match the reality and hurt the performance. This only makes
4961 sense when scheduling after register allocation, i.e.@: with
4962 @option{-fschedule-insns2} or at @option{-O2} or higher.
4964 @item -freschedule-modulo-scheduled-loops
4965 @opindex fscheduling-in-modulo-scheduled-loops
4966 The modulo scheduling comes before the traditional scheduling, if a loop was modulo scheduled
4967 we may want to prevent the later scheduling passes from changing its schedule, we use this
4968 option to control that.
4970 @item -fcaller-saves
4971 @opindex fcaller-saves
4972 Enable values to be allocated in registers that will be clobbered by
4973 function calls, by emitting extra instructions to save and restore the
4974 registers around such calls. Such allocation is done only when it
4975 seems to result in better code than would otherwise be produced.
4977 This option is always enabled by default on certain machines, usually
4978 those which have no call-preserved registers to use instead.
4980 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4983 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
4984 enabled by default at @option{-O2} and @option{-O3}.
4987 Perform Full Redundancy Elimination (FRE) on trees. The difference
4988 between FRE and PRE is that FRE only considers expressions
4989 that are computed on all paths leading to the redundant computation.
4990 This analysis faster than PRE, though it exposes fewer redundancies.
4991 This flag is enabled by default at @option{-O} and higher.
4993 @item -ftree-copy-prop
4994 Perform copy propagation on trees. This pass eliminates unnecessary
4995 copy operations. This flag is enabled by default at @option{-O} and
4998 @item -ftree-store-copy-prop
4999 Perform copy propagation of memory loads and stores. This pass
5000 eliminates unnecessary copy operations in memory references
5001 (structures, global variables, arrays, etc). This flag is enabled by
5002 default at @option{-O2} and higher.
5005 Perform structural alias analysis on trees. This flag
5006 is enabled by default at @option{-O} and higher.
5009 Perform forward store motion on trees. This flag is
5010 enabled by default at @option{-O} and higher.
5013 Perform sparse conditional constant propagation (CCP) on trees. This
5014 pass only operates on local scalar variables and is enabled by default
5015 at @option{-O} and higher.
5017 @item -ftree-store-ccp
5018 Perform sparse conditional constant propagation (CCP) on trees. This
5019 pass operates on both local scalar variables and memory stores and
5020 loads (global variables, structures, arrays, etc). This flag is
5021 enabled by default at @option{-O2} and higher.
5024 Perform dead code elimination (DCE) on trees. This flag is enabled by
5025 default at @option{-O} and higher.
5027 @item -ftree-dominator-opts
5028 Perform a variety of simple scalar cleanups (constant/copy
5029 propagation, redundancy elimination, range propagation and expression
5030 simplification) based on a dominator tree traversal. This also
5031 performs jump threading (to reduce jumps to jumps). This flag is
5032 enabled by default at @option{-O} and higher.
5035 Perform loop header copying on trees. This is beneficial since it increases
5036 effectiveness of code motion optimizations. It also saves one jump. This flag
5037 is enabled by default at @option{-O} and higher. It is not enabled
5038 for @option{-Os}, since it usually increases code size.
5040 @item -ftree-loop-optimize
5041 Perform loop optimizations on trees. This flag is enabled by default
5042 at @option{-O} and higher.
5044 @item -ftree-loop-linear
5045 Perform linear loop transformations on tree. This flag can improve cache
5046 performance and allow further loop optimizations to take place.
5048 @item -ftree-loop-im
5049 Perform loop invariant motion on trees. This pass moves only invariants that
5050 would be hard to handle at RTL level (function calls, operations that expand to
5051 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5052 operands of conditions that are invariant out of the loop, so that we can use
5053 just trivial invariantness analysis in loop unswitching. The pass also includes
5056 @item -ftree-loop-ivcanon
5057 Create a canonical counter for number of iterations in the loop for that
5058 determining number of iterations requires complicated analysis. Later
5059 optimizations then may determine the number easily. Useful especially
5060 in connection with unrolling.
5063 Perform induction variable optimizations (strength reduction, induction
5064 variable merging and induction variable elimination) on trees.
5067 Perform scalar replacement of aggregates. This pass replaces structure
5068 references with scalars to prevent committing structures to memory too
5069 early. This flag is enabled by default at @option{-O} and higher.
5071 @item -ftree-copyrename
5072 Perform copy renaming on trees. This pass attempts to rename compiler
5073 temporaries to other variables at copy locations, usually resulting in
5074 variable names which more closely resemble the original variables. This flag
5075 is enabled by default at @option{-O} and higher.
5078 Perform temporary expression replacement during the SSA->normal phase. Single
5079 use/single def temporaries are replaced at their use location with their
5080 defining expression. This results in non-GIMPLE code, but gives the expanders
5081 much more complex trees to work on resulting in better RTL generation. This is
5082 enabled by default at @option{-O} and higher.
5085 Perform live range splitting during the SSA->normal phase. Distinct live
5086 ranges of a variable are split into unique variables, allowing for better
5087 optimization later. This is enabled by default at @option{-O} and higher.
5089 @item -ftree-vectorize
5090 Perform loop vectorization on trees.
5092 @item -ftree-vect-loop-version
5093 @opindex ftree-vect-loop-version
5094 Perform loop versioning when doing loop vectorization on trees. When a loop
5095 appears to be vectorizable except that data alignment or data dependence cannot
5096 be determined at compile time then vectorized and non-vectorized versions of
5097 the loop are generated along with runtime checks for alignment or dependence
5098 to control which version is executed. This option is enabled by default
5099 except at level @option{-Os} where it is disabled.
5102 Perform Value Range Propagation on trees. This is similar to the
5103 constant propagation pass, but instead of values, ranges of values are
5104 propagated. This allows the optimizers to remove unnecessary range
5105 checks like array bound checks and null pointer checks. This is
5106 enabled by default at @option{-O2} and higher. Null pointer check
5107 elimination is only done if @option{-fdelete-null-pointer-checks} is
5112 Perform tail duplication to enlarge superblock size. This transformation
5113 simplifies the control flow of the function allowing other optimizations to do
5116 @item -funroll-loops
5117 @opindex funroll-loops
5118 Unroll loops whose number of iterations can be determined at compile
5119 time or upon entry to the loop. @option{-funroll-loops} implies both
5120 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
5121 option makes code larger, and may or may not make it run faster.
5123 @item -funroll-all-loops
5124 @opindex funroll-all-loops
5125 Unroll all loops, even if their number of iterations is uncertain when
5126 the loop is entered. This usually makes programs run more slowly.
5127 @option{-funroll-all-loops} implies the same options as
5128 @option{-funroll-loops},
5130 @item -fsplit-ivs-in-unroller
5131 @opindex -fsplit-ivs-in-unroller
5132 Enables expressing of values of induction variables in later iterations
5133 of the unrolled loop using the value in the first iteration. This breaks
5134 long dependency chains, thus improving efficiency of the scheduling passes.
5136 Combination of @option{-fweb} and CSE is often sufficient to obtain the
5137 same effect. However in cases the loop body is more complicated than
5138 a single basic block, this is not reliable. It also does not work at all
5139 on some of the architectures due to restrictions in the CSE pass.
5141 This optimization is enabled by default.
5143 @item -fvariable-expansion-in-unroller
5144 @opindex -fvariable-expansion-in-unroller
5145 With this option, the compiler will create multiple copies of some
5146 local variables when unrolling a loop which can result in superior code.
5148 @item -fprefetch-loop-arrays
5149 @opindex fprefetch-loop-arrays
5150 If supported by the target machine, generate instructions to prefetch
5151 memory to improve the performance of loops that access large arrays.
5153 These options may generate better or worse code; results are highly
5154 dependent on the structure of loops within the source code.
5157 @itemx -fno-peephole2
5158 @opindex fno-peephole
5159 @opindex fno-peephole2
5160 Disable any machine-specific peephole optimizations. The difference
5161 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
5162 are implemented in the compiler; some targets use one, some use the
5163 other, a few use both.
5165 @option{-fpeephole} is enabled by default.
5166 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5168 @item -fno-guess-branch-probability
5169 @opindex fno-guess-branch-probability
5170 Do not guess branch probabilities using heuristics.
5172 GCC will use heuristics to guess branch probabilities if they are
5173 not provided by profiling feedback (@option{-fprofile-arcs}). These
5174 heuristics are based on the control flow graph. If some branch probabilities
5175 are specified by @samp{__builtin_expect}, then the heuristics will be
5176 used to guess branch probabilities for the rest of the control flow graph,
5177 taking the @samp{__builtin_expect} info into account. The interactions
5178 between the heuristics and @samp{__builtin_expect} can be complex, and in
5179 some cases, it may be useful to disable the heuristics so that the effects
5180 of @samp{__builtin_expect} are easier to understand.
5182 The default is @option{-fguess-branch-probability} at levels
5183 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5185 @item -freorder-blocks
5186 @opindex freorder-blocks
5187 Reorder basic blocks in the compiled function in order to reduce number of
5188 taken branches and improve code locality.
5190 Enabled at levels @option{-O2}, @option{-O3}.
5192 @item -freorder-blocks-and-partition
5193 @opindex freorder-blocks-and-partition
5194 In addition to reordering basic blocks in the compiled function, in order
5195 to reduce number of taken branches, partitions hot and cold basic blocks
5196 into separate sections of the assembly and .o files, to improve
5197 paging and cache locality performance.
5199 This optimization is automatically turned off in the presence of
5200 exception handling, for linkonce sections, for functions with a user-defined
5201 section attribute and on any architecture that does not support named
5204 @item -freorder-functions
5205 @opindex freorder-functions
5206 Reorder functions in the object file in order to
5207 improve code locality. This is implemented by using special
5208 subsections @code{.text.hot} for most frequently executed functions and
5209 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
5210 the linker so object file format must support named sections and linker must
5211 place them in a reasonable way.
5213 Also profile feedback must be available in to make this option effective. See
5214 @option{-fprofile-arcs} for details.
5216 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5218 @item -fstrict-aliasing
5219 @opindex fstrict-aliasing
5220 Allows the compiler to assume the strictest aliasing rules applicable to
5221 the language being compiled. For C (and C++), this activates
5222 optimizations based on the type of expressions. In particular, an
5223 object of one type is assumed never to reside at the same address as an
5224 object of a different type, unless the types are almost the same. For
5225 example, an @code{unsigned int} can alias an @code{int}, but not a
5226 @code{void*} or a @code{double}. A character type may alias any other
5229 Pay special attention to code like this:
5242 The practice of reading from a different union member than the one most
5243 recently written to (called ``type-punning'') is common. Even with
5244 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
5245 is accessed through the union type. So, the code above will work as
5246 expected. However, this code might not:
5257 Every language that wishes to perform language-specific alias analysis
5258 should define a function that computes, given an @code{tree}
5259 node, an alias set for the node. Nodes in different alias sets are not
5260 allowed to alias. For an example, see the C front-end function
5261 @code{c_get_alias_set}.
5263 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5265 @item -falign-functions
5266 @itemx -falign-functions=@var{n}
5267 @opindex falign-functions
5268 Align the start of functions to the next power-of-two greater than
5269 @var{n}, skipping up to @var{n} bytes. For instance,
5270 @option{-falign-functions=32} aligns functions to the next 32-byte
5271 boundary, but @option{-falign-functions=24} would align to the next
5272 32-byte boundary only if this can be done by skipping 23 bytes or less.
5274 @option{-fno-align-functions} and @option{-falign-functions=1} are
5275 equivalent and mean that functions will not be aligned.
5277 Some assemblers only support this flag when @var{n} is a power of two;
5278 in that case, it is rounded up.
5280 If @var{n} is not specified or is zero, use a machine-dependent default.
5282 Enabled at levels @option{-O2}, @option{-O3}.
5284 @item -falign-labels
5285 @itemx -falign-labels=@var{n}
5286 @opindex falign-labels
5287 Align all branch targets to a power-of-two boundary, skipping up to
5288 @var{n} bytes like @option{-falign-functions}. This option can easily
5289 make code slower, because it must insert dummy operations for when the
5290 branch target is reached in the usual flow of the code.
5292 @option{-fno-align-labels} and @option{-falign-labels=1} are
5293 equivalent and mean that labels will not be aligned.
5295 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
5296 are greater than this value, then their values are used instead.
5298 If @var{n} is not specified or is zero, use a machine-dependent default
5299 which is very likely to be @samp{1}, meaning no alignment.
5301 Enabled at levels @option{-O2}, @option{-O3}.
5304 @itemx -falign-loops=@var{n}
5305 @opindex falign-loops
5306 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
5307 like @option{-falign-functions}. The hope is that the loop will be
5308 executed many times, which will make up for any execution of the dummy
5311 @option{-fno-align-loops} and @option{-falign-loops=1} are
5312 equivalent and mean that loops will not be aligned.
5314 If @var{n} is not specified or is zero, use a machine-dependent default.
5316 Enabled at levels @option{-O2}, @option{-O3}.
5319 @itemx -falign-jumps=@var{n}
5320 @opindex falign-jumps
5321 Align branch targets to a power-of-two boundary, for branch targets
5322 where the targets can only be reached by jumping, skipping up to @var{n}
5323 bytes like @option{-falign-functions}. In this case, no dummy operations
5326 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
5327 equivalent and mean that loops will not be aligned.
5329 If @var{n} is not specified or is zero, use a machine-dependent default.
5331 Enabled at levels @option{-O2}, @option{-O3}.
5333 @item -funit-at-a-time
5334 @opindex funit-at-a-time
5335 Parse the whole compilation unit before starting to produce code.
5336 This allows some extra optimizations to take place but consumes
5337 more memory (in general). There are some compatibility issues
5338 with @emph{unit-at-at-time} mode:
5341 enabling @emph{unit-at-a-time} mode may change the order
5342 in which functions, variables, and top-level @code{asm} statements
5343 are emitted, and will likely break code relying on some particular
5344 ordering. The majority of such top-level @code{asm} statements,
5345 though, can be replaced by @code{section} attributes.
5348 @emph{unit-at-a-time} mode removes unreferenced static variables
5349 and functions. This may result in undefined references
5350 when an @code{asm} statement refers directly to variables or functions
5351 that are otherwise unused. In that case either the variable/function
5352 shall be listed as an operand of the @code{asm} statement operand or,
5353 in the case of top-level @code{asm} statements the attribute @code{used}
5354 shall be used on the declaration.
5357 Static functions now can use non-standard passing conventions that
5358 may break @code{asm} statements calling functions directly. Again,
5359 attribute @code{used} will prevent this behavior.
5362 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
5363 but this scheme may not be supported by future releases of GCC@.
5365 Enabled at levels @option{-O2}, @option{-O3}.
5369 Constructs webs as commonly used for register allocation purposes and assign
5370 each web individual pseudo register. This allows the register allocation pass
5371 to operate on pseudos directly, but also strengthens several other optimization
5372 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
5373 however, make debugging impossible, since variables will no longer stay in a
5376 Enabled by default with @option{-funroll-loops}.
5378 @item -fwhole-program
5379 @opindex fwhole-program
5380 Assume that the current compilation unit represents whole program being
5381 compiled. All public functions and variables with the exception of @code{main}
5382 and those merged by attribute @code{externally_visible} become static functions
5383 and in a affect gets more aggressively optimized by interprocedural optimizers.
5384 While this option is equivalent to proper use of @code{static} keyword for
5385 programs consisting of single file, in combination with option
5386 @option{--combine} this flag can be used to compile most of smaller scale C
5387 programs since the functions and variables become local for the whole combined
5388 compilation unit, not for the single source file itself.
5391 @item -fno-cprop-registers
5392 @opindex fno-cprop-registers
5393 After register allocation and post-register allocation instruction splitting,
5394 we perform a copy-propagation pass to try to reduce scheduling dependencies
5395 and occasionally eliminate the copy.
5397 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5399 @item -fprofile-generate
5400 @opindex fprofile-generate
5402 Enable options usually used for instrumenting application to produce
5403 profile useful for later recompilation with profile feedback based
5404 optimization. You must use @option{-fprofile-generate} both when
5405 compiling and when linking your program.
5407 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
5410 @opindex fprofile-use
5411 Enable profile feedback directed optimizations, and optimizations
5412 generally profitable only with profile feedback available.
5414 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
5415 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer},
5416 @code{-fno-loop-optimize}.
5420 The following options control compiler behavior regarding floating
5421 point arithmetic. These options trade off between speed and
5422 correctness. All must be specifically enabled.
5426 @opindex ffloat-store
5427 Do not store floating point variables in registers, and inhibit other
5428 options that might change whether a floating point value is taken from a
5431 @cindex floating point precision
5432 This option prevents undesirable excess precision on machines such as
5433 the 68000 where the floating registers (of the 68881) keep more
5434 precision than a @code{double} is supposed to have. Similarly for the
5435 x86 architecture. For most programs, the excess precision does only
5436 good, but a few programs rely on the precise definition of IEEE floating
5437 point. Use @option{-ffloat-store} for such programs, after modifying
5438 them to store all pertinent intermediate computations into variables.
5442 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
5443 @option{-fno-trapping-math}, @option{-ffinite-math-only},
5444 @option{-fno-rounding-math}, @option{-fno-signaling-nans}
5445 and @option{fcx-limited-range}.
5447 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
5449 This option should never be turned on by any @option{-O} option since
5450 it can result in incorrect output for programs which depend on
5451 an exact implementation of IEEE or ISO rules/specifications for
5454 @item -fno-math-errno
5455 @opindex fno-math-errno
5456 Do not set ERRNO after calling math functions that are executed
5457 with a single instruction, e.g., sqrt. A program that relies on
5458 IEEE exceptions for math error handling may want to use this flag
5459 for speed while maintaining IEEE arithmetic compatibility.
5461 This option should never be turned on by any @option{-O} option since
5462 it can result in incorrect output for programs which depend on
5463 an exact implementation of IEEE or ISO rules/specifications for
5466 The default is @option{-fmath-errno}.
5468 On Darwin systems, the math library never sets @code{errno}. There is therefore
5469 no reason for the compiler to consider the possibility that it might,
5470 and @option{-fno-math-errno} is the default.
5472 @item -funsafe-math-optimizations
5473 @opindex funsafe-math-optimizations
5474 Allow optimizations for floating-point arithmetic that (a) assume
5475 that arguments and results are valid and (b) may violate IEEE or
5476 ANSI standards. When used at link-time, it may include libraries
5477 or startup files that change the default FPU control word or other
5478 similar optimizations.
5480 This option should never be turned on by any @option{-O} option since
5481 it can result in incorrect output for programs which depend on
5482 an exact implementation of IEEE or ISO rules/specifications for
5485 The default is @option{-fno-unsafe-math-optimizations}.
5487 @item -ffinite-math-only
5488 @opindex ffinite-math-only
5489 Allow optimizations for floating-point arithmetic that assume
5490 that arguments and results are not NaNs or +-Infs.
5492 This option should never be turned on by any @option{-O} option since
5493 it can result in incorrect output for programs which depend on
5494 an exact implementation of IEEE or ISO rules/specifications.
5496 The default is @option{-fno-finite-math-only}.
5498 @item -fno-trapping-math
5499 @opindex fno-trapping-math
5500 Compile code assuming that floating-point operations cannot generate
5501 user-visible traps. These traps include division by zero, overflow,
5502 underflow, inexact result and invalid operation. This option implies
5503 @option{-fno-signaling-nans}. Setting this option may allow faster
5504 code if one relies on ``non-stop'' IEEE arithmetic, for example.
5506 This option should never be turned on by any @option{-O} option since
5507 it can result in incorrect output for programs which depend on
5508 an exact implementation of IEEE or ISO rules/specifications for
5511 The default is @option{-ftrapping-math}.
5513 @item -frounding-math
5514 @opindex frounding-math
5515 Disable transformations and optimizations that assume default floating
5516 point rounding behavior. This is round-to-zero for all floating point
5517 to integer conversions, and round-to-nearest for all other arithmetic
5518 truncations. This option should be specified for programs that change
5519 the FP rounding mode dynamically, or that may be executed with a
5520 non-default rounding mode. This option disables constant folding of
5521 floating point expressions at compile-time (which may be affected by
5522 rounding mode) and arithmetic transformations that are unsafe in the
5523 presence of sign-dependent rounding modes.
5525 The default is @option{-fno-rounding-math}.
5527 This option is experimental and does not currently guarantee to
5528 disable all GCC optimizations that are affected by rounding mode.
5529 Future versions of GCC may provide finer control of this setting
5530 using C99's @code{FENV_ACCESS} pragma. This command line option
5531 will be used to specify the default state for @code{FENV_ACCESS}.
5533 @item -fsignaling-nans
5534 @opindex fsignaling-nans
5535 Compile code assuming that IEEE signaling NaNs may generate user-visible
5536 traps during floating-point operations. Setting this option disables
5537 optimizations that may change the number of exceptions visible with
5538 signaling NaNs. This option implies @option{-ftrapping-math}.
5540 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
5543 The default is @option{-fno-signaling-nans}.
5545 This option is experimental and does not currently guarantee to
5546 disable all GCC optimizations that affect signaling NaN behavior.
5548 @item -fsingle-precision-constant
5549 @opindex fsingle-precision-constant
5550 Treat floating point constant as single precision constant instead of
5551 implicitly converting it to double precision constant.
5553 @item -fcx-limited-range
5554 @itemx -fno-cx-limited-range
5555 @opindex fcx-limited-range
5556 @opindex fno-cx-limited-range
5557 When enabled, this option states that a range reduction step is not
5558 needed when performing complex division. The default is
5559 @option{-fno-cx-limited-range}, but is enabled by @option{-ffast-math}.
5561 This option controls the default setting of the ISO C99
5562 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
5567 The following options control optimizations that may improve
5568 performance, but are not enabled by any @option{-O} options. This
5569 section includes experimental options that may produce broken code.
5572 @item -fbranch-probabilities
5573 @opindex fbranch-probabilities
5574 After running a program compiled with @option{-fprofile-arcs}
5575 (@pxref{Debugging Options,, Options for Debugging Your Program or
5576 @command{gcc}}), you can compile it a second time using
5577 @option{-fbranch-probabilities}, to improve optimizations based on
5578 the number of times each branch was taken. When the program
5579 compiled with @option{-fprofile-arcs} exits it saves arc execution
5580 counts to a file called @file{@var{sourcename}.gcda} for each source
5581 file The information in this data file is very dependent on the
5582 structure of the generated code, so you must use the same source code
5583 and the same optimization options for both compilations.
5585 With @option{-fbranch-probabilities}, GCC puts a
5586 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
5587 These can be used to improve optimization. Currently, they are only
5588 used in one place: in @file{reorg.c}, instead of guessing which path a
5589 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
5590 exactly determine which path is taken more often.
5592 @item -fprofile-values
5593 @opindex fprofile-values
5594 If combined with @option{-fprofile-arcs}, it adds code so that some
5595 data about values of expressions in the program is gathered.
5597 With @option{-fbranch-probabilities}, it reads back the data gathered
5598 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
5599 notes to instructions for their later usage in optimizations.
5601 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
5605 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
5606 a code to gather information about values of expressions.
5608 With @option{-fbranch-probabilities}, it reads back the data gathered
5609 and actually performs the optimizations based on them.
5610 Currently the optimizations include specialization of division operation
5611 using the knowledge about the value of the denominator.
5613 @item -frename-registers
5614 @opindex frename-registers
5615 Attempt to avoid false dependencies in scheduled code by making use
5616 of registers left over after register allocation. This optimization
5617 will most benefit processors with lots of registers. Depending on the
5618 debug information format adopted by the target, however, it can
5619 make debugging impossible, since variables will no longer stay in
5620 a ``home register''.
5622 Enabled by default with @option{-funroll-loops}.
5626 Perform tail duplication to enlarge superblock size. This transformation
5627 simplifies the control flow of the function allowing other optimizations to do
5630 Enabled with @option{-fprofile-use}.
5632 @item -funroll-loops
5633 @opindex funroll-loops
5634 Unroll loops whose number of iterations can be determined at compile time or
5635 upon entry to the loop. @option{-funroll-loops} implies
5636 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
5637 It also turns on complete loop peeling (i.e.@: complete removal of loops with
5638 small constant number of iterations). This option makes code larger, and may
5639 or may not make it run faster.
5641 Enabled with @option{-fprofile-use}.
5643 @item -funroll-all-loops
5644 @opindex funroll-all-loops
5645 Unroll all loops, even if their number of iterations is uncertain when
5646 the loop is entered. This usually makes programs run more slowly.
5647 @option{-funroll-all-loops} implies the same options as
5648 @option{-funroll-loops}.
5651 @opindex fpeel-loops
5652 Peels the loops for that there is enough information that they do not
5653 roll much (from profile feedback). It also turns on complete loop peeling
5654 (i.e.@: complete removal of loops with small constant number of iterations).
5656 Enabled with @option{-fprofile-use}.
5658 @item -fmove-loop-invariants
5659 @opindex fmove-loop-invariants
5660 Enables the loop invariant motion pass in the new loop optimizer. Enabled
5661 at level @option{-O1}
5663 @item -funswitch-loops
5664 @opindex funswitch-loops
5665 Move branches with loop invariant conditions out of the loop, with duplicates
5666 of the loop on both branches (modified according to result of the condition).
5668 @item -fprefetch-loop-arrays
5669 @opindex fprefetch-loop-arrays
5670 If supported by the target machine, generate instructions to prefetch
5671 memory to improve the performance of loops that access large arrays.
5673 Disabled at level @option{-Os}.
5675 @item -ffunction-sections
5676 @itemx -fdata-sections
5677 @opindex ffunction-sections
5678 @opindex fdata-sections
5679 Place each function or data item into its own section in the output
5680 file if the target supports arbitrary sections. The name of the
5681 function or the name of the data item determines the section's name
5684 Use these options on systems where the linker can perform optimizations
5685 to improve locality of reference in the instruction space. Most systems
5686 using the ELF object format and SPARC processors running Solaris 2 have
5687 linkers with such optimizations. AIX may have these optimizations in
5690 Only use these options when there are significant benefits from doing
5691 so. When you specify these options, the assembler and linker will
5692 create larger object and executable files and will also be slower.
5693 You will not be able to use @code{gprof} on all systems if you
5694 specify this option and you may have problems with debugging if
5695 you specify both this option and @option{-g}.
5697 @item -fbranch-target-load-optimize
5698 @opindex fbranch-target-load-optimize
5699 Perform branch target register load optimization before prologue / epilogue
5701 The use of target registers can typically be exposed only during reload,
5702 thus hoisting loads out of loops and doing inter-block scheduling needs
5703 a separate optimization pass.
5705 @item -fbranch-target-load-optimize2
5706 @opindex fbranch-target-load-optimize2
5707 Perform branch target register load optimization after prologue / epilogue
5710 @item -fbtr-bb-exclusive
5711 @opindex fbtr-bb-exclusive
5712 When performing branch target register load optimization, don't reuse
5713 branch target registers in within any basic block.
5715 @item -fstack-protector
5716 Emit extra code to check for buffer overflows, such as stack smashing
5717 attacks. This is done by adding a guard variable to functions with
5718 vulnerable objects. This includes functions that call alloca, and
5719 functions with buffers larger than 8 bytes. The guards are initialized
5720 when a function is entered and then checked when the function exits.
5721 If a guard check fails, an error message is printed and the program exits.
5723 @item -fstack-protector-all
5724 Like @option{-fstack-protector} except that all functions are protected.
5726 @item --param @var{name}=@var{value}
5728 In some places, GCC uses various constants to control the amount of
5729 optimization that is done. For example, GCC will not inline functions
5730 that contain more that a certain number of instructions. You can
5731 control some of these constants on the command-line using the
5732 @option{--param} option.
5734 The names of specific parameters, and the meaning of the values, are
5735 tied to the internals of the compiler, and are subject to change
5736 without notice in future releases.
5738 In each case, the @var{value} is an integer. The allowable choices for
5739 @var{name} are given in the following table:
5742 @item salias-max-implicit-fields
5743 The maximum number of fields in a variable without direct
5744 structure accesses for which structure aliasing will consider trying
5745 to track each field. The default is 5
5747 @item sra-max-structure-size
5748 The maximum structure size, in bytes, at which the scalar replacement
5749 of aggregates (SRA) optimization will perform block copies. The
5750 default value, 0, implies that GCC will select the most appropriate
5753 @item sra-field-structure-ratio
5754 The threshold ratio (as a percentage) between instantiated fields and
5755 the complete structure size. We say that if the ratio of the number
5756 of bytes in instantiated fields to the number of bytes in the complete
5757 structure exceeds this parameter, then block copies are not used. The
5760 @item max-crossjump-edges
5761 The maximum number of incoming edges to consider for crossjumping.
5762 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
5763 the number of edges incoming to each block. Increasing values mean
5764 more aggressive optimization, making the compile time increase with
5765 probably small improvement in executable size.
5767 @item min-crossjump-insns
5768 The minimum number of instructions which must be matched at the end
5769 of two blocks before crossjumping will be performed on them. This
5770 value is ignored in the case where all instructions in the block being
5771 crossjumped from are matched. The default value is 5.
5773 @item max-grow-copy-bb-insns
5774 The maximum code size expansion factor when copying basic blocks
5775 instead of jumping. The expansion is relative to a jump instruction.
5776 The default value is 8.
5778 @item max-goto-duplication-insns
5779 The maximum number of instructions to duplicate to a block that jumps
5780 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
5781 passes, GCC factors computed gotos early in the compilation process,
5782 and unfactors them as late as possible. Only computed jumps at the
5783 end of a basic blocks with no more than max-goto-duplication-insns are
5784 unfactored. The default value is 8.
5786 @item max-delay-slot-insn-search
5787 The maximum number of instructions to consider when looking for an
5788 instruction to fill a delay slot. If more than this arbitrary number of
5789 instructions is searched, the time savings from filling the delay slot
5790 will be minimal so stop searching. Increasing values mean more
5791 aggressive optimization, making the compile time increase with probably
5792 small improvement in executable run time.
5794 @item max-delay-slot-live-search
5795 When trying to fill delay slots, the maximum number of instructions to
5796 consider when searching for a block with valid live register
5797 information. Increasing this arbitrarily chosen value means more
5798 aggressive optimization, increasing the compile time. This parameter
5799 should be removed when the delay slot code is rewritten to maintain the
5802 @item max-gcse-memory
5803 The approximate maximum amount of memory that will be allocated in
5804 order to perform the global common subexpression elimination
5805 optimization. If more memory than specified is required, the
5806 optimization will not be done.
5808 @item max-gcse-passes
5809 The maximum number of passes of GCSE to run. The default is 1.
5811 @item max-pending-list-length
5812 The maximum number of pending dependencies scheduling will allow
5813 before flushing the current state and starting over. Large functions
5814 with few branches or calls can create excessively large lists which
5815 needlessly consume memory and resources.
5817 @item max-inline-insns-single
5818 Several parameters control the tree inliner used in gcc.
5819 This number sets the maximum number of instructions (counted in GCC's
5820 internal representation) in a single function that the tree inliner
5821 will consider for inlining. This only affects functions declared
5822 inline and methods implemented in a class declaration (C++).
5823 The default value is 450.
5825 @item max-inline-insns-auto
5826 When you use @option{-finline-functions} (included in @option{-O3}),
5827 a lot of functions that would otherwise not be considered for inlining
5828 by the compiler will be investigated. To those functions, a different
5829 (more restrictive) limit compared to functions declared inline can
5831 The default value is 90.
5833 @item large-function-insns
5834 The limit specifying really large functions. For functions larger than this
5835 limit after inlining inlining is constrained by
5836 @option{--param large-function-growth}. This parameter is useful primarily
5837 to avoid extreme compilation time caused by non-linear algorithms used by the
5839 This parameter is ignored when @option{-funit-at-a-time} is not used.
5840 The default value is 2700.
5842 @item large-function-growth
5843 Specifies maximal growth of large function caused by inlining in percents.
5844 This parameter is ignored when @option{-funit-at-a-time} is not used.
5845 The default value is 100 which limits large function growth to 2.0 times
5848 @item large-unit-insns
5849 The limit specifying large translation unit. Growth caused by inlining of
5850 units larger than this limit is limited by @option{--param inline-unit-growth}.
5851 For small units this might be too tight (consider unit consisting of function A
5852 that is inline and B that just calls A three time. If B is small relative to
5853 A, the growth of unit is 300\% and yet such inlining is very sane. For very
5854 large units consisting of small inlininable functions however the overall unit
5855 growth limit is needed to avoid exponential explosion of code size. Thus for
5856 smaller units, the size is increased to @option{--param large-unit-insns}
5857 before applying @option{--param inline-unit-growth}. The default is 10000
5859 @item inline-unit-growth
5860 Specifies maximal overall growth of the compilation unit caused by inlining.
5861 This parameter is ignored when @option{-funit-at-a-time} is not used.
5862 The default value is 50 which limits unit growth to 1.5 times the original
5865 @item max-inline-insns-recursive
5866 @itemx max-inline-insns-recursive-auto
5867 Specifies maximum number of instructions out-of-line copy of self recursive inline
5868 function can grow into by performing recursive inlining.
5870 For functions declared inline @option{--param max-inline-insns-recursive} is
5871 taken into acount. For function not declared inline, recursive inlining
5872 happens only when @option{-finline-functions} (included in @option{-O3}) is
5873 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
5874 default value is 450.
5876 @item max-inline-recursive-depth
5877 @itemx max-inline-recursive-depth-auto
5878 Specifies maximum recursion depth used by the recursive inlining.
5880 For functions declared inline @option{--param max-inline-recursive-depth} is
5881 taken into acount. For function not declared inline, recursive inlining
5882 happens only when @option{-finline-functions} (included in @option{-O3}) is
5883 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
5884 default value is 450.
5886 @item min-inline-recursive-probability
5887 Recursive inlining is profitable only for function having deep recursion
5888 in average and can hurt for function having little recursion depth by
5889 increasing the prologue size or complexity of function body to other
5892 When profile feedback is available (see @option{-fprofile-generate}) the actual
5893 recursion depth can be guessed from probability that function will recurse via
5894 given call expression. This parameter limits inlining only to call expression
5895 whose probability exceeds given threshold (in percents). The default value is
5898 @item inline-call-cost
5899 Specify cost of call instruction relative to simple arithmetics operations
5900 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
5901 functions and at the same time increases size of leaf function that is believed to
5902 reduce function size by being inlined. In effect it increases amount of
5903 inlining for code having large abstraction penalty (many functions that just
5904 pass the arguments to other functions) and decrease inlining for code with low
5905 abstraction penalty. The default value is 16.
5907 @item max-unrolled-insns
5908 The maximum number of instructions that a loop should have if that loop
5909 is unrolled, and if the loop is unrolled, it determines how many times
5910 the loop code is unrolled.
5912 @item max-average-unrolled-insns
5913 The maximum number of instructions biased by probabilities of their execution
5914 that a loop should have if that loop is unrolled, and if the loop is unrolled,
5915 it determines how many times the loop code is unrolled.
5917 @item max-unroll-times
5918 The maximum number of unrollings of a single loop.
5920 @item max-peeled-insns
5921 The maximum number of instructions that a loop should have if that loop
5922 is peeled, and if the loop is peeled, it determines how many times
5923 the loop code is peeled.
5925 @item max-peel-times
5926 The maximum number of peelings of a single loop.
5928 @item max-completely-peeled-insns
5929 The maximum number of insns of a completely peeled loop.
5931 @item max-completely-peel-times
5932 The maximum number of iterations of a loop to be suitable for complete peeling.
5934 @item max-unswitch-insns
5935 The maximum number of insns of an unswitched loop.
5937 @item max-unswitch-level
5938 The maximum number of branches unswitched in a single loop.
5941 The minimum cost of an expensive expression in the loop invariant motion.
5943 @item iv-consider-all-candidates-bound
5944 Bound on number of candidates for induction variables below that
5945 all candidates are considered for each use in induction variable
5946 optimizations. Only the most relevant candidates are considered
5947 if there are more candidates, to avoid quadratic time complexity.
5949 @item iv-max-considered-uses
5950 The induction variable optimizations give up on loops that contain more
5951 induction variable uses.
5953 @item iv-always-prune-cand-set-bound
5954 If number of candidates in the set is smaller than this value,
5955 we always try to remove unnecessary ivs from the set during its
5956 optimization when a new iv is added to the set.
5958 @item scev-max-expr-size
5959 Bound on size of expressions used in the scalar evolutions analyzer.
5960 Large expressions slow the analyzer.
5962 @item vect-max-version-checks
5963 The maximum number of runtime checks that can be performed when doing
5964 loop versioning in the vectorizer. See option ftree-vect-loop-version
5965 for more information.
5967 @item max-iterations-to-track
5969 The maximum number of iterations of a loop the brute force algorithm
5970 for analysis of # of iterations of the loop tries to evaluate.
5972 @item hot-bb-count-fraction
5973 Select fraction of the maximal count of repetitions of basic block in program
5974 given basic block needs to have to be considered hot.
5976 @item hot-bb-frequency-fraction
5977 Select fraction of the maximal frequency of executions of basic block in
5978 function given basic block needs to have to be considered hot
5980 @item max-predicted-iterations
5981 The maximum number of loop iterations we predict statically. This is useful
5982 in cases where function contain single loop with known bound and other loop
5983 with unknown. We predict the known number of iterations correctly, while
5984 the unknown number of iterations average to roughly 10. This means that the
5985 loop without bounds would appear artificially cold relative to the other one.
5987 @item tracer-dynamic-coverage
5988 @itemx tracer-dynamic-coverage-feedback
5990 This value is used to limit superblock formation once the given percentage of
5991 executed instructions is covered. This limits unnecessary code size
5994 The @option{tracer-dynamic-coverage-feedback} is used only when profile
5995 feedback is available. The real profiles (as opposed to statically estimated
5996 ones) are much less balanced allowing the threshold to be larger value.
5998 @item tracer-max-code-growth
5999 Stop tail duplication once code growth has reached given percentage. This is
6000 rather hokey argument, as most of the duplicates will be eliminated later in
6001 cross jumping, so it may be set to much higher values than is the desired code
6004 @item tracer-min-branch-ratio
6006 Stop reverse growth when the reverse probability of best edge is less than this
6007 threshold (in percent).
6009 @item tracer-min-branch-ratio
6010 @itemx tracer-min-branch-ratio-feedback
6012 Stop forward growth if the best edge do have probability lower than this
6015 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
6016 compilation for profile feedback and one for compilation without. The value
6017 for compilation with profile feedback needs to be more conservative (higher) in
6018 order to make tracer effective.
6020 @item max-cse-path-length
6022 Maximum number of basic blocks on path that cse considers. The default is 10.
6025 The maximum instructions CSE process before flushing. The default is 1000.
6027 @item global-var-threshold
6029 Counts the number of function calls (@var{n}) and the number of
6030 call-clobbered variables (@var{v}). If @var{n}x@var{v} is larger than this limit, a
6031 single artificial variable will be created to represent all the
6032 call-clobbered variables at function call sites. This artificial
6033 variable will then be made to alias every call-clobbered variable.
6034 (done as @code{int * size_t} on the host machine; beware overflow).
6036 @item max-aliased-vops
6038 Maximum number of virtual operands allowed to represent aliases
6039 before triggering the alias grouping heuristic. Alias grouping
6040 reduces compile times and memory consumption needed for aliasing at
6041 the expense of precision loss in alias information.
6043 @item ggc-min-expand
6045 GCC uses a garbage collector to manage its own memory allocation. This
6046 parameter specifies the minimum percentage by which the garbage
6047 collector's heap should be allowed to expand between collections.
6048 Tuning this may improve compilation speed; it has no effect on code
6051 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
6052 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
6053 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
6054 GCC is not able to calculate RAM on a particular platform, the lower
6055 bound of 30% is used. Setting this parameter and
6056 @option{ggc-min-heapsize} to zero causes a full collection to occur at
6057 every opportunity. This is extremely slow, but can be useful for
6060 @item ggc-min-heapsize
6062 Minimum size of the garbage collector's heap before it begins bothering
6063 to collect garbage. The first collection occurs after the heap expands
6064 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
6065 tuning this may improve compilation speed, and has no effect on code
6068 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
6069 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
6070 with a lower bound of 4096 (four megabytes) and an upper bound of
6071 131072 (128 megabytes). If GCC is not able to calculate RAM on a
6072 particular platform, the lower bound is used. Setting this parameter
6073 very large effectively disables garbage collection. Setting this
6074 parameter and @option{ggc-min-expand} to zero causes a full collection
6075 to occur at every opportunity.
6077 @item max-reload-search-insns
6078 The maximum number of instruction reload should look backward for equivalent
6079 register. Increasing values mean more aggressive optimization, making the
6080 compile time increase with probably slightly better performance. The default
6083 @item max-cselib-memory-location
6084 The maximum number of memory locations cselib should take into acount.
6085 Increasing values mean more aggressive optimization, making the compile time
6086 increase with probably slightly better performance. The default value is 500.
6088 @item max-flow-memory-location
6089 Similar as @option{max-cselib-memory-location} but for dataflow liveness.
6090 The default value is 100.
6092 @item reorder-blocks-duplicate
6093 @itemx reorder-blocks-duplicate-feedback
6095 Used by basic block reordering pass to decide whether to use unconditional
6096 branch or duplicate the code on its destination. Code is duplicated when its
6097 estimated size is smaller than this value multiplied by the estimated size of
6098 unconditional jump in the hot spots of the program.
6100 The @option{reorder-block-duplicate-feedback} is used only when profile
6101 feedback is available and may be set to higher values than
6102 @option{reorder-block-duplicate} since information about the hot spots is more
6105 @item max-sched-region-blocks
6106 The maximum number of blocks in a region to be considered for
6107 interblock scheduling. The default value is 10.
6109 @item max-sched-region-insns
6110 The maximum number of insns in a region to be considered for
6111 interblock scheduling. The default value is 100.
6113 @item min-sched-prob
6114 The minimum probability of reaching a source block for interblock
6115 speculative scheduling. The default value is 40.
6117 @item max-last-value-rtl
6119 The maximum size measured as number of RTLs that can be recorded in an expression
6120 in combiner for a pseudo register as last known value of that register. The default
6123 @item integer-share-limit
6124 Small integer constants can use a shared data structure, reducing the
6125 compiler's memory usage and increasing its speed. This sets the maximum
6126 value of a shared integer constant's. The default value is 256.
6128 @item min-virtual-mappings
6129 Specifies the minimum number of virtual mappings in the incremental
6130 SSA updater that should be registered to trigger the virtual mappings
6131 heuristic defined by virtual-mappings-ratio. The default value is
6134 @item virtual-mappings-ratio
6135 If the number of virtual mappings is virtual-mappings-ratio bigger
6136 than the number of virtual symbols to be updated, then the incremental
6137 SSA updater switches to a full update for those symbols. The default
6140 @item ssp-buffer-size
6141 The minimum size of buffers (i.e. arrays) that will receive stack smashing
6142 protection when @option{-fstack-protection} is used.
6144 @item max-jump-thread-duplication-stmts
6145 Maximum number of statements allowed in a block that needs to be
6146 duplicated when threading jumps.
6150 @node Preprocessor Options
6151 @section Options Controlling the Preprocessor
6152 @cindex preprocessor options
6153 @cindex options, preprocessor
6155 These options control the C preprocessor, which is run on each C source
6156 file before actual compilation.
6158 If you use the @option{-E} option, nothing is done except preprocessing.
6159 Some of these options make sense only together with @option{-E} because
6160 they cause the preprocessor output to be unsuitable for actual
6165 You can use @option{-Wp,@var{option}} to bypass the compiler driver
6166 and pass @var{option} directly through to the preprocessor. If
6167 @var{option} contains commas, it is split into multiple options at the
6168 commas. However, many options are modified, translated or interpreted
6169 by the compiler driver before being passed to the preprocessor, and
6170 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
6171 interface is undocumented and subject to change, so whenever possible
6172 you should avoid using @option{-Wp} and let the driver handle the
6175 @item -Xpreprocessor @var{option}
6176 @opindex preprocessor
6177 Pass @var{option} as an option to the preprocessor. You can use this to
6178 supply system-specific preprocessor options which GCC does not know how to
6181 If you want to pass an option that takes an argument, you must use
6182 @option{-Xpreprocessor} twice, once for the option and once for the argument.
6185 @include cppopts.texi
6187 @node Assembler Options
6188 @section Passing Options to the Assembler
6190 @c prevent bad page break with this line
6191 You can pass options to the assembler.
6194 @item -Wa,@var{option}
6196 Pass @var{option} as an option to the assembler. If @var{option}
6197 contains commas, it is split into multiple options at the commas.
6199 @item -Xassembler @var{option}
6201 Pass @var{option} as an option to the assembler. You can use this to
6202 supply system-specific assembler options which GCC does not know how to
6205 If you want to pass an option that takes an argument, you must use
6206 @option{-Xassembler} twice, once for the option and once for the argument.
6211 @section Options for Linking
6212 @cindex link options
6213 @cindex options, linking
6215 These options come into play when the compiler links object files into
6216 an executable output file. They are meaningless if the compiler is
6217 not doing a link step.
6221 @item @var{object-file-name}
6222 A file name that does not end in a special recognized suffix is
6223 considered to name an object file or library. (Object files are
6224 distinguished from libraries by the linker according to the file
6225 contents.) If linking is done, these object files are used as input
6234 If any of these options is used, then the linker is not run, and
6235 object file names should not be used as arguments. @xref{Overall
6239 @item -l@var{library}
6240 @itemx -l @var{library}
6242 Search the library named @var{library} when linking. (The second
6243 alternative with the library as a separate argument is only for
6244 POSIX compliance and is not recommended.)
6246 It makes a difference where in the command you write this option; the
6247 linker searches and processes libraries and object files in the order they
6248 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
6249 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
6250 to functions in @samp{z}, those functions may not be loaded.
6252 The linker searches a standard list of directories for the library,
6253 which is actually a file named @file{lib@var{library}.a}. The linker
6254 then uses this file as if it had been specified precisely by name.
6256 The directories searched include several standard system directories
6257 plus any that you specify with @option{-L}.
6259 Normally the files found this way are library files---archive files
6260 whose members are object files. The linker handles an archive file by
6261 scanning through it for members which define symbols that have so far
6262 been referenced but not defined. But if the file that is found is an
6263 ordinary object file, it is linked in the usual fashion. The only
6264 difference between using an @option{-l} option and specifying a file name
6265 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
6266 and searches several directories.
6270 You need this special case of the @option{-l} option in order to
6271 link an Objective-C or Objective-C++ program.
6274 @opindex nostartfiles
6275 Do not use the standard system startup files when linking.
6276 The standard system libraries are used normally, unless @option{-nostdlib}
6277 or @option{-nodefaultlibs} is used.
6279 @item -nodefaultlibs
6280 @opindex nodefaultlibs
6281 Do not use the standard system libraries when linking.
6282 Only the libraries you specify will be passed to the linker.
6283 The standard startup files are used normally, unless @option{-nostartfiles}
6284 is used. The compiler may generate calls to @code{memcmp},
6285 @code{memset}, @code{memcpy} and @code{memmove}.
6286 These entries are usually resolved by entries in
6287 libc. These entry points should be supplied through some other
6288 mechanism when this option is specified.
6292 Do not use the standard system startup files or libraries when linking.
6293 No startup files and only the libraries you specify will be passed to
6294 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
6295 @code{memcpy} and @code{memmove}.
6296 These entries are usually resolved by entries in
6297 libc. These entry points should be supplied through some other
6298 mechanism when this option is specified.
6300 @cindex @option{-lgcc}, use with @option{-nostdlib}
6301 @cindex @option{-nostdlib} and unresolved references
6302 @cindex unresolved references and @option{-nostdlib}
6303 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
6304 @cindex @option{-nodefaultlibs} and unresolved references
6305 @cindex unresolved references and @option{-nodefaultlibs}
6306 One of the standard libraries bypassed by @option{-nostdlib} and
6307 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
6308 that GCC uses to overcome shortcomings of particular machines, or special
6309 needs for some languages.
6310 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
6311 Collection (GCC) Internals},
6312 for more discussion of @file{libgcc.a}.)
6313 In most cases, you need @file{libgcc.a} even when you want to avoid
6314 other standard libraries. In other words, when you specify @option{-nostdlib}
6315 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
6316 This ensures that you have no unresolved references to internal GCC
6317 library subroutines. (For example, @samp{__main}, used to ensure C++
6318 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
6319 GNU Compiler Collection (GCC) Internals}.)
6323 Produce a position independent executable on targets which support it.
6324 For predictable results, you must also specify the same set of options
6325 that were used to generate code (@option{-fpie}, @option{-fPIE},
6326 or model suboptions) when you specify this option.
6330 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
6331 that support it. This instructs the linker to add all symbols, not
6332 only used ones, to the dynamic symbol table. This option is needed
6333 for some uses of @code{dlopen} or to allow obtaining backtraces
6334 from within a program.
6338 Remove all symbol table and relocation information from the executable.
6342 On systems that support dynamic linking, this prevents linking with the shared
6343 libraries. On other systems, this option has no effect.
6347 Produce a shared object which can then be linked with other objects to
6348 form an executable. Not all systems support this option. For predictable
6349 results, you must also specify the same set of options that were used to
6350 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
6351 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
6352 needs to build supplementary stub code for constructors to work. On
6353 multi-libbed systems, @samp{gcc -shared} must select the correct support
6354 libraries to link against. Failing to supply the correct flags may lead
6355 to subtle defects. Supplying them in cases where they are not necessary
6358 @item -shared-libgcc
6359 @itemx -static-libgcc
6360 @opindex shared-libgcc
6361 @opindex static-libgcc
6362 On systems that provide @file{libgcc} as a shared library, these options
6363 force the use of either the shared or static version respectively.
6364 If no shared version of @file{libgcc} was built when the compiler was
6365 configured, these options have no effect.
6367 There are several situations in which an application should use the
6368 shared @file{libgcc} instead of the static version. The most common
6369 of these is when the application wishes to throw and catch exceptions
6370 across different shared libraries. In that case, each of the libraries
6371 as well as the application itself should use the shared @file{libgcc}.
6373 Therefore, the G++ and GCJ drivers automatically add
6374 @option{-shared-libgcc} whenever you build a shared library or a main
6375 executable, because C++ and Java programs typically use exceptions, so
6376 this is the right thing to do.
6378 If, instead, you use the GCC driver to create shared libraries, you may
6379 find that they will not always be linked with the shared @file{libgcc}.
6380 If GCC finds, at its configuration time, that you have a non-GNU linker
6381 or a GNU linker that does not support option @option{--eh-frame-hdr},
6382 it will link the shared version of @file{libgcc} into shared libraries
6383 by default. Otherwise, it will take advantage of the linker and optimize
6384 away the linking with the shared version of @file{libgcc}, linking with
6385 the static version of libgcc by default. This allows exceptions to
6386 propagate through such shared libraries, without incurring relocation
6387 costs at library load time.
6389 However, if a library or main executable is supposed to throw or catch
6390 exceptions, you must link it using the G++ or GCJ driver, as appropriate
6391 for the languages used in the program, or using the option
6392 @option{-shared-libgcc}, such that it is linked with the shared
6397 Bind references to global symbols when building a shared object. Warn
6398 about any unresolved references (unless overridden by the link editor
6399 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
6402 @item -Xlinker @var{option}
6404 Pass @var{option} as an option to the linker. You can use this to
6405 supply system-specific linker options which GCC does not know how to
6408 If you want to pass an option that takes an argument, you must use
6409 @option{-Xlinker} twice, once for the option and once for the argument.
6410 For example, to pass @option{-assert definitions}, you must write
6411 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
6412 @option{-Xlinker "-assert definitions"}, because this passes the entire
6413 string as a single argument, which is not what the linker expects.
6415 @item -Wl,@var{option}
6417 Pass @var{option} as an option to the linker. If @var{option} contains
6418 commas, it is split into multiple options at the commas.
6420 @item -u @var{symbol}
6422 Pretend the symbol @var{symbol} is undefined, to force linking of
6423 library modules to define it. You can use @option{-u} multiple times with
6424 different symbols to force loading of additional library modules.
6427 @node Directory Options
6428 @section Options for Directory Search
6429 @cindex directory options
6430 @cindex options, directory search
6433 These options specify directories to search for header files, for
6434 libraries and for parts of the compiler:
6439 Add the directory @var{dir} to the head of the list of directories to be
6440 searched for header files. This can be used to override a system header
6441 file, substituting your own version, since these directories are
6442 searched before the system header file directories. However, you should
6443 not use this option to add directories that contain vendor-supplied
6444 system header files (use @option{-isystem} for that). If you use more than
6445 one @option{-I} option, the directories are scanned in left-to-right
6446 order; the standard system directories come after.
6448 If a standard system include directory, or a directory specified with
6449 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
6450 option will be ignored. The directory will still be searched but as a
6451 system directory at its normal position in the system include chain.
6452 This is to ensure that GCC's procedure to fix buggy system headers and
6453 the ordering for the include_next directive are not inadvertently changed.
6454 If you really need to change the search order for system directories,
6455 use the @option{-nostdinc} and/or @option{-isystem} options.
6457 @item -iquote@var{dir}
6459 Add the directory @var{dir} to the head of the list of directories to
6460 be searched for header files only for the case of @samp{#include
6461 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
6462 otherwise just like @option{-I}.
6466 Add directory @var{dir} to the list of directories to be searched
6469 @item -B@var{prefix}
6471 This option specifies where to find the executables, libraries,
6472 include files, and data files of the compiler itself.
6474 The compiler driver program runs one or more of the subprograms
6475 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
6476 @var{prefix} as a prefix for each program it tries to run, both with and
6477 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
6479 For each subprogram to be run, the compiler driver first tries the
6480 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
6481 was not specified, the driver tries two standard prefixes, which are
6482 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
6483 those results in a file name that is found, the unmodified program
6484 name is searched for using the directories specified in your
6485 @env{PATH} environment variable.
6487 The compiler will check to see if the path provided by the @option{-B}
6488 refers to a directory, and if necessary it will add a directory
6489 separator character at the end of the path.
6491 @option{-B} prefixes that effectively specify directory names also apply
6492 to libraries in the linker, because the compiler translates these
6493 options into @option{-L} options for the linker. They also apply to
6494 includes files in the preprocessor, because the compiler translates these
6495 options into @option{-isystem} options for the preprocessor. In this case,
6496 the compiler appends @samp{include} to the prefix.
6498 The run-time support file @file{libgcc.a} can also be searched for using
6499 the @option{-B} prefix, if needed. If it is not found there, the two
6500 standard prefixes above are tried, and that is all. The file is left
6501 out of the link if it is not found by those means.
6503 Another way to specify a prefix much like the @option{-B} prefix is to use
6504 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
6507 As a special kludge, if the path provided by @option{-B} is
6508 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
6509 9, then it will be replaced by @file{[dir/]include}. This is to help
6510 with boot-strapping the compiler.
6512 @item -specs=@var{file}
6514 Process @var{file} after the compiler reads in the standard @file{specs}
6515 file, in order to override the defaults that the @file{gcc} driver
6516 program uses when determining what switches to pass to @file{cc1},
6517 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
6518 @option{-specs=@var{file}} can be specified on the command line, and they
6519 are processed in order, from left to right.
6521 @item --sysroot=@var{dir}
6523 Use @var{dir} as the logical root directory for headers and libraries.
6524 For example, if the compiler would normally search for headers in
6525 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
6526 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
6528 If you use both this option and the @option{-isysroot} option, then
6529 the @option{--sysroot} option will apply to libraries, but the
6530 @option{-isysroot} option will apply to header files.
6532 The GNU linker (beginning with version 2.16) has the necessary support
6533 for this option. If your linker does not support this option, the
6534 header file aspect of @option{--sysroot} will still work, but the
6535 library aspect will not.
6539 This option has been deprecated. Please use @option{-iquote} instead for
6540 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
6541 Any directories you specify with @option{-I} options before the @option{-I-}
6542 option are searched only for the case of @samp{#include "@var{file}"};
6543 they are not searched for @samp{#include <@var{file}>}.
6545 If additional directories are specified with @option{-I} options after
6546 the @option{-I-}, these directories are searched for all @samp{#include}
6547 directives. (Ordinarily @emph{all} @option{-I} directories are used
6550 In addition, the @option{-I-} option inhibits the use of the current
6551 directory (where the current input file came from) as the first search
6552 directory for @samp{#include "@var{file}"}. There is no way to
6553 override this effect of @option{-I-}. With @option{-I.} you can specify
6554 searching the directory which was current when the compiler was
6555 invoked. That is not exactly the same as what the preprocessor does
6556 by default, but it is often satisfactory.
6558 @option{-I-} does not inhibit the use of the standard system directories
6559 for header files. Thus, @option{-I-} and @option{-nostdinc} are
6566 @section Specifying subprocesses and the switches to pass to them
6569 @command{gcc} is a driver program. It performs its job by invoking a
6570 sequence of other programs to do the work of compiling, assembling and
6571 linking. GCC interprets its command-line parameters and uses these to
6572 deduce which programs it should invoke, and which command-line options
6573 it ought to place on their command lines. This behavior is controlled
6574 by @dfn{spec strings}. In most cases there is one spec string for each
6575 program that GCC can invoke, but a few programs have multiple spec
6576 strings to control their behavior. The spec strings built into GCC can
6577 be overridden by using the @option{-specs=} command-line switch to specify
6580 @dfn{Spec files} are plaintext files that are used to construct spec
6581 strings. They consist of a sequence of directives separated by blank
6582 lines. The type of directive is determined by the first non-whitespace
6583 character on the line and it can be one of the following:
6586 @item %@var{command}
6587 Issues a @var{command} to the spec file processor. The commands that can
6591 @item %include <@var{file}>
6593 Search for @var{file} and insert its text at the current point in the
6596 @item %include_noerr <@var{file}>
6597 @cindex %include_noerr
6598 Just like @samp{%include}, but do not generate an error message if the include
6599 file cannot be found.
6601 @item %rename @var{old_name} @var{new_name}
6603 Rename the spec string @var{old_name} to @var{new_name}.
6607 @item *[@var{spec_name}]:
6608 This tells the compiler to create, override or delete the named spec
6609 string. All lines after this directive up to the next directive or
6610 blank line are considered to be the text for the spec string. If this
6611 results in an empty string then the spec will be deleted. (Or, if the
6612 spec did not exist, then nothing will happened.) Otherwise, if the spec
6613 does not currently exist a new spec will be created. If the spec does
6614 exist then its contents will be overridden by the text of this
6615 directive, unless the first character of that text is the @samp{+}
6616 character, in which case the text will be appended to the spec.
6618 @item [@var{suffix}]:
6619 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
6620 and up to the next directive or blank line are considered to make up the
6621 spec string for the indicated suffix. When the compiler encounters an
6622 input file with the named suffix, it will processes the spec string in
6623 order to work out how to compile that file. For example:
6630 This says that any input file whose name ends in @samp{.ZZ} should be
6631 passed to the program @samp{z-compile}, which should be invoked with the
6632 command-line switch @option{-input} and with the result of performing the
6633 @samp{%i} substitution. (See below.)
6635 As an alternative to providing a spec string, the text that follows a
6636 suffix directive can be one of the following:
6639 @item @@@var{language}
6640 This says that the suffix is an alias for a known @var{language}. This is
6641 similar to using the @option{-x} command-line switch to GCC to specify a
6642 language explicitly. For example:
6649 Says that .ZZ files are, in fact, C++ source files.
6652 This causes an error messages saying:
6655 @var{name} compiler not installed on this system.
6659 GCC already has an extensive list of suffixes built into it.
6660 This directive will add an entry to the end of the list of suffixes, but
6661 since the list is searched from the end backwards, it is effectively
6662 possible to override earlier entries using this technique.
6666 GCC has the following spec strings built into it. Spec files can
6667 override these strings or create their own. Note that individual
6668 targets can also add their own spec strings to this list.
6671 asm Options to pass to the assembler
6672 asm_final Options to pass to the assembler post-processor
6673 cpp Options to pass to the C preprocessor
6674 cc1 Options to pass to the C compiler
6675 cc1plus Options to pass to the C++ compiler
6676 endfile Object files to include at the end of the link
6677 link Options to pass to the linker
6678 lib Libraries to include on the command line to the linker
6679 libgcc Decides which GCC support library to pass to the linker
6680 linker Sets the name of the linker
6681 predefines Defines to be passed to the C preprocessor
6682 signed_char Defines to pass to CPP to say whether @code{char} is signed
6684 startfile Object files to include at the start of the link
6687 Here is a small example of a spec file:
6693 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
6696 This example renames the spec called @samp{lib} to @samp{old_lib} and
6697 then overrides the previous definition of @samp{lib} with a new one.
6698 The new definition adds in some extra command-line options before
6699 including the text of the old definition.
6701 @dfn{Spec strings} are a list of command-line options to be passed to their
6702 corresponding program. In addition, the spec strings can contain
6703 @samp{%}-prefixed sequences to substitute variable text or to
6704 conditionally insert text into the command line. Using these constructs
6705 it is possible to generate quite complex command lines.
6707 Here is a table of all defined @samp{%}-sequences for spec
6708 strings. Note that spaces are not generated automatically around the
6709 results of expanding these sequences. Therefore you can concatenate them
6710 together or combine them with constant text in a single argument.
6714 Substitute one @samp{%} into the program name or argument.
6717 Substitute the name of the input file being processed.
6720 Substitute the basename of the input file being processed.
6721 This is the substring up to (and not including) the last period
6722 and not including the directory.
6725 This is the same as @samp{%b}, but include the file suffix (text after
6729 Marks the argument containing or following the @samp{%d} as a
6730 temporary file name, so that that file will be deleted if GCC exits
6731 successfully. Unlike @samp{%g}, this contributes no text to the
6734 @item %g@var{suffix}
6735 Substitute a file name that has suffix @var{suffix} and is chosen
6736 once per compilation, and mark the argument in the same way as
6737 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
6738 name is now chosen in a way that is hard to predict even when previously
6739 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
6740 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
6741 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
6742 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
6743 was simply substituted with a file name chosen once per compilation,
6744 without regard to any appended suffix (which was therefore treated
6745 just like ordinary text), making such attacks more likely to succeed.
6747 @item %u@var{suffix}
6748 Like @samp{%g}, but generates a new temporary file name even if
6749 @samp{%u@var{suffix}} was already seen.
6751 @item %U@var{suffix}
6752 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
6753 new one if there is no such last file name. In the absence of any
6754 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
6755 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
6756 would involve the generation of two distinct file names, one
6757 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
6758 simply substituted with a file name chosen for the previous @samp{%u},
6759 without regard to any appended suffix.
6761 @item %j@var{suffix}
6762 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
6763 writable, and if save-temps is off; otherwise, substitute the name
6764 of a temporary file, just like @samp{%u}. This temporary file is not
6765 meant for communication between processes, but rather as a junk
6768 @item %|@var{suffix}
6769 @itemx %m@var{suffix}
6770 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
6771 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
6772 all. These are the two most common ways to instruct a program that it
6773 should read from standard input or write to standard output. If you
6774 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
6775 construct: see for example @file{f/lang-specs.h}.
6777 @item %.@var{SUFFIX}
6778 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
6779 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
6780 terminated by the next space or %.
6783 Marks the argument containing or following the @samp{%w} as the
6784 designated output file of this compilation. This puts the argument
6785 into the sequence of arguments that @samp{%o} will substitute later.
6788 Substitutes the names of all the output files, with spaces
6789 automatically placed around them. You should write spaces
6790 around the @samp{%o} as well or the results are undefined.
6791 @samp{%o} is for use in the specs for running the linker.
6792 Input files whose names have no recognized suffix are not compiled
6793 at all, but they are included among the output files, so they will
6797 Substitutes the suffix for object files. Note that this is
6798 handled specially when it immediately follows @samp{%g, %u, or %U},
6799 because of the need for those to form complete file names. The
6800 handling is such that @samp{%O} is treated exactly as if it had already
6801 been substituted, except that @samp{%g, %u, and %U} do not currently
6802 support additional @var{suffix} characters following @samp{%O} as they would
6803 following, for example, @samp{.o}.
6806 Substitutes the standard macro predefinitions for the
6807 current target machine. Use this when running @code{cpp}.
6810 Like @samp{%p}, but puts @samp{__} before and after the name of each
6811 predefined macro, except for macros that start with @samp{__} or with
6812 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
6816 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
6817 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), and
6818 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
6822 Current argument is the name of a library or startup file of some sort.
6823 Search for that file in a standard list of directories and substitute
6824 the full name found.
6827 Print @var{str} as an error message. @var{str} is terminated by a newline.
6828 Use this when inconsistent options are detected.
6831 Substitute the contents of spec string @var{name} at this point.
6834 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
6836 @item %x@{@var{option}@}
6837 Accumulate an option for @samp{%X}.
6840 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
6844 Output the accumulated assembler options specified by @option{-Wa}.
6847 Output the accumulated preprocessor options specified by @option{-Wp}.
6850 Process the @code{asm} spec. This is used to compute the
6851 switches to be passed to the assembler.
6854 Process the @code{asm_final} spec. This is a spec string for
6855 passing switches to an assembler post-processor, if such a program is
6859 Process the @code{link} spec. This is the spec for computing the
6860 command line passed to the linker. Typically it will make use of the
6861 @samp{%L %G %S %D and %E} sequences.
6864 Dump out a @option{-L} option for each directory that GCC believes might
6865 contain startup files. If the target supports multilibs then the
6866 current multilib directory will be prepended to each of these paths.
6869 Process the @code{lib} spec. This is a spec string for deciding which
6870 libraries should be included on the command line to the linker.
6873 Process the @code{libgcc} spec. This is a spec string for deciding
6874 which GCC support library should be included on the command line to the linker.
6877 Process the @code{startfile} spec. This is a spec for deciding which
6878 object files should be the first ones passed to the linker. Typically
6879 this might be a file named @file{crt0.o}.
6882 Process the @code{endfile} spec. This is a spec string that specifies
6883 the last object files that will be passed to the linker.
6886 Process the @code{cpp} spec. This is used to construct the arguments
6887 to be passed to the C preprocessor.
6890 Process the @code{cc1} spec. This is used to construct the options to be
6891 passed to the actual C compiler (@samp{cc1}).
6894 Process the @code{cc1plus} spec. This is used to construct the options to be
6895 passed to the actual C++ compiler (@samp{cc1plus}).
6898 Substitute the variable part of a matched option. See below.
6899 Note that each comma in the substituted string is replaced by
6903 Remove all occurrences of @code{-S} from the command line. Note---this
6904 command is position dependent. @samp{%} commands in the spec string
6905 before this one will see @code{-S}, @samp{%} commands in the spec string
6906 after this one will not.
6908 @item %:@var{function}(@var{args})
6909 Call the named function @var{function}, passing it @var{args}.
6910 @var{args} is first processed as a nested spec string, then split
6911 into an argument vector in the usual fashion. The function returns
6912 a string which is processed as if it had appeared literally as part
6913 of the current spec.
6915 The following built-in spec functions are provided:
6918 @item @code{if-exists}
6919 The @code{if-exists} spec function takes one argument, an absolute
6920 pathname to a file. If the file exists, @code{if-exists} returns the
6921 pathname. Here is a small example of its usage:
6925 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
6928 @item @code{if-exists-else}
6929 The @code{if-exists-else} spec function is similar to the @code{if-exists}
6930 spec function, except that it takes two arguments. The first argument is
6931 an absolute pathname to a file. If the file exists, @code{if-exists-else}
6932 returns the pathname. If it does not exist, it returns the second argument.
6933 This way, @code{if-exists-else} can be used to select one file or another,
6934 based on the existence of the first. Here is a small example of its usage:
6938 crt0%O%s %:if-exists(crti%O%s) \
6939 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
6942 @item @code{replace-outfile}
6943 The @code{replace-outfile} spec function takes two arguments. It looks for the
6944 first argument in the outfiles array and replaces it with the second argument. Here
6945 is a small example of its usage:
6948 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
6954 Substitutes the @code{-S} switch, if that switch was given to GCC@.
6955 If that switch was not specified, this substitutes nothing. Note that
6956 the leading dash is omitted when specifying this option, and it is
6957 automatically inserted if the substitution is performed. Thus the spec
6958 string @samp{%@{foo@}} would match the command-line option @option{-foo}
6959 and would output the command line option @option{-foo}.
6961 @item %W@{@code{S}@}
6962 Like %@{@code{S}@} but mark last argument supplied within as a file to be
6965 @item %@{@code{S}*@}
6966 Substitutes all the switches specified to GCC whose names start
6967 with @code{-S}, but which also take an argument. This is used for
6968 switches like @option{-o}, @option{-D}, @option{-I}, etc.
6969 GCC considers @option{-o foo} as being
6970 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
6971 text, including the space. Thus two arguments would be generated.
6973 @item %@{@code{S}*&@code{T}*@}
6974 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
6975 (the order of @code{S} and @code{T} in the spec is not significant).
6976 There can be any number of ampersand-separated variables; for each the
6977 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
6979 @item %@{@code{S}:@code{X}@}
6980 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
6982 @item %@{!@code{S}:@code{X}@}
6983 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
6985 @item %@{@code{S}*:@code{X}@}
6986 Substitutes @code{X} if one or more switches whose names start with
6987 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
6988 once, no matter how many such switches appeared. However, if @code{%*}
6989 appears somewhere in @code{X}, then @code{X} will be substituted once
6990 for each matching switch, with the @code{%*} replaced by the part of
6991 that switch that matched the @code{*}.
6993 @item %@{.@code{S}:@code{X}@}
6994 Substitutes @code{X}, if processing a file with suffix @code{S}.
6996 @item %@{!.@code{S}:@code{X}@}
6997 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
6999 @item %@{@code{S}|@code{P}:@code{X}@}
7000 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
7001 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
7002 although they have a stronger binding than the @samp{|}. If @code{%*}
7003 appears in @code{X}, all of the alternatives must be starred, and only
7004 the first matching alternative is substituted.
7006 For example, a spec string like this:
7009 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
7012 will output the following command-line options from the following input
7013 command-line options:
7018 -d fred.c -foo -baz -boggle
7019 -d jim.d -bar -baz -boggle
7022 @item %@{S:X; T:Y; :D@}
7024 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
7025 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
7026 be as many clauses as you need. This may be combined with @code{.},
7027 @code{!}, @code{|}, and @code{*} as needed.
7032 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
7033 construct may contain other nested @samp{%} constructs or spaces, or
7034 even newlines. They are processed as usual, as described above.
7035 Trailing white space in @code{X} is ignored. White space may also
7036 appear anywhere on the left side of the colon in these constructs,
7037 except between @code{.} or @code{*} and the corresponding word.
7039 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
7040 handled specifically in these constructs. If another value of
7041 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
7042 @option{-W} switch is found later in the command line, the earlier
7043 switch value is ignored, except with @{@code{S}*@} where @code{S} is
7044 just one letter, which passes all matching options.
7046 The character @samp{|} at the beginning of the predicate text is used to
7047 indicate that a command should be piped to the following command, but
7048 only if @option{-pipe} is specified.
7050 It is built into GCC which switches take arguments and which do not.
7051 (You might think it would be useful to generalize this to allow each
7052 compiler's spec to say which switches take arguments. But this cannot
7053 be done in a consistent fashion. GCC cannot even decide which input
7054 files have been specified without knowing which switches take arguments,
7055 and it must know which input files to compile in order to tell which
7058 GCC also knows implicitly that arguments starting in @option{-l} are to be
7059 treated as compiler output files, and passed to the linker in their
7060 proper position among the other output files.
7062 @c man begin OPTIONS
7064 @node Target Options
7065 @section Specifying Target Machine and Compiler Version
7066 @cindex target options
7067 @cindex cross compiling
7068 @cindex specifying machine version
7069 @cindex specifying compiler version and target machine
7070 @cindex compiler version, specifying
7071 @cindex target machine, specifying
7073 The usual way to run GCC is to run the executable called @file{gcc}, or
7074 @file{<machine>-gcc} when cross-compiling, or
7075 @file{<machine>-gcc-<version>} to run a version other than the one that
7076 was installed last. Sometimes this is inconvenient, so GCC provides
7077 options that will switch to another cross-compiler or version.
7080 @item -b @var{machine}
7082 The argument @var{machine} specifies the target machine for compilation.
7084 The value to use for @var{machine} is the same as was specified as the
7085 machine type when configuring GCC as a cross-compiler. For
7086 example, if a cross-compiler was configured with @samp{configure
7087 arm-elf}, meaning to compile for an arm processor with elf binaries,
7088 then you would specify @option{-b arm-elf} to run that cross compiler.
7089 Because there are other options beginning with @option{-b}, the
7090 configuration must contain a hyphen.
7092 @item -V @var{version}
7094 The argument @var{version} specifies which version of GCC to run.
7095 This is useful when multiple versions are installed. For example,
7096 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
7099 The @option{-V} and @option{-b} options work by running the
7100 @file{<machine>-gcc-<version>} executable, so there's no real reason to
7101 use them if you can just run that directly.
7103 @node Submodel Options
7104 @section Hardware Models and Configurations
7105 @cindex submodel options
7106 @cindex specifying hardware config
7107 @cindex hardware models and configurations, specifying
7108 @cindex machine dependent options
7110 Earlier we discussed the standard option @option{-b} which chooses among
7111 different installed compilers for completely different target
7112 machines, such as VAX vs.@: 68000 vs.@: 80386.
7114 In addition, each of these target machine types can have its own
7115 special options, starting with @samp{-m}, to choose among various
7116 hardware models or configurations---for example, 68010 vs 68020,
7117 floating coprocessor or none. A single installed version of the
7118 compiler can compile for any model or configuration, according to the
7121 Some configurations of the compiler also support additional special
7122 options, usually for compatibility with other compilers on the same
7125 @c This list is ordered alphanumerically by subsection name.
7126 @c It should be the same order and spelling as these options are listed
7127 @c in Machine Dependent Options
7133 * Blackfin Options::
7137 * DEC Alpha Options::
7138 * DEC Alpha/VMS Options::
7142 * i386 and x86-64 Options::
7155 * RS/6000 and PowerPC Options::
7156 * S/390 and zSeries Options::
7159 * System V Options::
7160 * TMS320C3x/C4x Options::
7164 * Xstormy16 Options::
7170 @subsection ARC Options
7173 These options are defined for ARC implementations:
7178 Compile code for little endian mode. This is the default.
7182 Compile code for big endian mode.
7185 @opindex mmangle-cpu
7186 Prepend the name of the cpu to all public symbol names.
7187 In multiple-processor systems, there are many ARC variants with different
7188 instruction and register set characteristics. This flag prevents code
7189 compiled for one cpu to be linked with code compiled for another.
7190 No facility exists for handling variants that are ``almost identical''.
7191 This is an all or nothing option.
7193 @item -mcpu=@var{cpu}
7195 Compile code for ARC variant @var{cpu}.
7196 Which variants are supported depend on the configuration.
7197 All variants support @option{-mcpu=base}, this is the default.
7199 @item -mtext=@var{text-section}
7200 @itemx -mdata=@var{data-section}
7201 @itemx -mrodata=@var{readonly-data-section}
7205 Put functions, data, and readonly data in @var{text-section},
7206 @var{data-section}, and @var{readonly-data-section} respectively
7207 by default. This can be overridden with the @code{section} attribute.
7208 @xref{Variable Attributes}.
7213 @subsection ARM Options
7216 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
7220 @item -mabi=@var{name}
7222 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
7223 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
7226 @opindex mapcs-frame
7227 Generate a stack frame that is compliant with the ARM Procedure Call
7228 Standard for all functions, even if this is not strictly necessary for
7229 correct execution of the code. Specifying @option{-fomit-frame-pointer}
7230 with this option will cause the stack frames not to be generated for
7231 leaf functions. The default is @option{-mno-apcs-frame}.
7235 This is a synonym for @option{-mapcs-frame}.
7238 @c not currently implemented
7239 @item -mapcs-stack-check
7240 @opindex mapcs-stack-check
7241 Generate code to check the amount of stack space available upon entry to
7242 every function (that actually uses some stack space). If there is
7243 insufficient space available then either the function
7244 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
7245 called, depending upon the amount of stack space required. The run time
7246 system is required to provide these functions. The default is
7247 @option{-mno-apcs-stack-check}, since this produces smaller code.
7249 @c not currently implemented
7251 @opindex mapcs-float
7252 Pass floating point arguments using the float point registers. This is
7253 one of the variants of the APCS@. This option is recommended if the
7254 target hardware has a floating point unit or if a lot of floating point
7255 arithmetic is going to be performed by the code. The default is
7256 @option{-mno-apcs-float}, since integer only code is slightly increased in
7257 size if @option{-mapcs-float} is used.
7259 @c not currently implemented
7260 @item -mapcs-reentrant
7261 @opindex mapcs-reentrant
7262 Generate reentrant, position independent code. The default is
7263 @option{-mno-apcs-reentrant}.
7266 @item -mthumb-interwork
7267 @opindex mthumb-interwork
7268 Generate code which supports calling between the ARM and Thumb
7269 instruction sets. Without this option the two instruction sets cannot
7270 be reliably used inside one program. The default is
7271 @option{-mno-thumb-interwork}, since slightly larger code is generated
7272 when @option{-mthumb-interwork} is specified.
7274 @item -mno-sched-prolog
7275 @opindex mno-sched-prolog
7276 Prevent the reordering of instructions in the function prolog, or the
7277 merging of those instruction with the instructions in the function's
7278 body. This means that all functions will start with a recognizable set
7279 of instructions (or in fact one of a choice from a small set of
7280 different function prologues), and this information can be used to
7281 locate the start if functions inside an executable piece of code. The
7282 default is @option{-msched-prolog}.
7285 @opindex mhard-float
7286 Generate output containing floating point instructions. This is the
7290 @opindex msoft-float
7291 Generate output containing library calls for floating point.
7292 @strong{Warning:} the requisite libraries are not available for all ARM
7293 targets. Normally the facilities of the machine's usual C compiler are
7294 used, but this cannot be done directly in cross-compilation. You must make
7295 your own arrangements to provide suitable library functions for
7298 @option{-msoft-float} changes the calling convention in the output file;
7299 therefore, it is only useful if you compile @emph{all} of a program with
7300 this option. In particular, you need to compile @file{libgcc.a}, the
7301 library that comes with GCC, with @option{-msoft-float} in order for
7304 @item -mfloat-abi=@var{name}
7306 Specifies which ABI to use for floating point values. Permissible values
7307 are: @samp{soft}, @samp{softfp} and @samp{hard}.
7309 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
7310 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
7311 of floating point instructions, but still uses the soft-float calling
7314 @item -mlittle-endian
7315 @opindex mlittle-endian
7316 Generate code for a processor running in little-endian mode. This is
7317 the default for all standard configurations.
7320 @opindex mbig-endian
7321 Generate code for a processor running in big-endian mode; the default is
7322 to compile code for a little-endian processor.
7324 @item -mwords-little-endian
7325 @opindex mwords-little-endian
7326 This option only applies when generating code for big-endian processors.
7327 Generate code for a little-endian word order but a big-endian byte
7328 order. That is, a byte order of the form @samp{32107654}. Note: this
7329 option should only be used if you require compatibility with code for
7330 big-endian ARM processors generated by versions of the compiler prior to
7333 @item -mcpu=@var{name}
7335 This specifies the name of the target ARM processor. GCC uses this name
7336 to determine what kind of instructions it can emit when generating
7337 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
7338 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
7339 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
7340 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
7341 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
7342 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
7343 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
7344 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
7345 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
7346 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
7347 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
7348 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
7349 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
7350 @samp{arm1176jz-s}, @samp{arm1176jzf-s}, @samp{xscale}, @samp{iwmmxt},
7353 @itemx -mtune=@var{name}
7355 This option is very similar to the @option{-mcpu=} option, except that
7356 instead of specifying the actual target processor type, and hence
7357 restricting which instructions can be used, it specifies that GCC should
7358 tune the performance of the code as if the target were of the type
7359 specified in this option, but still choosing the instructions that it
7360 will generate based on the cpu specified by a @option{-mcpu=} option.
7361 For some ARM implementations better performance can be obtained by using
7364 @item -march=@var{name}
7366 This specifies the name of the target ARM architecture. GCC uses this
7367 name to determine what kind of instructions it can emit when generating
7368 assembly code. This option can be used in conjunction with or instead
7369 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
7370 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
7371 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
7372 @samp{iwmmxt}, @samp{ep9312}.
7374 @item -mfpu=@var{name}
7375 @itemx -mfpe=@var{number}
7376 @itemx -mfp=@var{number}
7380 This specifies what floating point hardware (or hardware emulation) is
7381 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
7382 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
7383 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
7384 with older versions of GCC@.
7386 If @option{-msoft-float} is specified this specifies the format of
7387 floating point values.
7389 @item -mstructure-size-boundary=@var{n}
7390 @opindex mstructure-size-boundary
7391 The size of all structures and unions will be rounded up to a multiple
7392 of the number of bits set by this option. Permissible values are 8, 32
7393 and 64. The default value varies for different toolchains. For the COFF
7394 targeted toolchain the default value is 8. A value of 64 is only allowed
7395 if the underlying ABI supports it.
7397 Specifying the larger number can produce faster, more efficient code, but
7398 can also increase the size of the program. Different values are potentially
7399 incompatible. Code compiled with one value cannot necessarily expect to
7400 work with code or libraries compiled with another value, if they exchange
7401 information using structures or unions.
7403 @item -mabort-on-noreturn
7404 @opindex mabort-on-noreturn
7405 Generate a call to the function @code{abort} at the end of a
7406 @code{noreturn} function. It will be executed if the function tries to
7410 @itemx -mno-long-calls
7411 @opindex mlong-calls
7412 @opindex mno-long-calls
7413 Tells the compiler to perform function calls by first loading the
7414 address of the function into a register and then performing a subroutine
7415 call on this register. This switch is needed if the target function
7416 will lie outside of the 64 megabyte addressing range of the offset based
7417 version of subroutine call instruction.
7419 Even if this switch is enabled, not all function calls will be turned
7420 into long calls. The heuristic is that static functions, functions
7421 which have the @samp{short-call} attribute, functions that are inside
7422 the scope of a @samp{#pragma no_long_calls} directive and functions whose
7423 definitions have already been compiled within the current compilation
7424 unit, will not be turned into long calls. The exception to this rule is
7425 that weak function definitions, functions with the @samp{long-call}
7426 attribute or the @samp{section} attribute, and functions that are within
7427 the scope of a @samp{#pragma long_calls} directive, will always be
7428 turned into long calls.
7430 This feature is not enabled by default. Specifying
7431 @option{-mno-long-calls} will restore the default behavior, as will
7432 placing the function calls within the scope of a @samp{#pragma
7433 long_calls_off} directive. Note these switches have no effect on how
7434 the compiler generates code to handle function calls via function
7437 @item -mnop-fun-dllimport
7438 @opindex mnop-fun-dllimport
7439 Disable support for the @code{dllimport} attribute.
7441 @item -msingle-pic-base
7442 @opindex msingle-pic-base
7443 Treat the register used for PIC addressing as read-only, rather than
7444 loading it in the prologue for each function. The run-time system is
7445 responsible for initializing this register with an appropriate value
7446 before execution begins.
7448 @item -mpic-register=@var{reg}
7449 @opindex mpic-register
7450 Specify the register to be used for PIC addressing. The default is R10
7451 unless stack-checking is enabled, when R9 is used.
7453 @item -mcirrus-fix-invalid-insns
7454 @opindex mcirrus-fix-invalid-insns
7455 @opindex mno-cirrus-fix-invalid-insns
7456 Insert NOPs into the instruction stream to in order to work around
7457 problems with invalid Maverick instruction combinations. This option
7458 is only valid if the @option{-mcpu=ep9312} option has been used to
7459 enable generation of instructions for the Cirrus Maverick floating
7460 point co-processor. This option is not enabled by default, since the
7461 problem is only present in older Maverick implementations. The default
7462 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
7465 @item -mpoke-function-name
7466 @opindex mpoke-function-name
7467 Write the name of each function into the text section, directly
7468 preceding the function prologue. The generated code is similar to this:
7472 .ascii "arm_poke_function_name", 0
7475 .word 0xff000000 + (t1 - t0)
7476 arm_poke_function_name
7478 stmfd sp!, @{fp, ip, lr, pc@}
7482 When performing a stack backtrace, code can inspect the value of
7483 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
7484 location @code{pc - 12} and the top 8 bits are set, then we know that
7485 there is a function name embedded immediately preceding this location
7486 and has length @code{((pc[-3]) & 0xff000000)}.
7490 Generate code for the 16-bit Thumb instruction set. The default is to
7491 use the 32-bit ARM instruction set.
7494 @opindex mtpcs-frame
7495 Generate a stack frame that is compliant with the Thumb Procedure Call
7496 Standard for all non-leaf functions. (A leaf function is one that does
7497 not call any other functions.) The default is @option{-mno-tpcs-frame}.
7499 @item -mtpcs-leaf-frame
7500 @opindex mtpcs-leaf-frame
7501 Generate a stack frame that is compliant with the Thumb Procedure Call
7502 Standard for all leaf functions. (A leaf function is one that does
7503 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
7505 @item -mcallee-super-interworking
7506 @opindex mcallee-super-interworking
7507 Gives all externally visible functions in the file being compiled an ARM
7508 instruction set header which switches to Thumb mode before executing the
7509 rest of the function. This allows these functions to be called from
7510 non-interworking code.
7512 @item -mcaller-super-interworking
7513 @opindex mcaller-super-interworking
7514 Allows calls via function pointers (including virtual functions) to
7515 execute correctly regardless of whether the target code has been
7516 compiled for interworking or not. There is a small overhead in the cost
7517 of executing a function pointer if this option is enabled.
7519 @item -mtp=@var{name}
7521 Specify the access model for the thread local storage pointer. The valid
7522 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
7523 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
7524 (supported in the arm6k architecture), and @option{auto}, which uses the
7525 best available method for the selected processor. The default setting is
7531 @subsection AVR Options
7534 These options are defined for AVR implementations:
7537 @item -mmcu=@var{mcu}
7539 Specify ATMEL AVR instruction set or MCU type.
7541 Instruction set avr1 is for the minimal AVR core, not supported by the C
7542 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
7543 attiny11, attiny12, attiny15, attiny28).
7545 Instruction set avr2 (default) is for the classic AVR core with up to
7546 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
7547 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
7548 at90c8534, at90s8535).
7550 Instruction set avr3 is for the classic AVR core with up to 128K program
7551 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
7553 Instruction set avr4 is for the enhanced AVR core with up to 8K program
7554 memory space (MCU types: atmega8, atmega83, atmega85).
7556 Instruction set avr5 is for the enhanced AVR core with up to 128K program
7557 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
7558 atmega64, atmega128, at43usb355, at94k).
7562 Output instruction sizes to the asm file.
7564 @item -minit-stack=@var{N}
7565 @opindex minit-stack
7566 Specify the initial stack address, which may be a symbol or numeric value,
7567 @samp{__stack} is the default.
7569 @item -mno-interrupts
7570 @opindex mno-interrupts
7571 Generated code is not compatible with hardware interrupts.
7572 Code size will be smaller.
7574 @item -mcall-prologues
7575 @opindex mcall-prologues
7576 Functions prologues/epilogues expanded as call to appropriate
7577 subroutines. Code size will be smaller.
7579 @item -mno-tablejump
7580 @opindex mno-tablejump
7581 Do not generate tablejump insns which sometimes increase code size.
7584 @opindex mtiny-stack
7585 Change only the low 8 bits of the stack pointer.
7589 Assume int to be 8 bit integer. This affects the sizes of all types: A
7590 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
7591 and long long will be 4 bytes. Please note that this option does not
7592 comply to the C standards, but it will provide you with smaller code
7596 @node Blackfin Options
7597 @subsection Blackfin Options
7598 @cindex Blackfin Options
7601 @item -momit-leaf-frame-pointer
7602 @opindex momit-leaf-frame-pointer
7603 Don't keep the frame pointer in a register for leaf functions. This
7604 avoids the instructions to save, set up and restore frame pointers and
7605 makes an extra register available in leaf functions. The option
7606 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7607 which might make debugging harder.
7609 @item -mspecld-anomaly
7610 @opindex mspecld-anomaly
7611 When enabled, the compiler will ensure that the generated code does not
7612 contain speculative loads after jump instructions. This option is enabled
7615 @item -mno-specld-anomaly
7616 @opindex mno-specld-anomaly
7617 Don't generate extra code to prevent speculative loads from occurring.
7619 @item -mcsync-anomaly
7620 @opindex mcsync-anomaly
7621 When enabled, the compiler will ensure that the generated code does not
7622 contain CSYNC or SSYNC instructions too soon after conditional branches.
7623 This option is enabled by default.
7625 @item -mno-csync-anomaly
7626 @opindex mno-csync-anomaly
7627 Don't generate extra code to prevent CSYNC or SSYNC instructions from
7628 occurring too soon after a conditional branch.
7632 When enabled, the compiler is free to take advantage of the knowledge that
7633 the entire program fits into the low 64k of memory.
7636 @opindex mno-low-64k
7637 Assume that the program is arbitrarily large. This is the default.
7639 @item -mid-shared-library
7640 @opindex mid-shared-library
7641 Generate code that supports shared libraries via the library ID method.
7642 This allows for execute in place and shared libraries in an environment
7643 without virtual memory management. This option implies @option{-fPIC}.
7645 @item -mno-id-shared-library
7646 @opindex mno-id-shared-library
7647 Generate code that doesn't assume ID based shared libraries are being used.
7648 This is the default.
7650 @item -mshared-library-id=n
7651 @opindex mshared-library-id
7652 Specified the identification number of the ID based shared library being
7653 compiled. Specifying a value of 0 will generate more compact code, specifying
7654 other values will force the allocation of that number to the current
7655 library but is no more space or time efficient than omitting this option.
7658 @itemx -mno-long-calls
7659 @opindex mlong-calls
7660 @opindex mno-long-calls
7661 Tells the compiler to perform function calls by first loading the
7662 address of the function into a register and then performing a subroutine
7663 call on this register. This switch is needed if the target function
7664 will lie outside of the 24 bit addressing range of the offset based
7665 version of subroutine call instruction.
7667 This feature is not enabled by default. Specifying
7668 @option{-mno-long-calls} will restore the default behavior. Note these
7669 switches have no effect on how the compiler generates code to handle
7670 function calls via function pointers.
7674 @subsection CRIS Options
7675 @cindex CRIS Options
7677 These options are defined specifically for the CRIS ports.
7680 @item -march=@var{architecture-type}
7681 @itemx -mcpu=@var{architecture-type}
7684 Generate code for the specified architecture. The choices for
7685 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
7686 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
7687 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
7690 @item -mtune=@var{architecture-type}
7692 Tune to @var{architecture-type} everything applicable about the generated
7693 code, except for the ABI and the set of available instructions. The
7694 choices for @var{architecture-type} are the same as for
7695 @option{-march=@var{architecture-type}}.
7697 @item -mmax-stack-frame=@var{n}
7698 @opindex mmax-stack-frame
7699 Warn when the stack frame of a function exceeds @var{n} bytes.
7701 @item -melinux-stacksize=@var{n}
7702 @opindex melinux-stacksize
7703 Only available with the @samp{cris-axis-aout} target. Arranges for
7704 indications in the program to the kernel loader that the stack of the
7705 program should be set to @var{n} bytes.
7711 The options @option{-metrax4} and @option{-metrax100} are synonyms for
7712 @option{-march=v3} and @option{-march=v8} respectively.
7714 @item -mmul-bug-workaround
7715 @itemx -mno-mul-bug-workaround
7716 @opindex mmul-bug-workaround
7717 @opindex mno-mul-bug-workaround
7718 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
7719 models where it applies. This option is active by default.
7723 Enable CRIS-specific verbose debug-related information in the assembly
7724 code. This option also has the effect to turn off the @samp{#NO_APP}
7725 formatted-code indicator to the assembler at the beginning of the
7730 Do not use condition-code results from previous instruction; always emit
7731 compare and test instructions before use of condition codes.
7733 @item -mno-side-effects
7734 @opindex mno-side-effects
7735 Do not emit instructions with side-effects in addressing modes other than
7739 @itemx -mno-stack-align
7741 @itemx -mno-data-align
7742 @itemx -mconst-align
7743 @itemx -mno-const-align
7744 @opindex mstack-align
7745 @opindex mno-stack-align
7746 @opindex mdata-align
7747 @opindex mno-data-align
7748 @opindex mconst-align
7749 @opindex mno-const-align
7750 These options (no-options) arranges (eliminate arrangements) for the
7751 stack-frame, individual data and constants to be aligned for the maximum
7752 single data access size for the chosen CPU model. The default is to
7753 arrange for 32-bit alignment. ABI details such as structure layout are
7754 not affected by these options.
7762 Similar to the stack- data- and const-align options above, these options
7763 arrange for stack-frame, writable data and constants to all be 32-bit,
7764 16-bit or 8-bit aligned. The default is 32-bit alignment.
7766 @item -mno-prologue-epilogue
7767 @itemx -mprologue-epilogue
7768 @opindex mno-prologue-epilogue
7769 @opindex mprologue-epilogue
7770 With @option{-mno-prologue-epilogue}, the normal function prologue and
7771 epilogue that sets up the stack-frame are omitted and no return
7772 instructions or return sequences are generated in the code. Use this
7773 option only together with visual inspection of the compiled code: no
7774 warnings or errors are generated when call-saved registers must be saved,
7775 or storage for local variable needs to be allocated.
7781 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
7782 instruction sequences that load addresses for functions from the PLT part
7783 of the GOT rather than (traditional on other architectures) calls to the
7784 PLT@. The default is @option{-mgotplt}.
7788 Legacy no-op option only recognized with the cris-axis-aout target.
7792 Legacy no-op option only recognized with the cris-axis-elf and
7793 cris-axis-linux-gnu targets.
7797 Only recognized with the cris-axis-aout target, where it selects a
7798 GNU/linux-like multilib, include files and instruction set for
7803 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
7807 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
7808 to link with input-output functions from a simulator library. Code,
7809 initialized data and zero-initialized data are allocated consecutively.
7813 Like @option{-sim}, but pass linker options to locate initialized data at
7814 0x40000000 and zero-initialized data at 0x80000000.
7818 @subsection CRX Options
7821 These options are defined specifically for the CRX ports.
7827 Enable the use of multiply-accumulate instructions. Disabled by default.
7831 Push instructions will be used to pass outgoing arguments when functions
7832 are called. Enabled by default.
7835 @node Darwin Options
7836 @subsection Darwin Options
7837 @cindex Darwin options
7839 These options are defined for all architectures running the Darwin operating
7842 FSF GCC on Darwin does not create ``fat'' object files; it will create
7843 an object file for the single architecture that it was built to
7844 target. Apple's GCC on Darwin does create ``fat'' files if multiple
7845 @option{-arch} options are used; it does so by running the compiler or
7846 linker multiple times and joining the results together with
7849 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
7850 @samp{i686}) is determined by the flags that specify the ISA
7851 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
7852 @option{-force_cpusubtype_ALL} option can be used to override this.
7854 The Darwin tools vary in their behavior when presented with an ISA
7855 mismatch. The assembler, @file{as}, will only permit instructions to
7856 be used that are valid for the subtype of the file it is generating,
7857 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
7858 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
7859 and print an error if asked to create a shared library with a less
7860 restrictive subtype than its input files (for instance, trying to put
7861 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
7862 for executables, @file{ld}, will quietly give the executable the most
7863 restrictive subtype of any of its input files.
7868 Add the framework directory @var{dir} to the head of the list of
7869 directories to be searched for header files. These directories are
7870 interleaved with those specified by @option{-I} options and are
7871 scanned in a left-to-right order.
7873 A framework directory is a directory with frameworks in it. A
7874 framework is a directory with a @samp{"Headers"} and/or
7875 @samp{"PrivateHeaders"} directory contained directly in it that ends
7876 in @samp{".framework"}. The name of a framework is the name of this
7877 directory excluding the @samp{".framework"}. Headers associated with
7878 the framework are found in one of those two directories, with
7879 @samp{"Headers"} being searched first. A subframework is a framework
7880 directory that is in a framework's @samp{"Frameworks"} directory.
7881 Includes of subframework headers can only appear in a header of a
7882 framework that contains the subframework, or in a sibling subframework
7883 header. Two subframeworks are siblings if they occur in the same
7884 framework. A subframework should not have the same name as a
7885 framework, a warning will be issued if this is violated. Currently a
7886 subframework cannot have subframeworks, in the future, the mechanism
7887 may be extended to support this. The standard frameworks can be found
7888 in @samp{"/System/Library/Frameworks"} and
7889 @samp{"/Library/Frameworks"}. An example include looks like
7890 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
7891 the name of the framework and header.h is found in the
7892 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
7896 Emit debugging information for symbols that are used. For STABS
7897 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
7898 This is by default ON@.
7902 Emit debugging information for all symbols and types.
7904 @item -mmacosx-version-min=@var{version}
7905 The earliest version of MacOS X that this executable will run on
7906 is @var{version}. Typical values of @var{version} include @code{10.1},
7907 @code{10.2}, and @code{10.3.9}.
7909 The default for this option is to make choices that seem to be most
7912 @item -mone-byte-bool
7913 @opindex -mone-byte-bool
7914 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
7915 By default @samp{sizeof(bool)} is @samp{4} when compiling for
7916 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
7917 option has no effect on x86.
7919 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
7920 to generate code that is not binary compatible with code generated
7921 without that switch. Using this switch may require recompiling all
7922 other modules in a program, including system libraries. Use this
7923 switch to conform to a non-default data model.
7925 @item -mfix-and-continue
7926 @itemx -ffix-and-continue
7927 @itemx -findirect-data
7928 @opindex mfix-and-continue
7929 @opindex ffix-and-continue
7930 @opindex findirect-data
7931 Generate code suitable for fast turn around development. Needed to
7932 enable gdb to dynamically load @code{.o} files into already running
7933 programs. @option{-findirect-data} and @option{-ffix-and-continue}
7934 are provided for backwards compatibility.
7938 Loads all members of static archive libraries.
7939 See man ld(1) for more information.
7941 @item -arch_errors_fatal
7942 @opindex arch_errors_fatal
7943 Cause the errors having to do with files that have the wrong architecture
7947 @opindex bind_at_load
7948 Causes the output file to be marked such that the dynamic linker will
7949 bind all undefined references when the file is loaded or launched.
7953 Produce a Mach-o bundle format file.
7954 See man ld(1) for more information.
7956 @item -bundle_loader @var{executable}
7957 @opindex bundle_loader
7958 This option specifies the @var{executable} that will be loading the build
7959 output file being linked. See man ld(1) for more information.
7962 @opindex -dynamiclib
7963 When passed this option, GCC will produce a dynamic library instead of
7964 an executable when linking, using the Darwin @file{libtool} command.
7966 @item -force_cpusubtype_ALL
7967 @opindex -force_cpusubtype_ALL
7968 This causes GCC's output file to have the @var{ALL} subtype, instead of
7969 one controlled by the @option{-mcpu} or @option{-march} option.
7971 @item -allowable_client @var{client_name}
7973 @itemx -compatibility_version
7974 @itemx -current_version
7976 @itemx -dependency-file
7978 @itemx -dylinker_install_name
7980 @itemx -exported_symbols_list
7982 @itemx -flat_namespace
7983 @itemx -force_flat_namespace
7984 @itemx -headerpad_max_install_names
7987 @itemx -install_name
7988 @itemx -keep_private_externs
7989 @itemx -multi_module
7990 @itemx -multiply_defined
7991 @itemx -multiply_defined_unused
7993 @itemx -no_dead_strip_inits_and_terms
7994 @itemx -nofixprebinding
7997 @itemx -noseglinkedit
7998 @itemx -pagezero_size
8000 @itemx -prebind_all_twolevel_modules
8001 @itemx -private_bundle
8002 @itemx -read_only_relocs
8004 @itemx -sectobjectsymbols
8008 @itemx -sectobjectsymbols
8011 @itemx -segs_read_only_addr
8012 @itemx -segs_read_write_addr
8013 @itemx -seg_addr_table
8014 @itemx -seg_addr_table_filename
8017 @itemx -segs_read_only_addr
8018 @itemx -segs_read_write_addr
8019 @itemx -single_module
8022 @itemx -sub_umbrella
8023 @itemx -twolevel_namespace
8026 @itemx -unexported_symbols_list
8027 @itemx -weak_reference_mismatches
8030 @opindex allowable_client
8031 @opindex client_name
8032 @opindex compatibility_version
8033 @opindex current_version
8035 @opindex dependency-file
8037 @opindex dylinker_install_name
8039 @opindex exported_symbols_list
8041 @opindex flat_namespace
8042 @opindex force_flat_namespace
8043 @opindex headerpad_max_install_names
8046 @opindex install_name
8047 @opindex keep_private_externs
8048 @opindex multi_module
8049 @opindex multiply_defined
8050 @opindex multiply_defined_unused
8052 @opindex no_dead_strip_inits_and_terms
8053 @opindex nofixprebinding
8054 @opindex nomultidefs
8056 @opindex noseglinkedit
8057 @opindex pagezero_size
8059 @opindex prebind_all_twolevel_modules
8060 @opindex private_bundle
8061 @opindex read_only_relocs
8063 @opindex sectobjectsymbols
8067 @opindex sectobjectsymbols
8070 @opindex segs_read_only_addr
8071 @opindex segs_read_write_addr
8072 @opindex seg_addr_table
8073 @opindex seg_addr_table_filename
8074 @opindex seglinkedit
8076 @opindex segs_read_only_addr
8077 @opindex segs_read_write_addr
8078 @opindex single_module
8080 @opindex sub_library
8081 @opindex sub_umbrella
8082 @opindex twolevel_namespace
8085 @opindex unexported_symbols_list
8086 @opindex weak_reference_mismatches
8087 @opindex whatsloaded
8089 These options are passed to the Darwin linker. The Darwin linker man page
8090 describes them in detail.
8093 @node DEC Alpha Options
8094 @subsection DEC Alpha Options
8096 These @samp{-m} options are defined for the DEC Alpha implementations:
8099 @item -mno-soft-float
8101 @opindex mno-soft-float
8102 @opindex msoft-float
8103 Use (do not use) the hardware floating-point instructions for
8104 floating-point operations. When @option{-msoft-float} is specified,
8105 functions in @file{libgcc.a} will be used to perform floating-point
8106 operations. Unless they are replaced by routines that emulate the
8107 floating-point operations, or compiled in such a way as to call such
8108 emulations routines, these routines will issue floating-point
8109 operations. If you are compiling for an Alpha without floating-point
8110 operations, you must ensure that the library is built so as not to call
8113 Note that Alpha implementations without floating-point operations are
8114 required to have floating-point registers.
8119 @opindex mno-fp-regs
8120 Generate code that uses (does not use) the floating-point register set.
8121 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8122 register set is not used, floating point operands are passed in integer
8123 registers as if they were integers and floating-point results are passed
8124 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8125 so any function with a floating-point argument or return value called by code
8126 compiled with @option{-mno-fp-regs} must also be compiled with that
8129 A typical use of this option is building a kernel that does not use,
8130 and hence need not save and restore, any floating-point registers.
8134 The Alpha architecture implements floating-point hardware optimized for
8135 maximum performance. It is mostly compliant with the IEEE floating
8136 point standard. However, for full compliance, software assistance is
8137 required. This option generates code fully IEEE compliant code
8138 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8139 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8140 defined during compilation. The resulting code is less efficient but is
8141 able to correctly support denormalized numbers and exceptional IEEE
8142 values such as not-a-number and plus/minus infinity. Other Alpha
8143 compilers call this option @option{-ieee_with_no_inexact}.
8145 @item -mieee-with-inexact
8146 @opindex mieee-with-inexact
8147 This is like @option{-mieee} except the generated code also maintains
8148 the IEEE @var{inexact-flag}. Turning on this option causes the
8149 generated code to implement fully-compliant IEEE math. In addition to
8150 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8151 macro. On some Alpha implementations the resulting code may execute
8152 significantly slower than the code generated by default. Since there is
8153 very little code that depends on the @var{inexact-flag}, you should
8154 normally not specify this option. Other Alpha compilers call this
8155 option @option{-ieee_with_inexact}.
8157 @item -mfp-trap-mode=@var{trap-mode}
8158 @opindex mfp-trap-mode
8159 This option controls what floating-point related traps are enabled.
8160 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8161 The trap mode can be set to one of four values:
8165 This is the default (normal) setting. The only traps that are enabled
8166 are the ones that cannot be disabled in software (e.g., division by zero
8170 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8174 Like @samp{su}, but the instructions are marked to be safe for software
8175 completion (see Alpha architecture manual for details).
8178 Like @samp{su}, but inexact traps are enabled as well.
8181 @item -mfp-rounding-mode=@var{rounding-mode}
8182 @opindex mfp-rounding-mode
8183 Selects the IEEE rounding mode. Other Alpha compilers call this option
8184 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8189 Normal IEEE rounding mode. Floating point numbers are rounded towards
8190 the nearest machine number or towards the even machine number in case
8194 Round towards minus infinity.
8197 Chopped rounding mode. Floating point numbers are rounded towards zero.
8200 Dynamic rounding mode. A field in the floating point control register
8201 (@var{fpcr}, see Alpha architecture reference manual) controls the
8202 rounding mode in effect. The C library initializes this register for
8203 rounding towards plus infinity. Thus, unless your program modifies the
8204 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8207 @item -mtrap-precision=@var{trap-precision}
8208 @opindex mtrap-precision
8209 In the Alpha architecture, floating point traps are imprecise. This
8210 means without software assistance it is impossible to recover from a
8211 floating trap and program execution normally needs to be terminated.
8212 GCC can generate code that can assist operating system trap handlers
8213 in determining the exact location that caused a floating point trap.
8214 Depending on the requirements of an application, different levels of
8215 precisions can be selected:
8219 Program precision. This option is the default and means a trap handler
8220 can only identify which program caused a floating point exception.
8223 Function precision. The trap handler can determine the function that
8224 caused a floating point exception.
8227 Instruction precision. The trap handler can determine the exact
8228 instruction that caused a floating point exception.
8231 Other Alpha compilers provide the equivalent options called
8232 @option{-scope_safe} and @option{-resumption_safe}.
8234 @item -mieee-conformant
8235 @opindex mieee-conformant
8236 This option marks the generated code as IEEE conformant. You must not
8237 use this option unless you also specify @option{-mtrap-precision=i} and either
8238 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8239 is to emit the line @samp{.eflag 48} in the function prologue of the
8240 generated assembly file. Under DEC Unix, this has the effect that
8241 IEEE-conformant math library routines will be linked in.
8243 @item -mbuild-constants
8244 @opindex mbuild-constants
8245 Normally GCC examines a 32- or 64-bit integer constant to
8246 see if it can construct it from smaller constants in two or three
8247 instructions. If it cannot, it will output the constant as a literal and
8248 generate code to load it from the data segment at runtime.
8250 Use this option to require GCC to construct @emph{all} integer constants
8251 using code, even if it takes more instructions (the maximum is six).
8253 You would typically use this option to build a shared library dynamic
8254 loader. Itself a shared library, it must relocate itself in memory
8255 before it can find the variables and constants in its own data segment.
8261 Select whether to generate code to be assembled by the vendor-supplied
8262 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8280 Indicate whether GCC should generate code to use the optional BWX,
8281 CIX, FIX and MAX instruction sets. The default is to use the instruction
8282 sets supported by the CPU type specified via @option{-mcpu=} option or that
8283 of the CPU on which GCC was built if none was specified.
8288 @opindex mfloat-ieee
8289 Generate code that uses (does not use) VAX F and G floating point
8290 arithmetic instead of IEEE single and double precision.
8292 @item -mexplicit-relocs
8293 @itemx -mno-explicit-relocs
8294 @opindex mexplicit-relocs
8295 @opindex mno-explicit-relocs
8296 Older Alpha assemblers provided no way to generate symbol relocations
8297 except via assembler macros. Use of these macros does not allow
8298 optimal instruction scheduling. GNU binutils as of version 2.12
8299 supports a new syntax that allows the compiler to explicitly mark
8300 which relocations should apply to which instructions. This option
8301 is mostly useful for debugging, as GCC detects the capabilities of
8302 the assembler when it is built and sets the default accordingly.
8306 @opindex msmall-data
8307 @opindex mlarge-data
8308 When @option{-mexplicit-relocs} is in effect, static data is
8309 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8310 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8311 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8312 16-bit relocations off of the @code{$gp} register. This limits the
8313 size of the small data area to 64KB, but allows the variables to be
8314 directly accessed via a single instruction.
8316 The default is @option{-mlarge-data}. With this option the data area
8317 is limited to just below 2GB@. Programs that require more than 2GB of
8318 data must use @code{malloc} or @code{mmap} to allocate the data in the
8319 heap instead of in the program's data segment.
8321 When generating code for shared libraries, @option{-fpic} implies
8322 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8326 @opindex msmall-text
8327 @opindex mlarge-text
8328 When @option{-msmall-text} is used, the compiler assumes that the
8329 code of the entire program (or shared library) fits in 4MB, and is
8330 thus reachable with a branch instruction. When @option{-msmall-data}
8331 is used, the compiler can assume that all local symbols share the
8332 same @code{$gp} value, and thus reduce the number of instructions
8333 required for a function call from 4 to 1.
8335 The default is @option{-mlarge-text}.
8337 @item -mcpu=@var{cpu_type}
8339 Set the instruction set and instruction scheduling parameters for
8340 machine type @var{cpu_type}. You can specify either the @samp{EV}
8341 style name or the corresponding chip number. GCC supports scheduling
8342 parameters for the EV4, EV5 and EV6 family of processors and will
8343 choose the default values for the instruction set from the processor
8344 you specify. If you do not specify a processor type, GCC will default
8345 to the processor on which the compiler was built.
8347 Supported values for @var{cpu_type} are
8353 Schedules as an EV4 and has no instruction set extensions.
8357 Schedules as an EV5 and has no instruction set extensions.
8361 Schedules as an EV5 and supports the BWX extension.
8366 Schedules as an EV5 and supports the BWX and MAX extensions.
8370 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8374 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8377 @item -mtune=@var{cpu_type}
8379 Set only the instruction scheduling parameters for machine type
8380 @var{cpu_type}. The instruction set is not changed.
8382 @item -mmemory-latency=@var{time}
8383 @opindex mmemory-latency
8384 Sets the latency the scheduler should assume for typical memory
8385 references as seen by the application. This number is highly
8386 dependent on the memory access patterns used by the application
8387 and the size of the external cache on the machine.
8389 Valid options for @var{time} are
8393 A decimal number representing clock cycles.
8399 The compiler contains estimates of the number of clock cycles for
8400 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8401 (also called Dcache, Scache, and Bcache), as well as to main memory.
8402 Note that L3 is only valid for EV5.
8407 @node DEC Alpha/VMS Options
8408 @subsection DEC Alpha/VMS Options
8410 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8413 @item -mvms-return-codes
8414 @opindex mvms-return-codes
8415 Return VMS condition codes from main. The default is to return POSIX
8416 style condition (e.g.@ error) codes.
8420 @subsection FRV Options
8427 Only use the first 32 general purpose registers.
8432 Use all 64 general purpose registers.
8437 Use only the first 32 floating point registers.
8442 Use all 64 floating point registers
8445 @opindex mhard-float
8447 Use hardware instructions for floating point operations.
8450 @opindex msoft-float
8452 Use library routines for floating point operations.
8457 Dynamically allocate condition code registers.
8462 Do not try to dynamically allocate condition code registers, only
8463 use @code{icc0} and @code{fcc0}.
8468 Change ABI to use double word insns.
8473 Do not use double word instructions.
8478 Use floating point double instructions.
8483 Do not use floating point double instructions.
8488 Use media instructions.
8493 Do not use media instructions.
8498 Use multiply and add/subtract instructions.
8503 Do not use multiply and add/subtract instructions.
8508 Select the FDPIC ABI, that uses function descriptors to represent
8509 pointers to functions. Without any PIC/PIE-related options, it
8510 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
8511 assumes GOT entries and small data are within a 12-bit range from the
8512 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
8513 are computed with 32 bits.
8516 @opindex minline-plt
8518 Enable inlining of PLT entries in function calls to functions that are
8519 not known to bind locally. It has no effect without @option{-mfdpic}.
8520 It's enabled by default if optimizing for speed and compiling for
8521 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
8522 optimization option such as @option{-O3} or above is present in the
8528 Assume a large TLS segment when generating thread-local code.
8533 Do not assume a large TLS segment when generating thread-local code.
8538 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
8539 that is known to be in read-only sections. It's enabled by default,
8540 except for @option{-fpic} or @option{-fpie}: even though it may help
8541 make the global offset table smaller, it trades 1 instruction for 4.
8542 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
8543 one of which may be shared by multiple symbols, and it avoids the need
8544 for a GOT entry for the referenced symbol, so it's more likely to be a
8545 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
8547 @item -multilib-library-pic
8548 @opindex multilib-library-pic
8550 Link with the (library, not FD) pic libraries. It's implied by
8551 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
8552 @option{-fpic} without @option{-mfdpic}. You should never have to use
8558 Follow the EABI requirement of always creating a frame pointer whenever
8559 a stack frame is allocated. This option is enabled by default and can
8560 be disabled with @option{-mno-linked-fp}.
8563 @opindex mlong-calls
8565 Use indirect addressing to call functions outside the current
8566 compilation unit. This allows the functions to be placed anywhere
8567 within the 32-bit address space.
8569 @item -malign-labels
8570 @opindex malign-labels
8572 Try to align labels to an 8-byte boundary by inserting nops into the
8573 previous packet. This option only has an effect when VLIW packing
8574 is enabled. It doesn't create new packets; it merely adds nops to
8578 @opindex mlibrary-pic
8580 Generate position-independent EABI code.
8585 Use only the first four media accumulator registers.
8590 Use all eight media accumulator registers.
8595 Pack VLIW instructions.
8600 Do not pack VLIW instructions.
8605 Do not mark ABI switches in e_flags.
8610 Enable the use of conditional-move instructions (default).
8612 This switch is mainly for debugging the compiler and will likely be removed
8613 in a future version.
8615 @item -mno-cond-move
8616 @opindex mno-cond-move
8618 Disable the use of conditional-move instructions.
8620 This switch is mainly for debugging the compiler and will likely be removed
8621 in a future version.
8626 Enable the use of conditional set instructions (default).
8628 This switch is mainly for debugging the compiler and will likely be removed
8629 in a future version.
8634 Disable the use of conditional set instructions.
8636 This switch is mainly for debugging the compiler and will likely be removed
8637 in a future version.
8642 Enable the use of conditional execution (default).
8644 This switch is mainly for debugging the compiler and will likely be removed
8645 in a future version.
8647 @item -mno-cond-exec
8648 @opindex mno-cond-exec
8650 Disable the use of conditional execution.
8652 This switch is mainly for debugging the compiler and will likely be removed
8653 in a future version.
8656 @opindex mvliw-branch
8658 Run a pass to pack branches into VLIW instructions (default).
8660 This switch is mainly for debugging the compiler and will likely be removed
8661 in a future version.
8663 @item -mno-vliw-branch
8664 @opindex mno-vliw-branch
8666 Do not run a pass to pack branches into VLIW instructions.
8668 This switch is mainly for debugging the compiler and will likely be removed
8669 in a future version.
8671 @item -mmulti-cond-exec
8672 @opindex mmulti-cond-exec
8674 Enable optimization of @code{&&} and @code{||} in conditional execution
8677 This switch is mainly for debugging the compiler and will likely be removed
8678 in a future version.
8680 @item -mno-multi-cond-exec
8681 @opindex mno-multi-cond-exec
8683 Disable optimization of @code{&&} and @code{||} in conditional execution.
8685 This switch is mainly for debugging the compiler and will likely be removed
8686 in a future version.
8688 @item -mnested-cond-exec
8689 @opindex mnested-cond-exec
8691 Enable nested conditional execution optimizations (default).
8693 This switch is mainly for debugging the compiler and will likely be removed
8694 in a future version.
8696 @item -mno-nested-cond-exec
8697 @opindex mno-nested-cond-exec
8699 Disable nested conditional execution optimizations.
8701 This switch is mainly for debugging the compiler and will likely be removed
8702 in a future version.
8704 @item -moptimize-membar
8705 @opindex moptimize-membar
8707 This switch removes redundant @code{membar} instructions from the
8708 compiler generated code. It is enabled by default.
8710 @item -mno-optimize-membar
8711 @opindex mno-optimize-membar
8713 This switch disables the automatic removal of redundant @code{membar}
8714 instructions from the generated code.
8716 @item -mtomcat-stats
8717 @opindex mtomcat-stats
8719 Cause gas to print out tomcat statistics.
8721 @item -mcpu=@var{cpu}
8724 Select the processor type for which to generate code. Possible values are
8725 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
8726 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
8730 @node H8/300 Options
8731 @subsection H8/300 Options
8733 These @samp{-m} options are defined for the H8/300 implementations:
8738 Shorten some address references at link time, when possible; uses the
8739 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8740 ld, Using ld}, for a fuller description.
8744 Generate code for the H8/300H@.
8748 Generate code for the H8S@.
8752 Generate code for the H8S and H8/300H in the normal mode. This switch
8753 must be used either with @option{-mh} or @option{-ms}.
8757 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
8761 Make @code{int} data 32 bits by default.
8765 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
8766 The default for the H8/300H and H8S is to align longs and floats on 4
8768 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8769 This option has no effect on the H8/300.
8773 @subsection HPPA Options
8774 @cindex HPPA Options
8776 These @samp{-m} options are defined for the HPPA family of computers:
8779 @item -march=@var{architecture-type}
8781 Generate code for the specified architecture. The choices for
8782 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8783 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8784 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8785 architecture option for your machine. Code compiled for lower numbered
8786 architectures will run on higher numbered architectures, but not the
8790 @itemx -mpa-risc-1-1
8791 @itemx -mpa-risc-2-0
8792 @opindex mpa-risc-1-0
8793 @opindex mpa-risc-1-1
8794 @opindex mpa-risc-2-0
8795 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8798 @opindex mbig-switch
8799 Generate code suitable for big switch tables. Use this option only if
8800 the assembler/linker complain about out of range branches within a switch
8803 @item -mjump-in-delay
8804 @opindex mjump-in-delay
8805 Fill delay slots of function calls with unconditional jump instructions
8806 by modifying the return pointer for the function call to be the target
8807 of the conditional jump.
8809 @item -mdisable-fpregs
8810 @opindex mdisable-fpregs
8811 Prevent floating point registers from being used in any manner. This is
8812 necessary for compiling kernels which perform lazy context switching of
8813 floating point registers. If you use this option and attempt to perform
8814 floating point operations, the compiler will abort.
8816 @item -mdisable-indexing
8817 @opindex mdisable-indexing
8818 Prevent the compiler from using indexing address modes. This avoids some
8819 rather obscure problems when compiling MIG generated code under MACH@.
8821 @item -mno-space-regs
8822 @opindex mno-space-regs
8823 Generate code that assumes the target has no space registers. This allows
8824 GCC to generate faster indirect calls and use unscaled index address modes.
8826 Such code is suitable for level 0 PA systems and kernels.
8828 @item -mfast-indirect-calls
8829 @opindex mfast-indirect-calls
8830 Generate code that assumes calls never cross space boundaries. This
8831 allows GCC to emit code which performs faster indirect calls.
8833 This option will not work in the presence of shared libraries or nested
8836 @item -mfixed-range=@var{register-range}
8837 @opindex mfixed-range
8838 Generate code treating the given register range as fixed registers.
8839 A fixed register is one that the register allocator can not use. This is
8840 useful when compiling kernel code. A register range is specified as
8841 two registers separated by a dash. Multiple register ranges can be
8842 specified separated by a comma.
8844 @item -mlong-load-store
8845 @opindex mlong-load-store
8846 Generate 3-instruction load and store sequences as sometimes required by
8847 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8850 @item -mportable-runtime
8851 @opindex mportable-runtime
8852 Use the portable calling conventions proposed by HP for ELF systems.
8856 Enable the use of assembler directives only GAS understands.
8858 @item -mschedule=@var{cpu-type}
8860 Schedule code according to the constraints for the machine type
8861 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8862 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8863 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8864 proper scheduling option for your machine. The default scheduling is
8868 @opindex mlinker-opt
8869 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8870 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8871 linkers in which they give bogus error messages when linking some programs.
8874 @opindex msoft-float
8875 Generate output containing library calls for floating point.
8876 @strong{Warning:} the requisite libraries are not available for all HPPA
8877 targets. Normally the facilities of the machine's usual C compiler are
8878 used, but this cannot be done directly in cross-compilation. You must make
8879 your own arrangements to provide suitable library functions for
8880 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8881 does provide software floating point support.
8883 @option{-msoft-float} changes the calling convention in the output file;
8884 therefore, it is only useful if you compile @emph{all} of a program with
8885 this option. In particular, you need to compile @file{libgcc.a}, the
8886 library that comes with GCC, with @option{-msoft-float} in order for
8891 Generate the predefine, @code{_SIO}, for server IO@. The default is
8892 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
8893 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
8894 options are available under HP-UX and HI-UX@.
8898 Use GNU ld specific options. This passes @option{-shared} to ld when
8899 building a shared library. It is the default when GCC is configured,
8900 explicitly or implicitly, with the GNU linker. This option does not
8901 have any affect on which ld is called, it only changes what parameters
8902 are passed to that ld. The ld that is called is determined by the
8903 @option{--with-ld} configure option, GCC's program search path, and
8904 finally by the user's @env{PATH}. The linker used by GCC can be printed
8905 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
8906 on the 64 bit HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
8910 Use HP ld specific options. This passes @option{-b} to ld when building
8911 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8912 links. It is the default when GCC is configured, explicitly or
8913 implicitly, with the HP linker. This option does not have any affect on
8914 which ld is called, it only changes what parameters are passed to that
8915 ld. The ld that is called is determined by the @option{--with-ld}
8916 configure option, GCC's program search path, and finally by the user's
8917 @env{PATH}. The linker used by GCC can be printed using @samp{which
8918 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
8919 HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
8922 @opindex mno-long-calls
8923 Generate code that uses long call sequences. This ensures that a call
8924 is always able to reach linker generated stubs. The default is to generate
8925 long calls only when the distance from the call site to the beginning
8926 of the function or translation unit, as the case may be, exceeds a
8927 predefined limit set by the branch type being used. The limits for
8928 normal calls are 7,600,000 and 240,000 bytes, respectively for the
8929 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
8932 Distances are measured from the beginning of functions when using the
8933 @option{-ffunction-sections} option, or when using the @option{-mgas}
8934 and @option{-mno-portable-runtime} options together under HP-UX with
8937 It is normally not desirable to use this option as it will degrade
8938 performance. However, it may be useful in large applications,
8939 particularly when partial linking is used to build the application.
8941 The types of long calls used depends on the capabilities of the
8942 assembler and linker, and the type of code being generated. The
8943 impact on systems that support long absolute calls, and long pic
8944 symbol-difference or pc-relative calls should be relatively small.
8945 However, an indirect call is used on 32-bit ELF systems in pic code
8946 and it is quite long.
8948 @item -munix=@var{unix-std}
8950 Generate compiler predefines and select a startfile for the specified
8951 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
8952 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
8953 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
8954 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
8955 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
8958 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
8959 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
8960 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
8961 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
8962 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
8963 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
8965 It is @emph{important} to note that this option changes the interfaces
8966 for various library routines. It also affects the operational behavior
8967 of the C library. Thus, @emph{extreme} care is needed in using this
8970 Library code that is intended to operate with more than one UNIX
8971 standard must test, set and restore the variable @var{__xpg4_extended_mask}
8972 as appropriate. Most GNU software doesn't provide this capability.
8976 Suppress the generation of link options to search libdld.sl when the
8977 @option{-static} option is specified on HP-UX 10 and later.
8981 The HP-UX implementation of setlocale in libc has a dependency on
8982 libdld.sl. There isn't an archive version of libdld.sl. Thus,
8983 when the @option{-static} option is specified, special link options
8984 are needed to resolve this dependency.
8986 On HP-UX 10 and later, the GCC driver adds the necessary options to
8987 link with libdld.sl when the @option{-static} option is specified.
8988 This causes the resulting binary to be dynamic. On the 64-bit port,
8989 the linkers generate dynamic binaries by default in any case. The
8990 @option{-nolibdld} option can be used to prevent the GCC driver from
8991 adding these link options.
8995 Add support for multithreading with the @dfn{dce thread} library
8996 under HP-UX@. This option sets flags for both the preprocessor and
9000 @node i386 and x86-64 Options
9001 @subsection Intel 386 and AMD x86-64 Options
9002 @cindex i386 Options
9003 @cindex x86-64 Options
9004 @cindex Intel 386 Options
9005 @cindex AMD x86-64 Options
9007 These @samp{-m} options are defined for the i386 and x86-64 family of
9011 @item -mtune=@var{cpu-type}
9013 Tune to @var{cpu-type} everything applicable about the generated code, except
9014 for the ABI and the set of available instructions. The choices for
9018 Original Intel's i386 CPU@.
9020 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
9022 Intel Pentium CPU with no MMX support.
9024 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
9025 @item i686, pentiumpro
9026 Intel PentiumPro CPU@.
9028 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
9029 @item pentium3, pentium3m
9030 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
9033 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
9034 support. Used by Centrino notebooks.
9035 @item pentium4, pentium4m
9036 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
9038 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
9041 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
9042 SSE2 and SSE3 instruction set support.
9044 AMD K6 CPU with MMX instruction set support.
9046 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
9047 @item athlon, athlon-tbird
9048 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
9050 @item athlon-4, athlon-xp, athlon-mp
9051 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
9052 instruction set support.
9053 @item k8, opteron, athlon64, athlon-fx
9054 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
9055 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
9057 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
9060 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
9061 instruction set support.
9063 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
9064 implemented for this chip.)
9066 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
9067 implemented for this chip.)
9070 While picking a specific @var{cpu-type} will schedule things appropriately
9071 for that particular chip, the compiler will not generate any code that
9072 does not run on the i386 without the @option{-march=@var{cpu-type}} option
9075 @item -march=@var{cpu-type}
9077 Generate instructions for the machine type @var{cpu-type}. The choices
9078 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
9079 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
9081 @item -mcpu=@var{cpu-type}
9083 A deprecated synonym for @option{-mtune}.
9092 @opindex mpentiumpro
9093 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
9094 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
9095 These synonyms are deprecated.
9097 @item -mfpmath=@var{unit}
9099 Generate floating point arithmetics for selected unit @var{unit}. The choices
9104 Use the standard 387 floating point coprocessor present majority of chips and
9105 emulated otherwise. Code compiled with this option will run almost everywhere.
9106 The temporary results are computed in 80bit precision instead of precision
9107 specified by the type resulting in slightly different results compared to most
9108 of other chips. See @option{-ffloat-store} for more detailed description.
9110 This is the default choice for i386 compiler.
9113 Use scalar floating point instructions present in the SSE instruction set.
9114 This instruction set is supported by Pentium3 and newer chips, in the AMD line
9115 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
9116 instruction set supports only single precision arithmetics, thus the double and
9117 extended precision arithmetics is still done using 387. Later version, present
9118 only in Pentium4 and the future AMD x86-64 chips supports double precision
9121 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
9122 or @option{-msse2} switches to enable SSE extensions and make this option
9123 effective. For the x86-64 compiler, these extensions are enabled by default.
9125 The resulting code should be considerably faster in the majority of cases and avoid
9126 the numerical instability problems of 387 code, but may break some existing
9127 code that expects temporaries to be 80bit.
9129 This is the default choice for the x86-64 compiler.
9132 Attempt to utilize both instruction sets at once. This effectively double the
9133 amount of available registers and on chips with separate execution units for
9134 387 and SSE the execution resources too. Use this option with care, as it is
9135 still experimental, because the GCC register allocator does not model separate
9136 functional units well resulting in instable performance.
9139 @item -masm=@var{dialect}
9140 @opindex masm=@var{dialect}
9141 Output asm instructions using selected @var{dialect}. Supported
9142 choices are @samp{intel} or @samp{att} (the default one). Darwin does
9143 not support @samp{intel}.
9148 @opindex mno-ieee-fp
9149 Control whether or not the compiler uses IEEE floating point
9150 comparisons. These handle correctly the case where the result of a
9151 comparison is unordered.
9154 @opindex msoft-float
9155 Generate output containing library calls for floating point.
9156 @strong{Warning:} the requisite libraries are not part of GCC@.
9157 Normally the facilities of the machine's usual C compiler are used, but
9158 this can't be done directly in cross-compilation. You must make your
9159 own arrangements to provide suitable library functions for
9162 On machines where a function returns floating point results in the 80387
9163 register stack, some floating point opcodes may be emitted even if
9164 @option{-msoft-float} is used.
9166 @item -mno-fp-ret-in-387
9167 @opindex mno-fp-ret-in-387
9168 Do not use the FPU registers for return values of functions.
9170 The usual calling convention has functions return values of types
9171 @code{float} and @code{double} in an FPU register, even if there
9172 is no FPU@. The idea is that the operating system should emulate
9175 The option @option{-mno-fp-ret-in-387} causes such values to be returned
9176 in ordinary CPU registers instead.
9178 @item -mno-fancy-math-387
9179 @opindex mno-fancy-math-387
9180 Some 387 emulators do not support the @code{sin}, @code{cos} and
9181 @code{sqrt} instructions for the 387. Specify this option to avoid
9182 generating those instructions. This option is the default on FreeBSD,
9183 OpenBSD and NetBSD@. This option is overridden when @option{-march}
9184 indicates that the target cpu will always have an FPU and so the
9185 instruction will not need emulation. As of revision 2.6.1, these
9186 instructions are not generated unless you also use the
9187 @option{-funsafe-math-optimizations} switch.
9189 @item -malign-double
9190 @itemx -mno-align-double
9191 @opindex malign-double
9192 @opindex mno-align-double
9193 Control whether GCC aligns @code{double}, @code{long double}, and
9194 @code{long long} variables on a two word boundary or a one word
9195 boundary. Aligning @code{double} variables on a two word boundary will
9196 produce code that runs somewhat faster on a @samp{Pentium} at the
9197 expense of more memory.
9199 @strong{Warning:} if you use the @option{-malign-double} switch,
9200 structures containing the above types will be aligned differently than
9201 the published application binary interface specifications for the 386
9202 and will not be binary compatible with structures in code compiled
9203 without that switch.
9205 @item -m96bit-long-double
9206 @itemx -m128bit-long-double
9207 @opindex m96bit-long-double
9208 @opindex m128bit-long-double
9209 These switches control the size of @code{long double} type. The i386
9210 application binary interface specifies the size to be 96 bits,
9211 so @option{-m96bit-long-double} is the default in 32 bit mode.
9213 Modern architectures (Pentium and newer) would prefer @code{long double}
9214 to be aligned to an 8 or 16 byte boundary. In arrays or structures
9215 conforming to the ABI, this would not be possible. So specifying a
9216 @option{-m128bit-long-double} will align @code{long double}
9217 to a 16 byte boundary by padding the @code{long double} with an additional
9220 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
9221 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
9223 Notice that neither of these options enable any extra precision over the x87
9224 standard of 80 bits for a @code{long double}.
9226 @strong{Warning:} if you override the default value for your target ABI, the
9227 structures and arrays containing @code{long double} variables will change
9228 their size as well as function calling convention for function taking
9229 @code{long double} will be modified. Hence they will not be binary
9230 compatible with arrays or structures in code compiled without that switch.
9232 @item -mmlarge-data-threshold=@var{number}
9233 @opindex mlarge-data-threshold=@var{number}
9234 When @option{-mcmodel=medium} is specified, the data greater than
9235 @var{threshold} are placed in large data section. This value must be the
9236 same across all object linked into the binary and defaults to 65535.
9239 @itemx -mno-svr3-shlib
9240 @opindex msvr3-shlib
9241 @opindex mno-svr3-shlib
9242 Control whether GCC places uninitialized local variables into the
9243 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
9244 into @code{bss}. These options are meaningful only on System V Release 3.
9248 Use a different function-calling convention, in which functions that
9249 take a fixed number of arguments return with the @code{ret} @var{num}
9250 instruction, which pops their arguments while returning. This saves one
9251 instruction in the caller since there is no need to pop the arguments
9254 You can specify that an individual function is called with this calling
9255 sequence with the function attribute @samp{stdcall}. You can also
9256 override the @option{-mrtd} option by using the function attribute
9257 @samp{cdecl}. @xref{Function Attributes}.
9259 @strong{Warning:} this calling convention is incompatible with the one
9260 normally used on Unix, so you cannot use it if you need to call
9261 libraries compiled with the Unix compiler.
9263 Also, you must provide function prototypes for all functions that
9264 take variable numbers of arguments (including @code{printf});
9265 otherwise incorrect code will be generated for calls to those
9268 In addition, seriously incorrect code will result if you call a
9269 function with too many arguments. (Normally, extra arguments are
9270 harmlessly ignored.)
9272 @item -mregparm=@var{num}
9274 Control how many registers are used to pass integer arguments. By
9275 default, no registers are used to pass arguments, and at most 3
9276 registers can be used. You can control this behavior for a specific
9277 function by using the function attribute @samp{regparm}.
9278 @xref{Function Attributes}.
9280 @strong{Warning:} if you use this switch, and
9281 @var{num} is nonzero, then you must build all modules with the same
9282 value, including any libraries. This includes the system libraries and
9286 @opindex msseregparm
9287 Use SSE register passing conventions for float and double arguments
9288 and return values. You can control this behavior for a specific
9289 function by using the function attribute @samp{sseregparm}.
9290 @xref{Function Attributes}.
9292 @strong{Warning:} if you use this switch then you must build all
9293 modules with the same value, including any libraries. This includes
9294 the system libraries and startup modules.
9296 @item -mpreferred-stack-boundary=@var{num}
9297 @opindex mpreferred-stack-boundary
9298 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
9299 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
9300 the default is 4 (16 bytes or 128 bits), except when optimizing for code
9301 size (@option{-Os}), in which case the default is the minimum correct
9302 alignment (4 bytes for x86, and 8 bytes for x86-64).
9304 On Pentium and PentiumPro, @code{double} and @code{long double} values
9305 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
9306 suffer significant run time performance penalties. On Pentium III, the
9307 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
9308 penalties if it is not 16 byte aligned.
9310 To ensure proper alignment of this values on the stack, the stack boundary
9311 must be as aligned as that required by any value stored on the stack.
9312 Further, every function must be generated such that it keeps the stack
9313 aligned. Thus calling a function compiled with a higher preferred
9314 stack boundary from a function compiled with a lower preferred stack
9315 boundary will most likely misalign the stack. It is recommended that
9316 libraries that use callbacks always use the default setting.
9318 This extra alignment does consume extra stack space, and generally
9319 increases code size. Code that is sensitive to stack space usage, such
9320 as embedded systems and operating system kernels, may want to reduce the
9321 preferred alignment to @option{-mpreferred-stack-boundary=2}.
9339 These switches enable or disable the use of instructions in the MMX,
9340 SSE, SSE2 or 3DNow! extended instruction sets. These extensions are
9341 also available as built-in functions: see @ref{X86 Built-in Functions},
9342 for details of the functions enabled and disabled by these switches.
9344 To have SSE/SSE2 instructions generated automatically from floating-point
9345 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
9347 These options will enable GCC to use these extended instructions in
9348 generated code, even without @option{-mfpmath=sse}. Applications which
9349 perform runtime CPU detection must compile separate files for each
9350 supported architecture, using the appropriate flags. In particular,
9351 the file containing the CPU detection code should be compiled without
9355 @itemx -mno-push-args
9357 @opindex mno-push-args
9358 Use PUSH operations to store outgoing parameters. This method is shorter
9359 and usually equally fast as method using SUB/MOV operations and is enabled
9360 by default. In some cases disabling it may improve performance because of
9361 improved scheduling and reduced dependencies.
9363 @item -maccumulate-outgoing-args
9364 @opindex maccumulate-outgoing-args
9365 If enabled, the maximum amount of space required for outgoing arguments will be
9366 computed in the function prologue. This is faster on most modern CPUs
9367 because of reduced dependencies, improved scheduling and reduced stack usage
9368 when preferred stack boundary is not equal to 2. The drawback is a notable
9369 increase in code size. This switch implies @option{-mno-push-args}.
9373 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
9374 on thread-safe exception handling must compile and link all code with the
9375 @option{-mthreads} option. When compiling, @option{-mthreads} defines
9376 @option{-D_MT}; when linking, it links in a special thread helper library
9377 @option{-lmingwthrd} which cleans up per thread exception handling data.
9379 @item -mno-align-stringops
9380 @opindex mno-align-stringops
9381 Do not align destination of inlined string operations. This switch reduces
9382 code size and improves performance in case the destination is already aligned,
9383 but GCC doesn't know about it.
9385 @item -minline-all-stringops
9386 @opindex minline-all-stringops
9387 By default GCC inlines string operations only when destination is known to be
9388 aligned at least to 4 byte boundary. This enables more inlining, increase code
9389 size, but may improve performance of code that depends on fast memcpy, strlen
9390 and memset for short lengths.
9392 @item -momit-leaf-frame-pointer
9393 @opindex momit-leaf-frame-pointer
9394 Don't keep the frame pointer in a register for leaf functions. This
9395 avoids the instructions to save, set up and restore frame pointers and
9396 makes an extra register available in leaf functions. The option
9397 @option{-fomit-frame-pointer} removes the frame pointer for all functions
9398 which might make debugging harder.
9400 @item -mtls-direct-seg-refs
9401 @itemx -mno-tls-direct-seg-refs
9402 @opindex mtls-direct-seg-refs
9403 Controls whether TLS variables may be accessed with offsets from the
9404 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
9405 or whether the thread base pointer must be added. Whether or not this
9406 is legal depends on the operating system, and whether it maps the
9407 segment to cover the entire TLS area.
9409 For systems that use GNU libc, the default is on.
9412 These @samp{-m} switches are supported in addition to the above
9413 on AMD x86-64 processors in 64-bit environments.
9420 Generate code for a 32-bit or 64-bit environment.
9421 The 32-bit environment sets int, long and pointer to 32 bits and
9422 generates code that runs on any i386 system.
9423 The 64-bit environment sets int to 32 bits and long and pointer
9424 to 64 bits and generates code for AMD's x86-64 architecture.
9427 @opindex no-red-zone
9428 Do not use a so called red zone for x86-64 code. The red zone is mandated
9429 by the x86-64 ABI, it is a 128-byte area beyond the location of the
9430 stack pointer that will not be modified by signal or interrupt handlers
9431 and therefore can be used for temporary data without adjusting the stack
9432 pointer. The flag @option{-mno-red-zone} disables this red zone.
9434 @item -mcmodel=small
9435 @opindex mcmodel=small
9436 Generate code for the small code model: the program and its symbols must
9437 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
9438 Programs can be statically or dynamically linked. This is the default
9441 @item -mcmodel=kernel
9442 @opindex mcmodel=kernel
9443 Generate code for the kernel code model. The kernel runs in the
9444 negative 2 GB of the address space.
9445 This model has to be used for Linux kernel code.
9447 @item -mcmodel=medium
9448 @opindex mcmodel=medium
9449 Generate code for the medium model: The program is linked in the lower 2
9450 GB of the address space but symbols can be located anywhere in the
9451 address space. Programs can be statically or dynamically linked, but
9452 building of shared libraries are not supported with the medium model.
9454 @item -mcmodel=large
9455 @opindex mcmodel=large
9456 Generate code for the large model: This model makes no assumptions
9457 about addresses and sizes of sections. Currently GCC does not implement
9462 @subsection IA-64 Options
9463 @cindex IA-64 Options
9465 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9469 @opindex mbig-endian
9470 Generate code for a big endian target. This is the default for HP-UX@.
9472 @item -mlittle-endian
9473 @opindex mlittle-endian
9474 Generate code for a little endian target. This is the default for AIX5
9481 Generate (or don't) code for the GNU assembler. This is the default.
9482 @c Also, this is the default if the configure option @option{--with-gnu-as}
9489 Generate (or don't) code for the GNU linker. This is the default.
9490 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9495 Generate code that does not use a global pointer register. The result
9496 is not position independent code, and violates the IA-64 ABI@.
9498 @item -mvolatile-asm-stop
9499 @itemx -mno-volatile-asm-stop
9500 @opindex mvolatile-asm-stop
9501 @opindex mno-volatile-asm-stop
9502 Generate (or don't) a stop bit immediately before and after volatile asm
9505 @item -mregister-names
9506 @itemx -mno-register-names
9507 @opindex mregister-names
9508 @opindex mno-register-names
9509 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9510 the stacked registers. This may make assembler output more readable.
9516 Disable (or enable) optimizations that use the small data section. This may
9517 be useful for working around optimizer bugs.
9520 @opindex mconstant-gp
9521 Generate code that uses a single constant global pointer value. This is
9522 useful when compiling kernel code.
9526 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9527 This is useful when compiling firmware code.
9529 @item -minline-float-divide-min-latency
9530 @opindex minline-float-divide-min-latency
9531 Generate code for inline divides of floating point values
9532 using the minimum latency algorithm.
9534 @item -minline-float-divide-max-throughput
9535 @opindex minline-float-divide-max-throughput
9536 Generate code for inline divides of floating point values
9537 using the maximum throughput algorithm.
9539 @item -minline-int-divide-min-latency
9540 @opindex minline-int-divide-min-latency
9541 Generate code for inline divides of integer values
9542 using the minimum latency algorithm.
9544 @item -minline-int-divide-max-throughput
9545 @opindex minline-int-divide-max-throughput
9546 Generate code for inline divides of integer values
9547 using the maximum throughput algorithm.
9549 @item -minline-sqrt-min-latency
9550 @opindex minline-sqrt-min-latency
9551 Generate code for inline square roots
9552 using the minimum latency algorithm.
9554 @item -minline-sqrt-max-throughput
9555 @opindex minline-sqrt-max-throughput
9556 Generate code for inline square roots
9557 using the maximum throughput algorithm.
9559 @item -mno-dwarf2-asm
9561 @opindex mno-dwarf2-asm
9562 @opindex mdwarf2-asm
9563 Don't (or do) generate assembler code for the DWARF2 line number debugging
9564 info. This may be useful when not using the GNU assembler.
9566 @item -mearly-stop-bits
9567 @itemx -mno-early-stop-bits
9568 @opindex mearly-stop-bits
9569 @opindex mno-early-stop-bits
9570 Allow stop bits to be placed earlier than immediately preceding the
9571 instruction that triggered the stop bit. This can improve instruction
9572 scheduling, but does not always do so.
9574 @item -mfixed-range=@var{register-range}
9575 @opindex mfixed-range
9576 Generate code treating the given register range as fixed registers.
9577 A fixed register is one that the register allocator can not use. This is
9578 useful when compiling kernel code. A register range is specified as
9579 two registers separated by a dash. Multiple register ranges can be
9580 specified separated by a comma.
9582 @item -mtls-size=@var{tls-size}
9584 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
9587 @item -mtune=@var{cpu-type}
9589 Tune the instruction scheduling for a particular CPU, Valid values are
9590 itanium, itanium1, merced, itanium2, and mckinley.
9596 Add support for multithreading using the POSIX threads library. This
9597 option sets flags for both the preprocessor and linker. It does
9598 not affect the thread safety of object code produced by the compiler or
9599 that of libraries supplied with it. These are HP-UX specific flags.
9605 Generate code for a 32-bit or 64-bit environment.
9606 The 32-bit environment sets int, long and pointer to 32 bits.
9607 The 64-bit environment sets int to 32 bits and long and pointer
9608 to 64 bits. These are HP-UX specific flags.
9613 @subsection M32C Options
9614 @cindex M32C options
9617 @item -mcpu=@var{name}
9619 Select the CPU for which code is generated. @var{name} may be one of
9620 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
9621 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
9626 Specifies that the program will be run on the simulator. This causes
9627 an alternate runtime library to be linked in which supports, for
9628 example, file I/O. You must not use this option when generating
9629 programs that will run on real hardware; you must provide your own
9630 runtime library for whatever I/O functions are needed.
9632 @item -memregs=@var{number}
9634 Specifies the number of memory-based pseudo-registers GCC will use
9635 during code generation. These pseudo-registers will be used like real
9636 registers, so there is a tradeoff between GCC's ability to fit the
9637 code into available registers, and the performance penalty of using
9638 memory instead of registers. Note that all modules in a program must
9639 be compiled with the same value for this option. Because of that, you
9640 must not use this option with the default runtime libraries gcc
9645 @node M32R/D Options
9646 @subsection M32R/D Options
9647 @cindex M32R/D options
9649 These @option{-m} options are defined for Renesas M32R/D architectures:
9654 Generate code for the M32R/2@.
9658 Generate code for the M32R/X@.
9662 Generate code for the M32R@. This is the default.
9665 @opindex mmodel=small
9666 Assume all objects live in the lower 16MB of memory (so that their addresses
9667 can be loaded with the @code{ld24} instruction), and assume all subroutines
9668 are reachable with the @code{bl} instruction.
9669 This is the default.
9671 The addressability of a particular object can be set with the
9672 @code{model} attribute.
9674 @item -mmodel=medium
9675 @opindex mmodel=medium
9676 Assume objects may be anywhere in the 32-bit address space (the compiler
9677 will generate @code{seth/add3} instructions to load their addresses), and
9678 assume all subroutines are reachable with the @code{bl} instruction.
9681 @opindex mmodel=large
9682 Assume objects may be anywhere in the 32-bit address space (the compiler
9683 will generate @code{seth/add3} instructions to load their addresses), and
9684 assume subroutines may not be reachable with the @code{bl} instruction
9685 (the compiler will generate the much slower @code{seth/add3/jl}
9686 instruction sequence).
9689 @opindex msdata=none
9690 Disable use of the small data area. Variables will be put into
9691 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
9692 @code{section} attribute has been specified).
9693 This is the default.
9695 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
9696 Objects may be explicitly put in the small data area with the
9697 @code{section} attribute using one of these sections.
9700 @opindex msdata=sdata
9701 Put small global and static data in the small data area, but do not
9702 generate special code to reference them.
9706 Put small global and static data in the small data area, and generate
9707 special instructions to reference them.
9711 @cindex smaller data references
9712 Put global and static objects less than or equal to @var{num} bytes
9713 into the small data or bss sections instead of the normal data or bss
9714 sections. The default value of @var{num} is 8.
9715 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
9716 for this option to have any effect.
9718 All modules should be compiled with the same @option{-G @var{num}} value.
9719 Compiling with different values of @var{num} may or may not work; if it
9720 doesn't the linker will give an error message---incorrect code will not be
9725 Makes the M32R specific code in the compiler display some statistics
9726 that might help in debugging programs.
9729 @opindex malign-loops
9730 Align all loops to a 32-byte boundary.
9732 @item -mno-align-loops
9733 @opindex mno-align-loops
9734 Do not enforce a 32-byte alignment for loops. This is the default.
9736 @item -missue-rate=@var{number}
9737 @opindex missue-rate=@var{number}
9738 Issue @var{number} instructions per cycle. @var{number} can only be 1
9741 @item -mbranch-cost=@var{number}
9742 @opindex mbranch-cost=@var{number}
9743 @var{number} can only be 1 or 2. If it is 1 then branches will be
9744 preferred over conditional code, if it is 2, then the opposite will
9747 @item -mflush-trap=@var{number}
9748 @opindex mflush-trap=@var{number}
9749 Specifies the trap number to use to flush the cache. The default is
9750 12. Valid numbers are between 0 and 15 inclusive.
9752 @item -mno-flush-trap
9753 @opindex mno-flush-trap
9754 Specifies that the cache cannot be flushed by using a trap.
9756 @item -mflush-func=@var{name}
9757 @opindex mflush-func=@var{name}
9758 Specifies the name of the operating system function to call to flush
9759 the cache. The default is @emph{_flush_cache}, but a function call
9760 will only be used if a trap is not available.
9762 @item -mno-flush-func
9763 @opindex mno-flush-func
9764 Indicates that there is no OS function for flushing the cache.
9768 @node M680x0 Options
9769 @subsection M680x0 Options
9770 @cindex M680x0 options
9772 These are the @samp{-m} options defined for the 68000 series. The default
9773 values for these options depends on which style of 68000 was selected when
9774 the compiler was configured; the defaults for the most common choices are
9782 Generate output for a 68000. This is the default
9783 when the compiler is configured for 68000-based systems.
9785 Use this option for microcontrollers with a 68000 or EC000 core,
9786 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
9792 Generate output for a 68020. This is the default
9793 when the compiler is configured for 68020-based systems.
9797 Generate output containing 68881 instructions for floating point.
9798 This is the default for most 68020 systems unless @option{--nfp} was
9799 specified when the compiler was configured.
9803 Generate output for a 68030. This is the default when the compiler is
9804 configured for 68030-based systems.
9808 Generate output for a 68040. This is the default when the compiler is
9809 configured for 68040-based systems.
9811 This option inhibits the use of 68881/68882 instructions that have to be
9812 emulated by software on the 68040. Use this option if your 68040 does not
9813 have code to emulate those instructions.
9817 Generate output for a 68060. This is the default when the compiler is
9818 configured for 68060-based systems.
9820 This option inhibits the use of 68020 and 68881/68882 instructions that
9821 have to be emulated by software on the 68060. Use this option if your 68060
9822 does not have code to emulate those instructions.
9826 Generate output for a CPU32. This is the default
9827 when the compiler is configured for CPU32-based systems.
9829 Use this option for microcontrollers with a
9830 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
9831 68336, 68340, 68341, 68349 and 68360.
9835 Generate output for a 520X ``coldfire'' family cpu. This is the default
9836 when the compiler is configured for 520X-based systems.
9838 Use this option for microcontroller with a 5200 core, including
9839 the MCF5202, MCF5203, MCF5204 and MCF5202.
9844 Generate output for a 68040, without using any of the new instructions.
9845 This results in code which can run relatively efficiently on either a
9846 68020/68881 or a 68030 or a 68040. The generated code does use the
9847 68881 instructions that are emulated on the 68040.
9851 Generate output for a 68060, without using any of the new instructions.
9852 This results in code which can run relatively efficiently on either a
9853 68020/68881 or a 68030 or a 68040. The generated code does use the
9854 68881 instructions that are emulated on the 68060.
9857 @opindex msoft-float
9858 Generate output containing library calls for floating point.
9859 @strong{Warning:} the requisite libraries are not available for all m68k
9860 targets. Normally the facilities of the machine's usual C compiler are
9861 used, but this can't be done directly in cross-compilation. You must
9862 make your own arrangements to provide suitable library functions for
9863 cross-compilation. The embedded targets @samp{m68k-*-aout} and
9864 @samp{m68k-*-coff} do provide software floating point support.
9868 Consider type @code{int} to be 16 bits wide, like @code{short int}.
9869 Additionally, parameters passed on the stack are also aligned to a
9870 16-bit boundary even on targets whose API mandates promotion to 32-bit.
9873 @opindex mnobitfield
9874 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
9875 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
9879 Do use the bit-field instructions. The @option{-m68020} option implies
9880 @option{-mbitfield}. This is the default if you use a configuration
9881 designed for a 68020.
9885 Use a different function-calling convention, in which functions
9886 that take a fixed number of arguments return with the @code{rtd}
9887 instruction, which pops their arguments while returning. This
9888 saves one instruction in the caller since there is no need to pop
9889 the arguments there.
9891 This calling convention is incompatible with the one normally
9892 used on Unix, so you cannot use it if you need to call libraries
9893 compiled with the Unix compiler.
9895 Also, you must provide function prototypes for all functions that
9896 take variable numbers of arguments (including @code{printf});
9897 otherwise incorrect code will be generated for calls to those
9900 In addition, seriously incorrect code will result if you call a
9901 function with too many arguments. (Normally, extra arguments are
9902 harmlessly ignored.)
9904 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
9905 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
9908 @itemx -mno-align-int
9910 @opindex mno-align-int
9911 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
9912 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
9913 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
9914 Aligning variables on 32-bit boundaries produces code that runs somewhat
9915 faster on processors with 32-bit busses at the expense of more memory.
9917 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
9918 align structures containing the above types differently than
9919 most published application binary interface specifications for the m68k.
9923 Use the pc-relative addressing mode of the 68000 directly, instead of
9924 using a global offset table. At present, this option implies @option{-fpic},
9925 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
9926 not presently supported with @option{-mpcrel}, though this could be supported for
9927 68020 and higher processors.
9929 @item -mno-strict-align
9930 @itemx -mstrict-align
9931 @opindex mno-strict-align
9932 @opindex mstrict-align
9933 Do not (do) assume that unaligned memory references will be handled by
9937 Generate code that allows the data segment to be located in a different
9938 area of memory from the text segment. This allows for execute in place in
9939 an environment without virtual memory management. This option implies
9943 Generate code that assumes that the data segment follows the text segment.
9944 This is the default.
9946 @item -mid-shared-library
9947 Generate code that supports shared libraries via the library ID method.
9948 This allows for execute in place and shared libraries in an environment
9949 without virtual memory management. This option implies @option{-fPIC}.
9951 @item -mno-id-shared-library
9952 Generate code that doesn't assume ID based shared libraries are being used.
9953 This is the default.
9955 @item -mshared-library-id=n
9956 Specified the identification number of the ID based shared library being
9957 compiled. Specifying a value of 0 will generate more compact code, specifying
9958 other values will force the allocation of that number to the current
9959 library but is no more space or time efficient than omitting this option.
9963 @node M68hc1x Options
9964 @subsection M68hc1x Options
9965 @cindex M68hc1x options
9967 These are the @samp{-m} options defined for the 68hc11 and 68hc12
9968 microcontrollers. The default values for these options depends on
9969 which style of microcontroller was selected when the compiler was configured;
9970 the defaults for the most common choices are given below.
9977 Generate output for a 68HC11. This is the default
9978 when the compiler is configured for 68HC11-based systems.
9984 Generate output for a 68HC12. This is the default
9985 when the compiler is configured for 68HC12-based systems.
9991 Generate output for a 68HCS12.
9994 @opindex mauto-incdec
9995 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
10002 Enable the use of 68HC12 min and max instructions.
10005 @itemx -mno-long-calls
10006 @opindex mlong-calls
10007 @opindex mno-long-calls
10008 Treat all calls as being far away (near). If calls are assumed to be
10009 far away, the compiler will use the @code{call} instruction to
10010 call a function and the @code{rtc} instruction for returning.
10014 Consider type @code{int} to be 16 bits wide, like @code{short int}.
10016 @item -msoft-reg-count=@var{count}
10017 @opindex msoft-reg-count
10018 Specify the number of pseudo-soft registers which are used for the
10019 code generation. The maximum number is 32. Using more pseudo-soft
10020 register may or may not result in better code depending on the program.
10021 The default is 4 for 68HC11 and 2 for 68HC12.
10025 @node MCore Options
10026 @subsection MCore Options
10027 @cindex MCore options
10029 These are the @samp{-m} options defined for the Motorola M*Core
10035 @itemx -mno-hardlit
10037 @opindex mno-hardlit
10038 Inline constants into the code stream if it can be done in two
10039 instructions or less.
10045 Use the divide instruction. (Enabled by default).
10047 @item -mrelax-immediate
10048 @itemx -mno-relax-immediate
10049 @opindex mrelax-immediate
10050 @opindex mno-relax-immediate
10051 Allow arbitrary sized immediates in bit operations.
10053 @item -mwide-bitfields
10054 @itemx -mno-wide-bitfields
10055 @opindex mwide-bitfields
10056 @opindex mno-wide-bitfields
10057 Always treat bit-fields as int-sized.
10059 @item -m4byte-functions
10060 @itemx -mno-4byte-functions
10061 @opindex m4byte-functions
10062 @opindex mno-4byte-functions
10063 Force all functions to be aligned to a four byte boundary.
10065 @item -mcallgraph-data
10066 @itemx -mno-callgraph-data
10067 @opindex mcallgraph-data
10068 @opindex mno-callgraph-data
10069 Emit callgraph information.
10072 @itemx -mno-slow-bytes
10073 @opindex mslow-bytes
10074 @opindex mno-slow-bytes
10075 Prefer word access when reading byte quantities.
10077 @item -mlittle-endian
10078 @itemx -mbig-endian
10079 @opindex mlittle-endian
10080 @opindex mbig-endian
10081 Generate code for a little endian target.
10087 Generate code for the 210 processor.
10091 @subsection MIPS Options
10092 @cindex MIPS options
10098 Generate big-endian code.
10102 Generate little-endian code. This is the default for @samp{mips*el-*-*}
10105 @item -march=@var{arch}
10107 Generate code that will run on @var{arch}, which can be the name of a
10108 generic MIPS ISA, or the name of a particular processor.
10110 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
10111 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
10112 The processor names are:
10113 @samp{4kc}, @samp{4km}, @samp{4kp},
10114 @samp{5kc}, @samp{5kf},
10116 @samp{24k}, @samp{24kc}, @samp{24kf}, @samp{24kx},
10119 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
10120 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
10121 @samp{rm7000}, @samp{rm9000},
10124 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
10125 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
10126 The special value @samp{from-abi} selects the
10127 most compatible architecture for the selected ABI (that is,
10128 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
10130 In processor names, a final @samp{000} can be abbreviated as @samp{k}
10131 (for example, @samp{-march=r2k}). Prefixes are optional, and
10132 @samp{vr} may be written @samp{r}.
10134 GCC defines two macros based on the value of this option. The first
10135 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
10136 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
10137 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
10138 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
10139 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
10141 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
10142 above. In other words, it will have the full prefix and will not
10143 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
10144 the macro names the resolved architecture (either @samp{"mips1"} or
10145 @samp{"mips3"}). It names the default architecture when no
10146 @option{-march} option is given.
10148 @item -mtune=@var{arch}
10150 Optimize for @var{arch}. Among other things, this option controls
10151 the way instructions are scheduled, and the perceived cost of arithmetic
10152 operations. The list of @var{arch} values is the same as for
10155 When this option is not used, GCC will optimize for the processor
10156 specified by @option{-march}. By using @option{-march} and
10157 @option{-mtune} together, it is possible to generate code that will
10158 run on a family of processors, but optimize the code for one
10159 particular member of that family.
10161 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
10162 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
10163 @samp{-march} ones described above.
10167 Equivalent to @samp{-march=mips1}.
10171 Equivalent to @samp{-march=mips2}.
10175 Equivalent to @samp{-march=mips3}.
10179 Equivalent to @samp{-march=mips4}.
10183 Equivalent to @samp{-march=mips32}.
10187 Equivalent to @samp{-march=mips32r2}.
10191 Equivalent to @samp{-march=mips64}.
10196 @opindex mno-mips16
10197 Generate (do not generate) MIPS16 code. If GCC is targetting a
10198 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
10210 Generate code for the given ABI@.
10212 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
10213 generates 64-bit code when you select a 64-bit architecture, but you
10214 can use @option{-mgp32} to get 32-bit code instead.
10216 For information about the O64 ABI, see
10217 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
10220 @itemx -mno-abicalls
10222 @opindex mno-abicalls
10223 Generate (do not generate) SVR4-style position-independent code.
10224 @option{-mabicalls} is the default for SVR4-based systems.
10230 Lift (do not lift) the usual restrictions on the size of the global
10233 GCC normally uses a single instruction to load values from the GOT@.
10234 While this is relatively efficient, it will only work if the GOT
10235 is smaller than about 64k. Anything larger will cause the linker
10236 to report an error such as:
10238 @cindex relocation truncated to fit (MIPS)
10240 relocation truncated to fit: R_MIPS_GOT16 foobar
10243 If this happens, you should recompile your code with @option{-mxgot}.
10244 It should then work with very large GOTs, although it will also be
10245 less efficient, since it will take three instructions to fetch the
10246 value of a global symbol.
10248 Note that some linkers can create multiple GOTs. If you have such a
10249 linker, you should only need to use @option{-mxgot} when a single object
10250 file accesses more than 64k's worth of GOT entries. Very few do.
10252 These options have no effect unless GCC is generating position
10257 Assume that general-purpose registers are 32 bits wide.
10261 Assume that general-purpose registers are 64 bits wide.
10265 Assume that floating-point registers are 32 bits wide.
10269 Assume that floating-point registers are 64 bits wide.
10272 @opindex mhard-float
10273 Use floating-point coprocessor instructions.
10276 @opindex msoft-float
10277 Do not use floating-point coprocessor instructions. Implement
10278 floating-point calculations using library calls instead.
10280 @item -msingle-float
10281 @opindex msingle-float
10282 Assume that the floating-point coprocessor only supports single-precision
10285 @itemx -mdouble-float
10286 @opindex mdouble-float
10287 Assume that the floating-point coprocessor supports double-precision
10288 operations. This is the default.
10294 Use (do not use) the MIPS DSP ASE. @xref{MIPS DSP Built-in Functions}.
10296 @itemx -mpaired-single
10297 @itemx -mno-paired-single
10298 @opindex mpaired-single
10299 @opindex mno-paired-single
10300 Use (do not use) paired-single floating-point instructions.
10301 @xref{MIPS Paired-Single Support}. This option can only be used
10302 when generating 64-bit code and requires hardware floating-point
10303 support to be enabled.
10308 @opindex mno-mips3d
10309 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
10310 The option @option{-mips3d} implies @option{-mpaired-single}.
10314 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
10315 an explanation of the default and the way that the pointer size is
10320 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
10322 The default size of @code{int}s, @code{long}s and pointers depends on
10323 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
10324 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
10325 32-bit @code{long}s. Pointers are the same size as @code{long}s,
10326 or the same size as integer registers, whichever is smaller.
10332 Assume (do not assume) that all symbols have 32-bit values, regardless
10333 of the selected ABI@. This option is useful in combination with
10334 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
10335 to generate shorter and faster references to symbolic addresses.
10339 @cindex smaller data references (MIPS)
10340 @cindex gp-relative references (MIPS)
10341 Put global and static items less than or equal to @var{num} bytes into
10342 the small data or bss section instead of the normal data or bss section.
10343 This allows the data to be accessed using a single instruction.
10345 All modules should be compiled with the same @option{-G @var{num}}
10348 @item -membedded-data
10349 @itemx -mno-embedded-data
10350 @opindex membedded-data
10351 @opindex mno-embedded-data
10352 Allocate variables to the read-only data section first if possible, then
10353 next in the small data section if possible, otherwise in data. This gives
10354 slightly slower code than the default, but reduces the amount of RAM required
10355 when executing, and thus may be preferred for some embedded systems.
10357 @item -muninit-const-in-rodata
10358 @itemx -mno-uninit-const-in-rodata
10359 @opindex muninit-const-in-rodata
10360 @opindex mno-uninit-const-in-rodata
10361 Put uninitialized @code{const} variables in the read-only data section.
10362 This option is only meaningful in conjunction with @option{-membedded-data}.
10364 @item -msplit-addresses
10365 @itemx -mno-split-addresses
10366 @opindex msplit-addresses
10367 @opindex mno-split-addresses
10368 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
10369 relocation operators. This option has been superseded by
10370 @option{-mexplicit-relocs} but is retained for backwards compatibility.
10372 @item -mexplicit-relocs
10373 @itemx -mno-explicit-relocs
10374 @opindex mexplicit-relocs
10375 @opindex mno-explicit-relocs
10376 Use (do not use) assembler relocation operators when dealing with symbolic
10377 addresses. The alternative, selected by @option{-mno-explicit-relocs},
10378 is to use assembler macros instead.
10380 @option{-mexplicit-relocs} is the default if GCC was configured
10381 to use an assembler that supports relocation operators.
10383 @item -mcheck-zero-division
10384 @itemx -mno-check-zero-division
10385 @opindex mcheck-zero-division
10386 @opindex mno-check-zero-division
10387 Trap (do not trap) on integer division by zero. The default is
10388 @option{-mcheck-zero-division}.
10390 @item -mdivide-traps
10391 @itemx -mdivide-breaks
10392 @opindex mdivide-traps
10393 @opindex mdivide-breaks
10394 MIPS systems check for division by zero by generating either a
10395 conditional trap or a break instruction. Using traps results in
10396 smaller code, but is only supported on MIPS II and later. Also, some
10397 versions of the Linux kernel have a bug that prevents trap from
10398 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
10399 allow conditional traps on architectures that support them and
10400 @option{-mdivide-breaks} to force the use of breaks.
10402 The default is usually @option{-mdivide-traps}, but this can be
10403 overridden at configure time using @option{--with-divide=breaks}.
10404 Divide-by-zero checks can be completely disabled using
10405 @option{-mno-check-zero-division}.
10410 @opindex mno-memcpy
10411 Force (do not force) the use of @code{memcpy()} for non-trivial block
10412 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
10413 most constant-sized copies.
10416 @itemx -mno-long-calls
10417 @opindex mlong-calls
10418 @opindex mno-long-calls
10419 Disable (do not disable) use of the @code{jal} instruction. Calling
10420 functions using @code{jal} is more efficient but requires the caller
10421 and callee to be in the same 256 megabyte segment.
10423 This option has no effect on abicalls code. The default is
10424 @option{-mno-long-calls}.
10430 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
10431 instructions, as provided by the R4650 ISA@.
10434 @itemx -mno-fused-madd
10435 @opindex mfused-madd
10436 @opindex mno-fused-madd
10437 Enable (disable) use of the floating point multiply-accumulate
10438 instructions, when they are available. The default is
10439 @option{-mfused-madd}.
10441 When multiply-accumulate instructions are used, the intermediate
10442 product is calculated to infinite precision and is not subject to
10443 the FCSR Flush to Zero bit. This may be undesirable in some
10448 Tell the MIPS assembler to not run its preprocessor over user
10449 assembler files (with a @samp{.s} suffix) when assembling them.
10452 @itemx -mno-fix-r4000
10453 @opindex mfix-r4000
10454 @opindex mno-fix-r4000
10455 Work around certain R4000 CPU errata:
10458 A double-word or a variable shift may give an incorrect result if executed
10459 immediately after starting an integer division.
10461 A double-word or a variable shift may give an incorrect result if executed
10462 while an integer multiplication is in progress.
10464 An integer division may give an incorrect result if started in a delay slot
10465 of a taken branch or a jump.
10469 @itemx -mno-fix-r4400
10470 @opindex mfix-r4400
10471 @opindex mno-fix-r4400
10472 Work around certain R4400 CPU errata:
10475 A double-word or a variable shift may give an incorrect result if executed
10476 immediately after starting an integer division.
10480 @itemx -mno-fix-vr4120
10481 @opindex mfix-vr4120
10482 Work around certain VR4120 errata:
10485 @code{dmultu} does not always produce the correct result.
10487 @code{div} and @code{ddiv} do not always produce the correct result if one
10488 of the operands is negative.
10490 The workarounds for the division errata rely on special functions in
10491 @file{libgcc.a}. At present, these functions are only provided by
10492 the @code{mips64vr*-elf} configurations.
10494 Other VR4120 errata require a nop to be inserted between certain pairs of
10495 instructions. These errata are handled by the assembler, not by GCC itself.
10498 @opindex mfix-vr4130
10499 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
10500 workarounds are implemented by the assembler rather than by GCC,
10501 although GCC will avoid using @code{mflo} and @code{mfhi} if the
10502 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
10503 instructions are available instead.
10506 @itemx -mno-fix-sb1
10508 Work around certain SB-1 CPU core errata.
10509 (This flag currently works around the SB-1 revision 2
10510 ``F1'' and ``F2'' floating point errata.)
10512 @item -mflush-func=@var{func}
10513 @itemx -mno-flush-func
10514 @opindex mflush-func
10515 Specifies the function to call to flush the I and D caches, or to not
10516 call any such function. If called, the function must take the same
10517 arguments as the common @code{_flush_func()}, that is, the address of the
10518 memory range for which the cache is being flushed, the size of the
10519 memory range, and the number 3 (to flush both caches). The default
10520 depends on the target GCC was configured for, but commonly is either
10521 @samp{_flush_func} or @samp{__cpu_flush}.
10523 @item -mbranch-likely
10524 @itemx -mno-branch-likely
10525 @opindex mbranch-likely
10526 @opindex mno-branch-likely
10527 Enable or disable use of Branch Likely instructions, regardless of the
10528 default for the selected architecture. By default, Branch Likely
10529 instructions may be generated if they are supported by the selected
10530 architecture. An exception is for the MIPS32 and MIPS64 architectures
10531 and processors which implement those architectures; for those, Branch
10532 Likely instructions will not be generated by default because the MIPS32
10533 and MIPS64 architectures specifically deprecate their use.
10535 @item -mfp-exceptions
10536 @itemx -mno-fp-exceptions
10537 @opindex mfp-exceptions
10538 Specifies whether FP exceptions are enabled. This affects how we schedule
10539 FP instructions for some processors. The default is that FP exceptions are
10542 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
10543 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
10546 @item -mvr4130-align
10547 @itemx -mno-vr4130-align
10548 @opindex mvr4130-align
10549 The VR4130 pipeline is two-way superscalar, but can only issue two
10550 instructions together if the first one is 8-byte aligned. When this
10551 option is enabled, GCC will align pairs of instructions that it
10552 thinks should execute in parallel.
10554 This option only has an effect when optimizing for the VR4130.
10555 It normally makes code faster, but at the expense of making it bigger.
10556 It is enabled by default at optimization level @option{-O3}.
10560 @subsection MMIX Options
10561 @cindex MMIX Options
10563 These options are defined for the MMIX:
10567 @itemx -mno-libfuncs
10569 @opindex mno-libfuncs
10570 Specify that intrinsic library functions are being compiled, passing all
10571 values in registers, no matter the size.
10574 @itemx -mno-epsilon
10576 @opindex mno-epsilon
10577 Generate floating-point comparison instructions that compare with respect
10578 to the @code{rE} epsilon register.
10580 @item -mabi=mmixware
10582 @opindex mabi-mmixware
10584 Generate code that passes function parameters and return values that (in
10585 the called function) are seen as registers @code{$0} and up, as opposed to
10586 the GNU ABI which uses global registers @code{$231} and up.
10588 @item -mzero-extend
10589 @itemx -mno-zero-extend
10590 @opindex mzero-extend
10591 @opindex mno-zero-extend
10592 When reading data from memory in sizes shorter than 64 bits, use (do not
10593 use) zero-extending load instructions by default, rather than
10594 sign-extending ones.
10597 @itemx -mno-knuthdiv
10599 @opindex mno-knuthdiv
10600 Make the result of a division yielding a remainder have the same sign as
10601 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
10602 remainder follows the sign of the dividend. Both methods are
10603 arithmetically valid, the latter being almost exclusively used.
10605 @item -mtoplevel-symbols
10606 @itemx -mno-toplevel-symbols
10607 @opindex mtoplevel-symbols
10608 @opindex mno-toplevel-symbols
10609 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
10610 code can be used with the @code{PREFIX} assembly directive.
10614 Generate an executable in the ELF format, rather than the default
10615 @samp{mmo} format used by the @command{mmix} simulator.
10617 @item -mbranch-predict
10618 @itemx -mno-branch-predict
10619 @opindex mbranch-predict
10620 @opindex mno-branch-predict
10621 Use (do not use) the probable-branch instructions, when static branch
10622 prediction indicates a probable branch.
10624 @item -mbase-addresses
10625 @itemx -mno-base-addresses
10626 @opindex mbase-addresses
10627 @opindex mno-base-addresses
10628 Generate (do not generate) code that uses @emph{base addresses}. Using a
10629 base address automatically generates a request (handled by the assembler
10630 and the linker) for a constant to be set up in a global register. The
10631 register is used for one or more base address requests within the range 0
10632 to 255 from the value held in the register. The generally leads to short
10633 and fast code, but the number of different data items that can be
10634 addressed is limited. This means that a program that uses lots of static
10635 data may require @option{-mno-base-addresses}.
10637 @item -msingle-exit
10638 @itemx -mno-single-exit
10639 @opindex msingle-exit
10640 @opindex mno-single-exit
10641 Force (do not force) generated code to have a single exit point in each
10645 @node MN10300 Options
10646 @subsection MN10300 Options
10647 @cindex MN10300 options
10649 These @option{-m} options are defined for Matsushita MN10300 architectures:
10654 Generate code to avoid bugs in the multiply instructions for the MN10300
10655 processors. This is the default.
10657 @item -mno-mult-bug
10658 @opindex mno-mult-bug
10659 Do not generate code to avoid bugs in the multiply instructions for the
10660 MN10300 processors.
10664 Generate code which uses features specific to the AM33 processor.
10668 Do not generate code which uses features specific to the AM33 processor. This
10671 @item -mreturn-pointer-on-d0
10672 @opindex mreturn-pointer-on-d0
10673 When generating a function which returns a pointer, return the pointer
10674 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
10675 only in a0, and attempts to call such functions without a prototype
10676 would result in errors. Note that this option is on by default; use
10677 @option{-mno-return-pointer-on-d0} to disable it.
10681 Do not link in the C run-time initialization object file.
10685 Indicate to the linker that it should perform a relaxation optimization pass
10686 to shorten branches, calls and absolute memory addresses. This option only
10687 has an effect when used on the command line for the final link step.
10689 This option makes symbolic debugging impossible.
10693 @subsection MT Options
10696 These @option{-m} options are defined for Morpho MT architectures:
10700 @item -march=@var{cpu-type}
10702 Generate code that will run on @var{cpu-type}, which is the name of a system
10703 representing a certain processor type. Possible values for
10704 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
10705 @samp{ms1-16-003} and @samp{ms2}.
10707 When this option is not used, the default is @option{-march=ms1-16-002}.
10711 Use byte loads and stores when generating code.
10715 Do not use byte loads and stores when generating code.
10719 Use simulator runtime
10723 Do not link in the C run-time initialization object file
10724 @file{crti.o}. Other run-time initialization and termination files
10725 such as @file{startup.o} and @file{exit.o} are still included on the
10726 linker command line.
10730 @node PDP-11 Options
10731 @subsection PDP-11 Options
10732 @cindex PDP-11 Options
10734 These options are defined for the PDP-11:
10739 Use hardware FPP floating point. This is the default. (FIS floating
10740 point on the PDP-11/40 is not supported.)
10743 @opindex msoft-float
10744 Do not use hardware floating point.
10748 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
10752 Return floating-point results in memory. This is the default.
10756 Generate code for a PDP-11/40.
10760 Generate code for a PDP-11/45. This is the default.
10764 Generate code for a PDP-11/10.
10766 @item -mbcopy-builtin
10767 @opindex bcopy-builtin
10768 Use inline @code{movmemhi} patterns for copying memory. This is the
10773 Do not use inline @code{movmemhi} patterns for copying memory.
10779 Use 16-bit @code{int}. This is the default.
10785 Use 32-bit @code{int}.
10788 @itemx -mno-float32
10790 @opindex mno-float32
10791 Use 64-bit @code{float}. This is the default.
10794 @itemx -mno-float64
10796 @opindex mno-float64
10797 Use 32-bit @code{float}.
10801 Use @code{abshi2} pattern. This is the default.
10805 Do not use @code{abshi2} pattern.
10807 @item -mbranch-expensive
10808 @opindex mbranch-expensive
10809 Pretend that branches are expensive. This is for experimenting with
10810 code generation only.
10812 @item -mbranch-cheap
10813 @opindex mbranch-cheap
10814 Do not pretend that branches are expensive. This is the default.
10818 Generate code for a system with split I&D@.
10822 Generate code for a system without split I&D@. This is the default.
10826 Use Unix assembler syntax. This is the default when configured for
10827 @samp{pdp11-*-bsd}.
10831 Use DEC assembler syntax. This is the default when configured for any
10832 PDP-11 target other than @samp{pdp11-*-bsd}.
10835 @node PowerPC Options
10836 @subsection PowerPC Options
10837 @cindex PowerPC options
10839 These are listed under @xref{RS/6000 and PowerPC Options}.
10841 @node RS/6000 and PowerPC Options
10842 @subsection IBM RS/6000 and PowerPC Options
10843 @cindex RS/6000 and PowerPC Options
10844 @cindex IBM RS/6000 and PowerPC Options
10846 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
10853 @itemx -mno-powerpc
10854 @itemx -mpowerpc-gpopt
10855 @itemx -mno-powerpc-gpopt
10856 @itemx -mpowerpc-gfxopt
10857 @itemx -mno-powerpc-gfxopt
10859 @itemx -mno-powerpc64
10863 @itemx -mno-popcntb
10869 @opindex mno-power2
10871 @opindex mno-powerpc
10872 @opindex mpowerpc-gpopt
10873 @opindex mno-powerpc-gpopt
10874 @opindex mpowerpc-gfxopt
10875 @opindex mno-powerpc-gfxopt
10876 @opindex mpowerpc64
10877 @opindex mno-powerpc64
10881 @opindex mno-popcntb
10884 GCC supports two related instruction set architectures for the
10885 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
10886 instructions supported by the @samp{rios} chip set used in the original
10887 RS/6000 systems and the @dfn{PowerPC} instruction set is the
10888 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
10889 the IBM 4xx, 6xx, and follow-on microprocessors.
10891 Neither architecture is a subset of the other. However there is a
10892 large common subset of instructions supported by both. An MQ
10893 register is included in processors supporting the POWER architecture.
10895 You use these options to specify which instructions are available on the
10896 processor you are using. The default value of these options is
10897 determined when configuring GCC@. Specifying the
10898 @option{-mcpu=@var{cpu_type}} overrides the specification of these
10899 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
10900 rather than the options listed above.
10902 The @option{-mpower} option allows GCC to generate instructions that
10903 are found only in the POWER architecture and to use the MQ register.
10904 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
10905 to generate instructions that are present in the POWER2 architecture but
10906 not the original POWER architecture.
10908 The @option{-mpowerpc} option allows GCC to generate instructions that
10909 are found only in the 32-bit subset of the PowerPC architecture.
10910 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
10911 GCC to use the optional PowerPC architecture instructions in the
10912 General Purpose group, including floating-point square root. Specifying
10913 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
10914 use the optional PowerPC architecture instructions in the Graphics
10915 group, including floating-point select.
10917 The @option{-mmfcrf} option allows GCC to generate the move from
10918 condition register field instruction implemented on the POWER4
10919 processor and other processors that support the PowerPC V2.01
10921 The @option{-mpopcntb} option allows GCC to generate the popcount and
10922 double precision FP reciprocal estimate instruction implemented on the
10923 POWER5 processor and other processors that support the PowerPC V2.02
10925 The @option{-mfprnd} option allows GCC to generate the FP round to
10926 integer instructions implemented on the POWER5+ processor and other
10927 processors that support the PowerPC V2.03 architecture.
10929 The @option{-mpowerpc64} option allows GCC to generate the additional
10930 64-bit instructions that are found in the full PowerPC64 architecture
10931 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
10932 @option{-mno-powerpc64}.
10934 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
10935 will use only the instructions in the common subset of both
10936 architectures plus some special AIX common-mode calls, and will not use
10937 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
10938 permits GCC to use any instruction from either architecture and to
10939 allow use of the MQ register; specify this for the Motorola MPC601.
10941 @item -mnew-mnemonics
10942 @itemx -mold-mnemonics
10943 @opindex mnew-mnemonics
10944 @opindex mold-mnemonics
10945 Select which mnemonics to use in the generated assembler code. With
10946 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
10947 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
10948 assembler mnemonics defined for the POWER architecture. Instructions
10949 defined in only one architecture have only one mnemonic; GCC uses that
10950 mnemonic irrespective of which of these options is specified.
10952 GCC defaults to the mnemonics appropriate for the architecture in
10953 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
10954 value of these option. Unless you are building a cross-compiler, you
10955 should normally not specify either @option{-mnew-mnemonics} or
10956 @option{-mold-mnemonics}, but should instead accept the default.
10958 @item -mcpu=@var{cpu_type}
10960 Set architecture type, register usage, choice of mnemonics, and
10961 instruction scheduling parameters for machine type @var{cpu_type}.
10962 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
10963 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
10964 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
10965 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
10966 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
10967 @samp{860}, @samp{970}, @samp{8540}, @samp{ec603e}, @samp{G3},
10968 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
10969 @samp{power4}, @samp{power5}, @samp{power5+},
10970 @samp{common}, @samp{powerpc}, @samp{powerpc64},
10971 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
10973 @option{-mcpu=common} selects a completely generic processor. Code
10974 generated under this option will run on any POWER or PowerPC processor.
10975 GCC will use only the instructions in the common subset of both
10976 architectures, and will not use the MQ register. GCC assumes a generic
10977 processor model for scheduling purposes.
10979 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
10980 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
10981 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
10982 types, with an appropriate, generic processor model assumed for
10983 scheduling purposes.
10985 The other options specify a specific processor. Code generated under
10986 those options will run best on that processor, and may not run at all on
10989 The @option{-mcpu} options automatically enable or disable the
10990 following options: @option{-maltivec}, @option{-mfprnd},
10991 @option{-mhard-float}, @option{-mmfcrf}, @option{-mmultiple},
10992 @option{-mnew-mnemonics}, @option{-mpopcntb}, @option{-mpower},
10993 @option{-mpower2}, @option{-mpowerpc64}, @option{-mpowerpc-gpopt},
10994 @option{-mpowerpc-gfxopt}, @option{-mstring}, @option{-mmulhw}.
10995 The particular options
10996 set for any particular CPU will vary between compiler versions,
10997 depending on what setting seems to produce optimal code for that CPU;
10998 it doesn't necessarily reflect the actual hardware's capabilities. If
10999 you wish to set an individual option to a particular value, you may
11000 specify it after the @option{-mcpu} option, like @samp{-mcpu=970
11003 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
11004 not enabled or disabled by the @option{-mcpu} option at present because
11005 AIX does not have full support for these options. You may still
11006 enable or disable them individually if you're sure it'll work in your
11009 @item -mtune=@var{cpu_type}
11011 Set the instruction scheduling parameters for machine type
11012 @var{cpu_type}, but do not set the architecture type, register usage, or
11013 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
11014 values for @var{cpu_type} are used for @option{-mtune} as for
11015 @option{-mcpu}. If both are specified, the code generated will use the
11016 architecture, registers, and mnemonics set by @option{-mcpu}, but the
11017 scheduling parameters set by @option{-mtune}.
11023 Generate code to compute division as reciprocal estimate and iterative
11024 refinement, creating opportunities for increased throughput. This
11025 feature requires: optional PowerPC Graphics instruction set for single
11026 precision and FRE instruction for double precision, assuming divides
11027 cannot generate user-visible traps, and the domain values not include
11028 Infinities, denormals or zero denominator.
11031 @itemx -mno-altivec
11033 @opindex mno-altivec
11034 Generate code that uses (does not use) AltiVec instructions, and also
11035 enable the use of built-in functions that allow more direct access to
11036 the AltiVec instruction set. You may also need to set
11037 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
11043 @opindex mno-vrsave
11044 Generate VRSAVE instructions when generating AltiVec code.
11048 Extend the current ABI with SPE ABI extensions. This does not change
11049 the default ABI, instead it adds the SPE ABI extensions to the current
11053 @opindex mabi=no-spe
11054 Disable Booke SPE ABI extensions for the current ABI@.
11057 @opindex msecure-plt
11058 Generate code that allows ld and ld.so to build executables and shared
11059 libraries with non-exec .plt and .got sections. This is a PowerPC
11060 32-bit SYSV ABI option.
11064 Generate code that uses a BSS .plt section that ld.so fills in, and
11065 requires .plt and .got sections that are both writable and executable.
11066 This is a PowerPC 32-bit SYSV ABI option.
11072 This switch enables or disables the generation of ISEL instructions.
11074 @item -misel=@var{yes/no}
11075 This switch has been deprecated. Use @option{-misel} and
11076 @option{-mno-isel} instead.
11082 This switch enables or disables the generation of SPE simd
11085 @item -mspe=@var{yes/no}
11086 This option has been deprecated. Use @option{-mspe} and
11087 @option{-mno-spe} instead.
11089 @item -mfloat-gprs=@var{yes/single/double/no}
11090 @itemx -mfloat-gprs
11091 @opindex mfloat-gprs
11092 This switch enables or disables the generation of floating point
11093 operations on the general purpose registers for architectures that
11096 The argument @var{yes} or @var{single} enables the use of
11097 single-precision floating point operations.
11099 The argument @var{double} enables the use of single and
11100 double-precision floating point operations.
11102 The argument @var{no} disables floating point operations on the
11103 general purpose registers.
11105 This option is currently only available on the MPC854x.
11111 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
11112 targets (including GNU/Linux). The 32-bit environment sets int, long
11113 and pointer to 32 bits and generates code that runs on any PowerPC
11114 variant. The 64-bit environment sets int to 32 bits and long and
11115 pointer to 64 bits, and generates code for PowerPC64, as for
11116 @option{-mpowerpc64}.
11119 @itemx -mno-fp-in-toc
11120 @itemx -mno-sum-in-toc
11121 @itemx -mminimal-toc
11123 @opindex mno-fp-in-toc
11124 @opindex mno-sum-in-toc
11125 @opindex mminimal-toc
11126 Modify generation of the TOC (Table Of Contents), which is created for
11127 every executable file. The @option{-mfull-toc} option is selected by
11128 default. In that case, GCC will allocate at least one TOC entry for
11129 each unique non-automatic variable reference in your program. GCC
11130 will also place floating-point constants in the TOC@. However, only
11131 16,384 entries are available in the TOC@.
11133 If you receive a linker error message that saying you have overflowed
11134 the available TOC space, you can reduce the amount of TOC space used
11135 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
11136 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
11137 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
11138 generate code to calculate the sum of an address and a constant at
11139 run-time instead of putting that sum into the TOC@. You may specify one
11140 or both of these options. Each causes GCC to produce very slightly
11141 slower and larger code at the expense of conserving TOC space.
11143 If you still run out of space in the TOC even when you specify both of
11144 these options, specify @option{-mminimal-toc} instead. This option causes
11145 GCC to make only one TOC entry for every file. When you specify this
11146 option, GCC will produce code that is slower and larger but which
11147 uses extremely little TOC space. You may wish to use this option
11148 only on files that contain less frequently executed code.
11154 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
11155 @code{long} type, and the infrastructure needed to support them.
11156 Specifying @option{-maix64} implies @option{-mpowerpc64} and
11157 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
11158 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
11161 @itemx -mno-xl-compat
11162 @opindex mxl-compat
11163 @opindex mno-xl-compat
11164 Produce code that conforms more closely to IBM XLC semantics when using
11165 AIX-compatible ABI. Pass floating-point arguments to prototyped
11166 functions beyond the register save area (RSA) on the stack in addition
11167 to argument FPRs. Do not assume that most significant double in 128
11168 bit long double value is properly rounded when comparing values.
11170 The AIX calling convention was extended but not initially documented to
11171 handle an obscure K&R C case of calling a function that takes the
11172 address of its arguments with fewer arguments than declared. AIX XL
11173 compilers access floating point arguments which do not fit in the
11174 RSA from the stack when a subroutine is compiled without
11175 optimization. Because always storing floating-point arguments on the
11176 stack is inefficient and rarely needed, this option is not enabled by
11177 default and only is necessary when calling subroutines compiled by AIX
11178 XL compilers without optimization.
11182 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
11183 application written to use message passing with special startup code to
11184 enable the application to run. The system must have PE installed in the
11185 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
11186 must be overridden with the @option{-specs=} option to specify the
11187 appropriate directory location. The Parallel Environment does not
11188 support threads, so the @option{-mpe} option and the @option{-pthread}
11189 option are incompatible.
11191 @item -malign-natural
11192 @itemx -malign-power
11193 @opindex malign-natural
11194 @opindex malign-power
11195 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
11196 @option{-malign-natural} overrides the ABI-defined alignment of larger
11197 types, such as floating-point doubles, on their natural size-based boundary.
11198 The option @option{-malign-power} instructs GCC to follow the ABI-specified
11199 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
11201 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
11205 @itemx -mhard-float
11206 @opindex msoft-float
11207 @opindex mhard-float
11208 Generate code that does not use (uses) the floating-point register set.
11209 Software floating point emulation is provided if you use the
11210 @option{-msoft-float} option, and pass the option to GCC when linking.
11213 @itemx -mno-multiple
11215 @opindex mno-multiple
11216 Generate code that uses (does not use) the load multiple word
11217 instructions and the store multiple word instructions. These
11218 instructions are generated by default on POWER systems, and not
11219 generated on PowerPC systems. Do not use @option{-mmultiple} on little
11220 endian PowerPC systems, since those instructions do not work when the
11221 processor is in little endian mode. The exceptions are PPC740 and
11222 PPC750 which permit the instructions usage in little endian mode.
11227 @opindex mno-string
11228 Generate code that uses (does not use) the load string instructions
11229 and the store string word instructions to save multiple registers and
11230 do small block moves. These instructions are generated by default on
11231 POWER systems, and not generated on PowerPC systems. Do not use
11232 @option{-mstring} on little endian PowerPC systems, since those
11233 instructions do not work when the processor is in little endian mode.
11234 The exceptions are PPC740 and PPC750 which permit the instructions
11235 usage in little endian mode.
11240 @opindex mno-update
11241 Generate code that uses (does not use) the load or store instructions
11242 that update the base register to the address of the calculated memory
11243 location. These instructions are generated by default. If you use
11244 @option{-mno-update}, there is a small window between the time that the
11245 stack pointer is updated and the address of the previous frame is
11246 stored, which means code that walks the stack frame across interrupts or
11247 signals may get corrupted data.
11250 @itemx -mno-fused-madd
11251 @opindex mfused-madd
11252 @opindex mno-fused-madd
11253 Generate code that uses (does not use) the floating point multiply and
11254 accumulate instructions. These instructions are generated by default if
11255 hardware floating is used.
11261 Generate code that uses (does not use) the half-word multiply and
11262 multiply-accumulate instructions on the IBM 405 and 440 processors.
11263 These instructions are generated by default when targetting those
11266 @item -mno-bit-align
11268 @opindex mno-bit-align
11269 @opindex mbit-align
11270 On System V.4 and embedded PowerPC systems do not (do) force structures
11271 and unions that contain bit-fields to be aligned to the base type of the
11274 For example, by default a structure containing nothing but 8
11275 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
11276 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
11277 the structure would be aligned to a 1 byte boundary and be one byte in
11280 @item -mno-strict-align
11281 @itemx -mstrict-align
11282 @opindex mno-strict-align
11283 @opindex mstrict-align
11284 On System V.4 and embedded PowerPC systems do not (do) assume that
11285 unaligned memory references will be handled by the system.
11287 @item -mrelocatable
11288 @itemx -mno-relocatable
11289 @opindex mrelocatable
11290 @opindex mno-relocatable
11291 On embedded PowerPC systems generate code that allows (does not allow)
11292 the program to be relocated to a different address at runtime. If you
11293 use @option{-mrelocatable} on any module, all objects linked together must
11294 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
11296 @item -mrelocatable-lib
11297 @itemx -mno-relocatable-lib
11298 @opindex mrelocatable-lib
11299 @opindex mno-relocatable-lib
11300 On embedded PowerPC systems generate code that allows (does not allow)
11301 the program to be relocated to a different address at runtime. Modules
11302 compiled with @option{-mrelocatable-lib} can be linked with either modules
11303 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
11304 with modules compiled with the @option{-mrelocatable} options.
11310 On System V.4 and embedded PowerPC systems do not (do) assume that
11311 register 2 contains a pointer to a global area pointing to the addresses
11312 used in the program.
11315 @itemx -mlittle-endian
11317 @opindex mlittle-endian
11318 On System V.4 and embedded PowerPC systems compile code for the
11319 processor in little endian mode. The @option{-mlittle-endian} option is
11320 the same as @option{-mlittle}.
11323 @itemx -mbig-endian
11325 @opindex mbig-endian
11326 On System V.4 and embedded PowerPC systems compile code for the
11327 processor in big endian mode. The @option{-mbig-endian} option is
11328 the same as @option{-mbig}.
11330 @item -mdynamic-no-pic
11331 @opindex mdynamic-no-pic
11332 On Darwin and Mac OS X systems, compile code so that it is not
11333 relocatable, but that its external references are relocatable. The
11334 resulting code is suitable for applications, but not shared
11337 @item -mprioritize-restricted-insns=@var{priority}
11338 @opindex mprioritize-restricted-insns
11339 This option controls the priority that is assigned to
11340 dispatch-slot restricted instructions during the second scheduling
11341 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
11342 @var{no/highest/second-highest} priority to dispatch slot restricted
11345 @item -msched-costly-dep=@var{dependence_type}
11346 @opindex msched-costly-dep
11347 This option controls which dependences are considered costly
11348 by the target during instruction scheduling. The argument
11349 @var{dependence_type} takes one of the following values:
11350 @var{no}: no dependence is costly,
11351 @var{all}: all dependences are costly,
11352 @var{true_store_to_load}: a true dependence from store to load is costly,
11353 @var{store_to_load}: any dependence from store to load is costly,
11354 @var{number}: any dependence which latency >= @var{number} is costly.
11356 @item -minsert-sched-nops=@var{scheme}
11357 @opindex minsert-sched-nops
11358 This option controls which nop insertion scheme will be used during
11359 the second scheduling pass. The argument @var{scheme} takes one of the
11361 @var{no}: Don't insert nops.
11362 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
11363 according to the scheduler's grouping.
11364 @var{regroup_exact}: Insert nops to force costly dependent insns into
11365 separate groups. Insert exactly as many nops as needed to force an insn
11366 to a new group, according to the estimated processor grouping.
11367 @var{number}: Insert nops to force costly dependent insns into
11368 separate groups. Insert @var{number} nops to force an insn to a new group.
11371 @opindex mcall-sysv
11372 On System V.4 and embedded PowerPC systems compile code using calling
11373 conventions that adheres to the March 1995 draft of the System V
11374 Application Binary Interface, PowerPC processor supplement. This is the
11375 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
11377 @item -mcall-sysv-eabi
11378 @opindex mcall-sysv-eabi
11379 Specify both @option{-mcall-sysv} and @option{-meabi} options.
11381 @item -mcall-sysv-noeabi
11382 @opindex mcall-sysv-noeabi
11383 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
11385 @item -mcall-solaris
11386 @opindex mcall-solaris
11387 On System V.4 and embedded PowerPC systems compile code for the Solaris
11391 @opindex mcall-linux
11392 On System V.4 and embedded PowerPC systems compile code for the
11393 Linux-based GNU system.
11397 On System V.4 and embedded PowerPC systems compile code for the
11398 Hurd-based GNU system.
11400 @item -mcall-netbsd
11401 @opindex mcall-netbsd
11402 On System V.4 and embedded PowerPC systems compile code for the
11403 NetBSD operating system.
11405 @item -maix-struct-return
11406 @opindex maix-struct-return
11407 Return all structures in memory (as specified by the AIX ABI)@.
11409 @item -msvr4-struct-return
11410 @opindex msvr4-struct-return
11411 Return structures smaller than 8 bytes in registers (as specified by the
11414 @item -mabi=@var{abi-type}
11416 Extend the current ABI with a particular extension, or remove such extension.
11417 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
11421 @itemx -mno-prototype
11422 @opindex mprototype
11423 @opindex mno-prototype
11424 On System V.4 and embedded PowerPC systems assume that all calls to
11425 variable argument functions are properly prototyped. Otherwise, the
11426 compiler must insert an instruction before every non prototyped call to
11427 set or clear bit 6 of the condition code register (@var{CR}) to
11428 indicate whether floating point values were passed in the floating point
11429 registers in case the function takes a variable arguments. With
11430 @option{-mprototype}, only calls to prototyped variable argument functions
11431 will set or clear the bit.
11435 On embedded PowerPC systems, assume that the startup module is called
11436 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
11437 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
11442 On embedded PowerPC systems, assume that the startup module is called
11443 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
11448 On embedded PowerPC systems, assume that the startup module is called
11449 @file{crt0.o} and the standard C libraries are @file{libads.a} and
11452 @item -myellowknife
11453 @opindex myellowknife
11454 On embedded PowerPC systems, assume that the startup module is called
11455 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
11460 On System V.4 and embedded PowerPC systems, specify that you are
11461 compiling for a VxWorks system.
11465 Specify that you are compiling for the WindISS simulation environment.
11469 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
11470 header to indicate that @samp{eabi} extended relocations are used.
11476 On System V.4 and embedded PowerPC systems do (do not) adhere to the
11477 Embedded Applications Binary Interface (eabi) which is a set of
11478 modifications to the System V.4 specifications. Selecting @option{-meabi}
11479 means that the stack is aligned to an 8 byte boundary, a function
11480 @code{__eabi} is called to from @code{main} to set up the eabi
11481 environment, and the @option{-msdata} option can use both @code{r2} and
11482 @code{r13} to point to two separate small data areas. Selecting
11483 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
11484 do not call an initialization function from @code{main}, and the
11485 @option{-msdata} option will only use @code{r13} to point to a single
11486 small data area. The @option{-meabi} option is on by default if you
11487 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
11490 @opindex msdata=eabi
11491 On System V.4 and embedded PowerPC systems, put small initialized
11492 @code{const} global and static data in the @samp{.sdata2} section, which
11493 is pointed to by register @code{r2}. Put small initialized
11494 non-@code{const} global and static data in the @samp{.sdata} section,
11495 which is pointed to by register @code{r13}. Put small uninitialized
11496 global and static data in the @samp{.sbss} section, which is adjacent to
11497 the @samp{.sdata} section. The @option{-msdata=eabi} option is
11498 incompatible with the @option{-mrelocatable} option. The
11499 @option{-msdata=eabi} option also sets the @option{-memb} option.
11502 @opindex msdata=sysv
11503 On System V.4 and embedded PowerPC systems, put small global and static
11504 data in the @samp{.sdata} section, which is pointed to by register
11505 @code{r13}. Put small uninitialized global and static data in the
11506 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
11507 The @option{-msdata=sysv} option is incompatible with the
11508 @option{-mrelocatable} option.
11510 @item -msdata=default
11512 @opindex msdata=default
11514 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
11515 compile code the same as @option{-msdata=eabi}, otherwise compile code the
11516 same as @option{-msdata=sysv}.
11519 @opindex msdata-data
11520 On System V.4 and embedded PowerPC systems, put small global
11521 data in the @samp{.sdata} section. Put small uninitialized global
11522 data in the @samp{.sbss} section. Do not use register @code{r13}
11523 to address small data however. This is the default behavior unless
11524 other @option{-msdata} options are used.
11528 @opindex msdata=none
11530 On embedded PowerPC systems, put all initialized global and static data
11531 in the @samp{.data} section, and all uninitialized data in the
11532 @samp{.bss} section.
11536 @cindex smaller data references (PowerPC)
11537 @cindex .sdata/.sdata2 references (PowerPC)
11538 On embedded PowerPC systems, put global and static items less than or
11539 equal to @var{num} bytes into the small data or bss sections instead of
11540 the normal data or bss section. By default, @var{num} is 8. The
11541 @option{-G @var{num}} switch is also passed to the linker.
11542 All modules should be compiled with the same @option{-G @var{num}} value.
11545 @itemx -mno-regnames
11547 @opindex mno-regnames
11548 On System V.4 and embedded PowerPC systems do (do not) emit register
11549 names in the assembly language output using symbolic forms.
11552 @itemx -mno-longcall
11554 @opindex mno-longcall
11555 Default to making all function calls indirectly, using a register, so
11556 that functions which reside further than 32 megabytes (33,554,432
11557 bytes) from the current location can be called. This setting can be
11558 overridden by the @code{shortcall} function attribute, or by
11559 @code{#pragma longcall(0)}.
11561 Some linkers are capable of detecting out-of-range calls and generating
11562 glue code on the fly. On these systems, long calls are unnecessary and
11563 generate slower code. As of this writing, the AIX linker can do this,
11564 as can the GNU linker for PowerPC/64. It is planned to add this feature
11565 to the GNU linker for 32-bit PowerPC systems as well.
11567 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
11568 callee, L42'', plus a ``branch island'' (glue code). The two target
11569 addresses represent the callee and the ``branch island''. The
11570 Darwin/PPC linker will prefer the first address and generate a ``bl
11571 callee'' if the PPC ``bl'' instruction will reach the callee directly;
11572 otherwise, the linker will generate ``bl L42'' to call the ``branch
11573 island''. The ``branch island'' is appended to the body of the
11574 calling function; it computes the full 32-bit address of the callee
11577 On Mach-O (Darwin) systems, this option directs the compiler emit to
11578 the glue for every direct call, and the Darwin linker decides whether
11579 to use or discard it.
11581 In the future, we may cause GCC to ignore all longcall specifications
11582 when the linker is known to generate glue.
11586 Adds support for multithreading with the @dfn{pthreads} library.
11587 This option sets flags for both the preprocessor and linker.
11591 @node S/390 and zSeries Options
11592 @subsection S/390 and zSeries Options
11593 @cindex S/390 and zSeries Options
11595 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
11599 @itemx -msoft-float
11600 @opindex mhard-float
11601 @opindex msoft-float
11602 Use (do not use) the hardware floating-point instructions and registers
11603 for floating-point operations. When @option{-msoft-float} is specified,
11604 functions in @file{libgcc.a} will be used to perform floating-point
11605 operations. When @option{-mhard-float} is specified, the compiler
11606 generates IEEE floating-point instructions. This is the default.
11609 @itemx -mno-backchain
11610 @opindex mbackchain
11611 @opindex mno-backchain
11612 Store (do not store) the address of the caller's frame as backchain pointer
11613 into the callee's stack frame.
11614 A backchain may be needed to allow debugging using tools that do not understand
11615 DWARF-2 call frame information.
11616 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
11617 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
11618 the backchain is placed into the topmost word of the 96/160 byte register
11621 In general, code compiled with @option{-mbackchain} is call-compatible with
11622 code compiled with @option{-mmo-backchain}; however, use of the backchain
11623 for debugging purposes usually requires that the whole binary is built with
11624 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
11625 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
11626 to build a linux kernel use @option{-msoft-float}.
11628 The default is to not maintain the backchain.
11630 @item -mpacked-stack
11631 @item -mno-packed-stack
11632 @opindex mpacked-stack
11633 @opindex mno-packed-stack
11634 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
11635 specified, the compiler uses the all fields of the 96/160 byte register save
11636 area only for their default purpose; unused fields still take up stack space.
11637 When @option{-mpacked-stack} is specified, register save slots are densely
11638 packed at the top of the register save area; unused space is reused for other
11639 purposes, allowing for more efficient use of the available stack space.
11640 However, when @option{-mbackchain} is also in effect, the topmost word of
11641 the save area is always used to store the backchain, and the return address
11642 register is always saved two words below the backchain.
11644 As long as the stack frame backchain is not used, code generated with
11645 @option{-mpacked-stack} is call-compatible with code generated with
11646 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
11647 S/390 or zSeries generated code that uses the stack frame backchain at run
11648 time, not just for debugging purposes. Such code is not call-compatible
11649 with code compiled with @option{-mpacked-stack}. Also, note that the
11650 combination of @option{-mbackchain},
11651 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
11652 to build a linux kernel use @option{-msoft-float}.
11654 The default is to not use the packed stack layout.
11657 @itemx -mno-small-exec
11658 @opindex msmall-exec
11659 @opindex mno-small-exec
11660 Generate (or do not generate) code using the @code{bras} instruction
11661 to do subroutine calls.
11662 This only works reliably if the total executable size does not
11663 exceed 64k. The default is to use the @code{basr} instruction instead,
11664 which does not have this limitation.
11670 When @option{-m31} is specified, generate code compliant to the
11671 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
11672 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
11673 particular to generate 64-bit instructions. For the @samp{s390}
11674 targets, the default is @option{-m31}, while the @samp{s390x}
11675 targets default to @option{-m64}.
11681 When @option{-mzarch} is specified, generate code using the
11682 instructions available on z/Architecture.
11683 When @option{-mesa} is specified, generate code using the
11684 instructions available on ESA/390. Note that @option{-mesa} is
11685 not possible with @option{-m64}.
11686 When generating code compliant to the GNU/Linux for S/390 ABI,
11687 the default is @option{-mesa}. When generating code compliant
11688 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
11694 Generate (or do not generate) code using the @code{mvcle} instruction
11695 to perform block moves. When @option{-mno-mvcle} is specified,
11696 use a @code{mvc} loop instead. This is the default unless optimizing for
11703 Print (or do not print) additional debug information when compiling.
11704 The default is to not print debug information.
11706 @item -march=@var{cpu-type}
11708 Generate code that will run on @var{cpu-type}, which is the name of a system
11709 representing a certain processor type. Possible values for
11710 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
11711 When generating code using the instructions available on z/Architecture,
11712 the default is @option{-march=z900}. Otherwise, the default is
11713 @option{-march=g5}.
11715 @item -mtune=@var{cpu-type}
11717 Tune to @var{cpu-type} everything applicable about the generated code,
11718 except for the ABI and the set of available instructions.
11719 The list of @var{cpu-type} values is the same as for @option{-march}.
11720 The default is the value used for @option{-march}.
11723 @itemx -mno-tpf-trace
11724 @opindex mtpf-trace
11725 @opindex mno-tpf-trace
11726 Generate code that adds (does not add) in TPF OS specific branches to trace
11727 routines in the operating system. This option is off by default, even
11728 when compiling for the TPF OS@.
11731 @itemx -mno-fused-madd
11732 @opindex mfused-madd
11733 @opindex mno-fused-madd
11734 Generate code that uses (does not use) the floating point multiply and
11735 accumulate instructions. These instructions are generated by default if
11736 hardware floating point is used.
11738 @item -mwarn-framesize=@var{framesize}
11739 @opindex mwarn-framesize
11740 Emit a warning if the current function exceeds the given frame size. Because
11741 this is a compile time check it doesn't need to be a real problem when the program
11742 runs. It is intended to identify functions which most probably cause
11743 a stack overflow. It is useful to be used in an environment with limited stack
11744 size e.g.@: the linux kernel.
11746 @item -mwarn-dynamicstack
11747 @opindex mwarn-dynamicstack
11748 Emit a warning if the function calls alloca or uses dynamically
11749 sized arrays. This is generally a bad idea with a limited stack size.
11751 @item -mstack-guard=@var{stack-guard}
11752 @item -mstack-size=@var{stack-size}
11753 @opindex mstack-guard
11754 @opindex mstack-size
11755 These arguments always have to be used in conjunction. If they are present the s390
11756 back end emits additional instructions in the function prologue which trigger a trap
11757 if the stack size is @var{stack-guard} bytes above the @var{stack-size}
11758 (remember that the stack on s390 grows downward). These options are intended to
11759 be used to help debugging stack overflow problems. The additionally emitted code
11760 causes only little overhead and hence can also be used in production like systems
11761 without greater performance degradation. The given values have to be exact
11762 powers of 2 and @var{stack-size} has to be greater than @var{stack-guard} without
11764 In order to be efficient the extra code makes the assumption that the stack starts
11765 at an address aligned to the value given by @var{stack-size}.
11769 @subsection SH Options
11771 These @samp{-m} options are defined for the SH implementations:
11776 Generate code for the SH1.
11780 Generate code for the SH2.
11783 Generate code for the SH2e.
11787 Generate code for the SH3.
11791 Generate code for the SH3e.
11795 Generate code for the SH4 without a floating-point unit.
11797 @item -m4-single-only
11798 @opindex m4-single-only
11799 Generate code for the SH4 with a floating-point unit that only
11800 supports single-precision arithmetic.
11804 Generate code for the SH4 assuming the floating-point unit is in
11805 single-precision mode by default.
11809 Generate code for the SH4.
11813 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
11814 floating-point unit is not used.
11816 @item -m4a-single-only
11817 @opindex m4a-single-only
11818 Generate code for the SH4a, in such a way that no double-precision
11819 floating point operations are used.
11822 @opindex m4a-single
11823 Generate code for the SH4a assuming the floating-point unit is in
11824 single-precision mode by default.
11828 Generate code for the SH4a.
11832 Same as @option{-m4a-nofpu}, except that it implicitly passes
11833 @option{-dsp} to the assembler. GCC doesn't generate any DSP
11834 instructions at the moment.
11838 Compile code for the processor in big endian mode.
11842 Compile code for the processor in little endian mode.
11846 Align doubles at 64-bit boundaries. Note that this changes the calling
11847 conventions, and thus some functions from the standard C library will
11848 not work unless you recompile it first with @option{-mdalign}.
11852 Shorten some address references at link time, when possible; uses the
11853 linker option @option{-relax}.
11857 Use 32-bit offsets in @code{switch} tables. The default is to use
11862 Enable the use of the instruction @code{fmovd}.
11866 Comply with the calling conventions defined by Renesas.
11870 Comply with the calling conventions defined by Renesas.
11874 Comply with the calling conventions defined for GCC before the Renesas
11875 conventions were available. This option is the default for all
11876 targets of the SH toolchain except for @samp{sh-symbianelf}.
11879 @opindex mnomacsave
11880 Mark the @code{MAC} register as call-clobbered, even if
11881 @option{-mhitachi} is given.
11885 Increase IEEE-compliance of floating-point code.
11886 At the moment, this is equivalent to @option{-fno-finite-math-only}.
11887 When generating 16 bit SH opcodes, getting IEEE-conforming results for
11888 comparisons of NANs / infinities incurs extra overhead in every
11889 floating point comparison, therefore the default is set to
11890 @option{-ffinite-math-only}.
11894 Dump instruction size and location in the assembly code.
11897 @opindex mpadstruct
11898 This option is deprecated. It pads structures to multiple of 4 bytes,
11899 which is incompatible with the SH ABI@.
11903 Optimize for space instead of speed. Implied by @option{-Os}.
11906 @opindex mprefergot
11907 When generating position-independent code, emit function calls using
11908 the Global Offset Table instead of the Procedure Linkage Table.
11912 Generate a library function call to invalidate instruction cache
11913 entries, after fixing up a trampoline. This library function call
11914 doesn't assume it can write to the whole memory address space. This
11915 is the default when the target is @code{sh-*-linux*}.
11917 @item -multcost=@var{number}
11918 @opindex multcost=@var{number}
11919 Set the cost to assume for a multiply insn.
11921 @item -mdiv=@var{strategy}
11922 @opindex mdiv=@var{strategy}
11923 Set the division strategy to use for SHmedia code. @var{strategy} must be
11924 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
11925 inv:call2, inv:fp .
11926 "fp" performs the operation in floating point. This has a very high latency,
11927 but needs only a few instructions, so it might be a good choice if
11928 your code has enough easily exploitable ILP to allow the compiler to
11929 schedule the floating point instructions together with other instructions.
11930 Division by zero causes a floating point exception.
11931 "inv" uses integer operations to calculate the inverse of the divisor,
11932 and then multiplies the dividend with the inverse. This strategy allows
11933 cse and hoisting of the inverse calculation. Division by zero calculates
11934 an unspecified result, but does not trap.
11935 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
11936 have been found, or if the entire operation has been hoisted to the same
11937 place, the last stages of the inverse calculation are intertwined with the
11938 final multiply to reduce the overall latency, at the expense of using a few
11939 more instructions, and thus offering fewer scheduling opportunities with
11941 "call" calls a library function that usually implements the inv:minlat
11943 This gives high code density for m5-*media-nofpu compilations.
11944 "call2" uses a different entry point of the same library function, where it
11945 assumes that a pointer to a lookup table has already been set up, which
11946 exposes the pointer load to cse / code hoisting optimizations.
11947 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
11948 code generation, but if the code stays unoptimized, revert to the "call",
11949 "call2", or "fp" strategies, respectively. Note that the
11950 potentially-trapping side effect of division by zero is carried by a
11951 separate instruction, so it is possible that all the integer instructions
11952 are hoisted out, but the marker for the side effect stays where it is.
11953 A recombination to fp operations or a call is not possible in that case.
11954 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
11955 that the inverse calculation was nor separated from the multiply, they speed
11956 up division where the dividend fits into 20 bits (plus sign where applicable),
11957 by inserting a test to skip a number of operations in this case; this test
11958 slows down the case of larger dividends. inv20u assumes the case of a such
11959 a small dividend to be unlikely, and inv20l assumes it to be likely.
11961 @item -mdivsi3_libfunc=@var{name}
11962 @opindex mdivsi3_libfunc=@var{name}
11963 Set the name of the library function used for 32 bit signed division to
11964 @var{name}. This only affect the name used in the call and inv:call
11965 division strategies, and the compiler will still expect the same
11966 sets of input/output/clobbered registers as if this option was not present.
11968 @item -madjust-unroll
11969 @opindex madjust-unroll
11970 Throttle unrolling to avoid thrashing target registers.
11971 This option only has an effect if the gcc code base supports the
11972 TARGET_ADJUST_UNROLL_MAX target hook.
11974 @item -mindexed-addressing
11975 @opindex mindexed-addressing
11976 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
11977 This is only safe if the hardware and/or OS implement 32 bit wrap-around
11978 semantics for the indexed addressing mode. The architecture allows the
11979 implementation of processors with 64 bit MMU, which the OS could use to
11980 get 32 bit addressing, but since no current hardware implementation supports
11981 this or any other way to make the indexed addressing mode safe to use in
11982 the 32 bit ABI, the default is -mno-indexed-addressing.
11984 @item -mgettrcost=@var{number}
11985 @opindex mgettrcost=@var{number}
11986 Set the cost assumed for the gettr instruction to @var{number}.
11987 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
11991 Assume pt* instructions won't trap. This will generally generate better
11992 scheduled code, but is unsafe on current hardware. The current architecture
11993 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
11994 This has the unintentional effect of making it unsafe to schedule ptabs /
11995 ptrel before a branch, or hoist it out of a loop. For example,
11996 __do_global_ctors, a part of libgcc that runs constructors at program
11997 startup, calls functions in a list which is delimited by -1. With the
11998 -mpt-fixed option, the ptabs will be done before testing against -1.
11999 That means that all the constructors will be run a bit quicker, but when
12000 the loop comes to the end of the list, the program crashes because ptabs
12001 loads -1 into a target register. Since this option is unsafe for any
12002 hardware implementing the current architecture specification, the default
12003 is -mno-pt-fixed. Unless the user specifies a specific cost with
12004 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
12005 this deters register allocation using target registers for storing
12008 @item -minvalid-symbols
12009 @opindex minvalid-symbols
12010 Assume symbols might be invalid. Ordinary function symbols generated by
12011 the compiler will always be valid to load with movi/shori/ptabs or
12012 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
12013 to generate symbols that will cause ptabs / ptrel to trap.
12014 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
12015 It will then prevent cross-basic-block cse, hoisting and most scheduling
12016 of symbol loads. The default is @option{-mno-invalid-symbols}.
12019 @node SPARC Options
12020 @subsection SPARC Options
12021 @cindex SPARC options
12023 These @samp{-m} options are supported on the SPARC:
12026 @item -mno-app-regs
12028 @opindex mno-app-regs
12030 Specify @option{-mapp-regs} to generate output using the global registers
12031 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
12034 To be fully SVR4 ABI compliant at the cost of some performance loss,
12035 specify @option{-mno-app-regs}. You should compile libraries and system
12036 software with this option.
12039 @itemx -mhard-float
12041 @opindex mhard-float
12042 Generate output containing floating point instructions. This is the
12046 @itemx -msoft-float
12048 @opindex msoft-float
12049 Generate output containing library calls for floating point.
12050 @strong{Warning:} the requisite libraries are not available for all SPARC
12051 targets. Normally the facilities of the machine's usual C compiler are
12052 used, but this cannot be done directly in cross-compilation. You must make
12053 your own arrangements to provide suitable library functions for
12054 cross-compilation. The embedded targets @samp{sparc-*-aout} and
12055 @samp{sparclite-*-*} do provide software floating point support.
12057 @option{-msoft-float} changes the calling convention in the output file;
12058 therefore, it is only useful if you compile @emph{all} of a program with
12059 this option. In particular, you need to compile @file{libgcc.a}, the
12060 library that comes with GCC, with @option{-msoft-float} in order for
12063 @item -mhard-quad-float
12064 @opindex mhard-quad-float
12065 Generate output containing quad-word (long double) floating point
12068 @item -msoft-quad-float
12069 @opindex msoft-quad-float
12070 Generate output containing library calls for quad-word (long double)
12071 floating point instructions. The functions called are those specified
12072 in the SPARC ABI@. This is the default.
12074 As of this writing, there are no SPARC implementations that have hardware
12075 support for the quad-word floating point instructions. They all invoke
12076 a trap handler for one of these instructions, and then the trap handler
12077 emulates the effect of the instruction. Because of the trap handler overhead,
12078 this is much slower than calling the ABI library routines. Thus the
12079 @option{-msoft-quad-float} option is the default.
12081 @item -mno-unaligned-doubles
12082 @itemx -munaligned-doubles
12083 @opindex mno-unaligned-doubles
12084 @opindex munaligned-doubles
12085 Assume that doubles have 8 byte alignment. This is the default.
12087 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
12088 alignment only if they are contained in another type, or if they have an
12089 absolute address. Otherwise, it assumes they have 4 byte alignment.
12090 Specifying this option avoids some rare compatibility problems with code
12091 generated by other compilers. It is not the default because it results
12092 in a performance loss, especially for floating point code.
12094 @item -mno-faster-structs
12095 @itemx -mfaster-structs
12096 @opindex mno-faster-structs
12097 @opindex mfaster-structs
12098 With @option{-mfaster-structs}, the compiler assumes that structures
12099 should have 8 byte alignment. This enables the use of pairs of
12100 @code{ldd} and @code{std} instructions for copies in structure
12101 assignment, in place of twice as many @code{ld} and @code{st} pairs.
12102 However, the use of this changed alignment directly violates the SPARC
12103 ABI@. Thus, it's intended only for use on targets where the developer
12104 acknowledges that their resulting code will not be directly in line with
12105 the rules of the ABI@.
12107 @item -mimpure-text
12108 @opindex mimpure-text
12109 @option{-mimpure-text}, used in addition to @option{-shared}, tells
12110 the compiler to not pass @option{-z text} to the linker when linking a
12111 shared object. Using this option, you can link position-dependent
12112 code into a shared object.
12114 @option{-mimpure-text} suppresses the ``relocations remain against
12115 allocatable but non-writable sections'' linker error message.
12116 However, the necessary relocations will trigger copy-on-write, and the
12117 shared object is not actually shared across processes. Instead of
12118 using @option{-mimpure-text}, you should compile all source code with
12119 @option{-fpic} or @option{-fPIC}.
12121 This option is only available on SunOS and Solaris.
12123 @item -mcpu=@var{cpu_type}
12125 Set the instruction set, register set, and instruction scheduling parameters
12126 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
12127 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
12128 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
12129 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
12130 @samp{ultrasparc3}.
12132 Default instruction scheduling parameters are used for values that select
12133 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
12134 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
12136 Here is a list of each supported architecture and their supported
12141 v8: supersparc, hypersparc
12142 sparclite: f930, f934, sparclite86x
12144 v9: ultrasparc, ultrasparc3
12147 By default (unless configured otherwise), GCC generates code for the V7
12148 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
12149 additionally optimizes it for the Cypress CY7C602 chip, as used in the
12150 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
12151 SPARCStation 1, 2, IPX etc.
12153 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
12154 architecture. The only difference from V7 code is that the compiler emits
12155 the integer multiply and integer divide instructions which exist in SPARC-V8
12156 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
12157 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
12160 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
12161 the SPARC architecture. This adds the integer multiply, integer divide step
12162 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
12163 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
12164 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
12165 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
12166 MB86934 chip, which is the more recent SPARClite with FPU@.
12168 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
12169 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
12170 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
12171 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
12172 optimizes it for the TEMIC SPARClet chip.
12174 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
12175 architecture. This adds 64-bit integer and floating-point move instructions,
12176 3 additional floating-point condition code registers and conditional move
12177 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
12178 optimizes it for the Sun UltraSPARC I/II chips. With
12179 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
12180 Sun UltraSPARC III chip.
12182 @item -mtune=@var{cpu_type}
12184 Set the instruction scheduling parameters for machine type
12185 @var{cpu_type}, but do not set the instruction set or register set that the
12186 option @option{-mcpu=@var{cpu_type}} would.
12188 The same values for @option{-mcpu=@var{cpu_type}} can be used for
12189 @option{-mtune=@var{cpu_type}}, but the only useful values are those
12190 that select a particular cpu implementation. Those are @samp{cypress},
12191 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
12192 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
12193 @samp{ultrasparc3}.
12198 @opindex mno-v8plus
12199 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
12200 difference from the V8 ABI is that the global and out registers are
12201 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
12202 mode for all SPARC-V9 processors.
12208 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
12209 Visual Instruction Set extensions. The default is @option{-mno-vis}.
12212 These @samp{-m} options are supported in addition to the above
12213 on SPARC-V9 processors in 64-bit environments:
12216 @item -mlittle-endian
12217 @opindex mlittle-endian
12218 Generate code for a processor running in little-endian mode. It is only
12219 available for a few configurations and most notably not on Solaris and Linux.
12225 Generate code for a 32-bit or 64-bit environment.
12226 The 32-bit environment sets int, long and pointer to 32 bits.
12227 The 64-bit environment sets int to 32 bits and long and pointer
12230 @item -mcmodel=medlow
12231 @opindex mcmodel=medlow
12232 Generate code for the Medium/Low code model: 64-bit addresses, programs
12233 must be linked in the low 32 bits of memory. Programs can be statically
12234 or dynamically linked.
12236 @item -mcmodel=medmid
12237 @opindex mcmodel=medmid
12238 Generate code for the Medium/Middle code model: 64-bit addresses, programs
12239 must be linked in the low 44 bits of memory, the text and data segments must
12240 be less than 2GB in size and the data segment must be located within 2GB of
12243 @item -mcmodel=medany
12244 @opindex mcmodel=medany
12245 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
12246 may be linked anywhere in memory, the text and data segments must be less
12247 than 2GB in size and the data segment must be located within 2GB of the
12250 @item -mcmodel=embmedany
12251 @opindex mcmodel=embmedany
12252 Generate code for the Medium/Anywhere code model for embedded systems:
12253 64-bit addresses, the text and data segments must be less than 2GB in
12254 size, both starting anywhere in memory (determined at link time). The
12255 global register %g4 points to the base of the data segment. Programs
12256 are statically linked and PIC is not supported.
12259 @itemx -mno-stack-bias
12260 @opindex mstack-bias
12261 @opindex mno-stack-bias
12262 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
12263 frame pointer if present, are offset by @minus{}2047 which must be added back
12264 when making stack frame references. This is the default in 64-bit mode.
12265 Otherwise, assume no such offset is present.
12268 These switches are supported in addition to the above on Solaris:
12273 Add support for multithreading using the Solaris threads library. This
12274 option sets flags for both the preprocessor and linker. This option does
12275 not affect the thread safety of object code produced by the compiler or
12276 that of libraries supplied with it.
12280 Add support for multithreading using the POSIX threads library. This
12281 option sets flags for both the preprocessor and linker. This option does
12282 not affect the thread safety of object code produced by the compiler or
12283 that of libraries supplied with it.
12286 @node System V Options
12287 @subsection Options for System V
12289 These additional options are available on System V Release 4 for
12290 compatibility with other compilers on those systems:
12295 Create a shared object.
12296 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
12300 Identify the versions of each tool used by the compiler, in a
12301 @code{.ident} assembler directive in the output.
12305 Refrain from adding @code{.ident} directives to the output file (this is
12308 @item -YP,@var{dirs}
12310 Search the directories @var{dirs}, and no others, for libraries
12311 specified with @option{-l}.
12313 @item -Ym,@var{dir}
12315 Look in the directory @var{dir} to find the M4 preprocessor.
12316 The assembler uses this option.
12317 @c This is supposed to go with a -Yd for predefined M4 macro files, but
12318 @c the generic assembler that comes with Solaris takes just -Ym.
12321 @node TMS320C3x/C4x Options
12322 @subsection TMS320C3x/C4x Options
12323 @cindex TMS320C3x/C4x Options
12325 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
12329 @item -mcpu=@var{cpu_type}
12331 Set the instruction set, register set, and instruction scheduling
12332 parameters for machine type @var{cpu_type}. Supported values for
12333 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
12334 @samp{c44}. The default is @samp{c40} to generate code for the
12339 @itemx -msmall-memory
12341 @opindex mbig-memory
12343 @opindex msmall-memory
12345 Generates code for the big or small memory model. The small memory
12346 model assumed that all data fits into one 64K word page. At run-time
12347 the data page (DP) register must be set to point to the 64K page
12348 containing the .bss and .data program sections. The big memory model is
12349 the default and requires reloading of the DP register for every direct
12356 Allow (disallow) allocation of general integer operands into the block
12357 count register BK@.
12363 Enable (disable) generation of code using decrement and branch,
12364 DBcond(D), instructions. This is enabled by default for the C4x. To be
12365 on the safe side, this is disabled for the C3x, since the maximum
12366 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
12367 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
12368 that it can utilize the decrement and branch instruction, but will give
12369 up if there is more than one memory reference in the loop. Thus a loop
12370 where the loop counter is decremented can generate slightly more
12371 efficient code, in cases where the RPTB instruction cannot be utilized.
12373 @item -mdp-isr-reload
12375 @opindex mdp-isr-reload
12377 Force the DP register to be saved on entry to an interrupt service
12378 routine (ISR), reloaded to point to the data section, and restored on
12379 exit from the ISR@. This should not be required unless someone has
12380 violated the small memory model by modifying the DP register, say within
12387 For the C3x use the 24-bit MPYI instruction for integer multiplies
12388 instead of a library call to guarantee 32-bit results. Note that if one
12389 of the operands is a constant, then the multiplication will be performed
12390 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
12391 then squaring operations are performed inline instead of a library call.
12394 @itemx -mno-fast-fix
12396 @opindex mno-fast-fix
12397 The C3x/C4x FIX instruction to convert a floating point value to an
12398 integer value chooses the nearest integer less than or equal to the
12399 floating point value rather than to the nearest integer. Thus if the
12400 floating point number is negative, the result will be incorrectly
12401 truncated an additional code is necessary to detect and correct this
12402 case. This option can be used to disable generation of the additional
12403 code required to correct the result.
12409 Enable (disable) generation of repeat block sequences using the RPTB
12410 instruction for zero overhead looping. The RPTB construct is only used
12411 for innermost loops that do not call functions or jump across the loop
12412 boundaries. There is no advantage having nested RPTB loops due to the
12413 overhead required to save and restore the RC, RS, and RE registers.
12414 This is enabled by default with @option{-O2}.
12416 @item -mrpts=@var{count}
12420 Enable (disable) the use of the single instruction repeat instruction
12421 RPTS@. If a repeat block contains a single instruction, and the loop
12422 count can be guaranteed to be less than the value @var{count}, GCC will
12423 emit a RPTS instruction instead of a RPTB@. If no value is specified,
12424 then a RPTS will be emitted even if the loop count cannot be determined
12425 at compile time. Note that the repeated instruction following RPTS does
12426 not have to be reloaded from memory each iteration, thus freeing up the
12427 CPU buses for operands. However, since interrupts are blocked by this
12428 instruction, it is disabled by default.
12430 @item -mloop-unsigned
12431 @itemx -mno-loop-unsigned
12432 @opindex mloop-unsigned
12433 @opindex mno-loop-unsigned
12434 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
12435 is @math{2^{31} + 1} since these instructions test if the iteration count is
12436 negative to terminate the loop. If the iteration count is unsigned
12437 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
12438 exceeded. This switch allows an unsigned iteration count.
12442 Try to emit an assembler syntax that the TI assembler (asm30) is happy
12443 with. This also enforces compatibility with the API employed by the TI
12444 C3x C compiler. For example, long doubles are passed as structures
12445 rather than in floating point registers.
12451 Generate code that uses registers (stack) for passing arguments to functions.
12452 By default, arguments are passed in registers where possible rather
12453 than by pushing arguments on to the stack.
12455 @item -mparallel-insns
12456 @itemx -mno-parallel-insns
12457 @opindex mparallel-insns
12458 @opindex mno-parallel-insns
12459 Allow the generation of parallel instructions. This is enabled by
12460 default with @option{-O2}.
12462 @item -mparallel-mpy
12463 @itemx -mno-parallel-mpy
12464 @opindex mparallel-mpy
12465 @opindex mno-parallel-mpy
12466 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
12467 provided @option{-mparallel-insns} is also specified. These instructions have
12468 tight register constraints which can pessimize the code generation
12469 of large functions.
12474 @subsection V850 Options
12475 @cindex V850 Options
12477 These @samp{-m} options are defined for V850 implementations:
12481 @itemx -mno-long-calls
12482 @opindex mlong-calls
12483 @opindex mno-long-calls
12484 Treat all calls as being far away (near). If calls are assumed to be
12485 far away, the compiler will always load the functions address up into a
12486 register, and call indirect through the pointer.
12492 Do not optimize (do optimize) basic blocks that use the same index
12493 pointer 4 or more times to copy pointer into the @code{ep} register, and
12494 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
12495 option is on by default if you optimize.
12497 @item -mno-prolog-function
12498 @itemx -mprolog-function
12499 @opindex mno-prolog-function
12500 @opindex mprolog-function
12501 Do not use (do use) external functions to save and restore registers
12502 at the prologue and epilogue of a function. The external functions
12503 are slower, but use less code space if more than one function saves
12504 the same number of registers. The @option{-mprolog-function} option
12505 is on by default if you optimize.
12509 Try to make the code as small as possible. At present, this just turns
12510 on the @option{-mep} and @option{-mprolog-function} options.
12512 @item -mtda=@var{n}
12514 Put static or global variables whose size is @var{n} bytes or less into
12515 the tiny data area that register @code{ep} points to. The tiny data
12516 area can hold up to 256 bytes in total (128 bytes for byte references).
12518 @item -msda=@var{n}
12520 Put static or global variables whose size is @var{n} bytes or less into
12521 the small data area that register @code{gp} points to. The small data
12522 area can hold up to 64 kilobytes.
12524 @item -mzda=@var{n}
12526 Put static or global variables whose size is @var{n} bytes or less into
12527 the first 32 kilobytes of memory.
12531 Specify that the target processor is the V850.
12534 @opindex mbig-switch
12535 Generate code suitable for big switch tables. Use this option only if
12536 the assembler/linker complain about out of range branches within a switch
12541 This option will cause r2 and r5 to be used in the code generated by
12542 the compiler. This setting is the default.
12544 @item -mno-app-regs
12545 @opindex mno-app-regs
12546 This option will cause r2 and r5 to be treated as fixed registers.
12550 Specify that the target processor is the V850E1. The preprocessor
12551 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
12552 this option is used.
12556 Specify that the target processor is the V850E@. The preprocessor
12557 constant @samp{__v850e__} will be defined if this option is used.
12559 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
12560 are defined then a default target processor will be chosen and the
12561 relevant @samp{__v850*__} preprocessor constant will be defined.
12563 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
12564 defined, regardless of which processor variant is the target.
12566 @item -mdisable-callt
12567 @opindex mdisable-callt
12568 This option will suppress generation of the CALLT instruction for the
12569 v850e and v850e1 flavors of the v850 architecture. The default is
12570 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
12575 @subsection VAX Options
12576 @cindex VAX options
12578 These @samp{-m} options are defined for the VAX:
12583 Do not output certain jump instructions (@code{aobleq} and so on)
12584 that the Unix assembler for the VAX cannot handle across long
12589 Do output those jump instructions, on the assumption that you
12590 will assemble with the GNU assembler.
12594 Output code for g-format floating point numbers instead of d-format.
12597 @node x86-64 Options
12598 @subsection x86-64 Options
12599 @cindex x86-64 options
12601 These are listed under @xref{i386 and x86-64 Options}.
12603 @node Xstormy16 Options
12604 @subsection Xstormy16 Options
12605 @cindex Xstormy16 Options
12607 These options are defined for Xstormy16:
12612 Choose startup files and linker script suitable for the simulator.
12615 @node Xtensa Options
12616 @subsection Xtensa Options
12617 @cindex Xtensa Options
12619 These options are supported for Xtensa targets:
12623 @itemx -mno-const16
12625 @opindex mno-const16
12626 Enable or disable use of @code{CONST16} instructions for loading
12627 constant values. The @code{CONST16} instruction is currently not a
12628 standard option from Tensilica. When enabled, @code{CONST16}
12629 instructions are always used in place of the standard @code{L32R}
12630 instructions. The use of @code{CONST16} is enabled by default only if
12631 the @code{L32R} instruction is not available.
12634 @itemx -mno-fused-madd
12635 @opindex mfused-madd
12636 @opindex mno-fused-madd
12637 Enable or disable use of fused multiply/add and multiply/subtract
12638 instructions in the floating-point option. This has no effect if the
12639 floating-point option is not also enabled. Disabling fused multiply/add
12640 and multiply/subtract instructions forces the compiler to use separate
12641 instructions for the multiply and add/subtract operations. This may be
12642 desirable in some cases where strict IEEE 754-compliant results are
12643 required: the fused multiply add/subtract instructions do not round the
12644 intermediate result, thereby producing results with @emph{more} bits of
12645 precision than specified by the IEEE standard. Disabling fused multiply
12646 add/subtract instructions also ensures that the program output is not
12647 sensitive to the compiler's ability to combine multiply and add/subtract
12650 @item -mtext-section-literals
12651 @itemx -mno-text-section-literals
12652 @opindex mtext-section-literals
12653 @opindex mno-text-section-literals
12654 Control the treatment of literal pools. The default is
12655 @option{-mno-text-section-literals}, which places literals in a separate
12656 section in the output file. This allows the literal pool to be placed
12657 in a data RAM/ROM, and it also allows the linker to combine literal
12658 pools from separate object files to remove redundant literals and
12659 improve code size. With @option{-mtext-section-literals}, the literals
12660 are interspersed in the text section in order to keep them as close as
12661 possible to their references. This may be necessary for large assembly
12664 @item -mtarget-align
12665 @itemx -mno-target-align
12666 @opindex mtarget-align
12667 @opindex mno-target-align
12668 When this option is enabled, GCC instructs the assembler to
12669 automatically align instructions to reduce branch penalties at the
12670 expense of some code density. The assembler attempts to widen density
12671 instructions to align branch targets and the instructions following call
12672 instructions. If there are not enough preceding safe density
12673 instructions to align a target, no widening will be performed. The
12674 default is @option{-mtarget-align}. These options do not affect the
12675 treatment of auto-aligned instructions like @code{LOOP}, which the
12676 assembler will always align, either by widening density instructions or
12677 by inserting no-op instructions.
12680 @itemx -mno-longcalls
12681 @opindex mlongcalls
12682 @opindex mno-longcalls
12683 When this option is enabled, GCC instructs the assembler to translate
12684 direct calls to indirect calls unless it can determine that the target
12685 of a direct call is in the range allowed by the call instruction. This
12686 translation typically occurs for calls to functions in other source
12687 files. Specifically, the assembler translates a direct @code{CALL}
12688 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
12689 The default is @option{-mno-longcalls}. This option should be used in
12690 programs where the call target can potentially be out of range. This
12691 option is implemented in the assembler, not the compiler, so the
12692 assembly code generated by GCC will still show direct call
12693 instructions---look at the disassembled object code to see the actual
12694 instructions. Note that the assembler will use an indirect call for
12695 every cross-file call, not just those that really will be out of range.
12698 @node zSeries Options
12699 @subsection zSeries Options
12700 @cindex zSeries options
12702 These are listed under @xref{S/390 and zSeries Options}.
12704 @node Code Gen Options
12705 @section Options for Code Generation Conventions
12706 @cindex code generation conventions
12707 @cindex options, code generation
12708 @cindex run-time options
12710 These machine-independent options control the interface conventions
12711 used in code generation.
12713 Most of them have both positive and negative forms; the negative form
12714 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
12715 one of the forms is listed---the one which is not the default. You
12716 can figure out the other form by either removing @samp{no-} or adding
12720 @item -fbounds-check
12721 @opindex fbounds-check
12722 For front-ends that support it, generate additional code to check that
12723 indices used to access arrays are within the declared range. This is
12724 currently only supported by the Java and Fortran 77 front-ends, where
12725 this option defaults to true and false respectively.
12729 This option generates traps for signed overflow on addition, subtraction,
12730 multiplication operations.
12734 This option instructs the compiler to assume that signed arithmetic
12735 overflow of addition, subtraction and multiplication wraps around
12736 using twos-complement representation. This flag enables some optimizations
12737 and disables others. This option is enabled by default for the Java
12738 front-end, as required by the Java language specification.
12741 @opindex fexceptions
12742 Enable exception handling. Generates extra code needed to propagate
12743 exceptions. For some targets, this implies GCC will generate frame
12744 unwind information for all functions, which can produce significant data
12745 size overhead, although it does not affect execution. If you do not
12746 specify this option, GCC will enable it by default for languages like
12747 C++ which normally require exception handling, and disable it for
12748 languages like C that do not normally require it. However, you may need
12749 to enable this option when compiling C code that needs to interoperate
12750 properly with exception handlers written in C++. You may also wish to
12751 disable this option if you are compiling older C++ programs that don't
12752 use exception handling.
12754 @item -fnon-call-exceptions
12755 @opindex fnon-call-exceptions
12756 Generate code that allows trapping instructions to throw exceptions.
12757 Note that this requires platform-specific runtime support that does
12758 not exist everywhere. Moreover, it only allows @emph{trapping}
12759 instructions to throw exceptions, i.e.@: memory references or floating
12760 point instructions. It does not allow exceptions to be thrown from
12761 arbitrary signal handlers such as @code{SIGALRM}.
12763 @item -funwind-tables
12764 @opindex funwind-tables
12765 Similar to @option{-fexceptions}, except that it will just generate any needed
12766 static data, but will not affect the generated code in any other way.
12767 You will normally not enable this option; instead, a language processor
12768 that needs this handling would enable it on your behalf.
12770 @item -fasynchronous-unwind-tables
12771 @opindex fasynchronous-unwind-tables
12772 Generate unwind table in dwarf2 format, if supported by target machine. The
12773 table is exact at each instruction boundary, so it can be used for stack
12774 unwinding from asynchronous events (such as debugger or garbage collector).
12776 @item -fpcc-struct-return
12777 @opindex fpcc-struct-return
12778 Return ``short'' @code{struct} and @code{union} values in memory like
12779 longer ones, rather than in registers. This convention is less
12780 efficient, but it has the advantage of allowing intercallability between
12781 GCC-compiled files and files compiled with other compilers, particularly
12782 the Portable C Compiler (pcc).
12784 The precise convention for returning structures in memory depends
12785 on the target configuration macros.
12787 Short structures and unions are those whose size and alignment match
12788 that of some integer type.
12790 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
12791 switch is not binary compatible with code compiled with the
12792 @option{-freg-struct-return} switch.
12793 Use it to conform to a non-default application binary interface.
12795 @item -freg-struct-return
12796 @opindex freg-struct-return
12797 Return @code{struct} and @code{union} values in registers when possible.
12798 This is more efficient for small structures than
12799 @option{-fpcc-struct-return}.
12801 If you specify neither @option{-fpcc-struct-return} nor
12802 @option{-freg-struct-return}, GCC defaults to whichever convention is
12803 standard for the target. If there is no standard convention, GCC
12804 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
12805 the principal compiler. In those cases, we can choose the standard, and
12806 we chose the more efficient register return alternative.
12808 @strong{Warning:} code compiled with the @option{-freg-struct-return}
12809 switch is not binary compatible with code compiled with the
12810 @option{-fpcc-struct-return} switch.
12811 Use it to conform to a non-default application binary interface.
12813 @item -fshort-enums
12814 @opindex fshort-enums
12815 Allocate to an @code{enum} type only as many bytes as it needs for the
12816 declared range of possible values. Specifically, the @code{enum} type
12817 will be equivalent to the smallest integer type which has enough room.
12819 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
12820 code that is not binary compatible with code generated without that switch.
12821 Use it to conform to a non-default application binary interface.
12823 @item -fshort-double
12824 @opindex fshort-double
12825 Use the same size for @code{double} as for @code{float}.
12827 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
12828 code that is not binary compatible with code generated without that switch.
12829 Use it to conform to a non-default application binary interface.
12831 @item -fshort-wchar
12832 @opindex fshort-wchar
12833 Override the underlying type for @samp{wchar_t} to be @samp{short
12834 unsigned int} instead of the default for the target. This option is
12835 useful for building programs to run under WINE@.
12837 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
12838 code that is not binary compatible with code generated without that switch.
12839 Use it to conform to a non-default application binary interface.
12841 @item -fshared-data
12842 @opindex fshared-data
12843 Requests that the data and non-@code{const} variables of this
12844 compilation be shared data rather than private data. The distinction
12845 makes sense only on certain operating systems, where shared data is
12846 shared between processes running the same program, while private data
12847 exists in one copy per process.
12850 @opindex fno-common
12851 In C, allocate even uninitialized global variables in the data section of the
12852 object file, rather than generating them as common blocks. This has the
12853 effect that if the same variable is declared (without @code{extern}) in
12854 two different compilations, you will get an error when you link them.
12855 The only reason this might be useful is if you wish to verify that the
12856 program will work on other systems which always work this way.
12860 Ignore the @samp{#ident} directive.
12862 @item -finhibit-size-directive
12863 @opindex finhibit-size-directive
12864 Don't output a @code{.size} assembler directive, or anything else that
12865 would cause trouble if the function is split in the middle, and the
12866 two halves are placed at locations far apart in memory. This option is
12867 used when compiling @file{crtstuff.c}; you should not need to use it
12870 @item -fverbose-asm
12871 @opindex fverbose-asm
12872 Put extra commentary information in the generated assembly code to
12873 make it more readable. This option is generally only of use to those
12874 who actually need to read the generated assembly code (perhaps while
12875 debugging the compiler itself).
12877 @option{-fno-verbose-asm}, the default, causes the
12878 extra information to be omitted and is useful when comparing two assembler
12883 @cindex global offset table
12885 Generate position-independent code (PIC) suitable for use in a shared
12886 library, if supported for the target machine. Such code accesses all
12887 constant addresses through a global offset table (GOT)@. The dynamic
12888 loader resolves the GOT entries when the program starts (the dynamic
12889 loader is not part of GCC; it is part of the operating system). If
12890 the GOT size for the linked executable exceeds a machine-specific
12891 maximum size, you get an error message from the linker indicating that
12892 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
12893 instead. (These maximums are 8k on the SPARC and 32k
12894 on the m68k and RS/6000. The 386 has no such limit.)
12896 Position-independent code requires special support, and therefore works
12897 only on certain machines. For the 386, GCC supports PIC for System V
12898 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
12899 position-independent.
12901 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
12906 If supported for the target machine, emit position-independent code,
12907 suitable for dynamic linking and avoiding any limit on the size of the
12908 global offset table. This option makes a difference on the m68k,
12909 PowerPC and SPARC@.
12911 Position-independent code requires special support, and therefore works
12912 only on certain machines.
12914 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
12921 These options are similar to @option{-fpic} and @option{-fPIC}, but
12922 generated position independent code can be only linked into executables.
12923 Usually these options are used when @option{-pie} GCC option will be
12924 used during linking.
12926 @item -fno-jump-tables
12927 @opindex fno-jump-tables
12928 Do not use jump tables for switch statements even where it would be
12929 more efficient than other code generation strategies. This option is
12930 of use in conjunction with @option{-fpic} or @option{-fPIC} for
12931 building code which forms part of a dynamic linker and cannot
12932 reference the address of a jump table. On some targets, jump tables
12933 do not require a GOT and this option is not needed.
12935 @item -ffixed-@var{reg}
12937 Treat the register named @var{reg} as a fixed register; generated code
12938 should never refer to it (except perhaps as a stack pointer, frame
12939 pointer or in some other fixed role).
12941 @var{reg} must be the name of a register. The register names accepted
12942 are machine-specific and are defined in the @code{REGISTER_NAMES}
12943 macro in the machine description macro file.
12945 This flag does not have a negative form, because it specifies a
12948 @item -fcall-used-@var{reg}
12949 @opindex fcall-used
12950 Treat the register named @var{reg} as an allocable register that is
12951 clobbered by function calls. It may be allocated for temporaries or
12952 variables that do not live across a call. Functions compiled this way
12953 will not save and restore the register @var{reg}.
12955 It is an error to used this flag with the frame pointer or stack pointer.
12956 Use of this flag for other registers that have fixed pervasive roles in
12957 the machine's execution model will produce disastrous results.
12959 This flag does not have a negative form, because it specifies a
12962 @item -fcall-saved-@var{reg}
12963 @opindex fcall-saved
12964 Treat the register named @var{reg} as an allocable register saved by
12965 functions. It may be allocated even for temporaries or variables that
12966 live across a call. Functions compiled this way will save and restore
12967 the register @var{reg} if they use it.
12969 It is an error to used this flag with the frame pointer or stack pointer.
12970 Use of this flag for other registers that have fixed pervasive roles in
12971 the machine's execution model will produce disastrous results.
12973 A different sort of disaster will result from the use of this flag for
12974 a register in which function values may be returned.
12976 This flag does not have a negative form, because it specifies a
12979 @item -fpack-struct[=@var{n}]
12980 @opindex fpack-struct
12981 Without a value specified, pack all structure members together without
12982 holes. When a value is specified (which must be a small power of two), pack
12983 structure members according to this value, representing the maximum
12984 alignment (that is, objects with default alignment requirements larger than
12985 this will be output potentially unaligned at the next fitting location.
12987 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
12988 code that is not binary compatible with code generated without that switch.
12989 Additionally, it makes the code suboptimal.
12990 Use it to conform to a non-default application binary interface.
12992 @item -finstrument-functions
12993 @opindex finstrument-functions
12994 Generate instrumentation calls for entry and exit to functions. Just
12995 after function entry and just before function exit, the following
12996 profiling functions will be called with the address of the current
12997 function and its call site. (On some platforms,
12998 @code{__builtin_return_address} does not work beyond the current
12999 function, so the call site information may not be available to the
13000 profiling functions otherwise.)
13003 void __cyg_profile_func_enter (void *this_fn,
13005 void __cyg_profile_func_exit (void *this_fn,
13009 The first argument is the address of the start of the current function,
13010 which may be looked up exactly in the symbol table.
13012 This instrumentation is also done for functions expanded inline in other
13013 functions. The profiling calls will indicate where, conceptually, the
13014 inline function is entered and exited. This means that addressable
13015 versions of such functions must be available. If all your uses of a
13016 function are expanded inline, this may mean an additional expansion of
13017 code size. If you use @samp{extern inline} in your C code, an
13018 addressable version of such functions must be provided. (This is
13019 normally the case anyways, but if you get lucky and the optimizer always
13020 expands the functions inline, you might have gotten away without
13021 providing static copies.)
13023 A function may be given the attribute @code{no_instrument_function}, in
13024 which case this instrumentation will not be done. This can be used, for
13025 example, for the profiling functions listed above, high-priority
13026 interrupt routines, and any functions from which the profiling functions
13027 cannot safely be called (perhaps signal handlers, if the profiling
13028 routines generate output or allocate memory).
13030 @item -fstack-check
13031 @opindex fstack-check
13032 Generate code to verify that you do not go beyond the boundary of the
13033 stack. You should specify this flag if you are running in an
13034 environment with multiple threads, but only rarely need to specify it in
13035 a single-threaded environment since stack overflow is automatically
13036 detected on nearly all systems if there is only one stack.
13038 Note that this switch does not actually cause checking to be done; the
13039 operating system must do that. The switch causes generation of code
13040 to ensure that the operating system sees the stack being extended.
13042 @item -fstack-limit-register=@var{reg}
13043 @itemx -fstack-limit-symbol=@var{sym}
13044 @itemx -fno-stack-limit
13045 @opindex fstack-limit-register
13046 @opindex fstack-limit-symbol
13047 @opindex fno-stack-limit
13048 Generate code to ensure that the stack does not grow beyond a certain value,
13049 either the value of a register or the address of a symbol. If the stack
13050 would grow beyond the value, a signal is raised. For most targets,
13051 the signal is raised before the stack overruns the boundary, so
13052 it is possible to catch the signal without taking special precautions.
13054 For instance, if the stack starts at absolute address @samp{0x80000000}
13055 and grows downwards, you can use the flags
13056 @option{-fstack-limit-symbol=__stack_limit} and
13057 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
13058 of 128KB@. Note that this may only work with the GNU linker.
13060 @cindex aliasing of parameters
13061 @cindex parameters, aliased
13062 @item -fargument-alias
13063 @itemx -fargument-noalias
13064 @itemx -fargument-noalias-global
13065 @opindex fargument-alias
13066 @opindex fargument-noalias
13067 @opindex fargument-noalias-global
13068 Specify the possible relationships among parameters and between
13069 parameters and global data.
13071 @option{-fargument-alias} specifies that arguments (parameters) may
13072 alias each other and may alias global storage.@*
13073 @option{-fargument-noalias} specifies that arguments do not alias
13074 each other, but may alias global storage.@*
13075 @option{-fargument-noalias-global} specifies that arguments do not
13076 alias each other and do not alias global storage.
13078 Each language will automatically use whatever option is required by
13079 the language standard. You should not need to use these options yourself.
13081 @item -fleading-underscore
13082 @opindex fleading-underscore
13083 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
13084 change the way C symbols are represented in the object file. One use
13085 is to help link with legacy assembly code.
13087 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
13088 generate code that is not binary compatible with code generated without that
13089 switch. Use it to conform to a non-default application binary interface.
13090 Not all targets provide complete support for this switch.
13092 @item -ftls-model=@var{model}
13093 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
13094 The @var{model} argument should be one of @code{global-dynamic},
13095 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
13097 The default without @option{-fpic} is @code{initial-exec}; with
13098 @option{-fpic} the default is @code{global-dynamic}.
13100 @item -fvisibility=@var{default|internal|hidden|protected}
13101 @opindex fvisibility
13102 Set the default ELF image symbol visibility to the specified option---all
13103 symbols will be marked with this unless overridden within the code.
13104 Using this feature can very substantially improve linking and
13105 load times of shared object libraries, produce more optimized
13106 code, provide near-perfect API export and prevent symbol clashes.
13107 It is @strong{strongly} recommended that you use this in any shared objects
13110 Despite the nomenclature, @code{default} always means public ie;
13111 available to be linked against from outside the shared object.
13112 @code{protected} and @code{internal} are pretty useless in real-world
13113 usage so the only other commonly used option will be @code{hidden}.
13114 The default if @option{-fvisibility} isn't specified is
13115 @code{default}, i.e., make every
13116 symbol public---this causes the same behavior as previous versions of
13119 A good explanation of the benefits offered by ensuring ELF
13120 symbols have the correct visibility is given by ``How To Write
13121 Shared Libraries'' by Ulrich Drepper (which can be found at
13122 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
13123 solution made possible by this option to marking things hidden when
13124 the default is public is to make the default hidden and mark things
13125 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
13126 and @code{__attribute__ ((visibility("default")))} instead of
13127 @code{__declspec(dllexport)} you get almost identical semantics with
13128 identical syntax. This is a great boon to those working with
13129 cross-platform projects.
13131 For those adding visibility support to existing code, you may find
13132 @samp{#pragma GCC visibility} of use. This works by you enclosing
13133 the declarations you wish to set visibility for with (for example)
13134 @samp{#pragma GCC visibility push(hidden)} and
13135 @samp{#pragma GCC visibility pop}.
13136 Bear in mind that symbol visibility should be viewed @strong{as
13137 part of the API interface contract} and thus all new code should
13138 always specify visibility when it is not the default ie; declarations
13139 only for use within the local DSO should @strong{always} be marked explicitly
13140 as hidden as so to avoid PLT indirection overheads---making this
13141 abundantly clear also aids readability and self-documentation of the code.
13142 Note that due to ISO C++ specification requirements, operator new and
13143 operator delete must always be of default visibility.
13145 An overview of these techniques, their benefits and how to use them
13146 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
13152 @node Environment Variables
13153 @section Environment Variables Affecting GCC
13154 @cindex environment variables
13156 @c man begin ENVIRONMENT
13157 This section describes several environment variables that affect how GCC
13158 operates. Some of them work by specifying directories or prefixes to use
13159 when searching for various kinds of files. Some are used to specify other
13160 aspects of the compilation environment.
13162 Note that you can also specify places to search using options such as
13163 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
13164 take precedence over places specified using environment variables, which
13165 in turn take precedence over those specified by the configuration of GCC@.
13166 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
13167 GNU Compiler Collection (GCC) Internals}.
13172 @c @itemx LC_COLLATE
13174 @c @itemx LC_MONETARY
13175 @c @itemx LC_NUMERIC
13180 @c @findex LC_COLLATE
13181 @findex LC_MESSAGES
13182 @c @findex LC_MONETARY
13183 @c @findex LC_NUMERIC
13187 These environment variables control the way that GCC uses
13188 localization information that allow GCC to work with different
13189 national conventions. GCC inspects the locale categories
13190 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
13191 so. These locale categories can be set to any value supported by your
13192 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
13193 Kingdom encoded in UTF-8.
13195 The @env{LC_CTYPE} environment variable specifies character
13196 classification. GCC uses it to determine the character boundaries in
13197 a string; this is needed for some multibyte encodings that contain quote
13198 and escape characters that would otherwise be interpreted as a string
13201 The @env{LC_MESSAGES} environment variable specifies the language to
13202 use in diagnostic messages.
13204 If the @env{LC_ALL} environment variable is set, it overrides the value
13205 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
13206 and @env{LC_MESSAGES} default to the value of the @env{LANG}
13207 environment variable. If none of these variables are set, GCC
13208 defaults to traditional C English behavior.
13212 If @env{TMPDIR} is set, it specifies the directory to use for temporary
13213 files. GCC uses temporary files to hold the output of one stage of
13214 compilation which is to be used as input to the next stage: for example,
13215 the output of the preprocessor, which is the input to the compiler
13218 @item GCC_EXEC_PREFIX
13219 @findex GCC_EXEC_PREFIX
13220 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
13221 names of the subprograms executed by the compiler. No slash is added
13222 when this prefix is combined with the name of a subprogram, but you can
13223 specify a prefix that ends with a slash if you wish.
13225 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
13226 an appropriate prefix to use based on the pathname it was invoked with.
13228 If GCC cannot find the subprogram using the specified prefix, it
13229 tries looking in the usual places for the subprogram.
13231 The default value of @env{GCC_EXEC_PREFIX} is
13232 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
13233 of @code{prefix} when you ran the @file{configure} script.
13235 Other prefixes specified with @option{-B} take precedence over this prefix.
13237 This prefix is also used for finding files such as @file{crt0.o} that are
13240 In addition, the prefix is used in an unusual way in finding the
13241 directories to search for header files. For each of the standard
13242 directories whose name normally begins with @samp{/usr/local/lib/gcc}
13243 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
13244 replacing that beginning with the specified prefix to produce an
13245 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
13246 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
13247 These alternate directories are searched first; the standard directories
13250 @item COMPILER_PATH
13251 @findex COMPILER_PATH
13252 The value of @env{COMPILER_PATH} is a colon-separated list of
13253 directories, much like @env{PATH}. GCC tries the directories thus
13254 specified when searching for subprograms, if it can't find the
13255 subprograms using @env{GCC_EXEC_PREFIX}.
13258 @findex LIBRARY_PATH
13259 The value of @env{LIBRARY_PATH} is a colon-separated list of
13260 directories, much like @env{PATH}. When configured as a native compiler,
13261 GCC tries the directories thus specified when searching for special
13262 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
13263 using GCC also uses these directories when searching for ordinary
13264 libraries for the @option{-l} option (but directories specified with
13265 @option{-L} come first).
13269 @cindex locale definition
13270 This variable is used to pass locale information to the compiler. One way in
13271 which this information is used is to determine the character set to be used
13272 when character literals, string literals and comments are parsed in C and C++.
13273 When the compiler is configured to allow multibyte characters,
13274 the following values for @env{LANG} are recognized:
13278 Recognize JIS characters.
13280 Recognize SJIS characters.
13282 Recognize EUCJP characters.
13285 If @env{LANG} is not defined, or if it has some other value, then the
13286 compiler will use mblen and mbtowc as defined by the default locale to
13287 recognize and translate multibyte characters.
13291 Some additional environments variables affect the behavior of the
13294 @include cppenv.texi
13298 @node Precompiled Headers
13299 @section Using Precompiled Headers
13300 @cindex precompiled headers
13301 @cindex speed of compilation
13303 Often large projects have many header files that are included in every
13304 source file. The time the compiler takes to process these header files
13305 over and over again can account for nearly all of the time required to
13306 build the project. To make builds faster, GCC allows users to
13307 `precompile' a header file; then, if builds can use the precompiled
13308 header file they will be much faster.
13310 To create a precompiled header file, simply compile it as you would any
13311 other file, if necessary using the @option{-x} option to make the driver
13312 treat it as a C or C++ header file. You will probably want to use a
13313 tool like @command{make} to keep the precompiled header up-to-date when
13314 the headers it contains change.
13316 A precompiled header file will be searched for when @code{#include} is
13317 seen in the compilation. As it searches for the included file
13318 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
13319 compiler looks for a precompiled header in each directory just before it
13320 looks for the include file in that directory. The name searched for is
13321 the name specified in the @code{#include} with @samp{.gch} appended. If
13322 the precompiled header file can't be used, it is ignored.
13324 For instance, if you have @code{#include "all.h"}, and you have
13325 @file{all.h.gch} in the same directory as @file{all.h}, then the
13326 precompiled header file will be used if possible, and the original
13327 header will be used otherwise.
13329 Alternatively, you might decide to put the precompiled header file in a
13330 directory and use @option{-I} to ensure that directory is searched
13331 before (or instead of) the directory containing the original header.
13332 Then, if you want to check that the precompiled header file is always
13333 used, you can put a file of the same name as the original header in this
13334 directory containing an @code{#error} command.
13336 This also works with @option{-include}. So yet another way to use
13337 precompiled headers, good for projects not designed with precompiled
13338 header files in mind, is to simply take most of the header files used by
13339 a project, include them from another header file, precompile that header
13340 file, and @option{-include} the precompiled header. If the header files
13341 have guards against multiple inclusion, they will be skipped because
13342 they've already been included (in the precompiled header).
13344 If you need to precompile the same header file for different
13345 languages, targets, or compiler options, you can instead make a
13346 @emph{directory} named like @file{all.h.gch}, and put each precompiled
13347 header in the directory, perhaps using @option{-o}. It doesn't matter
13348 what you call the files in the directory, every precompiled header in
13349 the directory will be considered. The first precompiled header
13350 encountered in the directory that is valid for this compilation will
13351 be used; they're searched in no particular order.
13353 There are many other possibilities, limited only by your imagination,
13354 good sense, and the constraints of your build system.
13356 A precompiled header file can be used only when these conditions apply:
13360 Only one precompiled header can be used in a particular compilation.
13363 A precompiled header can't be used once the first C token is seen. You
13364 can have preprocessor directives before a precompiled header; you can
13365 even include a precompiled header from inside another header, so long as
13366 there are no C tokens before the @code{#include}.
13369 The precompiled header file must be produced for the same language as
13370 the current compilation. You can't use a C precompiled header for a C++
13374 The precompiled header file must have been produced by the same compiler
13375 binary as the current compilation is using.
13378 Any macros defined before the precompiled header is included must
13379 either be defined in the same way as when the precompiled header was
13380 generated, or must not affect the precompiled header, which usually
13381 means that they don't appear in the precompiled header at all.
13383 The @option{-D} option is one way to define a macro before a
13384 precompiled header is included; using a @code{#define} can also do it.
13385 There are also some options that define macros implicitly, like
13386 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
13389 @item If debugging information is output when using the precompiled
13390 header, using @option{-g} or similar, the same kind of debugging information
13391 must have been output when building the precompiled header. However,
13392 a precompiled header built using @option{-g} can be used in a compilation
13393 when no debugging information is being output.
13395 @item The same @option{-m} options must generally be used when building
13396 and using the precompiled header. @xref{Submodel Options},
13397 for any cases where this rule is relaxed.
13399 @item Each of the following options must be the same when building and using
13400 the precompiled header:
13402 @gccoptlist{-fexceptions -funit-at-a-time}
13405 Some other command-line options starting with @option{-f},
13406 @option{-p}, or @option{-O} must be defined in the same way as when
13407 the precompiled header was generated. At present, it's not clear
13408 which options are safe to change and which are not; the safest choice
13409 is to use exactly the same options when generating and using the
13410 precompiled header. The following are known to be safe:
13412 @gccoptlist{-fmessage-length= -fpreprocessed
13413 -fsched-interblock -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous
13414 -fsched-verbose=<number> -fschedule-insns -fvisibility=
13419 For all of these except the last, the compiler will automatically
13420 ignore the precompiled header if the conditions aren't met. If you
13421 find an option combination that doesn't work and doesn't cause the
13422 precompiled header to be ignored, please consider filing a bug report,
13425 If you do use differing options when generating and using the
13426 precompiled header, the actual behavior will be a mixture of the
13427 behavior for the options. For instance, if you use @option{-g} to
13428 generate the precompiled header but not when using it, you may or may
13429 not get debugging information for routines in the precompiled header.
13431 @node Running Protoize
13432 @section Running Protoize
13434 The program @code{protoize} is an optional part of GCC@. You can use
13435 it to add prototypes to a program, thus converting the program to ISO
13436 C in one respect. The companion program @code{unprotoize} does the
13437 reverse: it removes argument types from any prototypes that are found.
13439 When you run these programs, you must specify a set of source files as
13440 command line arguments. The conversion programs start out by compiling
13441 these files to see what functions they define. The information gathered
13442 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
13444 After scanning comes actual conversion. The specified files are all
13445 eligible to be converted; any files they include (whether sources or
13446 just headers) are eligible as well.
13448 But not all the eligible files are converted. By default,
13449 @code{protoize} and @code{unprotoize} convert only source and header
13450 files in the current directory. You can specify additional directories
13451 whose files should be converted with the @option{-d @var{directory}}
13452 option. You can also specify particular files to exclude with the
13453 @option{-x @var{file}} option. A file is converted if it is eligible, its
13454 directory name matches one of the specified directory names, and its
13455 name within the directory has not been excluded.
13457 Basic conversion with @code{protoize} consists of rewriting most
13458 function definitions and function declarations to specify the types of
13459 the arguments. The only ones not rewritten are those for varargs
13462 @code{protoize} optionally inserts prototype declarations at the
13463 beginning of the source file, to make them available for any calls that
13464 precede the function's definition. Or it can insert prototype
13465 declarations with block scope in the blocks where undeclared functions
13468 Basic conversion with @code{unprotoize} consists of rewriting most
13469 function declarations to remove any argument types, and rewriting
13470 function definitions to the old-style pre-ISO form.
13472 Both conversion programs print a warning for any function declaration or
13473 definition that they can't convert. You can suppress these warnings
13476 The output from @code{protoize} or @code{unprotoize} replaces the
13477 original source file. The original file is renamed to a name ending
13478 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
13479 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
13480 for DOS) file already exists, then the source file is simply discarded.
13482 @code{protoize} and @code{unprotoize} both depend on GCC itself to
13483 scan the program and collect information about the functions it uses.
13484 So neither of these programs will work until GCC is installed.
13486 Here is a table of the options you can use with @code{protoize} and
13487 @code{unprotoize}. Each option works with both programs unless
13491 @item -B @var{directory}
13492 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
13493 usual directory (normally @file{/usr/local/lib}). This file contains
13494 prototype information about standard system functions. This option
13495 applies only to @code{protoize}.
13497 @item -c @var{compilation-options}
13498 Use @var{compilation-options} as the options when running @command{gcc} to
13499 produce the @samp{.X} files. The special option @option{-aux-info} is
13500 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
13502 Note that the compilation options must be given as a single argument to
13503 @code{protoize} or @code{unprotoize}. If you want to specify several
13504 @command{gcc} options, you must quote the entire set of compilation options
13505 to make them a single word in the shell.
13507 There are certain @command{gcc} arguments that you cannot use, because they
13508 would produce the wrong kind of output. These include @option{-g},
13509 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
13510 the @var{compilation-options}, they are ignored.
13513 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
13514 systems) instead of @samp{.c}. This is convenient if you are converting
13515 a C program to C++. This option applies only to @code{protoize}.
13518 Add explicit global declarations. This means inserting explicit
13519 declarations at the beginning of each source file for each function
13520 that is called in the file and was not declared. These declarations
13521 precede the first function definition that contains a call to an
13522 undeclared function. This option applies only to @code{protoize}.
13524 @item -i @var{string}
13525 Indent old-style parameter declarations with the string @var{string}.
13526 This option applies only to @code{protoize}.
13528 @code{unprotoize} converts prototyped function definitions to old-style
13529 function definitions, where the arguments are declared between the
13530 argument list and the initial @samp{@{}. By default, @code{unprotoize}
13531 uses five spaces as the indentation. If you want to indent with just
13532 one space instead, use @option{-i " "}.
13535 Keep the @samp{.X} files. Normally, they are deleted after conversion
13539 Add explicit local declarations. @code{protoize} with @option{-l} inserts
13540 a prototype declaration for each function in each block which calls the
13541 function without any declaration. This option applies only to
13545 Make no real changes. This mode just prints information about the conversions
13546 that would have been done without @option{-n}.
13549 Make no @samp{.save} files. The original files are simply deleted.
13550 Use this option with caution.
13552 @item -p @var{program}
13553 Use the program @var{program} as the compiler. Normally, the name
13554 @file{gcc} is used.
13557 Work quietly. Most warnings are suppressed.
13560 Print the version number, just like @option{-v} for @command{gcc}.
13563 If you need special compiler options to compile one of your program's
13564 source files, then you should generate that file's @samp{.X} file
13565 specially, by running @command{gcc} on that source file with the
13566 appropriate options and the option @option{-aux-info}. Then run
13567 @code{protoize} on the entire set of files. @code{protoize} will use
13568 the existing @samp{.X} file because it is newer than the source file.
13572 gcc -Dfoo=bar file1.c -aux-info file1.X
13577 You need to include the special files along with the rest in the
13578 @code{protoize} command, even though their @samp{.X} files already
13579 exist, because otherwise they won't get converted.
13581 @xref{Protoize Caveats}, for more information on how to use
13582 @code{protoize} successfully.