1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005, 2006 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, 1998,
13 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006 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 @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-* -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 -Woverlength-strings -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 -Wpointer-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 -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-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
323 -fomit-frame-pointer -foptimize-register-move @gol
324 -foptimize-sibling-calls -fprefetch-loop-arrays -fprefetch-loop-arrays-rtl @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 -frtl-abstract-sequences @gol
330 -fschedule-insns -fschedule-insns2 @gol
331 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
332 -fsched-spec-load-dangerous @gol
333 -fsched-stalled-insns=@var{n} -fsched-stalled-insns-dep=@var{n} @gol
334 -fsched2-use-superblocks @gol
335 -fsched2-use-traces -freschedule-modulo-scheduled-loops @gol
336 -fsection-anchors -fsignaling-nans -fsingle-precision-constant @gol
337 -fstack-protector -fstack-protector-all @gol
338 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
339 -funroll-all-loops -funroll-loops -fpeel-loops @gol
340 -fsplit-ivs-in-unroller -funswitch-loops @gol
341 -fvariable-expansion-in-unroller @gol
342 -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol
343 -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts @gol
344 -ftree-dominator-opts -ftree-dse -ftree-copyrename -ftree-sink @gol
345 -ftree-ch -ftree-sra -ftree-ter -ftree-lrs -ftree-fre -ftree-vectorize @gol
346 -ftree-vect-loop-version -ftree-salias -fipa-pta -fweb @gol
347 -ftree-copy-prop -ftree-store-ccp -ftree-store-copy-prop -fwhole-program @gol
348 --param @var{name}=@var{value}
349 -O -O0 -O1 -O2 -O3 -Os}
351 @item Preprocessor Options
352 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
353 @gccoptlist{-A@var{question}=@var{answer} @gol
354 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
355 -C -dD -dI -dM -dN @gol
356 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
357 -idirafter @var{dir} @gol
358 -include @var{file} -imacros @var{file} @gol
359 -iprefix @var{file} -iwithprefix @var{dir} @gol
360 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
361 -imultilib @var{dir} -isysroot @var{dir} @gol
362 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
363 -P -fworking-directory -remap @gol
364 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
365 -Xpreprocessor @var{option}}
367 @item Assembler Option
368 @xref{Assembler Options,,Passing Options to the Assembler}.
369 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
372 @xref{Link Options,,Options for Linking}.
373 @gccoptlist{@var{object-file-name} -l@var{library} @gol
374 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
375 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
376 -Wl,@var{option} -Xlinker @var{option} @gol
379 @item Directory Options
380 @xref{Directory Options,,Options for Directory Search}.
381 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
382 -specs=@var{file} -I- --sysroot=@var{dir}}
385 @c I wrote this xref this way to avoid overfull hbox. -- rms
386 @xref{Target Options}.
387 @gccoptlist{-V @var{version} -b @var{machine}}
389 @item Machine Dependent Options
390 @xref{Submodel Options,,Hardware Models and Configurations}.
391 @c This list is ordered alphanumerically by subsection name.
392 @c Try and put the significant identifier (CPU or system) first,
393 @c so users have a clue at guessing where the ones they want will be.
396 @gccoptlist{-EB -EL @gol
397 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
398 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
401 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
402 -mabi=@var{name} @gol
403 -mapcs-stack-check -mno-apcs-stack-check @gol
404 -mapcs-float -mno-apcs-float @gol
405 -mapcs-reentrant -mno-apcs-reentrant @gol
406 -msched-prolog -mno-sched-prolog @gol
407 -mlittle-endian -mbig-endian -mwords-little-endian @gol
408 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
409 -mthumb-interwork -mno-thumb-interwork @gol
410 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
411 -mstructure-size-boundary=@var{n} @gol
412 -mabort-on-noreturn @gol
413 -mlong-calls -mno-long-calls @gol
414 -msingle-pic-base -mno-single-pic-base @gol
415 -mpic-register=@var{reg} @gol
416 -mnop-fun-dllimport @gol
417 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
418 -mpoke-function-name @gol
420 -mtpcs-frame -mtpcs-leaf-frame @gol
421 -mcaller-super-interworking -mcallee-super-interworking @gol
425 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
426 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
428 @emph{Blackfin Options}
429 @gccoptlist{-momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
430 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
431 -mlow-64k -mno-low64k -mid-shared-library @gol
432 -mno-id-shared-library -mshared-library-id=@var{n} @gol
433 -mlong-calls -mno-long-calls}
436 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
437 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
438 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
439 -mstack-align -mdata-align -mconst-align @gol
440 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
441 -melf -maout -melinux -mlinux -sim -sim2 @gol
442 -mmul-bug-workaround -mno-mul-bug-workaround}
445 @gccoptlist{-mmac -mpush-args}
447 @emph{Darwin Options}
448 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
449 -arch_only -bind_at_load -bundle -bundle_loader @gol
450 -client_name -compatibility_version -current_version @gol
452 -dependency-file -dylib_file -dylinker_install_name @gol
453 -dynamic -dynamiclib -exported_symbols_list @gol
454 -filelist -flat_namespace -force_cpusubtype_ALL @gol
455 -force_flat_namespace -headerpad_max_install_names @gol
456 -image_base -init -install_name -keep_private_externs @gol
457 -multi_module -multiply_defined -multiply_defined_unused @gol
458 -noall_load -no_dead_strip_inits_and_terms @gol
459 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
460 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
461 -private_bundle -read_only_relocs -sectalign @gol
462 -sectobjectsymbols -whyload -seg1addr @gol
463 -sectcreate -sectobjectsymbols -sectorder @gol
464 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
465 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
466 -segprot -segs_read_only_addr -segs_read_write_addr @gol
467 -single_module -static -sub_library -sub_umbrella @gol
468 -twolevel_namespace -umbrella -undefined @gol
469 -unexported_symbols_list -weak_reference_mismatches @gol
470 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
473 @emph{DEC Alpha Options}
474 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
475 -mieee -mieee-with-inexact -mieee-conformant @gol
476 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
477 -mtrap-precision=@var{mode} -mbuild-constants @gol
478 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
479 -mbwx -mmax -mfix -mcix @gol
480 -mfloat-vax -mfloat-ieee @gol
481 -mexplicit-relocs -msmall-data -mlarge-data @gol
482 -msmall-text -mlarge-text @gol
483 -mmemory-latency=@var{time}}
485 @emph{DEC Alpha/VMS Options}
486 @gccoptlist{-mvms-return-codes}
489 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
490 -mhard-float -msoft-float @gol
491 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
492 -mdouble -mno-double @gol
493 -mmedia -mno-media -mmuladd -mno-muladd @gol
494 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
495 -mlinked-fp -mlong-calls -malign-labels @gol
496 -mlibrary-pic -macc-4 -macc-8 @gol
497 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
498 -moptimize-membar -mno-optimize-membar @gol
499 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
500 -mvliw-branch -mno-vliw-branch @gol
501 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
502 -mno-nested-cond-exec -mtomcat-stats @gol
506 @emph{GNU/Linux Options}
507 @gccoptlist{-muclibc}
509 @emph{H8/300 Options}
510 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
513 @gccoptlist{-march=@var{architecture-type} @gol
514 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
515 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
516 -mfixed-range=@var{register-range} @gol
517 -mjump-in-delay -mlinker-opt -mlong-calls @gol
518 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
519 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
520 -mno-jump-in-delay -mno-long-load-store @gol
521 -mno-portable-runtime -mno-soft-float @gol
522 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
523 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
524 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
525 -munix=@var{unix-std} -nolibdld -static -threads}
527 @emph{i386 and x86-64 Options}
528 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
529 -mfpmath=@var{unit} @gol
530 -masm=@var{dialect} -mno-fancy-math-387 @gol
531 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
532 -mno-wide-multiply -mrtd -malign-double @gol
533 -mpreferred-stack-boundary=@var{num} @gol
534 -mmmx -msse -msse2 -msse3 -m3dnow -msselibm @gol
535 -mthreads -mno-align-stringops -minline-all-stringops @gol
536 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
537 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
538 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
539 -mcmodel=@var{code-model} @gol
540 -m32 -m64 -mlarge-data-threshold=@var{num}}
543 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
544 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
545 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
546 -minline-float-divide-max-throughput @gol
547 -minline-int-divide-min-latency @gol
548 -minline-int-divide-max-throughput @gol
549 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
550 -mno-dwarf2-asm -mearly-stop-bits @gol
551 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
552 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64}
554 @emph{M32R/D Options}
555 @gccoptlist{-m32r2 -m32rx -m32r @gol
557 -malign-loops -mno-align-loops @gol
558 -missue-rate=@var{number} @gol
559 -mbranch-cost=@var{number} @gol
560 -mmodel=@var{code-size-model-type} @gol
561 -msdata=@var{sdata-type} @gol
562 -mno-flush-func -mflush-func=@var{name} @gol
563 -mno-flush-trap -mflush-trap=@var{number} @gol
567 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
569 @emph{M680x0 Options}
570 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
571 -m68060 -mcpu32 -m5200 -mcfv4e -m68881 -mbitfield @gol
572 -mc68000 -mc68020 @gol
573 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
574 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
575 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
577 @emph{M68hc1x Options}
578 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
579 -mauto-incdec -minmax -mlong-calls -mshort @gol
580 -msoft-reg-count=@var{count}}
583 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
584 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
585 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
586 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
587 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
590 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
591 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
592 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
593 -mxgot -mno-xgot -mgp32 -mgp64 -mfp32 -mfp64 @gol
594 -mhard-float -msoft-float -msingle-float -mdouble-float @gol
595 -mdsp -mpaired-single -mips3d @gol
596 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
597 -G@var{num} -membedded-data -mno-embedded-data @gol
598 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
599 -msplit-addresses -mno-split-addresses @gol
600 -mexplicit-relocs -mno-explicit-relocs @gol
601 -mcheck-zero-division -mno-check-zero-division @gol
602 -mdivide-traps -mdivide-breaks @gol
603 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
604 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
605 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
606 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 @gol
607 -mfix-sb1 -mno-fix-sb1 @gol
608 -mflush-func=@var{func} -mno-flush-func @gol
609 -mbranch-likely -mno-branch-likely @gol
610 -mfp-exceptions -mno-fp-exceptions @gol
611 -mvr4130-align -mno-vr4130-align}
614 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
615 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
616 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
617 -mno-base-addresses -msingle-exit -mno-single-exit}
619 @emph{MN10300 Options}
620 @gccoptlist{-mmult-bug -mno-mult-bug @gol
621 -mam33 -mno-am33 @gol
622 -mam33-2 -mno-am33-2 @gol
623 -mreturn-pointer-on-d0 @gol
627 @gccoptlist{-mno-crt0 -mbacc -msim @gol
628 -march=@var{cpu-type} }
630 @emph{PDP-11 Options}
631 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
632 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
633 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
634 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
635 -mbranch-expensive -mbranch-cheap @gol
636 -msplit -mno-split -munix-asm -mdec-asm}
638 @emph{PowerPC Options}
639 See RS/6000 and PowerPC Options.
641 @emph{RS/6000 and PowerPC Options}
642 @gccoptlist{-mcpu=@var{cpu-type} @gol
643 -mtune=@var{cpu-type} @gol
644 -mpower -mno-power -mpower2 -mno-power2 @gol
645 -mpowerpc -mpowerpc64 -mno-powerpc @gol
646 -maltivec -mno-altivec @gol
647 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
648 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
649 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
650 -mnew-mnemonics -mold-mnemonics @gol
651 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
652 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
653 -malign-power -malign-natural @gol
654 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
655 -mstring -mno-string -mupdate -mno-update @gol
656 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
657 -mstrict-align -mno-strict-align -mrelocatable @gol
658 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
659 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
660 -mdynamic-no-pic -maltivec -mswdiv @gol
661 -mprioritize-restricted-insns=@var{priority} @gol
662 -msched-costly-dep=@var{dependence_type} @gol
663 -minsert-sched-nops=@var{scheme} @gol
664 -mcall-sysv -mcall-netbsd @gol
665 -maix-struct-return -msvr4-struct-return @gol
666 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
667 -misel -mno-isel @gol
668 -misel=yes -misel=no @gol
670 -mspe=yes -mspe=no @gol
671 -mvrsave -mno-vrsave @gol
672 -mmulhw -mno-mulhw @gol
673 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
674 -mprototype -mno-prototype @gol
675 -msim -mmvme -mads -myellowknife -memb -msdata @gol
676 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
678 @emph{S/390 and zSeries Options}
679 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
680 -mhard-float -msoft-float -mlong-double-64 -mlong-double-128 @gol
681 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
682 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
683 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
684 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
685 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
688 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
689 -m4-nofpu -m4-single-only -m4-single -m4 @gol
690 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
691 -m5-64media -m5-64media-nofpu @gol
692 -m5-32media -m5-32media-nofpu @gol
693 -m5-compact -m5-compact-nofpu @gol
694 -mb -ml -mdalign -mrelax @gol
695 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
696 -mieee -misize -mpadstruct -mspace @gol
697 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
698 -mdivsi3_libfunc=@var{name} @gol
699 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
703 @gccoptlist{-mcpu=@var{cpu-type} @gol
704 -mtune=@var{cpu-type} @gol
705 -mcmodel=@var{code-model} @gol
706 -m32 -m64 -mapp-regs -mno-app-regs @gol
707 -mfaster-structs -mno-faster-structs @gol
708 -mfpu -mno-fpu -mhard-float -msoft-float @gol
709 -mhard-quad-float -msoft-quad-float @gol
710 -mimpure-text -mno-impure-text -mlittle-endian @gol
711 -mstack-bias -mno-stack-bias @gol
712 -munaligned-doubles -mno-unaligned-doubles @gol
713 -mv8plus -mno-v8plus -mvis -mno-vis
714 -threads -pthreads -pthread}
716 @emph{System V Options}
717 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
719 @emph{TMS320C3x/C4x Options}
720 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
721 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
722 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
723 -mparallel-insns -mparallel-mpy -mpreserve-float}
726 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
727 -mprolog-function -mno-prolog-function -mspace @gol
728 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
729 -mapp-regs -mno-app-regs @gol
730 -mdisable-callt -mno-disable-callt @gol
736 @gccoptlist{-mg -mgnu -munix}
738 @emph{x86-64 Options}
739 See i386 and x86-64 Options.
741 @emph{Xstormy16 Options}
744 @emph{Xtensa Options}
745 @gccoptlist{-mconst16 -mno-const16 @gol
746 -mfused-madd -mno-fused-madd @gol
747 -mtext-section-literals -mno-text-section-literals @gol
748 -mtarget-align -mno-target-align @gol
749 -mlongcalls -mno-longcalls}
751 @emph{zSeries Options}
752 See S/390 and zSeries Options.
754 @item Code Generation Options
755 @xref{Code Gen Options,,Options for Code Generation Conventions}.
756 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
757 -ffixed-@var{reg} -fexceptions @gol
758 -fnon-call-exceptions -funwind-tables @gol
759 -fasynchronous-unwind-tables @gol
760 -finhibit-size-directive -finstrument-functions @gol
761 -fno-common -fno-ident @gol
762 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
763 -fno-jump-tables @gol
764 -freg-struct-return -fshort-enums @gol
765 -fshort-double -fshort-wchar @gol
766 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
767 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
768 -fargument-alias -fargument-noalias @gol
769 -fargument-noalias-global -fleading-underscore @gol
770 -ftls-model=@var{model} @gol
771 -ftrapv -fwrapv -fbounds-check @gol
772 -fvisibility -fopenmp}
776 * Overall Options:: Controlling the kind of output:
777 an executable, object files, assembler files,
778 or preprocessed source.
779 * C Dialect Options:: Controlling the variant of C language compiled.
780 * C++ Dialect Options:: Variations on C++.
781 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
783 * Language Independent Options:: Controlling how diagnostics should be
785 * Warning Options:: How picky should the compiler be?
786 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
787 * Optimize Options:: How much optimization?
788 * Preprocessor Options:: Controlling header files and macro definitions.
789 Also, getting dependency information for Make.
790 * Assembler Options:: Passing options to the assembler.
791 * Link Options:: Specifying libraries and so on.
792 * Directory Options:: Where to find header files and libraries.
793 Where to find the compiler executable files.
794 * Spec Files:: How to pass switches to sub-processes.
795 * Target Options:: Running a cross-compiler, or an old version of GCC.
798 @node Overall Options
799 @section Options Controlling the Kind of Output
801 Compilation can involve up to four stages: preprocessing, compilation
802 proper, assembly and linking, always in that order. GCC is capable of
803 preprocessing and compiling several files either into several
804 assembler input files, or into one assembler input file; then each
805 assembler input file produces an object file, and linking combines all
806 the object files (those newly compiled, and those specified as input)
807 into an executable file.
809 @cindex file name suffix
810 For any given input file, the file name suffix determines what kind of
815 C source code which must be preprocessed.
818 C source code which should not be preprocessed.
821 C++ source code which should not be preprocessed.
824 Objective-C source code. Note that you must link with the @file{libobjc}
825 library to make an Objective-C program work.
828 Objective-C source code which should not be preprocessed.
832 Objective-C++ source code. Note that you must link with the @file{libobjc}
833 library to make an Objective-C++ program work. Note that @samp{.M} refers
834 to a literal capital M@.
837 Objective-C++ source code which should not be preprocessed.
840 C, C++, Objective-C or Objective-C++ header file to be turned into a
845 @itemx @var{file}.cxx
846 @itemx @var{file}.cpp
847 @itemx @var{file}.CPP
848 @itemx @var{file}.c++
850 C++ source code which must be preprocessed. Note that in @samp{.cxx},
851 the last two letters must both be literally @samp{x}. Likewise,
852 @samp{.C} refers to a literal capital C@.
856 Objective-C++ source code which must be preprocessed.
859 Objective-C++ source code which should not be preprocessed.
863 C++ header file to be turned into a precompiled header.
866 @itemx @var{file}.for
867 @itemx @var{file}.FOR
868 Fixed form Fortran source code which should not be preprocessed.
871 @itemx @var{file}.fpp
872 @itemx @var{file}.FPP
873 Fixed form Fortran source code which must be preprocessed (with the traditional
877 @itemx @var{file}.f95
878 Free form Fortran source code which should not be preprocessed.
881 @itemx @var{file}.F95
882 Free form Fortran source code which must be preprocessed (with the
883 traditional preprocessor).
885 @c FIXME: Descriptions of Java file types.
892 Ada source code file which contains a library unit declaration (a
893 declaration of a package, subprogram, or generic, or a generic
894 instantiation), or a library unit renaming declaration (a package,
895 generic, or subprogram renaming declaration). Such files are also
898 @itemx @var{file}.adb
899 Ada source code file containing a library unit body (a subprogram or
900 package body). Such files are also called @dfn{bodies}.
902 @c GCC also knows about some suffixes for languages not yet included:
913 Assembler code which must be preprocessed.
916 An object file to be fed straight into linking.
917 Any file name with no recognized suffix is treated this way.
921 You can specify the input language explicitly with the @option{-x} option:
924 @item -x @var{language}
925 Specify explicitly the @var{language} for the following input files
926 (rather than letting the compiler choose a default based on the file
927 name suffix). This option applies to all following input files until
928 the next @option{-x} option. Possible values for @var{language} are:
930 c c-header c-cpp-output
931 c++ c++-header c++-cpp-output
932 objective-c objective-c-header objective-c-cpp-output
933 objective-c++ objective-c++-header objective-c++-cpp-output
934 assembler assembler-with-cpp
943 Turn off any specification of a language, so that subsequent files are
944 handled according to their file name suffixes (as they are if @option{-x}
945 has not been used at all).
947 @item -pass-exit-codes
948 @opindex pass-exit-codes
949 Normally the @command{gcc} program will exit with the code of 1 if any
950 phase of the compiler returns a non-success return code. If you specify
951 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
952 numerically highest error produced by any phase that returned an error
956 If you only want some of the stages of compilation, you can use
957 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
958 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
959 @command{gcc} is to stop. Note that some combinations (for example,
960 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
965 Compile or assemble the source files, but do not link. The linking
966 stage simply is not done. The ultimate output is in the form of an
967 object file for each source file.
969 By default, the object file name for a source file is made by replacing
970 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
972 Unrecognized input files, not requiring compilation or assembly, are
977 Stop after the stage of compilation proper; do not assemble. The output
978 is in the form of an assembler code file for each non-assembler input
981 By default, the assembler file name for a source file is made by
982 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
984 Input files that don't require compilation are ignored.
988 Stop after the preprocessing stage; do not run the compiler proper. The
989 output is in the form of preprocessed source code, which is sent to the
992 Input files which don't require preprocessing are ignored.
994 @cindex output file option
997 Place output in file @var{file}. This applies regardless to whatever
998 sort of output is being produced, whether it be an executable file,
999 an object file, an assembler file or preprocessed C code.
1001 If @option{-o} is not specified, the default is to put an executable
1002 file in @file{a.out}, the object file for
1003 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1004 assembler file in @file{@var{source}.s}, a precompiled header file in
1005 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1010 Print (on standard error output) the commands executed to run the stages
1011 of compilation. Also print the version number of the compiler driver
1012 program and of the preprocessor and the compiler proper.
1016 Like @option{-v} except the commands are not executed and all command
1017 arguments are quoted. This is useful for shell scripts to capture the
1018 driver-generated command lines.
1022 Use pipes rather than temporary files for communication between the
1023 various stages of compilation. This fails to work on some systems where
1024 the assembler is unable to read from a pipe; but the GNU assembler has
1029 If you are compiling multiple source files, this option tells the driver
1030 to pass all the source files to the compiler at once (for those
1031 languages for which the compiler can handle this). This will allow
1032 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1033 language for which this is supported is C@. If you pass source files for
1034 multiple languages to the driver, using this option, the driver will invoke
1035 the compiler(s) that support IMA once each, passing each compiler all the
1036 source files appropriate for it. For those languages that do not support
1037 IMA this option will be ignored, and the compiler will be invoked once for
1038 each source file in that language. If you use this option in conjunction
1039 with @option{-save-temps}, the compiler will generate multiple
1041 (one for each source file), but only one (combined) @file{.o} or
1046 Print (on the standard output) a description of the command line options
1047 understood by @command{gcc}. If the @option{-v} option is also specified
1048 then @option{--help} will also be passed on to the various processes
1049 invoked by @command{gcc}, so that they can display the command line options
1050 they accept. If the @option{-Wextra} option is also specified then command
1051 line options which have no documentation associated with them will also
1055 @opindex target-help
1056 Print (on the standard output) a description of target specific command
1057 line options for each tool.
1061 Display the version number and copyrights of the invoked GCC@.
1063 @include @value{srcdir}/../libiberty/at-file.texi
1067 @section Compiling C++ Programs
1069 @cindex suffixes for C++ source
1070 @cindex C++ source file suffixes
1071 C++ source files conventionally use one of the suffixes @samp{.C},
1072 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1073 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
1074 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1075 files with these names and compiles them as C++ programs even if you
1076 call the compiler the same way as for compiling C programs (usually
1077 with the name @command{gcc}).
1081 However, C++ programs often require class libraries as well as a
1082 compiler that understands the C++ language---and under some
1083 circumstances, you might want to compile programs or header files from
1084 standard input, or otherwise without a suffix that flags them as C++
1085 programs. You might also like to precompile a C header file with a
1086 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
1087 program that calls GCC with the default language set to C++, and
1088 automatically specifies linking against the C++ library. On many
1089 systems, @command{g++} is also installed with the name @command{c++}.
1091 @cindex invoking @command{g++}
1092 When you compile C++ programs, you may specify many of the same
1093 command-line options that you use for compiling programs in any
1094 language; or command-line options meaningful for C and related
1095 languages; or options that are meaningful only for C++ programs.
1096 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1097 explanations of options for languages related to C@.
1098 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1099 explanations of options that are meaningful only for C++ programs.
1101 @node C Dialect Options
1102 @section Options Controlling C Dialect
1103 @cindex dialect options
1104 @cindex language dialect options
1105 @cindex options, dialect
1107 The following options control the dialect of C (or languages derived
1108 from C, such as C++, Objective-C and Objective-C++) that the compiler
1112 @cindex ANSI support
1116 In C mode, support all ISO C90 programs. In C++ mode,
1117 remove GNU extensions that conflict with ISO C++.
1119 This turns off certain features of GCC that are incompatible with ISO
1120 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1121 such as the @code{asm} and @code{typeof} keywords, and
1122 predefined macros such as @code{unix} and @code{vax} that identify the
1123 type of system you are using. It also enables the undesirable and
1124 rarely used ISO trigraph feature. For the C compiler,
1125 it disables recognition of C++ style @samp{//} comments as well as
1126 the @code{inline} keyword.
1128 The alternate keywords @code{__asm__}, @code{__extension__},
1129 @code{__inline__} and @code{__typeof__} continue to work despite
1130 @option{-ansi}. You would not want to use them in an ISO C program, of
1131 course, but it is useful to put them in header files that might be included
1132 in compilations done with @option{-ansi}. Alternate predefined macros
1133 such as @code{__unix__} and @code{__vax__} are also available, with or
1134 without @option{-ansi}.
1136 The @option{-ansi} option does not cause non-ISO programs to be
1137 rejected gratuitously. For that, @option{-pedantic} is required in
1138 addition to @option{-ansi}. @xref{Warning Options}.
1140 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1141 option is used. Some header files may notice this macro and refrain
1142 from declaring certain functions or defining certain macros that the
1143 ISO standard doesn't call for; this is to avoid interfering with any
1144 programs that might use these names for other things.
1146 Functions which would normally be built in but do not have semantics
1147 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1148 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1149 built-in functions provided by GCC}, for details of the functions
1154 Determine the language standard. This option is currently only
1155 supported when compiling C or C++. A value for this option must be
1156 provided; possible values are
1161 ISO C90 (same as @option{-ansi}).
1163 @item iso9899:199409
1164 ISO C90 as modified in amendment 1.
1170 ISO C99. Note that this standard is not yet fully supported; see
1171 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1172 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1175 Default, ISO C90 plus GNU extensions (including some C99 features).
1179 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1180 this will become the default. The name @samp{gnu9x} is deprecated.
1183 The 1998 ISO C++ standard plus amendments.
1186 The same as @option{-std=c++98} plus GNU extensions. This is the
1187 default for C++ code.
1190 Even when this option is not specified, you can still use some of the
1191 features of newer standards in so far as they do not conflict with
1192 previous C standards. For example, you may use @code{__restrict__} even
1193 when @option{-std=c99} is not specified.
1195 The @option{-std} options specifying some version of ISO C have the same
1196 effects as @option{-ansi}, except that features that were not in ISO C90
1197 but are in the specified version (for example, @samp{//} comments and
1198 the @code{inline} keyword in ISO C99) are not disabled.
1200 @xref{Standards,,Language Standards Supported by GCC}, for details of
1201 these standard versions.
1203 @item -aux-info @var{filename}
1205 Output to the given filename prototyped declarations for all functions
1206 declared and/or defined in a translation unit, including those in header
1207 files. This option is silently ignored in any language other than C@.
1209 Besides declarations, the file indicates, in comments, the origin of
1210 each declaration (source file and line), whether the declaration was
1211 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1212 @samp{O} for old, respectively, in the first character after the line
1213 number and the colon), and whether it came from a declaration or a
1214 definition (@samp{C} or @samp{F}, respectively, in the following
1215 character). In the case of function definitions, a K&R-style list of
1216 arguments followed by their declarations is also provided, inside
1217 comments, after the declaration.
1221 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1222 keyword, so that code can use these words as identifiers. You can use
1223 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1224 instead. @option{-ansi} implies @option{-fno-asm}.
1226 In C++, this switch only affects the @code{typeof} keyword, since
1227 @code{asm} and @code{inline} are standard keywords. You may want to
1228 use the @option{-fno-gnu-keywords} flag instead, which has the same
1229 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1230 switch only affects the @code{asm} and @code{typeof} keywords, since
1231 @code{inline} is a standard keyword in ISO C99.
1234 @itemx -fno-builtin-@var{function}
1235 @opindex fno-builtin
1236 @cindex built-in functions
1237 Don't recognize built-in functions that do not begin with
1238 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1239 functions provided by GCC}, for details of the functions affected,
1240 including those which are not built-in functions when @option{-ansi} or
1241 @option{-std} options for strict ISO C conformance are used because they
1242 do not have an ISO standard meaning.
1244 GCC normally generates special code to handle certain built-in functions
1245 more efficiently; for instance, calls to @code{alloca} may become single
1246 instructions that adjust the stack directly, and calls to @code{memcpy}
1247 may become inline copy loops. The resulting code is often both smaller
1248 and faster, but since the function calls no longer appear as such, you
1249 cannot set a breakpoint on those calls, nor can you change the behavior
1250 of the functions by linking with a different library. In addition,
1251 when a function is recognized as a built-in function, GCC may use
1252 information about that function to warn about problems with calls to
1253 that function, or to generate more efficient code, even if the
1254 resulting code still contains calls to that function. For example,
1255 warnings are given with @option{-Wformat} for bad calls to
1256 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1257 known not to modify global memory.
1259 With the @option{-fno-builtin-@var{function}} option
1260 only the built-in function @var{function} is
1261 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1262 function is named this is not built-in in this version of GCC, this
1263 option is ignored. There is no corresponding
1264 @option{-fbuiltin-@var{function}} option; if you wish to enable
1265 built-in functions selectively when using @option{-fno-builtin} or
1266 @option{-ffreestanding}, you may define macros such as:
1269 #define abs(n) __builtin_abs ((n))
1270 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1275 @cindex hosted environment
1277 Assert that compilation takes place in a hosted environment. This implies
1278 @option{-fbuiltin}. A hosted environment is one in which the
1279 entire standard library is available, and in which @code{main} has a return
1280 type of @code{int}. Examples are nearly everything except a kernel.
1281 This is equivalent to @option{-fno-freestanding}.
1283 @item -ffreestanding
1284 @opindex ffreestanding
1285 @cindex hosted environment
1287 Assert that compilation takes place in a freestanding environment. This
1288 implies @option{-fno-builtin}. A freestanding environment
1289 is one in which the standard library may not exist, and program startup may
1290 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1291 This is equivalent to @option{-fno-hosted}.
1293 @xref{Standards,,Language Standards Supported by GCC}, for details of
1294 freestanding and hosted environments.
1296 @item -fms-extensions
1297 @opindex fms-extensions
1298 Accept some non-standard constructs used in Microsoft header files.
1300 Some cases of unnamed fields in structures and unions are only
1301 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1302 fields within structs/unions}, for details.
1306 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1307 options for strict ISO C conformance) implies @option{-trigraphs}.
1309 @item -no-integrated-cpp
1310 @opindex no-integrated-cpp
1311 Performs a compilation in two passes: preprocessing and compiling. This
1312 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1313 @option{-B} option. The user supplied compilation step can then add in
1314 an additional preprocessing step after normal preprocessing but before
1315 compiling. The default is to use the integrated cpp (internal cpp)
1317 The semantics of this option will change if "cc1", "cc1plus", and
1318 "cc1obj" are merged.
1320 @cindex traditional C language
1321 @cindex C language, traditional
1323 @itemx -traditional-cpp
1324 @opindex traditional-cpp
1325 @opindex traditional
1326 Formerly, these options caused GCC to attempt to emulate a pre-standard
1327 C compiler. They are now only supported with the @option{-E} switch.
1328 The preprocessor continues to support a pre-standard mode. See the GNU
1329 CPP manual for details.
1331 @item -fcond-mismatch
1332 @opindex fcond-mismatch
1333 Allow conditional expressions with mismatched types in the second and
1334 third arguments. The value of such an expression is void. This option
1335 is not supported for C++.
1337 @item -funsigned-char
1338 @opindex funsigned-char
1339 Let the type @code{char} be unsigned, like @code{unsigned char}.
1341 Each kind of machine has a default for what @code{char} should
1342 be. It is either like @code{unsigned char} by default or like
1343 @code{signed char} by default.
1345 Ideally, a portable program should always use @code{signed char} or
1346 @code{unsigned char} when it depends on the signedness of an object.
1347 But many programs have been written to use plain @code{char} and
1348 expect it to be signed, or expect it to be unsigned, depending on the
1349 machines they were written for. This option, and its inverse, let you
1350 make such a program work with the opposite default.
1352 The type @code{char} is always a distinct type from each of
1353 @code{signed char} or @code{unsigned char}, even though its behavior
1354 is always just like one of those two.
1357 @opindex fsigned-char
1358 Let the type @code{char} be signed, like @code{signed char}.
1360 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1361 the negative form of @option{-funsigned-char}. Likewise, the option
1362 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1364 @item -fsigned-bitfields
1365 @itemx -funsigned-bitfields
1366 @itemx -fno-signed-bitfields
1367 @itemx -fno-unsigned-bitfields
1368 @opindex fsigned-bitfields
1369 @opindex funsigned-bitfields
1370 @opindex fno-signed-bitfields
1371 @opindex fno-unsigned-bitfields
1372 These options control whether a bit-field is signed or unsigned, when the
1373 declaration does not use either @code{signed} or @code{unsigned}. By
1374 default, such a bit-field is signed, because this is consistent: the
1375 basic integer types such as @code{int} are signed types.
1378 @node C++ Dialect Options
1379 @section Options Controlling C++ Dialect
1381 @cindex compiler options, C++
1382 @cindex C++ options, command line
1383 @cindex options, C++
1384 This section describes the command-line options that are only meaningful
1385 for C++ programs; but you can also use most of the GNU compiler options
1386 regardless of what language your program is in. For example, you
1387 might compile a file @code{firstClass.C} like this:
1390 g++ -g -frepo -O -c firstClass.C
1394 In this example, only @option{-frepo} is an option meant
1395 only for C++ programs; you can use the other options with any
1396 language supported by GCC@.
1398 Here is a list of options that are @emph{only} for compiling C++ programs:
1402 @item -fabi-version=@var{n}
1403 @opindex fabi-version
1404 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1405 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1406 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1407 the version that conforms most closely to the C++ ABI specification.
1408 Therefore, the ABI obtained using version 0 will change as ABI bugs
1411 The default is version 2.
1413 @item -fno-access-control
1414 @opindex fno-access-control
1415 Turn off all access checking. This switch is mainly useful for working
1416 around bugs in the access control code.
1420 Check that the pointer returned by @code{operator new} is non-null
1421 before attempting to modify the storage allocated. This check is
1422 normally unnecessary because the C++ standard specifies that
1423 @code{operator new} will only return @code{0} if it is declared
1424 @samp{throw()}, in which case the compiler will always check the
1425 return value even without this option. In all other cases, when
1426 @code{operator new} has a non-empty exception specification, memory
1427 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1428 @samp{new (nothrow)}.
1430 @item -fconserve-space
1431 @opindex fconserve-space
1432 Put uninitialized or runtime-initialized global variables into the
1433 common segment, as C does. This saves space in the executable at the
1434 cost of not diagnosing duplicate definitions. If you compile with this
1435 flag and your program mysteriously crashes after @code{main()} has
1436 completed, you may have an object that is being destroyed twice because
1437 two definitions were merged.
1439 This option is no longer useful on most targets, now that support has
1440 been added for putting variables into BSS without making them common.
1442 @item -ffriend-injection
1443 @opindex ffriend-injection
1444 Inject friend functions into the enclosing namespace, so that they are
1445 visible outside the scope of the class in which they are declared.
1446 Friend functions were documented to work this way in the old Annotated
1447 C++ Reference Manual, and versions of G++ before 4.1 always worked
1448 that way. However, in ISO C++ a friend function which is not declared
1449 in an enclosing scope can only be found using argument dependent
1450 lookup. This option causes friends to be injected as they were in
1453 This option is for compatibility, and may be removed in a future
1456 @item -fno-elide-constructors
1457 @opindex fno-elide-constructors
1458 The C++ standard allows an implementation to omit creating a temporary
1459 which is only used to initialize another object of the same type.
1460 Specifying this option disables that optimization, and forces G++ to
1461 call the copy constructor in all cases.
1463 @item -fno-enforce-eh-specs
1464 @opindex fno-enforce-eh-specs
1465 Don't generate code to check for violation of exception specifications
1466 at runtime. This option violates the C++ standard, but may be useful
1467 for reducing code size in production builds, much like defining
1468 @samp{NDEBUG}. This does not give user code permission to throw
1469 exceptions in violation of the exception specifications; the compiler
1470 will still optimize based on the specifications, so throwing an
1471 unexpected exception will result in undefined behavior.
1474 @itemx -fno-for-scope
1476 @opindex fno-for-scope
1477 If @option{-ffor-scope} is specified, the scope of variables declared in
1478 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1479 as specified by the C++ standard.
1480 If @option{-fno-for-scope} is specified, the scope of variables declared in
1481 a @i{for-init-statement} extends to the end of the enclosing scope,
1482 as was the case in old versions of G++, and other (traditional)
1483 implementations of C++.
1485 The default if neither flag is given to follow the standard,
1486 but to allow and give a warning for old-style code that would
1487 otherwise be invalid, or have different behavior.
1489 @item -fno-gnu-keywords
1490 @opindex fno-gnu-keywords
1491 Do not recognize @code{typeof} as a keyword, so that code can use this
1492 word as an identifier. You can use the keyword @code{__typeof__} instead.
1493 @option{-ansi} implies @option{-fno-gnu-keywords}.
1495 @item -fno-implicit-templates
1496 @opindex fno-implicit-templates
1497 Never emit code for non-inline templates which are instantiated
1498 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1499 @xref{Template Instantiation}, for more information.
1501 @item -fno-implicit-inline-templates
1502 @opindex fno-implicit-inline-templates
1503 Don't emit code for implicit instantiations of inline templates, either.
1504 The default is to handle inlines differently so that compiles with and
1505 without optimization will need the same set of explicit instantiations.
1507 @item -fno-implement-inlines
1508 @opindex fno-implement-inlines
1509 To save space, do not emit out-of-line copies of inline functions
1510 controlled by @samp{#pragma implementation}. This will cause linker
1511 errors if these functions are not inlined everywhere they are called.
1513 @item -fms-extensions
1514 @opindex fms-extensions
1515 Disable pedantic warnings about constructs used in MFC, such as implicit
1516 int and getting a pointer to member function via non-standard syntax.
1518 @item -fno-nonansi-builtins
1519 @opindex fno-nonansi-builtins
1520 Disable built-in declarations of functions that are not mandated by
1521 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1522 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1524 @item -fno-operator-names
1525 @opindex fno-operator-names
1526 Do not treat the operator name keywords @code{and}, @code{bitand},
1527 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1528 synonyms as keywords.
1530 @item -fno-optional-diags
1531 @opindex fno-optional-diags
1532 Disable diagnostics that the standard says a compiler does not need to
1533 issue. Currently, the only such diagnostic issued by G++ is the one for
1534 a name having multiple meanings within a class.
1537 @opindex fpermissive
1538 Downgrade some diagnostics about nonconformant code from errors to
1539 warnings. Thus, using @option{-fpermissive} will allow some
1540 nonconforming code to compile.
1544 Enable automatic template instantiation at link time. This option also
1545 implies @option{-fno-implicit-templates}. @xref{Template
1546 Instantiation}, for more information.
1550 Disable generation of information about every class with virtual
1551 functions for use by the C++ runtime type identification features
1552 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1553 of the language, you can save some space by using this flag. Note that
1554 exception handling uses the same information, but it will generate it as
1559 Emit statistics about front-end processing at the end of the compilation.
1560 This information is generally only useful to the G++ development team.
1562 @item -ftemplate-depth-@var{n}
1563 @opindex ftemplate-depth
1564 Set the maximum instantiation depth for template classes to @var{n}.
1565 A limit on the template instantiation depth is needed to detect
1566 endless recursions during template class instantiation. ANSI/ISO C++
1567 conforming programs must not rely on a maximum depth greater than 17.
1569 @item -fno-threadsafe-statics
1570 @opindex fno-threadsafe-statics
1571 Do not emit the extra code to use the routines specified in the C++
1572 ABI for thread-safe initialization of local statics. You can use this
1573 option to reduce code size slightly in code that doesn't need to be
1576 @item -fuse-cxa-atexit
1577 @opindex fuse-cxa-atexit
1578 Register destructors for objects with static storage duration with the
1579 @code{__cxa_atexit} function rather than the @code{atexit} function.
1580 This option is required for fully standards-compliant handling of static
1581 destructors, but will only work if your C library supports
1582 @code{__cxa_atexit}.
1584 @item -fvisibility-inlines-hidden
1585 @opindex fvisibility-inlines-hidden
1586 Causes all inlined methods to be marked with
1587 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1588 appear in the export table of a DSO and do not require a PLT indirection
1589 when used within the DSO@. Enabling this option can have a dramatic effect
1590 on load and link times of a DSO as it massively reduces the size of the
1591 dynamic export table when the library makes heavy use of templates. While
1592 it can cause bloating through duplication of code within each DSO where
1593 it is used, often the wastage is less than the considerable space occupied
1594 by a long symbol name in the export table which is typical when using
1595 templates and namespaces. For even more savings, combine with the
1596 @option{-fvisibility=hidden} switch.
1600 Do not use weak symbol support, even if it is provided by the linker.
1601 By default, G++ will use weak symbols if they are available. This
1602 option exists only for testing, and should not be used by end-users;
1603 it will result in inferior code and has no benefits. This option may
1604 be removed in a future release of G++.
1608 Do not search for header files in the standard directories specific to
1609 C++, but do still search the other standard directories. (This option
1610 is used when building the C++ library.)
1613 In addition, these optimization, warning, and code generation options
1614 have meanings only for C++ programs:
1617 @item -fno-default-inline
1618 @opindex fno-default-inline
1619 Do not assume @samp{inline} for functions defined inside a class scope.
1620 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1621 functions will have linkage like inline functions; they just won't be
1624 @item -Wabi @r{(C++ only)}
1626 Warn when G++ generates code that is probably not compatible with the
1627 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1628 all such cases, there are probably some cases that are not warned about,
1629 even though G++ is generating incompatible code. There may also be
1630 cases where warnings are emitted even though the code that is generated
1633 You should rewrite your code to avoid these warnings if you are
1634 concerned about the fact that code generated by G++ may not be binary
1635 compatible with code generated by other compilers.
1637 The known incompatibilities at this point include:
1642 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1643 pack data into the same byte as a base class. For example:
1646 struct A @{ virtual void f(); int f1 : 1; @};
1647 struct B : public A @{ int f2 : 1; @};
1651 In this case, G++ will place @code{B::f2} into the same byte
1652 as@code{A::f1}; other compilers will not. You can avoid this problem
1653 by explicitly padding @code{A} so that its size is a multiple of the
1654 byte size on your platform; that will cause G++ and other compilers to
1655 layout @code{B} identically.
1658 Incorrect handling of tail-padding for virtual bases. G++ does not use
1659 tail padding when laying out virtual bases. For example:
1662 struct A @{ virtual void f(); char c1; @};
1663 struct B @{ B(); char c2; @};
1664 struct C : public A, public virtual B @{@};
1668 In this case, G++ will not place @code{B} into the tail-padding for
1669 @code{A}; other compilers will. You can avoid this problem by
1670 explicitly padding @code{A} so that its size is a multiple of its
1671 alignment (ignoring virtual base classes); that will cause G++ and other
1672 compilers to layout @code{C} identically.
1675 Incorrect handling of bit-fields with declared widths greater than that
1676 of their underlying types, when the bit-fields appear in a union. For
1680 union U @{ int i : 4096; @};
1684 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1685 union too small by the number of bits in an @code{int}.
1688 Empty classes can be placed at incorrect offsets. For example:
1698 struct C : public B, public A @{@};
1702 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1703 it should be placed at offset zero. G++ mistakenly believes that the
1704 @code{A} data member of @code{B} is already at offset zero.
1707 Names of template functions whose types involve @code{typename} or
1708 template template parameters can be mangled incorrectly.
1711 template <typename Q>
1712 void f(typename Q::X) @{@}
1714 template <template <typename> class Q>
1715 void f(typename Q<int>::X) @{@}
1719 Instantiations of these templates may be mangled incorrectly.
1723 @item -Wctor-dtor-privacy @r{(C++ only)}
1724 @opindex Wctor-dtor-privacy
1725 Warn when a class seems unusable because all the constructors or
1726 destructors in that class are private, and it has neither friends nor
1727 public static member functions.
1729 @item -Wnon-virtual-dtor @r{(C++ only)}
1730 @opindex Wnon-virtual-dtor
1731 Warn when a class appears to be polymorphic, thereby requiring a virtual
1732 destructor, yet it declares a non-virtual one. This warning is also
1733 enabled if -Weffc++ is specified.
1735 @item -Wreorder @r{(C++ only)}
1737 @cindex reordering, warning
1738 @cindex warning for reordering of member initializers
1739 Warn when the order of member initializers given in the code does not
1740 match the order in which they must be executed. For instance:
1746 A(): j (0), i (1) @{ @}
1750 The compiler will rearrange the member initializers for @samp{i}
1751 and @samp{j} to match the declaration order of the members, emitting
1752 a warning to that effect. This warning is enabled by @option{-Wall}.
1755 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1758 @item -Weffc++ @r{(C++ only)}
1760 Warn about violations of the following style guidelines from Scott Meyers'
1761 @cite{Effective C++} book:
1765 Item 11: Define a copy constructor and an assignment operator for classes
1766 with dynamically allocated memory.
1769 Item 12: Prefer initialization to assignment in constructors.
1772 Item 14: Make destructors virtual in base classes.
1775 Item 15: Have @code{operator=} return a reference to @code{*this}.
1778 Item 23: Don't try to return a reference when you must return an object.
1782 Also warn about violations of the following style guidelines from
1783 Scott Meyers' @cite{More Effective C++} book:
1787 Item 6: Distinguish between prefix and postfix forms of increment and
1788 decrement operators.
1791 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1795 When selecting this option, be aware that the standard library
1796 headers do not obey all of these guidelines; use @samp{grep -v}
1797 to filter out those warnings.
1799 @item -Wno-deprecated @r{(C++ only)}
1800 @opindex Wno-deprecated
1801 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1803 @item -Wstrict-null-sentinel @r{(C++ only)}
1804 @opindex Wstrict-null-sentinel
1805 Warn also about the use of an uncasted @code{NULL} as sentinel. When
1806 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
1807 to @code{__null}. Although it is a null pointer constant not a null pointer,
1808 it is guaranteed to of the same size as a pointer. But this use is
1809 not portable across different compilers.
1811 @item -Wno-non-template-friend @r{(C++ only)}
1812 @opindex Wno-non-template-friend
1813 Disable warnings when non-templatized friend functions are declared
1814 within a template. Since the advent of explicit template specification
1815 support in G++, if the name of the friend is an unqualified-id (i.e.,
1816 @samp{friend foo(int)}), the C++ language specification demands that the
1817 friend declare or define an ordinary, nontemplate function. (Section
1818 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1819 could be interpreted as a particular specialization of a templatized
1820 function. Because this non-conforming behavior is no longer the default
1821 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1822 check existing code for potential trouble spots and is on by default.
1823 This new compiler behavior can be turned off with
1824 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1825 but disables the helpful warning.
1827 @item -Wold-style-cast @r{(C++ only)}
1828 @opindex Wold-style-cast
1829 Warn if an old-style (C-style) cast to a non-void type is used within
1830 a C++ program. The new-style casts (@samp{dynamic_cast},
1831 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
1832 less vulnerable to unintended effects and much easier to search for.
1834 @item -Woverloaded-virtual @r{(C++ only)}
1835 @opindex Woverloaded-virtual
1836 @cindex overloaded virtual fn, warning
1837 @cindex warning for overloaded virtual fn
1838 Warn when a function declaration hides virtual functions from a
1839 base class. For example, in:
1846 struct B: public A @{
1851 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1859 will fail to compile.
1861 @item -Wno-pmf-conversions @r{(C++ only)}
1862 @opindex Wno-pmf-conversions
1863 Disable the diagnostic for converting a bound pointer to member function
1866 @item -Wsign-promo @r{(C++ only)}
1867 @opindex Wsign-promo
1868 Warn when overload resolution chooses a promotion from unsigned or
1869 enumerated type to a signed type, over a conversion to an unsigned type of
1870 the same size. Previous versions of G++ would try to preserve
1871 unsignedness, but the standard mandates the current behavior.
1876 A& operator = (int);
1886 In this example, G++ will synthesize a default @samp{A& operator =
1887 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1890 @node Objective-C and Objective-C++ Dialect Options
1891 @section Options Controlling Objective-C and Objective-C++ Dialects
1893 @cindex compiler options, Objective-C and Objective-C++
1894 @cindex Objective-C and Objective-C++ options, command line
1895 @cindex options, Objective-C and Objective-C++
1896 (NOTE: This manual does not describe the Objective-C and Objective-C++
1897 languages themselves. See @xref{Standards,,Language Standards
1898 Supported by GCC}, for references.)
1900 This section describes the command-line options that are only meaningful
1901 for Objective-C and Objective-C++ programs, but you can also use most of
1902 the language-independent GNU compiler options.
1903 For example, you might compile a file @code{some_class.m} like this:
1906 gcc -g -fgnu-runtime -O -c some_class.m
1910 In this example, @option{-fgnu-runtime} is an option meant only for
1911 Objective-C and Objective-C++ programs; you can use the other options with
1912 any language supported by GCC@.
1914 Note that since Objective-C is an extension of the C language, Objective-C
1915 compilations may also use options specific to the C front-end (e.g.,
1916 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
1917 C++-specific options (e.g., @option{-Wabi}).
1919 Here is a list of options that are @emph{only} for compiling Objective-C
1920 and Objective-C++ programs:
1923 @item -fconstant-string-class=@var{class-name}
1924 @opindex fconstant-string-class
1925 Use @var{class-name} as the name of the class to instantiate for each
1926 literal string specified with the syntax @code{@@"@dots{}"}. The default
1927 class name is @code{NXConstantString} if the GNU runtime is being used, and
1928 @code{NSConstantString} if the NeXT runtime is being used (see below). The
1929 @option{-fconstant-cfstrings} option, if also present, will override the
1930 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
1931 to be laid out as constant CoreFoundation strings.
1934 @opindex fgnu-runtime
1935 Generate object code compatible with the standard GNU Objective-C
1936 runtime. This is the default for most types of systems.
1938 @item -fnext-runtime
1939 @opindex fnext-runtime
1940 Generate output compatible with the NeXT runtime. This is the default
1941 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1942 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1945 @item -fno-nil-receivers
1946 @opindex fno-nil-receivers
1947 Assume that all Objective-C message dispatches (e.g.,
1948 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
1949 is not @code{nil}. This allows for more efficient entry points in the runtime
1950 to be used. Currently, this option is only available in conjunction with
1951 the NeXT runtime on Mac OS X 10.3 and later.
1953 @item -fobjc-call-cxx-cdtors
1954 @opindex fobjc-call-cxx-cdtors
1955 For each Objective-C class, check if any of its instance variables is a
1956 C++ object with a non-trivial default constructor. If so, synthesize a
1957 special @code{- (id) .cxx_construct} instance method that will run
1958 non-trivial default constructors on any such instance variables, in order,
1959 and then return @code{self}. Similarly, check if any instance variable
1960 is a C++ object with a non-trivial destructor, and if so, synthesize a
1961 special @code{- (void) .cxx_destruct} method that will run
1962 all such default destructors, in reverse order.
1964 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
1965 thusly generated will only operate on instance variables declared in the
1966 current Objective-C class, and not those inherited from superclasses. It
1967 is the responsibility of the Objective-C runtime to invoke all such methods
1968 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
1969 will be invoked by the runtime immediately after a new object
1970 instance is allocated; the @code{- (void) .cxx_destruct} methods will
1971 be invoked immediately before the runtime deallocates an object instance.
1973 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
1974 support for invoking the @code{- (id) .cxx_construct} and
1975 @code{- (void) .cxx_destruct} methods.
1977 @item -fobjc-direct-dispatch
1978 @opindex fobjc-direct-dispatch
1979 Allow fast jumps to the message dispatcher. On Darwin this is
1980 accomplished via the comm page.
1982 @item -fobjc-exceptions
1983 @opindex fobjc-exceptions
1984 Enable syntactic support for structured exception handling in Objective-C,
1985 similar to what is offered by C++ and Java. This option is
1986 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
1995 @@catch (AnObjCClass *exc) @{
2002 @@catch (AnotherClass *exc) @{
2005 @@catch (id allOthers) @{
2015 The @code{@@throw} statement may appear anywhere in an Objective-C or
2016 Objective-C++ program; when used inside of a @code{@@catch} block, the
2017 @code{@@throw} may appear without an argument (as shown above), in which case
2018 the object caught by the @code{@@catch} will be rethrown.
2020 Note that only (pointers to) Objective-C objects may be thrown and
2021 caught using this scheme. When an object is thrown, it will be caught
2022 by the nearest @code{@@catch} clause capable of handling objects of that type,
2023 analogously to how @code{catch} blocks work in C++ and Java. A
2024 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2025 any and all Objective-C exceptions not caught by previous @code{@@catch}
2028 The @code{@@finally} clause, if present, will be executed upon exit from the
2029 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2030 regardless of whether any exceptions are thrown, caught or rethrown
2031 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2032 of the @code{finally} clause in Java.
2034 There are several caveats to using the new exception mechanism:
2038 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2039 idioms provided by the @code{NSException} class, the new
2040 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2041 systems, due to additional functionality needed in the (NeXT) Objective-C
2045 As mentioned above, the new exceptions do not support handling
2046 types other than Objective-C objects. Furthermore, when used from
2047 Objective-C++, the Objective-C exception model does not interoperate with C++
2048 exceptions at this time. This means you cannot @code{@@throw} an exception
2049 from Objective-C and @code{catch} it in C++, or vice versa
2050 (i.e., @code{throw @dots{} @@catch}).
2053 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2054 blocks for thread-safe execution:
2057 @@synchronized (ObjCClass *guard) @{
2062 Upon entering the @code{@@synchronized} block, a thread of execution shall
2063 first check whether a lock has been placed on the corresponding @code{guard}
2064 object by another thread. If it has, the current thread shall wait until
2065 the other thread relinquishes its lock. Once @code{guard} becomes available,
2066 the current thread will place its own lock on it, execute the code contained in
2067 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2068 making @code{guard} available to other threads).
2070 Unlike Java, Objective-C does not allow for entire methods to be marked
2071 @code{@@synchronized}. Note that throwing exceptions out of
2072 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2073 to be unlocked properly.
2077 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2079 @item -freplace-objc-classes
2080 @opindex freplace-objc-classes
2081 Emit a special marker instructing @command{ld(1)} not to statically link in
2082 the resulting object file, and allow @command{dyld(1)} to load it in at
2083 run time instead. This is used in conjunction with the Fix-and-Continue
2084 debugging mode, where the object file in question may be recompiled and
2085 dynamically reloaded in the course of program execution, without the need
2086 to restart the program itself. Currently, Fix-and-Continue functionality
2087 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2092 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2093 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2094 compile time) with static class references that get initialized at load time,
2095 which improves run-time performance. Specifying the @option{-fzero-link} flag
2096 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2097 to be retained. This is useful in Zero-Link debugging mode, since it allows
2098 for individual class implementations to be modified during program execution.
2102 Dump interface declarations for all classes seen in the source file to a
2103 file named @file{@var{sourcename}.decl}.
2105 @item -Wassign-intercept
2106 @opindex Wassign-intercept
2107 Warn whenever an Objective-C assignment is being intercepted by the
2111 @opindex Wno-protocol
2112 If a class is declared to implement a protocol, a warning is issued for
2113 every method in the protocol that is not implemented by the class. The
2114 default behavior is to issue a warning for every method not explicitly
2115 implemented in the class, even if a method implementation is inherited
2116 from the superclass. If you use the @option{-Wno-protocol} option, then
2117 methods inherited from the superclass are considered to be implemented,
2118 and no warning is issued for them.
2122 Warn if multiple methods of different types for the same selector are
2123 found during compilation. The check is performed on the list of methods
2124 in the final stage of compilation. Additionally, a check is performed
2125 for each selector appearing in a @code{@@selector(@dots{})}
2126 expression, and a corresponding method for that selector has been found
2127 during compilation. Because these checks scan the method table only at
2128 the end of compilation, these warnings are not produced if the final
2129 stage of compilation is not reached, for example because an error is
2130 found during compilation, or because the @option{-fsyntax-only} option is
2133 @item -Wstrict-selector-match
2134 @opindex Wstrict-selector-match
2135 Warn if multiple methods with differing argument and/or return types are
2136 found for a given selector when attempting to send a message using this
2137 selector to a receiver of type @code{id} or @code{Class}. When this flag
2138 is off (which is the default behavior), the compiler will omit such warnings
2139 if any differences found are confined to types which share the same size
2142 @item -Wundeclared-selector
2143 @opindex Wundeclared-selector
2144 Warn if a @code{@@selector(@dots{})} expression referring to an
2145 undeclared selector is found. A selector is considered undeclared if no
2146 method with that name has been declared before the
2147 @code{@@selector(@dots{})} expression, either explicitly in an
2148 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2149 an @code{@@implementation} section. This option always performs its
2150 checks as soon as a @code{@@selector(@dots{})} expression is found,
2151 while @option{-Wselector} only performs its checks in the final stage of
2152 compilation. This also enforces the coding style convention
2153 that methods and selectors must be declared before being used.
2155 @item -print-objc-runtime-info
2156 @opindex print-objc-runtime-info
2157 Generate C header describing the largest structure that is passed by
2162 @node Language Independent Options
2163 @section Options to Control Diagnostic Messages Formatting
2164 @cindex options to control diagnostics formatting
2165 @cindex diagnostic messages
2166 @cindex message formatting
2168 Traditionally, diagnostic messages have been formatted irrespective of
2169 the output device's aspect (e.g.@: its width, @dots{}). The options described
2170 below can be used to control the diagnostic messages formatting
2171 algorithm, e.g.@: how many characters per line, how often source location
2172 information should be reported. Right now, only the C++ front end can
2173 honor these options. However it is expected, in the near future, that
2174 the remaining front ends would be able to digest them correctly.
2177 @item -fmessage-length=@var{n}
2178 @opindex fmessage-length
2179 Try to format error messages so that they fit on lines of about @var{n}
2180 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2181 the front ends supported by GCC@. If @var{n} is zero, then no
2182 line-wrapping will be done; each error message will appear on a single
2185 @opindex fdiagnostics-show-location
2186 @item -fdiagnostics-show-location=once
2187 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2188 reporter to emit @emph{once} source location information; that is, in
2189 case the message is too long to fit on a single physical line and has to
2190 be wrapped, the source location won't be emitted (as prefix) again,
2191 over and over, in subsequent continuation lines. This is the default
2194 @item -fdiagnostics-show-location=every-line
2195 Only meaningful in line-wrapping mode. Instructs the diagnostic
2196 messages reporter to emit the same source location information (as
2197 prefix) for physical lines that result from the process of breaking
2198 a message which is too long to fit on a single line.
2200 @item -fdiagnostics-show-options
2201 @opindex fdiagnostics-show-options
2202 This option instructs the diagnostic machinery to add text to each
2203 diagnostic emitted, which indicates which command line option directly
2204 controls that diagnostic, when such an option is known to the
2205 diagnostic machinery.
2209 @node Warning Options
2210 @section Options to Request or Suppress Warnings
2211 @cindex options to control warnings
2212 @cindex warning messages
2213 @cindex messages, warning
2214 @cindex suppressing warnings
2216 Warnings are diagnostic messages that report constructions which
2217 are not inherently erroneous but which are risky or suggest there
2218 may have been an error.
2220 You can request many specific warnings with options beginning @samp{-W},
2221 for example @option{-Wimplicit} to request warnings on implicit
2222 declarations. Each of these specific warning options also has a
2223 negative form beginning @samp{-Wno-} to turn off warnings;
2224 for example, @option{-Wno-implicit}. This manual lists only one of the
2225 two forms, whichever is not the default.
2227 The following options control the amount and kinds of warnings produced
2228 by GCC; for further, language-specific options also refer to
2229 @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect
2233 @cindex syntax checking
2235 @opindex fsyntax-only
2236 Check the code for syntax errors, but don't do anything beyond that.
2240 Issue all the warnings demanded by strict ISO C and ISO C++;
2241 reject all programs that use forbidden extensions, and some other
2242 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2243 version of the ISO C standard specified by any @option{-std} option used.
2245 Valid ISO C and ISO C++ programs should compile properly with or without
2246 this option (though a rare few will require @option{-ansi} or a
2247 @option{-std} option specifying the required version of ISO C)@. However,
2248 without this option, certain GNU extensions and traditional C and C++
2249 features are supported as well. With this option, they are rejected.
2251 @option{-pedantic} does not cause warning messages for use of the
2252 alternate keywords whose names begin and end with @samp{__}. Pedantic
2253 warnings are also disabled in the expression that follows
2254 @code{__extension__}. However, only system header files should use
2255 these escape routes; application programs should avoid them.
2256 @xref{Alternate Keywords}.
2258 Some users try to use @option{-pedantic} to check programs for strict ISO
2259 C conformance. They soon find that it does not do quite what they want:
2260 it finds some non-ISO practices, but not all---only those for which
2261 ISO C @emph{requires} a diagnostic, and some others for which
2262 diagnostics have been added.
2264 A feature to report any failure to conform to ISO C might be useful in
2265 some instances, but would require considerable additional work and would
2266 be quite different from @option{-pedantic}. We don't have plans to
2267 support such a feature in the near future.
2269 Where the standard specified with @option{-std} represents a GNU
2270 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2271 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2272 extended dialect is based. Warnings from @option{-pedantic} are given
2273 where they are required by the base standard. (It would not make sense
2274 for such warnings to be given only for features not in the specified GNU
2275 C dialect, since by definition the GNU dialects of C include all
2276 features the compiler supports with the given option, and there would be
2277 nothing to warn about.)
2279 @item -pedantic-errors
2280 @opindex pedantic-errors
2281 Like @option{-pedantic}, except that errors are produced rather than
2286 Inhibit all warning messages.
2290 Inhibit warning messages about the use of @samp{#import}.
2292 @item -Wchar-subscripts
2293 @opindex Wchar-subscripts
2294 Warn if an array subscript has type @code{char}. This is a common cause
2295 of error, as programmers often forget that this type is signed on some
2297 This warning is enabled by @option{-Wall}.
2301 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2302 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2303 This warning is enabled by @option{-Wall}.
2305 @item -Wfatal-errors
2306 @opindex Wfatal-errors
2307 This option causes the compiler to abort compilation on the first error
2308 occurred rather than trying to keep going and printing further error
2313 @opindex ffreestanding
2314 @opindex fno-builtin
2315 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2316 the arguments supplied have types appropriate to the format string
2317 specified, and that the conversions specified in the format string make
2318 sense. This includes standard functions, and others specified by format
2319 attributes (@pxref{Function Attributes}), in the @code{printf},
2320 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2321 not in the C standard) families (or other target-specific families).
2322 Which functions are checked without format attributes having been
2323 specified depends on the standard version selected, and such checks of
2324 functions without the attribute specified are disabled by
2325 @option{-ffreestanding} or @option{-fno-builtin}.
2327 The formats are checked against the format features supported by GNU
2328 libc version 2.2. These include all ISO C90 and C99 features, as well
2329 as features from the Single Unix Specification and some BSD and GNU
2330 extensions. Other library implementations may not support all these
2331 features; GCC does not support warning about features that go beyond a
2332 particular library's limitations. However, if @option{-pedantic} is used
2333 with @option{-Wformat}, warnings will be given about format features not
2334 in the selected standard version (but not for @code{strfmon} formats,
2335 since those are not in any version of the C standard). @xref{C Dialect
2336 Options,,Options Controlling C Dialect}.
2338 Since @option{-Wformat} also checks for null format arguments for
2339 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2341 @option{-Wformat} is included in @option{-Wall}. For more control over some
2342 aspects of format checking, the options @option{-Wformat-y2k},
2343 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2344 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2345 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2348 @opindex Wformat-y2k
2349 If @option{-Wformat} is specified, also warn about @code{strftime}
2350 formats which may yield only a two-digit year.
2352 @item -Wno-format-extra-args
2353 @opindex Wno-format-extra-args
2354 If @option{-Wformat} is specified, do not warn about excess arguments to a
2355 @code{printf} or @code{scanf} format function. The C standard specifies
2356 that such arguments are ignored.
2358 Where the unused arguments lie between used arguments that are
2359 specified with @samp{$} operand number specifications, normally
2360 warnings are still given, since the implementation could not know what
2361 type to pass to @code{va_arg} to skip the unused arguments. However,
2362 in the case of @code{scanf} formats, this option will suppress the
2363 warning if the unused arguments are all pointers, since the Single
2364 Unix Specification says that such unused arguments are allowed.
2366 @item -Wno-format-zero-length
2367 @opindex Wno-format-zero-length
2368 If @option{-Wformat} is specified, do not warn about zero-length formats.
2369 The C standard specifies that zero-length formats are allowed.
2371 @item -Wformat-nonliteral
2372 @opindex Wformat-nonliteral
2373 If @option{-Wformat} is specified, also warn if the format string is not a
2374 string literal and so cannot be checked, unless the format function
2375 takes its format arguments as a @code{va_list}.
2377 @item -Wformat-security
2378 @opindex Wformat-security
2379 If @option{-Wformat} is specified, also warn about uses of format
2380 functions that represent possible security problems. At present, this
2381 warns about calls to @code{printf} and @code{scanf} functions where the
2382 format string is not a string literal and there are no format arguments,
2383 as in @code{printf (foo);}. This may be a security hole if the format
2384 string came from untrusted input and contains @samp{%n}. (This is
2385 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2386 in future warnings may be added to @option{-Wformat-security} that are not
2387 included in @option{-Wformat-nonliteral}.)
2391 Enable @option{-Wformat} plus format checks not included in
2392 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2393 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2397 Warn about passing a null pointer for arguments marked as
2398 requiring a non-null value by the @code{nonnull} function attribute.
2400 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2401 can be disabled with the @option{-Wno-nonnull} option.
2403 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2405 Warn about uninitialized variables which are initialized with themselves.
2406 Note this option can only be used with the @option{-Wuninitialized} option,
2407 which in turn only works with @option{-O1} and above.
2409 For example, GCC will warn about @code{i} being uninitialized in the
2410 following snippet only when @option{-Winit-self} has been specified:
2421 @item -Wimplicit-int
2422 @opindex Wimplicit-int
2423 Warn when a declaration does not specify a type.
2424 This warning is enabled by @option{-Wall}.
2426 @item -Wimplicit-function-declaration
2427 @itemx -Werror-implicit-function-declaration
2428 @opindex Wimplicit-function-declaration
2429 @opindex Werror-implicit-function-declaration
2430 Give a warning (or error) whenever a function is used before being
2431 declared. The form @option{-Wno-error-implicit-function-declaration}
2433 This warning is enabled by @option{-Wall} (as a warning, not an error).
2437 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2438 This warning is enabled by @option{-Wall}.
2442 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2443 function with external linkage, returning int, taking either zero
2444 arguments, two, or three arguments of appropriate types.
2445 This warning is enabled by @option{-Wall}.
2447 @item -Wmissing-braces
2448 @opindex Wmissing-braces
2449 Warn if an aggregate or union initializer is not fully bracketed. In
2450 the following example, the initializer for @samp{a} is not fully
2451 bracketed, but that for @samp{b} is fully bracketed.
2454 int a[2][2] = @{ 0, 1, 2, 3 @};
2455 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2458 This warning is enabled by @option{-Wall}.
2460 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2461 @opindex Wmissing-include-dirs
2462 Warn if a user-supplied include directory does not exist.
2465 @opindex Wparentheses
2466 Warn if parentheses are omitted in certain contexts, such
2467 as when there is an assignment in a context where a truth value
2468 is expected, or when operators are nested whose precedence people
2469 often get confused about. Only the warning for an assignment used as
2470 a truth value is supported when compiling C++; the other warnings are
2471 only supported when compiling C@.
2473 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2474 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2475 interpretation from that of ordinary mathematical notation.
2477 Also warn about constructions where there may be confusion to which
2478 @code{if} statement an @code{else} branch belongs. Here is an example of
2493 In C, every @code{else} branch belongs to the innermost possible @code{if}
2494 statement, which in this example is @code{if (b)}. This is often not
2495 what the programmer expected, as illustrated in the above example by
2496 indentation the programmer chose. When there is the potential for this
2497 confusion, GCC will issue a warning when this flag is specified.
2498 To eliminate the warning, add explicit braces around the innermost
2499 @code{if} statement so there is no way the @code{else} could belong to
2500 the enclosing @code{if}. The resulting code would look like this:
2516 This warning is enabled by @option{-Wall}.
2518 @item -Wsequence-point
2519 @opindex Wsequence-point
2520 Warn about code that may have undefined semantics because of violations
2521 of sequence point rules in the C and C++ standards.
2523 The C and C++ standards defines the order in which expressions in a C/C++
2524 program are evaluated in terms of @dfn{sequence points}, which represent
2525 a partial ordering between the execution of parts of the program: those
2526 executed before the sequence point, and those executed after it. These
2527 occur after the evaluation of a full expression (one which is not part
2528 of a larger expression), after the evaluation of the first operand of a
2529 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2530 function is called (but after the evaluation of its arguments and the
2531 expression denoting the called function), and in certain other places.
2532 Other than as expressed by the sequence point rules, the order of
2533 evaluation of subexpressions of an expression is not specified. All
2534 these rules describe only a partial order rather than a total order,
2535 since, for example, if two functions are called within one expression
2536 with no sequence point between them, the order in which the functions
2537 are called is not specified. However, the standards committee have
2538 ruled that function calls do not overlap.
2540 It is not specified when between sequence points modifications to the
2541 values of objects take effect. Programs whose behavior depends on this
2542 have undefined behavior; the C and C++ standards specify that ``Between
2543 the previous and next sequence point an object shall have its stored
2544 value modified at most once by the evaluation of an expression.
2545 Furthermore, the prior value shall be read only to determine the value
2546 to be stored.''. If a program breaks these rules, the results on any
2547 particular implementation are entirely unpredictable.
2549 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2550 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2551 diagnosed by this option, and it may give an occasional false positive
2552 result, but in general it has been found fairly effective at detecting
2553 this sort of problem in programs.
2555 The standard is worded confusingly, therefore there is some debate
2556 over the precise meaning of the sequence point rules in subtle cases.
2557 Links to discussions of the problem, including proposed formal
2558 definitions, may be found on the GCC readings page, at
2559 @w{@uref{http://gcc.gnu.org/readings.html}}.
2561 This warning is enabled by @option{-Wall} for C and C++.
2564 @opindex Wreturn-type
2565 Warn whenever a function is defined with a return-type that defaults to
2566 @code{int}. Also warn about any @code{return} statement with no
2567 return-value in a function whose return-type is not @code{void}.
2569 For C, also warn if the return type of a function has a type qualifier
2570 such as @code{const}. Such a type qualifier has no effect, since the
2571 value returned by a function is not an lvalue. ISO C prohibits
2572 qualified @code{void} return types on function definitions, so such
2573 return types always receive a warning even without this option.
2575 For C++, a function without return type always produces a diagnostic
2576 message, even when @option{-Wno-return-type} is specified. The only
2577 exceptions are @samp{main} and functions defined in system headers.
2579 This warning is enabled by @option{-Wall}.
2583 Warn whenever a @code{switch} statement has an index of enumerated type
2584 and lacks a @code{case} for one or more of the named codes of that
2585 enumeration. (The presence of a @code{default} label prevents this
2586 warning.) @code{case} labels outside the enumeration range also
2587 provoke warnings when this option is used.
2588 This warning is enabled by @option{-Wall}.
2590 @item -Wswitch-default
2591 @opindex Wswitch-switch
2592 Warn whenever a @code{switch} statement does not have a @code{default}
2596 @opindex Wswitch-enum
2597 Warn whenever a @code{switch} statement has an index of enumerated type
2598 and lacks a @code{case} for one or more of the named codes of that
2599 enumeration. @code{case} labels outside the enumeration range also
2600 provoke warnings when this option is used.
2604 Warn if any trigraphs are encountered that might change the meaning of
2605 the program (trigraphs within comments are not warned about).
2606 This warning is enabled by @option{-Wall}.
2608 @item -Wunused-function
2609 @opindex Wunused-function
2610 Warn whenever a static function is declared but not defined or a
2611 non-inline static function is unused.
2612 This warning is enabled by @option{-Wall}.
2614 @item -Wunused-label
2615 @opindex Wunused-label
2616 Warn whenever a label is declared but not used.
2617 This warning is enabled by @option{-Wall}.
2619 To suppress this warning use the @samp{unused} attribute
2620 (@pxref{Variable Attributes}).
2622 @item -Wunused-parameter
2623 @opindex Wunused-parameter
2624 Warn whenever a function parameter is unused aside from its declaration.
2626 To suppress this warning use the @samp{unused} attribute
2627 (@pxref{Variable Attributes}).
2629 @item -Wunused-variable
2630 @opindex Wunused-variable
2631 Warn whenever a local variable or non-constant static variable is unused
2632 aside from its declaration
2633 This warning is enabled by @option{-Wall}.
2635 To suppress this warning use the @samp{unused} attribute
2636 (@pxref{Variable Attributes}).
2638 @item -Wunused-value
2639 @opindex Wunused-value
2640 Warn whenever a statement computes a result that is explicitly not used.
2641 This warning is enabled by @option{-Wall}.
2643 To suppress this warning cast the expression to @samp{void}.
2647 All the above @option{-Wunused} options combined.
2649 In order to get a warning about an unused function parameter, you must
2650 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2651 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2653 @item -Wuninitialized
2654 @opindex Wuninitialized
2655 Warn if an automatic variable is used without first being initialized or
2656 if a variable may be clobbered by a @code{setjmp} call.
2658 These warnings are possible only in optimizing compilation,
2659 because they require data flow information that is computed only
2660 when optimizing. If you don't specify @option{-O}, you simply won't
2663 If you want to warn about code which uses the uninitialized value of the
2664 variable in its own initializer, use the @option{-Winit-self} option.
2666 These warnings occur for individual uninitialized or clobbered
2667 elements of structure, union or array variables as well as for
2668 variables which are uninitialized or clobbered as a whole. They do
2669 not occur for variables or elements declared @code{volatile}. Because
2670 these warnings depend on optimization, the exact variables or elements
2671 for which there are warnings will depend on the precise optimization
2672 options and version of GCC used.
2674 Note that there may be no warning about a variable that is used only
2675 to compute a value that itself is never used, because such
2676 computations may be deleted by data flow analysis before the warnings
2679 These warnings are made optional because GCC is not smart
2680 enough to see all the reasons why the code might be correct
2681 despite appearing to have an error. Here is one example of how
2702 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2703 always initialized, but GCC doesn't know this. Here is
2704 another common case:
2709 if (change_y) save_y = y, y = new_y;
2711 if (change_y) y = save_y;
2716 This has no bug because @code{save_y} is used only if it is set.
2718 @cindex @code{longjmp} warnings
2719 This option also warns when a non-volatile automatic variable might be
2720 changed by a call to @code{longjmp}. These warnings as well are possible
2721 only in optimizing compilation.
2723 The compiler sees only the calls to @code{setjmp}. It cannot know
2724 where @code{longjmp} will be called; in fact, a signal handler could
2725 call it at any point in the code. As a result, you may get a warning
2726 even when there is in fact no problem because @code{longjmp} cannot
2727 in fact be called at the place which would cause a problem.
2729 Some spurious warnings can be avoided if you declare all the functions
2730 you use that never return as @code{noreturn}. @xref{Function
2733 This warning is enabled by @option{-Wall}.
2735 @item -Wunknown-pragmas
2736 @opindex Wunknown-pragmas
2737 @cindex warning for unknown pragmas
2738 @cindex unknown pragmas, warning
2739 @cindex pragmas, warning of unknown
2740 Warn when a #pragma directive is encountered which is not understood by
2741 GCC@. If this command line option is used, warnings will even be issued
2742 for unknown pragmas in system header files. This is not the case if
2743 the warnings were only enabled by the @option{-Wall} command line option.
2746 @opindex Wno-pragmas
2748 Do not warn about misuses of pragmas, such as incorrect parameters,
2749 invalid syntax, or conflicts between pragmas. See also
2750 @samp{-Wunknown-pragmas}.
2752 @item -Wstrict-aliasing
2753 @opindex Wstrict-aliasing
2754 This option is only active when @option{-fstrict-aliasing} is active.
2755 It warns about code which might break the strict aliasing rules that the
2756 compiler is using for optimization. The warning does not catch all
2757 cases, but does attempt to catch the more common pitfalls. It is
2758 included in @option{-Wall}.
2760 @item -Wstrict-aliasing=2
2761 @opindex Wstrict-aliasing=2
2762 This option is only active when @option{-fstrict-aliasing} is active.
2763 It warns about code which might break the strict aliasing rules that the
2764 compiler is using for optimization. This warning catches more cases than
2765 @option{-Wstrict-aliasing}, but it will also give a warning for some ambiguous
2766 cases that are safe.
2770 All of the above @samp{-W} options combined. This enables all the
2771 warnings about constructions that some users consider questionable, and
2772 that are easy to avoid (or modify to prevent the warning), even in
2773 conjunction with macros. This also enables some language-specific
2774 warnings described in @ref{C++ Dialect Options} and
2775 @ref{Objective-C and Objective-C++ Dialect Options}.
2778 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2779 Some of them warn about constructions that users generally do not
2780 consider questionable, but which occasionally you might wish to check
2781 for; others warn about constructions that are necessary or hard to avoid
2782 in some cases, and there is no simple way to modify the code to suppress
2789 (This option used to be called @option{-W}. The older name is still
2790 supported, but the newer name is more descriptive.) Print extra warning
2791 messages for these events:
2795 A function can return either with or without a value. (Falling
2796 off the end of the function body is considered returning without
2797 a value.) For example, this function would evoke such a
2811 An expression-statement or the left-hand side of a comma expression
2812 contains no side effects.
2813 To suppress the warning, cast the unused expression to void.
2814 For example, an expression such as @samp{x[i,j]} will cause a warning,
2815 but @samp{x[(void)i,j]} will not.
2818 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2821 Storage-class specifiers like @code{static} are not the first things in
2822 a declaration. According to the C Standard, this usage is obsolescent.
2825 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2829 A comparison between signed and unsigned values could produce an
2830 incorrect result when the signed value is converted to unsigned.
2831 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2834 An aggregate has an initializer which does not initialize all members.
2835 This warning can be independently controlled by
2836 @option{-Wmissing-field-initializers}.
2839 A function parameter is declared without a type specifier in K&R-style
2847 An empty body occurs in an @samp{if} or @samp{else} statement.
2850 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2851 @samp{>}, or @samp{>=}.
2854 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2857 Any of several floating-point events that often indicate errors, such as
2858 overflow, underflow, loss of precision, etc.
2860 @item @r{(C++ only)}
2861 An enumerator and a non-enumerator both appear in a conditional expression.
2863 @item @r{(C++ only)}
2864 A non-static reference or non-static @samp{const} member appears in a
2865 class without constructors.
2867 @item @r{(C++ only)}
2868 Ambiguous virtual bases.
2870 @item @r{(C++ only)}
2871 Subscripting an array which has been declared @samp{register}.
2873 @item @r{(C++ only)}
2874 Taking the address of a variable which has been declared @samp{register}.
2876 @item @r{(C++ only)}
2877 A base class is not initialized in a derived class' copy constructor.
2880 @item -Wno-div-by-zero
2881 @opindex Wno-div-by-zero
2882 @opindex Wdiv-by-zero
2883 Do not warn about compile-time integer division by zero. Floating point
2884 division by zero is not warned about, as it can be a legitimate way of
2885 obtaining infinities and NaNs.
2887 @item -Wsystem-headers
2888 @opindex Wsystem-headers
2889 @cindex warnings from system headers
2890 @cindex system headers, warnings from
2891 Print warning messages for constructs found in system header files.
2892 Warnings from system headers are normally suppressed, on the assumption
2893 that they usually do not indicate real problems and would only make the
2894 compiler output harder to read. Using this command line option tells
2895 GCC to emit warnings from system headers as if they occurred in user
2896 code. However, note that using @option{-Wall} in conjunction with this
2897 option will @emph{not} warn about unknown pragmas in system
2898 headers---for that, @option{-Wunknown-pragmas} must also be used.
2901 @opindex Wfloat-equal
2902 Warn if floating point values are used in equality comparisons.
2904 The idea behind this is that sometimes it is convenient (for the
2905 programmer) to consider floating-point values as approximations to
2906 infinitely precise real numbers. If you are doing this, then you need
2907 to compute (by analyzing the code, or in some other way) the maximum or
2908 likely maximum error that the computation introduces, and allow for it
2909 when performing comparisons (and when producing output, but that's a
2910 different problem). In particular, instead of testing for equality, you
2911 would check to see whether the two values have ranges that overlap; and
2912 this is done with the relational operators, so equality comparisons are
2915 @item -Wtraditional @r{(C only)}
2916 @opindex Wtraditional
2917 Warn about certain constructs that behave differently in traditional and
2918 ISO C@. Also warn about ISO C constructs that have no traditional C
2919 equivalent, and/or problematic constructs which should be avoided.
2923 Macro parameters that appear within string literals in the macro body.
2924 In traditional C macro replacement takes place within string literals,
2925 but does not in ISO C@.
2928 In traditional C, some preprocessor directives did not exist.
2929 Traditional preprocessors would only consider a line to be a directive
2930 if the @samp{#} appeared in column 1 on the line. Therefore
2931 @option{-Wtraditional} warns about directives that traditional C
2932 understands but would ignore because the @samp{#} does not appear as the
2933 first character on the line. It also suggests you hide directives like
2934 @samp{#pragma} not understood by traditional C by indenting them. Some
2935 traditional implementations would not recognize @samp{#elif}, so it
2936 suggests avoiding it altogether.
2939 A function-like macro that appears without arguments.
2942 The unary plus operator.
2945 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2946 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2947 constants.) Note, these suffixes appear in macros defined in the system
2948 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2949 Use of these macros in user code might normally lead to spurious
2950 warnings, however GCC's integrated preprocessor has enough context to
2951 avoid warning in these cases.
2954 A function declared external in one block and then used after the end of
2958 A @code{switch} statement has an operand of type @code{long}.
2961 A non-@code{static} function declaration follows a @code{static} one.
2962 This construct is not accepted by some traditional C compilers.
2965 The ISO type of an integer constant has a different width or
2966 signedness from its traditional type. This warning is only issued if
2967 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2968 typically represent bit patterns, are not warned about.
2971 Usage of ISO string concatenation is detected.
2974 Initialization of automatic aggregates.
2977 Identifier conflicts with labels. Traditional C lacks a separate
2978 namespace for labels.
2981 Initialization of unions. If the initializer is zero, the warning is
2982 omitted. This is done under the assumption that the zero initializer in
2983 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2984 initializer warnings and relies on default initialization to zero in the
2988 Conversions by prototypes between fixed/floating point values and vice
2989 versa. The absence of these prototypes when compiling with traditional
2990 C would cause serious problems. This is a subset of the possible
2991 conversion warnings, for the full set use @option{-Wconversion}.
2994 Use of ISO C style function definitions. This warning intentionally is
2995 @emph{not} issued for prototype declarations or variadic functions
2996 because these ISO C features will appear in your code when using
2997 libiberty's traditional C compatibility macros, @code{PARAMS} and
2998 @code{VPARAMS}. This warning is also bypassed for nested functions
2999 because that feature is already a GCC extension and thus not relevant to
3000 traditional C compatibility.
3003 @item -Wdeclaration-after-statement @r{(C only)}
3004 @opindex Wdeclaration-after-statement
3005 Warn when a declaration is found after a statement in a block. This
3006 construct, known from C++, was introduced with ISO C99 and is by default
3007 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3008 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3012 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3014 @item -Wno-endif-labels
3015 @opindex Wno-endif-labels
3016 @opindex Wendif-labels
3017 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3021 Warn whenever a local variable shadows another local variable, parameter or
3022 global variable or whenever a built-in function is shadowed.
3024 @item -Wlarger-than-@var{len}
3025 @opindex Wlarger-than
3026 Warn whenever an object of larger than @var{len} bytes is defined.
3028 @item -Wunsafe-loop-optimizations
3029 @opindex Wunsafe-loop-optimizations
3030 Warn if the loop cannot be optimized because the compiler could not
3031 assume anything on the bounds of the loop indices. With
3032 @option{-funsafe-loop-optimizations} warn if the compiler made
3035 @item -Wpointer-arith
3036 @opindex Wpointer-arith
3037 Warn about anything that depends on the ``size of'' a function type or
3038 of @code{void}. GNU C assigns these types a size of 1, for
3039 convenience in calculations with @code{void *} pointers and pointers
3042 @item -Wbad-function-cast @r{(C only)}
3043 @opindex Wbad-function-cast
3044 Warn whenever a function call is cast to a non-matching type.
3045 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3048 Warn about ISO C constructs that are outside of the common subset of
3049 ISO C and ISO C++, e.g.@: request for implicit conversion from
3050 @code{void *} to a pointer to non-@code{void} type.
3054 Warn whenever a pointer is cast so as to remove a type qualifier from
3055 the target type. For example, warn if a @code{const char *} is cast
3056 to an ordinary @code{char *}.
3059 @opindex Wcast-align
3060 Warn whenever a pointer is cast such that the required alignment of the
3061 target is increased. For example, warn if a @code{char *} is cast to
3062 an @code{int *} on machines where integers can only be accessed at
3063 two- or four-byte boundaries.
3065 @item -Wwrite-strings
3066 @opindex Wwrite-strings
3067 When compiling C, give string constants the type @code{const
3068 char[@var{length}]} so that
3069 copying the address of one into a non-@code{const} @code{char *}
3070 pointer will get a warning; when compiling C++, warn about the
3071 deprecated conversion from string literals to @code{char *}. This
3072 warning, by default, is enabled for C++ programs.
3073 These warnings will help you find at
3074 compile time code that can try to write into a string constant, but
3075 only if you have been very careful about using @code{const} in
3076 declarations and prototypes. Otherwise, it will just be a nuisance;
3077 this is why we did not make @option{-Wall} request these warnings.
3080 @opindex Wconversion
3081 Warn if a prototype causes a type conversion that is different from what
3082 would happen to the same argument in the absence of a prototype. This
3083 includes conversions of fixed point to floating and vice versa, and
3084 conversions changing the width or signedness of a fixed point argument
3085 except when the same as the default promotion.
3087 Also, warn if a negative integer constant expression is implicitly
3088 converted to an unsigned type. For example, warn about the assignment
3089 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
3090 casts like @code{(unsigned) -1}.
3092 @item -Wsign-compare
3093 @opindex Wsign-compare
3094 @cindex warning for comparison of signed and unsigned values
3095 @cindex comparison of signed and unsigned values, warning
3096 @cindex signed and unsigned values, comparison warning
3097 Warn when a comparison between signed and unsigned values could produce
3098 an incorrect result when the signed value is converted to unsigned.
3099 This warning is also enabled by @option{-Wextra}; to get the other warnings
3100 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3102 @item -Waggregate-return
3103 @opindex Waggregate-return
3104 Warn if any functions that return structures or unions are defined or
3105 called. (In languages where you can return an array, this also elicits
3109 @opindex Walways-true
3110 Warn about comparisons which are always true such as testing if
3111 unsigned values are greater than or equal to zero. This warning is
3112 enabled by @option{-Wall}.
3114 @item -Wno-attributes
3115 @opindex Wno-attributes
3116 @opindex Wattributes
3117 Do not warn if an unexpected @code{__attribute__} is used, such as
3118 unrecognized attributes, function attributes applied to variables,
3119 etc. This will not stop errors for incorrect use of supported
3122 @item -Wstrict-prototypes @r{(C only)}
3123 @opindex Wstrict-prototypes
3124 Warn if a function is declared or defined without specifying the
3125 argument types. (An old-style function definition is permitted without
3126 a warning if preceded by a declaration which specifies the argument
3129 @item -Wold-style-definition @r{(C only)}
3130 @opindex Wold-style-definition
3131 Warn if an old-style function definition is used. A warning is given
3132 even if there is a previous prototype.
3134 @item -Wmissing-prototypes @r{(C only)}
3135 @opindex Wmissing-prototypes
3136 Warn if a global function is defined without a previous prototype
3137 declaration. This warning is issued even if the definition itself
3138 provides a prototype. The aim is to detect global functions that fail
3139 to be declared in header files.
3141 @item -Wmissing-declarations @r{(C only)}
3142 @opindex Wmissing-declarations
3143 Warn if a global function is defined without a previous declaration.
3144 Do so even if the definition itself provides a prototype.
3145 Use this option to detect global functions that are not declared in
3148 @item -Wmissing-field-initializers
3149 @opindex Wmissing-field-initializers
3152 Warn if a structure's initializer has some fields missing. For
3153 example, the following code would cause such a warning, because
3154 @code{x.h} is implicitly zero:
3157 struct s @{ int f, g, h; @};
3158 struct s x = @{ 3, 4 @};
3161 This option does not warn about designated initializers, so the following
3162 modification would not trigger a warning:
3165 struct s @{ int f, g, h; @};
3166 struct s x = @{ .f = 3, .g = 4 @};
3169 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3170 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3172 @item -Wmissing-noreturn
3173 @opindex Wmissing-noreturn
3174 Warn about functions which might be candidates for attribute @code{noreturn}.
3175 Note these are only possible candidates, not absolute ones. Care should
3176 be taken to manually verify functions actually do not ever return before
3177 adding the @code{noreturn} attribute, otherwise subtle code generation
3178 bugs could be introduced. You will not get a warning for @code{main} in
3179 hosted C environments.
3181 @item -Wmissing-format-attribute
3182 @opindex Wmissing-format-attribute
3184 Warn about function pointers which might be candidates for @code{format}
3185 attributes. Note these are only possible candidates, not absolute ones.
3186 GCC will guess that function pointers with @code{format} attributes that
3187 are used in assignment, initialization, parameter passing or return
3188 statements should have a corresponding @code{format} attribute in the
3189 resulting type. I.e.@: the left-hand side of the assignment or
3190 initialization, the type of the parameter variable, or the return type
3191 of the containing function respectively should also have a @code{format}
3192 attribute to avoid the warning.
3194 GCC will also warn about function definitions which might be
3195 candidates for @code{format} attributes. Again, these are only
3196 possible candidates. GCC will guess that @code{format} attributes
3197 might be appropriate for any function that calls a function like
3198 @code{vprintf} or @code{vscanf}, but this might not always be the
3199 case, and some functions for which @code{format} attributes are
3200 appropriate may not be detected.
3202 @item -Wno-multichar
3203 @opindex Wno-multichar
3205 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3206 Usually they indicate a typo in the user's code, as they have
3207 implementation-defined values, and should not be used in portable code.
3209 @item -Wnormalized=<none|id|nfc|nfkc>
3210 @opindex Wnormalized
3213 @cindex character set, input normalization
3214 In ISO C and ISO C++, two identifiers are different if they are
3215 different sequences of characters. However, sometimes when characters
3216 outside the basic ASCII character set are used, you can have two
3217 different character sequences that look the same. To avoid confusion,
3218 the ISO 10646 standard sets out some @dfn{normalization rules} which
3219 when applied ensure that two sequences that look the same are turned into
3220 the same sequence. GCC can warn you if you are using identifiers which
3221 have not been normalized; this option controls that warning.
3223 There are four levels of warning that GCC supports. The default is
3224 @option{-Wnormalized=nfc}, which warns about any identifier which is
3225 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3226 recommended form for most uses.
3228 Unfortunately, there are some characters which ISO C and ISO C++ allow
3229 in identifiers that when turned into NFC aren't allowable as
3230 identifiers. That is, there's no way to use these symbols in portable
3231 ISO C or C++ and have all your identifiers in NFC.
3232 @option{-Wnormalized=id} suppresses the warning for these characters.
3233 It is hoped that future versions of the standards involved will correct
3234 this, which is why this option is not the default.
3236 You can switch the warning off for all characters by writing
3237 @option{-Wnormalized=none}. You would only want to do this if you
3238 were using some other normalization scheme (like ``D''), because
3239 otherwise you can easily create bugs that are literally impossible to see.
3241 Some characters in ISO 10646 have distinct meanings but look identical
3242 in some fonts or display methodologies, especially once formatting has
3243 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3244 LETTER N'', will display just like a regular @code{n} which has been
3245 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3246 normalisation scheme to convert all these into a standard form as
3247 well, and GCC will warn if your code is not in NFKC if you use
3248 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3249 about every identifier that contains the letter O because it might be
3250 confused with the digit 0, and so is not the default, but may be
3251 useful as a local coding convention if the programming environment is
3252 unable to be fixed to display these characters distinctly.
3254 @item -Wno-deprecated-declarations
3255 @opindex Wno-deprecated-declarations
3256 Do not warn about uses of functions, variables, and types marked as
3257 deprecated by using the @code{deprecated} attribute.
3258 (@pxref{Function Attributes}, @pxref{Variable Attributes},
3259 @pxref{Type Attributes}.)
3263 Warn if a structure is given the packed attribute, but the packed
3264 attribute has no effect on the layout or size of the structure.
3265 Such structures may be mis-aligned for little benefit. For
3266 instance, in this code, the variable @code{f.x} in @code{struct bar}
3267 will be misaligned even though @code{struct bar} does not itself
3268 have the packed attribute:
3275 @} __attribute__((packed));
3285 Warn if padding is included in a structure, either to align an element
3286 of the structure or to align the whole structure. Sometimes when this
3287 happens it is possible to rearrange the fields of the structure to
3288 reduce the padding and so make the structure smaller.
3290 @item -Wredundant-decls
3291 @opindex Wredundant-decls
3292 Warn if anything is declared more than once in the same scope, even in
3293 cases where multiple declaration is valid and changes nothing.
3295 @item -Wnested-externs @r{(C only)}
3296 @opindex Wnested-externs
3297 Warn if an @code{extern} declaration is encountered within a function.
3299 @item -Wunreachable-code
3300 @opindex Wunreachable-code
3301 Warn if the compiler detects that code will never be executed.
3303 This option is intended to warn when the compiler detects that at
3304 least a whole line of source code will never be executed, because
3305 some condition is never satisfied or because it is after a
3306 procedure that never returns.
3308 It is possible for this option to produce a warning even though there
3309 are circumstances under which part of the affected line can be executed,
3310 so care should be taken when removing apparently-unreachable code.
3312 For instance, when a function is inlined, a warning may mean that the
3313 line is unreachable in only one inlined copy of the function.
3315 This option is not made part of @option{-Wall} because in a debugging
3316 version of a program there is often substantial code which checks
3317 correct functioning of the program and is, hopefully, unreachable
3318 because the program does work. Another common use of unreachable
3319 code is to provide behavior which is selectable at compile-time.
3323 Warn if a function can not be inlined and it was declared as inline.
3324 Even with this option, the compiler will not warn about failures to
3325 inline functions declared in system headers.
3327 The compiler uses a variety of heuristics to determine whether or not
3328 to inline a function. For example, the compiler takes into account
3329 the size of the function being inlined and the amount of inlining
3330 that has already been done in the current function. Therefore,
3331 seemingly insignificant changes in the source program can cause the
3332 warnings produced by @option{-Winline} to appear or disappear.
3334 @item -Wno-invalid-offsetof @r{(C++ only)}
3335 @opindex Wno-invalid-offsetof
3336 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3337 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3338 to a non-POD type is undefined. In existing C++ implementations,
3339 however, @samp{offsetof} typically gives meaningful results even when
3340 applied to certain kinds of non-POD types. (Such as a simple
3341 @samp{struct} that fails to be a POD type only by virtue of having a
3342 constructor.) This flag is for users who are aware that they are
3343 writing nonportable code and who have deliberately chosen to ignore the
3346 The restrictions on @samp{offsetof} may be relaxed in a future version
3347 of the C++ standard.
3349 @item -Wno-int-to-pointer-cast @r{(C only)}
3350 @opindex Wno-int-to-pointer-cast
3351 Suppress warnings from casts to pointer type of an integer of a
3354 @item -Wno-pointer-to-int-cast @r{(C only)}
3355 @opindex Wno-pointer-to-int-cast
3356 Suppress warnings from casts from a pointer to an integer type of a
3360 @opindex Winvalid-pch
3361 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3362 the search path but can't be used.
3366 @opindex Wno-long-long
3367 Warn if @samp{long long} type is used. This is default. To inhibit
3368 the warning messages, use @option{-Wno-long-long}. Flags
3369 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3370 only when @option{-pedantic} flag is used.
3372 @item -Wvariadic-macros
3373 @opindex Wvariadic-macros
3374 @opindex Wno-variadic-macros
3375 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3376 alternate syntax when in pedantic ISO C99 mode. This is default.
3377 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3379 @item -Wvolatile-register-var
3380 @opindex Wvolatile-register-var
3381 @opindex Wno-volatile-register-var
3382 Warn if a register variable is declared volatile. The volatile
3383 modifier does not inhibit all optimizations that may eliminate reads
3384 and/or writes to register variables.
3386 @item -Wdisabled-optimization
3387 @opindex Wdisabled-optimization
3388 Warn if a requested optimization pass is disabled. This warning does
3389 not generally indicate that there is anything wrong with your code; it
3390 merely indicates that GCC's optimizers were unable to handle the code
3391 effectively. Often, the problem is that your code is too big or too
3392 complex; GCC will refuse to optimize programs when the optimization
3393 itself is likely to take inordinate amounts of time.
3395 @item -Wpointer-sign
3396 @opindex Wpointer-sign
3397 @opindex Wno-pointer-sign
3398 Warn for pointer argument passing or assignment with different signedness.
3399 This option is only supported for C and Objective-C@. It is implied by
3400 @option{-Wall} and by @option{-pedantic}, which can be disabled with
3401 @option{-Wno-pointer-sign}.
3405 Make all warnings into errors.
3409 Make the specified warning into an errors. The specifier for a
3410 warning is appended, for example @option{-Werror=switch} turns the
3411 warnings controlled by @option{-Wswitch} into errors. This switch
3412 takes a negative form, to be used to negate @option{-Werror} for
3413 specific warnings, for example @option{-Wno-error=switch} makes
3414 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
3415 is in effect. You can use the @option{-fdiagnostics-show-option}
3416 option to have each controllable warning amended with the option which
3417 controls it, to determine what to use with this option.
3419 Note that specifying @option{-Werror=}@var{foo} automatically implies
3420 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
3423 @item -Wstack-protector
3424 @opindex Wstack-protector
3425 This option is only active when @option{-fstack-protector} is active. It
3426 warns about functions that will not be protected against stack smashing.
3428 @item -Wstring-literal-comparison
3429 @opindex Wstring-literal-comparison
3430 Warn about suspicious comparisons to string literal constants. In C,
3431 direct comparisons against the memory address of a string literal, such
3432 as @code{if (x == "abc")}, typically indicate a programmer error, and
3433 even when intentional, result in unspecified behavior and are not portable.
3434 Usually these warnings alert that the programmer intended to use
3435 @code{strcmp}. This warning is enabled by @option{-Wall}.
3437 @item -Woverlength-strings
3438 @opindex Woverlength-strings
3439 Warn about string constants which are longer than the ``minimum
3440 maximum'' length specified in the C standard. Modern compilers
3441 generally allow string constants which are much longer than the
3442 standard's minimum limit, but very portable programs should avoid
3443 using longer strings.
3445 The limit applies @emph{after} string constant concatenation, and does
3446 not count the trailing NUL@. In C89, the limit was 509 characters; in
3447 C99, it was raised to 4095. C++98 does not specify a normative
3448 minimum maximum, so we do not diagnose overlength strings in C++@.
3450 This option is implied by @option{-pedantic}, and can be disabled with
3451 @option{-Wno-overlength-strings}.
3454 @node Debugging Options
3455 @section Options for Debugging Your Program or GCC
3456 @cindex options, debugging
3457 @cindex debugging information options
3459 GCC has various special options that are used for debugging
3460 either your program or GCC:
3465 Produce debugging information in the operating system's native format
3466 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
3469 On most systems that use stabs format, @option{-g} enables use of extra
3470 debugging information that only GDB can use; this extra information
3471 makes debugging work better in GDB but will probably make other debuggers
3473 refuse to read the program. If you want to control for certain whether
3474 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3475 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3477 GCC allows you to use @option{-g} with
3478 @option{-O}. The shortcuts taken by optimized code may occasionally
3479 produce surprising results: some variables you declared may not exist
3480 at all; flow of control may briefly move where you did not expect it;
3481 some statements may not be executed because they compute constant
3482 results or their values were already at hand; some statements may
3483 execute in different places because they were moved out of loops.
3485 Nevertheless it proves possible to debug optimized output. This makes
3486 it reasonable to use the optimizer for programs that might have bugs.
3488 The following options are useful when GCC is generated with the
3489 capability for more than one debugging format.
3493 Produce debugging information for use by GDB@. This means to use the
3494 most expressive format available (DWARF 2, stabs, or the native format
3495 if neither of those are supported), including GDB extensions if at all
3500 Produce debugging information in stabs format (if that is supported),
3501 without GDB extensions. This is the format used by DBX on most BSD
3502 systems. On MIPS, Alpha and System V Release 4 systems this option
3503 produces stabs debugging output which is not understood by DBX or SDB@.
3504 On System V Release 4 systems this option requires the GNU assembler.
3506 @item -feliminate-unused-debug-symbols
3507 @opindex feliminate-unused-debug-symbols
3508 Produce debugging information in stabs format (if that is supported),
3509 for only symbols that are actually used.
3513 Produce debugging information in stabs format (if that is supported),
3514 using GNU extensions understood only by the GNU debugger (GDB)@. The
3515 use of these extensions is likely to make other debuggers crash or
3516 refuse to read the program.
3520 Produce debugging information in COFF format (if that is supported).
3521 This is the format used by SDB on most System V systems prior to
3526 Produce debugging information in XCOFF format (if that is supported).
3527 This is the format used by the DBX debugger on IBM RS/6000 systems.
3531 Produce debugging information in XCOFF format (if that is supported),
3532 using GNU extensions understood only by the GNU debugger (GDB)@. The
3533 use of these extensions is likely to make other debuggers crash or
3534 refuse to read the program, and may cause assemblers other than the GNU
3535 assembler (GAS) to fail with an error.
3539 Produce debugging information in DWARF version 2 format (if that is
3540 supported). This is the format used by DBX on IRIX 6. With this
3541 option, GCC uses features of DWARF version 3 when they are useful;
3542 version 3 is upward compatible with version 2, but may still cause
3543 problems for older debuggers.
3547 Produce debugging information in VMS debug format (if that is
3548 supported). This is the format used by DEBUG on VMS systems.
3551 @itemx -ggdb@var{level}
3552 @itemx -gstabs@var{level}
3553 @itemx -gcoff@var{level}
3554 @itemx -gxcoff@var{level}
3555 @itemx -gvms@var{level}
3556 Request debugging information and also use @var{level} to specify how
3557 much information. The default level is 2.
3559 Level 1 produces minimal information, enough for making backtraces in
3560 parts of the program that you don't plan to debug. This includes
3561 descriptions of functions and external variables, but no information
3562 about local variables and no line numbers.
3564 Level 3 includes extra information, such as all the macro definitions
3565 present in the program. Some debuggers support macro expansion when
3566 you use @option{-g3}.
3568 @option{-gdwarf-2} does not accept a concatenated debug level, because
3569 GCC used to support an option @option{-gdwarf} that meant to generate
3570 debug information in version 1 of the DWARF format (which is very
3571 different from version 2), and it would have been too confusing. That
3572 debug format is long obsolete, but the option cannot be changed now.
3573 Instead use an additional @option{-g@var{level}} option to change the
3574 debug level for DWARF2.
3576 @item -feliminate-dwarf2-dups
3577 @opindex feliminate-dwarf2-dups
3578 Compress DWARF2 debugging information by eliminating duplicated
3579 information about each symbol. This option only makes sense when
3580 generating DWARF2 debugging information with @option{-gdwarf-2}.
3582 @cindex @command{prof}
3585 Generate extra code to write profile information suitable for the
3586 analysis program @command{prof}. You must use this option when compiling
3587 the source files you want data about, and you must also use it when
3590 @cindex @command{gprof}
3593 Generate extra code to write profile information suitable for the
3594 analysis program @command{gprof}. You must use this option when compiling
3595 the source files you want data about, and you must also use it when
3600 Makes the compiler print out each function name as it is compiled, and
3601 print some statistics about each pass when it finishes.
3604 @opindex ftime-report
3605 Makes the compiler print some statistics about the time consumed by each
3606 pass when it finishes.
3609 @opindex fmem-report
3610 Makes the compiler print some statistics about permanent memory
3611 allocation when it finishes.
3613 @item -fprofile-arcs
3614 @opindex fprofile-arcs
3615 Add code so that program flow @dfn{arcs} are instrumented. During
3616 execution the program records how many times each branch and call is
3617 executed and how many times it is taken or returns. When the compiled
3618 program exits it saves this data to a file called
3619 @file{@var{auxname}.gcda} for each source file. The data may be used for
3620 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3621 test coverage analysis (@option{-ftest-coverage}). Each object file's
3622 @var{auxname} is generated from the name of the output file, if
3623 explicitly specified and it is not the final executable, otherwise it is
3624 the basename of the source file. In both cases any suffix is removed
3625 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
3626 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3627 @xref{Cross-profiling}.
3629 @cindex @command{gcov}
3633 This option is used to compile and link code instrumented for coverage
3634 analysis. The option is a synonym for @option{-fprofile-arcs}
3635 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
3636 linking). See the documentation for those options for more details.
3641 Compile the source files with @option{-fprofile-arcs} plus optimization
3642 and code generation options. For test coverage analysis, use the
3643 additional @option{-ftest-coverage} option. You do not need to profile
3644 every source file in a program.
3647 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3648 (the latter implies the former).
3651 Run the program on a representative workload to generate the arc profile
3652 information. This may be repeated any number of times. You can run
3653 concurrent instances of your program, and provided that the file system
3654 supports locking, the data files will be correctly updated. Also
3655 @code{fork} calls are detected and correctly handled (double counting
3659 For profile-directed optimizations, compile the source files again with
3660 the same optimization and code generation options plus
3661 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3662 Control Optimization}).
3665 For test coverage analysis, use @command{gcov} to produce human readable
3666 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3667 @command{gcov} documentation for further information.
3671 With @option{-fprofile-arcs}, for each function of your program GCC
3672 creates a program flow graph, then finds a spanning tree for the graph.
3673 Only arcs that are not on the spanning tree have to be instrumented: the
3674 compiler adds code to count the number of times that these arcs are
3675 executed. When an arc is the only exit or only entrance to a block, the
3676 instrumentation code can be added to the block; otherwise, a new basic
3677 block must be created to hold the instrumentation code.
3680 @item -ftest-coverage
3681 @opindex ftest-coverage
3682 Produce a notes file that the @command{gcov} code-coverage utility
3683 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3684 show program coverage. Each source file's note file is called
3685 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3686 above for a description of @var{auxname} and instructions on how to
3687 generate test coverage data. Coverage data will match the source files
3688 more closely, if you do not optimize.
3690 @item -d@var{letters}
3691 @item -fdump-rtl-@var{pass}
3693 Says to make debugging dumps during compilation at times specified by
3694 @var{letters}. This is used for debugging the RTL-based passes of the
3695 compiler. The file names for most of the dumps are made by appending a
3696 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
3697 from the name of the output file, if explicitly specified and it is not
3698 an executable, otherwise it is the basename of the source file.
3700 Most debug dumps can be enabled either passing a letter to the @option{-d}
3701 option, or with a long @option{-fdump-rtl} switch; here are the possible
3702 letters for use in @var{letters} and @var{pass}, and their meanings:
3707 Annotate the assembler output with miscellaneous debugging information.
3710 @itemx -fdump-rtl-bp
3712 @opindex fdump-rtl-bp
3713 Dump after computing branch probabilities, to @file{@var{file}.09.bp}.
3716 @itemx -fdump-rtl-bbro
3718 @opindex fdump-rtl-bbro
3719 Dump after block reordering, to @file{@var{file}.30.bbro}.
3722 @itemx -fdump-rtl-combine
3724 @opindex fdump-rtl-combine
3725 Dump after instruction combination, to the file @file{@var{file}.17.combine}.
3728 @itemx -fdump-rtl-ce1
3729 @itemx -fdump-rtl-ce2
3731 @opindex fdump-rtl-ce1
3732 @opindex fdump-rtl-ce2
3733 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
3734 first if conversion, to the file @file{@var{file}.11.ce1}. @option{-dC}
3735 and @option{-fdump-rtl-ce2} enable dumping after the second if
3736 conversion, to the file @file{@var{file}.18.ce2}.
3739 @itemx -fdump-rtl-btl
3740 @itemx -fdump-rtl-dbr
3742 @opindex fdump-rtl-btl
3743 @opindex fdump-rtl-dbr
3744 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
3745 target load optimization, to @file{@var{file}.31.btl}. @option{-dd}
3746 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
3747 scheduling, to @file{@var{file}.36.dbr}.
3751 Dump all macro definitions, at the end of preprocessing, in addition to
3755 @itemx -fdump-rtl-ce3
3757 @opindex fdump-rtl-ce3
3758 Dump after the third if conversion, to @file{@var{file}.28.ce3}.
3761 @itemx -fdump-rtl-cfg
3762 @itemx -fdump-rtl-life
3764 @opindex fdump-rtl-cfg
3765 @opindex fdump-rtl-life
3766 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
3767 and data flow analysis, to @file{@var{file}.08.cfg}. @option{-df}
3768 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
3769 to @file{@var{file}.16.life}.
3772 @itemx -fdump-rtl-greg
3774 @opindex fdump-rtl-greg
3775 Dump after global register allocation, to @file{@var{file}.23.greg}.
3778 @itemx -fdump-rtl-gcse
3779 @itemx -fdump-rtl-bypass
3781 @opindex fdump-rtl-gcse
3782 @opindex fdump-rtl-bypass
3783 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
3784 @file{@var{file}.05.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
3785 enable dumping after jump bypassing and control flow optimizations, to
3786 @file{@var{file}.07.bypass}.
3789 @itemx -fdump-rtl-eh
3791 @opindex fdump-rtl-eh
3792 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
3795 @itemx -fdump-rtl-sibling
3797 @opindex fdump-rtl-sibling
3798 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
3801 @itemx -fdump-rtl-jump
3803 @opindex fdump-rtl-jump
3804 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
3807 @itemx -fdump-rtl-stack
3809 @opindex fdump-rtl-stack
3810 Dump after conversion from registers to stack, to @file{@var{file}.33.stack}.
3813 @itemx -fdump-rtl-lreg
3815 @opindex fdump-rtl-lreg
3816 Dump after local register allocation, to @file{@var{file}.22.lreg}.
3819 @itemx -fdump-rtl-loop
3820 @itemx -fdump-rtl-loop2
3822 @opindex fdump-rtl-loop
3823 @opindex fdump-rtl-loop2
3824 @option{-dL} and @option{-fdump-rtl-loop} enable dumping after the first
3825 loop optimization pass, to @file{@var{file}.06.loop}. @option{-dL} and
3826 @option{-fdump-rtl-loop2} enable dumping after the second pass, to
3827 @file{@var{file}.13.loop2}.
3830 @itemx -fdump-rtl-sms
3832 @opindex fdump-rtl-sms
3833 Dump after modulo scheduling, to @file{@var{file}.20.sms}.
3836 @itemx -fdump-rtl-mach
3838 @opindex fdump-rtl-mach
3839 Dump after performing the machine dependent reorganization pass, to
3840 @file{@var{file}.35.mach}.
3843 @itemx -fdump-rtl-rnreg
3845 @opindex fdump-rtl-rnreg
3846 Dump after register renumbering, to @file{@var{file}.29.rnreg}.
3849 @itemx -fdump-rtl-regmove
3851 @opindex fdump-rtl-regmove
3852 Dump after the register move pass, to @file{@var{file}.19.regmove}.
3855 @itemx -fdump-rtl-postreload
3857 @opindex fdump-rtl-postreload
3858 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
3861 @itemx -fdump-rtl-expand
3863 @opindex fdump-rtl-expand
3864 Dump after RTL generation, to @file{@var{file}.00.expand}.
3867 @itemx -fdump-rtl-sched2
3869 @opindex fdump-rtl-sched2
3870 Dump after the second scheduling pass, to @file{@var{file}.32.sched2}.
3873 @itemx -fdump-rtl-cse
3875 @opindex fdump-rtl-cse
3876 Dump after CSE (including the jump optimization that sometimes follows
3877 CSE), to @file{@var{file}.04.cse}.
3880 @itemx -fdump-rtl-sched
3882 @opindex fdump-rtl-sched
3883 Dump after the first scheduling pass, to @file{@var{file}.21.sched}.
3886 @itemx -fdump-rtl-cse2
3888 @opindex fdump-rtl-cse2
3889 Dump after the second CSE pass (including the jump optimization that
3890 sometimes follows CSE), to @file{@var{file}.15.cse2}.
3893 @itemx -fdump-rtl-tracer
3895 @opindex fdump-rtl-tracer
3896 Dump after running tracer, to @file{@var{file}.12.tracer}.
3899 @itemx -fdump-rtl-vpt
3900 @itemx -fdump-rtl-vartrack
3902 @opindex fdump-rtl-vpt
3903 @opindex fdump-rtl-vartrack
3904 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
3905 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
3906 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
3907 to @file{@var{file}.34.vartrack}.
3910 @itemx -fdump-rtl-flow2
3912 @opindex fdump-rtl-flow2
3913 Dump after the second flow pass, to @file{@var{file}.26.flow2}.
3916 @itemx -fdump-rtl-peephole2
3918 @opindex fdump-rtl-peephole2
3919 Dump after the peephole pass, to @file{@var{file}.27.peephole2}.
3922 @itemx -fdump-rtl-web
3924 @opindex fdump-rtl-web
3925 Dump after live range splitting, to @file{@var{file}.14.web}.
3928 @itemx -fdump-rtl-all
3930 @opindex fdump-rtl-all
3931 Produce all the dumps listed above.
3935 Produce a core dump whenever an error occurs.
3939 Print statistics on memory usage, at the end of the run, to
3944 Annotate the assembler output with a comment indicating which
3945 pattern and alternative was used. The length of each instruction is
3950 Dump the RTL in the assembler output as a comment before each instruction.
3951 Also turns on @option{-dp} annotation.
3955 For each of the other indicated dump files (either with @option{-d} or
3956 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
3957 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3961 Just generate RTL for a function instead of compiling it. Usually used
3962 with @samp{r} (@option{-fdump-rtl-expand}).
3966 Dump debugging information during parsing, to standard error.
3969 @item -fdump-unnumbered
3970 @opindex fdump-unnumbered
3971 When doing debugging dumps (see @option{-d} option above), suppress instruction
3972 numbers and line number note output. This makes it more feasible to
3973 use diff on debugging dumps for compiler invocations with different
3974 options, in particular with and without @option{-g}.
3976 @item -fdump-translation-unit @r{(C++ only)}
3977 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
3978 @opindex fdump-translation-unit
3979 Dump a representation of the tree structure for the entire translation
3980 unit to a file. The file name is made by appending @file{.tu} to the
3981 source file name. If the @samp{-@var{options}} form is used, @var{options}
3982 controls the details of the dump as described for the
3983 @option{-fdump-tree} options.
3985 @item -fdump-class-hierarchy @r{(C++ only)}
3986 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3987 @opindex fdump-class-hierarchy
3988 Dump a representation of each class's hierarchy and virtual function
3989 table layout to a file. The file name is made by appending @file{.class}
3990 to the source file name. If the @samp{-@var{options}} form is used,
3991 @var{options} controls the details of the dump as described for the
3992 @option{-fdump-tree} options.
3994 @item -fdump-ipa-@var{switch}
3996 Control the dumping at various stages of inter-procedural analysis
3997 language tree to a file. The file name is generated by appending a switch
3998 specific suffix to the source file name. The following dumps are possible:
4002 Enables all inter-procedural analysis dumps; currently the only produced
4003 dump is the @samp{cgraph} dump.
4006 Dumps information about call-graph optimization, unused function removal,
4007 and inlining decisions.
4010 @item -fdump-tree-@var{switch}
4011 @itemx -fdump-tree-@var{switch}-@var{options}
4013 Control the dumping at various stages of processing the intermediate
4014 language tree to a file. The file name is generated by appending a switch
4015 specific suffix to the source file name. If the @samp{-@var{options}}
4016 form is used, @var{options} is a list of @samp{-} separated options that
4017 control the details of the dump. Not all options are applicable to all
4018 dumps, those which are not meaningful will be ignored. The following
4019 options are available
4023 Print the address of each node. Usually this is not meaningful as it
4024 changes according to the environment and source file. Its primary use
4025 is for tying up a dump file with a debug environment.
4027 Inhibit dumping of members of a scope or body of a function merely
4028 because that scope has been reached. Only dump such items when they
4029 are directly reachable by some other path. When dumping pretty-printed
4030 trees, this option inhibits dumping the bodies of control structures.
4032 Print a raw representation of the tree. By default, trees are
4033 pretty-printed into a C-like representation.
4035 Enable more detailed dumps (not honored by every dump option).
4037 Enable dumping various statistics about the pass (not honored by every dump
4040 Enable showing basic block boundaries (disabled in raw dumps).
4042 Enable showing virtual operands for every statement.
4044 Enable showing line numbers for statements.
4046 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4048 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4051 The following tree dumps are possible:
4055 Dump before any tree based optimization, to @file{@var{file}.original}.
4058 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4061 Dump after function inlining, to @file{@var{file}.inlined}.
4064 @opindex fdump-tree-gimple
4065 Dump each function before and after the gimplification pass to a file. The
4066 file name is made by appending @file{.gimple} to the source file name.
4069 @opindex fdump-tree-cfg
4070 Dump the control flow graph of each function to a file. The file name is
4071 made by appending @file{.cfg} to the source file name.
4074 @opindex fdump-tree-vcg
4075 Dump the control flow graph of each function to a file in VCG format. The
4076 file name is made by appending @file{.vcg} to the source file name. Note
4077 that if the file contains more than one function, the generated file cannot
4078 be used directly by VCG@. You will need to cut and paste each function's
4079 graph into its own separate file first.
4082 @opindex fdump-tree-ch
4083 Dump each function after copying loop headers. The file name is made by
4084 appending @file{.ch} to the source file name.
4087 @opindex fdump-tree-ssa
4088 Dump SSA related information to a file. The file name is made by appending
4089 @file{.ssa} to the source file name.
4092 @opindex fdump-tree-salias
4093 Dump structure aliasing variable information to a file. This file name
4094 is made by appending @file{.salias} to the source file name.
4097 @opindex fdump-tree-alias
4098 Dump aliasing information for each function. The file name is made by
4099 appending @file{.alias} to the source file name.
4102 @opindex fdump-tree-ccp
4103 Dump each function after CCP@. The file name is made by appending
4104 @file{.ccp} to the source file name.
4107 @opindex fdump-tree-storeccp
4108 Dump each function after STORE-CCP. The file name is made by appending
4109 @file{.storeccp} to the source file name.
4112 @opindex fdump-tree-pre
4113 Dump trees after partial redundancy elimination. The file name is made
4114 by appending @file{.pre} to the source file name.
4117 @opindex fdump-tree-fre
4118 Dump trees after full redundancy elimination. The file name is made
4119 by appending @file{.fre} to the source file name.
4122 @opindex fdump-tree-copyprop
4123 Dump trees after copy propagation. The file name is made
4124 by appending @file{.copyprop} to the source file name.
4126 @item store_copyprop
4127 @opindex fdump-tree-store_copyprop
4128 Dump trees after store copy-propagation. The file name is made
4129 by appending @file{.store_copyprop} to the source file name.
4132 @opindex fdump-tree-dce
4133 Dump each function after dead code elimination. The file name is made by
4134 appending @file{.dce} to the source file name.
4137 @opindex fdump-tree-mudflap
4138 Dump each function after adding mudflap instrumentation. The file name is
4139 made by appending @file{.mudflap} to the source file name.
4142 @opindex fdump-tree-sra
4143 Dump each function after performing scalar replacement of aggregates. The
4144 file name is made by appending @file{.sra} to the source file name.
4147 @opindex fdump-tree-sink
4148 Dump each function after performing code sinking. The file name is made
4149 by appending @file{.sink} to the source file name.
4152 @opindex fdump-tree-dom
4153 Dump each function after applying dominator tree optimizations. The file
4154 name is made by appending @file{.dom} to the source file name.
4157 @opindex fdump-tree-dse
4158 Dump each function after applying dead store elimination. The file
4159 name is made by appending @file{.dse} to the source file name.
4162 @opindex fdump-tree-phiopt
4163 Dump each function after optimizing PHI nodes into straightline code. The file
4164 name is made by appending @file{.phiopt} to the source file name.
4167 @opindex fdump-tree-forwprop
4168 Dump each function after forward propagating single use variables. The file
4169 name is made by appending @file{.forwprop} to the source file name.
4172 @opindex fdump-tree-copyrename
4173 Dump each function after applying the copy rename optimization. The file
4174 name is made by appending @file{.copyrename} to the source file name.
4177 @opindex fdump-tree-nrv
4178 Dump each function after applying the named return value optimization on
4179 generic trees. The file name is made by appending @file{.nrv} to the source
4183 @opindex fdump-tree-vect
4184 Dump each function after applying vectorization of loops. The file name is
4185 made by appending @file{.vect} to the source file name.
4188 @opindex fdump-tree-vrp
4189 Dump each function after Value Range Propagation (VRP). The file name
4190 is made by appending @file{.vrp} to the source file name.
4193 @opindex fdump-tree-all
4194 Enable all the available tree dumps with the flags provided in this option.
4197 @item -ftree-vectorizer-verbose=@var{n}
4198 @opindex ftree-vectorizer-verbose
4199 This option controls the amount of debugging output the vectorizer prints.
4200 This information is written to standard error, unless @option{-fdump-tree-all}
4201 or @option{-fdump-tree-vect} is specified, in which case it is output to the
4202 usual dump listing file, @file{.vect}.
4204 @item -frandom-seed=@var{string}
4205 @opindex frandom-string
4206 This option provides a seed that GCC uses when it would otherwise use
4207 random numbers. It is used to generate certain symbol names
4208 that have to be different in every compiled file. It is also used to
4209 place unique stamps in coverage data files and the object files that
4210 produce them. You can use the @option{-frandom-seed} option to produce
4211 reproducibly identical object files.
4213 The @var{string} should be different for every file you compile.
4215 @item -fsched-verbose=@var{n}
4216 @opindex fsched-verbose
4217 On targets that use instruction scheduling, this option controls the
4218 amount of debugging output the scheduler prints. This information is
4219 written to standard error, unless @option{-dS} or @option{-dR} is
4220 specified, in which case it is output to the usual dump
4221 listing file, @file{.sched} or @file{.sched2} respectively. However
4222 for @var{n} greater than nine, the output is always printed to standard
4225 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4226 same information as @option{-dRS}. For @var{n} greater than one, it
4227 also output basic block probabilities, detailed ready list information
4228 and unit/insn info. For @var{n} greater than two, it includes RTL
4229 at abort point, control-flow and regions info. And for @var{n} over
4230 four, @option{-fsched-verbose} also includes dependence info.
4234 Store the usual ``temporary'' intermediate files permanently; place them
4235 in the current directory and name them based on the source file. Thus,
4236 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4237 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4238 preprocessed @file{foo.i} output file even though the compiler now
4239 normally uses an integrated preprocessor.
4241 When used in combination with the @option{-x} command line option,
4242 @option{-save-temps} is sensible enough to avoid over writing an
4243 input source file with the same extension as an intermediate file.
4244 The corresponding intermediate file may be obtained by renaming the
4245 source file before using @option{-save-temps}.
4249 Report the CPU time taken by each subprocess in the compilation
4250 sequence. For C source files, this is the compiler proper and assembler
4251 (plus the linker if linking is done). The output looks like this:
4258 The first number on each line is the ``user time'', that is time spent
4259 executing the program itself. The second number is ``system time'',
4260 time spent executing operating system routines on behalf of the program.
4261 Both numbers are in seconds.
4263 @item -fvar-tracking
4264 @opindex fvar-tracking
4265 Run variable tracking pass. It computes where variables are stored at each
4266 position in code. Better debugging information is then generated
4267 (if the debugging information format supports this information).
4269 It is enabled by default when compiling with optimization (@option{-Os},
4270 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
4271 the debug info format supports it.
4273 @item -print-file-name=@var{library}
4274 @opindex print-file-name
4275 Print the full absolute name of the library file @var{library} that
4276 would be used when linking---and don't do anything else. With this
4277 option, GCC does not compile or link anything; it just prints the
4280 @item -print-multi-directory
4281 @opindex print-multi-directory
4282 Print the directory name corresponding to the multilib selected by any
4283 other switches present in the command line. This directory is supposed
4284 to exist in @env{GCC_EXEC_PREFIX}.
4286 @item -print-multi-lib
4287 @opindex print-multi-lib
4288 Print the mapping from multilib directory names to compiler switches
4289 that enable them. The directory name is separated from the switches by
4290 @samp{;}, and each switch starts with an @samp{@@} instead of the
4291 @samp{-}, without spaces between multiple switches. This is supposed to
4292 ease shell-processing.
4294 @item -print-prog-name=@var{program}
4295 @opindex print-prog-name
4296 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
4298 @item -print-libgcc-file-name
4299 @opindex print-libgcc-file-name
4300 Same as @option{-print-file-name=libgcc.a}.
4302 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
4303 but you do want to link with @file{libgcc.a}. You can do
4306 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
4309 @item -print-search-dirs
4310 @opindex print-search-dirs
4311 Print the name of the configured installation directory and a list of
4312 program and library directories @command{gcc} will search---and don't do anything else.
4314 This is useful when @command{gcc} prints the error message
4315 @samp{installation problem, cannot exec cpp0: No such file or directory}.
4316 To resolve this you either need to put @file{cpp0} and the other compiler
4317 components where @command{gcc} expects to find them, or you can set the environment
4318 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
4319 Don't forget the trailing @samp{/}.
4320 @xref{Environment Variables}.
4323 @opindex dumpmachine
4324 Print the compiler's target machine (for example,
4325 @samp{i686-pc-linux-gnu})---and don't do anything else.
4328 @opindex dumpversion
4329 Print the compiler version (for example, @samp{3.0})---and don't do
4334 Print the compiler's built-in specs---and don't do anything else. (This
4335 is used when GCC itself is being built.) @xref{Spec Files}.
4337 @item -feliminate-unused-debug-types
4338 @opindex feliminate-unused-debug-types
4339 Normally, when producing DWARF2 output, GCC will emit debugging
4340 information for all types declared in a compilation
4341 unit, regardless of whether or not they are actually used
4342 in that compilation unit. Sometimes this is useful, such as
4343 if, in the debugger, you want to cast a value to a type that is
4344 not actually used in your program (but is declared). More often,
4345 however, this results in a significant amount of wasted space.
4346 With this option, GCC will avoid producing debug symbol output
4347 for types that are nowhere used in the source file being compiled.
4350 @node Optimize Options
4351 @section Options That Control Optimization
4352 @cindex optimize options
4353 @cindex options, optimization
4355 These options control various sorts of optimizations.
4357 Without any optimization option, the compiler's goal is to reduce the
4358 cost of compilation and to make debugging produce the expected
4359 results. Statements are independent: if you stop the program with a
4360 breakpoint between statements, you can then assign a new value to any
4361 variable or change the program counter to any other statement in the
4362 function and get exactly the results you would expect from the source
4365 Turning on optimization flags makes the compiler attempt to improve
4366 the performance and/or code size at the expense of compilation time
4367 and possibly the ability to debug the program.
4369 The compiler performs optimization based on the knowledge it has of
4370 the program. Optimization levels @option{-O2} and above, in
4371 particular, enable @emph{unit-at-a-time} mode, which allows the
4372 compiler to consider information gained from later functions in
4373 the file when compiling a function. Compiling multiple files at
4374 once to a single output file in @emph{unit-at-a-time} mode allows
4375 the compiler to use information gained from all of the files when
4376 compiling each of them.
4378 Not all optimizations are controlled directly by a flag. Only
4379 optimizations that have a flag are listed.
4386 Optimize. Optimizing compilation takes somewhat more time, and a lot
4387 more memory for a large function.
4389 With @option{-O}, the compiler tries to reduce code size and execution
4390 time, without performing any optimizations that take a great deal of
4393 @option{-O} turns on the following optimization flags:
4394 @gccoptlist{-fdefer-pop @gol
4395 -fdelayed-branch @gol
4396 -fguess-branch-probability @gol
4397 -fcprop-registers @gol
4398 -floop-optimize @gol
4399 -fif-conversion @gol
4400 -fif-conversion2 @gol
4403 -ftree-dominator-opts @gol
4408 -ftree-copyrename @gol
4413 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
4414 where doing so does not interfere with debugging.
4418 Optimize even more. GCC performs nearly all supported optimizations
4419 that do not involve a space-speed tradeoff. The compiler does not
4420 perform loop unrolling or function inlining when you specify @option{-O2}.
4421 As compared to @option{-O}, this option increases both compilation time
4422 and the performance of the generated code.
4424 @option{-O2} turns on all optimization flags specified by @option{-O}. It
4425 also turns on the following optimization flags:
4426 @gccoptlist{-fthread-jumps @gol
4428 -foptimize-sibling-calls @gol
4429 -fcse-follow-jumps -fcse-skip-blocks @gol
4430 -fgcse -fgcse-lm @gol
4431 -fexpensive-optimizations @gol
4432 -fstrength-reduce @gol
4433 -frerun-cse-after-loop -frerun-loop-opt @gol
4436 -fschedule-insns -fschedule-insns2 @gol
4437 -fsched-interblock -fsched-spec @gol
4439 -fstrict-aliasing @gol
4440 -fdelete-null-pointer-checks @gol
4441 -freorder-blocks -freorder-functions @gol
4442 -funit-at-a-time @gol
4443 -falign-functions -falign-jumps @gol
4444 -falign-loops -falign-labels @gol
4448 Please note the warning under @option{-fgcse} about
4449 invoking @option{-O2} on programs that use computed gotos.
4453 Optimize yet more. @option{-O3} turns on all optimizations specified by
4454 @option{-O2} and also turns on the @option{-finline-functions},
4455 @option{-funswitch-loops} and @option{-fgcse-after-reload} options.
4459 Do not optimize. This is the default.
4463 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
4464 do not typically increase code size. It also performs further
4465 optimizations designed to reduce code size.
4467 @option{-Os} disables the following optimization flags:
4468 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
4469 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
4470 -fprefetch-loop-arrays -ftree-vect-loop-version}
4472 If you use multiple @option{-O} options, with or without level numbers,
4473 the last such option is the one that is effective.
4476 Options of the form @option{-f@var{flag}} specify machine-independent
4477 flags. Most flags have both positive and negative forms; the negative
4478 form of @option{-ffoo} would be @option{-fno-foo}. In the table
4479 below, only one of the forms is listed---the one you typically will
4480 use. You can figure out the other form by either removing @samp{no-}
4483 The following options control specific optimizations. They are either
4484 activated by @option{-O} options or are related to ones that are. You
4485 can use the following flags in the rare cases when ``fine-tuning'' of
4486 optimizations to be performed is desired.
4489 @item -fno-default-inline
4490 @opindex fno-default-inline
4491 Do not make member functions inline by default merely because they are
4492 defined inside the class scope (C++ only). Otherwise, when you specify
4493 @w{@option{-O}}, member functions defined inside class scope are compiled
4494 inline by default; i.e., you don't need to add @samp{inline} in front of
4495 the member function name.
4497 @item -fno-defer-pop
4498 @opindex fno-defer-pop
4499 Always pop the arguments to each function call as soon as that function
4500 returns. For machines which must pop arguments after a function call,
4501 the compiler normally lets arguments accumulate on the stack for several
4502 function calls and pops them all at once.
4504 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4508 Force memory operands to be copied into registers before doing
4509 arithmetic on them. This produces better code by making all memory
4510 references potential common subexpressions. When they are not common
4511 subexpressions, instruction combination should eliminate the separate
4512 register-load. This option is now a nop and will be removed in 4.2.
4515 @opindex fforce-addr
4516 Force memory address constants to be copied into registers before
4517 doing arithmetic on them.
4519 @item -fomit-frame-pointer
4520 @opindex fomit-frame-pointer
4521 Don't keep the frame pointer in a register for functions that
4522 don't need one. This avoids the instructions to save, set up and
4523 restore frame pointers; it also makes an extra register available
4524 in many functions. @strong{It also makes debugging impossible on
4527 On some machines, such as the VAX, this flag has no effect, because
4528 the standard calling sequence automatically handles the frame pointer
4529 and nothing is saved by pretending it doesn't exist. The
4530 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
4531 whether a target machine supports this flag. @xref{Registers,,Register
4532 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
4534 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4536 @item -foptimize-sibling-calls
4537 @opindex foptimize-sibling-calls
4538 Optimize sibling and tail recursive calls.
4540 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4544 Don't pay attention to the @code{inline} keyword. Normally this option
4545 is used to keep the compiler from expanding any functions inline.
4546 Note that if you are not optimizing, no functions can be expanded inline.
4548 @item -finline-functions
4549 @opindex finline-functions
4550 Integrate all simple functions into their callers. The compiler
4551 heuristically decides which functions are simple enough to be worth
4552 integrating in this way.
4554 If all calls to a given function are integrated, and the function is
4555 declared @code{static}, then the function is normally not output as
4556 assembler code in its own right.
4558 Enabled at level @option{-O3}.
4560 @item -finline-functions-called-once
4561 @opindex finline-functions-called-once
4562 Consider all @code{static} functions called once for inlining into their
4563 caller even if they are not marked @code{inline}. If a call to a given
4564 function is integrated, then the function is not output as assembler code
4567 Enabled if @option{-funit-at-a-time} is enabled.
4569 @item -fearly-inlining
4570 @opindex fearly-inlining
4571 Inline functions marked by @code{always_inline} and functions whose body seems
4572 smaller than the function call overhead early before doing
4573 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
4574 makes profiling significantly cheaper and usually inlining faster on programs
4575 having large chains of nested wrapper functions.
4579 @item -finline-limit=@var{n}
4580 @opindex finline-limit
4581 By default, GCC limits the size of functions that can be inlined. This flag
4582 allows the control of this limit for functions that are explicitly marked as
4583 inline (i.e., marked with the inline keyword or defined within the class
4584 definition in c++). @var{n} is the size of functions that can be inlined in
4585 number of pseudo instructions (not counting parameter handling). The default
4586 value of @var{n} is 600.
4587 Increasing this value can result in more inlined code at
4588 the cost of compilation time and memory consumption. Decreasing usually makes
4589 the compilation faster and less code will be inlined (which presumably
4590 means slower programs). This option is particularly useful for programs that
4591 use inlining heavily such as those based on recursive templates with C++.
4593 Inlining is actually controlled by a number of parameters, which may be
4594 specified individually by using @option{--param @var{name}=@var{value}}.
4595 The @option{-finline-limit=@var{n}} option sets some of these parameters
4599 @item max-inline-insns-single
4600 is set to @var{n}/2.
4601 @item max-inline-insns-auto
4602 is set to @var{n}/2.
4603 @item min-inline-insns
4604 is set to 130 or @var{n}/4, whichever is smaller.
4605 @item max-inline-insns-rtl
4609 See below for a documentation of the individual
4610 parameters controlling inlining.
4612 @emph{Note:} pseudo instruction represents, in this particular context, an
4613 abstract measurement of function's size. In no way does it represent a count
4614 of assembly instructions and as such its exact meaning might change from one
4615 release to an another.
4617 @item -fkeep-inline-functions
4618 @opindex fkeep-inline-functions
4619 In C, emit @code{static} functions that are declared @code{inline}
4620 into the object file, even if the function has been inlined into all
4621 of its callers. This switch does not affect functions using the
4622 @code{extern inline} extension in GNU C@. In C++, emit any and all
4623 inline functions into the object file.
4625 @item -fkeep-static-consts
4626 @opindex fkeep-static-consts
4627 Emit variables declared @code{static const} when optimization isn't turned
4628 on, even if the variables aren't referenced.
4630 GCC enables this option by default. If you want to force the compiler to
4631 check if the variable was referenced, regardless of whether or not
4632 optimization is turned on, use the @option{-fno-keep-static-consts} option.
4634 @item -fmerge-constants
4635 Attempt to merge identical constants (string constants and floating point
4636 constants) across compilation units.
4638 This option is the default for optimized compilation if the assembler and
4639 linker support it. Use @option{-fno-merge-constants} to inhibit this
4642 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4644 @item -fmerge-all-constants
4645 Attempt to merge identical constants and identical variables.
4647 This option implies @option{-fmerge-constants}. In addition to
4648 @option{-fmerge-constants} this considers e.g.@: even constant initialized
4649 arrays or initialized constant variables with integral or floating point
4650 types. Languages like C or C++ require each non-automatic variable to
4651 have distinct location, so using this option will result in non-conforming
4654 @item -fmodulo-sched
4655 @opindex fmodulo-sched
4656 Perform swing modulo scheduling immediately before the first scheduling
4657 pass. This pass looks at innermost loops and reorders their
4658 instructions by overlapping different iterations.
4660 @item -fno-branch-count-reg
4661 @opindex fno-branch-count-reg
4662 Do not use ``decrement and branch'' instructions on a count register,
4663 but instead generate a sequence of instructions that decrement a
4664 register, compare it against zero, then branch based upon the result.
4665 This option is only meaningful on architectures that support such
4666 instructions, which include x86, PowerPC, IA-64 and S/390.
4668 The default is @option{-fbranch-count-reg}, enabled when
4669 @option{-fstrength-reduce} is enabled.
4671 @item -fno-function-cse
4672 @opindex fno-function-cse
4673 Do not put function addresses in registers; make each instruction that
4674 calls a constant function contain the function's address explicitly.
4676 This option results in less efficient code, but some strange hacks
4677 that alter the assembler output may be confused by the optimizations
4678 performed when this option is not used.
4680 The default is @option{-ffunction-cse}
4682 @item -fno-zero-initialized-in-bss
4683 @opindex fno-zero-initialized-in-bss
4684 If the target supports a BSS section, GCC by default puts variables that
4685 are initialized to zero into BSS@. This can save space in the resulting
4688 This option turns off this behavior because some programs explicitly
4689 rely on variables going to the data section. E.g., so that the
4690 resulting executable can find the beginning of that section and/or make
4691 assumptions based on that.
4693 The default is @option{-fzero-initialized-in-bss}.
4695 @item -fbounds-check
4696 @opindex fbounds-check
4697 For front-ends that support it, generate additional code to check that
4698 indices used to access arrays are within the declared range. This is
4699 currently only supported by the Java and Fortran front-ends, where
4700 this option defaults to true and false respectively.
4702 @item -fmudflap -fmudflapth -fmudflapir
4706 @cindex bounds checking
4708 For front-ends that support it (C and C++), instrument all risky
4709 pointer/array dereferencing operations, some standard library
4710 string/heap functions, and some other associated constructs with
4711 range/validity tests. Modules so instrumented should be immune to
4712 buffer overflows, invalid heap use, and some other classes of C/C++
4713 programming errors. The instrumentation relies on a separate runtime
4714 library (@file{libmudflap}), which will be linked into a program if
4715 @option{-fmudflap} is given at link time. Run-time behavior of the
4716 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
4717 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
4720 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
4721 link if your program is multi-threaded. Use @option{-fmudflapir}, in
4722 addition to @option{-fmudflap} or @option{-fmudflapth}, if
4723 instrumentation should ignore pointer reads. This produces less
4724 instrumentation (and therefore faster execution) and still provides
4725 some protection against outright memory corrupting writes, but allows
4726 erroneously read data to propagate within a program.
4730 @cindex openmp parallel
4731 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
4732 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
4733 compiler generates parallel code according to the OpenMP Application
4734 Program Interface v2.5. To generate the final exectuable, the runtime
4735 library @code{libgomp} must be linked in using @option{-lgomp}.
4737 @item -fstrength-reduce
4738 @opindex fstrength-reduce
4739 Perform the optimizations of loop strength reduction and
4740 elimination of iteration variables.
4742 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4744 @item -fthread-jumps
4745 @opindex fthread-jumps
4746 Perform optimizations where we check to see if a jump branches to a
4747 location where another comparison subsumed by the first is found. If
4748 so, the first branch is redirected to either the destination of the
4749 second branch or a point immediately following it, depending on whether
4750 the condition is known to be true or false.
4752 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4754 @item -fcse-follow-jumps
4755 @opindex fcse-follow-jumps
4756 In common subexpression elimination, scan through jump instructions
4757 when the target of the jump is not reached by any other path. For
4758 example, when CSE encounters an @code{if} statement with an
4759 @code{else} clause, CSE will follow the jump when the condition
4762 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4764 @item -fcse-skip-blocks
4765 @opindex fcse-skip-blocks
4766 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
4767 follow jumps which conditionally skip over blocks. When CSE
4768 encounters a simple @code{if} statement with no else clause,
4769 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
4770 body of the @code{if}.
4772 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4774 @item -frerun-cse-after-loop
4775 @opindex frerun-cse-after-loop
4776 Re-run common subexpression elimination after loop optimizations has been
4779 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4781 @item -frerun-loop-opt
4782 @opindex frerun-loop-opt
4783 Run the loop optimizer twice.
4785 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4789 Perform a global common subexpression elimination pass.
4790 This pass also performs global constant and copy propagation.
4792 @emph{Note:} When compiling a program using computed gotos, a GCC
4793 extension, you may get better runtime performance if you disable
4794 the global common subexpression elimination pass by adding
4795 @option{-fno-gcse} to the command line.
4797 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4801 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
4802 attempt to move loads which are only killed by stores into themselves. This
4803 allows a loop containing a load/store sequence to be changed to a load outside
4804 the loop, and a copy/store within the loop.
4806 Enabled by default when gcse is enabled.
4810 When @option{-fgcse-sm} is enabled, a store motion pass is run after
4811 global common subexpression elimination. This pass will attempt to move
4812 stores out of loops. When used in conjunction with @option{-fgcse-lm},
4813 loops containing a load/store sequence can be changed to a load before
4814 the loop and a store after the loop.
4816 Not enabled at any optimization level.
4820 When @option{-fgcse-las} is enabled, the global common subexpression
4821 elimination pass eliminates redundant loads that come after stores to the
4822 same memory location (both partial and full redundancies).
4824 Not enabled at any optimization level.
4826 @item -fgcse-after-reload
4827 @opindex fgcse-after-reload
4828 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
4829 pass is performed after reload. The purpose of this pass is to cleanup
4832 @item -floop-optimize
4833 @opindex floop-optimize
4834 Perform loop optimizations: move constant expressions out of loops, simplify
4835 exit test conditions and optionally do strength-reduction as well.
4837 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4839 @item -funsafe-loop-optimizations
4840 @opindex funsafe-loop-optimizations
4841 If given, the loop optimizer will assume that loop indices do not
4842 overflow, and that the loops with nontrivial exit condition are not
4843 infinite. This enables a wider range of loop optimizations even if
4844 the loop optimizer itself cannot prove that these assumptions are valid.
4845 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
4846 if it finds this kind of loop.
4848 @item -fcrossjumping
4849 @opindex crossjumping
4850 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
4851 resulting code may or may not perform better than without cross-jumping.
4853 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4855 @item -fif-conversion
4856 @opindex if-conversion
4857 Attempt to transform conditional jumps into branch-less equivalents. This
4858 include use of conditional moves, min, max, set flags and abs instructions, and
4859 some tricks doable by standard arithmetics. The use of conditional execution
4860 on chips where it is available is controlled by @code{if-conversion2}.
4862 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4864 @item -fif-conversion2
4865 @opindex if-conversion2
4866 Use conditional execution (where available) to transform conditional jumps into
4867 branch-less equivalents.
4869 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4871 @item -fdelete-null-pointer-checks
4872 @opindex fdelete-null-pointer-checks
4873 Use global dataflow analysis to identify and eliminate useless checks
4874 for null pointers. The compiler assumes that dereferencing a null
4875 pointer would have halted the program. If a pointer is checked after
4876 it has already been dereferenced, it cannot be null.
4878 In some environments, this assumption is not true, and programs can
4879 safely dereference null pointers. Use
4880 @option{-fno-delete-null-pointer-checks} to disable this optimization
4881 for programs which depend on that behavior.
4883 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4885 @item -fexpensive-optimizations
4886 @opindex fexpensive-optimizations
4887 Perform a number of minor optimizations that are relatively expensive.
4889 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4891 @item -foptimize-register-move
4893 @opindex foptimize-register-move
4895 Attempt to reassign register numbers in move instructions and as
4896 operands of other simple instructions in order to maximize the amount of
4897 register tying. This is especially helpful on machines with two-operand
4900 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
4903 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4905 @item -fdelayed-branch
4906 @opindex fdelayed-branch
4907 If supported for the target machine, attempt to reorder instructions
4908 to exploit instruction slots available after delayed branch
4911 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4913 @item -fschedule-insns
4914 @opindex fschedule-insns
4915 If supported for the target machine, attempt to reorder instructions to
4916 eliminate execution stalls due to required data being unavailable. This
4917 helps machines that have slow floating point or memory load instructions
4918 by allowing other instructions to be issued until the result of the load
4919 or floating point instruction is required.
4921 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4923 @item -fschedule-insns2
4924 @opindex fschedule-insns2
4925 Similar to @option{-fschedule-insns}, but requests an additional pass of
4926 instruction scheduling after register allocation has been done. This is
4927 especially useful on machines with a relatively small number of
4928 registers and where memory load instructions take more than one cycle.
4930 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4932 @item -fno-sched-interblock
4933 @opindex fno-sched-interblock
4934 Don't schedule instructions across basic blocks. This is normally
4935 enabled by default when scheduling before register allocation, i.e.@:
4936 with @option{-fschedule-insns} or at @option{-O2} or higher.
4938 @item -fno-sched-spec
4939 @opindex fno-sched-spec
4940 Don't allow speculative motion of non-load instructions. This is normally
4941 enabled by default when scheduling before register allocation, i.e.@:
4942 with @option{-fschedule-insns} or at @option{-O2} or higher.
4944 @item -fsched-spec-load
4945 @opindex fsched-spec-load
4946 Allow speculative motion of some load instructions. This only makes
4947 sense when scheduling before register allocation, i.e.@: with
4948 @option{-fschedule-insns} or at @option{-O2} or higher.
4950 @item -fsched-spec-load-dangerous
4951 @opindex fsched-spec-load-dangerous
4952 Allow speculative motion of more load instructions. This only makes
4953 sense when scheduling before register allocation, i.e.@: with
4954 @option{-fschedule-insns} or at @option{-O2} or higher.
4956 @item -fsched-stalled-insns=@var{n}
4957 @opindex fsched-stalled-insns
4958 Define how many insns (if any) can be moved prematurely from the queue
4959 of stalled insns into the ready list, during the second scheduling pass.
4961 @item -fsched-stalled-insns-dep=@var{n}
4962 @opindex fsched-stalled-insns-dep
4963 Define how many insn groups (cycles) will be examined for a dependency
4964 on a stalled insn that is candidate for premature removal from the queue
4965 of stalled insns. Has an effect only during the second scheduling pass,
4966 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
4968 @item -fsched2-use-superblocks
4969 @opindex fsched2-use-superblocks
4970 When scheduling after register allocation, do use superblock scheduling
4971 algorithm. Superblock scheduling allows motion across basic block boundaries
4972 resulting on faster schedules. This option is experimental, as not all machine
4973 descriptions used by GCC model the CPU closely enough to avoid unreliable
4974 results from the algorithm.
4976 This only makes sense when scheduling after register allocation, i.e.@: with
4977 @option{-fschedule-insns2} or at @option{-O2} or higher.
4979 @item -fsched2-use-traces
4980 @opindex fsched2-use-traces
4981 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4982 allocation and additionally perform code duplication in order to increase the
4983 size of superblocks using tracer pass. See @option{-ftracer} for details on
4986 This mode should produce faster but significantly longer programs. Also
4987 without @option{-fbranch-probabilities} the traces constructed may not
4988 match the reality and hurt the performance. This only makes
4989 sense when scheduling after register allocation, i.e.@: with
4990 @option{-fschedule-insns2} or at @option{-O2} or higher.
4992 @item -freschedule-modulo-scheduled-loops
4993 @opindex fscheduling-in-modulo-scheduled-loops
4994 The modulo scheduling comes before the traditional scheduling, if a loop was modulo scheduled
4995 we may want to prevent the later scheduling passes from changing its schedule, we use this
4996 option to control that.
4998 @item -fcaller-saves
4999 @opindex fcaller-saves
5000 Enable values to be allocated in registers that will be clobbered by
5001 function calls, by emitting extra instructions to save and restore the
5002 registers around such calls. Such allocation is done only when it
5003 seems to result in better code than would otherwise be produced.
5005 This option is always enabled by default on certain machines, usually
5006 those which have no call-preserved registers to use instead.
5008 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5011 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
5012 enabled by default at @option{-O2} and @option{-O3}.
5015 Perform Full Redundancy Elimination (FRE) on trees. The difference
5016 between FRE and PRE is that FRE only considers expressions
5017 that are computed on all paths leading to the redundant computation.
5018 This analysis faster than PRE, though it exposes fewer redundancies.
5019 This flag is enabled by default at @option{-O} and higher.
5021 @item -ftree-copy-prop
5022 Perform copy propagation on trees. This pass eliminates unnecessary
5023 copy operations. This flag is enabled by default at @option{-O} and
5026 @item -ftree-store-copy-prop
5027 Perform copy propagation of memory loads and stores. This pass
5028 eliminates unnecessary copy operations in memory references
5029 (structures, global variables, arrays, etc). This flag is enabled by
5030 default at @option{-O2} and higher.
5033 Perform structural alias analysis on trees. This flag
5034 is enabled by default at @option{-O} and higher.
5037 Perform interprocedural pointer analysis.
5040 Perform forward store motion on trees. This flag is
5041 enabled by default at @option{-O} and higher.
5044 Perform sparse conditional constant propagation (CCP) on trees. This
5045 pass only operates on local scalar variables and is enabled by default
5046 at @option{-O} and higher.
5048 @item -ftree-store-ccp
5049 Perform sparse conditional constant propagation (CCP) on trees. This
5050 pass operates on both local scalar variables and memory stores and
5051 loads (global variables, structures, arrays, etc). This flag is
5052 enabled by default at @option{-O2} and higher.
5055 Perform dead code elimination (DCE) on trees. This flag is enabled by
5056 default at @option{-O} and higher.
5058 @item -ftree-dominator-opts
5059 Perform a variety of simple scalar cleanups (constant/copy
5060 propagation, redundancy elimination, range propagation and expression
5061 simplification) based on a dominator tree traversal. This also
5062 performs jump threading (to reduce jumps to jumps). This flag is
5063 enabled by default at @option{-O} and higher.
5066 Perform loop header copying on trees. This is beneficial since it increases
5067 effectiveness of code motion optimizations. It also saves one jump. This flag
5068 is enabled by default at @option{-O} and higher. It is not enabled
5069 for @option{-Os}, since it usually increases code size.
5071 @item -ftree-loop-optimize
5072 Perform loop optimizations on trees. This flag is enabled by default
5073 at @option{-O} and higher.
5075 @item -ftree-loop-linear
5076 Perform linear loop transformations on tree. This flag can improve cache
5077 performance and allow further loop optimizations to take place.
5079 @item -ftree-loop-im
5080 Perform loop invariant motion on trees. This pass moves only invariants that
5081 would be hard to handle at RTL level (function calls, operations that expand to
5082 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5083 operands of conditions that are invariant out of the loop, so that we can use
5084 just trivial invariantness analysis in loop unswitching. The pass also includes
5087 @item -ftree-loop-ivcanon
5088 Create a canonical counter for number of iterations in the loop for that
5089 determining number of iterations requires complicated analysis. Later
5090 optimizations then may determine the number easily. Useful especially
5091 in connection with unrolling.
5094 Perform induction variable optimizations (strength reduction, induction
5095 variable merging and induction variable elimination) on trees.
5098 Perform scalar replacement of aggregates. This pass replaces structure
5099 references with scalars to prevent committing structures to memory too
5100 early. This flag is enabled by default at @option{-O} and higher.
5102 @item -ftree-copyrename
5103 Perform copy renaming on trees. This pass attempts to rename compiler
5104 temporaries to other variables at copy locations, usually resulting in
5105 variable names which more closely resemble the original variables. This flag
5106 is enabled by default at @option{-O} and higher.
5109 Perform temporary expression replacement during the SSA->normal phase. Single
5110 use/single def temporaries are replaced at their use location with their
5111 defining expression. This results in non-GIMPLE code, but gives the expanders
5112 much more complex trees to work on resulting in better RTL generation. This is
5113 enabled by default at @option{-O} and higher.
5116 Perform live range splitting during the SSA->normal phase. Distinct live
5117 ranges of a variable are split into unique variables, allowing for better
5118 optimization later. This is enabled by default at @option{-O} and higher.
5120 @item -ftree-vectorize
5121 Perform loop vectorization on trees.
5123 @item -ftree-vect-loop-version
5124 @opindex ftree-vect-loop-version
5125 Perform loop versioning when doing loop vectorization on trees. When a loop
5126 appears to be vectorizable except that data alignment or data dependence cannot
5127 be determined at compile time then vectorized and non-vectorized versions of
5128 the loop are generated along with runtime checks for alignment or dependence
5129 to control which version is executed. This option is enabled by default
5130 except at level @option{-Os} where it is disabled.
5133 Perform Value Range Propagation on trees. This is similar to the
5134 constant propagation pass, but instead of values, ranges of values are
5135 propagated. This allows the optimizers to remove unnecessary range
5136 checks like array bound checks and null pointer checks. This is
5137 enabled by default at @option{-O2} and higher. Null pointer check
5138 elimination is only done if @option{-fdelete-null-pointer-checks} is
5143 Perform tail duplication to enlarge superblock size. This transformation
5144 simplifies the control flow of the function allowing other optimizations to do
5147 @item -funroll-loops
5148 @opindex funroll-loops
5149 Unroll loops whose number of iterations can be determined at compile
5150 time or upon entry to the loop. @option{-funroll-loops} implies both
5151 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
5152 option makes code larger, and may or may not make it run faster.
5154 @item -funroll-all-loops
5155 @opindex funroll-all-loops
5156 Unroll all loops, even if their number of iterations is uncertain when
5157 the loop is entered. This usually makes programs run more slowly.
5158 @option{-funroll-all-loops} implies the same options as
5159 @option{-funroll-loops},
5161 @item -fsplit-ivs-in-unroller
5162 @opindex -fsplit-ivs-in-unroller
5163 Enables expressing of values of induction variables in later iterations
5164 of the unrolled loop using the value in the first iteration. This breaks
5165 long dependency chains, thus improving efficiency of the scheduling passes.
5167 Combination of @option{-fweb} and CSE is often sufficient to obtain the
5168 same effect. However in cases the loop body is more complicated than
5169 a single basic block, this is not reliable. It also does not work at all
5170 on some of the architectures due to restrictions in the CSE pass.
5172 This optimization is enabled by default.
5174 @item -fvariable-expansion-in-unroller
5175 @opindex -fvariable-expansion-in-unroller
5176 With this option, the compiler will create multiple copies of some
5177 local variables when unrolling a loop which can result in superior code.
5179 @item -fprefetch-loop-arrays
5180 @itemx -fprefetch-loop-arrays-rtl
5181 @opindex fprefetch-loop-arrays
5182 @opindex fprefetch-loop-arrays-rtl
5183 If supported by the target machine, generate instructions to prefetch
5184 memory to improve the performance of loops that access large arrays.
5186 These options may generate better or worse code; results are highly
5187 dependent on the structure of loops within the source code.
5190 @itemx -fno-peephole2
5191 @opindex fno-peephole
5192 @opindex fno-peephole2
5193 Disable any machine-specific peephole optimizations. The difference
5194 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
5195 are implemented in the compiler; some targets use one, some use the
5196 other, a few use both.
5198 @option{-fpeephole} is enabled by default.
5199 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5201 @item -fno-guess-branch-probability
5202 @opindex fno-guess-branch-probability
5203 Do not guess branch probabilities using heuristics.
5205 GCC will use heuristics to guess branch probabilities if they are
5206 not provided by profiling feedback (@option{-fprofile-arcs}). These
5207 heuristics are based on the control flow graph. If some branch probabilities
5208 are specified by @samp{__builtin_expect}, then the heuristics will be
5209 used to guess branch probabilities for the rest of the control flow graph,
5210 taking the @samp{__builtin_expect} info into account. The interactions
5211 between the heuristics and @samp{__builtin_expect} can be complex, and in
5212 some cases, it may be useful to disable the heuristics so that the effects
5213 of @samp{__builtin_expect} are easier to understand.
5215 The default is @option{-fguess-branch-probability} at levels
5216 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5218 @item -freorder-blocks
5219 @opindex freorder-blocks
5220 Reorder basic blocks in the compiled function in order to reduce number of
5221 taken branches and improve code locality.
5223 Enabled at levels @option{-O2}, @option{-O3}.
5225 @item -freorder-blocks-and-partition
5226 @opindex freorder-blocks-and-partition
5227 In addition to reordering basic blocks in the compiled function, in order
5228 to reduce number of taken branches, partitions hot and cold basic blocks
5229 into separate sections of the assembly and .o files, to improve
5230 paging and cache locality performance.
5232 This optimization is automatically turned off in the presence of
5233 exception handling, for linkonce sections, for functions with a user-defined
5234 section attribute and on any architecture that does not support named
5237 @item -freorder-functions
5238 @opindex freorder-functions
5239 Reorder functions in the object file in order to
5240 improve code locality. This is implemented by using special
5241 subsections @code{.text.hot} for most frequently executed functions and
5242 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
5243 the linker so object file format must support named sections and linker must
5244 place them in a reasonable way.
5246 Also profile feedback must be available in to make this option effective. See
5247 @option{-fprofile-arcs} for details.
5249 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5251 @item -fstrict-aliasing
5252 @opindex fstrict-aliasing
5253 Allows the compiler to assume the strictest aliasing rules applicable to
5254 the language being compiled. For C (and C++), this activates
5255 optimizations based on the type of expressions. In particular, an
5256 object of one type is assumed never to reside at the same address as an
5257 object of a different type, unless the types are almost the same. For
5258 example, an @code{unsigned int} can alias an @code{int}, but not a
5259 @code{void*} or a @code{double}. A character type may alias any other
5262 Pay special attention to code like this:
5275 The practice of reading from a different union member than the one most
5276 recently written to (called ``type-punning'') is common. Even with
5277 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
5278 is accessed through the union type. So, the code above will work as
5279 expected. However, this code might not:
5290 Every language that wishes to perform language-specific alias analysis
5291 should define a function that computes, given an @code{tree}
5292 node, an alias set for the node. Nodes in different alias sets are not
5293 allowed to alias. For an example, see the C front-end function
5294 @code{c_get_alias_set}.
5296 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5298 @item -falign-functions
5299 @itemx -falign-functions=@var{n}
5300 @opindex falign-functions
5301 Align the start of functions to the next power-of-two greater than
5302 @var{n}, skipping up to @var{n} bytes. For instance,
5303 @option{-falign-functions=32} aligns functions to the next 32-byte
5304 boundary, but @option{-falign-functions=24} would align to the next
5305 32-byte boundary only if this can be done by skipping 23 bytes or less.
5307 @option{-fno-align-functions} and @option{-falign-functions=1} are
5308 equivalent and mean that functions will not be aligned.
5310 Some assemblers only support this flag when @var{n} is a power of two;
5311 in that case, it is rounded up.
5313 If @var{n} is not specified or is zero, use a machine-dependent default.
5315 Enabled at levels @option{-O2}, @option{-O3}.
5317 @item -falign-labels
5318 @itemx -falign-labels=@var{n}
5319 @opindex falign-labels
5320 Align all branch targets to a power-of-two boundary, skipping up to
5321 @var{n} bytes like @option{-falign-functions}. This option can easily
5322 make code slower, because it must insert dummy operations for when the
5323 branch target is reached in the usual flow of the code.
5325 @option{-fno-align-labels} and @option{-falign-labels=1} are
5326 equivalent and mean that labels will not be aligned.
5328 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
5329 are greater than this value, then their values are used instead.
5331 If @var{n} is not specified or is zero, use a machine-dependent default
5332 which is very likely to be @samp{1}, meaning no alignment.
5334 Enabled at levels @option{-O2}, @option{-O3}.
5337 @itemx -falign-loops=@var{n}
5338 @opindex falign-loops
5339 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
5340 like @option{-falign-functions}. The hope is that the loop will be
5341 executed many times, which will make up for any execution of the dummy
5344 @option{-fno-align-loops} and @option{-falign-loops=1} are
5345 equivalent and mean that loops will not be aligned.
5347 If @var{n} is not specified or is zero, use a machine-dependent default.
5349 Enabled at levels @option{-O2}, @option{-O3}.
5352 @itemx -falign-jumps=@var{n}
5353 @opindex falign-jumps
5354 Align branch targets to a power-of-two boundary, for branch targets
5355 where the targets can only be reached by jumping, skipping up to @var{n}
5356 bytes like @option{-falign-functions}. In this case, no dummy operations
5359 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
5360 equivalent and mean that loops will not be aligned.
5362 If @var{n} is not specified or is zero, use a machine-dependent default.
5364 Enabled at levels @option{-O2}, @option{-O3}.
5366 @item -funit-at-a-time
5367 @opindex funit-at-a-time
5368 Parse the whole compilation unit before starting to produce code.
5369 This allows some extra optimizations to take place but consumes
5370 more memory (in general). There are some compatibility issues
5371 with @emph{unit-at-a-time} mode:
5374 enabling @emph{unit-at-a-time} mode may change the order
5375 in which functions, variables, and top-level @code{asm} statements
5376 are emitted, and will likely break code relying on some particular
5377 ordering. The majority of such top-level @code{asm} statements,
5378 though, can be replaced by @code{section} attributes. The
5379 @option{fno-toplevel-reorder} option may be used to keep the ordering
5380 used in the input file, at the cost of some optimizations.
5383 @emph{unit-at-a-time} mode removes unreferenced static variables
5384 and functions. This may result in undefined references
5385 when an @code{asm} statement refers directly to variables or functions
5386 that are otherwise unused. In that case either the variable/function
5387 shall be listed as an operand of the @code{asm} statement operand or,
5388 in the case of top-level @code{asm} statements the attribute @code{used}
5389 shall be used on the declaration.
5392 Static functions now can use non-standard passing conventions that
5393 may break @code{asm} statements calling functions directly. Again,
5394 attribute @code{used} will prevent this behavior.
5397 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
5398 but this scheme may not be supported by future releases of GCC@.
5400 Enabled at levels @option{-O2}, @option{-O3}.
5402 @item -fno-toplevel-reorder
5403 Do not reorder top-level functions, variables, and @code{asm}
5404 statements. Output them in the same order that they appear in the
5405 input file. When this option is used, unreferenced static variables
5406 will not be removed. This option is intended to support existing code
5407 which relies on a particular ordering. For new code, it is better to
5412 Constructs webs as commonly used for register allocation purposes and assign
5413 each web individual pseudo register. This allows the register allocation pass
5414 to operate on pseudos directly, but also strengthens several other optimization
5415 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
5416 however, make debugging impossible, since variables will no longer stay in a
5419 Enabled by default with @option{-funroll-loops}.
5421 @item -fwhole-program
5422 @opindex fwhole-program
5423 Assume that the current compilation unit represents whole program being
5424 compiled. All public functions and variables with the exception of @code{main}
5425 and those merged by attribute @code{externally_visible} become static functions
5426 and in a affect gets more aggressively optimized by interprocedural optimizers.
5427 While this option is equivalent to proper use of @code{static} keyword for
5428 programs consisting of single file, in combination with option
5429 @option{--combine} this flag can be used to compile most of smaller scale C
5430 programs since the functions and variables become local for the whole combined
5431 compilation unit, not for the single source file itself.
5434 @item -fno-cprop-registers
5435 @opindex fno-cprop-registers
5436 After register allocation and post-register allocation instruction splitting,
5437 we perform a copy-propagation pass to try to reduce scheduling dependencies
5438 and occasionally eliminate the copy.
5440 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5442 @item -fprofile-generate
5443 @opindex fprofile-generate
5445 Enable options usually used for instrumenting application to produce
5446 profile useful for later recompilation with profile feedback based
5447 optimization. You must use @option{-fprofile-generate} both when
5448 compiling and when linking your program.
5450 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
5453 @opindex fprofile-use
5454 Enable profile feedback directed optimizations, and optimizations
5455 generally profitable only with profile feedback available.
5457 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
5458 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer},
5459 @code{-fno-loop-optimize}.
5463 The following options control compiler behavior regarding floating
5464 point arithmetic. These options trade off between speed and
5465 correctness. All must be specifically enabled.
5469 @opindex ffloat-store
5470 Do not store floating point variables in registers, and inhibit other
5471 options that might change whether a floating point value is taken from a
5474 @cindex floating point precision
5475 This option prevents undesirable excess precision on machines such as
5476 the 68000 where the floating registers (of the 68881) keep more
5477 precision than a @code{double} is supposed to have. Similarly for the
5478 x86 architecture. For most programs, the excess precision does only
5479 good, but a few programs rely on the precise definition of IEEE floating
5480 point. Use @option{-ffloat-store} for such programs, after modifying
5481 them to store all pertinent intermediate computations into variables.
5485 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
5486 @option{-fno-trapping-math}, @option{-ffinite-math-only},
5487 @option{-fno-rounding-math}, @option{-fno-signaling-nans}
5488 and @option{fcx-limited-range}.
5490 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
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 for
5497 @item -fno-math-errno
5498 @opindex fno-math-errno
5499 Do not set ERRNO after calling math functions that are executed
5500 with a single instruction, e.g., sqrt. A program that relies on
5501 IEEE exceptions for math error handling may want to use this flag
5502 for speed while maintaining IEEE arithmetic compatibility.
5504 This option should never be turned on by any @option{-O} option since
5505 it can result in incorrect output for programs which depend on
5506 an exact implementation of IEEE or ISO rules/specifications for
5509 The default is @option{-fmath-errno}.
5511 On Darwin systems, the math library never sets @code{errno}. There is therefore
5512 no reason for the compiler to consider the possibility that it might,
5513 and @option{-fno-math-errno} is the default.
5515 @item -funsafe-math-optimizations
5516 @opindex funsafe-math-optimizations
5517 Allow optimizations for floating-point arithmetic that (a) assume
5518 that arguments and results are valid and (b) may violate IEEE or
5519 ANSI standards. When used at link-time, it may include libraries
5520 or startup files that change the default FPU control word or other
5521 similar optimizations.
5523 This option should never be turned on by any @option{-O} option since
5524 it can result in incorrect output for programs which depend on
5525 an exact implementation of IEEE or ISO rules/specifications for
5528 The default is @option{-fno-unsafe-math-optimizations}.
5530 @item -ffinite-math-only
5531 @opindex ffinite-math-only
5532 Allow optimizations for floating-point arithmetic that assume
5533 that arguments and results are not NaNs or +-Infs.
5535 This option should never be turned on by any @option{-O} option since
5536 it can result in incorrect output for programs which depend on
5537 an exact implementation of IEEE or ISO rules/specifications.
5539 The default is @option{-fno-finite-math-only}.
5541 @item -fno-trapping-math
5542 @opindex fno-trapping-math
5543 Compile code assuming that floating-point operations cannot generate
5544 user-visible traps. These traps include division by zero, overflow,
5545 underflow, inexact result and invalid operation. This option implies
5546 @option{-fno-signaling-nans}. Setting this option may allow faster
5547 code if one relies on ``non-stop'' IEEE arithmetic, for example.
5549 This option should never be turned on by any @option{-O} option since
5550 it can result in incorrect output for programs which depend on
5551 an exact implementation of IEEE or ISO rules/specifications for
5554 The default is @option{-ftrapping-math}.
5556 @item -frounding-math
5557 @opindex frounding-math
5558 Disable transformations and optimizations that assume default floating
5559 point rounding behavior. This is round-to-zero for all floating point
5560 to integer conversions, and round-to-nearest for all other arithmetic
5561 truncations. This option should be specified for programs that change
5562 the FP rounding mode dynamically, or that may be executed with a
5563 non-default rounding mode. This option disables constant folding of
5564 floating point expressions at compile-time (which may be affected by
5565 rounding mode) and arithmetic transformations that are unsafe in the
5566 presence of sign-dependent rounding modes.
5568 The default is @option{-fno-rounding-math}.
5570 This option is experimental and does not currently guarantee to
5571 disable all GCC optimizations that are affected by rounding mode.
5572 Future versions of GCC may provide finer control of this setting
5573 using C99's @code{FENV_ACCESS} pragma. This command line option
5574 will be used to specify the default state for @code{FENV_ACCESS}.
5576 @item -frtl-abstract-sequences
5577 @opindex frtl-abstract-sequences
5578 It is a size optimization method. This option is to find identical
5579 sequences of code, which can be turned into pseudo-procedures and
5580 then replace all occurrences with calls to the newly created
5581 subroutine. It is kind of an opposite of @option{-finline-functions}.
5582 This optimization runs at RTL level.
5584 @item -fsignaling-nans
5585 @opindex fsignaling-nans
5586 Compile code assuming that IEEE signaling NaNs may generate user-visible
5587 traps during floating-point operations. Setting this option disables
5588 optimizations that may change the number of exceptions visible with
5589 signaling NaNs. This option implies @option{-ftrapping-math}.
5591 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
5594 The default is @option{-fno-signaling-nans}.
5596 This option is experimental and does not currently guarantee to
5597 disable all GCC optimizations that affect signaling NaN behavior.
5599 @item -fsingle-precision-constant
5600 @opindex fsingle-precision-constant
5601 Treat floating point constant as single precision constant instead of
5602 implicitly converting it to double precision constant.
5604 @item -fcx-limited-range
5605 @itemx -fno-cx-limited-range
5606 @opindex fcx-limited-range
5607 @opindex fno-cx-limited-range
5608 When enabled, this option states that a range reduction step is not
5609 needed when performing complex division. The default is
5610 @option{-fno-cx-limited-range}, but is enabled by @option{-ffast-math}.
5612 This option controls the default setting of the ISO C99
5613 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
5618 The following options control optimizations that may improve
5619 performance, but are not enabled by any @option{-O} options. This
5620 section includes experimental options that may produce broken code.
5623 @item -fbranch-probabilities
5624 @opindex fbranch-probabilities
5625 After running a program compiled with @option{-fprofile-arcs}
5626 (@pxref{Debugging Options,, Options for Debugging Your Program or
5627 @command{gcc}}), you can compile it a second time using
5628 @option{-fbranch-probabilities}, to improve optimizations based on
5629 the number of times each branch was taken. When the program
5630 compiled with @option{-fprofile-arcs} exits it saves arc execution
5631 counts to a file called @file{@var{sourcename}.gcda} for each source
5632 file The information in this data file is very dependent on the
5633 structure of the generated code, so you must use the same source code
5634 and the same optimization options for both compilations.
5636 With @option{-fbranch-probabilities}, GCC puts a
5637 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
5638 These can be used to improve optimization. Currently, they are only
5639 used in one place: in @file{reorg.c}, instead of guessing which path a
5640 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
5641 exactly determine which path is taken more often.
5643 @item -fprofile-values
5644 @opindex fprofile-values
5645 If combined with @option{-fprofile-arcs}, it adds code so that some
5646 data about values of expressions in the program is gathered.
5648 With @option{-fbranch-probabilities}, it reads back the data gathered
5649 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
5650 notes to instructions for their later usage in optimizations.
5652 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
5656 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
5657 a code to gather information about values of expressions.
5659 With @option{-fbranch-probabilities}, it reads back the data gathered
5660 and actually performs the optimizations based on them.
5661 Currently the optimizations include specialization of division operation
5662 using the knowledge about the value of the denominator.
5664 @item -frename-registers
5665 @opindex frename-registers
5666 Attempt to avoid false dependencies in scheduled code by making use
5667 of registers left over after register allocation. This optimization
5668 will most benefit processors with lots of registers. Depending on the
5669 debug information format adopted by the target, however, it can
5670 make debugging impossible, since variables will no longer stay in
5671 a ``home register''.
5673 Enabled by default with @option{-funroll-loops}.
5677 Perform tail duplication to enlarge superblock size. This transformation
5678 simplifies the control flow of the function allowing other optimizations to do
5681 Enabled with @option{-fprofile-use}.
5683 @item -funroll-loops
5684 @opindex funroll-loops
5685 Unroll loops whose number of iterations can be determined at compile time or
5686 upon entry to the loop. @option{-funroll-loops} implies
5687 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
5688 It also turns on complete loop peeling (i.e.@: complete removal of loops with
5689 small constant number of iterations). This option makes code larger, and may
5690 or may not make it run faster.
5692 Enabled with @option{-fprofile-use}.
5694 @item -funroll-all-loops
5695 @opindex funroll-all-loops
5696 Unroll all loops, even if their number of iterations is uncertain when
5697 the loop is entered. This usually makes programs run more slowly.
5698 @option{-funroll-all-loops} implies the same options as
5699 @option{-funroll-loops}.
5702 @opindex fpeel-loops
5703 Peels the loops for that there is enough information that they do not
5704 roll much (from profile feedback). It also turns on complete loop peeling
5705 (i.e.@: complete removal of loops with small constant number of iterations).
5707 Enabled with @option{-fprofile-use}.
5709 @item -fmove-loop-invariants
5710 @opindex fmove-loop-invariants
5711 Enables the loop invariant motion pass in the new loop optimizer. Enabled
5712 at level @option{-O1}
5714 @item -funswitch-loops
5715 @opindex funswitch-loops
5716 Move branches with loop invariant conditions out of the loop, with duplicates
5717 of the loop on both branches (modified according to result of the condition).
5719 @item -fprefetch-loop-arrays
5720 @itemx -fprefetch-loop-arrays-rtl
5721 @opindex fprefetch-loop-arrays
5722 @opindex fprefetch-loop-arrays-rtl
5723 If supported by the target machine, generate instructions to prefetch
5724 memory to improve the performance of loops that access large arrays.
5726 Disabled at level @option{-Os}.
5728 @item -ffunction-sections
5729 @itemx -fdata-sections
5730 @opindex ffunction-sections
5731 @opindex fdata-sections
5732 Place each function or data item into its own section in the output
5733 file if the target supports arbitrary sections. The name of the
5734 function or the name of the data item determines the section's name
5737 Use these options on systems where the linker can perform optimizations
5738 to improve locality of reference in the instruction space. Most systems
5739 using the ELF object format and SPARC processors running Solaris 2 have
5740 linkers with such optimizations. AIX may have these optimizations in
5743 Only use these options when there are significant benefits from doing
5744 so. When you specify these options, the assembler and linker will
5745 create larger object and executable files and will also be slower.
5746 You will not be able to use @code{gprof} on all systems if you
5747 specify this option and you may have problems with debugging if
5748 you specify both this option and @option{-g}.
5750 @item -fbranch-target-load-optimize
5751 @opindex fbranch-target-load-optimize
5752 Perform branch target register load optimization before prologue / epilogue
5754 The use of target registers can typically be exposed only during reload,
5755 thus hoisting loads out of loops and doing inter-block scheduling needs
5756 a separate optimization pass.
5758 @item -fbranch-target-load-optimize2
5759 @opindex fbranch-target-load-optimize2
5760 Perform branch target register load optimization after prologue / epilogue
5763 @item -fbtr-bb-exclusive
5764 @opindex fbtr-bb-exclusive
5765 When performing branch target register load optimization, don't reuse
5766 branch target registers in within any basic block.
5768 @item -fstack-protector
5769 Emit extra code to check for buffer overflows, such as stack smashing
5770 attacks. This is done by adding a guard variable to functions with
5771 vulnerable objects. This includes functions that call alloca, and
5772 functions with buffers larger than 8 bytes. The guards are initialized
5773 when a function is entered and then checked when the function exits.
5774 If a guard check fails, an error message is printed and the program exits.
5776 @item -fstack-protector-all
5777 Like @option{-fstack-protector} except that all functions are protected.
5779 @item -fsection-anchors
5780 @opindex fsection-anchors
5781 Try to reduce the number of symbolic address calculations by using
5782 shared ``anchor'' symbols to address nearby objects. This transformation
5783 can help to reduce the number of GOT entries and GOT accesses on some
5786 For example, the implementation of the following function @code{foo}:
5790 int foo (void) @{ return a + b + c; @}
5793 would usually calculate the addresses of all three variables, but if you
5794 compile it with @option{-fsection-anchors}, it will access the variables
5795 from a common anchor point instead. The effect is similar to the
5796 following pseudocode (which isn't valid C):
5801 register int *xr = &x;
5802 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
5806 Not all targets support this option.
5808 @item --param @var{name}=@var{value}
5810 In some places, GCC uses various constants to control the amount of
5811 optimization that is done. For example, GCC will not inline functions
5812 that contain more that a certain number of instructions. You can
5813 control some of these constants on the command-line using the
5814 @option{--param} option.
5816 The names of specific parameters, and the meaning of the values, are
5817 tied to the internals of the compiler, and are subject to change
5818 without notice in future releases.
5820 In each case, the @var{value} is an integer. The allowable choices for
5821 @var{name} are given in the following table:
5824 @item salias-max-implicit-fields
5825 The maximum number of fields in a variable without direct
5826 structure accesses for which structure aliasing will consider trying
5827 to track each field. The default is 5
5829 @item salias-max-array-elements
5830 The maximum number of elements an array can have and its elements
5831 still be tracked individually by structure aliasing. The default is 4
5833 @item sra-max-structure-size
5834 The maximum structure size, in bytes, at which the scalar replacement
5835 of aggregates (SRA) optimization will perform block copies. The
5836 default value, 0, implies that GCC will select the most appropriate
5839 @item sra-field-structure-ratio
5840 The threshold ratio (as a percentage) between instantiated fields and
5841 the complete structure size. We say that if the ratio of the number
5842 of bytes in instantiated fields to the number of bytes in the complete
5843 structure exceeds this parameter, then block copies are not used. The
5846 @item max-crossjump-edges
5847 The maximum number of incoming edges to consider for crossjumping.
5848 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
5849 the number of edges incoming to each block. Increasing values mean
5850 more aggressive optimization, making the compile time increase with
5851 probably small improvement in executable size.
5853 @item min-crossjump-insns
5854 The minimum number of instructions which must be matched at the end
5855 of two blocks before crossjumping will be performed on them. This
5856 value is ignored in the case where all instructions in the block being
5857 crossjumped from are matched. The default value is 5.
5859 @item max-grow-copy-bb-insns
5860 The maximum code size expansion factor when copying basic blocks
5861 instead of jumping. The expansion is relative to a jump instruction.
5862 The default value is 8.
5864 @item max-goto-duplication-insns
5865 The maximum number of instructions to duplicate to a block that jumps
5866 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
5867 passes, GCC factors computed gotos early in the compilation process,
5868 and unfactors them as late as possible. Only computed jumps at the
5869 end of a basic blocks with no more than max-goto-duplication-insns are
5870 unfactored. The default value is 8.
5872 @item max-delay-slot-insn-search
5873 The maximum number of instructions to consider when looking for an
5874 instruction to fill a delay slot. If more than this arbitrary number of
5875 instructions is searched, the time savings from filling the delay slot
5876 will be minimal so stop searching. Increasing values mean more
5877 aggressive optimization, making the compile time increase with probably
5878 small improvement in executable run time.
5880 @item max-delay-slot-live-search
5881 When trying to fill delay slots, the maximum number of instructions to
5882 consider when searching for a block with valid live register
5883 information. Increasing this arbitrarily chosen value means more
5884 aggressive optimization, increasing the compile time. This parameter
5885 should be removed when the delay slot code is rewritten to maintain the
5888 @item max-gcse-memory
5889 The approximate maximum amount of memory that will be allocated in
5890 order to perform the global common subexpression elimination
5891 optimization. If more memory than specified is required, the
5892 optimization will not be done.
5894 @item max-gcse-passes
5895 The maximum number of passes of GCSE to run. The default is 1.
5897 @item max-pending-list-length
5898 The maximum number of pending dependencies scheduling will allow
5899 before flushing the current state and starting over. Large functions
5900 with few branches or calls can create excessively large lists which
5901 needlessly consume memory and resources.
5903 @item max-inline-insns-single
5904 Several parameters control the tree inliner used in gcc.
5905 This number sets the maximum number of instructions (counted in GCC's
5906 internal representation) in a single function that the tree inliner
5907 will consider for inlining. This only affects functions declared
5908 inline and methods implemented in a class declaration (C++).
5909 The default value is 450.
5911 @item max-inline-insns-auto
5912 When you use @option{-finline-functions} (included in @option{-O3}),
5913 a lot of functions that would otherwise not be considered for inlining
5914 by the compiler will be investigated. To those functions, a different
5915 (more restrictive) limit compared to functions declared inline can
5917 The default value is 90.
5919 @item large-function-insns
5920 The limit specifying really large functions. For functions larger than this
5921 limit after inlining inlining is constrained by
5922 @option{--param large-function-growth}. This parameter is useful primarily
5923 to avoid extreme compilation time caused by non-linear algorithms used by the
5925 This parameter is ignored when @option{-funit-at-a-time} is not used.
5926 The default value is 2700.
5928 @item large-function-growth
5929 Specifies maximal growth of large function caused by inlining in percents.
5930 This parameter is ignored when @option{-funit-at-a-time} is not used.
5931 The default value is 100 which limits large function growth to 2.0 times
5934 @item large-unit-insns
5935 The limit specifying large translation unit. Growth caused by inlining of
5936 units larger than this limit is limited by @option{--param inline-unit-growth}.
5937 For small units this might be too tight (consider unit consisting of function A
5938 that is inline and B that just calls A three time. If B is small relative to
5939 A, the growth of unit is 300\% and yet such inlining is very sane. For very
5940 large units consisting of small inlininable functions however the overall unit
5941 growth limit is needed to avoid exponential explosion of code size. Thus for
5942 smaller units, the size is increased to @option{--param large-unit-insns}
5943 before applying @option{--param inline-unit-growth}. The default is 10000
5945 @item inline-unit-growth
5946 Specifies maximal overall growth of the compilation unit caused by inlining.
5947 This parameter is ignored when @option{-funit-at-a-time} is not used.
5948 The default value is 50 which limits unit growth to 1.5 times the original
5951 @item max-inline-insns-recursive
5952 @itemx max-inline-insns-recursive-auto
5953 Specifies maximum number of instructions out-of-line copy of self recursive inline
5954 function can grow into by performing recursive inlining.
5956 For functions declared inline @option{--param max-inline-insns-recursive} is
5957 taken into acount. For function not declared inline, recursive inlining
5958 happens only when @option{-finline-functions} (included in @option{-O3}) is
5959 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
5960 default value is 450.
5962 @item max-inline-recursive-depth
5963 @itemx max-inline-recursive-depth-auto
5964 Specifies maximum recursion depth used by the recursive inlining.
5966 For functions declared inline @option{--param max-inline-recursive-depth} is
5967 taken into acount. For function not declared inline, recursive inlining
5968 happens only when @option{-finline-functions} (included in @option{-O3}) is
5969 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
5970 default value is 450.
5972 @item min-inline-recursive-probability
5973 Recursive inlining is profitable only for function having deep recursion
5974 in average and can hurt for function having little recursion depth by
5975 increasing the prologue size or complexity of function body to other
5978 When profile feedback is available (see @option{-fprofile-generate}) the actual
5979 recursion depth can be guessed from probability that function will recurse via
5980 given call expression. This parameter limits inlining only to call expression
5981 whose probability exceeds given threshold (in percents). The default value is
5984 @item inline-call-cost
5985 Specify cost of call instruction relative to simple arithmetics operations
5986 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
5987 functions and at the same time increases size of leaf function that is believed to
5988 reduce function size by being inlined. In effect it increases amount of
5989 inlining for code having large abstraction penalty (many functions that just
5990 pass the arguments to other functions) and decrease inlining for code with low
5991 abstraction penalty. The default value is 16.
5993 @item max-unrolled-insns
5994 The maximum number of instructions that a loop should have if that loop
5995 is unrolled, and if the loop is unrolled, it determines how many times
5996 the loop code is unrolled.
5998 @item max-average-unrolled-insns
5999 The maximum number of instructions biased by probabilities of their execution
6000 that a loop should have if that loop is unrolled, and if the loop is unrolled,
6001 it determines how many times the loop code is unrolled.
6003 @item max-unroll-times
6004 The maximum number of unrollings of a single loop.
6006 @item max-peeled-insns
6007 The maximum number of instructions that a loop should have if that loop
6008 is peeled, and if the loop is peeled, it determines how many times
6009 the loop code is peeled.
6011 @item max-peel-times
6012 The maximum number of peelings of a single loop.
6014 @item max-completely-peeled-insns
6015 The maximum number of insns of a completely peeled loop.
6017 @item max-completely-peel-times
6018 The maximum number of iterations of a loop to be suitable for complete peeling.
6020 @item max-unswitch-insns
6021 The maximum number of insns of an unswitched loop.
6023 @item max-unswitch-level
6024 The maximum number of branches unswitched in a single loop.
6027 The minimum cost of an expensive expression in the loop invariant motion.
6029 @item iv-consider-all-candidates-bound
6030 Bound on number of candidates for induction variables below that
6031 all candidates are considered for each use in induction variable
6032 optimizations. Only the most relevant candidates are considered
6033 if there are more candidates, to avoid quadratic time complexity.
6035 @item iv-max-considered-uses
6036 The induction variable optimizations give up on loops that contain more
6037 induction variable uses.
6039 @item iv-always-prune-cand-set-bound
6040 If number of candidates in the set is smaller than this value,
6041 we always try to remove unnecessary ivs from the set during its
6042 optimization when a new iv is added to the set.
6044 @item scev-max-expr-size
6045 Bound on size of expressions used in the scalar evolutions analyzer.
6046 Large expressions slow the analyzer.
6048 @item vect-max-version-checks
6049 The maximum number of runtime checks that can be performed when doing
6050 loop versioning in the vectorizer. See option ftree-vect-loop-version
6051 for more information.
6053 @item max-iterations-to-track
6055 The maximum number of iterations of a loop the brute force algorithm
6056 for analysis of # of iterations of the loop tries to evaluate.
6058 @item hot-bb-count-fraction
6059 Select fraction of the maximal count of repetitions of basic block in program
6060 given basic block needs to have to be considered hot.
6062 @item hot-bb-frequency-fraction
6063 Select fraction of the maximal frequency of executions of basic block in
6064 function given basic block needs to have to be considered hot
6066 @item max-predicted-iterations
6067 The maximum number of loop iterations we predict statically. This is useful
6068 in cases where function contain single loop with known bound and other loop
6069 with unknown. We predict the known number of iterations correctly, while
6070 the unknown number of iterations average to roughly 10. This means that the
6071 loop without bounds would appear artificially cold relative to the other one.
6073 @item tracer-dynamic-coverage
6074 @itemx tracer-dynamic-coverage-feedback
6076 This value is used to limit superblock formation once the given percentage of
6077 executed instructions is covered. This limits unnecessary code size
6080 The @option{tracer-dynamic-coverage-feedback} is used only when profile
6081 feedback is available. The real profiles (as opposed to statically estimated
6082 ones) are much less balanced allowing the threshold to be larger value.
6084 @item tracer-max-code-growth
6085 Stop tail duplication once code growth has reached given percentage. This is
6086 rather hokey argument, as most of the duplicates will be eliminated later in
6087 cross jumping, so it may be set to much higher values than is the desired code
6090 @item tracer-min-branch-ratio
6092 Stop reverse growth when the reverse probability of best edge is less than this
6093 threshold (in percent).
6095 @item tracer-min-branch-ratio
6096 @itemx tracer-min-branch-ratio-feedback
6098 Stop forward growth if the best edge do have probability lower than this
6101 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
6102 compilation for profile feedback and one for compilation without. The value
6103 for compilation with profile feedback needs to be more conservative (higher) in
6104 order to make tracer effective.
6106 @item max-cse-path-length
6108 Maximum number of basic blocks on path that cse considers. The default is 10.
6111 The maximum instructions CSE process before flushing. The default is 1000.
6113 @item global-var-threshold
6115 Counts the number of function calls (@var{n}) and the number of
6116 call-clobbered variables (@var{v}). If @var{n}x@var{v} is larger than this limit, a
6117 single artificial variable will be created to represent all the
6118 call-clobbered variables at function call sites. This artificial
6119 variable will then be made to alias every call-clobbered variable.
6120 (done as @code{int * size_t} on the host machine; beware overflow).
6122 @item max-aliased-vops
6124 Maximum number of virtual operands allowed to represent aliases
6125 before triggering the alias grouping heuristic. Alias grouping
6126 reduces compile times and memory consumption needed for aliasing at
6127 the expense of precision loss in alias information.
6129 @item ggc-min-expand
6131 GCC uses a garbage collector to manage its own memory allocation. This
6132 parameter specifies the minimum percentage by which the garbage
6133 collector's heap should be allowed to expand between collections.
6134 Tuning this may improve compilation speed; it has no effect on code
6137 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
6138 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
6139 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
6140 GCC is not able to calculate RAM on a particular platform, the lower
6141 bound of 30% is used. Setting this parameter and
6142 @option{ggc-min-heapsize} to zero causes a full collection to occur at
6143 every opportunity. This is extremely slow, but can be useful for
6146 @item ggc-min-heapsize
6148 Minimum size of the garbage collector's heap before it begins bothering
6149 to collect garbage. The first collection occurs after the heap expands
6150 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
6151 tuning this may improve compilation speed, and has no effect on code
6154 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
6155 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
6156 with a lower bound of 4096 (four megabytes) and an upper bound of
6157 131072 (128 megabytes). If GCC is not able to calculate RAM on a
6158 particular platform, the lower bound is used. Setting this parameter
6159 very large effectively disables garbage collection. Setting this
6160 parameter and @option{ggc-min-expand} to zero causes a full collection
6161 to occur at every opportunity.
6163 @item max-reload-search-insns
6164 The maximum number of instruction reload should look backward for equivalent
6165 register. Increasing values mean more aggressive optimization, making the
6166 compile time increase with probably slightly better performance. The default
6169 @item max-cselib-memory-location
6170 The maximum number of memory locations cselib should take into acount.
6171 Increasing values mean more aggressive optimization, making the compile time
6172 increase with probably slightly better performance. The default value is 500.
6174 @item max-flow-memory-location
6175 Similar as @option{max-cselib-memory-location} but for dataflow liveness.
6176 The default value is 100.
6178 @item reorder-blocks-duplicate
6179 @itemx reorder-blocks-duplicate-feedback
6181 Used by basic block reordering pass to decide whether to use unconditional
6182 branch or duplicate the code on its destination. Code is duplicated when its
6183 estimated size is smaller than this value multiplied by the estimated size of
6184 unconditional jump in the hot spots of the program.
6186 The @option{reorder-block-duplicate-feedback} is used only when profile
6187 feedback is available and may be set to higher values than
6188 @option{reorder-block-duplicate} since information about the hot spots is more
6191 @item max-sched-region-blocks
6192 The maximum number of blocks in a region to be considered for
6193 interblock scheduling. The default value is 10.
6195 @item max-sched-region-insns
6196 The maximum number of insns in a region to be considered for
6197 interblock scheduling. The default value is 100.
6200 The minimum probability (in percents) of reaching a source block
6201 for interblock speculative scheduling. The default value is 40.
6203 @item max-last-value-rtl
6205 The maximum size measured as number of RTLs that can be recorded in an expression
6206 in combiner for a pseudo register as last known value of that register. The default
6209 @item integer-share-limit
6210 Small integer constants can use a shared data structure, reducing the
6211 compiler's memory usage and increasing its speed. This sets the maximum
6212 value of a shared integer constant's. The default value is 256.
6214 @item min-virtual-mappings
6215 Specifies the minimum number of virtual mappings in the incremental
6216 SSA updater that should be registered to trigger the virtual mappings
6217 heuristic defined by virtual-mappings-ratio. The default value is
6220 @item virtual-mappings-ratio
6221 If the number of virtual mappings is virtual-mappings-ratio bigger
6222 than the number of virtual symbols to be updated, then the incremental
6223 SSA updater switches to a full update for those symbols. The default
6226 @item ssp-buffer-size
6227 The minimum size of buffers (i.e. arrays) that will receive stack smashing
6228 protection when @option{-fstack-protection} is used.
6230 @item max-jump-thread-duplication-stmts
6231 Maximum number of statements allowed in a block that needs to be
6232 duplicated when threading jumps.
6234 @item max-fields-for-field-sensitive
6235 Maximum number of fields in a structure we will treat in
6236 a field sensitive manner during pointer analysis.
6241 @node Preprocessor Options
6242 @section Options Controlling the Preprocessor
6243 @cindex preprocessor options
6244 @cindex options, preprocessor
6246 These options control the C preprocessor, which is run on each C source
6247 file before actual compilation.
6249 If you use the @option{-E} option, nothing is done except preprocessing.
6250 Some of these options make sense only together with @option{-E} because
6251 they cause the preprocessor output to be unsuitable for actual
6256 You can use @option{-Wp,@var{option}} to bypass the compiler driver
6257 and pass @var{option} directly through to the preprocessor. If
6258 @var{option} contains commas, it is split into multiple options at the
6259 commas. However, many options are modified, translated or interpreted
6260 by the compiler driver before being passed to the preprocessor, and
6261 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
6262 interface is undocumented and subject to change, so whenever possible
6263 you should avoid using @option{-Wp} and let the driver handle the
6266 @item -Xpreprocessor @var{option}
6267 @opindex preprocessor
6268 Pass @var{option} as an option to the preprocessor. You can use this to
6269 supply system-specific preprocessor options which GCC does not know how to
6272 If you want to pass an option that takes an argument, you must use
6273 @option{-Xpreprocessor} twice, once for the option and once for the argument.
6276 @include cppopts.texi
6278 @node Assembler Options
6279 @section Passing Options to the Assembler
6281 @c prevent bad page break with this line
6282 You can pass options to the assembler.
6285 @item -Wa,@var{option}
6287 Pass @var{option} as an option to the assembler. If @var{option}
6288 contains commas, it is split into multiple options at the commas.
6290 @item -Xassembler @var{option}
6292 Pass @var{option} as an option to the assembler. You can use this to
6293 supply system-specific assembler options which GCC does not know how to
6296 If you want to pass an option that takes an argument, you must use
6297 @option{-Xassembler} twice, once for the option and once for the argument.
6302 @section Options for Linking
6303 @cindex link options
6304 @cindex options, linking
6306 These options come into play when the compiler links object files into
6307 an executable output file. They are meaningless if the compiler is
6308 not doing a link step.
6312 @item @var{object-file-name}
6313 A file name that does not end in a special recognized suffix is
6314 considered to name an object file or library. (Object files are
6315 distinguished from libraries by the linker according to the file
6316 contents.) If linking is done, these object files are used as input
6325 If any of these options is used, then the linker is not run, and
6326 object file names should not be used as arguments. @xref{Overall
6330 @item -l@var{library}
6331 @itemx -l @var{library}
6333 Search the library named @var{library} when linking. (The second
6334 alternative with the library as a separate argument is only for
6335 POSIX compliance and is not recommended.)
6337 It makes a difference where in the command you write this option; the
6338 linker searches and processes libraries and object files in the order they
6339 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
6340 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
6341 to functions in @samp{z}, those functions may not be loaded.
6343 The linker searches a standard list of directories for the library,
6344 which is actually a file named @file{lib@var{library}.a}. The linker
6345 then uses this file as if it had been specified precisely by name.
6347 The directories searched include several standard system directories
6348 plus any that you specify with @option{-L}.
6350 Normally the files found this way are library files---archive files
6351 whose members are object files. The linker handles an archive file by
6352 scanning through it for members which define symbols that have so far
6353 been referenced but not defined. But if the file that is found is an
6354 ordinary object file, it is linked in the usual fashion. The only
6355 difference between using an @option{-l} option and specifying a file name
6356 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
6357 and searches several directories.
6361 You need this special case of the @option{-l} option in order to
6362 link an Objective-C or Objective-C++ program.
6365 @opindex nostartfiles
6366 Do not use the standard system startup files when linking.
6367 The standard system libraries are used normally, unless @option{-nostdlib}
6368 or @option{-nodefaultlibs} is used.
6370 @item -nodefaultlibs
6371 @opindex nodefaultlibs
6372 Do not use the standard system libraries when linking.
6373 Only the libraries you specify will be passed to the linker.
6374 The standard startup files are used normally, unless @option{-nostartfiles}
6375 is used. The compiler may generate calls to @code{memcmp},
6376 @code{memset}, @code{memcpy} and @code{memmove}.
6377 These entries are usually resolved by entries in
6378 libc. These entry points should be supplied through some other
6379 mechanism when this option is specified.
6383 Do not use the standard system startup files or libraries when linking.
6384 No startup files and only the libraries you specify will be passed to
6385 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
6386 @code{memcpy} and @code{memmove}.
6387 These entries are usually resolved by entries in
6388 libc. These entry points should be supplied through some other
6389 mechanism when this option is specified.
6391 @cindex @option{-lgcc}, use with @option{-nostdlib}
6392 @cindex @option{-nostdlib} and unresolved references
6393 @cindex unresolved references and @option{-nostdlib}
6394 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
6395 @cindex @option{-nodefaultlibs} and unresolved references
6396 @cindex unresolved references and @option{-nodefaultlibs}
6397 One of the standard libraries bypassed by @option{-nostdlib} and
6398 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
6399 that GCC uses to overcome shortcomings of particular machines, or special
6400 needs for some languages.
6401 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
6402 Collection (GCC) Internals},
6403 for more discussion of @file{libgcc.a}.)
6404 In most cases, you need @file{libgcc.a} even when you want to avoid
6405 other standard libraries. In other words, when you specify @option{-nostdlib}
6406 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
6407 This ensures that you have no unresolved references to internal GCC
6408 library subroutines. (For example, @samp{__main}, used to ensure C++
6409 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
6410 GNU Compiler Collection (GCC) Internals}.)
6414 Produce a position independent executable on targets which support it.
6415 For predictable results, you must also specify the same set of options
6416 that were used to generate code (@option{-fpie}, @option{-fPIE},
6417 or model suboptions) when you specify this option.
6421 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
6422 that support it. This instructs the linker to add all symbols, not
6423 only used ones, to the dynamic symbol table. This option is needed
6424 for some uses of @code{dlopen} or to allow obtaining backtraces
6425 from within a program.
6429 Remove all symbol table and relocation information from the executable.
6433 On systems that support dynamic linking, this prevents linking with the shared
6434 libraries. On other systems, this option has no effect.
6438 Produce a shared object which can then be linked with other objects to
6439 form an executable. Not all systems support this option. For predictable
6440 results, you must also specify the same set of options that were used to
6441 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
6442 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
6443 needs to build supplementary stub code for constructors to work. On
6444 multi-libbed systems, @samp{gcc -shared} must select the correct support
6445 libraries to link against. Failing to supply the correct flags may lead
6446 to subtle defects. Supplying them in cases where they are not necessary
6449 @item -shared-libgcc
6450 @itemx -static-libgcc
6451 @opindex shared-libgcc
6452 @opindex static-libgcc
6453 On systems that provide @file{libgcc} as a shared library, these options
6454 force the use of either the shared or static version respectively.
6455 If no shared version of @file{libgcc} was built when the compiler was
6456 configured, these options have no effect.
6458 There are several situations in which an application should use the
6459 shared @file{libgcc} instead of the static version. The most common
6460 of these is when the application wishes to throw and catch exceptions
6461 across different shared libraries. In that case, each of the libraries
6462 as well as the application itself should use the shared @file{libgcc}.
6464 Therefore, the G++ and GCJ drivers automatically add
6465 @option{-shared-libgcc} whenever you build a shared library or a main
6466 executable, because C++ and Java programs typically use exceptions, so
6467 this is the right thing to do.
6469 If, instead, you use the GCC driver to create shared libraries, you may
6470 find that they will not always be linked with the shared @file{libgcc}.
6471 If GCC finds, at its configuration time, that you have a non-GNU linker
6472 or a GNU linker that does not support option @option{--eh-frame-hdr},
6473 it will link the shared version of @file{libgcc} into shared libraries
6474 by default. Otherwise, it will take advantage of the linker and optimize
6475 away the linking with the shared version of @file{libgcc}, linking with
6476 the static version of libgcc by default. This allows exceptions to
6477 propagate through such shared libraries, without incurring relocation
6478 costs at library load time.
6480 However, if a library or main executable is supposed to throw or catch
6481 exceptions, you must link it using the G++ or GCJ driver, as appropriate
6482 for the languages used in the program, or using the option
6483 @option{-shared-libgcc}, such that it is linked with the shared
6488 Bind references to global symbols when building a shared object. Warn
6489 about any unresolved references (unless overridden by the link editor
6490 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
6493 @item -Xlinker @var{option}
6495 Pass @var{option} as an option to the linker. You can use this to
6496 supply system-specific linker options which GCC does not know how to
6499 If you want to pass an option that takes an argument, you must use
6500 @option{-Xlinker} twice, once for the option and once for the argument.
6501 For example, to pass @option{-assert definitions}, you must write
6502 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
6503 @option{-Xlinker "-assert definitions"}, because this passes the entire
6504 string as a single argument, which is not what the linker expects.
6506 @item -Wl,@var{option}
6508 Pass @var{option} as an option to the linker. If @var{option} contains
6509 commas, it is split into multiple options at the commas.
6511 @item -u @var{symbol}
6513 Pretend the symbol @var{symbol} is undefined, to force linking of
6514 library modules to define it. You can use @option{-u} multiple times with
6515 different symbols to force loading of additional library modules.
6518 @node Directory Options
6519 @section Options for Directory Search
6520 @cindex directory options
6521 @cindex options, directory search
6524 These options specify directories to search for header files, for
6525 libraries and for parts of the compiler:
6530 Add the directory @var{dir} to the head of the list of directories to be
6531 searched for header files. This can be used to override a system header
6532 file, substituting your own version, since these directories are
6533 searched before the system header file directories. However, you should
6534 not use this option to add directories that contain vendor-supplied
6535 system header files (use @option{-isystem} for that). If you use more than
6536 one @option{-I} option, the directories are scanned in left-to-right
6537 order; the standard system directories come after.
6539 If a standard system include directory, or a directory specified with
6540 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
6541 option will be ignored. The directory will still be searched but as a
6542 system directory at its normal position in the system include chain.
6543 This is to ensure that GCC's procedure to fix buggy system headers and
6544 the ordering for the include_next directive are not inadvertently changed.
6545 If you really need to change the search order for system directories,
6546 use the @option{-nostdinc} and/or @option{-isystem} options.
6548 @item -iquote@var{dir}
6550 Add the directory @var{dir} to the head of the list of directories to
6551 be searched for header files only for the case of @samp{#include
6552 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
6553 otherwise just like @option{-I}.
6557 Add directory @var{dir} to the list of directories to be searched
6560 @item -B@var{prefix}
6562 This option specifies where to find the executables, libraries,
6563 include files, and data files of the compiler itself.
6565 The compiler driver program runs one or more of the subprograms
6566 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
6567 @var{prefix} as a prefix for each program it tries to run, both with and
6568 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
6570 For each subprogram to be run, the compiler driver first tries the
6571 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
6572 was not specified, the driver tries two standard prefixes, which are
6573 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
6574 those results in a file name that is found, the unmodified program
6575 name is searched for using the directories specified in your
6576 @env{PATH} environment variable.
6578 The compiler will check to see if the path provided by the @option{-B}
6579 refers to a directory, and if necessary it will add a directory
6580 separator character at the end of the path.
6582 @option{-B} prefixes that effectively specify directory names also apply
6583 to libraries in the linker, because the compiler translates these
6584 options into @option{-L} options for the linker. They also apply to
6585 includes files in the preprocessor, because the compiler translates these
6586 options into @option{-isystem} options for the preprocessor. In this case,
6587 the compiler appends @samp{include} to the prefix.
6589 The run-time support file @file{libgcc.a} can also be searched for using
6590 the @option{-B} prefix, if needed. If it is not found there, the two
6591 standard prefixes above are tried, and that is all. The file is left
6592 out of the link if it is not found by those means.
6594 Another way to specify a prefix much like the @option{-B} prefix is to use
6595 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
6598 As a special kludge, if the path provided by @option{-B} is
6599 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
6600 9, then it will be replaced by @file{[dir/]include}. This is to help
6601 with boot-strapping the compiler.
6603 @item -specs=@var{file}
6605 Process @var{file} after the compiler reads in the standard @file{specs}
6606 file, in order to override the defaults that the @file{gcc} driver
6607 program uses when determining what switches to pass to @file{cc1},
6608 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
6609 @option{-specs=@var{file}} can be specified on the command line, and they
6610 are processed in order, from left to right.
6612 @item --sysroot=@var{dir}
6614 Use @var{dir} as the logical root directory for headers and libraries.
6615 For example, if the compiler would normally search for headers in
6616 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
6617 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
6619 If you use both this option and the @option{-isysroot} option, then
6620 the @option{--sysroot} option will apply to libraries, but the
6621 @option{-isysroot} option will apply to header files.
6623 The GNU linker (beginning with version 2.16) has the necessary support
6624 for this option. If your linker does not support this option, the
6625 header file aspect of @option{--sysroot} will still work, but the
6626 library aspect will not.
6630 This option has been deprecated. Please use @option{-iquote} instead for
6631 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
6632 Any directories you specify with @option{-I} options before the @option{-I-}
6633 option are searched only for the case of @samp{#include "@var{file}"};
6634 they are not searched for @samp{#include <@var{file}>}.
6636 If additional directories are specified with @option{-I} options after
6637 the @option{-I-}, these directories are searched for all @samp{#include}
6638 directives. (Ordinarily @emph{all} @option{-I} directories are used
6641 In addition, the @option{-I-} option inhibits the use of the current
6642 directory (where the current input file came from) as the first search
6643 directory for @samp{#include "@var{file}"}. There is no way to
6644 override this effect of @option{-I-}. With @option{-I.} you can specify
6645 searching the directory which was current when the compiler was
6646 invoked. That is not exactly the same as what the preprocessor does
6647 by default, but it is often satisfactory.
6649 @option{-I-} does not inhibit the use of the standard system directories
6650 for header files. Thus, @option{-I-} and @option{-nostdinc} are
6657 @section Specifying subprocesses and the switches to pass to them
6660 @command{gcc} is a driver program. It performs its job by invoking a
6661 sequence of other programs to do the work of compiling, assembling and
6662 linking. GCC interprets its command-line parameters and uses these to
6663 deduce which programs it should invoke, and which command-line options
6664 it ought to place on their command lines. This behavior is controlled
6665 by @dfn{spec strings}. In most cases there is one spec string for each
6666 program that GCC can invoke, but a few programs have multiple spec
6667 strings to control their behavior. The spec strings built into GCC can
6668 be overridden by using the @option{-specs=} command-line switch to specify
6671 @dfn{Spec files} are plaintext files that are used to construct spec
6672 strings. They consist of a sequence of directives separated by blank
6673 lines. The type of directive is determined by the first non-whitespace
6674 character on the line and it can be one of the following:
6677 @item %@var{command}
6678 Issues a @var{command} to the spec file processor. The commands that can
6682 @item %include <@var{file}>
6684 Search for @var{file} and insert its text at the current point in the
6687 @item %include_noerr <@var{file}>
6688 @cindex %include_noerr
6689 Just like @samp{%include}, but do not generate an error message if the include
6690 file cannot be found.
6692 @item %rename @var{old_name} @var{new_name}
6694 Rename the spec string @var{old_name} to @var{new_name}.
6698 @item *[@var{spec_name}]:
6699 This tells the compiler to create, override or delete the named spec
6700 string. All lines after this directive up to the next directive or
6701 blank line are considered to be the text for the spec string. If this
6702 results in an empty string then the spec will be deleted. (Or, if the
6703 spec did not exist, then nothing will happened.) Otherwise, if the spec
6704 does not currently exist a new spec will be created. If the spec does
6705 exist then its contents will be overridden by the text of this
6706 directive, unless the first character of that text is the @samp{+}
6707 character, in which case the text will be appended to the spec.
6709 @item [@var{suffix}]:
6710 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
6711 and up to the next directive or blank line are considered to make up the
6712 spec string for the indicated suffix. When the compiler encounters an
6713 input file with the named suffix, it will processes the spec string in
6714 order to work out how to compile that file. For example:
6721 This says that any input file whose name ends in @samp{.ZZ} should be
6722 passed to the program @samp{z-compile}, which should be invoked with the
6723 command-line switch @option{-input} and with the result of performing the
6724 @samp{%i} substitution. (See below.)
6726 As an alternative to providing a spec string, the text that follows a
6727 suffix directive can be one of the following:
6730 @item @@@var{language}
6731 This says that the suffix is an alias for a known @var{language}. This is
6732 similar to using the @option{-x} command-line switch to GCC to specify a
6733 language explicitly. For example:
6740 Says that .ZZ files are, in fact, C++ source files.
6743 This causes an error messages saying:
6746 @var{name} compiler not installed on this system.
6750 GCC already has an extensive list of suffixes built into it.
6751 This directive will add an entry to the end of the list of suffixes, but
6752 since the list is searched from the end backwards, it is effectively
6753 possible to override earlier entries using this technique.
6757 GCC has the following spec strings built into it. Spec files can
6758 override these strings or create their own. Note that individual
6759 targets can also add their own spec strings to this list.
6762 asm Options to pass to the assembler
6763 asm_final Options to pass to the assembler post-processor
6764 cpp Options to pass to the C preprocessor
6765 cc1 Options to pass to the C compiler
6766 cc1plus Options to pass to the C++ compiler
6767 endfile Object files to include at the end of the link
6768 link Options to pass to the linker
6769 lib Libraries to include on the command line to the linker
6770 libgcc Decides which GCC support library to pass to the linker
6771 linker Sets the name of the linker
6772 predefines Defines to be passed to the C preprocessor
6773 signed_char Defines to pass to CPP to say whether @code{char} is signed
6775 startfile Object files to include at the start of the link
6778 Here is a small example of a spec file:
6784 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
6787 This example renames the spec called @samp{lib} to @samp{old_lib} and
6788 then overrides the previous definition of @samp{lib} with a new one.
6789 The new definition adds in some extra command-line options before
6790 including the text of the old definition.
6792 @dfn{Spec strings} are a list of command-line options to be passed to their
6793 corresponding program. In addition, the spec strings can contain
6794 @samp{%}-prefixed sequences to substitute variable text or to
6795 conditionally insert text into the command line. Using these constructs
6796 it is possible to generate quite complex command lines.
6798 Here is a table of all defined @samp{%}-sequences for spec
6799 strings. Note that spaces are not generated automatically around the
6800 results of expanding these sequences. Therefore you can concatenate them
6801 together or combine them with constant text in a single argument.
6805 Substitute one @samp{%} into the program name or argument.
6808 Substitute the name of the input file being processed.
6811 Substitute the basename of the input file being processed.
6812 This is the substring up to (and not including) the last period
6813 and not including the directory.
6816 This is the same as @samp{%b}, but include the file suffix (text after
6820 Marks the argument containing or following the @samp{%d} as a
6821 temporary file name, so that that file will be deleted if GCC exits
6822 successfully. Unlike @samp{%g}, this contributes no text to the
6825 @item %g@var{suffix}
6826 Substitute a file name that has suffix @var{suffix} and is chosen
6827 once per compilation, and mark the argument in the same way as
6828 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
6829 name is now chosen in a way that is hard to predict even when previously
6830 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
6831 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
6832 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
6833 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
6834 was simply substituted with a file name chosen once per compilation,
6835 without regard to any appended suffix (which was therefore treated
6836 just like ordinary text), making such attacks more likely to succeed.
6838 @item %u@var{suffix}
6839 Like @samp{%g}, but generates a new temporary file name even if
6840 @samp{%u@var{suffix}} was already seen.
6842 @item %U@var{suffix}
6843 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
6844 new one if there is no such last file name. In the absence of any
6845 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
6846 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
6847 would involve the generation of two distinct file names, one
6848 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
6849 simply substituted with a file name chosen for the previous @samp{%u},
6850 without regard to any appended suffix.
6852 @item %j@var{suffix}
6853 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
6854 writable, and if save-temps is off; otherwise, substitute the name
6855 of a temporary file, just like @samp{%u}. This temporary file is not
6856 meant for communication between processes, but rather as a junk
6859 @item %|@var{suffix}
6860 @itemx %m@var{suffix}
6861 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
6862 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
6863 all. These are the two most common ways to instruct a program that it
6864 should read from standard input or write to standard output. If you
6865 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
6866 construct: see for example @file{f/lang-specs.h}.
6868 @item %.@var{SUFFIX}
6869 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
6870 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
6871 terminated by the next space or %.
6874 Marks the argument containing or following the @samp{%w} as the
6875 designated output file of this compilation. This puts the argument
6876 into the sequence of arguments that @samp{%o} will substitute later.
6879 Substitutes the names of all the output files, with spaces
6880 automatically placed around them. You should write spaces
6881 around the @samp{%o} as well or the results are undefined.
6882 @samp{%o} is for use in the specs for running the linker.
6883 Input files whose names have no recognized suffix are not compiled
6884 at all, but they are included among the output files, so they will
6888 Substitutes the suffix for object files. Note that this is
6889 handled specially when it immediately follows @samp{%g, %u, or %U},
6890 because of the need for those to form complete file names. The
6891 handling is such that @samp{%O} is treated exactly as if it had already
6892 been substituted, except that @samp{%g, %u, and %U} do not currently
6893 support additional @var{suffix} characters following @samp{%O} as they would
6894 following, for example, @samp{.o}.
6897 Substitutes the standard macro predefinitions for the
6898 current target machine. Use this when running @code{cpp}.
6901 Like @samp{%p}, but puts @samp{__} before and after the name of each
6902 predefined macro, except for macros that start with @samp{__} or with
6903 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
6907 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
6908 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
6909 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
6910 and @option{-imultilib} as necessary.
6913 Current argument is the name of a library or startup file of some sort.
6914 Search for that file in a standard list of directories and substitute
6915 the full name found.
6918 Print @var{str} as an error message. @var{str} is terminated by a newline.
6919 Use this when inconsistent options are detected.
6922 Substitute the contents of spec string @var{name} at this point.
6925 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
6927 @item %x@{@var{option}@}
6928 Accumulate an option for @samp{%X}.
6931 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
6935 Output the accumulated assembler options specified by @option{-Wa}.
6938 Output the accumulated preprocessor options specified by @option{-Wp}.
6941 Process the @code{asm} spec. This is used to compute the
6942 switches to be passed to the assembler.
6945 Process the @code{asm_final} spec. This is a spec string for
6946 passing switches to an assembler post-processor, if such a program is
6950 Process the @code{link} spec. This is the spec for computing the
6951 command line passed to the linker. Typically it will make use of the
6952 @samp{%L %G %S %D and %E} sequences.
6955 Dump out a @option{-L} option for each directory that GCC believes might
6956 contain startup files. If the target supports multilibs then the
6957 current multilib directory will be prepended to each of these paths.
6960 Process the @code{lib} spec. This is a spec string for deciding which
6961 libraries should be included on the command line to the linker.
6964 Process the @code{libgcc} spec. This is a spec string for deciding
6965 which GCC support library should be included on the command line to the linker.
6968 Process the @code{startfile} spec. This is a spec for deciding which
6969 object files should be the first ones passed to the linker. Typically
6970 this might be a file named @file{crt0.o}.
6973 Process the @code{endfile} spec. This is a spec string that specifies
6974 the last object files that will be passed to the linker.
6977 Process the @code{cpp} spec. This is used to construct the arguments
6978 to be passed to the C preprocessor.
6981 Process the @code{cc1} spec. This is used to construct the options to be
6982 passed to the actual C compiler (@samp{cc1}).
6985 Process the @code{cc1plus} spec. This is used to construct the options to be
6986 passed to the actual C++ compiler (@samp{cc1plus}).
6989 Substitute the variable part of a matched option. See below.
6990 Note that each comma in the substituted string is replaced by
6994 Remove all occurrences of @code{-S} from the command line. Note---this
6995 command is position dependent. @samp{%} commands in the spec string
6996 before this one will see @code{-S}, @samp{%} commands in the spec string
6997 after this one will not.
6999 @item %:@var{function}(@var{args})
7000 Call the named function @var{function}, passing it @var{args}.
7001 @var{args} is first processed as a nested spec string, then split
7002 into an argument vector in the usual fashion. The function returns
7003 a string which is processed as if it had appeared literally as part
7004 of the current spec.
7006 The following built-in spec functions are provided:
7009 @item @code{if-exists}
7010 The @code{if-exists} spec function takes one argument, an absolute
7011 pathname to a file. If the file exists, @code{if-exists} returns the
7012 pathname. Here is a small example of its usage:
7016 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
7019 @item @code{if-exists-else}
7020 The @code{if-exists-else} spec function is similar to the @code{if-exists}
7021 spec function, except that it takes two arguments. The first argument is
7022 an absolute pathname to a file. If the file exists, @code{if-exists-else}
7023 returns the pathname. If it does not exist, it returns the second argument.
7024 This way, @code{if-exists-else} can be used to select one file or another,
7025 based on the existence of the first. Here is a small example of its usage:
7029 crt0%O%s %:if-exists(crti%O%s) \
7030 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
7033 @item @code{replace-outfile}
7034 The @code{replace-outfile} spec function takes two arguments. It looks for the
7035 first argument in the outfiles array and replaces it with the second argument. Here
7036 is a small example of its usage:
7039 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
7045 Substitutes the @code{-S} switch, if that switch was given to GCC@.
7046 If that switch was not specified, this substitutes nothing. Note that
7047 the leading dash is omitted when specifying this option, and it is
7048 automatically inserted if the substitution is performed. Thus the spec
7049 string @samp{%@{foo@}} would match the command-line option @option{-foo}
7050 and would output the command line option @option{-foo}.
7052 @item %W@{@code{S}@}
7053 Like %@{@code{S}@} but mark last argument supplied within as a file to be
7056 @item %@{@code{S}*@}
7057 Substitutes all the switches specified to GCC whose names start
7058 with @code{-S}, but which also take an argument. This is used for
7059 switches like @option{-o}, @option{-D}, @option{-I}, etc.
7060 GCC considers @option{-o foo} as being
7061 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
7062 text, including the space. Thus two arguments would be generated.
7064 @item %@{@code{S}*&@code{T}*@}
7065 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
7066 (the order of @code{S} and @code{T} in the spec is not significant).
7067 There can be any number of ampersand-separated variables; for each the
7068 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
7070 @item %@{@code{S}:@code{X}@}
7071 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
7073 @item %@{!@code{S}:@code{X}@}
7074 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
7076 @item %@{@code{S}*:@code{X}@}
7077 Substitutes @code{X} if one or more switches whose names start with
7078 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
7079 once, no matter how many such switches appeared. However, if @code{%*}
7080 appears somewhere in @code{X}, then @code{X} will be substituted once
7081 for each matching switch, with the @code{%*} replaced by the part of
7082 that switch that matched the @code{*}.
7084 @item %@{.@code{S}:@code{X}@}
7085 Substitutes @code{X}, if processing a file with suffix @code{S}.
7087 @item %@{!.@code{S}:@code{X}@}
7088 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
7090 @item %@{@code{S}|@code{P}:@code{X}@}
7091 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
7092 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
7093 although they have a stronger binding than the @samp{|}. If @code{%*}
7094 appears in @code{X}, all of the alternatives must be starred, and only
7095 the first matching alternative is substituted.
7097 For example, a spec string like this:
7100 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
7103 will output the following command-line options from the following input
7104 command-line options:
7109 -d fred.c -foo -baz -boggle
7110 -d jim.d -bar -baz -boggle
7113 @item %@{S:X; T:Y; :D@}
7115 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
7116 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
7117 be as many clauses as you need. This may be combined with @code{.},
7118 @code{!}, @code{|}, and @code{*} as needed.
7123 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
7124 construct may contain other nested @samp{%} constructs or spaces, or
7125 even newlines. They are processed as usual, as described above.
7126 Trailing white space in @code{X} is ignored. White space may also
7127 appear anywhere on the left side of the colon in these constructs,
7128 except between @code{.} or @code{*} and the corresponding word.
7130 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
7131 handled specifically in these constructs. If another value of
7132 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
7133 @option{-W} switch is found later in the command line, the earlier
7134 switch value is ignored, except with @{@code{S}*@} where @code{S} is
7135 just one letter, which passes all matching options.
7137 The character @samp{|} at the beginning of the predicate text is used to
7138 indicate that a command should be piped to the following command, but
7139 only if @option{-pipe} is specified.
7141 It is built into GCC which switches take arguments and which do not.
7142 (You might think it would be useful to generalize this to allow each
7143 compiler's spec to say which switches take arguments. But this cannot
7144 be done in a consistent fashion. GCC cannot even decide which input
7145 files have been specified without knowing which switches take arguments,
7146 and it must know which input files to compile in order to tell which
7149 GCC also knows implicitly that arguments starting in @option{-l} are to be
7150 treated as compiler output files, and passed to the linker in their
7151 proper position among the other output files.
7153 @c man begin OPTIONS
7155 @node Target Options
7156 @section Specifying Target Machine and Compiler Version
7157 @cindex target options
7158 @cindex cross compiling
7159 @cindex specifying machine version
7160 @cindex specifying compiler version and target machine
7161 @cindex compiler version, specifying
7162 @cindex target machine, specifying
7164 The usual way to run GCC is to run the executable called @file{gcc}, or
7165 @file{<machine>-gcc} when cross-compiling, or
7166 @file{<machine>-gcc-<version>} to run a version other than the one that
7167 was installed last. Sometimes this is inconvenient, so GCC provides
7168 options that will switch to another cross-compiler or version.
7171 @item -b @var{machine}
7173 The argument @var{machine} specifies the target machine for compilation.
7175 The value to use for @var{machine} is the same as was specified as the
7176 machine type when configuring GCC as a cross-compiler. For
7177 example, if a cross-compiler was configured with @samp{configure
7178 arm-elf}, meaning to compile for an arm processor with elf binaries,
7179 then you would specify @option{-b arm-elf} to run that cross compiler.
7180 Because there are other options beginning with @option{-b}, the
7181 configuration must contain a hyphen.
7183 @item -V @var{version}
7185 The argument @var{version} specifies which version of GCC to run.
7186 This is useful when multiple versions are installed. For example,
7187 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
7190 The @option{-V} and @option{-b} options work by running the
7191 @file{<machine>-gcc-<version>} executable, so there's no real reason to
7192 use them if you can just run that directly.
7194 @node Submodel Options
7195 @section Hardware Models and Configurations
7196 @cindex submodel options
7197 @cindex specifying hardware config
7198 @cindex hardware models and configurations, specifying
7199 @cindex machine dependent options
7201 Earlier we discussed the standard option @option{-b} which chooses among
7202 different installed compilers for completely different target
7203 machines, such as VAX vs.@: 68000 vs.@: 80386.
7205 In addition, each of these target machine types can have its own
7206 special options, starting with @samp{-m}, to choose among various
7207 hardware models or configurations---for example, 68010 vs 68020,
7208 floating coprocessor or none. A single installed version of the
7209 compiler can compile for any model or configuration, according to the
7212 Some configurations of the compiler also support additional special
7213 options, usually for compatibility with other compilers on the same
7216 @c This list is ordered alphanumerically by subsection name.
7217 @c It should be the same order and spelling as these options are listed
7218 @c in Machine Dependent Options
7224 * Blackfin Options::
7228 * DEC Alpha Options::
7229 * DEC Alpha/VMS Options::
7231 * GNU/Linux Options::
7234 * i386 and x86-64 Options::
7247 * RS/6000 and PowerPC Options::
7248 * S/390 and zSeries Options::
7251 * System V Options::
7252 * TMS320C3x/C4x Options::
7256 * Xstormy16 Options::
7262 @subsection ARC Options
7265 These options are defined for ARC implementations:
7270 Compile code for little endian mode. This is the default.
7274 Compile code for big endian mode.
7277 @opindex mmangle-cpu
7278 Prepend the name of the cpu to all public symbol names.
7279 In multiple-processor systems, there are many ARC variants with different
7280 instruction and register set characteristics. This flag prevents code
7281 compiled for one cpu to be linked with code compiled for another.
7282 No facility exists for handling variants that are ``almost identical''.
7283 This is an all or nothing option.
7285 @item -mcpu=@var{cpu}
7287 Compile code for ARC variant @var{cpu}.
7288 Which variants are supported depend on the configuration.
7289 All variants support @option{-mcpu=base}, this is the default.
7291 @item -mtext=@var{text-section}
7292 @itemx -mdata=@var{data-section}
7293 @itemx -mrodata=@var{readonly-data-section}
7297 Put functions, data, and readonly data in @var{text-section},
7298 @var{data-section}, and @var{readonly-data-section} respectively
7299 by default. This can be overridden with the @code{section} attribute.
7300 @xref{Variable Attributes}.
7305 @subsection ARM Options
7308 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
7312 @item -mabi=@var{name}
7314 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
7315 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
7318 @opindex mapcs-frame
7319 Generate a stack frame that is compliant with the ARM Procedure Call
7320 Standard for all functions, even if this is not strictly necessary for
7321 correct execution of the code. Specifying @option{-fomit-frame-pointer}
7322 with this option will cause the stack frames not to be generated for
7323 leaf functions. The default is @option{-mno-apcs-frame}.
7327 This is a synonym for @option{-mapcs-frame}.
7330 @c not currently implemented
7331 @item -mapcs-stack-check
7332 @opindex mapcs-stack-check
7333 Generate code to check the amount of stack space available upon entry to
7334 every function (that actually uses some stack space). If there is
7335 insufficient space available then either the function
7336 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
7337 called, depending upon the amount of stack space required. The run time
7338 system is required to provide these functions. The default is
7339 @option{-mno-apcs-stack-check}, since this produces smaller code.
7341 @c not currently implemented
7343 @opindex mapcs-float
7344 Pass floating point arguments using the float point registers. This is
7345 one of the variants of the APCS@. This option is recommended if the
7346 target hardware has a floating point unit or if a lot of floating point
7347 arithmetic is going to be performed by the code. The default is
7348 @option{-mno-apcs-float}, since integer only code is slightly increased in
7349 size if @option{-mapcs-float} is used.
7351 @c not currently implemented
7352 @item -mapcs-reentrant
7353 @opindex mapcs-reentrant
7354 Generate reentrant, position independent code. The default is
7355 @option{-mno-apcs-reentrant}.
7358 @item -mthumb-interwork
7359 @opindex mthumb-interwork
7360 Generate code which supports calling between the ARM and Thumb
7361 instruction sets. Without this option the two instruction sets cannot
7362 be reliably used inside one program. The default is
7363 @option{-mno-thumb-interwork}, since slightly larger code is generated
7364 when @option{-mthumb-interwork} is specified.
7366 @item -mno-sched-prolog
7367 @opindex mno-sched-prolog
7368 Prevent the reordering of instructions in the function prolog, or the
7369 merging of those instruction with the instructions in the function's
7370 body. This means that all functions will start with a recognizable set
7371 of instructions (or in fact one of a choice from a small set of
7372 different function prologues), and this information can be used to
7373 locate the start if functions inside an executable piece of code. The
7374 default is @option{-msched-prolog}.
7377 @opindex mhard-float
7378 Generate output containing floating point instructions. This is the
7382 @opindex msoft-float
7383 Generate output containing library calls for floating point.
7384 @strong{Warning:} the requisite libraries are not available for all ARM
7385 targets. Normally the facilities of the machine's usual C compiler are
7386 used, but this cannot be done directly in cross-compilation. You must make
7387 your own arrangements to provide suitable library functions for
7390 @option{-msoft-float} changes the calling convention in the output file;
7391 therefore, it is only useful if you compile @emph{all} of a program with
7392 this option. In particular, you need to compile @file{libgcc.a}, the
7393 library that comes with GCC, with @option{-msoft-float} in order for
7396 @item -mfloat-abi=@var{name}
7398 Specifies which ABI to use for floating point values. Permissible values
7399 are: @samp{soft}, @samp{softfp} and @samp{hard}.
7401 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
7402 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
7403 of floating point instructions, but still uses the soft-float calling
7406 @item -mlittle-endian
7407 @opindex mlittle-endian
7408 Generate code for a processor running in little-endian mode. This is
7409 the default for all standard configurations.
7412 @opindex mbig-endian
7413 Generate code for a processor running in big-endian mode; the default is
7414 to compile code for a little-endian processor.
7416 @item -mwords-little-endian
7417 @opindex mwords-little-endian
7418 This option only applies when generating code for big-endian processors.
7419 Generate code for a little-endian word order but a big-endian byte
7420 order. That is, a byte order of the form @samp{32107654}. Note: this
7421 option should only be used if you require compatibility with code for
7422 big-endian ARM processors generated by versions of the compiler prior to
7425 @item -mcpu=@var{name}
7427 This specifies the name of the target ARM processor. GCC uses this name
7428 to determine what kind of instructions it can emit when generating
7429 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
7430 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
7431 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
7432 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
7433 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
7434 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
7435 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
7436 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
7437 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
7438 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
7439 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
7440 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
7441 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
7442 @samp{arm1176jz-s}, @samp{arm1176jzf-s}, @samp{xscale}, @samp{iwmmxt},
7445 @itemx -mtune=@var{name}
7447 This option is very similar to the @option{-mcpu=} option, except that
7448 instead of specifying the actual target processor type, and hence
7449 restricting which instructions can be used, it specifies that GCC should
7450 tune the performance of the code as if the target were of the type
7451 specified in this option, but still choosing the instructions that it
7452 will generate based on the cpu specified by a @option{-mcpu=} option.
7453 For some ARM implementations better performance can be obtained by using
7456 @item -march=@var{name}
7458 This specifies the name of the target ARM architecture. GCC uses this
7459 name to determine what kind of instructions it can emit when generating
7460 assembly code. This option can be used in conjunction with or instead
7461 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
7462 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
7463 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
7464 @samp{iwmmxt}, @samp{ep9312}.
7466 @item -mfpu=@var{name}
7467 @itemx -mfpe=@var{number}
7468 @itemx -mfp=@var{number}
7472 This specifies what floating point hardware (or hardware emulation) is
7473 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
7474 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
7475 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
7476 with older versions of GCC@.
7478 If @option{-msoft-float} is specified this specifies the format of
7479 floating point values.
7481 @item -mstructure-size-boundary=@var{n}
7482 @opindex mstructure-size-boundary
7483 The size of all structures and unions will be rounded up to a multiple
7484 of the number of bits set by this option. Permissible values are 8, 32
7485 and 64. The default value varies for different toolchains. For the COFF
7486 targeted toolchain the default value is 8. A value of 64 is only allowed
7487 if the underlying ABI supports it.
7489 Specifying the larger number can produce faster, more efficient code, but
7490 can also increase the size of the program. Different values are potentially
7491 incompatible. Code compiled with one value cannot necessarily expect to
7492 work with code or libraries compiled with another value, if they exchange
7493 information using structures or unions.
7495 @item -mabort-on-noreturn
7496 @opindex mabort-on-noreturn
7497 Generate a call to the function @code{abort} at the end of a
7498 @code{noreturn} function. It will be executed if the function tries to
7502 @itemx -mno-long-calls
7503 @opindex mlong-calls
7504 @opindex mno-long-calls
7505 Tells the compiler to perform function calls by first loading the
7506 address of the function into a register and then performing a subroutine
7507 call on this register. This switch is needed if the target function
7508 will lie outside of the 64 megabyte addressing range of the offset based
7509 version of subroutine call instruction.
7511 Even if this switch is enabled, not all function calls will be turned
7512 into long calls. The heuristic is that static functions, functions
7513 which have the @samp{short-call} attribute, functions that are inside
7514 the scope of a @samp{#pragma no_long_calls} directive and functions whose
7515 definitions have already been compiled within the current compilation
7516 unit, will not be turned into long calls. The exception to this rule is
7517 that weak function definitions, functions with the @samp{long-call}
7518 attribute or the @samp{section} attribute, and functions that are within
7519 the scope of a @samp{#pragma long_calls} directive, will always be
7520 turned into long calls.
7522 This feature is not enabled by default. Specifying
7523 @option{-mno-long-calls} will restore the default behavior, as will
7524 placing the function calls within the scope of a @samp{#pragma
7525 long_calls_off} directive. Note these switches have no effect on how
7526 the compiler generates code to handle function calls via function
7529 @item -mnop-fun-dllimport
7530 @opindex mnop-fun-dllimport
7531 Disable support for the @code{dllimport} attribute.
7533 @item -msingle-pic-base
7534 @opindex msingle-pic-base
7535 Treat the register used for PIC addressing as read-only, rather than
7536 loading it in the prologue for each function. The run-time system is
7537 responsible for initializing this register with an appropriate value
7538 before execution begins.
7540 @item -mpic-register=@var{reg}
7541 @opindex mpic-register
7542 Specify the register to be used for PIC addressing. The default is R10
7543 unless stack-checking is enabled, when R9 is used.
7545 @item -mcirrus-fix-invalid-insns
7546 @opindex mcirrus-fix-invalid-insns
7547 @opindex mno-cirrus-fix-invalid-insns
7548 Insert NOPs into the instruction stream to in order to work around
7549 problems with invalid Maverick instruction combinations. This option
7550 is only valid if the @option{-mcpu=ep9312} option has been used to
7551 enable generation of instructions for the Cirrus Maverick floating
7552 point co-processor. This option is not enabled by default, since the
7553 problem is only present in older Maverick implementations. The default
7554 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
7557 @item -mpoke-function-name
7558 @opindex mpoke-function-name
7559 Write the name of each function into the text section, directly
7560 preceding the function prologue. The generated code is similar to this:
7564 .ascii "arm_poke_function_name", 0
7567 .word 0xff000000 + (t1 - t0)
7568 arm_poke_function_name
7570 stmfd sp!, @{fp, ip, lr, pc@}
7574 When performing a stack backtrace, code can inspect the value of
7575 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
7576 location @code{pc - 12} and the top 8 bits are set, then we know that
7577 there is a function name embedded immediately preceding this location
7578 and has length @code{((pc[-3]) & 0xff000000)}.
7582 Generate code for the 16-bit Thumb instruction set. The default is to
7583 use the 32-bit ARM instruction set.
7586 @opindex mtpcs-frame
7587 Generate a stack frame that is compliant with the Thumb Procedure Call
7588 Standard for all non-leaf functions. (A leaf function is one that does
7589 not call any other functions.) The default is @option{-mno-tpcs-frame}.
7591 @item -mtpcs-leaf-frame
7592 @opindex mtpcs-leaf-frame
7593 Generate a stack frame that is compliant with the Thumb Procedure Call
7594 Standard for all leaf functions. (A leaf function is one that does
7595 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
7597 @item -mcallee-super-interworking
7598 @opindex mcallee-super-interworking
7599 Gives all externally visible functions in the file being compiled an ARM
7600 instruction set header which switches to Thumb mode before executing the
7601 rest of the function. This allows these functions to be called from
7602 non-interworking code.
7604 @item -mcaller-super-interworking
7605 @opindex mcaller-super-interworking
7606 Allows calls via function pointers (including virtual functions) to
7607 execute correctly regardless of whether the target code has been
7608 compiled for interworking or not. There is a small overhead in the cost
7609 of executing a function pointer if this option is enabled.
7611 @item -mtp=@var{name}
7613 Specify the access model for the thread local storage pointer. The valid
7614 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
7615 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
7616 (supported in the arm6k architecture), and @option{auto}, which uses the
7617 best available method for the selected processor. The default setting is
7623 @subsection AVR Options
7626 These options are defined for AVR implementations:
7629 @item -mmcu=@var{mcu}
7631 Specify ATMEL AVR instruction set or MCU type.
7633 Instruction set avr1 is for the minimal AVR core, not supported by the C
7634 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
7635 attiny11, attiny12, attiny15, attiny28).
7637 Instruction set avr2 (default) is for the classic AVR core with up to
7638 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
7639 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
7640 at90c8534, at90s8535).
7642 Instruction set avr3 is for the classic AVR core with up to 128K program
7643 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
7645 Instruction set avr4 is for the enhanced AVR core with up to 8K program
7646 memory space (MCU types: atmega8, atmega83, atmega85).
7648 Instruction set avr5 is for the enhanced AVR core with up to 128K program
7649 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
7650 atmega64, atmega128, at43usb355, at94k).
7654 Output instruction sizes to the asm file.
7656 @item -minit-stack=@var{N}
7657 @opindex minit-stack
7658 Specify the initial stack address, which may be a symbol or numeric value,
7659 @samp{__stack} is the default.
7661 @item -mno-interrupts
7662 @opindex mno-interrupts
7663 Generated code is not compatible with hardware interrupts.
7664 Code size will be smaller.
7666 @item -mcall-prologues
7667 @opindex mcall-prologues
7668 Functions prologues/epilogues expanded as call to appropriate
7669 subroutines. Code size will be smaller.
7671 @item -mno-tablejump
7672 @opindex mno-tablejump
7673 Do not generate tablejump insns which sometimes increase code size.
7676 @opindex mtiny-stack
7677 Change only the low 8 bits of the stack pointer.
7681 Assume int to be 8 bit integer. This affects the sizes of all types: A
7682 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
7683 and long long will be 4 bytes. Please note that this option does not
7684 comply to the C standards, but it will provide you with smaller code
7688 @node Blackfin Options
7689 @subsection Blackfin Options
7690 @cindex Blackfin Options
7693 @item -momit-leaf-frame-pointer
7694 @opindex momit-leaf-frame-pointer
7695 Don't keep the frame pointer in a register for leaf functions. This
7696 avoids the instructions to save, set up and restore frame pointers and
7697 makes an extra register available in leaf functions. The option
7698 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7699 which might make debugging harder.
7701 @item -mspecld-anomaly
7702 @opindex mspecld-anomaly
7703 When enabled, the compiler will ensure that the generated code does not
7704 contain speculative loads after jump instructions. This option is enabled
7707 @item -mno-specld-anomaly
7708 @opindex mno-specld-anomaly
7709 Don't generate extra code to prevent speculative loads from occurring.
7711 @item -mcsync-anomaly
7712 @opindex mcsync-anomaly
7713 When enabled, the compiler will ensure that the generated code does not
7714 contain CSYNC or SSYNC instructions too soon after conditional branches.
7715 This option is enabled by default.
7717 @item -mno-csync-anomaly
7718 @opindex mno-csync-anomaly
7719 Don't generate extra code to prevent CSYNC or SSYNC instructions from
7720 occurring too soon after a conditional branch.
7724 When enabled, the compiler is free to take advantage of the knowledge that
7725 the entire program fits into the low 64k of memory.
7728 @opindex mno-low-64k
7729 Assume that the program is arbitrarily large. This is the default.
7731 @item -mid-shared-library
7732 @opindex mid-shared-library
7733 Generate code that supports shared libraries via the library ID method.
7734 This allows for execute in place and shared libraries in an environment
7735 without virtual memory management. This option implies @option{-fPIC}.
7737 @item -mno-id-shared-library
7738 @opindex mno-id-shared-library
7739 Generate code that doesn't assume ID based shared libraries are being used.
7740 This is the default.
7742 @item -mshared-library-id=n
7743 @opindex mshared-library-id
7744 Specified the identification number of the ID based shared library being
7745 compiled. Specifying a value of 0 will generate more compact code, specifying
7746 other values will force the allocation of that number to the current
7747 library but is no more space or time efficient than omitting this option.
7750 @itemx -mno-long-calls
7751 @opindex mlong-calls
7752 @opindex mno-long-calls
7753 Tells the compiler to perform function calls by first loading the
7754 address of the function into a register and then performing a subroutine
7755 call on this register. This switch is needed if the target function
7756 will lie outside of the 24 bit addressing range of the offset based
7757 version of subroutine call instruction.
7759 This feature is not enabled by default. Specifying
7760 @option{-mno-long-calls} will restore the default behavior. Note these
7761 switches have no effect on how the compiler generates code to handle
7762 function calls via function pointers.
7766 @subsection CRIS Options
7767 @cindex CRIS Options
7769 These options are defined specifically for the CRIS ports.
7772 @item -march=@var{architecture-type}
7773 @itemx -mcpu=@var{architecture-type}
7776 Generate code for the specified architecture. The choices for
7777 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
7778 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
7779 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
7782 @item -mtune=@var{architecture-type}
7784 Tune to @var{architecture-type} everything applicable about the generated
7785 code, except for the ABI and the set of available instructions. The
7786 choices for @var{architecture-type} are the same as for
7787 @option{-march=@var{architecture-type}}.
7789 @item -mmax-stack-frame=@var{n}
7790 @opindex mmax-stack-frame
7791 Warn when the stack frame of a function exceeds @var{n} bytes.
7793 @item -melinux-stacksize=@var{n}
7794 @opindex melinux-stacksize
7795 Only available with the @samp{cris-axis-aout} target. Arranges for
7796 indications in the program to the kernel loader that the stack of the
7797 program should be set to @var{n} bytes.
7803 The options @option{-metrax4} and @option{-metrax100} are synonyms for
7804 @option{-march=v3} and @option{-march=v8} respectively.
7806 @item -mmul-bug-workaround
7807 @itemx -mno-mul-bug-workaround
7808 @opindex mmul-bug-workaround
7809 @opindex mno-mul-bug-workaround
7810 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
7811 models where it applies. This option is active by default.
7815 Enable CRIS-specific verbose debug-related information in the assembly
7816 code. This option also has the effect to turn off the @samp{#NO_APP}
7817 formatted-code indicator to the assembler at the beginning of the
7822 Do not use condition-code results from previous instruction; always emit
7823 compare and test instructions before use of condition codes.
7825 @item -mno-side-effects
7826 @opindex mno-side-effects
7827 Do not emit instructions with side-effects in addressing modes other than
7831 @itemx -mno-stack-align
7833 @itemx -mno-data-align
7834 @itemx -mconst-align
7835 @itemx -mno-const-align
7836 @opindex mstack-align
7837 @opindex mno-stack-align
7838 @opindex mdata-align
7839 @opindex mno-data-align
7840 @opindex mconst-align
7841 @opindex mno-const-align
7842 These options (no-options) arranges (eliminate arrangements) for the
7843 stack-frame, individual data and constants to be aligned for the maximum
7844 single data access size for the chosen CPU model. The default is to
7845 arrange for 32-bit alignment. ABI details such as structure layout are
7846 not affected by these options.
7854 Similar to the stack- data- and const-align options above, these options
7855 arrange for stack-frame, writable data and constants to all be 32-bit,
7856 16-bit or 8-bit aligned. The default is 32-bit alignment.
7858 @item -mno-prologue-epilogue
7859 @itemx -mprologue-epilogue
7860 @opindex mno-prologue-epilogue
7861 @opindex mprologue-epilogue
7862 With @option{-mno-prologue-epilogue}, the normal function prologue and
7863 epilogue that sets up the stack-frame are omitted and no return
7864 instructions or return sequences are generated in the code. Use this
7865 option only together with visual inspection of the compiled code: no
7866 warnings or errors are generated when call-saved registers must be saved,
7867 or storage for local variable needs to be allocated.
7873 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
7874 instruction sequences that load addresses for functions from the PLT part
7875 of the GOT rather than (traditional on other architectures) calls to the
7876 PLT@. The default is @option{-mgotplt}.
7880 Legacy no-op option only recognized with the cris-axis-aout target.
7884 Legacy no-op option only recognized with the cris-axis-elf and
7885 cris-axis-linux-gnu targets.
7889 Only recognized with the cris-axis-aout target, where it selects a
7890 GNU/linux-like multilib, include files and instruction set for
7895 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
7899 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
7900 to link with input-output functions from a simulator library. Code,
7901 initialized data and zero-initialized data are allocated consecutively.
7905 Like @option{-sim}, but pass linker options to locate initialized data at
7906 0x40000000 and zero-initialized data at 0x80000000.
7910 @subsection CRX Options
7913 These options are defined specifically for the CRX ports.
7919 Enable the use of multiply-accumulate instructions. Disabled by default.
7923 Push instructions will be used to pass outgoing arguments when functions
7924 are called. Enabled by default.
7927 @node Darwin Options
7928 @subsection Darwin Options
7929 @cindex Darwin options
7931 These options are defined for all architectures running the Darwin operating
7934 FSF GCC on Darwin does not create ``fat'' object files; it will create
7935 an object file for the single architecture that it was built to
7936 target. Apple's GCC on Darwin does create ``fat'' files if multiple
7937 @option{-arch} options are used; it does so by running the compiler or
7938 linker multiple times and joining the results together with
7941 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
7942 @samp{i686}) is determined by the flags that specify the ISA
7943 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
7944 @option{-force_cpusubtype_ALL} option can be used to override this.
7946 The Darwin tools vary in their behavior when presented with an ISA
7947 mismatch. The assembler, @file{as}, will only permit instructions to
7948 be used that are valid for the subtype of the file it is generating,
7949 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
7950 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
7951 and print an error if asked to create a shared library with a less
7952 restrictive subtype than its input files (for instance, trying to put
7953 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
7954 for executables, @file{ld}, will quietly give the executable the most
7955 restrictive subtype of any of its input files.
7960 Add the framework directory @var{dir} to the head of the list of
7961 directories to be searched for header files. These directories are
7962 interleaved with those specified by @option{-I} options and are
7963 scanned in a left-to-right order.
7965 A framework directory is a directory with frameworks in it. A
7966 framework is a directory with a @samp{"Headers"} and/or
7967 @samp{"PrivateHeaders"} directory contained directly in it that ends
7968 in @samp{".framework"}. The name of a framework is the name of this
7969 directory excluding the @samp{".framework"}. Headers associated with
7970 the framework are found in one of those two directories, with
7971 @samp{"Headers"} being searched first. A subframework is a framework
7972 directory that is in a framework's @samp{"Frameworks"} directory.
7973 Includes of subframework headers can only appear in a header of a
7974 framework that contains the subframework, or in a sibling subframework
7975 header. Two subframeworks are siblings if they occur in the same
7976 framework. A subframework should not have the same name as a
7977 framework, a warning will be issued if this is violated. Currently a
7978 subframework cannot have subframeworks, in the future, the mechanism
7979 may be extended to support this. The standard frameworks can be found
7980 in @samp{"/System/Library/Frameworks"} and
7981 @samp{"/Library/Frameworks"}. An example include looks like
7982 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
7983 the name of the framework and header.h is found in the
7984 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
7988 Emit debugging information for symbols that are used. For STABS
7989 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
7990 This is by default ON@.
7994 Emit debugging information for all symbols and types.
7996 @item -mmacosx-version-min=@var{version}
7997 The earliest version of MacOS X that this executable will run on
7998 is @var{version}. Typical values of @var{version} include @code{10.1},
7999 @code{10.2}, and @code{10.3.9}.
8001 The default for this option is to make choices that seem to be most
8004 @item -mone-byte-bool
8005 @opindex -mone-byte-bool
8006 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
8007 By default @samp{sizeof(bool)} is @samp{4} when compiling for
8008 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
8009 option has no effect on x86.
8011 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
8012 to generate code that is not binary compatible with code generated
8013 without that switch. Using this switch may require recompiling all
8014 other modules in a program, including system libraries. Use this
8015 switch to conform to a non-default data model.
8017 @item -mfix-and-continue
8018 @itemx -ffix-and-continue
8019 @itemx -findirect-data
8020 @opindex mfix-and-continue
8021 @opindex ffix-and-continue
8022 @opindex findirect-data
8023 Generate code suitable for fast turn around development. Needed to
8024 enable gdb to dynamically load @code{.o} files into already running
8025 programs. @option{-findirect-data} and @option{-ffix-and-continue}
8026 are provided for backwards compatibility.
8030 Loads all members of static archive libraries.
8031 See man ld(1) for more information.
8033 @item -arch_errors_fatal
8034 @opindex arch_errors_fatal
8035 Cause the errors having to do with files that have the wrong architecture
8039 @opindex bind_at_load
8040 Causes the output file to be marked such that the dynamic linker will
8041 bind all undefined references when the file is loaded or launched.
8045 Produce a Mach-o bundle format file.
8046 See man ld(1) for more information.
8048 @item -bundle_loader @var{executable}
8049 @opindex bundle_loader
8050 This option specifies the @var{executable} that will be loading the build
8051 output file being linked. See man ld(1) for more information.
8054 @opindex -dynamiclib
8055 When passed this option, GCC will produce a dynamic library instead of
8056 an executable when linking, using the Darwin @file{libtool} command.
8058 @item -force_cpusubtype_ALL
8059 @opindex -force_cpusubtype_ALL
8060 This causes GCC's output file to have the @var{ALL} subtype, instead of
8061 one controlled by the @option{-mcpu} or @option{-march} option.
8063 @item -allowable_client @var{client_name}
8065 @itemx -compatibility_version
8066 @itemx -current_version
8068 @itemx -dependency-file
8070 @itemx -dylinker_install_name
8072 @itemx -exported_symbols_list
8074 @itemx -flat_namespace
8075 @itemx -force_flat_namespace
8076 @itemx -headerpad_max_install_names
8079 @itemx -install_name
8080 @itemx -keep_private_externs
8081 @itemx -multi_module
8082 @itemx -multiply_defined
8083 @itemx -multiply_defined_unused
8085 @itemx -no_dead_strip_inits_and_terms
8086 @itemx -nofixprebinding
8089 @itemx -noseglinkedit
8090 @itemx -pagezero_size
8092 @itemx -prebind_all_twolevel_modules
8093 @itemx -private_bundle
8094 @itemx -read_only_relocs
8096 @itemx -sectobjectsymbols
8100 @itemx -sectobjectsymbols
8103 @itemx -segs_read_only_addr
8104 @itemx -segs_read_write_addr
8105 @itemx -seg_addr_table
8106 @itemx -seg_addr_table_filename
8109 @itemx -segs_read_only_addr
8110 @itemx -segs_read_write_addr
8111 @itemx -single_module
8114 @itemx -sub_umbrella
8115 @itemx -twolevel_namespace
8118 @itemx -unexported_symbols_list
8119 @itemx -weak_reference_mismatches
8122 @opindex allowable_client
8123 @opindex client_name
8124 @opindex compatibility_version
8125 @opindex current_version
8127 @opindex dependency-file
8129 @opindex dylinker_install_name
8131 @opindex exported_symbols_list
8133 @opindex flat_namespace
8134 @opindex force_flat_namespace
8135 @opindex headerpad_max_install_names
8138 @opindex install_name
8139 @opindex keep_private_externs
8140 @opindex multi_module
8141 @opindex multiply_defined
8142 @opindex multiply_defined_unused
8144 @opindex no_dead_strip_inits_and_terms
8145 @opindex nofixprebinding
8146 @opindex nomultidefs
8148 @opindex noseglinkedit
8149 @opindex pagezero_size
8151 @opindex prebind_all_twolevel_modules
8152 @opindex private_bundle
8153 @opindex read_only_relocs
8155 @opindex sectobjectsymbols
8159 @opindex sectobjectsymbols
8162 @opindex segs_read_only_addr
8163 @opindex segs_read_write_addr
8164 @opindex seg_addr_table
8165 @opindex seg_addr_table_filename
8166 @opindex seglinkedit
8168 @opindex segs_read_only_addr
8169 @opindex segs_read_write_addr
8170 @opindex single_module
8172 @opindex sub_library
8173 @opindex sub_umbrella
8174 @opindex twolevel_namespace
8177 @opindex unexported_symbols_list
8178 @opindex weak_reference_mismatches
8179 @opindex whatsloaded
8181 These options are passed to the Darwin linker. The Darwin linker man page
8182 describes them in detail.
8185 @node DEC Alpha Options
8186 @subsection DEC Alpha Options
8188 These @samp{-m} options are defined for the DEC Alpha implementations:
8191 @item -mno-soft-float
8193 @opindex mno-soft-float
8194 @opindex msoft-float
8195 Use (do not use) the hardware floating-point instructions for
8196 floating-point operations. When @option{-msoft-float} is specified,
8197 functions in @file{libgcc.a} will be used to perform floating-point
8198 operations. Unless they are replaced by routines that emulate the
8199 floating-point operations, or compiled in such a way as to call such
8200 emulations routines, these routines will issue floating-point
8201 operations. If you are compiling for an Alpha without floating-point
8202 operations, you must ensure that the library is built so as not to call
8205 Note that Alpha implementations without floating-point operations are
8206 required to have floating-point registers.
8211 @opindex mno-fp-regs
8212 Generate code that uses (does not use) the floating-point register set.
8213 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8214 register set is not used, floating point operands are passed in integer
8215 registers as if they were integers and floating-point results are passed
8216 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8217 so any function with a floating-point argument or return value called by code
8218 compiled with @option{-mno-fp-regs} must also be compiled with that
8221 A typical use of this option is building a kernel that does not use,
8222 and hence need not save and restore, any floating-point registers.
8226 The Alpha architecture implements floating-point hardware optimized for
8227 maximum performance. It is mostly compliant with the IEEE floating
8228 point standard. However, for full compliance, software assistance is
8229 required. This option generates code fully IEEE compliant code
8230 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8231 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8232 defined during compilation. The resulting code is less efficient but is
8233 able to correctly support denormalized numbers and exceptional IEEE
8234 values such as not-a-number and plus/minus infinity. Other Alpha
8235 compilers call this option @option{-ieee_with_no_inexact}.
8237 @item -mieee-with-inexact
8238 @opindex mieee-with-inexact
8239 This is like @option{-mieee} except the generated code also maintains
8240 the IEEE @var{inexact-flag}. Turning on this option causes the
8241 generated code to implement fully-compliant IEEE math. In addition to
8242 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8243 macro. On some Alpha implementations the resulting code may execute
8244 significantly slower than the code generated by default. Since there is
8245 very little code that depends on the @var{inexact-flag}, you should
8246 normally not specify this option. Other Alpha compilers call this
8247 option @option{-ieee_with_inexact}.
8249 @item -mfp-trap-mode=@var{trap-mode}
8250 @opindex mfp-trap-mode
8251 This option controls what floating-point related traps are enabled.
8252 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8253 The trap mode can be set to one of four values:
8257 This is the default (normal) setting. The only traps that are enabled
8258 are the ones that cannot be disabled in software (e.g., division by zero
8262 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8266 Like @samp{su}, but the instructions are marked to be safe for software
8267 completion (see Alpha architecture manual for details).
8270 Like @samp{su}, but inexact traps are enabled as well.
8273 @item -mfp-rounding-mode=@var{rounding-mode}
8274 @opindex mfp-rounding-mode
8275 Selects the IEEE rounding mode. Other Alpha compilers call this option
8276 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8281 Normal IEEE rounding mode. Floating point numbers are rounded towards
8282 the nearest machine number or towards the even machine number in case
8286 Round towards minus infinity.
8289 Chopped rounding mode. Floating point numbers are rounded towards zero.
8292 Dynamic rounding mode. A field in the floating point control register
8293 (@var{fpcr}, see Alpha architecture reference manual) controls the
8294 rounding mode in effect. The C library initializes this register for
8295 rounding towards plus infinity. Thus, unless your program modifies the
8296 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8299 @item -mtrap-precision=@var{trap-precision}
8300 @opindex mtrap-precision
8301 In the Alpha architecture, floating point traps are imprecise. This
8302 means without software assistance it is impossible to recover from a
8303 floating trap and program execution normally needs to be terminated.
8304 GCC can generate code that can assist operating system trap handlers
8305 in determining the exact location that caused a floating point trap.
8306 Depending on the requirements of an application, different levels of
8307 precisions can be selected:
8311 Program precision. This option is the default and means a trap handler
8312 can only identify which program caused a floating point exception.
8315 Function precision. The trap handler can determine the function that
8316 caused a floating point exception.
8319 Instruction precision. The trap handler can determine the exact
8320 instruction that caused a floating point exception.
8323 Other Alpha compilers provide the equivalent options called
8324 @option{-scope_safe} and @option{-resumption_safe}.
8326 @item -mieee-conformant
8327 @opindex mieee-conformant
8328 This option marks the generated code as IEEE conformant. You must not
8329 use this option unless you also specify @option{-mtrap-precision=i} and either
8330 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8331 is to emit the line @samp{.eflag 48} in the function prologue of the
8332 generated assembly file. Under DEC Unix, this has the effect that
8333 IEEE-conformant math library routines will be linked in.
8335 @item -mbuild-constants
8336 @opindex mbuild-constants
8337 Normally GCC examines a 32- or 64-bit integer constant to
8338 see if it can construct it from smaller constants in two or three
8339 instructions. If it cannot, it will output the constant as a literal and
8340 generate code to load it from the data segment at runtime.
8342 Use this option to require GCC to construct @emph{all} integer constants
8343 using code, even if it takes more instructions (the maximum is six).
8345 You would typically use this option to build a shared library dynamic
8346 loader. Itself a shared library, it must relocate itself in memory
8347 before it can find the variables and constants in its own data segment.
8353 Select whether to generate code to be assembled by the vendor-supplied
8354 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8372 Indicate whether GCC should generate code to use the optional BWX,
8373 CIX, FIX and MAX instruction sets. The default is to use the instruction
8374 sets supported by the CPU type specified via @option{-mcpu=} option or that
8375 of the CPU on which GCC was built if none was specified.
8380 @opindex mfloat-ieee
8381 Generate code that uses (does not use) VAX F and G floating point
8382 arithmetic instead of IEEE single and double precision.
8384 @item -mexplicit-relocs
8385 @itemx -mno-explicit-relocs
8386 @opindex mexplicit-relocs
8387 @opindex mno-explicit-relocs
8388 Older Alpha assemblers provided no way to generate symbol relocations
8389 except via assembler macros. Use of these macros does not allow
8390 optimal instruction scheduling. GNU binutils as of version 2.12
8391 supports a new syntax that allows the compiler to explicitly mark
8392 which relocations should apply to which instructions. This option
8393 is mostly useful for debugging, as GCC detects the capabilities of
8394 the assembler when it is built and sets the default accordingly.
8398 @opindex msmall-data
8399 @opindex mlarge-data
8400 When @option{-mexplicit-relocs} is in effect, static data is
8401 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8402 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8403 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8404 16-bit relocations off of the @code{$gp} register. This limits the
8405 size of the small data area to 64KB, but allows the variables to be
8406 directly accessed via a single instruction.
8408 The default is @option{-mlarge-data}. With this option the data area
8409 is limited to just below 2GB@. Programs that require more than 2GB of
8410 data must use @code{malloc} or @code{mmap} to allocate the data in the
8411 heap instead of in the program's data segment.
8413 When generating code for shared libraries, @option{-fpic} implies
8414 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8418 @opindex msmall-text
8419 @opindex mlarge-text
8420 When @option{-msmall-text} is used, the compiler assumes that the
8421 code of the entire program (or shared library) fits in 4MB, and is
8422 thus reachable with a branch instruction. When @option{-msmall-data}
8423 is used, the compiler can assume that all local symbols share the
8424 same @code{$gp} value, and thus reduce the number of instructions
8425 required for a function call from 4 to 1.
8427 The default is @option{-mlarge-text}.
8429 @item -mcpu=@var{cpu_type}
8431 Set the instruction set and instruction scheduling parameters for
8432 machine type @var{cpu_type}. You can specify either the @samp{EV}
8433 style name or the corresponding chip number. GCC supports scheduling
8434 parameters for the EV4, EV5 and EV6 family of processors and will
8435 choose the default values for the instruction set from the processor
8436 you specify. If you do not specify a processor type, GCC will default
8437 to the processor on which the compiler was built.
8439 Supported values for @var{cpu_type} are
8445 Schedules as an EV4 and has no instruction set extensions.
8449 Schedules as an EV5 and has no instruction set extensions.
8453 Schedules as an EV5 and supports the BWX extension.
8458 Schedules as an EV5 and supports the BWX and MAX extensions.
8462 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8466 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8469 @item -mtune=@var{cpu_type}
8471 Set only the instruction scheduling parameters for machine type
8472 @var{cpu_type}. The instruction set is not changed.
8474 @item -mmemory-latency=@var{time}
8475 @opindex mmemory-latency
8476 Sets the latency the scheduler should assume for typical memory
8477 references as seen by the application. This number is highly
8478 dependent on the memory access patterns used by the application
8479 and the size of the external cache on the machine.
8481 Valid options for @var{time} are
8485 A decimal number representing clock cycles.
8491 The compiler contains estimates of the number of clock cycles for
8492 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8493 (also called Dcache, Scache, and Bcache), as well as to main memory.
8494 Note that L3 is only valid for EV5.
8499 @node DEC Alpha/VMS Options
8500 @subsection DEC Alpha/VMS Options
8502 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8505 @item -mvms-return-codes
8506 @opindex mvms-return-codes
8507 Return VMS condition codes from main. The default is to return POSIX
8508 style condition (e.g.@ error) codes.
8512 @subsection FRV Options
8519 Only use the first 32 general purpose registers.
8524 Use all 64 general purpose registers.
8529 Use only the first 32 floating point registers.
8534 Use all 64 floating point registers
8537 @opindex mhard-float
8539 Use hardware instructions for floating point operations.
8542 @opindex msoft-float
8544 Use library routines for floating point operations.
8549 Dynamically allocate condition code registers.
8554 Do not try to dynamically allocate condition code registers, only
8555 use @code{icc0} and @code{fcc0}.
8560 Change ABI to use double word insns.
8565 Do not use double word instructions.
8570 Use floating point double instructions.
8575 Do not use floating point double instructions.
8580 Use media instructions.
8585 Do not use media instructions.
8590 Use multiply and add/subtract instructions.
8595 Do not use multiply and add/subtract instructions.
8600 Select the FDPIC ABI, that uses function descriptors to represent
8601 pointers to functions. Without any PIC/PIE-related options, it
8602 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
8603 assumes GOT entries and small data are within a 12-bit range from the
8604 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
8605 are computed with 32 bits.
8608 @opindex minline-plt
8610 Enable inlining of PLT entries in function calls to functions that are
8611 not known to bind locally. It has no effect without @option{-mfdpic}.
8612 It's enabled by default if optimizing for speed and compiling for
8613 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
8614 optimization option such as @option{-O3} or above is present in the
8620 Assume a large TLS segment when generating thread-local code.
8625 Do not assume a large TLS segment when generating thread-local code.
8630 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
8631 that is known to be in read-only sections. It's enabled by default,
8632 except for @option{-fpic} or @option{-fpie}: even though it may help
8633 make the global offset table smaller, it trades 1 instruction for 4.
8634 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
8635 one of which may be shared by multiple symbols, and it avoids the need
8636 for a GOT entry for the referenced symbol, so it's more likely to be a
8637 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
8639 @item -multilib-library-pic
8640 @opindex multilib-library-pic
8642 Link with the (library, not FD) pic libraries. It's implied by
8643 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
8644 @option{-fpic} without @option{-mfdpic}. You should never have to use
8650 Follow the EABI requirement of always creating a frame pointer whenever
8651 a stack frame is allocated. This option is enabled by default and can
8652 be disabled with @option{-mno-linked-fp}.
8655 @opindex mlong-calls
8657 Use indirect addressing to call functions outside the current
8658 compilation unit. This allows the functions to be placed anywhere
8659 within the 32-bit address space.
8661 @item -malign-labels
8662 @opindex malign-labels
8664 Try to align labels to an 8-byte boundary by inserting nops into the
8665 previous packet. This option only has an effect when VLIW packing
8666 is enabled. It doesn't create new packets; it merely adds nops to
8670 @opindex mlibrary-pic
8672 Generate position-independent EABI code.
8677 Use only the first four media accumulator registers.
8682 Use all eight media accumulator registers.
8687 Pack VLIW instructions.
8692 Do not pack VLIW instructions.
8697 Do not mark ABI switches in e_flags.
8702 Enable the use of conditional-move instructions (default).
8704 This switch is mainly for debugging the compiler and will likely be removed
8705 in a future version.
8707 @item -mno-cond-move
8708 @opindex mno-cond-move
8710 Disable the use of conditional-move instructions.
8712 This switch is mainly for debugging the compiler and will likely be removed
8713 in a future version.
8718 Enable the use of conditional set instructions (default).
8720 This switch is mainly for debugging the compiler and will likely be removed
8721 in a future version.
8726 Disable the use of conditional set instructions.
8728 This switch is mainly for debugging the compiler and will likely be removed
8729 in a future version.
8734 Enable the use of conditional execution (default).
8736 This switch is mainly for debugging the compiler and will likely be removed
8737 in a future version.
8739 @item -mno-cond-exec
8740 @opindex mno-cond-exec
8742 Disable the use of conditional execution.
8744 This switch is mainly for debugging the compiler and will likely be removed
8745 in a future version.
8748 @opindex mvliw-branch
8750 Run a pass to pack branches into VLIW instructions (default).
8752 This switch is mainly for debugging the compiler and will likely be removed
8753 in a future version.
8755 @item -mno-vliw-branch
8756 @opindex mno-vliw-branch
8758 Do not run a pass to pack branches into VLIW instructions.
8760 This switch is mainly for debugging the compiler and will likely be removed
8761 in a future version.
8763 @item -mmulti-cond-exec
8764 @opindex mmulti-cond-exec
8766 Enable optimization of @code{&&} and @code{||} in conditional execution
8769 This switch is mainly for debugging the compiler and will likely be removed
8770 in a future version.
8772 @item -mno-multi-cond-exec
8773 @opindex mno-multi-cond-exec
8775 Disable optimization of @code{&&} and @code{||} in conditional execution.
8777 This switch is mainly for debugging the compiler and will likely be removed
8778 in a future version.
8780 @item -mnested-cond-exec
8781 @opindex mnested-cond-exec
8783 Enable nested conditional execution optimizations (default).
8785 This switch is mainly for debugging the compiler and will likely be removed
8786 in a future version.
8788 @item -mno-nested-cond-exec
8789 @opindex mno-nested-cond-exec
8791 Disable nested conditional execution optimizations.
8793 This switch is mainly for debugging the compiler and will likely be removed
8794 in a future version.
8796 @item -moptimize-membar
8797 @opindex moptimize-membar
8799 This switch removes redundant @code{membar} instructions from the
8800 compiler generated code. It is enabled by default.
8802 @item -mno-optimize-membar
8803 @opindex mno-optimize-membar
8805 This switch disables the automatic removal of redundant @code{membar}
8806 instructions from the generated code.
8808 @item -mtomcat-stats
8809 @opindex mtomcat-stats
8811 Cause gas to print out tomcat statistics.
8813 @item -mcpu=@var{cpu}
8816 Select the processor type for which to generate code. Possible values are
8817 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
8818 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
8822 @node GNU/Linux Options
8823 @subsection GNU/Linux Options
8825 These @samp{-m} options are defined for GNU/Linux targets:
8830 Use the GNU C library instead of uClibc. This is the default except
8831 on @samp{*-*-linux-*uclibc*} targets.
8835 Use uClibc instead of the GNU C library. This is the default on
8836 @samp{*-*-linux-*uclibc*} targets.
8839 @node H8/300 Options
8840 @subsection H8/300 Options
8842 These @samp{-m} options are defined for the H8/300 implementations:
8847 Shorten some address references at link time, when possible; uses the
8848 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8849 ld, Using ld}, for a fuller description.
8853 Generate code for the H8/300H@.
8857 Generate code for the H8S@.
8861 Generate code for the H8S and H8/300H in the normal mode. This switch
8862 must be used either with @option{-mh} or @option{-ms}.
8866 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
8870 Make @code{int} data 32 bits by default.
8874 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
8875 The default for the H8/300H and H8S is to align longs and floats on 4
8877 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8878 This option has no effect on the H8/300.
8882 @subsection HPPA Options
8883 @cindex HPPA Options
8885 These @samp{-m} options are defined for the HPPA family of computers:
8888 @item -march=@var{architecture-type}
8890 Generate code for the specified architecture. The choices for
8891 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8892 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8893 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8894 architecture option for your machine. Code compiled for lower numbered
8895 architectures will run on higher numbered architectures, but not the
8899 @itemx -mpa-risc-1-1
8900 @itemx -mpa-risc-2-0
8901 @opindex mpa-risc-1-0
8902 @opindex mpa-risc-1-1
8903 @opindex mpa-risc-2-0
8904 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8907 @opindex mbig-switch
8908 Generate code suitable for big switch tables. Use this option only if
8909 the assembler/linker complain about out of range branches within a switch
8912 @item -mjump-in-delay
8913 @opindex mjump-in-delay
8914 Fill delay slots of function calls with unconditional jump instructions
8915 by modifying the return pointer for the function call to be the target
8916 of the conditional jump.
8918 @item -mdisable-fpregs
8919 @opindex mdisable-fpregs
8920 Prevent floating point registers from being used in any manner. This is
8921 necessary for compiling kernels which perform lazy context switching of
8922 floating point registers. If you use this option and attempt to perform
8923 floating point operations, the compiler will abort.
8925 @item -mdisable-indexing
8926 @opindex mdisable-indexing
8927 Prevent the compiler from using indexing address modes. This avoids some
8928 rather obscure problems when compiling MIG generated code under MACH@.
8930 @item -mno-space-regs
8931 @opindex mno-space-regs
8932 Generate code that assumes the target has no space registers. This allows
8933 GCC to generate faster indirect calls and use unscaled index address modes.
8935 Such code is suitable for level 0 PA systems and kernels.
8937 @item -mfast-indirect-calls
8938 @opindex mfast-indirect-calls
8939 Generate code that assumes calls never cross space boundaries. This
8940 allows GCC to emit code which performs faster indirect calls.
8942 This option will not work in the presence of shared libraries or nested
8945 @item -mfixed-range=@var{register-range}
8946 @opindex mfixed-range
8947 Generate code treating the given register range as fixed registers.
8948 A fixed register is one that the register allocator can not use. This is
8949 useful when compiling kernel code. A register range is specified as
8950 two registers separated by a dash. Multiple register ranges can be
8951 specified separated by a comma.
8953 @item -mlong-load-store
8954 @opindex mlong-load-store
8955 Generate 3-instruction load and store sequences as sometimes required by
8956 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8959 @item -mportable-runtime
8960 @opindex mportable-runtime
8961 Use the portable calling conventions proposed by HP for ELF systems.
8965 Enable the use of assembler directives only GAS understands.
8967 @item -mschedule=@var{cpu-type}
8969 Schedule code according to the constraints for the machine type
8970 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8971 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8972 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8973 proper scheduling option for your machine. The default scheduling is
8977 @opindex mlinker-opt
8978 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8979 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8980 linkers in which they give bogus error messages when linking some programs.
8983 @opindex msoft-float
8984 Generate output containing library calls for floating point.
8985 @strong{Warning:} the requisite libraries are not available for all HPPA
8986 targets. Normally the facilities of the machine's usual C compiler are
8987 used, but this cannot be done directly in cross-compilation. You must make
8988 your own arrangements to provide suitable library functions for
8989 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8990 does provide software floating point support.
8992 @option{-msoft-float} changes the calling convention in the output file;
8993 therefore, it is only useful if you compile @emph{all} of a program with
8994 this option. In particular, you need to compile @file{libgcc.a}, the
8995 library that comes with GCC, with @option{-msoft-float} in order for
9000 Generate the predefine, @code{_SIO}, for server IO@. The default is
9001 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
9002 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
9003 options are available under HP-UX and HI-UX@.
9007 Use GNU ld specific options. This passes @option{-shared} to ld when
9008 building a shared library. It is the default when GCC is configured,
9009 explicitly or implicitly, with the GNU linker. This option does not
9010 have any affect on which ld is called, it only changes what parameters
9011 are passed to that ld. The ld that is called is determined by the
9012 @option{--with-ld} configure option, GCC's program search path, and
9013 finally by the user's @env{PATH}. The linker used by GCC can be printed
9014 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
9015 on the 64 bit HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9019 Use HP ld specific options. This passes @option{-b} to ld when building
9020 a shared library and passes @option{+Accept TypeMismatch} to ld on all
9021 links. It is the default when GCC is configured, explicitly or
9022 implicitly, with the HP linker. This option does not have any affect on
9023 which ld is called, it only changes what parameters are passed to that
9024 ld. The ld that is called is determined by the @option{--with-ld}
9025 configure option, GCC's program search path, and finally by the user's
9026 @env{PATH}. The linker used by GCC can be printed using @samp{which
9027 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
9028 HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9031 @opindex mno-long-calls
9032 Generate code that uses long call sequences. This ensures that a call
9033 is always able to reach linker generated stubs. The default is to generate
9034 long calls only when the distance from the call site to the beginning
9035 of the function or translation unit, as the case may be, exceeds a
9036 predefined limit set by the branch type being used. The limits for
9037 normal calls are 7,600,000 and 240,000 bytes, respectively for the
9038 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
9041 Distances are measured from the beginning of functions when using the
9042 @option{-ffunction-sections} option, or when using the @option{-mgas}
9043 and @option{-mno-portable-runtime} options together under HP-UX with
9046 It is normally not desirable to use this option as it will degrade
9047 performance. However, it may be useful in large applications,
9048 particularly when partial linking is used to build the application.
9050 The types of long calls used depends on the capabilities of the
9051 assembler and linker, and the type of code being generated. The
9052 impact on systems that support long absolute calls, and long pic
9053 symbol-difference or pc-relative calls should be relatively small.
9054 However, an indirect call is used on 32-bit ELF systems in pic code
9055 and it is quite long.
9057 @item -munix=@var{unix-std}
9059 Generate compiler predefines and select a startfile for the specified
9060 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
9061 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
9062 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
9063 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
9064 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
9067 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
9068 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
9069 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
9070 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
9071 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
9072 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
9074 It is @emph{important} to note that this option changes the interfaces
9075 for various library routines. It also affects the operational behavior
9076 of the C library. Thus, @emph{extreme} care is needed in using this
9079 Library code that is intended to operate with more than one UNIX
9080 standard must test, set and restore the variable @var{__xpg4_extended_mask}
9081 as appropriate. Most GNU software doesn't provide this capability.
9085 Suppress the generation of link options to search libdld.sl when the
9086 @option{-static} option is specified on HP-UX 10 and later.
9090 The HP-UX implementation of setlocale in libc has a dependency on
9091 libdld.sl. There isn't an archive version of libdld.sl. Thus,
9092 when the @option{-static} option is specified, special link options
9093 are needed to resolve this dependency.
9095 On HP-UX 10 and later, the GCC driver adds the necessary options to
9096 link with libdld.sl when the @option{-static} option is specified.
9097 This causes the resulting binary to be dynamic. On the 64-bit port,
9098 the linkers generate dynamic binaries by default in any case. The
9099 @option{-nolibdld} option can be used to prevent the GCC driver from
9100 adding these link options.
9104 Add support for multithreading with the @dfn{dce thread} library
9105 under HP-UX@. This option sets flags for both the preprocessor and
9109 @node i386 and x86-64 Options
9110 @subsection Intel 386 and AMD x86-64 Options
9111 @cindex i386 Options
9112 @cindex x86-64 Options
9113 @cindex Intel 386 Options
9114 @cindex AMD x86-64 Options
9116 These @samp{-m} options are defined for the i386 and x86-64 family of
9120 @item -mtune=@var{cpu-type}
9122 Tune to @var{cpu-type} everything applicable about the generated code, except
9123 for the ABI and the set of available instructions. The choices for
9127 Produce code optimized for the most common IA32/AMD64/EM64T processors.
9128 If you know the CPU on which your code will run, then you should use
9129 the corresponding @option{-mtune} option instead of
9130 @option{-mtune=generic}. But, if you do not know exactly what CPU users
9131 of your application will have, then you should use this option.
9133 As new processors are deployed in the marketplace, the behavior of this
9134 option will change. Therefore, if you upgrade to a newer version of
9135 GCC, the code generated option will change to reflect the processors
9136 that were most common when that version of GCC was released.
9138 There is no @option{-march=generic} option because @option{-march}
9139 indicates the instruction set the compiler can use, and there is no
9140 generic instruction set applicable to all processors. In contrast,
9141 @option{-mtune} indicates the processor (or, in this case, collection of
9142 processors) for which the code is optimized.
9144 Original Intel's i386 CPU@.
9146 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
9148 Intel Pentium CPU with no MMX support.
9150 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
9152 Intel PentiumPro CPU@.
9154 Same as @code{generic}, but when used as @code{march} option, PentiumPro
9155 instruction set will be used, so the code will run on all i686 familly chips.
9157 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
9158 @item pentium3, pentium3m
9159 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
9162 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
9163 support. Used by Centrino notebooks.
9164 @item pentium4, pentium4m
9165 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
9167 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
9170 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
9171 SSE2 and SSE3 instruction set support.
9173 AMD K6 CPU with MMX instruction set support.
9175 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
9176 @item athlon, athlon-tbird
9177 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
9179 @item athlon-4, athlon-xp, athlon-mp
9180 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
9181 instruction set support.
9182 @item k8, opteron, athlon64, athlon-fx
9183 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
9184 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
9186 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
9189 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
9190 instruction set support.
9192 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
9193 implemented for this chip.)
9195 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
9196 implemented for this chip.)
9199 While picking a specific @var{cpu-type} will schedule things appropriately
9200 for that particular chip, the compiler will not generate any code that
9201 does not run on the i386 without the @option{-march=@var{cpu-type}} option
9204 @item -march=@var{cpu-type}
9206 Generate instructions for the machine type @var{cpu-type}. The choices
9207 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
9208 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
9210 @item -mcpu=@var{cpu-type}
9212 A deprecated synonym for @option{-mtune}.
9221 @opindex mpentiumpro
9222 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
9223 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
9224 These synonyms are deprecated.
9226 @item -mfpmath=@var{unit}
9228 Generate floating point arithmetics for selected unit @var{unit}. The choices
9233 Use the standard 387 floating point coprocessor present majority of chips and
9234 emulated otherwise. Code compiled with this option will run almost everywhere.
9235 The temporary results are computed in 80bit precision instead of precision
9236 specified by the type resulting in slightly different results compared to most
9237 of other chips. See @option{-ffloat-store} for more detailed description.
9239 This is the default choice for i386 compiler.
9242 Use scalar floating point instructions present in the SSE instruction set.
9243 This instruction set is supported by Pentium3 and newer chips, in the AMD line
9244 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
9245 instruction set supports only single precision arithmetics, thus the double and
9246 extended precision arithmetics is still done using 387. Later version, present
9247 only in Pentium4 and the future AMD x86-64 chips supports double precision
9250 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
9251 or @option{-msse2} switches to enable SSE extensions and make this option
9252 effective. For the x86-64 compiler, these extensions are enabled by default.
9254 The resulting code should be considerably faster in the majority of cases and avoid
9255 the numerical instability problems of 387 code, but may break some existing
9256 code that expects temporaries to be 80bit.
9258 This is the default choice for the x86-64 compiler.
9261 Attempt to utilize both instruction sets at once. This effectively double the
9262 amount of available registers and on chips with separate execution units for
9263 387 and SSE the execution resources too. Use this option with care, as it is
9264 still experimental, because the GCC register allocator does not model separate
9265 functional units well resulting in instable performance.
9268 @item -masm=@var{dialect}
9269 @opindex masm=@var{dialect}
9270 Output asm instructions using selected @var{dialect}. Supported
9271 choices are @samp{intel} or @samp{att} (the default one). Darwin does
9272 not support @samp{intel}.
9277 @opindex mno-ieee-fp
9278 Control whether or not the compiler uses IEEE floating point
9279 comparisons. These handle correctly the case where the result of a
9280 comparison is unordered.
9283 @opindex msoft-float
9284 Generate output containing library calls for floating point.
9285 @strong{Warning:} the requisite libraries are not part of GCC@.
9286 Normally the facilities of the machine's usual C compiler are used, but
9287 this can't be done directly in cross-compilation. You must make your
9288 own arrangements to provide suitable library functions for
9291 On machines where a function returns floating point results in the 80387
9292 register stack, some floating point opcodes may be emitted even if
9293 @option{-msoft-float} is used.
9295 @item -mno-fp-ret-in-387
9296 @opindex mno-fp-ret-in-387
9297 Do not use the FPU registers for return values of functions.
9299 The usual calling convention has functions return values of types
9300 @code{float} and @code{double} in an FPU register, even if there
9301 is no FPU@. The idea is that the operating system should emulate
9304 The option @option{-mno-fp-ret-in-387} causes such values to be returned
9305 in ordinary CPU registers instead.
9307 @item -mno-fancy-math-387
9308 @opindex mno-fancy-math-387
9309 Some 387 emulators do not support the @code{sin}, @code{cos} and
9310 @code{sqrt} instructions for the 387. Specify this option to avoid
9311 generating those instructions. This option is the default on FreeBSD,
9312 OpenBSD and NetBSD@. This option is overridden when @option{-march}
9313 indicates that the target cpu will always have an FPU and so the
9314 instruction will not need emulation. As of revision 2.6.1, these
9315 instructions are not generated unless you also use the
9316 @option{-funsafe-math-optimizations} switch.
9318 @item -malign-double
9319 @itemx -mno-align-double
9320 @opindex malign-double
9321 @opindex mno-align-double
9322 Control whether GCC aligns @code{double}, @code{long double}, and
9323 @code{long long} variables on a two word boundary or a one word
9324 boundary. Aligning @code{double} variables on a two word boundary will
9325 produce code that runs somewhat faster on a @samp{Pentium} at the
9326 expense of more memory.
9328 @strong{Warning:} if you use the @option{-malign-double} switch,
9329 structures containing the above types will be aligned differently than
9330 the published application binary interface specifications for the 386
9331 and will not be binary compatible with structures in code compiled
9332 without that switch.
9334 @item -m96bit-long-double
9335 @itemx -m128bit-long-double
9336 @opindex m96bit-long-double
9337 @opindex m128bit-long-double
9338 These switches control the size of @code{long double} type. The i386
9339 application binary interface specifies the size to be 96 bits,
9340 so @option{-m96bit-long-double} is the default in 32 bit mode.
9342 Modern architectures (Pentium and newer) would prefer @code{long double}
9343 to be aligned to an 8 or 16 byte boundary. In arrays or structures
9344 conforming to the ABI, this would not be possible. So specifying a
9345 @option{-m128bit-long-double} will align @code{long double}
9346 to a 16 byte boundary by padding the @code{long double} with an additional
9349 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
9350 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
9352 Notice that neither of these options enable any extra precision over the x87
9353 standard of 80 bits for a @code{long double}.
9355 @strong{Warning:} if you override the default value for your target ABI, the
9356 structures and arrays containing @code{long double} variables will change
9357 their size as well as function calling convention for function taking
9358 @code{long double} will be modified. Hence they will not be binary
9359 compatible with arrays or structures in code compiled without that switch.
9361 @item -mmlarge-data-threshold=@var{number}
9362 @opindex mlarge-data-threshold=@var{number}
9363 When @option{-mcmodel=medium} is specified, the data greater than
9364 @var{threshold} are placed in large data section. This value must be the
9365 same across all object linked into the binary and defaults to 65535.
9368 @itemx -mno-svr3-shlib
9369 @opindex msvr3-shlib
9370 @opindex mno-svr3-shlib
9371 Control whether GCC places uninitialized local variables into the
9372 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
9373 into @code{bss}. These options are meaningful only on System V Release 3.
9377 Use a different function-calling convention, in which functions that
9378 take a fixed number of arguments return with the @code{ret} @var{num}
9379 instruction, which pops their arguments while returning. This saves one
9380 instruction in the caller since there is no need to pop the arguments
9383 You can specify that an individual function is called with this calling
9384 sequence with the function attribute @samp{stdcall}. You can also
9385 override the @option{-mrtd} option by using the function attribute
9386 @samp{cdecl}. @xref{Function Attributes}.
9388 @strong{Warning:} this calling convention is incompatible with the one
9389 normally used on Unix, so you cannot use it if you need to call
9390 libraries compiled with the Unix compiler.
9392 Also, you must provide function prototypes for all functions that
9393 take variable numbers of arguments (including @code{printf});
9394 otherwise incorrect code will be generated for calls to those
9397 In addition, seriously incorrect code will result if you call a
9398 function with too many arguments. (Normally, extra arguments are
9399 harmlessly ignored.)
9401 @item -mregparm=@var{num}
9403 Control how many registers are used to pass integer arguments. By
9404 default, no registers are used to pass arguments, and at most 3
9405 registers can be used. You can control this behavior for a specific
9406 function by using the function attribute @samp{regparm}.
9407 @xref{Function Attributes}.
9409 @strong{Warning:} if you use this switch, and
9410 @var{num} is nonzero, then you must build all modules with the same
9411 value, including any libraries. This includes the system libraries and
9415 @opindex msseregparm
9416 Use SSE register passing conventions for float and double arguments
9417 and return values. You can control this behavior for a specific
9418 function by using the function attribute @samp{sseregparm}.
9419 @xref{Function Attributes}.
9421 @strong{Warning:} if you use this switch then you must build all
9422 modules with the same value, including any libraries. This includes
9423 the system libraries and startup modules.
9425 @item -mpreferred-stack-boundary=@var{num}
9426 @opindex mpreferred-stack-boundary
9427 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
9428 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
9429 the default is 4 (16 bytes or 128 bits), except when optimizing for code
9430 size (@option{-Os}), in which case the default is the minimum correct
9431 alignment (4 bytes for x86, and 8 bytes for x86-64).
9433 On Pentium and PentiumPro, @code{double} and @code{long double} values
9434 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
9435 suffer significant run time performance penalties. On Pentium III, the
9436 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
9437 penalties if it is not 16 byte aligned.
9439 To ensure proper alignment of this values on the stack, the stack boundary
9440 must be as aligned as that required by any value stored on the stack.
9441 Further, every function must be generated such that it keeps the stack
9442 aligned. Thus calling a function compiled with a higher preferred
9443 stack boundary from a function compiled with a lower preferred stack
9444 boundary will most likely misalign the stack. It is recommended that
9445 libraries that use callbacks always use the default setting.
9447 This extra alignment does consume extra stack space, and generally
9448 increases code size. Code that is sensitive to stack space usage, such
9449 as embedded systems and operating system kernels, may want to reduce the
9450 preferred alignment to @option{-mpreferred-stack-boundary=2}.
9468 These switches enable or disable the use of instructions in the MMX,
9469 SSE, SSE2 or 3DNow! extended instruction sets. These extensions are
9470 also available as built-in functions: see @ref{X86 Built-in Functions},
9471 for details of the functions enabled and disabled by these switches.
9473 To have SSE/SSE2 instructions generated automatically from floating-point
9474 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
9476 These options will enable GCC to use these extended instructions in
9477 generated code, even without @option{-mfpmath=sse}. Applications which
9478 perform runtime CPU detection must compile separate files for each
9479 supported architecture, using the appropriate flags. In particular,
9480 the file containing the CPU detection code should be compiled without
9485 Use special versions of certain libm routines that come with an SSE
9486 ABI and an SSE implementation. Useful together with @option{-mfpmath=sse}
9487 to avoid moving values between SSE registers and the x87 FP stack.
9490 @itemx -mno-push-args
9492 @opindex mno-push-args
9493 Use PUSH operations to store outgoing parameters. This method is shorter
9494 and usually equally fast as method using SUB/MOV operations and is enabled
9495 by default. In some cases disabling it may improve performance because of
9496 improved scheduling and reduced dependencies.
9498 @item -maccumulate-outgoing-args
9499 @opindex maccumulate-outgoing-args
9500 If enabled, the maximum amount of space required for outgoing arguments will be
9501 computed in the function prologue. This is faster on most modern CPUs
9502 because of reduced dependencies, improved scheduling and reduced stack usage
9503 when preferred stack boundary is not equal to 2. The drawback is a notable
9504 increase in code size. This switch implies @option{-mno-push-args}.
9508 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
9509 on thread-safe exception handling must compile and link all code with the
9510 @option{-mthreads} option. When compiling, @option{-mthreads} defines
9511 @option{-D_MT}; when linking, it links in a special thread helper library
9512 @option{-lmingwthrd} which cleans up per thread exception handling data.
9514 @item -mno-align-stringops
9515 @opindex mno-align-stringops
9516 Do not align destination of inlined string operations. This switch reduces
9517 code size and improves performance in case the destination is already aligned,
9518 but GCC doesn't know about it.
9520 @item -minline-all-stringops
9521 @opindex minline-all-stringops
9522 By default GCC inlines string operations only when destination is known to be
9523 aligned at least to 4 byte boundary. This enables more inlining, increase code
9524 size, but may improve performance of code that depends on fast memcpy, strlen
9525 and memset for short lengths.
9527 @item -momit-leaf-frame-pointer
9528 @opindex momit-leaf-frame-pointer
9529 Don't keep the frame pointer in a register for leaf functions. This
9530 avoids the instructions to save, set up and restore frame pointers and
9531 makes an extra register available in leaf functions. The option
9532 @option{-fomit-frame-pointer} removes the frame pointer for all functions
9533 which might make debugging harder.
9535 @item -mtls-direct-seg-refs
9536 @itemx -mno-tls-direct-seg-refs
9537 @opindex mtls-direct-seg-refs
9538 Controls whether TLS variables may be accessed with offsets from the
9539 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
9540 or whether the thread base pointer must be added. Whether or not this
9541 is legal depends on the operating system, and whether it maps the
9542 segment to cover the entire TLS area.
9544 For systems that use GNU libc, the default is on.
9547 These @samp{-m} switches are supported in addition to the above
9548 on AMD x86-64 processors in 64-bit environments.
9555 Generate code for a 32-bit or 64-bit environment.
9556 The 32-bit environment sets int, long and pointer to 32 bits and
9557 generates code that runs on any i386 system.
9558 The 64-bit environment sets int to 32 bits and long and pointer
9559 to 64 bits and generates code for AMD's x86-64 architecture.
9562 @opindex no-red-zone
9563 Do not use a so called red zone for x86-64 code. The red zone is mandated
9564 by the x86-64 ABI, it is a 128-byte area beyond the location of the
9565 stack pointer that will not be modified by signal or interrupt handlers
9566 and therefore can be used for temporary data without adjusting the stack
9567 pointer. The flag @option{-mno-red-zone} disables this red zone.
9569 @item -mcmodel=small
9570 @opindex mcmodel=small
9571 Generate code for the small code model: the program and its symbols must
9572 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
9573 Programs can be statically or dynamically linked. This is the default
9576 @item -mcmodel=kernel
9577 @opindex mcmodel=kernel
9578 Generate code for the kernel code model. The kernel runs in the
9579 negative 2 GB of the address space.
9580 This model has to be used for Linux kernel code.
9582 @item -mcmodel=medium
9583 @opindex mcmodel=medium
9584 Generate code for the medium model: The program is linked in the lower 2
9585 GB of the address space but symbols can be located anywhere in the
9586 address space. Programs can be statically or dynamically linked, but
9587 building of shared libraries are not supported with the medium model.
9589 @item -mcmodel=large
9590 @opindex mcmodel=large
9591 Generate code for the large model: This model makes no assumptions
9592 about addresses and sizes of sections. Currently GCC does not implement
9597 @subsection IA-64 Options
9598 @cindex IA-64 Options
9600 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9604 @opindex mbig-endian
9605 Generate code for a big endian target. This is the default for HP-UX@.
9607 @item -mlittle-endian
9608 @opindex mlittle-endian
9609 Generate code for a little endian target. This is the default for AIX5
9616 Generate (or don't) code for the GNU assembler. This is the default.
9617 @c Also, this is the default if the configure option @option{--with-gnu-as}
9624 Generate (or don't) code for the GNU linker. This is the default.
9625 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9630 Generate code that does not use a global pointer register. The result
9631 is not position independent code, and violates the IA-64 ABI@.
9633 @item -mvolatile-asm-stop
9634 @itemx -mno-volatile-asm-stop
9635 @opindex mvolatile-asm-stop
9636 @opindex mno-volatile-asm-stop
9637 Generate (or don't) a stop bit immediately before and after volatile asm
9640 @item -mregister-names
9641 @itemx -mno-register-names
9642 @opindex mregister-names
9643 @opindex mno-register-names
9644 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9645 the stacked registers. This may make assembler output more readable.
9651 Disable (or enable) optimizations that use the small data section. This may
9652 be useful for working around optimizer bugs.
9655 @opindex mconstant-gp
9656 Generate code that uses a single constant global pointer value. This is
9657 useful when compiling kernel code.
9661 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9662 This is useful when compiling firmware code.
9664 @item -minline-float-divide-min-latency
9665 @opindex minline-float-divide-min-latency
9666 Generate code for inline divides of floating point values
9667 using the minimum latency algorithm.
9669 @item -minline-float-divide-max-throughput
9670 @opindex minline-float-divide-max-throughput
9671 Generate code for inline divides of floating point values
9672 using the maximum throughput algorithm.
9674 @item -minline-int-divide-min-latency
9675 @opindex minline-int-divide-min-latency
9676 Generate code for inline divides of integer values
9677 using the minimum latency algorithm.
9679 @item -minline-int-divide-max-throughput
9680 @opindex minline-int-divide-max-throughput
9681 Generate code for inline divides of integer values
9682 using the maximum throughput algorithm.
9684 @item -minline-sqrt-min-latency
9685 @opindex minline-sqrt-min-latency
9686 Generate code for inline square roots
9687 using the minimum latency algorithm.
9689 @item -minline-sqrt-max-throughput
9690 @opindex minline-sqrt-max-throughput
9691 Generate code for inline square roots
9692 using the maximum throughput algorithm.
9694 @item -mno-dwarf2-asm
9696 @opindex mno-dwarf2-asm
9697 @opindex mdwarf2-asm
9698 Don't (or do) generate assembler code for the DWARF2 line number debugging
9699 info. This may be useful when not using the GNU assembler.
9701 @item -mearly-stop-bits
9702 @itemx -mno-early-stop-bits
9703 @opindex mearly-stop-bits
9704 @opindex mno-early-stop-bits
9705 Allow stop bits to be placed earlier than immediately preceding the
9706 instruction that triggered the stop bit. This can improve instruction
9707 scheduling, but does not always do so.
9709 @item -mfixed-range=@var{register-range}
9710 @opindex mfixed-range
9711 Generate code treating the given register range as fixed registers.
9712 A fixed register is one that the register allocator can not use. This is
9713 useful when compiling kernel code. A register range is specified as
9714 two registers separated by a dash. Multiple register ranges can be
9715 specified separated by a comma.
9717 @item -mtls-size=@var{tls-size}
9719 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
9722 @item -mtune=@var{cpu-type}
9724 Tune the instruction scheduling for a particular CPU, Valid values are
9725 itanium, itanium1, merced, itanium2, and mckinley.
9731 Add support for multithreading using the POSIX threads library. This
9732 option sets flags for both the preprocessor and linker. It does
9733 not affect the thread safety of object code produced by the compiler or
9734 that of libraries supplied with it. These are HP-UX specific flags.
9740 Generate code for a 32-bit or 64-bit environment.
9741 The 32-bit environment sets int, long and pointer to 32 bits.
9742 The 64-bit environment sets int to 32 bits and long and pointer
9743 to 64 bits. These are HP-UX specific flags.
9748 @subsection M32C Options
9749 @cindex M32C options
9752 @item -mcpu=@var{name}
9754 Select the CPU for which code is generated. @var{name} may be one of
9755 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
9756 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
9761 Specifies that the program will be run on the simulator. This causes
9762 an alternate runtime library to be linked in which supports, for
9763 example, file I/O. You must not use this option when generating
9764 programs that will run on real hardware; you must provide your own
9765 runtime library for whatever I/O functions are needed.
9767 @item -memregs=@var{number}
9769 Specifies the number of memory-based pseudo-registers GCC will use
9770 during code generation. These pseudo-registers will be used like real
9771 registers, so there is a tradeoff between GCC's ability to fit the
9772 code into available registers, and the performance penalty of using
9773 memory instead of registers. Note that all modules in a program must
9774 be compiled with the same value for this option. Because of that, you
9775 must not use this option with the default runtime libraries gcc
9780 @node M32R/D Options
9781 @subsection M32R/D Options
9782 @cindex M32R/D options
9784 These @option{-m} options are defined for Renesas M32R/D architectures:
9789 Generate code for the M32R/2@.
9793 Generate code for the M32R/X@.
9797 Generate code for the M32R@. This is the default.
9800 @opindex mmodel=small
9801 Assume all objects live in the lower 16MB of memory (so that their addresses
9802 can be loaded with the @code{ld24} instruction), and assume all subroutines
9803 are reachable with the @code{bl} instruction.
9804 This is the default.
9806 The addressability of a particular object can be set with the
9807 @code{model} attribute.
9809 @item -mmodel=medium
9810 @opindex mmodel=medium
9811 Assume objects may be anywhere in the 32-bit address space (the compiler
9812 will generate @code{seth/add3} instructions to load their addresses), and
9813 assume all subroutines are reachable with the @code{bl} instruction.
9816 @opindex mmodel=large
9817 Assume objects may be anywhere in the 32-bit address space (the compiler
9818 will generate @code{seth/add3} instructions to load their addresses), and
9819 assume subroutines may not be reachable with the @code{bl} instruction
9820 (the compiler will generate the much slower @code{seth/add3/jl}
9821 instruction sequence).
9824 @opindex msdata=none
9825 Disable use of the small data area. Variables will be put into
9826 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
9827 @code{section} attribute has been specified).
9828 This is the default.
9830 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
9831 Objects may be explicitly put in the small data area with the
9832 @code{section} attribute using one of these sections.
9835 @opindex msdata=sdata
9836 Put small global and static data in the small data area, but do not
9837 generate special code to reference them.
9841 Put small global and static data in the small data area, and generate
9842 special instructions to reference them.
9846 @cindex smaller data references
9847 Put global and static objects less than or equal to @var{num} bytes
9848 into the small data or bss sections instead of the normal data or bss
9849 sections. The default value of @var{num} is 8.
9850 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
9851 for this option to have any effect.
9853 All modules should be compiled with the same @option{-G @var{num}} value.
9854 Compiling with different values of @var{num} may or may not work; if it
9855 doesn't the linker will give an error message---incorrect code will not be
9860 Makes the M32R specific code in the compiler display some statistics
9861 that might help in debugging programs.
9864 @opindex malign-loops
9865 Align all loops to a 32-byte boundary.
9867 @item -mno-align-loops
9868 @opindex mno-align-loops
9869 Do not enforce a 32-byte alignment for loops. This is the default.
9871 @item -missue-rate=@var{number}
9872 @opindex missue-rate=@var{number}
9873 Issue @var{number} instructions per cycle. @var{number} can only be 1
9876 @item -mbranch-cost=@var{number}
9877 @opindex mbranch-cost=@var{number}
9878 @var{number} can only be 1 or 2. If it is 1 then branches will be
9879 preferred over conditional code, if it is 2, then the opposite will
9882 @item -mflush-trap=@var{number}
9883 @opindex mflush-trap=@var{number}
9884 Specifies the trap number to use to flush the cache. The default is
9885 12. Valid numbers are between 0 and 15 inclusive.
9887 @item -mno-flush-trap
9888 @opindex mno-flush-trap
9889 Specifies that the cache cannot be flushed by using a trap.
9891 @item -mflush-func=@var{name}
9892 @opindex mflush-func=@var{name}
9893 Specifies the name of the operating system function to call to flush
9894 the cache. The default is @emph{_flush_cache}, but a function call
9895 will only be used if a trap is not available.
9897 @item -mno-flush-func
9898 @opindex mno-flush-func
9899 Indicates that there is no OS function for flushing the cache.
9903 @node M680x0 Options
9904 @subsection M680x0 Options
9905 @cindex M680x0 options
9907 These are the @samp{-m} options defined for the 68000 series. The default
9908 values for these options depends on which style of 68000 was selected when
9909 the compiler was configured; the defaults for the most common choices are
9917 Generate output for a 68000. This is the default
9918 when the compiler is configured for 68000-based systems.
9920 Use this option for microcontrollers with a 68000 or EC000 core,
9921 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
9927 Generate output for a 68020. This is the default
9928 when the compiler is configured for 68020-based systems.
9932 Generate output containing 68881 instructions for floating point.
9933 This is the default for most 68020 systems unless @option{--nfp} was
9934 specified when the compiler was configured.
9938 Generate output for a 68030. This is the default when the compiler is
9939 configured for 68030-based systems.
9943 Generate output for a 68040. This is the default when the compiler is
9944 configured for 68040-based systems.
9946 This option inhibits the use of 68881/68882 instructions that have to be
9947 emulated by software on the 68040. Use this option if your 68040 does not
9948 have code to emulate those instructions.
9952 Generate output for a 68060. This is the default when the compiler is
9953 configured for 68060-based systems.
9955 This option inhibits the use of 68020 and 68881/68882 instructions that
9956 have to be emulated by software on the 68060. Use this option if your 68060
9957 does not have code to emulate those instructions.
9961 Generate output for a CPU32. This is the default
9962 when the compiler is configured for CPU32-based systems.
9964 Use this option for microcontrollers with a
9965 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
9966 68336, 68340, 68341, 68349 and 68360.
9970 Generate output for a 520X ``coldfire'' family cpu. This is the default
9971 when the compiler is configured for 520X-based systems.
9973 Use this option for microcontroller with a 5200 core, including
9974 the MCF5202, MCF5203, MCF5204 and MCF5202.
9978 Generate output for a ColdFire V4e family cpu (e.g.@: 547x/548x).
9979 This includes use of hardware floating point instructions.
9983 Generate output for a 68040, without using any of the new instructions.
9984 This results in code which can run relatively efficiently on either a
9985 68020/68881 or a 68030 or a 68040. The generated code does use the
9986 68881 instructions that are emulated on the 68040.
9990 Generate output for a 68060, without using any of the new instructions.
9991 This results in code which can run relatively efficiently on either a
9992 68020/68881 or a 68030 or a 68040. The generated code does use the
9993 68881 instructions that are emulated on the 68060.
9996 @opindex msoft-float
9997 Generate output containing library calls for floating point.
9998 @strong{Warning:} the requisite libraries are not available for all m68k
9999 targets. Normally the facilities of the machine's usual C compiler are
10000 used, but this can't be done directly in cross-compilation. You must
10001 make your own arrangements to provide suitable library functions for
10002 cross-compilation. The embedded targets @samp{m68k-*-aout} and
10003 @samp{m68k-*-coff} do provide software floating point support.
10007 Consider type @code{int} to be 16 bits wide, like @code{short int}.
10008 Additionally, parameters passed on the stack are also aligned to a
10009 16-bit boundary even on targets whose API mandates promotion to 32-bit.
10012 @opindex mnobitfield
10013 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
10014 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
10018 Do use the bit-field instructions. The @option{-m68020} option implies
10019 @option{-mbitfield}. This is the default if you use a configuration
10020 designed for a 68020.
10024 Use a different function-calling convention, in which functions
10025 that take a fixed number of arguments return with the @code{rtd}
10026 instruction, which pops their arguments while returning. This
10027 saves one instruction in the caller since there is no need to pop
10028 the arguments there.
10030 This calling convention is incompatible with the one normally
10031 used on Unix, so you cannot use it if you need to call libraries
10032 compiled with the Unix compiler.
10034 Also, you must provide function prototypes for all functions that
10035 take variable numbers of arguments (including @code{printf});
10036 otherwise incorrect code will be generated for calls to those
10039 In addition, seriously incorrect code will result if you call a
10040 function with too many arguments. (Normally, extra arguments are
10041 harmlessly ignored.)
10043 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
10044 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
10047 @itemx -mno-align-int
10048 @opindex malign-int
10049 @opindex mno-align-int
10050 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
10051 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
10052 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
10053 Aligning variables on 32-bit boundaries produces code that runs somewhat
10054 faster on processors with 32-bit busses at the expense of more memory.
10056 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
10057 align structures containing the above types differently than
10058 most published application binary interface specifications for the m68k.
10062 Use the pc-relative addressing mode of the 68000 directly, instead of
10063 using a global offset table. At present, this option implies @option{-fpic},
10064 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
10065 not presently supported with @option{-mpcrel}, though this could be supported for
10066 68020 and higher processors.
10068 @item -mno-strict-align
10069 @itemx -mstrict-align
10070 @opindex mno-strict-align
10071 @opindex mstrict-align
10072 Do not (do) assume that unaligned memory references will be handled by
10076 Generate code that allows the data segment to be located in a different
10077 area of memory from the text segment. This allows for execute in place in
10078 an environment without virtual memory management. This option implies
10081 @item -mno-sep-data
10082 Generate code that assumes that the data segment follows the text segment.
10083 This is the default.
10085 @item -mid-shared-library
10086 Generate code that supports shared libraries via the library ID method.
10087 This allows for execute in place and shared libraries in an environment
10088 without virtual memory management. This option implies @option{-fPIC}.
10090 @item -mno-id-shared-library
10091 Generate code that doesn't assume ID based shared libraries are being used.
10092 This is the default.
10094 @item -mshared-library-id=n
10095 Specified the identification number of the ID based shared library being
10096 compiled. Specifying a value of 0 will generate more compact code, specifying
10097 other values will force the allocation of that number to the current
10098 library but is no more space or time efficient than omitting this option.
10102 @node M68hc1x Options
10103 @subsection M68hc1x Options
10104 @cindex M68hc1x options
10106 These are the @samp{-m} options defined for the 68hc11 and 68hc12
10107 microcontrollers. The default values for these options depends on
10108 which style of microcontroller was selected when the compiler was configured;
10109 the defaults for the most common choices are given below.
10116 Generate output for a 68HC11. This is the default
10117 when the compiler is configured for 68HC11-based systems.
10123 Generate output for a 68HC12. This is the default
10124 when the compiler is configured for 68HC12-based systems.
10130 Generate output for a 68HCS12.
10132 @item -mauto-incdec
10133 @opindex mauto-incdec
10134 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
10141 Enable the use of 68HC12 min and max instructions.
10144 @itemx -mno-long-calls
10145 @opindex mlong-calls
10146 @opindex mno-long-calls
10147 Treat all calls as being far away (near). If calls are assumed to be
10148 far away, the compiler will use the @code{call} instruction to
10149 call a function and the @code{rtc} instruction for returning.
10153 Consider type @code{int} to be 16 bits wide, like @code{short int}.
10155 @item -msoft-reg-count=@var{count}
10156 @opindex msoft-reg-count
10157 Specify the number of pseudo-soft registers which are used for the
10158 code generation. The maximum number is 32. Using more pseudo-soft
10159 register may or may not result in better code depending on the program.
10160 The default is 4 for 68HC11 and 2 for 68HC12.
10164 @node MCore Options
10165 @subsection MCore Options
10166 @cindex MCore options
10168 These are the @samp{-m} options defined for the Motorola M*Core
10174 @itemx -mno-hardlit
10176 @opindex mno-hardlit
10177 Inline constants into the code stream if it can be done in two
10178 instructions or less.
10184 Use the divide instruction. (Enabled by default).
10186 @item -mrelax-immediate
10187 @itemx -mno-relax-immediate
10188 @opindex mrelax-immediate
10189 @opindex mno-relax-immediate
10190 Allow arbitrary sized immediates in bit operations.
10192 @item -mwide-bitfields
10193 @itemx -mno-wide-bitfields
10194 @opindex mwide-bitfields
10195 @opindex mno-wide-bitfields
10196 Always treat bit-fields as int-sized.
10198 @item -m4byte-functions
10199 @itemx -mno-4byte-functions
10200 @opindex m4byte-functions
10201 @opindex mno-4byte-functions
10202 Force all functions to be aligned to a four byte boundary.
10204 @item -mcallgraph-data
10205 @itemx -mno-callgraph-data
10206 @opindex mcallgraph-data
10207 @opindex mno-callgraph-data
10208 Emit callgraph information.
10211 @itemx -mno-slow-bytes
10212 @opindex mslow-bytes
10213 @opindex mno-slow-bytes
10214 Prefer word access when reading byte quantities.
10216 @item -mlittle-endian
10217 @itemx -mbig-endian
10218 @opindex mlittle-endian
10219 @opindex mbig-endian
10220 Generate code for a little endian target.
10226 Generate code for the 210 processor.
10230 @subsection MIPS Options
10231 @cindex MIPS options
10237 Generate big-endian code.
10241 Generate little-endian code. This is the default for @samp{mips*el-*-*}
10244 @item -march=@var{arch}
10246 Generate code that will run on @var{arch}, which can be the name of a
10247 generic MIPS ISA, or the name of a particular processor.
10249 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
10250 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
10251 The processor names are:
10252 @samp{4kc}, @samp{4km}, @samp{4kp},
10253 @samp{5kc}, @samp{5kf},
10255 @samp{24k}, @samp{24kc}, @samp{24kf}, @samp{24kx},
10258 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
10259 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
10260 @samp{rm7000}, @samp{rm9000},
10263 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
10264 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
10265 The special value @samp{from-abi} selects the
10266 most compatible architecture for the selected ABI (that is,
10267 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
10269 In processor names, a final @samp{000} can be abbreviated as @samp{k}
10270 (for example, @samp{-march=r2k}). Prefixes are optional, and
10271 @samp{vr} may be written @samp{r}.
10273 GCC defines two macros based on the value of this option. The first
10274 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
10275 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
10276 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
10277 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
10278 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
10280 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
10281 above. In other words, it will have the full prefix and will not
10282 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
10283 the macro names the resolved architecture (either @samp{"mips1"} or
10284 @samp{"mips3"}). It names the default architecture when no
10285 @option{-march} option is given.
10287 @item -mtune=@var{arch}
10289 Optimize for @var{arch}. Among other things, this option controls
10290 the way instructions are scheduled, and the perceived cost of arithmetic
10291 operations. The list of @var{arch} values is the same as for
10294 When this option is not used, GCC will optimize for the processor
10295 specified by @option{-march}. By using @option{-march} and
10296 @option{-mtune} together, it is possible to generate code that will
10297 run on a family of processors, but optimize the code for one
10298 particular member of that family.
10300 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
10301 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
10302 @samp{-march} ones described above.
10306 Equivalent to @samp{-march=mips1}.
10310 Equivalent to @samp{-march=mips2}.
10314 Equivalent to @samp{-march=mips3}.
10318 Equivalent to @samp{-march=mips4}.
10322 Equivalent to @samp{-march=mips32}.
10326 Equivalent to @samp{-march=mips32r2}.
10330 Equivalent to @samp{-march=mips64}.
10335 @opindex mno-mips16
10336 Generate (do not generate) MIPS16 code. If GCC is targetting a
10337 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
10349 Generate code for the given ABI@.
10351 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
10352 generates 64-bit code when you select a 64-bit architecture, but you
10353 can use @option{-mgp32} to get 32-bit code instead.
10355 For information about the O64 ABI, see
10356 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
10359 @itemx -mno-abicalls
10361 @opindex mno-abicalls
10362 Generate (do not generate) SVR4-style position-independent code.
10363 @option{-mabicalls} is the default for SVR4-based systems.
10369 Lift (do not lift) the usual restrictions on the size of the global
10372 GCC normally uses a single instruction to load values from the GOT@.
10373 While this is relatively efficient, it will only work if the GOT
10374 is smaller than about 64k. Anything larger will cause the linker
10375 to report an error such as:
10377 @cindex relocation truncated to fit (MIPS)
10379 relocation truncated to fit: R_MIPS_GOT16 foobar
10382 If this happens, you should recompile your code with @option{-mxgot}.
10383 It should then work with very large GOTs, although it will also be
10384 less efficient, since it will take three instructions to fetch the
10385 value of a global symbol.
10387 Note that some linkers can create multiple GOTs. If you have such a
10388 linker, you should only need to use @option{-mxgot} when a single object
10389 file accesses more than 64k's worth of GOT entries. Very few do.
10391 These options have no effect unless GCC is generating position
10396 Assume that general-purpose registers are 32 bits wide.
10400 Assume that general-purpose registers are 64 bits wide.
10404 Assume that floating-point registers are 32 bits wide.
10408 Assume that floating-point registers are 64 bits wide.
10411 @opindex mhard-float
10412 Use floating-point coprocessor instructions.
10415 @opindex msoft-float
10416 Do not use floating-point coprocessor instructions. Implement
10417 floating-point calculations using library calls instead.
10419 @item -msingle-float
10420 @opindex msingle-float
10421 Assume that the floating-point coprocessor only supports single-precision
10424 @itemx -mdouble-float
10425 @opindex mdouble-float
10426 Assume that the floating-point coprocessor supports double-precision
10427 operations. This is the default.
10433 Use (do not use) the MIPS DSP ASE. @xref{MIPS DSP Built-in Functions}.
10435 @itemx -mpaired-single
10436 @itemx -mno-paired-single
10437 @opindex mpaired-single
10438 @opindex mno-paired-single
10439 Use (do not use) paired-single floating-point instructions.
10440 @xref{MIPS Paired-Single Support}. This option can only be used
10441 when generating 64-bit code and requires hardware floating-point
10442 support to be enabled.
10447 @opindex mno-mips3d
10448 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
10449 The option @option{-mips3d} implies @option{-mpaired-single}.
10453 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
10454 an explanation of the default and the way that the pointer size is
10459 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
10461 The default size of @code{int}s, @code{long}s and pointers depends on
10462 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
10463 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
10464 32-bit @code{long}s. Pointers are the same size as @code{long}s,
10465 or the same size as integer registers, whichever is smaller.
10471 Assume (do not assume) that all symbols have 32-bit values, regardless
10472 of the selected ABI@. This option is useful in combination with
10473 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
10474 to generate shorter and faster references to symbolic addresses.
10478 @cindex smaller data references (MIPS)
10479 @cindex gp-relative references (MIPS)
10480 Put global and static items less than or equal to @var{num} bytes into
10481 the small data or bss section instead of the normal data or bss section.
10482 This allows the data to be accessed using a single instruction.
10484 All modules should be compiled with the same @option{-G @var{num}}
10487 @item -membedded-data
10488 @itemx -mno-embedded-data
10489 @opindex membedded-data
10490 @opindex mno-embedded-data
10491 Allocate variables to the read-only data section first if possible, then
10492 next in the small data section if possible, otherwise in data. This gives
10493 slightly slower code than the default, but reduces the amount of RAM required
10494 when executing, and thus may be preferred for some embedded systems.
10496 @item -muninit-const-in-rodata
10497 @itemx -mno-uninit-const-in-rodata
10498 @opindex muninit-const-in-rodata
10499 @opindex mno-uninit-const-in-rodata
10500 Put uninitialized @code{const} variables in the read-only data section.
10501 This option is only meaningful in conjunction with @option{-membedded-data}.
10503 @item -msplit-addresses
10504 @itemx -mno-split-addresses
10505 @opindex msplit-addresses
10506 @opindex mno-split-addresses
10507 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
10508 relocation operators. This option has been superseded by
10509 @option{-mexplicit-relocs} but is retained for backwards compatibility.
10511 @item -mexplicit-relocs
10512 @itemx -mno-explicit-relocs
10513 @opindex mexplicit-relocs
10514 @opindex mno-explicit-relocs
10515 Use (do not use) assembler relocation operators when dealing with symbolic
10516 addresses. The alternative, selected by @option{-mno-explicit-relocs},
10517 is to use assembler macros instead.
10519 @option{-mexplicit-relocs} is the default if GCC was configured
10520 to use an assembler that supports relocation operators.
10522 @item -mcheck-zero-division
10523 @itemx -mno-check-zero-division
10524 @opindex mcheck-zero-division
10525 @opindex mno-check-zero-division
10526 Trap (do not trap) on integer division by zero. The default is
10527 @option{-mcheck-zero-division}.
10529 @item -mdivide-traps
10530 @itemx -mdivide-breaks
10531 @opindex mdivide-traps
10532 @opindex mdivide-breaks
10533 MIPS systems check for division by zero by generating either a
10534 conditional trap or a break instruction. Using traps results in
10535 smaller code, but is only supported on MIPS II and later. Also, some
10536 versions of the Linux kernel have a bug that prevents trap from
10537 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
10538 allow conditional traps on architectures that support them and
10539 @option{-mdivide-breaks} to force the use of breaks.
10541 The default is usually @option{-mdivide-traps}, but this can be
10542 overridden at configure time using @option{--with-divide=breaks}.
10543 Divide-by-zero checks can be completely disabled using
10544 @option{-mno-check-zero-division}.
10549 @opindex mno-memcpy
10550 Force (do not force) the use of @code{memcpy()} for non-trivial block
10551 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
10552 most constant-sized copies.
10555 @itemx -mno-long-calls
10556 @opindex mlong-calls
10557 @opindex mno-long-calls
10558 Disable (do not disable) use of the @code{jal} instruction. Calling
10559 functions using @code{jal} is more efficient but requires the caller
10560 and callee to be in the same 256 megabyte segment.
10562 This option has no effect on abicalls code. The default is
10563 @option{-mno-long-calls}.
10569 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
10570 instructions, as provided by the R4650 ISA@.
10573 @itemx -mno-fused-madd
10574 @opindex mfused-madd
10575 @opindex mno-fused-madd
10576 Enable (disable) use of the floating point multiply-accumulate
10577 instructions, when they are available. The default is
10578 @option{-mfused-madd}.
10580 When multiply-accumulate instructions are used, the intermediate
10581 product is calculated to infinite precision and is not subject to
10582 the FCSR Flush to Zero bit. This may be undesirable in some
10587 Tell the MIPS assembler to not run its preprocessor over user
10588 assembler files (with a @samp{.s} suffix) when assembling them.
10591 @itemx -mno-fix-r4000
10592 @opindex mfix-r4000
10593 @opindex mno-fix-r4000
10594 Work around certain R4000 CPU errata:
10597 A double-word or a variable shift may give an incorrect result if executed
10598 immediately after starting an integer division.
10600 A double-word or a variable shift may give an incorrect result if executed
10601 while an integer multiplication is in progress.
10603 An integer division may give an incorrect result if started in a delay slot
10604 of a taken branch or a jump.
10608 @itemx -mno-fix-r4400
10609 @opindex mfix-r4400
10610 @opindex mno-fix-r4400
10611 Work around certain R4400 CPU errata:
10614 A double-word or a variable shift may give an incorrect result if executed
10615 immediately after starting an integer division.
10619 @itemx -mno-fix-vr4120
10620 @opindex mfix-vr4120
10621 Work around certain VR4120 errata:
10624 @code{dmultu} does not always produce the correct result.
10626 @code{div} and @code{ddiv} do not always produce the correct result if one
10627 of the operands is negative.
10629 The workarounds for the division errata rely on special functions in
10630 @file{libgcc.a}. At present, these functions are only provided by
10631 the @code{mips64vr*-elf} configurations.
10633 Other VR4120 errata require a nop to be inserted between certain pairs of
10634 instructions. These errata are handled by the assembler, not by GCC itself.
10637 @opindex mfix-vr4130
10638 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
10639 workarounds are implemented by the assembler rather than by GCC,
10640 although GCC will avoid using @code{mflo} and @code{mfhi} if the
10641 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
10642 instructions are available instead.
10645 @itemx -mno-fix-sb1
10647 Work around certain SB-1 CPU core errata.
10648 (This flag currently works around the SB-1 revision 2
10649 ``F1'' and ``F2'' floating point errata.)
10651 @item -mflush-func=@var{func}
10652 @itemx -mno-flush-func
10653 @opindex mflush-func
10654 Specifies the function to call to flush the I and D caches, or to not
10655 call any such function. If called, the function must take the same
10656 arguments as the common @code{_flush_func()}, that is, the address of the
10657 memory range for which the cache is being flushed, the size of the
10658 memory range, and the number 3 (to flush both caches). The default
10659 depends on the target GCC was configured for, but commonly is either
10660 @samp{_flush_func} or @samp{__cpu_flush}.
10662 @item -mbranch-likely
10663 @itemx -mno-branch-likely
10664 @opindex mbranch-likely
10665 @opindex mno-branch-likely
10666 Enable or disable use of Branch Likely instructions, regardless of the
10667 default for the selected architecture. By default, Branch Likely
10668 instructions may be generated if they are supported by the selected
10669 architecture. An exception is for the MIPS32 and MIPS64 architectures
10670 and processors which implement those architectures; for those, Branch
10671 Likely instructions will not be generated by default because the MIPS32
10672 and MIPS64 architectures specifically deprecate their use.
10674 @item -mfp-exceptions
10675 @itemx -mno-fp-exceptions
10676 @opindex mfp-exceptions
10677 Specifies whether FP exceptions are enabled. This affects how we schedule
10678 FP instructions for some processors. The default is that FP exceptions are
10681 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
10682 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
10685 @item -mvr4130-align
10686 @itemx -mno-vr4130-align
10687 @opindex mvr4130-align
10688 The VR4130 pipeline is two-way superscalar, but can only issue two
10689 instructions together if the first one is 8-byte aligned. When this
10690 option is enabled, GCC will align pairs of instructions that it
10691 thinks should execute in parallel.
10693 This option only has an effect when optimizing for the VR4130.
10694 It normally makes code faster, but at the expense of making it bigger.
10695 It is enabled by default at optimization level @option{-O3}.
10699 @subsection MMIX Options
10700 @cindex MMIX Options
10702 These options are defined for the MMIX:
10706 @itemx -mno-libfuncs
10708 @opindex mno-libfuncs
10709 Specify that intrinsic library functions are being compiled, passing all
10710 values in registers, no matter the size.
10713 @itemx -mno-epsilon
10715 @opindex mno-epsilon
10716 Generate floating-point comparison instructions that compare with respect
10717 to the @code{rE} epsilon register.
10719 @item -mabi=mmixware
10721 @opindex mabi-mmixware
10723 Generate code that passes function parameters and return values that (in
10724 the called function) are seen as registers @code{$0} and up, as opposed to
10725 the GNU ABI which uses global registers @code{$231} and up.
10727 @item -mzero-extend
10728 @itemx -mno-zero-extend
10729 @opindex mzero-extend
10730 @opindex mno-zero-extend
10731 When reading data from memory in sizes shorter than 64 bits, use (do not
10732 use) zero-extending load instructions by default, rather than
10733 sign-extending ones.
10736 @itemx -mno-knuthdiv
10738 @opindex mno-knuthdiv
10739 Make the result of a division yielding a remainder have the same sign as
10740 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
10741 remainder follows the sign of the dividend. Both methods are
10742 arithmetically valid, the latter being almost exclusively used.
10744 @item -mtoplevel-symbols
10745 @itemx -mno-toplevel-symbols
10746 @opindex mtoplevel-symbols
10747 @opindex mno-toplevel-symbols
10748 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
10749 code can be used with the @code{PREFIX} assembly directive.
10753 Generate an executable in the ELF format, rather than the default
10754 @samp{mmo} format used by the @command{mmix} simulator.
10756 @item -mbranch-predict
10757 @itemx -mno-branch-predict
10758 @opindex mbranch-predict
10759 @opindex mno-branch-predict
10760 Use (do not use) the probable-branch instructions, when static branch
10761 prediction indicates a probable branch.
10763 @item -mbase-addresses
10764 @itemx -mno-base-addresses
10765 @opindex mbase-addresses
10766 @opindex mno-base-addresses
10767 Generate (do not generate) code that uses @emph{base addresses}. Using a
10768 base address automatically generates a request (handled by the assembler
10769 and the linker) for a constant to be set up in a global register. The
10770 register is used for one or more base address requests within the range 0
10771 to 255 from the value held in the register. The generally leads to short
10772 and fast code, but the number of different data items that can be
10773 addressed is limited. This means that a program that uses lots of static
10774 data may require @option{-mno-base-addresses}.
10776 @item -msingle-exit
10777 @itemx -mno-single-exit
10778 @opindex msingle-exit
10779 @opindex mno-single-exit
10780 Force (do not force) generated code to have a single exit point in each
10784 @node MN10300 Options
10785 @subsection MN10300 Options
10786 @cindex MN10300 options
10788 These @option{-m} options are defined for Matsushita MN10300 architectures:
10793 Generate code to avoid bugs in the multiply instructions for the MN10300
10794 processors. This is the default.
10796 @item -mno-mult-bug
10797 @opindex mno-mult-bug
10798 Do not generate code to avoid bugs in the multiply instructions for the
10799 MN10300 processors.
10803 Generate code which uses features specific to the AM33 processor.
10807 Do not generate code which uses features specific to the AM33 processor. This
10810 @item -mreturn-pointer-on-d0
10811 @opindex mreturn-pointer-on-d0
10812 When generating a function which returns a pointer, return the pointer
10813 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
10814 only in a0, and attempts to call such functions without a prototype
10815 would result in errors. Note that this option is on by default; use
10816 @option{-mno-return-pointer-on-d0} to disable it.
10820 Do not link in the C run-time initialization object file.
10824 Indicate to the linker that it should perform a relaxation optimization pass
10825 to shorten branches, calls and absolute memory addresses. This option only
10826 has an effect when used on the command line for the final link step.
10828 This option makes symbolic debugging impossible.
10832 @subsection MT Options
10835 These @option{-m} options are defined for Morpho MT architectures:
10839 @item -march=@var{cpu-type}
10841 Generate code that will run on @var{cpu-type}, which is the name of a system
10842 representing a certain processor type. Possible values for
10843 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
10844 @samp{ms1-16-003} and @samp{ms2}.
10846 When this option is not used, the default is @option{-march=ms1-16-002}.
10850 Use byte loads and stores when generating code.
10854 Do not use byte loads and stores when generating code.
10858 Use simulator runtime
10862 Do not link in the C run-time initialization object file
10863 @file{crti.o}. Other run-time initialization and termination files
10864 such as @file{startup.o} and @file{exit.o} are still included on the
10865 linker command line.
10869 @node PDP-11 Options
10870 @subsection PDP-11 Options
10871 @cindex PDP-11 Options
10873 These options are defined for the PDP-11:
10878 Use hardware FPP floating point. This is the default. (FIS floating
10879 point on the PDP-11/40 is not supported.)
10882 @opindex msoft-float
10883 Do not use hardware floating point.
10887 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
10891 Return floating-point results in memory. This is the default.
10895 Generate code for a PDP-11/40.
10899 Generate code for a PDP-11/45. This is the default.
10903 Generate code for a PDP-11/10.
10905 @item -mbcopy-builtin
10906 @opindex bcopy-builtin
10907 Use inline @code{movmemhi} patterns for copying memory. This is the
10912 Do not use inline @code{movmemhi} patterns for copying memory.
10918 Use 16-bit @code{int}. This is the default.
10924 Use 32-bit @code{int}.
10927 @itemx -mno-float32
10929 @opindex mno-float32
10930 Use 64-bit @code{float}. This is the default.
10933 @itemx -mno-float64
10935 @opindex mno-float64
10936 Use 32-bit @code{float}.
10940 Use @code{abshi2} pattern. This is the default.
10944 Do not use @code{abshi2} pattern.
10946 @item -mbranch-expensive
10947 @opindex mbranch-expensive
10948 Pretend that branches are expensive. This is for experimenting with
10949 code generation only.
10951 @item -mbranch-cheap
10952 @opindex mbranch-cheap
10953 Do not pretend that branches are expensive. This is the default.
10957 Generate code for a system with split I&D@.
10961 Generate code for a system without split I&D@. This is the default.
10965 Use Unix assembler syntax. This is the default when configured for
10966 @samp{pdp11-*-bsd}.
10970 Use DEC assembler syntax. This is the default when configured for any
10971 PDP-11 target other than @samp{pdp11-*-bsd}.
10974 @node PowerPC Options
10975 @subsection PowerPC Options
10976 @cindex PowerPC options
10978 These are listed under @xref{RS/6000 and PowerPC Options}.
10980 @node RS/6000 and PowerPC Options
10981 @subsection IBM RS/6000 and PowerPC Options
10982 @cindex RS/6000 and PowerPC Options
10983 @cindex IBM RS/6000 and PowerPC Options
10985 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
10992 @itemx -mno-powerpc
10993 @itemx -mpowerpc-gpopt
10994 @itemx -mno-powerpc-gpopt
10995 @itemx -mpowerpc-gfxopt
10996 @itemx -mno-powerpc-gfxopt
10998 @itemx -mno-powerpc64
11002 @itemx -mno-popcntb
11008 @opindex mno-power2
11010 @opindex mno-powerpc
11011 @opindex mpowerpc-gpopt
11012 @opindex mno-powerpc-gpopt
11013 @opindex mpowerpc-gfxopt
11014 @opindex mno-powerpc-gfxopt
11015 @opindex mpowerpc64
11016 @opindex mno-powerpc64
11020 @opindex mno-popcntb
11023 GCC supports two related instruction set architectures for the
11024 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
11025 instructions supported by the @samp{rios} chip set used in the original
11026 RS/6000 systems and the @dfn{PowerPC} instruction set is the
11027 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
11028 the IBM 4xx, 6xx, and follow-on microprocessors.
11030 Neither architecture is a subset of the other. However there is a
11031 large common subset of instructions supported by both. An MQ
11032 register is included in processors supporting the POWER architecture.
11034 You use these options to specify which instructions are available on the
11035 processor you are using. The default value of these options is
11036 determined when configuring GCC@. Specifying the
11037 @option{-mcpu=@var{cpu_type}} overrides the specification of these
11038 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
11039 rather than the options listed above.
11041 The @option{-mpower} option allows GCC to generate instructions that
11042 are found only in the POWER architecture and to use the MQ register.
11043 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
11044 to generate instructions that are present in the POWER2 architecture but
11045 not the original POWER architecture.
11047 The @option{-mpowerpc} option allows GCC to generate instructions that
11048 are found only in the 32-bit subset of the PowerPC architecture.
11049 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
11050 GCC to use the optional PowerPC architecture instructions in the
11051 General Purpose group, including floating-point square root. Specifying
11052 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
11053 use the optional PowerPC architecture instructions in the Graphics
11054 group, including floating-point select.
11056 The @option{-mmfcrf} option allows GCC to generate the move from
11057 condition register field instruction implemented on the POWER4
11058 processor and other processors that support the PowerPC V2.01
11060 The @option{-mpopcntb} option allows GCC to generate the popcount and
11061 double precision FP reciprocal estimate instruction implemented on the
11062 POWER5 processor and other processors that support the PowerPC V2.02
11064 The @option{-mfprnd} option allows GCC to generate the FP round to
11065 integer instructions implemented on the POWER5+ processor and other
11066 processors that support the PowerPC V2.03 architecture.
11068 The @option{-mpowerpc64} option allows GCC to generate the additional
11069 64-bit instructions that are found in the full PowerPC64 architecture
11070 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
11071 @option{-mno-powerpc64}.
11073 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
11074 will use only the instructions in the common subset of both
11075 architectures plus some special AIX common-mode calls, and will not use
11076 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
11077 permits GCC to use any instruction from either architecture and to
11078 allow use of the MQ register; specify this for the Motorola MPC601.
11080 @item -mnew-mnemonics
11081 @itemx -mold-mnemonics
11082 @opindex mnew-mnemonics
11083 @opindex mold-mnemonics
11084 Select which mnemonics to use in the generated assembler code. With
11085 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
11086 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
11087 assembler mnemonics defined for the POWER architecture. Instructions
11088 defined in only one architecture have only one mnemonic; GCC uses that
11089 mnemonic irrespective of which of these options is specified.
11091 GCC defaults to the mnemonics appropriate for the architecture in
11092 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
11093 value of these option. Unless you are building a cross-compiler, you
11094 should normally not specify either @option{-mnew-mnemonics} or
11095 @option{-mold-mnemonics}, but should instead accept the default.
11097 @item -mcpu=@var{cpu_type}
11099 Set architecture type, register usage, choice of mnemonics, and
11100 instruction scheduling parameters for machine type @var{cpu_type}.
11101 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
11102 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
11103 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
11104 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
11105 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
11106 @samp{860}, @samp{970}, @samp{8540}, @samp{ec603e}, @samp{G3},
11107 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
11108 @samp{power4}, @samp{power5}, @samp{power5+},
11109 @samp{common}, @samp{powerpc}, @samp{powerpc64},
11110 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
11112 @option{-mcpu=common} selects a completely generic processor. Code
11113 generated under this option will run on any POWER or PowerPC processor.
11114 GCC will use only the instructions in the common subset of both
11115 architectures, and will not use the MQ register. GCC assumes a generic
11116 processor model for scheduling purposes.
11118 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
11119 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
11120 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
11121 types, with an appropriate, generic processor model assumed for
11122 scheduling purposes.
11124 The other options specify a specific processor. Code generated under
11125 those options will run best on that processor, and may not run at all on
11128 The @option{-mcpu} options automatically enable or disable the
11129 following options: @option{-maltivec}, @option{-mfprnd},
11130 @option{-mhard-float}, @option{-mmfcrf}, @option{-mmultiple},
11131 @option{-mnew-mnemonics}, @option{-mpopcntb}, @option{-mpower},
11132 @option{-mpower2}, @option{-mpowerpc64}, @option{-mpowerpc-gpopt},
11133 @option{-mpowerpc-gfxopt}, @option{-mstring}, @option{-mmulhw}.
11134 The particular options
11135 set for any particular CPU will vary between compiler versions,
11136 depending on what setting seems to produce optimal code for that CPU;
11137 it doesn't necessarily reflect the actual hardware's capabilities. If
11138 you wish to set an individual option to a particular value, you may
11139 specify it after the @option{-mcpu} option, like @samp{-mcpu=970
11142 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
11143 not enabled or disabled by the @option{-mcpu} option at present because
11144 AIX does not have full support for these options. You may still
11145 enable or disable them individually if you're sure it'll work in your
11148 @item -mtune=@var{cpu_type}
11150 Set the instruction scheduling parameters for machine type
11151 @var{cpu_type}, but do not set the architecture type, register usage, or
11152 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
11153 values for @var{cpu_type} are used for @option{-mtune} as for
11154 @option{-mcpu}. If both are specified, the code generated will use the
11155 architecture, registers, and mnemonics set by @option{-mcpu}, but the
11156 scheduling parameters set by @option{-mtune}.
11162 Generate code to compute division as reciprocal estimate and iterative
11163 refinement, creating opportunities for increased throughput. This
11164 feature requires: optional PowerPC Graphics instruction set for single
11165 precision and FRE instruction for double precision, assuming divides
11166 cannot generate user-visible traps, and the domain values not include
11167 Infinities, denormals or zero denominator.
11170 @itemx -mno-altivec
11172 @opindex mno-altivec
11173 Generate code that uses (does not use) AltiVec instructions, and also
11174 enable the use of built-in functions that allow more direct access to
11175 the AltiVec instruction set. You may also need to set
11176 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
11182 @opindex mno-vrsave
11183 Generate VRSAVE instructions when generating AltiVec code.
11186 @opindex msecure-plt
11187 Generate code that allows ld and ld.so to build executables and shared
11188 libraries with non-exec .plt and .got sections. This is a PowerPC
11189 32-bit SYSV ABI option.
11193 Generate code that uses a BSS .plt section that ld.so fills in, and
11194 requires .plt and .got sections that are both writable and executable.
11195 This is a PowerPC 32-bit SYSV ABI option.
11201 This switch enables or disables the generation of ISEL instructions.
11203 @item -misel=@var{yes/no}
11204 This switch has been deprecated. Use @option{-misel} and
11205 @option{-mno-isel} instead.
11211 This switch enables or disables the generation of SPE simd
11214 @item -mspe=@var{yes/no}
11215 This option has been deprecated. Use @option{-mspe} and
11216 @option{-mno-spe} instead.
11218 @item -mfloat-gprs=@var{yes/single/double/no}
11219 @itemx -mfloat-gprs
11220 @opindex mfloat-gprs
11221 This switch enables or disables the generation of floating point
11222 operations on the general purpose registers for architectures that
11225 The argument @var{yes} or @var{single} enables the use of
11226 single-precision floating point operations.
11228 The argument @var{double} enables the use of single and
11229 double-precision floating point operations.
11231 The argument @var{no} disables floating point operations on the
11232 general purpose registers.
11234 This option is currently only available on the MPC854x.
11240 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
11241 targets (including GNU/Linux). The 32-bit environment sets int, long
11242 and pointer to 32 bits and generates code that runs on any PowerPC
11243 variant. The 64-bit environment sets int to 32 bits and long and
11244 pointer to 64 bits, and generates code for PowerPC64, as for
11245 @option{-mpowerpc64}.
11248 @itemx -mno-fp-in-toc
11249 @itemx -mno-sum-in-toc
11250 @itemx -mminimal-toc
11252 @opindex mno-fp-in-toc
11253 @opindex mno-sum-in-toc
11254 @opindex mminimal-toc
11255 Modify generation of the TOC (Table Of Contents), which is created for
11256 every executable file. The @option{-mfull-toc} option is selected by
11257 default. In that case, GCC will allocate at least one TOC entry for
11258 each unique non-automatic variable reference in your program. GCC
11259 will also place floating-point constants in the TOC@. However, only
11260 16,384 entries are available in the TOC@.
11262 If you receive a linker error message that saying you have overflowed
11263 the available TOC space, you can reduce the amount of TOC space used
11264 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
11265 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
11266 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
11267 generate code to calculate the sum of an address and a constant at
11268 run-time instead of putting that sum into the TOC@. You may specify one
11269 or both of these options. Each causes GCC to produce very slightly
11270 slower and larger code at the expense of conserving TOC space.
11272 If you still run out of space in the TOC even when you specify both of
11273 these options, specify @option{-mminimal-toc} instead. This option causes
11274 GCC to make only one TOC entry for every file. When you specify this
11275 option, GCC will produce code that is slower and larger but which
11276 uses extremely little TOC space. You may wish to use this option
11277 only on files that contain less frequently executed code.
11283 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
11284 @code{long} type, and the infrastructure needed to support them.
11285 Specifying @option{-maix64} implies @option{-mpowerpc64} and
11286 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
11287 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
11290 @itemx -mno-xl-compat
11291 @opindex mxl-compat
11292 @opindex mno-xl-compat
11293 Produce code that conforms more closely to IBM XL compiler semantics
11294 when using AIX-compatible ABI. Pass floating-point arguments to
11295 prototyped functions beyond the register save area (RSA) on the stack
11296 in addition to argument FPRs. Do not assume that most significant
11297 double in 128-bit long double value is properly rounded when comparing
11298 values and converting to double. Use XL symbol names for long double
11301 The AIX calling convention was extended but not initially documented to
11302 handle an obscure K&R C case of calling a function that takes the
11303 address of its arguments with fewer arguments than declared. IBM XL
11304 compilers access floating point arguments which do not fit in the
11305 RSA from the stack when a subroutine is compiled without
11306 optimization. Because always storing floating-point arguments on the
11307 stack is inefficient and rarely needed, this option is not enabled by
11308 default and only is necessary when calling subroutines compiled by IBM
11309 XL compilers without optimization.
11313 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
11314 application written to use message passing with special startup code to
11315 enable the application to run. The system must have PE installed in the
11316 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
11317 must be overridden with the @option{-specs=} option to specify the
11318 appropriate directory location. The Parallel Environment does not
11319 support threads, so the @option{-mpe} option and the @option{-pthread}
11320 option are incompatible.
11322 @item -malign-natural
11323 @itemx -malign-power
11324 @opindex malign-natural
11325 @opindex malign-power
11326 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
11327 @option{-malign-natural} overrides the ABI-defined alignment of larger
11328 types, such as floating-point doubles, on their natural size-based boundary.
11329 The option @option{-malign-power} instructs GCC to follow the ABI-specified
11330 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
11332 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
11336 @itemx -mhard-float
11337 @opindex msoft-float
11338 @opindex mhard-float
11339 Generate code that does not use (uses) the floating-point register set.
11340 Software floating point emulation is provided if you use the
11341 @option{-msoft-float} option, and pass the option to GCC when linking.
11344 @itemx -mno-multiple
11346 @opindex mno-multiple
11347 Generate code that uses (does not use) the load multiple word
11348 instructions and the store multiple word instructions. These
11349 instructions are generated by default on POWER systems, and not
11350 generated on PowerPC systems. Do not use @option{-mmultiple} on little
11351 endian PowerPC systems, since those instructions do not work when the
11352 processor is in little endian mode. The exceptions are PPC740 and
11353 PPC750 which permit the instructions usage in little endian mode.
11358 @opindex mno-string
11359 Generate code that uses (does not use) the load string instructions
11360 and the store string word instructions to save multiple registers and
11361 do small block moves. These instructions are generated by default on
11362 POWER systems, and not generated on PowerPC systems. Do not use
11363 @option{-mstring} on little endian PowerPC systems, since those
11364 instructions do not work when the processor is in little endian mode.
11365 The exceptions are PPC740 and PPC750 which permit the instructions
11366 usage in little endian mode.
11371 @opindex mno-update
11372 Generate code that uses (does not use) the load or store instructions
11373 that update the base register to the address of the calculated memory
11374 location. These instructions are generated by default. If you use
11375 @option{-mno-update}, there is a small window between the time that the
11376 stack pointer is updated and the address of the previous frame is
11377 stored, which means code that walks the stack frame across interrupts or
11378 signals may get corrupted data.
11381 @itemx -mno-fused-madd
11382 @opindex mfused-madd
11383 @opindex mno-fused-madd
11384 Generate code that uses (does not use) the floating point multiply and
11385 accumulate instructions. These instructions are generated by default if
11386 hardware floating is used.
11392 Generate code that uses (does not use) the half-word multiply and
11393 multiply-accumulate instructions on the IBM 405 and 440 processors.
11394 These instructions are generated by default when targetting those
11397 @item -mno-bit-align
11399 @opindex mno-bit-align
11400 @opindex mbit-align
11401 On System V.4 and embedded PowerPC systems do not (do) force structures
11402 and unions that contain bit-fields to be aligned to the base type of the
11405 For example, by default a structure containing nothing but 8
11406 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
11407 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
11408 the structure would be aligned to a 1 byte boundary and be one byte in
11411 @item -mno-strict-align
11412 @itemx -mstrict-align
11413 @opindex mno-strict-align
11414 @opindex mstrict-align
11415 On System V.4 and embedded PowerPC systems do not (do) assume that
11416 unaligned memory references will be handled by the system.
11418 @item -mrelocatable
11419 @itemx -mno-relocatable
11420 @opindex mrelocatable
11421 @opindex mno-relocatable
11422 On embedded PowerPC systems generate code that allows (does not allow)
11423 the program to be relocated to a different address at runtime. If you
11424 use @option{-mrelocatable} on any module, all objects linked together must
11425 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
11427 @item -mrelocatable-lib
11428 @itemx -mno-relocatable-lib
11429 @opindex mrelocatable-lib
11430 @opindex mno-relocatable-lib
11431 On embedded PowerPC systems generate code that allows (does not allow)
11432 the program to be relocated to a different address at runtime. Modules
11433 compiled with @option{-mrelocatable-lib} can be linked with either modules
11434 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
11435 with modules compiled with the @option{-mrelocatable} options.
11441 On System V.4 and embedded PowerPC systems do not (do) assume that
11442 register 2 contains a pointer to a global area pointing to the addresses
11443 used in the program.
11446 @itemx -mlittle-endian
11448 @opindex mlittle-endian
11449 On System V.4 and embedded PowerPC systems compile code for the
11450 processor in little endian mode. The @option{-mlittle-endian} option is
11451 the same as @option{-mlittle}.
11454 @itemx -mbig-endian
11456 @opindex mbig-endian
11457 On System V.4 and embedded PowerPC systems compile code for the
11458 processor in big endian mode. The @option{-mbig-endian} option is
11459 the same as @option{-mbig}.
11461 @item -mdynamic-no-pic
11462 @opindex mdynamic-no-pic
11463 On Darwin and Mac OS X systems, compile code so that it is not
11464 relocatable, but that its external references are relocatable. The
11465 resulting code is suitable for applications, but not shared
11468 @item -mprioritize-restricted-insns=@var{priority}
11469 @opindex mprioritize-restricted-insns
11470 This option controls the priority that is assigned to
11471 dispatch-slot restricted instructions during the second scheduling
11472 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
11473 @var{no/highest/second-highest} priority to dispatch slot restricted
11476 @item -msched-costly-dep=@var{dependence_type}
11477 @opindex msched-costly-dep
11478 This option controls which dependences are considered costly
11479 by the target during instruction scheduling. The argument
11480 @var{dependence_type} takes one of the following values:
11481 @var{no}: no dependence is costly,
11482 @var{all}: all dependences are costly,
11483 @var{true_store_to_load}: a true dependence from store to load is costly,
11484 @var{store_to_load}: any dependence from store to load is costly,
11485 @var{number}: any dependence which latency >= @var{number} is costly.
11487 @item -minsert-sched-nops=@var{scheme}
11488 @opindex minsert-sched-nops
11489 This option controls which nop insertion scheme will be used during
11490 the second scheduling pass. The argument @var{scheme} takes one of the
11492 @var{no}: Don't insert nops.
11493 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
11494 according to the scheduler's grouping.
11495 @var{regroup_exact}: Insert nops to force costly dependent insns into
11496 separate groups. Insert exactly as many nops as needed to force an insn
11497 to a new group, according to the estimated processor grouping.
11498 @var{number}: Insert nops to force costly dependent insns into
11499 separate groups. Insert @var{number} nops to force an insn to a new group.
11502 @opindex mcall-sysv
11503 On System V.4 and embedded PowerPC systems compile code using calling
11504 conventions that adheres to the March 1995 draft of the System V
11505 Application Binary Interface, PowerPC processor supplement. This is the
11506 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
11508 @item -mcall-sysv-eabi
11509 @opindex mcall-sysv-eabi
11510 Specify both @option{-mcall-sysv} and @option{-meabi} options.
11512 @item -mcall-sysv-noeabi
11513 @opindex mcall-sysv-noeabi
11514 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
11516 @item -mcall-solaris
11517 @opindex mcall-solaris
11518 On System V.4 and embedded PowerPC systems compile code for the Solaris
11522 @opindex mcall-linux
11523 On System V.4 and embedded PowerPC systems compile code for the
11524 Linux-based GNU system.
11528 On System V.4 and embedded PowerPC systems compile code for the
11529 Hurd-based GNU system.
11531 @item -mcall-netbsd
11532 @opindex mcall-netbsd
11533 On System V.4 and embedded PowerPC systems compile code for the
11534 NetBSD operating system.
11536 @item -maix-struct-return
11537 @opindex maix-struct-return
11538 Return all structures in memory (as specified by the AIX ABI)@.
11540 @item -msvr4-struct-return
11541 @opindex msvr4-struct-return
11542 Return structures smaller than 8 bytes in registers (as specified by the
11545 @item -mabi=@var{abi-type}
11547 Extend the current ABI with a particular extension, or remove such extension.
11548 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
11549 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
11553 Extend the current ABI with SPE ABI extensions. This does not change
11554 the default ABI, instead it adds the SPE ABI extensions to the current
11558 @opindex mabi=no-spe
11559 Disable Booke SPE ABI extensions for the current ABI@.
11561 @item -mabi=ibmlongdouble
11562 @opindex mabi=ibmlongdouble
11563 Change the current ABI to use IBM extended precision long double.
11564 This is a PowerPC 32-bit SYSV ABI option.
11566 @item -mabi=ieeelongdouble
11567 @opindex mabi=ieeelongdouble
11568 Change the current ABI to use IEEE extended precision long double.
11569 This is a PowerPC 32-bit Linux ABI option.
11572 @itemx -mno-prototype
11573 @opindex mprototype
11574 @opindex mno-prototype
11575 On System V.4 and embedded PowerPC systems assume that all calls to
11576 variable argument functions are properly prototyped. Otherwise, the
11577 compiler must insert an instruction before every non prototyped call to
11578 set or clear bit 6 of the condition code register (@var{CR}) to
11579 indicate whether floating point values were passed in the floating point
11580 registers in case the function takes a variable arguments. With
11581 @option{-mprototype}, only calls to prototyped variable argument functions
11582 will set or clear the bit.
11586 On embedded PowerPC systems, assume that the startup module is called
11587 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
11588 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
11593 On embedded PowerPC systems, assume that the startup module is called
11594 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
11599 On embedded PowerPC systems, assume that the startup module is called
11600 @file{crt0.o} and the standard C libraries are @file{libads.a} and
11603 @item -myellowknife
11604 @opindex myellowknife
11605 On embedded PowerPC systems, assume that the startup module is called
11606 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
11611 On System V.4 and embedded PowerPC systems, specify that you are
11612 compiling for a VxWorks system.
11616 Specify that you are compiling for the WindISS simulation environment.
11620 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
11621 header to indicate that @samp{eabi} extended relocations are used.
11627 On System V.4 and embedded PowerPC systems do (do not) adhere to the
11628 Embedded Applications Binary Interface (eabi) which is a set of
11629 modifications to the System V.4 specifications. Selecting @option{-meabi}
11630 means that the stack is aligned to an 8 byte boundary, a function
11631 @code{__eabi} is called to from @code{main} to set up the eabi
11632 environment, and the @option{-msdata} option can use both @code{r2} and
11633 @code{r13} to point to two separate small data areas. Selecting
11634 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
11635 do not call an initialization function from @code{main}, and the
11636 @option{-msdata} option will only use @code{r13} to point to a single
11637 small data area. The @option{-meabi} option is on by default if you
11638 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
11641 @opindex msdata=eabi
11642 On System V.4 and embedded PowerPC systems, put small initialized
11643 @code{const} global and static data in the @samp{.sdata2} section, which
11644 is pointed to by register @code{r2}. Put small initialized
11645 non-@code{const} global and static data in the @samp{.sdata} section,
11646 which is pointed to by register @code{r13}. Put small uninitialized
11647 global and static data in the @samp{.sbss} section, which is adjacent to
11648 the @samp{.sdata} section. The @option{-msdata=eabi} option is
11649 incompatible with the @option{-mrelocatable} option. The
11650 @option{-msdata=eabi} option also sets the @option{-memb} option.
11653 @opindex msdata=sysv
11654 On System V.4 and embedded PowerPC systems, put small global and static
11655 data in the @samp{.sdata} section, which is pointed to by register
11656 @code{r13}. Put small uninitialized global and static data in the
11657 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
11658 The @option{-msdata=sysv} option is incompatible with the
11659 @option{-mrelocatable} option.
11661 @item -msdata=default
11663 @opindex msdata=default
11665 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
11666 compile code the same as @option{-msdata=eabi}, otherwise compile code the
11667 same as @option{-msdata=sysv}.
11670 @opindex msdata-data
11671 On System V.4 and embedded PowerPC systems, put small global
11672 data in the @samp{.sdata} section. Put small uninitialized global
11673 data in the @samp{.sbss} section. Do not use register @code{r13}
11674 to address small data however. This is the default behavior unless
11675 other @option{-msdata} options are used.
11679 @opindex msdata=none
11681 On embedded PowerPC systems, put all initialized global and static data
11682 in the @samp{.data} section, and all uninitialized data in the
11683 @samp{.bss} section.
11687 @cindex smaller data references (PowerPC)
11688 @cindex .sdata/.sdata2 references (PowerPC)
11689 On embedded PowerPC systems, put global and static items less than or
11690 equal to @var{num} bytes into the small data or bss sections instead of
11691 the normal data or bss section. By default, @var{num} is 8. The
11692 @option{-G @var{num}} switch is also passed to the linker.
11693 All modules should be compiled with the same @option{-G @var{num}} value.
11696 @itemx -mno-regnames
11698 @opindex mno-regnames
11699 On System V.4 and embedded PowerPC systems do (do not) emit register
11700 names in the assembly language output using symbolic forms.
11703 @itemx -mno-longcall
11705 @opindex mno-longcall
11706 Default to making all function calls indirectly, using a register, so
11707 that functions which reside further than 32 megabytes (33,554,432
11708 bytes) from the current location can be called. This setting can be
11709 overridden by the @code{shortcall} function attribute, or by
11710 @code{#pragma longcall(0)}.
11712 Some linkers are capable of detecting out-of-range calls and generating
11713 glue code on the fly. On these systems, long calls are unnecessary and
11714 generate slower code. As of this writing, the AIX linker can do this,
11715 as can the GNU linker for PowerPC/64. It is planned to add this feature
11716 to the GNU linker for 32-bit PowerPC systems as well.
11718 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
11719 callee, L42'', plus a ``branch island'' (glue code). The two target
11720 addresses represent the callee and the ``branch island''. The
11721 Darwin/PPC linker will prefer the first address and generate a ``bl
11722 callee'' if the PPC ``bl'' instruction will reach the callee directly;
11723 otherwise, the linker will generate ``bl L42'' to call the ``branch
11724 island''. The ``branch island'' is appended to the body of the
11725 calling function; it computes the full 32-bit address of the callee
11728 On Mach-O (Darwin) systems, this option directs the compiler emit to
11729 the glue for every direct call, and the Darwin linker decides whether
11730 to use or discard it.
11732 In the future, we may cause GCC to ignore all longcall specifications
11733 when the linker is known to generate glue.
11737 Adds support for multithreading with the @dfn{pthreads} library.
11738 This option sets flags for both the preprocessor and linker.
11742 @node S/390 and zSeries Options
11743 @subsection S/390 and zSeries Options
11744 @cindex S/390 and zSeries Options
11746 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
11750 @itemx -msoft-float
11751 @opindex mhard-float
11752 @opindex msoft-float
11753 Use (do not use) the hardware floating-point instructions and registers
11754 for floating-point operations. When @option{-msoft-float} is specified,
11755 functions in @file{libgcc.a} will be used to perform floating-point
11756 operations. When @option{-mhard-float} is specified, the compiler
11757 generates IEEE floating-point instructions. This is the default.
11759 @item -mlong-double-64
11760 @itemx -mlong-double-128
11761 @opindex mlong-double-64
11762 @opindex mlong-double-128
11763 These switches control the size of @code{long double} type. A size
11764 of 64bit makes the @code{long double} type equivalent to the @code{double}
11765 type. This is the default.
11768 @itemx -mno-backchain
11769 @opindex mbackchain
11770 @opindex mno-backchain
11771 Store (do not store) the address of the caller's frame as backchain pointer
11772 into the callee's stack frame.
11773 A backchain may be needed to allow debugging using tools that do not understand
11774 DWARF-2 call frame information.
11775 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
11776 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
11777 the backchain is placed into the topmost word of the 96/160 byte register
11780 In general, code compiled with @option{-mbackchain} is call-compatible with
11781 code compiled with @option{-mmo-backchain}; however, use of the backchain
11782 for debugging purposes usually requires that the whole binary is built with
11783 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
11784 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
11785 to build a linux kernel use @option{-msoft-float}.
11787 The default is to not maintain the backchain.
11789 @item -mpacked-stack
11790 @item -mno-packed-stack
11791 @opindex mpacked-stack
11792 @opindex mno-packed-stack
11793 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
11794 specified, the compiler uses the all fields of the 96/160 byte register save
11795 area only for their default purpose; unused fields still take up stack space.
11796 When @option{-mpacked-stack} is specified, register save slots are densely
11797 packed at the top of the register save area; unused space is reused for other
11798 purposes, allowing for more efficient use of the available stack space.
11799 However, when @option{-mbackchain} is also in effect, the topmost word of
11800 the save area is always used to store the backchain, and the return address
11801 register is always saved two words below the backchain.
11803 As long as the stack frame backchain is not used, code generated with
11804 @option{-mpacked-stack} is call-compatible with code generated with
11805 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
11806 S/390 or zSeries generated code that uses the stack frame backchain at run
11807 time, not just for debugging purposes. Such code is not call-compatible
11808 with code compiled with @option{-mpacked-stack}. Also, note that the
11809 combination of @option{-mbackchain},
11810 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
11811 to build a linux kernel use @option{-msoft-float}.
11813 The default is to not use the packed stack layout.
11816 @itemx -mno-small-exec
11817 @opindex msmall-exec
11818 @opindex mno-small-exec
11819 Generate (or do not generate) code using the @code{bras} instruction
11820 to do subroutine calls.
11821 This only works reliably if the total executable size does not
11822 exceed 64k. The default is to use the @code{basr} instruction instead,
11823 which does not have this limitation.
11829 When @option{-m31} is specified, generate code compliant to the
11830 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
11831 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
11832 particular to generate 64-bit instructions. For the @samp{s390}
11833 targets, the default is @option{-m31}, while the @samp{s390x}
11834 targets default to @option{-m64}.
11840 When @option{-mzarch} is specified, generate code using the
11841 instructions available on z/Architecture.
11842 When @option{-mesa} is specified, generate code using the
11843 instructions available on ESA/390. Note that @option{-mesa} is
11844 not possible with @option{-m64}.
11845 When generating code compliant to the GNU/Linux for S/390 ABI,
11846 the default is @option{-mesa}. When generating code compliant
11847 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
11853 Generate (or do not generate) code using the @code{mvcle} instruction
11854 to perform block moves. When @option{-mno-mvcle} is specified,
11855 use a @code{mvc} loop instead. This is the default unless optimizing for
11862 Print (or do not print) additional debug information when compiling.
11863 The default is to not print debug information.
11865 @item -march=@var{cpu-type}
11867 Generate code that will run on @var{cpu-type}, which is the name of a system
11868 representing a certain processor type. Possible values for
11869 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
11870 When generating code using the instructions available on z/Architecture,
11871 the default is @option{-march=z900}. Otherwise, the default is
11872 @option{-march=g5}.
11874 @item -mtune=@var{cpu-type}
11876 Tune to @var{cpu-type} everything applicable about the generated code,
11877 except for the ABI and the set of available instructions.
11878 The list of @var{cpu-type} values is the same as for @option{-march}.
11879 The default is the value used for @option{-march}.
11882 @itemx -mno-tpf-trace
11883 @opindex mtpf-trace
11884 @opindex mno-tpf-trace
11885 Generate code that adds (does not add) in TPF OS specific branches to trace
11886 routines in the operating system. This option is off by default, even
11887 when compiling for the TPF OS@.
11890 @itemx -mno-fused-madd
11891 @opindex mfused-madd
11892 @opindex mno-fused-madd
11893 Generate code that uses (does not use) the floating point multiply and
11894 accumulate instructions. These instructions are generated by default if
11895 hardware floating point is used.
11897 @item -mwarn-framesize=@var{framesize}
11898 @opindex mwarn-framesize
11899 Emit a warning if the current function exceeds the given frame size. Because
11900 this is a compile time check it doesn't need to be a real problem when the program
11901 runs. It is intended to identify functions which most probably cause
11902 a stack overflow. It is useful to be used in an environment with limited stack
11903 size e.g.@: the linux kernel.
11905 @item -mwarn-dynamicstack
11906 @opindex mwarn-dynamicstack
11907 Emit a warning if the function calls alloca or uses dynamically
11908 sized arrays. This is generally a bad idea with a limited stack size.
11910 @item -mstack-guard=@var{stack-guard}
11911 @item -mstack-size=@var{stack-size}
11912 @opindex mstack-guard
11913 @opindex mstack-size
11914 These arguments always have to be used in conjunction. If they are present the s390
11915 back end emits additional instructions in the function prologue which trigger a trap
11916 if the stack size is @var{stack-guard} bytes above the @var{stack-size}
11917 (remember that the stack on s390 grows downward). These options are intended to
11918 be used to help debugging stack overflow problems. The additionally emitted code
11919 causes only little overhead and hence can also be used in production like systems
11920 without greater performance degradation. The given values have to be exact
11921 powers of 2 and @var{stack-size} has to be greater than @var{stack-guard} without
11923 In order to be efficient the extra code makes the assumption that the stack starts
11924 at an address aligned to the value given by @var{stack-size}.
11928 @subsection SH Options
11930 These @samp{-m} options are defined for the SH implementations:
11935 Generate code for the SH1.
11939 Generate code for the SH2.
11942 Generate code for the SH2e.
11946 Generate code for the SH3.
11950 Generate code for the SH3e.
11954 Generate code for the SH4 without a floating-point unit.
11956 @item -m4-single-only
11957 @opindex m4-single-only
11958 Generate code for the SH4 with a floating-point unit that only
11959 supports single-precision arithmetic.
11963 Generate code for the SH4 assuming the floating-point unit is in
11964 single-precision mode by default.
11968 Generate code for the SH4.
11972 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
11973 floating-point unit is not used.
11975 @item -m4a-single-only
11976 @opindex m4a-single-only
11977 Generate code for the SH4a, in such a way that no double-precision
11978 floating point operations are used.
11981 @opindex m4a-single
11982 Generate code for the SH4a assuming the floating-point unit is in
11983 single-precision mode by default.
11987 Generate code for the SH4a.
11991 Same as @option{-m4a-nofpu}, except that it implicitly passes
11992 @option{-dsp} to the assembler. GCC doesn't generate any DSP
11993 instructions at the moment.
11997 Compile code for the processor in big endian mode.
12001 Compile code for the processor in little endian mode.
12005 Align doubles at 64-bit boundaries. Note that this changes the calling
12006 conventions, and thus some functions from the standard C library will
12007 not work unless you recompile it first with @option{-mdalign}.
12011 Shorten some address references at link time, when possible; uses the
12012 linker option @option{-relax}.
12016 Use 32-bit offsets in @code{switch} tables. The default is to use
12021 Enable the use of the instruction @code{fmovd}.
12025 Comply with the calling conventions defined by Renesas.
12029 Comply with the calling conventions defined by Renesas.
12033 Comply with the calling conventions defined for GCC before the Renesas
12034 conventions were available. This option is the default for all
12035 targets of the SH toolchain except for @samp{sh-symbianelf}.
12038 @opindex mnomacsave
12039 Mark the @code{MAC} register as call-clobbered, even if
12040 @option{-mhitachi} is given.
12044 Increase IEEE-compliance of floating-point code.
12045 At the moment, this is equivalent to @option{-fno-finite-math-only}.
12046 When generating 16 bit SH opcodes, getting IEEE-conforming results for
12047 comparisons of NANs / infinities incurs extra overhead in every
12048 floating point comparison, therefore the default is set to
12049 @option{-ffinite-math-only}.
12053 Dump instruction size and location in the assembly code.
12056 @opindex mpadstruct
12057 This option is deprecated. It pads structures to multiple of 4 bytes,
12058 which is incompatible with the SH ABI@.
12062 Optimize for space instead of speed. Implied by @option{-Os}.
12065 @opindex mprefergot
12066 When generating position-independent code, emit function calls using
12067 the Global Offset Table instead of the Procedure Linkage Table.
12071 Generate a library function call to invalidate instruction cache
12072 entries, after fixing up a trampoline. This library function call
12073 doesn't assume it can write to the whole memory address space. This
12074 is the default when the target is @code{sh-*-linux*}.
12076 @item -multcost=@var{number}
12077 @opindex multcost=@var{number}
12078 Set the cost to assume for a multiply insn.
12080 @item -mdiv=@var{strategy}
12081 @opindex mdiv=@var{strategy}
12082 Set the division strategy to use for SHmedia code. @var{strategy} must be
12083 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
12084 inv:call2, inv:fp .
12085 "fp" performs the operation in floating point. This has a very high latency,
12086 but needs only a few instructions, so it might be a good choice if
12087 your code has enough easily exploitable ILP to allow the compiler to
12088 schedule the floating point instructions together with other instructions.
12089 Division by zero causes a floating point exception.
12090 "inv" uses integer operations to calculate the inverse of the divisor,
12091 and then multiplies the dividend with the inverse. This strategy allows
12092 cse and hoisting of the inverse calculation. Division by zero calculates
12093 an unspecified result, but does not trap.
12094 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
12095 have been found, or if the entire operation has been hoisted to the same
12096 place, the last stages of the inverse calculation are intertwined with the
12097 final multiply to reduce the overall latency, at the expense of using a few
12098 more instructions, and thus offering fewer scheduling opportunities with
12100 "call" calls a library function that usually implements the inv:minlat
12102 This gives high code density for m5-*media-nofpu compilations.
12103 "call2" uses a different entry point of the same library function, where it
12104 assumes that a pointer to a lookup table has already been set up, which
12105 exposes the pointer load to cse / code hoisting optimizations.
12106 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
12107 code generation, but if the code stays unoptimized, revert to the "call",
12108 "call2", or "fp" strategies, respectively. Note that the
12109 potentially-trapping side effect of division by zero is carried by a
12110 separate instruction, so it is possible that all the integer instructions
12111 are hoisted out, but the marker for the side effect stays where it is.
12112 A recombination to fp operations or a call is not possible in that case.
12113 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
12114 that the inverse calculation was nor separated from the multiply, they speed
12115 up division where the dividend fits into 20 bits (plus sign where applicable),
12116 by inserting a test to skip a number of operations in this case; this test
12117 slows down the case of larger dividends. inv20u assumes the case of a such
12118 a small dividend to be unlikely, and inv20l assumes it to be likely.
12120 @item -mdivsi3_libfunc=@var{name}
12121 @opindex mdivsi3_libfunc=@var{name}
12122 Set the name of the library function used for 32 bit signed division to
12123 @var{name}. This only affect the name used in the call and inv:call
12124 division strategies, and the compiler will still expect the same
12125 sets of input/output/clobbered registers as if this option was not present.
12127 @item -madjust-unroll
12128 @opindex madjust-unroll
12129 Throttle unrolling to avoid thrashing target registers.
12130 This option only has an effect if the gcc code base supports the
12131 TARGET_ADJUST_UNROLL_MAX target hook.
12133 @item -mindexed-addressing
12134 @opindex mindexed-addressing
12135 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
12136 This is only safe if the hardware and/or OS implement 32 bit wrap-around
12137 semantics for the indexed addressing mode. The architecture allows the
12138 implementation of processors with 64 bit MMU, which the OS could use to
12139 get 32 bit addressing, but since no current hardware implementation supports
12140 this or any other way to make the indexed addressing mode safe to use in
12141 the 32 bit ABI, the default is -mno-indexed-addressing.
12143 @item -mgettrcost=@var{number}
12144 @opindex mgettrcost=@var{number}
12145 Set the cost assumed for the gettr instruction to @var{number}.
12146 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
12150 Assume pt* instructions won't trap. This will generally generate better
12151 scheduled code, but is unsafe on current hardware. The current architecture
12152 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
12153 This has the unintentional effect of making it unsafe to schedule ptabs /
12154 ptrel before a branch, or hoist it out of a loop. For example,
12155 __do_global_ctors, a part of libgcc that runs constructors at program
12156 startup, calls functions in a list which is delimited by -1. With the
12157 -mpt-fixed option, the ptabs will be done before testing against -1.
12158 That means that all the constructors will be run a bit quicker, but when
12159 the loop comes to the end of the list, the program crashes because ptabs
12160 loads -1 into a target register. Since this option is unsafe for any
12161 hardware implementing the current architecture specification, the default
12162 is -mno-pt-fixed. Unless the user specifies a specific cost with
12163 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
12164 this deters register allocation using target registers for storing
12167 @item -minvalid-symbols
12168 @opindex minvalid-symbols
12169 Assume symbols might be invalid. Ordinary function symbols generated by
12170 the compiler will always be valid to load with movi/shori/ptabs or
12171 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
12172 to generate symbols that will cause ptabs / ptrel to trap.
12173 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
12174 It will then prevent cross-basic-block cse, hoisting and most scheduling
12175 of symbol loads. The default is @option{-mno-invalid-symbols}.
12178 @node SPARC Options
12179 @subsection SPARC Options
12180 @cindex SPARC options
12182 These @samp{-m} options are supported on the SPARC:
12185 @item -mno-app-regs
12187 @opindex mno-app-regs
12189 Specify @option{-mapp-regs} to generate output using the global registers
12190 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
12193 To be fully SVR4 ABI compliant at the cost of some performance loss,
12194 specify @option{-mno-app-regs}. You should compile libraries and system
12195 software with this option.
12198 @itemx -mhard-float
12200 @opindex mhard-float
12201 Generate output containing floating point instructions. This is the
12205 @itemx -msoft-float
12207 @opindex msoft-float
12208 Generate output containing library calls for floating point.
12209 @strong{Warning:} the requisite libraries are not available for all SPARC
12210 targets. Normally the facilities of the machine's usual C compiler are
12211 used, but this cannot be done directly in cross-compilation. You must make
12212 your own arrangements to provide suitable library functions for
12213 cross-compilation. The embedded targets @samp{sparc-*-aout} and
12214 @samp{sparclite-*-*} do provide software floating point support.
12216 @option{-msoft-float} changes the calling convention in the output file;
12217 therefore, it is only useful if you compile @emph{all} of a program with
12218 this option. In particular, you need to compile @file{libgcc.a}, the
12219 library that comes with GCC, with @option{-msoft-float} in order for
12222 @item -mhard-quad-float
12223 @opindex mhard-quad-float
12224 Generate output containing quad-word (long double) floating point
12227 @item -msoft-quad-float
12228 @opindex msoft-quad-float
12229 Generate output containing library calls for quad-word (long double)
12230 floating point instructions. The functions called are those specified
12231 in the SPARC ABI@. This is the default.
12233 As of this writing, there are no SPARC implementations that have hardware
12234 support for the quad-word floating point instructions. They all invoke
12235 a trap handler for one of these instructions, and then the trap handler
12236 emulates the effect of the instruction. Because of the trap handler overhead,
12237 this is much slower than calling the ABI library routines. Thus the
12238 @option{-msoft-quad-float} option is the default.
12240 @item -mno-unaligned-doubles
12241 @itemx -munaligned-doubles
12242 @opindex mno-unaligned-doubles
12243 @opindex munaligned-doubles
12244 Assume that doubles have 8 byte alignment. This is the default.
12246 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
12247 alignment only if they are contained in another type, or if they have an
12248 absolute address. Otherwise, it assumes they have 4 byte alignment.
12249 Specifying this option avoids some rare compatibility problems with code
12250 generated by other compilers. It is not the default because it results
12251 in a performance loss, especially for floating point code.
12253 @item -mno-faster-structs
12254 @itemx -mfaster-structs
12255 @opindex mno-faster-structs
12256 @opindex mfaster-structs
12257 With @option{-mfaster-structs}, the compiler assumes that structures
12258 should have 8 byte alignment. This enables the use of pairs of
12259 @code{ldd} and @code{std} instructions for copies in structure
12260 assignment, in place of twice as many @code{ld} and @code{st} pairs.
12261 However, the use of this changed alignment directly violates the SPARC
12262 ABI@. Thus, it's intended only for use on targets where the developer
12263 acknowledges that their resulting code will not be directly in line with
12264 the rules of the ABI@.
12266 @item -mimpure-text
12267 @opindex mimpure-text
12268 @option{-mimpure-text}, used in addition to @option{-shared}, tells
12269 the compiler to not pass @option{-z text} to the linker when linking a
12270 shared object. Using this option, you can link position-dependent
12271 code into a shared object.
12273 @option{-mimpure-text} suppresses the ``relocations remain against
12274 allocatable but non-writable sections'' linker error message.
12275 However, the necessary relocations will trigger copy-on-write, and the
12276 shared object is not actually shared across processes. Instead of
12277 using @option{-mimpure-text}, you should compile all source code with
12278 @option{-fpic} or @option{-fPIC}.
12280 This option is only available on SunOS and Solaris.
12282 @item -mcpu=@var{cpu_type}
12284 Set the instruction set, register set, and instruction scheduling parameters
12285 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
12286 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
12287 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
12288 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
12289 @samp{ultrasparc3}.
12291 Default instruction scheduling parameters are used for values that select
12292 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
12293 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
12295 Here is a list of each supported architecture and their supported
12300 v8: supersparc, hypersparc
12301 sparclite: f930, f934, sparclite86x
12303 v9: ultrasparc, ultrasparc3
12306 By default (unless configured otherwise), GCC generates code for the V7
12307 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
12308 additionally optimizes it for the Cypress CY7C602 chip, as used in the
12309 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
12310 SPARCStation 1, 2, IPX etc.
12312 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
12313 architecture. The only difference from V7 code is that the compiler emits
12314 the integer multiply and integer divide instructions which exist in SPARC-V8
12315 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
12316 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
12319 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
12320 the SPARC architecture. This adds the integer multiply, integer divide step
12321 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
12322 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
12323 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
12324 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
12325 MB86934 chip, which is the more recent SPARClite with FPU@.
12327 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
12328 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
12329 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
12330 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
12331 optimizes it for the TEMIC SPARClet chip.
12333 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
12334 architecture. This adds 64-bit integer and floating-point move instructions,
12335 3 additional floating-point condition code registers and conditional move
12336 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
12337 optimizes it for the Sun UltraSPARC I/II chips. With
12338 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
12339 Sun UltraSPARC III chip.
12341 @item -mtune=@var{cpu_type}
12343 Set the instruction scheduling parameters for machine type
12344 @var{cpu_type}, but do not set the instruction set or register set that the
12345 option @option{-mcpu=@var{cpu_type}} would.
12347 The same values for @option{-mcpu=@var{cpu_type}} can be used for
12348 @option{-mtune=@var{cpu_type}}, but the only useful values are those
12349 that select a particular cpu implementation. Those are @samp{cypress},
12350 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
12351 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
12352 @samp{ultrasparc3}.
12357 @opindex mno-v8plus
12358 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
12359 difference from the V8 ABI is that the global and out registers are
12360 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
12361 mode for all SPARC-V9 processors.
12367 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
12368 Visual Instruction Set extensions. The default is @option{-mno-vis}.
12371 These @samp{-m} options are supported in addition to the above
12372 on SPARC-V9 processors in 64-bit environments:
12375 @item -mlittle-endian
12376 @opindex mlittle-endian
12377 Generate code for a processor running in little-endian mode. It is only
12378 available for a few configurations and most notably not on Solaris and Linux.
12384 Generate code for a 32-bit or 64-bit environment.
12385 The 32-bit environment sets int, long and pointer to 32 bits.
12386 The 64-bit environment sets int to 32 bits and long and pointer
12389 @item -mcmodel=medlow
12390 @opindex mcmodel=medlow
12391 Generate code for the Medium/Low code model: 64-bit addresses, programs
12392 must be linked in the low 32 bits of memory. Programs can be statically
12393 or dynamically linked.
12395 @item -mcmodel=medmid
12396 @opindex mcmodel=medmid
12397 Generate code for the Medium/Middle code model: 64-bit addresses, programs
12398 must be linked in the low 44 bits of memory, the text and data segments must
12399 be less than 2GB in size and the data segment must be located within 2GB of
12402 @item -mcmodel=medany
12403 @opindex mcmodel=medany
12404 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
12405 may be linked anywhere in memory, the text and data segments must be less
12406 than 2GB in size and the data segment must be located within 2GB of the
12409 @item -mcmodel=embmedany
12410 @opindex mcmodel=embmedany
12411 Generate code for the Medium/Anywhere code model for embedded systems:
12412 64-bit addresses, the text and data segments must be less than 2GB in
12413 size, both starting anywhere in memory (determined at link time). The
12414 global register %g4 points to the base of the data segment. Programs
12415 are statically linked and PIC is not supported.
12418 @itemx -mno-stack-bias
12419 @opindex mstack-bias
12420 @opindex mno-stack-bias
12421 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
12422 frame pointer if present, are offset by @minus{}2047 which must be added back
12423 when making stack frame references. This is the default in 64-bit mode.
12424 Otherwise, assume no such offset is present.
12427 These switches are supported in addition to the above on Solaris:
12432 Add support for multithreading using the Solaris threads library. This
12433 option sets flags for both the preprocessor and linker. This option does
12434 not affect the thread safety of object code produced by the compiler or
12435 that of libraries supplied with it.
12439 Add support for multithreading using the POSIX threads library. This
12440 option sets flags for both the preprocessor and linker. This option does
12441 not affect the thread safety of object code produced by the compiler or
12442 that of libraries supplied with it.
12446 This is a synonym for @option{-pthreads}.
12449 @node System V Options
12450 @subsection Options for System V
12452 These additional options are available on System V Release 4 for
12453 compatibility with other compilers on those systems:
12458 Create a shared object.
12459 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
12463 Identify the versions of each tool used by the compiler, in a
12464 @code{.ident} assembler directive in the output.
12468 Refrain from adding @code{.ident} directives to the output file (this is
12471 @item -YP,@var{dirs}
12473 Search the directories @var{dirs}, and no others, for libraries
12474 specified with @option{-l}.
12476 @item -Ym,@var{dir}
12478 Look in the directory @var{dir} to find the M4 preprocessor.
12479 The assembler uses this option.
12480 @c This is supposed to go with a -Yd for predefined M4 macro files, but
12481 @c the generic assembler that comes with Solaris takes just -Ym.
12484 @node TMS320C3x/C4x Options
12485 @subsection TMS320C3x/C4x Options
12486 @cindex TMS320C3x/C4x Options
12488 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
12492 @item -mcpu=@var{cpu_type}
12494 Set the instruction set, register set, and instruction scheduling
12495 parameters for machine type @var{cpu_type}. Supported values for
12496 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
12497 @samp{c44}. The default is @samp{c40} to generate code for the
12502 @itemx -msmall-memory
12504 @opindex mbig-memory
12506 @opindex msmall-memory
12508 Generates code for the big or small memory model. The small memory
12509 model assumed that all data fits into one 64K word page. At run-time
12510 the data page (DP) register must be set to point to the 64K page
12511 containing the .bss and .data program sections. The big memory model is
12512 the default and requires reloading of the DP register for every direct
12519 Allow (disallow) allocation of general integer operands into the block
12520 count register BK@.
12526 Enable (disable) generation of code using decrement and branch,
12527 DBcond(D), instructions. This is enabled by default for the C4x. To be
12528 on the safe side, this is disabled for the C3x, since the maximum
12529 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
12530 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
12531 that it can utilize the decrement and branch instruction, but will give
12532 up if there is more than one memory reference in the loop. Thus a loop
12533 where the loop counter is decremented can generate slightly more
12534 efficient code, in cases where the RPTB instruction cannot be utilized.
12536 @item -mdp-isr-reload
12538 @opindex mdp-isr-reload
12540 Force the DP register to be saved on entry to an interrupt service
12541 routine (ISR), reloaded to point to the data section, and restored on
12542 exit from the ISR@. This should not be required unless someone has
12543 violated the small memory model by modifying the DP register, say within
12550 For the C3x use the 24-bit MPYI instruction for integer multiplies
12551 instead of a library call to guarantee 32-bit results. Note that if one
12552 of the operands is a constant, then the multiplication will be performed
12553 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
12554 then squaring operations are performed inline instead of a library call.
12557 @itemx -mno-fast-fix
12559 @opindex mno-fast-fix
12560 The C3x/C4x FIX instruction to convert a floating point value to an
12561 integer value chooses the nearest integer less than or equal to the
12562 floating point value rather than to the nearest integer. Thus if the
12563 floating point number is negative, the result will be incorrectly
12564 truncated an additional code is necessary to detect and correct this
12565 case. This option can be used to disable generation of the additional
12566 code required to correct the result.
12572 Enable (disable) generation of repeat block sequences using the RPTB
12573 instruction for zero overhead looping. The RPTB construct is only used
12574 for innermost loops that do not call functions or jump across the loop
12575 boundaries. There is no advantage having nested RPTB loops due to the
12576 overhead required to save and restore the RC, RS, and RE registers.
12577 This is enabled by default with @option{-O2}.
12579 @item -mrpts=@var{count}
12583 Enable (disable) the use of the single instruction repeat instruction
12584 RPTS@. If a repeat block contains a single instruction, and the loop
12585 count can be guaranteed to be less than the value @var{count}, GCC will
12586 emit a RPTS instruction instead of a RPTB@. If no value is specified,
12587 then a RPTS will be emitted even if the loop count cannot be determined
12588 at compile time. Note that the repeated instruction following RPTS does
12589 not have to be reloaded from memory each iteration, thus freeing up the
12590 CPU buses for operands. However, since interrupts are blocked by this
12591 instruction, it is disabled by default.
12593 @item -mloop-unsigned
12594 @itemx -mno-loop-unsigned
12595 @opindex mloop-unsigned
12596 @opindex mno-loop-unsigned
12597 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
12598 is @math{2^{31} + 1} since these instructions test if the iteration count is
12599 negative to terminate the loop. If the iteration count is unsigned
12600 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
12601 exceeded. This switch allows an unsigned iteration count.
12605 Try to emit an assembler syntax that the TI assembler (asm30) is happy
12606 with. This also enforces compatibility with the API employed by the TI
12607 C3x C compiler. For example, long doubles are passed as structures
12608 rather than in floating point registers.
12614 Generate code that uses registers (stack) for passing arguments to functions.
12615 By default, arguments are passed in registers where possible rather
12616 than by pushing arguments on to the stack.
12618 @item -mparallel-insns
12619 @itemx -mno-parallel-insns
12620 @opindex mparallel-insns
12621 @opindex mno-parallel-insns
12622 Allow the generation of parallel instructions. This is enabled by
12623 default with @option{-O2}.
12625 @item -mparallel-mpy
12626 @itemx -mno-parallel-mpy
12627 @opindex mparallel-mpy
12628 @opindex mno-parallel-mpy
12629 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
12630 provided @option{-mparallel-insns} is also specified. These instructions have
12631 tight register constraints which can pessimize the code generation
12632 of large functions.
12637 @subsection V850 Options
12638 @cindex V850 Options
12640 These @samp{-m} options are defined for V850 implementations:
12644 @itemx -mno-long-calls
12645 @opindex mlong-calls
12646 @opindex mno-long-calls
12647 Treat all calls as being far away (near). If calls are assumed to be
12648 far away, the compiler will always load the functions address up into a
12649 register, and call indirect through the pointer.
12655 Do not optimize (do optimize) basic blocks that use the same index
12656 pointer 4 or more times to copy pointer into the @code{ep} register, and
12657 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
12658 option is on by default if you optimize.
12660 @item -mno-prolog-function
12661 @itemx -mprolog-function
12662 @opindex mno-prolog-function
12663 @opindex mprolog-function
12664 Do not use (do use) external functions to save and restore registers
12665 at the prologue and epilogue of a function. The external functions
12666 are slower, but use less code space if more than one function saves
12667 the same number of registers. The @option{-mprolog-function} option
12668 is on by default if you optimize.
12672 Try to make the code as small as possible. At present, this just turns
12673 on the @option{-mep} and @option{-mprolog-function} options.
12675 @item -mtda=@var{n}
12677 Put static or global variables whose size is @var{n} bytes or less into
12678 the tiny data area that register @code{ep} points to. The tiny data
12679 area can hold up to 256 bytes in total (128 bytes for byte references).
12681 @item -msda=@var{n}
12683 Put static or global variables whose size is @var{n} bytes or less into
12684 the small data area that register @code{gp} points to. The small data
12685 area can hold up to 64 kilobytes.
12687 @item -mzda=@var{n}
12689 Put static or global variables whose size is @var{n} bytes or less into
12690 the first 32 kilobytes of memory.
12694 Specify that the target processor is the V850.
12697 @opindex mbig-switch
12698 Generate code suitable for big switch tables. Use this option only if
12699 the assembler/linker complain about out of range branches within a switch
12704 This option will cause r2 and r5 to be used in the code generated by
12705 the compiler. This setting is the default.
12707 @item -mno-app-regs
12708 @opindex mno-app-regs
12709 This option will cause r2 and r5 to be treated as fixed registers.
12713 Specify that the target processor is the V850E1. The preprocessor
12714 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
12715 this option is used.
12719 Specify that the target processor is the V850E@. The preprocessor
12720 constant @samp{__v850e__} will be defined if this option is used.
12722 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
12723 are defined then a default target processor will be chosen and the
12724 relevant @samp{__v850*__} preprocessor constant will be defined.
12726 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
12727 defined, regardless of which processor variant is the target.
12729 @item -mdisable-callt
12730 @opindex mdisable-callt
12731 This option will suppress generation of the CALLT instruction for the
12732 v850e and v850e1 flavors of the v850 architecture. The default is
12733 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
12738 @subsection VAX Options
12739 @cindex VAX options
12741 These @samp{-m} options are defined for the VAX:
12746 Do not output certain jump instructions (@code{aobleq} and so on)
12747 that the Unix assembler for the VAX cannot handle across long
12752 Do output those jump instructions, on the assumption that you
12753 will assemble with the GNU assembler.
12757 Output code for g-format floating point numbers instead of d-format.
12760 @node x86-64 Options
12761 @subsection x86-64 Options
12762 @cindex x86-64 options
12764 These are listed under @xref{i386 and x86-64 Options}.
12766 @node Xstormy16 Options
12767 @subsection Xstormy16 Options
12768 @cindex Xstormy16 Options
12770 These options are defined for Xstormy16:
12775 Choose startup files and linker script suitable for the simulator.
12778 @node Xtensa Options
12779 @subsection Xtensa Options
12780 @cindex Xtensa Options
12782 These options are supported for Xtensa targets:
12786 @itemx -mno-const16
12788 @opindex mno-const16
12789 Enable or disable use of @code{CONST16} instructions for loading
12790 constant values. The @code{CONST16} instruction is currently not a
12791 standard option from Tensilica. When enabled, @code{CONST16}
12792 instructions are always used in place of the standard @code{L32R}
12793 instructions. The use of @code{CONST16} is enabled by default only if
12794 the @code{L32R} instruction is not available.
12797 @itemx -mno-fused-madd
12798 @opindex mfused-madd
12799 @opindex mno-fused-madd
12800 Enable or disable use of fused multiply/add and multiply/subtract
12801 instructions in the floating-point option. This has no effect if the
12802 floating-point option is not also enabled. Disabling fused multiply/add
12803 and multiply/subtract instructions forces the compiler to use separate
12804 instructions for the multiply and add/subtract operations. This may be
12805 desirable in some cases where strict IEEE 754-compliant results are
12806 required: the fused multiply add/subtract instructions do not round the
12807 intermediate result, thereby producing results with @emph{more} bits of
12808 precision than specified by the IEEE standard. Disabling fused multiply
12809 add/subtract instructions also ensures that the program output is not
12810 sensitive to the compiler's ability to combine multiply and add/subtract
12813 @item -mtext-section-literals
12814 @itemx -mno-text-section-literals
12815 @opindex mtext-section-literals
12816 @opindex mno-text-section-literals
12817 Control the treatment of literal pools. The default is
12818 @option{-mno-text-section-literals}, which places literals in a separate
12819 section in the output file. This allows the literal pool to be placed
12820 in a data RAM/ROM, and it also allows the linker to combine literal
12821 pools from separate object files to remove redundant literals and
12822 improve code size. With @option{-mtext-section-literals}, the literals
12823 are interspersed in the text section in order to keep them as close as
12824 possible to their references. This may be necessary for large assembly
12827 @item -mtarget-align
12828 @itemx -mno-target-align
12829 @opindex mtarget-align
12830 @opindex mno-target-align
12831 When this option is enabled, GCC instructs the assembler to
12832 automatically align instructions to reduce branch penalties at the
12833 expense of some code density. The assembler attempts to widen density
12834 instructions to align branch targets and the instructions following call
12835 instructions. If there are not enough preceding safe density
12836 instructions to align a target, no widening will be performed. The
12837 default is @option{-mtarget-align}. These options do not affect the
12838 treatment of auto-aligned instructions like @code{LOOP}, which the
12839 assembler will always align, either by widening density instructions or
12840 by inserting no-op instructions.
12843 @itemx -mno-longcalls
12844 @opindex mlongcalls
12845 @opindex mno-longcalls
12846 When this option is enabled, GCC instructs the assembler to translate
12847 direct calls to indirect calls unless it can determine that the target
12848 of a direct call is in the range allowed by the call instruction. This
12849 translation typically occurs for calls to functions in other source
12850 files. Specifically, the assembler translates a direct @code{CALL}
12851 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
12852 The default is @option{-mno-longcalls}. This option should be used in
12853 programs where the call target can potentially be out of range. This
12854 option is implemented in the assembler, not the compiler, so the
12855 assembly code generated by GCC will still show direct call
12856 instructions---look at the disassembled object code to see the actual
12857 instructions. Note that the assembler will use an indirect call for
12858 every cross-file call, not just those that really will be out of range.
12861 @node zSeries Options
12862 @subsection zSeries Options
12863 @cindex zSeries options
12865 These are listed under @xref{S/390 and zSeries Options}.
12867 @node Code Gen Options
12868 @section Options for Code Generation Conventions
12869 @cindex code generation conventions
12870 @cindex options, code generation
12871 @cindex run-time options
12873 These machine-independent options control the interface conventions
12874 used in code generation.
12876 Most of them have both positive and negative forms; the negative form
12877 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
12878 one of the forms is listed---the one which is not the default. You
12879 can figure out the other form by either removing @samp{no-} or adding
12883 @item -fbounds-check
12884 @opindex fbounds-check
12885 For front-ends that support it, generate additional code to check that
12886 indices used to access arrays are within the declared range. This is
12887 currently only supported by the Java and Fortran 77 front-ends, where
12888 this option defaults to true and false respectively.
12892 This option generates traps for signed overflow on addition, subtraction,
12893 multiplication operations.
12897 This option instructs the compiler to assume that signed arithmetic
12898 overflow of addition, subtraction and multiplication wraps around
12899 using twos-complement representation. This flag enables some optimizations
12900 and disables others. This option is enabled by default for the Java
12901 front-end, as required by the Java language specification.
12904 @opindex fexceptions
12905 Enable exception handling. Generates extra code needed to propagate
12906 exceptions. For some targets, this implies GCC will generate frame
12907 unwind information for all functions, which can produce significant data
12908 size overhead, although it does not affect execution. If you do not
12909 specify this option, GCC will enable it by default for languages like
12910 C++ which normally require exception handling, and disable it for
12911 languages like C that do not normally require it. However, you may need
12912 to enable this option when compiling C code that needs to interoperate
12913 properly with exception handlers written in C++. You may also wish to
12914 disable this option if you are compiling older C++ programs that don't
12915 use exception handling.
12917 @item -fnon-call-exceptions
12918 @opindex fnon-call-exceptions
12919 Generate code that allows trapping instructions to throw exceptions.
12920 Note that this requires platform-specific runtime support that does
12921 not exist everywhere. Moreover, it only allows @emph{trapping}
12922 instructions to throw exceptions, i.e.@: memory references or floating
12923 point instructions. It does not allow exceptions to be thrown from
12924 arbitrary signal handlers such as @code{SIGALRM}.
12926 @item -funwind-tables
12927 @opindex funwind-tables
12928 Similar to @option{-fexceptions}, except that it will just generate any needed
12929 static data, but will not affect the generated code in any other way.
12930 You will normally not enable this option; instead, a language processor
12931 that needs this handling would enable it on your behalf.
12933 @item -fasynchronous-unwind-tables
12934 @opindex fasynchronous-unwind-tables
12935 Generate unwind table in dwarf2 format, if supported by target machine. The
12936 table is exact at each instruction boundary, so it can be used for stack
12937 unwinding from asynchronous events (such as debugger or garbage collector).
12939 @item -fpcc-struct-return
12940 @opindex fpcc-struct-return
12941 Return ``short'' @code{struct} and @code{union} values in memory like
12942 longer ones, rather than in registers. This convention is less
12943 efficient, but it has the advantage of allowing intercallability between
12944 GCC-compiled files and files compiled with other compilers, particularly
12945 the Portable C Compiler (pcc).
12947 The precise convention for returning structures in memory depends
12948 on the target configuration macros.
12950 Short structures and unions are those whose size and alignment match
12951 that of some integer type.
12953 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
12954 switch is not binary compatible with code compiled with the
12955 @option{-freg-struct-return} switch.
12956 Use it to conform to a non-default application binary interface.
12958 @item -freg-struct-return
12959 @opindex freg-struct-return
12960 Return @code{struct} and @code{union} values in registers when possible.
12961 This is more efficient for small structures than
12962 @option{-fpcc-struct-return}.
12964 If you specify neither @option{-fpcc-struct-return} nor
12965 @option{-freg-struct-return}, GCC defaults to whichever convention is
12966 standard for the target. If there is no standard convention, GCC
12967 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
12968 the principal compiler. In those cases, we can choose the standard, and
12969 we chose the more efficient register return alternative.
12971 @strong{Warning:} code compiled with the @option{-freg-struct-return}
12972 switch is not binary compatible with code compiled with the
12973 @option{-fpcc-struct-return} switch.
12974 Use it to conform to a non-default application binary interface.
12976 @item -fshort-enums
12977 @opindex fshort-enums
12978 Allocate to an @code{enum} type only as many bytes as it needs for the
12979 declared range of possible values. Specifically, the @code{enum} type
12980 will be equivalent to the smallest integer type which has enough room.
12982 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
12983 code that is not binary compatible with code generated without that switch.
12984 Use it to conform to a non-default application binary interface.
12986 @item -fshort-double
12987 @opindex fshort-double
12988 Use the same size for @code{double} as for @code{float}.
12990 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
12991 code that is not binary compatible with code generated without that switch.
12992 Use it to conform to a non-default application binary interface.
12994 @item -fshort-wchar
12995 @opindex fshort-wchar
12996 Override the underlying type for @samp{wchar_t} to be @samp{short
12997 unsigned int} instead of the default for the target. This option is
12998 useful for building programs to run under WINE@.
13000 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
13001 code that is not binary compatible with code generated without that switch.
13002 Use it to conform to a non-default application binary interface.
13005 @opindex fno-common
13006 In C, allocate even uninitialized global variables in the data section of the
13007 object file, rather than generating them as common blocks. This has the
13008 effect that if the same variable is declared (without @code{extern}) in
13009 two different compilations, you will get an error when you link them.
13010 The only reason this might be useful is if you wish to verify that the
13011 program will work on other systems which always work this way.
13015 Ignore the @samp{#ident} directive.
13017 @item -finhibit-size-directive
13018 @opindex finhibit-size-directive
13019 Don't output a @code{.size} assembler directive, or anything else that
13020 would cause trouble if the function is split in the middle, and the
13021 two halves are placed at locations far apart in memory. This option is
13022 used when compiling @file{crtstuff.c}; you should not need to use it
13025 @item -fverbose-asm
13026 @opindex fverbose-asm
13027 Put extra commentary information in the generated assembly code to
13028 make it more readable. This option is generally only of use to those
13029 who actually need to read the generated assembly code (perhaps while
13030 debugging the compiler itself).
13032 @option{-fno-verbose-asm}, the default, causes the
13033 extra information to be omitted and is useful when comparing two assembler
13038 @cindex global offset table
13040 Generate position-independent code (PIC) suitable for use in a shared
13041 library, if supported for the target machine. Such code accesses all
13042 constant addresses through a global offset table (GOT)@. The dynamic
13043 loader resolves the GOT entries when the program starts (the dynamic
13044 loader is not part of GCC; it is part of the operating system). If
13045 the GOT size for the linked executable exceeds a machine-specific
13046 maximum size, you get an error message from the linker indicating that
13047 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
13048 instead. (These maximums are 8k on the SPARC and 32k
13049 on the m68k and RS/6000. The 386 has no such limit.)
13051 Position-independent code requires special support, and therefore works
13052 only on certain machines. For the 386, GCC supports PIC for System V
13053 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
13054 position-independent.
13056 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
13061 If supported for the target machine, emit position-independent code,
13062 suitable for dynamic linking and avoiding any limit on the size of the
13063 global offset table. This option makes a difference on the m68k,
13064 PowerPC and SPARC@.
13066 Position-independent code requires special support, and therefore works
13067 only on certain machines.
13069 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
13076 These options are similar to @option{-fpic} and @option{-fPIC}, but
13077 generated position independent code can be only linked into executables.
13078 Usually these options are used when @option{-pie} GCC option will be
13079 used during linking.
13081 @item -fno-jump-tables
13082 @opindex fno-jump-tables
13083 Do not use jump tables for switch statements even where it would be
13084 more efficient than other code generation strategies. This option is
13085 of use in conjunction with @option{-fpic} or @option{-fPIC} for
13086 building code which forms part of a dynamic linker and cannot
13087 reference the address of a jump table. On some targets, jump tables
13088 do not require a GOT and this option is not needed.
13090 @item -ffixed-@var{reg}
13092 Treat the register named @var{reg} as a fixed register; generated code
13093 should never refer to it (except perhaps as a stack pointer, frame
13094 pointer or in some other fixed role).
13096 @var{reg} must be the name of a register. The register names accepted
13097 are machine-specific and are defined in the @code{REGISTER_NAMES}
13098 macro in the machine description macro file.
13100 This flag does not have a negative form, because it specifies a
13103 @item -fcall-used-@var{reg}
13104 @opindex fcall-used
13105 Treat the register named @var{reg} as an allocable register that is
13106 clobbered by function calls. It may be allocated for temporaries or
13107 variables that do not live across a call. Functions compiled this way
13108 will not save and restore the register @var{reg}.
13110 It is an error to used this flag with the frame pointer or stack pointer.
13111 Use of this flag for other registers that have fixed pervasive roles in
13112 the machine's execution model will produce disastrous results.
13114 This flag does not have a negative form, because it specifies a
13117 @item -fcall-saved-@var{reg}
13118 @opindex fcall-saved
13119 Treat the register named @var{reg} as an allocable register saved by
13120 functions. It may be allocated even for temporaries or variables that
13121 live across a call. Functions compiled this way will save and restore
13122 the register @var{reg} if they use it.
13124 It is an error to used this flag with the frame pointer or stack pointer.
13125 Use of this flag for other registers that have fixed pervasive roles in
13126 the machine's execution model will produce disastrous results.
13128 A different sort of disaster will result from the use of this flag for
13129 a register in which function values may be returned.
13131 This flag does not have a negative form, because it specifies a
13134 @item -fpack-struct[=@var{n}]
13135 @opindex fpack-struct
13136 Without a value specified, pack all structure members together without
13137 holes. When a value is specified (which must be a small power of two), pack
13138 structure members according to this value, representing the maximum
13139 alignment (that is, objects with default alignment requirements larger than
13140 this will be output potentially unaligned at the next fitting location.
13142 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
13143 code that is not binary compatible with code generated without that switch.
13144 Additionally, it makes the code suboptimal.
13145 Use it to conform to a non-default application binary interface.
13147 @item -finstrument-functions
13148 @opindex finstrument-functions
13149 Generate instrumentation calls for entry and exit to functions. Just
13150 after function entry and just before function exit, the following
13151 profiling functions will be called with the address of the current
13152 function and its call site. (On some platforms,
13153 @code{__builtin_return_address} does not work beyond the current
13154 function, so the call site information may not be available to the
13155 profiling functions otherwise.)
13158 void __cyg_profile_func_enter (void *this_fn,
13160 void __cyg_profile_func_exit (void *this_fn,
13164 The first argument is the address of the start of the current function,
13165 which may be looked up exactly in the symbol table.
13167 This instrumentation is also done for functions expanded inline in other
13168 functions. The profiling calls will indicate where, conceptually, the
13169 inline function is entered and exited. This means that addressable
13170 versions of such functions must be available. If all your uses of a
13171 function are expanded inline, this may mean an additional expansion of
13172 code size. If you use @samp{extern inline} in your C code, an
13173 addressable version of such functions must be provided. (This is
13174 normally the case anyways, but if you get lucky and the optimizer always
13175 expands the functions inline, you might have gotten away without
13176 providing static copies.)
13178 A function may be given the attribute @code{no_instrument_function}, in
13179 which case this instrumentation will not be done. This can be used, for
13180 example, for the profiling functions listed above, high-priority
13181 interrupt routines, and any functions from which the profiling functions
13182 cannot safely be called (perhaps signal handlers, if the profiling
13183 routines generate output or allocate memory).
13185 @item -fstack-check
13186 @opindex fstack-check
13187 Generate code to verify that you do not go beyond the boundary of the
13188 stack. You should specify this flag if you are running in an
13189 environment with multiple threads, but only rarely need to specify it in
13190 a single-threaded environment since stack overflow is automatically
13191 detected on nearly all systems if there is only one stack.
13193 Note that this switch does not actually cause checking to be done; the
13194 operating system must do that. The switch causes generation of code
13195 to ensure that the operating system sees the stack being extended.
13197 @item -fstack-limit-register=@var{reg}
13198 @itemx -fstack-limit-symbol=@var{sym}
13199 @itemx -fno-stack-limit
13200 @opindex fstack-limit-register
13201 @opindex fstack-limit-symbol
13202 @opindex fno-stack-limit
13203 Generate code to ensure that the stack does not grow beyond a certain value,
13204 either the value of a register or the address of a symbol. If the stack
13205 would grow beyond the value, a signal is raised. For most targets,
13206 the signal is raised before the stack overruns the boundary, so
13207 it is possible to catch the signal without taking special precautions.
13209 For instance, if the stack starts at absolute address @samp{0x80000000}
13210 and grows downwards, you can use the flags
13211 @option{-fstack-limit-symbol=__stack_limit} and
13212 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
13213 of 128KB@. Note that this may only work with the GNU linker.
13215 @cindex aliasing of parameters
13216 @cindex parameters, aliased
13217 @item -fargument-alias
13218 @itemx -fargument-noalias
13219 @itemx -fargument-noalias-global
13220 @opindex fargument-alias
13221 @opindex fargument-noalias
13222 @opindex fargument-noalias-global
13223 Specify the possible relationships among parameters and between
13224 parameters and global data.
13226 @option{-fargument-alias} specifies that arguments (parameters) may
13227 alias each other and may alias global storage.@*
13228 @option{-fargument-noalias} specifies that arguments do not alias
13229 each other, but may alias global storage.@*
13230 @option{-fargument-noalias-global} specifies that arguments do not
13231 alias each other and do not alias global storage.
13233 Each language will automatically use whatever option is required by
13234 the language standard. You should not need to use these options yourself.
13236 @item -fleading-underscore
13237 @opindex fleading-underscore
13238 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
13239 change the way C symbols are represented in the object file. One use
13240 is to help link with legacy assembly code.
13242 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
13243 generate code that is not binary compatible with code generated without that
13244 switch. Use it to conform to a non-default application binary interface.
13245 Not all targets provide complete support for this switch.
13247 @item -ftls-model=@var{model}
13248 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
13249 The @var{model} argument should be one of @code{global-dynamic},
13250 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
13252 The default without @option{-fpic} is @code{initial-exec}; with
13253 @option{-fpic} the default is @code{global-dynamic}.
13255 @item -fvisibility=@var{default|internal|hidden|protected}
13256 @opindex fvisibility
13257 Set the default ELF image symbol visibility to the specified option---all
13258 symbols will be marked with this unless overridden within the code.
13259 Using this feature can very substantially improve linking and
13260 load times of shared object libraries, produce more optimized
13261 code, provide near-perfect API export and prevent symbol clashes.
13262 It is @strong{strongly} recommended that you use this in any shared objects
13265 Despite the nomenclature, @code{default} always means public ie;
13266 available to be linked against from outside the shared object.
13267 @code{protected} and @code{internal} are pretty useless in real-world
13268 usage so the only other commonly used option will be @code{hidden}.
13269 The default if @option{-fvisibility} isn't specified is
13270 @code{default}, i.e., make every
13271 symbol public---this causes the same behavior as previous versions of
13274 A good explanation of the benefits offered by ensuring ELF
13275 symbols have the correct visibility is given by ``How To Write
13276 Shared Libraries'' by Ulrich Drepper (which can be found at
13277 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
13278 solution made possible by this option to marking things hidden when
13279 the default is public is to make the default hidden and mark things
13280 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
13281 and @code{__attribute__ ((visibility("default")))} instead of
13282 @code{__declspec(dllexport)} you get almost identical semantics with
13283 identical syntax. This is a great boon to those working with
13284 cross-platform projects.
13286 For those adding visibility support to existing code, you may find
13287 @samp{#pragma GCC visibility} of use. This works by you enclosing
13288 the declarations you wish to set visibility for with (for example)
13289 @samp{#pragma GCC visibility push(hidden)} and
13290 @samp{#pragma GCC visibility pop}.
13291 Bear in mind that symbol visibility should be viewed @strong{as
13292 part of the API interface contract} and thus all new code should
13293 always specify visibility when it is not the default ie; declarations
13294 only for use within the local DSO should @strong{always} be marked explicitly
13295 as hidden as so to avoid PLT indirection overheads---making this
13296 abundantly clear also aids readability and self-documentation of the code.
13297 Note that due to ISO C++ specification requirements, operator new and
13298 operator delete must always be of default visibility.
13300 An overview of these techniques, their benefits and how to use them
13301 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
13307 @node Environment Variables
13308 @section Environment Variables Affecting GCC
13309 @cindex environment variables
13311 @c man begin ENVIRONMENT
13312 This section describes several environment variables that affect how GCC
13313 operates. Some of them work by specifying directories or prefixes to use
13314 when searching for various kinds of files. Some are used to specify other
13315 aspects of the compilation environment.
13317 Note that you can also specify places to search using options such as
13318 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
13319 take precedence over places specified using environment variables, which
13320 in turn take precedence over those specified by the configuration of GCC@.
13321 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
13322 GNU Compiler Collection (GCC) Internals}.
13327 @c @itemx LC_COLLATE
13329 @c @itemx LC_MONETARY
13330 @c @itemx LC_NUMERIC
13335 @c @findex LC_COLLATE
13336 @findex LC_MESSAGES
13337 @c @findex LC_MONETARY
13338 @c @findex LC_NUMERIC
13342 These environment variables control the way that GCC uses
13343 localization information that allow GCC to work with different
13344 national conventions. GCC inspects the locale categories
13345 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
13346 so. These locale categories can be set to any value supported by your
13347 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
13348 Kingdom encoded in UTF-8.
13350 The @env{LC_CTYPE} environment variable specifies character
13351 classification. GCC uses it to determine the character boundaries in
13352 a string; this is needed for some multibyte encodings that contain quote
13353 and escape characters that would otherwise be interpreted as a string
13356 The @env{LC_MESSAGES} environment variable specifies the language to
13357 use in diagnostic messages.
13359 If the @env{LC_ALL} environment variable is set, it overrides the value
13360 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
13361 and @env{LC_MESSAGES} default to the value of the @env{LANG}
13362 environment variable. If none of these variables are set, GCC
13363 defaults to traditional C English behavior.
13367 If @env{TMPDIR} is set, it specifies the directory to use for temporary
13368 files. GCC uses temporary files to hold the output of one stage of
13369 compilation which is to be used as input to the next stage: for example,
13370 the output of the preprocessor, which is the input to the compiler
13373 @item GCC_EXEC_PREFIX
13374 @findex GCC_EXEC_PREFIX
13375 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
13376 names of the subprograms executed by the compiler. No slash is added
13377 when this prefix is combined with the name of a subprogram, but you can
13378 specify a prefix that ends with a slash if you wish.
13380 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
13381 an appropriate prefix to use based on the pathname it was invoked with.
13383 If GCC cannot find the subprogram using the specified prefix, it
13384 tries looking in the usual places for the subprogram.
13386 The default value of @env{GCC_EXEC_PREFIX} is
13387 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
13388 of @code{prefix} when you ran the @file{configure} script.
13390 Other prefixes specified with @option{-B} take precedence over this prefix.
13392 This prefix is also used for finding files such as @file{crt0.o} that are
13395 In addition, the prefix is used in an unusual way in finding the
13396 directories to search for header files. For each of the standard
13397 directories whose name normally begins with @samp{/usr/local/lib/gcc}
13398 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
13399 replacing that beginning with the specified prefix to produce an
13400 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
13401 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
13402 These alternate directories are searched first; the standard directories
13405 @item COMPILER_PATH
13406 @findex COMPILER_PATH
13407 The value of @env{COMPILER_PATH} is a colon-separated list of
13408 directories, much like @env{PATH}. GCC tries the directories thus
13409 specified when searching for subprograms, if it can't find the
13410 subprograms using @env{GCC_EXEC_PREFIX}.
13413 @findex LIBRARY_PATH
13414 The value of @env{LIBRARY_PATH} is a colon-separated list of
13415 directories, much like @env{PATH}. When configured as a native compiler,
13416 GCC tries the directories thus specified when searching for special
13417 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
13418 using GCC also uses these directories when searching for ordinary
13419 libraries for the @option{-l} option (but directories specified with
13420 @option{-L} come first).
13424 @cindex locale definition
13425 This variable is used to pass locale information to the compiler. One way in
13426 which this information is used is to determine the character set to be used
13427 when character literals, string literals and comments are parsed in C and C++.
13428 When the compiler is configured to allow multibyte characters,
13429 the following values for @env{LANG} are recognized:
13433 Recognize JIS characters.
13435 Recognize SJIS characters.
13437 Recognize EUCJP characters.
13440 If @env{LANG} is not defined, or if it has some other value, then the
13441 compiler will use mblen and mbtowc as defined by the default locale to
13442 recognize and translate multibyte characters.
13446 Some additional environments variables affect the behavior of the
13449 @include cppenv.texi
13453 @node Precompiled Headers
13454 @section Using Precompiled Headers
13455 @cindex precompiled headers
13456 @cindex speed of compilation
13458 Often large projects have many header files that are included in every
13459 source file. The time the compiler takes to process these header files
13460 over and over again can account for nearly all of the time required to
13461 build the project. To make builds faster, GCC allows users to
13462 `precompile' a header file; then, if builds can use the precompiled
13463 header file they will be much faster.
13465 To create a precompiled header file, simply compile it as you would any
13466 other file, if necessary using the @option{-x} option to make the driver
13467 treat it as a C or C++ header file. You will probably want to use a
13468 tool like @command{make} to keep the precompiled header up-to-date when
13469 the headers it contains change.
13471 A precompiled header file will be searched for when @code{#include} is
13472 seen in the compilation. As it searches for the included file
13473 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
13474 compiler looks for a precompiled header in each directory just before it
13475 looks for the include file in that directory. The name searched for is
13476 the name specified in the @code{#include} with @samp{.gch} appended. If
13477 the precompiled header file can't be used, it is ignored.
13479 For instance, if you have @code{#include "all.h"}, and you have
13480 @file{all.h.gch} in the same directory as @file{all.h}, then the
13481 precompiled header file will be used if possible, and the original
13482 header will be used otherwise.
13484 Alternatively, you might decide to put the precompiled header file in a
13485 directory and use @option{-I} to ensure that directory is searched
13486 before (or instead of) the directory containing the original header.
13487 Then, if you want to check that the precompiled header file is always
13488 used, you can put a file of the same name as the original header in this
13489 directory containing an @code{#error} command.
13491 This also works with @option{-include}. So yet another way to use
13492 precompiled headers, good for projects not designed with precompiled
13493 header files in mind, is to simply take most of the header files used by
13494 a project, include them from another header file, precompile that header
13495 file, and @option{-include} the precompiled header. If the header files
13496 have guards against multiple inclusion, they will be skipped because
13497 they've already been included (in the precompiled header).
13499 If you need to precompile the same header file for different
13500 languages, targets, or compiler options, you can instead make a
13501 @emph{directory} named like @file{all.h.gch}, and put each precompiled
13502 header in the directory, perhaps using @option{-o}. It doesn't matter
13503 what you call the files in the directory, every precompiled header in
13504 the directory will be considered. The first precompiled header
13505 encountered in the directory that is valid for this compilation will
13506 be used; they're searched in no particular order.
13508 There are many other possibilities, limited only by your imagination,
13509 good sense, and the constraints of your build system.
13511 A precompiled header file can be used only when these conditions apply:
13515 Only one precompiled header can be used in a particular compilation.
13518 A precompiled header can't be used once the first C token is seen. You
13519 can have preprocessor directives before a precompiled header; you can
13520 even include a precompiled header from inside another header, so long as
13521 there are no C tokens before the @code{#include}.
13524 The precompiled header file must be produced for the same language as
13525 the current compilation. You can't use a C precompiled header for a C++
13529 The precompiled header file must have been produced by the same compiler
13530 binary as the current compilation is using.
13533 Any macros defined before the precompiled header is included must
13534 either be defined in the same way as when the precompiled header was
13535 generated, or must not affect the precompiled header, which usually
13536 means that they don't appear in the precompiled header at all.
13538 The @option{-D} option is one way to define a macro before a
13539 precompiled header is included; using a @code{#define} can also do it.
13540 There are also some options that define macros implicitly, like
13541 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
13544 @item If debugging information is output when using the precompiled
13545 header, using @option{-g} or similar, the same kind of debugging information
13546 must have been output when building the precompiled header. However,
13547 a precompiled header built using @option{-g} can be used in a compilation
13548 when no debugging information is being output.
13550 @item The same @option{-m} options must generally be used when building
13551 and using the precompiled header. @xref{Submodel Options},
13552 for any cases where this rule is relaxed.
13554 @item Each of the following options must be the same when building and using
13555 the precompiled header:
13557 @gccoptlist{-fexceptions -funit-at-a-time}
13560 Some other command-line options starting with @option{-f},
13561 @option{-p}, or @option{-O} must be defined in the same way as when
13562 the precompiled header was generated. At present, it's not clear
13563 which options are safe to change and which are not; the safest choice
13564 is to use exactly the same options when generating and using the
13565 precompiled header. The following are known to be safe:
13567 @gccoptlist{-fmessage-length= -fpreprocessed
13568 -fsched-interblock -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous
13569 -fsched-verbose=<number> -fschedule-insns -fvisibility=
13574 For all of these except the last, the compiler will automatically
13575 ignore the precompiled header if the conditions aren't met. If you
13576 find an option combination that doesn't work and doesn't cause the
13577 precompiled header to be ignored, please consider filing a bug report,
13580 If you do use differing options when generating and using the
13581 precompiled header, the actual behavior will be a mixture of the
13582 behavior for the options. For instance, if you use @option{-g} to
13583 generate the precompiled header but not when using it, you may or may
13584 not get debugging information for routines in the precompiled header.
13586 @node Running Protoize
13587 @section Running Protoize
13589 The program @code{protoize} is an optional part of GCC@. You can use
13590 it to add prototypes to a program, thus converting the program to ISO
13591 C in one respect. The companion program @code{unprotoize} does the
13592 reverse: it removes argument types from any prototypes that are found.
13594 When you run these programs, you must specify a set of source files as
13595 command line arguments. The conversion programs start out by compiling
13596 these files to see what functions they define. The information gathered
13597 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
13599 After scanning comes actual conversion. The specified files are all
13600 eligible to be converted; any files they include (whether sources or
13601 just headers) are eligible as well.
13603 But not all the eligible files are converted. By default,
13604 @code{protoize} and @code{unprotoize} convert only source and header
13605 files in the current directory. You can specify additional directories
13606 whose files should be converted with the @option{-d @var{directory}}
13607 option. You can also specify particular files to exclude with the
13608 @option{-x @var{file}} option. A file is converted if it is eligible, its
13609 directory name matches one of the specified directory names, and its
13610 name within the directory has not been excluded.
13612 Basic conversion with @code{protoize} consists of rewriting most
13613 function definitions and function declarations to specify the types of
13614 the arguments. The only ones not rewritten are those for varargs
13617 @code{protoize} optionally inserts prototype declarations at the
13618 beginning of the source file, to make them available for any calls that
13619 precede the function's definition. Or it can insert prototype
13620 declarations with block scope in the blocks where undeclared functions
13623 Basic conversion with @code{unprotoize} consists of rewriting most
13624 function declarations to remove any argument types, and rewriting
13625 function definitions to the old-style pre-ISO form.
13627 Both conversion programs print a warning for any function declaration or
13628 definition that they can't convert. You can suppress these warnings
13631 The output from @code{protoize} or @code{unprotoize} replaces the
13632 original source file. The original file is renamed to a name ending
13633 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
13634 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
13635 for DOS) file already exists, then the source file is simply discarded.
13637 @code{protoize} and @code{unprotoize} both depend on GCC itself to
13638 scan the program and collect information about the functions it uses.
13639 So neither of these programs will work until GCC is installed.
13641 Here is a table of the options you can use with @code{protoize} and
13642 @code{unprotoize}. Each option works with both programs unless
13646 @item -B @var{directory}
13647 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
13648 usual directory (normally @file{/usr/local/lib}). This file contains
13649 prototype information about standard system functions. This option
13650 applies only to @code{protoize}.
13652 @item -c @var{compilation-options}
13653 Use @var{compilation-options} as the options when running @command{gcc} to
13654 produce the @samp{.X} files. The special option @option{-aux-info} is
13655 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
13657 Note that the compilation options must be given as a single argument to
13658 @code{protoize} or @code{unprotoize}. If you want to specify several
13659 @command{gcc} options, you must quote the entire set of compilation options
13660 to make them a single word in the shell.
13662 There are certain @command{gcc} arguments that you cannot use, because they
13663 would produce the wrong kind of output. These include @option{-g},
13664 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
13665 the @var{compilation-options}, they are ignored.
13668 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
13669 systems) instead of @samp{.c}. This is convenient if you are converting
13670 a C program to C++. This option applies only to @code{protoize}.
13673 Add explicit global declarations. This means inserting explicit
13674 declarations at the beginning of each source file for each function
13675 that is called in the file and was not declared. These declarations
13676 precede the first function definition that contains a call to an
13677 undeclared function. This option applies only to @code{protoize}.
13679 @item -i @var{string}
13680 Indent old-style parameter declarations with the string @var{string}.
13681 This option applies only to @code{protoize}.
13683 @code{unprotoize} converts prototyped function definitions to old-style
13684 function definitions, where the arguments are declared between the
13685 argument list and the initial @samp{@{}. By default, @code{unprotoize}
13686 uses five spaces as the indentation. If you want to indent with just
13687 one space instead, use @option{-i " "}.
13690 Keep the @samp{.X} files. Normally, they are deleted after conversion
13694 Add explicit local declarations. @code{protoize} with @option{-l} inserts
13695 a prototype declaration for each function in each block which calls the
13696 function without any declaration. This option applies only to
13700 Make no real changes. This mode just prints information about the conversions
13701 that would have been done without @option{-n}.
13704 Make no @samp{.save} files. The original files are simply deleted.
13705 Use this option with caution.
13707 @item -p @var{program}
13708 Use the program @var{program} as the compiler. Normally, the name
13709 @file{gcc} is used.
13712 Work quietly. Most warnings are suppressed.
13715 Print the version number, just like @option{-v} for @command{gcc}.
13718 If you need special compiler options to compile one of your program's
13719 source files, then you should generate that file's @samp{.X} file
13720 specially, by running @command{gcc} on that source file with the
13721 appropriate options and the option @option{-aux-info}. Then run
13722 @code{protoize} on the entire set of files. @code{protoize} will use
13723 the existing @samp{.X} file because it is newer than the source file.
13727 gcc -Dfoo=bar file1.c -aux-info file1.X
13732 You need to include the special files along with the rest in the
13733 @code{protoize} command, even though their @samp{.X} files already
13734 exist, because otherwise they won't get converted.
13736 @xref{Protoize Caveats}, for more information on how to use
13737 @code{protoize} successfully.