1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007
3 @c Free Software Foundation, Inc.
4 @c This is part of the GCC manual.
5 @c For copying conditions, see the file gcc.texi.
12 @c man begin COPYRIGHT
13 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
14 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006,
15 2007 Free Software Foundation, Inc.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.2 or
19 any later version published by the Free Software Foundation; with the
20 Invariant Sections being ``GNU General Public License'' and ``Funding
21 Free Software'', the Front-Cover texts being (a) (see below), and with
22 the Back-Cover Texts being (b) (see below). A copy of the license is
23 included in the gfdl(7) man page.
25 (a) The FSF's Front-Cover Text is:
29 (b) The FSF's Back-Cover Text is:
31 You have freedom to copy and modify this GNU Manual, like GNU
32 software. Copies published by the Free Software Foundation raise
33 funds for GNU development.
35 @c Set file name and title for the man page.
37 @settitle GNU project C and C++ compiler
39 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
40 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
41 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
42 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
43 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
44 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
45 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
47 Only the most useful options are listed here; see below for the
48 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
51 gpl(7), gfdl(7), fsf-funding(7),
52 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
53 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
54 @file{ld}, @file{binutils} and @file{gdb}.
57 For instructions on reporting bugs, see
61 See the Info entry for @command{gcc}, or
62 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
63 for contributors to GCC@.
68 @chapter GCC Command Options
69 @cindex GCC command options
70 @cindex command options
71 @cindex options, GCC command
73 @c man begin DESCRIPTION
74 When you invoke GCC, it normally does preprocessing, compilation,
75 assembly and linking. The ``overall options'' allow you to stop this
76 process at an intermediate stage. For example, the @option{-c} option
77 says not to run the linker. Then the output consists of object files
78 output by the assembler.
80 Other options are passed on to one stage of processing. Some options
81 control the preprocessor and others the compiler itself. Yet other
82 options control the assembler and linker; most of these are not
83 documented here, since you rarely need to use any of them.
85 @cindex C compilation options
86 Most of the command line options that you can use with GCC are useful
87 for C programs; when an option is only useful with another language
88 (usually C++), the explanation says so explicitly. If the description
89 for a particular option does not mention a source language, you can use
90 that option with all supported languages.
92 @cindex C++ compilation options
93 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
94 options for compiling C++ programs.
96 @cindex grouping options
97 @cindex options, grouping
98 The @command{gcc} program accepts options and file names as operands. Many
99 options have multi-letter names; therefore multiple single-letter options
100 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
103 @cindex order of options
104 @cindex options, order
105 You can mix options and other arguments. For the most part, the order
106 you use doesn't matter. Order does matter when you use several
107 options of the same kind; for example, if you specify @option{-L} more
108 than once, the directories are searched in the order specified. Also,
109 the placement of the @option{-l} option is significant.
111 Many options have long names starting with @samp{-f} or with
112 @samp{-W}---for example,
113 @option{-fmove-loop-invariants}, @option{-Wformat} and so on. Most of
114 these have both positive and negative forms; the negative form of
115 @option{-ffoo} would be @option{-fno-foo}. This manual documents
116 only one of these two forms, whichever one is not the default.
120 @xref{Option Index}, for an index to GCC's options.
123 * Option Summary:: Brief list of all options, without explanations.
124 * Overall Options:: Controlling the kind of output:
125 an executable, object files, assembler files,
126 or preprocessed source.
127 * Invoking G++:: Compiling C++ programs.
128 * C Dialect Options:: Controlling the variant of C language compiled.
129 * C++ Dialect Options:: Variations on C++.
130 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
132 * Language Independent Options:: Controlling how diagnostics should be
134 * Warning Options:: How picky should the compiler be?
135 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
136 * Optimize Options:: How much optimization?
137 * Preprocessor Options:: Controlling header files and macro definitions.
138 Also, getting dependency information for Make.
139 * Assembler Options:: Passing options to the assembler.
140 * Link Options:: Specifying libraries and so on.
141 * Directory Options:: Where to find header files and libraries.
142 Where to find the compiler executable files.
143 * Spec Files:: How to pass switches to sub-processes.
144 * Target Options:: Running a cross-compiler, or an old version of GCC.
145 * Submodel Options:: Specifying minor hardware or convention variations,
146 such as 68010 vs 68020.
147 * Code Gen Options:: Specifying conventions for function calls, data layout
149 * Environment Variables:: Env vars that affect GCC.
150 * Precompiled Headers:: Compiling a header once, and using it many times.
151 * Running Protoize:: Automatically adding or removing function prototypes.
157 @section Option Summary
159 Here is a summary of all the options, grouped by type. Explanations are
160 in the following sections.
163 @item Overall Options
164 @xref{Overall Options,,Options Controlling the Kind of Output}.
165 @gccoptlist{-c -S -E -o @var{file} -combine -pipe -pass-exit-codes @gol
166 -x @var{language} -v -### --help@r{[}=@var{class}@r{]} --target-help @gol
167 --version @@@var{file}}
169 @item C Language Options
170 @xref{C Dialect Options,,Options Controlling C Dialect}.
171 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
172 -aux-info @var{filename} @gol
173 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
174 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
175 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
176 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
177 -fsigned-bitfields -fsigned-char @gol
178 -funsigned-bitfields -funsigned-char}
180 @item C++ Language Options
181 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
182 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
183 -fconserve-space -ffriend-injection @gol
184 -fno-elide-constructors @gol
185 -fno-enforce-eh-specs @gol
186 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
187 -fno-implicit-templates @gol
188 -fno-implicit-inline-templates @gol
189 -fno-implement-inlines -fms-extensions @gol
190 -fno-nonansi-builtins -fno-operator-names @gol
191 -fno-optional-diags -fpermissive @gol
192 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
193 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
194 -fno-default-inline -fvisibility-inlines-hidden @gol
195 -fvisibility-ms-compat @gol
196 -Wabi -Wctor-dtor-privacy @gol
197 -Wnon-virtual-dtor -Wreorder @gol
198 -Weffc++ -Wno-deprecated -Wstrict-null-sentinel @gol
199 -Wno-non-template-friend -Wold-style-cast @gol
200 -Woverloaded-virtual -Wno-pmf-conversions @gol
203 @item Objective-C and Objective-C++ Language Options
204 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
205 Objective-C and Objective-C++ Dialects}.
206 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
207 -fgnu-runtime -fnext-runtime @gol
208 -fno-nil-receivers @gol
209 -fobjc-call-cxx-cdtors @gol
210 -fobjc-direct-dispatch @gol
211 -fobjc-exceptions @gol
213 -freplace-objc-classes @gol
216 -Wassign-intercept @gol
217 -Wno-protocol -Wselector @gol
218 -Wstrict-selector-match @gol
219 -Wundeclared-selector}
221 @item Language Independent Options
222 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
223 @gccoptlist{-fmessage-length=@var{n} @gol
224 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
225 -fdiagnostics-show-option}
227 @item Warning Options
228 @xref{Warning Options,,Options to Request or Suppress Warnings}.
229 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
230 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
231 -Wno-attributes -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
232 -Wchar-subscripts -Wclobbered -Wcomment @gol
233 -Wconversion -Wcoverage-mismatch -Wno-deprecated-declarations @gol
234 -Wdisabled-optimization -Wno-div-by-zero @gol
235 -Wempty-body -Wno-endif-labels @gol
236 -Werror -Werror=* @gol
237 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
238 -Wno-format-extra-args -Wformat-nonliteral @gol
239 -Wformat-security -Wformat-y2k @gol
240 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
241 -Wimport -Wno-import -Winit-self -Winline @gol
242 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
243 -Winvalid-pch -Wlarger-than-@var{len} -Wunsafe-loop-optimizations @gol
244 -Wlogical-op -Wlong-long @gol
245 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
246 -Wmissing-format-attribute -Wmissing-include-dirs @gol
247 -Wmissing-noreturn @gol
248 -Wno-multichar -Wnonnull -Wno-overflow @gol
249 -Woverlength-strings -Wpacked -Wpadded @gol
250 -Wparentheses -Wpointer-arith -Wno-pointer-to-int-cast @gol
251 -Wredundant-decls @gol
252 -Wreturn-type -Wsequence-point -Wshadow @gol
253 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
254 -Wstrict-aliasing -Wstrict-aliasing=n @gol
255 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
256 -Wswitch -Wswitch-default -Wswitch-enum @gol
257 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
258 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
259 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
260 -Wunused-value -Wunused-variable @gol
261 -Wvariadic-macros -Wvla @gol
262 -Wvolatile-register-var -Wwrite-strings}
264 @item C-only Warning Options
265 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
266 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
267 -Wold-style-declaration -Wold-style-definition @gol
268 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
269 -Wdeclaration-after-statement -Wpointer-sign}
271 @item Debugging Options
272 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
273 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
274 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
275 -fdump-noaddr -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
276 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
277 -fdump-ipa-all -fdump-ipa-cgraph @gol
279 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
280 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
281 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
282 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
284 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
285 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
287 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
288 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
289 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
290 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
291 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-nrv -fdump-tree-vect @gol
293 -fdump-tree-sink @gol
294 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
295 -fdump-tree-salias @gol
296 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
297 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
298 -ftree-vectorizer-verbose=@var{n} @gol
299 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
300 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
301 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
302 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
303 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
304 -ftest-coverage -ftime-report -fvar-tracking @gol
305 -g -g@var{level} -gcoff -gdwarf-2 @gol
306 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
307 -fdebug-prefix-map=@var{old}=@var{new} @gol
308 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
309 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
310 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
311 -print-multi-directory -print-multi-lib @gol
312 -print-prog-name=@var{program} -print-search-dirs -Q @gol
313 -print-sysroot-headers-suffix @gol
316 @item Optimization Options
317 @xref{Optimize Options,,Options that Control Optimization}.
318 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
319 -falign-labels=@var{n} -falign-loops=@var{n} @gol
320 -fbounds-check -fmudflap -fmudflapth -fmudflapir @gol
321 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
322 -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol
323 -fcaller-saves -fcprop-registers -fcse-follow-jumps @gol
324 -fcse-skip-blocks -fcx-limited-range -fdata-sections @gol
325 -fdelayed-branch -fdelete-null-pointer-checks -fearly-inlining @gol
326 -fexpensive-optimizations -ffast-math -ffloat-store @gol
327 -fforce-addr -fforward-propagate -ffunction-sections @gol
328 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -fgcse-after-reload @gol
329 -fcrossjumping -fif-conversion -fif-conversion2 @gol
330 -finline-functions -finline-functions-called-once @gol
331 -finline-limit=@var{n} -fkeep-inline-functions @gol
332 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
333 -fmodulo-sched -fmodulo-sched-allow-regmoves -fno-branch-count-reg @gol
334 -fno-default-inline -fno-defer-pop -fmove-loop-invariants @gol
335 -fno-function-cse -fno-guess-branch-probability @gol
336 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
337 -funsafe-math-optimizations -funsafe-loop-optimizations @gol
338 -ffinite-math-only -fno-signed-zeros @gol
339 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
340 -fomit-frame-pointer -foptimize-register-move @gol
341 -foptimize-sibling-calls -fpredictive-commoning -fprefetch-loop-arrays @gol
342 -fprofile-generate -fprofile-use @gol
343 -fregmove -frename-registers @gol
344 -freorder-blocks -freorder-blocks-and-partition -freorder-functions @gol
345 -frerun-cse-after-loop @gol
346 -frounding-math -frtl-abstract-sequences @gol
347 -fschedule-insns -fschedule-insns2 @gol
348 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
349 -fsched-spec-load-dangerous @gol
350 -fsched-stalled-insns=@var{n} -fsched-stalled-insns-dep=@var{n} @gol
351 -fsched2-use-superblocks @gol
352 -fsched2-use-traces -fsee -freschedule-modulo-scheduled-loops @gol
353 -fsection-anchors -fsignaling-nans -fsingle-precision-constant @gol
354 -fno-split-wide-types -fstack-protector -fstack-protector-all @gol
355 -fstrict-aliasing -fstrict-overflow -ftracer -fthread-jumps @gol
356 -funroll-all-loops -funroll-loops -fpeel-loops @gol
357 -fsplit-ivs-in-unroller -funswitch-loops @gol
358 -fvariable-expansion-in-unroller -ftree-reassoc @gol
359 -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol
360 -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts @gol
361 -fcheck-data-deps @gol
362 -ftree-dominator-opts -ftree-dse -ftree-copyrename -ftree-sink @gol
363 -ftree-ch -ftree-sra -ftree-ter -ftree-fre -ftree-vectorize @gol
364 -ftree-vect-loop-version -fvect-cost-model -ftree-salias -fipa-pta -fweb @gol
365 -ftree-copy-prop -ftree-store-ccp -ftree-store-copy-prop -fwhole-program @gol
366 --param @var{name}=@var{value}
367 -O -O0 -O1 -O2 -O3 -Os}
369 @item Preprocessor Options
370 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
371 @gccoptlist{-A@var{question}=@var{answer} @gol
372 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
373 -C -dD -dI -dM -dN @gol
374 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
375 -idirafter @var{dir} @gol
376 -include @var{file} -imacros @var{file} @gol
377 -iprefix @var{file} -iwithprefix @var{dir} @gol
378 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
379 -imultilib @var{dir} -isysroot @var{dir} @gol
380 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
381 -P -fworking-directory -remap @gol
382 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
383 -Xpreprocessor @var{option}}
385 @item Assembler Option
386 @xref{Assembler Options,,Passing Options to the Assembler}.
387 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
390 @xref{Link Options,,Options for Linking}.
391 @gccoptlist{@var{object-file-name} -l@var{library} @gol
392 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
393 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
394 -Wl,@var{option} -Xlinker @var{option} @gol
397 @item Directory Options
398 @xref{Directory Options,,Options for Directory Search}.
399 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
400 -specs=@var{file} -I- --sysroot=@var{dir}}
403 @c I wrote this xref this way to avoid overfull hbox. -- rms
404 @xref{Target Options}.
405 @gccoptlist{-V @var{version} -b @var{machine}}
407 @item Machine Dependent Options
408 @xref{Submodel Options,,Hardware Models and Configurations}.
409 @c This list is ordered alphanumerically by subsection name.
410 @c Try and put the significant identifier (CPU or system) first,
411 @c so users have a clue at guessing where the ones they want will be.
414 @gccoptlist{-EB -EL @gol
415 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
416 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
419 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
420 -mabi=@var{name} @gol
421 -mapcs-stack-check -mno-apcs-stack-check @gol
422 -mapcs-float -mno-apcs-float @gol
423 -mapcs-reentrant -mno-apcs-reentrant @gol
424 -msched-prolog -mno-sched-prolog @gol
425 -mlittle-endian -mbig-endian -mwords-little-endian @gol
426 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
427 -mthumb-interwork -mno-thumb-interwork @gol
428 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
429 -mstructure-size-boundary=@var{n} @gol
430 -mabort-on-noreturn @gol
431 -mlong-calls -mno-long-calls @gol
432 -msingle-pic-base -mno-single-pic-base @gol
433 -mpic-register=@var{reg} @gol
434 -mnop-fun-dllimport @gol
435 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
436 -mpoke-function-name @gol
438 -mtpcs-frame -mtpcs-leaf-frame @gol
439 -mcaller-super-interworking -mcallee-super-interworking @gol
443 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
444 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
446 @emph{Blackfin Options}
447 @gccoptlist{-mcpu=@var{cpu} -msim -momit-leaf-frame-pointer @gol
448 -mno-omit-leaf-frame-pointer -mspecld-anomaly -mno-specld-anomaly @gol
449 -mcsync-anomaly -mno-csync-anomaly -mlow-64k -mno-low64k @gol
450 -mstack-check-l1 -mid-shared-library -mno-id-shared-library @gol
451 -mshared-library-id=@var{n} -mleaf-id-shared-library @gol
452 -mno-leaf-id-shared-library -msep-data -mno-sep-data -mlong-calls @gol
453 -mno-long-calls -mfast-fp -minline-plt}
456 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
457 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
458 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
459 -mstack-align -mdata-align -mconst-align @gol
460 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
461 -melf -maout -melinux -mlinux -sim -sim2 @gol
462 -mmul-bug-workaround -mno-mul-bug-workaround}
465 @gccoptlist{-mmac -mpush-args}
467 @emph{Darwin Options}
468 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
469 -arch_only -bind_at_load -bundle -bundle_loader @gol
470 -client_name -compatibility_version -current_version @gol
472 -dependency-file -dylib_file -dylinker_install_name @gol
473 -dynamic -dynamiclib -exported_symbols_list @gol
474 -filelist -flat_namespace -force_cpusubtype_ALL @gol
475 -force_flat_namespace -headerpad_max_install_names @gol
477 -image_base -init -install_name -keep_private_externs @gol
478 -multi_module -multiply_defined -multiply_defined_unused @gol
479 -noall_load -no_dead_strip_inits_and_terms @gol
480 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
481 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
482 -private_bundle -read_only_relocs -sectalign @gol
483 -sectobjectsymbols -whyload -seg1addr @gol
484 -sectcreate -sectobjectsymbols -sectorder @gol
485 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
486 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
487 -segprot -segs_read_only_addr -segs_read_write_addr @gol
488 -single_module -static -sub_library -sub_umbrella @gol
489 -twolevel_namespace -umbrella -undefined @gol
490 -unexported_symbols_list -weak_reference_mismatches @gol
491 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
492 -mkernel -mone-byte-bool}
494 @emph{DEC Alpha Options}
495 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
496 -mieee -mieee-with-inexact -mieee-conformant @gol
497 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
498 -mtrap-precision=@var{mode} -mbuild-constants @gol
499 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
500 -mbwx -mmax -mfix -mcix @gol
501 -mfloat-vax -mfloat-ieee @gol
502 -mexplicit-relocs -msmall-data -mlarge-data @gol
503 -msmall-text -mlarge-text @gol
504 -mmemory-latency=@var{time}}
506 @emph{DEC Alpha/VMS Options}
507 @gccoptlist{-mvms-return-codes}
510 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
511 -mhard-float -msoft-float @gol
512 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
513 -mdouble -mno-double @gol
514 -mmedia -mno-media -mmuladd -mno-muladd @gol
515 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
516 -mlinked-fp -mlong-calls -malign-labels @gol
517 -mlibrary-pic -macc-4 -macc-8 @gol
518 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
519 -moptimize-membar -mno-optimize-membar @gol
520 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
521 -mvliw-branch -mno-vliw-branch @gol
522 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
523 -mno-nested-cond-exec -mtomcat-stats @gol
527 @emph{GNU/Linux Options}
528 @gccoptlist{-muclibc}
530 @emph{H8/300 Options}
531 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
534 @gccoptlist{-march=@var{architecture-type} @gol
535 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
536 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
537 -mfixed-range=@var{register-range} @gol
538 -mjump-in-delay -mlinker-opt -mlong-calls @gol
539 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
540 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
541 -mno-jump-in-delay -mno-long-load-store @gol
542 -mno-portable-runtime -mno-soft-float @gol
543 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
544 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
545 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
546 -munix=@var{unix-std} -nolibdld -static -threads}
548 @emph{i386 and x86-64 Options}
549 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
550 -mfpmath=@var{unit} @gol
551 -masm=@var{dialect} -mno-fancy-math-387 @gol
552 -mno-fp-ret-in-387 -msoft-float @gol
553 -mno-wide-multiply -mrtd -malign-double @gol
554 -mpreferred-stack-boundary=@var{num} -mcx16 -msahf -mrecip @gol
555 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 @gol
556 -msse4a -m3dnow -mpopcnt -mabm @gol
557 -mthreads -mno-align-stringops -minline-all-stringops @gol
558 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
559 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
560 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
561 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
562 -mcmodel=@var{code-model} @gol
563 -m32 -m64 -mlarge-data-threshold=@var{num}}
566 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
567 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
568 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
569 -minline-float-divide-max-throughput @gol
570 -minline-int-divide-min-latency @gol
571 -minline-int-divide-max-throughput @gol
572 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
573 -mno-dwarf2-asm -mearly-stop-bits @gol
574 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
575 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64 @gol
576 -mno-sched-br-data-spec -msched-ar-data-spec -mno-sched-control-spec @gol
577 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
578 -msched-ldc -mno-sched-control-ldc -mno-sched-spec-verbose @gol
579 -mno-sched-prefer-non-data-spec-insns @gol
580 -mno-sched-prefer-non-control-spec-insns @gol
581 -mno-sched-count-spec-in-critical-path}
583 @emph{M32R/D Options}
584 @gccoptlist{-m32r2 -m32rx -m32r @gol
586 -malign-loops -mno-align-loops @gol
587 -missue-rate=@var{number} @gol
588 -mbranch-cost=@var{number} @gol
589 -mmodel=@var{code-size-model-type} @gol
590 -msdata=@var{sdata-type} @gol
591 -mno-flush-func -mflush-func=@var{name} @gol
592 -mno-flush-trap -mflush-trap=@var{number} @gol
596 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
598 @emph{M680x0 Options}
599 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
600 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
601 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
602 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
603 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
604 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
605 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
606 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
608 @emph{M68hc1x Options}
609 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
610 -mauto-incdec -minmax -mlong-calls -mshort @gol
611 -msoft-reg-count=@var{count}}
614 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
615 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
616 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
617 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
618 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
621 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
622 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
623 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
624 -mshared -mno-shared -mxgot -mno-xgot -mgp32 -mgp64 @gol
625 -mfp32 -mfp64 -mhard-float -msoft-float @gol
626 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
627 -msmartmips -mno-smartmips @gol
628 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
629 -mips3d -mno-mips3d -mmt -mno-mt @gol
630 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
631 -G@var{num} -membedded-data -mno-embedded-data @gol
632 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
633 -mcode-readable=@var{setting} @gol
634 -msplit-addresses -mno-split-addresses @gol
635 -mexplicit-relocs -mno-explicit-relocs @gol
636 -mcheck-zero-division -mno-check-zero-division @gol
637 -mdivide-traps -mdivide-breaks @gol
638 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
639 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
640 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
641 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 -mno-fix-vr4130 @gol
642 -mfix-sb1 -mno-fix-sb1 @gol
643 -mflush-func=@var{func} -mno-flush-func @gol
644 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
645 -mfp-exceptions -mno-fp-exceptions @gol
646 -mvr4130-align -mno-vr4130-align}
649 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
650 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
651 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
652 -mno-base-addresses -msingle-exit -mno-single-exit}
654 @emph{MN10300 Options}
655 @gccoptlist{-mmult-bug -mno-mult-bug @gol
656 -mam33 -mno-am33 @gol
657 -mam33-2 -mno-am33-2 @gol
658 -mreturn-pointer-on-d0 @gol
662 @gccoptlist{-mno-crt0 -mbacc -msim @gol
663 -march=@var{cpu-type} }
665 @emph{PDP-11 Options}
666 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
667 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
668 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
669 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
670 -mbranch-expensive -mbranch-cheap @gol
671 -msplit -mno-split -munix-asm -mdec-asm}
673 @emph{PowerPC Options}
674 See RS/6000 and PowerPC Options.
676 @emph{RS/6000 and PowerPC Options}
677 @gccoptlist{-mcpu=@var{cpu-type} @gol
678 -mtune=@var{cpu-type} @gol
679 -mpower -mno-power -mpower2 -mno-power2 @gol
680 -mpowerpc -mpowerpc64 -mno-powerpc @gol
681 -maltivec -mno-altivec @gol
682 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
683 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
684 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
685 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mdfp -mno-dfp @gol
686 -mnew-mnemonics -mold-mnemonics @gol
687 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
688 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
689 -malign-power -malign-natural @gol
690 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
691 -mstring -mno-string -mupdate -mno-update @gol
692 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
693 -mstrict-align -mno-strict-align -mrelocatable @gol
694 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
695 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
696 -mdynamic-no-pic -maltivec -mswdiv @gol
697 -mprioritize-restricted-insns=@var{priority} @gol
698 -msched-costly-dep=@var{dependence_type} @gol
699 -minsert-sched-nops=@var{scheme} @gol
700 -mcall-sysv -mcall-netbsd @gol
701 -maix-struct-return -msvr4-struct-return @gol
702 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
703 -misel -mno-isel @gol
704 -misel=yes -misel=no @gol
706 -mspe=yes -mspe=no @gol
708 -mvrsave -mno-vrsave @gol
709 -mmulhw -mno-mulhw @gol
710 -mdlmzb -mno-dlmzb @gol
711 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
712 -mprototype -mno-prototype @gol
713 -msim -mmvme -mads -myellowknife -memb -msdata @gol
714 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
716 @emph{S/390 and zSeries Options}
717 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
718 -mhard-float -msoft-float -mlong-double-64 -mlong-double-128 @gol
719 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
720 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
721 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
722 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
723 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
726 @gccoptlist{-meb -mel @gol
730 -mscore5 -mscore5u -mscore7 -mscore7d}
733 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
734 -m4-nofpu -m4-single-only -m4-single -m4 @gol
735 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
736 -m5-64media -m5-64media-nofpu @gol
737 -m5-32media -m5-32media-nofpu @gol
738 -m5-compact -m5-compact-nofpu @gol
739 -mb -ml -mdalign -mrelax @gol
740 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
741 -mieee -misize -minline-ic_invalidate -mpadstruct -mspace @gol
742 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
743 -mdivsi3_libfunc=@var{name} @gol
744 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
748 @gccoptlist{-mcpu=@var{cpu-type} @gol
749 -mtune=@var{cpu-type} @gol
750 -mcmodel=@var{code-model} @gol
751 -m32 -m64 -mapp-regs -mno-app-regs @gol
752 -mfaster-structs -mno-faster-structs @gol
753 -mfpu -mno-fpu -mhard-float -msoft-float @gol
754 -mhard-quad-float -msoft-quad-float @gol
755 -mimpure-text -mno-impure-text -mlittle-endian @gol
756 -mstack-bias -mno-stack-bias @gol
757 -munaligned-doubles -mno-unaligned-doubles @gol
758 -mv8plus -mno-v8plus -mvis -mno-vis
759 -threads -pthreads -pthread}
762 @gccoptlist{-mwarn-reloc -merror-reloc @gol
763 -msafe-dma -munsafe-dma @gol
765 -msmall-mem -mlarge-mem -mstdmain @gol
766 -mfixed-range=@var{register-range}}
768 @emph{System V Options}
769 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
771 @emph{TMS320C3x/C4x Options}
772 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
773 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
774 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
775 -mparallel-insns -mparallel-mpy -mpreserve-float}
778 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
779 -mprolog-function -mno-prolog-function -mspace @gol
780 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
781 -mapp-regs -mno-app-regs @gol
782 -mdisable-callt -mno-disable-callt @gol
788 @gccoptlist{-mg -mgnu -munix}
790 @emph{VxWorks Options}
791 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
792 -Xbind-lazy -Xbind-now}
794 @emph{x86-64 Options}
795 See i386 and x86-64 Options.
797 @emph{Xstormy16 Options}
800 @emph{Xtensa Options}
801 @gccoptlist{-mconst16 -mno-const16 @gol
802 -mfused-madd -mno-fused-madd @gol
803 -mtext-section-literals -mno-text-section-literals @gol
804 -mtarget-align -mno-target-align @gol
805 -mlongcalls -mno-longcalls}
807 @emph{zSeries Options}
808 See S/390 and zSeries Options.
810 @item Code Generation Options
811 @xref{Code Gen Options,,Options for Code Generation Conventions}.
812 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
813 -ffixed-@var{reg} -fexceptions @gol
814 -fnon-call-exceptions -funwind-tables @gol
815 -fasynchronous-unwind-tables @gol
816 -finhibit-size-directive -finstrument-functions @gol
817 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
818 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
819 -fno-common -fno-ident @gol
820 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
821 -fno-jump-tables @gol
822 -frecord-gcc-switches @gol
823 -freg-struct-return -fshort-enums @gol
824 -fshort-double -fshort-wchar @gol
825 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
826 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
827 -fargument-alias -fargument-noalias @gol
828 -fargument-noalias-global -fargument-noalias-anything
829 -fleading-underscore -ftls-model=@var{model} @gol
830 -ftrapv -fwrapv -fbounds-check @gol
835 * Overall Options:: Controlling the kind of output:
836 an executable, object files, assembler files,
837 or preprocessed source.
838 * C Dialect Options:: Controlling the variant of C language compiled.
839 * C++ Dialect Options:: Variations on C++.
840 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
842 * Language Independent Options:: Controlling how diagnostics should be
844 * Warning Options:: How picky should the compiler be?
845 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
846 * Optimize Options:: How much optimization?
847 * Preprocessor Options:: Controlling header files and macro definitions.
848 Also, getting dependency information for Make.
849 * Assembler Options:: Passing options to the assembler.
850 * Link Options:: Specifying libraries and so on.
851 * Directory Options:: Where to find header files and libraries.
852 Where to find the compiler executable files.
853 * Spec Files:: How to pass switches to sub-processes.
854 * Target Options:: Running a cross-compiler, or an old version of GCC.
857 @node Overall Options
858 @section Options Controlling the Kind of Output
860 Compilation can involve up to four stages: preprocessing, compilation
861 proper, assembly and linking, always in that order. GCC is capable of
862 preprocessing and compiling several files either into several
863 assembler input files, or into one assembler input file; then each
864 assembler input file produces an object file, and linking combines all
865 the object files (those newly compiled, and those specified as input)
866 into an executable file.
868 @cindex file name suffix
869 For any given input file, the file name suffix determines what kind of
874 C source code which must be preprocessed.
877 C source code which should not be preprocessed.
880 C++ source code which should not be preprocessed.
883 Objective-C source code. Note that you must link with the @file{libobjc}
884 library to make an Objective-C program work.
887 Objective-C source code which should not be preprocessed.
891 Objective-C++ source code. Note that you must link with the @file{libobjc}
892 library to make an Objective-C++ program work. Note that @samp{.M} refers
893 to a literal capital M@.
896 Objective-C++ source code which should not be preprocessed.
899 C, C++, Objective-C or Objective-C++ header file to be turned into a
904 @itemx @var{file}.cxx
905 @itemx @var{file}.cpp
906 @itemx @var{file}.CPP
907 @itemx @var{file}.c++
909 C++ source code which must be preprocessed. Note that in @samp{.cxx},
910 the last two letters must both be literally @samp{x}. Likewise,
911 @samp{.C} refers to a literal capital C@.
915 Objective-C++ source code which must be preprocessed.
918 Objective-C++ source code which should not be preprocessed.
923 @itemx @var{file}.hxx
924 @itemx @var{file}.hpp
925 @itemx @var{file}.HPP
926 @itemx @var{file}.h++
927 @itemx @var{file}.tcc
928 C++ header file to be turned into a precompiled header.
931 @itemx @var{file}.for
932 @itemx @var{file}.FOR
933 Fixed form Fortran source code which should not be preprocessed.
936 @itemx @var{file}.fpp
937 @itemx @var{file}.FPP
938 Fixed form Fortran source code which must be preprocessed (with the traditional
942 @itemx @var{file}.f95
943 Free form Fortran source code which should not be preprocessed.
946 @itemx @var{file}.F95
947 Free form Fortran source code which must be preprocessed (with the
948 traditional preprocessor).
950 @c FIXME: Descriptions of Java file types.
957 Ada source code file which contains a library unit declaration (a
958 declaration of a package, subprogram, or generic, or a generic
959 instantiation), or a library unit renaming declaration (a package,
960 generic, or subprogram renaming declaration). Such files are also
963 @itemx @var{file}.adb
964 Ada source code file containing a library unit body (a subprogram or
965 package body). Such files are also called @dfn{bodies}.
967 @c GCC also knows about some suffixes for languages not yet included:
978 Assembler code which must be preprocessed.
981 An object file to be fed straight into linking.
982 Any file name with no recognized suffix is treated this way.
986 You can specify the input language explicitly with the @option{-x} option:
989 @item -x @var{language}
990 Specify explicitly the @var{language} for the following input files
991 (rather than letting the compiler choose a default based on the file
992 name suffix). This option applies to all following input files until
993 the next @option{-x} option. Possible values for @var{language} are:
995 c c-header c-cpp-output
996 c++ c++-header c++-cpp-output
997 objective-c objective-c-header objective-c-cpp-output
998 objective-c++ objective-c++-header objective-c++-cpp-output
999 assembler assembler-with-cpp
1007 Turn off any specification of a language, so that subsequent files are
1008 handled according to their file name suffixes (as they are if @option{-x}
1009 has not been used at all).
1011 @item -pass-exit-codes
1012 @opindex pass-exit-codes
1013 Normally the @command{gcc} program will exit with the code of 1 if any
1014 phase of the compiler returns a non-success return code. If you specify
1015 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1016 numerically highest error produced by any phase that returned an error
1017 indication. The C, C++, and Fortran frontends return 4, if an internal
1018 compiler error is encountered.
1021 If you only want some of the stages of compilation, you can use
1022 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1023 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1024 @command{gcc} is to stop. Note that some combinations (for example,
1025 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1030 Compile or assemble the source files, but do not link. The linking
1031 stage simply is not done. The ultimate output is in the form of an
1032 object file for each source file.
1034 By default, the object file name for a source file is made by replacing
1035 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1037 Unrecognized input files, not requiring compilation or assembly, are
1042 Stop after the stage of compilation proper; do not assemble. The output
1043 is in the form of an assembler code file for each non-assembler input
1046 By default, the assembler file name for a source file is made by
1047 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1049 Input files that don't require compilation are ignored.
1053 Stop after the preprocessing stage; do not run the compiler proper. The
1054 output is in the form of preprocessed source code, which is sent to the
1057 Input files which don't require preprocessing are ignored.
1059 @cindex output file option
1062 Place output in file @var{file}. This applies regardless to whatever
1063 sort of output is being produced, whether it be an executable file,
1064 an object file, an assembler file or preprocessed C code.
1066 If @option{-o} is not specified, the default is to put an executable
1067 file in @file{a.out}, the object file for
1068 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1069 assembler file in @file{@var{source}.s}, a precompiled header file in
1070 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1075 Print (on standard error output) the commands executed to run the stages
1076 of compilation. Also print the version number of the compiler driver
1077 program and of the preprocessor and the compiler proper.
1081 Like @option{-v} except the commands are not executed and all command
1082 arguments are quoted. This is useful for shell scripts to capture the
1083 driver-generated command lines.
1087 Use pipes rather than temporary files for communication between the
1088 various stages of compilation. This fails to work on some systems where
1089 the assembler is unable to read from a pipe; but the GNU assembler has
1094 If you are compiling multiple source files, this option tells the driver
1095 to pass all the source files to the compiler at once (for those
1096 languages for which the compiler can handle this). This will allow
1097 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1098 language for which this is supported is C@. If you pass source files for
1099 multiple languages to the driver, using this option, the driver will invoke
1100 the compiler(s) that support IMA once each, passing each compiler all the
1101 source files appropriate for it. For those languages that do not support
1102 IMA this option will be ignored, and the compiler will be invoked once for
1103 each source file in that language. If you use this option in conjunction
1104 with @option{-save-temps}, the compiler will generate multiple
1106 (one for each source file), but only one (combined) @file{.o} or
1111 Print (on the standard output) a description of the command line options
1112 understood by @command{gcc}. If the @option{-v} option is also specified
1113 then @option{--help} will also be passed on to the various processes
1114 invoked by @command{gcc}, so that they can display the command line options
1115 they accept. If the @option{-Wextra} option has also been specified
1116 (prior to the @option{--help} option), then command line options which
1117 have no documentation associated with them will also be displayed.
1120 @opindex target-help
1121 Print (on the standard output) a description of target-specific command
1122 line options for each tool. For some targets extra target-specific
1123 information may also be printed.
1125 @item --help=@var{class}@r{[},@var{qualifier}@r{]}
1126 Print (on the standard output) a description of the command line
1127 options understood by the compiler that fit into a specific class.
1128 The class can be one of @samp{optimizers}, @samp{warnings}, @samp{target},
1129 @samp{params}, or @var{language}:
1132 @item @samp{optimizers}
1133 This will display all of the optimization options supported by the
1136 @item @samp{warnings}
1137 This will display all of the options controlling warning messages
1138 produced by the compiler.
1141 This will display target-specific options. Unlike the
1142 @option{--target-help} option however, target-specific options of the
1143 linker and assembler will not be displayed. This is because those
1144 tools do not currently support the extended @option{--help=} syntax.
1147 This will display the values recognized by the @option{--param}
1150 @item @var{language}
1151 This will display the options supported for @var{language}, where
1152 @var{language} is the name of one of the languages supported in this
1156 This will display the options that are common to all languages.
1159 It is possible to further refine the output of the @option{--help=}
1160 option by adding a comma separated list of qualifiers after the
1161 class. These can be any from the following list:
1164 @item @samp{undocumented}
1165 Display only those options which are undocumented.
1168 Display options which take an argument that appears after an equal
1169 sign in the same continuous piece of text, such as:
1170 @samp{--help=target}.
1172 @item @samp{separate}
1173 Display options which take an argument that appears as a separate word
1174 following the original option, such as: @samp{-o output-file}.
1177 Thus for example to display all the undocumented target-specific
1178 switches supported by the compiler the following can be used:
1181 --help=target,undocumented
1184 The sense of a qualifier can be inverted by prefixing it with the
1185 @var{^} character, so for example to display all binary warning
1186 options (i.e. ones that are either on or off and that do not take an
1187 argument), which have a description the following can be used:
1190 --help=warnings,^joined,^undocumented
1193 A class can also be used as a qualifier, although this usually
1194 restricts the output by so much that there is nothing to display. One
1195 case where it does work however is when one of the classes is
1196 @var{target}. So for example to display all the target-specific
1197 optimization options the following can be used:
1200 --help=target,optimizers
1203 The @option{--help=} option can be repeated on the command line. Each
1204 successive use will display its requested class of options, skipping
1205 those that have already been displayed.
1207 If the @option{-Q} option appears on the command line before the
1208 @option{--help=} option, then the descriptive text displayed by
1209 @option{--help=} is changed. Instead of describing the displayed
1210 options, an indication is given as to whether the option is enabled,
1211 disabled or set to a specific value (assuming that the compiler
1212 knows this at the point where the @option{--help=} option is used).
1214 Here is a truncated example from the ARM port of @command{gcc}:
1217 % gcc -Q -mabi=2 --help=target -c
1218 The following options are target specific:
1220 -mabort-on-noreturn [disabled]
1224 The output is sensitive to the effects of previous command line
1225 options, so for example it is possible to find out which optimizations
1226 are enabled at @option{-O2} by using:
1229 -O2 --help=optimizers
1232 Alternatively you can discover which binary optimizations are enabled
1233 by @option{-O3} by using:
1236 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1237 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1238 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1243 Display the version number and copyrights of the invoked GCC@.
1245 @include @value{srcdir}/../libiberty/at-file.texi
1249 @section Compiling C++ Programs
1251 @cindex suffixes for C++ source
1252 @cindex C++ source file suffixes
1253 C++ source files conventionally use one of the suffixes @samp{.C},
1254 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1255 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1256 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1257 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1258 files with these names and compiles them as C++ programs even if you
1259 call the compiler the same way as for compiling C programs (usually
1260 with the name @command{gcc}).
1264 However, the use of @command{gcc} does not add the C++ library.
1265 @command{g++} is a program that calls GCC and treats @samp{.c},
1266 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1267 files unless @option{-x} is used, and automatically specifies linking
1268 against the C++ library. This program is also useful when
1269 precompiling a C header file with a @samp{.h} extension for use in C++
1270 compilations. On many systems, @command{g++} is also installed with
1271 the name @command{c++}.
1273 @cindex invoking @command{g++}
1274 When you compile C++ programs, you may specify many of the same
1275 command-line options that you use for compiling programs in any
1276 language; or command-line options meaningful for C and related
1277 languages; or options that are meaningful only for C++ programs.
1278 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1279 explanations of options for languages related to C@.
1280 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1281 explanations of options that are meaningful only for C++ programs.
1283 @node C Dialect Options
1284 @section Options Controlling C Dialect
1285 @cindex dialect options
1286 @cindex language dialect options
1287 @cindex options, dialect
1289 The following options control the dialect of C (or languages derived
1290 from C, such as C++, Objective-C and Objective-C++) that the compiler
1294 @cindex ANSI support
1298 In C mode, support all ISO C90 programs. In C++ mode,
1299 remove GNU extensions that conflict with ISO C++.
1301 This turns off certain features of GCC that are incompatible with ISO
1302 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1303 such as the @code{asm} and @code{typeof} keywords, and
1304 predefined macros such as @code{unix} and @code{vax} that identify the
1305 type of system you are using. It also enables the undesirable and
1306 rarely used ISO trigraph feature. For the C compiler,
1307 it disables recognition of C++ style @samp{//} comments as well as
1308 the @code{inline} keyword.
1310 The alternate keywords @code{__asm__}, @code{__extension__},
1311 @code{__inline__} and @code{__typeof__} continue to work despite
1312 @option{-ansi}. You would not want to use them in an ISO C program, of
1313 course, but it is useful to put them in header files that might be included
1314 in compilations done with @option{-ansi}. Alternate predefined macros
1315 such as @code{__unix__} and @code{__vax__} are also available, with or
1316 without @option{-ansi}.
1318 The @option{-ansi} option does not cause non-ISO programs to be
1319 rejected gratuitously. For that, @option{-pedantic} is required in
1320 addition to @option{-ansi}. @xref{Warning Options}.
1322 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1323 option is used. Some header files may notice this macro and refrain
1324 from declaring certain functions or defining certain macros that the
1325 ISO standard doesn't call for; this is to avoid interfering with any
1326 programs that might use these names for other things.
1328 Functions which would normally be built in but do not have semantics
1329 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1330 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1331 built-in functions provided by GCC}, for details of the functions
1336 Determine the language standard. This option is currently only
1337 supported when compiling C or C++. A value for this option must be
1338 provided; possible values are
1343 ISO C90 (same as @option{-ansi}).
1345 @item iso9899:199409
1346 ISO C90 as modified in amendment 1.
1352 ISO C99. Note that this standard is not yet fully supported; see
1353 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1354 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1357 Default, ISO C90 plus GNU extensions (including some C99 features).
1361 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1362 this will become the default. The name @samp{gnu9x} is deprecated.
1365 The 1998 ISO C++ standard plus amendments.
1368 The same as @option{-std=c++98} plus GNU extensions. This is the
1369 default for C++ code.
1372 The working draft of the upcoming ISO C++0x standard. This option
1373 enables experimental features that are likely to be included in
1374 C++0x. The working draft is constantly changing, and any feature that is
1375 enabled by this flag may be removed from future versions of GCC if it is
1376 not part of the C++0x standard.
1379 The same as @option{-std=c++0x} plus GNU extensions. As with
1380 @option{-std=c++0x}, this option enables experimental features that may
1381 be removed in future versions of GCC.
1384 Even when this option is not specified, you can still use some of the
1385 features of newer standards in so far as they do not conflict with
1386 previous C standards. For example, you may use @code{__restrict__} even
1387 when @option{-std=c99} is not specified.
1389 The @option{-std} options specifying some version of ISO C have the same
1390 effects as @option{-ansi}, except that features that were not in ISO C90
1391 but are in the specified version (for example, @samp{//} comments and
1392 the @code{inline} keyword in ISO C99) are not disabled.
1394 @xref{Standards,,Language Standards Supported by GCC}, for details of
1395 these standard versions.
1397 @item -fgnu89-inline
1398 @opindex fgnu89-inline
1399 The option @option{-fgnu89-inline} tells GCC to use the traditional
1400 GNU semantics for @code{inline} functions when in C99 mode.
1401 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1402 is accepted and ignored by GCC versions 4.1.3 up to but not including
1403 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1404 C99 mode. Using this option is roughly equivalent to adding the
1405 @code{gnu_inline} function attribute to all inline functions
1406 (@pxref{Function Attributes}).
1408 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1409 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1410 specifies the default behavior). This option was first supported in
1411 GCC 4.3. This option is not supported in C89 or gnu89 mode.
1413 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1414 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1415 in effect for @code{inline} functions. @xref{Common Predefined
1416 Macros,,,cpp,The C Preprocessor}.
1418 @item -aux-info @var{filename}
1420 Output to the given filename prototyped declarations for all functions
1421 declared and/or defined in a translation unit, including those in header
1422 files. This option is silently ignored in any language other than C@.
1424 Besides declarations, the file indicates, in comments, the origin of
1425 each declaration (source file and line), whether the declaration was
1426 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1427 @samp{O} for old, respectively, in the first character after the line
1428 number and the colon), and whether it came from a declaration or a
1429 definition (@samp{C} or @samp{F}, respectively, in the following
1430 character). In the case of function definitions, a K&R-style list of
1431 arguments followed by their declarations is also provided, inside
1432 comments, after the declaration.
1436 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1437 keyword, so that code can use these words as identifiers. You can use
1438 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1439 instead. @option{-ansi} implies @option{-fno-asm}.
1441 In C++, this switch only affects the @code{typeof} keyword, since
1442 @code{asm} and @code{inline} are standard keywords. You may want to
1443 use the @option{-fno-gnu-keywords} flag instead, which has the same
1444 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1445 switch only affects the @code{asm} and @code{typeof} keywords, since
1446 @code{inline} is a standard keyword in ISO C99.
1449 @itemx -fno-builtin-@var{function}
1450 @opindex fno-builtin
1451 @cindex built-in functions
1452 Don't recognize built-in functions that do not begin with
1453 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1454 functions provided by GCC}, for details of the functions affected,
1455 including those which are not built-in functions when @option{-ansi} or
1456 @option{-std} options for strict ISO C conformance are used because they
1457 do not have an ISO standard meaning.
1459 GCC normally generates special code to handle certain built-in functions
1460 more efficiently; for instance, calls to @code{alloca} may become single
1461 instructions that adjust the stack directly, and calls to @code{memcpy}
1462 may become inline copy loops. The resulting code is often both smaller
1463 and faster, but since the function calls no longer appear as such, you
1464 cannot set a breakpoint on those calls, nor can you change the behavior
1465 of the functions by linking with a different library. In addition,
1466 when a function is recognized as a built-in function, GCC may use
1467 information about that function to warn about problems with calls to
1468 that function, or to generate more efficient code, even if the
1469 resulting code still contains calls to that function. For example,
1470 warnings are given with @option{-Wformat} for bad calls to
1471 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1472 known not to modify global memory.
1474 With the @option{-fno-builtin-@var{function}} option
1475 only the built-in function @var{function} is
1476 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1477 function is named this is not built-in in this version of GCC, this
1478 option is ignored. There is no corresponding
1479 @option{-fbuiltin-@var{function}} option; if you wish to enable
1480 built-in functions selectively when using @option{-fno-builtin} or
1481 @option{-ffreestanding}, you may define macros such as:
1484 #define abs(n) __builtin_abs ((n))
1485 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1490 @cindex hosted environment
1492 Assert that compilation takes place in a hosted environment. This implies
1493 @option{-fbuiltin}. A hosted environment is one in which the
1494 entire standard library is available, and in which @code{main} has a return
1495 type of @code{int}. Examples are nearly everything except a kernel.
1496 This is equivalent to @option{-fno-freestanding}.
1498 @item -ffreestanding
1499 @opindex ffreestanding
1500 @cindex hosted environment
1502 Assert that compilation takes place in a freestanding environment. This
1503 implies @option{-fno-builtin}. A freestanding environment
1504 is one in which the standard library may not exist, and program startup may
1505 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1506 This is equivalent to @option{-fno-hosted}.
1508 @xref{Standards,,Language Standards Supported by GCC}, for details of
1509 freestanding and hosted environments.
1513 @cindex openmp parallel
1514 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1515 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1516 compiler generates parallel code according to the OpenMP Application
1517 Program Interface v2.5 @w{@uref{http://www.openmp.org/}}.
1519 @item -fms-extensions
1520 @opindex fms-extensions
1521 Accept some non-standard constructs used in Microsoft header files.
1523 Some cases of unnamed fields in structures and unions are only
1524 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1525 fields within structs/unions}, for details.
1529 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1530 options for strict ISO C conformance) implies @option{-trigraphs}.
1532 @item -no-integrated-cpp
1533 @opindex no-integrated-cpp
1534 Performs a compilation in two passes: preprocessing and compiling. This
1535 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1536 @option{-B} option. The user supplied compilation step can then add in
1537 an additional preprocessing step after normal preprocessing but before
1538 compiling. The default is to use the integrated cpp (internal cpp)
1540 The semantics of this option will change if "cc1", "cc1plus", and
1541 "cc1obj" are merged.
1543 @cindex traditional C language
1544 @cindex C language, traditional
1546 @itemx -traditional-cpp
1547 @opindex traditional-cpp
1548 @opindex traditional
1549 Formerly, these options caused GCC to attempt to emulate a pre-standard
1550 C compiler. They are now only supported with the @option{-E} switch.
1551 The preprocessor continues to support a pre-standard mode. See the GNU
1552 CPP manual for details.
1554 @item -fcond-mismatch
1555 @opindex fcond-mismatch
1556 Allow conditional expressions with mismatched types in the second and
1557 third arguments. The value of such an expression is void. This option
1558 is not supported for C++.
1560 @item -flax-vector-conversions
1561 @opindex flax-vector-conversions
1562 Allow implicit conversions between vectors with differing numbers of
1563 elements and/or incompatible element types. This option should not be
1566 @item -funsigned-char
1567 @opindex funsigned-char
1568 Let the type @code{char} be unsigned, like @code{unsigned char}.
1570 Each kind of machine has a default for what @code{char} should
1571 be. It is either like @code{unsigned char} by default or like
1572 @code{signed char} by default.
1574 Ideally, a portable program should always use @code{signed char} or
1575 @code{unsigned char} when it depends on the signedness of an object.
1576 But many programs have been written to use plain @code{char} and
1577 expect it to be signed, or expect it to be unsigned, depending on the
1578 machines they were written for. This option, and its inverse, let you
1579 make such a program work with the opposite default.
1581 The type @code{char} is always a distinct type from each of
1582 @code{signed char} or @code{unsigned char}, even though its behavior
1583 is always just like one of those two.
1586 @opindex fsigned-char
1587 Let the type @code{char} be signed, like @code{signed char}.
1589 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1590 the negative form of @option{-funsigned-char}. Likewise, the option
1591 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1593 @item -fsigned-bitfields
1594 @itemx -funsigned-bitfields
1595 @itemx -fno-signed-bitfields
1596 @itemx -fno-unsigned-bitfields
1597 @opindex fsigned-bitfields
1598 @opindex funsigned-bitfields
1599 @opindex fno-signed-bitfields
1600 @opindex fno-unsigned-bitfields
1601 These options control whether a bit-field is signed or unsigned, when the
1602 declaration does not use either @code{signed} or @code{unsigned}. By
1603 default, such a bit-field is signed, because this is consistent: the
1604 basic integer types such as @code{int} are signed types.
1607 @node C++ Dialect Options
1608 @section Options Controlling C++ Dialect
1610 @cindex compiler options, C++
1611 @cindex C++ options, command line
1612 @cindex options, C++
1613 This section describes the command-line options that are only meaningful
1614 for C++ programs; but you can also use most of the GNU compiler options
1615 regardless of what language your program is in. For example, you
1616 might compile a file @code{firstClass.C} like this:
1619 g++ -g -frepo -O -c firstClass.C
1623 In this example, only @option{-frepo} is an option meant
1624 only for C++ programs; you can use the other options with any
1625 language supported by GCC@.
1627 Here is a list of options that are @emph{only} for compiling C++ programs:
1631 @item -fabi-version=@var{n}
1632 @opindex fabi-version
1633 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1634 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1635 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1636 the version that conforms most closely to the C++ ABI specification.
1637 Therefore, the ABI obtained using version 0 will change as ABI bugs
1640 The default is version 2.
1642 @item -fno-access-control
1643 @opindex fno-access-control
1644 Turn off all access checking. This switch is mainly useful for working
1645 around bugs in the access control code.
1649 Check that the pointer returned by @code{operator new} is non-null
1650 before attempting to modify the storage allocated. This check is
1651 normally unnecessary because the C++ standard specifies that
1652 @code{operator new} will only return @code{0} if it is declared
1653 @samp{throw()}, in which case the compiler will always check the
1654 return value even without this option. In all other cases, when
1655 @code{operator new} has a non-empty exception specification, memory
1656 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1657 @samp{new (nothrow)}.
1659 @item -fconserve-space
1660 @opindex fconserve-space
1661 Put uninitialized or runtime-initialized global variables into the
1662 common segment, as C does. This saves space in the executable at the
1663 cost of not diagnosing duplicate definitions. If you compile with this
1664 flag and your program mysteriously crashes after @code{main()} has
1665 completed, you may have an object that is being destroyed twice because
1666 two definitions were merged.
1668 This option is no longer useful on most targets, now that support has
1669 been added for putting variables into BSS without making them common.
1671 @item -ffriend-injection
1672 @opindex ffriend-injection
1673 Inject friend functions into the enclosing namespace, so that they are
1674 visible outside the scope of the class in which they are declared.
1675 Friend functions were documented to work this way in the old Annotated
1676 C++ Reference Manual, and versions of G++ before 4.1 always worked
1677 that way. However, in ISO C++ a friend function which is not declared
1678 in an enclosing scope can only be found using argument dependent
1679 lookup. This option causes friends to be injected as they were in
1682 This option is for compatibility, and may be removed in a future
1685 @item -fno-elide-constructors
1686 @opindex fno-elide-constructors
1687 The C++ standard allows an implementation to omit creating a temporary
1688 which is only used to initialize another object of the same type.
1689 Specifying this option disables that optimization, and forces G++ to
1690 call the copy constructor in all cases.
1692 @item -fno-enforce-eh-specs
1693 @opindex fno-enforce-eh-specs
1694 Don't generate code to check for violation of exception specifications
1695 at runtime. This option violates the C++ standard, but may be useful
1696 for reducing code size in production builds, much like defining
1697 @samp{NDEBUG}. This does not give user code permission to throw
1698 exceptions in violation of the exception specifications; the compiler
1699 will still optimize based on the specifications, so throwing an
1700 unexpected exception will result in undefined behavior.
1703 @itemx -fno-for-scope
1705 @opindex fno-for-scope
1706 If @option{-ffor-scope} is specified, the scope of variables declared in
1707 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1708 as specified by the C++ standard.
1709 If @option{-fno-for-scope} is specified, the scope of variables declared in
1710 a @i{for-init-statement} extends to the end of the enclosing scope,
1711 as was the case in old versions of G++, and other (traditional)
1712 implementations of C++.
1714 The default if neither flag is given to follow the standard,
1715 but to allow and give a warning for old-style code that would
1716 otherwise be invalid, or have different behavior.
1718 @item -fno-gnu-keywords
1719 @opindex fno-gnu-keywords
1720 Do not recognize @code{typeof} as a keyword, so that code can use this
1721 word as an identifier. You can use the keyword @code{__typeof__} instead.
1722 @option{-ansi} implies @option{-fno-gnu-keywords}.
1724 @item -fno-implicit-templates
1725 @opindex fno-implicit-templates
1726 Never emit code for non-inline templates which are instantiated
1727 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1728 @xref{Template Instantiation}, for more information.
1730 @item -fno-implicit-inline-templates
1731 @opindex fno-implicit-inline-templates
1732 Don't emit code for implicit instantiations of inline templates, either.
1733 The default is to handle inlines differently so that compiles with and
1734 without optimization will need the same set of explicit instantiations.
1736 @item -fno-implement-inlines
1737 @opindex fno-implement-inlines
1738 To save space, do not emit out-of-line copies of inline functions
1739 controlled by @samp{#pragma implementation}. This will cause linker
1740 errors if these functions are not inlined everywhere they are called.
1742 @item -fms-extensions
1743 @opindex fms-extensions
1744 Disable pedantic warnings about constructs used in MFC, such as implicit
1745 int and getting a pointer to member function via non-standard syntax.
1747 @item -fno-nonansi-builtins
1748 @opindex fno-nonansi-builtins
1749 Disable built-in declarations of functions that are not mandated by
1750 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1751 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1753 @item -fno-operator-names
1754 @opindex fno-operator-names
1755 Do not treat the operator name keywords @code{and}, @code{bitand},
1756 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1757 synonyms as keywords.
1759 @item -fno-optional-diags
1760 @opindex fno-optional-diags
1761 Disable diagnostics that the standard says a compiler does not need to
1762 issue. Currently, the only such diagnostic issued by G++ is the one for
1763 a name having multiple meanings within a class.
1766 @opindex fpermissive
1767 Downgrade some diagnostics about nonconformant code from errors to
1768 warnings. Thus, using @option{-fpermissive} will allow some
1769 nonconforming code to compile.
1773 Enable automatic template instantiation at link time. This option also
1774 implies @option{-fno-implicit-templates}. @xref{Template
1775 Instantiation}, for more information.
1779 Disable generation of information about every class with virtual
1780 functions for use by the C++ runtime type identification features
1781 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1782 of the language, you can save some space by using this flag. Note that
1783 exception handling uses the same information, but it will generate it as
1784 needed. The @samp{dynamic_cast} operator can still be used for casts that
1785 do not require runtime type information, i.e. casts to @code{void *} or to
1786 unambiguous base classes.
1790 Emit statistics about front-end processing at the end of the compilation.
1791 This information is generally only useful to the G++ development team.
1793 @item -ftemplate-depth-@var{n}
1794 @opindex ftemplate-depth
1795 Set the maximum instantiation depth for template classes to @var{n}.
1796 A limit on the template instantiation depth is needed to detect
1797 endless recursions during template class instantiation. ANSI/ISO C++
1798 conforming programs must not rely on a maximum depth greater than 17.
1800 @item -fno-threadsafe-statics
1801 @opindex fno-threadsafe-statics
1802 Do not emit the extra code to use the routines specified in the C++
1803 ABI for thread-safe initialization of local statics. You can use this
1804 option to reduce code size slightly in code that doesn't need to be
1807 @item -fuse-cxa-atexit
1808 @opindex fuse-cxa-atexit
1809 Register destructors for objects with static storage duration with the
1810 @code{__cxa_atexit} function rather than the @code{atexit} function.
1811 This option is required for fully standards-compliant handling of static
1812 destructors, but will only work if your C library supports
1813 @code{__cxa_atexit}.
1815 @item -fno-use-cxa-get-exception-ptr
1816 @opindex fno-use-cxa-get-exception-ptr
1817 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1818 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1819 if the runtime routine is not available.
1821 @item -fvisibility-inlines-hidden
1822 @opindex fvisibility-inlines-hidden
1823 This switch declares that the user does not attempt to compare
1824 pointers to inline methods where the addresses of the two functions
1825 were taken in different shared objects.
1827 The effect of this is that GCC may, effectively, mark inline methods with
1828 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1829 appear in the export table of a DSO and do not require a PLT indirection
1830 when used within the DSO@. Enabling this option can have a dramatic effect
1831 on load and link times of a DSO as it massively reduces the size of the
1832 dynamic export table when the library makes heavy use of templates.
1834 The behavior of this switch is not quite the same as marking the
1835 methods as hidden directly, because it does not affect static variables
1836 local to the function or cause the compiler to deduce that
1837 the function is defined in only one shared object.
1839 You may mark a method as having a visibility explicitly to negate the
1840 effect of the switch for that method. For example, if you do want to
1841 compare pointers to a particular inline method, you might mark it as
1842 having default visibility. Marking the enclosing class with explicit
1843 visibility will have no effect.
1845 Explicitly instantiated inline methods are unaffected by this option
1846 as their linkage might otherwise cross a shared library boundary.
1847 @xref{Template Instantiation}.
1849 @item -fvisibility-ms-compat
1850 @opindex fvisibility-ms-compat
1851 This flag attempts to use visibility settings to make GCC's C++
1852 linkage model compatible with that of Microsoft Visual Studio.
1854 The flag makes these changes to GCC's linkage model:
1858 It sets the default visibility to @code{hidden}, like
1859 @option{-fvisibility=hidden}.
1862 Types, but not their members, are not hidden by default.
1865 The One Definition Rule is relaxed for types without explicit
1866 visibility specifications which are defined in more than one different
1867 shared object: those declarations are permitted if they would have
1868 been permitted when this option was not used.
1871 In new code it is better to use @option{-fvisibility=hidden} and
1872 export those classes which are intended to be externally visible.
1873 Unfortunately it is possible for code to rely, perhaps accidentally,
1874 on the Visual Studio behavior.
1876 Among the consequences of these changes are that static data members
1877 of the same type with the same name but defined in different shared
1878 objects will be different, so changing one will not change the other;
1879 and that pointers to function members defined in different shared
1880 objects may not compare equal. When this flag is given, it is a
1881 violation of the ODR to define types with the same name differently.
1885 Do not use weak symbol support, even if it is provided by the linker.
1886 By default, G++ will use weak symbols if they are available. This
1887 option exists only for testing, and should not be used by end-users;
1888 it will result in inferior code and has no benefits. This option may
1889 be removed in a future release of G++.
1893 Do not search for header files in the standard directories specific to
1894 C++, but do still search the other standard directories. (This option
1895 is used when building the C++ library.)
1898 In addition, these optimization, warning, and code generation options
1899 have meanings only for C++ programs:
1902 @item -fno-default-inline
1903 @opindex fno-default-inline
1904 Do not assume @samp{inline} for functions defined inside a class scope.
1905 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1906 functions will have linkage like inline functions; they just won't be
1909 @item -Wabi @r{(C++ only)}
1911 Warn when G++ generates code that is probably not compatible with the
1912 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1913 all such cases, there are probably some cases that are not warned about,
1914 even though G++ is generating incompatible code. There may also be
1915 cases where warnings are emitted even though the code that is generated
1918 You should rewrite your code to avoid these warnings if you are
1919 concerned about the fact that code generated by G++ may not be binary
1920 compatible with code generated by other compilers.
1922 The known incompatibilities at this point include:
1927 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1928 pack data into the same byte as a base class. For example:
1931 struct A @{ virtual void f(); int f1 : 1; @};
1932 struct B : public A @{ int f2 : 1; @};
1936 In this case, G++ will place @code{B::f2} into the same byte
1937 as@code{A::f1}; other compilers will not. You can avoid this problem
1938 by explicitly padding @code{A} so that its size is a multiple of the
1939 byte size on your platform; that will cause G++ and other compilers to
1940 layout @code{B} identically.
1943 Incorrect handling of tail-padding for virtual bases. G++ does not use
1944 tail padding when laying out virtual bases. For example:
1947 struct A @{ virtual void f(); char c1; @};
1948 struct B @{ B(); char c2; @};
1949 struct C : public A, public virtual B @{@};
1953 In this case, G++ will not place @code{B} into the tail-padding for
1954 @code{A}; other compilers will. You can avoid this problem by
1955 explicitly padding @code{A} so that its size is a multiple of its
1956 alignment (ignoring virtual base classes); that will cause G++ and other
1957 compilers to layout @code{C} identically.
1960 Incorrect handling of bit-fields with declared widths greater than that
1961 of their underlying types, when the bit-fields appear in a union. For
1965 union U @{ int i : 4096; @};
1969 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1970 union too small by the number of bits in an @code{int}.
1973 Empty classes can be placed at incorrect offsets. For example:
1983 struct C : public B, public A @{@};
1987 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1988 it should be placed at offset zero. G++ mistakenly believes that the
1989 @code{A} data member of @code{B} is already at offset zero.
1992 Names of template functions whose types involve @code{typename} or
1993 template template parameters can be mangled incorrectly.
1996 template <typename Q>
1997 void f(typename Q::X) @{@}
1999 template <template <typename> class Q>
2000 void f(typename Q<int>::X) @{@}
2004 Instantiations of these templates may be mangled incorrectly.
2008 @item -Wctor-dtor-privacy @r{(C++ only)}
2009 @opindex Wctor-dtor-privacy
2010 Warn when a class seems unusable because all the constructors or
2011 destructors in that class are private, and it has neither friends nor
2012 public static member functions.
2014 @item -Wnon-virtual-dtor @r{(C++ only)}
2015 @opindex Wnon-virtual-dtor
2016 Warn when a class has virtual functions and accessible non-virtual
2017 destructor, in which case it would be possible but unsafe to delete
2018 an instance of a derived class through a pointer to the base class.
2019 This warning is also enabled if -Weffc++ is specified.
2021 @item -Wreorder @r{(C++ only)}
2023 @cindex reordering, warning
2024 @cindex warning for reordering of member initializers
2025 Warn when the order of member initializers given in the code does not
2026 match the order in which they must be executed. For instance:
2032 A(): j (0), i (1) @{ @}
2036 The compiler will rearrange the member initializers for @samp{i}
2037 and @samp{j} to match the declaration order of the members, emitting
2038 a warning to that effect. This warning is enabled by @option{-Wall}.
2041 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2044 @item -Weffc++ @r{(C++ only)}
2046 Warn about violations of the following style guidelines from Scott Meyers'
2047 @cite{Effective C++} book:
2051 Item 11: Define a copy constructor and an assignment operator for classes
2052 with dynamically allocated memory.
2055 Item 12: Prefer initialization to assignment in constructors.
2058 Item 14: Make destructors virtual in base classes.
2061 Item 15: Have @code{operator=} return a reference to @code{*this}.
2064 Item 23: Don't try to return a reference when you must return an object.
2068 Also warn about violations of the following style guidelines from
2069 Scott Meyers' @cite{More Effective C++} book:
2073 Item 6: Distinguish between prefix and postfix forms of increment and
2074 decrement operators.
2077 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2081 When selecting this option, be aware that the standard library
2082 headers do not obey all of these guidelines; use @samp{grep -v}
2083 to filter out those warnings.
2085 @item -Wno-deprecated @r{(C++ only)}
2086 @opindex Wno-deprecated
2087 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
2089 @item -Wstrict-null-sentinel @r{(C++ only)}
2090 @opindex Wstrict-null-sentinel
2091 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2092 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2093 to @code{__null}. Although it is a null pointer constant not a null pointer,
2094 it is guaranteed to of the same size as a pointer. But this use is
2095 not portable across different compilers.
2097 @item -Wno-non-template-friend @r{(C++ only)}
2098 @opindex Wno-non-template-friend
2099 Disable warnings when non-templatized friend functions are declared
2100 within a template. Since the advent of explicit template specification
2101 support in G++, if the name of the friend is an unqualified-id (i.e.,
2102 @samp{friend foo(int)}), the C++ language specification demands that the
2103 friend declare or define an ordinary, nontemplate function. (Section
2104 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2105 could be interpreted as a particular specialization of a templatized
2106 function. Because this non-conforming behavior is no longer the default
2107 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2108 check existing code for potential trouble spots and is on by default.
2109 This new compiler behavior can be turned off with
2110 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2111 but disables the helpful warning.
2113 @item -Wold-style-cast @r{(C++ only)}
2114 @opindex Wold-style-cast
2115 Warn if an old-style (C-style) cast to a non-void type is used within
2116 a C++ program. The new-style casts (@samp{dynamic_cast},
2117 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2118 less vulnerable to unintended effects and much easier to search for.
2120 @item -Woverloaded-virtual @r{(C++ only)}
2121 @opindex Woverloaded-virtual
2122 @cindex overloaded virtual fn, warning
2123 @cindex warning for overloaded virtual fn
2124 Warn when a function declaration hides virtual functions from a
2125 base class. For example, in:
2132 struct B: public A @{
2137 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2145 will fail to compile.
2147 @item -Wno-pmf-conversions @r{(C++ only)}
2148 @opindex Wno-pmf-conversions
2149 Disable the diagnostic for converting a bound pointer to member function
2152 @item -Wsign-promo @r{(C++ only)}
2153 @opindex Wsign-promo
2154 Warn when overload resolution chooses a promotion from unsigned or
2155 enumerated type to a signed type, over a conversion to an unsigned type of
2156 the same size. Previous versions of G++ would try to preserve
2157 unsignedness, but the standard mandates the current behavior.
2162 A& operator = (int);
2172 In this example, G++ will synthesize a default @samp{A& operator =
2173 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2176 @node Objective-C and Objective-C++ Dialect Options
2177 @section Options Controlling Objective-C and Objective-C++ Dialects
2179 @cindex compiler options, Objective-C and Objective-C++
2180 @cindex Objective-C and Objective-C++ options, command line
2181 @cindex options, Objective-C and Objective-C++
2182 (NOTE: This manual does not describe the Objective-C and Objective-C++
2183 languages themselves. See @xref{Standards,,Language Standards
2184 Supported by GCC}, for references.)
2186 This section describes the command-line options that are only meaningful
2187 for Objective-C and Objective-C++ programs, but you can also use most of
2188 the language-independent GNU compiler options.
2189 For example, you might compile a file @code{some_class.m} like this:
2192 gcc -g -fgnu-runtime -O -c some_class.m
2196 In this example, @option{-fgnu-runtime} is an option meant only for
2197 Objective-C and Objective-C++ programs; you can use the other options with
2198 any language supported by GCC@.
2200 Note that since Objective-C is an extension of the C language, Objective-C
2201 compilations may also use options specific to the C front-end (e.g.,
2202 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2203 C++-specific options (e.g., @option{-Wabi}).
2205 Here is a list of options that are @emph{only} for compiling Objective-C
2206 and Objective-C++ programs:
2209 @item -fconstant-string-class=@var{class-name}
2210 @opindex fconstant-string-class
2211 Use @var{class-name} as the name of the class to instantiate for each
2212 literal string specified with the syntax @code{@@"@dots{}"}. The default
2213 class name is @code{NXConstantString} if the GNU runtime is being used, and
2214 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2215 @option{-fconstant-cfstrings} option, if also present, will override the
2216 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2217 to be laid out as constant CoreFoundation strings.
2220 @opindex fgnu-runtime
2221 Generate object code compatible with the standard GNU Objective-C
2222 runtime. This is the default for most types of systems.
2224 @item -fnext-runtime
2225 @opindex fnext-runtime
2226 Generate output compatible with the NeXT runtime. This is the default
2227 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2228 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2231 @item -fno-nil-receivers
2232 @opindex fno-nil-receivers
2233 Assume that all Objective-C message dispatches (e.g.,
2234 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2235 is not @code{nil}. This allows for more efficient entry points in the runtime
2236 to be used. Currently, this option is only available in conjunction with
2237 the NeXT runtime on Mac OS X 10.3 and later.
2239 @item -fobjc-call-cxx-cdtors
2240 @opindex fobjc-call-cxx-cdtors
2241 For each Objective-C class, check if any of its instance variables is a
2242 C++ object with a non-trivial default constructor. If so, synthesize a
2243 special @code{- (id) .cxx_construct} instance method that will run
2244 non-trivial default constructors on any such instance variables, in order,
2245 and then return @code{self}. Similarly, check if any instance variable
2246 is a C++ object with a non-trivial destructor, and if so, synthesize a
2247 special @code{- (void) .cxx_destruct} method that will run
2248 all such default destructors, in reverse order.
2250 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2251 thusly generated will only operate on instance variables declared in the
2252 current Objective-C class, and not those inherited from superclasses. It
2253 is the responsibility of the Objective-C runtime to invoke all such methods
2254 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2255 will be invoked by the runtime immediately after a new object
2256 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2257 be invoked immediately before the runtime deallocates an object instance.
2259 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2260 support for invoking the @code{- (id) .cxx_construct} and
2261 @code{- (void) .cxx_destruct} methods.
2263 @item -fobjc-direct-dispatch
2264 @opindex fobjc-direct-dispatch
2265 Allow fast jumps to the message dispatcher. On Darwin this is
2266 accomplished via the comm page.
2268 @item -fobjc-exceptions
2269 @opindex fobjc-exceptions
2270 Enable syntactic support for structured exception handling in Objective-C,
2271 similar to what is offered by C++ and Java. This option is
2272 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2281 @@catch (AnObjCClass *exc) @{
2288 @@catch (AnotherClass *exc) @{
2291 @@catch (id allOthers) @{
2301 The @code{@@throw} statement may appear anywhere in an Objective-C or
2302 Objective-C++ program; when used inside of a @code{@@catch} block, the
2303 @code{@@throw} may appear without an argument (as shown above), in which case
2304 the object caught by the @code{@@catch} will be rethrown.
2306 Note that only (pointers to) Objective-C objects may be thrown and
2307 caught using this scheme. When an object is thrown, it will be caught
2308 by the nearest @code{@@catch} clause capable of handling objects of that type,
2309 analogously to how @code{catch} blocks work in C++ and Java. A
2310 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2311 any and all Objective-C exceptions not caught by previous @code{@@catch}
2314 The @code{@@finally} clause, if present, will be executed upon exit from the
2315 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2316 regardless of whether any exceptions are thrown, caught or rethrown
2317 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2318 of the @code{finally} clause in Java.
2320 There are several caveats to using the new exception mechanism:
2324 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2325 idioms provided by the @code{NSException} class, the new
2326 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2327 systems, due to additional functionality needed in the (NeXT) Objective-C
2331 As mentioned above, the new exceptions do not support handling
2332 types other than Objective-C objects. Furthermore, when used from
2333 Objective-C++, the Objective-C exception model does not interoperate with C++
2334 exceptions at this time. This means you cannot @code{@@throw} an exception
2335 from Objective-C and @code{catch} it in C++, or vice versa
2336 (i.e., @code{throw @dots{} @@catch}).
2339 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2340 blocks for thread-safe execution:
2343 @@synchronized (ObjCClass *guard) @{
2348 Upon entering the @code{@@synchronized} block, a thread of execution shall
2349 first check whether a lock has been placed on the corresponding @code{guard}
2350 object by another thread. If it has, the current thread shall wait until
2351 the other thread relinquishes its lock. Once @code{guard} becomes available,
2352 the current thread will place its own lock on it, execute the code contained in
2353 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2354 making @code{guard} available to other threads).
2356 Unlike Java, Objective-C does not allow for entire methods to be marked
2357 @code{@@synchronized}. Note that throwing exceptions out of
2358 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2359 to be unlocked properly.
2363 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2365 @item -freplace-objc-classes
2366 @opindex freplace-objc-classes
2367 Emit a special marker instructing @command{ld(1)} not to statically link in
2368 the resulting object file, and allow @command{dyld(1)} to load it in at
2369 run time instead. This is used in conjunction with the Fix-and-Continue
2370 debugging mode, where the object file in question may be recompiled and
2371 dynamically reloaded in the course of program execution, without the need
2372 to restart the program itself. Currently, Fix-and-Continue functionality
2373 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2378 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2379 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2380 compile time) with static class references that get initialized at load time,
2381 which improves run-time performance. Specifying the @option{-fzero-link} flag
2382 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2383 to be retained. This is useful in Zero-Link debugging mode, since it allows
2384 for individual class implementations to be modified during program execution.
2388 Dump interface declarations for all classes seen in the source file to a
2389 file named @file{@var{sourcename}.decl}.
2391 @item -Wassign-intercept
2392 @opindex Wassign-intercept
2393 Warn whenever an Objective-C assignment is being intercepted by the
2397 @opindex Wno-protocol
2398 If a class is declared to implement a protocol, a warning is issued for
2399 every method in the protocol that is not implemented by the class. The
2400 default behavior is to issue a warning for every method not explicitly
2401 implemented in the class, even if a method implementation is inherited
2402 from the superclass. If you use the @option{-Wno-protocol} option, then
2403 methods inherited from the superclass are considered to be implemented,
2404 and no warning is issued for them.
2408 Warn if multiple methods of different types for the same selector are
2409 found during compilation. The check is performed on the list of methods
2410 in the final stage of compilation. Additionally, a check is performed
2411 for each selector appearing in a @code{@@selector(@dots{})}
2412 expression, and a corresponding method for that selector has been found
2413 during compilation. Because these checks scan the method table only at
2414 the end of compilation, these warnings are not produced if the final
2415 stage of compilation is not reached, for example because an error is
2416 found during compilation, or because the @option{-fsyntax-only} option is
2419 @item -Wstrict-selector-match
2420 @opindex Wstrict-selector-match
2421 Warn if multiple methods with differing argument and/or return types are
2422 found for a given selector when attempting to send a message using this
2423 selector to a receiver of type @code{id} or @code{Class}. When this flag
2424 is off (which is the default behavior), the compiler will omit such warnings
2425 if any differences found are confined to types which share the same size
2428 @item -Wundeclared-selector
2429 @opindex Wundeclared-selector
2430 Warn if a @code{@@selector(@dots{})} expression referring to an
2431 undeclared selector is found. A selector is considered undeclared if no
2432 method with that name has been declared before the
2433 @code{@@selector(@dots{})} expression, either explicitly in an
2434 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2435 an @code{@@implementation} section. This option always performs its
2436 checks as soon as a @code{@@selector(@dots{})} expression is found,
2437 while @option{-Wselector} only performs its checks in the final stage of
2438 compilation. This also enforces the coding style convention
2439 that methods and selectors must be declared before being used.
2441 @item -print-objc-runtime-info
2442 @opindex print-objc-runtime-info
2443 Generate C header describing the largest structure that is passed by
2448 @node Language Independent Options
2449 @section Options to Control Diagnostic Messages Formatting
2450 @cindex options to control diagnostics formatting
2451 @cindex diagnostic messages
2452 @cindex message formatting
2454 Traditionally, diagnostic messages have been formatted irrespective of
2455 the output device's aspect (e.g.@: its width, @dots{}). The options described
2456 below can be used to control the diagnostic messages formatting
2457 algorithm, e.g.@: how many characters per line, how often source location
2458 information should be reported. Right now, only the C++ front end can
2459 honor these options. However it is expected, in the near future, that
2460 the remaining front ends would be able to digest them correctly.
2463 @item -fmessage-length=@var{n}
2464 @opindex fmessage-length
2465 Try to format error messages so that they fit on lines of about @var{n}
2466 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2467 the front ends supported by GCC@. If @var{n} is zero, then no
2468 line-wrapping will be done; each error message will appear on a single
2471 @opindex fdiagnostics-show-location
2472 @item -fdiagnostics-show-location=once
2473 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2474 reporter to emit @emph{once} source location information; that is, in
2475 case the message is too long to fit on a single physical line and has to
2476 be wrapped, the source location won't be emitted (as prefix) again,
2477 over and over, in subsequent continuation lines. This is the default
2480 @item -fdiagnostics-show-location=every-line
2481 Only meaningful in line-wrapping mode. Instructs the diagnostic
2482 messages reporter to emit the same source location information (as
2483 prefix) for physical lines that result from the process of breaking
2484 a message which is too long to fit on a single line.
2486 @item -fdiagnostics-show-option
2487 @opindex fdiagnostics-show-option
2488 This option instructs the diagnostic machinery to add text to each
2489 diagnostic emitted, which indicates which command line option directly
2490 controls that diagnostic, when such an option is known to the
2491 diagnostic machinery.
2493 @item -Wcoverage-mismatch
2494 @opindex Wcoverage-mismatch
2495 Warn if feedback profiles do not match when using the
2496 @option{-fprofile-use} option.
2497 If a source file was changed between @option{-fprofile-gen} and
2498 @option{-fprofile-use}, the files with the profile feedback can fail
2499 to match the source file and GCC can not use the profile feedback
2500 information. By default, GCC emits an error message in this case.
2501 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2502 error. GCC does not use appropriate feedback profiles, so using this
2503 option can result in poorly optimized code. This option is useful
2504 only in the case of very minor changes such as bug fixes to an
2509 @node Warning Options
2510 @section Options to Request or Suppress Warnings
2511 @cindex options to control warnings
2512 @cindex warning messages
2513 @cindex messages, warning
2514 @cindex suppressing warnings
2516 Warnings are diagnostic messages that report constructions which
2517 are not inherently erroneous but which are risky or suggest there
2518 may have been an error.
2520 You can request many specific warnings with options beginning @samp{-W},
2521 for example @option{-Wimplicit} to request warnings on implicit
2522 declarations. Each of these specific warning options also has a
2523 negative form beginning @samp{-Wno-} to turn off warnings;
2524 for example, @option{-Wno-implicit}. This manual lists only one of the
2525 two forms, whichever is not the default.
2527 The following options control the amount and kinds of warnings produced
2528 by GCC; for further, language-specific options also refer to
2529 @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect
2533 @cindex syntax checking
2535 @opindex fsyntax-only
2536 Check the code for syntax errors, but don't do anything beyond that.
2540 Issue all the warnings demanded by strict ISO C and ISO C++;
2541 reject all programs that use forbidden extensions, and some other
2542 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2543 version of the ISO C standard specified by any @option{-std} option used.
2545 Valid ISO C and ISO C++ programs should compile properly with or without
2546 this option (though a rare few will require @option{-ansi} or a
2547 @option{-std} option specifying the required version of ISO C)@. However,
2548 without this option, certain GNU extensions and traditional C and C++
2549 features are supported as well. With this option, they are rejected.
2551 @option{-pedantic} does not cause warning messages for use of the
2552 alternate keywords whose names begin and end with @samp{__}. Pedantic
2553 warnings are also disabled in the expression that follows
2554 @code{__extension__}. However, only system header files should use
2555 these escape routes; application programs should avoid them.
2556 @xref{Alternate Keywords}.
2558 Some users try to use @option{-pedantic} to check programs for strict ISO
2559 C conformance. They soon find that it does not do quite what they want:
2560 it finds some non-ISO practices, but not all---only those for which
2561 ISO C @emph{requires} a diagnostic, and some others for which
2562 diagnostics have been added.
2564 A feature to report any failure to conform to ISO C might be useful in
2565 some instances, but would require considerable additional work and would
2566 be quite different from @option{-pedantic}. We don't have plans to
2567 support such a feature in the near future.
2569 Where the standard specified with @option{-std} represents a GNU
2570 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2571 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2572 extended dialect is based. Warnings from @option{-pedantic} are given
2573 where they are required by the base standard. (It would not make sense
2574 for such warnings to be given only for features not in the specified GNU
2575 C dialect, since by definition the GNU dialects of C include all
2576 features the compiler supports with the given option, and there would be
2577 nothing to warn about.)
2579 @item -pedantic-errors
2580 @opindex pedantic-errors
2581 Like @option{-pedantic}, except that errors are produced rather than
2586 Inhibit all warning messages.
2590 Inhibit warning messages about the use of @samp{#import}.
2592 @item -Wchar-subscripts
2593 @opindex Wchar-subscripts
2594 Warn if an array subscript has type @code{char}. This is a common cause
2595 of error, as programmers often forget that this type is signed on some
2597 This warning is enabled by @option{-Wall}.
2601 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2602 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2603 This warning is enabled by @option{-Wall}.
2605 @item -Wfatal-errors
2606 @opindex Wfatal-errors
2607 This option causes the compiler to abort compilation on the first error
2608 occurred rather than trying to keep going and printing further error
2613 @opindex ffreestanding
2614 @opindex fno-builtin
2615 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2616 the arguments supplied have types appropriate to the format string
2617 specified, and that the conversions specified in the format string make
2618 sense. This includes standard functions, and others specified by format
2619 attributes (@pxref{Function Attributes}), in the @code{printf},
2620 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2621 not in the C standard) families (or other target-specific families).
2622 Which functions are checked without format attributes having been
2623 specified depends on the standard version selected, and such checks of
2624 functions without the attribute specified are disabled by
2625 @option{-ffreestanding} or @option{-fno-builtin}.
2627 The formats are checked against the format features supported by GNU
2628 libc version 2.2. These include all ISO C90 and C99 features, as well
2629 as features from the Single Unix Specification and some BSD and GNU
2630 extensions. Other library implementations may not support all these
2631 features; GCC does not support warning about features that go beyond a
2632 particular library's limitations. However, if @option{-pedantic} is used
2633 with @option{-Wformat}, warnings will be given about format features not
2634 in the selected standard version (but not for @code{strfmon} formats,
2635 since those are not in any version of the C standard). @xref{C Dialect
2636 Options,,Options Controlling C Dialect}.
2638 Since @option{-Wformat} also checks for null format arguments for
2639 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2641 @option{-Wformat} is included in @option{-Wall}. For more control over some
2642 aspects of format checking, the options @option{-Wformat-y2k},
2643 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2644 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2645 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2648 @opindex Wformat-y2k
2649 If @option{-Wformat} is specified, also warn about @code{strftime}
2650 formats which may yield only a two-digit year.
2652 @item -Wno-format-extra-args
2653 @opindex Wno-format-extra-args
2654 If @option{-Wformat} is specified, do not warn about excess arguments to a
2655 @code{printf} or @code{scanf} format function. The C standard specifies
2656 that such arguments are ignored.
2658 Where the unused arguments lie between used arguments that are
2659 specified with @samp{$} operand number specifications, normally
2660 warnings are still given, since the implementation could not know what
2661 type to pass to @code{va_arg} to skip the unused arguments. However,
2662 in the case of @code{scanf} formats, this option will suppress the
2663 warning if the unused arguments are all pointers, since the Single
2664 Unix Specification says that such unused arguments are allowed.
2666 @item -Wno-format-zero-length
2667 @opindex Wno-format-zero-length
2668 If @option{-Wformat} is specified, do not warn about zero-length formats.
2669 The C standard specifies that zero-length formats are allowed.
2671 @item -Wformat-nonliteral
2672 @opindex Wformat-nonliteral
2673 If @option{-Wformat} is specified, also warn if the format string is not a
2674 string literal and so cannot be checked, unless the format function
2675 takes its format arguments as a @code{va_list}.
2677 @item -Wformat-security
2678 @opindex Wformat-security
2679 If @option{-Wformat} is specified, also warn about uses of format
2680 functions that represent possible security problems. At present, this
2681 warns about calls to @code{printf} and @code{scanf} functions where the
2682 format string is not a string literal and there are no format arguments,
2683 as in @code{printf (foo);}. This may be a security hole if the format
2684 string came from untrusted input and contains @samp{%n}. (This is
2685 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2686 in future warnings may be added to @option{-Wformat-security} that are not
2687 included in @option{-Wformat-nonliteral}.)
2691 Enable @option{-Wformat} plus format checks not included in
2692 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2693 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2697 Warn about passing a null pointer for arguments marked as
2698 requiring a non-null value by the @code{nonnull} function attribute.
2700 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2701 can be disabled with the @option{-Wno-nonnull} option.
2703 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2705 Warn about uninitialized variables which are initialized with themselves.
2706 Note this option can only be used with the @option{-Wuninitialized} option,
2707 which in turn only works with @option{-O1} and above.
2709 For example, GCC will warn about @code{i} being uninitialized in the
2710 following snippet only when @option{-Winit-self} has been specified:
2721 @item -Wimplicit-int
2722 @opindex Wimplicit-int
2723 Warn when a declaration does not specify a type.
2724 This warning is enabled by @option{-Wall}.
2726 @item -Wimplicit-function-declaration
2727 @opindex Wimplicit-function-declaration
2728 @opindex Wno-implicit-function-declaration
2729 Give a warning whenever a function is used before being declared. In
2730 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
2731 enabled by default and it is made into an error by
2732 @option{-pedantic-errors}. This warning is also enabled by
2737 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2738 This warning is enabled by @option{-Wall}.
2742 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2743 function with external linkage, returning int, taking either zero
2744 arguments, two, or three arguments of appropriate types.
2745 This warning is enabled by @option{-Wall}.
2747 @item -Wmissing-braces
2748 @opindex Wmissing-braces
2749 Warn if an aggregate or union initializer is not fully bracketed. In
2750 the following example, the initializer for @samp{a} is not fully
2751 bracketed, but that for @samp{b} is fully bracketed.
2754 int a[2][2] = @{ 0, 1, 2, 3 @};
2755 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2758 This warning is enabled by @option{-Wall}.
2760 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2761 @opindex Wmissing-include-dirs
2762 Warn if a user-supplied include directory does not exist.
2765 @opindex Wparentheses
2766 Warn if parentheses are omitted in certain contexts, such
2767 as when there is an assignment in a context where a truth value
2768 is expected, or when operators are nested whose precedence people
2769 often get confused about.
2771 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2772 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2773 interpretation from that of ordinary mathematical notation.
2775 Also warn about constructions where there may be confusion to which
2776 @code{if} statement an @code{else} branch belongs. Here is an example of
2791 In C/C++, every @code{else} branch belongs to the innermost possible
2792 @code{if} statement, which in this example is @code{if (b)}. This is
2793 often not what the programmer expected, as illustrated in the above
2794 example by indentation the programmer chose. When there is the
2795 potential for this confusion, GCC will issue a warning when this flag
2796 is specified. To eliminate the warning, add explicit braces around
2797 the innermost @code{if} statement so there is no way the @code{else}
2798 could belong to the enclosing @code{if}. The resulting code would
2815 This warning is enabled by @option{-Wall}.
2817 @item -Wsequence-point
2818 @opindex Wsequence-point
2819 Warn about code that may have undefined semantics because of violations
2820 of sequence point rules in the C and C++ standards.
2822 The C and C++ standards defines the order in which expressions in a C/C++
2823 program are evaluated in terms of @dfn{sequence points}, which represent
2824 a partial ordering between the execution of parts of the program: those
2825 executed before the sequence point, and those executed after it. These
2826 occur after the evaluation of a full expression (one which is not part
2827 of a larger expression), after the evaluation of the first operand of a
2828 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2829 function is called (but after the evaluation of its arguments and the
2830 expression denoting the called function), and in certain other places.
2831 Other than as expressed by the sequence point rules, the order of
2832 evaluation of subexpressions of an expression is not specified. All
2833 these rules describe only a partial order rather than a total order,
2834 since, for example, if two functions are called within one expression
2835 with no sequence point between them, the order in which the functions
2836 are called is not specified. However, the standards committee have
2837 ruled that function calls do not overlap.
2839 It is not specified when between sequence points modifications to the
2840 values of objects take effect. Programs whose behavior depends on this
2841 have undefined behavior; the C and C++ standards specify that ``Between
2842 the previous and next sequence point an object shall have its stored
2843 value modified at most once by the evaluation of an expression.
2844 Furthermore, the prior value shall be read only to determine the value
2845 to be stored.''. If a program breaks these rules, the results on any
2846 particular implementation are entirely unpredictable.
2848 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2849 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2850 diagnosed by this option, and it may give an occasional false positive
2851 result, but in general it has been found fairly effective at detecting
2852 this sort of problem in programs.
2854 The standard is worded confusingly, therefore there is some debate
2855 over the precise meaning of the sequence point rules in subtle cases.
2856 Links to discussions of the problem, including proposed formal
2857 definitions, may be found on the GCC readings page, at
2858 @w{@uref{http://gcc.gnu.org/readings.html}}.
2860 This warning is enabled by @option{-Wall} for C and C++.
2863 @opindex Wreturn-type
2864 @opindex Wno-return-type
2865 Warn whenever a function is defined with a return-type that defaults
2866 to @code{int}. Also warn about any @code{return} statement with no
2867 return-value in a function whose return-type is not @code{void}
2868 (falling off the end of the function body is considered returning
2869 without a value), and about a @code{return} statement with a
2870 expression in a function whose return-type is @code{void}.
2872 Also warn if the return type of a function has a type qualifier
2873 such as @code{const}. For ISO C such a type qualifier has no effect,
2874 since the value returned by a function is not an lvalue.
2875 For C++, the warning is only emitted for scalar types or @code{void}.
2876 ISO C prohibits qualified @code{void} return types on function
2877 definitions, so such return types always receive a warning
2878 even without this option.
2880 For C++, a function without return type always produces a diagnostic
2881 message, even when @option{-Wno-return-type} is specified. The only
2882 exceptions are @samp{main} and functions defined in system headers.
2884 This warning is enabled by @option{-Wall}.
2888 Warn whenever a @code{switch} statement has an index of enumerated type
2889 and lacks a @code{case} for one or more of the named codes of that
2890 enumeration. (The presence of a @code{default} label prevents this
2891 warning.) @code{case} labels outside the enumeration range also
2892 provoke warnings when this option is used.
2893 This warning is enabled by @option{-Wall}.
2895 @item -Wswitch-default
2896 @opindex Wswitch-switch
2897 Warn whenever a @code{switch} statement does not have a @code{default}
2901 @opindex Wswitch-enum
2902 Warn whenever a @code{switch} statement has an index of enumerated type
2903 and lacks a @code{case} for one or more of the named codes of that
2904 enumeration. @code{case} labels outside the enumeration range also
2905 provoke warnings when this option is used.
2909 Warn if any trigraphs are encountered that might change the meaning of
2910 the program (trigraphs within comments are not warned about).
2911 This warning is enabled by @option{-Wall}.
2913 @item -Wunused-function
2914 @opindex Wunused-function
2915 Warn whenever a static function is declared but not defined or a
2916 non-inline static function is unused.
2917 This warning is enabled by @option{-Wall}.
2919 @item -Wunused-label
2920 @opindex Wunused-label
2921 Warn whenever a label is declared but not used.
2922 This warning is enabled by @option{-Wall}.
2924 To suppress this warning use the @samp{unused} attribute
2925 (@pxref{Variable Attributes}).
2927 @item -Wunused-parameter
2928 @opindex Wunused-parameter
2929 Warn whenever a function parameter is unused aside from its declaration.
2931 To suppress this warning use the @samp{unused} attribute
2932 (@pxref{Variable Attributes}).
2934 @item -Wunused-variable
2935 @opindex Wunused-variable
2936 Warn whenever a local variable or non-constant static variable is unused
2937 aside from its declaration.
2938 This warning is enabled by @option{-Wall}.
2940 To suppress this warning use the @samp{unused} attribute
2941 (@pxref{Variable Attributes}).
2943 @item -Wunused-value
2944 @opindex Wunused-value
2945 Warn whenever a statement computes a result that is explicitly not
2946 used. To suppress this warning cast the unused expression to
2947 @samp{void}. This includes an expression-statement or the left-hand
2948 side of a comma expression that contains no side effects. For example,
2949 an expression such as @samp{x[i,j]} will cause a warning, while
2950 @samp{x[(void)i,j]} will not.
2952 This warning is enabled by @option{-Wall}.
2956 All the above @option{-Wunused} options combined.
2958 In order to get a warning about an unused function parameter, you must
2959 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2960 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2962 @item -Wuninitialized
2963 @opindex Wuninitialized
2964 Warn if an automatic variable is used without first being initialized or
2965 if a variable may be clobbered by a @code{setjmp} call.
2967 These warnings are possible only in optimizing compilation,
2968 because they require data flow information that is computed only
2969 when optimizing. If you do not specify @option{-O}, you will not get
2970 these warnings. Instead, GCC will issue a warning about @option{-Wuninitialized}
2971 requiring @option{-O}.
2973 If you want to warn about code which uses the uninitialized value of the
2974 variable in its own initializer, use the @option{-Winit-self} option.
2976 These warnings occur for individual uninitialized or clobbered
2977 elements of structure, union or array variables as well as for
2978 variables which are uninitialized or clobbered as a whole. They do
2979 not occur for variables or elements declared @code{volatile}. Because
2980 these warnings depend on optimization, the exact variables or elements
2981 for which there are warnings will depend on the precise optimization
2982 options and version of GCC used.
2984 Note that there may be no warning about a variable that is used only
2985 to compute a value that itself is never used, because such
2986 computations may be deleted by data flow analysis before the warnings
2989 These warnings are made optional because GCC is not smart
2990 enough to see all the reasons why the code might be correct
2991 despite appearing to have an error. Here is one example of how
3012 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3013 always initialized, but GCC doesn't know this. Here is
3014 another common case:
3019 if (change_y) save_y = y, y = new_y;
3021 if (change_y) y = save_y;
3026 This has no bug because @code{save_y} is used only if it is set.
3028 @cindex @code{longjmp} warnings
3029 This option also warns when a non-volatile automatic variable might be
3030 changed by a call to @code{longjmp}. These warnings as well are possible
3031 only in optimizing compilation.
3033 The compiler sees only the calls to @code{setjmp}. It cannot know
3034 where @code{longjmp} will be called; in fact, a signal handler could
3035 call it at any point in the code. As a result, you may get a warning
3036 even when there is in fact no problem because @code{longjmp} cannot
3037 in fact be called at the place which would cause a problem.
3039 Some spurious warnings can be avoided if you declare all the functions
3040 you use that never return as @code{noreturn}. @xref{Function
3043 This warning is enabled by @option{-Wall}.
3045 @item -Wunknown-pragmas
3046 @opindex Wunknown-pragmas
3047 @cindex warning for unknown pragmas
3048 @cindex unknown pragmas, warning
3049 @cindex pragmas, warning of unknown
3050 Warn when a #pragma directive is encountered which is not understood by
3051 GCC@. If this command line option is used, warnings will even be issued
3052 for unknown pragmas in system header files. This is not the case if
3053 the warnings were only enabled by the @option{-Wall} command line option.
3056 @opindex Wno-pragmas
3058 Do not warn about misuses of pragmas, such as incorrect parameters,
3059 invalid syntax, or conflicts between pragmas. See also
3060 @samp{-Wunknown-pragmas}.
3062 @item -Wstrict-aliasing
3063 @opindex Wstrict-aliasing
3064 This option is only active when @option{-fstrict-aliasing} is active.
3065 It warns about code which might break the strict aliasing rules that the
3066 compiler is using for optimization. The warning does not catch all
3067 cases, but does attempt to catch the more common pitfalls. It is
3068 included in @option{-Wall}.
3069 It is equivalent to -Wstrict-aliasing=3
3071 @item -Wstrict-aliasing=n
3072 @opindex Wstrict-aliasing=n
3073 This option is only active when @option{-fstrict-aliasing} is active.
3074 It warns about code which might break the strict aliasing rules that the
3075 compiler is using for optimization.
3076 Higher levels correspond to higher accuracy (fewer false positives).
3077 Higher levels also correspond to more effort, similar to the way -O works.
3078 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3081 Level 1: Most aggressive, quick, least accurate.
3082 Possibly useful when higher levels
3083 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3084 false negatives. However, it has many false positives.
3085 Warns for all pointer conversions between possibly incompatible types,
3086 even if never dereferenced. Runs in the frontend only.
3088 Level 2: Aggressive, quick, not too precise.
3089 May still have many false positives (not as many as level 1 though),
3090 and few false negatives (but possibly more than level 1).
3091 Unlike level 1, it only warns when an address is taken. Warns about
3092 incomplete types. Runs in the frontend only.
3094 Level 3 (default for @option{-Wstrict-aliasing}):
3095 Should have very few false positives and few false
3096 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3097 Takes care of the common punn+dereference pattern in the frontend:
3098 @code{*(int*)&some_float}.
3099 If optimization is enabled, it also runs in the backend, where it deals
3100 with multiple statement cases using flow-sensitive points-to information.
3101 Only warns when the converted pointer is dereferenced.
3102 Does not warn about incomplete types.
3104 @item -Wstrict-overflow
3105 @item -Wstrict-overflow=@var{n}
3106 @opindex Wstrict-overflow
3107 This option is only active when @option{-fstrict-overflow} is active.
3108 It warns about cases where the compiler optimizes based on the
3109 assumption that signed overflow does not occur. Note that it does not
3110 warn about all cases where the code might overflow: it only warns
3111 about cases where the compiler implements some optimization. Thus
3112 this warning depends on the optimization level.
3114 An optimization which assumes that signed overflow does not occur is
3115 perfectly safe if the values of the variables involved are such that
3116 overflow never does, in fact, occur. Therefore this warning can
3117 easily give a false positive: a warning about code which is not
3118 actually a problem. To help focus on important issues, several
3119 warning levels are defined. No warnings are issued for the use of
3120 undefined signed overflow when estimating how many iterations a loop
3121 will require, in particular when determining whether a loop will be
3125 @item -Wstrict-overflow=1
3126 Warn about cases which are both questionable and easy to avoid. For
3127 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3128 compiler will simplify this to @code{1}. This level of
3129 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3130 are not, and must be explicitly requested.
3132 @item -Wstrict-overflow=2
3133 Also warn about other cases where a comparison is simplified to a
3134 constant. For example: @code{abs (x) >= 0}. This can only be
3135 simplified when @option{-fstrict-overflow} is in effect, because
3136 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3137 zero. @option{-Wstrict-overflow} (with no level) is the same as
3138 @option{-Wstrict-overflow=2}.
3140 @item -Wstrict-overflow=3
3141 Also warn about other cases where a comparison is simplified. For
3142 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3144 @item -Wstrict-overflow=4
3145 Also warn about other simplifications not covered by the above cases.
3146 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3148 @item -Wstrict-overflow=5
3149 Also warn about cases where the compiler reduces the magnitude of a
3150 constant involved in a comparison. For example: @code{x + 2 > y} will
3151 be simplified to @code{x + 1 >= y}. This is reported only at the
3152 highest warning level because this simplification applies to many
3153 comparisons, so this warning level will give a very large number of
3157 @item -Warray-bounds
3158 @opindex Wno-array-bounds
3159 @opindex Warray-bounds
3160 This option is only active when @option{-ftree-vrp} is active
3161 (default for -O2 and above). It warns about subscripts to arrays
3162 that are always out of bounds. This warning is enabled by @option{-Wall}.
3166 All of the above @samp{-W} options combined. This enables all the
3167 warnings about constructions that some users consider questionable, and
3168 that are easy to avoid (or modify to prevent the warning), even in
3169 conjunction with macros. This also enables some language-specific
3170 warnings described in @ref{C++ Dialect Options} and
3171 @ref{Objective-C and Objective-C++ Dialect Options}.
3174 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
3175 Some of them warn about constructions that users generally do not
3176 consider questionable, but which occasionally you might wish to check
3177 for; others warn about constructions that are necessary or hard to avoid
3178 in some cases, and there is no simple way to modify the code to suppress
3185 (This option used to be called @option{-W}. The older name is still
3186 supported, but the newer name is more descriptive.) Print extra warning
3187 messages for these events:
3191 Warn if a comparison is always true or always false due to the limited
3192 range of the data type, but do not warn for constant expressions. For
3193 example, warn if an unsigned variable is compared against zero with
3194 @samp{<} or @samp{>=}. This warning can be independently controlled
3195 by @option{-Wtype-limits}.
3198 Storage-class specifiers like @code{static} are not the first things
3199 in a declaration. According to the C Standard, this usage is
3200 obsolescent. This warning can be independently controlled by
3201 @option{-Wold-style-declaration}.
3204 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
3208 A comparison between signed and unsigned values could produce an
3209 incorrect result when the signed value is converted to unsigned.
3210 (But don't warn if @option{-Wno-sign-compare} is also specified.)
3213 An aggregate has an initializer which does not initialize all members.
3214 This warning can be independently controlled by
3215 @option{-Wmissing-field-initializers}.
3218 An initialized field without side effects is overridden when using
3219 designated initializers (@pxref{Designated Inits, , Designated
3220 Initializers}). This warning can be independently controlled by
3221 @option{-Woverride-init}.
3224 A function parameter is declared without a type specifier in K&R-style
3225 functions. This warning can be independently controlled by
3226 @option{-Wmissing-parameter-type}.
3229 An empty body occurs in an @samp{if}, @samp{else} or
3230 @samp{do while} statement. This warning can be independently
3231 controlled by @option{-Wempty-body}.
3233 @item @r{(C++ only)}
3234 An empty body occurs in a @samp{while} or @samp{for} statement with no
3235 whitespacing before the semicolon. This warning can be independently
3236 controlled by @option{-Wempty-body}.
3239 A pointer is compared against integer zero with @samp{<}, @samp{<=},
3240 @samp{>}, or @samp{>=}.
3243 A variable might be changed by @samp{longjmp} or @samp{vfork}.
3244 This warning can be independently controlled by @option{-Wclobbered}.
3246 @item @r{(C++ only)}
3247 An enumerator and a non-enumerator both appear in a conditional expression.
3249 @item @r{(C++ only)}
3250 A non-static reference or non-static @samp{const} member appears in a
3251 class without constructors.
3253 @item @r{(C++ only)}
3254 Ambiguous virtual bases.
3256 @item @r{(C++ only)}
3257 Subscripting an array which has been declared @samp{register}.
3259 @item @r{(C++ only)}
3260 Taking the address of a variable which has been declared @samp{register}.
3262 @item @r{(C++ only)}
3263 A base class is not initialized in a derived class' copy constructor.
3266 @item -Wno-div-by-zero
3267 @opindex Wno-div-by-zero
3268 @opindex Wdiv-by-zero
3269 Do not warn about compile-time integer division by zero. Floating point
3270 division by zero is not warned about, as it can be a legitimate way of
3271 obtaining infinities and NaNs.
3273 @item -Wsystem-headers
3274 @opindex Wsystem-headers
3275 @cindex warnings from system headers
3276 @cindex system headers, warnings from
3277 Print warning messages for constructs found in system header files.
3278 Warnings from system headers are normally suppressed, on the assumption
3279 that they usually do not indicate real problems and would only make the
3280 compiler output harder to read. Using this command line option tells
3281 GCC to emit warnings from system headers as if they occurred in user
3282 code. However, note that using @option{-Wall} in conjunction with this
3283 option will @emph{not} warn about unknown pragmas in system
3284 headers---for that, @option{-Wunknown-pragmas} must also be used.
3287 @opindex Wfloat-equal
3288 Warn if floating point values are used in equality comparisons.
3290 The idea behind this is that sometimes it is convenient (for the
3291 programmer) to consider floating-point values as approximations to
3292 infinitely precise real numbers. If you are doing this, then you need
3293 to compute (by analyzing the code, or in some other way) the maximum or
3294 likely maximum error that the computation introduces, and allow for it
3295 when performing comparisons (and when producing output, but that's a
3296 different problem). In particular, instead of testing for equality, you
3297 would check to see whether the two values have ranges that overlap; and
3298 this is done with the relational operators, so equality comparisons are
3301 @item -Wtraditional @r{(C only)}
3302 @opindex Wtraditional
3303 Warn about certain constructs that behave differently in traditional and
3304 ISO C@. Also warn about ISO C constructs that have no traditional C
3305 equivalent, and/or problematic constructs which should be avoided.
3309 Macro parameters that appear within string literals in the macro body.
3310 In traditional C macro replacement takes place within string literals,
3311 but does not in ISO C@.
3314 In traditional C, some preprocessor directives did not exist.
3315 Traditional preprocessors would only consider a line to be a directive
3316 if the @samp{#} appeared in column 1 on the line. Therefore
3317 @option{-Wtraditional} warns about directives that traditional C
3318 understands but would ignore because the @samp{#} does not appear as the
3319 first character on the line. It also suggests you hide directives like
3320 @samp{#pragma} not understood by traditional C by indenting them. Some
3321 traditional implementations would not recognize @samp{#elif}, so it
3322 suggests avoiding it altogether.
3325 A function-like macro that appears without arguments.
3328 The unary plus operator.
3331 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3332 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3333 constants.) Note, these suffixes appear in macros defined in the system
3334 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3335 Use of these macros in user code might normally lead to spurious
3336 warnings, however GCC's integrated preprocessor has enough context to
3337 avoid warning in these cases.
3340 A function declared external in one block and then used after the end of
3344 A @code{switch} statement has an operand of type @code{long}.
3347 A non-@code{static} function declaration follows a @code{static} one.
3348 This construct is not accepted by some traditional C compilers.
3351 The ISO type of an integer constant has a different width or
3352 signedness from its traditional type. This warning is only issued if
3353 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3354 typically represent bit patterns, are not warned about.
3357 Usage of ISO string concatenation is detected.
3360 Initialization of automatic aggregates.
3363 Identifier conflicts with labels. Traditional C lacks a separate
3364 namespace for labels.
3367 Initialization of unions. If the initializer is zero, the warning is
3368 omitted. This is done under the assumption that the zero initializer in
3369 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3370 initializer warnings and relies on default initialization to zero in the
3374 Conversions by prototypes between fixed/floating point values and vice
3375 versa. The absence of these prototypes when compiling with traditional
3376 C would cause serious problems. This is a subset of the possible
3377 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3380 Use of ISO C style function definitions. This warning intentionally is
3381 @emph{not} issued for prototype declarations or variadic functions
3382 because these ISO C features will appear in your code when using
3383 libiberty's traditional C compatibility macros, @code{PARAMS} and
3384 @code{VPARAMS}. This warning is also bypassed for nested functions
3385 because that feature is already a GCC extension and thus not relevant to
3386 traditional C compatibility.
3389 @item -Wtraditional-conversion @r{(C only)}
3390 @opindex Wtraditional-conversion
3391 Warn if a prototype causes a type conversion that is different from what
3392 would happen to the same argument in the absence of a prototype. This
3393 includes conversions of fixed point to floating and vice versa, and
3394 conversions changing the width or signedness of a fixed point argument
3395 except when the same as the default promotion.
3397 @item -Wdeclaration-after-statement @r{(C only)}
3398 @opindex Wdeclaration-after-statement
3399 Warn when a declaration is found after a statement in a block. This
3400 construct, known from C++, was introduced with ISO C99 and is by default
3401 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3402 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3406 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3408 @item -Wno-endif-labels
3409 @opindex Wno-endif-labels
3410 @opindex Wendif-labels
3411 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3415 Warn whenever a local variable shadows another local variable, parameter or
3416 global variable or whenever a built-in function is shadowed.
3418 @item -Wlarger-than-@var{len}
3419 @opindex Wlarger-than
3420 Warn whenever an object of larger than @var{len} bytes is defined.
3422 @item -Wunsafe-loop-optimizations
3423 @opindex Wunsafe-loop-optimizations
3424 Warn if the loop cannot be optimized because the compiler could not
3425 assume anything on the bounds of the loop indices. With
3426 @option{-funsafe-loop-optimizations} warn if the compiler made
3429 @item -Wpointer-arith
3430 @opindex Wpointer-arith
3431 Warn about anything that depends on the ``size of'' a function type or
3432 of @code{void}. GNU C assigns these types a size of 1, for
3433 convenience in calculations with @code{void *} pointers and pointers
3434 to functions. In C++, warn also when an arithmetic operation involves
3435 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3438 @opindex Wtype-limits
3439 @opindex Wno-type-limits
3440 Warn if a comparison is always true or always false due to the limited
3441 range of the data type, but do not warn for constant expressions. For
3442 example, warn if an unsigned variable is compared against zero with
3443 @samp{<} or @samp{>=}. This warning is also enabled by
3446 @item -Wbad-function-cast @r{(C only)}
3447 @opindex Wbad-function-cast
3448 Warn whenever a function call is cast to a non-matching type.
3449 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3452 Warn about ISO C constructs that are outside of the common subset of
3453 ISO C and ISO C++, e.g.@: request for implicit conversion from
3454 @code{void *} to a pointer to non-@code{void} type.
3456 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3457 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3458 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3459 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3463 Warn whenever a pointer is cast so as to remove a type qualifier from
3464 the target type. For example, warn if a @code{const char *} is cast
3465 to an ordinary @code{char *}.
3468 @opindex Wcast-align
3469 Warn whenever a pointer is cast such that the required alignment of the
3470 target is increased. For example, warn if a @code{char *} is cast to
3471 an @code{int *} on machines where integers can only be accessed at
3472 two- or four-byte boundaries.
3474 @item -Wwrite-strings
3475 @opindex Wwrite-strings
3476 When compiling C, give string constants the type @code{const
3477 char[@var{length}]} so that
3478 copying the address of one into a non-@code{const} @code{char *}
3479 pointer will get a warning; when compiling C++, warn about the
3480 deprecated conversion from string literals to @code{char *}. This
3481 warning, by default, is enabled for C++ programs.
3482 These warnings will help you find at
3483 compile time code that can try to write into a string constant, but
3484 only if you have been very careful about using @code{const} in
3485 declarations and prototypes. Otherwise, it will just be a nuisance;
3486 this is why we did not make @option{-Wall} request these warnings.
3490 Warn for variables that might be changed by @samp{longjmp} or
3491 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3494 @opindex Wconversion
3495 @opindex Wno-conversion
3496 Warn for implicit conversions that may alter a value. This includes
3497 conversions between real and integer, like @code{abs (x)} when
3498 @code{x} is @code{double}; conversions between signed and unsigned,
3499 like @code{unsigned ui = -1}; and conversions to smaller types, like
3500 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3501 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3502 changed by the conversion like in @code{abs (2.0)}. Warnings about
3503 conversions between signed and unsigned integers can be disabled by
3504 using @option{-Wno-sign-conversion}.
3506 For C++, also warn for conversions between @code{NULL} and non-pointer
3507 types; confusing overload resolution for user-defined conversions; and
3508 conversions that will never use a type conversion operator:
3509 conversions to @code{void}, the same type, a base class or a reference
3510 to them. Warnings about conversions between signed and unsigned
3511 integers are disabled by default in C++ unless
3512 @option{-Wsign-conversion} is explicitly enabled.
3515 @opindex Wempty-body
3516 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3517 while} statement. Additionally, in C++, warn when an empty body occurs
3518 in a @samp{while} or @samp{for} statement with no whitespacing before
3519 the semicolon. This warning is also enabled by @option{-Wextra}.
3521 @item -Wsign-compare
3522 @opindex Wsign-compare
3523 @cindex warning for comparison of signed and unsigned values
3524 @cindex comparison of signed and unsigned values, warning
3525 @cindex signed and unsigned values, comparison warning
3526 Warn when a comparison between signed and unsigned values could produce
3527 an incorrect result when the signed value is converted to unsigned.
3528 This warning is also enabled by @option{-Wextra}; to get the other warnings
3529 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3531 @item -Wsign-conversion
3532 @opindex Wsign-conversion
3533 @opindex Wno-sign-conversion
3534 Warn for implicit conversions that may change the sign of an integer
3535 value, like assigning a signed integer expression to an unsigned
3536 integer variable. An explicit cast silences the warning. In C, this
3537 option is enabled also by @option{-Wconversion}.
3541 @opindex Wno-address
3542 Warn about suspicious uses of memory addresses. These include using
3543 the address of a function in a conditional expression, such as
3544 @code{void func(void); if (func)}, and comparisons against the memory
3545 address of a string literal, such as @code{if (x == "abc")}. Such
3546 uses typically indicate a programmer error: the address of a function
3547 always evaluates to true, so their use in a conditional usually
3548 indicate that the programmer forgot the parentheses in a function
3549 call; and comparisons against string literals result in unspecified
3550 behavior and are not portable in C, so they usually indicate that the
3551 programmer intended to use @code{strcmp}. This warning is enabled by
3555 @opindex Wlogical-op
3556 @opindex Wno-logical-op
3557 Warn about suspicious uses of logical operators in expressions.
3558 This includes using logical operators in contexts where a
3559 bit-wise operator is likely to be expected.
3561 @item -Waggregate-return
3562 @opindex Waggregate-return
3563 Warn if any functions that return structures or unions are defined or
3564 called. (In languages where you can return an array, this also elicits
3567 @item -Wno-attributes
3568 @opindex Wno-attributes
3569 @opindex Wattributes
3570 Do not warn if an unexpected @code{__attribute__} is used, such as
3571 unrecognized attributes, function attributes applied to variables,
3572 etc. This will not stop errors for incorrect use of supported
3575 @item -Wstrict-prototypes @r{(C only)}
3576 @opindex Wstrict-prototypes
3577 Warn if a function is declared or defined without specifying the
3578 argument types. (An old-style function definition is permitted without
3579 a warning if preceded by a declaration which specifies the argument
3582 @item -Wold-style-declaration @r{(C only)}
3583 @opindex Wold-style-declaration
3584 Warn for obsolescent usages, according to the C Standard, in a
3585 declaration. For example, warn if storage-class specifiers like
3586 @code{static} are not the first things in a declaration. This warning
3587 is also enabled by @option{-Wextra}.
3589 @item -Wold-style-definition @r{(C only)}
3590 @opindex Wold-style-definition
3591 Warn if an old-style function definition is used. A warning is given
3592 even if there is a previous prototype.
3594 @item -Wmissing-parameter-type @r{(C only)}
3595 @opindex Wmissing-parameter-type
3596 A function parameter is declared without a type specifier in K&R-style
3603 This warning is also enabled by @option{-Wextra}.
3605 @item -Wmissing-prototypes @r{(C only)}
3606 @opindex Wmissing-prototypes
3607 Warn if a global function is defined without a previous prototype
3608 declaration. This warning is issued even if the definition itself
3609 provides a prototype. The aim is to detect global functions that fail
3610 to be declared in header files.
3612 @item -Wmissing-declarations @r{(C and C++ only)}
3613 @opindex Wmissing-declarations
3614 Warn if a global function is defined without a previous declaration.
3615 Do so even if the definition itself provides a prototype.
3616 Use this option to detect global functions that are not declared in
3617 header files. In C++, no warnings are issued for function templates,
3618 or for inline functions, or for functions in anonymous namespaces.
3620 @item -Wmissing-field-initializers
3621 @opindex Wmissing-field-initializers
3624 Warn if a structure's initializer has some fields missing. For
3625 example, the following code would cause such a warning, because
3626 @code{x.h} is implicitly zero:
3629 struct s @{ int f, g, h; @};
3630 struct s x = @{ 3, 4 @};
3633 This option does not warn about designated initializers, so the following
3634 modification would not trigger a warning:
3637 struct s @{ int f, g, h; @};
3638 struct s x = @{ .f = 3, .g = 4 @};
3641 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3642 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3644 @item -Wmissing-noreturn
3645 @opindex Wmissing-noreturn
3646 Warn about functions which might be candidates for attribute @code{noreturn}.
3647 Note these are only possible candidates, not absolute ones. Care should
3648 be taken to manually verify functions actually do not ever return before
3649 adding the @code{noreturn} attribute, otherwise subtle code generation
3650 bugs could be introduced. You will not get a warning for @code{main} in
3651 hosted C environments.
3653 @item -Wmissing-format-attribute
3654 @opindex Wmissing-format-attribute
3656 Warn about function pointers which might be candidates for @code{format}
3657 attributes. Note these are only possible candidates, not absolute ones.
3658 GCC will guess that function pointers with @code{format} attributes that
3659 are used in assignment, initialization, parameter passing or return
3660 statements should have a corresponding @code{format} attribute in the
3661 resulting type. I.e.@: the left-hand side of the assignment or
3662 initialization, the type of the parameter variable, or the return type
3663 of the containing function respectively should also have a @code{format}
3664 attribute to avoid the warning.
3666 GCC will also warn about function definitions which might be
3667 candidates for @code{format} attributes. Again, these are only
3668 possible candidates. GCC will guess that @code{format} attributes
3669 might be appropriate for any function that calls a function like
3670 @code{vprintf} or @code{vscanf}, but this might not always be the
3671 case, and some functions for which @code{format} attributes are
3672 appropriate may not be detected.
3674 @item -Wno-multichar
3675 @opindex Wno-multichar
3677 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3678 Usually they indicate a typo in the user's code, as they have
3679 implementation-defined values, and should not be used in portable code.
3681 @item -Wnormalized=<none|id|nfc|nfkc>
3682 @opindex Wnormalized
3685 @cindex character set, input normalization
3686 In ISO C and ISO C++, two identifiers are different if they are
3687 different sequences of characters. However, sometimes when characters
3688 outside the basic ASCII character set are used, you can have two
3689 different character sequences that look the same. To avoid confusion,
3690 the ISO 10646 standard sets out some @dfn{normalization rules} which
3691 when applied ensure that two sequences that look the same are turned into
3692 the same sequence. GCC can warn you if you are using identifiers which
3693 have not been normalized; this option controls that warning.
3695 There are four levels of warning that GCC supports. The default is
3696 @option{-Wnormalized=nfc}, which warns about any identifier which is
3697 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3698 recommended form for most uses.
3700 Unfortunately, there are some characters which ISO C and ISO C++ allow
3701 in identifiers that when turned into NFC aren't allowable as
3702 identifiers. That is, there's no way to use these symbols in portable
3703 ISO C or C++ and have all your identifiers in NFC.
3704 @option{-Wnormalized=id} suppresses the warning for these characters.
3705 It is hoped that future versions of the standards involved will correct
3706 this, which is why this option is not the default.
3708 You can switch the warning off for all characters by writing
3709 @option{-Wnormalized=none}. You would only want to do this if you
3710 were using some other normalization scheme (like ``D''), because
3711 otherwise you can easily create bugs that are literally impossible to see.
3713 Some characters in ISO 10646 have distinct meanings but look identical
3714 in some fonts or display methodologies, especially once formatting has
3715 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3716 LETTER N'', will display just like a regular @code{n} which has been
3717 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3718 normalization scheme to convert all these into a standard form as
3719 well, and GCC will warn if your code is not in NFKC if you use
3720 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3721 about every identifier that contains the letter O because it might be
3722 confused with the digit 0, and so is not the default, but may be
3723 useful as a local coding convention if the programming environment is
3724 unable to be fixed to display these characters distinctly.
3726 @item -Wno-deprecated-declarations
3727 @opindex Wno-deprecated-declarations
3728 Do not warn about uses of functions (@pxref{Function Attributes}),
3729 variables (@pxref{Variable Attributes}), and types (@pxref{Type
3730 Attributes}) marked as deprecated by using the @code{deprecated}
3734 @opindex Wno-overflow
3735 Do not warn about compile-time overflow in constant expressions.
3737 @item -Woverride-init
3738 @opindex Woverride-init
3741 Warn if an initialized field without side effects is overridden when
3742 using designated initializers (@pxref{Designated Inits, , Designated
3745 This warning is included in @option{-Wextra}. To get other
3746 @option{-Wextra} warnings without this one, use @samp{-Wextra
3747 -Wno-override-init}.
3751 Warn if a structure is given the packed attribute, but the packed
3752 attribute has no effect on the layout or size of the structure.
3753 Such structures may be mis-aligned for little benefit. For
3754 instance, in this code, the variable @code{f.x} in @code{struct bar}
3755 will be misaligned even though @code{struct bar} does not itself
3756 have the packed attribute:
3763 @} __attribute__((packed));
3773 Warn if padding is included in a structure, either to align an element
3774 of the structure or to align the whole structure. Sometimes when this
3775 happens it is possible to rearrange the fields of the structure to
3776 reduce the padding and so make the structure smaller.
3778 @item -Wredundant-decls
3779 @opindex Wredundant-decls
3780 Warn if anything is declared more than once in the same scope, even in
3781 cases where multiple declaration is valid and changes nothing.
3783 @item -Wnested-externs @r{(C only)}
3784 @opindex Wnested-externs
3785 Warn if an @code{extern} declaration is encountered within a function.
3787 @item -Wunreachable-code
3788 @opindex Wunreachable-code
3789 Warn if the compiler detects that code will never be executed.
3791 This option is intended to warn when the compiler detects that at
3792 least a whole line of source code will never be executed, because
3793 some condition is never satisfied or because it is after a
3794 procedure that never returns.
3796 It is possible for this option to produce a warning even though there
3797 are circumstances under which part of the affected line can be executed,
3798 so care should be taken when removing apparently-unreachable code.
3800 For instance, when a function is inlined, a warning may mean that the
3801 line is unreachable in only one inlined copy of the function.
3803 This option is not made part of @option{-Wall} because in a debugging
3804 version of a program there is often substantial code which checks
3805 correct functioning of the program and is, hopefully, unreachable
3806 because the program does work. Another common use of unreachable
3807 code is to provide behavior which is selectable at compile-time.
3811 Warn if a function can not be inlined and it was declared as inline.
3812 Even with this option, the compiler will not warn about failures to
3813 inline functions declared in system headers.
3815 The compiler uses a variety of heuristics to determine whether or not
3816 to inline a function. For example, the compiler takes into account
3817 the size of the function being inlined and the amount of inlining
3818 that has already been done in the current function. Therefore,
3819 seemingly insignificant changes in the source program can cause the
3820 warnings produced by @option{-Winline} to appear or disappear.
3822 @item -Wno-invalid-offsetof @r{(C++ only)}
3823 @opindex Wno-invalid-offsetof
3824 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3825 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3826 to a non-POD type is undefined. In existing C++ implementations,
3827 however, @samp{offsetof} typically gives meaningful results even when
3828 applied to certain kinds of non-POD types. (Such as a simple
3829 @samp{struct} that fails to be a POD type only by virtue of having a
3830 constructor.) This flag is for users who are aware that they are
3831 writing nonportable code and who have deliberately chosen to ignore the
3834 The restrictions on @samp{offsetof} may be relaxed in a future version
3835 of the C++ standard.
3837 @item -Wno-int-to-pointer-cast @r{(C only)}
3838 @opindex Wno-int-to-pointer-cast
3839 Suppress warnings from casts to pointer type of an integer of a
3842 @item -Wno-pointer-to-int-cast @r{(C only)}
3843 @opindex Wno-pointer-to-int-cast
3844 Suppress warnings from casts from a pointer to an integer type of a
3848 @opindex Winvalid-pch
3849 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3850 the search path but can't be used.
3854 @opindex Wno-long-long
3855 Warn if @samp{long long} type is used. This is default. To inhibit
3856 the warning messages, use @option{-Wno-long-long}. Flags
3857 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3858 only when @option{-pedantic} flag is used.
3860 @item -Wvariadic-macros
3861 @opindex Wvariadic-macros
3862 @opindex Wno-variadic-macros
3863 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3864 alternate syntax when in pedantic ISO C99 mode. This is default.
3865 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3870 Warn if variable length array is used in the code.
3871 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
3872 the variable length array.
3874 @item -Wvolatile-register-var
3875 @opindex Wvolatile-register-var
3876 @opindex Wno-volatile-register-var
3877 Warn if a register variable is declared volatile. The volatile
3878 modifier does not inhibit all optimizations that may eliminate reads
3879 and/or writes to register variables.
3881 @item -Wdisabled-optimization
3882 @opindex Wdisabled-optimization
3883 Warn if a requested optimization pass is disabled. This warning does
3884 not generally indicate that there is anything wrong with your code; it
3885 merely indicates that GCC's optimizers were unable to handle the code
3886 effectively. Often, the problem is that your code is too big or too
3887 complex; GCC will refuse to optimize programs when the optimization
3888 itself is likely to take inordinate amounts of time.
3890 @item -Wpointer-sign
3891 @opindex Wpointer-sign
3892 @opindex Wno-pointer-sign
3893 Warn for pointer argument passing or assignment with different signedness.
3894 This option is only supported for C and Objective-C@. It is implied by
3895 @option{-Wall} and by @option{-pedantic}, which can be disabled with
3896 @option{-Wno-pointer-sign}.
3900 Make all warnings into errors.
3904 Make the specified warning into an errors. The specifier for a
3905 warning is appended, for example @option{-Werror=switch} turns the
3906 warnings controlled by @option{-Wswitch} into errors. This switch
3907 takes a negative form, to be used to negate @option{-Werror} for
3908 specific warnings, for example @option{-Wno-error=switch} makes
3909 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
3910 is in effect. You can use the @option{-fdiagnostics-show-option}
3911 option to have each controllable warning amended with the option which
3912 controls it, to determine what to use with this option.
3914 Note that specifying @option{-Werror=}@var{foo} automatically implies
3915 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
3918 @item -Wstack-protector
3919 @opindex Wstack-protector
3920 This option is only active when @option{-fstack-protector} is active. It
3921 warns about functions that will not be protected against stack smashing.
3923 @item -Woverlength-strings
3924 @opindex Woverlength-strings
3925 Warn about string constants which are longer than the ``minimum
3926 maximum'' length specified in the C standard. Modern compilers
3927 generally allow string constants which are much longer than the
3928 standard's minimum limit, but very portable programs should avoid
3929 using longer strings.
3931 The limit applies @emph{after} string constant concatenation, and does
3932 not count the trailing NUL@. In C89, the limit was 509 characters; in
3933 C99, it was raised to 4095. C++98 does not specify a normative
3934 minimum maximum, so we do not diagnose overlength strings in C++@.
3936 This option is implied by @option{-pedantic}, and can be disabled with
3937 @option{-Wno-overlength-strings}.
3940 @node Debugging Options
3941 @section Options for Debugging Your Program or GCC
3942 @cindex options, debugging
3943 @cindex debugging information options
3945 GCC has various special options that are used for debugging
3946 either your program or GCC:
3951 Produce debugging information in the operating system's native format
3952 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
3955 On most systems that use stabs format, @option{-g} enables use of extra
3956 debugging information that only GDB can use; this extra information
3957 makes debugging work better in GDB but will probably make other debuggers
3959 refuse to read the program. If you want to control for certain whether
3960 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3961 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3963 GCC allows you to use @option{-g} with
3964 @option{-O}. The shortcuts taken by optimized code may occasionally
3965 produce surprising results: some variables you declared may not exist
3966 at all; flow of control may briefly move where you did not expect it;
3967 some statements may not be executed because they compute constant
3968 results or their values were already at hand; some statements may
3969 execute in different places because they were moved out of loops.
3971 Nevertheless it proves possible to debug optimized output. This makes
3972 it reasonable to use the optimizer for programs that might have bugs.
3974 The following options are useful when GCC is generated with the
3975 capability for more than one debugging format.
3979 Produce debugging information for use by GDB@. This means to use the
3980 most expressive format available (DWARF 2, stabs, or the native format
3981 if neither of those are supported), including GDB extensions if at all
3986 Produce debugging information in stabs format (if that is supported),
3987 without GDB extensions. This is the format used by DBX on most BSD
3988 systems. On MIPS, Alpha and System V Release 4 systems this option
3989 produces stabs debugging output which is not understood by DBX or SDB@.
3990 On System V Release 4 systems this option requires the GNU assembler.
3992 @item -feliminate-unused-debug-symbols
3993 @opindex feliminate-unused-debug-symbols
3994 Produce debugging information in stabs format (if that is supported),
3995 for only symbols that are actually used.
3997 @item -femit-class-debug-always
3998 Instead of emitting debugging information for a C++ class in only one
3999 object file, emit it in all object files using the class. This option
4000 should be used only with debuggers that are unable to handle the way GCC
4001 normally emits debugging information for classes because using this
4002 option will increase the size of debugging information by as much as a
4007 Produce debugging information in stabs format (if that is supported),
4008 using GNU extensions understood only by the GNU debugger (GDB)@. The
4009 use of these extensions is likely to make other debuggers crash or
4010 refuse to read the program.
4014 Produce debugging information in COFF format (if that is supported).
4015 This is the format used by SDB on most System V systems prior to
4020 Produce debugging information in XCOFF format (if that is supported).
4021 This is the format used by the DBX debugger on IBM RS/6000 systems.
4025 Produce debugging information in XCOFF format (if that is supported),
4026 using GNU extensions understood only by the GNU debugger (GDB)@. The
4027 use of these extensions is likely to make other debuggers crash or
4028 refuse to read the program, and may cause assemblers other than the GNU
4029 assembler (GAS) to fail with an error.
4033 Produce debugging information in DWARF version 2 format (if that is
4034 supported). This is the format used by DBX on IRIX 6. With this
4035 option, GCC uses features of DWARF version 3 when they are useful;
4036 version 3 is upward compatible with version 2, but may still cause
4037 problems for older debuggers.
4041 Produce debugging information in VMS debug format (if that is
4042 supported). This is the format used by DEBUG on VMS systems.
4045 @itemx -ggdb@var{level}
4046 @itemx -gstabs@var{level}
4047 @itemx -gcoff@var{level}
4048 @itemx -gxcoff@var{level}
4049 @itemx -gvms@var{level}
4050 Request debugging information and also use @var{level} to specify how
4051 much information. The default level is 2.
4053 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4056 Level 1 produces minimal information, enough for making backtraces in
4057 parts of the program that you don't plan to debug. This includes
4058 descriptions of functions and external variables, but no information
4059 about local variables and no line numbers.
4061 Level 3 includes extra information, such as all the macro definitions
4062 present in the program. Some debuggers support macro expansion when
4063 you use @option{-g3}.
4065 @option{-gdwarf-2} does not accept a concatenated debug level, because
4066 GCC used to support an option @option{-gdwarf} that meant to generate
4067 debug information in version 1 of the DWARF format (which is very
4068 different from version 2), and it would have been too confusing. That
4069 debug format is long obsolete, but the option cannot be changed now.
4070 Instead use an additional @option{-g@var{level}} option to change the
4071 debug level for DWARF2.
4073 @item -feliminate-dwarf2-dups
4074 @opindex feliminate-dwarf2-dups
4075 Compress DWARF2 debugging information by eliminating duplicated
4076 information about each symbol. This option only makes sense when
4077 generating DWARF2 debugging information with @option{-gdwarf-2}.
4079 @item -femit-struct-debug-baseonly
4080 Emit debug information for struct-like types
4081 only when the base name of the compilation source file
4082 matches the base name of file in which the struct was defined.
4084 This option substantially reduces the size of debugging information,
4085 but at significant potential loss in type information to the debugger.
4086 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4087 See @option{-femit-struct-debug-detailed} for more detailed control.
4089 This option works only with DWARF 2.
4091 @item -femit-struct-debug-reduced
4092 Emit debug information for struct-like types
4093 only when the base name of the compilation source file
4094 matches the base name of file in which the type was defined,
4095 unless the struct is a template or defined in a system header.
4097 This option significantly reduces the size of debugging information,
4098 with some potential loss in type information to the debugger.
4099 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4100 See @option{-femit-struct-debug-detailed} for more detailed control.
4102 This option works only with DWARF 2.
4104 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4105 Specify the struct-like types
4106 for which the compiler will generate debug information.
4107 The intent is to reduce duplicate struct debug information
4108 between different object files within the same program.
4110 This option is a detailed version of
4111 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4112 which will serve for most needs.
4114 A specification has the syntax
4115 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4117 The optional first word limits the specification to
4118 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4119 A struct type is used directly when it is the type of a variable, member.
4120 Indirect uses arise through pointers to structs.
4121 That is, when use of an incomplete struct would be legal, the use is indirect.
4123 @samp{struct one direct; struct two * indirect;}.
4125 The optional second word limits the specification to
4126 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4127 Generic structs are a bit complicated to explain.
4128 For C++, these are non-explicit specializations of template classes,
4129 or non-template classes within the above.
4130 Other programming languages have generics,
4131 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4133 The third word specifies the source files for those
4134 structs for which the compiler will emit debug information.
4135 The values @samp{none} and @samp{any} have the normal meaning.
4136 The value @samp{base} means that
4137 the base of name of the file in which the type declaration appears
4138 must match the base of the name of the main compilation file.
4139 In practice, this means that
4140 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4141 but types declared in other header will not.
4142 The value @samp{sys} means those types satisfying @samp{base}
4143 or declared in system or compiler headers.
4145 You may need to experiment to determine the best settings for your application.
4147 The default is @samp{-femit-struct-debug-detailed=all}.
4149 This option works only with DWARF 2.
4151 @item -fdebug-prefix-map=@var{old}=@var{new}
4152 @opindex fdebug-prefix-map
4153 When compiling files in directory @file{@var{old}}, record debugging
4154 information describing them as in @file{@var{new}} instead.
4156 @cindex @command{prof}
4159 Generate extra code to write profile information suitable for the
4160 analysis program @command{prof}. You must use this option when compiling
4161 the source files you want data about, and you must also use it when
4164 @cindex @command{gprof}
4167 Generate extra code to write profile information suitable for the
4168 analysis program @command{gprof}. You must use this option when compiling
4169 the source files you want data about, and you must also use it when
4174 Makes the compiler print out each function name as it is compiled, and
4175 print some statistics about each pass when it finishes.
4178 @opindex ftime-report
4179 Makes the compiler print some statistics about the time consumed by each
4180 pass when it finishes.
4183 @opindex fmem-report
4184 Makes the compiler print some statistics about permanent memory
4185 allocation when it finishes.
4187 @item -fpre-ipa-mem-report
4188 @opindex fpre-ipa-mem-report
4189 @item -fpost-ipa-mem-report
4190 @opindex fpost-ipa-mem-report
4191 Makes the compiler print some statistics about permanent memory
4192 allocation before or after interprocedural optimization.
4194 @item -fprofile-arcs
4195 @opindex fprofile-arcs
4196 Add code so that program flow @dfn{arcs} are instrumented. During
4197 execution the program records how many times each branch and call is
4198 executed and how many times it is taken or returns. When the compiled
4199 program exits it saves this data to a file called
4200 @file{@var{auxname}.gcda} for each source file. The data may be used for
4201 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4202 test coverage analysis (@option{-ftest-coverage}). Each object file's
4203 @var{auxname} is generated from the name of the output file, if
4204 explicitly specified and it is not the final executable, otherwise it is
4205 the basename of the source file. In both cases any suffix is removed
4206 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4207 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4208 @xref{Cross-profiling}.
4210 @cindex @command{gcov}
4214 This option is used to compile and link code instrumented for coverage
4215 analysis. The option is a synonym for @option{-fprofile-arcs}
4216 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4217 linking). See the documentation for those options for more details.
4222 Compile the source files with @option{-fprofile-arcs} plus optimization
4223 and code generation options. For test coverage analysis, use the
4224 additional @option{-ftest-coverage} option. You do not need to profile
4225 every source file in a program.
4228 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4229 (the latter implies the former).
4232 Run the program on a representative workload to generate the arc profile
4233 information. This may be repeated any number of times. You can run
4234 concurrent instances of your program, and provided that the file system
4235 supports locking, the data files will be correctly updated. Also
4236 @code{fork} calls are detected and correctly handled (double counting
4240 For profile-directed optimizations, compile the source files again with
4241 the same optimization and code generation options plus
4242 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4243 Control Optimization}).
4246 For test coverage analysis, use @command{gcov} to produce human readable
4247 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4248 @command{gcov} documentation for further information.
4252 With @option{-fprofile-arcs}, for each function of your program GCC
4253 creates a program flow graph, then finds a spanning tree for the graph.
4254 Only arcs that are not on the spanning tree have to be instrumented: the
4255 compiler adds code to count the number of times that these arcs are
4256 executed. When an arc is the only exit or only entrance to a block, the
4257 instrumentation code can be added to the block; otherwise, a new basic
4258 block must be created to hold the instrumentation code.
4261 @item -ftest-coverage
4262 @opindex ftest-coverage
4263 Produce a notes file that the @command{gcov} code-coverage utility
4264 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4265 show program coverage. Each source file's note file is called
4266 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4267 above for a description of @var{auxname} and instructions on how to
4268 generate test coverage data. Coverage data will match the source files
4269 more closely, if you do not optimize.
4271 @item -fdbg-cnt-list
4272 @opindex fdbg-cnt-list
4273 Print the name and the counter upperbound for all debug counters.
4275 @item -fdbg-cnt=@var{counter-value-list}
4277 Set the internal debug counter upperbound. @var{counter-value-list}
4278 is a comma-separated list of @var{name}:@var{value} pairs
4279 which sets the upperbound of each debug counter @var{name} to @var{value}.
4280 All debug counters have the initial upperbound of @var{UINT_MAX},
4281 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4282 e.g. With -fdbg-cnt=dce:10,tail_call:0
4283 dbg_cnt(dce) will return true only for first 10 invocations
4284 and dbg_cnt(tail_call) will return false always.
4286 @item -d@var{letters}
4287 @item -fdump-rtl-@var{pass}
4289 Says to make debugging dumps during compilation at times specified by
4290 @var{letters}. This is used for debugging the RTL-based passes of the
4291 compiler. The file names for most of the dumps are made by appending a
4292 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
4293 from the name of the output file, if explicitly specified and it is not
4294 an executable, otherwise it is the basename of the source file. These
4295 switches may have different effects when @option{-E} is used for
4298 Most debug dumps can be enabled either passing a letter to the @option{-d}
4299 option, or with a long @option{-fdump-rtl} switch; here are the possible
4300 letters for use in @var{letters} and @var{pass}, and their meanings:
4305 Annotate the assembler output with miscellaneous debugging information.
4308 @itemx -fdump-rtl-bbro
4310 @opindex fdump-rtl-bbro
4311 Dump after block reordering, to @file{@var{file}.148r.bbro}.
4314 @itemx -fdump-rtl-combine
4316 @opindex fdump-rtl-combine
4317 Dump after the RTL instruction combination pass, to the file
4318 @file{@var{file}.129r.combine}.
4321 @itemx -fdump-rtl-ce1
4322 @itemx -fdump-rtl-ce2
4324 @opindex fdump-rtl-ce1
4325 @opindex fdump-rtl-ce2
4326 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
4327 first if conversion, to the file @file{@var{file}.117r.ce1}. @option{-dC}
4328 and @option{-fdump-rtl-ce2} enable dumping after the second if
4329 conversion, to the file @file{@var{file}.130r.ce2}.
4332 @itemx -fdump-rtl-btl
4333 @itemx -fdump-rtl-dbr
4335 @opindex fdump-rtl-btl
4336 @opindex fdump-rtl-dbr
4337 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
4338 target load optimization, to @file{@var{file}.31.btl}. @option{-dd}
4339 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
4340 scheduling, to @file{@var{file}.36.dbr}.
4344 Dump all macro definitions, at the end of preprocessing, in addition to
4348 @itemx -fdump-rtl-ce3
4350 @opindex fdump-rtl-ce3
4351 Dump after the third if conversion, to @file{@var{file}.146r.ce3}.
4354 @itemx -fdump-rtl-cfg
4355 @itemx -fdump-rtl-life
4357 @opindex fdump-rtl-cfg
4358 @opindex fdump-rtl-life
4359 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
4360 and data flow analysis, to @file{@var{file}.116r.cfg}. @option{-df}
4361 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
4362 to @file{@var{file}.128r.life1} and @file{@var{file}.135r.life2}.
4365 @itemx -fdump-rtl-greg
4367 @opindex fdump-rtl-greg
4368 Dump after global register allocation, to @file{@var{file}.139r.greg}.
4371 @itemx -fdump-rtl-gcse
4372 @itemx -fdump-rtl-bypass
4374 @opindex fdump-rtl-gcse
4375 @opindex fdump-rtl-bypass
4376 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
4377 @file{@var{file}.114r.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
4378 enable dumping after jump bypassing and control flow optimizations, to
4379 @file{@var{file}.115r.bypass}.
4382 @itemx -fdump-rtl-eh
4384 @opindex fdump-rtl-eh
4385 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
4388 @itemx -fdump-rtl-sibling
4390 @opindex fdump-rtl-sibling
4391 Dump after sibling call optimizations, to @file{@var{file}.106r.sibling}.
4394 @itemx -fdump-rtl-jump
4396 @opindex fdump-rtl-jump
4397 Dump after the first jump optimization, to @file{@var{file}.112r.jump}.
4400 @itemx -fdump-rtl-stack
4402 @opindex fdump-rtl-stack
4403 Dump after conversion from GCC's "flat register file" registers to the
4404 x87's stack-like registers, to @file{@var{file}.152r.stack}.
4407 @itemx -fdump-rtl-lreg
4409 @opindex fdump-rtl-lreg
4410 Dump after local register allocation, to @file{@var{file}.138r.lreg}.
4413 @itemx -fdump-rtl-loop2
4415 @opindex fdump-rtl-loop2
4416 @option{-dL} and @option{-fdump-rtl-loop2} enable dumping after the
4417 loop optimization pass, to @file{@var{file}.119r.loop2},
4418 @file{@var{file}.120r.loop2_init},
4419 @file{@var{file}.121r.loop2_invariant}, and
4420 @file{@var{file}.125r.loop2_done}.
4423 @itemx -fdump-rtl-sms
4425 @opindex fdump-rtl-sms
4426 Dump after modulo scheduling, to @file{@var{file}.136r.sms}.
4429 @itemx -fdump-rtl-mach
4431 @opindex fdump-rtl-mach
4432 Dump after performing the machine dependent reorganization pass, to
4433 @file{@var{file}.155r.mach} if that pass exists.
4436 @itemx -fdump-rtl-rnreg
4438 @opindex fdump-rtl-rnreg
4439 Dump after register renumbering, to @file{@var{file}.147r.rnreg}.
4442 @itemx -fdump-rtl-regmove
4444 @opindex fdump-rtl-regmove
4445 Dump after the register move pass, to @file{@var{file}.132r.regmove}.
4448 @itemx -fdump-rtl-postreload
4450 @opindex fdump-rtl-postreload
4451 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
4454 @itemx -fdump-rtl-expand
4456 @opindex fdump-rtl-expand
4457 Dump after RTL generation, to @file{@var{file}.104r.expand}.
4460 @itemx -fdump-rtl-sched2
4462 @opindex fdump-rtl-sched2
4463 Dump after the second scheduling pass, to @file{@var{file}.149r.sched2}.
4466 @itemx -fdump-rtl-cse
4468 @opindex fdump-rtl-cse
4469 Dump after CSE (including the jump optimization that sometimes follows
4470 CSE), to @file{@var{file}.113r.cse}.
4473 @itemx -fdump-rtl-sched1
4475 @opindex fdump-rtl-sched1
4476 Dump after the first scheduling pass, to @file{@var{file}.136r.sched1}.
4479 @itemx -fdump-rtl-cse2
4481 @opindex fdump-rtl-cse2
4482 Dump after the second CSE pass (including the jump optimization that
4483 sometimes follows CSE), to @file{@var{file}.127r.cse2}.
4486 @itemx -fdump-rtl-tracer
4488 @opindex fdump-rtl-tracer
4489 Dump after running tracer, to @file{@var{file}.118r.tracer}.
4492 @itemx -fdump-rtl-vpt
4493 @itemx -fdump-rtl-vartrack
4495 @opindex fdump-rtl-vpt
4496 @opindex fdump-rtl-vartrack
4497 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
4498 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
4499 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
4500 to @file{@var{file}.154r.vartrack}.
4503 @itemx -fdump-rtl-flow2
4505 @opindex fdump-rtl-flow2
4506 Dump after the second flow pass, to @file{@var{file}.142r.flow2}.
4509 @itemx -fdump-rtl-peephole2
4511 @opindex fdump-rtl-peephole2
4512 Dump after the peephole pass, to @file{@var{file}.145r.peephole2}.
4515 @itemx -fdump-rtl-web
4517 @opindex fdump-rtl-web
4518 Dump after live range splitting, to @file{@var{file}.126r.web}.
4521 @itemx -fdump-rtl-all
4523 @opindex fdump-rtl-all
4524 Produce all the dumps listed above.
4528 Produce a core dump whenever an error occurs.
4532 Print statistics on memory usage, at the end of the run, to
4537 Annotate the assembler output with a comment indicating which
4538 pattern and alternative was used. The length of each instruction is
4543 Dump the RTL in the assembler output as a comment before each instruction.
4544 Also turns on @option{-dp} annotation.
4548 For each of the other indicated dump files (either with @option{-d} or
4549 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
4550 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
4554 Just generate RTL for a function instead of compiling it. Usually used
4555 with @samp{r} (@option{-fdump-rtl-expand}).
4559 Dump debugging information during parsing, to standard error.
4563 @opindex fdump-noaddr
4564 When doing debugging dumps (see @option{-d} option above), suppress
4565 address output. This makes it more feasible to use diff on debugging
4566 dumps for compiler invocations with different compiler binaries and/or
4567 different text / bss / data / heap / stack / dso start locations.
4569 @item -fdump-unnumbered
4570 @opindex fdump-unnumbered
4571 When doing debugging dumps (see @option{-d} option above), suppress instruction
4572 numbers and address output. This makes it more feasible to
4573 use diff on debugging dumps for compiler invocations with different
4574 options, in particular with and without @option{-g}.
4576 @item -fdump-translation-unit @r{(C++ only)}
4577 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
4578 @opindex fdump-translation-unit
4579 Dump a representation of the tree structure for the entire translation
4580 unit to a file. The file name is made by appending @file{.tu} to the
4581 source file name. If the @samp{-@var{options}} form is used, @var{options}
4582 controls the details of the dump as described for the
4583 @option{-fdump-tree} options.
4585 @item -fdump-class-hierarchy @r{(C++ only)}
4586 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
4587 @opindex fdump-class-hierarchy
4588 Dump a representation of each class's hierarchy and virtual function
4589 table layout to a file. The file name is made by appending @file{.class}
4590 to the source file name. If the @samp{-@var{options}} form is used,
4591 @var{options} controls the details of the dump as described for the
4592 @option{-fdump-tree} options.
4594 @item -fdump-ipa-@var{switch}
4596 Control the dumping at various stages of inter-procedural analysis
4597 language tree to a file. The file name is generated by appending a switch
4598 specific suffix to the source file name. The following dumps are possible:
4602 Enables all inter-procedural analysis dumps; currently the only produced
4603 dump is the @samp{cgraph} dump.
4606 Dumps information about call-graph optimization, unused function removal,
4607 and inlining decisions.
4610 @item -fdump-tree-@var{switch}
4611 @itemx -fdump-tree-@var{switch}-@var{options}
4613 Control the dumping at various stages of processing the intermediate
4614 language tree to a file. The file name is generated by appending a switch
4615 specific suffix to the source file name. If the @samp{-@var{options}}
4616 form is used, @var{options} is a list of @samp{-} separated options that
4617 control the details of the dump. Not all options are applicable to all
4618 dumps, those which are not meaningful will be ignored. The following
4619 options are available
4623 Print the address of each node. Usually this is not meaningful as it
4624 changes according to the environment and source file. Its primary use
4625 is for tying up a dump file with a debug environment.
4627 Inhibit dumping of members of a scope or body of a function merely
4628 because that scope has been reached. Only dump such items when they
4629 are directly reachable by some other path. When dumping pretty-printed
4630 trees, this option inhibits dumping the bodies of control structures.
4632 Print a raw representation of the tree. By default, trees are
4633 pretty-printed into a C-like representation.
4635 Enable more detailed dumps (not honored by every dump option).
4637 Enable dumping various statistics about the pass (not honored by every dump
4640 Enable showing basic block boundaries (disabled in raw dumps).
4642 Enable showing virtual operands for every statement.
4644 Enable showing line numbers for statements.
4646 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4648 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4651 The following tree dumps are possible:
4655 Dump before any tree based optimization, to @file{@var{file}.original}.
4658 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4661 Dump after function inlining, to @file{@var{file}.inlined}.
4664 @opindex fdump-tree-gimple
4665 Dump each function before and after the gimplification pass to a file. The
4666 file name is made by appending @file{.gimple} to the source file name.
4669 @opindex fdump-tree-cfg
4670 Dump the control flow graph of each function to a file. The file name is
4671 made by appending @file{.cfg} to the source file name.
4674 @opindex fdump-tree-vcg
4675 Dump the control flow graph of each function to a file in VCG format. The
4676 file name is made by appending @file{.vcg} to the source file name. Note
4677 that if the file contains more than one function, the generated file cannot
4678 be used directly by VCG@. You will need to cut and paste each function's
4679 graph into its own separate file first.
4682 @opindex fdump-tree-ch
4683 Dump each function after copying loop headers. The file name is made by
4684 appending @file{.ch} to the source file name.
4687 @opindex fdump-tree-ssa
4688 Dump SSA related information to a file. The file name is made by appending
4689 @file{.ssa} to the source file name.
4692 @opindex fdump-tree-salias
4693 Dump structure aliasing variable information to a file. This file name
4694 is made by appending @file{.salias} to the source file name.
4697 @opindex fdump-tree-alias
4698 Dump aliasing information for each function. The file name is made by
4699 appending @file{.alias} to the source file name.
4702 @opindex fdump-tree-ccp
4703 Dump each function after CCP@. The file name is made by appending
4704 @file{.ccp} to the source file name.
4707 @opindex fdump-tree-storeccp
4708 Dump each function after STORE-CCP. The file name is made by appending
4709 @file{.storeccp} to the source file name.
4712 @opindex fdump-tree-pre
4713 Dump trees after partial redundancy elimination. The file name is made
4714 by appending @file{.pre} to the source file name.
4717 @opindex fdump-tree-fre
4718 Dump trees after full redundancy elimination. The file name is made
4719 by appending @file{.fre} to the source file name.
4722 @opindex fdump-tree-copyprop
4723 Dump trees after copy propagation. The file name is made
4724 by appending @file{.copyprop} to the source file name.
4726 @item store_copyprop
4727 @opindex fdump-tree-store_copyprop
4728 Dump trees after store copy-propagation. The file name is made
4729 by appending @file{.store_copyprop} to the source file name.
4732 @opindex fdump-tree-dce
4733 Dump each function after dead code elimination. The file name is made by
4734 appending @file{.dce} to the source file name.
4737 @opindex fdump-tree-mudflap
4738 Dump each function after adding mudflap instrumentation. The file name is
4739 made by appending @file{.mudflap} to the source file name.
4742 @opindex fdump-tree-sra
4743 Dump each function after performing scalar replacement of aggregates. The
4744 file name is made by appending @file{.sra} to the source file name.
4747 @opindex fdump-tree-sink
4748 Dump each function after performing code sinking. The file name is made
4749 by appending @file{.sink} to the source file name.
4752 @opindex fdump-tree-dom
4753 Dump each function after applying dominator tree optimizations. The file
4754 name is made by appending @file{.dom} to the source file name.
4757 @opindex fdump-tree-dse
4758 Dump each function after applying dead store elimination. The file
4759 name is made by appending @file{.dse} to the source file name.
4762 @opindex fdump-tree-phiopt
4763 Dump each function after optimizing PHI nodes into straightline code. The file
4764 name is made by appending @file{.phiopt} to the source file name.
4767 @opindex fdump-tree-forwprop
4768 Dump each function after forward propagating single use variables. The file
4769 name is made by appending @file{.forwprop} to the source file name.
4772 @opindex fdump-tree-copyrename
4773 Dump each function after applying the copy rename optimization. The file
4774 name is made by appending @file{.copyrename} to the source file name.
4777 @opindex fdump-tree-nrv
4778 Dump each function after applying the named return value optimization on
4779 generic trees. The file name is made by appending @file{.nrv} to the source
4783 @opindex fdump-tree-vect
4784 Dump each function after applying vectorization of loops. The file name is
4785 made by appending @file{.vect} to the source file name.
4788 @opindex fdump-tree-vrp
4789 Dump each function after Value Range Propagation (VRP). The file name
4790 is made by appending @file{.vrp} to the source file name.
4793 @opindex fdump-tree-all
4794 Enable all the available tree dumps with the flags provided in this option.
4797 @item -ftree-vectorizer-verbose=@var{n}
4798 @opindex ftree-vectorizer-verbose
4799 This option controls the amount of debugging output the vectorizer prints.
4800 This information is written to standard error, unless
4801 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
4802 in which case it is output to the usual dump listing file, @file{.vect}.
4803 For @var{n}=0 no diagnostic information is reported.
4804 If @var{n}=1 the vectorizer reports each loop that got vectorized,
4805 and the total number of loops that got vectorized.
4806 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
4807 the first analysis phase (vect_analyze_loop_form) - i.e. countable,
4808 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
4809 level that @option{-fdump-tree-vect-stats} uses.
4810 Higher verbosity levels mean either more information dumped for each
4811 reported loop, or same amount of information reported for more loops:
4812 If @var{n}=3, alignment related information is added to the reports.
4813 If @var{n}=4, data-references related information (e.g. memory dependences,
4814 memory access-patterns) is added to the reports.
4815 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
4816 that did not pass the first analysis phase (i.e. may not be countable, or
4817 may have complicated control-flow).
4818 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
4819 For @var{n}=7, all the information the vectorizer generates during its
4820 analysis and transformation is reported. This is the same verbosity level
4821 that @option{-fdump-tree-vect-details} uses.
4823 @item -frandom-seed=@var{string}
4824 @opindex frandom-string
4825 This option provides a seed that GCC uses when it would otherwise use
4826 random numbers. It is used to generate certain symbol names
4827 that have to be different in every compiled file. It is also used to
4828 place unique stamps in coverage data files and the object files that
4829 produce them. You can use the @option{-frandom-seed} option to produce
4830 reproducibly identical object files.
4832 The @var{string} should be different for every file you compile.
4834 @item -fsched-verbose=@var{n}
4835 @opindex fsched-verbose
4836 On targets that use instruction scheduling, this option controls the
4837 amount of debugging output the scheduler prints. This information is
4838 written to standard error, unless @option{-dS} or @option{-dR} is
4839 specified, in which case it is output to the usual dump
4840 listing file, @file{.sched} or @file{.sched2} respectively. However
4841 for @var{n} greater than nine, the output is always printed to standard
4844 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4845 same information as @option{-dRS}. For @var{n} greater than one, it
4846 also output basic block probabilities, detailed ready list information
4847 and unit/insn info. For @var{n} greater than two, it includes RTL
4848 at abort point, control-flow and regions info. And for @var{n} over
4849 four, @option{-fsched-verbose} also includes dependence info.
4853 Store the usual ``temporary'' intermediate files permanently; place them
4854 in the current directory and name them based on the source file. Thus,
4855 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4856 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4857 preprocessed @file{foo.i} output file even though the compiler now
4858 normally uses an integrated preprocessor.
4860 When used in combination with the @option{-x} command line option,
4861 @option{-save-temps} is sensible enough to avoid over writing an
4862 input source file with the same extension as an intermediate file.
4863 The corresponding intermediate file may be obtained by renaming the
4864 source file before using @option{-save-temps}.
4868 Report the CPU time taken by each subprocess in the compilation
4869 sequence. For C source files, this is the compiler proper and assembler
4870 (plus the linker if linking is done). The output looks like this:
4877 The first number on each line is the ``user time'', that is time spent
4878 executing the program itself. The second number is ``system time'',
4879 time spent executing operating system routines on behalf of the program.
4880 Both numbers are in seconds.
4882 @item -fvar-tracking
4883 @opindex fvar-tracking
4884 Run variable tracking pass. It computes where variables are stored at each
4885 position in code. Better debugging information is then generated
4886 (if the debugging information format supports this information).
4888 It is enabled by default when compiling with optimization (@option{-Os},
4889 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
4890 the debug info format supports it.
4892 @item -print-file-name=@var{library}
4893 @opindex print-file-name
4894 Print the full absolute name of the library file @var{library} that
4895 would be used when linking---and don't do anything else. With this
4896 option, GCC does not compile or link anything; it just prints the
4899 @item -print-multi-directory
4900 @opindex print-multi-directory
4901 Print the directory name corresponding to the multilib selected by any
4902 other switches present in the command line. This directory is supposed
4903 to exist in @env{GCC_EXEC_PREFIX}.
4905 @item -print-multi-lib
4906 @opindex print-multi-lib
4907 Print the mapping from multilib directory names to compiler switches
4908 that enable them. The directory name is separated from the switches by
4909 @samp{;}, and each switch starts with an @samp{@@} instead of the
4910 @samp{-}, without spaces between multiple switches. This is supposed to
4911 ease shell-processing.
4913 @item -print-prog-name=@var{program}
4914 @opindex print-prog-name
4915 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
4917 @item -print-libgcc-file-name
4918 @opindex print-libgcc-file-name
4919 Same as @option{-print-file-name=libgcc.a}.
4921 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
4922 but you do want to link with @file{libgcc.a}. You can do
4925 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
4928 @item -print-search-dirs
4929 @opindex print-search-dirs
4930 Print the name of the configured installation directory and a list of
4931 program and library directories @command{gcc} will search---and don't do anything else.
4933 This is useful when @command{gcc} prints the error message
4934 @samp{installation problem, cannot exec cpp0: No such file or directory}.
4935 To resolve this you either need to put @file{cpp0} and the other compiler
4936 components where @command{gcc} expects to find them, or you can set the environment
4937 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
4938 Don't forget the trailing @samp{/}.
4939 @xref{Environment Variables}.
4941 @item -print-sysroot-headers-suffix
4942 @opindex print-sysroot-headers-suffix
4943 Print the suffix added to the target sysroot when searching for
4944 headers, or give an error if the compiler is not configured with such
4945 a suffix---and don't do anything else.
4948 @opindex dumpmachine
4949 Print the compiler's target machine (for example,
4950 @samp{i686-pc-linux-gnu})---and don't do anything else.
4953 @opindex dumpversion
4954 Print the compiler version (for example, @samp{3.0})---and don't do
4959 Print the compiler's built-in specs---and don't do anything else. (This
4960 is used when GCC itself is being built.) @xref{Spec Files}.
4962 @item -feliminate-unused-debug-types
4963 @opindex feliminate-unused-debug-types
4964 Normally, when producing DWARF2 output, GCC will emit debugging
4965 information for all types declared in a compilation
4966 unit, regardless of whether or not they are actually used
4967 in that compilation unit. Sometimes this is useful, such as
4968 if, in the debugger, you want to cast a value to a type that is
4969 not actually used in your program (but is declared). More often,
4970 however, this results in a significant amount of wasted space.
4971 With this option, GCC will avoid producing debug symbol output
4972 for types that are nowhere used in the source file being compiled.
4975 @node Optimize Options
4976 @section Options That Control Optimization
4977 @cindex optimize options
4978 @cindex options, optimization
4980 These options control various sorts of optimizations.
4982 Without any optimization option, the compiler's goal is to reduce the
4983 cost of compilation and to make debugging produce the expected
4984 results. Statements are independent: if you stop the program with a
4985 breakpoint between statements, you can then assign a new value to any
4986 variable or change the program counter to any other statement in the
4987 function and get exactly the results you would expect from the source
4990 Turning on optimization flags makes the compiler attempt to improve
4991 the performance and/or code size at the expense of compilation time
4992 and possibly the ability to debug the program.
4994 The compiler performs optimization based on the knowledge it has of
4995 the program. Optimization levels @option{-O} and above, in
4996 particular, enable @emph{unit-at-a-time} mode, which allows the
4997 compiler to consider information gained from later functions in
4998 the file when compiling a function. Compiling multiple files at
4999 once to a single output file in @emph{unit-at-a-time} mode allows
5000 the compiler to use information gained from all of the files when
5001 compiling each of them.
5003 Not all optimizations are controlled directly by a flag. Only
5004 optimizations that have a flag are listed.
5011 Optimize. Optimizing compilation takes somewhat more time, and a lot
5012 more memory for a large function.
5014 With @option{-O}, the compiler tries to reduce code size and execution
5015 time, without performing any optimizations that take a great deal of
5018 @option{-O} turns on the following optimization flags:
5019 @gccoptlist{-fdefer-pop @gol
5020 -fdelayed-branch @gol
5021 -fguess-branch-probability @gol
5022 -fcprop-registers @gol
5023 -fif-conversion @gol
5024 -fif-conversion2 @gol
5025 -fsplit-wide-types @gol
5028 -ftree-dominator-opts @gol
5032 -ftree-copyrename @gol
5035 -funit-at-a-time @gol
5038 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5039 where doing so does not interfere with debugging.
5043 Optimize even more. GCC performs nearly all supported optimizations
5044 that do not involve a space-speed tradeoff. The compiler does not
5045 perform loop unrolling or function inlining when you specify @option{-O2}.
5046 As compared to @option{-O}, this option increases both compilation time
5047 and the performance of the generated code.
5049 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5050 also turns on the following optimization flags:
5051 @gccoptlist{-fthread-jumps @gol
5053 -foptimize-sibling-calls @gol
5054 -fcse-follow-jumps -fcse-skip-blocks @gol
5055 -fgcse -fgcse-lm @gol
5056 -fexpensive-optimizations @gol
5057 -frerun-cse-after-loop @gol
5060 -fschedule-insns -fschedule-insns2 @gol
5061 -fsched-interblock -fsched-spec @gol
5063 -fstrict-aliasing -fstrict-overflow @gol
5064 -fdelete-null-pointer-checks @gol
5065 -freorder-blocks -freorder-functions @gol
5066 -falign-functions -falign-jumps @gol
5067 -falign-loops -falign-labels @gol
5071 Please note the warning under @option{-fgcse} about
5072 invoking @option{-O2} on programs that use computed gotos.
5076 Optimize yet more. @option{-O3} turns on all optimizations specified by
5077 @option{-O2} and also turns on the @option{-finline-functions},
5078 @option{-funswitch-loops}, @option{-fpredictive-commoning} and
5079 @option{-fgcse-after-reload} options.
5083 Reduce compilation time and make debugging produce the expected
5084 results. This is the default.
5088 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5089 do not typically increase code size. It also performs further
5090 optimizations designed to reduce code size.
5092 @option{-Os} disables the following optimization flags:
5093 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5094 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5095 -fprefetch-loop-arrays -ftree-vect-loop-version}
5097 If you use multiple @option{-O} options, with or without level numbers,
5098 the last such option is the one that is effective.
5101 Options of the form @option{-f@var{flag}} specify machine-independent
5102 flags. Most flags have both positive and negative forms; the negative
5103 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5104 below, only one of the forms is listed---the one you typically will
5105 use. You can figure out the other form by either removing @samp{no-}
5108 The following options control specific optimizations. They are either
5109 activated by @option{-O} options or are related to ones that are. You
5110 can use the following flags in the rare cases when ``fine-tuning'' of
5111 optimizations to be performed is desired.
5114 @item -fno-default-inline
5115 @opindex fno-default-inline
5116 Do not make member functions inline by default merely because they are
5117 defined inside the class scope (C++ only). Otherwise, when you specify
5118 @w{@option{-O}}, member functions defined inside class scope are compiled
5119 inline by default; i.e., you don't need to add @samp{inline} in front of
5120 the member function name.
5122 @item -fno-defer-pop
5123 @opindex fno-defer-pop
5124 Always pop the arguments to each function call as soon as that function
5125 returns. For machines which must pop arguments after a function call,
5126 the compiler normally lets arguments accumulate on the stack for several
5127 function calls and pops them all at once.
5129 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5132 @opindex fforce-addr
5133 Force memory address constants to be copied into registers before
5134 doing arithmetic on them.
5136 @item -fforward-propagate
5137 @opindex fforward-propagate
5138 Perform a forward propagation pass on RTL. The pass tries to combine two
5139 instructions and checks if the result can be simplified. If loop unrolling
5140 is active, two passes are performed and the second is scheduled after
5143 This option is enabled by default at optimization levels @option{-O2},
5144 @option{-O3}, @option{-Os}.
5146 @item -fomit-frame-pointer
5147 @opindex fomit-frame-pointer
5148 Don't keep the frame pointer in a register for functions that
5149 don't need one. This avoids the instructions to save, set up and
5150 restore frame pointers; it also makes an extra register available
5151 in many functions. @strong{It also makes debugging impossible on
5154 On some machines, such as the VAX, this flag has no effect, because
5155 the standard calling sequence automatically handles the frame pointer
5156 and nothing is saved by pretending it doesn't exist. The
5157 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5158 whether a target machine supports this flag. @xref{Registers,,Register
5159 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5161 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5163 @item -foptimize-sibling-calls
5164 @opindex foptimize-sibling-calls
5165 Optimize sibling and tail recursive calls.
5167 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5171 Don't pay attention to the @code{inline} keyword. Normally this option
5172 is used to keep the compiler from expanding any functions inline.
5173 Note that if you are not optimizing, no functions can be expanded inline.
5175 @item -finline-functions
5176 @opindex finline-functions
5177 Integrate all simple functions into their callers. The compiler
5178 heuristically decides which functions are simple enough to be worth
5179 integrating in this way.
5181 If all calls to a given function are integrated, and the function is
5182 declared @code{static}, then the function is normally not output as
5183 assembler code in its own right.
5185 Enabled at level @option{-O3}.
5187 @item -finline-functions-called-once
5188 @opindex finline-functions-called-once
5189 Consider all @code{static} functions called once for inlining into their
5190 caller even if they are not marked @code{inline}. If a call to a given
5191 function is integrated, then the function is not output as assembler code
5194 Enabled if @option{-funit-at-a-time} is enabled.
5196 @item -fearly-inlining
5197 @opindex fearly-inlining
5198 Inline functions marked by @code{always_inline} and functions whose body seems
5199 smaller than the function call overhead early before doing
5200 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5201 makes profiling significantly cheaper and usually inlining faster on programs
5202 having large chains of nested wrapper functions.
5206 @item -finline-limit=@var{n}
5207 @opindex finline-limit
5208 By default, GCC limits the size of functions that can be inlined. This flag
5209 allows the control of this limit for functions that are explicitly marked as
5210 inline (i.e., marked with the inline keyword or defined within the class
5211 definition in c++). @var{n} is the size of functions that can be inlined in
5212 number of pseudo instructions (not counting parameter handling). The default
5213 value of @var{n} is 600.
5214 Increasing this value can result in more inlined code at
5215 the cost of compilation time and memory consumption. Decreasing usually makes
5216 the compilation faster and less code will be inlined (which presumably
5217 means slower programs). This option is particularly useful for programs that
5218 use inlining heavily such as those based on recursive templates with C++.
5220 Inlining is actually controlled by a number of parameters, which may be
5221 specified individually by using @option{--param @var{name}=@var{value}}.
5222 The @option{-finline-limit=@var{n}} option sets some of these parameters
5226 @item max-inline-insns-single
5227 is set to @var{n}/2.
5228 @item max-inline-insns-auto
5229 is set to @var{n}/2.
5230 @item min-inline-insns
5231 is set to 130 or @var{n}/4, whichever is smaller.
5232 @item max-inline-insns-rtl
5236 See below for a documentation of the individual
5237 parameters controlling inlining.
5239 @emph{Note:} pseudo instruction represents, in this particular context, an
5240 abstract measurement of function's size. In no way does it represent a count
5241 of assembly instructions and as such its exact meaning might change from one
5242 release to an another.
5244 @item -fkeep-inline-functions
5245 @opindex fkeep-inline-functions
5246 In C, emit @code{static} functions that are declared @code{inline}
5247 into the object file, even if the function has been inlined into all
5248 of its callers. This switch does not affect functions using the
5249 @code{extern inline} extension in GNU C89@. In C++, emit any and all
5250 inline functions into the object file.
5252 @item -fkeep-static-consts
5253 @opindex fkeep-static-consts
5254 Emit variables declared @code{static const} when optimization isn't turned
5255 on, even if the variables aren't referenced.
5257 GCC enables this option by default. If you want to force the compiler to
5258 check if the variable was referenced, regardless of whether or not
5259 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5261 @item -fmerge-constants
5262 Attempt to merge identical constants (string constants and floating point
5263 constants) across compilation units.
5265 This option is the default for optimized compilation if the assembler and
5266 linker support it. Use @option{-fno-merge-constants} to inhibit this
5269 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5271 @item -fmerge-all-constants
5272 Attempt to merge identical constants and identical variables.
5274 This option implies @option{-fmerge-constants}. In addition to
5275 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5276 arrays or initialized constant variables with integral or floating point
5277 types. Languages like C or C++ require each non-automatic variable to
5278 have distinct location, so using this option will result in non-conforming
5281 @item -fmodulo-sched
5282 @opindex fmodulo-sched
5283 Perform swing modulo scheduling immediately before the first scheduling
5284 pass. This pass looks at innermost loops and reorders their
5285 instructions by overlapping different iterations.
5287 @item -fmodulo-sched-allow-regmoves
5288 @opindex fmodulo-sched-allow-regmoves
5289 Perform more aggressive SMS based modulo scheduling with register moves
5290 allowed. By setting this flag certain anti-dependences edges will be
5291 deleted which will trigger the generation of reg-moves based on the
5292 life-range analysis. This option is effective only with
5293 @option{-fmodulo-sched} enabled.
5295 @item -fno-branch-count-reg
5296 @opindex fno-branch-count-reg
5297 Do not use ``decrement and branch'' instructions on a count register,
5298 but instead generate a sequence of instructions that decrement a
5299 register, compare it against zero, then branch based upon the result.
5300 This option is only meaningful on architectures that support such
5301 instructions, which include x86, PowerPC, IA-64 and S/390.
5303 The default is @option{-fbranch-count-reg}.
5305 @item -fno-function-cse
5306 @opindex fno-function-cse
5307 Do not put function addresses in registers; make each instruction that
5308 calls a constant function contain the function's address explicitly.
5310 This option results in less efficient code, but some strange hacks
5311 that alter the assembler output may be confused by the optimizations
5312 performed when this option is not used.
5314 The default is @option{-ffunction-cse}
5316 @item -fno-zero-initialized-in-bss
5317 @opindex fno-zero-initialized-in-bss
5318 If the target supports a BSS section, GCC by default puts variables that
5319 are initialized to zero into BSS@. This can save space in the resulting
5322 This option turns off this behavior because some programs explicitly
5323 rely on variables going to the data section. E.g., so that the
5324 resulting executable can find the beginning of that section and/or make
5325 assumptions based on that.
5327 The default is @option{-fzero-initialized-in-bss}.
5329 @item -fbounds-check
5330 @opindex fbounds-check
5331 For front-ends that support it, generate additional code to check that
5332 indices used to access arrays are within the declared range. This is
5333 currently only supported by the Java and Fortran front-ends, where
5334 this option defaults to true and false respectively.
5336 @item -fmudflap -fmudflapth -fmudflapir
5340 @cindex bounds checking
5342 For front-ends that support it (C and C++), instrument all risky
5343 pointer/array dereferencing operations, some standard library
5344 string/heap functions, and some other associated constructs with
5345 range/validity tests. Modules so instrumented should be immune to
5346 buffer overflows, invalid heap use, and some other classes of C/C++
5347 programming errors. The instrumentation relies on a separate runtime
5348 library (@file{libmudflap}), which will be linked into a program if
5349 @option{-fmudflap} is given at link time. Run-time behavior of the
5350 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
5351 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
5354 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
5355 link if your program is multi-threaded. Use @option{-fmudflapir}, in
5356 addition to @option{-fmudflap} or @option{-fmudflapth}, if
5357 instrumentation should ignore pointer reads. This produces less
5358 instrumentation (and therefore faster execution) and still provides
5359 some protection against outright memory corrupting writes, but allows
5360 erroneously read data to propagate within a program.
5362 @item -fthread-jumps
5363 @opindex fthread-jumps
5364 Perform optimizations where we check to see if a jump branches to a
5365 location where another comparison subsumed by the first is found. If
5366 so, the first branch is redirected to either the destination of the
5367 second branch or a point immediately following it, depending on whether
5368 the condition is known to be true or false.
5370 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5372 @item -fsplit-wide-types
5373 @opindex fsplit-wide-types
5374 When using a type that occupies multiple registers, such as @code{long
5375 long} on a 32-bit system, split the registers apart and allocate them
5376 independently. This normally generates better code for those types,
5377 but may make debugging more difficult.
5379 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
5382 @item -fcse-follow-jumps
5383 @opindex fcse-follow-jumps
5384 In common subexpression elimination, scan through jump instructions
5385 when the target of the jump is not reached by any other path. For
5386 example, when CSE encounters an @code{if} statement with an
5387 @code{else} clause, CSE will follow the jump when the condition
5390 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5392 @item -fcse-skip-blocks
5393 @opindex fcse-skip-blocks
5394 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
5395 follow jumps which conditionally skip over blocks. When CSE
5396 encounters a simple @code{if} statement with no else clause,
5397 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
5398 body of the @code{if}.
5400 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5402 @item -frerun-cse-after-loop
5403 @opindex frerun-cse-after-loop
5404 Re-run common subexpression elimination after loop optimizations has been
5407 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5411 Perform a global common subexpression elimination pass.
5412 This pass also performs global constant and copy propagation.
5414 @emph{Note:} When compiling a program using computed gotos, a GCC
5415 extension, you may get better runtime performance if you disable
5416 the global common subexpression elimination pass by adding
5417 @option{-fno-gcse} to the command line.
5419 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5423 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
5424 attempt to move loads which are only killed by stores into themselves. This
5425 allows a loop containing a load/store sequence to be changed to a load outside
5426 the loop, and a copy/store within the loop.
5428 Enabled by default when gcse is enabled.
5432 When @option{-fgcse-sm} is enabled, a store motion pass is run after
5433 global common subexpression elimination. This pass will attempt to move
5434 stores out of loops. When used in conjunction with @option{-fgcse-lm},
5435 loops containing a load/store sequence can be changed to a load before
5436 the loop and a store after the loop.
5438 Not enabled at any optimization level.
5442 When @option{-fgcse-las} is enabled, the global common subexpression
5443 elimination pass eliminates redundant loads that come after stores to the
5444 same memory location (both partial and full redundancies).
5446 Not enabled at any optimization level.
5448 @item -fgcse-after-reload
5449 @opindex fgcse-after-reload
5450 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
5451 pass is performed after reload. The purpose of this pass is to cleanup
5454 @item -funsafe-loop-optimizations
5455 @opindex funsafe-loop-optimizations
5456 If given, the loop optimizer will assume that loop indices do not
5457 overflow, and that the loops with nontrivial exit condition are not
5458 infinite. This enables a wider range of loop optimizations even if
5459 the loop optimizer itself cannot prove that these assumptions are valid.
5460 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
5461 if it finds this kind of loop.
5463 @item -fcrossjumping
5464 @opindex crossjumping
5465 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
5466 resulting code may or may not perform better than without cross-jumping.
5468 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5470 @item -fif-conversion
5471 @opindex if-conversion
5472 Attempt to transform conditional jumps into branch-less equivalents. This
5473 include use of conditional moves, min, max, set flags and abs instructions, and
5474 some tricks doable by standard arithmetics. The use of conditional execution
5475 on chips where it is available is controlled by @code{if-conversion2}.
5477 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5479 @item -fif-conversion2
5480 @opindex if-conversion2
5481 Use conditional execution (where available) to transform conditional jumps into
5482 branch-less equivalents.
5484 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5486 @item -fdelete-null-pointer-checks
5487 @opindex fdelete-null-pointer-checks
5488 Use global dataflow analysis to identify and eliminate useless checks
5489 for null pointers. The compiler assumes that dereferencing a null
5490 pointer would have halted the program. If a pointer is checked after
5491 it has already been dereferenced, it cannot be null.
5493 In some environments, this assumption is not true, and programs can
5494 safely dereference null pointers. Use
5495 @option{-fno-delete-null-pointer-checks} to disable this optimization
5496 for programs which depend on that behavior.
5498 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5500 @item -fexpensive-optimizations
5501 @opindex fexpensive-optimizations
5502 Perform a number of minor optimizations that are relatively expensive.
5504 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5506 @item -foptimize-register-move
5508 @opindex foptimize-register-move
5510 Attempt to reassign register numbers in move instructions and as
5511 operands of other simple instructions in order to maximize the amount of
5512 register tying. This is especially helpful on machines with two-operand
5515 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
5518 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5520 @item -fdelayed-branch
5521 @opindex fdelayed-branch
5522 If supported for the target machine, attempt to reorder instructions
5523 to exploit instruction slots available after delayed branch
5526 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5528 @item -fschedule-insns
5529 @opindex fschedule-insns
5530 If supported for the target machine, attempt to reorder instructions to
5531 eliminate execution stalls due to required data being unavailable. This
5532 helps machines that have slow floating point or memory load instructions
5533 by allowing other instructions to be issued until the result of the load
5534 or floating point instruction is required.
5536 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5538 @item -fschedule-insns2
5539 @opindex fschedule-insns2
5540 Similar to @option{-fschedule-insns}, but requests an additional pass of
5541 instruction scheduling after register allocation has been done. This is
5542 especially useful on machines with a relatively small number of
5543 registers and where memory load instructions take more than one cycle.
5545 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5547 @item -fno-sched-interblock
5548 @opindex fno-sched-interblock
5549 Don't schedule instructions across basic blocks. This is normally
5550 enabled by default when scheduling before register allocation, i.e.@:
5551 with @option{-fschedule-insns} or at @option{-O2} or higher.
5553 @item -fno-sched-spec
5554 @opindex fno-sched-spec
5555 Don't allow speculative motion of non-load instructions. This is normally
5556 enabled by default when scheduling before register allocation, i.e.@:
5557 with @option{-fschedule-insns} or at @option{-O2} or higher.
5559 @item -fsched-spec-load
5560 @opindex fsched-spec-load
5561 Allow speculative motion of some load instructions. This only makes
5562 sense when scheduling before register allocation, i.e.@: with
5563 @option{-fschedule-insns} or at @option{-O2} or higher.
5565 @item -fsched-spec-load-dangerous
5566 @opindex fsched-spec-load-dangerous
5567 Allow speculative motion of more load instructions. This only makes
5568 sense when scheduling before register allocation, i.e.@: with
5569 @option{-fschedule-insns} or at @option{-O2} or higher.
5571 @item -fsched-stalled-insns=@var{n}
5572 @opindex fsched-stalled-insns
5573 Define how many insns (if any) can be moved prematurely from the queue
5574 of stalled insns into the ready list, during the second scheduling pass.
5576 @item -fsched-stalled-insns-dep=@var{n}
5577 @opindex fsched-stalled-insns-dep
5578 Define how many insn groups (cycles) will be examined for a dependency
5579 on a stalled insn that is candidate for premature removal from the queue
5580 of stalled insns. Has an effect only during the second scheduling pass,
5581 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
5583 @item -fsched2-use-superblocks
5584 @opindex fsched2-use-superblocks
5585 When scheduling after register allocation, do use superblock scheduling
5586 algorithm. Superblock scheduling allows motion across basic block boundaries
5587 resulting on faster schedules. This option is experimental, as not all machine
5588 descriptions used by GCC model the CPU closely enough to avoid unreliable
5589 results from the algorithm.
5591 This only makes sense when scheduling after register allocation, i.e.@: with
5592 @option{-fschedule-insns2} or at @option{-O2} or higher.
5594 @item -fsched2-use-traces
5595 @opindex fsched2-use-traces
5596 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
5597 allocation and additionally perform code duplication in order to increase the
5598 size of superblocks using tracer pass. See @option{-ftracer} for details on
5601 This mode should produce faster but significantly longer programs. Also
5602 without @option{-fbranch-probabilities} the traces constructed may not
5603 match the reality and hurt the performance. This only makes
5604 sense when scheduling after register allocation, i.e.@: with
5605 @option{-fschedule-insns2} or at @option{-O2} or higher.
5609 Eliminates redundant extension instructions and move the non redundant
5610 ones to optimal placement using LCM.
5612 @item -freschedule-modulo-scheduled-loops
5613 @opindex fscheduling-in-modulo-scheduled-loops
5614 The modulo scheduling comes before the traditional scheduling, if a loop
5615 was modulo scheduled we may want to prevent the later scheduling passes
5616 from changing its schedule, we use this option to control that.
5618 @item -fcaller-saves
5619 @opindex fcaller-saves
5620 Enable values to be allocated in registers that will be clobbered by
5621 function calls, by emitting extra instructions to save and restore the
5622 registers around such calls. Such allocation is done only when it
5623 seems to result in better code than would otherwise be produced.
5625 This option is always enabled by default on certain machines, usually
5626 those which have no call-preserved registers to use instead.
5628 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5630 @item -ftree-reassoc
5631 Perform Reassociation on trees This flag is enabled by default
5632 at @option{-O} and higher.
5635 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
5636 enabled by default at @option{-O2} and @option{-O3}.
5639 Perform Full Redundancy Elimination (FRE) on trees. The difference
5640 between FRE and PRE is that FRE only considers expressions
5641 that are computed on all paths leading to the redundant computation.
5642 This analysis is faster than PRE, though it exposes fewer redundancies.
5643 This flag is enabled by default at @option{-O} and higher.
5645 @item -ftree-copy-prop
5646 Perform copy propagation on trees. This pass eliminates unnecessary
5647 copy operations. This flag is enabled by default at @option{-O} and
5650 @item -ftree-store-copy-prop
5651 Perform copy propagation of memory loads and stores. This pass
5652 eliminates unnecessary copy operations in memory references
5653 (structures, global variables, arrays, etc). This flag is enabled by
5654 default at @option{-O2} and higher.
5657 Perform structural alias analysis on trees. This flag
5658 is enabled by default at @option{-O} and higher.
5661 Perform interprocedural pointer analysis.
5664 Perform forward store motion on trees. This flag is
5665 enabled by default at @option{-O} and higher.
5668 Perform sparse conditional constant propagation (CCP) on trees. This
5669 pass only operates on local scalar variables and is enabled by default
5670 at @option{-O} and higher.
5672 @item -ftree-store-ccp
5673 Perform sparse conditional constant propagation (CCP) on trees. This
5674 pass operates on both local scalar variables and memory stores and
5675 loads (global variables, structures, arrays, etc). This flag is
5676 enabled by default at @option{-O2} and higher.
5679 Perform dead code elimination (DCE) on trees. This flag is enabled by
5680 default at @option{-O} and higher.
5682 @item -ftree-dominator-opts
5683 Perform a variety of simple scalar cleanups (constant/copy
5684 propagation, redundancy elimination, range propagation and expression
5685 simplification) based on a dominator tree traversal. This also
5686 performs jump threading (to reduce jumps to jumps). This flag is
5687 enabled by default at @option{-O} and higher.
5690 Perform loop header copying on trees. This is beneficial since it increases
5691 effectiveness of code motion optimizations. It also saves one jump. This flag
5692 is enabled by default at @option{-O} and higher. It is not enabled
5693 for @option{-Os}, since it usually increases code size.
5695 @item -ftree-loop-optimize
5696 Perform loop optimizations on trees. This flag is enabled by default
5697 at @option{-O} and higher.
5699 @item -ftree-loop-linear
5700 Perform linear loop transformations on tree. This flag can improve cache
5701 performance and allow further loop optimizations to take place.
5703 @item -fcheck-data-deps
5704 Compare the results of several data dependence analyzers. This option
5705 is used for debugging the data dependence analyzers.
5707 @item -ftree-loop-im
5708 Perform loop invariant motion on trees. This pass moves only invariants that
5709 would be hard to handle at RTL level (function calls, operations that expand to
5710 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5711 operands of conditions that are invariant out of the loop, so that we can use
5712 just trivial invariantness analysis in loop unswitching. The pass also includes
5715 @item -ftree-loop-ivcanon
5716 Create a canonical counter for number of iterations in the loop for that
5717 determining number of iterations requires complicated analysis. Later
5718 optimizations then may determine the number easily. Useful especially
5719 in connection with unrolling.
5722 Perform induction variable optimizations (strength reduction, induction
5723 variable merging and induction variable elimination) on trees.
5726 Perform scalar replacement of aggregates. This pass replaces structure
5727 references with scalars to prevent committing structures to memory too
5728 early. This flag is enabled by default at @option{-O} and higher.
5730 @item -ftree-copyrename
5731 Perform copy renaming on trees. This pass attempts to rename compiler
5732 temporaries to other variables at copy locations, usually resulting in
5733 variable names which more closely resemble the original variables. This flag
5734 is enabled by default at @option{-O} and higher.
5737 Perform temporary expression replacement during the SSA->normal phase. Single
5738 use/single def temporaries are replaced at their use location with their
5739 defining expression. This results in non-GIMPLE code, but gives the expanders
5740 much more complex trees to work on resulting in better RTL generation. This is
5741 enabled by default at @option{-O} and higher.
5743 @item -ftree-vectorize
5744 Perform loop vectorization on trees.
5746 @item -ftree-vect-loop-version
5747 @opindex ftree-vect-loop-version
5748 Perform loop versioning when doing loop vectorization on trees. When a loop
5749 appears to be vectorizable except that data alignment or data dependence cannot
5750 be determined at compile time then vectorized and non-vectorized versions of
5751 the loop are generated along with runtime checks for alignment or dependence
5752 to control which version is executed. This option is enabled by default
5753 except at level @option{-Os} where it is disabled.
5755 @item -fvect-cost-model
5756 Enable cost model for vectorization.
5759 Perform Value Range Propagation on trees. This is similar to the
5760 constant propagation pass, but instead of values, ranges of values are
5761 propagated. This allows the optimizers to remove unnecessary range
5762 checks like array bound checks and null pointer checks. This is
5763 enabled by default at @option{-O2} and higher. Null pointer check
5764 elimination is only done if @option{-fdelete-null-pointer-checks} is
5769 Perform tail duplication to enlarge superblock size. This transformation
5770 simplifies the control flow of the function allowing other optimizations to do
5773 @item -funroll-loops
5774 @opindex funroll-loops
5775 Unroll loops whose number of iterations can be determined at compile
5776 time or upon entry to the loop. @option{-funroll-loops} implies
5777 @option{-frerun-cse-after-loop}. This option makes code larger,
5778 and may or may not make it run faster.
5780 @item -funroll-all-loops
5781 @opindex funroll-all-loops
5782 Unroll all loops, even if their number of iterations is uncertain when
5783 the loop is entered. This usually makes programs run more slowly.
5784 @option{-funroll-all-loops} implies the same options as
5785 @option{-funroll-loops},
5787 @item -fsplit-ivs-in-unroller
5788 @opindex fsplit-ivs-in-unroller
5789 Enables expressing of values of induction variables in later iterations
5790 of the unrolled loop using the value in the first iteration. This breaks
5791 long dependency chains, thus improving efficiency of the scheduling passes.
5793 Combination of @option{-fweb} and CSE is often sufficient to obtain the
5794 same effect. However in cases the loop body is more complicated than
5795 a single basic block, this is not reliable. It also does not work at all
5796 on some of the architectures due to restrictions in the CSE pass.
5798 This optimization is enabled by default.
5800 @item -fvariable-expansion-in-unroller
5801 @opindex fvariable-expansion-in-unroller
5802 With this option, the compiler will create multiple copies of some
5803 local variables when unrolling a loop which can result in superior code.
5805 @item -fpredictive-commoning
5806 @opindex fpredictive-commoning
5807 Perform predictive commoning optimization, i.e., reusing computations
5808 (especially memory loads and stores) performed in previous
5809 iterations of loops.
5811 This option is enabled at level @option{-O3}.
5813 @item -fprefetch-loop-arrays
5814 @opindex fprefetch-loop-arrays
5815 If supported by the target machine, generate instructions to prefetch
5816 memory to improve the performance of loops that access large arrays.
5818 This option may generate better or worse code; results are highly
5819 dependent on the structure of loops within the source code.
5821 Disabled at level @option{-Os}.
5824 @itemx -fno-peephole2
5825 @opindex fno-peephole
5826 @opindex fno-peephole2
5827 Disable any machine-specific peephole optimizations. The difference
5828 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
5829 are implemented in the compiler; some targets use one, some use the
5830 other, a few use both.
5832 @option{-fpeephole} is enabled by default.
5833 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5835 @item -fno-guess-branch-probability
5836 @opindex fno-guess-branch-probability
5837 Do not guess branch probabilities using heuristics.
5839 GCC will use heuristics to guess branch probabilities if they are
5840 not provided by profiling feedback (@option{-fprofile-arcs}). These
5841 heuristics are based on the control flow graph. If some branch probabilities
5842 are specified by @samp{__builtin_expect}, then the heuristics will be
5843 used to guess branch probabilities for the rest of the control flow graph,
5844 taking the @samp{__builtin_expect} info into account. The interactions
5845 between the heuristics and @samp{__builtin_expect} can be complex, and in
5846 some cases, it may be useful to disable the heuristics so that the effects
5847 of @samp{__builtin_expect} are easier to understand.
5849 The default is @option{-fguess-branch-probability} at levels
5850 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5852 @item -freorder-blocks
5853 @opindex freorder-blocks
5854 Reorder basic blocks in the compiled function in order to reduce number of
5855 taken branches and improve code locality.
5857 Enabled at levels @option{-O2}, @option{-O3}.
5859 @item -freorder-blocks-and-partition
5860 @opindex freorder-blocks-and-partition
5861 In addition to reordering basic blocks in the compiled function, in order
5862 to reduce number of taken branches, partitions hot and cold basic blocks
5863 into separate sections of the assembly and .o files, to improve
5864 paging and cache locality performance.
5866 This optimization is automatically turned off in the presence of
5867 exception handling, for linkonce sections, for functions with a user-defined
5868 section attribute and on any architecture that does not support named
5871 @item -freorder-functions
5872 @opindex freorder-functions
5873 Reorder functions in the object file in order to
5874 improve code locality. This is implemented by using special
5875 subsections @code{.text.hot} for most frequently executed functions and
5876 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
5877 the linker so object file format must support named sections and linker must
5878 place them in a reasonable way.
5880 Also profile feedback must be available in to make this option effective. See
5881 @option{-fprofile-arcs} for details.
5883 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5885 @item -fstrict-aliasing
5886 @opindex fstrict-aliasing
5887 Allows the compiler to assume the strictest aliasing rules applicable to
5888 the language being compiled. For C (and C++), this activates
5889 optimizations based on the type of expressions. In particular, an
5890 object of one type is assumed never to reside at the same address as an
5891 object of a different type, unless the types are almost the same. For
5892 example, an @code{unsigned int} can alias an @code{int}, but not a
5893 @code{void*} or a @code{double}. A character type may alias any other
5896 Pay special attention to code like this:
5909 The practice of reading from a different union member than the one most
5910 recently written to (called ``type-punning'') is common. Even with
5911 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
5912 is accessed through the union type. So, the code above will work as
5913 expected. However, this code might not:
5924 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5926 @item -fstrict-overflow
5927 @opindex fstrict-overflow
5928 Allow the compiler to assume strict signed overflow rules, depending
5929 on the language being compiled. For C (and C++) this means that
5930 overflow when doing arithmetic with signed numbers is undefined, which
5931 means that the compiler may assume that it will not happen. This
5932 permits various optimizations. For example, the compiler will assume
5933 that an expression like @code{i + 10 > i} will always be true for
5934 signed @code{i}. This assumption is only valid if signed overflow is
5935 undefined, as the expression is false if @code{i + 10} overflows when
5936 using twos complement arithmetic. When this option is in effect any
5937 attempt to determine whether an operation on signed numbers will
5938 overflow must be written carefully to not actually involve overflow.
5940 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
5941 that signed overflow is fully defined: it wraps. When
5942 @option{-fwrapv} is used, there is no difference between
5943 @option{-fstrict-overflow} and @option{-fno-strict-overflow}. With
5944 @option{-fwrapv} certain types of overflow are permitted. For
5945 example, if the compiler gets an overflow when doing arithmetic on
5946 constants, the overflowed value can still be used with
5947 @option{-fwrapv}, but not otherwise.
5949 The @option{-fstrict-overflow} option is enabled at levels
5950 @option{-O2}, @option{-O3}, @option{-Os}.
5952 @item -falign-functions
5953 @itemx -falign-functions=@var{n}
5954 @opindex falign-functions
5955 Align the start of functions to the next power-of-two greater than
5956 @var{n}, skipping up to @var{n} bytes. For instance,
5957 @option{-falign-functions=32} aligns functions to the next 32-byte
5958 boundary, but @option{-falign-functions=24} would align to the next
5959 32-byte boundary only if this can be done by skipping 23 bytes or less.
5961 @option{-fno-align-functions} and @option{-falign-functions=1} are
5962 equivalent and mean that functions will not be aligned.
5964 Some assemblers only support this flag when @var{n} is a power of two;
5965 in that case, it is rounded up.
5967 If @var{n} is not specified or is zero, use a machine-dependent default.
5969 Enabled at levels @option{-O2}, @option{-O3}.
5971 @item -falign-labels
5972 @itemx -falign-labels=@var{n}
5973 @opindex falign-labels
5974 Align all branch targets to a power-of-two boundary, skipping up to
5975 @var{n} bytes like @option{-falign-functions}. This option can easily
5976 make code slower, because it must insert dummy operations for when the
5977 branch target is reached in the usual flow of the code.
5979 @option{-fno-align-labels} and @option{-falign-labels=1} are
5980 equivalent and mean that labels will not be aligned.
5982 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
5983 are greater than this value, then their values are used instead.
5985 If @var{n} is not specified or is zero, use a machine-dependent default
5986 which is very likely to be @samp{1}, meaning no alignment.
5988 Enabled at levels @option{-O2}, @option{-O3}.
5991 @itemx -falign-loops=@var{n}
5992 @opindex falign-loops
5993 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
5994 like @option{-falign-functions}. The hope is that the loop will be
5995 executed many times, which will make up for any execution of the dummy
5998 @option{-fno-align-loops} and @option{-falign-loops=1} are
5999 equivalent and mean that loops will not be aligned.
6001 If @var{n} is not specified or is zero, use a machine-dependent default.
6003 Enabled at levels @option{-O2}, @option{-O3}.
6006 @itemx -falign-jumps=@var{n}
6007 @opindex falign-jumps
6008 Align branch targets to a power-of-two boundary, for branch targets
6009 where the targets can only be reached by jumping, skipping up to @var{n}
6010 bytes like @option{-falign-functions}. In this case, no dummy operations
6013 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
6014 equivalent and mean that loops will not be aligned.
6016 If @var{n} is not specified or is zero, use a machine-dependent default.
6018 Enabled at levels @option{-O2}, @option{-O3}.
6020 @item -funit-at-a-time
6021 @opindex funit-at-a-time
6022 Parse the whole compilation unit before starting to produce code.
6023 This allows some extra optimizations to take place but consumes
6024 more memory (in general). There are some compatibility issues
6025 with @emph{unit-at-a-time} mode:
6028 enabling @emph{unit-at-a-time} mode may change the order
6029 in which functions, variables, and top-level @code{asm} statements
6030 are emitted, and will likely break code relying on some particular
6031 ordering. The majority of such top-level @code{asm} statements,
6032 though, can be replaced by @code{section} attributes. The
6033 @option{fno-toplevel-reorder} option may be used to keep the ordering
6034 used in the input file, at the cost of some optimizations.
6037 @emph{unit-at-a-time} mode removes unreferenced static variables
6038 and functions. This may result in undefined references
6039 when an @code{asm} statement refers directly to variables or functions
6040 that are otherwise unused. In that case either the variable/function
6041 shall be listed as an operand of the @code{asm} statement operand or,
6042 in the case of top-level @code{asm} statements the attribute @code{used}
6043 shall be used on the declaration.
6046 Static functions now can use non-standard passing conventions that
6047 may break @code{asm} statements calling functions directly. Again,
6048 attribute @code{used} will prevent this behavior.
6051 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
6052 but this scheme may not be supported by future releases of GCC@.
6054 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6056 @item -fno-toplevel-reorder
6057 Do not reorder top-level functions, variables, and @code{asm}
6058 statements. Output them in the same order that they appear in the
6059 input file. When this option is used, unreferenced static variables
6060 will not be removed. This option is intended to support existing code
6061 which relies on a particular ordering. For new code, it is better to
6066 Constructs webs as commonly used for register allocation purposes and assign
6067 each web individual pseudo register. This allows the register allocation pass
6068 to operate on pseudos directly, but also strengthens several other optimization
6069 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
6070 however, make debugging impossible, since variables will no longer stay in a
6073 Enabled by default with @option{-funroll-loops}.
6075 @item -fwhole-program
6076 @opindex fwhole-program
6077 Assume that the current compilation unit represents whole program being
6078 compiled. All public functions and variables with the exception of @code{main}
6079 and those merged by attribute @code{externally_visible} become static functions
6080 and in a affect gets more aggressively optimized by interprocedural optimizers.
6081 While this option is equivalent to proper use of @code{static} keyword for
6082 programs consisting of single file, in combination with option
6083 @option{--combine} this flag can be used to compile most of smaller scale C
6084 programs since the functions and variables become local for the whole combined
6085 compilation unit, not for the single source file itself.
6088 @item -fno-cprop-registers
6089 @opindex fno-cprop-registers
6090 After register allocation and post-register allocation instruction splitting,
6091 we perform a copy-propagation pass to try to reduce scheduling dependencies
6092 and occasionally eliminate the copy.
6094 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6096 @item -fprofile-generate
6097 @opindex fprofile-generate
6099 Enable options usually used for instrumenting application to produce
6100 profile useful for later recompilation with profile feedback based
6101 optimization. You must use @option{-fprofile-generate} both when
6102 compiling and when linking your program.
6104 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
6107 @opindex fprofile-use
6108 Enable profile feedback directed optimizations, and optimizations
6109 generally profitable only with profile feedback available.
6111 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
6112 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
6114 By default, GCC emits an error message if the feedback profiles do not
6115 match the source code. This error can be turned into a warning by using
6116 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
6120 The following options control compiler behavior regarding floating
6121 point arithmetic. These options trade off between speed and
6122 correctness. All must be specifically enabled.
6126 @opindex ffloat-store
6127 Do not store floating point variables in registers, and inhibit other
6128 options that might change whether a floating point value is taken from a
6131 @cindex floating point precision
6132 This option prevents undesirable excess precision on machines such as
6133 the 68000 where the floating registers (of the 68881) keep more
6134 precision than a @code{double} is supposed to have. Similarly for the
6135 x86 architecture. For most programs, the excess precision does only
6136 good, but a few programs rely on the precise definition of IEEE floating
6137 point. Use @option{-ffloat-store} for such programs, after modifying
6138 them to store all pertinent intermediate computations into variables.
6142 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
6143 @option{-fno-trapping-math}, @option{-ffinite-math-only},
6144 @option{-fno-rounding-math}, @option{-fno-signaling-nans},
6145 @option{-fno-signed-zeros} and @option{fcx-limited-range}.
6147 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
6149 This option is not turned on by any @option{-O} option since
6150 it can result in incorrect output for programs which depend on
6151 an exact implementation of IEEE or ISO rules/specifications for
6152 math functions. It may, however, yield faster code for programs
6153 that do not require the guarantees of these specifications.
6155 @item -fno-math-errno
6156 @opindex fno-math-errno
6157 Do not set ERRNO after calling math functions that are executed
6158 with a single instruction, e.g., sqrt. A program that relies on
6159 IEEE exceptions for math error handling may want to use this flag
6160 for speed while maintaining IEEE arithmetic compatibility.
6162 This option is not turned on by any @option{-O} option since
6163 it can result in incorrect output for programs which depend on
6164 an exact implementation of IEEE or ISO rules/specifications for
6165 math functions. It may, however, yield faster code for programs
6166 that do not require the guarantees of these specifications.
6168 The default is @option{-fmath-errno}.
6170 On Darwin systems, the math library never sets @code{errno}. There is
6171 therefore no reason for the compiler to consider the possibility that
6172 it might, and @option{-fno-math-errno} is the default.
6174 @item -funsafe-math-optimizations
6175 @opindex funsafe-math-optimizations
6177 Allow optimizations for floating-point arithmetic that (a) assume
6178 that arguments and results are valid and (b) may violate IEEE or
6179 ANSI standards. When used at link-time, it may include libraries
6180 or startup files that change the default FPU control word or other
6181 similar optimizations.
6183 This option is not turned on by any @option{-O} option since
6184 it can result in incorrect output for programs which depend on
6185 an exact implementation of IEEE or ISO rules/specifications for
6186 math functions. It may, however, yield faster code for programs
6187 that do not require the guarantees of these specifications.
6188 Enables @option{-freciprocal-math} and @option{-fassociative-math}.
6190 The default is @option{-fno-unsafe-math-optimizations}.
6192 @item -fassociative-math
6193 @opindex -fassociative-math
6195 Allow re-association of operands in series of floating-point operations.
6196 This violates the ISO C and C++ language standard by possibly changing
6197 computation result. NOTE: re-ordering may change the sign of zero as
6198 well as ignore NaNs and inhibit or create underflow or overflow (and
6199 thus cannot be used on a code which relies on rounding behavior like
6200 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
6201 and thus may not be used when ordered comparisons are required.
6202 This flag doesn't make much sense without @option{-fno-signed-zeros}
6203 or @option{-fno-trapping-math} or with @option{-frounding-math}.
6205 The default is @option{-fno-associative-math}.
6207 @item -freciprocal-math
6208 @opindex -freciprocal-math
6210 Allow the reciprocal of a value to be used instead of dividing by
6211 the value if this enables optimizations. For example @code{x / y}
6212 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
6213 is subject to common subexpression elimination. Note that this loses
6214 precision and increases the number of flops operating on the value.
6216 The default is @option{-fno-reciprocal-math}.
6218 @item -ffinite-math-only
6219 @opindex ffinite-math-only
6220 Allow optimizations for floating-point arithmetic that assume
6221 that arguments and results are not NaNs or +-Infs.
6223 This option is not turned on by any @option{-O} option since
6224 it can result in incorrect output for programs which depend on
6225 an exact implementation of IEEE or ISO rules/specifications for
6226 math functions. It may, however, yield faster code for programs
6227 that do not require the guarantees of these specifications.
6229 The default is @option{-fno-finite-math-only}.
6231 @item -fno-signed-zeros
6232 @opindex fno-signed-zeros
6233 Allow optimizations for floating point arithmetic that ignore the
6234 signedness of zero. IEEE arithmetic specifies the behavior of
6235 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
6236 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
6237 This option implies that the sign of a zero result isn't significant.
6239 The default is @option{-fsigned-zeros}.
6241 @item -fno-trapping-math
6242 @opindex fno-trapping-math
6243 Compile code assuming that floating-point operations cannot generate
6244 user-visible traps. These traps include division by zero, overflow,
6245 underflow, inexact result and invalid operation. This option implies
6246 @option{-fno-signaling-nans}. Setting this option may allow faster
6247 code if one relies on ``non-stop'' IEEE arithmetic, for example.
6249 This option should never be turned on by any @option{-O} option since
6250 it can result in incorrect output for programs which depend on
6251 an exact implementation of IEEE or ISO rules/specifications for
6254 The default is @option{-ftrapping-math}.
6256 @item -frounding-math
6257 @opindex frounding-math
6258 Disable transformations and optimizations that assume default floating
6259 point rounding behavior. This is round-to-zero for all floating point
6260 to integer conversions, and round-to-nearest for all other arithmetic
6261 truncations. This option should be specified for programs that change
6262 the FP rounding mode dynamically, or that may be executed with a
6263 non-default rounding mode. This option disables constant folding of
6264 floating point expressions at compile-time (which may be affected by
6265 rounding mode) and arithmetic transformations that are unsafe in the
6266 presence of sign-dependent rounding modes.
6268 The default is @option{-fno-rounding-math}.
6270 This option is experimental and does not currently guarantee to
6271 disable all GCC optimizations that are affected by rounding mode.
6272 Future versions of GCC may provide finer control of this setting
6273 using C99's @code{FENV_ACCESS} pragma. This command line option
6274 will be used to specify the default state for @code{FENV_ACCESS}.
6276 @item -frtl-abstract-sequences
6277 @opindex frtl-abstract-sequences
6278 It is a size optimization method. This option is to find identical
6279 sequences of code, which can be turned into pseudo-procedures and
6280 then replace all occurrences with calls to the newly created
6281 subroutine. It is kind of an opposite of @option{-finline-functions}.
6282 This optimization runs at RTL level.
6284 @item -fsignaling-nans
6285 @opindex fsignaling-nans
6286 Compile code assuming that IEEE signaling NaNs may generate user-visible
6287 traps during floating-point operations. Setting this option disables
6288 optimizations that may change the number of exceptions visible with
6289 signaling NaNs. This option implies @option{-ftrapping-math}.
6291 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
6294 The default is @option{-fno-signaling-nans}.
6296 This option is experimental and does not currently guarantee to
6297 disable all GCC optimizations that affect signaling NaN behavior.
6299 @item -fsingle-precision-constant
6300 @opindex fsingle-precision-constant
6301 Treat floating point constant as single precision constant instead of
6302 implicitly converting it to double precision constant.
6304 @item -fcx-limited-range
6305 @itemx -fno-cx-limited-range
6306 @opindex fcx-limited-range
6307 @opindex fno-cx-limited-range
6308 When enabled, this option states that a range reduction step is not
6309 needed when performing complex division. The default is
6310 @option{-fno-cx-limited-range}, but is enabled by @option{-ffast-math}.
6312 This option controls the default setting of the ISO C99
6313 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
6318 The following options control optimizations that may improve
6319 performance, but are not enabled by any @option{-O} options. This
6320 section includes experimental options that may produce broken code.
6323 @item -fbranch-probabilities
6324 @opindex fbranch-probabilities
6325 After running a program compiled with @option{-fprofile-arcs}
6326 (@pxref{Debugging Options,, Options for Debugging Your Program or
6327 @command{gcc}}), you can compile it a second time using
6328 @option{-fbranch-probabilities}, to improve optimizations based on
6329 the number of times each branch was taken. When the program
6330 compiled with @option{-fprofile-arcs} exits it saves arc execution
6331 counts to a file called @file{@var{sourcename}.gcda} for each source
6332 file. The information in this data file is very dependent on the
6333 structure of the generated code, so you must use the same source code
6334 and the same optimization options for both compilations.
6336 With @option{-fbranch-probabilities}, GCC puts a
6337 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
6338 These can be used to improve optimization. Currently, they are only
6339 used in one place: in @file{reorg.c}, instead of guessing which path a
6340 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
6341 exactly determine which path is taken more often.
6343 @item -fprofile-values
6344 @opindex fprofile-values
6345 If combined with @option{-fprofile-arcs}, it adds code so that some
6346 data about values of expressions in the program is gathered.
6348 With @option{-fbranch-probabilities}, it reads back the data gathered
6349 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
6350 notes to instructions for their later usage in optimizations.
6352 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
6356 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
6357 a code to gather information about values of expressions.
6359 With @option{-fbranch-probabilities}, it reads back the data gathered
6360 and actually performs the optimizations based on them.
6361 Currently the optimizations include specialization of division operation
6362 using the knowledge about the value of the denominator.
6364 @item -frename-registers
6365 @opindex frename-registers
6366 Attempt to avoid false dependencies in scheduled code by making use
6367 of registers left over after register allocation. This optimization
6368 will most benefit processors with lots of registers. Depending on the
6369 debug information format adopted by the target, however, it can
6370 make debugging impossible, since variables will no longer stay in
6371 a ``home register''.
6373 Enabled by default with @option{-funroll-loops}.
6377 Perform tail duplication to enlarge superblock size. This transformation
6378 simplifies the control flow of the function allowing other optimizations to do
6381 Enabled with @option{-fprofile-use}.
6383 @item -funroll-loops
6384 @opindex funroll-loops
6385 Unroll loops whose number of iterations can be determined at compile time or
6386 upon entry to the loop. @option{-funroll-loops} implies
6387 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
6388 It also turns on complete loop peeling (i.e.@: complete removal of loops with
6389 small constant number of iterations). This option makes code larger, and may
6390 or may not make it run faster.
6392 Enabled with @option{-fprofile-use}.
6394 @item -funroll-all-loops
6395 @opindex funroll-all-loops
6396 Unroll all loops, even if their number of iterations is uncertain when
6397 the loop is entered. This usually makes programs run more slowly.
6398 @option{-funroll-all-loops} implies the same options as
6399 @option{-funroll-loops}.
6402 @opindex fpeel-loops
6403 Peels the loops for that there is enough information that they do not
6404 roll much (from profile feedback). It also turns on complete loop peeling
6405 (i.e.@: complete removal of loops with small constant number of iterations).
6407 Enabled with @option{-fprofile-use}.
6409 @item -fmove-loop-invariants
6410 @opindex fmove-loop-invariants
6411 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
6412 at level @option{-O1}
6414 @item -funswitch-loops
6415 @opindex funswitch-loops
6416 Move branches with loop invariant conditions out of the loop, with duplicates
6417 of the loop on both branches (modified according to result of the condition).
6419 @item -ffunction-sections
6420 @itemx -fdata-sections
6421 @opindex ffunction-sections
6422 @opindex fdata-sections
6423 Place each function or data item into its own section in the output
6424 file if the target supports arbitrary sections. The name of the
6425 function or the name of the data item determines the section's name
6428 Use these options on systems where the linker can perform optimizations
6429 to improve locality of reference in the instruction space. Most systems
6430 using the ELF object format and SPARC processors running Solaris 2 have
6431 linkers with such optimizations. AIX may have these optimizations in
6434 Only use these options when there are significant benefits from doing
6435 so. When you specify these options, the assembler and linker will
6436 create larger object and executable files and will also be slower.
6437 You will not be able to use @code{gprof} on all systems if you
6438 specify this option and you may have problems with debugging if
6439 you specify both this option and @option{-g}.
6441 @item -fbranch-target-load-optimize
6442 @opindex fbranch-target-load-optimize
6443 Perform branch target register load optimization before prologue / epilogue
6445 The use of target registers can typically be exposed only during reload,
6446 thus hoisting loads out of loops and doing inter-block scheduling needs
6447 a separate optimization pass.
6449 @item -fbranch-target-load-optimize2
6450 @opindex fbranch-target-load-optimize2
6451 Perform branch target register load optimization after prologue / epilogue
6454 @item -fbtr-bb-exclusive
6455 @opindex fbtr-bb-exclusive
6456 When performing branch target register load optimization, don't reuse
6457 branch target registers in within any basic block.
6459 @item -fstack-protector
6460 Emit extra code to check for buffer overflows, such as stack smashing
6461 attacks. This is done by adding a guard variable to functions with
6462 vulnerable objects. This includes functions that call alloca, and
6463 functions with buffers larger than 8 bytes. The guards are initialized
6464 when a function is entered and then checked when the function exits.
6465 If a guard check fails, an error message is printed and the program exits.
6467 @item -fstack-protector-all
6468 Like @option{-fstack-protector} except that all functions are protected.
6470 @item -fsection-anchors
6471 @opindex fsection-anchors
6472 Try to reduce the number of symbolic address calculations by using
6473 shared ``anchor'' symbols to address nearby objects. This transformation
6474 can help to reduce the number of GOT entries and GOT accesses on some
6477 For example, the implementation of the following function @code{foo}:
6481 int foo (void) @{ return a + b + c; @}
6484 would usually calculate the addresses of all three variables, but if you
6485 compile it with @option{-fsection-anchors}, it will access the variables
6486 from a common anchor point instead. The effect is similar to the
6487 following pseudocode (which isn't valid C):
6492 register int *xr = &x;
6493 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
6497 Not all targets support this option.
6499 @item --param @var{name}=@var{value}
6501 In some places, GCC uses various constants to control the amount of
6502 optimization that is done. For example, GCC will not inline functions
6503 that contain more that a certain number of instructions. You can
6504 control some of these constants on the command-line using the
6505 @option{--param} option.
6507 The names of specific parameters, and the meaning of the values, are
6508 tied to the internals of the compiler, and are subject to change
6509 without notice in future releases.
6511 In each case, the @var{value} is an integer. The allowable choices for
6512 @var{name} are given in the following table:
6515 @item salias-max-implicit-fields
6516 The maximum number of fields in a variable without direct
6517 structure accesses for which structure aliasing will consider trying
6518 to track each field. The default is 5
6520 @item salias-max-array-elements
6521 The maximum number of elements an array can have and its elements
6522 still be tracked individually by structure aliasing. The default is 4
6524 @item sra-max-structure-size
6525 The maximum structure size, in bytes, at which the scalar replacement
6526 of aggregates (SRA) optimization will perform block copies. The
6527 default value, 0, implies that GCC will select the most appropriate
6530 @item sra-field-structure-ratio
6531 The threshold ratio (as a percentage) between instantiated fields and
6532 the complete structure size. We say that if the ratio of the number
6533 of bytes in instantiated fields to the number of bytes in the complete
6534 structure exceeds this parameter, then block copies are not used. The
6537 @item max-crossjump-edges
6538 The maximum number of incoming edges to consider for crossjumping.
6539 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
6540 the number of edges incoming to each block. Increasing values mean
6541 more aggressive optimization, making the compile time increase with
6542 probably small improvement in executable size.
6544 @item min-crossjump-insns
6545 The minimum number of instructions which must be matched at the end
6546 of two blocks before crossjumping will be performed on them. This
6547 value is ignored in the case where all instructions in the block being
6548 crossjumped from are matched. The default value is 5.
6550 @item max-grow-copy-bb-insns
6551 The maximum code size expansion factor when copying basic blocks
6552 instead of jumping. The expansion is relative to a jump instruction.
6553 The default value is 8.
6555 @item max-goto-duplication-insns
6556 The maximum number of instructions to duplicate to a block that jumps
6557 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
6558 passes, GCC factors computed gotos early in the compilation process,
6559 and unfactors them as late as possible. Only computed jumps at the
6560 end of a basic blocks with no more than max-goto-duplication-insns are
6561 unfactored. The default value is 8.
6563 @item max-delay-slot-insn-search
6564 The maximum number of instructions to consider when looking for an
6565 instruction to fill a delay slot. If more than this arbitrary number of
6566 instructions is searched, the time savings from filling the delay slot
6567 will be minimal so stop searching. Increasing values mean more
6568 aggressive optimization, making the compile time increase with probably
6569 small improvement in executable run time.
6571 @item max-delay-slot-live-search
6572 When trying to fill delay slots, the maximum number of instructions to
6573 consider when searching for a block with valid live register
6574 information. Increasing this arbitrarily chosen value means more
6575 aggressive optimization, increasing the compile time. This parameter
6576 should be removed when the delay slot code is rewritten to maintain the
6579 @item max-gcse-memory
6580 The approximate maximum amount of memory that will be allocated in
6581 order to perform the global common subexpression elimination
6582 optimization. If more memory than specified is required, the
6583 optimization will not be done.
6585 @item max-gcse-passes
6586 The maximum number of passes of GCSE to run. The default is 1.
6588 @item max-pending-list-length
6589 The maximum number of pending dependencies scheduling will allow
6590 before flushing the current state and starting over. Large functions
6591 with few branches or calls can create excessively large lists which
6592 needlessly consume memory and resources.
6594 @item max-inline-insns-single
6595 Several parameters control the tree inliner used in gcc.
6596 This number sets the maximum number of instructions (counted in GCC's
6597 internal representation) in a single function that the tree inliner
6598 will consider for inlining. This only affects functions declared
6599 inline and methods implemented in a class declaration (C++).
6600 The default value is 450.
6602 @item max-inline-insns-auto
6603 When you use @option{-finline-functions} (included in @option{-O3}),
6604 a lot of functions that would otherwise not be considered for inlining
6605 by the compiler will be investigated. To those functions, a different
6606 (more restrictive) limit compared to functions declared inline can
6608 The default value is 90.
6610 @item large-function-insns
6611 The limit specifying really large functions. For functions larger than this
6612 limit after inlining inlining is constrained by
6613 @option{--param large-function-growth}. This parameter is useful primarily
6614 to avoid extreme compilation time caused by non-linear algorithms used by the
6616 This parameter is ignored when @option{-funit-at-a-time} is not used.
6617 The default value is 2700.
6619 @item large-function-growth
6620 Specifies maximal growth of large function caused by inlining in percents.
6621 This parameter is ignored when @option{-funit-at-a-time} is not used.
6622 The default value is 100 which limits large function growth to 2.0 times
6625 @item large-unit-insns
6626 The limit specifying large translation unit. Growth caused by inlining of
6627 units larger than this limit is limited by @option{--param inline-unit-growth}.
6628 For small units this might be too tight (consider unit consisting of function A
6629 that is inline and B that just calls A three time. If B is small relative to
6630 A, the growth of unit is 300\% and yet such inlining is very sane. For very
6631 large units consisting of small inlininable functions however the overall unit
6632 growth limit is needed to avoid exponential explosion of code size. Thus for
6633 smaller units, the size is increased to @option{--param large-unit-insns}
6634 before applying @option{--param inline-unit-growth}. The default is 10000
6636 @item inline-unit-growth
6637 Specifies maximal overall growth of the compilation unit caused by inlining.
6638 This parameter is ignored when @option{-funit-at-a-time} is not used.
6639 The default value is 30 which limits unit growth to 1.3 times the original
6642 @item large-stack-frame
6643 The limit specifying large stack frames. While inlining the algorithm is trying
6644 to not grow past this limit too much. Default value is 256 bytes.
6646 @item large-stack-frame-growth
6647 Specifies maximal growth of large stack frames caused by inlining in percents.
6648 The default value is 1000 which limits large stack frame growth to 11 times
6651 @item max-inline-insns-recursive
6652 @itemx max-inline-insns-recursive-auto
6653 Specifies maximum number of instructions out-of-line copy of self recursive inline
6654 function can grow into by performing recursive inlining.
6656 For functions declared inline @option{--param max-inline-insns-recursive} is
6657 taken into account. For function not declared inline, recursive inlining
6658 happens only when @option{-finline-functions} (included in @option{-O3}) is
6659 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
6660 default value is 450.
6662 @item max-inline-recursive-depth
6663 @itemx max-inline-recursive-depth-auto
6664 Specifies maximum recursion depth used by the recursive inlining.
6666 For functions declared inline @option{--param max-inline-recursive-depth} is
6667 taken into account. For function not declared inline, recursive inlining
6668 happens only when @option{-finline-functions} (included in @option{-O3}) is
6669 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
6670 default value is 450.
6672 @item min-inline-recursive-probability
6673 Recursive inlining is profitable only for function having deep recursion
6674 in average and can hurt for function having little recursion depth by
6675 increasing the prologue size or complexity of function body to other
6678 When profile feedback is available (see @option{-fprofile-generate}) the actual
6679 recursion depth can be guessed from probability that function will recurse via
6680 given call expression. This parameter limits inlining only to call expression
6681 whose probability exceeds given threshold (in percents). The default value is
6684 @item inline-call-cost
6685 Specify cost of call instruction relative to simple arithmetics operations
6686 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
6687 functions and at the same time increases size of leaf function that is believed to
6688 reduce function size by being inlined. In effect it increases amount of
6689 inlining for code having large abstraction penalty (many functions that just
6690 pass the arguments to other functions) and decrease inlining for code with low
6691 abstraction penalty. The default value is 16.
6693 @item min-vect-loop-bound
6694 The minimum number of iterations under which a loop will not get vectorized
6695 when @option{-ftree-vectorize} is used. The number of iterations after
6696 vectorization needs to be greater than the value specified by this option
6697 to allow vectorization. The default value is 0.
6699 @item max-unrolled-insns
6700 The maximum number of instructions that a loop should have if that loop
6701 is unrolled, and if the loop is unrolled, it determines how many times
6702 the loop code is unrolled.
6704 @item max-average-unrolled-insns
6705 The maximum number of instructions biased by probabilities of their execution
6706 that a loop should have if that loop is unrolled, and if the loop is unrolled,
6707 it determines how many times the loop code is unrolled.
6709 @item max-unroll-times
6710 The maximum number of unrollings of a single loop.
6712 @item max-peeled-insns
6713 The maximum number of instructions that a loop should have if that loop
6714 is peeled, and if the loop is peeled, it determines how many times
6715 the loop code is peeled.
6717 @item max-peel-times
6718 The maximum number of peelings of a single loop.
6720 @item max-completely-peeled-insns
6721 The maximum number of insns of a completely peeled loop.
6723 @item max-completely-peel-times
6724 The maximum number of iterations of a loop to be suitable for complete peeling.
6726 @item max-unswitch-insns
6727 The maximum number of insns of an unswitched loop.
6729 @item max-unswitch-level
6730 The maximum number of branches unswitched in a single loop.
6733 The minimum cost of an expensive expression in the loop invariant motion.
6735 @item iv-consider-all-candidates-bound
6736 Bound on number of candidates for induction variables below that
6737 all candidates are considered for each use in induction variable
6738 optimizations. Only the most relevant candidates are considered
6739 if there are more candidates, to avoid quadratic time complexity.
6741 @item iv-max-considered-uses
6742 The induction variable optimizations give up on loops that contain more
6743 induction variable uses.
6745 @item iv-always-prune-cand-set-bound
6746 If number of candidates in the set is smaller than this value,
6747 we always try to remove unnecessary ivs from the set during its
6748 optimization when a new iv is added to the set.
6750 @item scev-max-expr-size
6751 Bound on size of expressions used in the scalar evolutions analyzer.
6752 Large expressions slow the analyzer.
6754 @item omega-max-vars
6755 The maximum number of variables in an Omega constraint system.
6756 The default value is 128.
6758 @item omega-max-geqs
6759 The maximum number of inequalities in an Omega constraint system.
6760 The default value is 256.
6763 The maximum number of equalities in an Omega constraint system.
6764 The default value is 128.
6766 @item omega-max-wild-cards
6767 The maximum number of wildcard variables that the Omega solver will
6768 be able to insert. The default value is 18.
6770 @item omega-hash-table-size
6771 The size of the hash table in the Omega solver. The default value is
6774 @item omega-max-keys
6775 The maximal number of keys used by the Omega solver. The default
6778 @item omega-eliminate-redundant-constraints
6779 When set to 1, use expensive methods to eliminate all redundant
6780 constraints. The default value is 0.
6782 @item vect-max-version-for-alignment-checks
6783 The maximum number of runtime checks that can be performed when
6784 doing loop versioning for alignment in the vectorizer. See option
6785 ftree-vect-loop-version for more information.
6787 @item vect-max-version-for-alias-checks
6788 The maximum number of runtime checks that can be performed when
6789 doing loop versioning for alias in the vectorizer. See option
6790 ftree-vect-loop-version for more information.
6792 @item max-iterations-to-track
6794 The maximum number of iterations of a loop the brute force algorithm
6795 for analysis of # of iterations of the loop tries to evaluate.
6797 @item hot-bb-count-fraction
6798 Select fraction of the maximal count of repetitions of basic block in program
6799 given basic block needs to have to be considered hot.
6801 @item hot-bb-frequency-fraction
6802 Select fraction of the maximal frequency of executions of basic block in
6803 function given basic block needs to have to be considered hot
6805 @item max-predicted-iterations
6806 The maximum number of loop iterations we predict statically. This is useful
6807 in cases where function contain single loop with known bound and other loop
6808 with unknown. We predict the known number of iterations correctly, while
6809 the unknown number of iterations average to roughly 10. This means that the
6810 loop without bounds would appear artificially cold relative to the other one.
6812 @item tracer-dynamic-coverage
6813 @itemx tracer-dynamic-coverage-feedback
6815 This value is used to limit superblock formation once the given percentage of
6816 executed instructions is covered. This limits unnecessary code size
6819 The @option{tracer-dynamic-coverage-feedback} is used only when profile
6820 feedback is available. The real profiles (as opposed to statically estimated
6821 ones) are much less balanced allowing the threshold to be larger value.
6823 @item tracer-max-code-growth
6824 Stop tail duplication once code growth has reached given percentage. This is
6825 rather hokey argument, as most of the duplicates will be eliminated later in
6826 cross jumping, so it may be set to much higher values than is the desired code
6829 @item tracer-min-branch-ratio
6831 Stop reverse growth when the reverse probability of best edge is less than this
6832 threshold (in percent).
6834 @item tracer-min-branch-ratio
6835 @itemx tracer-min-branch-ratio-feedback
6837 Stop forward growth if the best edge do have probability lower than this
6840 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
6841 compilation for profile feedback and one for compilation without. The value
6842 for compilation with profile feedback needs to be more conservative (higher) in
6843 order to make tracer effective.
6845 @item max-cse-path-length
6847 Maximum number of basic blocks on path that cse considers. The default is 10.
6850 The maximum instructions CSE process before flushing. The default is 1000.
6852 @item max-aliased-vops
6854 Maximum number of virtual operands per function allowed to represent
6855 aliases before triggering the alias partitioning heuristic. Alias
6856 partitioning reduces compile times and memory consumption needed for
6857 aliasing at the expense of precision loss in alias information. The
6858 default value for this parameter is 100 for -O1, 500 for -O2 and 1000
6861 Notice that if a function contains more memory statements than the
6862 value of this parameter, it is not really possible to achieve this
6863 reduction. In this case, the compiler will use the number of memory
6864 statements as the value for @option{max-aliased-vops}.
6866 @item avg-aliased-vops
6868 Average number of virtual operands per statement allowed to represent
6869 aliases before triggering the alias partitioning heuristic. This
6870 works in conjunction with @option{max-aliased-vops}. If a function
6871 contains more than @option{max-aliased-vops} virtual operators, then
6872 memory symbols will be grouped into memory partitions until either the
6873 total number of virtual operators is below @option{max-aliased-vops}
6874 or the average number of virtual operators per memory statement is
6875 below @option{avg-aliased-vops}. The default value for this parameter
6876 is 1 for -O1 and -O2, and 3 for -O3.
6878 @item ggc-min-expand
6880 GCC uses a garbage collector to manage its own memory allocation. This
6881 parameter specifies the minimum percentage by which the garbage
6882 collector's heap should be allowed to expand between collections.
6883 Tuning this may improve compilation speed; it has no effect on code
6886 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
6887 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
6888 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
6889 GCC is not able to calculate RAM on a particular platform, the lower
6890 bound of 30% is used. Setting this parameter and
6891 @option{ggc-min-heapsize} to zero causes a full collection to occur at
6892 every opportunity. This is extremely slow, but can be useful for
6895 @item ggc-min-heapsize
6897 Minimum size of the garbage collector's heap before it begins bothering
6898 to collect garbage. The first collection occurs after the heap expands
6899 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
6900 tuning this may improve compilation speed, and has no effect on code
6903 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
6904 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
6905 with a lower bound of 4096 (four megabytes) and an upper bound of
6906 131072 (128 megabytes). If GCC is not able to calculate RAM on a
6907 particular platform, the lower bound is used. Setting this parameter
6908 very large effectively disables garbage collection. Setting this
6909 parameter and @option{ggc-min-expand} to zero causes a full collection
6910 to occur at every opportunity.
6912 @item max-reload-search-insns
6913 The maximum number of instruction reload should look backward for equivalent
6914 register. Increasing values mean more aggressive optimization, making the
6915 compile time increase with probably slightly better performance. The default
6918 @item max-cselib-memory-locations
6919 The maximum number of memory locations cselib should take into account.
6920 Increasing values mean more aggressive optimization, making the compile time
6921 increase with probably slightly better performance. The default value is 500.
6923 @item max-flow-memory-locations
6924 Similar as @option{max-cselib-memory-locations} but for dataflow liveness.
6925 The default value is 100.
6927 @item reorder-blocks-duplicate
6928 @itemx reorder-blocks-duplicate-feedback
6930 Used by basic block reordering pass to decide whether to use unconditional
6931 branch or duplicate the code on its destination. Code is duplicated when its
6932 estimated size is smaller than this value multiplied by the estimated size of
6933 unconditional jump in the hot spots of the program.
6935 The @option{reorder-block-duplicate-feedback} is used only when profile
6936 feedback is available and may be set to higher values than
6937 @option{reorder-block-duplicate} since information about the hot spots is more
6940 @item max-sched-ready-insns
6941 The maximum number of instructions ready to be issued the scheduler should
6942 consider at any given time during the first scheduling pass. Increasing
6943 values mean more thorough searches, making the compilation time increase
6944 with probably little benefit. The default value is 100.
6946 @item max-sched-region-blocks
6947 The maximum number of blocks in a region to be considered for
6948 interblock scheduling. The default value is 10.
6950 @item max-sched-region-insns
6951 The maximum number of insns in a region to be considered for
6952 interblock scheduling. The default value is 100.
6955 The minimum probability (in percents) of reaching a source block
6956 for interblock speculative scheduling. The default value is 40.
6958 @item max-sched-extend-regions-iters
6959 The maximum number of iterations through CFG to extend regions.
6960 0 - disable region extension,
6961 N - do at most N iterations.
6962 The default value is 0.
6964 @item max-sched-insn-conflict-delay
6965 The maximum conflict delay for an insn to be considered for speculative motion.
6966 The default value is 3.
6968 @item sched-spec-prob-cutoff
6969 The minimal probability of speculation success (in percents), so that
6970 speculative insn will be scheduled.
6971 The default value is 40.
6973 @item max-last-value-rtl
6975 The maximum size measured as number of RTLs that can be recorded in an expression
6976 in combiner for a pseudo register as last known value of that register. The default
6979 @item integer-share-limit
6980 Small integer constants can use a shared data structure, reducing the
6981 compiler's memory usage and increasing its speed. This sets the maximum
6982 value of a shared integer constant's. The default value is 256.
6984 @item min-virtual-mappings
6985 Specifies the minimum number of virtual mappings in the incremental
6986 SSA updater that should be registered to trigger the virtual mappings
6987 heuristic defined by virtual-mappings-ratio. The default value is
6990 @item virtual-mappings-ratio
6991 If the number of virtual mappings is virtual-mappings-ratio bigger
6992 than the number of virtual symbols to be updated, then the incremental
6993 SSA updater switches to a full update for those symbols. The default
6996 @item ssp-buffer-size
6997 The minimum size of buffers (i.e. arrays) that will receive stack smashing
6998 protection when @option{-fstack-protection} is used.
7000 @item max-jump-thread-duplication-stmts
7001 Maximum number of statements allowed in a block that needs to be
7002 duplicated when threading jumps.
7004 @item max-fields-for-field-sensitive
7005 Maximum number of fields in a structure we will treat in
7006 a field sensitive manner during pointer analysis.
7008 @item prefetch-latency
7009 Estimate on average number of instructions that are executed before
7010 prefetch finishes. The distance we prefetch ahead is proportional
7011 to this constant. Increasing this number may also lead to less
7012 streams being prefetched (see @option{simultaneous-prefetches}).
7014 @item simultaneous-prefetches
7015 Maximum number of prefetches that can run at the same time.
7017 @item l1-cache-line-size
7018 The size of cache line in L1 cache, in bytes.
7021 The size of L1 cache, in kilobytes.
7024 The size of L2 cache, in kilobytes.
7026 @item use-canonical-types
7027 Whether the compiler should use the ``canonical'' type system. By
7028 default, this should always be 1, which uses a more efficient internal
7029 mechanism for comparing types in C++ and Objective-C++. However, if
7030 bugs in the canonical type system are causing compilation failures,
7031 set this value to 0 to disable canonical types.
7036 @node Preprocessor Options
7037 @section Options Controlling the Preprocessor
7038 @cindex preprocessor options
7039 @cindex options, preprocessor
7041 These options control the C preprocessor, which is run on each C source
7042 file before actual compilation.
7044 If you use the @option{-E} option, nothing is done except preprocessing.
7045 Some of these options make sense only together with @option{-E} because
7046 they cause the preprocessor output to be unsuitable for actual
7051 You can use @option{-Wp,@var{option}} to bypass the compiler driver
7052 and pass @var{option} directly through to the preprocessor. If
7053 @var{option} contains commas, it is split into multiple options at the
7054 commas. However, many options are modified, translated or interpreted
7055 by the compiler driver before being passed to the preprocessor, and
7056 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
7057 interface is undocumented and subject to change, so whenever possible
7058 you should avoid using @option{-Wp} and let the driver handle the
7061 @item -Xpreprocessor @var{option}
7062 @opindex preprocessor
7063 Pass @var{option} as an option to the preprocessor. You can use this to
7064 supply system-specific preprocessor options which GCC does not know how to
7067 If you want to pass an option that takes an argument, you must use
7068 @option{-Xpreprocessor} twice, once for the option and once for the argument.
7071 @include cppopts.texi
7073 @node Assembler Options
7074 @section Passing Options to the Assembler
7076 @c prevent bad page break with this line
7077 You can pass options to the assembler.
7080 @item -Wa,@var{option}
7082 Pass @var{option} as an option to the assembler. If @var{option}
7083 contains commas, it is split into multiple options at the commas.
7085 @item -Xassembler @var{option}
7087 Pass @var{option} as an option to the assembler. You can use this to
7088 supply system-specific assembler options which GCC does not know how to
7091 If you want to pass an option that takes an argument, you must use
7092 @option{-Xassembler} twice, once for the option and once for the argument.
7097 @section Options for Linking
7098 @cindex link options
7099 @cindex options, linking
7101 These options come into play when the compiler links object files into
7102 an executable output file. They are meaningless if the compiler is
7103 not doing a link step.
7107 @item @var{object-file-name}
7108 A file name that does not end in a special recognized suffix is
7109 considered to name an object file or library. (Object files are
7110 distinguished from libraries by the linker according to the file
7111 contents.) If linking is done, these object files are used as input
7120 If any of these options is used, then the linker is not run, and
7121 object file names should not be used as arguments. @xref{Overall
7125 @item -l@var{library}
7126 @itemx -l @var{library}
7128 Search the library named @var{library} when linking. (The second
7129 alternative with the library as a separate argument is only for
7130 POSIX compliance and is not recommended.)
7132 It makes a difference where in the command you write this option; the
7133 linker searches and processes libraries and object files in the order they
7134 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
7135 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
7136 to functions in @samp{z}, those functions may not be loaded.
7138 The linker searches a standard list of directories for the library,
7139 which is actually a file named @file{lib@var{library}.a}. The linker
7140 then uses this file as if it had been specified precisely by name.
7142 The directories searched include several standard system directories
7143 plus any that you specify with @option{-L}.
7145 Normally the files found this way are library files---archive files
7146 whose members are object files. The linker handles an archive file by
7147 scanning through it for members which define symbols that have so far
7148 been referenced but not defined. But if the file that is found is an
7149 ordinary object file, it is linked in the usual fashion. The only
7150 difference between using an @option{-l} option and specifying a file name
7151 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
7152 and searches several directories.
7156 You need this special case of the @option{-l} option in order to
7157 link an Objective-C or Objective-C++ program.
7160 @opindex nostartfiles
7161 Do not use the standard system startup files when linking.
7162 The standard system libraries are used normally, unless @option{-nostdlib}
7163 or @option{-nodefaultlibs} is used.
7165 @item -nodefaultlibs
7166 @opindex nodefaultlibs
7167 Do not use the standard system libraries when linking.
7168 Only the libraries you specify will be passed to the linker.
7169 The standard startup files are used normally, unless @option{-nostartfiles}
7170 is used. The compiler may generate calls to @code{memcmp},
7171 @code{memset}, @code{memcpy} and @code{memmove}.
7172 These entries are usually resolved by entries in
7173 libc. These entry points should be supplied through some other
7174 mechanism when this option is specified.
7178 Do not use the standard system startup files or libraries when linking.
7179 No startup files and only the libraries you specify will be passed to
7180 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
7181 @code{memcpy} and @code{memmove}.
7182 These entries are usually resolved by entries in
7183 libc. These entry points should be supplied through some other
7184 mechanism when this option is specified.
7186 @cindex @option{-lgcc}, use with @option{-nostdlib}
7187 @cindex @option{-nostdlib} and unresolved references
7188 @cindex unresolved references and @option{-nostdlib}
7189 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
7190 @cindex @option{-nodefaultlibs} and unresolved references
7191 @cindex unresolved references and @option{-nodefaultlibs}
7192 One of the standard libraries bypassed by @option{-nostdlib} and
7193 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
7194 that GCC uses to overcome shortcomings of particular machines, or special
7195 needs for some languages.
7196 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
7197 Collection (GCC) Internals},
7198 for more discussion of @file{libgcc.a}.)
7199 In most cases, you need @file{libgcc.a} even when you want to avoid
7200 other standard libraries. In other words, when you specify @option{-nostdlib}
7201 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
7202 This ensures that you have no unresolved references to internal GCC
7203 library subroutines. (For example, @samp{__main}, used to ensure C++
7204 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
7205 GNU Compiler Collection (GCC) Internals}.)
7209 Produce a position independent executable on targets which support it.
7210 For predictable results, you must also specify the same set of options
7211 that were used to generate code (@option{-fpie}, @option{-fPIE},
7212 or model suboptions) when you specify this option.
7216 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
7217 that support it. This instructs the linker to add all symbols, not
7218 only used ones, to the dynamic symbol table. This option is needed
7219 for some uses of @code{dlopen} or to allow obtaining backtraces
7220 from within a program.
7224 Remove all symbol table and relocation information from the executable.
7228 On systems that support dynamic linking, this prevents linking with the shared
7229 libraries. On other systems, this option has no effect.
7233 Produce a shared object which can then be linked with other objects to
7234 form an executable. Not all systems support this option. For predictable
7235 results, you must also specify the same set of options that were used to
7236 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
7237 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
7238 needs to build supplementary stub code for constructors to work. On
7239 multi-libbed systems, @samp{gcc -shared} must select the correct support
7240 libraries to link against. Failing to supply the correct flags may lead
7241 to subtle defects. Supplying them in cases where they are not necessary
7244 @item -shared-libgcc
7245 @itemx -static-libgcc
7246 @opindex shared-libgcc
7247 @opindex static-libgcc
7248 On systems that provide @file{libgcc} as a shared library, these options
7249 force the use of either the shared or static version respectively.
7250 If no shared version of @file{libgcc} was built when the compiler was
7251 configured, these options have no effect.
7253 There are several situations in which an application should use the
7254 shared @file{libgcc} instead of the static version. The most common
7255 of these is when the application wishes to throw and catch exceptions
7256 across different shared libraries. In that case, each of the libraries
7257 as well as the application itself should use the shared @file{libgcc}.
7259 Therefore, the G++ and GCJ drivers automatically add
7260 @option{-shared-libgcc} whenever you build a shared library or a main
7261 executable, because C++ and Java programs typically use exceptions, so
7262 this is the right thing to do.
7264 If, instead, you use the GCC driver to create shared libraries, you may
7265 find that they will not always be linked with the shared @file{libgcc}.
7266 If GCC finds, at its configuration time, that you have a non-GNU linker
7267 or a GNU linker that does not support option @option{--eh-frame-hdr},
7268 it will link the shared version of @file{libgcc} into shared libraries
7269 by default. Otherwise, it will take advantage of the linker and optimize
7270 away the linking with the shared version of @file{libgcc}, linking with
7271 the static version of libgcc by default. This allows exceptions to
7272 propagate through such shared libraries, without incurring relocation
7273 costs at library load time.
7275 However, if a library or main executable is supposed to throw or catch
7276 exceptions, you must link it using the G++ or GCJ driver, as appropriate
7277 for the languages used in the program, or using the option
7278 @option{-shared-libgcc}, such that it is linked with the shared
7283 Bind references to global symbols when building a shared object. Warn
7284 about any unresolved references (unless overridden by the link editor
7285 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
7288 @item -Xlinker @var{option}
7290 Pass @var{option} as an option to the linker. You can use this to
7291 supply system-specific linker options which GCC does not know how to
7294 If you want to pass an option that takes an argument, you must use
7295 @option{-Xlinker} twice, once for the option and once for the argument.
7296 For example, to pass @option{-assert definitions}, you must write
7297 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
7298 @option{-Xlinker "-assert definitions"}, because this passes the entire
7299 string as a single argument, which is not what the linker expects.
7301 @item -Wl,@var{option}
7303 Pass @var{option} as an option to the linker. If @var{option} contains
7304 commas, it is split into multiple options at the commas.
7306 @item -u @var{symbol}
7308 Pretend the symbol @var{symbol} is undefined, to force linking of
7309 library modules to define it. You can use @option{-u} multiple times with
7310 different symbols to force loading of additional library modules.
7313 @node Directory Options
7314 @section Options for Directory Search
7315 @cindex directory options
7316 @cindex options, directory search
7319 These options specify directories to search for header files, for
7320 libraries and for parts of the compiler:
7325 Add the directory @var{dir} to the head of the list of directories to be
7326 searched for header files. This can be used to override a system header
7327 file, substituting your own version, since these directories are
7328 searched before the system header file directories. However, you should
7329 not use this option to add directories that contain vendor-supplied
7330 system header files (use @option{-isystem} for that). If you use more than
7331 one @option{-I} option, the directories are scanned in left-to-right
7332 order; the standard system directories come after.
7334 If a standard system include directory, or a directory specified with
7335 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
7336 option will be ignored. The directory will still be searched but as a
7337 system directory at its normal position in the system include chain.
7338 This is to ensure that GCC's procedure to fix buggy system headers and
7339 the ordering for the include_next directive are not inadvertently changed.
7340 If you really need to change the search order for system directories,
7341 use the @option{-nostdinc} and/or @option{-isystem} options.
7343 @item -iquote@var{dir}
7345 Add the directory @var{dir} to the head of the list of directories to
7346 be searched for header files only for the case of @samp{#include
7347 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
7348 otherwise just like @option{-I}.
7352 Add directory @var{dir} to the list of directories to be searched
7355 @item -B@var{prefix}
7357 This option specifies where to find the executables, libraries,
7358 include files, and data files of the compiler itself.
7360 The compiler driver program runs one or more of the subprograms
7361 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
7362 @var{prefix} as a prefix for each program it tries to run, both with and
7363 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
7365 For each subprogram to be run, the compiler driver first tries the
7366 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
7367 was not specified, the driver tries two standard prefixes, which are
7368 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
7369 those results in a file name that is found, the unmodified program
7370 name is searched for using the directories specified in your
7371 @env{PATH} environment variable.
7373 The compiler will check to see if the path provided by the @option{-B}
7374 refers to a directory, and if necessary it will add a directory
7375 separator character at the end of the path.
7377 @option{-B} prefixes that effectively specify directory names also apply
7378 to libraries in the linker, because the compiler translates these
7379 options into @option{-L} options for the linker. They also apply to
7380 includes files in the preprocessor, because the compiler translates these
7381 options into @option{-isystem} options for the preprocessor. In this case,
7382 the compiler appends @samp{include} to the prefix.
7384 The run-time support file @file{libgcc.a} can also be searched for using
7385 the @option{-B} prefix, if needed. If it is not found there, the two
7386 standard prefixes above are tried, and that is all. The file is left
7387 out of the link if it is not found by those means.
7389 Another way to specify a prefix much like the @option{-B} prefix is to use
7390 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
7393 As a special kludge, if the path provided by @option{-B} is
7394 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
7395 9, then it will be replaced by @file{[dir/]include}. This is to help
7396 with boot-strapping the compiler.
7398 @item -specs=@var{file}
7400 Process @var{file} after the compiler reads in the standard @file{specs}
7401 file, in order to override the defaults that the @file{gcc} driver
7402 program uses when determining what switches to pass to @file{cc1},
7403 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
7404 @option{-specs=@var{file}} can be specified on the command line, and they
7405 are processed in order, from left to right.
7407 @item --sysroot=@var{dir}
7409 Use @var{dir} as the logical root directory for headers and libraries.
7410 For example, if the compiler would normally search for headers in
7411 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
7412 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
7414 If you use both this option and the @option{-isysroot} option, then
7415 the @option{--sysroot} option will apply to libraries, but the
7416 @option{-isysroot} option will apply to header files.
7418 The GNU linker (beginning with version 2.16) has the necessary support
7419 for this option. If your linker does not support this option, the
7420 header file aspect of @option{--sysroot} will still work, but the
7421 library aspect will not.
7425 This option has been deprecated. Please use @option{-iquote} instead for
7426 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
7427 Any directories you specify with @option{-I} options before the @option{-I-}
7428 option are searched only for the case of @samp{#include "@var{file}"};
7429 they are not searched for @samp{#include <@var{file}>}.
7431 If additional directories are specified with @option{-I} options after
7432 the @option{-I-}, these directories are searched for all @samp{#include}
7433 directives. (Ordinarily @emph{all} @option{-I} directories are used
7436 In addition, the @option{-I-} option inhibits the use of the current
7437 directory (where the current input file came from) as the first search
7438 directory for @samp{#include "@var{file}"}. There is no way to
7439 override this effect of @option{-I-}. With @option{-I.} you can specify
7440 searching the directory which was current when the compiler was
7441 invoked. That is not exactly the same as what the preprocessor does
7442 by default, but it is often satisfactory.
7444 @option{-I-} does not inhibit the use of the standard system directories
7445 for header files. Thus, @option{-I-} and @option{-nostdinc} are
7452 @section Specifying subprocesses and the switches to pass to them
7455 @command{gcc} is a driver program. It performs its job by invoking a
7456 sequence of other programs to do the work of compiling, assembling and
7457 linking. GCC interprets its command-line parameters and uses these to
7458 deduce which programs it should invoke, and which command-line options
7459 it ought to place on their command lines. This behavior is controlled
7460 by @dfn{spec strings}. In most cases there is one spec string for each
7461 program that GCC can invoke, but a few programs have multiple spec
7462 strings to control their behavior. The spec strings built into GCC can
7463 be overridden by using the @option{-specs=} command-line switch to specify
7466 @dfn{Spec files} are plaintext files that are used to construct spec
7467 strings. They consist of a sequence of directives separated by blank
7468 lines. The type of directive is determined by the first non-whitespace
7469 character on the line and it can be one of the following:
7472 @item %@var{command}
7473 Issues a @var{command} to the spec file processor. The commands that can
7477 @item %include <@var{file}>
7479 Search for @var{file} and insert its text at the current point in the
7482 @item %include_noerr <@var{file}>
7483 @cindex %include_noerr
7484 Just like @samp{%include}, but do not generate an error message if the include
7485 file cannot be found.
7487 @item %rename @var{old_name} @var{new_name}
7489 Rename the spec string @var{old_name} to @var{new_name}.
7493 @item *[@var{spec_name}]:
7494 This tells the compiler to create, override or delete the named spec
7495 string. All lines after this directive up to the next directive or
7496 blank line are considered to be the text for the spec string. If this
7497 results in an empty string then the spec will be deleted. (Or, if the
7498 spec did not exist, then nothing will happened.) Otherwise, if the spec
7499 does not currently exist a new spec will be created. If the spec does
7500 exist then its contents will be overridden by the text of this
7501 directive, unless the first character of that text is the @samp{+}
7502 character, in which case the text will be appended to the spec.
7504 @item [@var{suffix}]:
7505 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
7506 and up to the next directive or blank line are considered to make up the
7507 spec string for the indicated suffix. When the compiler encounters an
7508 input file with the named suffix, it will processes the spec string in
7509 order to work out how to compile that file. For example:
7516 This says that any input file whose name ends in @samp{.ZZ} should be
7517 passed to the program @samp{z-compile}, which should be invoked with the
7518 command-line switch @option{-input} and with the result of performing the
7519 @samp{%i} substitution. (See below.)
7521 As an alternative to providing a spec string, the text that follows a
7522 suffix directive can be one of the following:
7525 @item @@@var{language}
7526 This says that the suffix is an alias for a known @var{language}. This is
7527 similar to using the @option{-x} command-line switch to GCC to specify a
7528 language explicitly. For example:
7535 Says that .ZZ files are, in fact, C++ source files.
7538 This causes an error messages saying:
7541 @var{name} compiler not installed on this system.
7545 GCC already has an extensive list of suffixes built into it.
7546 This directive will add an entry to the end of the list of suffixes, but
7547 since the list is searched from the end backwards, it is effectively
7548 possible to override earlier entries using this technique.
7552 GCC has the following spec strings built into it. Spec files can
7553 override these strings or create their own. Note that individual
7554 targets can also add their own spec strings to this list.
7557 asm Options to pass to the assembler
7558 asm_final Options to pass to the assembler post-processor
7559 cpp Options to pass to the C preprocessor
7560 cc1 Options to pass to the C compiler
7561 cc1plus Options to pass to the C++ compiler
7562 endfile Object files to include at the end of the link
7563 link Options to pass to the linker
7564 lib Libraries to include on the command line to the linker
7565 libgcc Decides which GCC support library to pass to the linker
7566 linker Sets the name of the linker
7567 predefines Defines to be passed to the C preprocessor
7568 signed_char Defines to pass to CPP to say whether @code{char} is signed
7570 startfile Object files to include at the start of the link
7573 Here is a small example of a spec file:
7579 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
7582 This example renames the spec called @samp{lib} to @samp{old_lib} and
7583 then overrides the previous definition of @samp{lib} with a new one.
7584 The new definition adds in some extra command-line options before
7585 including the text of the old definition.
7587 @dfn{Spec strings} are a list of command-line options to be passed to their
7588 corresponding program. In addition, the spec strings can contain
7589 @samp{%}-prefixed sequences to substitute variable text or to
7590 conditionally insert text into the command line. Using these constructs
7591 it is possible to generate quite complex command lines.
7593 Here is a table of all defined @samp{%}-sequences for spec
7594 strings. Note that spaces are not generated automatically around the
7595 results of expanding these sequences. Therefore you can concatenate them
7596 together or combine them with constant text in a single argument.
7600 Substitute one @samp{%} into the program name or argument.
7603 Substitute the name of the input file being processed.
7606 Substitute the basename of the input file being processed.
7607 This is the substring up to (and not including) the last period
7608 and not including the directory.
7611 This is the same as @samp{%b}, but include the file suffix (text after
7615 Marks the argument containing or following the @samp{%d} as a
7616 temporary file name, so that that file will be deleted if GCC exits
7617 successfully. Unlike @samp{%g}, this contributes no text to the
7620 @item %g@var{suffix}
7621 Substitute a file name that has suffix @var{suffix} and is chosen
7622 once per compilation, and mark the argument in the same way as
7623 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
7624 name is now chosen in a way that is hard to predict even when previously
7625 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
7626 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
7627 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
7628 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
7629 was simply substituted with a file name chosen once per compilation,
7630 without regard to any appended suffix (which was therefore treated
7631 just like ordinary text), making such attacks more likely to succeed.
7633 @item %u@var{suffix}
7634 Like @samp{%g}, but generates a new temporary file name even if
7635 @samp{%u@var{suffix}} was already seen.
7637 @item %U@var{suffix}
7638 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
7639 new one if there is no such last file name. In the absence of any
7640 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
7641 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
7642 would involve the generation of two distinct file names, one
7643 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
7644 simply substituted with a file name chosen for the previous @samp{%u},
7645 without regard to any appended suffix.
7647 @item %j@var{suffix}
7648 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
7649 writable, and if save-temps is off; otherwise, substitute the name
7650 of a temporary file, just like @samp{%u}. This temporary file is not
7651 meant for communication between processes, but rather as a junk
7654 @item %|@var{suffix}
7655 @itemx %m@var{suffix}
7656 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
7657 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
7658 all. These are the two most common ways to instruct a program that it
7659 should read from standard input or write to standard output. If you
7660 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
7661 construct: see for example @file{f/lang-specs.h}.
7663 @item %.@var{SUFFIX}
7664 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
7665 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
7666 terminated by the next space or %.
7669 Marks the argument containing or following the @samp{%w} as the
7670 designated output file of this compilation. This puts the argument
7671 into the sequence of arguments that @samp{%o} will substitute later.
7674 Substitutes the names of all the output files, with spaces
7675 automatically placed around them. You should write spaces
7676 around the @samp{%o} as well or the results are undefined.
7677 @samp{%o} is for use in the specs for running the linker.
7678 Input files whose names have no recognized suffix are not compiled
7679 at all, but they are included among the output files, so they will
7683 Substitutes the suffix for object files. Note that this is
7684 handled specially when it immediately follows @samp{%g, %u, or %U},
7685 because of the need for those to form complete file names. The
7686 handling is such that @samp{%O} is treated exactly as if it had already
7687 been substituted, except that @samp{%g, %u, and %U} do not currently
7688 support additional @var{suffix} characters following @samp{%O} as they would
7689 following, for example, @samp{.o}.
7692 Substitutes the standard macro predefinitions for the
7693 current target machine. Use this when running @code{cpp}.
7696 Like @samp{%p}, but puts @samp{__} before and after the name of each
7697 predefined macro, except for macros that start with @samp{__} or with
7698 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
7702 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
7703 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
7704 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
7705 and @option{-imultilib} as necessary.
7708 Current argument is the name of a library or startup file of some sort.
7709 Search for that file in a standard list of directories and substitute
7710 the full name found.
7713 Print @var{str} as an error message. @var{str} is terminated by a newline.
7714 Use this when inconsistent options are detected.
7717 Substitute the contents of spec string @var{name} at this point.
7720 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
7722 @item %x@{@var{option}@}
7723 Accumulate an option for @samp{%X}.
7726 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
7730 Output the accumulated assembler options specified by @option{-Wa}.
7733 Output the accumulated preprocessor options specified by @option{-Wp}.
7736 Process the @code{asm} spec. This is used to compute the
7737 switches to be passed to the assembler.
7740 Process the @code{asm_final} spec. This is a spec string for
7741 passing switches to an assembler post-processor, if such a program is
7745 Process the @code{link} spec. This is the spec for computing the
7746 command line passed to the linker. Typically it will make use of the
7747 @samp{%L %G %S %D and %E} sequences.
7750 Dump out a @option{-L} option for each directory that GCC believes might
7751 contain startup files. If the target supports multilibs then the
7752 current multilib directory will be prepended to each of these paths.
7755 Process the @code{lib} spec. This is a spec string for deciding which
7756 libraries should be included on the command line to the linker.
7759 Process the @code{libgcc} spec. This is a spec string for deciding
7760 which GCC support library should be included on the command line to the linker.
7763 Process the @code{startfile} spec. This is a spec for deciding which
7764 object files should be the first ones passed to the linker. Typically
7765 this might be a file named @file{crt0.o}.
7768 Process the @code{endfile} spec. This is a spec string that specifies
7769 the last object files that will be passed to the linker.
7772 Process the @code{cpp} spec. This is used to construct the arguments
7773 to be passed to the C preprocessor.
7776 Process the @code{cc1} spec. This is used to construct the options to be
7777 passed to the actual C compiler (@samp{cc1}).
7780 Process the @code{cc1plus} spec. This is used to construct the options to be
7781 passed to the actual C++ compiler (@samp{cc1plus}).
7784 Substitute the variable part of a matched option. See below.
7785 Note that each comma in the substituted string is replaced by
7789 Remove all occurrences of @code{-S} from the command line. Note---this
7790 command is position dependent. @samp{%} commands in the spec string
7791 before this one will see @code{-S}, @samp{%} commands in the spec string
7792 after this one will not.
7794 @item %:@var{function}(@var{args})
7795 Call the named function @var{function}, passing it @var{args}.
7796 @var{args} is first processed as a nested spec string, then split
7797 into an argument vector in the usual fashion. The function returns
7798 a string which is processed as if it had appeared literally as part
7799 of the current spec.
7801 The following built-in spec functions are provided:
7805 The @code{getenv} spec function takes two arguments: an environment
7806 variable name and a string. If the environment variable is not
7807 defined, a fatal error is issued. Otherwise, the return value is the
7808 value of the environment variable concatenated with the string. For
7809 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
7812 %:getenv(TOPDIR /include)
7815 expands to @file{/path/to/top/include}.
7817 @item @code{if-exists}
7818 The @code{if-exists} spec function takes one argument, an absolute
7819 pathname to a file. If the file exists, @code{if-exists} returns the
7820 pathname. Here is a small example of its usage:
7824 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
7827 @item @code{if-exists-else}
7828 The @code{if-exists-else} spec function is similar to the @code{if-exists}
7829 spec function, except that it takes two arguments. The first argument is
7830 an absolute pathname to a file. If the file exists, @code{if-exists-else}
7831 returns the pathname. If it does not exist, it returns the second argument.
7832 This way, @code{if-exists-else} can be used to select one file or another,
7833 based on the existence of the first. Here is a small example of its usage:
7837 crt0%O%s %:if-exists(crti%O%s) \
7838 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
7841 @item @code{replace-outfile}
7842 The @code{replace-outfile} spec function takes two arguments. It looks for the
7843 first argument in the outfiles array and replaces it with the second argument. Here
7844 is a small example of its usage:
7847 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
7850 @item @code{print-asm-header}
7851 The @code{print-asm-header} function takes no arguments and simply
7852 prints a banner like:
7858 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
7861 It is used to separate compiler options from assembler options
7862 in the @option{--target-help} output.
7866 Substitutes the @code{-S} switch, if that switch was given to GCC@.
7867 If that switch was not specified, this substitutes nothing. Note that
7868 the leading dash is omitted when specifying this option, and it is
7869 automatically inserted if the substitution is performed. Thus the spec
7870 string @samp{%@{foo@}} would match the command-line option @option{-foo}
7871 and would output the command line option @option{-foo}.
7873 @item %W@{@code{S}@}
7874 Like %@{@code{S}@} but mark last argument supplied within as a file to be
7877 @item %@{@code{S}*@}
7878 Substitutes all the switches specified to GCC whose names start
7879 with @code{-S}, but which also take an argument. This is used for
7880 switches like @option{-o}, @option{-D}, @option{-I}, etc.
7881 GCC considers @option{-o foo} as being
7882 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
7883 text, including the space. Thus two arguments would be generated.
7885 @item %@{@code{S}*&@code{T}*@}
7886 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
7887 (the order of @code{S} and @code{T} in the spec is not significant).
7888 There can be any number of ampersand-separated variables; for each the
7889 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
7891 @item %@{@code{S}:@code{X}@}
7892 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
7894 @item %@{!@code{S}:@code{X}@}
7895 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
7897 @item %@{@code{S}*:@code{X}@}
7898 Substitutes @code{X} if one or more switches whose names start with
7899 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
7900 once, no matter how many such switches appeared. However, if @code{%*}
7901 appears somewhere in @code{X}, then @code{X} will be substituted once
7902 for each matching switch, with the @code{%*} replaced by the part of
7903 that switch that matched the @code{*}.
7905 @item %@{.@code{S}:@code{X}@}
7906 Substitutes @code{X}, if processing a file with suffix @code{S}.
7908 @item %@{!.@code{S}:@code{X}@}
7909 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
7911 @item %@{,@code{S}:@code{X}@}
7912 Substitutes @code{X}, if processing a file for language @code{S}.
7914 @item %@{!,@code{S}:@code{X}@}
7915 Substitutes @code{X}, if not processing a file for language @code{S}.
7917 @item %@{@code{S}|@code{P}:@code{X}@}
7918 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
7919 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
7920 @code{*} sequences as well, although they have a stronger binding than
7921 the @samp{|}. If @code{%*} appears in @code{X}, all of the
7922 alternatives must be starred, and only the first matching alternative
7925 For example, a spec string like this:
7928 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
7931 will output the following command-line options from the following input
7932 command-line options:
7937 -d fred.c -foo -baz -boggle
7938 -d jim.d -bar -baz -boggle
7941 @item %@{S:X; T:Y; :D@}
7943 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
7944 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
7945 be as many clauses as you need. This may be combined with @code{.},
7946 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
7951 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
7952 construct may contain other nested @samp{%} constructs or spaces, or
7953 even newlines. They are processed as usual, as described above.
7954 Trailing white space in @code{X} is ignored. White space may also
7955 appear anywhere on the left side of the colon in these constructs,
7956 except between @code{.} or @code{*} and the corresponding word.
7958 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
7959 handled specifically in these constructs. If another value of
7960 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
7961 @option{-W} switch is found later in the command line, the earlier
7962 switch value is ignored, except with @{@code{S}*@} where @code{S} is
7963 just one letter, which passes all matching options.
7965 The character @samp{|} at the beginning of the predicate text is used to
7966 indicate that a command should be piped to the following command, but
7967 only if @option{-pipe} is specified.
7969 It is built into GCC which switches take arguments and which do not.
7970 (You might think it would be useful to generalize this to allow each
7971 compiler's spec to say which switches take arguments. But this cannot
7972 be done in a consistent fashion. GCC cannot even decide which input
7973 files have been specified without knowing which switches take arguments,
7974 and it must know which input files to compile in order to tell which
7977 GCC also knows implicitly that arguments starting in @option{-l} are to be
7978 treated as compiler output files, and passed to the linker in their
7979 proper position among the other output files.
7981 @c man begin OPTIONS
7983 @node Target Options
7984 @section Specifying Target Machine and Compiler Version
7985 @cindex target options
7986 @cindex cross compiling
7987 @cindex specifying machine version
7988 @cindex specifying compiler version and target machine
7989 @cindex compiler version, specifying
7990 @cindex target machine, specifying
7992 The usual way to run GCC is to run the executable called @file{gcc}, or
7993 @file{<machine>-gcc} when cross-compiling, or
7994 @file{<machine>-gcc-<version>} to run a version other than the one that
7995 was installed last. Sometimes this is inconvenient, so GCC provides
7996 options that will switch to another cross-compiler or version.
7999 @item -b @var{machine}
8001 The argument @var{machine} specifies the target machine for compilation.
8003 The value to use for @var{machine} is the same as was specified as the
8004 machine type when configuring GCC as a cross-compiler. For
8005 example, if a cross-compiler was configured with @samp{configure
8006 arm-elf}, meaning to compile for an arm processor with elf binaries,
8007 then you would specify @option{-b arm-elf} to run that cross compiler.
8008 Because there are other options beginning with @option{-b}, the
8009 configuration must contain a hyphen.
8011 @item -V @var{version}
8013 The argument @var{version} specifies which version of GCC to run.
8014 This is useful when multiple versions are installed. For example,
8015 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
8018 The @option{-V} and @option{-b} options work by running the
8019 @file{<machine>-gcc-<version>} executable, so there's no real reason to
8020 use them if you can just run that directly.
8022 @node Submodel Options
8023 @section Hardware Models and Configurations
8024 @cindex submodel options
8025 @cindex specifying hardware config
8026 @cindex hardware models and configurations, specifying
8027 @cindex machine dependent options
8029 Earlier we discussed the standard option @option{-b} which chooses among
8030 different installed compilers for completely different target
8031 machines, such as VAX vs.@: 68000 vs.@: 80386.
8033 In addition, each of these target machine types can have its own
8034 special options, starting with @samp{-m}, to choose among various
8035 hardware models or configurations---for example, 68010 vs 68020,
8036 floating coprocessor or none. A single installed version of the
8037 compiler can compile for any model or configuration, according to the
8040 Some configurations of the compiler also support additional special
8041 options, usually for compatibility with other compilers on the same
8044 @c This list is ordered alphanumerically by subsection name.
8045 @c It should be the same order and spelling as these options are listed
8046 @c in Machine Dependent Options
8052 * Blackfin Options::
8056 * DEC Alpha Options::
8057 * DEC Alpha/VMS Options::
8059 * GNU/Linux Options::
8062 * i386 and x86-64 Options::
8075 * RS/6000 and PowerPC Options::
8076 * S/390 and zSeries Options::
8081 * System V Options::
8082 * TMS320C3x/C4x Options::
8087 * Xstormy16 Options::
8093 @subsection ARC Options
8096 These options are defined for ARC implementations:
8101 Compile code for little endian mode. This is the default.
8105 Compile code for big endian mode.
8108 @opindex mmangle-cpu
8109 Prepend the name of the cpu to all public symbol names.
8110 In multiple-processor systems, there are many ARC variants with different
8111 instruction and register set characteristics. This flag prevents code
8112 compiled for one cpu to be linked with code compiled for another.
8113 No facility exists for handling variants that are ``almost identical''.
8114 This is an all or nothing option.
8116 @item -mcpu=@var{cpu}
8118 Compile code for ARC variant @var{cpu}.
8119 Which variants are supported depend on the configuration.
8120 All variants support @option{-mcpu=base}, this is the default.
8122 @item -mtext=@var{text-section}
8123 @itemx -mdata=@var{data-section}
8124 @itemx -mrodata=@var{readonly-data-section}
8128 Put functions, data, and readonly data in @var{text-section},
8129 @var{data-section}, and @var{readonly-data-section} respectively
8130 by default. This can be overridden with the @code{section} attribute.
8131 @xref{Variable Attributes}.
8136 @subsection ARM Options
8139 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
8143 @item -mabi=@var{name}
8145 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
8146 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
8149 @opindex mapcs-frame
8150 Generate a stack frame that is compliant with the ARM Procedure Call
8151 Standard for all functions, even if this is not strictly necessary for
8152 correct execution of the code. Specifying @option{-fomit-frame-pointer}
8153 with this option will cause the stack frames not to be generated for
8154 leaf functions. The default is @option{-mno-apcs-frame}.
8158 This is a synonym for @option{-mapcs-frame}.
8161 @c not currently implemented
8162 @item -mapcs-stack-check
8163 @opindex mapcs-stack-check
8164 Generate code to check the amount of stack space available upon entry to
8165 every function (that actually uses some stack space). If there is
8166 insufficient space available then either the function
8167 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
8168 called, depending upon the amount of stack space required. The run time
8169 system is required to provide these functions. The default is
8170 @option{-mno-apcs-stack-check}, since this produces smaller code.
8172 @c not currently implemented
8174 @opindex mapcs-float
8175 Pass floating point arguments using the float point registers. This is
8176 one of the variants of the APCS@. This option is recommended if the
8177 target hardware has a floating point unit or if a lot of floating point
8178 arithmetic is going to be performed by the code. The default is
8179 @option{-mno-apcs-float}, since integer only code is slightly increased in
8180 size if @option{-mapcs-float} is used.
8182 @c not currently implemented
8183 @item -mapcs-reentrant
8184 @opindex mapcs-reentrant
8185 Generate reentrant, position independent code. The default is
8186 @option{-mno-apcs-reentrant}.
8189 @item -mthumb-interwork
8190 @opindex mthumb-interwork
8191 Generate code which supports calling between the ARM and Thumb
8192 instruction sets. Without this option the two instruction sets cannot
8193 be reliably used inside one program. The default is
8194 @option{-mno-thumb-interwork}, since slightly larger code is generated
8195 when @option{-mthumb-interwork} is specified.
8197 @item -mno-sched-prolog
8198 @opindex mno-sched-prolog
8199 Prevent the reordering of instructions in the function prolog, or the
8200 merging of those instruction with the instructions in the function's
8201 body. This means that all functions will start with a recognizable set
8202 of instructions (or in fact one of a choice from a small set of
8203 different function prologues), and this information can be used to
8204 locate the start if functions inside an executable piece of code. The
8205 default is @option{-msched-prolog}.
8208 @opindex mhard-float
8209 Generate output containing floating point instructions. This is the
8213 @opindex msoft-float
8214 Generate output containing library calls for floating point.
8215 @strong{Warning:} the requisite libraries are not available for all ARM
8216 targets. Normally the facilities of the machine's usual C compiler are
8217 used, but this cannot be done directly in cross-compilation. You must make
8218 your own arrangements to provide suitable library functions for
8221 @option{-msoft-float} changes the calling convention in the output file;
8222 therefore, it is only useful if you compile @emph{all} of a program with
8223 this option. In particular, you need to compile @file{libgcc.a}, the
8224 library that comes with GCC, with @option{-msoft-float} in order for
8227 @item -mfloat-abi=@var{name}
8229 Specifies which ABI to use for floating point values. Permissible values
8230 are: @samp{soft}, @samp{softfp} and @samp{hard}.
8232 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
8233 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
8234 of floating point instructions, but still uses the soft-float calling
8237 @item -mlittle-endian
8238 @opindex mlittle-endian
8239 Generate code for a processor running in little-endian mode. This is
8240 the default for all standard configurations.
8243 @opindex mbig-endian
8244 Generate code for a processor running in big-endian mode; the default is
8245 to compile code for a little-endian processor.
8247 @item -mwords-little-endian
8248 @opindex mwords-little-endian
8249 This option only applies when generating code for big-endian processors.
8250 Generate code for a little-endian word order but a big-endian byte
8251 order. That is, a byte order of the form @samp{32107654}. Note: this
8252 option should only be used if you require compatibility with code for
8253 big-endian ARM processors generated by versions of the compiler prior to
8256 @item -mcpu=@var{name}
8258 This specifies the name of the target ARM processor. GCC uses this name
8259 to determine what kind of instructions it can emit when generating
8260 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
8261 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
8262 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
8263 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
8264 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
8265 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
8266 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
8267 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
8268 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
8269 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
8270 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
8271 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
8272 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
8273 @samp{arm1156t2-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
8274 @samp{cortex-a8}, @samp{cortex-r4}, @samp{cortex-m3},
8275 @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
8277 @itemx -mtune=@var{name}
8279 This option is very similar to the @option{-mcpu=} option, except that
8280 instead of specifying the actual target processor type, and hence
8281 restricting which instructions can be used, it specifies that GCC should
8282 tune the performance of the code as if the target were of the type
8283 specified in this option, but still choosing the instructions that it
8284 will generate based on the cpu specified by a @option{-mcpu=} option.
8285 For some ARM implementations better performance can be obtained by using
8288 @item -march=@var{name}
8290 This specifies the name of the target ARM architecture. GCC uses this
8291 name to determine what kind of instructions it can emit when generating
8292 assembly code. This option can be used in conjunction with or instead
8293 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
8294 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
8295 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
8296 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv7}, @samp{armv7-a},
8297 @samp{armv7-r}, @samp{armv7-m}, @samp{iwmmxt}, @samp{ep9312}.
8299 @item -mfpu=@var{name}
8300 @itemx -mfpe=@var{number}
8301 @itemx -mfp=@var{number}
8305 This specifies what floating point hardware (or hardware emulation) is
8306 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
8307 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
8308 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
8309 with older versions of GCC@.
8311 If @option{-msoft-float} is specified this specifies the format of
8312 floating point values.
8314 @item -mstructure-size-boundary=@var{n}
8315 @opindex mstructure-size-boundary
8316 The size of all structures and unions will be rounded up to a multiple
8317 of the number of bits set by this option. Permissible values are 8, 32
8318 and 64. The default value varies for different toolchains. For the COFF
8319 targeted toolchain the default value is 8. A value of 64 is only allowed
8320 if the underlying ABI supports it.
8322 Specifying the larger number can produce faster, more efficient code, but
8323 can also increase the size of the program. Different values are potentially
8324 incompatible. Code compiled with one value cannot necessarily expect to
8325 work with code or libraries compiled with another value, if they exchange
8326 information using structures or unions.
8328 @item -mabort-on-noreturn
8329 @opindex mabort-on-noreturn
8330 Generate a call to the function @code{abort} at the end of a
8331 @code{noreturn} function. It will be executed if the function tries to
8335 @itemx -mno-long-calls
8336 @opindex mlong-calls
8337 @opindex mno-long-calls
8338 Tells the compiler to perform function calls by first loading the
8339 address of the function into a register and then performing a subroutine
8340 call on this register. This switch is needed if the target function
8341 will lie outside of the 64 megabyte addressing range of the offset based
8342 version of subroutine call instruction.
8344 Even if this switch is enabled, not all function calls will be turned
8345 into long calls. The heuristic is that static functions, functions
8346 which have the @samp{short-call} attribute, functions that are inside
8347 the scope of a @samp{#pragma no_long_calls} directive and functions whose
8348 definitions have already been compiled within the current compilation
8349 unit, will not be turned into long calls. The exception to this rule is
8350 that weak function definitions, functions with the @samp{long-call}
8351 attribute or the @samp{section} attribute, and functions that are within
8352 the scope of a @samp{#pragma long_calls} directive, will always be
8353 turned into long calls.
8355 This feature is not enabled by default. Specifying
8356 @option{-mno-long-calls} will restore the default behavior, as will
8357 placing the function calls within the scope of a @samp{#pragma
8358 long_calls_off} directive. Note these switches have no effect on how
8359 the compiler generates code to handle function calls via function
8362 @item -mnop-fun-dllimport
8363 @opindex mnop-fun-dllimport
8364 Disable support for the @code{dllimport} attribute.
8366 @item -msingle-pic-base
8367 @opindex msingle-pic-base
8368 Treat the register used for PIC addressing as read-only, rather than
8369 loading it in the prologue for each function. The run-time system is
8370 responsible for initializing this register with an appropriate value
8371 before execution begins.
8373 @item -mpic-register=@var{reg}
8374 @opindex mpic-register
8375 Specify the register to be used for PIC addressing. The default is R10
8376 unless stack-checking is enabled, when R9 is used.
8378 @item -mcirrus-fix-invalid-insns
8379 @opindex mcirrus-fix-invalid-insns
8380 @opindex mno-cirrus-fix-invalid-insns
8381 Insert NOPs into the instruction stream to in order to work around
8382 problems with invalid Maverick instruction combinations. This option
8383 is only valid if the @option{-mcpu=ep9312} option has been used to
8384 enable generation of instructions for the Cirrus Maverick floating
8385 point co-processor. This option is not enabled by default, since the
8386 problem is only present in older Maverick implementations. The default
8387 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
8390 @item -mpoke-function-name
8391 @opindex mpoke-function-name
8392 Write the name of each function into the text section, directly
8393 preceding the function prologue. The generated code is similar to this:
8397 .ascii "arm_poke_function_name", 0
8400 .word 0xff000000 + (t1 - t0)
8401 arm_poke_function_name
8403 stmfd sp!, @{fp, ip, lr, pc@}
8407 When performing a stack backtrace, code can inspect the value of
8408 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
8409 location @code{pc - 12} and the top 8 bits are set, then we know that
8410 there is a function name embedded immediately preceding this location
8411 and has length @code{((pc[-3]) & 0xff000000)}.
8415 Generate code for the Thumb instruction set. The default is to
8416 use the 32-bit ARM instruction set.
8417 This option automatically enables either 16-bit Thumb-1 or
8418 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
8419 and @option{-march=@var{name}} options.
8422 @opindex mtpcs-frame
8423 Generate a stack frame that is compliant with the Thumb Procedure Call
8424 Standard for all non-leaf functions. (A leaf function is one that does
8425 not call any other functions.) The default is @option{-mno-tpcs-frame}.
8427 @item -mtpcs-leaf-frame
8428 @opindex mtpcs-leaf-frame
8429 Generate a stack frame that is compliant with the Thumb Procedure Call
8430 Standard for all leaf functions. (A leaf function is one that does
8431 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
8433 @item -mcallee-super-interworking
8434 @opindex mcallee-super-interworking
8435 Gives all externally visible functions in the file being compiled an ARM
8436 instruction set header which switches to Thumb mode before executing the
8437 rest of the function. This allows these functions to be called from
8438 non-interworking code.
8440 @item -mcaller-super-interworking
8441 @opindex mcaller-super-interworking
8442 Allows calls via function pointers (including virtual functions) to
8443 execute correctly regardless of whether the target code has been
8444 compiled for interworking or not. There is a small overhead in the cost
8445 of executing a function pointer if this option is enabled.
8447 @item -mtp=@var{name}
8449 Specify the access model for the thread local storage pointer. The valid
8450 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
8451 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
8452 (supported in the arm6k architecture), and @option{auto}, which uses the
8453 best available method for the selected processor. The default setting is
8459 @subsection AVR Options
8462 These options are defined for AVR implementations:
8465 @item -mmcu=@var{mcu}
8467 Specify ATMEL AVR instruction set or MCU type.
8469 Instruction set avr1 is for the minimal AVR core, not supported by the C
8470 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8471 attiny11, attiny12, attiny15, attiny28).
8473 Instruction set avr2 (default) is for the classic AVR core with up to
8474 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8475 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8476 at90c8534, at90s8535).
8478 Instruction set avr3 is for the classic AVR core with up to 128K program
8479 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8481 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8482 memory space (MCU types: atmega8, atmega83, atmega85).
8484 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8485 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8486 atmega64, atmega128, at43usb355, at94k).
8490 Output instruction sizes to the asm file.
8492 @item -minit-stack=@var{N}
8493 @opindex minit-stack
8494 Specify the initial stack address, which may be a symbol or numeric value,
8495 @samp{__stack} is the default.
8497 @item -mno-interrupts
8498 @opindex mno-interrupts
8499 Generated code is not compatible with hardware interrupts.
8500 Code size will be smaller.
8502 @item -mcall-prologues
8503 @opindex mcall-prologues
8504 Functions prologues/epilogues expanded as call to appropriate
8505 subroutines. Code size will be smaller.
8507 @item -mno-tablejump
8508 @opindex mno-tablejump
8509 Do not generate tablejump insns which sometimes increase code size.
8512 @opindex mtiny-stack
8513 Change only the low 8 bits of the stack pointer.
8517 Assume int to be 8 bit integer. This affects the sizes of all types: A
8518 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
8519 and long long will be 4 bytes. Please note that this option does not
8520 comply to the C standards, but it will provide you with smaller code
8524 @node Blackfin Options
8525 @subsection Blackfin Options
8526 @cindex Blackfin Options
8529 @item -mcpu=@var{cpu}
8531 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
8532 can be one of @samp{bf531}, @samp{bf532}, @samp{bf533},
8533 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf561}.
8534 Without this option, @samp{bf532} is used as the processor by default.
8535 The corresponding predefined processor macros for @var{cpu} is to
8536 be defined. For the @samp{bfin-elf} toolchain, this causes the hardware
8537 BSP provided by libgloss to be linked in if @samp{-msim} is not given.
8538 Support for @samp{bf561} is incomplete; only the processor macro is defined.
8542 Specifies that the program will be run on the simulator. This causes
8543 the simulator BSP provided by libgloss to be linked in. This option
8544 has effect only for @samp{bfin-elf} toolchain.
8546 @item -momit-leaf-frame-pointer
8547 @opindex momit-leaf-frame-pointer
8548 Don't keep the frame pointer in a register for leaf functions. This
8549 avoids the instructions to save, set up and restore frame pointers and
8550 makes an extra register available in leaf functions. The option
8551 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8552 which might make debugging harder.
8554 @item -mspecld-anomaly
8555 @opindex mspecld-anomaly
8556 When enabled, the compiler will ensure that the generated code does not
8557 contain speculative loads after jump instructions. This option is enabled
8560 @item -mno-specld-anomaly
8561 @opindex mno-specld-anomaly
8562 Don't generate extra code to prevent speculative loads from occurring.
8564 @item -mcsync-anomaly
8565 @opindex mcsync-anomaly
8566 When enabled, the compiler will ensure that the generated code does not
8567 contain CSYNC or SSYNC instructions too soon after conditional branches.
8568 This option is enabled by default.
8570 @item -mno-csync-anomaly
8571 @opindex mno-csync-anomaly
8572 Don't generate extra code to prevent CSYNC or SSYNC instructions from
8573 occurring too soon after a conditional branch.
8577 When enabled, the compiler is free to take advantage of the knowledge that
8578 the entire program fits into the low 64k of memory.
8581 @opindex mno-low-64k
8582 Assume that the program is arbitrarily large. This is the default.
8584 @item -mstack-check-l1
8585 @opindex mstack-check-l1
8586 Do stack checking using information placed into L1 scratchpad memory by the
8589 @item -mid-shared-library
8590 @opindex mid-shared-library
8591 Generate code that supports shared libraries via the library ID method.
8592 This allows for execute in place and shared libraries in an environment
8593 without virtual memory management. This option implies @option{-fPIC}.
8595 @item -mno-id-shared-library
8596 @opindex mno-id-shared-library
8597 Generate code that doesn't assume ID based shared libraries are being used.
8598 This is the default.
8600 @item -mleaf-id-shared-library
8601 @opindex mleaf-id-shared-library
8602 Generate code that supports shared libraries via the library ID method,
8603 but assumes that this library or executable won't link against any other
8604 ID shared libraries. That allows the compiler to use faster code for jumps
8607 @item -mno-leaf-id-shared-library
8608 @opindex mno-leaf-id-shared-library
8609 Do not assume that the code being compiled won't link against any ID shared
8610 libraries. Slower code will be generated for jump and call insns.
8612 @item -mshared-library-id=n
8613 @opindex mshared-library-id
8614 Specified the identification number of the ID based shared library being
8615 compiled. Specifying a value of 0 will generate more compact code, specifying
8616 other values will force the allocation of that number to the current
8617 library but is no more space or time efficient than omitting this option.
8621 Generate code that allows the data segment to be located in a different
8622 area of memory from the text segment. This allows for execute in place in
8623 an environment without virtual memory management by eliminating relocations
8624 against the text section.
8627 @opindex mno-sep-data
8628 Generate code that assumes that the data segment follows the text segment.
8629 This is the default.
8632 @itemx -mno-long-calls
8633 @opindex mlong-calls
8634 @opindex mno-long-calls
8635 Tells the compiler to perform function calls by first loading the
8636 address of the function into a register and then performing a subroutine
8637 call on this register. This switch is needed if the target function
8638 will lie outside of the 24 bit addressing range of the offset based
8639 version of subroutine call instruction.
8641 This feature is not enabled by default. Specifying
8642 @option{-mno-long-calls} will restore the default behavior. Note these
8643 switches have no effect on how the compiler generates code to handle
8644 function calls via function pointers.
8648 Link with the fast floating-point library. This library relaxes some of
8649 the IEEE floating-point standard's rules for checking inputs against
8650 Not-a-Number (NAN), in the interest of performance.
8653 @opindex minline-plt
8654 Enable inlining of PLT entries in function calls to functions that are
8655 not known to bind locally. It has no effect without @option{-mfdpic}.
8659 @subsection CRIS Options
8660 @cindex CRIS Options
8662 These options are defined specifically for the CRIS ports.
8665 @item -march=@var{architecture-type}
8666 @itemx -mcpu=@var{architecture-type}
8669 Generate code for the specified architecture. The choices for
8670 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
8671 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
8672 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
8675 @item -mtune=@var{architecture-type}
8677 Tune to @var{architecture-type} everything applicable about the generated
8678 code, except for the ABI and the set of available instructions. The
8679 choices for @var{architecture-type} are the same as for
8680 @option{-march=@var{architecture-type}}.
8682 @item -mmax-stack-frame=@var{n}
8683 @opindex mmax-stack-frame
8684 Warn when the stack frame of a function exceeds @var{n} bytes.
8686 @item -melinux-stacksize=@var{n}
8687 @opindex melinux-stacksize
8688 Only available with the @samp{cris-axis-aout} target. Arranges for
8689 indications in the program to the kernel loader that the stack of the
8690 program should be set to @var{n} bytes.
8696 The options @option{-metrax4} and @option{-metrax100} are synonyms for
8697 @option{-march=v3} and @option{-march=v8} respectively.
8699 @item -mmul-bug-workaround
8700 @itemx -mno-mul-bug-workaround
8701 @opindex mmul-bug-workaround
8702 @opindex mno-mul-bug-workaround
8703 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
8704 models where it applies. This option is active by default.
8708 Enable CRIS-specific verbose debug-related information in the assembly
8709 code. This option also has the effect to turn off the @samp{#NO_APP}
8710 formatted-code indicator to the assembler at the beginning of the
8715 Do not use condition-code results from previous instruction; always emit
8716 compare and test instructions before use of condition codes.
8718 @item -mno-side-effects
8719 @opindex mno-side-effects
8720 Do not emit instructions with side-effects in addressing modes other than
8724 @itemx -mno-stack-align
8726 @itemx -mno-data-align
8727 @itemx -mconst-align
8728 @itemx -mno-const-align
8729 @opindex mstack-align
8730 @opindex mno-stack-align
8731 @opindex mdata-align
8732 @opindex mno-data-align
8733 @opindex mconst-align
8734 @opindex mno-const-align
8735 These options (no-options) arranges (eliminate arrangements) for the
8736 stack-frame, individual data and constants to be aligned for the maximum
8737 single data access size for the chosen CPU model. The default is to
8738 arrange for 32-bit alignment. ABI details such as structure layout are
8739 not affected by these options.
8747 Similar to the stack- data- and const-align options above, these options
8748 arrange for stack-frame, writable data and constants to all be 32-bit,
8749 16-bit or 8-bit aligned. The default is 32-bit alignment.
8751 @item -mno-prologue-epilogue
8752 @itemx -mprologue-epilogue
8753 @opindex mno-prologue-epilogue
8754 @opindex mprologue-epilogue
8755 With @option{-mno-prologue-epilogue}, the normal function prologue and
8756 epilogue that sets up the stack-frame are omitted and no return
8757 instructions or return sequences are generated in the code. Use this
8758 option only together with visual inspection of the compiled code: no
8759 warnings or errors are generated when call-saved registers must be saved,
8760 or storage for local variable needs to be allocated.
8766 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
8767 instruction sequences that load addresses for functions from the PLT part
8768 of the GOT rather than (traditional on other architectures) calls to the
8769 PLT@. The default is @option{-mgotplt}.
8773 Legacy no-op option only recognized with the cris-axis-aout target.
8777 Legacy no-op option only recognized with the cris-axis-elf and
8778 cris-axis-linux-gnu targets.
8782 Only recognized with the cris-axis-aout target, where it selects a
8783 GNU/linux-like multilib, include files and instruction set for
8788 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
8792 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
8793 to link with input-output functions from a simulator library. Code,
8794 initialized data and zero-initialized data are allocated consecutively.
8798 Like @option{-sim}, but pass linker options to locate initialized data at
8799 0x40000000 and zero-initialized data at 0x80000000.
8803 @subsection CRX Options
8806 These options are defined specifically for the CRX ports.
8812 Enable the use of multiply-accumulate instructions. Disabled by default.
8816 Push instructions will be used to pass outgoing arguments when functions
8817 are called. Enabled by default.
8820 @node Darwin Options
8821 @subsection Darwin Options
8822 @cindex Darwin options
8824 These options are defined for all architectures running the Darwin operating
8827 FSF GCC on Darwin does not create ``fat'' object files; it will create
8828 an object file for the single architecture that it was built to
8829 target. Apple's GCC on Darwin does create ``fat'' files if multiple
8830 @option{-arch} options are used; it does so by running the compiler or
8831 linker multiple times and joining the results together with
8834 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
8835 @samp{i686}) is determined by the flags that specify the ISA
8836 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
8837 @option{-force_cpusubtype_ALL} option can be used to override this.
8839 The Darwin tools vary in their behavior when presented with an ISA
8840 mismatch. The assembler, @file{as}, will only permit instructions to
8841 be used that are valid for the subtype of the file it is generating,
8842 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
8843 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
8844 and print an error if asked to create a shared library with a less
8845 restrictive subtype than its input files (for instance, trying to put
8846 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
8847 for executables, @file{ld}, will quietly give the executable the most
8848 restrictive subtype of any of its input files.
8853 Add the framework directory @var{dir} to the head of the list of
8854 directories to be searched for header files. These directories are
8855 interleaved with those specified by @option{-I} options and are
8856 scanned in a left-to-right order.
8858 A framework directory is a directory with frameworks in it. A
8859 framework is a directory with a @samp{"Headers"} and/or
8860 @samp{"PrivateHeaders"} directory contained directly in it that ends
8861 in @samp{".framework"}. The name of a framework is the name of this
8862 directory excluding the @samp{".framework"}. Headers associated with
8863 the framework are found in one of those two directories, with
8864 @samp{"Headers"} being searched first. A subframework is a framework
8865 directory that is in a framework's @samp{"Frameworks"} directory.
8866 Includes of subframework headers can only appear in a header of a
8867 framework that contains the subframework, or in a sibling subframework
8868 header. Two subframeworks are siblings if they occur in the same
8869 framework. A subframework should not have the same name as a
8870 framework, a warning will be issued if this is violated. Currently a
8871 subframework cannot have subframeworks, in the future, the mechanism
8872 may be extended to support this. The standard frameworks can be found
8873 in @samp{"/System/Library/Frameworks"} and
8874 @samp{"/Library/Frameworks"}. An example include looks like
8875 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
8876 the name of the framework and header.h is found in the
8877 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
8879 @item -iframework@var{dir}
8881 Like @option{-F} except the directory is a treated as a system
8882 directory. The main difference between this @option{-iframework} and
8883 @option{-F} is that with @option{-iframework} the compiler does not
8884 warn about constructs contained within header files found via
8885 @var{dir}. This option is valid only for the C family of languages.
8889 Emit debugging information for symbols that are used. For STABS
8890 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
8891 This is by default ON@.
8895 Emit debugging information for all symbols and types.
8897 @item -mmacosx-version-min=@var{version}
8898 The earliest version of MacOS X that this executable will run on
8899 is @var{version}. Typical values of @var{version} include @code{10.1},
8900 @code{10.2}, and @code{10.3.9}.
8902 If the compiler was built to use the system's headers by default,
8903 then the default for this option is the system version on which the
8904 compiler is running, otherwise the default is to make choices which
8905 are compatible with as many systems and code bases as possible.
8909 Enable kernel development mode. The @option{-mkernel} option sets
8910 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
8911 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
8912 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
8913 applicable. This mode also sets @option{-mno-altivec},
8914 @option{-msoft-float}, @option{-fno-builtin} and
8915 @option{-mlong-branch} for PowerPC targets.
8917 @item -mone-byte-bool
8918 @opindex mone-byte-bool
8919 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
8920 By default @samp{sizeof(bool)} is @samp{4} when compiling for
8921 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
8922 option has no effect on x86.
8924 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
8925 to generate code that is not binary compatible with code generated
8926 without that switch. Using this switch may require recompiling all
8927 other modules in a program, including system libraries. Use this
8928 switch to conform to a non-default data model.
8930 @item -mfix-and-continue
8931 @itemx -ffix-and-continue
8932 @itemx -findirect-data
8933 @opindex mfix-and-continue
8934 @opindex ffix-and-continue
8935 @opindex findirect-data
8936 Generate code suitable for fast turn around development. Needed to
8937 enable gdb to dynamically load @code{.o} files into already running
8938 programs. @option{-findirect-data} and @option{-ffix-and-continue}
8939 are provided for backwards compatibility.
8943 Loads all members of static archive libraries.
8944 See man ld(1) for more information.
8946 @item -arch_errors_fatal
8947 @opindex arch_errors_fatal
8948 Cause the errors having to do with files that have the wrong architecture
8952 @opindex bind_at_load
8953 Causes the output file to be marked such that the dynamic linker will
8954 bind all undefined references when the file is loaded or launched.
8958 Produce a Mach-o bundle format file.
8959 See man ld(1) for more information.
8961 @item -bundle_loader @var{executable}
8962 @opindex bundle_loader
8963 This option specifies the @var{executable} that will be loading the build
8964 output file being linked. See man ld(1) for more information.
8968 When passed this option, GCC will produce a dynamic library instead of
8969 an executable when linking, using the Darwin @file{libtool} command.
8971 @item -force_cpusubtype_ALL
8972 @opindex force_cpusubtype_ALL
8973 This causes GCC's output file to have the @var{ALL} subtype, instead of
8974 one controlled by the @option{-mcpu} or @option{-march} option.
8976 @item -allowable_client @var{client_name}
8978 @itemx -compatibility_version
8979 @itemx -current_version
8981 @itemx -dependency-file
8983 @itemx -dylinker_install_name
8985 @itemx -exported_symbols_list
8987 @itemx -flat_namespace
8988 @itemx -force_flat_namespace
8989 @itemx -headerpad_max_install_names
8992 @itemx -install_name
8993 @itemx -keep_private_externs
8994 @itemx -multi_module
8995 @itemx -multiply_defined
8996 @itemx -multiply_defined_unused
8998 @itemx -no_dead_strip_inits_and_terms
8999 @itemx -nofixprebinding
9002 @itemx -noseglinkedit
9003 @itemx -pagezero_size
9005 @itemx -prebind_all_twolevel_modules
9006 @itemx -private_bundle
9007 @itemx -read_only_relocs
9009 @itemx -sectobjectsymbols
9013 @itemx -sectobjectsymbols
9016 @itemx -segs_read_only_addr
9017 @itemx -segs_read_write_addr
9018 @itemx -seg_addr_table
9019 @itemx -seg_addr_table_filename
9022 @itemx -segs_read_only_addr
9023 @itemx -segs_read_write_addr
9024 @itemx -single_module
9027 @itemx -sub_umbrella
9028 @itemx -twolevel_namespace
9031 @itemx -unexported_symbols_list
9032 @itemx -weak_reference_mismatches
9035 @opindex allowable_client
9036 @opindex client_name
9037 @opindex compatibility_version
9038 @opindex current_version
9040 @opindex dependency-file
9042 @opindex dylinker_install_name
9044 @opindex exported_symbols_list
9046 @opindex flat_namespace
9047 @opindex force_flat_namespace
9048 @opindex headerpad_max_install_names
9051 @opindex install_name
9052 @opindex keep_private_externs
9053 @opindex multi_module
9054 @opindex multiply_defined
9055 @opindex multiply_defined_unused
9057 @opindex no_dead_strip_inits_and_terms
9058 @opindex nofixprebinding
9059 @opindex nomultidefs
9061 @opindex noseglinkedit
9062 @opindex pagezero_size
9064 @opindex prebind_all_twolevel_modules
9065 @opindex private_bundle
9066 @opindex read_only_relocs
9068 @opindex sectobjectsymbols
9072 @opindex sectobjectsymbols
9075 @opindex segs_read_only_addr
9076 @opindex segs_read_write_addr
9077 @opindex seg_addr_table
9078 @opindex seg_addr_table_filename
9079 @opindex seglinkedit
9081 @opindex segs_read_only_addr
9082 @opindex segs_read_write_addr
9083 @opindex single_module
9085 @opindex sub_library
9086 @opindex sub_umbrella
9087 @opindex twolevel_namespace
9090 @opindex unexported_symbols_list
9091 @opindex weak_reference_mismatches
9092 @opindex whatsloaded
9094 These options are passed to the Darwin linker. The Darwin linker man page
9095 describes them in detail.
9098 @node DEC Alpha Options
9099 @subsection DEC Alpha Options
9101 These @samp{-m} options are defined for the DEC Alpha implementations:
9104 @item -mno-soft-float
9106 @opindex mno-soft-float
9107 @opindex msoft-float
9108 Use (do not use) the hardware floating-point instructions for
9109 floating-point operations. When @option{-msoft-float} is specified,
9110 functions in @file{libgcc.a} will be used to perform floating-point
9111 operations. Unless they are replaced by routines that emulate the
9112 floating-point operations, or compiled in such a way as to call such
9113 emulations routines, these routines will issue floating-point
9114 operations. If you are compiling for an Alpha without floating-point
9115 operations, you must ensure that the library is built so as not to call
9118 Note that Alpha implementations without floating-point operations are
9119 required to have floating-point registers.
9124 @opindex mno-fp-regs
9125 Generate code that uses (does not use) the floating-point register set.
9126 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
9127 register set is not used, floating point operands are passed in integer
9128 registers as if they were integers and floating-point results are passed
9129 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
9130 so any function with a floating-point argument or return value called by code
9131 compiled with @option{-mno-fp-regs} must also be compiled with that
9134 A typical use of this option is building a kernel that does not use,
9135 and hence need not save and restore, any floating-point registers.
9139 The Alpha architecture implements floating-point hardware optimized for
9140 maximum performance. It is mostly compliant with the IEEE floating
9141 point standard. However, for full compliance, software assistance is
9142 required. This option generates code fully IEEE compliant code
9143 @emph{except} that the @var{inexact-flag} is not maintained (see below).
9144 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
9145 defined during compilation. The resulting code is less efficient but is
9146 able to correctly support denormalized numbers and exceptional IEEE
9147 values such as not-a-number and plus/minus infinity. Other Alpha
9148 compilers call this option @option{-ieee_with_no_inexact}.
9150 @item -mieee-with-inexact
9151 @opindex mieee-with-inexact
9152 This is like @option{-mieee} except the generated code also maintains
9153 the IEEE @var{inexact-flag}. Turning on this option causes the
9154 generated code to implement fully-compliant IEEE math. In addition to
9155 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
9156 macro. On some Alpha implementations the resulting code may execute
9157 significantly slower than the code generated by default. Since there is
9158 very little code that depends on the @var{inexact-flag}, you should
9159 normally not specify this option. Other Alpha compilers call this
9160 option @option{-ieee_with_inexact}.
9162 @item -mfp-trap-mode=@var{trap-mode}
9163 @opindex mfp-trap-mode
9164 This option controls what floating-point related traps are enabled.
9165 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9166 The trap mode can be set to one of four values:
9170 This is the default (normal) setting. The only traps that are enabled
9171 are the ones that cannot be disabled in software (e.g., division by zero
9175 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9179 Like @samp{u}, but the instructions are marked to be safe for software
9180 completion (see Alpha architecture manual for details).
9183 Like @samp{su}, but inexact traps are enabled as well.
9186 @item -mfp-rounding-mode=@var{rounding-mode}
9187 @opindex mfp-rounding-mode
9188 Selects the IEEE rounding mode. Other Alpha compilers call this option
9189 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9194 Normal IEEE rounding mode. Floating point numbers are rounded towards
9195 the nearest machine number or towards the even machine number in case
9199 Round towards minus infinity.
9202 Chopped rounding mode. Floating point numbers are rounded towards zero.
9205 Dynamic rounding mode. A field in the floating point control register
9206 (@var{fpcr}, see Alpha architecture reference manual) controls the
9207 rounding mode in effect. The C library initializes this register for
9208 rounding towards plus infinity. Thus, unless your program modifies the
9209 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9212 @item -mtrap-precision=@var{trap-precision}
9213 @opindex mtrap-precision
9214 In the Alpha architecture, floating point traps are imprecise. This
9215 means without software assistance it is impossible to recover from a
9216 floating trap and program execution normally needs to be terminated.
9217 GCC can generate code that can assist operating system trap handlers
9218 in determining the exact location that caused a floating point trap.
9219 Depending on the requirements of an application, different levels of
9220 precisions can be selected:
9224 Program precision. This option is the default and means a trap handler
9225 can only identify which program caused a floating point exception.
9228 Function precision. The trap handler can determine the function that
9229 caused a floating point exception.
9232 Instruction precision. The trap handler can determine the exact
9233 instruction that caused a floating point exception.
9236 Other Alpha compilers provide the equivalent options called
9237 @option{-scope_safe} and @option{-resumption_safe}.
9239 @item -mieee-conformant
9240 @opindex mieee-conformant
9241 This option marks the generated code as IEEE conformant. You must not
9242 use this option unless you also specify @option{-mtrap-precision=i} and either
9243 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9244 is to emit the line @samp{.eflag 48} in the function prologue of the
9245 generated assembly file. Under DEC Unix, this has the effect that
9246 IEEE-conformant math library routines will be linked in.
9248 @item -mbuild-constants
9249 @opindex mbuild-constants
9250 Normally GCC examines a 32- or 64-bit integer constant to
9251 see if it can construct it from smaller constants in two or three
9252 instructions. If it cannot, it will output the constant as a literal and
9253 generate code to load it from the data segment at runtime.
9255 Use this option to require GCC to construct @emph{all} integer constants
9256 using code, even if it takes more instructions (the maximum is six).
9258 You would typically use this option to build a shared library dynamic
9259 loader. Itself a shared library, it must relocate itself in memory
9260 before it can find the variables and constants in its own data segment.
9266 Select whether to generate code to be assembled by the vendor-supplied
9267 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9285 Indicate whether GCC should generate code to use the optional BWX,
9286 CIX, FIX and MAX instruction sets. The default is to use the instruction
9287 sets supported by the CPU type specified via @option{-mcpu=} option or that
9288 of the CPU on which GCC was built if none was specified.
9293 @opindex mfloat-ieee
9294 Generate code that uses (does not use) VAX F and G floating point
9295 arithmetic instead of IEEE single and double precision.
9297 @item -mexplicit-relocs
9298 @itemx -mno-explicit-relocs
9299 @opindex mexplicit-relocs
9300 @opindex mno-explicit-relocs
9301 Older Alpha assemblers provided no way to generate symbol relocations
9302 except via assembler macros. Use of these macros does not allow
9303 optimal instruction scheduling. GNU binutils as of version 2.12
9304 supports a new syntax that allows the compiler to explicitly mark
9305 which relocations should apply to which instructions. This option
9306 is mostly useful for debugging, as GCC detects the capabilities of
9307 the assembler when it is built and sets the default accordingly.
9311 @opindex msmall-data
9312 @opindex mlarge-data
9313 When @option{-mexplicit-relocs} is in effect, static data is
9314 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9315 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9316 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9317 16-bit relocations off of the @code{$gp} register. This limits the
9318 size of the small data area to 64KB, but allows the variables to be
9319 directly accessed via a single instruction.
9321 The default is @option{-mlarge-data}. With this option the data area
9322 is limited to just below 2GB@. Programs that require more than 2GB of
9323 data must use @code{malloc} or @code{mmap} to allocate the data in the
9324 heap instead of in the program's data segment.
9326 When generating code for shared libraries, @option{-fpic} implies
9327 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9331 @opindex msmall-text
9332 @opindex mlarge-text
9333 When @option{-msmall-text} is used, the compiler assumes that the
9334 code of the entire program (or shared library) fits in 4MB, and is
9335 thus reachable with a branch instruction. When @option{-msmall-data}
9336 is used, the compiler can assume that all local symbols share the
9337 same @code{$gp} value, and thus reduce the number of instructions
9338 required for a function call from 4 to 1.
9340 The default is @option{-mlarge-text}.
9342 @item -mcpu=@var{cpu_type}
9344 Set the instruction set and instruction scheduling parameters for
9345 machine type @var{cpu_type}. You can specify either the @samp{EV}
9346 style name or the corresponding chip number. GCC supports scheduling
9347 parameters for the EV4, EV5 and EV6 family of processors and will
9348 choose the default values for the instruction set from the processor
9349 you specify. If you do not specify a processor type, GCC will default
9350 to the processor on which the compiler was built.
9352 Supported values for @var{cpu_type} are
9358 Schedules as an EV4 and has no instruction set extensions.
9362 Schedules as an EV5 and has no instruction set extensions.
9366 Schedules as an EV5 and supports the BWX extension.
9371 Schedules as an EV5 and supports the BWX and MAX extensions.
9375 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9379 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9382 @item -mtune=@var{cpu_type}
9384 Set only the instruction scheduling parameters for machine type
9385 @var{cpu_type}. The instruction set is not changed.
9387 @item -mmemory-latency=@var{time}
9388 @opindex mmemory-latency
9389 Sets the latency the scheduler should assume for typical memory
9390 references as seen by the application. This number is highly
9391 dependent on the memory access patterns used by the application
9392 and the size of the external cache on the machine.
9394 Valid options for @var{time} are
9398 A decimal number representing clock cycles.
9404 The compiler contains estimates of the number of clock cycles for
9405 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9406 (also called Dcache, Scache, and Bcache), as well as to main memory.
9407 Note that L3 is only valid for EV5.
9412 @node DEC Alpha/VMS Options
9413 @subsection DEC Alpha/VMS Options
9415 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9418 @item -mvms-return-codes
9419 @opindex mvms-return-codes
9420 Return VMS condition codes from main. The default is to return POSIX
9421 style condition (e.g.@: error) codes.
9425 @subsection FRV Options
9432 Only use the first 32 general purpose registers.
9437 Use all 64 general purpose registers.
9442 Use only the first 32 floating point registers.
9447 Use all 64 floating point registers
9450 @opindex mhard-float
9452 Use hardware instructions for floating point operations.
9455 @opindex msoft-float
9457 Use library routines for floating point operations.
9462 Dynamically allocate condition code registers.
9467 Do not try to dynamically allocate condition code registers, only
9468 use @code{icc0} and @code{fcc0}.
9473 Change ABI to use double word insns.
9478 Do not use double word instructions.
9483 Use floating point double instructions.
9488 Do not use floating point double instructions.
9493 Use media instructions.
9498 Do not use media instructions.
9503 Use multiply and add/subtract instructions.
9508 Do not use multiply and add/subtract instructions.
9513 Select the FDPIC ABI, that uses function descriptors to represent
9514 pointers to functions. Without any PIC/PIE-related options, it
9515 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
9516 assumes GOT entries and small data are within a 12-bit range from the
9517 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
9518 are computed with 32 bits.
9521 @opindex minline-plt
9523 Enable inlining of PLT entries in function calls to functions that are
9524 not known to bind locally. It has no effect without @option{-mfdpic}.
9525 It's enabled by default if optimizing for speed and compiling for
9526 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
9527 optimization option such as @option{-O3} or above is present in the
9533 Assume a large TLS segment when generating thread-local code.
9538 Do not assume a large TLS segment when generating thread-local code.
9543 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
9544 that is known to be in read-only sections. It's enabled by default,
9545 except for @option{-fpic} or @option{-fpie}: even though it may help
9546 make the global offset table smaller, it trades 1 instruction for 4.
9547 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
9548 one of which may be shared by multiple symbols, and it avoids the need
9549 for a GOT entry for the referenced symbol, so it's more likely to be a
9550 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
9552 @item -multilib-library-pic
9553 @opindex multilib-library-pic
9555 Link with the (library, not FD) pic libraries. It's implied by
9556 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
9557 @option{-fpic} without @option{-mfdpic}. You should never have to use
9563 Follow the EABI requirement of always creating a frame pointer whenever
9564 a stack frame is allocated. This option is enabled by default and can
9565 be disabled with @option{-mno-linked-fp}.
9568 @opindex mlong-calls
9570 Use indirect addressing to call functions outside the current
9571 compilation unit. This allows the functions to be placed anywhere
9572 within the 32-bit address space.
9574 @item -malign-labels
9575 @opindex malign-labels
9577 Try to align labels to an 8-byte boundary by inserting nops into the
9578 previous packet. This option only has an effect when VLIW packing
9579 is enabled. It doesn't create new packets; it merely adds nops to
9583 @opindex mlibrary-pic
9585 Generate position-independent EABI code.
9590 Use only the first four media accumulator registers.
9595 Use all eight media accumulator registers.
9600 Pack VLIW instructions.
9605 Do not pack VLIW instructions.
9610 Do not mark ABI switches in e_flags.
9615 Enable the use of conditional-move instructions (default).
9617 This switch is mainly for debugging the compiler and will likely be removed
9618 in a future version.
9620 @item -mno-cond-move
9621 @opindex mno-cond-move
9623 Disable the use of conditional-move instructions.
9625 This switch is mainly for debugging the compiler and will likely be removed
9626 in a future version.
9631 Enable the use of conditional set instructions (default).
9633 This switch is mainly for debugging the compiler and will likely be removed
9634 in a future version.
9639 Disable the use of conditional set instructions.
9641 This switch is mainly for debugging the compiler and will likely be removed
9642 in a future version.
9647 Enable the use of conditional execution (default).
9649 This switch is mainly for debugging the compiler and will likely be removed
9650 in a future version.
9652 @item -mno-cond-exec
9653 @opindex mno-cond-exec
9655 Disable the use of conditional execution.
9657 This switch is mainly for debugging the compiler and will likely be removed
9658 in a future version.
9661 @opindex mvliw-branch
9663 Run a pass to pack branches into VLIW instructions (default).
9665 This switch is mainly for debugging the compiler and will likely be removed
9666 in a future version.
9668 @item -mno-vliw-branch
9669 @opindex mno-vliw-branch
9671 Do not run a pass to pack branches into VLIW instructions.
9673 This switch is mainly for debugging the compiler and will likely be removed
9674 in a future version.
9676 @item -mmulti-cond-exec
9677 @opindex mmulti-cond-exec
9679 Enable optimization of @code{&&} and @code{||} in conditional execution
9682 This switch is mainly for debugging the compiler and will likely be removed
9683 in a future version.
9685 @item -mno-multi-cond-exec
9686 @opindex mno-multi-cond-exec
9688 Disable optimization of @code{&&} and @code{||} in conditional execution.
9690 This switch is mainly for debugging the compiler and will likely be removed
9691 in a future version.
9693 @item -mnested-cond-exec
9694 @opindex mnested-cond-exec
9696 Enable nested conditional execution optimizations (default).
9698 This switch is mainly for debugging the compiler and will likely be removed
9699 in a future version.
9701 @item -mno-nested-cond-exec
9702 @opindex mno-nested-cond-exec
9704 Disable nested conditional execution optimizations.
9706 This switch is mainly for debugging the compiler and will likely be removed
9707 in a future version.
9709 @item -moptimize-membar
9710 @opindex moptimize-membar
9712 This switch removes redundant @code{membar} instructions from the
9713 compiler generated code. It is enabled by default.
9715 @item -mno-optimize-membar
9716 @opindex mno-optimize-membar
9718 This switch disables the automatic removal of redundant @code{membar}
9719 instructions from the generated code.
9721 @item -mtomcat-stats
9722 @opindex mtomcat-stats
9724 Cause gas to print out tomcat statistics.
9726 @item -mcpu=@var{cpu}
9729 Select the processor type for which to generate code. Possible values are
9730 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
9731 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
9735 @node GNU/Linux Options
9736 @subsection GNU/Linux Options
9738 These @samp{-m} options are defined for GNU/Linux targets:
9743 Use the GNU C library instead of uClibc. This is the default except
9744 on @samp{*-*-linux-*uclibc*} targets.
9748 Use uClibc instead of the GNU C library. This is the default on
9749 @samp{*-*-linux-*uclibc*} targets.
9752 @node H8/300 Options
9753 @subsection H8/300 Options
9755 These @samp{-m} options are defined for the H8/300 implementations:
9760 Shorten some address references at link time, when possible; uses the
9761 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9762 ld, Using ld}, for a fuller description.
9766 Generate code for the H8/300H@.
9770 Generate code for the H8S@.
9774 Generate code for the H8S and H8/300H in the normal mode. This switch
9775 must be used either with @option{-mh} or @option{-ms}.
9779 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9783 Make @code{int} data 32 bits by default.
9787 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9788 The default for the H8/300H and H8S is to align longs and floats on 4
9790 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9791 This option has no effect on the H8/300.
9795 @subsection HPPA Options
9796 @cindex HPPA Options
9798 These @samp{-m} options are defined for the HPPA family of computers:
9801 @item -march=@var{architecture-type}
9803 Generate code for the specified architecture. The choices for
9804 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
9805 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
9806 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
9807 architecture option for your machine. Code compiled for lower numbered
9808 architectures will run on higher numbered architectures, but not the
9812 @itemx -mpa-risc-1-1
9813 @itemx -mpa-risc-2-0
9814 @opindex mpa-risc-1-0
9815 @opindex mpa-risc-1-1
9816 @opindex mpa-risc-2-0
9817 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
9820 @opindex mbig-switch
9821 Generate code suitable for big switch tables. Use this option only if
9822 the assembler/linker complain about out of range branches within a switch
9825 @item -mjump-in-delay
9826 @opindex mjump-in-delay
9827 Fill delay slots of function calls with unconditional jump instructions
9828 by modifying the return pointer for the function call to be the target
9829 of the conditional jump.
9831 @item -mdisable-fpregs
9832 @opindex mdisable-fpregs
9833 Prevent floating point registers from being used in any manner. This is
9834 necessary for compiling kernels which perform lazy context switching of
9835 floating point registers. If you use this option and attempt to perform
9836 floating point operations, the compiler will abort.
9838 @item -mdisable-indexing
9839 @opindex mdisable-indexing
9840 Prevent the compiler from using indexing address modes. This avoids some
9841 rather obscure problems when compiling MIG generated code under MACH@.
9843 @item -mno-space-regs
9844 @opindex mno-space-regs
9845 Generate code that assumes the target has no space registers. This allows
9846 GCC to generate faster indirect calls and use unscaled index address modes.
9848 Such code is suitable for level 0 PA systems and kernels.
9850 @item -mfast-indirect-calls
9851 @opindex mfast-indirect-calls
9852 Generate code that assumes calls never cross space boundaries. This
9853 allows GCC to emit code which performs faster indirect calls.
9855 This option will not work in the presence of shared libraries or nested
9858 @item -mfixed-range=@var{register-range}
9859 @opindex mfixed-range
9860 Generate code treating the given register range as fixed registers.
9861 A fixed register is one that the register allocator can not use. This is
9862 useful when compiling kernel code. A register range is specified as
9863 two registers separated by a dash. Multiple register ranges can be
9864 specified separated by a comma.
9866 @item -mlong-load-store
9867 @opindex mlong-load-store
9868 Generate 3-instruction load and store sequences as sometimes required by
9869 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
9872 @item -mportable-runtime
9873 @opindex mportable-runtime
9874 Use the portable calling conventions proposed by HP for ELF systems.
9878 Enable the use of assembler directives only GAS understands.
9880 @item -mschedule=@var{cpu-type}
9882 Schedule code according to the constraints for the machine type
9883 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
9884 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
9885 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
9886 proper scheduling option for your machine. The default scheduling is
9890 @opindex mlinker-opt
9891 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
9892 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
9893 linkers in which they give bogus error messages when linking some programs.
9896 @opindex msoft-float
9897 Generate output containing library calls for floating point.
9898 @strong{Warning:} the requisite libraries are not available for all HPPA
9899 targets. Normally the facilities of the machine's usual C compiler are
9900 used, but this cannot be done directly in cross-compilation. You must make
9901 your own arrangements to provide suitable library functions for
9902 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
9903 does provide software floating point support.
9905 @option{-msoft-float} changes the calling convention in the output file;
9906 therefore, it is only useful if you compile @emph{all} of a program with
9907 this option. In particular, you need to compile @file{libgcc.a}, the
9908 library that comes with GCC, with @option{-msoft-float} in order for
9913 Generate the predefine, @code{_SIO}, for server IO@. The default is
9914 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
9915 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
9916 options are available under HP-UX and HI-UX@.
9920 Use GNU ld specific options. This passes @option{-shared} to ld when
9921 building a shared library. It is the default when GCC is configured,
9922 explicitly or implicitly, with the GNU linker. This option does not
9923 have any affect on which ld is called, it only changes what parameters
9924 are passed to that ld. The ld that is called is determined by the
9925 @option{--with-ld} configure option, GCC's program search path, and
9926 finally by the user's @env{PATH}. The linker used by GCC can be printed
9927 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
9928 on the 64 bit HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9932 Use HP ld specific options. This passes @option{-b} to ld when building
9933 a shared library and passes @option{+Accept TypeMismatch} to ld on all
9934 links. It is the default when GCC is configured, explicitly or
9935 implicitly, with the HP linker. This option does not have any affect on
9936 which ld is called, it only changes what parameters are passed to that
9937 ld. The ld that is called is determined by the @option{--with-ld}
9938 configure option, GCC's program search path, and finally by the user's
9939 @env{PATH}. The linker used by GCC can be printed using @samp{which
9940 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
9941 HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9944 @opindex mno-long-calls
9945 Generate code that uses long call sequences. This ensures that a call
9946 is always able to reach linker generated stubs. The default is to generate
9947 long calls only when the distance from the call site to the beginning
9948 of the function or translation unit, as the case may be, exceeds a
9949 predefined limit set by the branch type being used. The limits for
9950 normal calls are 7,600,000 and 240,000 bytes, respectively for the
9951 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
9954 Distances are measured from the beginning of functions when using the
9955 @option{-ffunction-sections} option, or when using the @option{-mgas}
9956 and @option{-mno-portable-runtime} options together under HP-UX with
9959 It is normally not desirable to use this option as it will degrade
9960 performance. However, it may be useful in large applications,
9961 particularly when partial linking is used to build the application.
9963 The types of long calls used depends on the capabilities of the
9964 assembler and linker, and the type of code being generated. The
9965 impact on systems that support long absolute calls, and long pic
9966 symbol-difference or pc-relative calls should be relatively small.
9967 However, an indirect call is used on 32-bit ELF systems in pic code
9968 and it is quite long.
9970 @item -munix=@var{unix-std}
9972 Generate compiler predefines and select a startfile for the specified
9973 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
9974 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
9975 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
9976 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
9977 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
9980 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
9981 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
9982 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
9983 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
9984 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
9985 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
9987 It is @emph{important} to note that this option changes the interfaces
9988 for various library routines. It also affects the operational behavior
9989 of the C library. Thus, @emph{extreme} care is needed in using this
9992 Library code that is intended to operate with more than one UNIX
9993 standard must test, set and restore the variable @var{__xpg4_extended_mask}
9994 as appropriate. Most GNU software doesn't provide this capability.
9998 Suppress the generation of link options to search libdld.sl when the
9999 @option{-static} option is specified on HP-UX 10 and later.
10003 The HP-UX implementation of setlocale in libc has a dependency on
10004 libdld.sl. There isn't an archive version of libdld.sl. Thus,
10005 when the @option{-static} option is specified, special link options
10006 are needed to resolve this dependency.
10008 On HP-UX 10 and later, the GCC driver adds the necessary options to
10009 link with libdld.sl when the @option{-static} option is specified.
10010 This causes the resulting binary to be dynamic. On the 64-bit port,
10011 the linkers generate dynamic binaries by default in any case. The
10012 @option{-nolibdld} option can be used to prevent the GCC driver from
10013 adding these link options.
10017 Add support for multithreading with the @dfn{dce thread} library
10018 under HP-UX@. This option sets flags for both the preprocessor and
10022 @node i386 and x86-64 Options
10023 @subsection Intel 386 and AMD x86-64 Options
10024 @cindex i386 Options
10025 @cindex x86-64 Options
10026 @cindex Intel 386 Options
10027 @cindex AMD x86-64 Options
10029 These @samp{-m} options are defined for the i386 and x86-64 family of
10033 @item -mtune=@var{cpu-type}
10035 Tune to @var{cpu-type} everything applicable about the generated code, except
10036 for the ABI and the set of available instructions. The choices for
10037 @var{cpu-type} are:
10040 Produce code optimized for the most common IA32/AMD64/EM64T processors.
10041 If you know the CPU on which your code will run, then you should use
10042 the corresponding @option{-mtune} option instead of
10043 @option{-mtune=generic}. But, if you do not know exactly what CPU users
10044 of your application will have, then you should use this option.
10046 As new processors are deployed in the marketplace, the behavior of this
10047 option will change. Therefore, if you upgrade to a newer version of
10048 GCC, the code generated option will change to reflect the processors
10049 that were most common when that version of GCC was released.
10051 There is no @option{-march=generic} option because @option{-march}
10052 indicates the instruction set the compiler can use, and there is no
10053 generic instruction set applicable to all processors. In contrast,
10054 @option{-mtune} indicates the processor (or, in this case, collection of
10055 processors) for which the code is optimized.
10057 This selects the CPU to tune for at compilation time by determining
10058 the processor type of the compiling machine. Using @option{-mtune=native}
10059 will produce code optimized for the local machine under the constraints
10060 of the selected instruction set. Using @option{-march=native} will
10061 enable all instruction subsets supported by the local machine (hence
10062 the result might not run on different machines).
10064 Original Intel's i386 CPU@.
10066 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
10067 @item i586, pentium
10068 Intel Pentium CPU with no MMX support.
10070 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
10072 Intel PentiumPro CPU@.
10074 Same as @code{generic}, but when used as @code{march} option, PentiumPro
10075 instruction set will be used, so the code will run on all i686 family chips.
10077 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
10078 @item pentium3, pentium3m
10079 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
10082 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
10083 support. Used by Centrino notebooks.
10084 @item pentium4, pentium4m
10085 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
10087 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
10090 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
10091 SSE2 and SSE3 instruction set support.
10093 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
10094 instruction set support.
10096 AMD K6 CPU with MMX instruction set support.
10098 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
10099 @item athlon, athlon-tbird
10100 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
10102 @item athlon-4, athlon-xp, athlon-mp
10103 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
10104 instruction set support.
10105 @item k8, opteron, athlon64, athlon-fx
10106 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
10107 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
10108 @item k8-sse3, opteron-sse3, athlon64-sse3
10109 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
10110 @item amdfam10, barcelona
10111 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
10112 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3dNOW!, enhanced 3dNOW!, ABM and 64-bit
10113 instruction set extensions.)
10115 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
10118 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
10119 instruction set support.
10121 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
10122 implemented for this chip.)
10124 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
10125 implemented for this chip.)
10127 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
10130 While picking a specific @var{cpu-type} will schedule things appropriately
10131 for that particular chip, the compiler will not generate any code that
10132 does not run on the i386 without the @option{-march=@var{cpu-type}} option
10135 @item -march=@var{cpu-type}
10137 Generate instructions for the machine type @var{cpu-type}. The choices
10138 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
10139 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
10141 @item -mcpu=@var{cpu-type}
10143 A deprecated synonym for @option{-mtune}.
10145 @item -mfpmath=@var{unit}
10147 Generate floating point arithmetics for selected unit @var{unit}. The choices
10148 for @var{unit} are:
10152 Use the standard 387 floating point coprocessor present majority of chips and
10153 emulated otherwise. Code compiled with this option will run almost everywhere.
10154 The temporary results are computed in 80bit precision instead of precision
10155 specified by the type resulting in slightly different results compared to most
10156 of other chips. See @option{-ffloat-store} for more detailed description.
10158 This is the default choice for i386 compiler.
10161 Use scalar floating point instructions present in the SSE instruction set.
10162 This instruction set is supported by Pentium3 and newer chips, in the AMD line
10163 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
10164 instruction set supports only single precision arithmetics, thus the double and
10165 extended precision arithmetics is still done using 387. Later version, present
10166 only in Pentium4 and the future AMD x86-64 chips supports double precision
10169 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
10170 or @option{-msse2} switches to enable SSE extensions and make this option
10171 effective. For the x86-64 compiler, these extensions are enabled by default.
10173 The resulting code should be considerably faster in the majority of cases and avoid
10174 the numerical instability problems of 387 code, but may break some existing
10175 code that expects temporaries to be 80bit.
10177 This is the default choice for the x86-64 compiler.
10180 Attempt to utilize both instruction sets at once. This effectively double the
10181 amount of available registers and on chips with separate execution units for
10182 387 and SSE the execution resources too. Use this option with care, as it is
10183 still experimental, because the GCC register allocator does not model separate
10184 functional units well resulting in instable performance.
10187 @item -masm=@var{dialect}
10188 @opindex masm=@var{dialect}
10189 Output asm instructions using selected @var{dialect}. Supported
10190 choices are @samp{intel} or @samp{att} (the default one). Darwin does
10191 not support @samp{intel}.
10194 @itemx -mno-ieee-fp
10196 @opindex mno-ieee-fp
10197 Control whether or not the compiler uses IEEE floating point
10198 comparisons. These handle correctly the case where the result of a
10199 comparison is unordered.
10202 @opindex msoft-float
10203 Generate output containing library calls for floating point.
10204 @strong{Warning:} the requisite libraries are not part of GCC@.
10205 Normally the facilities of the machine's usual C compiler are used, but
10206 this can't be done directly in cross-compilation. You must make your
10207 own arrangements to provide suitable library functions for
10210 On machines where a function returns floating point results in the 80387
10211 register stack, some floating point opcodes may be emitted even if
10212 @option{-msoft-float} is used.
10214 @item -mno-fp-ret-in-387
10215 @opindex mno-fp-ret-in-387
10216 Do not use the FPU registers for return values of functions.
10218 The usual calling convention has functions return values of types
10219 @code{float} and @code{double} in an FPU register, even if there
10220 is no FPU@. The idea is that the operating system should emulate
10223 The option @option{-mno-fp-ret-in-387} causes such values to be returned
10224 in ordinary CPU registers instead.
10226 @item -mno-fancy-math-387
10227 @opindex mno-fancy-math-387
10228 Some 387 emulators do not support the @code{sin}, @code{cos} and
10229 @code{sqrt} instructions for the 387. Specify this option to avoid
10230 generating those instructions. This option is the default on FreeBSD,
10231 OpenBSD and NetBSD@. This option is overridden when @option{-march}
10232 indicates that the target cpu will always have an FPU and so the
10233 instruction will not need emulation. As of revision 2.6.1, these
10234 instructions are not generated unless you also use the
10235 @option{-funsafe-math-optimizations} switch.
10237 @item -malign-double
10238 @itemx -mno-align-double
10239 @opindex malign-double
10240 @opindex mno-align-double
10241 Control whether GCC aligns @code{double}, @code{long double}, and
10242 @code{long long} variables on a two word boundary or a one word
10243 boundary. Aligning @code{double} variables on a two word boundary will
10244 produce code that runs somewhat faster on a @samp{Pentium} at the
10245 expense of more memory.
10247 On x86-64, @option{-malign-double} is enabled by default.
10249 @strong{Warning:} if you use the @option{-malign-double} switch,
10250 structures containing the above types will be aligned differently than
10251 the published application binary interface specifications for the 386
10252 and will not be binary compatible with structures in code compiled
10253 without that switch.
10255 @item -m96bit-long-double
10256 @itemx -m128bit-long-double
10257 @opindex m96bit-long-double
10258 @opindex m128bit-long-double
10259 These switches control the size of @code{long double} type. The i386
10260 application binary interface specifies the size to be 96 bits,
10261 so @option{-m96bit-long-double} is the default in 32 bit mode.
10263 Modern architectures (Pentium and newer) would prefer @code{long double}
10264 to be aligned to an 8 or 16 byte boundary. In arrays or structures
10265 conforming to the ABI, this would not be possible. So specifying a
10266 @option{-m128bit-long-double} will align @code{long double}
10267 to a 16 byte boundary by padding the @code{long double} with an additional
10270 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
10271 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
10273 Notice that neither of these options enable any extra precision over the x87
10274 standard of 80 bits for a @code{long double}.
10276 @strong{Warning:} if you override the default value for your target ABI, the
10277 structures and arrays containing @code{long double} variables will change
10278 their size as well as function calling convention for function taking
10279 @code{long double} will be modified. Hence they will not be binary
10280 compatible with arrays or structures in code compiled without that switch.
10282 @item -mmlarge-data-threshold=@var{number}
10283 @opindex mlarge-data-threshold=@var{number}
10284 When @option{-mcmodel=medium} is specified, the data greater than
10285 @var{threshold} are placed in large data section. This value must be the
10286 same across all object linked into the binary and defaults to 65535.
10290 Use a different function-calling convention, in which functions that
10291 take a fixed number of arguments return with the @code{ret} @var{num}
10292 instruction, which pops their arguments while returning. This saves one
10293 instruction in the caller since there is no need to pop the arguments
10296 You can specify that an individual function is called with this calling
10297 sequence with the function attribute @samp{stdcall}. You can also
10298 override the @option{-mrtd} option by using the function attribute
10299 @samp{cdecl}. @xref{Function Attributes}.
10301 @strong{Warning:} this calling convention is incompatible with the one
10302 normally used on Unix, so you cannot use it if you need to call
10303 libraries compiled with the Unix compiler.
10305 Also, you must provide function prototypes for all functions that
10306 take variable numbers of arguments (including @code{printf});
10307 otherwise incorrect code will be generated for calls to those
10310 In addition, seriously incorrect code will result if you call a
10311 function with too many arguments. (Normally, extra arguments are
10312 harmlessly ignored.)
10314 @item -mregparm=@var{num}
10316 Control how many registers are used to pass integer arguments. By
10317 default, no registers are used to pass arguments, and at most 3
10318 registers can be used. You can control this behavior for a specific
10319 function by using the function attribute @samp{regparm}.
10320 @xref{Function Attributes}.
10322 @strong{Warning:} if you use this switch, and
10323 @var{num} is nonzero, then you must build all modules with the same
10324 value, including any libraries. This includes the system libraries and
10328 @opindex msseregparm
10329 Use SSE register passing conventions for float and double arguments
10330 and return values. You can control this behavior for a specific
10331 function by using the function attribute @samp{sseregparm}.
10332 @xref{Function Attributes}.
10334 @strong{Warning:} if you use this switch then you must build all
10335 modules with the same value, including any libraries. This includes
10336 the system libraries and startup modules.
10345 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
10346 is specified, the significands of results of floating-point operations are
10347 rounded to 24 bits (single precision); @option{-mpc64} rounds the the
10348 significands of results of floating-point operations to 53 bits (double
10349 precision) and @option{-mpc80} rounds the significands of results of
10350 floating-point operations to 64 bits (extended double precision), which is
10351 the default. When this option is used, floating-point operations in higher
10352 precisions are not available to the programmer without setting the FPU
10353 control word explicitly.
10355 Setting the rounding of floating-point operations to less than the default
10356 80 bits can speed some programs by 2% or more. Note that some mathematical
10357 libraries assume that extended precision (80 bit) floating-point operations
10358 are enabled by default; routines in such libraries could suffer significant
10359 loss of accuracy, typically through so-called "catastrophic cancellation",
10360 when this option is used to set the precision to less than extended precision.
10362 @item -mstackrealign
10363 @opindex mstackrealign
10364 Realign the stack at entry. On the Intel x86, the
10365 @option{-mstackrealign} option will generate an alternate prologue and
10366 epilogue that realigns the runtime stack. This supports mixing legacy
10367 codes that keep a 4-byte aligned stack with modern codes that keep a
10368 16-byte stack for SSE compatibility. The alternate prologue and
10369 epilogue are slower and bigger than the regular ones, and the
10370 alternate prologue requires an extra scratch register; this lowers the
10371 number of registers available if used in conjunction with the
10372 @code{regparm} attribute. The @option{-mstackrealign} option is
10373 incompatible with the nested function prologue; this is considered a
10374 hard error. See also the attribute @code{force_align_arg_pointer},
10375 applicable to individual functions.
10377 @item -mpreferred-stack-boundary=@var{num}
10378 @opindex mpreferred-stack-boundary
10379 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
10380 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
10381 the default is 4 (16 bytes or 128 bits).
10383 On Pentium and PentiumPro, @code{double} and @code{long double} values
10384 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
10385 suffer significant run time performance penalties. On Pentium III, the
10386 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
10387 properly if it is not 16 byte aligned.
10389 To ensure proper alignment of this values on the stack, the stack boundary
10390 must be as aligned as that required by any value stored on the stack.
10391 Further, every function must be generated such that it keeps the stack
10392 aligned. Thus calling a function compiled with a higher preferred
10393 stack boundary from a function compiled with a lower preferred stack
10394 boundary will most likely misalign the stack. It is recommended that
10395 libraries that use callbacks always use the default setting.
10397 This extra alignment does consume extra stack space, and generally
10398 increases code size. Code that is sensitive to stack space usage, such
10399 as embedded systems and operating system kernels, may want to reduce the
10400 preferred alignment to @option{-mpreferred-stack-boundary=2}.
10432 These switches enable or disable the use of instructions in the MMX,
10433 SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4A, ABM or 3DNow! extended
10435 These extensions are also available as built-in functions: see
10436 @ref{X86 Built-in Functions}, for details of the functions enabled and
10437 disabled by these switches.
10439 To have SSE/SSE2 instructions generated automatically from floating-point
10440 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
10442 These options will enable GCC to use these extended instructions in
10443 generated code, even without @option{-mfpmath=sse}. Applications which
10444 perform runtime CPU detection must compile separate files for each
10445 supported architecture, using the appropriate flags. In particular,
10446 the file containing the CPU detection code should be compiled without
10451 This option will enable GCC to use CMPXCHG16B instruction in generated code.
10452 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
10453 data types. This is useful for high resolution counters that could be updated
10454 by multiple processors (or cores). This instruction is generated as part of
10455 atomic built-in functions: see @ref{Atomic Builtins} for details.
10459 This option will enable GCC to use SAHF instruction in generated 64-bit code.
10460 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
10461 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
10462 SAHF are load and store instructions, respectively, for certain status flags.
10463 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
10464 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
10468 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
10469 vectorized variants RCPPS and RSQRTPS) instead of DIVSS and SQRTSS (and their
10470 vectorized variants). These instructions will be generated only when
10471 @option{-funsafe-math-optimizations} is enabled.
10473 @item -mveclibabi=@var{type}
10474 @opindex mveclibabi
10475 Specifies the ABI type to use for vectorizing intrinsics using an
10476 external library. Supported types are @code{acml} for the AMD
10477 math core library style of interfacing. GCC will currently emit
10478 calls to @code{__vrd2_sin}, @code{__vrd2_cos}, @code{__vrd2_exp},
10479 @code{__vrd2_log}, @code{__vrd2_log2}, @code{__vrd2_log10},
10480 @code{__vrs4_sinf}, @code{__vrs4_cosf}, @code{__vrs4_expf},
10481 @code{__vrs4_logf}, @code{__vrs4_log2f}, @code{__vrs4_log10f}
10482 and @code{__vrs4_powf} when using this type and @option{-ftree-vectorize}
10483 is enabled. A ACML ABI compatible library will have to be specified
10487 @itemx -mno-push-args
10488 @opindex mpush-args
10489 @opindex mno-push-args
10490 Use PUSH operations to store outgoing parameters. This method is shorter
10491 and usually equally fast as method using SUB/MOV operations and is enabled
10492 by default. In some cases disabling it may improve performance because of
10493 improved scheduling and reduced dependencies.
10495 @item -maccumulate-outgoing-args
10496 @opindex maccumulate-outgoing-args
10497 If enabled, the maximum amount of space required for outgoing arguments will be
10498 computed in the function prologue. This is faster on most modern CPUs
10499 because of reduced dependencies, improved scheduling and reduced stack usage
10500 when preferred stack boundary is not equal to 2. The drawback is a notable
10501 increase in code size. This switch implies @option{-mno-push-args}.
10505 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
10506 on thread-safe exception handling must compile and link all code with the
10507 @option{-mthreads} option. When compiling, @option{-mthreads} defines
10508 @option{-D_MT}; when linking, it links in a special thread helper library
10509 @option{-lmingwthrd} which cleans up per thread exception handling data.
10511 @item -mno-align-stringops
10512 @opindex mno-align-stringops
10513 Do not align destination of inlined string operations. This switch reduces
10514 code size and improves performance in case the destination is already aligned,
10515 but GCC doesn't know about it.
10517 @item -minline-all-stringops
10518 @opindex minline-all-stringops
10519 By default GCC inlines string operations only when destination is known to be
10520 aligned at least to 4 byte boundary. This enables more inlining, increase code
10521 size, but may improve performance of code that depends on fast memcpy, strlen
10522 and memset for short lengths.
10524 @item -minline-stringops-dynamically
10525 @opindex minline-stringops-dynamically
10526 For string operation of unknown size, inline runtime checks so for small
10527 blocks inline code is used, while for large blocks library call is used.
10529 @item -mstringop-strategy=@var{alg}
10530 @opindex mstringop-strategy=@var{alg}
10531 Overwrite internal decision heuristic about particular algorithm to inline
10532 string operation with. The allowed values are @code{rep_byte},
10533 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
10534 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
10535 expanding inline loop, @code{libcall} for always expanding library call.
10537 @item -momit-leaf-frame-pointer
10538 @opindex momit-leaf-frame-pointer
10539 Don't keep the frame pointer in a register for leaf functions. This
10540 avoids the instructions to save, set up and restore frame pointers and
10541 makes an extra register available in leaf functions. The option
10542 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10543 which might make debugging harder.
10545 @item -mtls-direct-seg-refs
10546 @itemx -mno-tls-direct-seg-refs
10547 @opindex mtls-direct-seg-refs
10548 Controls whether TLS variables may be accessed with offsets from the
10549 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
10550 or whether the thread base pointer must be added. Whether or not this
10551 is legal depends on the operating system, and whether it maps the
10552 segment to cover the entire TLS area.
10554 For systems that use GNU libc, the default is on.
10557 These @samp{-m} switches are supported in addition to the above
10558 on AMD x86-64 processors in 64-bit environments.
10565 Generate code for a 32-bit or 64-bit environment.
10566 The 32-bit environment sets int, long and pointer to 32 bits and
10567 generates code that runs on any i386 system.
10568 The 64-bit environment sets int to 32 bits and long and pointer
10569 to 64 bits and generates code for AMD's x86-64 architecture. For
10570 darwin only the -m64 option turns off the @option{-fno-pic} and
10571 @option{-mdynamic-no-pic} options.
10573 @item -mno-red-zone
10574 @opindex no-red-zone
10575 Do not use a so called red zone for x86-64 code. The red zone is mandated
10576 by the x86-64 ABI, it is a 128-byte area beyond the location of the
10577 stack pointer that will not be modified by signal or interrupt handlers
10578 and therefore can be used for temporary data without adjusting the stack
10579 pointer. The flag @option{-mno-red-zone} disables this red zone.
10581 @item -mcmodel=small
10582 @opindex mcmodel=small
10583 Generate code for the small code model: the program and its symbols must
10584 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
10585 Programs can be statically or dynamically linked. This is the default
10588 @item -mcmodel=kernel
10589 @opindex mcmodel=kernel
10590 Generate code for the kernel code model. The kernel runs in the
10591 negative 2 GB of the address space.
10592 This model has to be used for Linux kernel code.
10594 @item -mcmodel=medium
10595 @opindex mcmodel=medium
10596 Generate code for the medium model: The program is linked in the lower 2
10597 GB of the address space but symbols can be located anywhere in the
10598 address space. Programs can be statically or dynamically linked, but
10599 building of shared libraries are not supported with the medium model.
10601 @item -mcmodel=large
10602 @opindex mcmodel=large
10603 Generate code for the large model: This model makes no assumptions
10604 about addresses and sizes of sections.
10607 @node IA-64 Options
10608 @subsection IA-64 Options
10609 @cindex IA-64 Options
10611 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10615 @opindex mbig-endian
10616 Generate code for a big endian target. This is the default for HP-UX@.
10618 @item -mlittle-endian
10619 @opindex mlittle-endian
10620 Generate code for a little endian target. This is the default for AIX5
10626 @opindex mno-gnu-as
10627 Generate (or don't) code for the GNU assembler. This is the default.
10628 @c Also, this is the default if the configure option @option{--with-gnu-as}
10634 @opindex mno-gnu-ld
10635 Generate (or don't) code for the GNU linker. This is the default.
10636 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10641 Generate code that does not use a global pointer register. The result
10642 is not position independent code, and violates the IA-64 ABI@.
10644 @item -mvolatile-asm-stop
10645 @itemx -mno-volatile-asm-stop
10646 @opindex mvolatile-asm-stop
10647 @opindex mno-volatile-asm-stop
10648 Generate (or don't) a stop bit immediately before and after volatile asm
10651 @item -mregister-names
10652 @itemx -mno-register-names
10653 @opindex mregister-names
10654 @opindex mno-register-names
10655 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
10656 the stacked registers. This may make assembler output more readable.
10662 Disable (or enable) optimizations that use the small data section. This may
10663 be useful for working around optimizer bugs.
10665 @item -mconstant-gp
10666 @opindex mconstant-gp
10667 Generate code that uses a single constant global pointer value. This is
10668 useful when compiling kernel code.
10672 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
10673 This is useful when compiling firmware code.
10675 @item -minline-float-divide-min-latency
10676 @opindex minline-float-divide-min-latency
10677 Generate code for inline divides of floating point values
10678 using the minimum latency algorithm.
10680 @item -minline-float-divide-max-throughput
10681 @opindex minline-float-divide-max-throughput
10682 Generate code for inline divides of floating point values
10683 using the maximum throughput algorithm.
10685 @item -minline-int-divide-min-latency
10686 @opindex minline-int-divide-min-latency
10687 Generate code for inline divides of integer values
10688 using the minimum latency algorithm.
10690 @item -minline-int-divide-max-throughput
10691 @opindex minline-int-divide-max-throughput
10692 Generate code for inline divides of integer values
10693 using the maximum throughput algorithm.
10695 @item -minline-sqrt-min-latency
10696 @opindex minline-sqrt-min-latency
10697 Generate code for inline square roots
10698 using the minimum latency algorithm.
10700 @item -minline-sqrt-max-throughput
10701 @opindex minline-sqrt-max-throughput
10702 Generate code for inline square roots
10703 using the maximum throughput algorithm.
10705 @item -mno-dwarf2-asm
10706 @itemx -mdwarf2-asm
10707 @opindex mno-dwarf2-asm
10708 @opindex mdwarf2-asm
10709 Don't (or do) generate assembler code for the DWARF2 line number debugging
10710 info. This may be useful when not using the GNU assembler.
10712 @item -mearly-stop-bits
10713 @itemx -mno-early-stop-bits
10714 @opindex mearly-stop-bits
10715 @opindex mno-early-stop-bits
10716 Allow stop bits to be placed earlier than immediately preceding the
10717 instruction that triggered the stop bit. This can improve instruction
10718 scheduling, but does not always do so.
10720 @item -mfixed-range=@var{register-range}
10721 @opindex mfixed-range
10722 Generate code treating the given register range as fixed registers.
10723 A fixed register is one that the register allocator can not use. This is
10724 useful when compiling kernel code. A register range is specified as
10725 two registers separated by a dash. Multiple register ranges can be
10726 specified separated by a comma.
10728 @item -mtls-size=@var{tls-size}
10730 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
10733 @item -mtune=@var{cpu-type}
10735 Tune the instruction scheduling for a particular CPU, Valid values are
10736 itanium, itanium1, merced, itanium2, and mckinley.
10742 Add support for multithreading using the POSIX threads library. This
10743 option sets flags for both the preprocessor and linker. It does
10744 not affect the thread safety of object code produced by the compiler or
10745 that of libraries supplied with it. These are HP-UX specific flags.
10751 Generate code for a 32-bit or 64-bit environment.
10752 The 32-bit environment sets int, long and pointer to 32 bits.
10753 The 64-bit environment sets int to 32 bits and long and pointer
10754 to 64 bits. These are HP-UX specific flags.
10756 @item -mno-sched-br-data-spec
10757 @itemx -msched-br-data-spec
10758 @opindex mno-sched-br-data-spec
10759 @opindex msched-br-data-spec
10760 (Dis/En)able data speculative scheduling before reload.
10761 This will result in generation of the ld.a instructions and
10762 the corresponding check instructions (ld.c / chk.a).
10763 The default is 'disable'.
10765 @item -msched-ar-data-spec
10766 @itemx -mno-sched-ar-data-spec
10767 @opindex msched-ar-data-spec
10768 @opindex mno-sched-ar-data-spec
10769 (En/Dis)able data speculative scheduling after reload.
10770 This will result in generation of the ld.a instructions and
10771 the corresponding check instructions (ld.c / chk.a).
10772 The default is 'enable'.
10774 @item -mno-sched-control-spec
10775 @itemx -msched-control-spec
10776 @opindex mno-sched-control-spec
10777 @opindex msched-control-spec
10778 (Dis/En)able control speculative scheduling. This feature is
10779 available only during region scheduling (i.e. before reload).
10780 This will result in generation of the ld.s instructions and
10781 the corresponding check instructions chk.s .
10782 The default is 'disable'.
10784 @item -msched-br-in-data-spec
10785 @itemx -mno-sched-br-in-data-spec
10786 @opindex msched-br-in-data-spec
10787 @opindex mno-sched-br-in-data-spec
10788 (En/Dis)able speculative scheduling of the instructions that
10789 are dependent on the data speculative loads before reload.
10790 This is effective only with @option{-msched-br-data-spec} enabled.
10791 The default is 'enable'.
10793 @item -msched-ar-in-data-spec
10794 @itemx -mno-sched-ar-in-data-spec
10795 @opindex msched-ar-in-data-spec
10796 @opindex mno-sched-ar-in-data-spec
10797 (En/Dis)able speculative scheduling of the instructions that
10798 are dependent on the data speculative loads after reload.
10799 This is effective only with @option{-msched-ar-data-spec} enabled.
10800 The default is 'enable'.
10802 @item -msched-in-control-spec
10803 @itemx -mno-sched-in-control-spec
10804 @opindex msched-in-control-spec
10805 @opindex mno-sched-in-control-spec
10806 (En/Dis)able speculative scheduling of the instructions that
10807 are dependent on the control speculative loads.
10808 This is effective only with @option{-msched-control-spec} enabled.
10809 The default is 'enable'.
10812 @itemx -mno-sched-ldc
10813 @opindex msched-ldc
10814 @opindex mno-sched-ldc
10815 (En/Dis)able use of simple data speculation checks ld.c .
10816 If disabled, only chk.a instructions will be emitted to check
10817 data speculative loads.
10818 The default is 'enable'.
10820 @item -mno-sched-control-ldc
10821 @itemx -msched-control-ldc
10822 @opindex mno-sched-control-ldc
10823 @opindex msched-control-ldc
10824 (Dis/En)able use of ld.c instructions to check control speculative loads.
10825 If enabled, in case of control speculative load with no speculatively
10826 scheduled dependent instructions this load will be emitted as ld.sa and
10827 ld.c will be used to check it.
10828 The default is 'disable'.
10830 @item -mno-sched-spec-verbose
10831 @itemx -msched-spec-verbose
10832 @opindex mno-sched-spec-verbose
10833 @opindex msched-spec-verbose
10834 (Dis/En)able printing of the information about speculative motions.
10836 @item -mno-sched-prefer-non-data-spec-insns
10837 @itemx -msched-prefer-non-data-spec-insns
10838 @opindex mno-sched-prefer-non-data-spec-insns
10839 @opindex msched-prefer-non-data-spec-insns
10840 If enabled, data speculative instructions will be chosen for schedule
10841 only if there are no other choices at the moment. This will make
10842 the use of the data speculation much more conservative.
10843 The default is 'disable'.
10845 @item -mno-sched-prefer-non-control-spec-insns
10846 @itemx -msched-prefer-non-control-spec-insns
10847 @opindex mno-sched-prefer-non-control-spec-insns
10848 @opindex msched-prefer-non-control-spec-insns
10849 If enabled, control speculative instructions will be chosen for schedule
10850 only if there are no other choices at the moment. This will make
10851 the use of the control speculation much more conservative.
10852 The default is 'disable'.
10854 @item -mno-sched-count-spec-in-critical-path
10855 @itemx -msched-count-spec-in-critical-path
10856 @opindex mno-sched-count-spec-in-critical-path
10857 @opindex msched-count-spec-in-critical-path
10858 If enabled, speculative dependencies will be considered during
10859 computation of the instructions priorities. This will make the use of the
10860 speculation a bit more conservative.
10861 The default is 'disable'.
10866 @subsection M32C Options
10867 @cindex M32C options
10870 @item -mcpu=@var{name}
10872 Select the CPU for which code is generated. @var{name} may be one of
10873 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
10874 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
10875 the M32C/80 series.
10879 Specifies that the program will be run on the simulator. This causes
10880 an alternate runtime library to be linked in which supports, for
10881 example, file I/O. You must not use this option when generating
10882 programs that will run on real hardware; you must provide your own
10883 runtime library for whatever I/O functions are needed.
10885 @item -memregs=@var{number}
10887 Specifies the number of memory-based pseudo-registers GCC will use
10888 during code generation. These pseudo-registers will be used like real
10889 registers, so there is a tradeoff between GCC's ability to fit the
10890 code into available registers, and the performance penalty of using
10891 memory instead of registers. Note that all modules in a program must
10892 be compiled with the same value for this option. Because of that, you
10893 must not use this option with the default runtime libraries gcc
10898 @node M32R/D Options
10899 @subsection M32R/D Options
10900 @cindex M32R/D options
10902 These @option{-m} options are defined for Renesas M32R/D architectures:
10907 Generate code for the M32R/2@.
10911 Generate code for the M32R/X@.
10915 Generate code for the M32R@. This is the default.
10917 @item -mmodel=small
10918 @opindex mmodel=small
10919 Assume all objects live in the lower 16MB of memory (so that their addresses
10920 can be loaded with the @code{ld24} instruction), and assume all subroutines
10921 are reachable with the @code{bl} instruction.
10922 This is the default.
10924 The addressability of a particular object can be set with the
10925 @code{model} attribute.
10927 @item -mmodel=medium
10928 @opindex mmodel=medium
10929 Assume objects may be anywhere in the 32-bit address space (the compiler
10930 will generate @code{seth/add3} instructions to load their addresses), and
10931 assume all subroutines are reachable with the @code{bl} instruction.
10933 @item -mmodel=large
10934 @opindex mmodel=large
10935 Assume objects may be anywhere in the 32-bit address space (the compiler
10936 will generate @code{seth/add3} instructions to load their addresses), and
10937 assume subroutines may not be reachable with the @code{bl} instruction
10938 (the compiler will generate the much slower @code{seth/add3/jl}
10939 instruction sequence).
10942 @opindex msdata=none
10943 Disable use of the small data area. Variables will be put into
10944 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
10945 @code{section} attribute has been specified).
10946 This is the default.
10948 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
10949 Objects may be explicitly put in the small data area with the
10950 @code{section} attribute using one of these sections.
10952 @item -msdata=sdata
10953 @opindex msdata=sdata
10954 Put small global and static data in the small data area, but do not
10955 generate special code to reference them.
10958 @opindex msdata=use
10959 Put small global and static data in the small data area, and generate
10960 special instructions to reference them.
10964 @cindex smaller data references
10965 Put global and static objects less than or equal to @var{num} bytes
10966 into the small data or bss sections instead of the normal data or bss
10967 sections. The default value of @var{num} is 8.
10968 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
10969 for this option to have any effect.
10971 All modules should be compiled with the same @option{-G @var{num}} value.
10972 Compiling with different values of @var{num} may or may not work; if it
10973 doesn't the linker will give an error message---incorrect code will not be
10978 Makes the M32R specific code in the compiler display some statistics
10979 that might help in debugging programs.
10981 @item -malign-loops
10982 @opindex malign-loops
10983 Align all loops to a 32-byte boundary.
10985 @item -mno-align-loops
10986 @opindex mno-align-loops
10987 Do not enforce a 32-byte alignment for loops. This is the default.
10989 @item -missue-rate=@var{number}
10990 @opindex missue-rate=@var{number}
10991 Issue @var{number} instructions per cycle. @var{number} can only be 1
10994 @item -mbranch-cost=@var{number}
10995 @opindex mbranch-cost=@var{number}
10996 @var{number} can only be 1 or 2. If it is 1 then branches will be
10997 preferred over conditional code, if it is 2, then the opposite will
11000 @item -mflush-trap=@var{number}
11001 @opindex mflush-trap=@var{number}
11002 Specifies the trap number to use to flush the cache. The default is
11003 12. Valid numbers are between 0 and 15 inclusive.
11005 @item -mno-flush-trap
11006 @opindex mno-flush-trap
11007 Specifies that the cache cannot be flushed by using a trap.
11009 @item -mflush-func=@var{name}
11010 @opindex mflush-func=@var{name}
11011 Specifies the name of the operating system function to call to flush
11012 the cache. The default is @emph{_flush_cache}, but a function call
11013 will only be used if a trap is not available.
11015 @item -mno-flush-func
11016 @opindex mno-flush-func
11017 Indicates that there is no OS function for flushing the cache.
11021 @node M680x0 Options
11022 @subsection M680x0 Options
11023 @cindex M680x0 options
11025 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
11026 The default settings depend on which architecture was selected when
11027 the compiler was configured; the defaults for the most common choices
11031 @item -march=@var{arch}
11033 Generate code for a specific M680x0 or ColdFire instruction set
11034 architecture. Permissible values of @var{arch} for M680x0
11035 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
11036 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
11037 architectures are selected according to Freescale's ISA classification
11038 and the permissible values are: @samp{isaa}, @samp{isaaplus},
11039 @samp{isab} and @samp{isac}.
11041 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
11042 code for a ColdFire target. The @var{arch} in this macro is one of the
11043 @option{-march} arguments given above.
11045 When used together, @option{-march} and @option{-mtune} select code
11046 that runs on a family of similar processors but that is optimized
11047 for a particular microarchitecture.
11049 @item -mcpu=@var{cpu}
11051 Generate code for a specific M680x0 or ColdFire processor.
11052 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
11053 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
11054 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
11055 below, which also classifies the CPUs into families:
11057 @multitable @columnfractions 0.20 0.80
11058 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
11059 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
11060 @item @samp{5206e} @tab @samp{5206e}
11061 @item @samp{5208} @tab @samp{5207} @samp{5208}
11062 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
11063 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
11064 @item @samp{5216} @tab @samp{5214} @samp{5216}
11065 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
11066 @item @samp{5225} @tab @samp{5224} @samp{5225}
11067 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
11068 @item @samp{5249} @tab @samp{5249}
11069 @item @samp{5250} @tab @samp{5250}
11070 @item @samp{5271} @tab @samp{5270} @samp{5271}
11071 @item @samp{5272} @tab @samp{5272}
11072 @item @samp{5275} @tab @samp{5274} @samp{5275}
11073 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
11074 @item @samp{5307} @tab @samp{5307}
11075 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
11076 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
11077 @item @samp{5407} @tab @samp{5407}
11078 @item @samp{5475} @tab @samp{5470} @samp{5471} @samp{5472} @samp{5473} @samp{5474} @samp{5475} @samp{547x} @samp{5480} @samp{5481} @samp{5482} @samp{5483} @samp{5484} @samp{5485}
11081 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
11082 @var{arch} is compatible with @var{cpu}. Other combinations of
11083 @option{-mcpu} and @option{-march} are rejected.
11085 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
11086 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
11087 where the value of @var{family} is given by the table above.
11089 @item -mtune=@var{tune}
11091 Tune the code for a particular microarchitecture, within the
11092 constraints set by @option{-march} and @option{-mcpu}.
11093 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
11094 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
11095 and @samp{cpu32}. The ColdFire microarchitectures
11096 are: @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
11098 You can also use @option{-mtune=68020-40} for code that needs
11099 to run relatively well on 68020, 68030 and 68040 targets.
11100 @option{-mtune=68020-60} is similar but includes 68060 targets
11101 as well. These two options select the same tuning decisions as
11102 @option{-m68020-40} and @option{-m68020-60} respectively.
11104 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
11105 when tuning for 680x0 architecture @var{arch}. It also defines
11106 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
11107 option is used. If gcc is tuning for a range of architectures,
11108 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
11109 it defines the macros for every architecture in the range.
11111 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
11112 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
11113 of the arguments given above.
11119 Generate output for a 68000. This is the default
11120 when the compiler is configured for 68000-based systems.
11121 It is equivalent to @option{-march=68000}.
11123 Use this option for microcontrollers with a 68000 or EC000 core,
11124 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
11128 Generate output for a 68010. This is the default
11129 when the compiler is configured for 68010-based systems.
11130 It is equivalent to @option{-march=68010}.
11136 Generate output for a 68020. This is the default
11137 when the compiler is configured for 68020-based systems.
11138 It is equivalent to @option{-march=68020}.
11142 Generate output for a 68030. This is the default when the compiler is
11143 configured for 68030-based systems. It is equivalent to
11144 @option{-march=68030}.
11148 Generate output for a 68040. This is the default when the compiler is
11149 configured for 68040-based systems. It is equivalent to
11150 @option{-march=68040}.
11152 This option inhibits the use of 68881/68882 instructions that have to be
11153 emulated by software on the 68040. Use this option if your 68040 does not
11154 have code to emulate those instructions.
11158 Generate output for a 68060. This is the default when the compiler is
11159 configured for 68060-based systems. It is equivalent to
11160 @option{-march=68060}.
11162 This option inhibits the use of 68020 and 68881/68882 instructions that
11163 have to be emulated by software on the 68060. Use this option if your 68060
11164 does not have code to emulate those instructions.
11168 Generate output for a CPU32. This is the default
11169 when the compiler is configured for CPU32-based systems.
11170 It is equivalent to @option{-march=cpu32}.
11172 Use this option for microcontrollers with a
11173 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
11174 68336, 68340, 68341, 68349 and 68360.
11178 Generate output for a 520X ColdFire CPU. This is the default
11179 when the compiler is configured for 520X-based systems.
11180 It is equivalent to @option{-mcpu=5206}, and is now deprecated
11181 in favor of that option.
11183 Use this option for microcontroller with a 5200 core, including
11184 the MCF5202, MCF5203, MCF5204 and MCF5206.
11188 Generate output for a 5206e ColdFire CPU. The option is now
11189 deprecated in favor of the equivalent @option{-mcpu=5206e}.
11193 Generate output for a member of the ColdFire 528X family.
11194 The option is now deprecated in favor of the equivalent
11195 @option{-mcpu=528x}.
11199 Generate output for a ColdFire 5307 CPU. The option is now deprecated
11200 in favor of the equivalent @option{-mcpu=5307}.
11204 Generate output for a ColdFire 5407 CPU. The option is now deprecated
11205 in favor of the equivalent @option{-mcpu=5407}.
11209 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
11210 This includes use of hardware floating point instructions.
11211 The option is equivalent to @option{-mcpu=547x}, and is now
11212 deprecated in favor of that option.
11216 Generate output for a 68040, without using any of the new instructions.
11217 This results in code which can run relatively efficiently on either a
11218 68020/68881 or a 68030 or a 68040. The generated code does use the
11219 68881 instructions that are emulated on the 68040.
11221 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
11225 Generate output for a 68060, without using any of the new instructions.
11226 This results in code which can run relatively efficiently on either a
11227 68020/68881 or a 68030 or a 68040. The generated code does use the
11228 68881 instructions that are emulated on the 68060.
11230 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
11234 @opindex mhard-float
11236 Generate floating-point instructions. This is the default for 68020
11237 and above, and for ColdFire devices that have an FPU. It defines the
11238 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
11239 on ColdFire targets.
11242 @opindex msoft-float
11243 Do not generate floating-point instructions; use library calls instead.
11244 This is the default for 68000, 68010, and 68832 targets. It is also
11245 the default for ColdFire devices that have no FPU.
11251 Generate (do not generate) ColdFire hardware divide and remainder
11252 instructions. If @option{-march} is used without @option{-mcpu},
11253 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
11254 architectures. Otherwise, the default is taken from the target CPU
11255 (either the default CPU, or the one specified by @option{-mcpu}). For
11256 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
11257 @option{-mcpu=5206e}.
11259 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
11263 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11264 Additionally, parameters passed on the stack are also aligned to a
11265 16-bit boundary even on targets whose API mandates promotion to 32-bit.
11269 Do not consider type @code{int} to be 16 bits wide. This is the default.
11272 @itemx -mno-bitfield
11273 @opindex mnobitfield
11274 @opindex mno-bitfield
11275 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
11276 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
11280 Do use the bit-field instructions. The @option{-m68020} option implies
11281 @option{-mbitfield}. This is the default if you use a configuration
11282 designed for a 68020.
11286 Use a different function-calling convention, in which functions
11287 that take a fixed number of arguments return with the @code{rtd}
11288 instruction, which pops their arguments while returning. This
11289 saves one instruction in the caller since there is no need to pop
11290 the arguments there.
11292 This calling convention is incompatible with the one normally
11293 used on Unix, so you cannot use it if you need to call libraries
11294 compiled with the Unix compiler.
11296 Also, you must provide function prototypes for all functions that
11297 take variable numbers of arguments (including @code{printf});
11298 otherwise incorrect code will be generated for calls to those
11301 In addition, seriously incorrect code will result if you call a
11302 function with too many arguments. (Normally, extra arguments are
11303 harmlessly ignored.)
11305 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
11306 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
11310 Do not use the calling conventions selected by @option{-mrtd}.
11311 This is the default.
11314 @itemx -mno-align-int
11315 @opindex malign-int
11316 @opindex mno-align-int
11317 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
11318 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
11319 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
11320 Aligning variables on 32-bit boundaries produces code that runs somewhat
11321 faster on processors with 32-bit busses at the expense of more memory.
11323 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
11324 align structures containing the above types differently than
11325 most published application binary interface specifications for the m68k.
11329 Use the pc-relative addressing mode of the 68000 directly, instead of
11330 using a global offset table. At present, this option implies @option{-fpic},
11331 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
11332 not presently supported with @option{-mpcrel}, though this could be supported for
11333 68020 and higher processors.
11335 @item -mno-strict-align
11336 @itemx -mstrict-align
11337 @opindex mno-strict-align
11338 @opindex mstrict-align
11339 Do not (do) assume that unaligned memory references will be handled by
11343 Generate code that allows the data segment to be located in a different
11344 area of memory from the text segment. This allows for execute in place in
11345 an environment without virtual memory management. This option implies
11348 @item -mno-sep-data
11349 Generate code that assumes that the data segment follows the text segment.
11350 This is the default.
11352 @item -mid-shared-library
11353 Generate code that supports shared libraries via the library ID method.
11354 This allows for execute in place and shared libraries in an environment
11355 without virtual memory management. This option implies @option{-fPIC}.
11357 @item -mno-id-shared-library
11358 Generate code that doesn't assume ID based shared libraries are being used.
11359 This is the default.
11361 @item -mshared-library-id=n
11362 Specified the identification number of the ID based shared library being
11363 compiled. Specifying a value of 0 will generate more compact code, specifying
11364 other values will force the allocation of that number to the current
11365 library but is no more space or time efficient than omitting this option.
11369 @node M68hc1x Options
11370 @subsection M68hc1x Options
11371 @cindex M68hc1x options
11373 These are the @samp{-m} options defined for the 68hc11 and 68hc12
11374 microcontrollers. The default values for these options depends on
11375 which style of microcontroller was selected when the compiler was configured;
11376 the defaults for the most common choices are given below.
11383 Generate output for a 68HC11. This is the default
11384 when the compiler is configured for 68HC11-based systems.
11390 Generate output for a 68HC12. This is the default
11391 when the compiler is configured for 68HC12-based systems.
11397 Generate output for a 68HCS12.
11399 @item -mauto-incdec
11400 @opindex mauto-incdec
11401 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
11408 Enable the use of 68HC12 min and max instructions.
11411 @itemx -mno-long-calls
11412 @opindex mlong-calls
11413 @opindex mno-long-calls
11414 Treat all calls as being far away (near). If calls are assumed to be
11415 far away, the compiler will use the @code{call} instruction to
11416 call a function and the @code{rtc} instruction for returning.
11420 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11422 @item -msoft-reg-count=@var{count}
11423 @opindex msoft-reg-count
11424 Specify the number of pseudo-soft registers which are used for the
11425 code generation. The maximum number is 32. Using more pseudo-soft
11426 register may or may not result in better code depending on the program.
11427 The default is 4 for 68HC11 and 2 for 68HC12.
11431 @node MCore Options
11432 @subsection MCore Options
11433 @cindex MCore options
11435 These are the @samp{-m} options defined for the Motorola M*Core
11441 @itemx -mno-hardlit
11443 @opindex mno-hardlit
11444 Inline constants into the code stream if it can be done in two
11445 instructions or less.
11451 Use the divide instruction. (Enabled by default).
11453 @item -mrelax-immediate
11454 @itemx -mno-relax-immediate
11455 @opindex mrelax-immediate
11456 @opindex mno-relax-immediate
11457 Allow arbitrary sized immediates in bit operations.
11459 @item -mwide-bitfields
11460 @itemx -mno-wide-bitfields
11461 @opindex mwide-bitfields
11462 @opindex mno-wide-bitfields
11463 Always treat bit-fields as int-sized.
11465 @item -m4byte-functions
11466 @itemx -mno-4byte-functions
11467 @opindex m4byte-functions
11468 @opindex mno-4byte-functions
11469 Force all functions to be aligned to a four byte boundary.
11471 @item -mcallgraph-data
11472 @itemx -mno-callgraph-data
11473 @opindex mcallgraph-data
11474 @opindex mno-callgraph-data
11475 Emit callgraph information.
11478 @itemx -mno-slow-bytes
11479 @opindex mslow-bytes
11480 @opindex mno-slow-bytes
11481 Prefer word access when reading byte quantities.
11483 @item -mlittle-endian
11484 @itemx -mbig-endian
11485 @opindex mlittle-endian
11486 @opindex mbig-endian
11487 Generate code for a little endian target.
11493 Generate code for the 210 processor.
11497 @subsection MIPS Options
11498 @cindex MIPS options
11504 Generate big-endian code.
11508 Generate little-endian code. This is the default for @samp{mips*el-*-*}
11511 @item -march=@var{arch}
11513 Generate code that will run on @var{arch}, which can be the name of a
11514 generic MIPS ISA, or the name of a particular processor.
11516 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
11517 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
11518 The processor names are:
11519 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
11520 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
11521 @samp{5kc}, @samp{5kf},
11523 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
11524 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
11525 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
11526 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
11529 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
11530 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
11531 @samp{rm7000}, @samp{rm9000},
11534 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
11535 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
11536 The special value @samp{from-abi} selects the
11537 most compatible architecture for the selected ABI (that is,
11538 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
11540 In processor names, a final @samp{000} can be abbreviated as @samp{k}
11541 (for example, @samp{-march=r2k}). Prefixes are optional, and
11542 @samp{vr} may be written @samp{r}.
11544 Names of the form @samp{@var{n}f2_1} refer to processors with
11545 FPUs clocked at half the rate of the core, names of the form
11546 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
11547 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
11548 processors with FPUs clocked a ratio of 3:2 with respect to the core.
11549 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
11550 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
11551 accepted as synonyms for @samp{@var{n}f1_1}.
11553 GCC defines two macros based on the value of this option. The first
11554 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
11555 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
11556 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
11557 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
11558 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
11560 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
11561 above. In other words, it will have the full prefix and will not
11562 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
11563 the macro names the resolved architecture (either @samp{"mips1"} or
11564 @samp{"mips3"}). It names the default architecture when no
11565 @option{-march} option is given.
11567 @item -mtune=@var{arch}
11569 Optimize for @var{arch}. Among other things, this option controls
11570 the way instructions are scheduled, and the perceived cost of arithmetic
11571 operations. The list of @var{arch} values is the same as for
11574 When this option is not used, GCC will optimize for the processor
11575 specified by @option{-march}. By using @option{-march} and
11576 @option{-mtune} together, it is possible to generate code that will
11577 run on a family of processors, but optimize the code for one
11578 particular member of that family.
11580 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
11581 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
11582 @samp{-march} ones described above.
11586 Equivalent to @samp{-march=mips1}.
11590 Equivalent to @samp{-march=mips2}.
11594 Equivalent to @samp{-march=mips3}.
11598 Equivalent to @samp{-march=mips4}.
11602 Equivalent to @samp{-march=mips32}.
11606 Equivalent to @samp{-march=mips32r2}.
11610 Equivalent to @samp{-march=mips64}.
11615 @opindex mno-mips16
11616 Generate (do not generate) MIPS16 code. If GCC is targetting a
11617 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
11629 Generate code for the given ABI@.
11631 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
11632 generates 64-bit code when you select a 64-bit architecture, but you
11633 can use @option{-mgp32} to get 32-bit code instead.
11635 For information about the O64 ABI, see
11636 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
11638 GCC supports a variant of the o32 ABI in which floating-point registers
11639 are 64 rather than 32 bits wide. You can select this combination with
11640 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
11641 and @samp{mfhc1} instructions and is therefore only supported for
11642 MIPS32R2 processors.
11644 The register assignments for arguments and return values remain the
11645 same, but each scalar value is passed in a single 64-bit register
11646 rather than a pair of 32-bit registers. For example, scalar
11647 floating-point values are returned in @samp{$f0} only, not a
11648 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
11649 remains the same, but all 64 bits are saved.
11652 @itemx -mno-abicalls
11654 @opindex mno-abicalls
11655 Generate (do not generate) code that is suitable for SVR4-style
11656 dynamic objects. @option{-mabicalls} is the default for SVR4-based
11661 Generate (do not generate) code that is fully position-independent,
11662 and that can therefore be linked into shared libraries. This option
11663 only affects @option{-mabicalls}.
11665 All @option{-mabicalls} code has traditionally been position-independent,
11666 regardless of options like @option{-fPIC} and @option{-fpic}. However,
11667 as an extension, the GNU toolchain allows executables to use absolute
11668 accesses for locally-binding symbols. It can also use shorter GP
11669 initialization sequences and generate direct calls to locally-defined
11670 functions. This mode is selected by @option{-mno-shared}.
11672 @option{-mno-shared} depends on binutils 2.16 or higher and generates
11673 objects that can only be linked by the GNU linker. However, the option
11674 does not affect the ABI of the final executable; it only affects the ABI
11675 of relocatable objects. Using @option{-mno-shared} will generally make
11676 executables both smaller and quicker.
11678 @option{-mshared} is the default.
11684 Lift (do not lift) the usual restrictions on the size of the global
11687 GCC normally uses a single instruction to load values from the GOT@.
11688 While this is relatively efficient, it will only work if the GOT
11689 is smaller than about 64k. Anything larger will cause the linker
11690 to report an error such as:
11692 @cindex relocation truncated to fit (MIPS)
11694 relocation truncated to fit: R_MIPS_GOT16 foobar
11697 If this happens, you should recompile your code with @option{-mxgot}.
11698 It should then work with very large GOTs, although it will also be
11699 less efficient, since it will take three instructions to fetch the
11700 value of a global symbol.
11702 Note that some linkers can create multiple GOTs. If you have such a
11703 linker, you should only need to use @option{-mxgot} when a single object
11704 file accesses more than 64k's worth of GOT entries. Very few do.
11706 These options have no effect unless GCC is generating position
11711 Assume that general-purpose registers are 32 bits wide.
11715 Assume that general-purpose registers are 64 bits wide.
11719 Assume that floating-point registers are 32 bits wide.
11723 Assume that floating-point registers are 64 bits wide.
11726 @opindex mhard-float
11727 Use floating-point coprocessor instructions.
11730 @opindex msoft-float
11731 Do not use floating-point coprocessor instructions. Implement
11732 floating-point calculations using library calls instead.
11734 @item -msingle-float
11735 @opindex msingle-float
11736 Assume that the floating-point coprocessor only supports single-precision
11739 @item -mdouble-float
11740 @opindex mdouble-float
11741 Assume that the floating-point coprocessor supports double-precision
11742 operations. This is the default.
11748 Use (do not use) revision 1 of the MIPS DSP ASE.
11749 @xref{MIPS DSP Built-in Functions}. This option defines the
11750 preprocessor macro @samp{__mips_dsp}. It also defines
11751 @samp{__mips_dsp_rev} to 1.
11757 Use (do not use) revision 2 of the MIPS DSP ASE.
11758 @xref{MIPS DSP Built-in Functions}. This option defines the
11759 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
11760 It also defines @samp{__mips_dsp_rev} to 2.
11763 @itemx -mno-smartmips
11764 @opindex msmartmips
11765 @opindex mno-smartmips
11766 Use (do not use) the MIPS SmartMIPS ASE.
11768 @item -mpaired-single
11769 @itemx -mno-paired-single
11770 @opindex mpaired-single
11771 @opindex mno-paired-single
11772 Use (do not use) paired-single floating-point instructions.
11773 @xref{MIPS Paired-Single Support}. This option can only be used
11774 when generating 64-bit code and requires hardware floating-point
11775 support to be enabled.
11781 Use (do not use) MIPS Digital Media Extension instructions.
11782 This option can only be used when generating 64-bit code and requires
11783 hardware floating-point support to be enabled.
11788 @opindex mno-mips3d
11789 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
11790 The option @option{-mips3d} implies @option{-mpaired-single}.
11796 Use (do not use) MT Multithreading instructions.
11800 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
11801 an explanation of the default and the way that the pointer size is
11806 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
11808 The default size of @code{int}s, @code{long}s and pointers depends on
11809 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
11810 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
11811 32-bit @code{long}s. Pointers are the same size as @code{long}s,
11812 or the same size as integer registers, whichever is smaller.
11818 Assume (do not assume) that all symbols have 32-bit values, regardless
11819 of the selected ABI@. This option is useful in combination with
11820 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
11821 to generate shorter and faster references to symbolic addresses.
11825 @cindex smaller data references (MIPS)
11826 @cindex gp-relative references (MIPS)
11827 Put global and static items less than or equal to @var{num} bytes into
11828 the small data or bss section instead of the normal data or bss section.
11829 This allows the data to be accessed using a single instruction.
11831 All modules should be compiled with the same @option{-G @var{num}}
11834 @item -membedded-data
11835 @itemx -mno-embedded-data
11836 @opindex membedded-data
11837 @opindex mno-embedded-data
11838 Allocate variables to the read-only data section first if possible, then
11839 next in the small data section if possible, otherwise in data. This gives
11840 slightly slower code than the default, but reduces the amount of RAM required
11841 when executing, and thus may be preferred for some embedded systems.
11843 @item -muninit-const-in-rodata
11844 @itemx -mno-uninit-const-in-rodata
11845 @opindex muninit-const-in-rodata
11846 @opindex mno-uninit-const-in-rodata
11847 Put uninitialized @code{const} variables in the read-only data section.
11848 This option is only meaningful in conjunction with @option{-membedded-data}.
11850 @item -mcode-readable=@var{setting}
11851 @opindex mcode-readable
11852 Specify whether GCC may generate code that reads from executable sections.
11853 There are three possible settings:
11856 @item -mcode-readable=yes
11857 Instructions may freely access executable sections. This is the
11860 @item -mcode-readable=pcrel
11861 MIPS16 PC-relative load instructions can access executable sections,
11862 but other instructions must not do so. This option is useful on 4KSc
11863 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
11864 It is also useful on processors that can be configured to have a dual
11865 instruction/data SRAM interface and that, like the M4K, automatically
11866 redirect PC-relative loads to the instruction RAM.
11868 @item -mcode-readable=no
11869 Instructions must not access executable sections. This option can be
11870 useful on targets that are configured to have a dual instruction/data
11871 SRAM interface but that (unlike the M4K) do not automatically redirect
11872 PC-relative loads to the instruction RAM.
11875 @item -msplit-addresses
11876 @itemx -mno-split-addresses
11877 @opindex msplit-addresses
11878 @opindex mno-split-addresses
11879 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
11880 relocation operators. This option has been superseded by
11881 @option{-mexplicit-relocs} but is retained for backwards compatibility.
11883 @item -mexplicit-relocs
11884 @itemx -mno-explicit-relocs
11885 @opindex mexplicit-relocs
11886 @opindex mno-explicit-relocs
11887 Use (do not use) assembler relocation operators when dealing with symbolic
11888 addresses. The alternative, selected by @option{-mno-explicit-relocs},
11889 is to use assembler macros instead.
11891 @option{-mexplicit-relocs} is the default if GCC was configured
11892 to use an assembler that supports relocation operators.
11894 @item -mcheck-zero-division
11895 @itemx -mno-check-zero-division
11896 @opindex mcheck-zero-division
11897 @opindex mno-check-zero-division
11898 Trap (do not trap) on integer division by zero.
11900 The default is @option{-mcheck-zero-division}.
11902 @item -mdivide-traps
11903 @itemx -mdivide-breaks
11904 @opindex mdivide-traps
11905 @opindex mdivide-breaks
11906 MIPS systems check for division by zero by generating either a
11907 conditional trap or a break instruction. Using traps results in
11908 smaller code, but is only supported on MIPS II and later. Also, some
11909 versions of the Linux kernel have a bug that prevents trap from
11910 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
11911 allow conditional traps on architectures that support them and
11912 @option{-mdivide-breaks} to force the use of breaks.
11914 The default is usually @option{-mdivide-traps}, but this can be
11915 overridden at configure time using @option{--with-divide=breaks}.
11916 Divide-by-zero checks can be completely disabled using
11917 @option{-mno-check-zero-division}.
11922 @opindex mno-memcpy
11923 Force (do not force) the use of @code{memcpy()} for non-trivial block
11924 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
11925 most constant-sized copies.
11928 @itemx -mno-long-calls
11929 @opindex mlong-calls
11930 @opindex mno-long-calls
11931 Disable (do not disable) use of the @code{jal} instruction. Calling
11932 functions using @code{jal} is more efficient but requires the caller
11933 and callee to be in the same 256 megabyte segment.
11935 This option has no effect on abicalls code. The default is
11936 @option{-mno-long-calls}.
11942 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
11943 instructions, as provided by the R4650 ISA@.
11946 @itemx -mno-fused-madd
11947 @opindex mfused-madd
11948 @opindex mno-fused-madd
11949 Enable (disable) use of the floating point multiply-accumulate
11950 instructions, when they are available. The default is
11951 @option{-mfused-madd}.
11953 When multiply-accumulate instructions are used, the intermediate
11954 product is calculated to infinite precision and is not subject to
11955 the FCSR Flush to Zero bit. This may be undesirable in some
11960 Tell the MIPS assembler to not run its preprocessor over user
11961 assembler files (with a @samp{.s} suffix) when assembling them.
11964 @itemx -mno-fix-r4000
11965 @opindex mfix-r4000
11966 @opindex mno-fix-r4000
11967 Work around certain R4000 CPU errata:
11970 A double-word or a variable shift may give an incorrect result if executed
11971 immediately after starting an integer division.
11973 A double-word or a variable shift may give an incorrect result if executed
11974 while an integer multiplication is in progress.
11976 An integer division may give an incorrect result if started in a delay slot
11977 of a taken branch or a jump.
11981 @itemx -mno-fix-r4400
11982 @opindex mfix-r4400
11983 @opindex mno-fix-r4400
11984 Work around certain R4400 CPU errata:
11987 A double-word or a variable shift may give an incorrect result if executed
11988 immediately after starting an integer division.
11992 @itemx -mno-fix-vr4120
11993 @opindex mfix-vr4120
11994 Work around certain VR4120 errata:
11997 @code{dmultu} does not always produce the correct result.
11999 @code{div} and @code{ddiv} do not always produce the correct result if one
12000 of the operands is negative.
12002 The workarounds for the division errata rely on special functions in
12003 @file{libgcc.a}. At present, these functions are only provided by
12004 the @code{mips64vr*-elf} configurations.
12006 Other VR4120 errata require a nop to be inserted between certain pairs of
12007 instructions. These errata are handled by the assembler, not by GCC itself.
12010 @opindex mfix-vr4130
12011 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
12012 workarounds are implemented by the assembler rather than by GCC,
12013 although GCC will avoid using @code{mflo} and @code{mfhi} if the
12014 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
12015 instructions are available instead.
12018 @itemx -mno-fix-sb1
12020 Work around certain SB-1 CPU core errata.
12021 (This flag currently works around the SB-1 revision 2
12022 ``F1'' and ``F2'' floating point errata.)
12024 @item -mflush-func=@var{func}
12025 @itemx -mno-flush-func
12026 @opindex mflush-func
12027 Specifies the function to call to flush the I and D caches, or to not
12028 call any such function. If called, the function must take the same
12029 arguments as the common @code{_flush_func()}, that is, the address of the
12030 memory range for which the cache is being flushed, the size of the
12031 memory range, and the number 3 (to flush both caches). The default
12032 depends on the target GCC was configured for, but commonly is either
12033 @samp{_flush_func} or @samp{__cpu_flush}.
12035 @item mbranch-cost=@var{num}
12036 @opindex mbranch-cost
12037 Set the cost of branches to roughly @var{num} ``simple'' instructions.
12038 This cost is only a heuristic and is not guaranteed to produce
12039 consistent results across releases. A zero cost redundantly selects
12040 the default, which is based on the @option{-mtune} setting.
12042 @item -mbranch-likely
12043 @itemx -mno-branch-likely
12044 @opindex mbranch-likely
12045 @opindex mno-branch-likely
12046 Enable or disable use of Branch Likely instructions, regardless of the
12047 default for the selected architecture. By default, Branch Likely
12048 instructions may be generated if they are supported by the selected
12049 architecture. An exception is for the MIPS32 and MIPS64 architectures
12050 and processors which implement those architectures; for those, Branch
12051 Likely instructions will not be generated by default because the MIPS32
12052 and MIPS64 architectures specifically deprecate their use.
12054 @item -mfp-exceptions
12055 @itemx -mno-fp-exceptions
12056 @opindex mfp-exceptions
12057 Specifies whether FP exceptions are enabled. This affects how we schedule
12058 FP instructions for some processors. The default is that FP exceptions are
12061 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
12062 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
12065 @item -mvr4130-align
12066 @itemx -mno-vr4130-align
12067 @opindex mvr4130-align
12068 The VR4130 pipeline is two-way superscalar, but can only issue two
12069 instructions together if the first one is 8-byte aligned. When this
12070 option is enabled, GCC will align pairs of instructions that it
12071 thinks should execute in parallel.
12073 This option only has an effect when optimizing for the VR4130.
12074 It normally makes code faster, but at the expense of making it bigger.
12075 It is enabled by default at optimization level @option{-O3}.
12079 @subsection MMIX Options
12080 @cindex MMIX Options
12082 These options are defined for the MMIX:
12086 @itemx -mno-libfuncs
12088 @opindex mno-libfuncs
12089 Specify that intrinsic library functions are being compiled, passing all
12090 values in registers, no matter the size.
12093 @itemx -mno-epsilon
12095 @opindex mno-epsilon
12096 Generate floating-point comparison instructions that compare with respect
12097 to the @code{rE} epsilon register.
12099 @item -mabi=mmixware
12101 @opindex mabi-mmixware
12103 Generate code that passes function parameters and return values that (in
12104 the called function) are seen as registers @code{$0} and up, as opposed to
12105 the GNU ABI which uses global registers @code{$231} and up.
12107 @item -mzero-extend
12108 @itemx -mno-zero-extend
12109 @opindex mzero-extend
12110 @opindex mno-zero-extend
12111 When reading data from memory in sizes shorter than 64 bits, use (do not
12112 use) zero-extending load instructions by default, rather than
12113 sign-extending ones.
12116 @itemx -mno-knuthdiv
12118 @opindex mno-knuthdiv
12119 Make the result of a division yielding a remainder have the same sign as
12120 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
12121 remainder follows the sign of the dividend. Both methods are
12122 arithmetically valid, the latter being almost exclusively used.
12124 @item -mtoplevel-symbols
12125 @itemx -mno-toplevel-symbols
12126 @opindex mtoplevel-symbols
12127 @opindex mno-toplevel-symbols
12128 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
12129 code can be used with the @code{PREFIX} assembly directive.
12133 Generate an executable in the ELF format, rather than the default
12134 @samp{mmo} format used by the @command{mmix} simulator.
12136 @item -mbranch-predict
12137 @itemx -mno-branch-predict
12138 @opindex mbranch-predict
12139 @opindex mno-branch-predict
12140 Use (do not use) the probable-branch instructions, when static branch
12141 prediction indicates a probable branch.
12143 @item -mbase-addresses
12144 @itemx -mno-base-addresses
12145 @opindex mbase-addresses
12146 @opindex mno-base-addresses
12147 Generate (do not generate) code that uses @emph{base addresses}. Using a
12148 base address automatically generates a request (handled by the assembler
12149 and the linker) for a constant to be set up in a global register. The
12150 register is used for one or more base address requests within the range 0
12151 to 255 from the value held in the register. The generally leads to short
12152 and fast code, but the number of different data items that can be
12153 addressed is limited. This means that a program that uses lots of static
12154 data may require @option{-mno-base-addresses}.
12156 @item -msingle-exit
12157 @itemx -mno-single-exit
12158 @opindex msingle-exit
12159 @opindex mno-single-exit
12160 Force (do not force) generated code to have a single exit point in each
12164 @node MN10300 Options
12165 @subsection MN10300 Options
12166 @cindex MN10300 options
12168 These @option{-m} options are defined for Matsushita MN10300 architectures:
12173 Generate code to avoid bugs in the multiply instructions for the MN10300
12174 processors. This is the default.
12176 @item -mno-mult-bug
12177 @opindex mno-mult-bug
12178 Do not generate code to avoid bugs in the multiply instructions for the
12179 MN10300 processors.
12183 Generate code which uses features specific to the AM33 processor.
12187 Do not generate code which uses features specific to the AM33 processor. This
12190 @item -mreturn-pointer-on-d0
12191 @opindex mreturn-pointer-on-d0
12192 When generating a function which returns a pointer, return the pointer
12193 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
12194 only in a0, and attempts to call such functions without a prototype
12195 would result in errors. Note that this option is on by default; use
12196 @option{-mno-return-pointer-on-d0} to disable it.
12200 Do not link in the C run-time initialization object file.
12204 Indicate to the linker that it should perform a relaxation optimization pass
12205 to shorten branches, calls and absolute memory addresses. This option only
12206 has an effect when used on the command line for the final link step.
12208 This option makes symbolic debugging impossible.
12212 @subsection MT Options
12215 These @option{-m} options are defined for Morpho MT architectures:
12219 @item -march=@var{cpu-type}
12221 Generate code that will run on @var{cpu-type}, which is the name of a system
12222 representing a certain processor type. Possible values for
12223 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
12224 @samp{ms1-16-003} and @samp{ms2}.
12226 When this option is not used, the default is @option{-march=ms1-16-002}.
12230 Use byte loads and stores when generating code.
12234 Do not use byte loads and stores when generating code.
12238 Use simulator runtime
12242 Do not link in the C run-time initialization object file
12243 @file{crti.o}. Other run-time initialization and termination files
12244 such as @file{startup.o} and @file{exit.o} are still included on the
12245 linker command line.
12249 @node PDP-11 Options
12250 @subsection PDP-11 Options
12251 @cindex PDP-11 Options
12253 These options are defined for the PDP-11:
12258 Use hardware FPP floating point. This is the default. (FIS floating
12259 point on the PDP-11/40 is not supported.)
12262 @opindex msoft-float
12263 Do not use hardware floating point.
12267 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
12271 Return floating-point results in memory. This is the default.
12275 Generate code for a PDP-11/40.
12279 Generate code for a PDP-11/45. This is the default.
12283 Generate code for a PDP-11/10.
12285 @item -mbcopy-builtin
12286 @opindex bcopy-builtin
12287 Use inline @code{movmemhi} patterns for copying memory. This is the
12292 Do not use inline @code{movmemhi} patterns for copying memory.
12298 Use 16-bit @code{int}. This is the default.
12304 Use 32-bit @code{int}.
12307 @itemx -mno-float32
12309 @opindex mno-float32
12310 Use 64-bit @code{float}. This is the default.
12313 @itemx -mno-float64
12315 @opindex mno-float64
12316 Use 32-bit @code{float}.
12320 Use @code{abshi2} pattern. This is the default.
12324 Do not use @code{abshi2} pattern.
12326 @item -mbranch-expensive
12327 @opindex mbranch-expensive
12328 Pretend that branches are expensive. This is for experimenting with
12329 code generation only.
12331 @item -mbranch-cheap
12332 @opindex mbranch-cheap
12333 Do not pretend that branches are expensive. This is the default.
12337 Generate code for a system with split I&D@.
12341 Generate code for a system without split I&D@. This is the default.
12345 Use Unix assembler syntax. This is the default when configured for
12346 @samp{pdp11-*-bsd}.
12350 Use DEC assembler syntax. This is the default when configured for any
12351 PDP-11 target other than @samp{pdp11-*-bsd}.
12354 @node PowerPC Options
12355 @subsection PowerPC Options
12356 @cindex PowerPC options
12358 These are listed under @xref{RS/6000 and PowerPC Options}.
12360 @node RS/6000 and PowerPC Options
12361 @subsection IBM RS/6000 and PowerPC Options
12362 @cindex RS/6000 and PowerPC Options
12363 @cindex IBM RS/6000 and PowerPC Options
12365 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
12372 @itemx -mno-powerpc
12373 @itemx -mpowerpc-gpopt
12374 @itemx -mno-powerpc-gpopt
12375 @itemx -mpowerpc-gfxopt
12376 @itemx -mno-powerpc-gfxopt
12378 @itemx -mno-powerpc64
12382 @itemx -mno-popcntb
12394 @opindex mno-power2
12396 @opindex mno-powerpc
12397 @opindex mpowerpc-gpopt
12398 @opindex mno-powerpc-gpopt
12399 @opindex mpowerpc-gfxopt
12400 @opindex mno-powerpc-gfxopt
12401 @opindex mpowerpc64
12402 @opindex mno-powerpc64
12406 @opindex mno-popcntb
12412 @opindex mno-mfpgpr
12415 GCC supports two related instruction set architectures for the
12416 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
12417 instructions supported by the @samp{rios} chip set used in the original
12418 RS/6000 systems and the @dfn{PowerPC} instruction set is the
12419 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
12420 the IBM 4xx, 6xx, and follow-on microprocessors.
12422 Neither architecture is a subset of the other. However there is a
12423 large common subset of instructions supported by both. An MQ
12424 register is included in processors supporting the POWER architecture.
12426 You use these options to specify which instructions are available on the
12427 processor you are using. The default value of these options is
12428 determined when configuring GCC@. Specifying the
12429 @option{-mcpu=@var{cpu_type}} overrides the specification of these
12430 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
12431 rather than the options listed above.
12433 The @option{-mpower} option allows GCC to generate instructions that
12434 are found only in the POWER architecture and to use the MQ register.
12435 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
12436 to generate instructions that are present in the POWER2 architecture but
12437 not the original POWER architecture.
12439 The @option{-mpowerpc} option allows GCC to generate instructions that
12440 are found only in the 32-bit subset of the PowerPC architecture.
12441 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
12442 GCC to use the optional PowerPC architecture instructions in the
12443 General Purpose group, including floating-point square root. Specifying
12444 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
12445 use the optional PowerPC architecture instructions in the Graphics
12446 group, including floating-point select.
12448 The @option{-mmfcrf} option allows GCC to generate the move from
12449 condition register field instruction implemented on the POWER4
12450 processor and other processors that support the PowerPC V2.01
12452 The @option{-mpopcntb} option allows GCC to generate the popcount and
12453 double precision FP reciprocal estimate instruction implemented on the
12454 POWER5 processor and other processors that support the PowerPC V2.02
12456 The @option{-mfprnd} option allows GCC to generate the FP round to
12457 integer instructions implemented on the POWER5+ processor and other
12458 processors that support the PowerPC V2.03 architecture.
12459 The @option{-mcmpb} option allows GCC to generate the compare bytes
12460 instruction implemented on the POWER6 processor and other processors
12461 that support the PowerPC V2.05 architecture.
12462 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
12463 general purpose register instructions implemented on the POWER6X
12464 processor and other processors that support the extended PowerPC V2.05
12466 The @option{-mdfp} option allows GCC to generate the decimal floating
12467 point instructions implemented on some POWER processors.
12469 The @option{-mpowerpc64} option allows GCC to generate the additional
12470 64-bit instructions that are found in the full PowerPC64 architecture
12471 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
12472 @option{-mno-powerpc64}.
12474 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
12475 will use only the instructions in the common subset of both
12476 architectures plus some special AIX common-mode calls, and will not use
12477 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
12478 permits GCC to use any instruction from either architecture and to
12479 allow use of the MQ register; specify this for the Motorola MPC601.
12481 @item -mnew-mnemonics
12482 @itemx -mold-mnemonics
12483 @opindex mnew-mnemonics
12484 @opindex mold-mnemonics
12485 Select which mnemonics to use in the generated assembler code. With
12486 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
12487 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
12488 assembler mnemonics defined for the POWER architecture. Instructions
12489 defined in only one architecture have only one mnemonic; GCC uses that
12490 mnemonic irrespective of which of these options is specified.
12492 GCC defaults to the mnemonics appropriate for the architecture in
12493 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
12494 value of these option. Unless you are building a cross-compiler, you
12495 should normally not specify either @option{-mnew-mnemonics} or
12496 @option{-mold-mnemonics}, but should instead accept the default.
12498 @item -mcpu=@var{cpu_type}
12500 Set architecture type, register usage, choice of mnemonics, and
12501 instruction scheduling parameters for machine type @var{cpu_type}.
12502 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
12503 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
12504 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
12505 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
12506 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
12507 @samp{860}, @samp{970}, @samp{8540}, @samp{ec603e}, @samp{G3},
12508 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
12509 @samp{power4}, @samp{power5}, @samp{power5+}, @samp{power6},
12510 @samp{power6x}, @samp{common}, @samp{powerpc}, @samp{powerpc64},
12511 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
12513 @option{-mcpu=common} selects a completely generic processor. Code
12514 generated under this option will run on any POWER or PowerPC processor.
12515 GCC will use only the instructions in the common subset of both
12516 architectures, and will not use the MQ register. GCC assumes a generic
12517 processor model for scheduling purposes.
12519 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
12520 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
12521 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
12522 types, with an appropriate, generic processor model assumed for
12523 scheduling purposes.
12525 The other options specify a specific processor. Code generated under
12526 those options will run best on that processor, and may not run at all on
12529 The @option{-mcpu} options automatically enable or disable the
12532 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
12533 -mnew-mnemonics -mpopcntb -mpower -mpower2 -mpowerpc64 @gol
12534 -mpowerpc-gpopt -mpowerpc-gfxopt -mstring -mmulhw -mdlmzb -mmfpgpr}
12536 The particular options set for any particular CPU will vary between
12537 compiler versions, depending on what setting seems to produce optimal
12538 code for that CPU; it doesn't necessarily reflect the actual hardware's
12539 capabilities. If you wish to set an individual option to a particular
12540 value, you may specify it after the @option{-mcpu} option, like
12541 @samp{-mcpu=970 -mno-altivec}.
12543 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
12544 not enabled or disabled by the @option{-mcpu} option at present because
12545 AIX does not have full support for these options. You may still
12546 enable or disable them individually if you're sure it'll work in your
12549 @item -mtune=@var{cpu_type}
12551 Set the instruction scheduling parameters for machine type
12552 @var{cpu_type}, but do not set the architecture type, register usage, or
12553 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
12554 values for @var{cpu_type} are used for @option{-mtune} as for
12555 @option{-mcpu}. If both are specified, the code generated will use the
12556 architecture, registers, and mnemonics set by @option{-mcpu}, but the
12557 scheduling parameters set by @option{-mtune}.
12563 Generate code to compute division as reciprocal estimate and iterative
12564 refinement, creating opportunities for increased throughput. This
12565 feature requires: optional PowerPC Graphics instruction set for single
12566 precision and FRE instruction for double precision, assuming divides
12567 cannot generate user-visible traps, and the domain values not include
12568 Infinities, denormals or zero denominator.
12571 @itemx -mno-altivec
12573 @opindex mno-altivec
12574 Generate code that uses (does not use) AltiVec instructions, and also
12575 enable the use of built-in functions that allow more direct access to
12576 the AltiVec instruction set. You may also need to set
12577 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
12583 @opindex mno-vrsave
12584 Generate VRSAVE instructions when generating AltiVec code.
12587 @opindex msecure-plt
12588 Generate code that allows ld and ld.so to build executables and shared
12589 libraries with non-exec .plt and .got sections. This is a PowerPC
12590 32-bit SYSV ABI option.
12594 Generate code that uses a BSS .plt section that ld.so fills in, and
12595 requires .plt and .got sections that are both writable and executable.
12596 This is a PowerPC 32-bit SYSV ABI option.
12602 This switch enables or disables the generation of ISEL instructions.
12604 @item -misel=@var{yes/no}
12605 This switch has been deprecated. Use @option{-misel} and
12606 @option{-mno-isel} instead.
12612 This switch enables or disables the generation of SPE simd
12618 @opindex mno-paired
12619 This switch enables or disables the generation of PAIRED simd
12622 @item -mspe=@var{yes/no}
12623 This option has been deprecated. Use @option{-mspe} and
12624 @option{-mno-spe} instead.
12626 @item -mfloat-gprs=@var{yes/single/double/no}
12627 @itemx -mfloat-gprs
12628 @opindex mfloat-gprs
12629 This switch enables or disables the generation of floating point
12630 operations on the general purpose registers for architectures that
12633 The argument @var{yes} or @var{single} enables the use of
12634 single-precision floating point operations.
12636 The argument @var{double} enables the use of single and
12637 double-precision floating point operations.
12639 The argument @var{no} disables floating point operations on the
12640 general purpose registers.
12642 This option is currently only available on the MPC854x.
12648 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
12649 targets (including GNU/Linux). The 32-bit environment sets int, long
12650 and pointer to 32 bits and generates code that runs on any PowerPC
12651 variant. The 64-bit environment sets int to 32 bits and long and
12652 pointer to 64 bits, and generates code for PowerPC64, as for
12653 @option{-mpowerpc64}.
12656 @itemx -mno-fp-in-toc
12657 @itemx -mno-sum-in-toc
12658 @itemx -mminimal-toc
12660 @opindex mno-fp-in-toc
12661 @opindex mno-sum-in-toc
12662 @opindex mminimal-toc
12663 Modify generation of the TOC (Table Of Contents), which is created for
12664 every executable file. The @option{-mfull-toc} option is selected by
12665 default. In that case, GCC will allocate at least one TOC entry for
12666 each unique non-automatic variable reference in your program. GCC
12667 will also place floating-point constants in the TOC@. However, only
12668 16,384 entries are available in the TOC@.
12670 If you receive a linker error message that saying you have overflowed
12671 the available TOC space, you can reduce the amount of TOC space used
12672 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
12673 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
12674 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
12675 generate code to calculate the sum of an address and a constant at
12676 run-time instead of putting that sum into the TOC@. You may specify one
12677 or both of these options. Each causes GCC to produce very slightly
12678 slower and larger code at the expense of conserving TOC space.
12680 If you still run out of space in the TOC even when you specify both of
12681 these options, specify @option{-mminimal-toc} instead. This option causes
12682 GCC to make only one TOC entry for every file. When you specify this
12683 option, GCC will produce code that is slower and larger but which
12684 uses extremely little TOC space. You may wish to use this option
12685 only on files that contain less frequently executed code.
12691 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
12692 @code{long} type, and the infrastructure needed to support them.
12693 Specifying @option{-maix64} implies @option{-mpowerpc64} and
12694 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
12695 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
12698 @itemx -mno-xl-compat
12699 @opindex mxl-compat
12700 @opindex mno-xl-compat
12701 Produce code that conforms more closely to IBM XL compiler semantics
12702 when using AIX-compatible ABI. Pass floating-point arguments to
12703 prototyped functions beyond the register save area (RSA) on the stack
12704 in addition to argument FPRs. Do not assume that most significant
12705 double in 128-bit long double value is properly rounded when comparing
12706 values and converting to double. Use XL symbol names for long double
12709 The AIX calling convention was extended but not initially documented to
12710 handle an obscure K&R C case of calling a function that takes the
12711 address of its arguments with fewer arguments than declared. IBM XL
12712 compilers access floating point arguments which do not fit in the
12713 RSA from the stack when a subroutine is compiled without
12714 optimization. Because always storing floating-point arguments on the
12715 stack is inefficient and rarely needed, this option is not enabled by
12716 default and only is necessary when calling subroutines compiled by IBM
12717 XL compilers without optimization.
12721 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
12722 application written to use message passing with special startup code to
12723 enable the application to run. The system must have PE installed in the
12724 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
12725 must be overridden with the @option{-specs=} option to specify the
12726 appropriate directory location. The Parallel Environment does not
12727 support threads, so the @option{-mpe} option and the @option{-pthread}
12728 option are incompatible.
12730 @item -malign-natural
12731 @itemx -malign-power
12732 @opindex malign-natural
12733 @opindex malign-power
12734 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
12735 @option{-malign-natural} overrides the ABI-defined alignment of larger
12736 types, such as floating-point doubles, on their natural size-based boundary.
12737 The option @option{-malign-power} instructs GCC to follow the ABI-specified
12738 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
12740 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
12744 @itemx -mhard-float
12745 @opindex msoft-float
12746 @opindex mhard-float
12747 Generate code that does not use (uses) the floating-point register set.
12748 Software floating point emulation is provided if you use the
12749 @option{-msoft-float} option, and pass the option to GCC when linking.
12752 @itemx -mno-multiple
12754 @opindex mno-multiple
12755 Generate code that uses (does not use) the load multiple word
12756 instructions and the store multiple word instructions. These
12757 instructions are generated by default on POWER systems, and not
12758 generated on PowerPC systems. Do not use @option{-mmultiple} on little
12759 endian PowerPC systems, since those instructions do not work when the
12760 processor is in little endian mode. The exceptions are PPC740 and
12761 PPC750 which permit the instructions usage in little endian mode.
12766 @opindex mno-string
12767 Generate code that uses (does not use) the load string instructions
12768 and the store string word instructions to save multiple registers and
12769 do small block moves. These instructions are generated by default on
12770 POWER systems, and not generated on PowerPC systems. Do not use
12771 @option{-mstring} on little endian PowerPC systems, since those
12772 instructions do not work when the processor is in little endian mode.
12773 The exceptions are PPC740 and PPC750 which permit the instructions
12774 usage in little endian mode.
12779 @opindex mno-update
12780 Generate code that uses (does not use) the load or store instructions
12781 that update the base register to the address of the calculated memory
12782 location. These instructions are generated by default. If you use
12783 @option{-mno-update}, there is a small window between the time that the
12784 stack pointer is updated and the address of the previous frame is
12785 stored, which means code that walks the stack frame across interrupts or
12786 signals may get corrupted data.
12789 @itemx -mno-fused-madd
12790 @opindex mfused-madd
12791 @opindex mno-fused-madd
12792 Generate code that uses (does not use) the floating point multiply and
12793 accumulate instructions. These instructions are generated by default if
12794 hardware floating is used.
12800 Generate code that uses (does not use) the half-word multiply and
12801 multiply-accumulate instructions on the IBM 405 and 440 processors.
12802 These instructions are generated by default when targetting those
12809 Generate code that uses (does not use) the string-search @samp{dlmzb}
12810 instruction on the IBM 405 and 440 processors. This instruction is
12811 generated by default when targetting those processors.
12813 @item -mno-bit-align
12815 @opindex mno-bit-align
12816 @opindex mbit-align
12817 On System V.4 and embedded PowerPC systems do not (do) force structures
12818 and unions that contain bit-fields to be aligned to the base type of the
12821 For example, by default a structure containing nothing but 8
12822 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
12823 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
12824 the structure would be aligned to a 1 byte boundary and be one byte in
12827 @item -mno-strict-align
12828 @itemx -mstrict-align
12829 @opindex mno-strict-align
12830 @opindex mstrict-align
12831 On System V.4 and embedded PowerPC systems do not (do) assume that
12832 unaligned memory references will be handled by the system.
12834 @item -mrelocatable
12835 @itemx -mno-relocatable
12836 @opindex mrelocatable
12837 @opindex mno-relocatable
12838 On embedded PowerPC systems generate code that allows (does not allow)
12839 the program to be relocated to a different address at runtime. If you
12840 use @option{-mrelocatable} on any module, all objects linked together must
12841 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
12843 @item -mrelocatable-lib
12844 @itemx -mno-relocatable-lib
12845 @opindex mrelocatable-lib
12846 @opindex mno-relocatable-lib
12847 On embedded PowerPC systems generate code that allows (does not allow)
12848 the program to be relocated to a different address at runtime. Modules
12849 compiled with @option{-mrelocatable-lib} can be linked with either modules
12850 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
12851 with modules compiled with the @option{-mrelocatable} options.
12857 On System V.4 and embedded PowerPC systems do not (do) assume that
12858 register 2 contains a pointer to a global area pointing to the addresses
12859 used in the program.
12862 @itemx -mlittle-endian
12864 @opindex mlittle-endian
12865 On System V.4 and embedded PowerPC systems compile code for the
12866 processor in little endian mode. The @option{-mlittle-endian} option is
12867 the same as @option{-mlittle}.
12870 @itemx -mbig-endian
12872 @opindex mbig-endian
12873 On System V.4 and embedded PowerPC systems compile code for the
12874 processor in big endian mode. The @option{-mbig-endian} option is
12875 the same as @option{-mbig}.
12877 @item -mdynamic-no-pic
12878 @opindex mdynamic-no-pic
12879 On Darwin and Mac OS X systems, compile code so that it is not
12880 relocatable, but that its external references are relocatable. The
12881 resulting code is suitable for applications, but not shared
12884 @item -mprioritize-restricted-insns=@var{priority}
12885 @opindex mprioritize-restricted-insns
12886 This option controls the priority that is assigned to
12887 dispatch-slot restricted instructions during the second scheduling
12888 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
12889 @var{no/highest/second-highest} priority to dispatch slot restricted
12892 @item -msched-costly-dep=@var{dependence_type}
12893 @opindex msched-costly-dep
12894 This option controls which dependences are considered costly
12895 by the target during instruction scheduling. The argument
12896 @var{dependence_type} takes one of the following values:
12897 @var{no}: no dependence is costly,
12898 @var{all}: all dependences are costly,
12899 @var{true_store_to_load}: a true dependence from store to load is costly,
12900 @var{store_to_load}: any dependence from store to load is costly,
12901 @var{number}: any dependence which latency >= @var{number} is costly.
12903 @item -minsert-sched-nops=@var{scheme}
12904 @opindex minsert-sched-nops
12905 This option controls which nop insertion scheme will be used during
12906 the second scheduling pass. The argument @var{scheme} takes one of the
12908 @var{no}: Don't insert nops.
12909 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
12910 according to the scheduler's grouping.
12911 @var{regroup_exact}: Insert nops to force costly dependent insns into
12912 separate groups. Insert exactly as many nops as needed to force an insn
12913 to a new group, according to the estimated processor grouping.
12914 @var{number}: Insert nops to force costly dependent insns into
12915 separate groups. Insert @var{number} nops to force an insn to a new group.
12918 @opindex mcall-sysv
12919 On System V.4 and embedded PowerPC systems compile code using calling
12920 conventions that adheres to the March 1995 draft of the System V
12921 Application Binary Interface, PowerPC processor supplement. This is the
12922 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
12924 @item -mcall-sysv-eabi
12925 @opindex mcall-sysv-eabi
12926 Specify both @option{-mcall-sysv} and @option{-meabi} options.
12928 @item -mcall-sysv-noeabi
12929 @opindex mcall-sysv-noeabi
12930 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
12932 @item -mcall-solaris
12933 @opindex mcall-solaris
12934 On System V.4 and embedded PowerPC systems compile code for the Solaris
12938 @opindex mcall-linux
12939 On System V.4 and embedded PowerPC systems compile code for the
12940 Linux-based GNU system.
12944 On System V.4 and embedded PowerPC systems compile code for the
12945 Hurd-based GNU system.
12947 @item -mcall-netbsd
12948 @opindex mcall-netbsd
12949 On System V.4 and embedded PowerPC systems compile code for the
12950 NetBSD operating system.
12952 @item -maix-struct-return
12953 @opindex maix-struct-return
12954 Return all structures in memory (as specified by the AIX ABI)@.
12956 @item -msvr4-struct-return
12957 @opindex msvr4-struct-return
12958 Return structures smaller than 8 bytes in registers (as specified by the
12961 @item -mabi=@var{abi-type}
12963 Extend the current ABI with a particular extension, or remove such extension.
12964 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
12965 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
12969 Extend the current ABI with SPE ABI extensions. This does not change
12970 the default ABI, instead it adds the SPE ABI extensions to the current
12974 @opindex mabi=no-spe
12975 Disable Booke SPE ABI extensions for the current ABI@.
12977 @item -mabi=ibmlongdouble
12978 @opindex mabi=ibmlongdouble
12979 Change the current ABI to use IBM extended precision long double.
12980 This is a PowerPC 32-bit SYSV ABI option.
12982 @item -mabi=ieeelongdouble
12983 @opindex mabi=ieeelongdouble
12984 Change the current ABI to use IEEE extended precision long double.
12985 This is a PowerPC 32-bit Linux ABI option.
12988 @itemx -mno-prototype
12989 @opindex mprototype
12990 @opindex mno-prototype
12991 On System V.4 and embedded PowerPC systems assume that all calls to
12992 variable argument functions are properly prototyped. Otherwise, the
12993 compiler must insert an instruction before every non prototyped call to
12994 set or clear bit 6 of the condition code register (@var{CR}) to
12995 indicate whether floating point values were passed in the floating point
12996 registers in case the function takes a variable arguments. With
12997 @option{-mprototype}, only calls to prototyped variable argument functions
12998 will set or clear the bit.
13002 On embedded PowerPC systems, assume that the startup module is called
13003 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
13004 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
13009 On embedded PowerPC systems, assume that the startup module is called
13010 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
13015 On embedded PowerPC systems, assume that the startup module is called
13016 @file{crt0.o} and the standard C libraries are @file{libads.a} and
13019 @item -myellowknife
13020 @opindex myellowknife
13021 On embedded PowerPC systems, assume that the startup module is called
13022 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
13027 On System V.4 and embedded PowerPC systems, specify that you are
13028 compiling for a VxWorks system.
13032 Specify that you are compiling for the WindISS simulation environment.
13036 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
13037 header to indicate that @samp{eabi} extended relocations are used.
13043 On System V.4 and embedded PowerPC systems do (do not) adhere to the
13044 Embedded Applications Binary Interface (eabi) which is a set of
13045 modifications to the System V.4 specifications. Selecting @option{-meabi}
13046 means that the stack is aligned to an 8 byte boundary, a function
13047 @code{__eabi} is called to from @code{main} to set up the eabi
13048 environment, and the @option{-msdata} option can use both @code{r2} and
13049 @code{r13} to point to two separate small data areas. Selecting
13050 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
13051 do not call an initialization function from @code{main}, and the
13052 @option{-msdata} option will only use @code{r13} to point to a single
13053 small data area. The @option{-meabi} option is on by default if you
13054 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
13057 @opindex msdata=eabi
13058 On System V.4 and embedded PowerPC systems, put small initialized
13059 @code{const} global and static data in the @samp{.sdata2} section, which
13060 is pointed to by register @code{r2}. Put small initialized
13061 non-@code{const} global and static data in the @samp{.sdata} section,
13062 which is pointed to by register @code{r13}. Put small uninitialized
13063 global and static data in the @samp{.sbss} section, which is adjacent to
13064 the @samp{.sdata} section. The @option{-msdata=eabi} option is
13065 incompatible with the @option{-mrelocatable} option. The
13066 @option{-msdata=eabi} option also sets the @option{-memb} option.
13069 @opindex msdata=sysv
13070 On System V.4 and embedded PowerPC systems, put small global and static
13071 data in the @samp{.sdata} section, which is pointed to by register
13072 @code{r13}. Put small uninitialized global and static data in the
13073 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
13074 The @option{-msdata=sysv} option is incompatible with the
13075 @option{-mrelocatable} option.
13077 @item -msdata=default
13079 @opindex msdata=default
13081 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
13082 compile code the same as @option{-msdata=eabi}, otherwise compile code the
13083 same as @option{-msdata=sysv}.
13086 @opindex msdata-data
13087 On System V.4 and embedded PowerPC systems, put small global
13088 data in the @samp{.sdata} section. Put small uninitialized global
13089 data in the @samp{.sbss} section. Do not use register @code{r13}
13090 to address small data however. This is the default behavior unless
13091 other @option{-msdata} options are used.
13095 @opindex msdata=none
13097 On embedded PowerPC systems, put all initialized global and static data
13098 in the @samp{.data} section, and all uninitialized data in the
13099 @samp{.bss} section.
13103 @cindex smaller data references (PowerPC)
13104 @cindex .sdata/.sdata2 references (PowerPC)
13105 On embedded PowerPC systems, put global and static items less than or
13106 equal to @var{num} bytes into the small data or bss sections instead of
13107 the normal data or bss section. By default, @var{num} is 8. The
13108 @option{-G @var{num}} switch is also passed to the linker.
13109 All modules should be compiled with the same @option{-G @var{num}} value.
13112 @itemx -mno-regnames
13114 @opindex mno-regnames
13115 On System V.4 and embedded PowerPC systems do (do not) emit register
13116 names in the assembly language output using symbolic forms.
13119 @itemx -mno-longcall
13121 @opindex mno-longcall
13122 By default assume that all calls are far away so that a longer more
13123 expensive calling sequence is required. This is required for calls
13124 further than 32 megabytes (33,554,432 bytes) from the current location.
13125 A short call will be generated if the compiler knows
13126 the call cannot be that far away. This setting can be overridden by
13127 the @code{shortcall} function attribute, or by @code{#pragma
13130 Some linkers are capable of detecting out-of-range calls and generating
13131 glue code on the fly. On these systems, long calls are unnecessary and
13132 generate slower code. As of this writing, the AIX linker can do this,
13133 as can the GNU linker for PowerPC/64. It is planned to add this feature
13134 to the GNU linker for 32-bit PowerPC systems as well.
13136 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
13137 callee, L42'', plus a ``branch island'' (glue code). The two target
13138 addresses represent the callee and the ``branch island''. The
13139 Darwin/PPC linker will prefer the first address and generate a ``bl
13140 callee'' if the PPC ``bl'' instruction will reach the callee directly;
13141 otherwise, the linker will generate ``bl L42'' to call the ``branch
13142 island''. The ``branch island'' is appended to the body of the
13143 calling function; it computes the full 32-bit address of the callee
13146 On Mach-O (Darwin) systems, this option directs the compiler emit to
13147 the glue for every direct call, and the Darwin linker decides whether
13148 to use or discard it.
13150 In the future, we may cause GCC to ignore all longcall specifications
13151 when the linker is known to generate glue.
13155 Adds support for multithreading with the @dfn{pthreads} library.
13156 This option sets flags for both the preprocessor and linker.
13160 @node S/390 and zSeries Options
13161 @subsection S/390 and zSeries Options
13162 @cindex S/390 and zSeries Options
13164 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
13168 @itemx -msoft-float
13169 @opindex mhard-float
13170 @opindex msoft-float
13171 Use (do not use) the hardware floating-point instructions and registers
13172 for floating-point operations. When @option{-msoft-float} is specified,
13173 functions in @file{libgcc.a} will be used to perform floating-point
13174 operations. When @option{-mhard-float} is specified, the compiler
13175 generates IEEE floating-point instructions. This is the default.
13177 @item -mlong-double-64
13178 @itemx -mlong-double-128
13179 @opindex mlong-double-64
13180 @opindex mlong-double-128
13181 These switches control the size of @code{long double} type. A size
13182 of 64bit makes the @code{long double} type equivalent to the @code{double}
13183 type. This is the default.
13186 @itemx -mno-backchain
13187 @opindex mbackchain
13188 @opindex mno-backchain
13189 Store (do not store) the address of the caller's frame as backchain pointer
13190 into the callee's stack frame.
13191 A backchain may be needed to allow debugging using tools that do not understand
13192 DWARF-2 call frame information.
13193 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
13194 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
13195 the backchain is placed into the topmost word of the 96/160 byte register
13198 In general, code compiled with @option{-mbackchain} is call-compatible with
13199 code compiled with @option{-mmo-backchain}; however, use of the backchain
13200 for debugging purposes usually requires that the whole binary is built with
13201 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
13202 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13203 to build a linux kernel use @option{-msoft-float}.
13205 The default is to not maintain the backchain.
13207 @item -mpacked-stack
13208 @item -mno-packed-stack
13209 @opindex mpacked-stack
13210 @opindex mno-packed-stack
13211 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
13212 specified, the compiler uses the all fields of the 96/160 byte register save
13213 area only for their default purpose; unused fields still take up stack space.
13214 When @option{-mpacked-stack} is specified, register save slots are densely
13215 packed at the top of the register save area; unused space is reused for other
13216 purposes, allowing for more efficient use of the available stack space.
13217 However, when @option{-mbackchain} is also in effect, the topmost word of
13218 the save area is always used to store the backchain, and the return address
13219 register is always saved two words below the backchain.
13221 As long as the stack frame backchain is not used, code generated with
13222 @option{-mpacked-stack} is call-compatible with code generated with
13223 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
13224 S/390 or zSeries generated code that uses the stack frame backchain at run
13225 time, not just for debugging purposes. Such code is not call-compatible
13226 with code compiled with @option{-mpacked-stack}. Also, note that the
13227 combination of @option{-mbackchain},
13228 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13229 to build a linux kernel use @option{-msoft-float}.
13231 The default is to not use the packed stack layout.
13234 @itemx -mno-small-exec
13235 @opindex msmall-exec
13236 @opindex mno-small-exec
13237 Generate (or do not generate) code using the @code{bras} instruction
13238 to do subroutine calls.
13239 This only works reliably if the total executable size does not
13240 exceed 64k. The default is to use the @code{basr} instruction instead,
13241 which does not have this limitation.
13247 When @option{-m31} is specified, generate code compliant to the
13248 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
13249 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
13250 particular to generate 64-bit instructions. For the @samp{s390}
13251 targets, the default is @option{-m31}, while the @samp{s390x}
13252 targets default to @option{-m64}.
13258 When @option{-mzarch} is specified, generate code using the
13259 instructions available on z/Architecture.
13260 When @option{-mesa} is specified, generate code using the
13261 instructions available on ESA/390. Note that @option{-mesa} is
13262 not possible with @option{-m64}.
13263 When generating code compliant to the GNU/Linux for S/390 ABI,
13264 the default is @option{-mesa}. When generating code compliant
13265 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
13271 Generate (or do not generate) code using the @code{mvcle} instruction
13272 to perform block moves. When @option{-mno-mvcle} is specified,
13273 use a @code{mvc} loop instead. This is the default unless optimizing for
13280 Print (or do not print) additional debug information when compiling.
13281 The default is to not print debug information.
13283 @item -march=@var{cpu-type}
13285 Generate code that will run on @var{cpu-type}, which is the name of a system
13286 representing a certain processor type. Possible values for
13287 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
13288 When generating code using the instructions available on z/Architecture,
13289 the default is @option{-march=z900}. Otherwise, the default is
13290 @option{-march=g5}.
13292 @item -mtune=@var{cpu-type}
13294 Tune to @var{cpu-type} everything applicable about the generated code,
13295 except for the ABI and the set of available instructions.
13296 The list of @var{cpu-type} values is the same as for @option{-march}.
13297 The default is the value used for @option{-march}.
13300 @itemx -mno-tpf-trace
13301 @opindex mtpf-trace
13302 @opindex mno-tpf-trace
13303 Generate code that adds (does not add) in TPF OS specific branches to trace
13304 routines in the operating system. This option is off by default, even
13305 when compiling for the TPF OS@.
13308 @itemx -mno-fused-madd
13309 @opindex mfused-madd
13310 @opindex mno-fused-madd
13311 Generate code that uses (does not use) the floating point multiply and
13312 accumulate instructions. These instructions are generated by default if
13313 hardware floating point is used.
13315 @item -mwarn-framesize=@var{framesize}
13316 @opindex mwarn-framesize
13317 Emit a warning if the current function exceeds the given frame size. Because
13318 this is a compile time check it doesn't need to be a real problem when the program
13319 runs. It is intended to identify functions which most probably cause
13320 a stack overflow. It is useful to be used in an environment with limited stack
13321 size e.g.@: the linux kernel.
13323 @item -mwarn-dynamicstack
13324 @opindex mwarn-dynamicstack
13325 Emit a warning if the function calls alloca or uses dynamically
13326 sized arrays. This is generally a bad idea with a limited stack size.
13328 @item -mstack-guard=@var{stack-guard}
13329 @item -mstack-size=@var{stack-size}
13330 @opindex mstack-guard
13331 @opindex mstack-size
13332 If these options are provided the s390 back end emits additional instructions in
13333 the function prologue which trigger a trap if the stack size is @var{stack-guard}
13334 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
13335 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
13336 the frame size of the compiled function is chosen.
13337 These options are intended to be used to help debugging stack overflow problems.
13338 The additionally emitted code causes only little overhead and hence can also be
13339 used in production like systems without greater performance degradation. The given
13340 values have to be exact powers of 2 and @var{stack-size} has to be greater than
13341 @var{stack-guard} without exceeding 64k.
13342 In order to be efficient the extra code makes the assumption that the stack starts
13343 at an address aligned to the value given by @var{stack-size}.
13344 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
13347 @node Score Options
13348 @subsection Score Options
13349 @cindex Score Options
13351 These options are defined for Score implementations:
13356 Compile code for big endian mode. This is the default.
13360 Compile code for little endian mode.
13364 Disable generate bcnz instruction.
13368 Enable generate unaligned load and store instruction.
13372 Enable the use of multiply-accumulate instructions. Disabled by default.
13376 Specify the SCORE5 as the target architecture.
13380 Specify the SCORE5U of the target architecture.
13384 Specify the SCORE7 as the target architecture. This is the default.
13388 Specify the SCORE7D as the target architecture.
13392 @subsection SH Options
13394 These @samp{-m} options are defined for the SH implementations:
13399 Generate code for the SH1.
13403 Generate code for the SH2.
13406 Generate code for the SH2e.
13410 Generate code for the SH3.
13414 Generate code for the SH3e.
13418 Generate code for the SH4 without a floating-point unit.
13420 @item -m4-single-only
13421 @opindex m4-single-only
13422 Generate code for the SH4 with a floating-point unit that only
13423 supports single-precision arithmetic.
13427 Generate code for the SH4 assuming the floating-point unit is in
13428 single-precision mode by default.
13432 Generate code for the SH4.
13436 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
13437 floating-point unit is not used.
13439 @item -m4a-single-only
13440 @opindex m4a-single-only
13441 Generate code for the SH4a, in such a way that no double-precision
13442 floating point operations are used.
13445 @opindex m4a-single
13446 Generate code for the SH4a assuming the floating-point unit is in
13447 single-precision mode by default.
13451 Generate code for the SH4a.
13455 Same as @option{-m4a-nofpu}, except that it implicitly passes
13456 @option{-dsp} to the assembler. GCC doesn't generate any DSP
13457 instructions at the moment.
13461 Compile code for the processor in big endian mode.
13465 Compile code for the processor in little endian mode.
13469 Align doubles at 64-bit boundaries. Note that this changes the calling
13470 conventions, and thus some functions from the standard C library will
13471 not work unless you recompile it first with @option{-mdalign}.
13475 Shorten some address references at link time, when possible; uses the
13476 linker option @option{-relax}.
13480 Use 32-bit offsets in @code{switch} tables. The default is to use
13485 Enable the use of the instruction @code{fmovd}.
13489 Comply with the calling conventions defined by Renesas.
13493 Comply with the calling conventions defined by Renesas.
13497 Comply with the calling conventions defined for GCC before the Renesas
13498 conventions were available. This option is the default for all
13499 targets of the SH toolchain except for @samp{sh-symbianelf}.
13502 @opindex mnomacsave
13503 Mark the @code{MAC} register as call-clobbered, even if
13504 @option{-mhitachi} is given.
13508 Increase IEEE-compliance of floating-point code.
13509 At the moment, this is equivalent to @option{-fno-finite-math-only}.
13510 When generating 16 bit SH opcodes, getting IEEE-conforming results for
13511 comparisons of NANs / infinities incurs extra overhead in every
13512 floating point comparison, therefore the default is set to
13513 @option{-ffinite-math-only}.
13515 @item -minline-ic_invalidate
13516 @opindex minline-ic_invalidate
13517 Inline code to invalidate instruction cache entries after setting up
13518 nested function trampolines.
13519 This option has no effect if -musermode is in effect and the selected
13520 code generation option (e.g. -m4) does not allow the use of the icbi
13522 If the selected code generation option does not allow the use of the icbi
13523 instruction, and -musermode is not in effect, the inlined code will
13524 manipulate the instruction cache address array directly with an associative
13525 write. This not only requires privileged mode, but it will also
13526 fail if the cache line had been mapped via the TLB and has become unmapped.
13530 Dump instruction size and location in the assembly code.
13533 @opindex mpadstruct
13534 This option is deprecated. It pads structures to multiple of 4 bytes,
13535 which is incompatible with the SH ABI@.
13539 Optimize for space instead of speed. Implied by @option{-Os}.
13542 @opindex mprefergot
13543 When generating position-independent code, emit function calls using
13544 the Global Offset Table instead of the Procedure Linkage Table.
13548 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
13549 if the inlined code would not work in user mode.
13550 This is the default when the target is @code{sh-*-linux*}.
13552 @item -multcost=@var{number}
13553 @opindex multcost=@var{number}
13554 Set the cost to assume for a multiply insn.
13556 @item -mdiv=@var{strategy}
13557 @opindex mdiv=@var{strategy}
13558 Set the division strategy to use for SHmedia code. @var{strategy} must be
13559 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
13560 inv:call2, inv:fp .
13561 "fp" performs the operation in floating point. This has a very high latency,
13562 but needs only a few instructions, so it might be a good choice if
13563 your code has enough easily exploitable ILP to allow the compiler to
13564 schedule the floating point instructions together with other instructions.
13565 Division by zero causes a floating point exception.
13566 "inv" uses integer operations to calculate the inverse of the divisor,
13567 and then multiplies the dividend with the inverse. This strategy allows
13568 cse and hoisting of the inverse calculation. Division by zero calculates
13569 an unspecified result, but does not trap.
13570 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
13571 have been found, or if the entire operation has been hoisted to the same
13572 place, the last stages of the inverse calculation are intertwined with the
13573 final multiply to reduce the overall latency, at the expense of using a few
13574 more instructions, and thus offering fewer scheduling opportunities with
13576 "call" calls a library function that usually implements the inv:minlat
13578 This gives high code density for m5-*media-nofpu compilations.
13579 "call2" uses a different entry point of the same library function, where it
13580 assumes that a pointer to a lookup table has already been set up, which
13581 exposes the pointer load to cse / code hoisting optimizations.
13582 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
13583 code generation, but if the code stays unoptimized, revert to the "call",
13584 "call2", or "fp" strategies, respectively. Note that the
13585 potentially-trapping side effect of division by zero is carried by a
13586 separate instruction, so it is possible that all the integer instructions
13587 are hoisted out, but the marker for the side effect stays where it is.
13588 A recombination to fp operations or a call is not possible in that case.
13589 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
13590 that the inverse calculation was nor separated from the multiply, they speed
13591 up division where the dividend fits into 20 bits (plus sign where applicable),
13592 by inserting a test to skip a number of operations in this case; this test
13593 slows down the case of larger dividends. inv20u assumes the case of a such
13594 a small dividend to be unlikely, and inv20l assumes it to be likely.
13596 @item -mdivsi3_libfunc=@var{name}
13597 @opindex mdivsi3_libfunc=@var{name}
13598 Set the name of the library function used for 32 bit signed division to
13599 @var{name}. This only affect the name used in the call and inv:call
13600 division strategies, and the compiler will still expect the same
13601 sets of input/output/clobbered registers as if this option was not present.
13603 @item -madjust-unroll
13604 @opindex madjust-unroll
13605 Throttle unrolling to avoid thrashing target registers.
13606 This option only has an effect if the gcc code base supports the
13607 TARGET_ADJUST_UNROLL_MAX target hook.
13609 @item -mindexed-addressing
13610 @opindex mindexed-addressing
13611 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
13612 This is only safe if the hardware and/or OS implement 32 bit wrap-around
13613 semantics for the indexed addressing mode. The architecture allows the
13614 implementation of processors with 64 bit MMU, which the OS could use to
13615 get 32 bit addressing, but since no current hardware implementation supports
13616 this or any other way to make the indexed addressing mode safe to use in
13617 the 32 bit ABI, the default is -mno-indexed-addressing.
13619 @item -mgettrcost=@var{number}
13620 @opindex mgettrcost=@var{number}
13621 Set the cost assumed for the gettr instruction to @var{number}.
13622 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
13626 Assume pt* instructions won't trap. This will generally generate better
13627 scheduled code, but is unsafe on current hardware. The current architecture
13628 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
13629 This has the unintentional effect of making it unsafe to schedule ptabs /
13630 ptrel before a branch, or hoist it out of a loop. For example,
13631 __do_global_ctors, a part of libgcc that runs constructors at program
13632 startup, calls functions in a list which is delimited by @minus{}1. With the
13633 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
13634 That means that all the constructors will be run a bit quicker, but when
13635 the loop comes to the end of the list, the program crashes because ptabs
13636 loads @minus{}1 into a target register. Since this option is unsafe for any
13637 hardware implementing the current architecture specification, the default
13638 is -mno-pt-fixed. Unless the user specifies a specific cost with
13639 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
13640 this deters register allocation using target registers for storing
13643 @item -minvalid-symbols
13644 @opindex minvalid-symbols
13645 Assume symbols might be invalid. Ordinary function symbols generated by
13646 the compiler will always be valid to load with movi/shori/ptabs or
13647 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
13648 to generate symbols that will cause ptabs / ptrel to trap.
13649 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
13650 It will then prevent cross-basic-block cse, hoisting and most scheduling
13651 of symbol loads. The default is @option{-mno-invalid-symbols}.
13654 @node SPARC Options
13655 @subsection SPARC Options
13656 @cindex SPARC options
13658 These @samp{-m} options are supported on the SPARC:
13661 @item -mno-app-regs
13663 @opindex mno-app-regs
13665 Specify @option{-mapp-regs} to generate output using the global registers
13666 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
13669 To be fully SVR4 ABI compliant at the cost of some performance loss,
13670 specify @option{-mno-app-regs}. You should compile libraries and system
13671 software with this option.
13674 @itemx -mhard-float
13676 @opindex mhard-float
13677 Generate output containing floating point instructions. This is the
13681 @itemx -msoft-float
13683 @opindex msoft-float
13684 Generate output containing library calls for floating point.
13685 @strong{Warning:} the requisite libraries are not available for all SPARC
13686 targets. Normally the facilities of the machine's usual C compiler are
13687 used, but this cannot be done directly in cross-compilation. You must make
13688 your own arrangements to provide suitable library functions for
13689 cross-compilation. The embedded targets @samp{sparc-*-aout} and
13690 @samp{sparclite-*-*} do provide software floating point support.
13692 @option{-msoft-float} changes the calling convention in the output file;
13693 therefore, it is only useful if you compile @emph{all} of a program with
13694 this option. In particular, you need to compile @file{libgcc.a}, the
13695 library that comes with GCC, with @option{-msoft-float} in order for
13698 @item -mhard-quad-float
13699 @opindex mhard-quad-float
13700 Generate output containing quad-word (long double) floating point
13703 @item -msoft-quad-float
13704 @opindex msoft-quad-float
13705 Generate output containing library calls for quad-word (long double)
13706 floating point instructions. The functions called are those specified
13707 in the SPARC ABI@. This is the default.
13709 As of this writing, there are no SPARC implementations that have hardware
13710 support for the quad-word floating point instructions. They all invoke
13711 a trap handler for one of these instructions, and then the trap handler
13712 emulates the effect of the instruction. Because of the trap handler overhead,
13713 this is much slower than calling the ABI library routines. Thus the
13714 @option{-msoft-quad-float} option is the default.
13716 @item -mno-unaligned-doubles
13717 @itemx -munaligned-doubles
13718 @opindex mno-unaligned-doubles
13719 @opindex munaligned-doubles
13720 Assume that doubles have 8 byte alignment. This is the default.
13722 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
13723 alignment only if they are contained in another type, or if they have an
13724 absolute address. Otherwise, it assumes they have 4 byte alignment.
13725 Specifying this option avoids some rare compatibility problems with code
13726 generated by other compilers. It is not the default because it results
13727 in a performance loss, especially for floating point code.
13729 @item -mno-faster-structs
13730 @itemx -mfaster-structs
13731 @opindex mno-faster-structs
13732 @opindex mfaster-structs
13733 With @option{-mfaster-structs}, the compiler assumes that structures
13734 should have 8 byte alignment. This enables the use of pairs of
13735 @code{ldd} and @code{std} instructions for copies in structure
13736 assignment, in place of twice as many @code{ld} and @code{st} pairs.
13737 However, the use of this changed alignment directly violates the SPARC
13738 ABI@. Thus, it's intended only for use on targets where the developer
13739 acknowledges that their resulting code will not be directly in line with
13740 the rules of the ABI@.
13742 @item -mimpure-text
13743 @opindex mimpure-text
13744 @option{-mimpure-text}, used in addition to @option{-shared}, tells
13745 the compiler to not pass @option{-z text} to the linker when linking a
13746 shared object. Using this option, you can link position-dependent
13747 code into a shared object.
13749 @option{-mimpure-text} suppresses the ``relocations remain against
13750 allocatable but non-writable sections'' linker error message.
13751 However, the necessary relocations will trigger copy-on-write, and the
13752 shared object is not actually shared across processes. Instead of
13753 using @option{-mimpure-text}, you should compile all source code with
13754 @option{-fpic} or @option{-fPIC}.
13756 This option is only available on SunOS and Solaris.
13758 @item -mcpu=@var{cpu_type}
13760 Set the instruction set, register set, and instruction scheduling parameters
13761 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
13762 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
13763 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
13764 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
13765 @samp{ultrasparc3}, and @samp{niagara}.
13767 Default instruction scheduling parameters are used for values that select
13768 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
13769 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
13771 Here is a list of each supported architecture and their supported
13776 v8: supersparc, hypersparc
13777 sparclite: f930, f934, sparclite86x
13779 v9: ultrasparc, ultrasparc3, niagara
13782 By default (unless configured otherwise), GCC generates code for the V7
13783 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
13784 additionally optimizes it for the Cypress CY7C602 chip, as used in the
13785 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
13786 SPARCStation 1, 2, IPX etc.
13788 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
13789 architecture. The only difference from V7 code is that the compiler emits
13790 the integer multiply and integer divide instructions which exist in SPARC-V8
13791 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
13792 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
13795 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
13796 the SPARC architecture. This adds the integer multiply, integer divide step
13797 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
13798 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
13799 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
13800 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
13801 MB86934 chip, which is the more recent SPARClite with FPU@.
13803 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
13804 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
13805 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
13806 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
13807 optimizes it for the TEMIC SPARClet chip.
13809 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
13810 architecture. This adds 64-bit integer and floating-point move instructions,
13811 3 additional floating-point condition code registers and conditional move
13812 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
13813 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
13814 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
13815 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
13816 @option{-mcpu=niagara}, the compiler additionally optimizes it for
13817 Sun UltraSPARC T1 chips.
13819 @item -mtune=@var{cpu_type}
13821 Set the instruction scheduling parameters for machine type
13822 @var{cpu_type}, but do not set the instruction set or register set that the
13823 option @option{-mcpu=@var{cpu_type}} would.
13825 The same values for @option{-mcpu=@var{cpu_type}} can be used for
13826 @option{-mtune=@var{cpu_type}}, but the only useful values are those
13827 that select a particular cpu implementation. Those are @samp{cypress},
13828 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
13829 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
13830 @samp{ultrasparc3}, and @samp{niagara}.
13835 @opindex mno-v8plus
13836 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
13837 difference from the V8 ABI is that the global and out registers are
13838 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
13839 mode for all SPARC-V9 processors.
13845 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
13846 Visual Instruction Set extensions. The default is @option{-mno-vis}.
13849 These @samp{-m} options are supported in addition to the above
13850 on SPARC-V9 processors in 64-bit environments:
13853 @item -mlittle-endian
13854 @opindex mlittle-endian
13855 Generate code for a processor running in little-endian mode. It is only
13856 available for a few configurations and most notably not on Solaris and Linux.
13862 Generate code for a 32-bit or 64-bit environment.
13863 The 32-bit environment sets int, long and pointer to 32 bits.
13864 The 64-bit environment sets int to 32 bits and long and pointer
13867 @item -mcmodel=medlow
13868 @opindex mcmodel=medlow
13869 Generate code for the Medium/Low code model: 64-bit addresses, programs
13870 must be linked in the low 32 bits of memory. Programs can be statically
13871 or dynamically linked.
13873 @item -mcmodel=medmid
13874 @opindex mcmodel=medmid
13875 Generate code for the Medium/Middle code model: 64-bit addresses, programs
13876 must be linked in the low 44 bits of memory, the text and data segments must
13877 be less than 2GB in size and the data segment must be located within 2GB of
13880 @item -mcmodel=medany
13881 @opindex mcmodel=medany
13882 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
13883 may be linked anywhere in memory, the text and data segments must be less
13884 than 2GB in size and the data segment must be located within 2GB of the
13887 @item -mcmodel=embmedany
13888 @opindex mcmodel=embmedany
13889 Generate code for the Medium/Anywhere code model for embedded systems:
13890 64-bit addresses, the text and data segments must be less than 2GB in
13891 size, both starting anywhere in memory (determined at link time). The
13892 global register %g4 points to the base of the data segment. Programs
13893 are statically linked and PIC is not supported.
13896 @itemx -mno-stack-bias
13897 @opindex mstack-bias
13898 @opindex mno-stack-bias
13899 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
13900 frame pointer if present, are offset by @minus{}2047 which must be added back
13901 when making stack frame references. This is the default in 64-bit mode.
13902 Otherwise, assume no such offset is present.
13905 These switches are supported in addition to the above on Solaris:
13910 Add support for multithreading using the Solaris threads library. This
13911 option sets flags for both the preprocessor and linker. This option does
13912 not affect the thread safety of object code produced by the compiler or
13913 that of libraries supplied with it.
13917 Add support for multithreading using the POSIX threads library. This
13918 option sets flags for both the preprocessor and linker. This option does
13919 not affect the thread safety of object code produced by the compiler or
13920 that of libraries supplied with it.
13924 This is a synonym for @option{-pthreads}.
13928 @subsection SPU Options
13929 @cindex SPU options
13931 These @samp{-m} options are supported on the SPU:
13935 @itemx -merror-reloc
13936 @opindex mwarn-reloc
13937 @opindex merror-reloc
13939 The loader for SPU does not handle dynamic relocations. By default, GCC
13940 will give an error when it generates code that requires a dynamic
13941 relocation. @option{-mno-error-reloc} disables the error,
13942 @option{-mwarn-reloc} will generate a warning instead.
13945 @itemx -munsafe-dma
13947 @opindex munsafe-dma
13949 Instructions which initiate or test completion of DMA must not be
13950 reordered with respect to loads and stores of the memory which is being
13951 accessed. Users typically address this problem using the volatile
13952 keyword, but that can lead to inefficient code in places where the
13953 memory is known to not change. Rather than mark the memory as volatile
13954 we treat the DMA instructions as potentially effecting all memory. With
13955 @option{-munsafe-dma} users must use the volatile keyword to protect
13958 @item -mbranch-hints
13959 @opindex mbranch-hints
13961 By default, GCC will generate a branch hint instruction to avoid
13962 pipeline stalls for always taken or probably taken branches. A hint
13963 will not be generated closer than 8 instructions away from its branch.
13964 There is little reason to disable them, except for debugging purposes,
13965 or to make an object a little bit smaller.
13969 @opindex msmall-mem
13970 @opindex mlarge-mem
13972 By default, GCC generates code assuming that addresses are never larger
13973 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
13974 a full 32 bit address.
13979 By default, GCC links against startup code that assumes the SPU-style
13980 main function interface (which has an unconventional parameter list).
13981 With @option{-mstdmain}, GCC will link your program against startup
13982 code that assumes a C99-style interface to @code{main}, including a
13983 local copy of @code{argv} strings.
13985 @item -mfixed-range=@var{register-range}
13986 @opindex mfixed-range
13987 Generate code treating the given register range as fixed registers.
13988 A fixed register is one that the register allocator can not use. This is
13989 useful when compiling kernel code. A register range is specified as
13990 two registers separated by a dash. Multiple register ranges can be
13991 specified separated by a comma.
13995 @node System V Options
13996 @subsection Options for System V
13998 These additional options are available on System V Release 4 for
13999 compatibility with other compilers on those systems:
14004 Create a shared object.
14005 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
14009 Identify the versions of each tool used by the compiler, in a
14010 @code{.ident} assembler directive in the output.
14014 Refrain from adding @code{.ident} directives to the output file (this is
14017 @item -YP,@var{dirs}
14019 Search the directories @var{dirs}, and no others, for libraries
14020 specified with @option{-l}.
14022 @item -Ym,@var{dir}
14024 Look in the directory @var{dir} to find the M4 preprocessor.
14025 The assembler uses this option.
14026 @c This is supposed to go with a -Yd for predefined M4 macro files, but
14027 @c the generic assembler that comes with Solaris takes just -Ym.
14030 @node TMS320C3x/C4x Options
14031 @subsection TMS320C3x/C4x Options
14032 @cindex TMS320C3x/C4x Options
14034 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
14038 @item -mcpu=@var{cpu_type}
14040 Set the instruction set, register set, and instruction scheduling
14041 parameters for machine type @var{cpu_type}. Supported values for
14042 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
14043 @samp{c44}. The default is @samp{c40} to generate code for the
14048 @itemx -msmall-memory
14050 @opindex mbig-memory
14052 @opindex msmall-memory
14054 Generates code for the big or small memory model. The small memory
14055 model assumed that all data fits into one 64K word page. At run-time
14056 the data page (DP) register must be set to point to the 64K page
14057 containing the .bss and .data program sections. The big memory model is
14058 the default and requires reloading of the DP register for every direct
14065 Allow (disallow) allocation of general integer operands into the block
14066 count register BK@.
14072 Enable (disable) generation of code using decrement and branch,
14073 DBcond(D), instructions. This is enabled by default for the C4x. To be
14074 on the safe side, this is disabled for the C3x, since the maximum
14075 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
14076 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
14077 that it can utilize the decrement and branch instruction, but will give
14078 up if there is more than one memory reference in the loop. Thus a loop
14079 where the loop counter is decremented can generate slightly more
14080 efficient code, in cases where the RPTB instruction cannot be utilized.
14082 @item -mdp-isr-reload
14084 @opindex mdp-isr-reload
14086 Force the DP register to be saved on entry to an interrupt service
14087 routine (ISR), reloaded to point to the data section, and restored on
14088 exit from the ISR@. This should not be required unless someone has
14089 violated the small memory model by modifying the DP register, say within
14096 For the C3x use the 24-bit MPYI instruction for integer multiplies
14097 instead of a library call to guarantee 32-bit results. Note that if one
14098 of the operands is a constant, then the multiplication will be performed
14099 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
14100 then squaring operations are performed inline instead of a library call.
14103 @itemx -mno-fast-fix
14105 @opindex mno-fast-fix
14106 The C3x/C4x FIX instruction to convert a floating point value to an
14107 integer value chooses the nearest integer less than or equal to the
14108 floating point value rather than to the nearest integer. Thus if the
14109 floating point number is negative, the result will be incorrectly
14110 truncated an additional code is necessary to detect and correct this
14111 case. This option can be used to disable generation of the additional
14112 code required to correct the result.
14118 Enable (disable) generation of repeat block sequences using the RPTB
14119 instruction for zero overhead looping. The RPTB construct is only used
14120 for innermost loops that do not call functions or jump across the loop
14121 boundaries. There is no advantage having nested RPTB loops due to the
14122 overhead required to save and restore the RC, RS, and RE registers.
14123 This is enabled by default with @option{-O2}.
14125 @item -mrpts=@var{count}
14129 Enable (disable) the use of the single instruction repeat instruction
14130 RPTS@. If a repeat block contains a single instruction, and the loop
14131 count can be guaranteed to be less than the value @var{count}, GCC will
14132 emit a RPTS instruction instead of a RPTB@. If no value is specified,
14133 then a RPTS will be emitted even if the loop count cannot be determined
14134 at compile time. Note that the repeated instruction following RPTS does
14135 not have to be reloaded from memory each iteration, thus freeing up the
14136 CPU buses for operands. However, since interrupts are blocked by this
14137 instruction, it is disabled by default.
14139 @item -mloop-unsigned
14140 @itemx -mno-loop-unsigned
14141 @opindex mloop-unsigned
14142 @opindex mno-loop-unsigned
14143 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
14144 is @math{2^{31} + 1} since these instructions test if the iteration count is
14145 negative to terminate the loop. If the iteration count is unsigned
14146 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
14147 exceeded. This switch allows an unsigned iteration count.
14151 Try to emit an assembler syntax that the TI assembler (asm30) is happy
14152 with. This also enforces compatibility with the API employed by the TI
14153 C3x C compiler. For example, long doubles are passed as structures
14154 rather than in floating point registers.
14160 Generate code that uses registers (stack) for passing arguments to functions.
14161 By default, arguments are passed in registers where possible rather
14162 than by pushing arguments on to the stack.
14164 @item -mparallel-insns
14165 @itemx -mno-parallel-insns
14166 @opindex mparallel-insns
14167 @opindex mno-parallel-insns
14168 Allow the generation of parallel instructions. This is enabled by
14169 default with @option{-O2}.
14171 @item -mparallel-mpy
14172 @itemx -mno-parallel-mpy
14173 @opindex mparallel-mpy
14174 @opindex mno-parallel-mpy
14175 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
14176 provided @option{-mparallel-insns} is also specified. These instructions have
14177 tight register constraints which can pessimize the code generation
14178 of large functions.
14183 @subsection V850 Options
14184 @cindex V850 Options
14186 These @samp{-m} options are defined for V850 implementations:
14190 @itemx -mno-long-calls
14191 @opindex mlong-calls
14192 @opindex mno-long-calls
14193 Treat all calls as being far away (near). If calls are assumed to be
14194 far away, the compiler will always load the functions address up into a
14195 register, and call indirect through the pointer.
14201 Do not optimize (do optimize) basic blocks that use the same index
14202 pointer 4 or more times to copy pointer into the @code{ep} register, and
14203 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
14204 option is on by default if you optimize.
14206 @item -mno-prolog-function
14207 @itemx -mprolog-function
14208 @opindex mno-prolog-function
14209 @opindex mprolog-function
14210 Do not use (do use) external functions to save and restore registers
14211 at the prologue and epilogue of a function. The external functions
14212 are slower, but use less code space if more than one function saves
14213 the same number of registers. The @option{-mprolog-function} option
14214 is on by default if you optimize.
14218 Try to make the code as small as possible. At present, this just turns
14219 on the @option{-mep} and @option{-mprolog-function} options.
14221 @item -mtda=@var{n}
14223 Put static or global variables whose size is @var{n} bytes or less into
14224 the tiny data area that register @code{ep} points to. The tiny data
14225 area can hold up to 256 bytes in total (128 bytes for byte references).
14227 @item -msda=@var{n}
14229 Put static or global variables whose size is @var{n} bytes or less into
14230 the small data area that register @code{gp} points to. The small data
14231 area can hold up to 64 kilobytes.
14233 @item -mzda=@var{n}
14235 Put static or global variables whose size is @var{n} bytes or less into
14236 the first 32 kilobytes of memory.
14240 Specify that the target processor is the V850.
14243 @opindex mbig-switch
14244 Generate code suitable for big switch tables. Use this option only if
14245 the assembler/linker complain about out of range branches within a switch
14250 This option will cause r2 and r5 to be used in the code generated by
14251 the compiler. This setting is the default.
14253 @item -mno-app-regs
14254 @opindex mno-app-regs
14255 This option will cause r2 and r5 to be treated as fixed registers.
14259 Specify that the target processor is the V850E1. The preprocessor
14260 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
14261 this option is used.
14265 Specify that the target processor is the V850E@. The preprocessor
14266 constant @samp{__v850e__} will be defined if this option is used.
14268 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
14269 are defined then a default target processor will be chosen and the
14270 relevant @samp{__v850*__} preprocessor constant will be defined.
14272 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
14273 defined, regardless of which processor variant is the target.
14275 @item -mdisable-callt
14276 @opindex mdisable-callt
14277 This option will suppress generation of the CALLT instruction for the
14278 v850e and v850e1 flavors of the v850 architecture. The default is
14279 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
14284 @subsection VAX Options
14285 @cindex VAX options
14287 These @samp{-m} options are defined for the VAX:
14292 Do not output certain jump instructions (@code{aobleq} and so on)
14293 that the Unix assembler for the VAX cannot handle across long
14298 Do output those jump instructions, on the assumption that you
14299 will assemble with the GNU assembler.
14303 Output code for g-format floating point numbers instead of d-format.
14306 @node VxWorks Options
14307 @subsection VxWorks Options
14308 @cindex VxWorks Options
14310 The options in this section are defined for all VxWorks targets.
14311 Options specific to the target hardware are listed with the other
14312 options for that target.
14317 GCC can generate code for both VxWorks kernels and real time processes
14318 (RTPs). This option switches from the former to the latter. It also
14319 defines the preprocessor macro @code{__RTP__}.
14322 @opindex non-static
14323 Link an RTP executable against shared libraries rather than static
14324 libraries. The options @option{-static} and @option{-shared} can
14325 also be used for RTPs (@pxref{Link Options}); @option{-static}
14332 These options are passed down to the linker. They are defined for
14333 compatibility with Diab.
14336 @opindex Xbind-lazy
14337 Enable lazy binding of function calls. This option is equivalent to
14338 @option{-Wl,-z,now} and is defined for compatibility with Diab.
14342 Disable lazy binding of function calls. This option is the default and
14343 is defined for compatibility with Diab.
14346 @node x86-64 Options
14347 @subsection x86-64 Options
14348 @cindex x86-64 options
14350 These are listed under @xref{i386 and x86-64 Options}.
14352 @node Xstormy16 Options
14353 @subsection Xstormy16 Options
14354 @cindex Xstormy16 Options
14356 These options are defined for Xstormy16:
14361 Choose startup files and linker script suitable for the simulator.
14364 @node Xtensa Options
14365 @subsection Xtensa Options
14366 @cindex Xtensa Options
14368 These options are supported for Xtensa targets:
14372 @itemx -mno-const16
14374 @opindex mno-const16
14375 Enable or disable use of @code{CONST16} instructions for loading
14376 constant values. The @code{CONST16} instruction is currently not a
14377 standard option from Tensilica. When enabled, @code{CONST16}
14378 instructions are always used in place of the standard @code{L32R}
14379 instructions. The use of @code{CONST16} is enabled by default only if
14380 the @code{L32R} instruction is not available.
14383 @itemx -mno-fused-madd
14384 @opindex mfused-madd
14385 @opindex mno-fused-madd
14386 Enable or disable use of fused multiply/add and multiply/subtract
14387 instructions in the floating-point option. This has no effect if the
14388 floating-point option is not also enabled. Disabling fused multiply/add
14389 and multiply/subtract instructions forces the compiler to use separate
14390 instructions for the multiply and add/subtract operations. This may be
14391 desirable in some cases where strict IEEE 754-compliant results are
14392 required: the fused multiply add/subtract instructions do not round the
14393 intermediate result, thereby producing results with @emph{more} bits of
14394 precision than specified by the IEEE standard. Disabling fused multiply
14395 add/subtract instructions also ensures that the program output is not
14396 sensitive to the compiler's ability to combine multiply and add/subtract
14399 @item -mtext-section-literals
14400 @itemx -mno-text-section-literals
14401 @opindex mtext-section-literals
14402 @opindex mno-text-section-literals
14403 Control the treatment of literal pools. The default is
14404 @option{-mno-text-section-literals}, which places literals in a separate
14405 section in the output file. This allows the literal pool to be placed
14406 in a data RAM/ROM, and it also allows the linker to combine literal
14407 pools from separate object files to remove redundant literals and
14408 improve code size. With @option{-mtext-section-literals}, the literals
14409 are interspersed in the text section in order to keep them as close as
14410 possible to their references. This may be necessary for large assembly
14413 @item -mtarget-align
14414 @itemx -mno-target-align
14415 @opindex mtarget-align
14416 @opindex mno-target-align
14417 When this option is enabled, GCC instructs the assembler to
14418 automatically align instructions to reduce branch penalties at the
14419 expense of some code density. The assembler attempts to widen density
14420 instructions to align branch targets and the instructions following call
14421 instructions. If there are not enough preceding safe density
14422 instructions to align a target, no widening will be performed. The
14423 default is @option{-mtarget-align}. These options do not affect the
14424 treatment of auto-aligned instructions like @code{LOOP}, which the
14425 assembler will always align, either by widening density instructions or
14426 by inserting no-op instructions.
14429 @itemx -mno-longcalls
14430 @opindex mlongcalls
14431 @opindex mno-longcalls
14432 When this option is enabled, GCC instructs the assembler to translate
14433 direct calls to indirect calls unless it can determine that the target
14434 of a direct call is in the range allowed by the call instruction. This
14435 translation typically occurs for calls to functions in other source
14436 files. Specifically, the assembler translates a direct @code{CALL}
14437 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
14438 The default is @option{-mno-longcalls}. This option should be used in
14439 programs where the call target can potentially be out of range. This
14440 option is implemented in the assembler, not the compiler, so the
14441 assembly code generated by GCC will still show direct call
14442 instructions---look at the disassembled object code to see the actual
14443 instructions. Note that the assembler will use an indirect call for
14444 every cross-file call, not just those that really will be out of range.
14447 @node zSeries Options
14448 @subsection zSeries Options
14449 @cindex zSeries options
14451 These are listed under @xref{S/390 and zSeries Options}.
14453 @node Code Gen Options
14454 @section Options for Code Generation Conventions
14455 @cindex code generation conventions
14456 @cindex options, code generation
14457 @cindex run-time options
14459 These machine-independent options control the interface conventions
14460 used in code generation.
14462 Most of them have both positive and negative forms; the negative form
14463 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
14464 one of the forms is listed---the one which is not the default. You
14465 can figure out the other form by either removing @samp{no-} or adding
14469 @item -fbounds-check
14470 @opindex fbounds-check
14471 For front-ends that support it, generate additional code to check that
14472 indices used to access arrays are within the declared range. This is
14473 currently only supported by the Java and Fortran front-ends, where
14474 this option defaults to true and false respectively.
14478 This option generates traps for signed overflow on addition, subtraction,
14479 multiplication operations.
14483 This option instructs the compiler to assume that signed arithmetic
14484 overflow of addition, subtraction and multiplication wraps around
14485 using twos-complement representation. This flag enables some optimizations
14486 and disables others. This option is enabled by default for the Java
14487 front-end, as required by the Java language specification.
14490 @opindex fexceptions
14491 Enable exception handling. Generates extra code needed to propagate
14492 exceptions. For some targets, this implies GCC will generate frame
14493 unwind information for all functions, which can produce significant data
14494 size overhead, although it does not affect execution. If you do not
14495 specify this option, GCC will enable it by default for languages like
14496 C++ which normally require exception handling, and disable it for
14497 languages like C that do not normally require it. However, you may need
14498 to enable this option when compiling C code that needs to interoperate
14499 properly with exception handlers written in C++. You may also wish to
14500 disable this option if you are compiling older C++ programs that don't
14501 use exception handling.
14503 @item -fnon-call-exceptions
14504 @opindex fnon-call-exceptions
14505 Generate code that allows trapping instructions to throw exceptions.
14506 Note that this requires platform-specific runtime support that does
14507 not exist everywhere. Moreover, it only allows @emph{trapping}
14508 instructions to throw exceptions, i.e.@: memory references or floating
14509 point instructions. It does not allow exceptions to be thrown from
14510 arbitrary signal handlers such as @code{SIGALRM}.
14512 @item -funwind-tables
14513 @opindex funwind-tables
14514 Similar to @option{-fexceptions}, except that it will just generate any needed
14515 static data, but will not affect the generated code in any other way.
14516 You will normally not enable this option; instead, a language processor
14517 that needs this handling would enable it on your behalf.
14519 @item -fasynchronous-unwind-tables
14520 @opindex fasynchronous-unwind-tables
14521 Generate unwind table in dwarf2 format, if supported by target machine. The
14522 table is exact at each instruction boundary, so it can be used for stack
14523 unwinding from asynchronous events (such as debugger or garbage collector).
14525 @item -fpcc-struct-return
14526 @opindex fpcc-struct-return
14527 Return ``short'' @code{struct} and @code{union} values in memory like
14528 longer ones, rather than in registers. This convention is less
14529 efficient, but it has the advantage of allowing intercallability between
14530 GCC-compiled files and files compiled with other compilers, particularly
14531 the Portable C Compiler (pcc).
14533 The precise convention for returning structures in memory depends
14534 on the target configuration macros.
14536 Short structures and unions are those whose size and alignment match
14537 that of some integer type.
14539 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
14540 switch is not binary compatible with code compiled with the
14541 @option{-freg-struct-return} switch.
14542 Use it to conform to a non-default application binary interface.
14544 @item -freg-struct-return
14545 @opindex freg-struct-return
14546 Return @code{struct} and @code{union} values in registers when possible.
14547 This is more efficient for small structures than
14548 @option{-fpcc-struct-return}.
14550 If you specify neither @option{-fpcc-struct-return} nor
14551 @option{-freg-struct-return}, GCC defaults to whichever convention is
14552 standard for the target. If there is no standard convention, GCC
14553 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
14554 the principal compiler. In those cases, we can choose the standard, and
14555 we chose the more efficient register return alternative.
14557 @strong{Warning:} code compiled with the @option{-freg-struct-return}
14558 switch is not binary compatible with code compiled with the
14559 @option{-fpcc-struct-return} switch.
14560 Use it to conform to a non-default application binary interface.
14562 @item -fshort-enums
14563 @opindex fshort-enums
14564 Allocate to an @code{enum} type only as many bytes as it needs for the
14565 declared range of possible values. Specifically, the @code{enum} type
14566 will be equivalent to the smallest integer type which has enough room.
14568 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
14569 code that is not binary compatible with code generated without that switch.
14570 Use it to conform to a non-default application binary interface.
14572 @item -fshort-double
14573 @opindex fshort-double
14574 Use the same size for @code{double} as for @code{float}.
14576 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
14577 code that is not binary compatible with code generated without that switch.
14578 Use it to conform to a non-default application binary interface.
14580 @item -fshort-wchar
14581 @opindex fshort-wchar
14582 Override the underlying type for @samp{wchar_t} to be @samp{short
14583 unsigned int} instead of the default for the target. This option is
14584 useful for building programs to run under WINE@.
14586 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
14587 code that is not binary compatible with code generated without that switch.
14588 Use it to conform to a non-default application binary interface.
14591 @opindex fno-common
14592 In C, allocate even uninitialized global variables in the data section of the
14593 object file, rather than generating them as common blocks. This has the
14594 effect that if the same variable is declared (without @code{extern}) in
14595 two different compilations, you will get an error when you link them.
14596 The only reason this might be useful is if you wish to verify that the
14597 program will work on other systems which always work this way.
14601 Ignore the @samp{#ident} directive.
14603 @item -finhibit-size-directive
14604 @opindex finhibit-size-directive
14605 Don't output a @code{.size} assembler directive, or anything else that
14606 would cause trouble if the function is split in the middle, and the
14607 two halves are placed at locations far apart in memory. This option is
14608 used when compiling @file{crtstuff.c}; you should not need to use it
14611 @item -fverbose-asm
14612 @opindex fverbose-asm
14613 Put extra commentary information in the generated assembly code to
14614 make it more readable. This option is generally only of use to those
14615 who actually need to read the generated assembly code (perhaps while
14616 debugging the compiler itself).
14618 @option{-fno-verbose-asm}, the default, causes the
14619 extra information to be omitted and is useful when comparing two assembler
14622 @item -frecord-gcc-switches
14623 @opindex frecord-gcc-switches
14624 This switch causes the command line that was used to invoke the
14625 compiler to be recorded into the object file that is being created.
14626 This switch is only implemented on some targets and the exact format
14627 of the recording is target and binary file format dependent, but it
14628 usually takes the form of a section containing ASCII text. This
14629 switch is related to the @option{-fverbose-asm} switch, but that
14630 switch only records information in the assembler output file as
14631 comments, so it never reaches the object file.
14635 @cindex global offset table
14637 Generate position-independent code (PIC) suitable for use in a shared
14638 library, if supported for the target machine. Such code accesses all
14639 constant addresses through a global offset table (GOT)@. The dynamic
14640 loader resolves the GOT entries when the program starts (the dynamic
14641 loader is not part of GCC; it is part of the operating system). If
14642 the GOT size for the linked executable exceeds a machine-specific
14643 maximum size, you get an error message from the linker indicating that
14644 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
14645 instead. (These maximums are 8k on the SPARC and 32k
14646 on the m68k and RS/6000. The 386 has no such limit.)
14648 Position-independent code requires special support, and therefore works
14649 only on certain machines. For the 386, GCC supports PIC for System V
14650 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
14651 position-independent.
14653 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14658 If supported for the target machine, emit position-independent code,
14659 suitable for dynamic linking and avoiding any limit on the size of the
14660 global offset table. This option makes a difference on the m68k,
14661 PowerPC and SPARC@.
14663 Position-independent code requires special support, and therefore works
14664 only on certain machines.
14666 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14673 These options are similar to @option{-fpic} and @option{-fPIC}, but
14674 generated position independent code can be only linked into executables.
14675 Usually these options are used when @option{-pie} GCC option will be
14676 used during linking.
14678 @option{-fpie} and @option{-fPIE} both define the macros
14679 @code{__pie__} and @code{__PIE__}. The macros have the value 1
14680 for @option{-fpie} and 2 for @option{-fPIE}.
14682 @item -fno-jump-tables
14683 @opindex fno-jump-tables
14684 Do not use jump tables for switch statements even where it would be
14685 more efficient than other code generation strategies. This option is
14686 of use in conjunction with @option{-fpic} or @option{-fPIC} for
14687 building code which forms part of a dynamic linker and cannot
14688 reference the address of a jump table. On some targets, jump tables
14689 do not require a GOT and this option is not needed.
14691 @item -ffixed-@var{reg}
14693 Treat the register named @var{reg} as a fixed register; generated code
14694 should never refer to it (except perhaps as a stack pointer, frame
14695 pointer or in some other fixed role).
14697 @var{reg} must be the name of a register. The register names accepted
14698 are machine-specific and are defined in the @code{REGISTER_NAMES}
14699 macro in the machine description macro file.
14701 This flag does not have a negative form, because it specifies a
14704 @item -fcall-used-@var{reg}
14705 @opindex fcall-used
14706 Treat the register named @var{reg} as an allocable register that is
14707 clobbered by function calls. It may be allocated for temporaries or
14708 variables that do not live across a call. Functions compiled this way
14709 will not save and restore the register @var{reg}.
14711 It is an error to used this flag with the frame pointer or stack pointer.
14712 Use of this flag for other registers that have fixed pervasive roles in
14713 the machine's execution model will produce disastrous results.
14715 This flag does not have a negative form, because it specifies a
14718 @item -fcall-saved-@var{reg}
14719 @opindex fcall-saved
14720 Treat the register named @var{reg} as an allocable register saved by
14721 functions. It may be allocated even for temporaries or variables that
14722 live across a call. Functions compiled this way will save and restore
14723 the register @var{reg} if they use it.
14725 It is an error to used this flag with the frame pointer or stack pointer.
14726 Use of this flag for other registers that have fixed pervasive roles in
14727 the machine's execution model will produce disastrous results.
14729 A different sort of disaster will result from the use of this flag for
14730 a register in which function values may be returned.
14732 This flag does not have a negative form, because it specifies a
14735 @item -fpack-struct[=@var{n}]
14736 @opindex fpack-struct
14737 Without a value specified, pack all structure members together without
14738 holes. When a value is specified (which must be a small power of two), pack
14739 structure members according to this value, representing the maximum
14740 alignment (that is, objects with default alignment requirements larger than
14741 this will be output potentially unaligned at the next fitting location.
14743 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
14744 code that is not binary compatible with code generated without that switch.
14745 Additionally, it makes the code suboptimal.
14746 Use it to conform to a non-default application binary interface.
14748 @item -finstrument-functions
14749 @opindex finstrument-functions
14750 Generate instrumentation calls for entry and exit to functions. Just
14751 after function entry and just before function exit, the following
14752 profiling functions will be called with the address of the current
14753 function and its call site. (On some platforms,
14754 @code{__builtin_return_address} does not work beyond the current
14755 function, so the call site information may not be available to the
14756 profiling functions otherwise.)
14759 void __cyg_profile_func_enter (void *this_fn,
14761 void __cyg_profile_func_exit (void *this_fn,
14765 The first argument is the address of the start of the current function,
14766 which may be looked up exactly in the symbol table.
14768 This instrumentation is also done for functions expanded inline in other
14769 functions. The profiling calls will indicate where, conceptually, the
14770 inline function is entered and exited. This means that addressable
14771 versions of such functions must be available. If all your uses of a
14772 function are expanded inline, this may mean an additional expansion of
14773 code size. If you use @samp{extern inline} in your C code, an
14774 addressable version of such functions must be provided. (This is
14775 normally the case anyways, but if you get lucky and the optimizer always
14776 expands the functions inline, you might have gotten away without
14777 providing static copies.)
14779 A function may be given the attribute @code{no_instrument_function}, in
14780 which case this instrumentation will not be done. This can be used, for
14781 example, for the profiling functions listed above, high-priority
14782 interrupt routines, and any functions from which the profiling functions
14783 cannot safely be called (perhaps signal handlers, if the profiling
14784 routines generate output or allocate memory).
14786 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
14787 @opindex finstrument-functions-exclude-file-list
14789 Set the list of functions that are excluded from instrumentation (see
14790 the description of @code{-finstrument-functions}). If the file that
14791 contains a function definition matches with one of @var{file}, then
14792 that function is not instrumented. The match is done on substrings:
14793 if the @var{file} parameter is a substring of the file name, it is
14794 considered to be a match.
14797 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
14798 will exclude any inline function defined in files whose pathnames
14799 contain @code{/bits/stl} or @code{include/sys}.
14801 If, for some reason, you want to include letter @code{','} in one of
14802 @var{sym}, write @code{'\,'}. For example,
14803 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
14804 (note the single quote surrounding the option).
14806 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
14807 @opindex finstrument-functions-exclude-function-list
14809 This is similar to @code{-finstrument-functions-exclude-file-list},
14810 but this option sets the list of function names to be excluded from
14811 instrumentation. The function name to be matched is its user-visible
14812 name, such as @code{vector<int> blah(const vector<int> &)}, not the
14813 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
14814 match is done on substrings: if the @var{sym} parameter is a substring
14815 of the function name, it is considered to be a match.
14817 @item -fstack-check
14818 @opindex fstack-check
14819 Generate code to verify that you do not go beyond the boundary of the
14820 stack. You should specify this flag if you are running in an
14821 environment with multiple threads, but only rarely need to specify it in
14822 a single-threaded environment since stack overflow is automatically
14823 detected on nearly all systems if there is only one stack.
14825 Note that this switch does not actually cause checking to be done; the
14826 operating system must do that. The switch causes generation of code
14827 to ensure that the operating system sees the stack being extended.
14829 @item -fstack-limit-register=@var{reg}
14830 @itemx -fstack-limit-symbol=@var{sym}
14831 @itemx -fno-stack-limit
14832 @opindex fstack-limit-register
14833 @opindex fstack-limit-symbol
14834 @opindex fno-stack-limit
14835 Generate code to ensure that the stack does not grow beyond a certain value,
14836 either the value of a register or the address of a symbol. If the stack
14837 would grow beyond the value, a signal is raised. For most targets,
14838 the signal is raised before the stack overruns the boundary, so
14839 it is possible to catch the signal without taking special precautions.
14841 For instance, if the stack starts at absolute address @samp{0x80000000}
14842 and grows downwards, you can use the flags
14843 @option{-fstack-limit-symbol=__stack_limit} and
14844 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
14845 of 128KB@. Note that this may only work with the GNU linker.
14847 @cindex aliasing of parameters
14848 @cindex parameters, aliased
14849 @item -fargument-alias
14850 @itemx -fargument-noalias
14851 @itemx -fargument-noalias-global
14852 @itemx -fargument-noalias-anything
14853 @opindex fargument-alias
14854 @opindex fargument-noalias
14855 @opindex fargument-noalias-global
14856 @opindex fargument-noalias-anything
14857 Specify the possible relationships among parameters and between
14858 parameters and global data.
14860 @option{-fargument-alias} specifies that arguments (parameters) may
14861 alias each other and may alias global storage.@*
14862 @option{-fargument-noalias} specifies that arguments do not alias
14863 each other, but may alias global storage.@*
14864 @option{-fargument-noalias-global} specifies that arguments do not
14865 alias each other and do not alias global storage.
14866 @option{-fargument-noalias-anything} specifies that arguments do not
14867 alias any other storage.
14869 Each language will automatically use whatever option is required by
14870 the language standard. You should not need to use these options yourself.
14872 @item -fleading-underscore
14873 @opindex fleading-underscore
14874 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
14875 change the way C symbols are represented in the object file. One use
14876 is to help link with legacy assembly code.
14878 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
14879 generate code that is not binary compatible with code generated without that
14880 switch. Use it to conform to a non-default application binary interface.
14881 Not all targets provide complete support for this switch.
14883 @item -ftls-model=@var{model}
14884 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
14885 The @var{model} argument should be one of @code{global-dynamic},
14886 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
14888 The default without @option{-fpic} is @code{initial-exec}; with
14889 @option{-fpic} the default is @code{global-dynamic}.
14891 @item -fvisibility=@var{default|internal|hidden|protected}
14892 @opindex fvisibility
14893 Set the default ELF image symbol visibility to the specified option---all
14894 symbols will be marked with this unless overridden within the code.
14895 Using this feature can very substantially improve linking and
14896 load times of shared object libraries, produce more optimized
14897 code, provide near-perfect API export and prevent symbol clashes.
14898 It is @strong{strongly} recommended that you use this in any shared objects
14901 Despite the nomenclature, @code{default} always means public ie;
14902 available to be linked against from outside the shared object.
14903 @code{protected} and @code{internal} are pretty useless in real-world
14904 usage so the only other commonly used option will be @code{hidden}.
14905 The default if @option{-fvisibility} isn't specified is
14906 @code{default}, i.e., make every
14907 symbol public---this causes the same behavior as previous versions of
14910 A good explanation of the benefits offered by ensuring ELF
14911 symbols have the correct visibility is given by ``How To Write
14912 Shared Libraries'' by Ulrich Drepper (which can be found at
14913 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
14914 solution made possible by this option to marking things hidden when
14915 the default is public is to make the default hidden and mark things
14916 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
14917 and @code{__attribute__ ((visibility("default")))} instead of
14918 @code{__declspec(dllexport)} you get almost identical semantics with
14919 identical syntax. This is a great boon to those working with
14920 cross-platform projects.
14922 For those adding visibility support to existing code, you may find
14923 @samp{#pragma GCC visibility} of use. This works by you enclosing
14924 the declarations you wish to set visibility for with (for example)
14925 @samp{#pragma GCC visibility push(hidden)} and
14926 @samp{#pragma GCC visibility pop}.
14927 Bear in mind that symbol visibility should be viewed @strong{as
14928 part of the API interface contract} and thus all new code should
14929 always specify visibility when it is not the default ie; declarations
14930 only for use within the local DSO should @strong{always} be marked explicitly
14931 as hidden as so to avoid PLT indirection overheads---making this
14932 abundantly clear also aids readability and self-documentation of the code.
14933 Note that due to ISO C++ specification requirements, operator new and
14934 operator delete must always be of default visibility.
14936 Be aware that headers from outside your project, in particular system
14937 headers and headers from any other library you use, may not be
14938 expecting to be compiled with visibility other than the default. You
14939 may need to explicitly say @samp{#pragma GCC visibility push(default)}
14940 before including any such headers.
14942 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
14943 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
14944 no modifications. However, this means that calls to @samp{extern}
14945 functions with no explicit visibility will use the PLT, so it is more
14946 effective to use @samp{__attribute ((visibility))} and/or
14947 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
14948 declarations should be treated as hidden.
14950 Note that @samp{-fvisibility} does affect C++ vague linkage
14951 entities. This means that, for instance, an exception class that will
14952 be thrown between DSOs must be explicitly marked with default
14953 visibility so that the @samp{type_info} nodes will be unified between
14956 An overview of these techniques, their benefits and how to use them
14957 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
14963 @node Environment Variables
14964 @section Environment Variables Affecting GCC
14965 @cindex environment variables
14967 @c man begin ENVIRONMENT
14968 This section describes several environment variables that affect how GCC
14969 operates. Some of them work by specifying directories or prefixes to use
14970 when searching for various kinds of files. Some are used to specify other
14971 aspects of the compilation environment.
14973 Note that you can also specify places to search using options such as
14974 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
14975 take precedence over places specified using environment variables, which
14976 in turn take precedence over those specified by the configuration of GCC@.
14977 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
14978 GNU Compiler Collection (GCC) Internals}.
14983 @c @itemx LC_COLLATE
14985 @c @itemx LC_MONETARY
14986 @c @itemx LC_NUMERIC
14991 @c @findex LC_COLLATE
14992 @findex LC_MESSAGES
14993 @c @findex LC_MONETARY
14994 @c @findex LC_NUMERIC
14998 These environment variables control the way that GCC uses
14999 localization information that allow GCC to work with different
15000 national conventions. GCC inspects the locale categories
15001 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
15002 so. These locale categories can be set to any value supported by your
15003 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
15004 Kingdom encoded in UTF-8.
15006 The @env{LC_CTYPE} environment variable specifies character
15007 classification. GCC uses it to determine the character boundaries in
15008 a string; this is needed for some multibyte encodings that contain quote
15009 and escape characters that would otherwise be interpreted as a string
15012 The @env{LC_MESSAGES} environment variable specifies the language to
15013 use in diagnostic messages.
15015 If the @env{LC_ALL} environment variable is set, it overrides the value
15016 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
15017 and @env{LC_MESSAGES} default to the value of the @env{LANG}
15018 environment variable. If none of these variables are set, GCC
15019 defaults to traditional C English behavior.
15023 If @env{TMPDIR} is set, it specifies the directory to use for temporary
15024 files. GCC uses temporary files to hold the output of one stage of
15025 compilation which is to be used as input to the next stage: for example,
15026 the output of the preprocessor, which is the input to the compiler
15029 @item GCC_EXEC_PREFIX
15030 @findex GCC_EXEC_PREFIX
15031 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
15032 names of the subprograms executed by the compiler. No slash is added
15033 when this prefix is combined with the name of a subprogram, but you can
15034 specify a prefix that ends with a slash if you wish.
15036 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
15037 an appropriate prefix to use based on the pathname it was invoked with.
15039 If GCC cannot find the subprogram using the specified prefix, it
15040 tries looking in the usual places for the subprogram.
15042 The default value of @env{GCC_EXEC_PREFIX} is
15043 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
15044 the installed compiler. In many cases @var{prefix} is the value
15045 of @code{prefix} when you ran the @file{configure} script.
15047 Other prefixes specified with @option{-B} take precedence over this prefix.
15049 This prefix is also used for finding files such as @file{crt0.o} that are
15052 In addition, the prefix is used in an unusual way in finding the
15053 directories to search for header files. For each of the standard
15054 directories whose name normally begins with @samp{/usr/local/lib/gcc}
15055 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
15056 replacing that beginning with the specified prefix to produce an
15057 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
15058 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
15059 These alternate directories are searched first; the standard directories
15060 come next. If a standard directory begins with the configured
15061 @var{prefix} then the value of @var{prefix} is replaced by
15062 @env{GCC_EXEC_PREFIX} when looking for header files.
15064 @item COMPILER_PATH
15065 @findex COMPILER_PATH
15066 The value of @env{COMPILER_PATH} is a colon-separated list of
15067 directories, much like @env{PATH}. GCC tries the directories thus
15068 specified when searching for subprograms, if it can't find the
15069 subprograms using @env{GCC_EXEC_PREFIX}.
15072 @findex LIBRARY_PATH
15073 The value of @env{LIBRARY_PATH} is a colon-separated list of
15074 directories, much like @env{PATH}. When configured as a native compiler,
15075 GCC tries the directories thus specified when searching for special
15076 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
15077 using GCC also uses these directories when searching for ordinary
15078 libraries for the @option{-l} option (but directories specified with
15079 @option{-L} come first).
15083 @cindex locale definition
15084 This variable is used to pass locale information to the compiler. One way in
15085 which this information is used is to determine the character set to be used
15086 when character literals, string literals and comments are parsed in C and C++.
15087 When the compiler is configured to allow multibyte characters,
15088 the following values for @env{LANG} are recognized:
15092 Recognize JIS characters.
15094 Recognize SJIS characters.
15096 Recognize EUCJP characters.
15099 If @env{LANG} is not defined, or if it has some other value, then the
15100 compiler will use mblen and mbtowc as defined by the default locale to
15101 recognize and translate multibyte characters.
15105 Some additional environments variables affect the behavior of the
15108 @include cppenv.texi
15112 @node Precompiled Headers
15113 @section Using Precompiled Headers
15114 @cindex precompiled headers
15115 @cindex speed of compilation
15117 Often large projects have many header files that are included in every
15118 source file. The time the compiler takes to process these header files
15119 over and over again can account for nearly all of the time required to
15120 build the project. To make builds faster, GCC allows users to
15121 `precompile' a header file; then, if builds can use the precompiled
15122 header file they will be much faster.
15124 To create a precompiled header file, simply compile it as you would any
15125 other file, if necessary using the @option{-x} option to make the driver
15126 treat it as a C or C++ header file. You will probably want to use a
15127 tool like @command{make} to keep the precompiled header up-to-date when
15128 the headers it contains change.
15130 A precompiled header file will be searched for when @code{#include} is
15131 seen in the compilation. As it searches for the included file
15132 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
15133 compiler looks for a precompiled header in each directory just before it
15134 looks for the include file in that directory. The name searched for is
15135 the name specified in the @code{#include} with @samp{.gch} appended. If
15136 the precompiled header file can't be used, it is ignored.
15138 For instance, if you have @code{#include "all.h"}, and you have
15139 @file{all.h.gch} in the same directory as @file{all.h}, then the
15140 precompiled header file will be used if possible, and the original
15141 header will be used otherwise.
15143 Alternatively, you might decide to put the precompiled header file in a
15144 directory and use @option{-I} to ensure that directory is searched
15145 before (or instead of) the directory containing the original header.
15146 Then, if you want to check that the precompiled header file is always
15147 used, you can put a file of the same name as the original header in this
15148 directory containing an @code{#error} command.
15150 This also works with @option{-include}. So yet another way to use
15151 precompiled headers, good for projects not designed with precompiled
15152 header files in mind, is to simply take most of the header files used by
15153 a project, include them from another header file, precompile that header
15154 file, and @option{-include} the precompiled header. If the header files
15155 have guards against multiple inclusion, they will be skipped because
15156 they've already been included (in the precompiled header).
15158 If you need to precompile the same header file for different
15159 languages, targets, or compiler options, you can instead make a
15160 @emph{directory} named like @file{all.h.gch}, and put each precompiled
15161 header in the directory, perhaps using @option{-o}. It doesn't matter
15162 what you call the files in the directory, every precompiled header in
15163 the directory will be considered. The first precompiled header
15164 encountered in the directory that is valid for this compilation will
15165 be used; they're searched in no particular order.
15167 There are many other possibilities, limited only by your imagination,
15168 good sense, and the constraints of your build system.
15170 A precompiled header file can be used only when these conditions apply:
15174 Only one precompiled header can be used in a particular compilation.
15177 A precompiled header can't be used once the first C token is seen. You
15178 can have preprocessor directives before a precompiled header; you can
15179 even include a precompiled header from inside another header, so long as
15180 there are no C tokens before the @code{#include}.
15183 The precompiled header file must be produced for the same language as
15184 the current compilation. You can't use a C precompiled header for a C++
15188 The precompiled header file must have been produced by the same compiler
15189 binary as the current compilation is using.
15192 Any macros defined before the precompiled header is included must
15193 either be defined in the same way as when the precompiled header was
15194 generated, or must not affect the precompiled header, which usually
15195 means that they don't appear in the precompiled header at all.
15197 The @option{-D} option is one way to define a macro before a
15198 precompiled header is included; using a @code{#define} can also do it.
15199 There are also some options that define macros implicitly, like
15200 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
15203 @item If debugging information is output when using the precompiled
15204 header, using @option{-g} or similar, the same kind of debugging information
15205 must have been output when building the precompiled header. However,
15206 a precompiled header built using @option{-g} can be used in a compilation
15207 when no debugging information is being output.
15209 @item The same @option{-m} options must generally be used when building
15210 and using the precompiled header. @xref{Submodel Options},
15211 for any cases where this rule is relaxed.
15213 @item Each of the following options must be the same when building and using
15214 the precompiled header:
15216 @gccoptlist{-fexceptions -funit-at-a-time}
15219 Some other command-line options starting with @option{-f},
15220 @option{-p}, or @option{-O} must be defined in the same way as when
15221 the precompiled header was generated. At present, it's not clear
15222 which options are safe to change and which are not; the safest choice
15223 is to use exactly the same options when generating and using the
15224 precompiled header. The following are known to be safe:
15226 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
15227 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
15228 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
15233 For all of these except the last, the compiler will automatically
15234 ignore the precompiled header if the conditions aren't met. If you
15235 find an option combination that doesn't work and doesn't cause the
15236 precompiled header to be ignored, please consider filing a bug report,
15239 If you do use differing options when generating and using the
15240 precompiled header, the actual behavior will be a mixture of the
15241 behavior for the options. For instance, if you use @option{-g} to
15242 generate the precompiled header but not when using it, you may or may
15243 not get debugging information for routines in the precompiled header.
15245 @node Running Protoize
15246 @section Running Protoize
15248 The program @code{protoize} is an optional part of GCC@. You can use
15249 it to add prototypes to a program, thus converting the program to ISO
15250 C in one respect. The companion program @code{unprotoize} does the
15251 reverse: it removes argument types from any prototypes that are found.
15253 When you run these programs, you must specify a set of source files as
15254 command line arguments. The conversion programs start out by compiling
15255 these files to see what functions they define. The information gathered
15256 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
15258 After scanning comes actual conversion. The specified files are all
15259 eligible to be converted; any files they include (whether sources or
15260 just headers) are eligible as well.
15262 But not all the eligible files are converted. By default,
15263 @code{protoize} and @code{unprotoize} convert only source and header
15264 files in the current directory. You can specify additional directories
15265 whose files should be converted with the @option{-d @var{directory}}
15266 option. You can also specify particular files to exclude with the
15267 @option{-x @var{file}} option. A file is converted if it is eligible, its
15268 directory name matches one of the specified directory names, and its
15269 name within the directory has not been excluded.
15271 Basic conversion with @code{protoize} consists of rewriting most
15272 function definitions and function declarations to specify the types of
15273 the arguments. The only ones not rewritten are those for varargs
15276 @code{protoize} optionally inserts prototype declarations at the
15277 beginning of the source file, to make them available for any calls that
15278 precede the function's definition. Or it can insert prototype
15279 declarations with block scope in the blocks where undeclared functions
15282 Basic conversion with @code{unprotoize} consists of rewriting most
15283 function declarations to remove any argument types, and rewriting
15284 function definitions to the old-style pre-ISO form.
15286 Both conversion programs print a warning for any function declaration or
15287 definition that they can't convert. You can suppress these warnings
15290 The output from @code{protoize} or @code{unprotoize} replaces the
15291 original source file. The original file is renamed to a name ending
15292 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
15293 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
15294 for DOS) file already exists, then the source file is simply discarded.
15296 @code{protoize} and @code{unprotoize} both depend on GCC itself to
15297 scan the program and collect information about the functions it uses.
15298 So neither of these programs will work until GCC is installed.
15300 Here is a table of the options you can use with @code{protoize} and
15301 @code{unprotoize}. Each option works with both programs unless
15305 @item -B @var{directory}
15306 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
15307 usual directory (normally @file{/usr/local/lib}). This file contains
15308 prototype information about standard system functions. This option
15309 applies only to @code{protoize}.
15311 @item -c @var{compilation-options}
15312 Use @var{compilation-options} as the options when running @command{gcc} to
15313 produce the @samp{.X} files. The special option @option{-aux-info} is
15314 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
15316 Note that the compilation options must be given as a single argument to
15317 @code{protoize} or @code{unprotoize}. If you want to specify several
15318 @command{gcc} options, you must quote the entire set of compilation options
15319 to make them a single word in the shell.
15321 There are certain @command{gcc} arguments that you cannot use, because they
15322 would produce the wrong kind of output. These include @option{-g},
15323 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
15324 the @var{compilation-options}, they are ignored.
15327 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
15328 systems) instead of @samp{.c}. This is convenient if you are converting
15329 a C program to C++. This option applies only to @code{protoize}.
15332 Add explicit global declarations. This means inserting explicit
15333 declarations at the beginning of each source file for each function
15334 that is called in the file and was not declared. These declarations
15335 precede the first function definition that contains a call to an
15336 undeclared function. This option applies only to @code{protoize}.
15338 @item -i @var{string}
15339 Indent old-style parameter declarations with the string @var{string}.
15340 This option applies only to @code{protoize}.
15342 @code{unprotoize} converts prototyped function definitions to old-style
15343 function definitions, where the arguments are declared between the
15344 argument list and the initial @samp{@{}. By default, @code{unprotoize}
15345 uses five spaces as the indentation. If you want to indent with just
15346 one space instead, use @option{-i " "}.
15349 Keep the @samp{.X} files. Normally, they are deleted after conversion
15353 Add explicit local declarations. @code{protoize} with @option{-l} inserts
15354 a prototype declaration for each function in each block which calls the
15355 function without any declaration. This option applies only to
15359 Make no real changes. This mode just prints information about the conversions
15360 that would have been done without @option{-n}.
15363 Make no @samp{.save} files. The original files are simply deleted.
15364 Use this option with caution.
15366 @item -p @var{program}
15367 Use the program @var{program} as the compiler. Normally, the name
15368 @file{gcc} is used.
15371 Work quietly. Most warnings are suppressed.
15374 Print the version number, just like @option{-v} for @command{gcc}.
15377 If you need special compiler options to compile one of your program's
15378 source files, then you should generate that file's @samp{.X} file
15379 specially, by running @command{gcc} on that source file with the
15380 appropriate options and the option @option{-aux-info}. Then run
15381 @code{protoize} on the entire set of files. @code{protoize} will use
15382 the existing @samp{.X} file because it is newer than the source file.
15386 gcc -Dfoo=bar file1.c -aux-info file1.X
15391 You need to include the special files along with the rest in the
15392 @code{protoize} command, even though their @samp{.X} files already
15393 exist, because otherwise they won't get converted.
15395 @xref{Protoize Caveats}, for more information on how to use
15396 @code{protoize} successfully.