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-small-functions -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 -ftree-parallelize-loops @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}@r{[}-@var{sirevision}@r{]} @gol
448 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
449 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
450 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
451 -mno-id-shared-library -mshared-library-id=@var{n} @gol
452 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
453 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
454 -mfast-fp -minline-plt}
457 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
458 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
459 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
460 -mstack-align -mdata-align -mconst-align @gol
461 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
462 -melf -maout -melinux -mlinux -sim -sim2 @gol
463 -mmul-bug-workaround -mno-mul-bug-workaround}
466 @gccoptlist{-mmac -mpush-args}
468 @emph{Darwin Options}
469 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
470 -arch_only -bind_at_load -bundle -bundle_loader @gol
471 -client_name -compatibility_version -current_version @gol
473 -dependency-file -dylib_file -dylinker_install_name @gol
474 -dynamic -dynamiclib -exported_symbols_list @gol
475 -filelist -flat_namespace -force_cpusubtype_ALL @gol
476 -force_flat_namespace -headerpad_max_install_names @gol
478 -image_base -init -install_name -keep_private_externs @gol
479 -multi_module -multiply_defined -multiply_defined_unused @gol
480 -noall_load -no_dead_strip_inits_and_terms @gol
481 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
482 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
483 -private_bundle -read_only_relocs -sectalign @gol
484 -sectobjectsymbols -whyload -seg1addr @gol
485 -sectcreate -sectobjectsymbols -sectorder @gol
486 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
487 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
488 -segprot -segs_read_only_addr -segs_read_write_addr @gol
489 -single_module -static -sub_library -sub_umbrella @gol
490 -twolevel_namespace -umbrella -undefined @gol
491 -unexported_symbols_list -weak_reference_mismatches @gol
492 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
493 -mkernel -mone-byte-bool}
495 @emph{DEC Alpha Options}
496 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
497 -mieee -mieee-with-inexact -mieee-conformant @gol
498 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
499 -mtrap-precision=@var{mode} -mbuild-constants @gol
500 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
501 -mbwx -mmax -mfix -mcix @gol
502 -mfloat-vax -mfloat-ieee @gol
503 -mexplicit-relocs -msmall-data -mlarge-data @gol
504 -msmall-text -mlarge-text @gol
505 -mmemory-latency=@var{time}}
507 @emph{DEC Alpha/VMS Options}
508 @gccoptlist{-mvms-return-codes}
511 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
512 -mhard-float -msoft-float @gol
513 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
514 -mdouble -mno-double @gol
515 -mmedia -mno-media -mmuladd -mno-muladd @gol
516 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
517 -mlinked-fp -mlong-calls -malign-labels @gol
518 -mlibrary-pic -macc-4 -macc-8 @gol
519 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
520 -moptimize-membar -mno-optimize-membar @gol
521 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
522 -mvliw-branch -mno-vliw-branch @gol
523 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
524 -mno-nested-cond-exec -mtomcat-stats @gol
528 @emph{GNU/Linux Options}
529 @gccoptlist{-muclibc}
531 @emph{H8/300 Options}
532 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
535 @gccoptlist{-march=@var{architecture-type} @gol
536 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
537 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
538 -mfixed-range=@var{register-range} @gol
539 -mjump-in-delay -mlinker-opt -mlong-calls @gol
540 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
541 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
542 -mno-jump-in-delay -mno-long-load-store @gol
543 -mno-portable-runtime -mno-soft-float @gol
544 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
545 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
546 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
547 -munix=@var{unix-std} -nolibdld -static -threads}
549 @emph{i386 and x86-64 Options}
550 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
551 -mfpmath=@var{unit} @gol
552 -masm=@var{dialect} -mno-fancy-math-387 @gol
553 -mno-fp-ret-in-387 -msoft-float @gol
554 -mno-wide-multiply -mrtd -malign-double @gol
555 -mpreferred-stack-boundary=@var{num} -mcx16 -msahf -mrecip @gol
556 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 @gol
557 -msse4a -m3dnow -mpopcnt -mabm -msse5 @gol
558 -mthreads -mno-align-stringops -minline-all-stringops @gol
559 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
560 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
561 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
562 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
563 -mcmodel=@var{code-model} @gol
564 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
565 -mfused-madd -mno-fused-madd}
568 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
569 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
570 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
571 -minline-float-divide-max-throughput @gol
572 -minline-int-divide-min-latency @gol
573 -minline-int-divide-max-throughput @gol
574 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
575 -mno-dwarf2-asm -mearly-stop-bits @gol
576 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
577 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64 @gol
578 -mno-sched-br-data-spec -msched-ar-data-spec -mno-sched-control-spec @gol
579 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
580 -msched-ldc -mno-sched-control-ldc -mno-sched-spec-verbose @gol
581 -mno-sched-prefer-non-data-spec-insns @gol
582 -mno-sched-prefer-non-control-spec-insns @gol
583 -mno-sched-count-spec-in-critical-path}
585 @emph{M32R/D Options}
586 @gccoptlist{-m32r2 -m32rx -m32r @gol
588 -malign-loops -mno-align-loops @gol
589 -missue-rate=@var{number} @gol
590 -mbranch-cost=@var{number} @gol
591 -mmodel=@var{code-size-model-type} @gol
592 -msdata=@var{sdata-type} @gol
593 -mno-flush-func -mflush-func=@var{name} @gol
594 -mno-flush-trap -mflush-trap=@var{number} @gol
598 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
600 @emph{M680x0 Options}
601 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
602 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
603 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
604 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
605 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
606 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
607 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
608 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
610 @emph{M68hc1x Options}
611 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
612 -mauto-incdec -minmax -mlong-calls -mshort @gol
613 -msoft-reg-count=@var{count}}
616 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
617 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
618 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
619 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
620 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
623 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
624 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
625 -mips16 -mno-mips16 -mflip-mips16 @gol
626 -minterlink-mips16 -mno-interlink-mips16 @gol
627 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
628 -mshared -mno-shared -mxgot -mno-xgot -mgp32 -mgp64 @gol
629 -mfp32 -mfp64 -mhard-float -msoft-float @gol
630 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
631 -msmartmips -mno-smartmips @gol
632 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
633 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
634 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
635 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
636 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
637 -membedded-data -mno-embedded-data @gol
638 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
639 -mcode-readable=@var{setting} @gol
640 -msplit-addresses -mno-split-addresses @gol
641 -mexplicit-relocs -mno-explicit-relocs @gol
642 -mcheck-zero-division -mno-check-zero-division @gol
643 -mdivide-traps -mdivide-breaks @gol
644 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
645 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
646 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
647 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 -mno-fix-vr4130 @gol
648 -mfix-sb1 -mno-fix-sb1 @gol
649 -mflush-func=@var{func} -mno-flush-func @gol
650 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
651 -mfp-exceptions -mno-fp-exceptions @gol
652 -mvr4130-align -mno-vr4130-align}
655 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
656 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
657 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
658 -mno-base-addresses -msingle-exit -mno-single-exit}
660 @emph{MN10300 Options}
661 @gccoptlist{-mmult-bug -mno-mult-bug @gol
662 -mam33 -mno-am33 @gol
663 -mam33-2 -mno-am33-2 @gol
664 -mreturn-pointer-on-d0 @gol
668 @gccoptlist{-mno-crt0 -mbacc -msim @gol
669 -march=@var{cpu-type} }
671 @emph{PDP-11 Options}
672 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
673 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
674 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
675 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
676 -mbranch-expensive -mbranch-cheap @gol
677 -msplit -mno-split -munix-asm -mdec-asm}
679 @emph{PowerPC Options}
680 See RS/6000 and PowerPC Options.
682 @emph{RS/6000 and PowerPC Options}
683 @gccoptlist{-mcpu=@var{cpu-type} @gol
684 -mtune=@var{cpu-type} @gol
685 -mpower -mno-power -mpower2 -mno-power2 @gol
686 -mpowerpc -mpowerpc64 -mno-powerpc @gol
687 -maltivec -mno-altivec @gol
688 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
689 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
690 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
691 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mdfp -mno-dfp @gol
692 -mnew-mnemonics -mold-mnemonics @gol
693 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
694 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
695 -malign-power -malign-natural @gol
696 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
697 -mstring -mno-string -mupdate -mno-update @gol
698 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
699 -mstrict-align -mno-strict-align -mrelocatable @gol
700 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
701 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
702 -mdynamic-no-pic -maltivec -mswdiv @gol
703 -mprioritize-restricted-insns=@var{priority} @gol
704 -msched-costly-dep=@var{dependence_type} @gol
705 -minsert-sched-nops=@var{scheme} @gol
706 -mcall-sysv -mcall-netbsd @gol
707 -maix-struct-return -msvr4-struct-return @gol
708 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
709 -misel -mno-isel @gol
710 -misel=yes -misel=no @gol
712 -mspe=yes -mspe=no @gol
714 -mvrsave -mno-vrsave @gol
715 -mmulhw -mno-mulhw @gol
716 -mdlmzb -mno-dlmzb @gol
717 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
718 -mprototype -mno-prototype @gol
719 -msim -mmvme -mads -myellowknife -memb -msdata @gol
720 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
722 @emph{S/390 and zSeries Options}
723 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
724 -mhard-float -msoft-float -mlong-double-64 -mlong-double-128 @gol
725 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
726 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
727 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
728 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
729 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
732 @gccoptlist{-meb -mel @gol
736 -mscore5 -mscore5u -mscore7 -mscore7d}
739 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
740 -m4-nofpu -m4-single-only -m4-single -m4 @gol
741 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
742 -m5-64media -m5-64media-nofpu @gol
743 -m5-32media -m5-32media-nofpu @gol
744 -m5-compact -m5-compact-nofpu @gol
745 -mb -ml -mdalign -mrelax @gol
746 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
747 -mieee -misize -minline-ic_invalidate -mpadstruct -mspace @gol
748 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
749 -mdivsi3_libfunc=@var{name} @gol
750 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
754 @gccoptlist{-mcpu=@var{cpu-type} @gol
755 -mtune=@var{cpu-type} @gol
756 -mcmodel=@var{code-model} @gol
757 -m32 -m64 -mapp-regs -mno-app-regs @gol
758 -mfaster-structs -mno-faster-structs @gol
759 -mfpu -mno-fpu -mhard-float -msoft-float @gol
760 -mhard-quad-float -msoft-quad-float @gol
761 -mimpure-text -mno-impure-text -mlittle-endian @gol
762 -mstack-bias -mno-stack-bias @gol
763 -munaligned-doubles -mno-unaligned-doubles @gol
764 -mv8plus -mno-v8plus -mvis -mno-vis
765 -threads -pthreads -pthread}
768 @gccoptlist{-mwarn-reloc -merror-reloc @gol
769 -msafe-dma -munsafe-dma @gol
771 -msmall-mem -mlarge-mem -mstdmain @gol
772 -mfixed-range=@var{register-range}}
774 @emph{System V Options}
775 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
777 @emph{TMS320C3x/C4x Options}
778 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
779 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
780 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
781 -mparallel-insns -mparallel-mpy -mpreserve-float}
784 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
785 -mprolog-function -mno-prolog-function -mspace @gol
786 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
787 -mapp-regs -mno-app-regs @gol
788 -mdisable-callt -mno-disable-callt @gol
794 @gccoptlist{-mg -mgnu -munix}
796 @emph{VxWorks Options}
797 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
798 -Xbind-lazy -Xbind-now}
800 @emph{x86-64 Options}
801 See i386 and x86-64 Options.
803 @emph{Xstormy16 Options}
806 @emph{Xtensa Options}
807 @gccoptlist{-mconst16 -mno-const16 @gol
808 -mfused-madd -mno-fused-madd @gol
809 -mtext-section-literals -mno-text-section-literals @gol
810 -mtarget-align -mno-target-align @gol
811 -mlongcalls -mno-longcalls}
813 @emph{zSeries Options}
814 See S/390 and zSeries Options.
816 @item Code Generation Options
817 @xref{Code Gen Options,,Options for Code Generation Conventions}.
818 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
819 -ffixed-@var{reg} -fexceptions @gol
820 -fnon-call-exceptions -funwind-tables @gol
821 -fasynchronous-unwind-tables @gol
822 -finhibit-size-directive -finstrument-functions @gol
823 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
824 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
825 -fno-common -fno-ident @gol
826 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
827 -fno-jump-tables @gol
828 -frecord-gcc-switches @gol
829 -freg-struct-return -fshort-enums @gol
830 -fshort-double -fshort-wchar @gol
831 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
832 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
833 -fargument-alias -fargument-noalias @gol
834 -fargument-noalias-global -fargument-noalias-anything
835 -fleading-underscore -ftls-model=@var{model} @gol
836 -ftrapv -fwrapv -fbounds-check @gol
841 * Overall Options:: Controlling the kind of output:
842 an executable, object files, assembler files,
843 or preprocessed source.
844 * C Dialect Options:: Controlling the variant of C language compiled.
845 * C++ Dialect Options:: Variations on C++.
846 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
848 * Language Independent Options:: Controlling how diagnostics should be
850 * Warning Options:: How picky should the compiler be?
851 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
852 * Optimize Options:: How much optimization?
853 * Preprocessor Options:: Controlling header files and macro definitions.
854 Also, getting dependency information for Make.
855 * Assembler Options:: Passing options to the assembler.
856 * Link Options:: Specifying libraries and so on.
857 * Directory Options:: Where to find header files and libraries.
858 Where to find the compiler executable files.
859 * Spec Files:: How to pass switches to sub-processes.
860 * Target Options:: Running a cross-compiler, or an old version of GCC.
863 @node Overall Options
864 @section Options Controlling the Kind of Output
866 Compilation can involve up to four stages: preprocessing, compilation
867 proper, assembly and linking, always in that order. GCC is capable of
868 preprocessing and compiling several files either into several
869 assembler input files, or into one assembler input file; then each
870 assembler input file produces an object file, and linking combines all
871 the object files (those newly compiled, and those specified as input)
872 into an executable file.
874 @cindex file name suffix
875 For any given input file, the file name suffix determines what kind of
880 C source code which must be preprocessed.
883 C source code which should not be preprocessed.
886 C++ source code which should not be preprocessed.
889 Objective-C source code. Note that you must link with the @file{libobjc}
890 library to make an Objective-C program work.
893 Objective-C source code which should not be preprocessed.
897 Objective-C++ source code. Note that you must link with the @file{libobjc}
898 library to make an Objective-C++ program work. Note that @samp{.M} refers
899 to a literal capital M@.
902 Objective-C++ source code which should not be preprocessed.
905 C, C++, Objective-C or Objective-C++ header file to be turned into a
910 @itemx @var{file}.cxx
911 @itemx @var{file}.cpp
912 @itemx @var{file}.CPP
913 @itemx @var{file}.c++
915 C++ source code which must be preprocessed. Note that in @samp{.cxx},
916 the last two letters must both be literally @samp{x}. Likewise,
917 @samp{.C} refers to a literal capital C@.
921 Objective-C++ source code which must be preprocessed.
924 Objective-C++ source code which should not be preprocessed.
929 @itemx @var{file}.hxx
930 @itemx @var{file}.hpp
931 @itemx @var{file}.HPP
932 @itemx @var{file}.h++
933 @itemx @var{file}.tcc
934 C++ header file to be turned into a precompiled header.
937 @itemx @var{file}.for
938 @itemx @var{file}.FOR
939 Fixed form Fortran source code which should not be preprocessed.
942 @itemx @var{file}.fpp
943 @itemx @var{file}.FPP
944 Fixed form Fortran source code which must be preprocessed (with the traditional
948 @itemx @var{file}.f95
949 Free form Fortran source code which should not be preprocessed.
952 @itemx @var{file}.F95
953 Free form Fortran source code which must be preprocessed (with the
954 traditional preprocessor).
956 @c FIXME: Descriptions of Java file types.
963 Ada source code file which contains a library unit declaration (a
964 declaration of a package, subprogram, or generic, or a generic
965 instantiation), or a library unit renaming declaration (a package,
966 generic, or subprogram renaming declaration). Such files are also
969 @itemx @var{file}.adb
970 Ada source code file containing a library unit body (a subprogram or
971 package body). Such files are also called @dfn{bodies}.
973 @c GCC also knows about some suffixes for languages not yet included:
985 Assembler code which must be preprocessed.
988 An object file to be fed straight into linking.
989 Any file name with no recognized suffix is treated this way.
993 You can specify the input language explicitly with the @option{-x} option:
996 @item -x @var{language}
997 Specify explicitly the @var{language} for the following input files
998 (rather than letting the compiler choose a default based on the file
999 name suffix). This option applies to all following input files until
1000 the next @option{-x} option. Possible values for @var{language} are:
1002 c c-header c-cpp-output
1003 c++ c++-header c++-cpp-output
1004 objective-c objective-c-header objective-c-cpp-output
1005 objective-c++ objective-c++-header objective-c++-cpp-output
1006 assembler assembler-with-cpp
1014 Turn off any specification of a language, so that subsequent files are
1015 handled according to their file name suffixes (as they are if @option{-x}
1016 has not been used at all).
1018 @item -pass-exit-codes
1019 @opindex pass-exit-codes
1020 Normally the @command{gcc} program will exit with the code of 1 if any
1021 phase of the compiler returns a non-success return code. If you specify
1022 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1023 numerically highest error produced by any phase that returned an error
1024 indication. The C, C++, and Fortran frontends return 4, if an internal
1025 compiler error is encountered.
1028 If you only want some of the stages of compilation, you can use
1029 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1030 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1031 @command{gcc} is to stop. Note that some combinations (for example,
1032 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1037 Compile or assemble the source files, but do not link. The linking
1038 stage simply is not done. The ultimate output is in the form of an
1039 object file for each source file.
1041 By default, the object file name for a source file is made by replacing
1042 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1044 Unrecognized input files, not requiring compilation or assembly, are
1049 Stop after the stage of compilation proper; do not assemble. The output
1050 is in the form of an assembler code file for each non-assembler input
1053 By default, the assembler file name for a source file is made by
1054 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1056 Input files that don't require compilation are ignored.
1060 Stop after the preprocessing stage; do not run the compiler proper. The
1061 output is in the form of preprocessed source code, which is sent to the
1064 Input files which don't require preprocessing are ignored.
1066 @cindex output file option
1069 Place output in file @var{file}. This applies regardless to whatever
1070 sort of output is being produced, whether it be an executable file,
1071 an object file, an assembler file or preprocessed C code.
1073 If @option{-o} is not specified, the default is to put an executable
1074 file in @file{a.out}, the object file for
1075 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1076 assembler file in @file{@var{source}.s}, a precompiled header file in
1077 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1082 Print (on standard error output) the commands executed to run the stages
1083 of compilation. Also print the version number of the compiler driver
1084 program and of the preprocessor and the compiler proper.
1088 Like @option{-v} except the commands are not executed and all command
1089 arguments are quoted. This is useful for shell scripts to capture the
1090 driver-generated command lines.
1094 Use pipes rather than temporary files for communication between the
1095 various stages of compilation. This fails to work on some systems where
1096 the assembler is unable to read from a pipe; but the GNU assembler has
1101 If you are compiling multiple source files, this option tells the driver
1102 to pass all the source files to the compiler at once (for those
1103 languages for which the compiler can handle this). This will allow
1104 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1105 language for which this is supported is C@. If you pass source files for
1106 multiple languages to the driver, using this option, the driver will invoke
1107 the compiler(s) that support IMA once each, passing each compiler all the
1108 source files appropriate for it. For those languages that do not support
1109 IMA this option will be ignored, and the compiler will be invoked once for
1110 each source file in that language. If you use this option in conjunction
1111 with @option{-save-temps}, the compiler will generate multiple
1113 (one for each source file), but only one (combined) @file{.o} or
1118 Print (on the standard output) a description of the command line options
1119 understood by @command{gcc}. If the @option{-v} option is also specified
1120 then @option{--help} will also be passed on to the various processes
1121 invoked by @command{gcc}, so that they can display the command line options
1122 they accept. If the @option{-Wextra} option has also been specified
1123 (prior to the @option{--help} option), then command line options which
1124 have no documentation associated with them will also be displayed.
1127 @opindex target-help
1128 Print (on the standard output) a description of target-specific command
1129 line options for each tool. For some targets extra target-specific
1130 information may also be printed.
1132 @item --help=@var{class}@r{[},@var{qualifier}@r{]}
1133 Print (on the standard output) a description of the command line
1134 options understood by the compiler that fit into a specific class.
1135 The class can be one of @samp{optimizers}, @samp{warnings}, @samp{target},
1136 @samp{params}, or @var{language}:
1139 @item @samp{optimizers}
1140 This will display all of the optimization options supported by the
1143 @item @samp{warnings}
1144 This will display all of the options controlling warning messages
1145 produced by the compiler.
1148 This will display target-specific options. Unlike the
1149 @option{--target-help} option however, target-specific options of the
1150 linker and assembler will not be displayed. This is because those
1151 tools do not currently support the extended @option{--help=} syntax.
1154 This will display the values recognized by the @option{--param}
1157 @item @var{language}
1158 This will display the options supported for @var{language}, where
1159 @var{language} is the name of one of the languages supported in this
1163 This will display the options that are common to all languages.
1166 It is possible to further refine the output of the @option{--help=}
1167 option by adding a comma separated list of qualifiers after the
1168 class. These can be any from the following list:
1171 @item @samp{undocumented}
1172 Display only those options which are undocumented.
1175 Display options which take an argument that appears after an equal
1176 sign in the same continuous piece of text, such as:
1177 @samp{--help=target}.
1179 @item @samp{separate}
1180 Display options which take an argument that appears as a separate word
1181 following the original option, such as: @samp{-o output-file}.
1184 Thus for example to display all the undocumented target-specific
1185 switches supported by the compiler the following can be used:
1188 --help=target,undocumented
1191 The sense of a qualifier can be inverted by prefixing it with the
1192 @var{^} character, so for example to display all binary warning
1193 options (i.e. ones that are either on or off and that do not take an
1194 argument), which have a description the following can be used:
1197 --help=warnings,^joined,^undocumented
1200 A class can also be used as a qualifier, although this usually
1201 restricts the output by so much that there is nothing to display. One
1202 case where it does work however is when one of the classes is
1203 @var{target}. So for example to display all the target-specific
1204 optimization options the following can be used:
1207 --help=target,optimizers
1210 The @option{--help=} option can be repeated on the command line. Each
1211 successive use will display its requested class of options, skipping
1212 those that have already been displayed.
1214 If the @option{-Q} option appears on the command line before the
1215 @option{--help=} option, then the descriptive text displayed by
1216 @option{--help=} is changed. Instead of describing the displayed
1217 options, an indication is given as to whether the option is enabled,
1218 disabled or set to a specific value (assuming that the compiler
1219 knows this at the point where the @option{--help=} option is used).
1221 Here is a truncated example from the ARM port of @command{gcc}:
1224 % gcc -Q -mabi=2 --help=target -c
1225 The following options are target specific:
1227 -mabort-on-noreturn [disabled]
1231 The output is sensitive to the effects of previous command line
1232 options, so for example it is possible to find out which optimizations
1233 are enabled at @option{-O2} by using:
1236 -O2 --help=optimizers
1239 Alternatively you can discover which binary optimizations are enabled
1240 by @option{-O3} by using:
1243 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1244 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1245 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1250 Display the version number and copyrights of the invoked GCC@.
1252 @include @value{srcdir}/../libiberty/at-file.texi
1256 @section Compiling C++ Programs
1258 @cindex suffixes for C++ source
1259 @cindex C++ source file suffixes
1260 C++ source files conventionally use one of the suffixes @samp{.C},
1261 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1262 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1263 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1264 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1265 files with these names and compiles them as C++ programs even if you
1266 call the compiler the same way as for compiling C programs (usually
1267 with the name @command{gcc}).
1271 However, the use of @command{gcc} does not add the C++ library.
1272 @command{g++} is a program that calls GCC and treats @samp{.c},
1273 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1274 files unless @option{-x} is used, and automatically specifies linking
1275 against the C++ library. This program is also useful when
1276 precompiling a C header file with a @samp{.h} extension for use in C++
1277 compilations. On many systems, @command{g++} is also installed with
1278 the name @command{c++}.
1280 @cindex invoking @command{g++}
1281 When you compile C++ programs, you may specify many of the same
1282 command-line options that you use for compiling programs in any
1283 language; or command-line options meaningful for C and related
1284 languages; or options that are meaningful only for C++ programs.
1285 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1286 explanations of options for languages related to C@.
1287 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1288 explanations of options that are meaningful only for C++ programs.
1290 @node C Dialect Options
1291 @section Options Controlling C Dialect
1292 @cindex dialect options
1293 @cindex language dialect options
1294 @cindex options, dialect
1296 The following options control the dialect of C (or languages derived
1297 from C, such as C++, Objective-C and Objective-C++) that the compiler
1301 @cindex ANSI support
1305 In C mode, support all ISO C90 programs. In C++ mode,
1306 remove GNU extensions that conflict with ISO C++.
1308 This turns off certain features of GCC that are incompatible with ISO
1309 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1310 such as the @code{asm} and @code{typeof} keywords, and
1311 predefined macros such as @code{unix} and @code{vax} that identify the
1312 type of system you are using. It also enables the undesirable and
1313 rarely used ISO trigraph feature. For the C compiler,
1314 it disables recognition of C++ style @samp{//} comments as well as
1315 the @code{inline} keyword.
1317 The alternate keywords @code{__asm__}, @code{__extension__},
1318 @code{__inline__} and @code{__typeof__} continue to work despite
1319 @option{-ansi}. You would not want to use them in an ISO C program, of
1320 course, but it is useful to put them in header files that might be included
1321 in compilations done with @option{-ansi}. Alternate predefined macros
1322 such as @code{__unix__} and @code{__vax__} are also available, with or
1323 without @option{-ansi}.
1325 The @option{-ansi} option does not cause non-ISO programs to be
1326 rejected gratuitously. For that, @option{-pedantic} is required in
1327 addition to @option{-ansi}. @xref{Warning Options}.
1329 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1330 option is used. Some header files may notice this macro and refrain
1331 from declaring certain functions or defining certain macros that the
1332 ISO standard doesn't call for; this is to avoid interfering with any
1333 programs that might use these names for other things.
1335 Functions which would normally be built in but do not have semantics
1336 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1337 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1338 built-in functions provided by GCC}, for details of the functions
1343 Determine the language standard. This option is currently only
1344 supported when compiling C or C++. A value for this option must be
1345 provided; possible values are
1350 ISO C90 (same as @option{-ansi}).
1352 @item iso9899:199409
1353 ISO C90 as modified in amendment 1.
1359 ISO C99. Note that this standard is not yet fully supported; see
1360 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1361 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1364 Default, ISO C90 plus GNU extensions (including some C99 features).
1368 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1369 this will become the default. The name @samp{gnu9x} is deprecated.
1372 The 1998 ISO C++ standard plus amendments.
1375 The same as @option{-std=c++98} plus GNU extensions. This is the
1376 default for C++ code.
1379 The working draft of the upcoming ISO C++0x standard. This option
1380 enables experimental features that are likely to be included in
1381 C++0x. The working draft is constantly changing, and any feature that is
1382 enabled by this flag may be removed from future versions of GCC if it is
1383 not part of the C++0x standard.
1386 The same as @option{-std=c++0x} plus GNU extensions. As with
1387 @option{-std=c++0x}, this option enables experimental features that may
1388 be removed in future versions of GCC.
1391 Even when this option is not specified, you can still use some of the
1392 features of newer standards in so far as they do not conflict with
1393 previous C standards. For example, you may use @code{__restrict__} even
1394 when @option{-std=c99} is not specified.
1396 The @option{-std} options specifying some version of ISO C have the same
1397 effects as @option{-ansi}, except that features that were not in ISO C90
1398 but are in the specified version (for example, @samp{//} comments and
1399 the @code{inline} keyword in ISO C99) are not disabled.
1401 @xref{Standards,,Language Standards Supported by GCC}, for details of
1402 these standard versions.
1404 @item -fgnu89-inline
1405 @opindex fgnu89-inline
1406 The option @option{-fgnu89-inline} tells GCC to use the traditional
1407 GNU semantics for @code{inline} functions when in C99 mode.
1408 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1409 is accepted and ignored by GCC versions 4.1.3 up to but not including
1410 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1411 C99 mode. Using this option is roughly equivalent to adding the
1412 @code{gnu_inline} function attribute to all inline functions
1413 (@pxref{Function Attributes}).
1415 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1416 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1417 specifies the default behavior). This option was first supported in
1418 GCC 4.3. This option is not supported in C89 or gnu89 mode.
1420 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1421 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1422 in effect for @code{inline} functions. @xref{Common Predefined
1423 Macros,,,cpp,The C Preprocessor}.
1425 @item -aux-info @var{filename}
1427 Output to the given filename prototyped declarations for all functions
1428 declared and/or defined in a translation unit, including those in header
1429 files. This option is silently ignored in any language other than C@.
1431 Besides declarations, the file indicates, in comments, the origin of
1432 each declaration (source file and line), whether the declaration was
1433 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1434 @samp{O} for old, respectively, in the first character after the line
1435 number and the colon), and whether it came from a declaration or a
1436 definition (@samp{C} or @samp{F}, respectively, in the following
1437 character). In the case of function definitions, a K&R-style list of
1438 arguments followed by their declarations is also provided, inside
1439 comments, after the declaration.
1443 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1444 keyword, so that code can use these words as identifiers. You can use
1445 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1446 instead. @option{-ansi} implies @option{-fno-asm}.
1448 In C++, this switch only affects the @code{typeof} keyword, since
1449 @code{asm} and @code{inline} are standard keywords. You may want to
1450 use the @option{-fno-gnu-keywords} flag instead, which has the same
1451 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1452 switch only affects the @code{asm} and @code{typeof} keywords, since
1453 @code{inline} is a standard keyword in ISO C99.
1456 @itemx -fno-builtin-@var{function}
1457 @opindex fno-builtin
1458 @cindex built-in functions
1459 Don't recognize built-in functions that do not begin with
1460 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1461 functions provided by GCC}, for details of the functions affected,
1462 including those which are not built-in functions when @option{-ansi} or
1463 @option{-std} options for strict ISO C conformance are used because they
1464 do not have an ISO standard meaning.
1466 GCC normally generates special code to handle certain built-in functions
1467 more efficiently; for instance, calls to @code{alloca} may become single
1468 instructions that adjust the stack directly, and calls to @code{memcpy}
1469 may become inline copy loops. The resulting code is often both smaller
1470 and faster, but since the function calls no longer appear as such, you
1471 cannot set a breakpoint on those calls, nor can you change the behavior
1472 of the functions by linking with a different library. In addition,
1473 when a function is recognized as a built-in function, GCC may use
1474 information about that function to warn about problems with calls to
1475 that function, or to generate more efficient code, even if the
1476 resulting code still contains calls to that function. For example,
1477 warnings are given with @option{-Wformat} for bad calls to
1478 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1479 known not to modify global memory.
1481 With the @option{-fno-builtin-@var{function}} option
1482 only the built-in function @var{function} is
1483 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1484 function is named this is not built-in in this version of GCC, this
1485 option is ignored. There is no corresponding
1486 @option{-fbuiltin-@var{function}} option; if you wish to enable
1487 built-in functions selectively when using @option{-fno-builtin} or
1488 @option{-ffreestanding}, you may define macros such as:
1491 #define abs(n) __builtin_abs ((n))
1492 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1497 @cindex hosted environment
1499 Assert that compilation takes place in a hosted environment. This implies
1500 @option{-fbuiltin}. A hosted environment is one in which the
1501 entire standard library is available, and in which @code{main} has a return
1502 type of @code{int}. Examples are nearly everything except a kernel.
1503 This is equivalent to @option{-fno-freestanding}.
1505 @item -ffreestanding
1506 @opindex ffreestanding
1507 @cindex hosted environment
1509 Assert that compilation takes place in a freestanding environment. This
1510 implies @option{-fno-builtin}. A freestanding environment
1511 is one in which the standard library may not exist, and program startup may
1512 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1513 This is equivalent to @option{-fno-hosted}.
1515 @xref{Standards,,Language Standards Supported by GCC}, for details of
1516 freestanding and hosted environments.
1520 @cindex openmp parallel
1521 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1522 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1523 compiler generates parallel code according to the OpenMP Application
1524 Program Interface v2.5 @w{@uref{http://www.openmp.org/}}.
1526 @item -fms-extensions
1527 @opindex fms-extensions
1528 Accept some non-standard constructs used in Microsoft header files.
1530 Some cases of unnamed fields in structures and unions are only
1531 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1532 fields within structs/unions}, for details.
1536 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1537 options for strict ISO C conformance) implies @option{-trigraphs}.
1539 @item -no-integrated-cpp
1540 @opindex no-integrated-cpp
1541 Performs a compilation in two passes: preprocessing and compiling. This
1542 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1543 @option{-B} option. The user supplied compilation step can then add in
1544 an additional preprocessing step after normal preprocessing but before
1545 compiling. The default is to use the integrated cpp (internal cpp)
1547 The semantics of this option will change if "cc1", "cc1plus", and
1548 "cc1obj" are merged.
1550 @cindex traditional C language
1551 @cindex C language, traditional
1553 @itemx -traditional-cpp
1554 @opindex traditional-cpp
1555 @opindex traditional
1556 Formerly, these options caused GCC to attempt to emulate a pre-standard
1557 C compiler. They are now only supported with the @option{-E} switch.
1558 The preprocessor continues to support a pre-standard mode. See the GNU
1559 CPP manual for details.
1561 @item -fcond-mismatch
1562 @opindex fcond-mismatch
1563 Allow conditional expressions with mismatched types in the second and
1564 third arguments. The value of such an expression is void. This option
1565 is not supported for C++.
1567 @item -flax-vector-conversions
1568 @opindex flax-vector-conversions
1569 Allow implicit conversions between vectors with differing numbers of
1570 elements and/or incompatible element types. This option should not be
1573 @item -funsigned-char
1574 @opindex funsigned-char
1575 Let the type @code{char} be unsigned, like @code{unsigned char}.
1577 Each kind of machine has a default for what @code{char} should
1578 be. It is either like @code{unsigned char} by default or like
1579 @code{signed char} by default.
1581 Ideally, a portable program should always use @code{signed char} or
1582 @code{unsigned char} when it depends on the signedness of an object.
1583 But many programs have been written to use plain @code{char} and
1584 expect it to be signed, or expect it to be unsigned, depending on the
1585 machines they were written for. This option, and its inverse, let you
1586 make such a program work with the opposite default.
1588 The type @code{char} is always a distinct type from each of
1589 @code{signed char} or @code{unsigned char}, even though its behavior
1590 is always just like one of those two.
1593 @opindex fsigned-char
1594 Let the type @code{char} be signed, like @code{signed char}.
1596 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1597 the negative form of @option{-funsigned-char}. Likewise, the option
1598 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1600 @item -fsigned-bitfields
1601 @itemx -funsigned-bitfields
1602 @itemx -fno-signed-bitfields
1603 @itemx -fno-unsigned-bitfields
1604 @opindex fsigned-bitfields
1605 @opindex funsigned-bitfields
1606 @opindex fno-signed-bitfields
1607 @opindex fno-unsigned-bitfields
1608 These options control whether a bit-field is signed or unsigned, when the
1609 declaration does not use either @code{signed} or @code{unsigned}. By
1610 default, such a bit-field is signed, because this is consistent: the
1611 basic integer types such as @code{int} are signed types.
1614 @node C++ Dialect Options
1615 @section Options Controlling C++ Dialect
1617 @cindex compiler options, C++
1618 @cindex C++ options, command line
1619 @cindex options, C++
1620 This section describes the command-line options that are only meaningful
1621 for C++ programs; but you can also use most of the GNU compiler options
1622 regardless of what language your program is in. For example, you
1623 might compile a file @code{firstClass.C} like this:
1626 g++ -g -frepo -O -c firstClass.C
1630 In this example, only @option{-frepo} is an option meant
1631 only for C++ programs; you can use the other options with any
1632 language supported by GCC@.
1634 Here is a list of options that are @emph{only} for compiling C++ programs:
1638 @item -fabi-version=@var{n}
1639 @opindex fabi-version
1640 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1641 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1642 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1643 the version that conforms most closely to the C++ ABI specification.
1644 Therefore, the ABI obtained using version 0 will change as ABI bugs
1647 The default is version 2.
1649 @item -fno-access-control
1650 @opindex fno-access-control
1651 Turn off all access checking. This switch is mainly useful for working
1652 around bugs in the access control code.
1656 Check that the pointer returned by @code{operator new} is non-null
1657 before attempting to modify the storage allocated. This check is
1658 normally unnecessary because the C++ standard specifies that
1659 @code{operator new} will only return @code{0} if it is declared
1660 @samp{throw()}, in which case the compiler will always check the
1661 return value even without this option. In all other cases, when
1662 @code{operator new} has a non-empty exception specification, memory
1663 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1664 @samp{new (nothrow)}.
1666 @item -fconserve-space
1667 @opindex fconserve-space
1668 Put uninitialized or runtime-initialized global variables into the
1669 common segment, as C does. This saves space in the executable at the
1670 cost of not diagnosing duplicate definitions. If you compile with this
1671 flag and your program mysteriously crashes after @code{main()} has
1672 completed, you may have an object that is being destroyed twice because
1673 two definitions were merged.
1675 This option is no longer useful on most targets, now that support has
1676 been added for putting variables into BSS without making them common.
1678 @item -ffriend-injection
1679 @opindex ffriend-injection
1680 Inject friend functions into the enclosing namespace, so that they are
1681 visible outside the scope of the class in which they are declared.
1682 Friend functions were documented to work this way in the old Annotated
1683 C++ Reference Manual, and versions of G++ before 4.1 always worked
1684 that way. However, in ISO C++ a friend function which is not declared
1685 in an enclosing scope can only be found using argument dependent
1686 lookup. This option causes friends to be injected as they were in
1689 This option is for compatibility, and may be removed in a future
1692 @item -fno-elide-constructors
1693 @opindex fno-elide-constructors
1694 The C++ standard allows an implementation to omit creating a temporary
1695 which is only used to initialize another object of the same type.
1696 Specifying this option disables that optimization, and forces G++ to
1697 call the copy constructor in all cases.
1699 @item -fno-enforce-eh-specs
1700 @opindex fno-enforce-eh-specs
1701 Don't generate code to check for violation of exception specifications
1702 at runtime. This option violates the C++ standard, but may be useful
1703 for reducing code size in production builds, much like defining
1704 @samp{NDEBUG}. This does not give user code permission to throw
1705 exceptions in violation of the exception specifications; the compiler
1706 will still optimize based on the specifications, so throwing an
1707 unexpected exception will result in undefined behavior.
1710 @itemx -fno-for-scope
1712 @opindex fno-for-scope
1713 If @option{-ffor-scope} is specified, the scope of variables declared in
1714 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1715 as specified by the C++ standard.
1716 If @option{-fno-for-scope} is specified, the scope of variables declared in
1717 a @i{for-init-statement} extends to the end of the enclosing scope,
1718 as was the case in old versions of G++, and other (traditional)
1719 implementations of C++.
1721 The default if neither flag is given to follow the standard,
1722 but to allow and give a warning for old-style code that would
1723 otherwise be invalid, or have different behavior.
1725 @item -fno-gnu-keywords
1726 @opindex fno-gnu-keywords
1727 Do not recognize @code{typeof} as a keyword, so that code can use this
1728 word as an identifier. You can use the keyword @code{__typeof__} instead.
1729 @option{-ansi} implies @option{-fno-gnu-keywords}.
1731 @item -fno-implicit-templates
1732 @opindex fno-implicit-templates
1733 Never emit code for non-inline templates which are instantiated
1734 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1735 @xref{Template Instantiation}, for more information.
1737 @item -fno-implicit-inline-templates
1738 @opindex fno-implicit-inline-templates
1739 Don't emit code for implicit instantiations of inline templates, either.
1740 The default is to handle inlines differently so that compiles with and
1741 without optimization will need the same set of explicit instantiations.
1743 @item -fno-implement-inlines
1744 @opindex fno-implement-inlines
1745 To save space, do not emit out-of-line copies of inline functions
1746 controlled by @samp{#pragma implementation}. This will cause linker
1747 errors if these functions are not inlined everywhere they are called.
1749 @item -fms-extensions
1750 @opindex fms-extensions
1751 Disable pedantic warnings about constructs used in MFC, such as implicit
1752 int and getting a pointer to member function via non-standard syntax.
1754 @item -fno-nonansi-builtins
1755 @opindex fno-nonansi-builtins
1756 Disable built-in declarations of functions that are not mandated by
1757 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1758 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1760 @item -fno-operator-names
1761 @opindex fno-operator-names
1762 Do not treat the operator name keywords @code{and}, @code{bitand},
1763 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1764 synonyms as keywords.
1766 @item -fno-optional-diags
1767 @opindex fno-optional-diags
1768 Disable diagnostics that the standard says a compiler does not need to
1769 issue. Currently, the only such diagnostic issued by G++ is the one for
1770 a name having multiple meanings within a class.
1773 @opindex fpermissive
1774 Downgrade some diagnostics about nonconformant code from errors to
1775 warnings. Thus, using @option{-fpermissive} will allow some
1776 nonconforming code to compile.
1780 Enable automatic template instantiation at link time. This option also
1781 implies @option{-fno-implicit-templates}. @xref{Template
1782 Instantiation}, for more information.
1786 Disable generation of information about every class with virtual
1787 functions for use by the C++ runtime type identification features
1788 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1789 of the language, you can save some space by using this flag. Note that
1790 exception handling uses the same information, but it will generate it as
1791 needed. The @samp{dynamic_cast} operator can still be used for casts that
1792 do not require runtime type information, i.e. casts to @code{void *} or to
1793 unambiguous base classes.
1797 Emit statistics about front-end processing at the end of the compilation.
1798 This information is generally only useful to the G++ development team.
1800 @item -ftemplate-depth-@var{n}
1801 @opindex ftemplate-depth
1802 Set the maximum instantiation depth for template classes to @var{n}.
1803 A limit on the template instantiation depth is needed to detect
1804 endless recursions during template class instantiation. ANSI/ISO C++
1805 conforming programs must not rely on a maximum depth greater than 17.
1807 @item -fno-threadsafe-statics
1808 @opindex fno-threadsafe-statics
1809 Do not emit the extra code to use the routines specified in the C++
1810 ABI for thread-safe initialization of local statics. You can use this
1811 option to reduce code size slightly in code that doesn't need to be
1814 @item -fuse-cxa-atexit
1815 @opindex fuse-cxa-atexit
1816 Register destructors for objects with static storage duration with the
1817 @code{__cxa_atexit} function rather than the @code{atexit} function.
1818 This option is required for fully standards-compliant handling of static
1819 destructors, but will only work if your C library supports
1820 @code{__cxa_atexit}.
1822 @item -fno-use-cxa-get-exception-ptr
1823 @opindex fno-use-cxa-get-exception-ptr
1824 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1825 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1826 if the runtime routine is not available.
1828 @item -fvisibility-inlines-hidden
1829 @opindex fvisibility-inlines-hidden
1830 This switch declares that the user does not attempt to compare
1831 pointers to inline methods where the addresses of the two functions
1832 were taken in different shared objects.
1834 The effect of this is that GCC may, effectively, mark inline methods with
1835 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1836 appear in the export table of a DSO and do not require a PLT indirection
1837 when used within the DSO@. Enabling this option can have a dramatic effect
1838 on load and link times of a DSO as it massively reduces the size of the
1839 dynamic export table when the library makes heavy use of templates.
1841 The behavior of this switch is not quite the same as marking the
1842 methods as hidden directly, because it does not affect static variables
1843 local to the function or cause the compiler to deduce that
1844 the function is defined in only one shared object.
1846 You may mark a method as having a visibility explicitly to negate the
1847 effect of the switch for that method. For example, if you do want to
1848 compare pointers to a particular inline method, you might mark it as
1849 having default visibility. Marking the enclosing class with explicit
1850 visibility will have no effect.
1852 Explicitly instantiated inline methods are unaffected by this option
1853 as their linkage might otherwise cross a shared library boundary.
1854 @xref{Template Instantiation}.
1856 @item -fvisibility-ms-compat
1857 @opindex fvisibility-ms-compat
1858 This flag attempts to use visibility settings to make GCC's C++
1859 linkage model compatible with that of Microsoft Visual Studio.
1861 The flag makes these changes to GCC's linkage model:
1865 It sets the default visibility to @code{hidden}, like
1866 @option{-fvisibility=hidden}.
1869 Types, but not their members, are not hidden by default.
1872 The One Definition Rule is relaxed for types without explicit
1873 visibility specifications which are defined in more than one different
1874 shared object: those declarations are permitted if they would have
1875 been permitted when this option was not used.
1878 In new code it is better to use @option{-fvisibility=hidden} and
1879 export those classes which are intended to be externally visible.
1880 Unfortunately it is possible for code to rely, perhaps accidentally,
1881 on the Visual Studio behavior.
1883 Among the consequences of these changes are that static data members
1884 of the same type with the same name but defined in different shared
1885 objects will be different, so changing one will not change the other;
1886 and that pointers to function members defined in different shared
1887 objects may not compare equal. When this flag is given, it is a
1888 violation of the ODR to define types with the same name differently.
1892 Do not use weak symbol support, even if it is provided by the linker.
1893 By default, G++ will use weak symbols if they are available. This
1894 option exists only for testing, and should not be used by end-users;
1895 it will result in inferior code and has no benefits. This option may
1896 be removed in a future release of G++.
1900 Do not search for header files in the standard directories specific to
1901 C++, but do still search the other standard directories. (This option
1902 is used when building the C++ library.)
1905 In addition, these optimization, warning, and code generation options
1906 have meanings only for C++ programs:
1909 @item -fno-default-inline
1910 @opindex fno-default-inline
1911 Do not assume @samp{inline} for functions defined inside a class scope.
1912 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1913 functions will have linkage like inline functions; they just won't be
1916 @item -Wabi @r{(C++ only)}
1918 Warn when G++ generates code that is probably not compatible with the
1919 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1920 all such cases, there are probably some cases that are not warned about,
1921 even though G++ is generating incompatible code. There may also be
1922 cases where warnings are emitted even though the code that is generated
1925 You should rewrite your code to avoid these warnings if you are
1926 concerned about the fact that code generated by G++ may not be binary
1927 compatible with code generated by other compilers.
1929 The known incompatibilities at this point include:
1934 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1935 pack data into the same byte as a base class. For example:
1938 struct A @{ virtual void f(); int f1 : 1; @};
1939 struct B : public A @{ int f2 : 1; @};
1943 In this case, G++ will place @code{B::f2} into the same byte
1944 as@code{A::f1}; other compilers will not. You can avoid this problem
1945 by explicitly padding @code{A} so that its size is a multiple of the
1946 byte size on your platform; that will cause G++ and other compilers to
1947 layout @code{B} identically.
1950 Incorrect handling of tail-padding for virtual bases. G++ does not use
1951 tail padding when laying out virtual bases. For example:
1954 struct A @{ virtual void f(); char c1; @};
1955 struct B @{ B(); char c2; @};
1956 struct C : public A, public virtual B @{@};
1960 In this case, G++ will not place @code{B} into the tail-padding for
1961 @code{A}; other compilers will. You can avoid this problem by
1962 explicitly padding @code{A} so that its size is a multiple of its
1963 alignment (ignoring virtual base classes); that will cause G++ and other
1964 compilers to layout @code{C} identically.
1967 Incorrect handling of bit-fields with declared widths greater than that
1968 of their underlying types, when the bit-fields appear in a union. For
1972 union U @{ int i : 4096; @};
1976 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1977 union too small by the number of bits in an @code{int}.
1980 Empty classes can be placed at incorrect offsets. For example:
1990 struct C : public B, public A @{@};
1994 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1995 it should be placed at offset zero. G++ mistakenly believes that the
1996 @code{A} data member of @code{B} is already at offset zero.
1999 Names of template functions whose types involve @code{typename} or
2000 template template parameters can be mangled incorrectly.
2003 template <typename Q>
2004 void f(typename Q::X) @{@}
2006 template <template <typename> class Q>
2007 void f(typename Q<int>::X) @{@}
2011 Instantiations of these templates may be mangled incorrectly.
2015 @item -Wctor-dtor-privacy @r{(C++ only)}
2016 @opindex Wctor-dtor-privacy
2017 Warn when a class seems unusable because all the constructors or
2018 destructors in that class are private, and it has neither friends nor
2019 public static member functions.
2021 @item -Wnon-virtual-dtor @r{(C++ only)}
2022 @opindex Wnon-virtual-dtor
2023 Warn when a class has virtual functions and accessible non-virtual
2024 destructor, in which case it would be possible but unsafe to delete
2025 an instance of a derived class through a pointer to the base class.
2026 This warning is also enabled if -Weffc++ is specified.
2028 @item -Wreorder @r{(C++ only)}
2030 @cindex reordering, warning
2031 @cindex warning for reordering of member initializers
2032 Warn when the order of member initializers given in the code does not
2033 match the order in which they must be executed. For instance:
2039 A(): j (0), i (1) @{ @}
2043 The compiler will rearrange the member initializers for @samp{i}
2044 and @samp{j} to match the declaration order of the members, emitting
2045 a warning to that effect. This warning is enabled by @option{-Wall}.
2048 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2051 @item -Weffc++ @r{(C++ only)}
2053 Warn about violations of the following style guidelines from Scott Meyers'
2054 @cite{Effective C++} book:
2058 Item 11: Define a copy constructor and an assignment operator for classes
2059 with dynamically allocated memory.
2062 Item 12: Prefer initialization to assignment in constructors.
2065 Item 14: Make destructors virtual in base classes.
2068 Item 15: Have @code{operator=} return a reference to @code{*this}.
2071 Item 23: Don't try to return a reference when you must return an object.
2075 Also warn about violations of the following style guidelines from
2076 Scott Meyers' @cite{More Effective C++} book:
2080 Item 6: Distinguish between prefix and postfix forms of increment and
2081 decrement operators.
2084 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2088 When selecting this option, be aware that the standard library
2089 headers do not obey all of these guidelines; use @samp{grep -v}
2090 to filter out those warnings.
2092 @item -Wno-deprecated @r{(C++ only)}
2093 @opindex Wno-deprecated
2094 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
2096 @item -Wstrict-null-sentinel @r{(C++ only)}
2097 @opindex Wstrict-null-sentinel
2098 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2099 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2100 to @code{__null}. Although it is a null pointer constant not a null pointer,
2101 it is guaranteed to of the same size as a pointer. But this use is
2102 not portable across different compilers.
2104 @item -Wno-non-template-friend @r{(C++ only)}
2105 @opindex Wno-non-template-friend
2106 Disable warnings when non-templatized friend functions are declared
2107 within a template. Since the advent of explicit template specification
2108 support in G++, if the name of the friend is an unqualified-id (i.e.,
2109 @samp{friend foo(int)}), the C++ language specification demands that the
2110 friend declare or define an ordinary, nontemplate function. (Section
2111 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2112 could be interpreted as a particular specialization of a templatized
2113 function. Because this non-conforming behavior is no longer the default
2114 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2115 check existing code for potential trouble spots and is on by default.
2116 This new compiler behavior can be turned off with
2117 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2118 but disables the helpful warning.
2120 @item -Wold-style-cast @r{(C++ only)}
2121 @opindex Wold-style-cast
2122 Warn if an old-style (C-style) cast to a non-void type is used within
2123 a C++ program. The new-style casts (@samp{dynamic_cast},
2124 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2125 less vulnerable to unintended effects and much easier to search for.
2127 @item -Woverloaded-virtual @r{(C++ only)}
2128 @opindex Woverloaded-virtual
2129 @cindex overloaded virtual fn, warning
2130 @cindex warning for overloaded virtual fn
2131 Warn when a function declaration hides virtual functions from a
2132 base class. For example, in:
2139 struct B: public A @{
2144 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2152 will fail to compile.
2154 @item -Wno-pmf-conversions @r{(C++ only)}
2155 @opindex Wno-pmf-conversions
2156 Disable the diagnostic for converting a bound pointer to member function
2159 @item -Wsign-promo @r{(C++ only)}
2160 @opindex Wsign-promo
2161 Warn when overload resolution chooses a promotion from unsigned or
2162 enumerated type to a signed type, over a conversion to an unsigned type of
2163 the same size. Previous versions of G++ would try to preserve
2164 unsignedness, but the standard mandates the current behavior.
2169 A& operator = (int);
2179 In this example, G++ will synthesize a default @samp{A& operator =
2180 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2183 @node Objective-C and Objective-C++ Dialect Options
2184 @section Options Controlling Objective-C and Objective-C++ Dialects
2186 @cindex compiler options, Objective-C and Objective-C++
2187 @cindex Objective-C and Objective-C++ options, command line
2188 @cindex options, Objective-C and Objective-C++
2189 (NOTE: This manual does not describe the Objective-C and Objective-C++
2190 languages themselves. See @xref{Standards,,Language Standards
2191 Supported by GCC}, for references.)
2193 This section describes the command-line options that are only meaningful
2194 for Objective-C and Objective-C++ programs, but you can also use most of
2195 the language-independent GNU compiler options.
2196 For example, you might compile a file @code{some_class.m} like this:
2199 gcc -g -fgnu-runtime -O -c some_class.m
2203 In this example, @option{-fgnu-runtime} is an option meant only for
2204 Objective-C and Objective-C++ programs; you can use the other options with
2205 any language supported by GCC@.
2207 Note that since Objective-C is an extension of the C language, Objective-C
2208 compilations may also use options specific to the C front-end (e.g.,
2209 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2210 C++-specific options (e.g., @option{-Wabi}).
2212 Here is a list of options that are @emph{only} for compiling Objective-C
2213 and Objective-C++ programs:
2216 @item -fconstant-string-class=@var{class-name}
2217 @opindex fconstant-string-class
2218 Use @var{class-name} as the name of the class to instantiate for each
2219 literal string specified with the syntax @code{@@"@dots{}"}. The default
2220 class name is @code{NXConstantString} if the GNU runtime is being used, and
2221 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2222 @option{-fconstant-cfstrings} option, if also present, will override the
2223 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2224 to be laid out as constant CoreFoundation strings.
2227 @opindex fgnu-runtime
2228 Generate object code compatible with the standard GNU Objective-C
2229 runtime. This is the default for most types of systems.
2231 @item -fnext-runtime
2232 @opindex fnext-runtime
2233 Generate output compatible with the NeXT runtime. This is the default
2234 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2235 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2238 @item -fno-nil-receivers
2239 @opindex fno-nil-receivers
2240 Assume that all Objective-C message dispatches (e.g.,
2241 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2242 is not @code{nil}. This allows for more efficient entry points in the runtime
2243 to be used. Currently, this option is only available in conjunction with
2244 the NeXT runtime on Mac OS X 10.3 and later.
2246 @item -fobjc-call-cxx-cdtors
2247 @opindex fobjc-call-cxx-cdtors
2248 For each Objective-C class, check if any of its instance variables is a
2249 C++ object with a non-trivial default constructor. If so, synthesize a
2250 special @code{- (id) .cxx_construct} instance method that will run
2251 non-trivial default constructors on any such instance variables, in order,
2252 and then return @code{self}. Similarly, check if any instance variable
2253 is a C++ object with a non-trivial destructor, and if so, synthesize a
2254 special @code{- (void) .cxx_destruct} method that will run
2255 all such default destructors, in reverse order.
2257 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2258 thusly generated will only operate on instance variables declared in the
2259 current Objective-C class, and not those inherited from superclasses. It
2260 is the responsibility of the Objective-C runtime to invoke all such methods
2261 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2262 will be invoked by the runtime immediately after a new object
2263 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2264 be invoked immediately before the runtime deallocates an object instance.
2266 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2267 support for invoking the @code{- (id) .cxx_construct} and
2268 @code{- (void) .cxx_destruct} methods.
2270 @item -fobjc-direct-dispatch
2271 @opindex fobjc-direct-dispatch
2272 Allow fast jumps to the message dispatcher. On Darwin this is
2273 accomplished via the comm page.
2275 @item -fobjc-exceptions
2276 @opindex fobjc-exceptions
2277 Enable syntactic support for structured exception handling in Objective-C,
2278 similar to what is offered by C++ and Java. This option is
2279 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2288 @@catch (AnObjCClass *exc) @{
2295 @@catch (AnotherClass *exc) @{
2298 @@catch (id allOthers) @{
2308 The @code{@@throw} statement may appear anywhere in an Objective-C or
2309 Objective-C++ program; when used inside of a @code{@@catch} block, the
2310 @code{@@throw} may appear without an argument (as shown above), in which case
2311 the object caught by the @code{@@catch} will be rethrown.
2313 Note that only (pointers to) Objective-C objects may be thrown and
2314 caught using this scheme. When an object is thrown, it will be caught
2315 by the nearest @code{@@catch} clause capable of handling objects of that type,
2316 analogously to how @code{catch} blocks work in C++ and Java. A
2317 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2318 any and all Objective-C exceptions not caught by previous @code{@@catch}
2321 The @code{@@finally} clause, if present, will be executed upon exit from the
2322 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2323 regardless of whether any exceptions are thrown, caught or rethrown
2324 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2325 of the @code{finally} clause in Java.
2327 There are several caveats to using the new exception mechanism:
2331 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2332 idioms provided by the @code{NSException} class, the new
2333 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2334 systems, due to additional functionality needed in the (NeXT) Objective-C
2338 As mentioned above, the new exceptions do not support handling
2339 types other than Objective-C objects. Furthermore, when used from
2340 Objective-C++, the Objective-C exception model does not interoperate with C++
2341 exceptions at this time. This means you cannot @code{@@throw} an exception
2342 from Objective-C and @code{catch} it in C++, or vice versa
2343 (i.e., @code{throw @dots{} @@catch}).
2346 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2347 blocks for thread-safe execution:
2350 @@synchronized (ObjCClass *guard) @{
2355 Upon entering the @code{@@synchronized} block, a thread of execution shall
2356 first check whether a lock has been placed on the corresponding @code{guard}
2357 object by another thread. If it has, the current thread shall wait until
2358 the other thread relinquishes its lock. Once @code{guard} becomes available,
2359 the current thread will place its own lock on it, execute the code contained in
2360 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2361 making @code{guard} available to other threads).
2363 Unlike Java, Objective-C does not allow for entire methods to be marked
2364 @code{@@synchronized}. Note that throwing exceptions out of
2365 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2366 to be unlocked properly.
2370 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2372 @item -freplace-objc-classes
2373 @opindex freplace-objc-classes
2374 Emit a special marker instructing @command{ld(1)} not to statically link in
2375 the resulting object file, and allow @command{dyld(1)} to load it in at
2376 run time instead. This is used in conjunction with the Fix-and-Continue
2377 debugging mode, where the object file in question may be recompiled and
2378 dynamically reloaded in the course of program execution, without the need
2379 to restart the program itself. Currently, Fix-and-Continue functionality
2380 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2385 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2386 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2387 compile time) with static class references that get initialized at load time,
2388 which improves run-time performance. Specifying the @option{-fzero-link} flag
2389 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2390 to be retained. This is useful in Zero-Link debugging mode, since it allows
2391 for individual class implementations to be modified during program execution.
2395 Dump interface declarations for all classes seen in the source file to a
2396 file named @file{@var{sourcename}.decl}.
2398 @item -Wassign-intercept
2399 @opindex Wassign-intercept
2400 Warn whenever an Objective-C assignment is being intercepted by the
2404 @opindex Wno-protocol
2405 If a class is declared to implement a protocol, a warning is issued for
2406 every method in the protocol that is not implemented by the class. The
2407 default behavior is to issue a warning for every method not explicitly
2408 implemented in the class, even if a method implementation is inherited
2409 from the superclass. If you use the @option{-Wno-protocol} option, then
2410 methods inherited from the superclass are considered to be implemented,
2411 and no warning is issued for them.
2415 Warn if multiple methods of different types for the same selector are
2416 found during compilation. The check is performed on the list of methods
2417 in the final stage of compilation. Additionally, a check is performed
2418 for each selector appearing in a @code{@@selector(@dots{})}
2419 expression, and a corresponding method for that selector has been found
2420 during compilation. Because these checks scan the method table only at
2421 the end of compilation, these warnings are not produced if the final
2422 stage of compilation is not reached, for example because an error is
2423 found during compilation, or because the @option{-fsyntax-only} option is
2426 @item -Wstrict-selector-match
2427 @opindex Wstrict-selector-match
2428 Warn if multiple methods with differing argument and/or return types are
2429 found for a given selector when attempting to send a message using this
2430 selector to a receiver of type @code{id} or @code{Class}. When this flag
2431 is off (which is the default behavior), the compiler will omit such warnings
2432 if any differences found are confined to types which share the same size
2435 @item -Wundeclared-selector
2436 @opindex Wundeclared-selector
2437 Warn if a @code{@@selector(@dots{})} expression referring to an
2438 undeclared selector is found. A selector is considered undeclared if no
2439 method with that name has been declared before the
2440 @code{@@selector(@dots{})} expression, either explicitly in an
2441 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2442 an @code{@@implementation} section. This option always performs its
2443 checks as soon as a @code{@@selector(@dots{})} expression is found,
2444 while @option{-Wselector} only performs its checks in the final stage of
2445 compilation. This also enforces the coding style convention
2446 that methods and selectors must be declared before being used.
2448 @item -print-objc-runtime-info
2449 @opindex print-objc-runtime-info
2450 Generate C header describing the largest structure that is passed by
2455 @node Language Independent Options
2456 @section Options to Control Diagnostic Messages Formatting
2457 @cindex options to control diagnostics formatting
2458 @cindex diagnostic messages
2459 @cindex message formatting
2461 Traditionally, diagnostic messages have been formatted irrespective of
2462 the output device's aspect (e.g.@: its width, @dots{}). The options described
2463 below can be used to control the diagnostic messages formatting
2464 algorithm, e.g.@: how many characters per line, how often source location
2465 information should be reported. Right now, only the C++ front end can
2466 honor these options. However it is expected, in the near future, that
2467 the remaining front ends would be able to digest them correctly.
2470 @item -fmessage-length=@var{n}
2471 @opindex fmessage-length
2472 Try to format error messages so that they fit on lines of about @var{n}
2473 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2474 the front ends supported by GCC@. If @var{n} is zero, then no
2475 line-wrapping will be done; each error message will appear on a single
2478 @opindex fdiagnostics-show-location
2479 @item -fdiagnostics-show-location=once
2480 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2481 reporter to emit @emph{once} source location information; that is, in
2482 case the message is too long to fit on a single physical line and has to
2483 be wrapped, the source location won't be emitted (as prefix) again,
2484 over and over, in subsequent continuation lines. This is the default
2487 @item -fdiagnostics-show-location=every-line
2488 Only meaningful in line-wrapping mode. Instructs the diagnostic
2489 messages reporter to emit the same source location information (as
2490 prefix) for physical lines that result from the process of breaking
2491 a message which is too long to fit on a single line.
2493 @item -fdiagnostics-show-option
2494 @opindex fdiagnostics-show-option
2495 This option instructs the diagnostic machinery to add text to each
2496 diagnostic emitted, which indicates which command line option directly
2497 controls that diagnostic, when such an option is known to the
2498 diagnostic machinery.
2500 @item -Wcoverage-mismatch
2501 @opindex Wcoverage-mismatch
2502 Warn if feedback profiles do not match when using the
2503 @option{-fprofile-use} option.
2504 If a source file was changed between @option{-fprofile-gen} and
2505 @option{-fprofile-use}, the files with the profile feedback can fail
2506 to match the source file and GCC can not use the profile feedback
2507 information. By default, GCC emits an error message in this case.
2508 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2509 error. GCC does not use appropriate feedback profiles, so using this
2510 option can result in poorly optimized code. This option is useful
2511 only in the case of very minor changes such as bug fixes to an
2516 @node Warning Options
2517 @section Options to Request or Suppress Warnings
2518 @cindex options to control warnings
2519 @cindex warning messages
2520 @cindex messages, warning
2521 @cindex suppressing warnings
2523 Warnings are diagnostic messages that report constructions which
2524 are not inherently erroneous but which are risky or suggest there
2525 may have been an error.
2527 You can request many specific warnings with options beginning @samp{-W},
2528 for example @option{-Wimplicit} to request warnings on implicit
2529 declarations. Each of these specific warning options also has a
2530 negative form beginning @samp{-Wno-} to turn off warnings;
2531 for example, @option{-Wno-implicit}. This manual lists only one of the
2532 two forms, whichever is not the default.
2534 The following options control the amount and kinds of warnings produced
2535 by GCC; for further, language-specific options also refer to
2536 @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect
2540 @cindex syntax checking
2542 @opindex fsyntax-only
2543 Check the code for syntax errors, but don't do anything beyond that.
2547 Issue all the warnings demanded by strict ISO C and ISO C++;
2548 reject all programs that use forbidden extensions, and some other
2549 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2550 version of the ISO C standard specified by any @option{-std} option used.
2552 Valid ISO C and ISO C++ programs should compile properly with or without
2553 this option (though a rare few will require @option{-ansi} or a
2554 @option{-std} option specifying the required version of ISO C)@. However,
2555 without this option, certain GNU extensions and traditional C and C++
2556 features are supported as well. With this option, they are rejected.
2558 @option{-pedantic} does not cause warning messages for use of the
2559 alternate keywords whose names begin and end with @samp{__}. Pedantic
2560 warnings are also disabled in the expression that follows
2561 @code{__extension__}. However, only system header files should use
2562 these escape routes; application programs should avoid them.
2563 @xref{Alternate Keywords}.
2565 Some users try to use @option{-pedantic} to check programs for strict ISO
2566 C conformance. They soon find that it does not do quite what they want:
2567 it finds some non-ISO practices, but not all---only those for which
2568 ISO C @emph{requires} a diagnostic, and some others for which
2569 diagnostics have been added.
2571 A feature to report any failure to conform to ISO C might be useful in
2572 some instances, but would require considerable additional work and would
2573 be quite different from @option{-pedantic}. We don't have plans to
2574 support such a feature in the near future.
2576 Where the standard specified with @option{-std} represents a GNU
2577 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2578 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2579 extended dialect is based. Warnings from @option{-pedantic} are given
2580 where they are required by the base standard. (It would not make sense
2581 for such warnings to be given only for features not in the specified GNU
2582 C dialect, since by definition the GNU dialects of C include all
2583 features the compiler supports with the given option, and there would be
2584 nothing to warn about.)
2586 @item -pedantic-errors
2587 @opindex pedantic-errors
2588 Like @option{-pedantic}, except that errors are produced rather than
2593 Inhibit all warning messages.
2597 Inhibit warning messages about the use of @samp{#import}.
2599 @item -Wchar-subscripts
2600 @opindex Wchar-subscripts
2601 Warn if an array subscript has type @code{char}. This is a common cause
2602 of error, as programmers often forget that this type is signed on some
2604 This warning is enabled by @option{-Wall}.
2608 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2609 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2610 This warning is enabled by @option{-Wall}.
2612 @item -Wfatal-errors
2613 @opindex Wfatal-errors
2614 This option causes the compiler to abort compilation on the first error
2615 occurred rather than trying to keep going and printing further error
2620 @opindex ffreestanding
2621 @opindex fno-builtin
2622 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2623 the arguments supplied have types appropriate to the format string
2624 specified, and that the conversions specified in the format string make
2625 sense. This includes standard functions, and others specified by format
2626 attributes (@pxref{Function Attributes}), in the @code{printf},
2627 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2628 not in the C standard) families (or other target-specific families).
2629 Which functions are checked without format attributes having been
2630 specified depends on the standard version selected, and such checks of
2631 functions without the attribute specified are disabled by
2632 @option{-ffreestanding} or @option{-fno-builtin}.
2634 The formats are checked against the format features supported by GNU
2635 libc version 2.2. These include all ISO C90 and C99 features, as well
2636 as features from the Single Unix Specification and some BSD and GNU
2637 extensions. Other library implementations may not support all these
2638 features; GCC does not support warning about features that go beyond a
2639 particular library's limitations. However, if @option{-pedantic} is used
2640 with @option{-Wformat}, warnings will be given about format features not
2641 in the selected standard version (but not for @code{strfmon} formats,
2642 since those are not in any version of the C standard). @xref{C Dialect
2643 Options,,Options Controlling C Dialect}.
2645 Since @option{-Wformat} also checks for null format arguments for
2646 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2648 @option{-Wformat} is included in @option{-Wall}. For more control over some
2649 aspects of format checking, the options @option{-Wformat-y2k},
2650 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2651 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2652 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2655 @opindex Wformat-y2k
2656 If @option{-Wformat} is specified, also warn about @code{strftime}
2657 formats which may yield only a two-digit year.
2659 @item -Wno-format-extra-args
2660 @opindex Wno-format-extra-args
2661 If @option{-Wformat} is specified, do not warn about excess arguments to a
2662 @code{printf} or @code{scanf} format function. The C standard specifies
2663 that such arguments are ignored.
2665 Where the unused arguments lie between used arguments that are
2666 specified with @samp{$} operand number specifications, normally
2667 warnings are still given, since the implementation could not know what
2668 type to pass to @code{va_arg} to skip the unused arguments. However,
2669 in the case of @code{scanf} formats, this option will suppress the
2670 warning if the unused arguments are all pointers, since the Single
2671 Unix Specification says that such unused arguments are allowed.
2673 @item -Wno-format-zero-length
2674 @opindex Wno-format-zero-length
2675 If @option{-Wformat} is specified, do not warn about zero-length formats.
2676 The C standard specifies that zero-length formats are allowed.
2678 @item -Wformat-nonliteral
2679 @opindex Wformat-nonliteral
2680 If @option{-Wformat} is specified, also warn if the format string is not a
2681 string literal and so cannot be checked, unless the format function
2682 takes its format arguments as a @code{va_list}.
2684 @item -Wformat-security
2685 @opindex Wformat-security
2686 If @option{-Wformat} is specified, also warn about uses of format
2687 functions that represent possible security problems. At present, this
2688 warns about calls to @code{printf} and @code{scanf} functions where the
2689 format string is not a string literal and there are no format arguments,
2690 as in @code{printf (foo);}. This may be a security hole if the format
2691 string came from untrusted input and contains @samp{%n}. (This is
2692 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2693 in future warnings may be added to @option{-Wformat-security} that are not
2694 included in @option{-Wformat-nonliteral}.)
2698 Enable @option{-Wformat} plus format checks not included in
2699 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2700 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2704 Warn about passing a null pointer for arguments marked as
2705 requiring a non-null value by the @code{nonnull} function attribute.
2707 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2708 can be disabled with the @option{-Wno-nonnull} option.
2710 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2712 Warn about uninitialized variables which are initialized with themselves.
2713 Note this option can only be used with the @option{-Wuninitialized} option,
2714 which in turn only works with @option{-O1} and above.
2716 For example, GCC will warn about @code{i} being uninitialized in the
2717 following snippet only when @option{-Winit-self} has been specified:
2728 @item -Wimplicit-int
2729 @opindex Wimplicit-int
2730 Warn when a declaration does not specify a type.
2731 This warning is enabled by @option{-Wall}.
2733 @item -Wimplicit-function-declaration
2734 @opindex Wimplicit-function-declaration
2735 @opindex Wno-implicit-function-declaration
2736 Give a warning whenever a function is used before being declared. In
2737 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
2738 enabled by default and it is made into an error by
2739 @option{-pedantic-errors}. This warning is also enabled by
2744 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2745 This warning is enabled by @option{-Wall}.
2749 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2750 function with external linkage, returning int, taking either zero
2751 arguments, two, or three arguments of appropriate types.
2752 This warning is enabled by @option{-Wall}.
2754 @item -Wmissing-braces
2755 @opindex Wmissing-braces
2756 Warn if an aggregate or union initializer is not fully bracketed. In
2757 the following example, the initializer for @samp{a} is not fully
2758 bracketed, but that for @samp{b} is fully bracketed.
2761 int a[2][2] = @{ 0, 1, 2, 3 @};
2762 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2765 This warning is enabled by @option{-Wall}.
2767 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2768 @opindex Wmissing-include-dirs
2769 Warn if a user-supplied include directory does not exist.
2772 @opindex Wparentheses
2773 Warn if parentheses are omitted in certain contexts, such
2774 as when there is an assignment in a context where a truth value
2775 is expected, or when operators are nested whose precedence people
2776 often get confused about.
2778 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2779 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2780 interpretation from that of ordinary mathematical notation.
2782 Also warn about constructions where there may be confusion to which
2783 @code{if} statement an @code{else} branch belongs. Here is an example of
2798 In C/C++, every @code{else} branch belongs to the innermost possible
2799 @code{if} statement, which in this example is @code{if (b)}. This is
2800 often not what the programmer expected, as illustrated in the above
2801 example by indentation the programmer chose. When there is the
2802 potential for this confusion, GCC will issue a warning when this flag
2803 is specified. To eliminate the warning, add explicit braces around
2804 the innermost @code{if} statement so there is no way the @code{else}
2805 could belong to the enclosing @code{if}. The resulting code would
2822 This warning is enabled by @option{-Wall}.
2824 @item -Wsequence-point
2825 @opindex Wsequence-point
2826 Warn about code that may have undefined semantics because of violations
2827 of sequence point rules in the C and C++ standards.
2829 The C and C++ standards defines the order in which expressions in a C/C++
2830 program are evaluated in terms of @dfn{sequence points}, which represent
2831 a partial ordering between the execution of parts of the program: those
2832 executed before the sequence point, and those executed after it. These
2833 occur after the evaluation of a full expression (one which is not part
2834 of a larger expression), after the evaluation of the first operand of a
2835 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2836 function is called (but after the evaluation of its arguments and the
2837 expression denoting the called function), and in certain other places.
2838 Other than as expressed by the sequence point rules, the order of
2839 evaluation of subexpressions of an expression is not specified. All
2840 these rules describe only a partial order rather than a total order,
2841 since, for example, if two functions are called within one expression
2842 with no sequence point between them, the order in which the functions
2843 are called is not specified. However, the standards committee have
2844 ruled that function calls do not overlap.
2846 It is not specified when between sequence points modifications to the
2847 values of objects take effect. Programs whose behavior depends on this
2848 have undefined behavior; the C and C++ standards specify that ``Between
2849 the previous and next sequence point an object shall have its stored
2850 value modified at most once by the evaluation of an expression.
2851 Furthermore, the prior value shall be read only to determine the value
2852 to be stored.''. If a program breaks these rules, the results on any
2853 particular implementation are entirely unpredictable.
2855 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2856 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2857 diagnosed by this option, and it may give an occasional false positive
2858 result, but in general it has been found fairly effective at detecting
2859 this sort of problem in programs.
2861 The standard is worded confusingly, therefore there is some debate
2862 over the precise meaning of the sequence point rules in subtle cases.
2863 Links to discussions of the problem, including proposed formal
2864 definitions, may be found on the GCC readings page, at
2865 @w{@uref{http://gcc.gnu.org/readings.html}}.
2867 This warning is enabled by @option{-Wall} for C and C++.
2870 @opindex Wreturn-type
2871 @opindex Wno-return-type
2872 Warn whenever a function is defined with a return-type that defaults
2873 to @code{int}. Also warn about any @code{return} statement with no
2874 return-value in a function whose return-type is not @code{void}
2875 (falling off the end of the function body is considered returning
2876 without a value), and about a @code{return} statement with a
2877 expression in a function whose return-type is @code{void}.
2879 Also warn if the return type of a function has a type qualifier
2880 such as @code{const}. For ISO C such a type qualifier has no effect,
2881 since the value returned by a function is not an lvalue.
2882 For C++, the warning is only emitted for scalar types or @code{void}.
2883 ISO C prohibits qualified @code{void} return types on function
2884 definitions, so such return types always receive a warning
2885 even without this option.
2887 For C++, a function without return type always produces a diagnostic
2888 message, even when @option{-Wno-return-type} is specified. The only
2889 exceptions are @samp{main} and functions defined in system headers.
2891 This warning is enabled by @option{-Wall}.
2895 Warn whenever a @code{switch} statement has an index of enumerated type
2896 and lacks a @code{case} for one or more of the named codes of that
2897 enumeration. (The presence of a @code{default} label prevents this
2898 warning.) @code{case} labels outside the enumeration range also
2899 provoke warnings when this option is used.
2900 This warning is enabled by @option{-Wall}.
2902 @item -Wswitch-default
2903 @opindex Wswitch-switch
2904 Warn whenever a @code{switch} statement does not have a @code{default}
2908 @opindex Wswitch-enum
2909 Warn whenever a @code{switch} statement has an index of enumerated type
2910 and lacks a @code{case} for one or more of the named codes of that
2911 enumeration. @code{case} labels outside the enumeration range also
2912 provoke warnings when this option is used.
2916 Warn if any trigraphs are encountered that might change the meaning of
2917 the program (trigraphs within comments are not warned about).
2918 This warning is enabled by @option{-Wall}.
2920 @item -Wunused-function
2921 @opindex Wunused-function
2922 Warn whenever a static function is declared but not defined or a
2923 non-inline static function is unused.
2924 This warning is enabled by @option{-Wall}.
2926 @item -Wunused-label
2927 @opindex Wunused-label
2928 Warn whenever a label is declared but not used.
2929 This warning is enabled by @option{-Wall}.
2931 To suppress this warning use the @samp{unused} attribute
2932 (@pxref{Variable Attributes}).
2934 @item -Wunused-parameter
2935 @opindex Wunused-parameter
2936 Warn whenever a function parameter is unused aside from its declaration.
2938 To suppress this warning use the @samp{unused} attribute
2939 (@pxref{Variable Attributes}).
2941 @item -Wunused-variable
2942 @opindex Wunused-variable
2943 Warn whenever a local variable or non-constant static variable is unused
2944 aside from its declaration.
2945 This warning is enabled by @option{-Wall}.
2947 To suppress this warning use the @samp{unused} attribute
2948 (@pxref{Variable Attributes}).
2950 @item -Wunused-value
2951 @opindex Wunused-value
2952 Warn whenever a statement computes a result that is explicitly not
2953 used. To suppress this warning cast the unused expression to
2954 @samp{void}. This includes an expression-statement or the left-hand
2955 side of a comma expression that contains no side effects. For example,
2956 an expression such as @samp{x[i,j]} will cause a warning, while
2957 @samp{x[(void)i,j]} will not.
2959 This warning is enabled by @option{-Wall}.
2963 All the above @option{-Wunused} options combined.
2965 In order to get a warning about an unused function parameter, you must
2966 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2967 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2969 @item -Wuninitialized
2970 @opindex Wuninitialized
2971 Warn if an automatic variable is used without first being initialized or
2972 if a variable may be clobbered by a @code{setjmp} call.
2974 These warnings are possible only in optimizing compilation,
2975 because they require data flow information that is computed only
2976 when optimizing. If you do not specify @option{-O}, you will not get
2977 these warnings. Instead, GCC will issue a warning about @option{-Wuninitialized}
2978 requiring @option{-O}.
2980 If you want to warn about code which uses the uninitialized value of the
2981 variable in its own initializer, use the @option{-Winit-self} option.
2983 These warnings occur for individual uninitialized or clobbered
2984 elements of structure, union or array variables as well as for
2985 variables which are uninitialized or clobbered as a whole. They do
2986 not occur for variables or elements declared @code{volatile}. Because
2987 these warnings depend on optimization, the exact variables or elements
2988 for which there are warnings will depend on the precise optimization
2989 options and version of GCC used.
2991 Note that there may be no warning about a variable that is used only
2992 to compute a value that itself is never used, because such
2993 computations may be deleted by data flow analysis before the warnings
2996 These warnings are made optional because GCC is not smart
2997 enough to see all the reasons why the code might be correct
2998 despite appearing to have an error. Here is one example of how
3019 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3020 always initialized, but GCC doesn't know this. Here is
3021 another common case:
3026 if (change_y) save_y = y, y = new_y;
3028 if (change_y) y = save_y;
3033 This has no bug because @code{save_y} is used only if it is set.
3035 @cindex @code{longjmp} warnings
3036 This option also warns when a non-volatile automatic variable might be
3037 changed by a call to @code{longjmp}. These warnings as well are possible
3038 only in optimizing compilation.
3040 The compiler sees only the calls to @code{setjmp}. It cannot know
3041 where @code{longjmp} will be called; in fact, a signal handler could
3042 call it at any point in the code. As a result, you may get a warning
3043 even when there is in fact no problem because @code{longjmp} cannot
3044 in fact be called at the place which would cause a problem.
3046 Some spurious warnings can be avoided if you declare all the functions
3047 you use that never return as @code{noreturn}. @xref{Function
3050 This warning is enabled by @option{-Wall}.
3052 @item -Wunknown-pragmas
3053 @opindex Wunknown-pragmas
3054 @cindex warning for unknown pragmas
3055 @cindex unknown pragmas, warning
3056 @cindex pragmas, warning of unknown
3057 Warn when a #pragma directive is encountered which is not understood by
3058 GCC@. If this command line option is used, warnings will even be issued
3059 for unknown pragmas in system header files. This is not the case if
3060 the warnings were only enabled by the @option{-Wall} command line option.
3063 @opindex Wno-pragmas
3065 Do not warn about misuses of pragmas, such as incorrect parameters,
3066 invalid syntax, or conflicts between pragmas. See also
3067 @samp{-Wunknown-pragmas}.
3069 @item -Wstrict-aliasing
3070 @opindex Wstrict-aliasing
3071 This option is only active when @option{-fstrict-aliasing} is active.
3072 It warns about code which might break the strict aliasing rules that the
3073 compiler is using for optimization. The warning does not catch all
3074 cases, but does attempt to catch the more common pitfalls. It is
3075 included in @option{-Wall}.
3076 It is equivalent to -Wstrict-aliasing=3
3078 @item -Wstrict-aliasing=n
3079 @opindex Wstrict-aliasing=n
3080 This option is only active when @option{-fstrict-aliasing} is active.
3081 It warns about code which might break the strict aliasing rules that the
3082 compiler is using for optimization.
3083 Higher levels correspond to higher accuracy (fewer false positives).
3084 Higher levels also correspond to more effort, similar to the way -O works.
3085 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3088 Level 1: Most aggressive, quick, least accurate.
3089 Possibly useful when higher levels
3090 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3091 false negatives. However, it has many false positives.
3092 Warns for all pointer conversions between possibly incompatible types,
3093 even if never dereferenced. Runs in the frontend only.
3095 Level 2: Aggressive, quick, not too precise.
3096 May still have many false positives (not as many as level 1 though),
3097 and few false negatives (but possibly more than level 1).
3098 Unlike level 1, it only warns when an address is taken. Warns about
3099 incomplete types. Runs in the frontend only.
3101 Level 3 (default for @option{-Wstrict-aliasing}):
3102 Should have very few false positives and few false
3103 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3104 Takes care of the common punn+dereference pattern in the frontend:
3105 @code{*(int*)&some_float}.
3106 If optimization is enabled, it also runs in the backend, where it deals
3107 with multiple statement cases using flow-sensitive points-to information.
3108 Only warns when the converted pointer is dereferenced.
3109 Does not warn about incomplete types.
3111 @item -Wstrict-overflow
3112 @item -Wstrict-overflow=@var{n}
3113 @opindex Wstrict-overflow
3114 This option is only active when @option{-fstrict-overflow} is active.
3115 It warns about cases where the compiler optimizes based on the
3116 assumption that signed overflow does not occur. Note that it does not
3117 warn about all cases where the code might overflow: it only warns
3118 about cases where the compiler implements some optimization. Thus
3119 this warning depends on the optimization level.
3121 An optimization which assumes that signed overflow does not occur is
3122 perfectly safe if the values of the variables involved are such that
3123 overflow never does, in fact, occur. Therefore this warning can
3124 easily give a false positive: a warning about code which is not
3125 actually a problem. To help focus on important issues, several
3126 warning levels are defined. No warnings are issued for the use of
3127 undefined signed overflow when estimating how many iterations a loop
3128 will require, in particular when determining whether a loop will be
3132 @item -Wstrict-overflow=1
3133 Warn about cases which are both questionable and easy to avoid. For
3134 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3135 compiler will simplify this to @code{1}. This level of
3136 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3137 are not, and must be explicitly requested.
3139 @item -Wstrict-overflow=2
3140 Also warn about other cases where a comparison is simplified to a
3141 constant. For example: @code{abs (x) >= 0}. This can only be
3142 simplified when @option{-fstrict-overflow} is in effect, because
3143 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3144 zero. @option{-Wstrict-overflow} (with no level) is the same as
3145 @option{-Wstrict-overflow=2}.
3147 @item -Wstrict-overflow=3
3148 Also warn about other cases where a comparison is simplified. For
3149 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3151 @item -Wstrict-overflow=4
3152 Also warn about other simplifications not covered by the above cases.
3153 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3155 @item -Wstrict-overflow=5
3156 Also warn about cases where the compiler reduces the magnitude of a
3157 constant involved in a comparison. For example: @code{x + 2 > y} will
3158 be simplified to @code{x + 1 >= y}. This is reported only at the
3159 highest warning level because this simplification applies to many
3160 comparisons, so this warning level will give a very large number of
3164 @item -Warray-bounds
3165 @opindex Wno-array-bounds
3166 @opindex Warray-bounds
3167 This option is only active when @option{-ftree-vrp} is active
3168 (default for -O2 and above). It warns about subscripts to arrays
3169 that are always out of bounds. This warning is enabled by @option{-Wall}.
3173 All of the above @samp{-W} options combined. This enables all the
3174 warnings about constructions that some users consider questionable, and
3175 that are easy to avoid (or modify to prevent the warning), even in
3176 conjunction with macros. This also enables some language-specific
3177 warnings described in @ref{C++ Dialect Options} and
3178 @ref{Objective-C and Objective-C++ Dialect Options}.
3181 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
3182 Some of them warn about constructions that users generally do not
3183 consider questionable, but which occasionally you might wish to check
3184 for; others warn about constructions that are necessary or hard to avoid
3185 in some cases, and there is no simple way to modify the code to suppress
3192 (This option used to be called @option{-W}. The older name is still
3193 supported, but the newer name is more descriptive.) Print extra warning
3194 messages for these events:
3198 Warn if a comparison is always true or always false due to the limited
3199 range of the data type, but do not warn for constant expressions. For
3200 example, warn if an unsigned variable is compared against zero with
3201 @samp{<} or @samp{>=}. This warning can be independently controlled
3202 by @option{-Wtype-limits}.
3205 Storage-class specifiers like @code{static} are not the first things
3206 in a declaration. According to the C Standard, this usage is
3207 obsolescent. This warning can be independently controlled by
3208 @option{-Wold-style-declaration}.
3211 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
3215 A comparison between signed and unsigned values could produce an
3216 incorrect result when the signed value is converted to unsigned.
3217 (But don't warn if @option{-Wno-sign-compare} is also specified.)
3220 An aggregate has an initializer which does not initialize all members.
3221 This warning can be independently controlled by
3222 @option{-Wmissing-field-initializers}.
3225 An initialized field without side effects is overridden when using
3226 designated initializers (@pxref{Designated Inits, , Designated
3227 Initializers}). This warning can be independently controlled by
3228 @option{-Woverride-init}.
3231 A function parameter is declared without a type specifier in K&R-style
3232 functions. This warning can be independently controlled by
3233 @option{-Wmissing-parameter-type}.
3236 An empty body occurs in an @samp{if}, @samp{else} or
3237 @samp{do while} statement. This warning can be independently
3238 controlled by @option{-Wempty-body}.
3240 @item @r{(C++ only)}
3241 An empty body occurs in a @samp{while} or @samp{for} statement with no
3242 whitespacing before the semicolon. This warning can be independently
3243 controlled by @option{-Wempty-body}.
3246 A pointer is compared against integer zero with @samp{<}, @samp{<=},
3247 @samp{>}, or @samp{>=}.
3250 A variable might be changed by @samp{longjmp} or @samp{vfork}.
3251 This warning can be independently controlled by @option{-Wclobbered}.
3253 @item @r{(C++ only)}
3254 An enumerator and a non-enumerator both appear in a conditional expression.
3256 @item @r{(C++ only)}
3257 A non-static reference or non-static @samp{const} member appears in a
3258 class without constructors.
3260 @item @r{(C++ only)}
3261 Ambiguous virtual bases.
3263 @item @r{(C++ only)}
3264 Subscripting an array which has been declared @samp{register}.
3266 @item @r{(C++ only)}
3267 Taking the address of a variable which has been declared @samp{register}.
3269 @item @r{(C++ only)}
3270 A base class is not initialized in a derived class' copy constructor.
3273 @item -Wno-div-by-zero
3274 @opindex Wno-div-by-zero
3275 @opindex Wdiv-by-zero
3276 Do not warn about compile-time integer division by zero. Floating point
3277 division by zero is not warned about, as it can be a legitimate way of
3278 obtaining infinities and NaNs.
3280 @item -Wsystem-headers
3281 @opindex Wsystem-headers
3282 @cindex warnings from system headers
3283 @cindex system headers, warnings from
3284 Print warning messages for constructs found in system header files.
3285 Warnings from system headers are normally suppressed, on the assumption
3286 that they usually do not indicate real problems and would only make the
3287 compiler output harder to read. Using this command line option tells
3288 GCC to emit warnings from system headers as if they occurred in user
3289 code. However, note that using @option{-Wall} in conjunction with this
3290 option will @emph{not} warn about unknown pragmas in system
3291 headers---for that, @option{-Wunknown-pragmas} must also be used.
3294 @opindex Wfloat-equal
3295 Warn if floating point values are used in equality comparisons.
3297 The idea behind this is that sometimes it is convenient (for the
3298 programmer) to consider floating-point values as approximations to
3299 infinitely precise real numbers. If you are doing this, then you need
3300 to compute (by analyzing the code, or in some other way) the maximum or
3301 likely maximum error that the computation introduces, and allow for it
3302 when performing comparisons (and when producing output, but that's a
3303 different problem). In particular, instead of testing for equality, you
3304 would check to see whether the two values have ranges that overlap; and
3305 this is done with the relational operators, so equality comparisons are
3308 @item -Wtraditional @r{(C only)}
3309 @opindex Wtraditional
3310 Warn about certain constructs that behave differently in traditional and
3311 ISO C@. Also warn about ISO C constructs that have no traditional C
3312 equivalent, and/or problematic constructs which should be avoided.
3316 Macro parameters that appear within string literals in the macro body.
3317 In traditional C macro replacement takes place within string literals,
3318 but does not in ISO C@.
3321 In traditional C, some preprocessor directives did not exist.
3322 Traditional preprocessors would only consider a line to be a directive
3323 if the @samp{#} appeared in column 1 on the line. Therefore
3324 @option{-Wtraditional} warns about directives that traditional C
3325 understands but would ignore because the @samp{#} does not appear as the
3326 first character on the line. It also suggests you hide directives like
3327 @samp{#pragma} not understood by traditional C by indenting them. Some
3328 traditional implementations would not recognize @samp{#elif}, so it
3329 suggests avoiding it altogether.
3332 A function-like macro that appears without arguments.
3335 The unary plus operator.
3338 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3339 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3340 constants.) Note, these suffixes appear in macros defined in the system
3341 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3342 Use of these macros in user code might normally lead to spurious
3343 warnings, however GCC's integrated preprocessor has enough context to
3344 avoid warning in these cases.
3347 A function declared external in one block and then used after the end of
3351 A @code{switch} statement has an operand of type @code{long}.
3354 A non-@code{static} function declaration follows a @code{static} one.
3355 This construct is not accepted by some traditional C compilers.
3358 The ISO type of an integer constant has a different width or
3359 signedness from its traditional type. This warning is only issued if
3360 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3361 typically represent bit patterns, are not warned about.
3364 Usage of ISO string concatenation is detected.
3367 Initialization of automatic aggregates.
3370 Identifier conflicts with labels. Traditional C lacks a separate
3371 namespace for labels.
3374 Initialization of unions. If the initializer is zero, the warning is
3375 omitted. This is done under the assumption that the zero initializer in
3376 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3377 initializer warnings and relies on default initialization to zero in the
3381 Conversions by prototypes between fixed/floating point values and vice
3382 versa. The absence of these prototypes when compiling with traditional
3383 C would cause serious problems. This is a subset of the possible
3384 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3387 Use of ISO C style function definitions. This warning intentionally is
3388 @emph{not} issued for prototype declarations or variadic functions
3389 because these ISO C features will appear in your code when using
3390 libiberty's traditional C compatibility macros, @code{PARAMS} and
3391 @code{VPARAMS}. This warning is also bypassed for nested functions
3392 because that feature is already a GCC extension and thus not relevant to
3393 traditional C compatibility.
3396 @item -Wtraditional-conversion @r{(C only)}
3397 @opindex Wtraditional-conversion
3398 Warn if a prototype causes a type conversion that is different from what
3399 would happen to the same argument in the absence of a prototype. This
3400 includes conversions of fixed point to floating and vice versa, and
3401 conversions changing the width or signedness of a fixed point argument
3402 except when the same as the default promotion.
3404 @item -Wdeclaration-after-statement @r{(C only)}
3405 @opindex Wdeclaration-after-statement
3406 Warn when a declaration is found after a statement in a block. This
3407 construct, known from C++, was introduced with ISO C99 and is by default
3408 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3409 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3413 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3415 @item -Wno-endif-labels
3416 @opindex Wno-endif-labels
3417 @opindex Wendif-labels
3418 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3422 Warn whenever a local variable shadows another local variable, parameter or
3423 global variable or whenever a built-in function is shadowed.
3425 @item -Wlarger-than-@var{len}
3426 @opindex Wlarger-than
3427 Warn whenever an object of larger than @var{len} bytes is defined.
3429 @item -Wunsafe-loop-optimizations
3430 @opindex Wunsafe-loop-optimizations
3431 Warn if the loop cannot be optimized because the compiler could not
3432 assume anything on the bounds of the loop indices. With
3433 @option{-funsafe-loop-optimizations} warn if the compiler made
3436 @item -Wpointer-arith
3437 @opindex Wpointer-arith
3438 Warn about anything that depends on the ``size of'' a function type or
3439 of @code{void}. GNU C assigns these types a size of 1, for
3440 convenience in calculations with @code{void *} pointers and pointers
3441 to functions. In C++, warn also when an arithmetic operation involves
3442 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3445 @opindex Wtype-limits
3446 @opindex Wno-type-limits
3447 Warn if a comparison is always true or always false due to the limited
3448 range of the data type, but do not warn for constant expressions. For
3449 example, warn if an unsigned variable is compared against zero with
3450 @samp{<} or @samp{>=}. This warning is also enabled by
3453 @item -Wbad-function-cast @r{(C only)}
3454 @opindex Wbad-function-cast
3455 Warn whenever a function call is cast to a non-matching type.
3456 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3459 Warn about ISO C constructs that are outside of the common subset of
3460 ISO C and ISO C++, e.g.@: request for implicit conversion from
3461 @code{void *} to a pointer to non-@code{void} type.
3463 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3464 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3465 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3466 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3470 Warn whenever a pointer is cast so as to remove a type qualifier from
3471 the target type. For example, warn if a @code{const char *} is cast
3472 to an ordinary @code{char *}.
3475 @opindex Wcast-align
3476 Warn whenever a pointer is cast such that the required alignment of the
3477 target is increased. For example, warn if a @code{char *} is cast to
3478 an @code{int *} on machines where integers can only be accessed at
3479 two- or four-byte boundaries.
3481 @item -Wwrite-strings
3482 @opindex Wwrite-strings
3483 When compiling C, give string constants the type @code{const
3484 char[@var{length}]} so that
3485 copying the address of one into a non-@code{const} @code{char *}
3486 pointer will get a warning; when compiling C++, warn about the
3487 deprecated conversion from string literals to @code{char *}. This
3488 warning, by default, is enabled for C++ programs.
3489 These warnings will help you find at
3490 compile time code that can try to write into a string constant, but
3491 only if you have been very careful about using @code{const} in
3492 declarations and prototypes. Otherwise, it will just be a nuisance;
3493 this is why we did not make @option{-Wall} request these warnings.
3497 Warn for variables that might be changed by @samp{longjmp} or
3498 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3501 @opindex Wconversion
3502 @opindex Wno-conversion
3503 Warn for implicit conversions that may alter a value. This includes
3504 conversions between real and integer, like @code{abs (x)} when
3505 @code{x} is @code{double}; conversions between signed and unsigned,
3506 like @code{unsigned ui = -1}; and conversions to smaller types, like
3507 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3508 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3509 changed by the conversion like in @code{abs (2.0)}. Warnings about
3510 conversions between signed and unsigned integers can be disabled by
3511 using @option{-Wno-sign-conversion}.
3513 For C++, also warn for conversions between @code{NULL} and non-pointer
3514 types; confusing overload resolution for user-defined conversions; and
3515 conversions that will never use a type conversion operator:
3516 conversions to @code{void}, the same type, a base class or a reference
3517 to them. Warnings about conversions between signed and unsigned
3518 integers are disabled by default in C++ unless
3519 @option{-Wsign-conversion} is explicitly enabled.
3522 @opindex Wempty-body
3523 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3524 while} statement. Additionally, in C++, warn when an empty body occurs
3525 in a @samp{while} or @samp{for} statement with no whitespacing before
3526 the semicolon. This warning is also enabled by @option{-Wextra}.
3528 @item -Wsign-compare
3529 @opindex Wsign-compare
3530 @cindex warning for comparison of signed and unsigned values
3531 @cindex comparison of signed and unsigned values, warning
3532 @cindex signed and unsigned values, comparison warning
3533 Warn when a comparison between signed and unsigned values could produce
3534 an incorrect result when the signed value is converted to unsigned.
3535 This warning is also enabled by @option{-Wextra}; to get the other warnings
3536 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3538 @item -Wsign-conversion
3539 @opindex Wsign-conversion
3540 @opindex Wno-sign-conversion
3541 Warn for implicit conversions that may change the sign of an integer
3542 value, like assigning a signed integer expression to an unsigned
3543 integer variable. An explicit cast silences the warning. In C, this
3544 option is enabled also by @option{-Wconversion}.
3548 @opindex Wno-address
3549 Warn about suspicious uses of memory addresses. These include using
3550 the address of a function in a conditional expression, such as
3551 @code{void func(void); if (func)}, and comparisons against the memory
3552 address of a string literal, such as @code{if (x == "abc")}. Such
3553 uses typically indicate a programmer error: the address of a function
3554 always evaluates to true, so their use in a conditional usually
3555 indicate that the programmer forgot the parentheses in a function
3556 call; and comparisons against string literals result in unspecified
3557 behavior and are not portable in C, so they usually indicate that the
3558 programmer intended to use @code{strcmp}. This warning is enabled by
3562 @opindex Wlogical-op
3563 @opindex Wno-logical-op
3564 Warn about suspicious uses of logical operators in expressions.
3565 This includes using logical operators in contexts where a
3566 bit-wise operator is likely to be expected.
3568 @item -Waggregate-return
3569 @opindex Waggregate-return
3570 Warn if any functions that return structures or unions are defined or
3571 called. (In languages where you can return an array, this also elicits
3574 @item -Wno-attributes
3575 @opindex Wno-attributes
3576 @opindex Wattributes
3577 Do not warn if an unexpected @code{__attribute__} is used, such as
3578 unrecognized attributes, function attributes applied to variables,
3579 etc. This will not stop errors for incorrect use of supported
3582 @item -Wstrict-prototypes @r{(C only)}
3583 @opindex Wstrict-prototypes
3584 Warn if a function is declared or defined without specifying the
3585 argument types. (An old-style function definition is permitted without
3586 a warning if preceded by a declaration which specifies the argument
3589 @item -Wold-style-declaration @r{(C only)}
3590 @opindex Wold-style-declaration
3591 Warn for obsolescent usages, according to the C Standard, in a
3592 declaration. For example, warn if storage-class specifiers like
3593 @code{static} are not the first things in a declaration. This warning
3594 is also enabled by @option{-Wextra}.
3596 @item -Wold-style-definition @r{(C only)}
3597 @opindex Wold-style-definition
3598 Warn if an old-style function definition is used. A warning is given
3599 even if there is a previous prototype.
3601 @item -Wmissing-parameter-type @r{(C only)}
3602 @opindex Wmissing-parameter-type
3603 A function parameter is declared without a type specifier in K&R-style
3610 This warning is also enabled by @option{-Wextra}.
3612 @item -Wmissing-prototypes @r{(C only)}
3613 @opindex Wmissing-prototypes
3614 Warn if a global function is defined without a previous prototype
3615 declaration. This warning is issued even if the definition itself
3616 provides a prototype. The aim is to detect global functions that fail
3617 to be declared in header files.
3619 @item -Wmissing-declarations @r{(C and C++ only)}
3620 @opindex Wmissing-declarations
3621 Warn if a global function is defined without a previous declaration.
3622 Do so even if the definition itself provides a prototype.
3623 Use this option to detect global functions that are not declared in
3624 header files. In C++, no warnings are issued for function templates,
3625 or for inline functions, or for functions in anonymous namespaces.
3627 @item -Wmissing-field-initializers
3628 @opindex Wmissing-field-initializers
3631 Warn if a structure's initializer has some fields missing. For
3632 example, the following code would cause such a warning, because
3633 @code{x.h} is implicitly zero:
3636 struct s @{ int f, g, h; @};
3637 struct s x = @{ 3, 4 @};
3640 This option does not warn about designated initializers, so the following
3641 modification would not trigger a warning:
3644 struct s @{ int f, g, h; @};
3645 struct s x = @{ .f = 3, .g = 4 @};
3648 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3649 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3651 @item -Wmissing-noreturn
3652 @opindex Wmissing-noreturn
3653 Warn about functions which might be candidates for attribute @code{noreturn}.
3654 Note these are only possible candidates, not absolute ones. Care should
3655 be taken to manually verify functions actually do not ever return before
3656 adding the @code{noreturn} attribute, otherwise subtle code generation
3657 bugs could be introduced. You will not get a warning for @code{main} in
3658 hosted C environments.
3660 @item -Wmissing-format-attribute
3661 @opindex Wmissing-format-attribute
3663 Warn about function pointers which might be candidates for @code{format}
3664 attributes. Note these are only possible candidates, not absolute ones.
3665 GCC will guess that function pointers with @code{format} attributes that
3666 are used in assignment, initialization, parameter passing or return
3667 statements should have a corresponding @code{format} attribute in the
3668 resulting type. I.e.@: the left-hand side of the assignment or
3669 initialization, the type of the parameter variable, or the return type
3670 of the containing function respectively should also have a @code{format}
3671 attribute to avoid the warning.
3673 GCC will also warn about function definitions which might be
3674 candidates for @code{format} attributes. Again, these are only
3675 possible candidates. GCC will guess that @code{format} attributes
3676 might be appropriate for any function that calls a function like
3677 @code{vprintf} or @code{vscanf}, but this might not always be the
3678 case, and some functions for which @code{format} attributes are
3679 appropriate may not be detected.
3681 @item -Wno-multichar
3682 @opindex Wno-multichar
3684 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3685 Usually they indicate a typo in the user's code, as they have
3686 implementation-defined values, and should not be used in portable code.
3688 @item -Wnormalized=<none|id|nfc|nfkc>
3689 @opindex Wnormalized
3692 @cindex character set, input normalization
3693 In ISO C and ISO C++, two identifiers are different if they are
3694 different sequences of characters. However, sometimes when characters
3695 outside the basic ASCII character set are used, you can have two
3696 different character sequences that look the same. To avoid confusion,
3697 the ISO 10646 standard sets out some @dfn{normalization rules} which
3698 when applied ensure that two sequences that look the same are turned into
3699 the same sequence. GCC can warn you if you are using identifiers which
3700 have not been normalized; this option controls that warning.
3702 There are four levels of warning that GCC supports. The default is
3703 @option{-Wnormalized=nfc}, which warns about any identifier which is
3704 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3705 recommended form for most uses.
3707 Unfortunately, there are some characters which ISO C and ISO C++ allow
3708 in identifiers that when turned into NFC aren't allowable as
3709 identifiers. That is, there's no way to use these symbols in portable
3710 ISO C or C++ and have all your identifiers in NFC.
3711 @option{-Wnormalized=id} suppresses the warning for these characters.
3712 It is hoped that future versions of the standards involved will correct
3713 this, which is why this option is not the default.
3715 You can switch the warning off for all characters by writing
3716 @option{-Wnormalized=none}. You would only want to do this if you
3717 were using some other normalization scheme (like ``D''), because
3718 otherwise you can easily create bugs that are literally impossible to see.
3720 Some characters in ISO 10646 have distinct meanings but look identical
3721 in some fonts or display methodologies, especially once formatting has
3722 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3723 LETTER N'', will display just like a regular @code{n} which has been
3724 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3725 normalization scheme to convert all these into a standard form as
3726 well, and GCC will warn if your code is not in NFKC if you use
3727 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3728 about every identifier that contains the letter O because it might be
3729 confused with the digit 0, and so is not the default, but may be
3730 useful as a local coding convention if the programming environment is
3731 unable to be fixed to display these characters distinctly.
3733 @item -Wno-deprecated-declarations
3734 @opindex Wno-deprecated-declarations
3735 Do not warn about uses of functions (@pxref{Function Attributes}),
3736 variables (@pxref{Variable Attributes}), and types (@pxref{Type
3737 Attributes}) marked as deprecated by using the @code{deprecated}
3741 @opindex Wno-overflow
3742 Do not warn about compile-time overflow in constant expressions.
3744 @item -Woverride-init
3745 @opindex Woverride-init
3748 Warn if an initialized field without side effects is overridden when
3749 using designated initializers (@pxref{Designated Inits, , Designated
3752 This warning is included in @option{-Wextra}. To get other
3753 @option{-Wextra} warnings without this one, use @samp{-Wextra
3754 -Wno-override-init}.
3758 Warn if a structure is given the packed attribute, but the packed
3759 attribute has no effect on the layout or size of the structure.
3760 Such structures may be mis-aligned for little benefit. For
3761 instance, in this code, the variable @code{f.x} in @code{struct bar}
3762 will be misaligned even though @code{struct bar} does not itself
3763 have the packed attribute:
3770 @} __attribute__((packed));
3780 Warn if padding is included in a structure, either to align an element
3781 of the structure or to align the whole structure. Sometimes when this
3782 happens it is possible to rearrange the fields of the structure to
3783 reduce the padding and so make the structure smaller.
3785 @item -Wredundant-decls
3786 @opindex Wredundant-decls
3787 Warn if anything is declared more than once in the same scope, even in
3788 cases where multiple declaration is valid and changes nothing.
3790 @item -Wnested-externs @r{(C only)}
3791 @opindex Wnested-externs
3792 Warn if an @code{extern} declaration is encountered within a function.
3794 @item -Wunreachable-code
3795 @opindex Wunreachable-code
3796 Warn if the compiler detects that code will never be executed.
3798 This option is intended to warn when the compiler detects that at
3799 least a whole line of source code will never be executed, because
3800 some condition is never satisfied or because it is after a
3801 procedure that never returns.
3803 It is possible for this option to produce a warning even though there
3804 are circumstances under which part of the affected line can be executed,
3805 so care should be taken when removing apparently-unreachable code.
3807 For instance, when a function is inlined, a warning may mean that the
3808 line is unreachable in only one inlined copy of the function.
3810 This option is not made part of @option{-Wall} because in a debugging
3811 version of a program there is often substantial code which checks
3812 correct functioning of the program and is, hopefully, unreachable
3813 because the program does work. Another common use of unreachable
3814 code is to provide behavior which is selectable at compile-time.
3818 Warn if a function can not be inlined and it was declared as inline.
3819 Even with this option, the compiler will not warn about failures to
3820 inline functions declared in system headers.
3822 The compiler uses a variety of heuristics to determine whether or not
3823 to inline a function. For example, the compiler takes into account
3824 the size of the function being inlined and the amount of inlining
3825 that has already been done in the current function. Therefore,
3826 seemingly insignificant changes in the source program can cause the
3827 warnings produced by @option{-Winline} to appear or disappear.
3829 @item -Wno-invalid-offsetof @r{(C++ only)}
3830 @opindex Wno-invalid-offsetof
3831 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3832 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3833 to a non-POD type is undefined. In existing C++ implementations,
3834 however, @samp{offsetof} typically gives meaningful results even when
3835 applied to certain kinds of non-POD types. (Such as a simple
3836 @samp{struct} that fails to be a POD type only by virtue of having a
3837 constructor.) This flag is for users who are aware that they are
3838 writing nonportable code and who have deliberately chosen to ignore the
3841 The restrictions on @samp{offsetof} may be relaxed in a future version
3842 of the C++ standard.
3844 @item -Wno-int-to-pointer-cast @r{(C only)}
3845 @opindex Wno-int-to-pointer-cast
3846 Suppress warnings from casts to pointer type of an integer of a
3849 @item -Wno-pointer-to-int-cast @r{(C only)}
3850 @opindex Wno-pointer-to-int-cast
3851 Suppress warnings from casts from a pointer to an integer type of a
3855 @opindex Winvalid-pch
3856 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3857 the search path but can't be used.
3861 @opindex Wno-long-long
3862 Warn if @samp{long long} type is used. This is default. To inhibit
3863 the warning messages, use @option{-Wno-long-long}. Flags
3864 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3865 only when @option{-pedantic} flag is used.
3867 @item -Wvariadic-macros
3868 @opindex Wvariadic-macros
3869 @opindex Wno-variadic-macros
3870 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3871 alternate syntax when in pedantic ISO C99 mode. This is default.
3872 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3877 Warn if variable length array is used in the code.
3878 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
3879 the variable length array.
3881 @item -Wvolatile-register-var
3882 @opindex Wvolatile-register-var
3883 @opindex Wno-volatile-register-var
3884 Warn if a register variable is declared volatile. The volatile
3885 modifier does not inhibit all optimizations that may eliminate reads
3886 and/or writes to register variables.
3888 @item -Wdisabled-optimization
3889 @opindex Wdisabled-optimization
3890 Warn if a requested optimization pass is disabled. This warning does
3891 not generally indicate that there is anything wrong with your code; it
3892 merely indicates that GCC's optimizers were unable to handle the code
3893 effectively. Often, the problem is that your code is too big or too
3894 complex; GCC will refuse to optimize programs when the optimization
3895 itself is likely to take inordinate amounts of time.
3897 @item -Wpointer-sign
3898 @opindex Wpointer-sign
3899 @opindex Wno-pointer-sign
3900 Warn for pointer argument passing or assignment with different signedness.
3901 This option is only supported for C and Objective-C@. It is implied by
3902 @option{-Wall} and by @option{-pedantic}, which can be disabled with
3903 @option{-Wno-pointer-sign}.
3907 Make all warnings into errors.
3911 Make the specified warning into an errors. The specifier for a
3912 warning is appended, for example @option{-Werror=switch} turns the
3913 warnings controlled by @option{-Wswitch} into errors. This switch
3914 takes a negative form, to be used to negate @option{-Werror} for
3915 specific warnings, for example @option{-Wno-error=switch} makes
3916 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
3917 is in effect. You can use the @option{-fdiagnostics-show-option}
3918 option to have each controllable warning amended with the option which
3919 controls it, to determine what to use with this option.
3921 Note that specifying @option{-Werror=}@var{foo} automatically implies
3922 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
3925 @item -Wstack-protector
3926 @opindex Wstack-protector
3927 This option is only active when @option{-fstack-protector} is active. It
3928 warns about functions that will not be protected against stack smashing.
3930 @item -Woverlength-strings
3931 @opindex Woverlength-strings
3932 Warn about string constants which are longer than the ``minimum
3933 maximum'' length specified in the C standard. Modern compilers
3934 generally allow string constants which are much longer than the
3935 standard's minimum limit, but very portable programs should avoid
3936 using longer strings.
3938 The limit applies @emph{after} string constant concatenation, and does
3939 not count the trailing NUL@. In C89, the limit was 509 characters; in
3940 C99, it was raised to 4095. C++98 does not specify a normative
3941 minimum maximum, so we do not diagnose overlength strings in C++@.
3943 This option is implied by @option{-pedantic}, and can be disabled with
3944 @option{-Wno-overlength-strings}.
3947 @node Debugging Options
3948 @section Options for Debugging Your Program or GCC
3949 @cindex options, debugging
3950 @cindex debugging information options
3952 GCC has various special options that are used for debugging
3953 either your program or GCC:
3958 Produce debugging information in the operating system's native format
3959 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
3962 On most systems that use stabs format, @option{-g} enables use of extra
3963 debugging information that only GDB can use; this extra information
3964 makes debugging work better in GDB but will probably make other debuggers
3966 refuse to read the program. If you want to control for certain whether
3967 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3968 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3970 GCC allows you to use @option{-g} with
3971 @option{-O}. The shortcuts taken by optimized code may occasionally
3972 produce surprising results: some variables you declared may not exist
3973 at all; flow of control may briefly move where you did not expect it;
3974 some statements may not be executed because they compute constant
3975 results or their values were already at hand; some statements may
3976 execute in different places because they were moved out of loops.
3978 Nevertheless it proves possible to debug optimized output. This makes
3979 it reasonable to use the optimizer for programs that might have bugs.
3981 The following options are useful when GCC is generated with the
3982 capability for more than one debugging format.
3986 Produce debugging information for use by GDB@. This means to use the
3987 most expressive format available (DWARF 2, stabs, or the native format
3988 if neither of those are supported), including GDB extensions if at all
3993 Produce debugging information in stabs format (if that is supported),
3994 without GDB extensions. This is the format used by DBX on most BSD
3995 systems. On MIPS, Alpha and System V Release 4 systems this option
3996 produces stabs debugging output which is not understood by DBX or SDB@.
3997 On System V Release 4 systems this option requires the GNU assembler.
3999 @item -feliminate-unused-debug-symbols
4000 @opindex feliminate-unused-debug-symbols
4001 Produce debugging information in stabs format (if that is supported),
4002 for only symbols that are actually used.
4004 @item -femit-class-debug-always
4005 Instead of emitting debugging information for a C++ class in only one
4006 object file, emit it in all object files using the class. This option
4007 should be used only with debuggers that are unable to handle the way GCC
4008 normally emits debugging information for classes because using this
4009 option will increase the size of debugging information by as much as a
4014 Produce debugging information in stabs format (if that is supported),
4015 using GNU extensions understood only by the GNU debugger (GDB)@. The
4016 use of these extensions is likely to make other debuggers crash or
4017 refuse to read the program.
4021 Produce debugging information in COFF format (if that is supported).
4022 This is the format used by SDB on most System V systems prior to
4027 Produce debugging information in XCOFF format (if that is supported).
4028 This is the format used by the DBX debugger on IBM RS/6000 systems.
4032 Produce debugging information in XCOFF format (if that is supported),
4033 using GNU extensions understood only by the GNU debugger (GDB)@. The
4034 use of these extensions is likely to make other debuggers crash or
4035 refuse to read the program, and may cause assemblers other than the GNU
4036 assembler (GAS) to fail with an error.
4040 Produce debugging information in DWARF version 2 format (if that is
4041 supported). This is the format used by DBX on IRIX 6. With this
4042 option, GCC uses features of DWARF version 3 when they are useful;
4043 version 3 is upward compatible with version 2, but may still cause
4044 problems for older debuggers.
4048 Produce debugging information in VMS debug format (if that is
4049 supported). This is the format used by DEBUG on VMS systems.
4052 @itemx -ggdb@var{level}
4053 @itemx -gstabs@var{level}
4054 @itemx -gcoff@var{level}
4055 @itemx -gxcoff@var{level}
4056 @itemx -gvms@var{level}
4057 Request debugging information and also use @var{level} to specify how
4058 much information. The default level is 2.
4060 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4063 Level 1 produces minimal information, enough for making backtraces in
4064 parts of the program that you don't plan to debug. This includes
4065 descriptions of functions and external variables, but no information
4066 about local variables and no line numbers.
4068 Level 3 includes extra information, such as all the macro definitions
4069 present in the program. Some debuggers support macro expansion when
4070 you use @option{-g3}.
4072 @option{-gdwarf-2} does not accept a concatenated debug level, because
4073 GCC used to support an option @option{-gdwarf} that meant to generate
4074 debug information in version 1 of the DWARF format (which is very
4075 different from version 2), and it would have been too confusing. That
4076 debug format is long obsolete, but the option cannot be changed now.
4077 Instead use an additional @option{-g@var{level}} option to change the
4078 debug level for DWARF2.
4080 @item -feliminate-dwarf2-dups
4081 @opindex feliminate-dwarf2-dups
4082 Compress DWARF2 debugging information by eliminating duplicated
4083 information about each symbol. This option only makes sense when
4084 generating DWARF2 debugging information with @option{-gdwarf-2}.
4086 @item -femit-struct-debug-baseonly
4087 Emit debug information for struct-like types
4088 only when the base name of the compilation source file
4089 matches the base name of file in which the struct was defined.
4091 This option substantially reduces the size of debugging information,
4092 but at significant potential loss in type information to the debugger.
4093 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4094 See @option{-femit-struct-debug-detailed} for more detailed control.
4096 This option works only with DWARF 2.
4098 @item -femit-struct-debug-reduced
4099 Emit debug information for struct-like types
4100 only when the base name of the compilation source file
4101 matches the base name of file in which the type was defined,
4102 unless the struct is a template or defined in a system header.
4104 This option significantly reduces the size of debugging information,
4105 with some potential loss in type information to the debugger.
4106 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4107 See @option{-femit-struct-debug-detailed} for more detailed control.
4109 This option works only with DWARF 2.
4111 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4112 Specify the struct-like types
4113 for which the compiler will generate debug information.
4114 The intent is to reduce duplicate struct debug information
4115 between different object files within the same program.
4117 This option is a detailed version of
4118 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4119 which will serve for most needs.
4121 A specification has the syntax
4122 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4124 The optional first word limits the specification to
4125 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4126 A struct type is used directly when it is the type of a variable, member.
4127 Indirect uses arise through pointers to structs.
4128 That is, when use of an incomplete struct would be legal, the use is indirect.
4130 @samp{struct one direct; struct two * indirect;}.
4132 The optional second word limits the specification to
4133 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4134 Generic structs are a bit complicated to explain.
4135 For C++, these are non-explicit specializations of template classes,
4136 or non-template classes within the above.
4137 Other programming languages have generics,
4138 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4140 The third word specifies the source files for those
4141 structs for which the compiler will emit debug information.
4142 The values @samp{none} and @samp{any} have the normal meaning.
4143 The value @samp{base} means that
4144 the base of name of the file in which the type declaration appears
4145 must match the base of the name of the main compilation file.
4146 In practice, this means that
4147 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4148 but types declared in other header will not.
4149 The value @samp{sys} means those types satisfying @samp{base}
4150 or declared in system or compiler headers.
4152 You may need to experiment to determine the best settings for your application.
4154 The default is @samp{-femit-struct-debug-detailed=all}.
4156 This option works only with DWARF 2.
4158 @item -fdebug-prefix-map=@var{old}=@var{new}
4159 @opindex fdebug-prefix-map
4160 When compiling files in directory @file{@var{old}}, record debugging
4161 information describing them as in @file{@var{new}} instead.
4163 @cindex @command{prof}
4166 Generate extra code to write profile information suitable for the
4167 analysis program @command{prof}. You must use this option when compiling
4168 the source files you want data about, and you must also use it when
4171 @cindex @command{gprof}
4174 Generate extra code to write profile information suitable for the
4175 analysis program @command{gprof}. You must use this option when compiling
4176 the source files you want data about, and you must also use it when
4181 Makes the compiler print out each function name as it is compiled, and
4182 print some statistics about each pass when it finishes.
4185 @opindex ftime-report
4186 Makes the compiler print some statistics about the time consumed by each
4187 pass when it finishes.
4190 @opindex fmem-report
4191 Makes the compiler print some statistics about permanent memory
4192 allocation when it finishes.
4194 @item -fpre-ipa-mem-report
4195 @opindex fpre-ipa-mem-report
4196 @item -fpost-ipa-mem-report
4197 @opindex fpost-ipa-mem-report
4198 Makes the compiler print some statistics about permanent memory
4199 allocation before or after interprocedural optimization.
4201 @item -fprofile-arcs
4202 @opindex fprofile-arcs
4203 Add code so that program flow @dfn{arcs} are instrumented. During
4204 execution the program records how many times each branch and call is
4205 executed and how many times it is taken or returns. When the compiled
4206 program exits it saves this data to a file called
4207 @file{@var{auxname}.gcda} for each source file. The data may be used for
4208 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4209 test coverage analysis (@option{-ftest-coverage}). Each object file's
4210 @var{auxname} is generated from the name of the output file, if
4211 explicitly specified and it is not the final executable, otherwise it is
4212 the basename of the source file. In both cases any suffix is removed
4213 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4214 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4215 @xref{Cross-profiling}.
4217 @cindex @command{gcov}
4221 This option is used to compile and link code instrumented for coverage
4222 analysis. The option is a synonym for @option{-fprofile-arcs}
4223 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4224 linking). See the documentation for those options for more details.
4229 Compile the source files with @option{-fprofile-arcs} plus optimization
4230 and code generation options. For test coverage analysis, use the
4231 additional @option{-ftest-coverage} option. You do not need to profile
4232 every source file in a program.
4235 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4236 (the latter implies the former).
4239 Run the program on a representative workload to generate the arc profile
4240 information. This may be repeated any number of times. You can run
4241 concurrent instances of your program, and provided that the file system
4242 supports locking, the data files will be correctly updated. Also
4243 @code{fork} calls are detected and correctly handled (double counting
4247 For profile-directed optimizations, compile the source files again with
4248 the same optimization and code generation options plus
4249 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4250 Control Optimization}).
4253 For test coverage analysis, use @command{gcov} to produce human readable
4254 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4255 @command{gcov} documentation for further information.
4259 With @option{-fprofile-arcs}, for each function of your program GCC
4260 creates a program flow graph, then finds a spanning tree for the graph.
4261 Only arcs that are not on the spanning tree have to be instrumented: the
4262 compiler adds code to count the number of times that these arcs are
4263 executed. When an arc is the only exit or only entrance to a block, the
4264 instrumentation code can be added to the block; otherwise, a new basic
4265 block must be created to hold the instrumentation code.
4268 @item -ftest-coverage
4269 @opindex ftest-coverage
4270 Produce a notes file that the @command{gcov} code-coverage utility
4271 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4272 show program coverage. Each source file's note file is called
4273 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4274 above for a description of @var{auxname} and instructions on how to
4275 generate test coverage data. Coverage data will match the source files
4276 more closely, if you do not optimize.
4278 @item -fdbg-cnt-list
4279 @opindex fdbg-cnt-list
4280 Print the name and the counter upperbound for all debug counters.
4282 @item -fdbg-cnt=@var{counter-value-list}
4284 Set the internal debug counter upperbound. @var{counter-value-list}
4285 is a comma-separated list of @var{name}:@var{value} pairs
4286 which sets the upperbound of each debug counter @var{name} to @var{value}.
4287 All debug counters have the initial upperbound of @var{UINT_MAX},
4288 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4289 e.g. With -fdbg-cnt=dce:10,tail_call:0
4290 dbg_cnt(dce) will return true only for first 10 invocations
4291 and dbg_cnt(tail_call) will return false always.
4293 @item -d@var{letters}
4294 @item -fdump-rtl-@var{pass}
4296 Says to make debugging dumps during compilation at times specified by
4297 @var{letters}. This is used for debugging the RTL-based passes of the
4298 compiler. The file names for most of the dumps are made by appending a
4299 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
4300 from the name of the output file, if explicitly specified and it is not
4301 an executable, otherwise it is the basename of the source file. These
4302 switches may have different effects when @option{-E} is used for
4305 Most debug dumps can be enabled either passing a letter to the @option{-d}
4306 option, or with a long @option{-fdump-rtl} switch; here are the possible
4307 letters for use in @var{letters} and @var{pass}, and their meanings:
4312 Annotate the assembler output with miscellaneous debugging information.
4315 @itemx -fdump-rtl-bbro
4317 @opindex fdump-rtl-bbro
4318 Dump after block reordering, to @file{@var{file}.148r.bbro}.
4321 @itemx -fdump-rtl-combine
4323 @opindex fdump-rtl-combine
4324 Dump after the RTL instruction combination pass, to the file
4325 @file{@var{file}.129r.combine}.
4328 @itemx -fdump-rtl-ce1
4329 @itemx -fdump-rtl-ce2
4331 @opindex fdump-rtl-ce1
4332 @opindex fdump-rtl-ce2
4333 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
4334 first if conversion, to the file @file{@var{file}.117r.ce1}. @option{-dC}
4335 and @option{-fdump-rtl-ce2} enable dumping after the second if
4336 conversion, to the file @file{@var{file}.130r.ce2}.
4339 @itemx -fdump-rtl-btl
4340 @itemx -fdump-rtl-dbr
4342 @opindex fdump-rtl-btl
4343 @opindex fdump-rtl-dbr
4344 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
4345 target load optimization, to @file{@var{file}.31.btl}. @option{-dd}
4346 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
4347 scheduling, to @file{@var{file}.36.dbr}.
4351 Dump all macro definitions, at the end of preprocessing, in addition to
4355 @itemx -fdump-rtl-ce3
4357 @opindex fdump-rtl-ce3
4358 Dump after the third if conversion, to @file{@var{file}.146r.ce3}.
4361 @itemx -fdump-rtl-cfg
4362 @itemx -fdump-rtl-life
4364 @opindex fdump-rtl-cfg
4365 @opindex fdump-rtl-life
4366 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
4367 and data flow analysis, to @file{@var{file}.116r.cfg}. @option{-df}
4368 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
4369 to @file{@var{file}.128r.life1} and @file{@var{file}.135r.life2}.
4372 @itemx -fdump-rtl-greg
4374 @opindex fdump-rtl-greg
4375 Dump after global register allocation, to @file{@var{file}.139r.greg}.
4378 @itemx -fdump-rtl-gcse
4379 @itemx -fdump-rtl-bypass
4381 @opindex fdump-rtl-gcse
4382 @opindex fdump-rtl-bypass
4383 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
4384 @file{@var{file}.114r.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
4385 enable dumping after jump bypassing and control flow optimizations, to
4386 @file{@var{file}.115r.bypass}.
4389 @itemx -fdump-rtl-eh
4391 @opindex fdump-rtl-eh
4392 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
4395 @itemx -fdump-rtl-sibling
4397 @opindex fdump-rtl-sibling
4398 Dump after sibling call optimizations, to @file{@var{file}.106r.sibling}.
4401 @itemx -fdump-rtl-jump
4403 @opindex fdump-rtl-jump
4404 Dump after the first jump optimization, to @file{@var{file}.112r.jump}.
4407 @itemx -fdump-rtl-stack
4409 @opindex fdump-rtl-stack
4410 Dump after conversion from GCC's "flat register file" registers to the
4411 x87's stack-like registers, to @file{@var{file}.152r.stack}.
4414 @itemx -fdump-rtl-lreg
4416 @opindex fdump-rtl-lreg
4417 Dump after local register allocation, to @file{@var{file}.138r.lreg}.
4420 @itemx -fdump-rtl-loop2
4422 @opindex fdump-rtl-loop2
4423 @option{-dL} and @option{-fdump-rtl-loop2} enable dumping after the
4424 loop optimization pass, to @file{@var{file}.119r.loop2},
4425 @file{@var{file}.120r.loop2_init},
4426 @file{@var{file}.121r.loop2_invariant}, and
4427 @file{@var{file}.125r.loop2_done}.
4430 @itemx -fdump-rtl-sms
4432 @opindex fdump-rtl-sms
4433 Dump after modulo scheduling, to @file{@var{file}.136r.sms}.
4436 @itemx -fdump-rtl-mach
4438 @opindex fdump-rtl-mach
4439 Dump after performing the machine dependent reorganization pass, to
4440 @file{@var{file}.155r.mach} if that pass exists.
4443 @itemx -fdump-rtl-rnreg
4445 @opindex fdump-rtl-rnreg
4446 Dump after register renumbering, to @file{@var{file}.147r.rnreg}.
4449 @itemx -fdump-rtl-regmove
4451 @opindex fdump-rtl-regmove
4452 Dump after the register move pass, to @file{@var{file}.132r.regmove}.
4455 @itemx -fdump-rtl-postreload
4457 @opindex fdump-rtl-postreload
4458 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
4461 @itemx -fdump-rtl-expand
4463 @opindex fdump-rtl-expand
4464 Dump after RTL generation, to @file{@var{file}.104r.expand}.
4467 @itemx -fdump-rtl-sched2
4469 @opindex fdump-rtl-sched2
4470 Dump after the second scheduling pass, to @file{@var{file}.149r.sched2}.
4473 @itemx -fdump-rtl-cse
4475 @opindex fdump-rtl-cse
4476 Dump after CSE (including the jump optimization that sometimes follows
4477 CSE), to @file{@var{file}.113r.cse}.
4480 @itemx -fdump-rtl-sched1
4482 @opindex fdump-rtl-sched1
4483 Dump after the first scheduling pass, to @file{@var{file}.136r.sched1}.
4486 @itemx -fdump-rtl-cse2
4488 @opindex fdump-rtl-cse2
4489 Dump after the second CSE pass (including the jump optimization that
4490 sometimes follows CSE), to @file{@var{file}.127r.cse2}.
4493 @itemx -fdump-rtl-tracer
4495 @opindex fdump-rtl-tracer
4496 Dump after running tracer, to @file{@var{file}.118r.tracer}.
4499 @itemx -fdump-rtl-vpt
4500 @itemx -fdump-rtl-vartrack
4502 @opindex fdump-rtl-vpt
4503 @opindex fdump-rtl-vartrack
4504 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
4505 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
4506 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
4507 to @file{@var{file}.154r.vartrack}.
4510 @itemx -fdump-rtl-flow2
4512 @opindex fdump-rtl-flow2
4513 Dump after the second flow pass, to @file{@var{file}.142r.flow2}.
4516 @itemx -fdump-rtl-peephole2
4518 @opindex fdump-rtl-peephole2
4519 Dump after the peephole pass, to @file{@var{file}.145r.peephole2}.
4522 @itemx -fdump-rtl-web
4524 @opindex fdump-rtl-web
4525 Dump after live range splitting, to @file{@var{file}.126r.web}.
4528 @itemx -fdump-rtl-all
4530 @opindex fdump-rtl-all
4531 Produce all the dumps listed above.
4535 Produce a core dump whenever an error occurs.
4539 Print statistics on memory usage, at the end of the run, to
4544 Annotate the assembler output with a comment indicating which
4545 pattern and alternative was used. The length of each instruction is
4550 Dump the RTL in the assembler output as a comment before each instruction.
4551 Also turns on @option{-dp} annotation.
4555 For each of the other indicated dump files (either with @option{-d} or
4556 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
4557 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
4561 Just generate RTL for a function instead of compiling it. Usually used
4562 with @samp{r} (@option{-fdump-rtl-expand}).
4566 Dump debugging information during parsing, to standard error.
4570 @opindex fdump-noaddr
4571 When doing debugging dumps (see @option{-d} option above), suppress
4572 address output. This makes it more feasible to use diff on debugging
4573 dumps for compiler invocations with different compiler binaries and/or
4574 different text / bss / data / heap / stack / dso start locations.
4576 @item -fdump-unnumbered
4577 @opindex fdump-unnumbered
4578 When doing debugging dumps (see @option{-d} option above), suppress instruction
4579 numbers and address output. This makes it more feasible to
4580 use diff on debugging dumps for compiler invocations with different
4581 options, in particular with and without @option{-g}.
4583 @item -fdump-translation-unit @r{(C++ only)}
4584 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
4585 @opindex fdump-translation-unit
4586 Dump a representation of the tree structure for the entire translation
4587 unit to a file. The file name is made by appending @file{.tu} to the
4588 source file name. If the @samp{-@var{options}} form is used, @var{options}
4589 controls the details of the dump as described for the
4590 @option{-fdump-tree} options.
4592 @item -fdump-class-hierarchy @r{(C++ only)}
4593 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
4594 @opindex fdump-class-hierarchy
4595 Dump a representation of each class's hierarchy and virtual function
4596 table layout to a file. The file name is made by appending @file{.class}
4597 to the source file name. If the @samp{-@var{options}} form is used,
4598 @var{options} controls the details of the dump as described for the
4599 @option{-fdump-tree} options.
4601 @item -fdump-ipa-@var{switch}
4603 Control the dumping at various stages of inter-procedural analysis
4604 language tree to a file. The file name is generated by appending a switch
4605 specific suffix to the source file name. The following dumps are possible:
4609 Enables all inter-procedural analysis dumps; currently the only produced
4610 dump is the @samp{cgraph} dump.
4613 Dumps information about call-graph optimization, unused function removal,
4614 and inlining decisions.
4617 @item -fdump-tree-@var{switch}
4618 @itemx -fdump-tree-@var{switch}-@var{options}
4620 Control the dumping at various stages of processing the intermediate
4621 language tree to a file. The file name is generated by appending a switch
4622 specific suffix to the source file name. If the @samp{-@var{options}}
4623 form is used, @var{options} is a list of @samp{-} separated options that
4624 control the details of the dump. Not all options are applicable to all
4625 dumps, those which are not meaningful will be ignored. The following
4626 options are available
4630 Print the address of each node. Usually this is not meaningful as it
4631 changes according to the environment and source file. Its primary use
4632 is for tying up a dump file with a debug environment.
4634 Inhibit dumping of members of a scope or body of a function merely
4635 because that scope has been reached. Only dump such items when they
4636 are directly reachable by some other path. When dumping pretty-printed
4637 trees, this option inhibits dumping the bodies of control structures.
4639 Print a raw representation of the tree. By default, trees are
4640 pretty-printed into a C-like representation.
4642 Enable more detailed dumps (not honored by every dump option).
4644 Enable dumping various statistics about the pass (not honored by every dump
4647 Enable showing basic block boundaries (disabled in raw dumps).
4649 Enable showing virtual operands for every statement.
4651 Enable showing line numbers for statements.
4653 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4655 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4658 The following tree dumps are possible:
4662 Dump before any tree based optimization, to @file{@var{file}.original}.
4665 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4668 Dump after function inlining, to @file{@var{file}.inlined}.
4671 @opindex fdump-tree-gimple
4672 Dump each function before and after the gimplification pass to a file. The
4673 file name is made by appending @file{.gimple} to the source file name.
4676 @opindex fdump-tree-cfg
4677 Dump the control flow graph of each function to a file. The file name is
4678 made by appending @file{.cfg} to the source file name.
4681 @opindex fdump-tree-vcg
4682 Dump the control flow graph of each function to a file in VCG format. The
4683 file name is made by appending @file{.vcg} to the source file name. Note
4684 that if the file contains more than one function, the generated file cannot
4685 be used directly by VCG@. You will need to cut and paste each function's
4686 graph into its own separate file first.
4689 @opindex fdump-tree-ch
4690 Dump each function after copying loop headers. The file name is made by
4691 appending @file{.ch} to the source file name.
4694 @opindex fdump-tree-ssa
4695 Dump SSA related information to a file. The file name is made by appending
4696 @file{.ssa} to the source file name.
4699 @opindex fdump-tree-salias
4700 Dump structure aliasing variable information to a file. This file name
4701 is made by appending @file{.salias} to the source file name.
4704 @opindex fdump-tree-alias
4705 Dump aliasing information for each function. The file name is made by
4706 appending @file{.alias} to the source file name.
4709 @opindex fdump-tree-ccp
4710 Dump each function after CCP@. The file name is made by appending
4711 @file{.ccp} to the source file name.
4714 @opindex fdump-tree-storeccp
4715 Dump each function after STORE-CCP. The file name is made by appending
4716 @file{.storeccp} to the source file name.
4719 @opindex fdump-tree-pre
4720 Dump trees after partial redundancy elimination. The file name is made
4721 by appending @file{.pre} to the source file name.
4724 @opindex fdump-tree-fre
4725 Dump trees after full redundancy elimination. The file name is made
4726 by appending @file{.fre} to the source file name.
4729 @opindex fdump-tree-copyprop
4730 Dump trees after copy propagation. The file name is made
4731 by appending @file{.copyprop} to the source file name.
4733 @item store_copyprop
4734 @opindex fdump-tree-store_copyprop
4735 Dump trees after store copy-propagation. The file name is made
4736 by appending @file{.store_copyprop} to the source file name.
4739 @opindex fdump-tree-dce
4740 Dump each function after dead code elimination. The file name is made by
4741 appending @file{.dce} to the source file name.
4744 @opindex fdump-tree-mudflap
4745 Dump each function after adding mudflap instrumentation. The file name is
4746 made by appending @file{.mudflap} to the source file name.
4749 @opindex fdump-tree-sra
4750 Dump each function after performing scalar replacement of aggregates. The
4751 file name is made by appending @file{.sra} to the source file name.
4754 @opindex fdump-tree-sink
4755 Dump each function after performing code sinking. The file name is made
4756 by appending @file{.sink} to the source file name.
4759 @opindex fdump-tree-dom
4760 Dump each function after applying dominator tree optimizations. The file
4761 name is made by appending @file{.dom} to the source file name.
4764 @opindex fdump-tree-dse
4765 Dump each function after applying dead store elimination. The file
4766 name is made by appending @file{.dse} to the source file name.
4769 @opindex fdump-tree-phiopt
4770 Dump each function after optimizing PHI nodes into straightline code. The file
4771 name is made by appending @file{.phiopt} to the source file name.
4774 @opindex fdump-tree-forwprop
4775 Dump each function after forward propagating single use variables. The file
4776 name is made by appending @file{.forwprop} to the source file name.
4779 @opindex fdump-tree-copyrename
4780 Dump each function after applying the copy rename optimization. The file
4781 name is made by appending @file{.copyrename} to the source file name.
4784 @opindex fdump-tree-nrv
4785 Dump each function after applying the named return value optimization on
4786 generic trees. The file name is made by appending @file{.nrv} to the source
4790 @opindex fdump-tree-vect
4791 Dump each function after applying vectorization of loops. The file name is
4792 made by appending @file{.vect} to the source file name.
4795 @opindex fdump-tree-vrp
4796 Dump each function after Value Range Propagation (VRP). The file name
4797 is made by appending @file{.vrp} to the source file name.
4800 @opindex fdump-tree-all
4801 Enable all the available tree dumps with the flags provided in this option.
4804 @item -ftree-vectorizer-verbose=@var{n}
4805 @opindex ftree-vectorizer-verbose
4806 This option controls the amount of debugging output the vectorizer prints.
4807 This information is written to standard error, unless
4808 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
4809 in which case it is output to the usual dump listing file, @file{.vect}.
4810 For @var{n}=0 no diagnostic information is reported.
4811 If @var{n}=1 the vectorizer reports each loop that got vectorized,
4812 and the total number of loops that got vectorized.
4813 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
4814 the first analysis phase (vect_analyze_loop_form) - i.e. countable,
4815 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
4816 level that @option{-fdump-tree-vect-stats} uses.
4817 Higher verbosity levels mean either more information dumped for each
4818 reported loop, or same amount of information reported for more loops:
4819 If @var{n}=3, alignment related information is added to the reports.
4820 If @var{n}=4, data-references related information (e.g. memory dependences,
4821 memory access-patterns) is added to the reports.
4822 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
4823 that did not pass the first analysis phase (i.e. may not be countable, or
4824 may have complicated control-flow).
4825 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
4826 For @var{n}=7, all the information the vectorizer generates during its
4827 analysis and transformation is reported. This is the same verbosity level
4828 that @option{-fdump-tree-vect-details} uses.
4830 @item -frandom-seed=@var{string}
4831 @opindex frandom-string
4832 This option provides a seed that GCC uses when it would otherwise use
4833 random numbers. It is used to generate certain symbol names
4834 that have to be different in every compiled file. It is also used to
4835 place unique stamps in coverage data files and the object files that
4836 produce them. You can use the @option{-frandom-seed} option to produce
4837 reproducibly identical object files.
4839 The @var{string} should be different for every file you compile.
4841 @item -fsched-verbose=@var{n}
4842 @opindex fsched-verbose
4843 On targets that use instruction scheduling, this option controls the
4844 amount of debugging output the scheduler prints. This information is
4845 written to standard error, unless @option{-dS} or @option{-dR} is
4846 specified, in which case it is output to the usual dump
4847 listing file, @file{.sched} or @file{.sched2} respectively. However
4848 for @var{n} greater than nine, the output is always printed to standard
4851 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4852 same information as @option{-dRS}. For @var{n} greater than one, it
4853 also output basic block probabilities, detailed ready list information
4854 and unit/insn info. For @var{n} greater than two, it includes RTL
4855 at abort point, control-flow and regions info. And for @var{n} over
4856 four, @option{-fsched-verbose} also includes dependence info.
4860 Store the usual ``temporary'' intermediate files permanently; place them
4861 in the current directory and name them based on the source file. Thus,
4862 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4863 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4864 preprocessed @file{foo.i} output file even though the compiler now
4865 normally uses an integrated preprocessor.
4867 When used in combination with the @option{-x} command line option,
4868 @option{-save-temps} is sensible enough to avoid over writing an
4869 input source file with the same extension as an intermediate file.
4870 The corresponding intermediate file may be obtained by renaming the
4871 source file before using @option{-save-temps}.
4875 Report the CPU time taken by each subprocess in the compilation
4876 sequence. For C source files, this is the compiler proper and assembler
4877 (plus the linker if linking is done). The output looks like this:
4884 The first number on each line is the ``user time'', that is time spent
4885 executing the program itself. The second number is ``system time'',
4886 time spent executing operating system routines on behalf of the program.
4887 Both numbers are in seconds.
4889 @item -fvar-tracking
4890 @opindex fvar-tracking
4891 Run variable tracking pass. It computes where variables are stored at each
4892 position in code. Better debugging information is then generated
4893 (if the debugging information format supports this information).
4895 It is enabled by default when compiling with optimization (@option{-Os},
4896 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
4897 the debug info format supports it.
4899 @item -print-file-name=@var{library}
4900 @opindex print-file-name
4901 Print the full absolute name of the library file @var{library} that
4902 would be used when linking---and don't do anything else. With this
4903 option, GCC does not compile or link anything; it just prints the
4906 @item -print-multi-directory
4907 @opindex print-multi-directory
4908 Print the directory name corresponding to the multilib selected by any
4909 other switches present in the command line. This directory is supposed
4910 to exist in @env{GCC_EXEC_PREFIX}.
4912 @item -print-multi-lib
4913 @opindex print-multi-lib
4914 Print the mapping from multilib directory names to compiler switches
4915 that enable them. The directory name is separated from the switches by
4916 @samp{;}, and each switch starts with an @samp{@@} instead of the
4917 @samp{-}, without spaces between multiple switches. This is supposed to
4918 ease shell-processing.
4920 @item -print-prog-name=@var{program}
4921 @opindex print-prog-name
4922 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
4924 @item -print-libgcc-file-name
4925 @opindex print-libgcc-file-name
4926 Same as @option{-print-file-name=libgcc.a}.
4928 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
4929 but you do want to link with @file{libgcc.a}. You can do
4932 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
4935 @item -print-search-dirs
4936 @opindex print-search-dirs
4937 Print the name of the configured installation directory and a list of
4938 program and library directories @command{gcc} will search---and don't do anything else.
4940 This is useful when @command{gcc} prints the error message
4941 @samp{installation problem, cannot exec cpp0: No such file or directory}.
4942 To resolve this you either need to put @file{cpp0} and the other compiler
4943 components where @command{gcc} expects to find them, or you can set the environment
4944 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
4945 Don't forget the trailing @samp{/}.
4946 @xref{Environment Variables}.
4948 @item -print-sysroot-headers-suffix
4949 @opindex print-sysroot-headers-suffix
4950 Print the suffix added to the target sysroot when searching for
4951 headers, or give an error if the compiler is not configured with such
4952 a suffix---and don't do anything else.
4955 @opindex dumpmachine
4956 Print the compiler's target machine (for example,
4957 @samp{i686-pc-linux-gnu})---and don't do anything else.
4960 @opindex dumpversion
4961 Print the compiler version (for example, @samp{3.0})---and don't do
4966 Print the compiler's built-in specs---and don't do anything else. (This
4967 is used when GCC itself is being built.) @xref{Spec Files}.
4969 @item -feliminate-unused-debug-types
4970 @opindex feliminate-unused-debug-types
4971 Normally, when producing DWARF2 output, GCC will emit debugging
4972 information for all types declared in a compilation
4973 unit, regardless of whether or not they are actually used
4974 in that compilation unit. Sometimes this is useful, such as
4975 if, in the debugger, you want to cast a value to a type that is
4976 not actually used in your program (but is declared). More often,
4977 however, this results in a significant amount of wasted space.
4978 With this option, GCC will avoid producing debug symbol output
4979 for types that are nowhere used in the source file being compiled.
4982 @node Optimize Options
4983 @section Options That Control Optimization
4984 @cindex optimize options
4985 @cindex options, optimization
4987 These options control various sorts of optimizations.
4989 Without any optimization option, the compiler's goal is to reduce the
4990 cost of compilation and to make debugging produce the expected
4991 results. Statements are independent: if you stop the program with a
4992 breakpoint between statements, you can then assign a new value to any
4993 variable or change the program counter to any other statement in the
4994 function and get exactly the results you would expect from the source
4997 Turning on optimization flags makes the compiler attempt to improve
4998 the performance and/or code size at the expense of compilation time
4999 and possibly the ability to debug the program.
5001 The compiler performs optimization based on the knowledge it has of
5002 the program. Optimization levels @option{-O} and above, in
5003 particular, enable @emph{unit-at-a-time} mode, which allows the
5004 compiler to consider information gained from later functions in
5005 the file when compiling a function. Compiling multiple files at
5006 once to a single output file in @emph{unit-at-a-time} mode allows
5007 the compiler to use information gained from all of the files when
5008 compiling each of them.
5010 Not all optimizations are controlled directly by a flag. Only
5011 optimizations that have a flag are listed.
5018 Optimize. Optimizing compilation takes somewhat more time, and a lot
5019 more memory for a large function.
5021 With @option{-O}, the compiler tries to reduce code size and execution
5022 time, without performing any optimizations that take a great deal of
5025 @option{-O} turns on the following optimization flags:
5026 @gccoptlist{-fdefer-pop @gol
5027 -fdelayed-branch @gol
5028 -fguess-branch-probability @gol
5029 -fcprop-registers @gol
5030 -fif-conversion @gol
5031 -fif-conversion2 @gol
5032 -fsplit-wide-types @gol
5035 -ftree-dominator-opts @gol
5039 -ftree-copyrename @gol
5042 -funit-at-a-time @gol
5043 -finline-small-functions @gol
5046 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5047 where doing so does not interfere with debugging.
5051 Optimize even more. GCC performs nearly all supported optimizations
5052 that do not involve a space-speed tradeoff. The compiler does not
5053 perform loop unrolling or function inlining when you specify @option{-O2}.
5054 As compared to @option{-O}, this option increases both compilation time
5055 and the performance of the generated code.
5057 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5058 also turns on the following optimization flags:
5059 @gccoptlist{-fthread-jumps @gol
5061 -foptimize-sibling-calls @gol
5062 -fcse-follow-jumps -fcse-skip-blocks @gol
5063 -fgcse -fgcse-lm @gol
5064 -fexpensive-optimizations @gol
5065 -frerun-cse-after-loop @gol
5068 -fschedule-insns -fschedule-insns2 @gol
5069 -fsched-interblock -fsched-spec @gol
5071 -fstrict-aliasing -fstrict-overflow @gol
5072 -fdelete-null-pointer-checks @gol
5073 -freorder-blocks -freorder-functions @gol
5074 -falign-functions -falign-jumps @gol
5075 -falign-loops -falign-labels @gol
5079 Please note the warning under @option{-fgcse} about
5080 invoking @option{-O2} on programs that use computed gotos.
5084 Optimize yet more. @option{-O3} turns on all optimizations specified by
5085 @option{-O2} and also turns on the @option{-finline-functions},
5086 @option{-funswitch-loops}, @option{-fpredictive-commoning} and
5087 @option{-fgcse-after-reload} options.
5091 Reduce compilation time and make debugging produce the expected
5092 results. This is the default.
5096 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5097 do not typically increase code size. It also performs further
5098 optimizations designed to reduce code size.
5100 @option{-Os} disables the following optimization flags:
5101 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5102 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5103 -fprefetch-loop-arrays -ftree-vect-loop-version}
5105 If you use multiple @option{-O} options, with or without level numbers,
5106 the last such option is the one that is effective.
5109 Options of the form @option{-f@var{flag}} specify machine-independent
5110 flags. Most flags have both positive and negative forms; the negative
5111 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5112 below, only one of the forms is listed---the one you typically will
5113 use. You can figure out the other form by either removing @samp{no-}
5116 The following options control specific optimizations. They are either
5117 activated by @option{-O} options or are related to ones that are. You
5118 can use the following flags in the rare cases when ``fine-tuning'' of
5119 optimizations to be performed is desired.
5122 @item -fno-default-inline
5123 @opindex fno-default-inline
5124 Do not make member functions inline by default merely because they are
5125 defined inside the class scope (C++ only). Otherwise, when you specify
5126 @w{@option{-O}}, member functions defined inside class scope are compiled
5127 inline by default; i.e., you don't need to add @samp{inline} in front of
5128 the member function name.
5130 @item -fno-defer-pop
5131 @opindex fno-defer-pop
5132 Always pop the arguments to each function call as soon as that function
5133 returns. For machines which must pop arguments after a function call,
5134 the compiler normally lets arguments accumulate on the stack for several
5135 function calls and pops them all at once.
5137 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5140 @opindex fforce-addr
5141 Force memory address constants to be copied into registers before
5142 doing arithmetic on them.
5144 @item -fforward-propagate
5145 @opindex fforward-propagate
5146 Perform a forward propagation pass on RTL. The pass tries to combine two
5147 instructions and checks if the result can be simplified. If loop unrolling
5148 is active, two passes are performed and the second is scheduled after
5151 This option is enabled by default at optimization levels @option{-O2},
5152 @option{-O3}, @option{-Os}.
5154 @item -fomit-frame-pointer
5155 @opindex fomit-frame-pointer
5156 Don't keep the frame pointer in a register for functions that
5157 don't need one. This avoids the instructions to save, set up and
5158 restore frame pointers; it also makes an extra register available
5159 in many functions. @strong{It also makes debugging impossible on
5162 On some machines, such as the VAX, this flag has no effect, because
5163 the standard calling sequence automatically handles the frame pointer
5164 and nothing is saved by pretending it doesn't exist. The
5165 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5166 whether a target machine supports this flag. @xref{Registers,,Register
5167 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5169 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5171 @item -foptimize-sibling-calls
5172 @opindex foptimize-sibling-calls
5173 Optimize sibling and tail recursive calls.
5175 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5179 Don't pay attention to the @code{inline} keyword. Normally this option
5180 is used to keep the compiler from expanding any functions inline.
5181 Note that if you are not optimizing, no functions can be expanded inline.
5183 @item -finline-small-functions
5184 @opindex finline-small-functions
5185 Integrate functions into their callers when their body is smaller than expected
5186 function call code (so overall size of program gets smaller). The compiler
5187 heuristically decides which functions are simple enough to be worth integrating
5190 Enabled at level @option{-O2}.
5192 @item -finline-functions
5193 @opindex finline-functions
5194 Integrate all simple functions into their callers. The compiler
5195 heuristically decides which functions are simple enough to be worth
5196 integrating in this way.
5198 If all calls to a given function are integrated, and the function is
5199 declared @code{static}, then the function is normally not output as
5200 assembler code in its own right.
5202 Enabled at level @option{-O3}.
5204 @item -finline-functions-called-once
5205 @opindex finline-functions-called-once
5206 Consider all @code{static} functions called once for inlining into their
5207 caller even if they are not marked @code{inline}. If a call to a given
5208 function is integrated, then the function is not output as assembler code
5211 Enabled if @option{-funit-at-a-time} is enabled.
5213 @item -fearly-inlining
5214 @opindex fearly-inlining
5215 Inline functions marked by @code{always_inline} and functions whose body seems
5216 smaller than the function call overhead early before doing
5217 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5218 makes profiling significantly cheaper and usually inlining faster on programs
5219 having large chains of nested wrapper functions.
5223 @item -finline-limit=@var{n}
5224 @opindex finline-limit
5225 By default, GCC limits the size of functions that can be inlined. This flag
5226 allows the control of this limit for functions that are explicitly marked as
5227 inline (i.e., marked with the inline keyword or defined within the class
5228 definition in c++). @var{n} is the size of functions that can be inlined in
5229 number of pseudo instructions (not counting parameter handling). The default
5230 value of @var{n} is 600.
5231 Increasing this value can result in more inlined code at
5232 the cost of compilation time and memory consumption. Decreasing usually makes
5233 the compilation faster and less code will be inlined (which presumably
5234 means slower programs). This option is particularly useful for programs that
5235 use inlining heavily such as those based on recursive templates with C++.
5237 Inlining is actually controlled by a number of parameters, which may be
5238 specified individually by using @option{--param @var{name}=@var{value}}.
5239 The @option{-finline-limit=@var{n}} option sets some of these parameters
5243 @item max-inline-insns-single
5244 is set to @var{n}/2.
5245 @item max-inline-insns-auto
5246 is set to @var{n}/2.
5247 @item min-inline-insns
5248 is set to 130 or @var{n}/4, whichever is smaller.
5249 @item max-inline-insns-rtl
5253 See below for a documentation of the individual
5254 parameters controlling inlining.
5256 @emph{Note:} pseudo instruction represents, in this particular context, an
5257 abstract measurement of function's size. In no way does it represent a count
5258 of assembly instructions and as such its exact meaning might change from one
5259 release to an another.
5261 @item -fkeep-inline-functions
5262 @opindex fkeep-inline-functions
5263 In C, emit @code{static} functions that are declared @code{inline}
5264 into the object file, even if the function has been inlined into all
5265 of its callers. This switch does not affect functions using the
5266 @code{extern inline} extension in GNU C89@. In C++, emit any and all
5267 inline functions into the object file.
5269 @item -fkeep-static-consts
5270 @opindex fkeep-static-consts
5271 Emit variables declared @code{static const} when optimization isn't turned
5272 on, even if the variables aren't referenced.
5274 GCC enables this option by default. If you want to force the compiler to
5275 check if the variable was referenced, regardless of whether or not
5276 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5278 @item -fmerge-constants
5279 Attempt to merge identical constants (string constants and floating point
5280 constants) across compilation units.
5282 This option is the default for optimized compilation if the assembler and
5283 linker support it. Use @option{-fno-merge-constants} to inhibit this
5286 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5288 @item -fmerge-all-constants
5289 Attempt to merge identical constants and identical variables.
5291 This option implies @option{-fmerge-constants}. In addition to
5292 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5293 arrays or initialized constant variables with integral or floating point
5294 types. Languages like C or C++ require each non-automatic variable to
5295 have distinct location, so using this option will result in non-conforming
5298 @item -fmodulo-sched
5299 @opindex fmodulo-sched
5300 Perform swing modulo scheduling immediately before the first scheduling
5301 pass. This pass looks at innermost loops and reorders their
5302 instructions by overlapping different iterations.
5304 @item -fmodulo-sched-allow-regmoves
5305 @opindex fmodulo-sched-allow-regmoves
5306 Perform more aggressive SMS based modulo scheduling with register moves
5307 allowed. By setting this flag certain anti-dependences edges will be
5308 deleted which will trigger the generation of reg-moves based on the
5309 life-range analysis. This option is effective only with
5310 @option{-fmodulo-sched} enabled.
5312 @item -fno-branch-count-reg
5313 @opindex fno-branch-count-reg
5314 Do not use ``decrement and branch'' instructions on a count register,
5315 but instead generate a sequence of instructions that decrement a
5316 register, compare it against zero, then branch based upon the result.
5317 This option is only meaningful on architectures that support such
5318 instructions, which include x86, PowerPC, IA-64 and S/390.
5320 The default is @option{-fbranch-count-reg}.
5322 @item -fno-function-cse
5323 @opindex fno-function-cse
5324 Do not put function addresses in registers; make each instruction that
5325 calls a constant function contain the function's address explicitly.
5327 This option results in less efficient code, but some strange hacks
5328 that alter the assembler output may be confused by the optimizations
5329 performed when this option is not used.
5331 The default is @option{-ffunction-cse}
5333 @item -fno-zero-initialized-in-bss
5334 @opindex fno-zero-initialized-in-bss
5335 If the target supports a BSS section, GCC by default puts variables that
5336 are initialized to zero into BSS@. This can save space in the resulting
5339 This option turns off this behavior because some programs explicitly
5340 rely on variables going to the data section. E.g., so that the
5341 resulting executable can find the beginning of that section and/or make
5342 assumptions based on that.
5344 The default is @option{-fzero-initialized-in-bss}.
5346 @item -fbounds-check
5347 @opindex fbounds-check
5348 For front-ends that support it, generate additional code to check that
5349 indices used to access arrays are within the declared range. This is
5350 currently only supported by the Java and Fortran front-ends, where
5351 this option defaults to true and false respectively.
5353 @item -fmudflap -fmudflapth -fmudflapir
5357 @cindex bounds checking
5359 For front-ends that support it (C and C++), instrument all risky
5360 pointer/array dereferencing operations, some standard library
5361 string/heap functions, and some other associated constructs with
5362 range/validity tests. Modules so instrumented should be immune to
5363 buffer overflows, invalid heap use, and some other classes of C/C++
5364 programming errors. The instrumentation relies on a separate runtime
5365 library (@file{libmudflap}), which will be linked into a program if
5366 @option{-fmudflap} is given at link time. Run-time behavior of the
5367 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
5368 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
5371 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
5372 link if your program is multi-threaded. Use @option{-fmudflapir}, in
5373 addition to @option{-fmudflap} or @option{-fmudflapth}, if
5374 instrumentation should ignore pointer reads. This produces less
5375 instrumentation (and therefore faster execution) and still provides
5376 some protection against outright memory corrupting writes, but allows
5377 erroneously read data to propagate within a program.
5379 @item -fthread-jumps
5380 @opindex fthread-jumps
5381 Perform optimizations where we check to see if a jump branches to a
5382 location where another comparison subsumed by the first is found. If
5383 so, the first branch is redirected to either the destination of the
5384 second branch or a point immediately following it, depending on whether
5385 the condition is known to be true or false.
5387 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5389 @item -fsplit-wide-types
5390 @opindex fsplit-wide-types
5391 When using a type that occupies multiple registers, such as @code{long
5392 long} on a 32-bit system, split the registers apart and allocate them
5393 independently. This normally generates better code for those types,
5394 but may make debugging more difficult.
5396 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
5399 @item -fcse-follow-jumps
5400 @opindex fcse-follow-jumps
5401 In common subexpression elimination, scan through jump instructions
5402 when the target of the jump is not reached by any other path. For
5403 example, when CSE encounters an @code{if} statement with an
5404 @code{else} clause, CSE will follow the jump when the condition
5407 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5409 @item -fcse-skip-blocks
5410 @opindex fcse-skip-blocks
5411 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
5412 follow jumps which conditionally skip over blocks. When CSE
5413 encounters a simple @code{if} statement with no else clause,
5414 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
5415 body of the @code{if}.
5417 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5419 @item -frerun-cse-after-loop
5420 @opindex frerun-cse-after-loop
5421 Re-run common subexpression elimination after loop optimizations has been
5424 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5428 Perform a global common subexpression elimination pass.
5429 This pass also performs global constant and copy propagation.
5431 @emph{Note:} When compiling a program using computed gotos, a GCC
5432 extension, you may get better runtime performance if you disable
5433 the global common subexpression elimination pass by adding
5434 @option{-fno-gcse} to the command line.
5436 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5440 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
5441 attempt to move loads which are only killed by stores into themselves. This
5442 allows a loop containing a load/store sequence to be changed to a load outside
5443 the loop, and a copy/store within the loop.
5445 Enabled by default when gcse is enabled.
5449 When @option{-fgcse-sm} is enabled, a store motion pass is run after
5450 global common subexpression elimination. This pass will attempt to move
5451 stores out of loops. When used in conjunction with @option{-fgcse-lm},
5452 loops containing a load/store sequence can be changed to a load before
5453 the loop and a store after the loop.
5455 Not enabled at any optimization level.
5459 When @option{-fgcse-las} is enabled, the global common subexpression
5460 elimination pass eliminates redundant loads that come after stores to the
5461 same memory location (both partial and full redundancies).
5463 Not enabled at any optimization level.
5465 @item -fgcse-after-reload
5466 @opindex fgcse-after-reload
5467 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
5468 pass is performed after reload. The purpose of this pass is to cleanup
5471 @item -funsafe-loop-optimizations
5472 @opindex funsafe-loop-optimizations
5473 If given, the loop optimizer will assume that loop indices do not
5474 overflow, and that the loops with nontrivial exit condition are not
5475 infinite. This enables a wider range of loop optimizations even if
5476 the loop optimizer itself cannot prove that these assumptions are valid.
5477 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
5478 if it finds this kind of loop.
5480 @item -fcrossjumping
5481 @opindex crossjumping
5482 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
5483 resulting code may or may not perform better than without cross-jumping.
5485 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5487 @item -fif-conversion
5488 @opindex if-conversion
5489 Attempt to transform conditional jumps into branch-less equivalents. This
5490 include use of conditional moves, min, max, set flags and abs instructions, and
5491 some tricks doable by standard arithmetics. The use of conditional execution
5492 on chips where it is available is controlled by @code{if-conversion2}.
5494 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5496 @item -fif-conversion2
5497 @opindex if-conversion2
5498 Use conditional execution (where available) to transform conditional jumps into
5499 branch-less equivalents.
5501 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5503 @item -fdelete-null-pointer-checks
5504 @opindex fdelete-null-pointer-checks
5505 Use global dataflow analysis to identify and eliminate useless checks
5506 for null pointers. The compiler assumes that dereferencing a null
5507 pointer would have halted the program. If a pointer is checked after
5508 it has already been dereferenced, it cannot be null.
5510 In some environments, this assumption is not true, and programs can
5511 safely dereference null pointers. Use
5512 @option{-fno-delete-null-pointer-checks} to disable this optimization
5513 for programs which depend on that behavior.
5515 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5517 @item -fexpensive-optimizations
5518 @opindex fexpensive-optimizations
5519 Perform a number of minor optimizations that are relatively expensive.
5521 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5523 @item -foptimize-register-move
5525 @opindex foptimize-register-move
5527 Attempt to reassign register numbers in move instructions and as
5528 operands of other simple instructions in order to maximize the amount of
5529 register tying. This is especially helpful on machines with two-operand
5532 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
5535 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5537 @item -fdelayed-branch
5538 @opindex fdelayed-branch
5539 If supported for the target machine, attempt to reorder instructions
5540 to exploit instruction slots available after delayed branch
5543 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5545 @item -fschedule-insns
5546 @opindex fschedule-insns
5547 If supported for the target machine, attempt to reorder instructions to
5548 eliminate execution stalls due to required data being unavailable. This
5549 helps machines that have slow floating point or memory load instructions
5550 by allowing other instructions to be issued until the result of the load
5551 or floating point instruction is required.
5553 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5555 @item -fschedule-insns2
5556 @opindex fschedule-insns2
5557 Similar to @option{-fschedule-insns}, but requests an additional pass of
5558 instruction scheduling after register allocation has been done. This is
5559 especially useful on machines with a relatively small number of
5560 registers and where memory load instructions take more than one cycle.
5562 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5564 @item -fno-sched-interblock
5565 @opindex fno-sched-interblock
5566 Don't schedule instructions across basic blocks. This is normally
5567 enabled by default when scheduling before register allocation, i.e.@:
5568 with @option{-fschedule-insns} or at @option{-O2} or higher.
5570 @item -fno-sched-spec
5571 @opindex fno-sched-spec
5572 Don't allow speculative motion of non-load instructions. This is normally
5573 enabled by default when scheduling before register allocation, i.e.@:
5574 with @option{-fschedule-insns} or at @option{-O2} or higher.
5576 @item -fsched-spec-load
5577 @opindex fsched-spec-load
5578 Allow speculative motion of some load instructions. This only makes
5579 sense when scheduling before register allocation, i.e.@: with
5580 @option{-fschedule-insns} or at @option{-O2} or higher.
5582 @item -fsched-spec-load-dangerous
5583 @opindex fsched-spec-load-dangerous
5584 Allow speculative motion of more load instructions. This only makes
5585 sense when scheduling before register allocation, i.e.@: with
5586 @option{-fschedule-insns} or at @option{-O2} or higher.
5588 @item -fsched-stalled-insns=@var{n}
5589 @opindex fsched-stalled-insns
5590 Define how many insns (if any) can be moved prematurely from the queue
5591 of stalled insns into the ready list, during the second scheduling pass.
5593 @item -fsched-stalled-insns-dep=@var{n}
5594 @opindex fsched-stalled-insns-dep
5595 Define how many insn groups (cycles) will be examined for a dependency
5596 on a stalled insn that is candidate for premature removal from the queue
5597 of stalled insns. Has an effect only during the second scheduling pass,
5598 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
5600 @item -fsched2-use-superblocks
5601 @opindex fsched2-use-superblocks
5602 When scheduling after register allocation, do use superblock scheduling
5603 algorithm. Superblock scheduling allows motion across basic block boundaries
5604 resulting on faster schedules. This option is experimental, as not all machine
5605 descriptions used by GCC model the CPU closely enough to avoid unreliable
5606 results from the algorithm.
5608 This only makes sense when scheduling after register allocation, i.e.@: with
5609 @option{-fschedule-insns2} or at @option{-O2} or higher.
5611 @item -fsched2-use-traces
5612 @opindex fsched2-use-traces
5613 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
5614 allocation and additionally perform code duplication in order to increase the
5615 size of superblocks using tracer pass. See @option{-ftracer} for details on
5618 This mode should produce faster but significantly longer programs. Also
5619 without @option{-fbranch-probabilities} the traces constructed may not
5620 match the reality and hurt the performance. This only makes
5621 sense when scheduling after register allocation, i.e.@: with
5622 @option{-fschedule-insns2} or at @option{-O2} or higher.
5626 Eliminates redundant extension instructions and move the non redundant
5627 ones to optimal placement using LCM.
5629 @item -freschedule-modulo-scheduled-loops
5630 @opindex fscheduling-in-modulo-scheduled-loops
5631 The modulo scheduling comes before the traditional scheduling, if a loop
5632 was modulo scheduled we may want to prevent the later scheduling passes
5633 from changing its schedule, we use this option to control that.
5635 @item -fcaller-saves
5636 @opindex fcaller-saves
5637 Enable values to be allocated in registers that will be clobbered by
5638 function calls, by emitting extra instructions to save and restore the
5639 registers around such calls. Such allocation is done only when it
5640 seems to result in better code than would otherwise be produced.
5642 This option is always enabled by default on certain machines, usually
5643 those which have no call-preserved registers to use instead.
5645 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5647 @item -ftree-reassoc
5648 Perform Reassociation on trees This flag is enabled by default
5649 at @option{-O} and higher.
5652 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
5653 enabled by default at @option{-O2} and @option{-O3}.
5656 Perform Full Redundancy Elimination (FRE) on trees. The difference
5657 between FRE and PRE is that FRE only considers expressions
5658 that are computed on all paths leading to the redundant computation.
5659 This analysis is faster than PRE, though it exposes fewer redundancies.
5660 This flag is enabled by default at @option{-O} and higher.
5662 @item -ftree-copy-prop
5663 Perform copy propagation on trees. This pass eliminates unnecessary
5664 copy operations. This flag is enabled by default at @option{-O} and
5667 @item -ftree-store-copy-prop
5668 Perform copy propagation of memory loads and stores. This pass
5669 eliminates unnecessary copy operations in memory references
5670 (structures, global variables, arrays, etc). This flag is enabled by
5671 default at @option{-O2} and higher.
5674 Perform structural alias analysis on trees. This flag
5675 is enabled by default at @option{-O} and higher.
5678 Perform interprocedural pointer analysis.
5681 Perform forward store motion on trees. This flag is
5682 enabled by default at @option{-O} and higher.
5685 Perform sparse conditional constant propagation (CCP) on trees. This
5686 pass only operates on local scalar variables and is enabled by default
5687 at @option{-O} and higher.
5689 @item -ftree-store-ccp
5690 Perform sparse conditional constant propagation (CCP) on trees. This
5691 pass operates on both local scalar variables and memory stores and
5692 loads (global variables, structures, arrays, etc). This flag is
5693 enabled by default at @option{-O2} and higher.
5696 Perform dead code elimination (DCE) on trees. This flag is enabled by
5697 default at @option{-O} and higher.
5699 @item -ftree-dominator-opts
5700 Perform a variety of simple scalar cleanups (constant/copy
5701 propagation, redundancy elimination, range propagation and expression
5702 simplification) based on a dominator tree traversal. This also
5703 performs jump threading (to reduce jumps to jumps). This flag is
5704 enabled by default at @option{-O} and higher.
5708 Perform dead store elimination (DSE) on trees. A dead store is a store into
5709 a memory location which will later be overwritten by another store without
5710 any intervening loads. In this case the earlier store can be deleted. This
5711 flag is enabled by default at @option{-O} and higher.
5714 Perform loop header copying on trees. This is beneficial since it increases
5715 effectiveness of code motion optimizations. It also saves one jump. This flag
5716 is enabled by default at @option{-O} and higher. It is not enabled
5717 for @option{-Os}, since it usually increases code size.
5719 @item -ftree-loop-optimize
5720 Perform loop optimizations on trees. This flag is enabled by default
5721 at @option{-O} and higher.
5723 @item -ftree-loop-linear
5724 Perform linear loop transformations on tree. This flag can improve cache
5725 performance and allow further loop optimizations to take place.
5727 @item -fcheck-data-deps
5728 Compare the results of several data dependence analyzers. This option
5729 is used for debugging the data dependence analyzers.
5731 @item -ftree-loop-im
5732 Perform loop invariant motion on trees. This pass moves only invariants that
5733 would be hard to handle at RTL level (function calls, operations that expand to
5734 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5735 operands of conditions that are invariant out of the loop, so that we can use
5736 just trivial invariantness analysis in loop unswitching. The pass also includes
5739 @item -ftree-loop-ivcanon
5740 Create a canonical counter for number of iterations in the loop for that
5741 determining number of iterations requires complicated analysis. Later
5742 optimizations then may determine the number easily. Useful especially
5743 in connection with unrolling.
5746 Perform induction variable optimizations (strength reduction, induction
5747 variable merging and induction variable elimination) on trees.
5749 @item -ftree-parallelize-loops=n
5750 @opindex ftree-parallelize-loops=n
5751 Parallelize loops, i.e., split their iteration space to run in n threads.
5752 This is only possible for loops whose iterations are independent
5753 and can be arbitrarily reordered. The optimization is only
5754 profitable on multiprocessor machines, for loops that are CPU-intensive,
5755 rather than constrained e.g. by memory bandwidth.
5758 Perform scalar replacement of aggregates. This pass replaces structure
5759 references with scalars to prevent committing structures to memory too
5760 early. This flag is enabled by default at @option{-O} and higher.
5762 @item -ftree-copyrename
5763 Perform copy renaming on trees. This pass attempts to rename compiler
5764 temporaries to other variables at copy locations, usually resulting in
5765 variable names which more closely resemble the original variables. This flag
5766 is enabled by default at @option{-O} and higher.
5769 Perform temporary expression replacement during the SSA->normal phase. Single
5770 use/single def temporaries are replaced at their use location with their
5771 defining expression. This results in non-GIMPLE code, but gives the expanders
5772 much more complex trees to work on resulting in better RTL generation. This is
5773 enabled by default at @option{-O} and higher.
5775 @item -ftree-vectorize
5776 Perform loop vectorization on trees.
5778 @item -ftree-vect-loop-version
5779 @opindex ftree-vect-loop-version
5780 Perform loop versioning when doing loop vectorization on trees. When a loop
5781 appears to be vectorizable except that data alignment or data dependence cannot
5782 be determined at compile time then vectorized and non-vectorized versions of
5783 the loop are generated along with runtime checks for alignment or dependence
5784 to control which version is executed. This option is enabled by default
5785 except at level @option{-Os} where it is disabled.
5787 @item -fvect-cost-model
5788 Enable cost model for vectorization.
5791 Perform Value Range Propagation on trees. This is similar to the
5792 constant propagation pass, but instead of values, ranges of values are
5793 propagated. This allows the optimizers to remove unnecessary range
5794 checks like array bound checks and null pointer checks. This is
5795 enabled by default at @option{-O2} and higher. Null pointer check
5796 elimination is only done if @option{-fdelete-null-pointer-checks} is
5801 Perform tail duplication to enlarge superblock size. This transformation
5802 simplifies the control flow of the function allowing other optimizations to do
5805 @item -funroll-loops
5806 @opindex funroll-loops
5807 Unroll loops whose number of iterations can be determined at compile
5808 time or upon entry to the loop. @option{-funroll-loops} implies
5809 @option{-frerun-cse-after-loop}. This option makes code larger,
5810 and may or may not make it run faster.
5812 @item -funroll-all-loops
5813 @opindex funroll-all-loops
5814 Unroll all loops, even if their number of iterations is uncertain when
5815 the loop is entered. This usually makes programs run more slowly.
5816 @option{-funroll-all-loops} implies the same options as
5817 @option{-funroll-loops},
5819 @item -fsplit-ivs-in-unroller
5820 @opindex fsplit-ivs-in-unroller
5821 Enables expressing of values of induction variables in later iterations
5822 of the unrolled loop using the value in the first iteration. This breaks
5823 long dependency chains, thus improving efficiency of the scheduling passes.
5825 Combination of @option{-fweb} and CSE is often sufficient to obtain the
5826 same effect. However in cases the loop body is more complicated than
5827 a single basic block, this is not reliable. It also does not work at all
5828 on some of the architectures due to restrictions in the CSE pass.
5830 This optimization is enabled by default.
5832 @item -fvariable-expansion-in-unroller
5833 @opindex fvariable-expansion-in-unroller
5834 With this option, the compiler will create multiple copies of some
5835 local variables when unrolling a loop which can result in superior code.
5837 @item -fpredictive-commoning
5838 @opindex fpredictive-commoning
5839 Perform predictive commoning optimization, i.e., reusing computations
5840 (especially memory loads and stores) performed in previous
5841 iterations of loops.
5843 This option is enabled at level @option{-O3}.
5845 @item -fprefetch-loop-arrays
5846 @opindex fprefetch-loop-arrays
5847 If supported by the target machine, generate instructions to prefetch
5848 memory to improve the performance of loops that access large arrays.
5850 This option may generate better or worse code; results are highly
5851 dependent on the structure of loops within the source code.
5853 Disabled at level @option{-Os}.
5856 @itemx -fno-peephole2
5857 @opindex fno-peephole
5858 @opindex fno-peephole2
5859 Disable any machine-specific peephole optimizations. The difference
5860 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
5861 are implemented in the compiler; some targets use one, some use the
5862 other, a few use both.
5864 @option{-fpeephole} is enabled by default.
5865 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5867 @item -fno-guess-branch-probability
5868 @opindex fno-guess-branch-probability
5869 Do not guess branch probabilities using heuristics.
5871 GCC will use heuristics to guess branch probabilities if they are
5872 not provided by profiling feedback (@option{-fprofile-arcs}). These
5873 heuristics are based on the control flow graph. If some branch probabilities
5874 are specified by @samp{__builtin_expect}, then the heuristics will be
5875 used to guess branch probabilities for the rest of the control flow graph,
5876 taking the @samp{__builtin_expect} info into account. The interactions
5877 between the heuristics and @samp{__builtin_expect} can be complex, and in
5878 some cases, it may be useful to disable the heuristics so that the effects
5879 of @samp{__builtin_expect} are easier to understand.
5881 The default is @option{-fguess-branch-probability} at levels
5882 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5884 @item -freorder-blocks
5885 @opindex freorder-blocks
5886 Reorder basic blocks in the compiled function in order to reduce number of
5887 taken branches and improve code locality.
5889 Enabled at levels @option{-O2}, @option{-O3}.
5891 @item -freorder-blocks-and-partition
5892 @opindex freorder-blocks-and-partition
5893 In addition to reordering basic blocks in the compiled function, in order
5894 to reduce number of taken branches, partitions hot and cold basic blocks
5895 into separate sections of the assembly and .o files, to improve
5896 paging and cache locality performance.
5898 This optimization is automatically turned off in the presence of
5899 exception handling, for linkonce sections, for functions with a user-defined
5900 section attribute and on any architecture that does not support named
5903 @item -freorder-functions
5904 @opindex freorder-functions
5905 Reorder functions in the object file in order to
5906 improve code locality. This is implemented by using special
5907 subsections @code{.text.hot} for most frequently executed functions and
5908 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
5909 the linker so object file format must support named sections and linker must
5910 place them in a reasonable way.
5912 Also profile feedback must be available in to make this option effective. See
5913 @option{-fprofile-arcs} for details.
5915 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5917 @item -fstrict-aliasing
5918 @opindex fstrict-aliasing
5919 Allows the compiler to assume the strictest aliasing rules applicable to
5920 the language being compiled. For C (and C++), this activates
5921 optimizations based on the type of expressions. In particular, an
5922 object of one type is assumed never to reside at the same address as an
5923 object of a different type, unless the types are almost the same. For
5924 example, an @code{unsigned int} can alias an @code{int}, but not a
5925 @code{void*} or a @code{double}. A character type may alias any other
5928 Pay special attention to code like this:
5941 The practice of reading from a different union member than the one most
5942 recently written to (called ``type-punning'') is common. Even with
5943 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
5944 is accessed through the union type. So, the code above will work as
5945 expected. However, this code might not:
5956 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5958 @item -fstrict-overflow
5959 @opindex fstrict-overflow
5960 Allow the compiler to assume strict signed overflow rules, depending
5961 on the language being compiled. For C (and C++) this means that
5962 overflow when doing arithmetic with signed numbers is undefined, which
5963 means that the compiler may assume that it will not happen. This
5964 permits various optimizations. For example, the compiler will assume
5965 that an expression like @code{i + 10 > i} will always be true for
5966 signed @code{i}. This assumption is only valid if signed overflow is
5967 undefined, as the expression is false if @code{i + 10} overflows when
5968 using twos complement arithmetic. When this option is in effect any
5969 attempt to determine whether an operation on signed numbers will
5970 overflow must be written carefully to not actually involve overflow.
5972 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
5973 that signed overflow is fully defined: it wraps. When
5974 @option{-fwrapv} is used, there is no difference between
5975 @option{-fstrict-overflow} and @option{-fno-strict-overflow}. With
5976 @option{-fwrapv} certain types of overflow are permitted. For
5977 example, if the compiler gets an overflow when doing arithmetic on
5978 constants, the overflowed value can still be used with
5979 @option{-fwrapv}, but not otherwise.
5981 The @option{-fstrict-overflow} option is enabled at levels
5982 @option{-O2}, @option{-O3}, @option{-Os}.
5984 @item -falign-functions
5985 @itemx -falign-functions=@var{n}
5986 @opindex falign-functions
5987 Align the start of functions to the next power-of-two greater than
5988 @var{n}, skipping up to @var{n} bytes. For instance,
5989 @option{-falign-functions=32} aligns functions to the next 32-byte
5990 boundary, but @option{-falign-functions=24} would align to the next
5991 32-byte boundary only if this can be done by skipping 23 bytes or less.
5993 @option{-fno-align-functions} and @option{-falign-functions=1} are
5994 equivalent and mean that functions will not be aligned.
5996 Some assemblers only support this flag when @var{n} is a power of two;
5997 in that case, it is rounded up.
5999 If @var{n} is not specified or is zero, use a machine-dependent default.
6001 Enabled at levels @option{-O2}, @option{-O3}.
6003 @item -falign-labels
6004 @itemx -falign-labels=@var{n}
6005 @opindex falign-labels
6006 Align all branch targets to a power-of-two boundary, skipping up to
6007 @var{n} bytes like @option{-falign-functions}. This option can easily
6008 make code slower, because it must insert dummy operations for when the
6009 branch target is reached in the usual flow of the code.
6011 @option{-fno-align-labels} and @option{-falign-labels=1} are
6012 equivalent and mean that labels will not be aligned.
6014 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
6015 are greater than this value, then their values are used instead.
6017 If @var{n} is not specified or is zero, use a machine-dependent default
6018 which is very likely to be @samp{1}, meaning no alignment.
6020 Enabled at levels @option{-O2}, @option{-O3}.
6023 @itemx -falign-loops=@var{n}
6024 @opindex falign-loops
6025 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
6026 like @option{-falign-functions}. The hope is that the loop will be
6027 executed many times, which will make up for any execution of the dummy
6030 @option{-fno-align-loops} and @option{-falign-loops=1} are
6031 equivalent and mean that loops will not be aligned.
6033 If @var{n} is not specified or is zero, use a machine-dependent default.
6035 Enabled at levels @option{-O2}, @option{-O3}.
6038 @itemx -falign-jumps=@var{n}
6039 @opindex falign-jumps
6040 Align branch targets to a power-of-two boundary, for branch targets
6041 where the targets can only be reached by jumping, skipping up to @var{n}
6042 bytes like @option{-falign-functions}. In this case, no dummy operations
6045 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
6046 equivalent and mean that loops will not be aligned.
6048 If @var{n} is not specified or is zero, use a machine-dependent default.
6050 Enabled at levels @option{-O2}, @option{-O3}.
6052 @item -funit-at-a-time
6053 @opindex funit-at-a-time
6054 Parse the whole compilation unit before starting to produce code.
6055 This allows some extra optimizations to take place but consumes
6056 more memory (in general). There are some compatibility issues
6057 with @emph{unit-at-a-time} mode:
6060 enabling @emph{unit-at-a-time} mode may change the order
6061 in which functions, variables, and top-level @code{asm} statements
6062 are emitted, and will likely break code relying on some particular
6063 ordering. The majority of such top-level @code{asm} statements,
6064 though, can be replaced by @code{section} attributes. The
6065 @option{fno-toplevel-reorder} option may be used to keep the ordering
6066 used in the input file, at the cost of some optimizations.
6069 @emph{unit-at-a-time} mode removes unreferenced static variables
6070 and functions. This may result in undefined references
6071 when an @code{asm} statement refers directly to variables or functions
6072 that are otherwise unused. In that case either the variable/function
6073 shall be listed as an operand of the @code{asm} statement operand or,
6074 in the case of top-level @code{asm} statements the attribute @code{used}
6075 shall be used on the declaration.
6078 Static functions now can use non-standard passing conventions that
6079 may break @code{asm} statements calling functions directly. Again,
6080 attribute @code{used} will prevent this behavior.
6083 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
6084 but this scheme may not be supported by future releases of GCC@.
6086 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6088 @item -fno-toplevel-reorder
6089 Do not reorder top-level functions, variables, and @code{asm}
6090 statements. Output them in the same order that they appear in the
6091 input file. When this option is used, unreferenced static variables
6092 will not be removed. This option is intended to support existing code
6093 which relies on a particular ordering. For new code, it is better to
6098 Constructs webs as commonly used for register allocation purposes and assign
6099 each web individual pseudo register. This allows the register allocation pass
6100 to operate on pseudos directly, but also strengthens several other optimization
6101 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
6102 however, make debugging impossible, since variables will no longer stay in a
6105 Enabled by default with @option{-funroll-loops}.
6107 @item -fwhole-program
6108 @opindex fwhole-program
6109 Assume that the current compilation unit represents whole program being
6110 compiled. All public functions and variables with the exception of @code{main}
6111 and those merged by attribute @code{externally_visible} become static functions
6112 and in a affect gets more aggressively optimized by interprocedural optimizers.
6113 While this option is equivalent to proper use of @code{static} keyword for
6114 programs consisting of single file, in combination with option
6115 @option{--combine} this flag can be used to compile most of smaller scale C
6116 programs since the functions and variables become local for the whole combined
6117 compilation unit, not for the single source file itself.
6120 @item -fno-cprop-registers
6121 @opindex fno-cprop-registers
6122 After register allocation and post-register allocation instruction splitting,
6123 we perform a copy-propagation pass to try to reduce scheduling dependencies
6124 and occasionally eliminate the copy.
6126 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6128 @item -fprofile-generate
6129 @opindex fprofile-generate
6131 Enable options usually used for instrumenting application to produce
6132 profile useful for later recompilation with profile feedback based
6133 optimization. You must use @option{-fprofile-generate} both when
6134 compiling and when linking your program.
6136 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
6139 @opindex fprofile-use
6140 Enable profile feedback directed optimizations, and optimizations
6141 generally profitable only with profile feedback available.
6143 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
6144 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
6146 By default, GCC emits an error message if the feedback profiles do not
6147 match the source code. This error can be turned into a warning by using
6148 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
6152 The following options control compiler behavior regarding floating
6153 point arithmetic. These options trade off between speed and
6154 correctness. All must be specifically enabled.
6158 @opindex ffloat-store
6159 Do not store floating point variables in registers, and inhibit other
6160 options that might change whether a floating point value is taken from a
6163 @cindex floating point precision
6164 This option prevents undesirable excess precision on machines such as
6165 the 68000 where the floating registers (of the 68881) keep more
6166 precision than a @code{double} is supposed to have. Similarly for the
6167 x86 architecture. For most programs, the excess precision does only
6168 good, but a few programs rely on the precise definition of IEEE floating
6169 point. Use @option{-ffloat-store} for such programs, after modifying
6170 them to store all pertinent intermediate computations into variables.
6174 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
6175 @option{-fno-trapping-math}, @option{-ffinite-math-only},
6176 @option{-fno-rounding-math}, @option{-fno-signaling-nans},
6177 @option{-fno-signed-zeros} and @option{fcx-limited-range}.
6179 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
6181 This option is not turned on by any @option{-O} option since
6182 it can result in incorrect output for programs which depend on
6183 an exact implementation of IEEE or ISO rules/specifications for
6184 math functions. It may, however, yield faster code for programs
6185 that do not require the guarantees of these specifications.
6187 @item -fno-math-errno
6188 @opindex fno-math-errno
6189 Do not set ERRNO after calling math functions that are executed
6190 with a single instruction, e.g., sqrt. A program that relies on
6191 IEEE exceptions for math error handling may want to use this flag
6192 for speed while maintaining IEEE arithmetic compatibility.
6194 This option is not turned on by any @option{-O} option since
6195 it can result in incorrect output for programs which depend on
6196 an exact implementation of IEEE or ISO rules/specifications for
6197 math functions. It may, however, yield faster code for programs
6198 that do not require the guarantees of these specifications.
6200 The default is @option{-fmath-errno}.
6202 On Darwin systems, the math library never sets @code{errno}. There is
6203 therefore no reason for the compiler to consider the possibility that
6204 it might, and @option{-fno-math-errno} is the default.
6206 @item -funsafe-math-optimizations
6207 @opindex funsafe-math-optimizations
6209 Allow optimizations for floating-point arithmetic that (a) assume
6210 that arguments and results are valid and (b) may violate IEEE or
6211 ANSI standards. When used at link-time, it may include libraries
6212 or startup files that change the default FPU control word or other
6213 similar optimizations.
6215 This option is not turned on by any @option{-O} option since
6216 it can result in incorrect output for programs which depend on
6217 an exact implementation of IEEE or ISO rules/specifications for
6218 math functions. It may, however, yield faster code for programs
6219 that do not require the guarantees of these specifications.
6220 Enables @option{-freciprocal-math} and @option{-fassociative-math}.
6222 The default is @option{-fno-unsafe-math-optimizations}.
6224 @item -fassociative-math
6225 @opindex -fassociative-math
6227 Allow re-association of operands in series of floating-point operations.
6228 This violates the ISO C and C++ language standard by possibly changing
6229 computation result. NOTE: re-ordering may change the sign of zero as
6230 well as ignore NaNs and inhibit or create underflow or overflow (and
6231 thus cannot be used on a code which relies on rounding behavior like
6232 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
6233 and thus may not be used when ordered comparisons are required.
6234 This flag doesn't make much sense without @option{-fno-signed-zeros}
6235 or @option{-fno-trapping-math} or with @option{-frounding-math}.
6237 The default is @option{-fno-associative-math}.
6239 @item -freciprocal-math
6240 @opindex -freciprocal-math
6242 Allow the reciprocal of a value to be used instead of dividing by
6243 the value if this enables optimizations. For example @code{x / y}
6244 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
6245 is subject to common subexpression elimination. Note that this loses
6246 precision and increases the number of flops operating on the value.
6248 The default is @option{-fno-reciprocal-math}.
6250 @item -ffinite-math-only
6251 @opindex ffinite-math-only
6252 Allow optimizations for floating-point arithmetic that assume
6253 that arguments and results are not NaNs or +-Infs.
6255 This option is not turned on by any @option{-O} option since
6256 it can result in incorrect output for programs which depend on
6257 an exact implementation of IEEE or ISO rules/specifications for
6258 math functions. It may, however, yield faster code for programs
6259 that do not require the guarantees of these specifications.
6261 The default is @option{-fno-finite-math-only}.
6263 @item -fno-signed-zeros
6264 @opindex fno-signed-zeros
6265 Allow optimizations for floating point arithmetic that ignore the
6266 signedness of zero. IEEE arithmetic specifies the behavior of
6267 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
6268 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
6269 This option implies that the sign of a zero result isn't significant.
6271 The default is @option{-fsigned-zeros}.
6273 @item -fno-trapping-math
6274 @opindex fno-trapping-math
6275 Compile code assuming that floating-point operations cannot generate
6276 user-visible traps. These traps include division by zero, overflow,
6277 underflow, inexact result and invalid operation. This option implies
6278 @option{-fno-signaling-nans}. Setting this option may allow faster
6279 code if one relies on ``non-stop'' IEEE arithmetic, for example.
6281 This option should never be turned on by any @option{-O} option since
6282 it can result in incorrect output for programs which depend on
6283 an exact implementation of IEEE or ISO rules/specifications for
6286 The default is @option{-ftrapping-math}.
6288 @item -frounding-math
6289 @opindex frounding-math
6290 Disable transformations and optimizations that assume default floating
6291 point rounding behavior. This is round-to-zero for all floating point
6292 to integer conversions, and round-to-nearest for all other arithmetic
6293 truncations. This option should be specified for programs that change
6294 the FP rounding mode dynamically, or that may be executed with a
6295 non-default rounding mode. This option disables constant folding of
6296 floating point expressions at compile-time (which may be affected by
6297 rounding mode) and arithmetic transformations that are unsafe in the
6298 presence of sign-dependent rounding modes.
6300 The default is @option{-fno-rounding-math}.
6302 This option is experimental and does not currently guarantee to
6303 disable all GCC optimizations that are affected by rounding mode.
6304 Future versions of GCC may provide finer control of this setting
6305 using C99's @code{FENV_ACCESS} pragma. This command line option
6306 will be used to specify the default state for @code{FENV_ACCESS}.
6308 @item -frtl-abstract-sequences
6309 @opindex frtl-abstract-sequences
6310 It is a size optimization method. This option is to find identical
6311 sequences of code, which can be turned into pseudo-procedures and
6312 then replace all occurrences with calls to the newly created
6313 subroutine. It is kind of an opposite of @option{-finline-functions}.
6314 This optimization runs at RTL level.
6316 @item -fsignaling-nans
6317 @opindex fsignaling-nans
6318 Compile code assuming that IEEE signaling NaNs may generate user-visible
6319 traps during floating-point operations. Setting this option disables
6320 optimizations that may change the number of exceptions visible with
6321 signaling NaNs. This option implies @option{-ftrapping-math}.
6323 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
6326 The default is @option{-fno-signaling-nans}.
6328 This option is experimental and does not currently guarantee to
6329 disable all GCC optimizations that affect signaling NaN behavior.
6331 @item -fsingle-precision-constant
6332 @opindex fsingle-precision-constant
6333 Treat floating point constant as single precision constant instead of
6334 implicitly converting it to double precision constant.
6336 @item -fcx-limited-range
6337 @itemx -fno-cx-limited-range
6338 @opindex fcx-limited-range
6339 @opindex fno-cx-limited-range
6340 When enabled, this option states that a range reduction step is not
6341 needed when performing complex division. The default is
6342 @option{-fno-cx-limited-range}, but is enabled by @option{-ffast-math}.
6344 This option controls the default setting of the ISO C99
6345 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
6350 The following options control optimizations that may improve
6351 performance, but are not enabled by any @option{-O} options. This
6352 section includes experimental options that may produce broken code.
6355 @item -fbranch-probabilities
6356 @opindex fbranch-probabilities
6357 After running a program compiled with @option{-fprofile-arcs}
6358 (@pxref{Debugging Options,, Options for Debugging Your Program or
6359 @command{gcc}}), you can compile it a second time using
6360 @option{-fbranch-probabilities}, to improve optimizations based on
6361 the number of times each branch was taken. When the program
6362 compiled with @option{-fprofile-arcs} exits it saves arc execution
6363 counts to a file called @file{@var{sourcename}.gcda} for each source
6364 file. The information in this data file is very dependent on the
6365 structure of the generated code, so you must use the same source code
6366 and the same optimization options for both compilations.
6368 With @option{-fbranch-probabilities}, GCC puts a
6369 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
6370 These can be used to improve optimization. Currently, they are only
6371 used in one place: in @file{reorg.c}, instead of guessing which path a
6372 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
6373 exactly determine which path is taken more often.
6375 @item -fprofile-values
6376 @opindex fprofile-values
6377 If combined with @option{-fprofile-arcs}, it adds code so that some
6378 data about values of expressions in the program is gathered.
6380 With @option{-fbranch-probabilities}, it reads back the data gathered
6381 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
6382 notes to instructions for their later usage in optimizations.
6384 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
6388 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
6389 a code to gather information about values of expressions.
6391 With @option{-fbranch-probabilities}, it reads back the data gathered
6392 and actually performs the optimizations based on them.
6393 Currently the optimizations include specialization of division operation
6394 using the knowledge about the value of the denominator.
6396 @item -frename-registers
6397 @opindex frename-registers
6398 Attempt to avoid false dependencies in scheduled code by making use
6399 of registers left over after register allocation. This optimization
6400 will most benefit processors with lots of registers. Depending on the
6401 debug information format adopted by the target, however, it can
6402 make debugging impossible, since variables will no longer stay in
6403 a ``home register''.
6405 Enabled by default with @option{-funroll-loops}.
6409 Perform tail duplication to enlarge superblock size. This transformation
6410 simplifies the control flow of the function allowing other optimizations to do
6413 Enabled with @option{-fprofile-use}.
6415 @item -funroll-loops
6416 @opindex funroll-loops
6417 Unroll loops whose number of iterations can be determined at compile time or
6418 upon entry to the loop. @option{-funroll-loops} implies
6419 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
6420 It also turns on complete loop peeling (i.e.@: complete removal of loops with
6421 small constant number of iterations). This option makes code larger, and may
6422 or may not make it run faster.
6424 Enabled with @option{-fprofile-use}.
6426 @item -funroll-all-loops
6427 @opindex funroll-all-loops
6428 Unroll all loops, even if their number of iterations is uncertain when
6429 the loop is entered. This usually makes programs run more slowly.
6430 @option{-funroll-all-loops} implies the same options as
6431 @option{-funroll-loops}.
6434 @opindex fpeel-loops
6435 Peels the loops for that there is enough information that they do not
6436 roll much (from profile feedback). It also turns on complete loop peeling
6437 (i.e.@: complete removal of loops with small constant number of iterations).
6439 Enabled with @option{-fprofile-use}.
6441 @item -fmove-loop-invariants
6442 @opindex fmove-loop-invariants
6443 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
6444 at level @option{-O1}
6446 @item -funswitch-loops
6447 @opindex funswitch-loops
6448 Move branches with loop invariant conditions out of the loop, with duplicates
6449 of the loop on both branches (modified according to result of the condition).
6451 @item -ffunction-sections
6452 @itemx -fdata-sections
6453 @opindex ffunction-sections
6454 @opindex fdata-sections
6455 Place each function or data item into its own section in the output
6456 file if the target supports arbitrary sections. The name of the
6457 function or the name of the data item determines the section's name
6460 Use these options on systems where the linker can perform optimizations
6461 to improve locality of reference in the instruction space. Most systems
6462 using the ELF object format and SPARC processors running Solaris 2 have
6463 linkers with such optimizations. AIX may have these optimizations in
6466 Only use these options when there are significant benefits from doing
6467 so. When you specify these options, the assembler and linker will
6468 create larger object and executable files and will also be slower.
6469 You will not be able to use @code{gprof} on all systems if you
6470 specify this option and you may have problems with debugging if
6471 you specify both this option and @option{-g}.
6473 @item -fbranch-target-load-optimize
6474 @opindex fbranch-target-load-optimize
6475 Perform branch target register load optimization before prologue / epilogue
6477 The use of target registers can typically be exposed only during reload,
6478 thus hoisting loads out of loops and doing inter-block scheduling needs
6479 a separate optimization pass.
6481 @item -fbranch-target-load-optimize2
6482 @opindex fbranch-target-load-optimize2
6483 Perform branch target register load optimization after prologue / epilogue
6486 @item -fbtr-bb-exclusive
6487 @opindex fbtr-bb-exclusive
6488 When performing branch target register load optimization, don't reuse
6489 branch target registers in within any basic block.
6491 @item -fstack-protector
6492 Emit extra code to check for buffer overflows, such as stack smashing
6493 attacks. This is done by adding a guard variable to functions with
6494 vulnerable objects. This includes functions that call alloca, and
6495 functions with buffers larger than 8 bytes. The guards are initialized
6496 when a function is entered and then checked when the function exits.
6497 If a guard check fails, an error message is printed and the program exits.
6499 @item -fstack-protector-all
6500 Like @option{-fstack-protector} except that all functions are protected.
6502 @item -fsection-anchors
6503 @opindex fsection-anchors
6504 Try to reduce the number of symbolic address calculations by using
6505 shared ``anchor'' symbols to address nearby objects. This transformation
6506 can help to reduce the number of GOT entries and GOT accesses on some
6509 For example, the implementation of the following function @code{foo}:
6513 int foo (void) @{ return a + b + c; @}
6516 would usually calculate the addresses of all three variables, but if you
6517 compile it with @option{-fsection-anchors}, it will access the variables
6518 from a common anchor point instead. The effect is similar to the
6519 following pseudocode (which isn't valid C):
6524 register int *xr = &x;
6525 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
6529 Not all targets support this option.
6531 @item --param @var{name}=@var{value}
6533 In some places, GCC uses various constants to control the amount of
6534 optimization that is done. For example, GCC will not inline functions
6535 that contain more that a certain number of instructions. You can
6536 control some of these constants on the command-line using the
6537 @option{--param} option.
6539 The names of specific parameters, and the meaning of the values, are
6540 tied to the internals of the compiler, and are subject to change
6541 without notice in future releases.
6543 In each case, the @var{value} is an integer. The allowable choices for
6544 @var{name} are given in the following table:
6547 @item salias-max-implicit-fields
6548 The maximum number of fields in a variable without direct
6549 structure accesses for which structure aliasing will consider trying
6550 to track each field. The default is 5
6552 @item salias-max-array-elements
6553 The maximum number of elements an array can have and its elements
6554 still be tracked individually by structure aliasing. The default is 4
6556 @item sra-max-structure-size
6557 The maximum structure size, in bytes, at which the scalar replacement
6558 of aggregates (SRA) optimization will perform block copies. The
6559 default value, 0, implies that GCC will select the most appropriate
6562 @item sra-field-structure-ratio
6563 The threshold ratio (as a percentage) between instantiated fields and
6564 the complete structure size. We say that if the ratio of the number
6565 of bytes in instantiated fields to the number of bytes in the complete
6566 structure exceeds this parameter, then block copies are not used. The
6569 @item max-crossjump-edges
6570 The maximum number of incoming edges to consider for crossjumping.
6571 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
6572 the number of edges incoming to each block. Increasing values mean
6573 more aggressive optimization, making the compile time increase with
6574 probably small improvement in executable size.
6576 @item min-crossjump-insns
6577 The minimum number of instructions which must be matched at the end
6578 of two blocks before crossjumping will be performed on them. This
6579 value is ignored in the case where all instructions in the block being
6580 crossjumped from are matched. The default value is 5.
6582 @item max-grow-copy-bb-insns
6583 The maximum code size expansion factor when copying basic blocks
6584 instead of jumping. The expansion is relative to a jump instruction.
6585 The default value is 8.
6587 @item max-goto-duplication-insns
6588 The maximum number of instructions to duplicate to a block that jumps
6589 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
6590 passes, GCC factors computed gotos early in the compilation process,
6591 and unfactors them as late as possible. Only computed jumps at the
6592 end of a basic blocks with no more than max-goto-duplication-insns are
6593 unfactored. The default value is 8.
6595 @item max-delay-slot-insn-search
6596 The maximum number of instructions to consider when looking for an
6597 instruction to fill a delay slot. If more than this arbitrary number of
6598 instructions is searched, the time savings from filling the delay slot
6599 will be minimal so stop searching. Increasing values mean more
6600 aggressive optimization, making the compile time increase with probably
6601 small improvement in executable run time.
6603 @item max-delay-slot-live-search
6604 When trying to fill delay slots, the maximum number of instructions to
6605 consider when searching for a block with valid live register
6606 information. Increasing this arbitrarily chosen value means more
6607 aggressive optimization, increasing the compile time. This parameter
6608 should be removed when the delay slot code is rewritten to maintain the
6611 @item max-gcse-memory
6612 The approximate maximum amount of memory that will be allocated in
6613 order to perform the global common subexpression elimination
6614 optimization. If more memory than specified is required, the
6615 optimization will not be done.
6617 @item max-gcse-passes
6618 The maximum number of passes of GCSE to run. The default is 1.
6620 @item max-pending-list-length
6621 The maximum number of pending dependencies scheduling will allow
6622 before flushing the current state and starting over. Large functions
6623 with few branches or calls can create excessively large lists which
6624 needlessly consume memory and resources.
6626 @item max-inline-insns-single
6627 Several parameters control the tree inliner used in gcc.
6628 This number sets the maximum number of instructions (counted in GCC's
6629 internal representation) in a single function that the tree inliner
6630 will consider for inlining. This only affects functions declared
6631 inline and methods implemented in a class declaration (C++).
6632 The default value is 450.
6634 @item max-inline-insns-auto
6635 When you use @option{-finline-functions} (included in @option{-O3}),
6636 a lot of functions that would otherwise not be considered for inlining
6637 by the compiler will be investigated. To those functions, a different
6638 (more restrictive) limit compared to functions declared inline can
6640 The default value is 90.
6642 @item large-function-insns
6643 The limit specifying really large functions. For functions larger than this
6644 limit after inlining inlining is constrained by
6645 @option{--param large-function-growth}. This parameter is useful primarily
6646 to avoid extreme compilation time caused by non-linear algorithms used by the
6648 This parameter is ignored when @option{-funit-at-a-time} is not used.
6649 The default value is 2700.
6651 @item large-function-growth
6652 Specifies maximal growth of large function caused by inlining in percents.
6653 This parameter is ignored when @option{-funit-at-a-time} is not used.
6654 The default value is 100 which limits large function growth to 2.0 times
6657 @item large-unit-insns
6658 The limit specifying large translation unit. Growth caused by inlining of
6659 units larger than this limit is limited by @option{--param inline-unit-growth}.
6660 For small units this might be too tight (consider unit consisting of function A
6661 that is inline and B that just calls A three time. If B is small relative to
6662 A, the growth of unit is 300\% and yet such inlining is very sane. For very
6663 large units consisting of small inlininable functions however the overall unit
6664 growth limit is needed to avoid exponential explosion of code size. Thus for
6665 smaller units, the size is increased to @option{--param large-unit-insns}
6666 before applying @option{--param inline-unit-growth}. The default is 10000
6668 @item inline-unit-growth
6669 Specifies maximal overall growth of the compilation unit caused by inlining.
6670 This parameter is ignored when @option{-funit-at-a-time} is not used.
6671 The default value is 30 which limits unit growth to 1.3 times the original
6674 @item large-stack-frame
6675 The limit specifying large stack frames. While inlining the algorithm is trying
6676 to not grow past this limit too much. Default value is 256 bytes.
6678 @item large-stack-frame-growth
6679 Specifies maximal growth of large stack frames caused by inlining in percents.
6680 The default value is 1000 which limits large stack frame growth to 11 times
6683 @item max-inline-insns-recursive
6684 @itemx max-inline-insns-recursive-auto
6685 Specifies maximum number of instructions out-of-line copy of self recursive inline
6686 function can grow into by performing recursive inlining.
6688 For functions declared inline @option{--param max-inline-insns-recursive} is
6689 taken into account. For function not declared inline, recursive inlining
6690 happens only when @option{-finline-functions} (included in @option{-O3}) is
6691 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
6692 default value is 450.
6694 @item max-inline-recursive-depth
6695 @itemx max-inline-recursive-depth-auto
6696 Specifies maximum recursion depth used by the recursive inlining.
6698 For functions declared inline @option{--param max-inline-recursive-depth} is
6699 taken into account. For function not declared inline, recursive inlining
6700 happens only when @option{-finline-functions} (included in @option{-O3}) is
6701 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
6702 default value is 450.
6704 @item min-inline-recursive-probability
6705 Recursive inlining is profitable only for function having deep recursion
6706 in average and can hurt for function having little recursion depth by
6707 increasing the prologue size or complexity of function body to other
6710 When profile feedback is available (see @option{-fprofile-generate}) the actual
6711 recursion depth can be guessed from probability that function will recurse via
6712 given call expression. This parameter limits inlining only to call expression
6713 whose probability exceeds given threshold (in percents). The default value is
6716 @item inline-call-cost
6717 Specify cost of call instruction relative to simple arithmetics operations
6718 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
6719 functions and at the same time increases size of leaf function that is believed to
6720 reduce function size by being inlined. In effect it increases amount of
6721 inlining for code having large abstraction penalty (many functions that just
6722 pass the arguments to other functions) and decrease inlining for code with low
6723 abstraction penalty. The default value is 16.
6725 @item min-vect-loop-bound
6726 The minimum number of iterations under which a loop will not get vectorized
6727 when @option{-ftree-vectorize} is used. The number of iterations after
6728 vectorization needs to be greater than the value specified by this option
6729 to allow vectorization. The default value is 0.
6731 @item max-unrolled-insns
6732 The maximum number of instructions that a loop should have if that loop
6733 is unrolled, and if the loop is unrolled, it determines how many times
6734 the loop code is unrolled.
6736 @item max-average-unrolled-insns
6737 The maximum number of instructions biased by probabilities of their execution
6738 that a loop should have if that loop is unrolled, and if the loop is unrolled,
6739 it determines how many times the loop code is unrolled.
6741 @item max-unroll-times
6742 The maximum number of unrollings of a single loop.
6744 @item max-peeled-insns
6745 The maximum number of instructions that a loop should have if that loop
6746 is peeled, and if the loop is peeled, it determines how many times
6747 the loop code is peeled.
6749 @item max-peel-times
6750 The maximum number of peelings of a single loop.
6752 @item max-completely-peeled-insns
6753 The maximum number of insns of a completely peeled loop.
6755 @item max-completely-peel-times
6756 The maximum number of iterations of a loop to be suitable for complete peeling.
6758 @item max-unswitch-insns
6759 The maximum number of insns of an unswitched loop.
6761 @item max-unswitch-level
6762 The maximum number of branches unswitched in a single loop.
6765 The minimum cost of an expensive expression in the loop invariant motion.
6767 @item iv-consider-all-candidates-bound
6768 Bound on number of candidates for induction variables below that
6769 all candidates are considered for each use in induction variable
6770 optimizations. Only the most relevant candidates are considered
6771 if there are more candidates, to avoid quadratic time complexity.
6773 @item iv-max-considered-uses
6774 The induction variable optimizations give up on loops that contain more
6775 induction variable uses.
6777 @item iv-always-prune-cand-set-bound
6778 If number of candidates in the set is smaller than this value,
6779 we always try to remove unnecessary ivs from the set during its
6780 optimization when a new iv is added to the set.
6782 @item scev-max-expr-size
6783 Bound on size of expressions used in the scalar evolutions analyzer.
6784 Large expressions slow the analyzer.
6786 @item omega-max-vars
6787 The maximum number of variables in an Omega constraint system.
6788 The default value is 128.
6790 @item omega-max-geqs
6791 The maximum number of inequalities in an Omega constraint system.
6792 The default value is 256.
6795 The maximum number of equalities in an Omega constraint system.
6796 The default value is 128.
6798 @item omega-max-wild-cards
6799 The maximum number of wildcard variables that the Omega solver will
6800 be able to insert. The default value is 18.
6802 @item omega-hash-table-size
6803 The size of the hash table in the Omega solver. The default value is
6806 @item omega-max-keys
6807 The maximal number of keys used by the Omega solver. The default
6810 @item omega-eliminate-redundant-constraints
6811 When set to 1, use expensive methods to eliminate all redundant
6812 constraints. The default value is 0.
6814 @item vect-max-version-for-alignment-checks
6815 The maximum number of runtime checks that can be performed when
6816 doing loop versioning for alignment in the vectorizer. See option
6817 ftree-vect-loop-version for more information.
6819 @item vect-max-version-for-alias-checks
6820 The maximum number of runtime checks that can be performed when
6821 doing loop versioning for alias in the vectorizer. See option
6822 ftree-vect-loop-version for more information.
6824 @item max-iterations-to-track
6826 The maximum number of iterations of a loop the brute force algorithm
6827 for analysis of # of iterations of the loop tries to evaluate.
6829 @item hot-bb-count-fraction
6830 Select fraction of the maximal count of repetitions of basic block in program
6831 given basic block needs to have to be considered hot.
6833 @item hot-bb-frequency-fraction
6834 Select fraction of the maximal frequency of executions of basic block in
6835 function given basic block needs to have to be considered hot
6837 @item max-predicted-iterations
6838 The maximum number of loop iterations we predict statically. This is useful
6839 in cases where function contain single loop with known bound and other loop
6840 with unknown. We predict the known number of iterations correctly, while
6841 the unknown number of iterations average to roughly 10. This means that the
6842 loop without bounds would appear artificially cold relative to the other one.
6844 @item tracer-dynamic-coverage
6845 @itemx tracer-dynamic-coverage-feedback
6847 This value is used to limit superblock formation once the given percentage of
6848 executed instructions is covered. This limits unnecessary code size
6851 The @option{tracer-dynamic-coverage-feedback} is used only when profile
6852 feedback is available. The real profiles (as opposed to statically estimated
6853 ones) are much less balanced allowing the threshold to be larger value.
6855 @item tracer-max-code-growth
6856 Stop tail duplication once code growth has reached given percentage. This is
6857 rather hokey argument, as most of the duplicates will be eliminated later in
6858 cross jumping, so it may be set to much higher values than is the desired code
6861 @item tracer-min-branch-ratio
6863 Stop reverse growth when the reverse probability of best edge is less than this
6864 threshold (in percent).
6866 @item tracer-min-branch-ratio
6867 @itemx tracer-min-branch-ratio-feedback
6869 Stop forward growth if the best edge do have probability lower than this
6872 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
6873 compilation for profile feedback and one for compilation without. The value
6874 for compilation with profile feedback needs to be more conservative (higher) in
6875 order to make tracer effective.
6877 @item max-cse-path-length
6879 Maximum number of basic blocks on path that cse considers. The default is 10.
6882 The maximum instructions CSE process before flushing. The default is 1000.
6884 @item max-aliased-vops
6886 Maximum number of virtual operands per function allowed to represent
6887 aliases before triggering the alias partitioning heuristic. Alias
6888 partitioning reduces compile times and memory consumption needed for
6889 aliasing at the expense of precision loss in alias information. The
6890 default value for this parameter is 100 for -O1, 500 for -O2 and 1000
6893 Notice that if a function contains more memory statements than the
6894 value of this parameter, it is not really possible to achieve this
6895 reduction. In this case, the compiler will use the number of memory
6896 statements as the value for @option{max-aliased-vops}.
6898 @item avg-aliased-vops
6900 Average number of virtual operands per statement allowed to represent
6901 aliases before triggering the alias partitioning heuristic. This
6902 works in conjunction with @option{max-aliased-vops}. If a function
6903 contains more than @option{max-aliased-vops} virtual operators, then
6904 memory symbols will be grouped into memory partitions until either the
6905 total number of virtual operators is below @option{max-aliased-vops}
6906 or the average number of virtual operators per memory statement is
6907 below @option{avg-aliased-vops}. The default value for this parameter
6908 is 1 for -O1 and -O2, and 3 for -O3.
6910 @item ggc-min-expand
6912 GCC uses a garbage collector to manage its own memory allocation. This
6913 parameter specifies the minimum percentage by which the garbage
6914 collector's heap should be allowed to expand between collections.
6915 Tuning this may improve compilation speed; it has no effect on code
6918 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
6919 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
6920 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
6921 GCC is not able to calculate RAM on a particular platform, the lower
6922 bound of 30% is used. Setting this parameter and
6923 @option{ggc-min-heapsize} to zero causes a full collection to occur at
6924 every opportunity. This is extremely slow, but can be useful for
6927 @item ggc-min-heapsize
6929 Minimum size of the garbage collector's heap before it begins bothering
6930 to collect garbage. The first collection occurs after the heap expands
6931 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
6932 tuning this may improve compilation speed, and has no effect on code
6935 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
6936 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
6937 with a lower bound of 4096 (four megabytes) and an upper bound of
6938 131072 (128 megabytes). If GCC is not able to calculate RAM on a
6939 particular platform, the lower bound is used. Setting this parameter
6940 very large effectively disables garbage collection. Setting this
6941 parameter and @option{ggc-min-expand} to zero causes a full collection
6942 to occur at every opportunity.
6944 @item max-reload-search-insns
6945 The maximum number of instruction reload should look backward for equivalent
6946 register. Increasing values mean more aggressive optimization, making the
6947 compile time increase with probably slightly better performance. The default
6950 @item max-cselib-memory-locations
6951 The maximum number of memory locations cselib should take into account.
6952 Increasing values mean more aggressive optimization, making the compile time
6953 increase with probably slightly better performance. The default value is 500.
6955 @item max-flow-memory-locations
6956 Similar as @option{max-cselib-memory-locations} but for dataflow liveness.
6957 The default value is 100.
6959 @item reorder-blocks-duplicate
6960 @itemx reorder-blocks-duplicate-feedback
6962 Used by basic block reordering pass to decide whether to use unconditional
6963 branch or duplicate the code on its destination. Code is duplicated when its
6964 estimated size is smaller than this value multiplied by the estimated size of
6965 unconditional jump in the hot spots of the program.
6967 The @option{reorder-block-duplicate-feedback} is used only when profile
6968 feedback is available and may be set to higher values than
6969 @option{reorder-block-duplicate} since information about the hot spots is more
6972 @item max-sched-ready-insns
6973 The maximum number of instructions ready to be issued the scheduler should
6974 consider at any given time during the first scheduling pass. Increasing
6975 values mean more thorough searches, making the compilation time increase
6976 with probably little benefit. The default value is 100.
6978 @item max-sched-region-blocks
6979 The maximum number of blocks in a region to be considered for
6980 interblock scheduling. The default value is 10.
6982 @item max-sched-region-insns
6983 The maximum number of insns in a region to be considered for
6984 interblock scheduling. The default value is 100.
6987 The minimum probability (in percents) of reaching a source block
6988 for interblock speculative scheduling. The default value is 40.
6990 @item max-sched-extend-regions-iters
6991 The maximum number of iterations through CFG to extend regions.
6992 0 - disable region extension,
6993 N - do at most N iterations.
6994 The default value is 0.
6996 @item max-sched-insn-conflict-delay
6997 The maximum conflict delay for an insn to be considered for speculative motion.
6998 The default value is 3.
7000 @item sched-spec-prob-cutoff
7001 The minimal probability of speculation success (in percents), so that
7002 speculative insn will be scheduled.
7003 The default value is 40.
7005 @item max-last-value-rtl
7007 The maximum size measured as number of RTLs that can be recorded in an expression
7008 in combiner for a pseudo register as last known value of that register. The default
7011 @item integer-share-limit
7012 Small integer constants can use a shared data structure, reducing the
7013 compiler's memory usage and increasing its speed. This sets the maximum
7014 value of a shared integer constant's. The default value is 256.
7016 @item min-virtual-mappings
7017 Specifies the minimum number of virtual mappings in the incremental
7018 SSA updater that should be registered to trigger the virtual mappings
7019 heuristic defined by virtual-mappings-ratio. The default value is
7022 @item virtual-mappings-ratio
7023 If the number of virtual mappings is virtual-mappings-ratio bigger
7024 than the number of virtual symbols to be updated, then the incremental
7025 SSA updater switches to a full update for those symbols. The default
7028 @item ssp-buffer-size
7029 The minimum size of buffers (i.e. arrays) that will receive stack smashing
7030 protection when @option{-fstack-protection} is used.
7032 @item max-jump-thread-duplication-stmts
7033 Maximum number of statements allowed in a block that needs to be
7034 duplicated when threading jumps.
7036 @item max-fields-for-field-sensitive
7037 Maximum number of fields in a structure we will treat in
7038 a field sensitive manner during pointer analysis.
7040 @item prefetch-latency
7041 Estimate on average number of instructions that are executed before
7042 prefetch finishes. The distance we prefetch ahead is proportional
7043 to this constant. Increasing this number may also lead to less
7044 streams being prefetched (see @option{simultaneous-prefetches}).
7046 @item simultaneous-prefetches
7047 Maximum number of prefetches that can run at the same time.
7049 @item l1-cache-line-size
7050 The size of cache line in L1 cache, in bytes.
7053 The size of L1 cache, in kilobytes.
7056 The size of L2 cache, in kilobytes.
7058 @item use-canonical-types
7059 Whether the compiler should use the ``canonical'' type system. By
7060 default, this should always be 1, which uses a more efficient internal
7061 mechanism for comparing types in C++ and Objective-C++. However, if
7062 bugs in the canonical type system are causing compilation failures,
7063 set this value to 0 to disable canonical types.
7068 @node Preprocessor Options
7069 @section Options Controlling the Preprocessor
7070 @cindex preprocessor options
7071 @cindex options, preprocessor
7073 These options control the C preprocessor, which is run on each C source
7074 file before actual compilation.
7076 If you use the @option{-E} option, nothing is done except preprocessing.
7077 Some of these options make sense only together with @option{-E} because
7078 they cause the preprocessor output to be unsuitable for actual
7083 You can use @option{-Wp,@var{option}} to bypass the compiler driver
7084 and pass @var{option} directly through to the preprocessor. If
7085 @var{option} contains commas, it is split into multiple options at the
7086 commas. However, many options are modified, translated or interpreted
7087 by the compiler driver before being passed to the preprocessor, and
7088 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
7089 interface is undocumented and subject to change, so whenever possible
7090 you should avoid using @option{-Wp} and let the driver handle the
7093 @item -Xpreprocessor @var{option}
7094 @opindex preprocessor
7095 Pass @var{option} as an option to the preprocessor. You can use this to
7096 supply system-specific preprocessor options which GCC does not know how to
7099 If you want to pass an option that takes an argument, you must use
7100 @option{-Xpreprocessor} twice, once for the option and once for the argument.
7103 @include cppopts.texi
7105 @node Assembler Options
7106 @section Passing Options to the Assembler
7108 @c prevent bad page break with this line
7109 You can pass options to the assembler.
7112 @item -Wa,@var{option}
7114 Pass @var{option} as an option to the assembler. If @var{option}
7115 contains commas, it is split into multiple options at the commas.
7117 @item -Xassembler @var{option}
7119 Pass @var{option} as an option to the assembler. You can use this to
7120 supply system-specific assembler options which GCC does not know how to
7123 If you want to pass an option that takes an argument, you must use
7124 @option{-Xassembler} twice, once for the option and once for the argument.
7129 @section Options for Linking
7130 @cindex link options
7131 @cindex options, linking
7133 These options come into play when the compiler links object files into
7134 an executable output file. They are meaningless if the compiler is
7135 not doing a link step.
7139 @item @var{object-file-name}
7140 A file name that does not end in a special recognized suffix is
7141 considered to name an object file or library. (Object files are
7142 distinguished from libraries by the linker according to the file
7143 contents.) If linking is done, these object files are used as input
7152 If any of these options is used, then the linker is not run, and
7153 object file names should not be used as arguments. @xref{Overall
7157 @item -l@var{library}
7158 @itemx -l @var{library}
7160 Search the library named @var{library} when linking. (The second
7161 alternative with the library as a separate argument is only for
7162 POSIX compliance and is not recommended.)
7164 It makes a difference where in the command you write this option; the
7165 linker searches and processes libraries and object files in the order they
7166 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
7167 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
7168 to functions in @samp{z}, those functions may not be loaded.
7170 The linker searches a standard list of directories for the library,
7171 which is actually a file named @file{lib@var{library}.a}. The linker
7172 then uses this file as if it had been specified precisely by name.
7174 The directories searched include several standard system directories
7175 plus any that you specify with @option{-L}.
7177 Normally the files found this way are library files---archive files
7178 whose members are object files. The linker handles an archive file by
7179 scanning through it for members which define symbols that have so far
7180 been referenced but not defined. But if the file that is found is an
7181 ordinary object file, it is linked in the usual fashion. The only
7182 difference between using an @option{-l} option and specifying a file name
7183 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
7184 and searches several directories.
7188 You need this special case of the @option{-l} option in order to
7189 link an Objective-C or Objective-C++ program.
7192 @opindex nostartfiles
7193 Do not use the standard system startup files when linking.
7194 The standard system libraries are used normally, unless @option{-nostdlib}
7195 or @option{-nodefaultlibs} is used.
7197 @item -nodefaultlibs
7198 @opindex nodefaultlibs
7199 Do not use the standard system libraries when linking.
7200 Only the libraries you specify will be passed to the linker.
7201 The standard startup files are used normally, unless @option{-nostartfiles}
7202 is used. The compiler may generate calls to @code{memcmp},
7203 @code{memset}, @code{memcpy} and @code{memmove}.
7204 These entries are usually resolved by entries in
7205 libc. These entry points should be supplied through some other
7206 mechanism when this option is specified.
7210 Do not use the standard system startup files or libraries when linking.
7211 No startup files and only the libraries you specify will be passed to
7212 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
7213 @code{memcpy} and @code{memmove}.
7214 These entries are usually resolved by entries in
7215 libc. These entry points should be supplied through some other
7216 mechanism when this option is specified.
7218 @cindex @option{-lgcc}, use with @option{-nostdlib}
7219 @cindex @option{-nostdlib} and unresolved references
7220 @cindex unresolved references and @option{-nostdlib}
7221 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
7222 @cindex @option{-nodefaultlibs} and unresolved references
7223 @cindex unresolved references and @option{-nodefaultlibs}
7224 One of the standard libraries bypassed by @option{-nostdlib} and
7225 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
7226 that GCC uses to overcome shortcomings of particular machines, or special
7227 needs for some languages.
7228 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
7229 Collection (GCC) Internals},
7230 for more discussion of @file{libgcc.a}.)
7231 In most cases, you need @file{libgcc.a} even when you want to avoid
7232 other standard libraries. In other words, when you specify @option{-nostdlib}
7233 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
7234 This ensures that you have no unresolved references to internal GCC
7235 library subroutines. (For example, @samp{__main}, used to ensure C++
7236 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
7237 GNU Compiler Collection (GCC) Internals}.)
7241 Produce a position independent executable on targets which support it.
7242 For predictable results, you must also specify the same set of options
7243 that were used to generate code (@option{-fpie}, @option{-fPIE},
7244 or model suboptions) when you specify this option.
7248 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
7249 that support it. This instructs the linker to add all symbols, not
7250 only used ones, to the dynamic symbol table. This option is needed
7251 for some uses of @code{dlopen} or to allow obtaining backtraces
7252 from within a program.
7256 Remove all symbol table and relocation information from the executable.
7260 On systems that support dynamic linking, this prevents linking with the shared
7261 libraries. On other systems, this option has no effect.
7265 Produce a shared object which can then be linked with other objects to
7266 form an executable. Not all systems support this option. For predictable
7267 results, you must also specify the same set of options that were used to
7268 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
7269 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
7270 needs to build supplementary stub code for constructors to work. On
7271 multi-libbed systems, @samp{gcc -shared} must select the correct support
7272 libraries to link against. Failing to supply the correct flags may lead
7273 to subtle defects. Supplying them in cases where they are not necessary
7276 @item -shared-libgcc
7277 @itemx -static-libgcc
7278 @opindex shared-libgcc
7279 @opindex static-libgcc
7280 On systems that provide @file{libgcc} as a shared library, these options
7281 force the use of either the shared or static version respectively.
7282 If no shared version of @file{libgcc} was built when the compiler was
7283 configured, these options have no effect.
7285 There are several situations in which an application should use the
7286 shared @file{libgcc} instead of the static version. The most common
7287 of these is when the application wishes to throw and catch exceptions
7288 across different shared libraries. In that case, each of the libraries
7289 as well as the application itself should use the shared @file{libgcc}.
7291 Therefore, the G++ and GCJ drivers automatically add
7292 @option{-shared-libgcc} whenever you build a shared library or a main
7293 executable, because C++ and Java programs typically use exceptions, so
7294 this is the right thing to do.
7296 If, instead, you use the GCC driver to create shared libraries, you may
7297 find that they will not always be linked with the shared @file{libgcc}.
7298 If GCC finds, at its configuration time, that you have a non-GNU linker
7299 or a GNU linker that does not support option @option{--eh-frame-hdr},
7300 it will link the shared version of @file{libgcc} into shared libraries
7301 by default. Otherwise, it will take advantage of the linker and optimize
7302 away the linking with the shared version of @file{libgcc}, linking with
7303 the static version of libgcc by default. This allows exceptions to
7304 propagate through such shared libraries, without incurring relocation
7305 costs at library load time.
7307 However, if a library or main executable is supposed to throw or catch
7308 exceptions, you must link it using the G++ or GCJ driver, as appropriate
7309 for the languages used in the program, or using the option
7310 @option{-shared-libgcc}, such that it is linked with the shared
7315 Bind references to global symbols when building a shared object. Warn
7316 about any unresolved references (unless overridden by the link editor
7317 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
7320 @item -Xlinker @var{option}
7322 Pass @var{option} as an option to the linker. You can use this to
7323 supply system-specific linker options which GCC does not know how to
7326 If you want to pass an option that takes an argument, you must use
7327 @option{-Xlinker} twice, once for the option and once for the argument.
7328 For example, to pass @option{-assert definitions}, you must write
7329 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
7330 @option{-Xlinker "-assert definitions"}, because this passes the entire
7331 string as a single argument, which is not what the linker expects.
7333 @item -Wl,@var{option}
7335 Pass @var{option} as an option to the linker. If @var{option} contains
7336 commas, it is split into multiple options at the commas.
7338 @item -u @var{symbol}
7340 Pretend the symbol @var{symbol} is undefined, to force linking of
7341 library modules to define it. You can use @option{-u} multiple times with
7342 different symbols to force loading of additional library modules.
7345 @node Directory Options
7346 @section Options for Directory Search
7347 @cindex directory options
7348 @cindex options, directory search
7351 These options specify directories to search for header files, for
7352 libraries and for parts of the compiler:
7357 Add the directory @var{dir} to the head of the list of directories to be
7358 searched for header files. This can be used to override a system header
7359 file, substituting your own version, since these directories are
7360 searched before the system header file directories. However, you should
7361 not use this option to add directories that contain vendor-supplied
7362 system header files (use @option{-isystem} for that). If you use more than
7363 one @option{-I} option, the directories are scanned in left-to-right
7364 order; the standard system directories come after.
7366 If a standard system include directory, or a directory specified with
7367 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
7368 option will be ignored. The directory will still be searched but as a
7369 system directory at its normal position in the system include chain.
7370 This is to ensure that GCC's procedure to fix buggy system headers and
7371 the ordering for the include_next directive are not inadvertently changed.
7372 If you really need to change the search order for system directories,
7373 use the @option{-nostdinc} and/or @option{-isystem} options.
7375 @item -iquote@var{dir}
7377 Add the directory @var{dir} to the head of the list of directories to
7378 be searched for header files only for the case of @samp{#include
7379 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
7380 otherwise just like @option{-I}.
7384 Add directory @var{dir} to the list of directories to be searched
7387 @item -B@var{prefix}
7389 This option specifies where to find the executables, libraries,
7390 include files, and data files of the compiler itself.
7392 The compiler driver program runs one or more of the subprograms
7393 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
7394 @var{prefix} as a prefix for each program it tries to run, both with and
7395 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
7397 For each subprogram to be run, the compiler driver first tries the
7398 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
7399 was not specified, the driver tries two standard prefixes, which are
7400 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
7401 those results in a file name that is found, the unmodified program
7402 name is searched for using the directories specified in your
7403 @env{PATH} environment variable.
7405 The compiler will check to see if the path provided by the @option{-B}
7406 refers to a directory, and if necessary it will add a directory
7407 separator character at the end of the path.
7409 @option{-B} prefixes that effectively specify directory names also apply
7410 to libraries in the linker, because the compiler translates these
7411 options into @option{-L} options for the linker. They also apply to
7412 includes files in the preprocessor, because the compiler translates these
7413 options into @option{-isystem} options for the preprocessor. In this case,
7414 the compiler appends @samp{include} to the prefix.
7416 The run-time support file @file{libgcc.a} can also be searched for using
7417 the @option{-B} prefix, if needed. If it is not found there, the two
7418 standard prefixes above are tried, and that is all. The file is left
7419 out of the link if it is not found by those means.
7421 Another way to specify a prefix much like the @option{-B} prefix is to use
7422 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
7425 As a special kludge, if the path provided by @option{-B} is
7426 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
7427 9, then it will be replaced by @file{[dir/]include}. This is to help
7428 with boot-strapping the compiler.
7430 @item -specs=@var{file}
7432 Process @var{file} after the compiler reads in the standard @file{specs}
7433 file, in order to override the defaults that the @file{gcc} driver
7434 program uses when determining what switches to pass to @file{cc1},
7435 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
7436 @option{-specs=@var{file}} can be specified on the command line, and they
7437 are processed in order, from left to right.
7439 @item --sysroot=@var{dir}
7441 Use @var{dir} as the logical root directory for headers and libraries.
7442 For example, if the compiler would normally search for headers in
7443 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
7444 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
7446 If you use both this option and the @option{-isysroot} option, then
7447 the @option{--sysroot} option will apply to libraries, but the
7448 @option{-isysroot} option will apply to header files.
7450 The GNU linker (beginning with version 2.16) has the necessary support
7451 for this option. If your linker does not support this option, the
7452 header file aspect of @option{--sysroot} will still work, but the
7453 library aspect will not.
7457 This option has been deprecated. Please use @option{-iquote} instead for
7458 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
7459 Any directories you specify with @option{-I} options before the @option{-I-}
7460 option are searched only for the case of @samp{#include "@var{file}"};
7461 they are not searched for @samp{#include <@var{file}>}.
7463 If additional directories are specified with @option{-I} options after
7464 the @option{-I-}, these directories are searched for all @samp{#include}
7465 directives. (Ordinarily @emph{all} @option{-I} directories are used
7468 In addition, the @option{-I-} option inhibits the use of the current
7469 directory (where the current input file came from) as the first search
7470 directory for @samp{#include "@var{file}"}. There is no way to
7471 override this effect of @option{-I-}. With @option{-I.} you can specify
7472 searching the directory which was current when the compiler was
7473 invoked. That is not exactly the same as what the preprocessor does
7474 by default, but it is often satisfactory.
7476 @option{-I-} does not inhibit the use of the standard system directories
7477 for header files. Thus, @option{-I-} and @option{-nostdinc} are
7484 @section Specifying subprocesses and the switches to pass to them
7487 @command{gcc} is a driver program. It performs its job by invoking a
7488 sequence of other programs to do the work of compiling, assembling and
7489 linking. GCC interprets its command-line parameters and uses these to
7490 deduce which programs it should invoke, and which command-line options
7491 it ought to place on their command lines. This behavior is controlled
7492 by @dfn{spec strings}. In most cases there is one spec string for each
7493 program that GCC can invoke, but a few programs have multiple spec
7494 strings to control their behavior. The spec strings built into GCC can
7495 be overridden by using the @option{-specs=} command-line switch to specify
7498 @dfn{Spec files} are plaintext files that are used to construct spec
7499 strings. They consist of a sequence of directives separated by blank
7500 lines. The type of directive is determined by the first non-whitespace
7501 character on the line and it can be one of the following:
7504 @item %@var{command}
7505 Issues a @var{command} to the spec file processor. The commands that can
7509 @item %include <@var{file}>
7511 Search for @var{file} and insert its text at the current point in the
7514 @item %include_noerr <@var{file}>
7515 @cindex %include_noerr
7516 Just like @samp{%include}, but do not generate an error message if the include
7517 file cannot be found.
7519 @item %rename @var{old_name} @var{new_name}
7521 Rename the spec string @var{old_name} to @var{new_name}.
7525 @item *[@var{spec_name}]:
7526 This tells the compiler to create, override or delete the named spec
7527 string. All lines after this directive up to the next directive or
7528 blank line are considered to be the text for the spec string. If this
7529 results in an empty string then the spec will be deleted. (Or, if the
7530 spec did not exist, then nothing will happened.) Otherwise, if the spec
7531 does not currently exist a new spec will be created. If the spec does
7532 exist then its contents will be overridden by the text of this
7533 directive, unless the first character of that text is the @samp{+}
7534 character, in which case the text will be appended to the spec.
7536 @item [@var{suffix}]:
7537 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
7538 and up to the next directive or blank line are considered to make up the
7539 spec string for the indicated suffix. When the compiler encounters an
7540 input file with the named suffix, it will processes the spec string in
7541 order to work out how to compile that file. For example:
7548 This says that any input file whose name ends in @samp{.ZZ} should be
7549 passed to the program @samp{z-compile}, which should be invoked with the
7550 command-line switch @option{-input} and with the result of performing the
7551 @samp{%i} substitution. (See below.)
7553 As an alternative to providing a spec string, the text that follows a
7554 suffix directive can be one of the following:
7557 @item @@@var{language}
7558 This says that the suffix is an alias for a known @var{language}. This is
7559 similar to using the @option{-x} command-line switch to GCC to specify a
7560 language explicitly. For example:
7567 Says that .ZZ files are, in fact, C++ source files.
7570 This causes an error messages saying:
7573 @var{name} compiler not installed on this system.
7577 GCC already has an extensive list of suffixes built into it.
7578 This directive will add an entry to the end of the list of suffixes, but
7579 since the list is searched from the end backwards, it is effectively
7580 possible to override earlier entries using this technique.
7584 GCC has the following spec strings built into it. Spec files can
7585 override these strings or create their own. Note that individual
7586 targets can also add their own spec strings to this list.
7589 asm Options to pass to the assembler
7590 asm_final Options to pass to the assembler post-processor
7591 cpp Options to pass to the C preprocessor
7592 cc1 Options to pass to the C compiler
7593 cc1plus Options to pass to the C++ compiler
7594 endfile Object files to include at the end of the link
7595 link Options to pass to the linker
7596 lib Libraries to include on the command line to the linker
7597 libgcc Decides which GCC support library to pass to the linker
7598 linker Sets the name of the linker
7599 predefines Defines to be passed to the C preprocessor
7600 signed_char Defines to pass to CPP to say whether @code{char} is signed
7602 startfile Object files to include at the start of the link
7605 Here is a small example of a spec file:
7611 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
7614 This example renames the spec called @samp{lib} to @samp{old_lib} and
7615 then overrides the previous definition of @samp{lib} with a new one.
7616 The new definition adds in some extra command-line options before
7617 including the text of the old definition.
7619 @dfn{Spec strings} are a list of command-line options to be passed to their
7620 corresponding program. In addition, the spec strings can contain
7621 @samp{%}-prefixed sequences to substitute variable text or to
7622 conditionally insert text into the command line. Using these constructs
7623 it is possible to generate quite complex command lines.
7625 Here is a table of all defined @samp{%}-sequences for spec
7626 strings. Note that spaces are not generated automatically around the
7627 results of expanding these sequences. Therefore you can concatenate them
7628 together or combine them with constant text in a single argument.
7632 Substitute one @samp{%} into the program name or argument.
7635 Substitute the name of the input file being processed.
7638 Substitute the basename of the input file being processed.
7639 This is the substring up to (and not including) the last period
7640 and not including the directory.
7643 This is the same as @samp{%b}, but include the file suffix (text after
7647 Marks the argument containing or following the @samp{%d} as a
7648 temporary file name, so that that file will be deleted if GCC exits
7649 successfully. Unlike @samp{%g}, this contributes no text to the
7652 @item %g@var{suffix}
7653 Substitute a file name that has suffix @var{suffix} and is chosen
7654 once per compilation, and mark the argument in the same way as
7655 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
7656 name is now chosen in a way that is hard to predict even when previously
7657 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
7658 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
7659 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
7660 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
7661 was simply substituted with a file name chosen once per compilation,
7662 without regard to any appended suffix (which was therefore treated
7663 just like ordinary text), making such attacks more likely to succeed.
7665 @item %u@var{suffix}
7666 Like @samp{%g}, but generates a new temporary file name even if
7667 @samp{%u@var{suffix}} was already seen.
7669 @item %U@var{suffix}
7670 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
7671 new one if there is no such last file name. In the absence of any
7672 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
7673 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
7674 would involve the generation of two distinct file names, one
7675 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
7676 simply substituted with a file name chosen for the previous @samp{%u},
7677 without regard to any appended suffix.
7679 @item %j@var{suffix}
7680 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
7681 writable, and if save-temps is off; otherwise, substitute the name
7682 of a temporary file, just like @samp{%u}. This temporary file is not
7683 meant for communication between processes, but rather as a junk
7686 @item %|@var{suffix}
7687 @itemx %m@var{suffix}
7688 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
7689 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
7690 all. These are the two most common ways to instruct a program that it
7691 should read from standard input or write to standard output. If you
7692 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
7693 construct: see for example @file{f/lang-specs.h}.
7695 @item %.@var{SUFFIX}
7696 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
7697 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
7698 terminated by the next space or %.
7701 Marks the argument containing or following the @samp{%w} as the
7702 designated output file of this compilation. This puts the argument
7703 into the sequence of arguments that @samp{%o} will substitute later.
7706 Substitutes the names of all the output files, with spaces
7707 automatically placed around them. You should write spaces
7708 around the @samp{%o} as well or the results are undefined.
7709 @samp{%o} is for use in the specs for running the linker.
7710 Input files whose names have no recognized suffix are not compiled
7711 at all, but they are included among the output files, so they will
7715 Substitutes the suffix for object files. Note that this is
7716 handled specially when it immediately follows @samp{%g, %u, or %U},
7717 because of the need for those to form complete file names. The
7718 handling is such that @samp{%O} is treated exactly as if it had already
7719 been substituted, except that @samp{%g, %u, and %U} do not currently
7720 support additional @var{suffix} characters following @samp{%O} as they would
7721 following, for example, @samp{.o}.
7724 Substitutes the standard macro predefinitions for the
7725 current target machine. Use this when running @code{cpp}.
7728 Like @samp{%p}, but puts @samp{__} before and after the name of each
7729 predefined macro, except for macros that start with @samp{__} or with
7730 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
7734 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
7735 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
7736 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
7737 and @option{-imultilib} as necessary.
7740 Current argument is the name of a library or startup file of some sort.
7741 Search for that file in a standard list of directories and substitute
7742 the full name found.
7745 Print @var{str} as an error message. @var{str} is terminated by a newline.
7746 Use this when inconsistent options are detected.
7749 Substitute the contents of spec string @var{name} at this point.
7752 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
7754 @item %x@{@var{option}@}
7755 Accumulate an option for @samp{%X}.
7758 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
7762 Output the accumulated assembler options specified by @option{-Wa}.
7765 Output the accumulated preprocessor options specified by @option{-Wp}.
7768 Process the @code{asm} spec. This is used to compute the
7769 switches to be passed to the assembler.
7772 Process the @code{asm_final} spec. This is a spec string for
7773 passing switches to an assembler post-processor, if such a program is
7777 Process the @code{link} spec. This is the spec for computing the
7778 command line passed to the linker. Typically it will make use of the
7779 @samp{%L %G %S %D and %E} sequences.
7782 Dump out a @option{-L} option for each directory that GCC believes might
7783 contain startup files. If the target supports multilibs then the
7784 current multilib directory will be prepended to each of these paths.
7787 Process the @code{lib} spec. This is a spec string for deciding which
7788 libraries should be included on the command line to the linker.
7791 Process the @code{libgcc} spec. This is a spec string for deciding
7792 which GCC support library should be included on the command line to the linker.
7795 Process the @code{startfile} spec. This is a spec for deciding which
7796 object files should be the first ones passed to the linker. Typically
7797 this might be a file named @file{crt0.o}.
7800 Process the @code{endfile} spec. This is a spec string that specifies
7801 the last object files that will be passed to the linker.
7804 Process the @code{cpp} spec. This is used to construct the arguments
7805 to be passed to the C preprocessor.
7808 Process the @code{cc1} spec. This is used to construct the options to be
7809 passed to the actual C compiler (@samp{cc1}).
7812 Process the @code{cc1plus} spec. This is used to construct the options to be
7813 passed to the actual C++ compiler (@samp{cc1plus}).
7816 Substitute the variable part of a matched option. See below.
7817 Note that each comma in the substituted string is replaced by
7821 Remove all occurrences of @code{-S} from the command line. Note---this
7822 command is position dependent. @samp{%} commands in the spec string
7823 before this one will see @code{-S}, @samp{%} commands in the spec string
7824 after this one will not.
7826 @item %:@var{function}(@var{args})
7827 Call the named function @var{function}, passing it @var{args}.
7828 @var{args} is first processed as a nested spec string, then split
7829 into an argument vector in the usual fashion. The function returns
7830 a string which is processed as if it had appeared literally as part
7831 of the current spec.
7833 The following built-in spec functions are provided:
7837 The @code{getenv} spec function takes two arguments: an environment
7838 variable name and a string. If the environment variable is not
7839 defined, a fatal error is issued. Otherwise, the return value is the
7840 value of the environment variable concatenated with the string. For
7841 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
7844 %:getenv(TOPDIR /include)
7847 expands to @file{/path/to/top/include}.
7849 @item @code{if-exists}
7850 The @code{if-exists} spec function takes one argument, an absolute
7851 pathname to a file. If the file exists, @code{if-exists} returns the
7852 pathname. Here is a small example of its usage:
7856 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
7859 @item @code{if-exists-else}
7860 The @code{if-exists-else} spec function is similar to the @code{if-exists}
7861 spec function, except that it takes two arguments. The first argument is
7862 an absolute pathname to a file. If the file exists, @code{if-exists-else}
7863 returns the pathname. If it does not exist, it returns the second argument.
7864 This way, @code{if-exists-else} can be used to select one file or another,
7865 based on the existence of the first. Here is a small example of its usage:
7869 crt0%O%s %:if-exists(crti%O%s) \
7870 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
7873 @item @code{replace-outfile}
7874 The @code{replace-outfile} spec function takes two arguments. It looks for the
7875 first argument in the outfiles array and replaces it with the second argument. Here
7876 is a small example of its usage:
7879 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
7882 @item @code{print-asm-header}
7883 The @code{print-asm-header} function takes no arguments and simply
7884 prints a banner like:
7890 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
7893 It is used to separate compiler options from assembler options
7894 in the @option{--target-help} output.
7898 Substitutes the @code{-S} switch, if that switch was given to GCC@.
7899 If that switch was not specified, this substitutes nothing. Note that
7900 the leading dash is omitted when specifying this option, and it is
7901 automatically inserted if the substitution is performed. Thus the spec
7902 string @samp{%@{foo@}} would match the command-line option @option{-foo}
7903 and would output the command line option @option{-foo}.
7905 @item %W@{@code{S}@}
7906 Like %@{@code{S}@} but mark last argument supplied within as a file to be
7909 @item %@{@code{S}*@}
7910 Substitutes all the switches specified to GCC whose names start
7911 with @code{-S}, but which also take an argument. This is used for
7912 switches like @option{-o}, @option{-D}, @option{-I}, etc.
7913 GCC considers @option{-o foo} as being
7914 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
7915 text, including the space. Thus two arguments would be generated.
7917 @item %@{@code{S}*&@code{T}*@}
7918 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
7919 (the order of @code{S} and @code{T} in the spec is not significant).
7920 There can be any number of ampersand-separated variables; for each the
7921 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
7923 @item %@{@code{S}:@code{X}@}
7924 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
7926 @item %@{!@code{S}:@code{X}@}
7927 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
7929 @item %@{@code{S}*:@code{X}@}
7930 Substitutes @code{X} if one or more switches whose names start with
7931 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
7932 once, no matter how many such switches appeared. However, if @code{%*}
7933 appears somewhere in @code{X}, then @code{X} will be substituted once
7934 for each matching switch, with the @code{%*} replaced by the part of
7935 that switch that matched the @code{*}.
7937 @item %@{.@code{S}:@code{X}@}
7938 Substitutes @code{X}, if processing a file with suffix @code{S}.
7940 @item %@{!.@code{S}:@code{X}@}
7941 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
7943 @item %@{,@code{S}:@code{X}@}
7944 Substitutes @code{X}, if processing a file for language @code{S}.
7946 @item %@{!,@code{S}:@code{X}@}
7947 Substitutes @code{X}, if not processing a file for language @code{S}.
7949 @item %@{@code{S}|@code{P}:@code{X}@}
7950 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
7951 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
7952 @code{*} sequences as well, although they have a stronger binding than
7953 the @samp{|}. If @code{%*} appears in @code{X}, all of the
7954 alternatives must be starred, and only the first matching alternative
7957 For example, a spec string like this:
7960 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
7963 will output the following command-line options from the following input
7964 command-line options:
7969 -d fred.c -foo -baz -boggle
7970 -d jim.d -bar -baz -boggle
7973 @item %@{S:X; T:Y; :D@}
7975 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
7976 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
7977 be as many clauses as you need. This may be combined with @code{.},
7978 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
7983 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
7984 construct may contain other nested @samp{%} constructs or spaces, or
7985 even newlines. They are processed as usual, as described above.
7986 Trailing white space in @code{X} is ignored. White space may also
7987 appear anywhere on the left side of the colon in these constructs,
7988 except between @code{.} or @code{*} and the corresponding word.
7990 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
7991 handled specifically in these constructs. If another value of
7992 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
7993 @option{-W} switch is found later in the command line, the earlier
7994 switch value is ignored, except with @{@code{S}*@} where @code{S} is
7995 just one letter, which passes all matching options.
7997 The character @samp{|} at the beginning of the predicate text is used to
7998 indicate that a command should be piped to the following command, but
7999 only if @option{-pipe} is specified.
8001 It is built into GCC which switches take arguments and which do not.
8002 (You might think it would be useful to generalize this to allow each
8003 compiler's spec to say which switches take arguments. But this cannot
8004 be done in a consistent fashion. GCC cannot even decide which input
8005 files have been specified without knowing which switches take arguments,
8006 and it must know which input files to compile in order to tell which
8009 GCC also knows implicitly that arguments starting in @option{-l} are to be
8010 treated as compiler output files, and passed to the linker in their
8011 proper position among the other output files.
8013 @c man begin OPTIONS
8015 @node Target Options
8016 @section Specifying Target Machine and Compiler Version
8017 @cindex target options
8018 @cindex cross compiling
8019 @cindex specifying machine version
8020 @cindex specifying compiler version and target machine
8021 @cindex compiler version, specifying
8022 @cindex target machine, specifying
8024 The usual way to run GCC is to run the executable called @file{gcc}, or
8025 @file{<machine>-gcc} when cross-compiling, or
8026 @file{<machine>-gcc-<version>} to run a version other than the one that
8027 was installed last. Sometimes this is inconvenient, so GCC provides
8028 options that will switch to another cross-compiler or version.
8031 @item -b @var{machine}
8033 The argument @var{machine} specifies the target machine for compilation.
8035 The value to use for @var{machine} is the same as was specified as the
8036 machine type when configuring GCC as a cross-compiler. For
8037 example, if a cross-compiler was configured with @samp{configure
8038 arm-elf}, meaning to compile for an arm processor with elf binaries,
8039 then you would specify @option{-b arm-elf} to run that cross compiler.
8040 Because there are other options beginning with @option{-b}, the
8041 configuration must contain a hyphen.
8043 @item -V @var{version}
8045 The argument @var{version} specifies which version of GCC to run.
8046 This is useful when multiple versions are installed. For example,
8047 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
8050 The @option{-V} and @option{-b} options work by running the
8051 @file{<machine>-gcc-<version>} executable, so there's no real reason to
8052 use them if you can just run that directly.
8054 @node Submodel Options
8055 @section Hardware Models and Configurations
8056 @cindex submodel options
8057 @cindex specifying hardware config
8058 @cindex hardware models and configurations, specifying
8059 @cindex machine dependent options
8061 Earlier we discussed the standard option @option{-b} which chooses among
8062 different installed compilers for completely different target
8063 machines, such as VAX vs.@: 68000 vs.@: 80386.
8065 In addition, each of these target machine types can have its own
8066 special options, starting with @samp{-m}, to choose among various
8067 hardware models or configurations---for example, 68010 vs 68020,
8068 floating coprocessor or none. A single installed version of the
8069 compiler can compile for any model or configuration, according to the
8072 Some configurations of the compiler also support additional special
8073 options, usually for compatibility with other compilers on the same
8076 @c This list is ordered alphanumerically by subsection name.
8077 @c It should be the same order and spelling as these options are listed
8078 @c in Machine Dependent Options
8084 * Blackfin Options::
8088 * DEC Alpha Options::
8089 * DEC Alpha/VMS Options::
8091 * GNU/Linux Options::
8094 * i386 and x86-64 Options::
8107 * RS/6000 and PowerPC Options::
8108 * S/390 and zSeries Options::
8113 * System V Options::
8114 * TMS320C3x/C4x Options::
8119 * Xstormy16 Options::
8125 @subsection ARC Options
8128 These options are defined for ARC implementations:
8133 Compile code for little endian mode. This is the default.
8137 Compile code for big endian mode.
8140 @opindex mmangle-cpu
8141 Prepend the name of the cpu to all public symbol names.
8142 In multiple-processor systems, there are many ARC variants with different
8143 instruction and register set characteristics. This flag prevents code
8144 compiled for one cpu to be linked with code compiled for another.
8145 No facility exists for handling variants that are ``almost identical''.
8146 This is an all or nothing option.
8148 @item -mcpu=@var{cpu}
8150 Compile code for ARC variant @var{cpu}.
8151 Which variants are supported depend on the configuration.
8152 All variants support @option{-mcpu=base}, this is the default.
8154 @item -mtext=@var{text-section}
8155 @itemx -mdata=@var{data-section}
8156 @itemx -mrodata=@var{readonly-data-section}
8160 Put functions, data, and readonly data in @var{text-section},
8161 @var{data-section}, and @var{readonly-data-section} respectively
8162 by default. This can be overridden with the @code{section} attribute.
8163 @xref{Variable Attributes}.
8168 @subsection ARM Options
8171 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
8175 @item -mabi=@var{name}
8177 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
8178 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
8181 @opindex mapcs-frame
8182 Generate a stack frame that is compliant with the ARM Procedure Call
8183 Standard for all functions, even if this is not strictly necessary for
8184 correct execution of the code. Specifying @option{-fomit-frame-pointer}
8185 with this option will cause the stack frames not to be generated for
8186 leaf functions. The default is @option{-mno-apcs-frame}.
8190 This is a synonym for @option{-mapcs-frame}.
8193 @c not currently implemented
8194 @item -mapcs-stack-check
8195 @opindex mapcs-stack-check
8196 Generate code to check the amount of stack space available upon entry to
8197 every function (that actually uses some stack space). If there is
8198 insufficient space available then either the function
8199 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
8200 called, depending upon the amount of stack space required. The run time
8201 system is required to provide these functions. The default is
8202 @option{-mno-apcs-stack-check}, since this produces smaller code.
8204 @c not currently implemented
8206 @opindex mapcs-float
8207 Pass floating point arguments using the float point registers. This is
8208 one of the variants of the APCS@. This option is recommended if the
8209 target hardware has a floating point unit or if a lot of floating point
8210 arithmetic is going to be performed by the code. The default is
8211 @option{-mno-apcs-float}, since integer only code is slightly increased in
8212 size if @option{-mapcs-float} is used.
8214 @c not currently implemented
8215 @item -mapcs-reentrant
8216 @opindex mapcs-reentrant
8217 Generate reentrant, position independent code. The default is
8218 @option{-mno-apcs-reentrant}.
8221 @item -mthumb-interwork
8222 @opindex mthumb-interwork
8223 Generate code which supports calling between the ARM and Thumb
8224 instruction sets. Without this option the two instruction sets cannot
8225 be reliably used inside one program. The default is
8226 @option{-mno-thumb-interwork}, since slightly larger code is generated
8227 when @option{-mthumb-interwork} is specified.
8229 @item -mno-sched-prolog
8230 @opindex mno-sched-prolog
8231 Prevent the reordering of instructions in the function prolog, or the
8232 merging of those instruction with the instructions in the function's
8233 body. This means that all functions will start with a recognizable set
8234 of instructions (or in fact one of a choice from a small set of
8235 different function prologues), and this information can be used to
8236 locate the start if functions inside an executable piece of code. The
8237 default is @option{-msched-prolog}.
8240 @opindex mhard-float
8241 Generate output containing floating point instructions. This is the
8245 @opindex msoft-float
8246 Generate output containing library calls for floating point.
8247 @strong{Warning:} the requisite libraries are not available for all ARM
8248 targets. Normally the facilities of the machine's usual C compiler are
8249 used, but this cannot be done directly in cross-compilation. You must make
8250 your own arrangements to provide suitable library functions for
8253 @option{-msoft-float} changes the calling convention in the output file;
8254 therefore, it is only useful if you compile @emph{all} of a program with
8255 this option. In particular, you need to compile @file{libgcc.a}, the
8256 library that comes with GCC, with @option{-msoft-float} in order for
8259 @item -mfloat-abi=@var{name}
8261 Specifies which ABI to use for floating point values. Permissible values
8262 are: @samp{soft}, @samp{softfp} and @samp{hard}.
8264 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
8265 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
8266 of floating point instructions, but still uses the soft-float calling
8269 @item -mlittle-endian
8270 @opindex mlittle-endian
8271 Generate code for a processor running in little-endian mode. This is
8272 the default for all standard configurations.
8275 @opindex mbig-endian
8276 Generate code for a processor running in big-endian mode; the default is
8277 to compile code for a little-endian processor.
8279 @item -mwords-little-endian
8280 @opindex mwords-little-endian
8281 This option only applies when generating code for big-endian processors.
8282 Generate code for a little-endian word order but a big-endian byte
8283 order. That is, a byte order of the form @samp{32107654}. Note: this
8284 option should only be used if you require compatibility with code for
8285 big-endian ARM processors generated by versions of the compiler prior to
8288 @item -mcpu=@var{name}
8290 This specifies the name of the target ARM processor. GCC uses this name
8291 to determine what kind of instructions it can emit when generating
8292 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
8293 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
8294 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
8295 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
8296 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
8297 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
8298 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
8299 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
8300 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
8301 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
8302 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
8303 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
8304 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
8305 @samp{arm1156t2-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
8306 @samp{cortex-a8}, @samp{cortex-r4}, @samp{cortex-m3},
8307 @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
8309 @itemx -mtune=@var{name}
8311 This option is very similar to the @option{-mcpu=} option, except that
8312 instead of specifying the actual target processor type, and hence
8313 restricting which instructions can be used, it specifies that GCC should
8314 tune the performance of the code as if the target were of the type
8315 specified in this option, but still choosing the instructions that it
8316 will generate based on the cpu specified by a @option{-mcpu=} option.
8317 For some ARM implementations better performance can be obtained by using
8320 @item -march=@var{name}
8322 This specifies the name of the target ARM architecture. GCC uses this
8323 name to determine what kind of instructions it can emit when generating
8324 assembly code. This option can be used in conjunction with or instead
8325 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
8326 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
8327 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
8328 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv7}, @samp{armv7-a},
8329 @samp{armv7-r}, @samp{armv7-m}, @samp{iwmmxt}, @samp{ep9312}.
8331 @item -mfpu=@var{name}
8332 @itemx -mfpe=@var{number}
8333 @itemx -mfp=@var{number}
8337 This specifies what floating point hardware (or hardware emulation) is
8338 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
8339 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
8340 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
8341 with older versions of GCC@.
8343 If @option{-msoft-float} is specified this specifies the format of
8344 floating point values.
8346 @item -mstructure-size-boundary=@var{n}
8347 @opindex mstructure-size-boundary
8348 The size of all structures and unions will be rounded up to a multiple
8349 of the number of bits set by this option. Permissible values are 8, 32
8350 and 64. The default value varies for different toolchains. For the COFF
8351 targeted toolchain the default value is 8. A value of 64 is only allowed
8352 if the underlying ABI supports it.
8354 Specifying the larger number can produce faster, more efficient code, but
8355 can also increase the size of the program. Different values are potentially
8356 incompatible. Code compiled with one value cannot necessarily expect to
8357 work with code or libraries compiled with another value, if they exchange
8358 information using structures or unions.
8360 @item -mabort-on-noreturn
8361 @opindex mabort-on-noreturn
8362 Generate a call to the function @code{abort} at the end of a
8363 @code{noreturn} function. It will be executed if the function tries to
8367 @itemx -mno-long-calls
8368 @opindex mlong-calls
8369 @opindex mno-long-calls
8370 Tells the compiler to perform function calls by first loading the
8371 address of the function into a register and then performing a subroutine
8372 call on this register. This switch is needed if the target function
8373 will lie outside of the 64 megabyte addressing range of the offset based
8374 version of subroutine call instruction.
8376 Even if this switch is enabled, not all function calls will be turned
8377 into long calls. The heuristic is that static functions, functions
8378 which have the @samp{short-call} attribute, functions that are inside
8379 the scope of a @samp{#pragma no_long_calls} directive and functions whose
8380 definitions have already been compiled within the current compilation
8381 unit, will not be turned into long calls. The exception to this rule is
8382 that weak function definitions, functions with the @samp{long-call}
8383 attribute or the @samp{section} attribute, and functions that are within
8384 the scope of a @samp{#pragma long_calls} directive, will always be
8385 turned into long calls.
8387 This feature is not enabled by default. Specifying
8388 @option{-mno-long-calls} will restore the default behavior, as will
8389 placing the function calls within the scope of a @samp{#pragma
8390 long_calls_off} directive. Note these switches have no effect on how
8391 the compiler generates code to handle function calls via function
8394 @item -mnop-fun-dllimport
8395 @opindex mnop-fun-dllimport
8396 Disable support for the @code{dllimport} attribute.
8398 @item -msingle-pic-base
8399 @opindex msingle-pic-base
8400 Treat the register used for PIC addressing as read-only, rather than
8401 loading it in the prologue for each function. The run-time system is
8402 responsible for initializing this register with an appropriate value
8403 before execution begins.
8405 @item -mpic-register=@var{reg}
8406 @opindex mpic-register
8407 Specify the register to be used for PIC addressing. The default is R10
8408 unless stack-checking is enabled, when R9 is used.
8410 @item -mcirrus-fix-invalid-insns
8411 @opindex mcirrus-fix-invalid-insns
8412 @opindex mno-cirrus-fix-invalid-insns
8413 Insert NOPs into the instruction stream to in order to work around
8414 problems with invalid Maverick instruction combinations. This option
8415 is only valid if the @option{-mcpu=ep9312} option has been used to
8416 enable generation of instructions for the Cirrus Maverick floating
8417 point co-processor. This option is not enabled by default, since the
8418 problem is only present in older Maverick implementations. The default
8419 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
8422 @item -mpoke-function-name
8423 @opindex mpoke-function-name
8424 Write the name of each function into the text section, directly
8425 preceding the function prologue. The generated code is similar to this:
8429 .ascii "arm_poke_function_name", 0
8432 .word 0xff000000 + (t1 - t0)
8433 arm_poke_function_name
8435 stmfd sp!, @{fp, ip, lr, pc@}
8439 When performing a stack backtrace, code can inspect the value of
8440 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
8441 location @code{pc - 12} and the top 8 bits are set, then we know that
8442 there is a function name embedded immediately preceding this location
8443 and has length @code{((pc[-3]) & 0xff000000)}.
8447 Generate code for the Thumb instruction set. The default is to
8448 use the 32-bit ARM instruction set.
8449 This option automatically enables either 16-bit Thumb-1 or
8450 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
8451 and @option{-march=@var{name}} options.
8454 @opindex mtpcs-frame
8455 Generate a stack frame that is compliant with the Thumb Procedure Call
8456 Standard for all non-leaf functions. (A leaf function is one that does
8457 not call any other functions.) The default is @option{-mno-tpcs-frame}.
8459 @item -mtpcs-leaf-frame
8460 @opindex mtpcs-leaf-frame
8461 Generate a stack frame that is compliant with the Thumb Procedure Call
8462 Standard for all leaf functions. (A leaf function is one that does
8463 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
8465 @item -mcallee-super-interworking
8466 @opindex mcallee-super-interworking
8467 Gives all externally visible functions in the file being compiled an ARM
8468 instruction set header which switches to Thumb mode before executing the
8469 rest of the function. This allows these functions to be called from
8470 non-interworking code.
8472 @item -mcaller-super-interworking
8473 @opindex mcaller-super-interworking
8474 Allows calls via function pointers (including virtual functions) to
8475 execute correctly regardless of whether the target code has been
8476 compiled for interworking or not. There is a small overhead in the cost
8477 of executing a function pointer if this option is enabled.
8479 @item -mtp=@var{name}
8481 Specify the access model for the thread local storage pointer. The valid
8482 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
8483 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
8484 (supported in the arm6k architecture), and @option{auto}, which uses the
8485 best available method for the selected processor. The default setting is
8491 @subsection AVR Options
8494 These options are defined for AVR implementations:
8497 @item -mmcu=@var{mcu}
8499 Specify ATMEL AVR instruction set or MCU type.
8501 Instruction set avr1 is for the minimal AVR core, not supported by the C
8502 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8503 attiny11, attiny12, attiny15, attiny28).
8505 Instruction set avr2 (default) is for the classic AVR core with up to
8506 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8507 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8508 at90c8534, at90s8535).
8510 Instruction set avr3 is for the classic AVR core with up to 128K program
8511 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8513 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8514 memory space (MCU types: atmega8, atmega83, atmega85).
8516 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8517 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8518 atmega64, atmega128, at43usb355, at94k).
8522 Output instruction sizes to the asm file.
8524 @item -minit-stack=@var{N}
8525 @opindex minit-stack
8526 Specify the initial stack address, which may be a symbol or numeric value,
8527 @samp{__stack} is the default.
8529 @item -mno-interrupts
8530 @opindex mno-interrupts
8531 Generated code is not compatible with hardware interrupts.
8532 Code size will be smaller.
8534 @item -mcall-prologues
8535 @opindex mcall-prologues
8536 Functions prologues/epilogues expanded as call to appropriate
8537 subroutines. Code size will be smaller.
8539 @item -mno-tablejump
8540 @opindex mno-tablejump
8541 Do not generate tablejump insns which sometimes increase code size.
8544 @opindex mtiny-stack
8545 Change only the low 8 bits of the stack pointer.
8549 Assume int to be 8 bit integer. This affects the sizes of all types: A
8550 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
8551 and long long will be 4 bytes. Please note that this option does not
8552 comply to the C standards, but it will provide you with smaller code
8556 @node Blackfin Options
8557 @subsection Blackfin Options
8558 @cindex Blackfin Options
8561 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
8563 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
8564 can be one of @samp{bf522}, @samp{bf525}, @samp{bf527},
8565 @samp{bf531}, @samp{bf532}, @samp{bf533}, @samp{bf534},
8566 @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
8567 @samp{bf542}, @samp{bf544}, @samp{bf548}, @samp{bf549},
8569 The optional @var{sirevision} specifies the silicon revision of the target
8570 Blackfin processor. Any workarounds available for the targeted silicon revision
8571 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
8572 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
8573 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
8574 hexadecimal digits representing the major and minor numbers in the silicon
8575 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
8576 is not defined. If @var{sirevision} is @samp{any}, the
8577 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
8578 If this optional @var{sirevision} is not used, GCC assumes the latest known
8579 silicon revision of the targeted Blackfin processor.
8581 Support for @samp{bf561} is incomplete. For @samp{bf561},
8582 Only the processor macro is defined.
8583 Without this option, @samp{bf532} is used as the processor by default.
8584 The corresponding predefined processor macros for @var{cpu} is to
8585 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
8586 provided by libgloss to be linked in if @option{-msim} is not given.
8590 Specifies that the program will be run on the simulator. This causes
8591 the simulator BSP provided by libgloss to be linked in. This option
8592 has effect only for @samp{bfin-elf} toolchain.
8594 @item -momit-leaf-frame-pointer
8595 @opindex momit-leaf-frame-pointer
8596 Don't keep the frame pointer in a register for leaf functions. This
8597 avoids the instructions to save, set up and restore frame pointers and
8598 makes an extra register available in leaf functions. The option
8599 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8600 which might make debugging harder.
8602 @item -mspecld-anomaly
8603 @opindex mspecld-anomaly
8604 When enabled, the compiler will ensure that the generated code does not
8605 contain speculative loads after jump instructions. If this option is used,
8606 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
8608 @item -mno-specld-anomaly
8609 @opindex mno-specld-anomaly
8610 Don't generate extra code to prevent speculative loads from occurring.
8612 @item -mcsync-anomaly
8613 @opindex mcsync-anomaly
8614 When enabled, the compiler will ensure that the generated code does not
8615 contain CSYNC or SSYNC instructions too soon after conditional branches.
8616 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
8618 @item -mno-csync-anomaly
8619 @opindex mno-csync-anomaly
8620 Don't generate extra code to prevent CSYNC or SSYNC instructions from
8621 occurring too soon after a conditional branch.
8625 When enabled, the compiler is free to take advantage of the knowledge that
8626 the entire program fits into the low 64k of memory.
8629 @opindex mno-low-64k
8630 Assume that the program is arbitrarily large. This is the default.
8632 @item -mstack-check-l1
8633 @opindex mstack-check-l1
8634 Do stack checking using information placed into L1 scratchpad memory by the
8637 @item -mid-shared-library
8638 @opindex mid-shared-library
8639 Generate code that supports shared libraries via the library ID method.
8640 This allows for execute in place and shared libraries in an environment
8641 without virtual memory management. This option implies @option{-fPIC}.
8643 @item -mno-id-shared-library
8644 @opindex mno-id-shared-library
8645 Generate code that doesn't assume ID based shared libraries are being used.
8646 This is the default.
8648 @item -mleaf-id-shared-library
8649 @opindex mleaf-id-shared-library
8650 Generate code that supports shared libraries via the library ID method,
8651 but assumes that this library or executable won't link against any other
8652 ID shared libraries. That allows the compiler to use faster code for jumps
8655 @item -mno-leaf-id-shared-library
8656 @opindex mno-leaf-id-shared-library
8657 Do not assume that the code being compiled won't link against any ID shared
8658 libraries. Slower code will be generated for jump and call insns.
8660 @item -mshared-library-id=n
8661 @opindex mshared-library-id
8662 Specified the identification number of the ID based shared library being
8663 compiled. Specifying a value of 0 will generate more compact code, specifying
8664 other values will force the allocation of that number to the current
8665 library but is no more space or time efficient than omitting this option.
8669 Generate code that allows the data segment to be located in a different
8670 area of memory from the text segment. This allows for execute in place in
8671 an environment without virtual memory management by eliminating relocations
8672 against the text section.
8675 @opindex mno-sep-data
8676 Generate code that assumes that the data segment follows the text segment.
8677 This is the default.
8680 @itemx -mno-long-calls
8681 @opindex mlong-calls
8682 @opindex mno-long-calls
8683 Tells the compiler to perform function calls by first loading the
8684 address of the function into a register and then performing a subroutine
8685 call on this register. This switch is needed if the target function
8686 will lie outside of the 24 bit addressing range of the offset based
8687 version of subroutine call instruction.
8689 This feature is not enabled by default. Specifying
8690 @option{-mno-long-calls} will restore the default behavior. Note these
8691 switches have no effect on how the compiler generates code to handle
8692 function calls via function pointers.
8696 Link with the fast floating-point library. This library relaxes some of
8697 the IEEE floating-point standard's rules for checking inputs against
8698 Not-a-Number (NAN), in the interest of performance.
8701 @opindex minline-plt
8702 Enable inlining of PLT entries in function calls to functions that are
8703 not known to bind locally. It has no effect without @option{-mfdpic}.
8707 @subsection CRIS Options
8708 @cindex CRIS Options
8710 These options are defined specifically for the CRIS ports.
8713 @item -march=@var{architecture-type}
8714 @itemx -mcpu=@var{architecture-type}
8717 Generate code for the specified architecture. The choices for
8718 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
8719 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
8720 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
8723 @item -mtune=@var{architecture-type}
8725 Tune to @var{architecture-type} everything applicable about the generated
8726 code, except for the ABI and the set of available instructions. The
8727 choices for @var{architecture-type} are the same as for
8728 @option{-march=@var{architecture-type}}.
8730 @item -mmax-stack-frame=@var{n}
8731 @opindex mmax-stack-frame
8732 Warn when the stack frame of a function exceeds @var{n} bytes.
8734 @item -melinux-stacksize=@var{n}
8735 @opindex melinux-stacksize
8736 Only available with the @samp{cris-axis-aout} target. Arranges for
8737 indications in the program to the kernel loader that the stack of the
8738 program should be set to @var{n} bytes.
8744 The options @option{-metrax4} and @option{-metrax100} are synonyms for
8745 @option{-march=v3} and @option{-march=v8} respectively.
8747 @item -mmul-bug-workaround
8748 @itemx -mno-mul-bug-workaround
8749 @opindex mmul-bug-workaround
8750 @opindex mno-mul-bug-workaround
8751 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
8752 models where it applies. This option is active by default.
8756 Enable CRIS-specific verbose debug-related information in the assembly
8757 code. This option also has the effect to turn off the @samp{#NO_APP}
8758 formatted-code indicator to the assembler at the beginning of the
8763 Do not use condition-code results from previous instruction; always emit
8764 compare and test instructions before use of condition codes.
8766 @item -mno-side-effects
8767 @opindex mno-side-effects
8768 Do not emit instructions with side-effects in addressing modes other than
8772 @itemx -mno-stack-align
8774 @itemx -mno-data-align
8775 @itemx -mconst-align
8776 @itemx -mno-const-align
8777 @opindex mstack-align
8778 @opindex mno-stack-align
8779 @opindex mdata-align
8780 @opindex mno-data-align
8781 @opindex mconst-align
8782 @opindex mno-const-align
8783 These options (no-options) arranges (eliminate arrangements) for the
8784 stack-frame, individual data and constants to be aligned for the maximum
8785 single data access size for the chosen CPU model. The default is to
8786 arrange for 32-bit alignment. ABI details such as structure layout are
8787 not affected by these options.
8795 Similar to the stack- data- and const-align options above, these options
8796 arrange for stack-frame, writable data and constants to all be 32-bit,
8797 16-bit or 8-bit aligned. The default is 32-bit alignment.
8799 @item -mno-prologue-epilogue
8800 @itemx -mprologue-epilogue
8801 @opindex mno-prologue-epilogue
8802 @opindex mprologue-epilogue
8803 With @option{-mno-prologue-epilogue}, the normal function prologue and
8804 epilogue that sets up the stack-frame are omitted and no return
8805 instructions or return sequences are generated in the code. Use this
8806 option only together with visual inspection of the compiled code: no
8807 warnings or errors are generated when call-saved registers must be saved,
8808 or storage for local variable needs to be allocated.
8814 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
8815 instruction sequences that load addresses for functions from the PLT part
8816 of the GOT rather than (traditional on other architectures) calls to the
8817 PLT@. The default is @option{-mgotplt}.
8821 Legacy no-op option only recognized with the cris-axis-aout target.
8825 Legacy no-op option only recognized with the cris-axis-elf and
8826 cris-axis-linux-gnu targets.
8830 Only recognized with the cris-axis-aout target, where it selects a
8831 GNU/linux-like multilib, include files and instruction set for
8836 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
8840 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
8841 to link with input-output functions from a simulator library. Code,
8842 initialized data and zero-initialized data are allocated consecutively.
8846 Like @option{-sim}, but pass linker options to locate initialized data at
8847 0x40000000 and zero-initialized data at 0x80000000.
8851 @subsection CRX Options
8854 These options are defined specifically for the CRX ports.
8860 Enable the use of multiply-accumulate instructions. Disabled by default.
8864 Push instructions will be used to pass outgoing arguments when functions
8865 are called. Enabled by default.
8868 @node Darwin Options
8869 @subsection Darwin Options
8870 @cindex Darwin options
8872 These options are defined for all architectures running the Darwin operating
8875 FSF GCC on Darwin does not create ``fat'' object files; it will create
8876 an object file for the single architecture that it was built to
8877 target. Apple's GCC on Darwin does create ``fat'' files if multiple
8878 @option{-arch} options are used; it does so by running the compiler or
8879 linker multiple times and joining the results together with
8882 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
8883 @samp{i686}) is determined by the flags that specify the ISA
8884 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
8885 @option{-force_cpusubtype_ALL} option can be used to override this.
8887 The Darwin tools vary in their behavior when presented with an ISA
8888 mismatch. The assembler, @file{as}, will only permit instructions to
8889 be used that are valid for the subtype of the file it is generating,
8890 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
8891 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
8892 and print an error if asked to create a shared library with a less
8893 restrictive subtype than its input files (for instance, trying to put
8894 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
8895 for executables, @file{ld}, will quietly give the executable the most
8896 restrictive subtype of any of its input files.
8901 Add the framework directory @var{dir} to the head of the list of
8902 directories to be searched for header files. These directories are
8903 interleaved with those specified by @option{-I} options and are
8904 scanned in a left-to-right order.
8906 A framework directory is a directory with frameworks in it. A
8907 framework is a directory with a @samp{"Headers"} and/or
8908 @samp{"PrivateHeaders"} directory contained directly in it that ends
8909 in @samp{".framework"}. The name of a framework is the name of this
8910 directory excluding the @samp{".framework"}. Headers associated with
8911 the framework are found in one of those two directories, with
8912 @samp{"Headers"} being searched first. A subframework is a framework
8913 directory that is in a framework's @samp{"Frameworks"} directory.
8914 Includes of subframework headers can only appear in a header of a
8915 framework that contains the subframework, or in a sibling subframework
8916 header. Two subframeworks are siblings if they occur in the same
8917 framework. A subframework should not have the same name as a
8918 framework, a warning will be issued if this is violated. Currently a
8919 subframework cannot have subframeworks, in the future, the mechanism
8920 may be extended to support this. The standard frameworks can be found
8921 in @samp{"/System/Library/Frameworks"} and
8922 @samp{"/Library/Frameworks"}. An example include looks like
8923 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
8924 the name of the framework and header.h is found in the
8925 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
8927 @item -iframework@var{dir}
8929 Like @option{-F} except the directory is a treated as a system
8930 directory. The main difference between this @option{-iframework} and
8931 @option{-F} is that with @option{-iframework} the compiler does not
8932 warn about constructs contained within header files found via
8933 @var{dir}. This option is valid only for the C family of languages.
8937 Emit debugging information for symbols that are used. For STABS
8938 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
8939 This is by default ON@.
8943 Emit debugging information for all symbols and types.
8945 @item -mmacosx-version-min=@var{version}
8946 The earliest version of MacOS X that this executable will run on
8947 is @var{version}. Typical values of @var{version} include @code{10.1},
8948 @code{10.2}, and @code{10.3.9}.
8950 If the compiler was built to use the system's headers by default,
8951 then the default for this option is the system version on which the
8952 compiler is running, otherwise the default is to make choices which
8953 are compatible with as many systems and code bases as possible.
8957 Enable kernel development mode. The @option{-mkernel} option sets
8958 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
8959 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
8960 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
8961 applicable. This mode also sets @option{-mno-altivec},
8962 @option{-msoft-float}, @option{-fno-builtin} and
8963 @option{-mlong-branch} for PowerPC targets.
8965 @item -mone-byte-bool
8966 @opindex mone-byte-bool
8967 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
8968 By default @samp{sizeof(bool)} is @samp{4} when compiling for
8969 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
8970 option has no effect on x86.
8972 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
8973 to generate code that is not binary compatible with code generated
8974 without that switch. Using this switch may require recompiling all
8975 other modules in a program, including system libraries. Use this
8976 switch to conform to a non-default data model.
8978 @item -mfix-and-continue
8979 @itemx -ffix-and-continue
8980 @itemx -findirect-data
8981 @opindex mfix-and-continue
8982 @opindex ffix-and-continue
8983 @opindex findirect-data
8984 Generate code suitable for fast turn around development. Needed to
8985 enable gdb to dynamically load @code{.o} files into already running
8986 programs. @option{-findirect-data} and @option{-ffix-and-continue}
8987 are provided for backwards compatibility.
8991 Loads all members of static archive libraries.
8992 See man ld(1) for more information.
8994 @item -arch_errors_fatal
8995 @opindex arch_errors_fatal
8996 Cause the errors having to do with files that have the wrong architecture
9000 @opindex bind_at_load
9001 Causes the output file to be marked such that the dynamic linker will
9002 bind all undefined references when the file is loaded or launched.
9006 Produce a Mach-o bundle format file.
9007 See man ld(1) for more information.
9009 @item -bundle_loader @var{executable}
9010 @opindex bundle_loader
9011 This option specifies the @var{executable} that will be loading the build
9012 output file being linked. See man ld(1) for more information.
9016 When passed this option, GCC will produce a dynamic library instead of
9017 an executable when linking, using the Darwin @file{libtool} command.
9019 @item -force_cpusubtype_ALL
9020 @opindex force_cpusubtype_ALL
9021 This causes GCC's output file to have the @var{ALL} subtype, instead of
9022 one controlled by the @option{-mcpu} or @option{-march} option.
9024 @item -allowable_client @var{client_name}
9026 @itemx -compatibility_version
9027 @itemx -current_version
9029 @itemx -dependency-file
9031 @itemx -dylinker_install_name
9033 @itemx -exported_symbols_list
9035 @itemx -flat_namespace
9036 @itemx -force_flat_namespace
9037 @itemx -headerpad_max_install_names
9040 @itemx -install_name
9041 @itemx -keep_private_externs
9042 @itemx -multi_module
9043 @itemx -multiply_defined
9044 @itemx -multiply_defined_unused
9046 @itemx -no_dead_strip_inits_and_terms
9047 @itemx -nofixprebinding
9050 @itemx -noseglinkedit
9051 @itemx -pagezero_size
9053 @itemx -prebind_all_twolevel_modules
9054 @itemx -private_bundle
9055 @itemx -read_only_relocs
9057 @itemx -sectobjectsymbols
9061 @itemx -sectobjectsymbols
9064 @itemx -segs_read_only_addr
9065 @itemx -segs_read_write_addr
9066 @itemx -seg_addr_table
9067 @itemx -seg_addr_table_filename
9070 @itemx -segs_read_only_addr
9071 @itemx -segs_read_write_addr
9072 @itemx -single_module
9075 @itemx -sub_umbrella
9076 @itemx -twolevel_namespace
9079 @itemx -unexported_symbols_list
9080 @itemx -weak_reference_mismatches
9083 @opindex allowable_client
9084 @opindex client_name
9085 @opindex compatibility_version
9086 @opindex current_version
9088 @opindex dependency-file
9090 @opindex dylinker_install_name
9092 @opindex exported_symbols_list
9094 @opindex flat_namespace
9095 @opindex force_flat_namespace
9096 @opindex headerpad_max_install_names
9099 @opindex install_name
9100 @opindex keep_private_externs
9101 @opindex multi_module
9102 @opindex multiply_defined
9103 @opindex multiply_defined_unused
9105 @opindex no_dead_strip_inits_and_terms
9106 @opindex nofixprebinding
9107 @opindex nomultidefs
9109 @opindex noseglinkedit
9110 @opindex pagezero_size
9112 @opindex prebind_all_twolevel_modules
9113 @opindex private_bundle
9114 @opindex read_only_relocs
9116 @opindex sectobjectsymbols
9120 @opindex sectobjectsymbols
9123 @opindex segs_read_only_addr
9124 @opindex segs_read_write_addr
9125 @opindex seg_addr_table
9126 @opindex seg_addr_table_filename
9127 @opindex seglinkedit
9129 @opindex segs_read_only_addr
9130 @opindex segs_read_write_addr
9131 @opindex single_module
9133 @opindex sub_library
9134 @opindex sub_umbrella
9135 @opindex twolevel_namespace
9138 @opindex unexported_symbols_list
9139 @opindex weak_reference_mismatches
9140 @opindex whatsloaded
9142 These options are passed to the Darwin linker. The Darwin linker man page
9143 describes them in detail.
9146 @node DEC Alpha Options
9147 @subsection DEC Alpha Options
9149 These @samp{-m} options are defined for the DEC Alpha implementations:
9152 @item -mno-soft-float
9154 @opindex mno-soft-float
9155 @opindex msoft-float
9156 Use (do not use) the hardware floating-point instructions for
9157 floating-point operations. When @option{-msoft-float} is specified,
9158 functions in @file{libgcc.a} will be used to perform floating-point
9159 operations. Unless they are replaced by routines that emulate the
9160 floating-point operations, or compiled in such a way as to call such
9161 emulations routines, these routines will issue floating-point
9162 operations. If you are compiling for an Alpha without floating-point
9163 operations, you must ensure that the library is built so as not to call
9166 Note that Alpha implementations without floating-point operations are
9167 required to have floating-point registers.
9172 @opindex mno-fp-regs
9173 Generate code that uses (does not use) the floating-point register set.
9174 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
9175 register set is not used, floating point operands are passed in integer
9176 registers as if they were integers and floating-point results are passed
9177 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
9178 so any function with a floating-point argument or return value called by code
9179 compiled with @option{-mno-fp-regs} must also be compiled with that
9182 A typical use of this option is building a kernel that does not use,
9183 and hence need not save and restore, any floating-point registers.
9187 The Alpha architecture implements floating-point hardware optimized for
9188 maximum performance. It is mostly compliant with the IEEE floating
9189 point standard. However, for full compliance, software assistance is
9190 required. This option generates code fully IEEE compliant code
9191 @emph{except} that the @var{inexact-flag} is not maintained (see below).
9192 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
9193 defined during compilation. The resulting code is less efficient but is
9194 able to correctly support denormalized numbers and exceptional IEEE
9195 values such as not-a-number and plus/minus infinity. Other Alpha
9196 compilers call this option @option{-ieee_with_no_inexact}.
9198 @item -mieee-with-inexact
9199 @opindex mieee-with-inexact
9200 This is like @option{-mieee} except the generated code also maintains
9201 the IEEE @var{inexact-flag}. Turning on this option causes the
9202 generated code to implement fully-compliant IEEE math. In addition to
9203 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
9204 macro. On some Alpha implementations the resulting code may execute
9205 significantly slower than the code generated by default. Since there is
9206 very little code that depends on the @var{inexact-flag}, you should
9207 normally not specify this option. Other Alpha compilers call this
9208 option @option{-ieee_with_inexact}.
9210 @item -mfp-trap-mode=@var{trap-mode}
9211 @opindex mfp-trap-mode
9212 This option controls what floating-point related traps are enabled.
9213 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9214 The trap mode can be set to one of four values:
9218 This is the default (normal) setting. The only traps that are enabled
9219 are the ones that cannot be disabled in software (e.g., division by zero
9223 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9227 Like @samp{u}, but the instructions are marked to be safe for software
9228 completion (see Alpha architecture manual for details).
9231 Like @samp{su}, but inexact traps are enabled as well.
9234 @item -mfp-rounding-mode=@var{rounding-mode}
9235 @opindex mfp-rounding-mode
9236 Selects the IEEE rounding mode. Other Alpha compilers call this option
9237 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9242 Normal IEEE rounding mode. Floating point numbers are rounded towards
9243 the nearest machine number or towards the even machine number in case
9247 Round towards minus infinity.
9250 Chopped rounding mode. Floating point numbers are rounded towards zero.
9253 Dynamic rounding mode. A field in the floating point control register
9254 (@var{fpcr}, see Alpha architecture reference manual) controls the
9255 rounding mode in effect. The C library initializes this register for
9256 rounding towards plus infinity. Thus, unless your program modifies the
9257 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9260 @item -mtrap-precision=@var{trap-precision}
9261 @opindex mtrap-precision
9262 In the Alpha architecture, floating point traps are imprecise. This
9263 means without software assistance it is impossible to recover from a
9264 floating trap and program execution normally needs to be terminated.
9265 GCC can generate code that can assist operating system trap handlers
9266 in determining the exact location that caused a floating point trap.
9267 Depending on the requirements of an application, different levels of
9268 precisions can be selected:
9272 Program precision. This option is the default and means a trap handler
9273 can only identify which program caused a floating point exception.
9276 Function precision. The trap handler can determine the function that
9277 caused a floating point exception.
9280 Instruction precision. The trap handler can determine the exact
9281 instruction that caused a floating point exception.
9284 Other Alpha compilers provide the equivalent options called
9285 @option{-scope_safe} and @option{-resumption_safe}.
9287 @item -mieee-conformant
9288 @opindex mieee-conformant
9289 This option marks the generated code as IEEE conformant. You must not
9290 use this option unless you also specify @option{-mtrap-precision=i} and either
9291 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9292 is to emit the line @samp{.eflag 48} in the function prologue of the
9293 generated assembly file. Under DEC Unix, this has the effect that
9294 IEEE-conformant math library routines will be linked in.
9296 @item -mbuild-constants
9297 @opindex mbuild-constants
9298 Normally GCC examines a 32- or 64-bit integer constant to
9299 see if it can construct it from smaller constants in two or three
9300 instructions. If it cannot, it will output the constant as a literal and
9301 generate code to load it from the data segment at runtime.
9303 Use this option to require GCC to construct @emph{all} integer constants
9304 using code, even if it takes more instructions (the maximum is six).
9306 You would typically use this option to build a shared library dynamic
9307 loader. Itself a shared library, it must relocate itself in memory
9308 before it can find the variables and constants in its own data segment.
9314 Select whether to generate code to be assembled by the vendor-supplied
9315 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9333 Indicate whether GCC should generate code to use the optional BWX,
9334 CIX, FIX and MAX instruction sets. The default is to use the instruction
9335 sets supported by the CPU type specified via @option{-mcpu=} option or that
9336 of the CPU on which GCC was built if none was specified.
9341 @opindex mfloat-ieee
9342 Generate code that uses (does not use) VAX F and G floating point
9343 arithmetic instead of IEEE single and double precision.
9345 @item -mexplicit-relocs
9346 @itemx -mno-explicit-relocs
9347 @opindex mexplicit-relocs
9348 @opindex mno-explicit-relocs
9349 Older Alpha assemblers provided no way to generate symbol relocations
9350 except via assembler macros. Use of these macros does not allow
9351 optimal instruction scheduling. GNU binutils as of version 2.12
9352 supports a new syntax that allows the compiler to explicitly mark
9353 which relocations should apply to which instructions. This option
9354 is mostly useful for debugging, as GCC detects the capabilities of
9355 the assembler when it is built and sets the default accordingly.
9359 @opindex msmall-data
9360 @opindex mlarge-data
9361 When @option{-mexplicit-relocs} is in effect, static data is
9362 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9363 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9364 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9365 16-bit relocations off of the @code{$gp} register. This limits the
9366 size of the small data area to 64KB, but allows the variables to be
9367 directly accessed via a single instruction.
9369 The default is @option{-mlarge-data}. With this option the data area
9370 is limited to just below 2GB@. Programs that require more than 2GB of
9371 data must use @code{malloc} or @code{mmap} to allocate the data in the
9372 heap instead of in the program's data segment.
9374 When generating code for shared libraries, @option{-fpic} implies
9375 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9379 @opindex msmall-text
9380 @opindex mlarge-text
9381 When @option{-msmall-text} is used, the compiler assumes that the
9382 code of the entire program (or shared library) fits in 4MB, and is
9383 thus reachable with a branch instruction. When @option{-msmall-data}
9384 is used, the compiler can assume that all local symbols share the
9385 same @code{$gp} value, and thus reduce the number of instructions
9386 required for a function call from 4 to 1.
9388 The default is @option{-mlarge-text}.
9390 @item -mcpu=@var{cpu_type}
9392 Set the instruction set and instruction scheduling parameters for
9393 machine type @var{cpu_type}. You can specify either the @samp{EV}
9394 style name or the corresponding chip number. GCC supports scheduling
9395 parameters for the EV4, EV5 and EV6 family of processors and will
9396 choose the default values for the instruction set from the processor
9397 you specify. If you do not specify a processor type, GCC will default
9398 to the processor on which the compiler was built.
9400 Supported values for @var{cpu_type} are
9406 Schedules as an EV4 and has no instruction set extensions.
9410 Schedules as an EV5 and has no instruction set extensions.
9414 Schedules as an EV5 and supports the BWX extension.
9419 Schedules as an EV5 and supports the BWX and MAX extensions.
9423 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9427 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9430 @item -mtune=@var{cpu_type}
9432 Set only the instruction scheduling parameters for machine type
9433 @var{cpu_type}. The instruction set is not changed.
9435 @item -mmemory-latency=@var{time}
9436 @opindex mmemory-latency
9437 Sets the latency the scheduler should assume for typical memory
9438 references as seen by the application. This number is highly
9439 dependent on the memory access patterns used by the application
9440 and the size of the external cache on the machine.
9442 Valid options for @var{time} are
9446 A decimal number representing clock cycles.
9452 The compiler contains estimates of the number of clock cycles for
9453 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9454 (also called Dcache, Scache, and Bcache), as well as to main memory.
9455 Note that L3 is only valid for EV5.
9460 @node DEC Alpha/VMS Options
9461 @subsection DEC Alpha/VMS Options
9463 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9466 @item -mvms-return-codes
9467 @opindex mvms-return-codes
9468 Return VMS condition codes from main. The default is to return POSIX
9469 style condition (e.g.@: error) codes.
9473 @subsection FRV Options
9480 Only use the first 32 general purpose registers.
9485 Use all 64 general purpose registers.
9490 Use only the first 32 floating point registers.
9495 Use all 64 floating point registers
9498 @opindex mhard-float
9500 Use hardware instructions for floating point operations.
9503 @opindex msoft-float
9505 Use library routines for floating point operations.
9510 Dynamically allocate condition code registers.
9515 Do not try to dynamically allocate condition code registers, only
9516 use @code{icc0} and @code{fcc0}.
9521 Change ABI to use double word insns.
9526 Do not use double word instructions.
9531 Use floating point double instructions.
9536 Do not use floating point double instructions.
9541 Use media instructions.
9546 Do not use media instructions.
9551 Use multiply and add/subtract instructions.
9556 Do not use multiply and add/subtract instructions.
9561 Select the FDPIC ABI, that uses function descriptors to represent
9562 pointers to functions. Without any PIC/PIE-related options, it
9563 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
9564 assumes GOT entries and small data are within a 12-bit range from the
9565 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
9566 are computed with 32 bits.
9569 @opindex minline-plt
9571 Enable inlining of PLT entries in function calls to functions that are
9572 not known to bind locally. It has no effect without @option{-mfdpic}.
9573 It's enabled by default if optimizing for speed and compiling for
9574 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
9575 optimization option such as @option{-O3} or above is present in the
9581 Assume a large TLS segment when generating thread-local code.
9586 Do not assume a large TLS segment when generating thread-local code.
9591 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
9592 that is known to be in read-only sections. It's enabled by default,
9593 except for @option{-fpic} or @option{-fpie}: even though it may help
9594 make the global offset table smaller, it trades 1 instruction for 4.
9595 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
9596 one of which may be shared by multiple symbols, and it avoids the need
9597 for a GOT entry for the referenced symbol, so it's more likely to be a
9598 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
9600 @item -multilib-library-pic
9601 @opindex multilib-library-pic
9603 Link with the (library, not FD) pic libraries. It's implied by
9604 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
9605 @option{-fpic} without @option{-mfdpic}. You should never have to use
9611 Follow the EABI requirement of always creating a frame pointer whenever
9612 a stack frame is allocated. This option is enabled by default and can
9613 be disabled with @option{-mno-linked-fp}.
9616 @opindex mlong-calls
9618 Use indirect addressing to call functions outside the current
9619 compilation unit. This allows the functions to be placed anywhere
9620 within the 32-bit address space.
9622 @item -malign-labels
9623 @opindex malign-labels
9625 Try to align labels to an 8-byte boundary by inserting nops into the
9626 previous packet. This option only has an effect when VLIW packing
9627 is enabled. It doesn't create new packets; it merely adds nops to
9631 @opindex mlibrary-pic
9633 Generate position-independent EABI code.
9638 Use only the first four media accumulator registers.
9643 Use all eight media accumulator registers.
9648 Pack VLIW instructions.
9653 Do not pack VLIW instructions.
9658 Do not mark ABI switches in e_flags.
9663 Enable the use of conditional-move instructions (default).
9665 This switch is mainly for debugging the compiler and will likely be removed
9666 in a future version.
9668 @item -mno-cond-move
9669 @opindex mno-cond-move
9671 Disable the use of conditional-move instructions.
9673 This switch is mainly for debugging the compiler and will likely be removed
9674 in a future version.
9679 Enable the use of conditional set instructions (default).
9681 This switch is mainly for debugging the compiler and will likely be removed
9682 in a future version.
9687 Disable the use of conditional set instructions.
9689 This switch is mainly for debugging the compiler and will likely be removed
9690 in a future version.
9695 Enable the use of conditional execution (default).
9697 This switch is mainly for debugging the compiler and will likely be removed
9698 in a future version.
9700 @item -mno-cond-exec
9701 @opindex mno-cond-exec
9703 Disable the use of conditional execution.
9705 This switch is mainly for debugging the compiler and will likely be removed
9706 in a future version.
9709 @opindex mvliw-branch
9711 Run a pass to pack branches into VLIW instructions (default).
9713 This switch is mainly for debugging the compiler and will likely be removed
9714 in a future version.
9716 @item -mno-vliw-branch
9717 @opindex mno-vliw-branch
9719 Do not run a pass to pack branches into VLIW instructions.
9721 This switch is mainly for debugging the compiler and will likely be removed
9722 in a future version.
9724 @item -mmulti-cond-exec
9725 @opindex mmulti-cond-exec
9727 Enable optimization of @code{&&} and @code{||} in conditional execution
9730 This switch is mainly for debugging the compiler and will likely be removed
9731 in a future version.
9733 @item -mno-multi-cond-exec
9734 @opindex mno-multi-cond-exec
9736 Disable optimization of @code{&&} and @code{||} in conditional execution.
9738 This switch is mainly for debugging the compiler and will likely be removed
9739 in a future version.
9741 @item -mnested-cond-exec
9742 @opindex mnested-cond-exec
9744 Enable nested conditional execution optimizations (default).
9746 This switch is mainly for debugging the compiler and will likely be removed
9747 in a future version.
9749 @item -mno-nested-cond-exec
9750 @opindex mno-nested-cond-exec
9752 Disable nested conditional execution optimizations.
9754 This switch is mainly for debugging the compiler and will likely be removed
9755 in a future version.
9757 @item -moptimize-membar
9758 @opindex moptimize-membar
9760 This switch removes redundant @code{membar} instructions from the
9761 compiler generated code. It is enabled by default.
9763 @item -mno-optimize-membar
9764 @opindex mno-optimize-membar
9766 This switch disables the automatic removal of redundant @code{membar}
9767 instructions from the generated code.
9769 @item -mtomcat-stats
9770 @opindex mtomcat-stats
9772 Cause gas to print out tomcat statistics.
9774 @item -mcpu=@var{cpu}
9777 Select the processor type for which to generate code. Possible values are
9778 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
9779 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
9783 @node GNU/Linux Options
9784 @subsection GNU/Linux Options
9786 These @samp{-m} options are defined for GNU/Linux targets:
9791 Use the GNU C library instead of uClibc. This is the default except
9792 on @samp{*-*-linux-*uclibc*} targets.
9796 Use uClibc instead of the GNU C library. This is the default on
9797 @samp{*-*-linux-*uclibc*} targets.
9800 @node H8/300 Options
9801 @subsection H8/300 Options
9803 These @samp{-m} options are defined for the H8/300 implementations:
9808 Shorten some address references at link time, when possible; uses the
9809 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9810 ld, Using ld}, for a fuller description.
9814 Generate code for the H8/300H@.
9818 Generate code for the H8S@.
9822 Generate code for the H8S and H8/300H in the normal mode. This switch
9823 must be used either with @option{-mh} or @option{-ms}.
9827 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9831 Make @code{int} data 32 bits by default.
9835 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9836 The default for the H8/300H and H8S is to align longs and floats on 4
9838 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9839 This option has no effect on the H8/300.
9843 @subsection HPPA Options
9844 @cindex HPPA Options
9846 These @samp{-m} options are defined for the HPPA family of computers:
9849 @item -march=@var{architecture-type}
9851 Generate code for the specified architecture. The choices for
9852 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
9853 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
9854 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
9855 architecture option for your machine. Code compiled for lower numbered
9856 architectures will run on higher numbered architectures, but not the
9860 @itemx -mpa-risc-1-1
9861 @itemx -mpa-risc-2-0
9862 @opindex mpa-risc-1-0
9863 @opindex mpa-risc-1-1
9864 @opindex mpa-risc-2-0
9865 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
9868 @opindex mbig-switch
9869 Generate code suitable for big switch tables. Use this option only if
9870 the assembler/linker complain about out of range branches within a switch
9873 @item -mjump-in-delay
9874 @opindex mjump-in-delay
9875 Fill delay slots of function calls with unconditional jump instructions
9876 by modifying the return pointer for the function call to be the target
9877 of the conditional jump.
9879 @item -mdisable-fpregs
9880 @opindex mdisable-fpregs
9881 Prevent floating point registers from being used in any manner. This is
9882 necessary for compiling kernels which perform lazy context switching of
9883 floating point registers. If you use this option and attempt to perform
9884 floating point operations, the compiler will abort.
9886 @item -mdisable-indexing
9887 @opindex mdisable-indexing
9888 Prevent the compiler from using indexing address modes. This avoids some
9889 rather obscure problems when compiling MIG generated code under MACH@.
9891 @item -mno-space-regs
9892 @opindex mno-space-regs
9893 Generate code that assumes the target has no space registers. This allows
9894 GCC to generate faster indirect calls and use unscaled index address modes.
9896 Such code is suitable for level 0 PA systems and kernels.
9898 @item -mfast-indirect-calls
9899 @opindex mfast-indirect-calls
9900 Generate code that assumes calls never cross space boundaries. This
9901 allows GCC to emit code which performs faster indirect calls.
9903 This option will not work in the presence of shared libraries or nested
9906 @item -mfixed-range=@var{register-range}
9907 @opindex mfixed-range
9908 Generate code treating the given register range as fixed registers.
9909 A fixed register is one that the register allocator can not use. This is
9910 useful when compiling kernel code. A register range is specified as
9911 two registers separated by a dash. Multiple register ranges can be
9912 specified separated by a comma.
9914 @item -mlong-load-store
9915 @opindex mlong-load-store
9916 Generate 3-instruction load and store sequences as sometimes required by
9917 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
9920 @item -mportable-runtime
9921 @opindex mportable-runtime
9922 Use the portable calling conventions proposed by HP for ELF systems.
9926 Enable the use of assembler directives only GAS understands.
9928 @item -mschedule=@var{cpu-type}
9930 Schedule code according to the constraints for the machine type
9931 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
9932 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
9933 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
9934 proper scheduling option for your machine. The default scheduling is
9938 @opindex mlinker-opt
9939 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
9940 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
9941 linkers in which they give bogus error messages when linking some programs.
9944 @opindex msoft-float
9945 Generate output containing library calls for floating point.
9946 @strong{Warning:} the requisite libraries are not available for all HPPA
9947 targets. Normally the facilities of the machine's usual C compiler are
9948 used, but this cannot be done directly in cross-compilation. You must make
9949 your own arrangements to provide suitable library functions for
9950 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
9951 does provide software floating point support.
9953 @option{-msoft-float} changes the calling convention in the output file;
9954 therefore, it is only useful if you compile @emph{all} of a program with
9955 this option. In particular, you need to compile @file{libgcc.a}, the
9956 library that comes with GCC, with @option{-msoft-float} in order for
9961 Generate the predefine, @code{_SIO}, for server IO@. The default is
9962 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
9963 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
9964 options are available under HP-UX and HI-UX@.
9968 Use GNU ld specific options. This passes @option{-shared} to ld when
9969 building a shared library. It is the default when GCC is configured,
9970 explicitly or implicitly, with the GNU linker. This option does not
9971 have any affect on which ld is called, it only changes what parameters
9972 are passed to that ld. The ld that is called is determined by the
9973 @option{--with-ld} configure option, GCC's program search path, and
9974 finally by the user's @env{PATH}. The linker used by GCC can be printed
9975 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
9976 on the 64 bit HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9980 Use HP ld specific options. This passes @option{-b} to ld when building
9981 a shared library and passes @option{+Accept TypeMismatch} to ld on all
9982 links. It is the default when GCC is configured, explicitly or
9983 implicitly, with the HP linker. This option does not have any affect on
9984 which ld is called, it only changes what parameters are passed to that
9985 ld. The ld that is called is determined by the @option{--with-ld}
9986 configure option, GCC's program search path, and finally by the user's
9987 @env{PATH}. The linker used by GCC can be printed using @samp{which
9988 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
9989 HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9992 @opindex mno-long-calls
9993 Generate code that uses long call sequences. This ensures that a call
9994 is always able to reach linker generated stubs. The default is to generate
9995 long calls only when the distance from the call site to the beginning
9996 of the function or translation unit, as the case may be, exceeds a
9997 predefined limit set by the branch type being used. The limits for
9998 normal calls are 7,600,000 and 240,000 bytes, respectively for the
9999 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
10002 Distances are measured from the beginning of functions when using the
10003 @option{-ffunction-sections} option, or when using the @option{-mgas}
10004 and @option{-mno-portable-runtime} options together under HP-UX with
10007 It is normally not desirable to use this option as it will degrade
10008 performance. However, it may be useful in large applications,
10009 particularly when partial linking is used to build the application.
10011 The types of long calls used depends on the capabilities of the
10012 assembler and linker, and the type of code being generated. The
10013 impact on systems that support long absolute calls, and long pic
10014 symbol-difference or pc-relative calls should be relatively small.
10015 However, an indirect call is used on 32-bit ELF systems in pic code
10016 and it is quite long.
10018 @item -munix=@var{unix-std}
10020 Generate compiler predefines and select a startfile for the specified
10021 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
10022 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
10023 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
10024 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
10025 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
10028 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
10029 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
10030 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
10031 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
10032 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
10033 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
10035 It is @emph{important} to note that this option changes the interfaces
10036 for various library routines. It also affects the operational behavior
10037 of the C library. Thus, @emph{extreme} care is needed in using this
10040 Library code that is intended to operate with more than one UNIX
10041 standard must test, set and restore the variable @var{__xpg4_extended_mask}
10042 as appropriate. Most GNU software doesn't provide this capability.
10046 Suppress the generation of link options to search libdld.sl when the
10047 @option{-static} option is specified on HP-UX 10 and later.
10051 The HP-UX implementation of setlocale in libc has a dependency on
10052 libdld.sl. There isn't an archive version of libdld.sl. Thus,
10053 when the @option{-static} option is specified, special link options
10054 are needed to resolve this dependency.
10056 On HP-UX 10 and later, the GCC driver adds the necessary options to
10057 link with libdld.sl when the @option{-static} option is specified.
10058 This causes the resulting binary to be dynamic. On the 64-bit port,
10059 the linkers generate dynamic binaries by default in any case. The
10060 @option{-nolibdld} option can be used to prevent the GCC driver from
10061 adding these link options.
10065 Add support for multithreading with the @dfn{dce thread} library
10066 under HP-UX@. This option sets flags for both the preprocessor and
10070 @node i386 and x86-64 Options
10071 @subsection Intel 386 and AMD x86-64 Options
10072 @cindex i386 Options
10073 @cindex x86-64 Options
10074 @cindex Intel 386 Options
10075 @cindex AMD x86-64 Options
10077 These @samp{-m} options are defined for the i386 and x86-64 family of
10081 @item -mtune=@var{cpu-type}
10083 Tune to @var{cpu-type} everything applicable about the generated code, except
10084 for the ABI and the set of available instructions. The choices for
10085 @var{cpu-type} are:
10088 Produce code optimized for the most common IA32/AMD64/EM64T processors.
10089 If you know the CPU on which your code will run, then you should use
10090 the corresponding @option{-mtune} option instead of
10091 @option{-mtune=generic}. But, if you do not know exactly what CPU users
10092 of your application will have, then you should use this option.
10094 As new processors are deployed in the marketplace, the behavior of this
10095 option will change. Therefore, if you upgrade to a newer version of
10096 GCC, the code generated option will change to reflect the processors
10097 that were most common when that version of GCC was released.
10099 There is no @option{-march=generic} option because @option{-march}
10100 indicates the instruction set the compiler can use, and there is no
10101 generic instruction set applicable to all processors. In contrast,
10102 @option{-mtune} indicates the processor (or, in this case, collection of
10103 processors) for which the code is optimized.
10105 This selects the CPU to tune for at compilation time by determining
10106 the processor type of the compiling machine. Using @option{-mtune=native}
10107 will produce code optimized for the local machine under the constraints
10108 of the selected instruction set. Using @option{-march=native} will
10109 enable all instruction subsets supported by the local machine (hence
10110 the result might not run on different machines).
10112 Original Intel's i386 CPU@.
10114 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
10115 @item i586, pentium
10116 Intel Pentium CPU with no MMX support.
10118 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
10120 Intel PentiumPro CPU@.
10122 Same as @code{generic}, but when used as @code{march} option, PentiumPro
10123 instruction set will be used, so the code will run on all i686 family chips.
10125 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
10126 @item pentium3, pentium3m
10127 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
10130 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
10131 support. Used by Centrino notebooks.
10132 @item pentium4, pentium4m
10133 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
10135 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
10138 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
10139 SSE2 and SSE3 instruction set support.
10141 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
10142 instruction set support.
10144 AMD K6 CPU with MMX instruction set support.
10146 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
10147 @item athlon, athlon-tbird
10148 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
10150 @item athlon-4, athlon-xp, athlon-mp
10151 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
10152 instruction set support.
10153 @item k8, opteron, athlon64, athlon-fx
10154 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
10155 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
10156 @item k8-sse3, opteron-sse3, athlon64-sse3
10157 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
10158 @item amdfam10, barcelona
10159 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
10160 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3dNOW!, enhanced 3dNOW!, ABM and 64-bit
10161 instruction set extensions.)
10163 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
10166 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
10167 instruction set support.
10169 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
10170 implemented for this chip.)
10172 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
10173 implemented for this chip.)
10175 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
10178 While picking a specific @var{cpu-type} will schedule things appropriately
10179 for that particular chip, the compiler will not generate any code that
10180 does not run on the i386 without the @option{-march=@var{cpu-type}} option
10183 @item -march=@var{cpu-type}
10185 Generate instructions for the machine type @var{cpu-type}. The choices
10186 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
10187 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
10189 @item -mcpu=@var{cpu-type}
10191 A deprecated synonym for @option{-mtune}.
10193 @item -mfpmath=@var{unit}
10195 Generate floating point arithmetics for selected unit @var{unit}. The choices
10196 for @var{unit} are:
10200 Use the standard 387 floating point coprocessor present majority of chips and
10201 emulated otherwise. Code compiled with this option will run almost everywhere.
10202 The temporary results are computed in 80bit precision instead of precision
10203 specified by the type resulting in slightly different results compared to most
10204 of other chips. See @option{-ffloat-store} for more detailed description.
10206 This is the default choice for i386 compiler.
10209 Use scalar floating point instructions present in the SSE instruction set.
10210 This instruction set is supported by Pentium3 and newer chips, in the AMD line
10211 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
10212 instruction set supports only single precision arithmetics, thus the double and
10213 extended precision arithmetics is still done using 387. Later version, present
10214 only in Pentium4 and the future AMD x86-64 chips supports double precision
10217 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
10218 or @option{-msse2} switches to enable SSE extensions and make this option
10219 effective. For the x86-64 compiler, these extensions are enabled by default.
10221 The resulting code should be considerably faster in the majority of cases and avoid
10222 the numerical instability problems of 387 code, but may break some existing
10223 code that expects temporaries to be 80bit.
10225 This is the default choice for the x86-64 compiler.
10228 Attempt to utilize both instruction sets at once. This effectively double the
10229 amount of available registers and on chips with separate execution units for
10230 387 and SSE the execution resources too. Use this option with care, as it is
10231 still experimental, because the GCC register allocator does not model separate
10232 functional units well resulting in instable performance.
10235 @item -masm=@var{dialect}
10236 @opindex masm=@var{dialect}
10237 Output asm instructions using selected @var{dialect}. Supported
10238 choices are @samp{intel} or @samp{att} (the default one). Darwin does
10239 not support @samp{intel}.
10242 @itemx -mno-ieee-fp
10244 @opindex mno-ieee-fp
10245 Control whether or not the compiler uses IEEE floating point
10246 comparisons. These handle correctly the case where the result of a
10247 comparison is unordered.
10250 @opindex msoft-float
10251 Generate output containing library calls for floating point.
10252 @strong{Warning:} the requisite libraries are not part of GCC@.
10253 Normally the facilities of the machine's usual C compiler are used, but
10254 this can't be done directly in cross-compilation. You must make your
10255 own arrangements to provide suitable library functions for
10258 On machines where a function returns floating point results in the 80387
10259 register stack, some floating point opcodes may be emitted even if
10260 @option{-msoft-float} is used.
10262 @item -mno-fp-ret-in-387
10263 @opindex mno-fp-ret-in-387
10264 Do not use the FPU registers for return values of functions.
10266 The usual calling convention has functions return values of types
10267 @code{float} and @code{double} in an FPU register, even if there
10268 is no FPU@. The idea is that the operating system should emulate
10271 The option @option{-mno-fp-ret-in-387} causes such values to be returned
10272 in ordinary CPU registers instead.
10274 @item -mno-fancy-math-387
10275 @opindex mno-fancy-math-387
10276 Some 387 emulators do not support the @code{sin}, @code{cos} and
10277 @code{sqrt} instructions for the 387. Specify this option to avoid
10278 generating those instructions. This option is the default on FreeBSD,
10279 OpenBSD and NetBSD@. This option is overridden when @option{-march}
10280 indicates that the target cpu will always have an FPU and so the
10281 instruction will not need emulation. As of revision 2.6.1, these
10282 instructions are not generated unless you also use the
10283 @option{-funsafe-math-optimizations} switch.
10285 @item -malign-double
10286 @itemx -mno-align-double
10287 @opindex malign-double
10288 @opindex mno-align-double
10289 Control whether GCC aligns @code{double}, @code{long double}, and
10290 @code{long long} variables on a two word boundary or a one word
10291 boundary. Aligning @code{double} variables on a two word boundary will
10292 produce code that runs somewhat faster on a @samp{Pentium} at the
10293 expense of more memory.
10295 On x86-64, @option{-malign-double} is enabled by default.
10297 @strong{Warning:} if you use the @option{-malign-double} switch,
10298 structures containing the above types will be aligned differently than
10299 the published application binary interface specifications for the 386
10300 and will not be binary compatible with structures in code compiled
10301 without that switch.
10303 @item -m96bit-long-double
10304 @itemx -m128bit-long-double
10305 @opindex m96bit-long-double
10306 @opindex m128bit-long-double
10307 These switches control the size of @code{long double} type. The i386
10308 application binary interface specifies the size to be 96 bits,
10309 so @option{-m96bit-long-double} is the default in 32 bit mode.
10311 Modern architectures (Pentium and newer) would prefer @code{long double}
10312 to be aligned to an 8 or 16 byte boundary. In arrays or structures
10313 conforming to the ABI, this would not be possible. So specifying a
10314 @option{-m128bit-long-double} will align @code{long double}
10315 to a 16 byte boundary by padding the @code{long double} with an additional
10318 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
10319 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
10321 Notice that neither of these options enable any extra precision over the x87
10322 standard of 80 bits for a @code{long double}.
10324 @strong{Warning:} if you override the default value for your target ABI, the
10325 structures and arrays containing @code{long double} variables will change
10326 their size as well as function calling convention for function taking
10327 @code{long double} will be modified. Hence they will not be binary
10328 compatible with arrays or structures in code compiled without that switch.
10330 @item -mmlarge-data-threshold=@var{number}
10331 @opindex mlarge-data-threshold=@var{number}
10332 When @option{-mcmodel=medium} is specified, the data greater than
10333 @var{threshold} are placed in large data section. This value must be the
10334 same across all object linked into the binary and defaults to 65535.
10338 Use a different function-calling convention, in which functions that
10339 take a fixed number of arguments return with the @code{ret} @var{num}
10340 instruction, which pops their arguments while returning. This saves one
10341 instruction in the caller since there is no need to pop the arguments
10344 You can specify that an individual function is called with this calling
10345 sequence with the function attribute @samp{stdcall}. You can also
10346 override the @option{-mrtd} option by using the function attribute
10347 @samp{cdecl}. @xref{Function Attributes}.
10349 @strong{Warning:} this calling convention is incompatible with the one
10350 normally used on Unix, so you cannot use it if you need to call
10351 libraries compiled with the Unix compiler.
10353 Also, you must provide function prototypes for all functions that
10354 take variable numbers of arguments (including @code{printf});
10355 otherwise incorrect code will be generated for calls to those
10358 In addition, seriously incorrect code will result if you call a
10359 function with too many arguments. (Normally, extra arguments are
10360 harmlessly ignored.)
10362 @item -mregparm=@var{num}
10364 Control how many registers are used to pass integer arguments. By
10365 default, no registers are used to pass arguments, and at most 3
10366 registers can be used. You can control this behavior for a specific
10367 function by using the function attribute @samp{regparm}.
10368 @xref{Function Attributes}.
10370 @strong{Warning:} if you use this switch, and
10371 @var{num} is nonzero, then you must build all modules with the same
10372 value, including any libraries. This includes the system libraries and
10376 @opindex msseregparm
10377 Use SSE register passing conventions for float and double arguments
10378 and return values. You can control this behavior for a specific
10379 function by using the function attribute @samp{sseregparm}.
10380 @xref{Function Attributes}.
10382 @strong{Warning:} if you use this switch then you must build all
10383 modules with the same value, including any libraries. This includes
10384 the system libraries and startup modules.
10393 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
10394 is specified, the significands of results of floating-point operations are
10395 rounded to 24 bits (single precision); @option{-mpc64} rounds the the
10396 significands of results of floating-point operations to 53 bits (double
10397 precision) and @option{-mpc80} rounds the significands of results of
10398 floating-point operations to 64 bits (extended double precision), which is
10399 the default. When this option is used, floating-point operations in higher
10400 precisions are not available to the programmer without setting the FPU
10401 control word explicitly.
10403 Setting the rounding of floating-point operations to less than the default
10404 80 bits can speed some programs by 2% or more. Note that some mathematical
10405 libraries assume that extended precision (80 bit) floating-point operations
10406 are enabled by default; routines in such libraries could suffer significant
10407 loss of accuracy, typically through so-called "catastrophic cancellation",
10408 when this option is used to set the precision to less than extended precision.
10410 @item -mstackrealign
10411 @opindex mstackrealign
10412 Realign the stack at entry. On the Intel x86, the
10413 @option{-mstackrealign} option will generate an alternate prologue and
10414 epilogue that realigns the runtime stack. This supports mixing legacy
10415 codes that keep a 4-byte aligned stack with modern codes that keep a
10416 16-byte stack for SSE compatibility. The alternate prologue and
10417 epilogue are slower and bigger than the regular ones, and the
10418 alternate prologue requires an extra scratch register; this lowers the
10419 number of registers available if used in conjunction with the
10420 @code{regparm} attribute. The @option{-mstackrealign} option is
10421 incompatible with the nested function prologue; this is considered a
10422 hard error. See also the attribute @code{force_align_arg_pointer},
10423 applicable to individual functions.
10425 @item -mpreferred-stack-boundary=@var{num}
10426 @opindex mpreferred-stack-boundary
10427 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
10428 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
10429 the default is 4 (16 bytes or 128 bits).
10431 On Pentium and PentiumPro, @code{double} and @code{long double} values
10432 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
10433 suffer significant run time performance penalties. On Pentium III, the
10434 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
10435 properly if it is not 16 byte aligned.
10437 To ensure proper alignment of this values on the stack, the stack boundary
10438 must be as aligned as that required by any value stored on the stack.
10439 Further, every function must be generated such that it keeps the stack
10440 aligned. Thus calling a function compiled with a higher preferred
10441 stack boundary from a function compiled with a lower preferred stack
10442 boundary will most likely misalign the stack. It is recommended that
10443 libraries that use callbacks always use the default setting.
10445 This extra alignment does consume extra stack space, and generally
10446 increases code size. Code that is sensitive to stack space usage, such
10447 as embedded systems and operating system kernels, may want to reduce the
10448 preferred alignment to @option{-mpreferred-stack-boundary=2}.
10482 These switches enable or disable the use of instructions in the MMX,
10483 SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4A, SSE5, ABM or 3DNow! extended
10485 These extensions are also available as built-in functions: see
10486 @ref{X86 Built-in Functions}, for details of the functions enabled and
10487 disabled by these switches.
10489 To have SSE/SSE2 instructions generated automatically from floating-point
10490 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
10492 These options will enable GCC to use these extended instructions in
10493 generated code, even without @option{-mfpmath=sse}. Applications which
10494 perform runtime CPU detection must compile separate files for each
10495 supported architecture, using the appropriate flags. In particular,
10496 the file containing the CPU detection code should be compiled without
10501 This option will enable GCC to use CMPXCHG16B instruction in generated code.
10502 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
10503 data types. This is useful for high resolution counters that could be updated
10504 by multiple processors (or cores). This instruction is generated as part of
10505 atomic built-in functions: see @ref{Atomic Builtins} for details.
10509 This option will enable GCC to use SAHF instruction in generated 64-bit code.
10510 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
10511 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
10512 SAHF are load and store instructions, respectively, for certain status flags.
10513 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
10514 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
10518 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
10519 vectorized variants RCPPS and RSQRTPS) instead of DIVSS and SQRTSS (and their
10520 vectorized variants). These instructions will be generated only when
10521 @option{-funsafe-math-optimizations} is enabled.
10523 @item -mveclibabi=@var{type}
10524 @opindex mveclibabi
10525 Specifies the ABI type to use for vectorizing intrinsics using an
10526 external library. Supported types are @code{acml} for the AMD
10527 math core library style of interfacing. GCC will currently emit
10528 calls to @code{__vrd2_sin}, @code{__vrd2_cos}, @code{__vrd2_exp},
10529 @code{__vrd2_log}, @code{__vrd2_log2}, @code{__vrd2_log10},
10530 @code{__vrs4_sinf}, @code{__vrs4_cosf}, @code{__vrs4_expf},
10531 @code{__vrs4_logf}, @code{__vrs4_log2f}, @code{__vrs4_log10f}
10532 and @code{__vrs4_powf} when using this type and @option{-ftree-vectorize}
10533 is enabled. A ACML ABI compatible library will have to be specified
10537 @itemx -mno-push-args
10538 @opindex mpush-args
10539 @opindex mno-push-args
10540 Use PUSH operations to store outgoing parameters. This method is shorter
10541 and usually equally fast as method using SUB/MOV operations and is enabled
10542 by default. In some cases disabling it may improve performance because of
10543 improved scheduling and reduced dependencies.
10545 @item -maccumulate-outgoing-args
10546 @opindex maccumulate-outgoing-args
10547 If enabled, the maximum amount of space required for outgoing arguments will be
10548 computed in the function prologue. This is faster on most modern CPUs
10549 because of reduced dependencies, improved scheduling and reduced stack usage
10550 when preferred stack boundary is not equal to 2. The drawback is a notable
10551 increase in code size. This switch implies @option{-mno-push-args}.
10555 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
10556 on thread-safe exception handling must compile and link all code with the
10557 @option{-mthreads} option. When compiling, @option{-mthreads} defines
10558 @option{-D_MT}; when linking, it links in a special thread helper library
10559 @option{-lmingwthrd} which cleans up per thread exception handling data.
10561 @item -mno-align-stringops
10562 @opindex mno-align-stringops
10563 Do not align destination of inlined string operations. This switch reduces
10564 code size and improves performance in case the destination is already aligned,
10565 but GCC doesn't know about it.
10567 @item -minline-all-stringops
10568 @opindex minline-all-stringops
10569 By default GCC inlines string operations only when destination is known to be
10570 aligned at least to 4 byte boundary. This enables more inlining, increase code
10571 size, but may improve performance of code that depends on fast memcpy, strlen
10572 and memset for short lengths.
10574 @item -minline-stringops-dynamically
10575 @opindex minline-stringops-dynamically
10576 For string operation of unknown size, inline runtime checks so for small
10577 blocks inline code is used, while for large blocks library call is used.
10579 @item -mstringop-strategy=@var{alg}
10580 @opindex mstringop-strategy=@var{alg}
10581 Overwrite internal decision heuristic about particular algorithm to inline
10582 string operation with. The allowed values are @code{rep_byte},
10583 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
10584 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
10585 expanding inline loop, @code{libcall} for always expanding library call.
10587 @item -momit-leaf-frame-pointer
10588 @opindex momit-leaf-frame-pointer
10589 Don't keep the frame pointer in a register for leaf functions. This
10590 avoids the instructions to save, set up and restore frame pointers and
10591 makes an extra register available in leaf functions. The option
10592 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10593 which might make debugging harder.
10595 @item -mtls-direct-seg-refs
10596 @itemx -mno-tls-direct-seg-refs
10597 @opindex mtls-direct-seg-refs
10598 Controls whether TLS variables may be accessed with offsets from the
10599 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
10600 or whether the thread base pointer must be added. Whether or not this
10601 is legal depends on the operating system, and whether it maps the
10602 segment to cover the entire TLS area.
10604 For systems that use GNU libc, the default is on.
10607 @itemx -mno-fused-madd
10608 @opindex mfused-madd
10609 Enable automatic generation of fused floating point multiply-add instructions
10610 if the ISA supports such instructions. The -mfused-madd option is on by
10611 default. The fused multiply-add instructions have a different
10612 rounding behavior compared to executing a multiply followed by an add.
10615 These @samp{-m} switches are supported in addition to the above
10616 on AMD x86-64 processors in 64-bit environments.
10623 Generate code for a 32-bit or 64-bit environment.
10624 The 32-bit environment sets int, long and pointer to 32 bits and
10625 generates code that runs on any i386 system.
10626 The 64-bit environment sets int to 32 bits and long and pointer
10627 to 64 bits and generates code for AMD's x86-64 architecture. For
10628 darwin only the -m64 option turns off the @option{-fno-pic} and
10629 @option{-mdynamic-no-pic} options.
10631 @item -mno-red-zone
10632 @opindex no-red-zone
10633 Do not use a so called red zone for x86-64 code. The red zone is mandated
10634 by the x86-64 ABI, it is a 128-byte area beyond the location of the
10635 stack pointer that will not be modified by signal or interrupt handlers
10636 and therefore can be used for temporary data without adjusting the stack
10637 pointer. The flag @option{-mno-red-zone} disables this red zone.
10639 @item -mcmodel=small
10640 @opindex mcmodel=small
10641 Generate code for the small code model: the program and its symbols must
10642 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
10643 Programs can be statically or dynamically linked. This is the default
10646 @item -mcmodel=kernel
10647 @opindex mcmodel=kernel
10648 Generate code for the kernel code model. The kernel runs in the
10649 negative 2 GB of the address space.
10650 This model has to be used for Linux kernel code.
10652 @item -mcmodel=medium
10653 @opindex mcmodel=medium
10654 Generate code for the medium model: The program is linked in the lower 2
10655 GB of the address space but symbols can be located anywhere in the
10656 address space. Programs can be statically or dynamically linked, but
10657 building of shared libraries are not supported with the medium model.
10659 @item -mcmodel=large
10660 @opindex mcmodel=large
10661 Generate code for the large model: This model makes no assumptions
10662 about addresses and sizes of sections.
10665 @node IA-64 Options
10666 @subsection IA-64 Options
10667 @cindex IA-64 Options
10669 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10673 @opindex mbig-endian
10674 Generate code for a big endian target. This is the default for HP-UX@.
10676 @item -mlittle-endian
10677 @opindex mlittle-endian
10678 Generate code for a little endian target. This is the default for AIX5
10684 @opindex mno-gnu-as
10685 Generate (or don't) code for the GNU assembler. This is the default.
10686 @c Also, this is the default if the configure option @option{--with-gnu-as}
10692 @opindex mno-gnu-ld
10693 Generate (or don't) code for the GNU linker. This is the default.
10694 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10699 Generate code that does not use a global pointer register. The result
10700 is not position independent code, and violates the IA-64 ABI@.
10702 @item -mvolatile-asm-stop
10703 @itemx -mno-volatile-asm-stop
10704 @opindex mvolatile-asm-stop
10705 @opindex mno-volatile-asm-stop
10706 Generate (or don't) a stop bit immediately before and after volatile asm
10709 @item -mregister-names
10710 @itemx -mno-register-names
10711 @opindex mregister-names
10712 @opindex mno-register-names
10713 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
10714 the stacked registers. This may make assembler output more readable.
10720 Disable (or enable) optimizations that use the small data section. This may
10721 be useful for working around optimizer bugs.
10723 @item -mconstant-gp
10724 @opindex mconstant-gp
10725 Generate code that uses a single constant global pointer value. This is
10726 useful when compiling kernel code.
10730 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
10731 This is useful when compiling firmware code.
10733 @item -minline-float-divide-min-latency
10734 @opindex minline-float-divide-min-latency
10735 Generate code for inline divides of floating point values
10736 using the minimum latency algorithm.
10738 @item -minline-float-divide-max-throughput
10739 @opindex minline-float-divide-max-throughput
10740 Generate code for inline divides of floating point values
10741 using the maximum throughput algorithm.
10743 @item -minline-int-divide-min-latency
10744 @opindex minline-int-divide-min-latency
10745 Generate code for inline divides of integer values
10746 using the minimum latency algorithm.
10748 @item -minline-int-divide-max-throughput
10749 @opindex minline-int-divide-max-throughput
10750 Generate code for inline divides of integer values
10751 using the maximum throughput algorithm.
10753 @item -minline-sqrt-min-latency
10754 @opindex minline-sqrt-min-latency
10755 Generate code for inline square roots
10756 using the minimum latency algorithm.
10758 @item -minline-sqrt-max-throughput
10759 @opindex minline-sqrt-max-throughput
10760 Generate code for inline square roots
10761 using the maximum throughput algorithm.
10763 @item -mno-dwarf2-asm
10764 @itemx -mdwarf2-asm
10765 @opindex mno-dwarf2-asm
10766 @opindex mdwarf2-asm
10767 Don't (or do) generate assembler code for the DWARF2 line number debugging
10768 info. This may be useful when not using the GNU assembler.
10770 @item -mearly-stop-bits
10771 @itemx -mno-early-stop-bits
10772 @opindex mearly-stop-bits
10773 @opindex mno-early-stop-bits
10774 Allow stop bits to be placed earlier than immediately preceding the
10775 instruction that triggered the stop bit. This can improve instruction
10776 scheduling, but does not always do so.
10778 @item -mfixed-range=@var{register-range}
10779 @opindex mfixed-range
10780 Generate code treating the given register range as fixed registers.
10781 A fixed register is one that the register allocator can not use. This is
10782 useful when compiling kernel code. A register range is specified as
10783 two registers separated by a dash. Multiple register ranges can be
10784 specified separated by a comma.
10786 @item -mtls-size=@var{tls-size}
10788 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
10791 @item -mtune=@var{cpu-type}
10793 Tune the instruction scheduling for a particular CPU, Valid values are
10794 itanium, itanium1, merced, itanium2, and mckinley.
10800 Add support for multithreading using the POSIX threads library. This
10801 option sets flags for both the preprocessor and linker. It does
10802 not affect the thread safety of object code produced by the compiler or
10803 that of libraries supplied with it. These are HP-UX specific flags.
10809 Generate code for a 32-bit or 64-bit environment.
10810 The 32-bit environment sets int, long and pointer to 32 bits.
10811 The 64-bit environment sets int to 32 bits and long and pointer
10812 to 64 bits. These are HP-UX specific flags.
10814 @item -mno-sched-br-data-spec
10815 @itemx -msched-br-data-spec
10816 @opindex mno-sched-br-data-spec
10817 @opindex msched-br-data-spec
10818 (Dis/En)able data speculative scheduling before reload.
10819 This will result in generation of the ld.a instructions and
10820 the corresponding check instructions (ld.c / chk.a).
10821 The default is 'disable'.
10823 @item -msched-ar-data-spec
10824 @itemx -mno-sched-ar-data-spec
10825 @opindex msched-ar-data-spec
10826 @opindex mno-sched-ar-data-spec
10827 (En/Dis)able data speculative scheduling after reload.
10828 This will result in generation of the ld.a instructions and
10829 the corresponding check instructions (ld.c / chk.a).
10830 The default is 'enable'.
10832 @item -mno-sched-control-spec
10833 @itemx -msched-control-spec
10834 @opindex mno-sched-control-spec
10835 @opindex msched-control-spec
10836 (Dis/En)able control speculative scheduling. This feature is
10837 available only during region scheduling (i.e. before reload).
10838 This will result in generation of the ld.s instructions and
10839 the corresponding check instructions chk.s .
10840 The default is 'disable'.
10842 @item -msched-br-in-data-spec
10843 @itemx -mno-sched-br-in-data-spec
10844 @opindex msched-br-in-data-spec
10845 @opindex mno-sched-br-in-data-spec
10846 (En/Dis)able speculative scheduling of the instructions that
10847 are dependent on the data speculative loads before reload.
10848 This is effective only with @option{-msched-br-data-spec} enabled.
10849 The default is 'enable'.
10851 @item -msched-ar-in-data-spec
10852 @itemx -mno-sched-ar-in-data-spec
10853 @opindex msched-ar-in-data-spec
10854 @opindex mno-sched-ar-in-data-spec
10855 (En/Dis)able speculative scheduling of the instructions that
10856 are dependent on the data speculative loads after reload.
10857 This is effective only with @option{-msched-ar-data-spec} enabled.
10858 The default is 'enable'.
10860 @item -msched-in-control-spec
10861 @itemx -mno-sched-in-control-spec
10862 @opindex msched-in-control-spec
10863 @opindex mno-sched-in-control-spec
10864 (En/Dis)able speculative scheduling of the instructions that
10865 are dependent on the control speculative loads.
10866 This is effective only with @option{-msched-control-spec} enabled.
10867 The default is 'enable'.
10870 @itemx -mno-sched-ldc
10871 @opindex msched-ldc
10872 @opindex mno-sched-ldc
10873 (En/Dis)able use of simple data speculation checks ld.c .
10874 If disabled, only chk.a instructions will be emitted to check
10875 data speculative loads.
10876 The default is 'enable'.
10878 @item -mno-sched-control-ldc
10879 @itemx -msched-control-ldc
10880 @opindex mno-sched-control-ldc
10881 @opindex msched-control-ldc
10882 (Dis/En)able use of ld.c instructions to check control speculative loads.
10883 If enabled, in case of control speculative load with no speculatively
10884 scheduled dependent instructions this load will be emitted as ld.sa and
10885 ld.c will be used to check it.
10886 The default is 'disable'.
10888 @item -mno-sched-spec-verbose
10889 @itemx -msched-spec-verbose
10890 @opindex mno-sched-spec-verbose
10891 @opindex msched-spec-verbose
10892 (Dis/En)able printing of the information about speculative motions.
10894 @item -mno-sched-prefer-non-data-spec-insns
10895 @itemx -msched-prefer-non-data-spec-insns
10896 @opindex mno-sched-prefer-non-data-spec-insns
10897 @opindex msched-prefer-non-data-spec-insns
10898 If enabled, data speculative instructions will be chosen for schedule
10899 only if there are no other choices at the moment. This will make
10900 the use of the data speculation much more conservative.
10901 The default is 'disable'.
10903 @item -mno-sched-prefer-non-control-spec-insns
10904 @itemx -msched-prefer-non-control-spec-insns
10905 @opindex mno-sched-prefer-non-control-spec-insns
10906 @opindex msched-prefer-non-control-spec-insns
10907 If enabled, control speculative instructions will be chosen for schedule
10908 only if there are no other choices at the moment. This will make
10909 the use of the control speculation much more conservative.
10910 The default is 'disable'.
10912 @item -mno-sched-count-spec-in-critical-path
10913 @itemx -msched-count-spec-in-critical-path
10914 @opindex mno-sched-count-spec-in-critical-path
10915 @opindex msched-count-spec-in-critical-path
10916 If enabled, speculative dependencies will be considered during
10917 computation of the instructions priorities. This will make the use of the
10918 speculation a bit more conservative.
10919 The default is 'disable'.
10924 @subsection M32C Options
10925 @cindex M32C options
10928 @item -mcpu=@var{name}
10930 Select the CPU for which code is generated. @var{name} may be one of
10931 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
10932 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
10933 the M32C/80 series.
10937 Specifies that the program will be run on the simulator. This causes
10938 an alternate runtime library to be linked in which supports, for
10939 example, file I/O. You must not use this option when generating
10940 programs that will run on real hardware; you must provide your own
10941 runtime library for whatever I/O functions are needed.
10943 @item -memregs=@var{number}
10945 Specifies the number of memory-based pseudo-registers GCC will use
10946 during code generation. These pseudo-registers will be used like real
10947 registers, so there is a tradeoff between GCC's ability to fit the
10948 code into available registers, and the performance penalty of using
10949 memory instead of registers. Note that all modules in a program must
10950 be compiled with the same value for this option. Because of that, you
10951 must not use this option with the default runtime libraries gcc
10956 @node M32R/D Options
10957 @subsection M32R/D Options
10958 @cindex M32R/D options
10960 These @option{-m} options are defined for Renesas M32R/D architectures:
10965 Generate code for the M32R/2@.
10969 Generate code for the M32R/X@.
10973 Generate code for the M32R@. This is the default.
10975 @item -mmodel=small
10976 @opindex mmodel=small
10977 Assume all objects live in the lower 16MB of memory (so that their addresses
10978 can be loaded with the @code{ld24} instruction), and assume all subroutines
10979 are reachable with the @code{bl} instruction.
10980 This is the default.
10982 The addressability of a particular object can be set with the
10983 @code{model} attribute.
10985 @item -mmodel=medium
10986 @opindex mmodel=medium
10987 Assume objects may be anywhere in the 32-bit address space (the compiler
10988 will generate @code{seth/add3} instructions to load their addresses), and
10989 assume all subroutines are reachable with the @code{bl} instruction.
10991 @item -mmodel=large
10992 @opindex mmodel=large
10993 Assume objects may be anywhere in the 32-bit address space (the compiler
10994 will generate @code{seth/add3} instructions to load their addresses), and
10995 assume subroutines may not be reachable with the @code{bl} instruction
10996 (the compiler will generate the much slower @code{seth/add3/jl}
10997 instruction sequence).
11000 @opindex msdata=none
11001 Disable use of the small data area. Variables will be put into
11002 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
11003 @code{section} attribute has been specified).
11004 This is the default.
11006 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
11007 Objects may be explicitly put in the small data area with the
11008 @code{section} attribute using one of these sections.
11010 @item -msdata=sdata
11011 @opindex msdata=sdata
11012 Put small global and static data in the small data area, but do not
11013 generate special code to reference them.
11016 @opindex msdata=use
11017 Put small global and static data in the small data area, and generate
11018 special instructions to reference them.
11022 @cindex smaller data references
11023 Put global and static objects less than or equal to @var{num} bytes
11024 into the small data or bss sections instead of the normal data or bss
11025 sections. The default value of @var{num} is 8.
11026 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
11027 for this option to have any effect.
11029 All modules should be compiled with the same @option{-G @var{num}} value.
11030 Compiling with different values of @var{num} may or may not work; if it
11031 doesn't the linker will give an error message---incorrect code will not be
11036 Makes the M32R specific code in the compiler display some statistics
11037 that might help in debugging programs.
11039 @item -malign-loops
11040 @opindex malign-loops
11041 Align all loops to a 32-byte boundary.
11043 @item -mno-align-loops
11044 @opindex mno-align-loops
11045 Do not enforce a 32-byte alignment for loops. This is the default.
11047 @item -missue-rate=@var{number}
11048 @opindex missue-rate=@var{number}
11049 Issue @var{number} instructions per cycle. @var{number} can only be 1
11052 @item -mbranch-cost=@var{number}
11053 @opindex mbranch-cost=@var{number}
11054 @var{number} can only be 1 or 2. If it is 1 then branches will be
11055 preferred over conditional code, if it is 2, then the opposite will
11058 @item -mflush-trap=@var{number}
11059 @opindex mflush-trap=@var{number}
11060 Specifies the trap number to use to flush the cache. The default is
11061 12. Valid numbers are between 0 and 15 inclusive.
11063 @item -mno-flush-trap
11064 @opindex mno-flush-trap
11065 Specifies that the cache cannot be flushed by using a trap.
11067 @item -mflush-func=@var{name}
11068 @opindex mflush-func=@var{name}
11069 Specifies the name of the operating system function to call to flush
11070 the cache. The default is @emph{_flush_cache}, but a function call
11071 will only be used if a trap is not available.
11073 @item -mno-flush-func
11074 @opindex mno-flush-func
11075 Indicates that there is no OS function for flushing the cache.
11079 @node M680x0 Options
11080 @subsection M680x0 Options
11081 @cindex M680x0 options
11083 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
11084 The default settings depend on which architecture was selected when
11085 the compiler was configured; the defaults for the most common choices
11089 @item -march=@var{arch}
11091 Generate code for a specific M680x0 or ColdFire instruction set
11092 architecture. Permissible values of @var{arch} for M680x0
11093 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
11094 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
11095 architectures are selected according to Freescale's ISA classification
11096 and the permissible values are: @samp{isaa}, @samp{isaaplus},
11097 @samp{isab} and @samp{isac}.
11099 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
11100 code for a ColdFire target. The @var{arch} in this macro is one of the
11101 @option{-march} arguments given above.
11103 When used together, @option{-march} and @option{-mtune} select code
11104 that runs on a family of similar processors but that is optimized
11105 for a particular microarchitecture.
11107 @item -mcpu=@var{cpu}
11109 Generate code for a specific M680x0 or ColdFire processor.
11110 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
11111 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
11112 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
11113 below, which also classifies the CPUs into families:
11115 @multitable @columnfractions 0.20 0.80
11116 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
11117 @item @samp{51qe} @tab @samp{51qe}
11118 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
11119 @item @samp{5206e} @tab @samp{5206e}
11120 @item @samp{5208} @tab @samp{5207} @samp{5208}
11121 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
11122 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
11123 @item @samp{5216} @tab @samp{5214} @samp{5216}
11124 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
11125 @item @samp{5225} @tab @samp{5224} @samp{5225}
11126 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
11127 @item @samp{5249} @tab @samp{5249}
11128 @item @samp{5250} @tab @samp{5250}
11129 @item @samp{5271} @tab @samp{5270} @samp{5271}
11130 @item @samp{5272} @tab @samp{5272}
11131 @item @samp{5275} @tab @samp{5274} @samp{5275}
11132 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
11133 @item @samp{5307} @tab @samp{5307}
11134 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
11135 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
11136 @item @samp{5407} @tab @samp{5407}
11137 @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}
11140 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
11141 @var{arch} is compatible with @var{cpu}. Other combinations of
11142 @option{-mcpu} and @option{-march} are rejected.
11144 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
11145 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
11146 where the value of @var{family} is given by the table above.
11148 @item -mtune=@var{tune}
11150 Tune the code for a particular microarchitecture, within the
11151 constraints set by @option{-march} and @option{-mcpu}.
11152 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
11153 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
11154 and @samp{cpu32}. The ColdFire microarchitectures
11155 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
11157 You can also use @option{-mtune=68020-40} for code that needs
11158 to run relatively well on 68020, 68030 and 68040 targets.
11159 @option{-mtune=68020-60} is similar but includes 68060 targets
11160 as well. These two options select the same tuning decisions as
11161 @option{-m68020-40} and @option{-m68020-60} respectively.
11163 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
11164 when tuning for 680x0 architecture @var{arch}. It also defines
11165 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
11166 option is used. If gcc is tuning for a range of architectures,
11167 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
11168 it defines the macros for every architecture in the range.
11170 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
11171 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
11172 of the arguments given above.
11178 Generate output for a 68000. This is the default
11179 when the compiler is configured for 68000-based systems.
11180 It is equivalent to @option{-march=68000}.
11182 Use this option for microcontrollers with a 68000 or EC000 core,
11183 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
11187 Generate output for a 68010. This is the default
11188 when the compiler is configured for 68010-based systems.
11189 It is equivalent to @option{-march=68010}.
11195 Generate output for a 68020. This is the default
11196 when the compiler is configured for 68020-based systems.
11197 It is equivalent to @option{-march=68020}.
11201 Generate output for a 68030. This is the default when the compiler is
11202 configured for 68030-based systems. It is equivalent to
11203 @option{-march=68030}.
11207 Generate output for a 68040. This is the default when the compiler is
11208 configured for 68040-based systems. It is equivalent to
11209 @option{-march=68040}.
11211 This option inhibits the use of 68881/68882 instructions that have to be
11212 emulated by software on the 68040. Use this option if your 68040 does not
11213 have code to emulate those instructions.
11217 Generate output for a 68060. This is the default when the compiler is
11218 configured for 68060-based systems. It is equivalent to
11219 @option{-march=68060}.
11221 This option inhibits the use of 68020 and 68881/68882 instructions that
11222 have to be emulated by software on the 68060. Use this option if your 68060
11223 does not have code to emulate those instructions.
11227 Generate output for a CPU32. This is the default
11228 when the compiler is configured for CPU32-based systems.
11229 It is equivalent to @option{-march=cpu32}.
11231 Use this option for microcontrollers with a
11232 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
11233 68336, 68340, 68341, 68349 and 68360.
11237 Generate output for a 520X ColdFire CPU. This is the default
11238 when the compiler is configured for 520X-based systems.
11239 It is equivalent to @option{-mcpu=5206}, and is now deprecated
11240 in favor of that option.
11242 Use this option for microcontroller with a 5200 core, including
11243 the MCF5202, MCF5203, MCF5204 and MCF5206.
11247 Generate output for a 5206e ColdFire CPU. The option is now
11248 deprecated in favor of the equivalent @option{-mcpu=5206e}.
11252 Generate output for a member of the ColdFire 528X family.
11253 The option is now deprecated in favor of the equivalent
11254 @option{-mcpu=528x}.
11258 Generate output for a ColdFire 5307 CPU. The option is now deprecated
11259 in favor of the equivalent @option{-mcpu=5307}.
11263 Generate output for a ColdFire 5407 CPU. The option is now deprecated
11264 in favor of the equivalent @option{-mcpu=5407}.
11268 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
11269 This includes use of hardware floating point instructions.
11270 The option is equivalent to @option{-mcpu=547x}, and is now
11271 deprecated in favor of that option.
11275 Generate output for a 68040, without using any of the new instructions.
11276 This results in code which can run relatively efficiently on either a
11277 68020/68881 or a 68030 or a 68040. The generated code does use the
11278 68881 instructions that are emulated on the 68040.
11280 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
11284 Generate output for a 68060, without using any of the new instructions.
11285 This results in code which can run relatively efficiently on either a
11286 68020/68881 or a 68030 or a 68040. The generated code does use the
11287 68881 instructions that are emulated on the 68060.
11289 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
11293 @opindex mhard-float
11295 Generate floating-point instructions. This is the default for 68020
11296 and above, and for ColdFire devices that have an FPU. It defines the
11297 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
11298 on ColdFire targets.
11301 @opindex msoft-float
11302 Do not generate floating-point instructions; use library calls instead.
11303 This is the default for 68000, 68010, and 68832 targets. It is also
11304 the default for ColdFire devices that have no FPU.
11310 Generate (do not generate) ColdFire hardware divide and remainder
11311 instructions. If @option{-march} is used without @option{-mcpu},
11312 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
11313 architectures. Otherwise, the default is taken from the target CPU
11314 (either the default CPU, or the one specified by @option{-mcpu}). For
11315 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
11316 @option{-mcpu=5206e}.
11318 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
11322 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11323 Additionally, parameters passed on the stack are also aligned to a
11324 16-bit boundary even on targets whose API mandates promotion to 32-bit.
11328 Do not consider type @code{int} to be 16 bits wide. This is the default.
11331 @itemx -mno-bitfield
11332 @opindex mnobitfield
11333 @opindex mno-bitfield
11334 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
11335 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
11339 Do use the bit-field instructions. The @option{-m68020} option implies
11340 @option{-mbitfield}. This is the default if you use a configuration
11341 designed for a 68020.
11345 Use a different function-calling convention, in which functions
11346 that take a fixed number of arguments return with the @code{rtd}
11347 instruction, which pops their arguments while returning. This
11348 saves one instruction in the caller since there is no need to pop
11349 the arguments there.
11351 This calling convention is incompatible with the one normally
11352 used on Unix, so you cannot use it if you need to call libraries
11353 compiled with the Unix compiler.
11355 Also, you must provide function prototypes for all functions that
11356 take variable numbers of arguments (including @code{printf});
11357 otherwise incorrect code will be generated for calls to those
11360 In addition, seriously incorrect code will result if you call a
11361 function with too many arguments. (Normally, extra arguments are
11362 harmlessly ignored.)
11364 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
11365 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
11369 Do not use the calling conventions selected by @option{-mrtd}.
11370 This is the default.
11373 @itemx -mno-align-int
11374 @opindex malign-int
11375 @opindex mno-align-int
11376 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
11377 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
11378 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
11379 Aligning variables on 32-bit boundaries produces code that runs somewhat
11380 faster on processors with 32-bit busses at the expense of more memory.
11382 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
11383 align structures containing the above types differently than
11384 most published application binary interface specifications for the m68k.
11388 Use the pc-relative addressing mode of the 68000 directly, instead of
11389 using a global offset table. At present, this option implies @option{-fpic},
11390 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
11391 not presently supported with @option{-mpcrel}, though this could be supported for
11392 68020 and higher processors.
11394 @item -mno-strict-align
11395 @itemx -mstrict-align
11396 @opindex mno-strict-align
11397 @opindex mstrict-align
11398 Do not (do) assume that unaligned memory references will be handled by
11402 Generate code that allows the data segment to be located in a different
11403 area of memory from the text segment. This allows for execute in place in
11404 an environment without virtual memory management. This option implies
11407 @item -mno-sep-data
11408 Generate code that assumes that the data segment follows the text segment.
11409 This is the default.
11411 @item -mid-shared-library
11412 Generate code that supports shared libraries via the library ID method.
11413 This allows for execute in place and shared libraries in an environment
11414 without virtual memory management. This option implies @option{-fPIC}.
11416 @item -mno-id-shared-library
11417 Generate code that doesn't assume ID based shared libraries are being used.
11418 This is the default.
11420 @item -mshared-library-id=n
11421 Specified the identification number of the ID based shared library being
11422 compiled. Specifying a value of 0 will generate more compact code, specifying
11423 other values will force the allocation of that number to the current
11424 library but is no more space or time efficient than omitting this option.
11428 @node M68hc1x Options
11429 @subsection M68hc1x Options
11430 @cindex M68hc1x options
11432 These are the @samp{-m} options defined for the 68hc11 and 68hc12
11433 microcontrollers. The default values for these options depends on
11434 which style of microcontroller was selected when the compiler was configured;
11435 the defaults for the most common choices are given below.
11442 Generate output for a 68HC11. This is the default
11443 when the compiler is configured for 68HC11-based systems.
11449 Generate output for a 68HC12. This is the default
11450 when the compiler is configured for 68HC12-based systems.
11456 Generate output for a 68HCS12.
11458 @item -mauto-incdec
11459 @opindex mauto-incdec
11460 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
11467 Enable the use of 68HC12 min and max instructions.
11470 @itemx -mno-long-calls
11471 @opindex mlong-calls
11472 @opindex mno-long-calls
11473 Treat all calls as being far away (near). If calls are assumed to be
11474 far away, the compiler will use the @code{call} instruction to
11475 call a function and the @code{rtc} instruction for returning.
11479 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11481 @item -msoft-reg-count=@var{count}
11482 @opindex msoft-reg-count
11483 Specify the number of pseudo-soft registers which are used for the
11484 code generation. The maximum number is 32. Using more pseudo-soft
11485 register may or may not result in better code depending on the program.
11486 The default is 4 for 68HC11 and 2 for 68HC12.
11490 @node MCore Options
11491 @subsection MCore Options
11492 @cindex MCore options
11494 These are the @samp{-m} options defined for the Motorola M*Core
11500 @itemx -mno-hardlit
11502 @opindex mno-hardlit
11503 Inline constants into the code stream if it can be done in two
11504 instructions or less.
11510 Use the divide instruction. (Enabled by default).
11512 @item -mrelax-immediate
11513 @itemx -mno-relax-immediate
11514 @opindex mrelax-immediate
11515 @opindex mno-relax-immediate
11516 Allow arbitrary sized immediates in bit operations.
11518 @item -mwide-bitfields
11519 @itemx -mno-wide-bitfields
11520 @opindex mwide-bitfields
11521 @opindex mno-wide-bitfields
11522 Always treat bit-fields as int-sized.
11524 @item -m4byte-functions
11525 @itemx -mno-4byte-functions
11526 @opindex m4byte-functions
11527 @opindex mno-4byte-functions
11528 Force all functions to be aligned to a four byte boundary.
11530 @item -mcallgraph-data
11531 @itemx -mno-callgraph-data
11532 @opindex mcallgraph-data
11533 @opindex mno-callgraph-data
11534 Emit callgraph information.
11537 @itemx -mno-slow-bytes
11538 @opindex mslow-bytes
11539 @opindex mno-slow-bytes
11540 Prefer word access when reading byte quantities.
11542 @item -mlittle-endian
11543 @itemx -mbig-endian
11544 @opindex mlittle-endian
11545 @opindex mbig-endian
11546 Generate code for a little endian target.
11552 Generate code for the 210 processor.
11556 @subsection MIPS Options
11557 @cindex MIPS options
11563 Generate big-endian code.
11567 Generate little-endian code. This is the default for @samp{mips*el-*-*}
11570 @item -march=@var{arch}
11572 Generate code that will run on @var{arch}, which can be the name of a
11573 generic MIPS ISA, or the name of a particular processor.
11575 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
11576 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
11577 The processor names are:
11578 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
11579 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
11580 @samp{5kc}, @samp{5kf},
11582 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
11583 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
11584 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
11585 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
11588 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
11589 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
11590 @samp{rm7000}, @samp{rm9000},
11593 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
11594 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
11595 The special value @samp{from-abi} selects the
11596 most compatible architecture for the selected ABI (that is,
11597 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
11599 In processor names, a final @samp{000} can be abbreviated as @samp{k}
11600 (for example, @samp{-march=r2k}). Prefixes are optional, and
11601 @samp{vr} may be written @samp{r}.
11603 Names of the form @samp{@var{n}f2_1} refer to processors with
11604 FPUs clocked at half the rate of the core, names of the form
11605 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
11606 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
11607 processors with FPUs clocked a ratio of 3:2 with respect to the core.
11608 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
11609 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
11610 accepted as synonyms for @samp{@var{n}f1_1}.
11612 GCC defines two macros based on the value of this option. The first
11613 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
11614 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
11615 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
11616 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
11617 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
11619 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
11620 above. In other words, it will have the full prefix and will not
11621 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
11622 the macro names the resolved architecture (either @samp{"mips1"} or
11623 @samp{"mips3"}). It names the default architecture when no
11624 @option{-march} option is given.
11626 @item -mtune=@var{arch}
11628 Optimize for @var{arch}. Among other things, this option controls
11629 the way instructions are scheduled, and the perceived cost of arithmetic
11630 operations. The list of @var{arch} values is the same as for
11633 When this option is not used, GCC will optimize for the processor
11634 specified by @option{-march}. By using @option{-march} and
11635 @option{-mtune} together, it is possible to generate code that will
11636 run on a family of processors, but optimize the code for one
11637 particular member of that family.
11639 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
11640 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
11641 @samp{-march} ones described above.
11645 Equivalent to @samp{-march=mips1}.
11649 Equivalent to @samp{-march=mips2}.
11653 Equivalent to @samp{-march=mips3}.
11657 Equivalent to @samp{-march=mips4}.
11661 Equivalent to @samp{-march=mips32}.
11665 Equivalent to @samp{-march=mips32r2}.
11669 Equivalent to @samp{-march=mips64}.
11674 @opindex mno-mips16
11675 Generate (do not generate) MIPS16 code. If GCC is targetting a
11676 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
11678 MIPS16 code generation can also be controlled on a per-function basis
11679 by means of @code{mips16} and @code{nomips16} attributes.
11680 @xref{Function Attributes}, for more information.
11682 @item -mflip-mips16
11683 @opindex mflip-mips16
11684 Generate MIPS16 code on alternating functions. This option is provided
11685 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
11686 not intended for ordinary use in compiling user code.
11688 @item -minterlink-mips16
11689 @itemx -mno-interlink-mips16
11690 @opindex minterlink-mips16
11691 @opindex mno-interlink-mips16
11692 Require (do not require) that non-MIPS16 code be link-compatible with
11695 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
11696 it must either use a call or an indirect jump. @option{-minterlink-mips16}
11697 therefore disables direct jumps unless GCC knows that the target of the
11698 jump is not MIPS16.
11710 Generate code for the given ABI@.
11712 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
11713 generates 64-bit code when you select a 64-bit architecture, but you
11714 can use @option{-mgp32} to get 32-bit code instead.
11716 For information about the O64 ABI, see
11717 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
11719 GCC supports a variant of the o32 ABI in which floating-point registers
11720 are 64 rather than 32 bits wide. You can select this combination with
11721 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
11722 and @samp{mfhc1} instructions and is therefore only supported for
11723 MIPS32R2 processors.
11725 The register assignments for arguments and return values remain the
11726 same, but each scalar value is passed in a single 64-bit register
11727 rather than a pair of 32-bit registers. For example, scalar
11728 floating-point values are returned in @samp{$f0} only, not a
11729 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
11730 remains the same, but all 64 bits are saved.
11733 @itemx -mno-abicalls
11735 @opindex mno-abicalls
11736 Generate (do not generate) code that is suitable for SVR4-style
11737 dynamic objects. @option{-mabicalls} is the default for SVR4-based
11742 Generate (do not generate) code that is fully position-independent,
11743 and that can therefore be linked into shared libraries. This option
11744 only affects @option{-mabicalls}.
11746 All @option{-mabicalls} code has traditionally been position-independent,
11747 regardless of options like @option{-fPIC} and @option{-fpic}. However,
11748 as an extension, the GNU toolchain allows executables to use absolute
11749 accesses for locally-binding symbols. It can also use shorter GP
11750 initialization sequences and generate direct calls to locally-defined
11751 functions. This mode is selected by @option{-mno-shared}.
11753 @option{-mno-shared} depends on binutils 2.16 or higher and generates
11754 objects that can only be linked by the GNU linker. However, the option
11755 does not affect the ABI of the final executable; it only affects the ABI
11756 of relocatable objects. Using @option{-mno-shared} will generally make
11757 executables both smaller and quicker.
11759 @option{-mshared} is the default.
11765 Lift (do not lift) the usual restrictions on the size of the global
11768 GCC normally uses a single instruction to load values from the GOT@.
11769 While this is relatively efficient, it will only work if the GOT
11770 is smaller than about 64k. Anything larger will cause the linker
11771 to report an error such as:
11773 @cindex relocation truncated to fit (MIPS)
11775 relocation truncated to fit: R_MIPS_GOT16 foobar
11778 If this happens, you should recompile your code with @option{-mxgot}.
11779 It should then work with very large GOTs, although it will also be
11780 less efficient, since it will take three instructions to fetch the
11781 value of a global symbol.
11783 Note that some linkers can create multiple GOTs. If you have such a
11784 linker, you should only need to use @option{-mxgot} when a single object
11785 file accesses more than 64k's worth of GOT entries. Very few do.
11787 These options have no effect unless GCC is generating position
11792 Assume that general-purpose registers are 32 bits wide.
11796 Assume that general-purpose registers are 64 bits wide.
11800 Assume that floating-point registers are 32 bits wide.
11804 Assume that floating-point registers are 64 bits wide.
11807 @opindex mhard-float
11808 Use floating-point coprocessor instructions.
11811 @opindex msoft-float
11812 Do not use floating-point coprocessor instructions. Implement
11813 floating-point calculations using library calls instead.
11815 @item -msingle-float
11816 @opindex msingle-float
11817 Assume that the floating-point coprocessor only supports single-precision
11820 @item -mdouble-float
11821 @opindex mdouble-float
11822 Assume that the floating-point coprocessor supports double-precision
11823 operations. This is the default.
11829 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
11830 implement atomic memory built-in functions. When neither option is
11831 specified, GCC will use the instructions if the target architecture
11834 @option{-mllsc} is useful if the runtime environment can emulate the
11835 instructions and @option{-mno-llsc} can be useful when compiling for
11836 nonstandard ISAs. You can make either option the default by
11837 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
11838 respectively. @option{--with-llsc} is the default for some
11839 configurations; see the installation documentation for details.
11845 Use (do not use) revision 1 of the MIPS DSP ASE.
11846 @xref{MIPS DSP Built-in Functions}. This option defines the
11847 preprocessor macro @samp{__mips_dsp}. It also defines
11848 @samp{__mips_dsp_rev} to 1.
11854 Use (do not use) revision 2 of the MIPS DSP ASE.
11855 @xref{MIPS DSP Built-in Functions}. This option defines the
11856 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
11857 It also defines @samp{__mips_dsp_rev} to 2.
11860 @itemx -mno-smartmips
11861 @opindex msmartmips
11862 @opindex mno-smartmips
11863 Use (do not use) the MIPS SmartMIPS ASE.
11865 @item -mpaired-single
11866 @itemx -mno-paired-single
11867 @opindex mpaired-single
11868 @opindex mno-paired-single
11869 Use (do not use) paired-single floating-point instructions.
11870 @xref{MIPS Paired-Single Support}. This option can only be used
11871 when generating 64-bit code and requires hardware floating-point
11872 support to be enabled.
11878 Use (do not use) MIPS Digital Media Extension instructions.
11879 This option can only be used when generating 64-bit code and requires
11880 hardware floating-point support to be enabled.
11885 @opindex mno-mips3d
11886 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
11887 The option @option{-mips3d} implies @option{-mpaired-single}.
11893 Use (do not use) MT Multithreading instructions.
11897 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
11898 an explanation of the default and the way that the pointer size is
11903 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
11905 The default size of @code{int}s, @code{long}s and pointers depends on
11906 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
11907 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
11908 32-bit @code{long}s. Pointers are the same size as @code{long}s,
11909 or the same size as integer registers, whichever is smaller.
11915 Assume (do not assume) that all symbols have 32-bit values, regardless
11916 of the selected ABI@. This option is useful in combination with
11917 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
11918 to generate shorter and faster references to symbolic addresses.
11922 Put definitions of externally-visible data in a small data section
11923 if that data is no bigger than @var{num} bytes. GCC can then access
11924 the data more efficiently; see @option{-mgpopt} for details.
11926 The default @option{-G} option depends on the configuration.
11928 @item -mlocal-sdata
11929 @itemx -mno-local-sdata
11930 @opindex mlocal-sdata
11931 @opindex mno-local-sdata
11932 Extend (do not extend) the @option{-G} behavior to local data too,
11933 such as to static variables in C. @option{-mlocal-sdata} is the
11934 default for all configurations.
11936 If the linker complains that an application is using too much small data,
11937 you might want to try rebuilding the less performance-critical parts with
11938 @option{-mno-local-sdata}. You might also want to build large
11939 libraries with @option{-mno-local-sdata}, so that the libraries leave
11940 more room for the main program.
11942 @item -mextern-sdata
11943 @itemx -mno-extern-sdata
11944 @opindex mextern-sdata
11945 @opindex mno-extern-sdata
11946 Assume (do not assume) that externally-defined data will be in
11947 a small data section if that data is within the @option{-G} limit.
11948 @option{-mextern-sdata} is the default for all configurations.
11950 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
11951 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
11952 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
11953 is placed in a small data section. If @var{Var} is defined by another
11954 module, you must either compile that module with a high-enough
11955 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
11956 definition. If @var{Var} is common, you must link the application
11957 with a high-enough @option{-G} setting.
11959 The easiest way of satisfying these restrictions is to compile
11960 and link every module with the same @option{-G} option. However,
11961 you may wish to build a library that supports several different
11962 small data limits. You can do this by compiling the library with
11963 the highest supported @option{-G} setting and additionally using
11964 @option{-mno-extern-sdata} to stop the library from making assumptions
11965 about externally-defined data.
11971 Use (do not use) GP-relative accesses for symbols that are known to be
11972 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
11973 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
11976 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
11977 might not hold the value of @code{_gp}. For example, if the code is
11978 part of a library that might be used in a boot monitor, programs that
11979 call boot monitor routines will pass an unknown value in @code{$gp}.
11980 (In such situations, the boot monitor itself would usually be compiled
11981 with @option{-G0}.)
11983 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
11984 @option{-mno-extern-sdata}.
11986 @item -membedded-data
11987 @itemx -mno-embedded-data
11988 @opindex membedded-data
11989 @opindex mno-embedded-data
11990 Allocate variables to the read-only data section first if possible, then
11991 next in the small data section if possible, otherwise in data. This gives
11992 slightly slower code than the default, but reduces the amount of RAM required
11993 when executing, and thus may be preferred for some embedded systems.
11995 @item -muninit-const-in-rodata
11996 @itemx -mno-uninit-const-in-rodata
11997 @opindex muninit-const-in-rodata
11998 @opindex mno-uninit-const-in-rodata
11999 Put uninitialized @code{const} variables in the read-only data section.
12000 This option is only meaningful in conjunction with @option{-membedded-data}.
12002 @item -mcode-readable=@var{setting}
12003 @opindex mcode-readable
12004 Specify whether GCC may generate code that reads from executable sections.
12005 There are three possible settings:
12008 @item -mcode-readable=yes
12009 Instructions may freely access executable sections. This is the
12012 @item -mcode-readable=pcrel
12013 MIPS16 PC-relative load instructions can access executable sections,
12014 but other instructions must not do so. This option is useful on 4KSc
12015 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
12016 It is also useful on processors that can be configured to have a dual
12017 instruction/data SRAM interface and that, like the M4K, automatically
12018 redirect PC-relative loads to the instruction RAM.
12020 @item -mcode-readable=no
12021 Instructions must not access executable sections. This option can be
12022 useful on targets that are configured to have a dual instruction/data
12023 SRAM interface but that (unlike the M4K) do not automatically redirect
12024 PC-relative loads to the instruction RAM.
12027 @item -msplit-addresses
12028 @itemx -mno-split-addresses
12029 @opindex msplit-addresses
12030 @opindex mno-split-addresses
12031 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
12032 relocation operators. This option has been superseded by
12033 @option{-mexplicit-relocs} but is retained for backwards compatibility.
12035 @item -mexplicit-relocs
12036 @itemx -mno-explicit-relocs
12037 @opindex mexplicit-relocs
12038 @opindex mno-explicit-relocs
12039 Use (do not use) assembler relocation operators when dealing with symbolic
12040 addresses. The alternative, selected by @option{-mno-explicit-relocs},
12041 is to use assembler macros instead.
12043 @option{-mexplicit-relocs} is the default if GCC was configured
12044 to use an assembler that supports relocation operators.
12046 @item -mcheck-zero-division
12047 @itemx -mno-check-zero-division
12048 @opindex mcheck-zero-division
12049 @opindex mno-check-zero-division
12050 Trap (do not trap) on integer division by zero.
12052 The default is @option{-mcheck-zero-division}.
12054 @item -mdivide-traps
12055 @itemx -mdivide-breaks
12056 @opindex mdivide-traps
12057 @opindex mdivide-breaks
12058 MIPS systems check for division by zero by generating either a
12059 conditional trap or a break instruction. Using traps results in
12060 smaller code, but is only supported on MIPS II and later. Also, some
12061 versions of the Linux kernel have a bug that prevents trap from
12062 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
12063 allow conditional traps on architectures that support them and
12064 @option{-mdivide-breaks} to force the use of breaks.
12066 The default is usually @option{-mdivide-traps}, but this can be
12067 overridden at configure time using @option{--with-divide=breaks}.
12068 Divide-by-zero checks can be completely disabled using
12069 @option{-mno-check-zero-division}.
12074 @opindex mno-memcpy
12075 Force (do not force) the use of @code{memcpy()} for non-trivial block
12076 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
12077 most constant-sized copies.
12080 @itemx -mno-long-calls
12081 @opindex mlong-calls
12082 @opindex mno-long-calls
12083 Disable (do not disable) use of the @code{jal} instruction. Calling
12084 functions using @code{jal} is more efficient but requires the caller
12085 and callee to be in the same 256 megabyte segment.
12087 This option has no effect on abicalls code. The default is
12088 @option{-mno-long-calls}.
12094 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
12095 instructions, as provided by the R4650 ISA@.
12098 @itemx -mno-fused-madd
12099 @opindex mfused-madd
12100 @opindex mno-fused-madd
12101 Enable (disable) use of the floating point multiply-accumulate
12102 instructions, when they are available. The default is
12103 @option{-mfused-madd}.
12105 When multiply-accumulate instructions are used, the intermediate
12106 product is calculated to infinite precision and is not subject to
12107 the FCSR Flush to Zero bit. This may be undesirable in some
12112 Tell the MIPS assembler to not run its preprocessor over user
12113 assembler files (with a @samp{.s} suffix) when assembling them.
12116 @itemx -mno-fix-r4000
12117 @opindex mfix-r4000
12118 @opindex mno-fix-r4000
12119 Work around certain R4000 CPU errata:
12122 A double-word or a variable shift may give an incorrect result if executed
12123 immediately after starting an integer division.
12125 A double-word or a variable shift may give an incorrect result if executed
12126 while an integer multiplication is in progress.
12128 An integer division may give an incorrect result if started in a delay slot
12129 of a taken branch or a jump.
12133 @itemx -mno-fix-r4400
12134 @opindex mfix-r4400
12135 @opindex mno-fix-r4400
12136 Work around certain R4400 CPU errata:
12139 A double-word or a variable shift may give an incorrect result if executed
12140 immediately after starting an integer division.
12144 @itemx -mno-fix-vr4120
12145 @opindex mfix-vr4120
12146 Work around certain VR4120 errata:
12149 @code{dmultu} does not always produce the correct result.
12151 @code{div} and @code{ddiv} do not always produce the correct result if one
12152 of the operands is negative.
12154 The workarounds for the division errata rely on special functions in
12155 @file{libgcc.a}. At present, these functions are only provided by
12156 the @code{mips64vr*-elf} configurations.
12158 Other VR4120 errata require a nop to be inserted between certain pairs of
12159 instructions. These errata are handled by the assembler, not by GCC itself.
12162 @opindex mfix-vr4130
12163 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
12164 workarounds are implemented by the assembler rather than by GCC,
12165 although GCC will avoid using @code{mflo} and @code{mfhi} if the
12166 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
12167 instructions are available instead.
12170 @itemx -mno-fix-sb1
12172 Work around certain SB-1 CPU core errata.
12173 (This flag currently works around the SB-1 revision 2
12174 ``F1'' and ``F2'' floating point errata.)
12176 @item -mflush-func=@var{func}
12177 @itemx -mno-flush-func
12178 @opindex mflush-func
12179 Specifies the function to call to flush the I and D caches, or to not
12180 call any such function. If called, the function must take the same
12181 arguments as the common @code{_flush_func()}, that is, the address of the
12182 memory range for which the cache is being flushed, the size of the
12183 memory range, and the number 3 (to flush both caches). The default
12184 depends on the target GCC was configured for, but commonly is either
12185 @samp{_flush_func} or @samp{__cpu_flush}.
12187 @item mbranch-cost=@var{num}
12188 @opindex mbranch-cost
12189 Set the cost of branches to roughly @var{num} ``simple'' instructions.
12190 This cost is only a heuristic and is not guaranteed to produce
12191 consistent results across releases. A zero cost redundantly selects
12192 the default, which is based on the @option{-mtune} setting.
12194 @item -mbranch-likely
12195 @itemx -mno-branch-likely
12196 @opindex mbranch-likely
12197 @opindex mno-branch-likely
12198 Enable or disable use of Branch Likely instructions, regardless of the
12199 default for the selected architecture. By default, Branch Likely
12200 instructions may be generated if they are supported by the selected
12201 architecture. An exception is for the MIPS32 and MIPS64 architectures
12202 and processors which implement those architectures; for those, Branch
12203 Likely instructions will not be generated by default because the MIPS32
12204 and MIPS64 architectures specifically deprecate their use.
12206 @item -mfp-exceptions
12207 @itemx -mno-fp-exceptions
12208 @opindex mfp-exceptions
12209 Specifies whether FP exceptions are enabled. This affects how we schedule
12210 FP instructions for some processors. The default is that FP exceptions are
12213 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
12214 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
12217 @item -mvr4130-align
12218 @itemx -mno-vr4130-align
12219 @opindex mvr4130-align
12220 The VR4130 pipeline is two-way superscalar, but can only issue two
12221 instructions together if the first one is 8-byte aligned. When this
12222 option is enabled, GCC will align pairs of instructions that it
12223 thinks should execute in parallel.
12225 This option only has an effect when optimizing for the VR4130.
12226 It normally makes code faster, but at the expense of making it bigger.
12227 It is enabled by default at optimization level @option{-O3}.
12231 @subsection MMIX Options
12232 @cindex MMIX Options
12234 These options are defined for the MMIX:
12238 @itemx -mno-libfuncs
12240 @opindex mno-libfuncs
12241 Specify that intrinsic library functions are being compiled, passing all
12242 values in registers, no matter the size.
12245 @itemx -mno-epsilon
12247 @opindex mno-epsilon
12248 Generate floating-point comparison instructions that compare with respect
12249 to the @code{rE} epsilon register.
12251 @item -mabi=mmixware
12253 @opindex mabi-mmixware
12255 Generate code that passes function parameters and return values that (in
12256 the called function) are seen as registers @code{$0} and up, as opposed to
12257 the GNU ABI which uses global registers @code{$231} and up.
12259 @item -mzero-extend
12260 @itemx -mno-zero-extend
12261 @opindex mzero-extend
12262 @opindex mno-zero-extend
12263 When reading data from memory in sizes shorter than 64 bits, use (do not
12264 use) zero-extending load instructions by default, rather than
12265 sign-extending ones.
12268 @itemx -mno-knuthdiv
12270 @opindex mno-knuthdiv
12271 Make the result of a division yielding a remainder have the same sign as
12272 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
12273 remainder follows the sign of the dividend. Both methods are
12274 arithmetically valid, the latter being almost exclusively used.
12276 @item -mtoplevel-symbols
12277 @itemx -mno-toplevel-symbols
12278 @opindex mtoplevel-symbols
12279 @opindex mno-toplevel-symbols
12280 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
12281 code can be used with the @code{PREFIX} assembly directive.
12285 Generate an executable in the ELF format, rather than the default
12286 @samp{mmo} format used by the @command{mmix} simulator.
12288 @item -mbranch-predict
12289 @itemx -mno-branch-predict
12290 @opindex mbranch-predict
12291 @opindex mno-branch-predict
12292 Use (do not use) the probable-branch instructions, when static branch
12293 prediction indicates a probable branch.
12295 @item -mbase-addresses
12296 @itemx -mno-base-addresses
12297 @opindex mbase-addresses
12298 @opindex mno-base-addresses
12299 Generate (do not generate) code that uses @emph{base addresses}. Using a
12300 base address automatically generates a request (handled by the assembler
12301 and the linker) for a constant to be set up in a global register. The
12302 register is used for one or more base address requests within the range 0
12303 to 255 from the value held in the register. The generally leads to short
12304 and fast code, but the number of different data items that can be
12305 addressed is limited. This means that a program that uses lots of static
12306 data may require @option{-mno-base-addresses}.
12308 @item -msingle-exit
12309 @itemx -mno-single-exit
12310 @opindex msingle-exit
12311 @opindex mno-single-exit
12312 Force (do not force) generated code to have a single exit point in each
12316 @node MN10300 Options
12317 @subsection MN10300 Options
12318 @cindex MN10300 options
12320 These @option{-m} options are defined for Matsushita MN10300 architectures:
12325 Generate code to avoid bugs in the multiply instructions for the MN10300
12326 processors. This is the default.
12328 @item -mno-mult-bug
12329 @opindex mno-mult-bug
12330 Do not generate code to avoid bugs in the multiply instructions for the
12331 MN10300 processors.
12335 Generate code which uses features specific to the AM33 processor.
12339 Do not generate code which uses features specific to the AM33 processor. This
12342 @item -mreturn-pointer-on-d0
12343 @opindex mreturn-pointer-on-d0
12344 When generating a function which returns a pointer, return the pointer
12345 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
12346 only in a0, and attempts to call such functions without a prototype
12347 would result in errors. Note that this option is on by default; use
12348 @option{-mno-return-pointer-on-d0} to disable it.
12352 Do not link in the C run-time initialization object file.
12356 Indicate to the linker that it should perform a relaxation optimization pass
12357 to shorten branches, calls and absolute memory addresses. This option only
12358 has an effect when used on the command line for the final link step.
12360 This option makes symbolic debugging impossible.
12364 @subsection MT Options
12367 These @option{-m} options are defined for Morpho MT architectures:
12371 @item -march=@var{cpu-type}
12373 Generate code that will run on @var{cpu-type}, which is the name of a system
12374 representing a certain processor type. Possible values for
12375 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
12376 @samp{ms1-16-003} and @samp{ms2}.
12378 When this option is not used, the default is @option{-march=ms1-16-002}.
12382 Use byte loads and stores when generating code.
12386 Do not use byte loads and stores when generating code.
12390 Use simulator runtime
12394 Do not link in the C run-time initialization object file
12395 @file{crti.o}. Other run-time initialization and termination files
12396 such as @file{startup.o} and @file{exit.o} are still included on the
12397 linker command line.
12401 @node PDP-11 Options
12402 @subsection PDP-11 Options
12403 @cindex PDP-11 Options
12405 These options are defined for the PDP-11:
12410 Use hardware FPP floating point. This is the default. (FIS floating
12411 point on the PDP-11/40 is not supported.)
12414 @opindex msoft-float
12415 Do not use hardware floating point.
12419 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
12423 Return floating-point results in memory. This is the default.
12427 Generate code for a PDP-11/40.
12431 Generate code for a PDP-11/45. This is the default.
12435 Generate code for a PDP-11/10.
12437 @item -mbcopy-builtin
12438 @opindex bcopy-builtin
12439 Use inline @code{movmemhi} patterns for copying memory. This is the
12444 Do not use inline @code{movmemhi} patterns for copying memory.
12450 Use 16-bit @code{int}. This is the default.
12456 Use 32-bit @code{int}.
12459 @itemx -mno-float32
12461 @opindex mno-float32
12462 Use 64-bit @code{float}. This is the default.
12465 @itemx -mno-float64
12467 @opindex mno-float64
12468 Use 32-bit @code{float}.
12472 Use @code{abshi2} pattern. This is the default.
12476 Do not use @code{abshi2} pattern.
12478 @item -mbranch-expensive
12479 @opindex mbranch-expensive
12480 Pretend that branches are expensive. This is for experimenting with
12481 code generation only.
12483 @item -mbranch-cheap
12484 @opindex mbranch-cheap
12485 Do not pretend that branches are expensive. This is the default.
12489 Generate code for a system with split I&D@.
12493 Generate code for a system without split I&D@. This is the default.
12497 Use Unix assembler syntax. This is the default when configured for
12498 @samp{pdp11-*-bsd}.
12502 Use DEC assembler syntax. This is the default when configured for any
12503 PDP-11 target other than @samp{pdp11-*-bsd}.
12506 @node PowerPC Options
12507 @subsection PowerPC Options
12508 @cindex PowerPC options
12510 These are listed under @xref{RS/6000 and PowerPC Options}.
12512 @node RS/6000 and PowerPC Options
12513 @subsection IBM RS/6000 and PowerPC Options
12514 @cindex RS/6000 and PowerPC Options
12515 @cindex IBM RS/6000 and PowerPC Options
12517 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
12524 @itemx -mno-powerpc
12525 @itemx -mpowerpc-gpopt
12526 @itemx -mno-powerpc-gpopt
12527 @itemx -mpowerpc-gfxopt
12528 @itemx -mno-powerpc-gfxopt
12530 @itemx -mno-powerpc64
12534 @itemx -mno-popcntb
12546 @opindex mno-power2
12548 @opindex mno-powerpc
12549 @opindex mpowerpc-gpopt
12550 @opindex mno-powerpc-gpopt
12551 @opindex mpowerpc-gfxopt
12552 @opindex mno-powerpc-gfxopt
12553 @opindex mpowerpc64
12554 @opindex mno-powerpc64
12558 @opindex mno-popcntb
12564 @opindex mno-mfpgpr
12567 GCC supports two related instruction set architectures for the
12568 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
12569 instructions supported by the @samp{rios} chip set used in the original
12570 RS/6000 systems and the @dfn{PowerPC} instruction set is the
12571 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
12572 the IBM 4xx, 6xx, and follow-on microprocessors.
12574 Neither architecture is a subset of the other. However there is a
12575 large common subset of instructions supported by both. An MQ
12576 register is included in processors supporting the POWER architecture.
12578 You use these options to specify which instructions are available on the
12579 processor you are using. The default value of these options is
12580 determined when configuring GCC@. Specifying the
12581 @option{-mcpu=@var{cpu_type}} overrides the specification of these
12582 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
12583 rather than the options listed above.
12585 The @option{-mpower} option allows GCC to generate instructions that
12586 are found only in the POWER architecture and to use the MQ register.
12587 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
12588 to generate instructions that are present in the POWER2 architecture but
12589 not the original POWER architecture.
12591 The @option{-mpowerpc} option allows GCC to generate instructions that
12592 are found only in the 32-bit subset of the PowerPC architecture.
12593 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
12594 GCC to use the optional PowerPC architecture instructions in the
12595 General Purpose group, including floating-point square root. Specifying
12596 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
12597 use the optional PowerPC architecture instructions in the Graphics
12598 group, including floating-point select.
12600 The @option{-mmfcrf} option allows GCC to generate the move from
12601 condition register field instruction implemented on the POWER4
12602 processor and other processors that support the PowerPC V2.01
12604 The @option{-mpopcntb} option allows GCC to generate the popcount and
12605 double precision FP reciprocal estimate instruction implemented on the
12606 POWER5 processor and other processors that support the PowerPC V2.02
12608 The @option{-mfprnd} option allows GCC to generate the FP round to
12609 integer instructions implemented on the POWER5+ processor and other
12610 processors that support the PowerPC V2.03 architecture.
12611 The @option{-mcmpb} option allows GCC to generate the compare bytes
12612 instruction implemented on the POWER6 processor and other processors
12613 that support the PowerPC V2.05 architecture.
12614 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
12615 general purpose register instructions implemented on the POWER6X
12616 processor and other processors that support the extended PowerPC V2.05
12618 The @option{-mdfp} option allows GCC to generate the decimal floating
12619 point instructions implemented on some POWER processors.
12621 The @option{-mpowerpc64} option allows GCC to generate the additional
12622 64-bit instructions that are found in the full PowerPC64 architecture
12623 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
12624 @option{-mno-powerpc64}.
12626 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
12627 will use only the instructions in the common subset of both
12628 architectures plus some special AIX common-mode calls, and will not use
12629 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
12630 permits GCC to use any instruction from either architecture and to
12631 allow use of the MQ register; specify this for the Motorola MPC601.
12633 @item -mnew-mnemonics
12634 @itemx -mold-mnemonics
12635 @opindex mnew-mnemonics
12636 @opindex mold-mnemonics
12637 Select which mnemonics to use in the generated assembler code. With
12638 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
12639 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
12640 assembler mnemonics defined for the POWER architecture. Instructions
12641 defined in only one architecture have only one mnemonic; GCC uses that
12642 mnemonic irrespective of which of these options is specified.
12644 GCC defaults to the mnemonics appropriate for the architecture in
12645 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
12646 value of these option. Unless you are building a cross-compiler, you
12647 should normally not specify either @option{-mnew-mnemonics} or
12648 @option{-mold-mnemonics}, but should instead accept the default.
12650 @item -mcpu=@var{cpu_type}
12652 Set architecture type, register usage, choice of mnemonics, and
12653 instruction scheduling parameters for machine type @var{cpu_type}.
12654 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
12655 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
12656 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
12657 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
12658 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
12659 @samp{860}, @samp{970}, @samp{8540}, @samp{ec603e}, @samp{G3},
12660 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
12661 @samp{power4}, @samp{power5}, @samp{power5+}, @samp{power6},
12662 @samp{power6x}, @samp{common}, @samp{powerpc}, @samp{powerpc64},
12663 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
12665 @option{-mcpu=common} selects a completely generic processor. Code
12666 generated under this option will run on any POWER or PowerPC processor.
12667 GCC will use only the instructions in the common subset of both
12668 architectures, and will not use the MQ register. GCC assumes a generic
12669 processor model for scheduling purposes.
12671 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
12672 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
12673 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
12674 types, with an appropriate, generic processor model assumed for
12675 scheduling purposes.
12677 The other options specify a specific processor. Code generated under
12678 those options will run best on that processor, and may not run at all on
12681 The @option{-mcpu} options automatically enable or disable the
12684 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
12685 -mnew-mnemonics -mpopcntb -mpower -mpower2 -mpowerpc64 @gol
12686 -mpowerpc-gpopt -mpowerpc-gfxopt -mstring -mmulhw -mdlmzb -mmfpgpr}
12688 The particular options set for any particular CPU will vary between
12689 compiler versions, depending on what setting seems to produce optimal
12690 code for that CPU; it doesn't necessarily reflect the actual hardware's
12691 capabilities. If you wish to set an individual option to a particular
12692 value, you may specify it after the @option{-mcpu} option, like
12693 @samp{-mcpu=970 -mno-altivec}.
12695 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
12696 not enabled or disabled by the @option{-mcpu} option at present because
12697 AIX does not have full support for these options. You may still
12698 enable or disable them individually if you're sure it'll work in your
12701 @item -mtune=@var{cpu_type}
12703 Set the instruction scheduling parameters for machine type
12704 @var{cpu_type}, but do not set the architecture type, register usage, or
12705 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
12706 values for @var{cpu_type} are used for @option{-mtune} as for
12707 @option{-mcpu}. If both are specified, the code generated will use the
12708 architecture, registers, and mnemonics set by @option{-mcpu}, but the
12709 scheduling parameters set by @option{-mtune}.
12715 Generate code to compute division as reciprocal estimate and iterative
12716 refinement, creating opportunities for increased throughput. This
12717 feature requires: optional PowerPC Graphics instruction set for single
12718 precision and FRE instruction for double precision, assuming divides
12719 cannot generate user-visible traps, and the domain values not include
12720 Infinities, denormals or zero denominator.
12723 @itemx -mno-altivec
12725 @opindex mno-altivec
12726 Generate code that uses (does not use) AltiVec instructions, and also
12727 enable the use of built-in functions that allow more direct access to
12728 the AltiVec instruction set. You may also need to set
12729 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
12735 @opindex mno-vrsave
12736 Generate VRSAVE instructions when generating AltiVec code.
12739 @opindex msecure-plt
12740 Generate code that allows ld and ld.so to build executables and shared
12741 libraries with non-exec .plt and .got sections. This is a PowerPC
12742 32-bit SYSV ABI option.
12746 Generate code that uses a BSS .plt section that ld.so fills in, and
12747 requires .plt and .got sections that are both writable and executable.
12748 This is a PowerPC 32-bit SYSV ABI option.
12754 This switch enables or disables the generation of ISEL instructions.
12756 @item -misel=@var{yes/no}
12757 This switch has been deprecated. Use @option{-misel} and
12758 @option{-mno-isel} instead.
12764 This switch enables or disables the generation of SPE simd
12770 @opindex mno-paired
12771 This switch enables or disables the generation of PAIRED simd
12774 @item -mspe=@var{yes/no}
12775 This option has been deprecated. Use @option{-mspe} and
12776 @option{-mno-spe} instead.
12778 @item -mfloat-gprs=@var{yes/single/double/no}
12779 @itemx -mfloat-gprs
12780 @opindex mfloat-gprs
12781 This switch enables or disables the generation of floating point
12782 operations on the general purpose registers for architectures that
12785 The argument @var{yes} or @var{single} enables the use of
12786 single-precision floating point operations.
12788 The argument @var{double} enables the use of single and
12789 double-precision floating point operations.
12791 The argument @var{no} disables floating point operations on the
12792 general purpose registers.
12794 This option is currently only available on the MPC854x.
12800 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
12801 targets (including GNU/Linux). The 32-bit environment sets int, long
12802 and pointer to 32 bits and generates code that runs on any PowerPC
12803 variant. The 64-bit environment sets int to 32 bits and long and
12804 pointer to 64 bits, and generates code for PowerPC64, as for
12805 @option{-mpowerpc64}.
12808 @itemx -mno-fp-in-toc
12809 @itemx -mno-sum-in-toc
12810 @itemx -mminimal-toc
12812 @opindex mno-fp-in-toc
12813 @opindex mno-sum-in-toc
12814 @opindex mminimal-toc
12815 Modify generation of the TOC (Table Of Contents), which is created for
12816 every executable file. The @option{-mfull-toc} option is selected by
12817 default. In that case, GCC will allocate at least one TOC entry for
12818 each unique non-automatic variable reference in your program. GCC
12819 will also place floating-point constants in the TOC@. However, only
12820 16,384 entries are available in the TOC@.
12822 If you receive a linker error message that saying you have overflowed
12823 the available TOC space, you can reduce the amount of TOC space used
12824 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
12825 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
12826 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
12827 generate code to calculate the sum of an address and a constant at
12828 run-time instead of putting that sum into the TOC@. You may specify one
12829 or both of these options. Each causes GCC to produce very slightly
12830 slower and larger code at the expense of conserving TOC space.
12832 If you still run out of space in the TOC even when you specify both of
12833 these options, specify @option{-mminimal-toc} instead. This option causes
12834 GCC to make only one TOC entry for every file. When you specify this
12835 option, GCC will produce code that is slower and larger but which
12836 uses extremely little TOC space. You may wish to use this option
12837 only on files that contain less frequently executed code.
12843 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
12844 @code{long} type, and the infrastructure needed to support them.
12845 Specifying @option{-maix64} implies @option{-mpowerpc64} and
12846 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
12847 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
12850 @itemx -mno-xl-compat
12851 @opindex mxl-compat
12852 @opindex mno-xl-compat
12853 Produce code that conforms more closely to IBM XL compiler semantics
12854 when using AIX-compatible ABI. Pass floating-point arguments to
12855 prototyped functions beyond the register save area (RSA) on the stack
12856 in addition to argument FPRs. Do not assume that most significant
12857 double in 128-bit long double value is properly rounded when comparing
12858 values and converting to double. Use XL symbol names for long double
12861 The AIX calling convention was extended but not initially documented to
12862 handle an obscure K&R C case of calling a function that takes the
12863 address of its arguments with fewer arguments than declared. IBM XL
12864 compilers access floating point arguments which do not fit in the
12865 RSA from the stack when a subroutine is compiled without
12866 optimization. Because always storing floating-point arguments on the
12867 stack is inefficient and rarely needed, this option is not enabled by
12868 default and only is necessary when calling subroutines compiled by IBM
12869 XL compilers without optimization.
12873 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
12874 application written to use message passing with special startup code to
12875 enable the application to run. The system must have PE installed in the
12876 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
12877 must be overridden with the @option{-specs=} option to specify the
12878 appropriate directory location. The Parallel Environment does not
12879 support threads, so the @option{-mpe} option and the @option{-pthread}
12880 option are incompatible.
12882 @item -malign-natural
12883 @itemx -malign-power
12884 @opindex malign-natural
12885 @opindex malign-power
12886 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
12887 @option{-malign-natural} overrides the ABI-defined alignment of larger
12888 types, such as floating-point doubles, on their natural size-based boundary.
12889 The option @option{-malign-power} instructs GCC to follow the ABI-specified
12890 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
12892 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
12896 @itemx -mhard-float
12897 @opindex msoft-float
12898 @opindex mhard-float
12899 Generate code that does not use (uses) the floating-point register set.
12900 Software floating point emulation is provided if you use the
12901 @option{-msoft-float} option, and pass the option to GCC when linking.
12904 @itemx -mno-multiple
12906 @opindex mno-multiple
12907 Generate code that uses (does not use) the load multiple word
12908 instructions and the store multiple word instructions. These
12909 instructions are generated by default on POWER systems, and not
12910 generated on PowerPC systems. Do not use @option{-mmultiple} on little
12911 endian PowerPC systems, since those instructions do not work when the
12912 processor is in little endian mode. The exceptions are PPC740 and
12913 PPC750 which permit the instructions usage in little endian mode.
12918 @opindex mno-string
12919 Generate code that uses (does not use) the load string instructions
12920 and the store string word instructions to save multiple registers and
12921 do small block moves. These instructions are generated by default on
12922 POWER systems, and not generated on PowerPC systems. Do not use
12923 @option{-mstring} on little endian PowerPC systems, since those
12924 instructions do not work when the processor is in little endian mode.
12925 The exceptions are PPC740 and PPC750 which permit the instructions
12926 usage in little endian mode.
12931 @opindex mno-update
12932 Generate code that uses (does not use) the load or store instructions
12933 that update the base register to the address of the calculated memory
12934 location. These instructions are generated by default. If you use
12935 @option{-mno-update}, there is a small window between the time that the
12936 stack pointer is updated and the address of the previous frame is
12937 stored, which means code that walks the stack frame across interrupts or
12938 signals may get corrupted data.
12941 @itemx -mno-fused-madd
12942 @opindex mfused-madd
12943 @opindex mno-fused-madd
12944 Generate code that uses (does not use) the floating point multiply and
12945 accumulate instructions. These instructions are generated by default if
12946 hardware floating is used.
12952 Generate code that uses (does not use) the half-word multiply and
12953 multiply-accumulate instructions on the IBM 405 and 440 processors.
12954 These instructions are generated by default when targetting those
12961 Generate code that uses (does not use) the string-search @samp{dlmzb}
12962 instruction on the IBM 405 and 440 processors. This instruction is
12963 generated by default when targetting those processors.
12965 @item -mno-bit-align
12967 @opindex mno-bit-align
12968 @opindex mbit-align
12969 On System V.4 and embedded PowerPC systems do not (do) force structures
12970 and unions that contain bit-fields to be aligned to the base type of the
12973 For example, by default a structure containing nothing but 8
12974 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
12975 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
12976 the structure would be aligned to a 1 byte boundary and be one byte in
12979 @item -mno-strict-align
12980 @itemx -mstrict-align
12981 @opindex mno-strict-align
12982 @opindex mstrict-align
12983 On System V.4 and embedded PowerPC systems do not (do) assume that
12984 unaligned memory references will be handled by the system.
12986 @item -mrelocatable
12987 @itemx -mno-relocatable
12988 @opindex mrelocatable
12989 @opindex mno-relocatable
12990 On embedded PowerPC systems generate code that allows (does not allow)
12991 the program to be relocated to a different address at runtime. If you
12992 use @option{-mrelocatable} on any module, all objects linked together must
12993 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
12995 @item -mrelocatable-lib
12996 @itemx -mno-relocatable-lib
12997 @opindex mrelocatable-lib
12998 @opindex mno-relocatable-lib
12999 On embedded PowerPC systems generate code that allows (does not allow)
13000 the program to be relocated to a different address at runtime. Modules
13001 compiled with @option{-mrelocatable-lib} can be linked with either modules
13002 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
13003 with modules compiled with the @option{-mrelocatable} options.
13009 On System V.4 and embedded PowerPC systems do not (do) assume that
13010 register 2 contains a pointer to a global area pointing to the addresses
13011 used in the program.
13014 @itemx -mlittle-endian
13016 @opindex mlittle-endian
13017 On System V.4 and embedded PowerPC systems compile code for the
13018 processor in little endian mode. The @option{-mlittle-endian} option is
13019 the same as @option{-mlittle}.
13022 @itemx -mbig-endian
13024 @opindex mbig-endian
13025 On System V.4 and embedded PowerPC systems compile code for the
13026 processor in big endian mode. The @option{-mbig-endian} option is
13027 the same as @option{-mbig}.
13029 @item -mdynamic-no-pic
13030 @opindex mdynamic-no-pic
13031 On Darwin and Mac OS X systems, compile code so that it is not
13032 relocatable, but that its external references are relocatable. The
13033 resulting code is suitable for applications, but not shared
13036 @item -mprioritize-restricted-insns=@var{priority}
13037 @opindex mprioritize-restricted-insns
13038 This option controls the priority that is assigned to
13039 dispatch-slot restricted instructions during the second scheduling
13040 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
13041 @var{no/highest/second-highest} priority to dispatch slot restricted
13044 @item -msched-costly-dep=@var{dependence_type}
13045 @opindex msched-costly-dep
13046 This option controls which dependences are considered costly
13047 by the target during instruction scheduling. The argument
13048 @var{dependence_type} takes one of the following values:
13049 @var{no}: no dependence is costly,
13050 @var{all}: all dependences are costly,
13051 @var{true_store_to_load}: a true dependence from store to load is costly,
13052 @var{store_to_load}: any dependence from store to load is costly,
13053 @var{number}: any dependence which latency >= @var{number} is costly.
13055 @item -minsert-sched-nops=@var{scheme}
13056 @opindex minsert-sched-nops
13057 This option controls which nop insertion scheme will be used during
13058 the second scheduling pass. The argument @var{scheme} takes one of the
13060 @var{no}: Don't insert nops.
13061 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
13062 according to the scheduler's grouping.
13063 @var{regroup_exact}: Insert nops to force costly dependent insns into
13064 separate groups. Insert exactly as many nops as needed to force an insn
13065 to a new group, according to the estimated processor grouping.
13066 @var{number}: Insert nops to force costly dependent insns into
13067 separate groups. Insert @var{number} nops to force an insn to a new group.
13070 @opindex mcall-sysv
13071 On System V.4 and embedded PowerPC systems compile code using calling
13072 conventions that adheres to the March 1995 draft of the System V
13073 Application Binary Interface, PowerPC processor supplement. This is the
13074 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
13076 @item -mcall-sysv-eabi
13077 @opindex mcall-sysv-eabi
13078 Specify both @option{-mcall-sysv} and @option{-meabi} options.
13080 @item -mcall-sysv-noeabi
13081 @opindex mcall-sysv-noeabi
13082 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
13084 @item -mcall-solaris
13085 @opindex mcall-solaris
13086 On System V.4 and embedded PowerPC systems compile code for the Solaris
13090 @opindex mcall-linux
13091 On System V.4 and embedded PowerPC systems compile code for the
13092 Linux-based GNU system.
13096 On System V.4 and embedded PowerPC systems compile code for the
13097 Hurd-based GNU system.
13099 @item -mcall-netbsd
13100 @opindex mcall-netbsd
13101 On System V.4 and embedded PowerPC systems compile code for the
13102 NetBSD operating system.
13104 @item -maix-struct-return
13105 @opindex maix-struct-return
13106 Return all structures in memory (as specified by the AIX ABI)@.
13108 @item -msvr4-struct-return
13109 @opindex msvr4-struct-return
13110 Return structures smaller than 8 bytes in registers (as specified by the
13113 @item -mabi=@var{abi-type}
13115 Extend the current ABI with a particular extension, or remove such extension.
13116 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
13117 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
13121 Extend the current ABI with SPE ABI extensions. This does not change
13122 the default ABI, instead it adds the SPE ABI extensions to the current
13126 @opindex mabi=no-spe
13127 Disable Booke SPE ABI extensions for the current ABI@.
13129 @item -mabi=ibmlongdouble
13130 @opindex mabi=ibmlongdouble
13131 Change the current ABI to use IBM extended precision long double.
13132 This is a PowerPC 32-bit SYSV ABI option.
13134 @item -mabi=ieeelongdouble
13135 @opindex mabi=ieeelongdouble
13136 Change the current ABI to use IEEE extended precision long double.
13137 This is a PowerPC 32-bit Linux ABI option.
13140 @itemx -mno-prototype
13141 @opindex mprototype
13142 @opindex mno-prototype
13143 On System V.4 and embedded PowerPC systems assume that all calls to
13144 variable argument functions are properly prototyped. Otherwise, the
13145 compiler must insert an instruction before every non prototyped call to
13146 set or clear bit 6 of the condition code register (@var{CR}) to
13147 indicate whether floating point values were passed in the floating point
13148 registers in case the function takes a variable arguments. With
13149 @option{-mprototype}, only calls to prototyped variable argument functions
13150 will set or clear the bit.
13154 On embedded PowerPC systems, assume that the startup module is called
13155 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
13156 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
13161 On embedded PowerPC systems, assume that the startup module is called
13162 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
13167 On embedded PowerPC systems, assume that the startup module is called
13168 @file{crt0.o} and the standard C libraries are @file{libads.a} and
13171 @item -myellowknife
13172 @opindex myellowknife
13173 On embedded PowerPC systems, assume that the startup module is called
13174 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
13179 On System V.4 and embedded PowerPC systems, specify that you are
13180 compiling for a VxWorks system.
13184 Specify that you are compiling for the WindISS simulation environment.
13188 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
13189 header to indicate that @samp{eabi} extended relocations are used.
13195 On System V.4 and embedded PowerPC systems do (do not) adhere to the
13196 Embedded Applications Binary Interface (eabi) which is a set of
13197 modifications to the System V.4 specifications. Selecting @option{-meabi}
13198 means that the stack is aligned to an 8 byte boundary, a function
13199 @code{__eabi} is called to from @code{main} to set up the eabi
13200 environment, and the @option{-msdata} option can use both @code{r2} and
13201 @code{r13} to point to two separate small data areas. Selecting
13202 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
13203 do not call an initialization function from @code{main}, and the
13204 @option{-msdata} option will only use @code{r13} to point to a single
13205 small data area. The @option{-meabi} option is on by default if you
13206 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
13209 @opindex msdata=eabi
13210 On System V.4 and embedded PowerPC systems, put small initialized
13211 @code{const} global and static data in the @samp{.sdata2} section, which
13212 is pointed to by register @code{r2}. Put small initialized
13213 non-@code{const} global and static data in the @samp{.sdata} section,
13214 which is pointed to by register @code{r13}. Put small uninitialized
13215 global and static data in the @samp{.sbss} section, which is adjacent to
13216 the @samp{.sdata} section. The @option{-msdata=eabi} option is
13217 incompatible with the @option{-mrelocatable} option. The
13218 @option{-msdata=eabi} option also sets the @option{-memb} option.
13221 @opindex msdata=sysv
13222 On System V.4 and embedded PowerPC systems, put small global and static
13223 data in the @samp{.sdata} section, which is pointed to by register
13224 @code{r13}. Put small uninitialized global and static data in the
13225 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
13226 The @option{-msdata=sysv} option is incompatible with the
13227 @option{-mrelocatable} option.
13229 @item -msdata=default
13231 @opindex msdata=default
13233 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
13234 compile code the same as @option{-msdata=eabi}, otherwise compile code the
13235 same as @option{-msdata=sysv}.
13238 @opindex msdata-data
13239 On System V.4 and embedded PowerPC systems, put small global
13240 data in the @samp{.sdata} section. Put small uninitialized global
13241 data in the @samp{.sbss} section. Do not use register @code{r13}
13242 to address small data however. This is the default behavior unless
13243 other @option{-msdata} options are used.
13247 @opindex msdata=none
13249 On embedded PowerPC systems, put all initialized global and static data
13250 in the @samp{.data} section, and all uninitialized data in the
13251 @samp{.bss} section.
13255 @cindex smaller data references (PowerPC)
13256 @cindex .sdata/.sdata2 references (PowerPC)
13257 On embedded PowerPC systems, put global and static items less than or
13258 equal to @var{num} bytes into the small data or bss sections instead of
13259 the normal data or bss section. By default, @var{num} is 8. The
13260 @option{-G @var{num}} switch is also passed to the linker.
13261 All modules should be compiled with the same @option{-G @var{num}} value.
13264 @itemx -mno-regnames
13266 @opindex mno-regnames
13267 On System V.4 and embedded PowerPC systems do (do not) emit register
13268 names in the assembly language output using symbolic forms.
13271 @itemx -mno-longcall
13273 @opindex mno-longcall
13274 By default assume that all calls are far away so that a longer more
13275 expensive calling sequence is required. This is required for calls
13276 further than 32 megabytes (33,554,432 bytes) from the current location.
13277 A short call will be generated if the compiler knows
13278 the call cannot be that far away. This setting can be overridden by
13279 the @code{shortcall} function attribute, or by @code{#pragma
13282 Some linkers are capable of detecting out-of-range calls and generating
13283 glue code on the fly. On these systems, long calls are unnecessary and
13284 generate slower code. As of this writing, the AIX linker can do this,
13285 as can the GNU linker for PowerPC/64. It is planned to add this feature
13286 to the GNU linker for 32-bit PowerPC systems as well.
13288 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
13289 callee, L42'', plus a ``branch island'' (glue code). The two target
13290 addresses represent the callee and the ``branch island''. The
13291 Darwin/PPC linker will prefer the first address and generate a ``bl
13292 callee'' if the PPC ``bl'' instruction will reach the callee directly;
13293 otherwise, the linker will generate ``bl L42'' to call the ``branch
13294 island''. The ``branch island'' is appended to the body of the
13295 calling function; it computes the full 32-bit address of the callee
13298 On Mach-O (Darwin) systems, this option directs the compiler emit to
13299 the glue for every direct call, and the Darwin linker decides whether
13300 to use or discard it.
13302 In the future, we may cause GCC to ignore all longcall specifications
13303 when the linker is known to generate glue.
13307 Adds support for multithreading with the @dfn{pthreads} library.
13308 This option sets flags for both the preprocessor and linker.
13312 @node S/390 and zSeries Options
13313 @subsection S/390 and zSeries Options
13314 @cindex S/390 and zSeries Options
13316 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
13320 @itemx -msoft-float
13321 @opindex mhard-float
13322 @opindex msoft-float
13323 Use (do not use) the hardware floating-point instructions and registers
13324 for floating-point operations. When @option{-msoft-float} is specified,
13325 functions in @file{libgcc.a} will be used to perform floating-point
13326 operations. When @option{-mhard-float} is specified, the compiler
13327 generates IEEE floating-point instructions. This is the default.
13329 @item -mlong-double-64
13330 @itemx -mlong-double-128
13331 @opindex mlong-double-64
13332 @opindex mlong-double-128
13333 These switches control the size of @code{long double} type. A size
13334 of 64bit makes the @code{long double} type equivalent to the @code{double}
13335 type. This is the default.
13338 @itemx -mno-backchain
13339 @opindex mbackchain
13340 @opindex mno-backchain
13341 Store (do not store) the address of the caller's frame as backchain pointer
13342 into the callee's stack frame.
13343 A backchain may be needed to allow debugging using tools that do not understand
13344 DWARF-2 call frame information.
13345 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
13346 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
13347 the backchain is placed into the topmost word of the 96/160 byte register
13350 In general, code compiled with @option{-mbackchain} is call-compatible with
13351 code compiled with @option{-mmo-backchain}; however, use of the backchain
13352 for debugging purposes usually requires that the whole binary is built with
13353 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
13354 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13355 to build a linux kernel use @option{-msoft-float}.
13357 The default is to not maintain the backchain.
13359 @item -mpacked-stack
13360 @item -mno-packed-stack
13361 @opindex mpacked-stack
13362 @opindex mno-packed-stack
13363 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
13364 specified, the compiler uses the all fields of the 96/160 byte register save
13365 area only for their default purpose; unused fields still take up stack space.
13366 When @option{-mpacked-stack} is specified, register save slots are densely
13367 packed at the top of the register save area; unused space is reused for other
13368 purposes, allowing for more efficient use of the available stack space.
13369 However, when @option{-mbackchain} is also in effect, the topmost word of
13370 the save area is always used to store the backchain, and the return address
13371 register is always saved two words below the backchain.
13373 As long as the stack frame backchain is not used, code generated with
13374 @option{-mpacked-stack} is call-compatible with code generated with
13375 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
13376 S/390 or zSeries generated code that uses the stack frame backchain at run
13377 time, not just for debugging purposes. Such code is not call-compatible
13378 with code compiled with @option{-mpacked-stack}. Also, note that the
13379 combination of @option{-mbackchain},
13380 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13381 to build a linux kernel use @option{-msoft-float}.
13383 The default is to not use the packed stack layout.
13386 @itemx -mno-small-exec
13387 @opindex msmall-exec
13388 @opindex mno-small-exec
13389 Generate (or do not generate) code using the @code{bras} instruction
13390 to do subroutine calls.
13391 This only works reliably if the total executable size does not
13392 exceed 64k. The default is to use the @code{basr} instruction instead,
13393 which does not have this limitation.
13399 When @option{-m31} is specified, generate code compliant to the
13400 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
13401 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
13402 particular to generate 64-bit instructions. For the @samp{s390}
13403 targets, the default is @option{-m31}, while the @samp{s390x}
13404 targets default to @option{-m64}.
13410 When @option{-mzarch} is specified, generate code using the
13411 instructions available on z/Architecture.
13412 When @option{-mesa} is specified, generate code using the
13413 instructions available on ESA/390. Note that @option{-mesa} is
13414 not possible with @option{-m64}.
13415 When generating code compliant to the GNU/Linux for S/390 ABI,
13416 the default is @option{-mesa}. When generating code compliant
13417 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
13423 Generate (or do not generate) code using the @code{mvcle} instruction
13424 to perform block moves. When @option{-mno-mvcle} is specified,
13425 use a @code{mvc} loop instead. This is the default unless optimizing for
13432 Print (or do not print) additional debug information when compiling.
13433 The default is to not print debug information.
13435 @item -march=@var{cpu-type}
13437 Generate code that will run on @var{cpu-type}, which is the name of a system
13438 representing a certain processor type. Possible values for
13439 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
13440 When generating code using the instructions available on z/Architecture,
13441 the default is @option{-march=z900}. Otherwise, the default is
13442 @option{-march=g5}.
13444 @item -mtune=@var{cpu-type}
13446 Tune to @var{cpu-type} everything applicable about the generated code,
13447 except for the ABI and the set of available instructions.
13448 The list of @var{cpu-type} values is the same as for @option{-march}.
13449 The default is the value used for @option{-march}.
13452 @itemx -mno-tpf-trace
13453 @opindex mtpf-trace
13454 @opindex mno-tpf-trace
13455 Generate code that adds (does not add) in TPF OS specific branches to trace
13456 routines in the operating system. This option is off by default, even
13457 when compiling for the TPF OS@.
13460 @itemx -mno-fused-madd
13461 @opindex mfused-madd
13462 @opindex mno-fused-madd
13463 Generate code that uses (does not use) the floating point multiply and
13464 accumulate instructions. These instructions are generated by default if
13465 hardware floating point is used.
13467 @item -mwarn-framesize=@var{framesize}
13468 @opindex mwarn-framesize
13469 Emit a warning if the current function exceeds the given frame size. Because
13470 this is a compile time check it doesn't need to be a real problem when the program
13471 runs. It is intended to identify functions which most probably cause
13472 a stack overflow. It is useful to be used in an environment with limited stack
13473 size e.g.@: the linux kernel.
13475 @item -mwarn-dynamicstack
13476 @opindex mwarn-dynamicstack
13477 Emit a warning if the function calls alloca or uses dynamically
13478 sized arrays. This is generally a bad idea with a limited stack size.
13480 @item -mstack-guard=@var{stack-guard}
13481 @item -mstack-size=@var{stack-size}
13482 @opindex mstack-guard
13483 @opindex mstack-size
13484 If these options are provided the s390 back end emits additional instructions in
13485 the function prologue which trigger a trap if the stack size is @var{stack-guard}
13486 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
13487 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
13488 the frame size of the compiled function is chosen.
13489 These options are intended to be used to help debugging stack overflow problems.
13490 The additionally emitted code causes only little overhead and hence can also be
13491 used in production like systems without greater performance degradation. The given
13492 values have to be exact powers of 2 and @var{stack-size} has to be greater than
13493 @var{stack-guard} without exceeding 64k.
13494 In order to be efficient the extra code makes the assumption that the stack starts
13495 at an address aligned to the value given by @var{stack-size}.
13496 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
13499 @node Score Options
13500 @subsection Score Options
13501 @cindex Score Options
13503 These options are defined for Score implementations:
13508 Compile code for big endian mode. This is the default.
13512 Compile code for little endian mode.
13516 Disable generate bcnz instruction.
13520 Enable generate unaligned load and store instruction.
13524 Enable the use of multiply-accumulate instructions. Disabled by default.
13528 Specify the SCORE5 as the target architecture.
13532 Specify the SCORE5U of the target architecture.
13536 Specify the SCORE7 as the target architecture. This is the default.
13540 Specify the SCORE7D as the target architecture.
13544 @subsection SH Options
13546 These @samp{-m} options are defined for the SH implementations:
13551 Generate code for the SH1.
13555 Generate code for the SH2.
13558 Generate code for the SH2e.
13562 Generate code for the SH3.
13566 Generate code for the SH3e.
13570 Generate code for the SH4 without a floating-point unit.
13572 @item -m4-single-only
13573 @opindex m4-single-only
13574 Generate code for the SH4 with a floating-point unit that only
13575 supports single-precision arithmetic.
13579 Generate code for the SH4 assuming the floating-point unit is in
13580 single-precision mode by default.
13584 Generate code for the SH4.
13588 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
13589 floating-point unit is not used.
13591 @item -m4a-single-only
13592 @opindex m4a-single-only
13593 Generate code for the SH4a, in such a way that no double-precision
13594 floating point operations are used.
13597 @opindex m4a-single
13598 Generate code for the SH4a assuming the floating-point unit is in
13599 single-precision mode by default.
13603 Generate code for the SH4a.
13607 Same as @option{-m4a-nofpu}, except that it implicitly passes
13608 @option{-dsp} to the assembler. GCC doesn't generate any DSP
13609 instructions at the moment.
13613 Compile code for the processor in big endian mode.
13617 Compile code for the processor in little endian mode.
13621 Align doubles at 64-bit boundaries. Note that this changes the calling
13622 conventions, and thus some functions from the standard C library will
13623 not work unless you recompile it first with @option{-mdalign}.
13627 Shorten some address references at link time, when possible; uses the
13628 linker option @option{-relax}.
13632 Use 32-bit offsets in @code{switch} tables. The default is to use
13637 Enable the use of the instruction @code{fmovd}.
13641 Comply with the calling conventions defined by Renesas.
13645 Comply with the calling conventions defined by Renesas.
13649 Comply with the calling conventions defined for GCC before the Renesas
13650 conventions were available. This option is the default for all
13651 targets of the SH toolchain except for @samp{sh-symbianelf}.
13654 @opindex mnomacsave
13655 Mark the @code{MAC} register as call-clobbered, even if
13656 @option{-mhitachi} is given.
13660 Increase IEEE-compliance of floating-point code.
13661 At the moment, this is equivalent to @option{-fno-finite-math-only}.
13662 When generating 16 bit SH opcodes, getting IEEE-conforming results for
13663 comparisons of NANs / infinities incurs extra overhead in every
13664 floating point comparison, therefore the default is set to
13665 @option{-ffinite-math-only}.
13667 @item -minline-ic_invalidate
13668 @opindex minline-ic_invalidate
13669 Inline code to invalidate instruction cache entries after setting up
13670 nested function trampolines.
13671 This option has no effect if -musermode is in effect and the selected
13672 code generation option (e.g. -m4) does not allow the use of the icbi
13674 If the selected code generation option does not allow the use of the icbi
13675 instruction, and -musermode is not in effect, the inlined code will
13676 manipulate the instruction cache address array directly with an associative
13677 write. This not only requires privileged mode, but it will also
13678 fail if the cache line had been mapped via the TLB and has become unmapped.
13682 Dump instruction size and location in the assembly code.
13685 @opindex mpadstruct
13686 This option is deprecated. It pads structures to multiple of 4 bytes,
13687 which is incompatible with the SH ABI@.
13691 Optimize for space instead of speed. Implied by @option{-Os}.
13694 @opindex mprefergot
13695 When generating position-independent code, emit function calls using
13696 the Global Offset Table instead of the Procedure Linkage Table.
13700 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
13701 if the inlined code would not work in user mode.
13702 This is the default when the target is @code{sh-*-linux*}.
13704 @item -multcost=@var{number}
13705 @opindex multcost=@var{number}
13706 Set the cost to assume for a multiply insn.
13708 @item -mdiv=@var{strategy}
13709 @opindex mdiv=@var{strategy}
13710 Set the division strategy to use for SHmedia code. @var{strategy} must be
13711 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
13712 inv:call2, inv:fp .
13713 "fp" performs the operation in floating point. This has a very high latency,
13714 but needs only a few instructions, so it might be a good choice if
13715 your code has enough easily exploitable ILP to allow the compiler to
13716 schedule the floating point instructions together with other instructions.
13717 Division by zero causes a floating point exception.
13718 "inv" uses integer operations to calculate the inverse of the divisor,
13719 and then multiplies the dividend with the inverse. This strategy allows
13720 cse and hoisting of the inverse calculation. Division by zero calculates
13721 an unspecified result, but does not trap.
13722 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
13723 have been found, or if the entire operation has been hoisted to the same
13724 place, the last stages of the inverse calculation are intertwined with the
13725 final multiply to reduce the overall latency, at the expense of using a few
13726 more instructions, and thus offering fewer scheduling opportunities with
13728 "call" calls a library function that usually implements the inv:minlat
13730 This gives high code density for m5-*media-nofpu compilations.
13731 "call2" uses a different entry point of the same library function, where it
13732 assumes that a pointer to a lookup table has already been set up, which
13733 exposes the pointer load to cse / code hoisting optimizations.
13734 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
13735 code generation, but if the code stays unoptimized, revert to the "call",
13736 "call2", or "fp" strategies, respectively. Note that the
13737 potentially-trapping side effect of division by zero is carried by a
13738 separate instruction, so it is possible that all the integer instructions
13739 are hoisted out, but the marker for the side effect stays where it is.
13740 A recombination to fp operations or a call is not possible in that case.
13741 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
13742 that the inverse calculation was nor separated from the multiply, they speed
13743 up division where the dividend fits into 20 bits (plus sign where applicable),
13744 by inserting a test to skip a number of operations in this case; this test
13745 slows down the case of larger dividends. inv20u assumes the case of a such
13746 a small dividend to be unlikely, and inv20l assumes it to be likely.
13748 @item -mdivsi3_libfunc=@var{name}
13749 @opindex mdivsi3_libfunc=@var{name}
13750 Set the name of the library function used for 32 bit signed division to
13751 @var{name}. This only affect the name used in the call and inv:call
13752 division strategies, and the compiler will still expect the same
13753 sets of input/output/clobbered registers as if this option was not present.
13755 @item -madjust-unroll
13756 @opindex madjust-unroll
13757 Throttle unrolling to avoid thrashing target registers.
13758 This option only has an effect if the gcc code base supports the
13759 TARGET_ADJUST_UNROLL_MAX target hook.
13761 @item -mindexed-addressing
13762 @opindex mindexed-addressing
13763 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
13764 This is only safe if the hardware and/or OS implement 32 bit wrap-around
13765 semantics for the indexed addressing mode. The architecture allows the
13766 implementation of processors with 64 bit MMU, which the OS could use to
13767 get 32 bit addressing, but since no current hardware implementation supports
13768 this or any other way to make the indexed addressing mode safe to use in
13769 the 32 bit ABI, the default is -mno-indexed-addressing.
13771 @item -mgettrcost=@var{number}
13772 @opindex mgettrcost=@var{number}
13773 Set the cost assumed for the gettr instruction to @var{number}.
13774 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
13778 Assume pt* instructions won't trap. This will generally generate better
13779 scheduled code, but is unsafe on current hardware. The current architecture
13780 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
13781 This has the unintentional effect of making it unsafe to schedule ptabs /
13782 ptrel before a branch, or hoist it out of a loop. For example,
13783 __do_global_ctors, a part of libgcc that runs constructors at program
13784 startup, calls functions in a list which is delimited by @minus{}1. With the
13785 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
13786 That means that all the constructors will be run a bit quicker, but when
13787 the loop comes to the end of the list, the program crashes because ptabs
13788 loads @minus{}1 into a target register. Since this option is unsafe for any
13789 hardware implementing the current architecture specification, the default
13790 is -mno-pt-fixed. Unless the user specifies a specific cost with
13791 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
13792 this deters register allocation using target registers for storing
13795 @item -minvalid-symbols
13796 @opindex minvalid-symbols
13797 Assume symbols might be invalid. Ordinary function symbols generated by
13798 the compiler will always be valid to load with movi/shori/ptabs or
13799 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
13800 to generate symbols that will cause ptabs / ptrel to trap.
13801 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
13802 It will then prevent cross-basic-block cse, hoisting and most scheduling
13803 of symbol loads. The default is @option{-mno-invalid-symbols}.
13806 @node SPARC Options
13807 @subsection SPARC Options
13808 @cindex SPARC options
13810 These @samp{-m} options are supported on the SPARC:
13813 @item -mno-app-regs
13815 @opindex mno-app-regs
13817 Specify @option{-mapp-regs} to generate output using the global registers
13818 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
13821 To be fully SVR4 ABI compliant at the cost of some performance loss,
13822 specify @option{-mno-app-regs}. You should compile libraries and system
13823 software with this option.
13826 @itemx -mhard-float
13828 @opindex mhard-float
13829 Generate output containing floating point instructions. This is the
13833 @itemx -msoft-float
13835 @opindex msoft-float
13836 Generate output containing library calls for floating point.
13837 @strong{Warning:} the requisite libraries are not available for all SPARC
13838 targets. Normally the facilities of the machine's usual C compiler are
13839 used, but this cannot be done directly in cross-compilation. You must make
13840 your own arrangements to provide suitable library functions for
13841 cross-compilation. The embedded targets @samp{sparc-*-aout} and
13842 @samp{sparclite-*-*} do provide software floating point support.
13844 @option{-msoft-float} changes the calling convention in the output file;
13845 therefore, it is only useful if you compile @emph{all} of a program with
13846 this option. In particular, you need to compile @file{libgcc.a}, the
13847 library that comes with GCC, with @option{-msoft-float} in order for
13850 @item -mhard-quad-float
13851 @opindex mhard-quad-float
13852 Generate output containing quad-word (long double) floating point
13855 @item -msoft-quad-float
13856 @opindex msoft-quad-float
13857 Generate output containing library calls for quad-word (long double)
13858 floating point instructions. The functions called are those specified
13859 in the SPARC ABI@. This is the default.
13861 As of this writing, there are no SPARC implementations that have hardware
13862 support for the quad-word floating point instructions. They all invoke
13863 a trap handler for one of these instructions, and then the trap handler
13864 emulates the effect of the instruction. Because of the trap handler overhead,
13865 this is much slower than calling the ABI library routines. Thus the
13866 @option{-msoft-quad-float} option is the default.
13868 @item -mno-unaligned-doubles
13869 @itemx -munaligned-doubles
13870 @opindex mno-unaligned-doubles
13871 @opindex munaligned-doubles
13872 Assume that doubles have 8 byte alignment. This is the default.
13874 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
13875 alignment only if they are contained in another type, or if they have an
13876 absolute address. Otherwise, it assumes they have 4 byte alignment.
13877 Specifying this option avoids some rare compatibility problems with code
13878 generated by other compilers. It is not the default because it results
13879 in a performance loss, especially for floating point code.
13881 @item -mno-faster-structs
13882 @itemx -mfaster-structs
13883 @opindex mno-faster-structs
13884 @opindex mfaster-structs
13885 With @option{-mfaster-structs}, the compiler assumes that structures
13886 should have 8 byte alignment. This enables the use of pairs of
13887 @code{ldd} and @code{std} instructions for copies in structure
13888 assignment, in place of twice as many @code{ld} and @code{st} pairs.
13889 However, the use of this changed alignment directly violates the SPARC
13890 ABI@. Thus, it's intended only for use on targets where the developer
13891 acknowledges that their resulting code will not be directly in line with
13892 the rules of the ABI@.
13894 @item -mimpure-text
13895 @opindex mimpure-text
13896 @option{-mimpure-text}, used in addition to @option{-shared}, tells
13897 the compiler to not pass @option{-z text} to the linker when linking a
13898 shared object. Using this option, you can link position-dependent
13899 code into a shared object.
13901 @option{-mimpure-text} suppresses the ``relocations remain against
13902 allocatable but non-writable sections'' linker error message.
13903 However, the necessary relocations will trigger copy-on-write, and the
13904 shared object is not actually shared across processes. Instead of
13905 using @option{-mimpure-text}, you should compile all source code with
13906 @option{-fpic} or @option{-fPIC}.
13908 This option is only available on SunOS and Solaris.
13910 @item -mcpu=@var{cpu_type}
13912 Set the instruction set, register set, and instruction scheduling parameters
13913 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
13914 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
13915 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
13916 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
13917 @samp{ultrasparc3}, and @samp{niagara}.
13919 Default instruction scheduling parameters are used for values that select
13920 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
13921 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
13923 Here is a list of each supported architecture and their supported
13928 v8: supersparc, hypersparc
13929 sparclite: f930, f934, sparclite86x
13931 v9: ultrasparc, ultrasparc3, niagara
13934 By default (unless configured otherwise), GCC generates code for the V7
13935 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
13936 additionally optimizes it for the Cypress CY7C602 chip, as used in the
13937 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
13938 SPARCStation 1, 2, IPX etc.
13940 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
13941 architecture. The only difference from V7 code is that the compiler emits
13942 the integer multiply and integer divide instructions which exist in SPARC-V8
13943 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
13944 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
13947 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
13948 the SPARC architecture. This adds the integer multiply, integer divide step
13949 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
13950 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
13951 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
13952 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
13953 MB86934 chip, which is the more recent SPARClite with FPU@.
13955 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
13956 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
13957 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
13958 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
13959 optimizes it for the TEMIC SPARClet chip.
13961 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
13962 architecture. This adds 64-bit integer and floating-point move instructions,
13963 3 additional floating-point condition code registers and conditional move
13964 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
13965 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
13966 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
13967 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
13968 @option{-mcpu=niagara}, the compiler additionally optimizes it for
13969 Sun UltraSPARC T1 chips.
13971 @item -mtune=@var{cpu_type}
13973 Set the instruction scheduling parameters for machine type
13974 @var{cpu_type}, but do not set the instruction set or register set that the
13975 option @option{-mcpu=@var{cpu_type}} would.
13977 The same values for @option{-mcpu=@var{cpu_type}} can be used for
13978 @option{-mtune=@var{cpu_type}}, but the only useful values are those
13979 that select a particular cpu implementation. Those are @samp{cypress},
13980 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
13981 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
13982 @samp{ultrasparc3}, and @samp{niagara}.
13987 @opindex mno-v8plus
13988 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
13989 difference from the V8 ABI is that the global and out registers are
13990 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
13991 mode for all SPARC-V9 processors.
13997 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
13998 Visual Instruction Set extensions. The default is @option{-mno-vis}.
14001 These @samp{-m} options are supported in addition to the above
14002 on SPARC-V9 processors in 64-bit environments:
14005 @item -mlittle-endian
14006 @opindex mlittle-endian
14007 Generate code for a processor running in little-endian mode. It is only
14008 available for a few configurations and most notably not on Solaris and Linux.
14014 Generate code for a 32-bit or 64-bit environment.
14015 The 32-bit environment sets int, long and pointer to 32 bits.
14016 The 64-bit environment sets int to 32 bits and long and pointer
14019 @item -mcmodel=medlow
14020 @opindex mcmodel=medlow
14021 Generate code for the Medium/Low code model: 64-bit addresses, programs
14022 must be linked in the low 32 bits of memory. Programs can be statically
14023 or dynamically linked.
14025 @item -mcmodel=medmid
14026 @opindex mcmodel=medmid
14027 Generate code for the Medium/Middle code model: 64-bit addresses, programs
14028 must be linked in the low 44 bits of memory, the text and data segments must
14029 be less than 2GB in size and the data segment must be located within 2GB of
14032 @item -mcmodel=medany
14033 @opindex mcmodel=medany
14034 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
14035 may be linked anywhere in memory, the text and data segments must be less
14036 than 2GB in size and the data segment must be located within 2GB of the
14039 @item -mcmodel=embmedany
14040 @opindex mcmodel=embmedany
14041 Generate code for the Medium/Anywhere code model for embedded systems:
14042 64-bit addresses, the text and data segments must be less than 2GB in
14043 size, both starting anywhere in memory (determined at link time). The
14044 global register %g4 points to the base of the data segment. Programs
14045 are statically linked and PIC is not supported.
14048 @itemx -mno-stack-bias
14049 @opindex mstack-bias
14050 @opindex mno-stack-bias
14051 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
14052 frame pointer if present, are offset by @minus{}2047 which must be added back
14053 when making stack frame references. This is the default in 64-bit mode.
14054 Otherwise, assume no such offset is present.
14057 These switches are supported in addition to the above on Solaris:
14062 Add support for multithreading using the Solaris threads library. This
14063 option sets flags for both the preprocessor and linker. This option does
14064 not affect the thread safety of object code produced by the compiler or
14065 that of libraries supplied with it.
14069 Add support for multithreading using the POSIX threads library. This
14070 option sets flags for both the preprocessor and linker. This option does
14071 not affect the thread safety of object code produced by the compiler or
14072 that of libraries supplied with it.
14076 This is a synonym for @option{-pthreads}.
14080 @subsection SPU Options
14081 @cindex SPU options
14083 These @samp{-m} options are supported on the SPU:
14087 @itemx -merror-reloc
14088 @opindex mwarn-reloc
14089 @opindex merror-reloc
14091 The loader for SPU does not handle dynamic relocations. By default, GCC
14092 will give an error when it generates code that requires a dynamic
14093 relocation. @option{-mno-error-reloc} disables the error,
14094 @option{-mwarn-reloc} will generate a warning instead.
14097 @itemx -munsafe-dma
14099 @opindex munsafe-dma
14101 Instructions which initiate or test completion of DMA must not be
14102 reordered with respect to loads and stores of the memory which is being
14103 accessed. Users typically address this problem using the volatile
14104 keyword, but that can lead to inefficient code in places where the
14105 memory is known to not change. Rather than mark the memory as volatile
14106 we treat the DMA instructions as potentially effecting all memory. With
14107 @option{-munsafe-dma} users must use the volatile keyword to protect
14110 @item -mbranch-hints
14111 @opindex mbranch-hints
14113 By default, GCC will generate a branch hint instruction to avoid
14114 pipeline stalls for always taken or probably taken branches. A hint
14115 will not be generated closer than 8 instructions away from its branch.
14116 There is little reason to disable them, except for debugging purposes,
14117 or to make an object a little bit smaller.
14121 @opindex msmall-mem
14122 @opindex mlarge-mem
14124 By default, GCC generates code assuming that addresses are never larger
14125 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
14126 a full 32 bit address.
14131 By default, GCC links against startup code that assumes the SPU-style
14132 main function interface (which has an unconventional parameter list).
14133 With @option{-mstdmain}, GCC will link your program against startup
14134 code that assumes a C99-style interface to @code{main}, including a
14135 local copy of @code{argv} strings.
14137 @item -mfixed-range=@var{register-range}
14138 @opindex mfixed-range
14139 Generate code treating the given register range as fixed registers.
14140 A fixed register is one that the register allocator can not use. This is
14141 useful when compiling kernel code. A register range is specified as
14142 two registers separated by a dash. Multiple register ranges can be
14143 specified separated by a comma.
14147 @node System V Options
14148 @subsection Options for System V
14150 These additional options are available on System V Release 4 for
14151 compatibility with other compilers on those systems:
14156 Create a shared object.
14157 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
14161 Identify the versions of each tool used by the compiler, in a
14162 @code{.ident} assembler directive in the output.
14166 Refrain from adding @code{.ident} directives to the output file (this is
14169 @item -YP,@var{dirs}
14171 Search the directories @var{dirs}, and no others, for libraries
14172 specified with @option{-l}.
14174 @item -Ym,@var{dir}
14176 Look in the directory @var{dir} to find the M4 preprocessor.
14177 The assembler uses this option.
14178 @c This is supposed to go with a -Yd for predefined M4 macro files, but
14179 @c the generic assembler that comes with Solaris takes just -Ym.
14182 @node TMS320C3x/C4x Options
14183 @subsection TMS320C3x/C4x Options
14184 @cindex TMS320C3x/C4x Options
14186 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
14190 @item -mcpu=@var{cpu_type}
14192 Set the instruction set, register set, and instruction scheduling
14193 parameters for machine type @var{cpu_type}. Supported values for
14194 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
14195 @samp{c44}. The default is @samp{c40} to generate code for the
14200 @itemx -msmall-memory
14202 @opindex mbig-memory
14204 @opindex msmall-memory
14206 Generates code for the big or small memory model. The small memory
14207 model assumed that all data fits into one 64K word page. At run-time
14208 the data page (DP) register must be set to point to the 64K page
14209 containing the .bss and .data program sections. The big memory model is
14210 the default and requires reloading of the DP register for every direct
14217 Allow (disallow) allocation of general integer operands into the block
14218 count register BK@.
14224 Enable (disable) generation of code using decrement and branch,
14225 DBcond(D), instructions. This is enabled by default for the C4x. To be
14226 on the safe side, this is disabled for the C3x, since the maximum
14227 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
14228 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
14229 that it can utilize the decrement and branch instruction, but will give
14230 up if there is more than one memory reference in the loop. Thus a loop
14231 where the loop counter is decremented can generate slightly more
14232 efficient code, in cases where the RPTB instruction cannot be utilized.
14234 @item -mdp-isr-reload
14236 @opindex mdp-isr-reload
14238 Force the DP register to be saved on entry to an interrupt service
14239 routine (ISR), reloaded to point to the data section, and restored on
14240 exit from the ISR@. This should not be required unless someone has
14241 violated the small memory model by modifying the DP register, say within
14248 For the C3x use the 24-bit MPYI instruction for integer multiplies
14249 instead of a library call to guarantee 32-bit results. Note that if one
14250 of the operands is a constant, then the multiplication will be performed
14251 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
14252 then squaring operations are performed inline instead of a library call.
14255 @itemx -mno-fast-fix
14257 @opindex mno-fast-fix
14258 The C3x/C4x FIX instruction to convert a floating point value to an
14259 integer value chooses the nearest integer less than or equal to the
14260 floating point value rather than to the nearest integer. Thus if the
14261 floating point number is negative, the result will be incorrectly
14262 truncated an additional code is necessary to detect and correct this
14263 case. This option can be used to disable generation of the additional
14264 code required to correct the result.
14270 Enable (disable) generation of repeat block sequences using the RPTB
14271 instruction for zero overhead looping. The RPTB construct is only used
14272 for innermost loops that do not call functions or jump across the loop
14273 boundaries. There is no advantage having nested RPTB loops due to the
14274 overhead required to save and restore the RC, RS, and RE registers.
14275 This is enabled by default with @option{-O2}.
14277 @item -mrpts=@var{count}
14281 Enable (disable) the use of the single instruction repeat instruction
14282 RPTS@. If a repeat block contains a single instruction, and the loop
14283 count can be guaranteed to be less than the value @var{count}, GCC will
14284 emit a RPTS instruction instead of a RPTB@. If no value is specified,
14285 then a RPTS will be emitted even if the loop count cannot be determined
14286 at compile time. Note that the repeated instruction following RPTS does
14287 not have to be reloaded from memory each iteration, thus freeing up the
14288 CPU buses for operands. However, since interrupts are blocked by this
14289 instruction, it is disabled by default.
14291 @item -mloop-unsigned
14292 @itemx -mno-loop-unsigned
14293 @opindex mloop-unsigned
14294 @opindex mno-loop-unsigned
14295 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
14296 is @math{2^{31} + 1} since these instructions test if the iteration count is
14297 negative to terminate the loop. If the iteration count is unsigned
14298 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
14299 exceeded. This switch allows an unsigned iteration count.
14303 Try to emit an assembler syntax that the TI assembler (asm30) is happy
14304 with. This also enforces compatibility with the API employed by the TI
14305 C3x C compiler. For example, long doubles are passed as structures
14306 rather than in floating point registers.
14312 Generate code that uses registers (stack) for passing arguments to functions.
14313 By default, arguments are passed in registers where possible rather
14314 than by pushing arguments on to the stack.
14316 @item -mparallel-insns
14317 @itemx -mno-parallel-insns
14318 @opindex mparallel-insns
14319 @opindex mno-parallel-insns
14320 Allow the generation of parallel instructions. This is enabled by
14321 default with @option{-O2}.
14323 @item -mparallel-mpy
14324 @itemx -mno-parallel-mpy
14325 @opindex mparallel-mpy
14326 @opindex mno-parallel-mpy
14327 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
14328 provided @option{-mparallel-insns} is also specified. These instructions have
14329 tight register constraints which can pessimize the code generation
14330 of large functions.
14335 @subsection V850 Options
14336 @cindex V850 Options
14338 These @samp{-m} options are defined for V850 implementations:
14342 @itemx -mno-long-calls
14343 @opindex mlong-calls
14344 @opindex mno-long-calls
14345 Treat all calls as being far away (near). If calls are assumed to be
14346 far away, the compiler will always load the functions address up into a
14347 register, and call indirect through the pointer.
14353 Do not optimize (do optimize) basic blocks that use the same index
14354 pointer 4 or more times to copy pointer into the @code{ep} register, and
14355 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
14356 option is on by default if you optimize.
14358 @item -mno-prolog-function
14359 @itemx -mprolog-function
14360 @opindex mno-prolog-function
14361 @opindex mprolog-function
14362 Do not use (do use) external functions to save and restore registers
14363 at the prologue and epilogue of a function. The external functions
14364 are slower, but use less code space if more than one function saves
14365 the same number of registers. The @option{-mprolog-function} option
14366 is on by default if you optimize.
14370 Try to make the code as small as possible. At present, this just turns
14371 on the @option{-mep} and @option{-mprolog-function} options.
14373 @item -mtda=@var{n}
14375 Put static or global variables whose size is @var{n} bytes or less into
14376 the tiny data area that register @code{ep} points to. The tiny data
14377 area can hold up to 256 bytes in total (128 bytes for byte references).
14379 @item -msda=@var{n}
14381 Put static or global variables whose size is @var{n} bytes or less into
14382 the small data area that register @code{gp} points to. The small data
14383 area can hold up to 64 kilobytes.
14385 @item -mzda=@var{n}
14387 Put static or global variables whose size is @var{n} bytes or less into
14388 the first 32 kilobytes of memory.
14392 Specify that the target processor is the V850.
14395 @opindex mbig-switch
14396 Generate code suitable for big switch tables. Use this option only if
14397 the assembler/linker complain about out of range branches within a switch
14402 This option will cause r2 and r5 to be used in the code generated by
14403 the compiler. This setting is the default.
14405 @item -mno-app-regs
14406 @opindex mno-app-regs
14407 This option will cause r2 and r5 to be treated as fixed registers.
14411 Specify that the target processor is the V850E1. The preprocessor
14412 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
14413 this option is used.
14417 Specify that the target processor is the V850E@. The preprocessor
14418 constant @samp{__v850e__} will be defined if this option is used.
14420 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
14421 are defined then a default target processor will be chosen and the
14422 relevant @samp{__v850*__} preprocessor constant will be defined.
14424 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
14425 defined, regardless of which processor variant is the target.
14427 @item -mdisable-callt
14428 @opindex mdisable-callt
14429 This option will suppress generation of the CALLT instruction for the
14430 v850e and v850e1 flavors of the v850 architecture. The default is
14431 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
14436 @subsection VAX Options
14437 @cindex VAX options
14439 These @samp{-m} options are defined for the VAX:
14444 Do not output certain jump instructions (@code{aobleq} and so on)
14445 that the Unix assembler for the VAX cannot handle across long
14450 Do output those jump instructions, on the assumption that you
14451 will assemble with the GNU assembler.
14455 Output code for g-format floating point numbers instead of d-format.
14458 @node VxWorks Options
14459 @subsection VxWorks Options
14460 @cindex VxWorks Options
14462 The options in this section are defined for all VxWorks targets.
14463 Options specific to the target hardware are listed with the other
14464 options for that target.
14469 GCC can generate code for both VxWorks kernels and real time processes
14470 (RTPs). This option switches from the former to the latter. It also
14471 defines the preprocessor macro @code{__RTP__}.
14474 @opindex non-static
14475 Link an RTP executable against shared libraries rather than static
14476 libraries. The options @option{-static} and @option{-shared} can
14477 also be used for RTPs (@pxref{Link Options}); @option{-static}
14484 These options are passed down to the linker. They are defined for
14485 compatibility with Diab.
14488 @opindex Xbind-lazy
14489 Enable lazy binding of function calls. This option is equivalent to
14490 @option{-Wl,-z,now} and is defined for compatibility with Diab.
14494 Disable lazy binding of function calls. This option is the default and
14495 is defined for compatibility with Diab.
14498 @node x86-64 Options
14499 @subsection x86-64 Options
14500 @cindex x86-64 options
14502 These are listed under @xref{i386 and x86-64 Options}.
14504 @node Xstormy16 Options
14505 @subsection Xstormy16 Options
14506 @cindex Xstormy16 Options
14508 These options are defined for Xstormy16:
14513 Choose startup files and linker script suitable for the simulator.
14516 @node Xtensa Options
14517 @subsection Xtensa Options
14518 @cindex Xtensa Options
14520 These options are supported for Xtensa targets:
14524 @itemx -mno-const16
14526 @opindex mno-const16
14527 Enable or disable use of @code{CONST16} instructions for loading
14528 constant values. The @code{CONST16} instruction is currently not a
14529 standard option from Tensilica. When enabled, @code{CONST16}
14530 instructions are always used in place of the standard @code{L32R}
14531 instructions. The use of @code{CONST16} is enabled by default only if
14532 the @code{L32R} instruction is not available.
14535 @itemx -mno-fused-madd
14536 @opindex mfused-madd
14537 @opindex mno-fused-madd
14538 Enable or disable use of fused multiply/add and multiply/subtract
14539 instructions in the floating-point option. This has no effect if the
14540 floating-point option is not also enabled. Disabling fused multiply/add
14541 and multiply/subtract instructions forces the compiler to use separate
14542 instructions for the multiply and add/subtract operations. This may be
14543 desirable in some cases where strict IEEE 754-compliant results are
14544 required: the fused multiply add/subtract instructions do not round the
14545 intermediate result, thereby producing results with @emph{more} bits of
14546 precision than specified by the IEEE standard. Disabling fused multiply
14547 add/subtract instructions also ensures that the program output is not
14548 sensitive to the compiler's ability to combine multiply and add/subtract
14551 @item -mtext-section-literals
14552 @itemx -mno-text-section-literals
14553 @opindex mtext-section-literals
14554 @opindex mno-text-section-literals
14555 Control the treatment of literal pools. The default is
14556 @option{-mno-text-section-literals}, which places literals in a separate
14557 section in the output file. This allows the literal pool to be placed
14558 in a data RAM/ROM, and it also allows the linker to combine literal
14559 pools from separate object files to remove redundant literals and
14560 improve code size. With @option{-mtext-section-literals}, the literals
14561 are interspersed in the text section in order to keep them as close as
14562 possible to their references. This may be necessary for large assembly
14565 @item -mtarget-align
14566 @itemx -mno-target-align
14567 @opindex mtarget-align
14568 @opindex mno-target-align
14569 When this option is enabled, GCC instructs the assembler to
14570 automatically align instructions to reduce branch penalties at the
14571 expense of some code density. The assembler attempts to widen density
14572 instructions to align branch targets and the instructions following call
14573 instructions. If there are not enough preceding safe density
14574 instructions to align a target, no widening will be performed. The
14575 default is @option{-mtarget-align}. These options do not affect the
14576 treatment of auto-aligned instructions like @code{LOOP}, which the
14577 assembler will always align, either by widening density instructions or
14578 by inserting no-op instructions.
14581 @itemx -mno-longcalls
14582 @opindex mlongcalls
14583 @opindex mno-longcalls
14584 When this option is enabled, GCC instructs the assembler to translate
14585 direct calls to indirect calls unless it can determine that the target
14586 of a direct call is in the range allowed by the call instruction. This
14587 translation typically occurs for calls to functions in other source
14588 files. Specifically, the assembler translates a direct @code{CALL}
14589 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
14590 The default is @option{-mno-longcalls}. This option should be used in
14591 programs where the call target can potentially be out of range. This
14592 option is implemented in the assembler, not the compiler, so the
14593 assembly code generated by GCC will still show direct call
14594 instructions---look at the disassembled object code to see the actual
14595 instructions. Note that the assembler will use an indirect call for
14596 every cross-file call, not just those that really will be out of range.
14599 @node zSeries Options
14600 @subsection zSeries Options
14601 @cindex zSeries options
14603 These are listed under @xref{S/390 and zSeries Options}.
14605 @node Code Gen Options
14606 @section Options for Code Generation Conventions
14607 @cindex code generation conventions
14608 @cindex options, code generation
14609 @cindex run-time options
14611 These machine-independent options control the interface conventions
14612 used in code generation.
14614 Most of them have both positive and negative forms; the negative form
14615 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
14616 one of the forms is listed---the one which is not the default. You
14617 can figure out the other form by either removing @samp{no-} or adding
14621 @item -fbounds-check
14622 @opindex fbounds-check
14623 For front-ends that support it, generate additional code to check that
14624 indices used to access arrays are within the declared range. This is
14625 currently only supported by the Java and Fortran front-ends, where
14626 this option defaults to true and false respectively.
14630 This option generates traps for signed overflow on addition, subtraction,
14631 multiplication operations.
14635 This option instructs the compiler to assume that signed arithmetic
14636 overflow of addition, subtraction and multiplication wraps around
14637 using twos-complement representation. This flag enables some optimizations
14638 and disables others. This option is enabled by default for the Java
14639 front-end, as required by the Java language specification.
14642 @opindex fexceptions
14643 Enable exception handling. Generates extra code needed to propagate
14644 exceptions. For some targets, this implies GCC will generate frame
14645 unwind information for all functions, which can produce significant data
14646 size overhead, although it does not affect execution. If you do not
14647 specify this option, GCC will enable it by default for languages like
14648 C++ which normally require exception handling, and disable it for
14649 languages like C that do not normally require it. However, you may need
14650 to enable this option when compiling C code that needs to interoperate
14651 properly with exception handlers written in C++. You may also wish to
14652 disable this option if you are compiling older C++ programs that don't
14653 use exception handling.
14655 @item -fnon-call-exceptions
14656 @opindex fnon-call-exceptions
14657 Generate code that allows trapping instructions to throw exceptions.
14658 Note that this requires platform-specific runtime support that does
14659 not exist everywhere. Moreover, it only allows @emph{trapping}
14660 instructions to throw exceptions, i.e.@: memory references or floating
14661 point instructions. It does not allow exceptions to be thrown from
14662 arbitrary signal handlers such as @code{SIGALRM}.
14664 @item -funwind-tables
14665 @opindex funwind-tables
14666 Similar to @option{-fexceptions}, except that it will just generate any needed
14667 static data, but will not affect the generated code in any other way.
14668 You will normally not enable this option; instead, a language processor
14669 that needs this handling would enable it on your behalf.
14671 @item -fasynchronous-unwind-tables
14672 @opindex fasynchronous-unwind-tables
14673 Generate unwind table in dwarf2 format, if supported by target machine. The
14674 table is exact at each instruction boundary, so it can be used for stack
14675 unwinding from asynchronous events (such as debugger or garbage collector).
14677 @item -fpcc-struct-return
14678 @opindex fpcc-struct-return
14679 Return ``short'' @code{struct} and @code{union} values in memory like
14680 longer ones, rather than in registers. This convention is less
14681 efficient, but it has the advantage of allowing intercallability between
14682 GCC-compiled files and files compiled with other compilers, particularly
14683 the Portable C Compiler (pcc).
14685 The precise convention for returning structures in memory depends
14686 on the target configuration macros.
14688 Short structures and unions are those whose size and alignment match
14689 that of some integer type.
14691 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
14692 switch is not binary compatible with code compiled with the
14693 @option{-freg-struct-return} switch.
14694 Use it to conform to a non-default application binary interface.
14696 @item -freg-struct-return
14697 @opindex freg-struct-return
14698 Return @code{struct} and @code{union} values in registers when possible.
14699 This is more efficient for small structures than
14700 @option{-fpcc-struct-return}.
14702 If you specify neither @option{-fpcc-struct-return} nor
14703 @option{-freg-struct-return}, GCC defaults to whichever convention is
14704 standard for the target. If there is no standard convention, GCC
14705 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
14706 the principal compiler. In those cases, we can choose the standard, and
14707 we chose the more efficient register return alternative.
14709 @strong{Warning:} code compiled with the @option{-freg-struct-return}
14710 switch is not binary compatible with code compiled with the
14711 @option{-fpcc-struct-return} switch.
14712 Use it to conform to a non-default application binary interface.
14714 @item -fshort-enums
14715 @opindex fshort-enums
14716 Allocate to an @code{enum} type only as many bytes as it needs for the
14717 declared range of possible values. Specifically, the @code{enum} type
14718 will be equivalent to the smallest integer type which has enough room.
14720 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
14721 code that is not binary compatible with code generated without that switch.
14722 Use it to conform to a non-default application binary interface.
14724 @item -fshort-double
14725 @opindex fshort-double
14726 Use the same size for @code{double} as for @code{float}.
14728 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
14729 code that is not binary compatible with code generated without that switch.
14730 Use it to conform to a non-default application binary interface.
14732 @item -fshort-wchar
14733 @opindex fshort-wchar
14734 Override the underlying type for @samp{wchar_t} to be @samp{short
14735 unsigned int} instead of the default for the target. This option is
14736 useful for building programs to run under WINE@.
14738 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
14739 code that is not binary compatible with code generated without that switch.
14740 Use it to conform to a non-default application binary interface.
14743 @opindex fno-common
14744 In C, allocate even uninitialized global variables in the data section of the
14745 object file, rather than generating them as common blocks. This has the
14746 effect that if the same variable is declared (without @code{extern}) in
14747 two different compilations, you will get an error when you link them.
14748 The only reason this might be useful is if you wish to verify that the
14749 program will work on other systems which always work this way.
14753 Ignore the @samp{#ident} directive.
14755 @item -finhibit-size-directive
14756 @opindex finhibit-size-directive
14757 Don't output a @code{.size} assembler directive, or anything else that
14758 would cause trouble if the function is split in the middle, and the
14759 two halves are placed at locations far apart in memory. This option is
14760 used when compiling @file{crtstuff.c}; you should not need to use it
14763 @item -fverbose-asm
14764 @opindex fverbose-asm
14765 Put extra commentary information in the generated assembly code to
14766 make it more readable. This option is generally only of use to those
14767 who actually need to read the generated assembly code (perhaps while
14768 debugging the compiler itself).
14770 @option{-fno-verbose-asm}, the default, causes the
14771 extra information to be omitted and is useful when comparing two assembler
14774 @item -frecord-gcc-switches
14775 @opindex frecord-gcc-switches
14776 This switch causes the command line that was used to invoke the
14777 compiler to be recorded into the object file that is being created.
14778 This switch is only implemented on some targets and the exact format
14779 of the recording is target and binary file format dependent, but it
14780 usually takes the form of a section containing ASCII text. This
14781 switch is related to the @option{-fverbose-asm} switch, but that
14782 switch only records information in the assembler output file as
14783 comments, so it never reaches the object file.
14787 @cindex global offset table
14789 Generate position-independent code (PIC) suitable for use in a shared
14790 library, if supported for the target machine. Such code accesses all
14791 constant addresses through a global offset table (GOT)@. The dynamic
14792 loader resolves the GOT entries when the program starts (the dynamic
14793 loader is not part of GCC; it is part of the operating system). If
14794 the GOT size for the linked executable exceeds a machine-specific
14795 maximum size, you get an error message from the linker indicating that
14796 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
14797 instead. (These maximums are 8k on the SPARC and 32k
14798 on the m68k and RS/6000. The 386 has no such limit.)
14800 Position-independent code requires special support, and therefore works
14801 only on certain machines. For the 386, GCC supports PIC for System V
14802 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
14803 position-independent.
14805 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14810 If supported for the target machine, emit position-independent code,
14811 suitable for dynamic linking and avoiding any limit on the size of the
14812 global offset table. This option makes a difference on the m68k,
14813 PowerPC and SPARC@.
14815 Position-independent code requires special support, and therefore works
14816 only on certain machines.
14818 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14825 These options are similar to @option{-fpic} and @option{-fPIC}, but
14826 generated position independent code can be only linked into executables.
14827 Usually these options are used when @option{-pie} GCC option will be
14828 used during linking.
14830 @option{-fpie} and @option{-fPIE} both define the macros
14831 @code{__pie__} and @code{__PIE__}. The macros have the value 1
14832 for @option{-fpie} and 2 for @option{-fPIE}.
14834 @item -fno-jump-tables
14835 @opindex fno-jump-tables
14836 Do not use jump tables for switch statements even where it would be
14837 more efficient than other code generation strategies. This option is
14838 of use in conjunction with @option{-fpic} or @option{-fPIC} for
14839 building code which forms part of a dynamic linker and cannot
14840 reference the address of a jump table. On some targets, jump tables
14841 do not require a GOT and this option is not needed.
14843 @item -ffixed-@var{reg}
14845 Treat the register named @var{reg} as a fixed register; generated code
14846 should never refer to it (except perhaps as a stack pointer, frame
14847 pointer or in some other fixed role).
14849 @var{reg} must be the name of a register. The register names accepted
14850 are machine-specific and are defined in the @code{REGISTER_NAMES}
14851 macro in the machine description macro file.
14853 This flag does not have a negative form, because it specifies a
14856 @item -fcall-used-@var{reg}
14857 @opindex fcall-used
14858 Treat the register named @var{reg} as an allocable register that is
14859 clobbered by function calls. It may be allocated for temporaries or
14860 variables that do not live across a call. Functions compiled this way
14861 will not save and restore the register @var{reg}.
14863 It is an error to used this flag with the frame pointer or stack pointer.
14864 Use of this flag for other registers that have fixed pervasive roles in
14865 the machine's execution model will produce disastrous results.
14867 This flag does not have a negative form, because it specifies a
14870 @item -fcall-saved-@var{reg}
14871 @opindex fcall-saved
14872 Treat the register named @var{reg} as an allocable register saved by
14873 functions. It may be allocated even for temporaries or variables that
14874 live across a call. Functions compiled this way will save and restore
14875 the register @var{reg} if they use it.
14877 It is an error to used this flag with the frame pointer or stack pointer.
14878 Use of this flag for other registers that have fixed pervasive roles in
14879 the machine's execution model will produce disastrous results.
14881 A different sort of disaster will result from the use of this flag for
14882 a register in which function values may be returned.
14884 This flag does not have a negative form, because it specifies a
14887 @item -fpack-struct[=@var{n}]
14888 @opindex fpack-struct
14889 Without a value specified, pack all structure members together without
14890 holes. When a value is specified (which must be a small power of two), pack
14891 structure members according to this value, representing the maximum
14892 alignment (that is, objects with default alignment requirements larger than
14893 this will be output potentially unaligned at the next fitting location.
14895 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
14896 code that is not binary compatible with code generated without that switch.
14897 Additionally, it makes the code suboptimal.
14898 Use it to conform to a non-default application binary interface.
14900 @item -finstrument-functions
14901 @opindex finstrument-functions
14902 Generate instrumentation calls for entry and exit to functions. Just
14903 after function entry and just before function exit, the following
14904 profiling functions will be called with the address of the current
14905 function and its call site. (On some platforms,
14906 @code{__builtin_return_address} does not work beyond the current
14907 function, so the call site information may not be available to the
14908 profiling functions otherwise.)
14911 void __cyg_profile_func_enter (void *this_fn,
14913 void __cyg_profile_func_exit (void *this_fn,
14917 The first argument is the address of the start of the current function,
14918 which may be looked up exactly in the symbol table.
14920 This instrumentation is also done for functions expanded inline in other
14921 functions. The profiling calls will indicate where, conceptually, the
14922 inline function is entered and exited. This means that addressable
14923 versions of such functions must be available. If all your uses of a
14924 function are expanded inline, this may mean an additional expansion of
14925 code size. If you use @samp{extern inline} in your C code, an
14926 addressable version of such functions must be provided. (This is
14927 normally the case anyways, but if you get lucky and the optimizer always
14928 expands the functions inline, you might have gotten away without
14929 providing static copies.)
14931 A function may be given the attribute @code{no_instrument_function}, in
14932 which case this instrumentation will not be done. This can be used, for
14933 example, for the profiling functions listed above, high-priority
14934 interrupt routines, and any functions from which the profiling functions
14935 cannot safely be called (perhaps signal handlers, if the profiling
14936 routines generate output or allocate memory).
14938 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
14939 @opindex finstrument-functions-exclude-file-list
14941 Set the list of functions that are excluded from instrumentation (see
14942 the description of @code{-finstrument-functions}). If the file that
14943 contains a function definition matches with one of @var{file}, then
14944 that function is not instrumented. The match is done on substrings:
14945 if the @var{file} parameter is a substring of the file name, it is
14946 considered to be a match.
14949 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
14950 will exclude any inline function defined in files whose pathnames
14951 contain @code{/bits/stl} or @code{include/sys}.
14953 If, for some reason, you want to include letter @code{','} in one of
14954 @var{sym}, write @code{'\,'}. For example,
14955 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
14956 (note the single quote surrounding the option).
14958 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
14959 @opindex finstrument-functions-exclude-function-list
14961 This is similar to @code{-finstrument-functions-exclude-file-list},
14962 but this option sets the list of function names to be excluded from
14963 instrumentation. The function name to be matched is its user-visible
14964 name, such as @code{vector<int> blah(const vector<int> &)}, not the
14965 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
14966 match is done on substrings: if the @var{sym} parameter is a substring
14967 of the function name, it is considered to be a match.
14969 @item -fstack-check
14970 @opindex fstack-check
14971 Generate code to verify that you do not go beyond the boundary of the
14972 stack. You should specify this flag if you are running in an
14973 environment with multiple threads, but only rarely need to specify it in
14974 a single-threaded environment since stack overflow is automatically
14975 detected on nearly all systems if there is only one stack.
14977 Note that this switch does not actually cause checking to be done; the
14978 operating system must do that. The switch causes generation of code
14979 to ensure that the operating system sees the stack being extended.
14981 @item -fstack-limit-register=@var{reg}
14982 @itemx -fstack-limit-symbol=@var{sym}
14983 @itemx -fno-stack-limit
14984 @opindex fstack-limit-register
14985 @opindex fstack-limit-symbol
14986 @opindex fno-stack-limit
14987 Generate code to ensure that the stack does not grow beyond a certain value,
14988 either the value of a register or the address of a symbol. If the stack
14989 would grow beyond the value, a signal is raised. For most targets,
14990 the signal is raised before the stack overruns the boundary, so
14991 it is possible to catch the signal without taking special precautions.
14993 For instance, if the stack starts at absolute address @samp{0x80000000}
14994 and grows downwards, you can use the flags
14995 @option{-fstack-limit-symbol=__stack_limit} and
14996 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
14997 of 128KB@. Note that this may only work with the GNU linker.
14999 @cindex aliasing of parameters
15000 @cindex parameters, aliased
15001 @item -fargument-alias
15002 @itemx -fargument-noalias
15003 @itemx -fargument-noalias-global
15004 @itemx -fargument-noalias-anything
15005 @opindex fargument-alias
15006 @opindex fargument-noalias
15007 @opindex fargument-noalias-global
15008 @opindex fargument-noalias-anything
15009 Specify the possible relationships among parameters and between
15010 parameters and global data.
15012 @option{-fargument-alias} specifies that arguments (parameters) may
15013 alias each other and may alias global storage.@*
15014 @option{-fargument-noalias} specifies that arguments do not alias
15015 each other, but may alias global storage.@*
15016 @option{-fargument-noalias-global} specifies that arguments do not
15017 alias each other and do not alias global storage.
15018 @option{-fargument-noalias-anything} specifies that arguments do not
15019 alias any other storage.
15021 Each language will automatically use whatever option is required by
15022 the language standard. You should not need to use these options yourself.
15024 @item -fleading-underscore
15025 @opindex fleading-underscore
15026 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
15027 change the way C symbols are represented in the object file. One use
15028 is to help link with legacy assembly code.
15030 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
15031 generate code that is not binary compatible with code generated without that
15032 switch. Use it to conform to a non-default application binary interface.
15033 Not all targets provide complete support for this switch.
15035 @item -ftls-model=@var{model}
15036 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
15037 The @var{model} argument should be one of @code{global-dynamic},
15038 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
15040 The default without @option{-fpic} is @code{initial-exec}; with
15041 @option{-fpic} the default is @code{global-dynamic}.
15043 @item -fvisibility=@var{default|internal|hidden|protected}
15044 @opindex fvisibility
15045 Set the default ELF image symbol visibility to the specified option---all
15046 symbols will be marked with this unless overridden within the code.
15047 Using this feature can very substantially improve linking and
15048 load times of shared object libraries, produce more optimized
15049 code, provide near-perfect API export and prevent symbol clashes.
15050 It is @strong{strongly} recommended that you use this in any shared objects
15053 Despite the nomenclature, @code{default} always means public ie;
15054 available to be linked against from outside the shared object.
15055 @code{protected} and @code{internal} are pretty useless in real-world
15056 usage so the only other commonly used option will be @code{hidden}.
15057 The default if @option{-fvisibility} isn't specified is
15058 @code{default}, i.e., make every
15059 symbol public---this causes the same behavior as previous versions of
15062 A good explanation of the benefits offered by ensuring ELF
15063 symbols have the correct visibility is given by ``How To Write
15064 Shared Libraries'' by Ulrich Drepper (which can be found at
15065 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
15066 solution made possible by this option to marking things hidden when
15067 the default is public is to make the default hidden and mark things
15068 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
15069 and @code{__attribute__ ((visibility("default")))} instead of
15070 @code{__declspec(dllexport)} you get almost identical semantics with
15071 identical syntax. This is a great boon to those working with
15072 cross-platform projects.
15074 For those adding visibility support to existing code, you may find
15075 @samp{#pragma GCC visibility} of use. This works by you enclosing
15076 the declarations you wish to set visibility for with (for example)
15077 @samp{#pragma GCC visibility push(hidden)} and
15078 @samp{#pragma GCC visibility pop}.
15079 Bear in mind that symbol visibility should be viewed @strong{as
15080 part of the API interface contract} and thus all new code should
15081 always specify visibility when it is not the default ie; declarations
15082 only for use within the local DSO should @strong{always} be marked explicitly
15083 as hidden as so to avoid PLT indirection overheads---making this
15084 abundantly clear also aids readability and self-documentation of the code.
15085 Note that due to ISO C++ specification requirements, operator new and
15086 operator delete must always be of default visibility.
15088 Be aware that headers from outside your project, in particular system
15089 headers and headers from any other library you use, may not be
15090 expecting to be compiled with visibility other than the default. You
15091 may need to explicitly say @samp{#pragma GCC visibility push(default)}
15092 before including any such headers.
15094 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
15095 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
15096 no modifications. However, this means that calls to @samp{extern}
15097 functions with no explicit visibility will use the PLT, so it is more
15098 effective to use @samp{__attribute ((visibility))} and/or
15099 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
15100 declarations should be treated as hidden.
15102 Note that @samp{-fvisibility} does affect C++ vague linkage
15103 entities. This means that, for instance, an exception class that will
15104 be thrown between DSOs must be explicitly marked with default
15105 visibility so that the @samp{type_info} nodes will be unified between
15108 An overview of these techniques, their benefits and how to use them
15109 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
15115 @node Environment Variables
15116 @section Environment Variables Affecting GCC
15117 @cindex environment variables
15119 @c man begin ENVIRONMENT
15120 This section describes several environment variables that affect how GCC
15121 operates. Some of them work by specifying directories or prefixes to use
15122 when searching for various kinds of files. Some are used to specify other
15123 aspects of the compilation environment.
15125 Note that you can also specify places to search using options such as
15126 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
15127 take precedence over places specified using environment variables, which
15128 in turn take precedence over those specified by the configuration of GCC@.
15129 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
15130 GNU Compiler Collection (GCC) Internals}.
15135 @c @itemx LC_COLLATE
15137 @c @itemx LC_MONETARY
15138 @c @itemx LC_NUMERIC
15143 @c @findex LC_COLLATE
15144 @findex LC_MESSAGES
15145 @c @findex LC_MONETARY
15146 @c @findex LC_NUMERIC
15150 These environment variables control the way that GCC uses
15151 localization information that allow GCC to work with different
15152 national conventions. GCC inspects the locale categories
15153 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
15154 so. These locale categories can be set to any value supported by your
15155 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
15156 Kingdom encoded in UTF-8.
15158 The @env{LC_CTYPE} environment variable specifies character
15159 classification. GCC uses it to determine the character boundaries in
15160 a string; this is needed for some multibyte encodings that contain quote
15161 and escape characters that would otherwise be interpreted as a string
15164 The @env{LC_MESSAGES} environment variable specifies the language to
15165 use in diagnostic messages.
15167 If the @env{LC_ALL} environment variable is set, it overrides the value
15168 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
15169 and @env{LC_MESSAGES} default to the value of the @env{LANG}
15170 environment variable. If none of these variables are set, GCC
15171 defaults to traditional C English behavior.
15175 If @env{TMPDIR} is set, it specifies the directory to use for temporary
15176 files. GCC uses temporary files to hold the output of one stage of
15177 compilation which is to be used as input to the next stage: for example,
15178 the output of the preprocessor, which is the input to the compiler
15181 @item GCC_EXEC_PREFIX
15182 @findex GCC_EXEC_PREFIX
15183 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
15184 names of the subprograms executed by the compiler. No slash is added
15185 when this prefix is combined with the name of a subprogram, but you can
15186 specify a prefix that ends with a slash if you wish.
15188 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
15189 an appropriate prefix to use based on the pathname it was invoked with.
15191 If GCC cannot find the subprogram using the specified prefix, it
15192 tries looking in the usual places for the subprogram.
15194 The default value of @env{GCC_EXEC_PREFIX} is
15195 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
15196 the installed compiler. In many cases @var{prefix} is the value
15197 of @code{prefix} when you ran the @file{configure} script.
15199 Other prefixes specified with @option{-B} take precedence over this prefix.
15201 This prefix is also used for finding files such as @file{crt0.o} that are
15204 In addition, the prefix is used in an unusual way in finding the
15205 directories to search for header files. For each of the standard
15206 directories whose name normally begins with @samp{/usr/local/lib/gcc}
15207 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
15208 replacing that beginning with the specified prefix to produce an
15209 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
15210 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
15211 These alternate directories are searched first; the standard directories
15212 come next. If a standard directory begins with the configured
15213 @var{prefix} then the value of @var{prefix} is replaced by
15214 @env{GCC_EXEC_PREFIX} when looking for header files.
15216 @item COMPILER_PATH
15217 @findex COMPILER_PATH
15218 The value of @env{COMPILER_PATH} is a colon-separated list of
15219 directories, much like @env{PATH}. GCC tries the directories thus
15220 specified when searching for subprograms, if it can't find the
15221 subprograms using @env{GCC_EXEC_PREFIX}.
15224 @findex LIBRARY_PATH
15225 The value of @env{LIBRARY_PATH} is a colon-separated list of
15226 directories, much like @env{PATH}. When configured as a native compiler,
15227 GCC tries the directories thus specified when searching for special
15228 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
15229 using GCC also uses these directories when searching for ordinary
15230 libraries for the @option{-l} option (but directories specified with
15231 @option{-L} come first).
15235 @cindex locale definition
15236 This variable is used to pass locale information to the compiler. One way in
15237 which this information is used is to determine the character set to be used
15238 when character literals, string literals and comments are parsed in C and C++.
15239 When the compiler is configured to allow multibyte characters,
15240 the following values for @env{LANG} are recognized:
15244 Recognize JIS characters.
15246 Recognize SJIS characters.
15248 Recognize EUCJP characters.
15251 If @env{LANG} is not defined, or if it has some other value, then the
15252 compiler will use mblen and mbtowc as defined by the default locale to
15253 recognize and translate multibyte characters.
15257 Some additional environments variables affect the behavior of the
15260 @include cppenv.texi
15264 @node Precompiled Headers
15265 @section Using Precompiled Headers
15266 @cindex precompiled headers
15267 @cindex speed of compilation
15269 Often large projects have many header files that are included in every
15270 source file. The time the compiler takes to process these header files
15271 over and over again can account for nearly all of the time required to
15272 build the project. To make builds faster, GCC allows users to
15273 `precompile' a header file; then, if builds can use the precompiled
15274 header file they will be much faster.
15276 To create a precompiled header file, simply compile it as you would any
15277 other file, if necessary using the @option{-x} option to make the driver
15278 treat it as a C or C++ header file. You will probably want to use a
15279 tool like @command{make} to keep the precompiled header up-to-date when
15280 the headers it contains change.
15282 A precompiled header file will be searched for when @code{#include} is
15283 seen in the compilation. As it searches for the included file
15284 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
15285 compiler looks for a precompiled header in each directory just before it
15286 looks for the include file in that directory. The name searched for is
15287 the name specified in the @code{#include} with @samp{.gch} appended. If
15288 the precompiled header file can't be used, it is ignored.
15290 For instance, if you have @code{#include "all.h"}, and you have
15291 @file{all.h.gch} in the same directory as @file{all.h}, then the
15292 precompiled header file will be used if possible, and the original
15293 header will be used otherwise.
15295 Alternatively, you might decide to put the precompiled header file in a
15296 directory and use @option{-I} to ensure that directory is searched
15297 before (or instead of) the directory containing the original header.
15298 Then, if you want to check that the precompiled header file is always
15299 used, you can put a file of the same name as the original header in this
15300 directory containing an @code{#error} command.
15302 This also works with @option{-include}. So yet another way to use
15303 precompiled headers, good for projects not designed with precompiled
15304 header files in mind, is to simply take most of the header files used by
15305 a project, include them from another header file, precompile that header
15306 file, and @option{-include} the precompiled header. If the header files
15307 have guards against multiple inclusion, they will be skipped because
15308 they've already been included (in the precompiled header).
15310 If you need to precompile the same header file for different
15311 languages, targets, or compiler options, you can instead make a
15312 @emph{directory} named like @file{all.h.gch}, and put each precompiled
15313 header in the directory, perhaps using @option{-o}. It doesn't matter
15314 what you call the files in the directory, every precompiled header in
15315 the directory will be considered. The first precompiled header
15316 encountered in the directory that is valid for this compilation will
15317 be used; they're searched in no particular order.
15319 There are many other possibilities, limited only by your imagination,
15320 good sense, and the constraints of your build system.
15322 A precompiled header file can be used only when these conditions apply:
15326 Only one precompiled header can be used in a particular compilation.
15329 A precompiled header can't be used once the first C token is seen. You
15330 can have preprocessor directives before a precompiled header; you can
15331 even include a precompiled header from inside another header, so long as
15332 there are no C tokens before the @code{#include}.
15335 The precompiled header file must be produced for the same language as
15336 the current compilation. You can't use a C precompiled header for a C++
15340 The precompiled header file must have been produced by the same compiler
15341 binary as the current compilation is using.
15344 Any macros defined before the precompiled header is included must
15345 either be defined in the same way as when the precompiled header was
15346 generated, or must not affect the precompiled header, which usually
15347 means that they don't appear in the precompiled header at all.
15349 The @option{-D} option is one way to define a macro before a
15350 precompiled header is included; using a @code{#define} can also do it.
15351 There are also some options that define macros implicitly, like
15352 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
15355 @item If debugging information is output when using the precompiled
15356 header, using @option{-g} or similar, the same kind of debugging information
15357 must have been output when building the precompiled header. However,
15358 a precompiled header built using @option{-g} can be used in a compilation
15359 when no debugging information is being output.
15361 @item The same @option{-m} options must generally be used when building
15362 and using the precompiled header. @xref{Submodel Options},
15363 for any cases where this rule is relaxed.
15365 @item Each of the following options must be the same when building and using
15366 the precompiled header:
15368 @gccoptlist{-fexceptions -funit-at-a-time}
15371 Some other command-line options starting with @option{-f},
15372 @option{-p}, or @option{-O} must be defined in the same way as when
15373 the precompiled header was generated. At present, it's not clear
15374 which options are safe to change and which are not; the safest choice
15375 is to use exactly the same options when generating and using the
15376 precompiled header. The following are known to be safe:
15378 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
15379 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
15380 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
15385 For all of these except the last, the compiler will automatically
15386 ignore the precompiled header if the conditions aren't met. If you
15387 find an option combination that doesn't work and doesn't cause the
15388 precompiled header to be ignored, please consider filing a bug report,
15391 If you do use differing options when generating and using the
15392 precompiled header, the actual behavior will be a mixture of the
15393 behavior for the options. For instance, if you use @option{-g} to
15394 generate the precompiled header but not when using it, you may or may
15395 not get debugging information for routines in the precompiled header.
15397 @node Running Protoize
15398 @section Running Protoize
15400 The program @code{protoize} is an optional part of GCC@. You can use
15401 it to add prototypes to a program, thus converting the program to ISO
15402 C in one respect. The companion program @code{unprotoize} does the
15403 reverse: it removes argument types from any prototypes that are found.
15405 When you run these programs, you must specify a set of source files as
15406 command line arguments. The conversion programs start out by compiling
15407 these files to see what functions they define. The information gathered
15408 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
15410 After scanning comes actual conversion. The specified files are all
15411 eligible to be converted; any files they include (whether sources or
15412 just headers) are eligible as well.
15414 But not all the eligible files are converted. By default,
15415 @code{protoize} and @code{unprotoize} convert only source and header
15416 files in the current directory. You can specify additional directories
15417 whose files should be converted with the @option{-d @var{directory}}
15418 option. You can also specify particular files to exclude with the
15419 @option{-x @var{file}} option. A file is converted if it is eligible, its
15420 directory name matches one of the specified directory names, and its
15421 name within the directory has not been excluded.
15423 Basic conversion with @code{protoize} consists of rewriting most
15424 function definitions and function declarations to specify the types of
15425 the arguments. The only ones not rewritten are those for varargs
15428 @code{protoize} optionally inserts prototype declarations at the
15429 beginning of the source file, to make them available for any calls that
15430 precede the function's definition. Or it can insert prototype
15431 declarations with block scope in the blocks where undeclared functions
15434 Basic conversion with @code{unprotoize} consists of rewriting most
15435 function declarations to remove any argument types, and rewriting
15436 function definitions to the old-style pre-ISO form.
15438 Both conversion programs print a warning for any function declaration or
15439 definition that they can't convert. You can suppress these warnings
15442 The output from @code{protoize} or @code{unprotoize} replaces the
15443 original source file. The original file is renamed to a name ending
15444 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
15445 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
15446 for DOS) file already exists, then the source file is simply discarded.
15448 @code{protoize} and @code{unprotoize} both depend on GCC itself to
15449 scan the program and collect information about the functions it uses.
15450 So neither of these programs will work until GCC is installed.
15452 Here is a table of the options you can use with @code{protoize} and
15453 @code{unprotoize}. Each option works with both programs unless
15457 @item -B @var{directory}
15458 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
15459 usual directory (normally @file{/usr/local/lib}). This file contains
15460 prototype information about standard system functions. This option
15461 applies only to @code{protoize}.
15463 @item -c @var{compilation-options}
15464 Use @var{compilation-options} as the options when running @command{gcc} to
15465 produce the @samp{.X} files. The special option @option{-aux-info} is
15466 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
15468 Note that the compilation options must be given as a single argument to
15469 @code{protoize} or @code{unprotoize}. If you want to specify several
15470 @command{gcc} options, you must quote the entire set of compilation options
15471 to make them a single word in the shell.
15473 There are certain @command{gcc} arguments that you cannot use, because they
15474 would produce the wrong kind of output. These include @option{-g},
15475 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
15476 the @var{compilation-options}, they are ignored.
15479 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
15480 systems) instead of @samp{.c}. This is convenient if you are converting
15481 a C program to C++. This option applies only to @code{protoize}.
15484 Add explicit global declarations. This means inserting explicit
15485 declarations at the beginning of each source file for each function
15486 that is called in the file and was not declared. These declarations
15487 precede the first function definition that contains a call to an
15488 undeclared function. This option applies only to @code{protoize}.
15490 @item -i @var{string}
15491 Indent old-style parameter declarations with the string @var{string}.
15492 This option applies only to @code{protoize}.
15494 @code{unprotoize} converts prototyped function definitions to old-style
15495 function definitions, where the arguments are declared between the
15496 argument list and the initial @samp{@{}. By default, @code{unprotoize}
15497 uses five spaces as the indentation. If you want to indent with just
15498 one space instead, use @option{-i " "}.
15501 Keep the @samp{.X} files. Normally, they are deleted after conversion
15505 Add explicit local declarations. @code{protoize} with @option{-l} inserts
15506 a prototype declaration for each function in each block which calls the
15507 function without any declaration. This option applies only to
15511 Make no real changes. This mode just prints information about the conversions
15512 that would have been done without @option{-n}.
15515 Make no @samp{.save} files. The original files are simply deleted.
15516 Use this option with caution.
15518 @item -p @var{program}
15519 Use the program @var{program} as the compiler. Normally, the name
15520 @file{gcc} is used.
15523 Work quietly. Most warnings are suppressed.
15526 Print the version number, just like @option{-v} for @command{gcc}.
15529 If you need special compiler options to compile one of your program's
15530 source files, then you should generate that file's @samp{.X} file
15531 specially, by running @command{gcc} on that source file with the
15532 appropriate options and the option @option{-aux-info}. Then run
15533 @code{protoize} on the entire set of files. @code{protoize} will use
15534 the existing @samp{.X} file because it is newer than the source file.
15538 gcc -Dfoo=bar file1.c -aux-info file1.X
15543 You need to include the special files along with the rest in the
15544 @code{protoize} command, even though their @samp{.X} files already
15545 exist, because otherwise they won't get converted.
15547 @xref{Protoize Caveats}, for more information on how to use
15548 @code{protoize} successfully.