1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
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, 2007, 2008
15 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 -wrapper@@@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++ -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 @gol
234 -Wno-deprecated-declarations -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-contains-nul -Wno-format-extra-args -Wformat-nonliteral @gol
239 -Wformat-security -Wformat-y2k @gol
240 -Wframe-larger-than=@var{len} -Wignored-qualifiers @gol
241 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
242 -Wimport -Wno-import -Winit-self -Winline @gol
243 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
244 -Winvalid-pch -Wlarger-than=@var{len} -Wunsafe-loop-optimizations @gol
245 -Wlogical-op -Wlong-long @gol
246 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
247 -Wmissing-format-attribute -Wmissing-include-dirs @gol
248 -Wmissing-noreturn -Wno-mudflap @gol
249 -Wno-multichar -Wnonnull -Wno-overflow @gol
250 -Woverlength-strings -Wpacked -Wpadded @gol
251 -Wparentheses -Wpointer-arith -Wno-pointer-to-int-cast @gol
252 -Wredundant-decls @gol
253 -Wreturn-type -Wsequence-point -Wshadow @gol
254 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
255 -Wstrict-aliasing -Wstrict-aliasing=n @gol
256 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
257 -Wswitch -Wswitch-default -Wswitch-enum @gol
258 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
259 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
260 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
261 -Wunused-value -Wunused-variable @gol
262 -Wvariadic-macros -Wvla @gol
263 -Wvolatile-register-var -Wwrite-strings}
265 @item C and Objective-C-only Warning Options
266 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
267 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
268 -Wold-style-declaration -Wold-style-definition @gol
269 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
270 -Wdeclaration-after-statement -Wpointer-sign}
272 @item Debugging Options
273 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
274 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
275 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
276 -fdump-noaddr -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
277 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
278 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
279 -fdump-statistics @gol
281 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
282 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
283 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
285 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
287 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
288 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
289 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
290 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
291 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
293 -fdump-tree-nrv -fdump-tree-vect @gol
294 -fdump-tree-sink @gol
295 -fdump-tree-sra@r{[}-@var{n}@r{]} @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 -fno-merge-debug-strings -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}.
319 -falign-functions[=@var{n}] -falign-jumps[=@var{n}] @gol
320 -falign-labels[=@var{n}] -falign-loops[=@var{n}] -fassociative-math @gol
321 -fauto-inc-dec -fbranch-probabilities -fbranch-target-load-optimize @gol
322 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fcaller-saves @gol
323 -fcheck-data-deps -fcprop-registers -fcrossjumping -fcse-follow-jumps @gol
324 -fcse-skip-blocks -fcx-fortran-rules -fcx-limited-range @gol
325 -fdata-sections -fdce -fdce @gol
326 -fdelayed-branch -fdelete-null-pointer-checks -fdse -fdse @gol
327 -fearly-inlining -fexpensive-optimizations -ffast-math @gol
328 -ffinite-math-only -ffloat-store -fforward-propagate @gol
329 -ffunction-sections -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol
330 -fgcse-sm -fif-conversion -fif-conversion2 -finline-functions @gol
331 -finline-functions-called-once -finline-limit=@var{n} @gol
332 -finline-small-functions -fipa-cp -fipa-marix-reorg -fipa-pta @gol
333 -fipa-pure-const -fipa-reference -fipa-struct-reorg @gol
334 -fipa-type-escape -fivopts -fkeep-inline-functions -fkeep-static-consts @gol
335 -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol
336 -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol
337 -fmudflapir -fmudflapth -fno-branch-count-reg -fno-default-inline @gol
338 -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
339 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
340 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
341 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
342 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
343 -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
344 -fprofile-dir=@var{path} -fprofile-generate -fprofile-generate=@var{path} @gol
345 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
346 -freciprocal-math -fregmove -frename-registers -freorder-blocks @gol
347 -freorder-blocks-and-partition -freorder-functions @gol
348 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
349 -frounding-math -frtl-abstract-sequences -fsched2-use-superblocks @gol
350 -fsched2-use-traces -fsched-spec-load -fsched-spec-load-dangerous @gol
351 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
352 -fschedule-insns -fschedule-insns2 -fsection-anchors -fsee @gol
353 -fsignaling-nans -fsingle-precision-constant -fsplit-ivs-in-unroller @gol
354 -fsplit-wide-types -fstack-protector -fstack-protector-all @gol
355 -fstrict-aliasing -fstrict-overflow -fthread-jumps -ftracer -ftree-ccp @gol
356 -ftree-ch -ftree-copy-prop -ftree-copyrename -ftree-dce @gol
357 -ftree-dominator-opts -ftree-dse -ftree-fre -ftree-loop-im @gol
358 -ftree-loop-distribution @gol
359 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
360 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-reassoc @gol
361 -ftree-sink -ftree-sra -ftree-store-ccp -ftree-ter @gol
362 -ftree-vect-loop-version -ftree-vectorize -ftree-vrp -funit-at-a-time @gol
363 -funroll-all-loops -funroll-loops -funsafe-loop-optimizations @gol
364 -funsafe-math-optimizations -funswitch-loops @gol
365 -fvariable-expansion-in-unroller -fvect-cost-model -fvpt -fweb @gol
367 --param @var{name}=@var{value}
368 -O -O0 -O1 -O2 -O3 -Os}
370 @item Preprocessor Options
371 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
372 @gccoptlist{-A@var{question}=@var{answer} @gol
373 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
374 -C -dD -dI -dM -dN @gol
375 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
376 -idirafter @var{dir} @gol
377 -include @var{file} -imacros @var{file} @gol
378 -iprefix @var{file} -iwithprefix @var{dir} @gol
379 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
380 -imultilib @var{dir} -isysroot @var{dir} @gol
381 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
382 -P -fworking-directory -remap @gol
383 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
384 -Xpreprocessor @var{option}}
386 @item Assembler Option
387 @xref{Assembler Options,,Passing Options to the Assembler}.
388 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
391 @xref{Link Options,,Options for Linking}.
392 @gccoptlist{@var{object-file-name} -l@var{library} @gol
393 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
394 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
395 -Wl,@var{option} -Xlinker @var{option} @gol
398 @item Directory Options
399 @xref{Directory Options,,Options for Directory Search}.
400 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
401 -specs=@var{file} -I- --sysroot=@var{dir}}
404 @c I wrote this xref this way to avoid overfull hbox. -- rms
405 @xref{Target Options}.
406 @gccoptlist{-V @var{version} -b @var{machine}}
408 @item Machine Dependent Options
409 @xref{Submodel Options,,Hardware Models and Configurations}.
410 @c This list is ordered alphanumerically by subsection name.
411 @c Try and put the significant identifier (CPU or system) first,
412 @c so users have a clue at guessing where the ones they want will be.
415 @gccoptlist{-EB -EL @gol
416 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
417 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
420 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
421 -mabi=@var{name} @gol
422 -mapcs-stack-check -mno-apcs-stack-check @gol
423 -mapcs-float -mno-apcs-float @gol
424 -mapcs-reentrant -mno-apcs-reentrant @gol
425 -msched-prolog -mno-sched-prolog @gol
426 -mlittle-endian -mbig-endian -mwords-little-endian @gol
427 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
428 -mthumb-interwork -mno-thumb-interwork @gol
429 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
430 -mstructure-size-boundary=@var{n} @gol
431 -mabort-on-noreturn @gol
432 -mlong-calls -mno-long-calls @gol
433 -msingle-pic-base -mno-single-pic-base @gol
434 -mpic-register=@var{reg} @gol
435 -mnop-fun-dllimport @gol
436 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
437 -mpoke-function-name @gol
439 -mtpcs-frame -mtpcs-leaf-frame @gol
440 -mcaller-super-interworking -mcallee-super-interworking @gol
444 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
445 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
447 @emph{Blackfin Options}
448 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
449 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
450 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
451 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
452 -mno-id-shared-library -mshared-library-id=@var{n} @gol
453 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
454 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
455 -mfast-fp -minline-plt}
458 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
459 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
460 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
461 -mstack-align -mdata-align -mconst-align @gol
462 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
463 -melf -maout -melinux -mlinux -sim -sim2 @gol
464 -mmul-bug-workaround -mno-mul-bug-workaround}
467 @gccoptlist{-mmac -mpush-args}
469 @emph{Darwin Options}
470 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
471 -arch_only -bind_at_load -bundle -bundle_loader @gol
472 -client_name -compatibility_version -current_version @gol
474 -dependency-file -dylib_file -dylinker_install_name @gol
475 -dynamic -dynamiclib -exported_symbols_list @gol
476 -filelist -flat_namespace -force_cpusubtype_ALL @gol
477 -force_flat_namespace -headerpad_max_install_names @gol
479 -image_base -init -install_name -keep_private_externs @gol
480 -multi_module -multiply_defined -multiply_defined_unused @gol
481 -noall_load -no_dead_strip_inits_and_terms @gol
482 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
483 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
484 -private_bundle -read_only_relocs -sectalign @gol
485 -sectobjectsymbols -whyload -seg1addr @gol
486 -sectcreate -sectobjectsymbols -sectorder @gol
487 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
488 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
489 -segprot -segs_read_only_addr -segs_read_write_addr @gol
490 -single_module -static -sub_library -sub_umbrella @gol
491 -twolevel_namespace -umbrella -undefined @gol
492 -unexported_symbols_list -weak_reference_mismatches @gol
493 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
494 -mkernel -mone-byte-bool}
496 @emph{DEC Alpha Options}
497 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
498 -mieee -mieee-with-inexact -mieee-conformant @gol
499 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
500 -mtrap-precision=@var{mode} -mbuild-constants @gol
501 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
502 -mbwx -mmax -mfix -mcix @gol
503 -mfloat-vax -mfloat-ieee @gol
504 -mexplicit-relocs -msmall-data -mlarge-data @gol
505 -msmall-text -mlarge-text @gol
506 -mmemory-latency=@var{time}}
508 @emph{DEC Alpha/VMS Options}
509 @gccoptlist{-mvms-return-codes}
512 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
513 -mhard-float -msoft-float @gol
514 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
515 -mdouble -mno-double @gol
516 -mmedia -mno-media -mmuladd -mno-muladd @gol
517 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
518 -mlinked-fp -mlong-calls -malign-labels @gol
519 -mlibrary-pic -macc-4 -macc-8 @gol
520 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
521 -moptimize-membar -mno-optimize-membar @gol
522 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
523 -mvliw-branch -mno-vliw-branch @gol
524 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
525 -mno-nested-cond-exec -mtomcat-stats @gol
529 @emph{GNU/Linux Options}
530 @gccoptlist{-muclibc}
532 @emph{H8/300 Options}
533 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
536 @gccoptlist{-march=@var{architecture-type} @gol
537 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
538 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
539 -mfixed-range=@var{register-range} @gol
540 -mjump-in-delay -mlinker-opt -mlong-calls @gol
541 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
542 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
543 -mno-jump-in-delay -mno-long-load-store @gol
544 -mno-portable-runtime -mno-soft-float @gol
545 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
546 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
547 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
548 -munix=@var{unix-std} -nolibdld -static -threads}
550 @emph{i386 and x86-64 Options}
551 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
552 -mfpmath=@var{unit} @gol
553 -masm=@var{dialect} -mno-fancy-math-387 @gol
554 -mno-fp-ret-in-387 -msoft-float @gol
555 -mno-wide-multiply -mrtd -malign-double @gol
556 -mpreferred-stack-boundary=@var{num} -mcx16 -msahf -mrecip @gol
557 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 @gol
559 -msse4a -m3dnow -mpopcnt -mabm -msse5 @gol
560 -mthreads -mno-align-stringops -minline-all-stringops @gol
561 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
562 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
563 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
564 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
565 -mcmodel=@var{code-model} @gol
566 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
567 -mfused-madd -mno-fused-madd}
570 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
571 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
572 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
573 -minline-float-divide-max-throughput @gol
574 -minline-int-divide-min-latency @gol
575 -minline-int-divide-max-throughput @gol
576 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
577 -mno-dwarf2-asm -mearly-stop-bits @gol
578 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
579 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64 @gol
580 -mno-sched-br-data-spec -msched-ar-data-spec -mno-sched-control-spec @gol
581 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
582 -msched-ldc -mno-sched-control-ldc -mno-sched-spec-verbose @gol
583 -mno-sched-prefer-non-data-spec-insns @gol
584 -mno-sched-prefer-non-control-spec-insns @gol
585 -mno-sched-count-spec-in-critical-path}
587 @emph{M32R/D Options}
588 @gccoptlist{-m32r2 -m32rx -m32r @gol
590 -malign-loops -mno-align-loops @gol
591 -missue-rate=@var{number} @gol
592 -mbranch-cost=@var{number} @gol
593 -mmodel=@var{code-size-model-type} @gol
594 -msdata=@var{sdata-type} @gol
595 -mno-flush-func -mflush-func=@var{name} @gol
596 -mno-flush-trap -mflush-trap=@var{number} @gol
600 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
602 @emph{M680x0 Options}
603 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
604 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
605 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
606 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
607 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
608 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
609 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
610 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
612 @emph{M68hc1x Options}
613 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
614 -mauto-incdec -minmax -mlong-calls -mshort @gol
615 -msoft-reg-count=@var{count}}
618 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
619 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
620 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
621 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
622 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
625 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
626 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
627 -mips16 -mno-mips16 -mflip-mips16 @gol
628 -minterlink-mips16 -mno-interlink-mips16 @gol
629 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
630 -mshared -mno-shared -mxgot -mno-xgot -mgp32 -mgp64 @gol
631 -mfp32 -mfp64 -mhard-float -msoft-float @gol
632 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
633 -msmartmips -mno-smartmips @gol
634 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
635 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
636 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
637 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
638 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
639 -membedded-data -mno-embedded-data @gol
640 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
641 -mcode-readable=@var{setting} @gol
642 -msplit-addresses -mno-split-addresses @gol
643 -mexplicit-relocs -mno-explicit-relocs @gol
644 -mcheck-zero-division -mno-check-zero-division @gol
645 -mdivide-traps -mdivide-breaks @gol
646 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
647 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
648 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
649 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 -mno-fix-vr4130 @gol
650 -mfix-sb1 -mno-fix-sb1 @gol
651 -mflush-func=@var{func} -mno-flush-func @gol
652 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
653 -mfp-exceptions -mno-fp-exceptions @gol
654 -mvr4130-align -mno-vr4130-align}
657 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
658 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
659 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
660 -mno-base-addresses -msingle-exit -mno-single-exit}
662 @emph{MN10300 Options}
663 @gccoptlist{-mmult-bug -mno-mult-bug @gol
664 -mam33 -mno-am33 @gol
665 -mam33-2 -mno-am33-2 @gol
666 -mreturn-pointer-on-d0 @gol
670 @gccoptlist{-mno-crt0 -mbacc -msim @gol
671 -march=@var{cpu-type} }
673 @emph{PDP-11 Options}
674 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
675 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
676 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
677 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
678 -mbranch-expensive -mbranch-cheap @gol
679 -msplit -mno-split -munix-asm -mdec-asm}
681 @emph{PowerPC Options}
682 See RS/6000 and PowerPC Options.
684 @emph{RS/6000 and PowerPC Options}
685 @gccoptlist{-mcpu=@var{cpu-type} @gol
686 -mtune=@var{cpu-type} @gol
687 -mpower -mno-power -mpower2 -mno-power2 @gol
688 -mpowerpc -mpowerpc64 -mno-powerpc @gol
689 -maltivec -mno-altivec @gol
690 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
691 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
692 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
693 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
694 -mnew-mnemonics -mold-mnemonics @gol
695 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
696 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
697 -malign-power -malign-natural @gol
698 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
699 -mstring -mno-string -mupdate -mno-update @gol
700 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
701 -mstrict-align -mno-strict-align -mrelocatable @gol
702 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
703 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
704 -mdynamic-no-pic -maltivec -mswdiv @gol
705 -mprioritize-restricted-insns=@var{priority} @gol
706 -msched-costly-dep=@var{dependence_type} @gol
707 -minsert-sched-nops=@var{scheme} @gol
708 -mcall-sysv -mcall-netbsd @gol
709 -maix-struct-return -msvr4-struct-return @gol
710 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
711 -misel -mno-isel @gol
712 -misel=yes -misel=no @gol
714 -mspe=yes -mspe=no @gol
716 -mvrsave -mno-vrsave @gol
717 -mmulhw -mno-mulhw @gol
718 -mdlmzb -mno-dlmzb @gol
719 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
720 -mprototype -mno-prototype @gol
721 -msim -mmvme -mads -myellowknife -memb -msdata @gol
722 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
724 @emph{S/390 and zSeries Options}
725 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
726 -mhard-float -msoft-float -mlong-double-64 -mlong-double-128 @gol
727 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
728 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
729 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
730 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
731 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
734 @gccoptlist{-meb -mel @gol
738 -mscore5 -mscore5u -mscore7 -mscore7d}
741 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
742 -m4-nofpu -m4-single-only -m4-single -m4 @gol
743 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
744 -m5-64media -m5-64media-nofpu @gol
745 -m5-32media -m5-32media-nofpu @gol
746 -m5-compact -m5-compact-nofpu @gol
747 -mb -ml -mdalign -mrelax @gol
748 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
749 -mieee -mbitops -misize -minline-ic_invalidate -mpadstruct -mspace @gol
750 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
751 -mdivsi3_libfunc=@var{name} @gol
752 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
756 @gccoptlist{-mcpu=@var{cpu-type} @gol
757 -mtune=@var{cpu-type} @gol
758 -mcmodel=@var{code-model} @gol
759 -m32 -m64 -mapp-regs -mno-app-regs @gol
760 -mfaster-structs -mno-faster-structs @gol
761 -mfpu -mno-fpu -mhard-float -msoft-float @gol
762 -mhard-quad-float -msoft-quad-float @gol
763 -mimpure-text -mno-impure-text -mlittle-endian @gol
764 -mstack-bias -mno-stack-bias @gol
765 -munaligned-doubles -mno-unaligned-doubles @gol
766 -mv8plus -mno-v8plus -mvis -mno-vis
767 -threads -pthreads -pthread}
770 @gccoptlist{-mwarn-reloc -merror-reloc @gol
771 -msafe-dma -munsafe-dma @gol
773 -msmall-mem -mlarge-mem -mstdmain @gol
774 -mfixed-range=@var{register-range}}
776 @emph{System V Options}
777 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
780 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
781 -mprolog-function -mno-prolog-function -mspace @gol
782 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
783 -mapp-regs -mno-app-regs @gol
784 -mdisable-callt -mno-disable-callt @gol
790 @gccoptlist{-mg -mgnu -munix}
792 @emph{VxWorks Options}
793 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
794 -Xbind-lazy -Xbind-now}
796 @emph{x86-64 Options}
797 See i386 and x86-64 Options.
799 @emph{Xstormy16 Options}
802 @emph{Xtensa Options}
803 @gccoptlist{-mconst16 -mno-const16 @gol
804 -mfused-madd -mno-fused-madd @gol
805 -mserialize-volatile -mno-serialize-volatile @gol
806 -mtext-section-literals -mno-text-section-literals @gol
807 -mtarget-align -mno-target-align @gol
808 -mlongcalls -mno-longcalls}
810 @emph{zSeries Options}
811 See S/390 and zSeries Options.
813 @item Code Generation Options
814 @xref{Code Gen Options,,Options for Code Generation Conventions}.
815 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
816 -ffixed-@var{reg} -fexceptions @gol
817 -fnon-call-exceptions -funwind-tables @gol
818 -fasynchronous-unwind-tables @gol
819 -finhibit-size-directive -finstrument-functions @gol
820 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
821 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
822 -fno-common -fno-ident @gol
823 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
824 -fno-jump-tables @gol
825 -frecord-gcc-switches @gol
826 -freg-struct-return -fshort-enums @gol
827 -fshort-double -fshort-wchar @gol
828 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
829 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
830 -fno-stack-limit -fargument-alias -fargument-noalias @gol
831 -fargument-noalias-global -fargument-noalias-anything @gol
832 -fleading-underscore -ftls-model=@var{model} @gol
833 -ftrapv -fwrapv -fbounds-check @gol
838 * Overall Options:: Controlling the kind of output:
839 an executable, object files, assembler files,
840 or preprocessed source.
841 * C Dialect Options:: Controlling the variant of C language compiled.
842 * C++ Dialect Options:: Variations on C++.
843 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
845 * Language Independent Options:: Controlling how diagnostics should be
847 * Warning Options:: How picky should the compiler be?
848 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
849 * Optimize Options:: How much optimization?
850 * Preprocessor Options:: Controlling header files and macro definitions.
851 Also, getting dependency information for Make.
852 * Assembler Options:: Passing options to the assembler.
853 * Link Options:: Specifying libraries and so on.
854 * Directory Options:: Where to find header files and libraries.
855 Where to find the compiler executable files.
856 * Spec Files:: How to pass switches to sub-processes.
857 * Target Options:: Running a cross-compiler, or an old version of GCC.
860 @node Overall Options
861 @section Options Controlling the Kind of Output
863 Compilation can involve up to four stages: preprocessing, compilation
864 proper, assembly and linking, always in that order. GCC is capable of
865 preprocessing and compiling several files either into several
866 assembler input files, or into one assembler input file; then each
867 assembler input file produces an object file, and linking combines all
868 the object files (those newly compiled, and those specified as input)
869 into an executable file.
871 @cindex file name suffix
872 For any given input file, the file name suffix determines what kind of
877 C source code which must be preprocessed.
880 C source code which should not be preprocessed.
883 C++ source code which should not be preprocessed.
886 Objective-C source code. Note that you must link with the @file{libobjc}
887 library to make an Objective-C program work.
890 Objective-C source code which should not be preprocessed.
894 Objective-C++ source code. Note that you must link with the @file{libobjc}
895 library to make an Objective-C++ program work. Note that @samp{.M} refers
896 to a literal capital M@.
899 Objective-C++ source code which should not be preprocessed.
902 C, C++, Objective-C or Objective-C++ header file to be turned into a
907 @itemx @var{file}.cxx
908 @itemx @var{file}.cpp
909 @itemx @var{file}.CPP
910 @itemx @var{file}.c++
912 C++ source code which must be preprocessed. Note that in @samp{.cxx},
913 the last two letters must both be literally @samp{x}. Likewise,
914 @samp{.C} refers to a literal capital C@.
918 Objective-C++ source code which must be preprocessed.
921 Objective-C++ source code which should not be preprocessed.
926 @itemx @var{file}.hxx
927 @itemx @var{file}.hpp
928 @itemx @var{file}.HPP
929 @itemx @var{file}.h++
930 @itemx @var{file}.tcc
931 C++ header file to be turned into a precompiled header.
934 @itemx @var{file}.for
935 @itemx @var{file}.ftn
936 Fixed form Fortran source code which should not be preprocessed.
939 @itemx @var{file}.FOR
940 @itemx @var{file}.fpp
941 @itemx @var{file}.FPP
942 @itemx @var{file}.FTN
943 Fixed form Fortran source code which must be preprocessed (with the traditional
947 @itemx @var{file}.f95
948 @itemx @var{file}.f03
949 @itemx @var{file}.f08
950 Free form Fortran source code which should not be preprocessed.
953 @itemx @var{file}.F95
954 @itemx @var{file}.F03
955 @itemx @var{file}.F08
956 Free form Fortran source code which must be preprocessed (with the
957 traditional preprocessor).
959 @c FIXME: Descriptions of Java file types.
966 Ada source code file which contains a library unit declaration (a
967 declaration of a package, subprogram, or generic, or a generic
968 instantiation), or a library unit renaming declaration (a package,
969 generic, or subprogram renaming declaration). Such files are also
973 Ada source code file containing a library unit body (a subprogram or
974 package body). Such files are also called @dfn{bodies}.
976 @c GCC also knows about some suffixes for languages not yet included:
988 Assembler code which must be preprocessed.
991 An object file to be fed straight into linking.
992 Any file name with no recognized suffix is treated this way.
996 You can specify the input language explicitly with the @option{-x} option:
999 @item -x @var{language}
1000 Specify explicitly the @var{language} for the following input files
1001 (rather than letting the compiler choose a default based on the file
1002 name suffix). This option applies to all following input files until
1003 the next @option{-x} option. Possible values for @var{language} are:
1005 c c-header c-cpp-output
1006 c++ c++-header c++-cpp-output
1007 objective-c objective-c-header objective-c-cpp-output
1008 objective-c++ objective-c++-header objective-c++-cpp-output
1009 assembler assembler-with-cpp
1016 Turn off any specification of a language, so that subsequent files are
1017 handled according to their file name suffixes (as they are if @option{-x}
1018 has not been used at all).
1020 @item -pass-exit-codes
1021 @opindex pass-exit-codes
1022 Normally the @command{gcc} program will exit with the code of 1 if any
1023 phase of the compiler returns a non-success return code. If you specify
1024 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1025 numerically highest error produced by any phase that returned an error
1026 indication. The C, C++, and Fortran frontends return 4, if an internal
1027 compiler error is encountered.
1030 If you only want some of the stages of compilation, you can use
1031 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1032 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1033 @command{gcc} is to stop. Note that some combinations (for example,
1034 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1039 Compile or assemble the source files, but do not link. The linking
1040 stage simply is not done. The ultimate output is in the form of an
1041 object file for each source file.
1043 By default, the object file name for a source file is made by replacing
1044 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1046 Unrecognized input files, not requiring compilation or assembly, are
1051 Stop after the stage of compilation proper; do not assemble. The output
1052 is in the form of an assembler code file for each non-assembler input
1055 By default, the assembler file name for a source file is made by
1056 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1058 Input files that don't require compilation are ignored.
1062 Stop after the preprocessing stage; do not run the compiler proper. The
1063 output is in the form of preprocessed source code, which is sent to the
1066 Input files which don't require preprocessing are ignored.
1068 @cindex output file option
1071 Place output in file @var{file}. This applies regardless to whatever
1072 sort of output is being produced, whether it be an executable file,
1073 an object file, an assembler file or preprocessed C code.
1075 If @option{-o} is not specified, the default is to put an executable
1076 file in @file{a.out}, the object file for
1077 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1078 assembler file in @file{@var{source}.s}, a precompiled header file in
1079 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1084 Print (on standard error output) the commands executed to run the stages
1085 of compilation. Also print the version number of the compiler driver
1086 program and of the preprocessor and the compiler proper.
1090 Like @option{-v} except the commands are not executed and all command
1091 arguments are quoted. This is useful for shell scripts to capture the
1092 driver-generated command lines.
1096 Use pipes rather than temporary files for communication between the
1097 various stages of compilation. This fails to work on some systems where
1098 the assembler is unable to read from a pipe; but the GNU assembler has
1103 If you are compiling multiple source files, this option tells the driver
1104 to pass all the source files to the compiler at once (for those
1105 languages for which the compiler can handle this). This will allow
1106 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1107 language for which this is supported is C@. If you pass source files for
1108 multiple languages to the driver, using this option, the driver will invoke
1109 the compiler(s) that support IMA once each, passing each compiler all the
1110 source files appropriate for it. For those languages that do not support
1111 IMA this option will be ignored, and the compiler will be invoked once for
1112 each source file in that language. If you use this option in conjunction
1113 with @option{-save-temps}, the compiler will generate multiple
1115 (one for each source file), but only one (combined) @file{.o} or
1120 Print (on the standard output) a description of the command line options
1121 understood by @command{gcc}. If the @option{-v} option is also specified
1122 then @option{--help} will also be passed on to the various processes
1123 invoked by @command{gcc}, so that they can display the command line options
1124 they accept. If the @option{-Wextra} option has also been specified
1125 (prior to the @option{--help} option), then command line options which
1126 have no documentation associated with them will also be displayed.
1129 @opindex target-help
1130 Print (on the standard output) a description of target-specific command
1131 line options for each tool. For some targets extra target-specific
1132 information may also be printed.
1134 @item --help=@var{class}@r{[},@var{qualifier}@r{]}
1135 Print (on the standard output) a description of the command line
1136 options understood by the compiler that fit into a specific class.
1137 The class can be one of @samp{optimizers}, @samp{warnings}, @samp{target},
1138 @samp{params}, or @var{language}:
1141 @item @samp{optimizers}
1142 This will display all of the optimization options supported by the
1145 @item @samp{warnings}
1146 This will display all of the options controlling warning messages
1147 produced by the compiler.
1150 This will display target-specific options. Unlike the
1151 @option{--target-help} option however, target-specific options of the
1152 linker and assembler will not be displayed. This is because those
1153 tools do not currently support the extended @option{--help=} syntax.
1156 This will display the values recognized by the @option{--param}
1159 @item @var{language}
1160 This will display the options supported for @var{language}, where
1161 @var{language} is the name of one of the languages supported in this
1165 This will display the options that are common to all languages.
1168 It is possible to further refine the output of the @option{--help=}
1169 option by adding a comma separated list of qualifiers after the
1170 class. These can be any from the following list:
1173 @item @samp{undocumented}
1174 Display only those options which are undocumented.
1177 Display options which take an argument that appears after an equal
1178 sign in the same continuous piece of text, such as:
1179 @samp{--help=target}.
1181 @item @samp{separate}
1182 Display options which take an argument that appears as a separate word
1183 following the original option, such as: @samp{-o output-file}.
1186 Thus for example to display all the undocumented target-specific
1187 switches supported by the compiler the following can be used:
1190 --help=target,undocumented
1193 The sense of a qualifier can be inverted by prefixing it with the
1194 @var{^} character, so for example to display all binary warning
1195 options (i.e., ones that are either on or off and that do not take an
1196 argument), which have a description the following can be used:
1199 --help=warnings,^joined,^undocumented
1202 A class can also be used as a qualifier, although this usually
1203 restricts the output by so much that there is nothing to display. One
1204 case where it does work however is when one of the classes is
1205 @var{target}. So for example to display all the target-specific
1206 optimization options the following can be used:
1209 --help=target,optimizers
1212 The @option{--help=} option can be repeated on the command line. Each
1213 successive use will display its requested class of options, skipping
1214 those that have already been displayed.
1216 If the @option{-Q} option appears on the command line before the
1217 @option{--help=} option, then the descriptive text displayed by
1218 @option{--help=} is changed. Instead of describing the displayed
1219 options, an indication is given as to whether the option is enabled,
1220 disabled or set to a specific value (assuming that the compiler
1221 knows this at the point where the @option{--help=} option is used).
1223 Here is a truncated example from the ARM port of @command{gcc}:
1226 % gcc -Q -mabi=2 --help=target -c
1227 The following options are target specific:
1229 -mabort-on-noreturn [disabled]
1233 The output is sensitive to the effects of previous command line
1234 options, so for example it is possible to find out which optimizations
1235 are enabled at @option{-O2} by using:
1238 -O2 --help=optimizers
1241 Alternatively you can discover which binary optimizations are enabled
1242 by @option{-O3} by using:
1245 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1246 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1247 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1252 Display the version number and copyrights of the invoked GCC@.
1256 Invoke all subcommands under a wrapper program. It takes a single
1257 comma separated list as an argument, which will be used to invoke
1261 gcc -c t.c -wrapper gdb,--args
1264 This will invoke all subprograms of gcc under "gdb --args",
1265 thus cc1 invocation will be "gdb --args cc1 ...".
1267 @include @value{srcdir}/../libiberty/at-file.texi
1271 @section Compiling C++ Programs
1273 @cindex suffixes for C++ source
1274 @cindex C++ source file suffixes
1275 C++ source files conventionally use one of the suffixes @samp{.C},
1276 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1277 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1278 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1279 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1280 files with these names and compiles them as C++ programs even if you
1281 call the compiler the same way as for compiling C programs (usually
1282 with the name @command{gcc}).
1286 However, the use of @command{gcc} does not add the C++ library.
1287 @command{g++} is a program that calls GCC and treats @samp{.c},
1288 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1289 files unless @option{-x} is used, and automatically specifies linking
1290 against the C++ library. This program is also useful when
1291 precompiling a C header file with a @samp{.h} extension for use in C++
1292 compilations. On many systems, @command{g++} is also installed with
1293 the name @command{c++}.
1295 @cindex invoking @command{g++}
1296 When you compile C++ programs, you may specify many of the same
1297 command-line options that you use for compiling programs in any
1298 language; or command-line options meaningful for C and related
1299 languages; or options that are meaningful only for C++ programs.
1300 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1301 explanations of options for languages related to C@.
1302 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1303 explanations of options that are meaningful only for C++ programs.
1305 @node C Dialect Options
1306 @section Options Controlling C Dialect
1307 @cindex dialect options
1308 @cindex language dialect options
1309 @cindex options, dialect
1311 The following options control the dialect of C (or languages derived
1312 from C, such as C++, Objective-C and Objective-C++) that the compiler
1316 @cindex ANSI support
1320 In C mode, this is equivalent to @samp{-std=c89}. In C++ mode, it is
1321 equivalent to @samp{-std=c++98}.
1323 This turns off certain features of GCC that are incompatible with ISO
1324 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1325 such as the @code{asm} and @code{typeof} keywords, and
1326 predefined macros such as @code{unix} and @code{vax} that identify the
1327 type of system you are using. It also enables the undesirable and
1328 rarely used ISO trigraph feature. For the C compiler,
1329 it disables recognition of C++ style @samp{//} comments as well as
1330 the @code{inline} keyword.
1332 The alternate keywords @code{__asm__}, @code{__extension__},
1333 @code{__inline__} and @code{__typeof__} continue to work despite
1334 @option{-ansi}. You would not want to use them in an ISO C program, of
1335 course, but it is useful to put them in header files that might be included
1336 in compilations done with @option{-ansi}. Alternate predefined macros
1337 such as @code{__unix__} and @code{__vax__} are also available, with or
1338 without @option{-ansi}.
1340 The @option{-ansi} option does not cause non-ISO programs to be
1341 rejected gratuitously. For that, @option{-pedantic} is required in
1342 addition to @option{-ansi}. @xref{Warning Options}.
1344 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1345 option is used. Some header files may notice this macro and refrain
1346 from declaring certain functions or defining certain macros that the
1347 ISO standard doesn't call for; this is to avoid interfering with any
1348 programs that might use these names for other things.
1350 Functions that would normally be built in but do not have semantics
1351 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1352 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1353 built-in functions provided by GCC}, for details of the functions
1358 Determine the language standard. @xref{Standards,,Language Standards
1359 Supported by GCC}, for details of these standard versions. This option
1360 is currently only supported when compiling C or C++.
1362 The compiler can accept several base standards, such as @samp{c89} or
1363 @samp{c++98}, and GNU dialects of those standards, such as
1364 @samp{gnu89} or @samp{gnu++98}. By specifing a base standard, the
1365 compiler will accept all programs following that standard and those
1366 using GNU extensions that do not contradict it. For example,
1367 @samp{-std=c89} turns off certain features of GCC that are
1368 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1369 keywords, but not other GNU extensions that do not have a meaning in
1370 ISO C90, such as omitting the middle term of a @code{?:}
1371 expression. On the other hand, by specifing a GNU dialect of a
1372 standard, all features the compiler support are enabled, even when
1373 those features change the meaning of the base standard and some
1374 strict-conforming programs may be rejected. The particular standard
1375 is used by @option{-pedantic} to identify which features are GNU
1376 extensions given that version of the standard. For example
1377 @samp{-std=gnu89 -pedantic} would warn about C++ style @samp{//}
1378 comments, while @samp{-std=gnu99 -pedantic} would not.
1380 A value for this option must be provided; possible values are
1385 Support all ISO C90 programs (certain GNU extensions that conflict
1386 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1388 @item iso9899:199409
1389 ISO C90 as modified in amendment 1.
1395 ISO C99. Note that this standard is not yet fully supported; see
1396 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1397 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1400 GNU dialect of ISO C90 (including some C99 features). This
1401 is the default for C code.
1405 GNU dialect of ISO C99. When ISO C99 is fully implemented in GCC,
1406 this will become the default. The name @samp{gnu9x} is deprecated.
1409 The 1998 ISO C++ standard plus amendments. Same as @option{-ansi} for
1413 GNU dialect of @option{-std=c++98}. This is the default for
1417 The working draft of the upcoming ISO C++0x standard. This option
1418 enables experimental features that are likely to be included in
1419 C++0x. The working draft is constantly changing, and any feature that is
1420 enabled by this flag may be removed from future versions of GCC if it is
1421 not part of the C++0x standard.
1424 GNU dialect of @option{-std=c++0x}. This option enables
1425 experimental features that may be removed in future versions of GCC.
1428 @item -fgnu89-inline
1429 @opindex fgnu89-inline
1430 The option @option{-fgnu89-inline} tells GCC to use the traditional
1431 GNU semantics for @code{inline} functions when in C99 mode.
1432 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1433 is accepted and ignored by GCC versions 4.1.3 up to but not including
1434 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1435 C99 mode. Using this option is roughly equivalent to adding the
1436 @code{gnu_inline} function attribute to all inline functions
1437 (@pxref{Function Attributes}).
1439 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1440 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1441 specifies the default behavior). This option was first supported in
1442 GCC 4.3. This option is not supported in C89 or gnu89 mode.
1444 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1445 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1446 in effect for @code{inline} functions. @xref{Common Predefined
1447 Macros,,,cpp,The C Preprocessor}.
1449 @item -aux-info @var{filename}
1451 Output to the given filename prototyped declarations for all functions
1452 declared and/or defined in a translation unit, including those in header
1453 files. This option is silently ignored in any language other than C@.
1455 Besides declarations, the file indicates, in comments, the origin of
1456 each declaration (source file and line), whether the declaration was
1457 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1458 @samp{O} for old, respectively, in the first character after the line
1459 number and the colon), and whether it came from a declaration or a
1460 definition (@samp{C} or @samp{F}, respectively, in the following
1461 character). In the case of function definitions, a K&R-style list of
1462 arguments followed by their declarations is also provided, inside
1463 comments, after the declaration.
1467 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1468 keyword, so that code can use these words as identifiers. You can use
1469 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1470 instead. @option{-ansi} implies @option{-fno-asm}.
1472 In C++, this switch only affects the @code{typeof} keyword, since
1473 @code{asm} and @code{inline} are standard keywords. You may want to
1474 use the @option{-fno-gnu-keywords} flag instead, which has the same
1475 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1476 switch only affects the @code{asm} and @code{typeof} keywords, since
1477 @code{inline} is a standard keyword in ISO C99.
1480 @itemx -fno-builtin-@var{function}
1481 @opindex fno-builtin
1482 @cindex built-in functions
1483 Don't recognize built-in functions that do not begin with
1484 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1485 functions provided by GCC}, for details of the functions affected,
1486 including those which are not built-in functions when @option{-ansi} or
1487 @option{-std} options for strict ISO C conformance are used because they
1488 do not have an ISO standard meaning.
1490 GCC normally generates special code to handle certain built-in functions
1491 more efficiently; for instance, calls to @code{alloca} may become single
1492 instructions that adjust the stack directly, and calls to @code{memcpy}
1493 may become inline copy loops. The resulting code is often both smaller
1494 and faster, but since the function calls no longer appear as such, you
1495 cannot set a breakpoint on those calls, nor can you change the behavior
1496 of the functions by linking with a different library. In addition,
1497 when a function is recognized as a built-in function, GCC may use
1498 information about that function to warn about problems with calls to
1499 that function, or to generate more efficient code, even if the
1500 resulting code still contains calls to that function. For example,
1501 warnings are given with @option{-Wformat} for bad calls to
1502 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1503 known not to modify global memory.
1505 With the @option{-fno-builtin-@var{function}} option
1506 only the built-in function @var{function} is
1507 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1508 function is named this is not built-in in this version of GCC, this
1509 option is ignored. There is no corresponding
1510 @option{-fbuiltin-@var{function}} option; if you wish to enable
1511 built-in functions selectively when using @option{-fno-builtin} or
1512 @option{-ffreestanding}, you may define macros such as:
1515 #define abs(n) __builtin_abs ((n))
1516 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1521 @cindex hosted environment
1523 Assert that compilation takes place in a hosted environment. This implies
1524 @option{-fbuiltin}. A hosted environment is one in which the
1525 entire standard library is available, and in which @code{main} has a return
1526 type of @code{int}. Examples are nearly everything except a kernel.
1527 This is equivalent to @option{-fno-freestanding}.
1529 @item -ffreestanding
1530 @opindex ffreestanding
1531 @cindex hosted environment
1533 Assert that compilation takes place in a freestanding environment. This
1534 implies @option{-fno-builtin}. A freestanding environment
1535 is one in which the standard library may not exist, and program startup may
1536 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1537 This is equivalent to @option{-fno-hosted}.
1539 @xref{Standards,,Language Standards Supported by GCC}, for details of
1540 freestanding and hosted environments.
1544 @cindex openmp parallel
1545 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1546 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1547 compiler generates parallel code according to the OpenMP Application
1548 Program Interface v2.5 @w{@uref{http://www.openmp.org/}}. This option
1549 implies @option{-pthread}, and thus is only supported on targets that
1550 have support for @option{-pthread}.
1552 @item -fms-extensions
1553 @opindex fms-extensions
1554 Accept some non-standard constructs used in Microsoft header files.
1556 Some cases of unnamed fields in structures and unions are only
1557 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1558 fields within structs/unions}, for details.
1562 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1563 options for strict ISO C conformance) implies @option{-trigraphs}.
1565 @item -no-integrated-cpp
1566 @opindex no-integrated-cpp
1567 Performs a compilation in two passes: preprocessing and compiling. This
1568 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1569 @option{-B} option. The user supplied compilation step can then add in
1570 an additional preprocessing step after normal preprocessing but before
1571 compiling. The default is to use the integrated cpp (internal cpp)
1573 The semantics of this option will change if "cc1", "cc1plus", and
1574 "cc1obj" are merged.
1576 @cindex traditional C language
1577 @cindex C language, traditional
1579 @itemx -traditional-cpp
1580 @opindex traditional-cpp
1581 @opindex traditional
1582 Formerly, these options caused GCC to attempt to emulate a pre-standard
1583 C compiler. They are now only supported with the @option{-E} switch.
1584 The preprocessor continues to support a pre-standard mode. See the GNU
1585 CPP manual for details.
1587 @item -fcond-mismatch
1588 @opindex fcond-mismatch
1589 Allow conditional expressions with mismatched types in the second and
1590 third arguments. The value of such an expression is void. This option
1591 is not supported for C++.
1593 @item -flax-vector-conversions
1594 @opindex flax-vector-conversions
1595 Allow implicit conversions between vectors with differing numbers of
1596 elements and/or incompatible element types. This option should not be
1599 @item -funsigned-char
1600 @opindex funsigned-char
1601 Let the type @code{char} be unsigned, like @code{unsigned char}.
1603 Each kind of machine has a default for what @code{char} should
1604 be. It is either like @code{unsigned char} by default or like
1605 @code{signed char} by default.
1607 Ideally, a portable program should always use @code{signed char} or
1608 @code{unsigned char} when it depends on the signedness of an object.
1609 But many programs have been written to use plain @code{char} and
1610 expect it to be signed, or expect it to be unsigned, depending on the
1611 machines they were written for. This option, and its inverse, let you
1612 make such a program work with the opposite default.
1614 The type @code{char} is always a distinct type from each of
1615 @code{signed char} or @code{unsigned char}, even though its behavior
1616 is always just like one of those two.
1619 @opindex fsigned-char
1620 Let the type @code{char} be signed, like @code{signed char}.
1622 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1623 the negative form of @option{-funsigned-char}. Likewise, the option
1624 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1626 @item -fsigned-bitfields
1627 @itemx -funsigned-bitfields
1628 @itemx -fno-signed-bitfields
1629 @itemx -fno-unsigned-bitfields
1630 @opindex fsigned-bitfields
1631 @opindex funsigned-bitfields
1632 @opindex fno-signed-bitfields
1633 @opindex fno-unsigned-bitfields
1634 These options control whether a bit-field is signed or unsigned, when the
1635 declaration does not use either @code{signed} or @code{unsigned}. By
1636 default, such a bit-field is signed, because this is consistent: the
1637 basic integer types such as @code{int} are signed types.
1640 @node C++ Dialect Options
1641 @section Options Controlling C++ Dialect
1643 @cindex compiler options, C++
1644 @cindex C++ options, command line
1645 @cindex options, C++
1646 This section describes the command-line options that are only meaningful
1647 for C++ programs; but you can also use most of the GNU compiler options
1648 regardless of what language your program is in. For example, you
1649 might compile a file @code{firstClass.C} like this:
1652 g++ -g -frepo -O -c firstClass.C
1656 In this example, only @option{-frepo} is an option meant
1657 only for C++ programs; you can use the other options with any
1658 language supported by GCC@.
1660 Here is a list of options that are @emph{only} for compiling C++ programs:
1664 @item -fabi-version=@var{n}
1665 @opindex fabi-version
1666 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1667 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1668 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1669 the version that conforms most closely to the C++ ABI specification.
1670 Therefore, the ABI obtained using version 0 will change as ABI bugs
1673 The default is version 2.
1675 @item -fno-access-control
1676 @opindex fno-access-control
1677 Turn off all access checking. This switch is mainly useful for working
1678 around bugs in the access control code.
1682 Check that the pointer returned by @code{operator new} is non-null
1683 before attempting to modify the storage allocated. This check is
1684 normally unnecessary because the C++ standard specifies that
1685 @code{operator new} will only return @code{0} if it is declared
1686 @samp{throw()}, in which case the compiler will always check the
1687 return value even without this option. In all other cases, when
1688 @code{operator new} has a non-empty exception specification, memory
1689 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1690 @samp{new (nothrow)}.
1692 @item -fconserve-space
1693 @opindex fconserve-space
1694 Put uninitialized or runtime-initialized global variables into the
1695 common segment, as C does. This saves space in the executable at the
1696 cost of not diagnosing duplicate definitions. If you compile with this
1697 flag and your program mysteriously crashes after @code{main()} has
1698 completed, you may have an object that is being destroyed twice because
1699 two definitions were merged.
1701 This option is no longer useful on most targets, now that support has
1702 been added for putting variables into BSS without making them common.
1704 @item -ffriend-injection
1705 @opindex ffriend-injection
1706 Inject friend functions into the enclosing namespace, so that they are
1707 visible outside the scope of the class in which they are declared.
1708 Friend functions were documented to work this way in the old Annotated
1709 C++ Reference Manual, and versions of G++ before 4.1 always worked
1710 that way. However, in ISO C++ a friend function which is not declared
1711 in an enclosing scope can only be found using argument dependent
1712 lookup. This option causes friends to be injected as they were in
1715 This option is for compatibility, and may be removed in a future
1718 @item -fno-elide-constructors
1719 @opindex fno-elide-constructors
1720 The C++ standard allows an implementation to omit creating a temporary
1721 which is only used to initialize another object of the same type.
1722 Specifying this option disables that optimization, and forces G++ to
1723 call the copy constructor in all cases.
1725 @item -fno-enforce-eh-specs
1726 @opindex fno-enforce-eh-specs
1727 Don't generate code to check for violation of exception specifications
1728 at runtime. This option violates the C++ standard, but may be useful
1729 for reducing code size in production builds, much like defining
1730 @samp{NDEBUG}. This does not give user code permission to throw
1731 exceptions in violation of the exception specifications; the compiler
1732 will still optimize based on the specifications, so throwing an
1733 unexpected exception will result in undefined behavior.
1736 @itemx -fno-for-scope
1738 @opindex fno-for-scope
1739 If @option{-ffor-scope} is specified, the scope of variables declared in
1740 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1741 as specified by the C++ standard.
1742 If @option{-fno-for-scope} is specified, the scope of variables declared in
1743 a @i{for-init-statement} extends to the end of the enclosing scope,
1744 as was the case in old versions of G++, and other (traditional)
1745 implementations of C++.
1747 The default if neither flag is given to follow the standard,
1748 but to allow and give a warning for old-style code that would
1749 otherwise be invalid, or have different behavior.
1751 @item -fno-gnu-keywords
1752 @opindex fno-gnu-keywords
1753 Do not recognize @code{typeof} as a keyword, so that code can use this
1754 word as an identifier. You can use the keyword @code{__typeof__} instead.
1755 @option{-ansi} implies @option{-fno-gnu-keywords}.
1757 @item -fno-implicit-templates
1758 @opindex fno-implicit-templates
1759 Never emit code for non-inline templates which are instantiated
1760 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1761 @xref{Template Instantiation}, for more information.
1763 @item -fno-implicit-inline-templates
1764 @opindex fno-implicit-inline-templates
1765 Don't emit code for implicit instantiations of inline templates, either.
1766 The default is to handle inlines differently so that compiles with and
1767 without optimization will need the same set of explicit instantiations.
1769 @item -fno-implement-inlines
1770 @opindex fno-implement-inlines
1771 To save space, do not emit out-of-line copies of inline functions
1772 controlled by @samp{#pragma implementation}. This will cause linker
1773 errors if these functions are not inlined everywhere they are called.
1775 @item -fms-extensions
1776 @opindex fms-extensions
1777 Disable pedantic warnings about constructs used in MFC, such as implicit
1778 int and getting a pointer to member function via non-standard syntax.
1780 @item -fno-nonansi-builtins
1781 @opindex fno-nonansi-builtins
1782 Disable built-in declarations of functions that are not mandated by
1783 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1784 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1786 @item -fno-operator-names
1787 @opindex fno-operator-names
1788 Do not treat the operator name keywords @code{and}, @code{bitand},
1789 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1790 synonyms as keywords.
1792 @item -fno-optional-diags
1793 @opindex fno-optional-diags
1794 Disable diagnostics that the standard says a compiler does not need to
1795 issue. Currently, the only such diagnostic issued by G++ is the one for
1796 a name having multiple meanings within a class.
1799 @opindex fpermissive
1800 Downgrade some diagnostics about nonconformant code from errors to
1801 warnings. Thus, using @option{-fpermissive} will allow some
1802 nonconforming code to compile.
1806 Enable automatic template instantiation at link time. This option also
1807 implies @option{-fno-implicit-templates}. @xref{Template
1808 Instantiation}, for more information.
1812 Disable generation of information about every class with virtual
1813 functions for use by the C++ runtime type identification features
1814 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1815 of the language, you can save some space by using this flag. Note that
1816 exception handling uses the same information, but it will generate it as
1817 needed. The @samp{dynamic_cast} operator can still be used for casts that
1818 do not require runtime type information, i.e.@: casts to @code{void *} or to
1819 unambiguous base classes.
1823 Emit statistics about front-end processing at the end of the compilation.
1824 This information is generally only useful to the G++ development team.
1826 @item -ftemplate-depth-@var{n}
1827 @opindex ftemplate-depth
1828 Set the maximum instantiation depth for template classes to @var{n}.
1829 A limit on the template instantiation depth is needed to detect
1830 endless recursions during template class instantiation. ANSI/ISO C++
1831 conforming programs must not rely on a maximum depth greater than 17.
1833 @item -fno-threadsafe-statics
1834 @opindex fno-threadsafe-statics
1835 Do not emit the extra code to use the routines specified in the C++
1836 ABI for thread-safe initialization of local statics. You can use this
1837 option to reduce code size slightly in code that doesn't need to be
1840 @item -fuse-cxa-atexit
1841 @opindex fuse-cxa-atexit
1842 Register destructors for objects with static storage duration with the
1843 @code{__cxa_atexit} function rather than the @code{atexit} function.
1844 This option is required for fully standards-compliant handling of static
1845 destructors, but will only work if your C library supports
1846 @code{__cxa_atexit}.
1848 @item -fno-use-cxa-get-exception-ptr
1849 @opindex fno-use-cxa-get-exception-ptr
1850 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1851 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1852 if the runtime routine is not available.
1854 @item -fvisibility-inlines-hidden
1855 @opindex fvisibility-inlines-hidden
1856 This switch declares that the user does not attempt to compare
1857 pointers to inline methods where the addresses of the two functions
1858 were taken in different shared objects.
1860 The effect of this is that GCC may, effectively, mark inline methods with
1861 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1862 appear in the export table of a DSO and do not require a PLT indirection
1863 when used within the DSO@. Enabling this option can have a dramatic effect
1864 on load and link times of a DSO as it massively reduces the size of the
1865 dynamic export table when the library makes heavy use of templates.
1867 The behavior of this switch is not quite the same as marking the
1868 methods as hidden directly, because it does not affect static variables
1869 local to the function or cause the compiler to deduce that
1870 the function is defined in only one shared object.
1872 You may mark a method as having a visibility explicitly to negate the
1873 effect of the switch for that method. For example, if you do want to
1874 compare pointers to a particular inline method, you might mark it as
1875 having default visibility. Marking the enclosing class with explicit
1876 visibility will have no effect.
1878 Explicitly instantiated inline methods are unaffected by this option
1879 as their linkage might otherwise cross a shared library boundary.
1880 @xref{Template Instantiation}.
1882 @item -fvisibility-ms-compat
1883 @opindex fvisibility-ms-compat
1884 This flag attempts to use visibility settings to make GCC's C++
1885 linkage model compatible with that of Microsoft Visual Studio.
1887 The flag makes these changes to GCC's linkage model:
1891 It sets the default visibility to @code{hidden}, like
1892 @option{-fvisibility=hidden}.
1895 Types, but not their members, are not hidden by default.
1898 The One Definition Rule is relaxed for types without explicit
1899 visibility specifications which are defined in more than one different
1900 shared object: those declarations are permitted if they would have
1901 been permitted when this option was not used.
1904 In new code it is better to use @option{-fvisibility=hidden} and
1905 export those classes which are intended to be externally visible.
1906 Unfortunately it is possible for code to rely, perhaps accidentally,
1907 on the Visual Studio behavior.
1909 Among the consequences of these changes are that static data members
1910 of the same type with the same name but defined in different shared
1911 objects will be different, so changing one will not change the other;
1912 and that pointers to function members defined in different shared
1913 objects may not compare equal. When this flag is given, it is a
1914 violation of the ODR to define types with the same name differently.
1918 Do not use weak symbol support, even if it is provided by the linker.
1919 By default, G++ will use weak symbols if they are available. This
1920 option exists only for testing, and should not be used by end-users;
1921 it will result in inferior code and has no benefits. This option may
1922 be removed in a future release of G++.
1926 Do not search for header files in the standard directories specific to
1927 C++, but do still search the other standard directories. (This option
1928 is used when building the C++ library.)
1931 In addition, these optimization, warning, and code generation options
1932 have meanings only for C++ programs:
1935 @item -fno-default-inline
1936 @opindex fno-default-inline
1937 Do not assume @samp{inline} for functions defined inside a class scope.
1938 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1939 functions will have linkage like inline functions; they just won't be
1942 @item -Wabi @r{(C++ and Objective-C++ only)}
1945 Warn when G++ generates code that is probably not compatible with the
1946 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1947 all such cases, there are probably some cases that are not warned about,
1948 even though G++ is generating incompatible code. There may also be
1949 cases where warnings are emitted even though the code that is generated
1952 You should rewrite your code to avoid these warnings if you are
1953 concerned about the fact that code generated by G++ may not be binary
1954 compatible with code generated by other compilers.
1956 The known incompatibilities at this point include:
1961 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1962 pack data into the same byte as a base class. For example:
1965 struct A @{ virtual void f(); int f1 : 1; @};
1966 struct B : public A @{ int f2 : 1; @};
1970 In this case, G++ will place @code{B::f2} into the same byte
1971 as@code{A::f1}; other compilers will not. You can avoid this problem
1972 by explicitly padding @code{A} so that its size is a multiple of the
1973 byte size on your platform; that will cause G++ and other compilers to
1974 layout @code{B} identically.
1977 Incorrect handling of tail-padding for virtual bases. G++ does not use
1978 tail padding when laying out virtual bases. For example:
1981 struct A @{ virtual void f(); char c1; @};
1982 struct B @{ B(); char c2; @};
1983 struct C : public A, public virtual B @{@};
1987 In this case, G++ will not place @code{B} into the tail-padding for
1988 @code{A}; other compilers will. You can avoid this problem by
1989 explicitly padding @code{A} so that its size is a multiple of its
1990 alignment (ignoring virtual base classes); that will cause G++ and other
1991 compilers to layout @code{C} identically.
1994 Incorrect handling of bit-fields with declared widths greater than that
1995 of their underlying types, when the bit-fields appear in a union. For
1999 union U @{ int i : 4096; @};
2003 Assuming that an @code{int} does not have 4096 bits, G++ will make the
2004 union too small by the number of bits in an @code{int}.
2007 Empty classes can be placed at incorrect offsets. For example:
2017 struct C : public B, public A @{@};
2021 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
2022 it should be placed at offset zero. G++ mistakenly believes that the
2023 @code{A} data member of @code{B} is already at offset zero.
2026 Names of template functions whose types involve @code{typename} or
2027 template template parameters can be mangled incorrectly.
2030 template <typename Q>
2031 void f(typename Q::X) @{@}
2033 template <template <typename> class Q>
2034 void f(typename Q<int>::X) @{@}
2038 Instantiations of these templates may be mangled incorrectly.
2042 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2043 @opindex Wctor-dtor-privacy
2044 @opindex Wno-ctor-dtor-privacy
2045 Warn when a class seems unusable because all the constructors or
2046 destructors in that class are private, and it has neither friends nor
2047 public static member functions.
2049 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2050 @opindex Wnon-virtual-dtor
2051 @opindex Wno-non-virtual-dtor
2052 Warn when a class has virtual functions and accessible non-virtual
2053 destructor, in which case it would be possible but unsafe to delete
2054 an instance of a derived class through a pointer to the base class.
2055 This warning is also enabled if -Weffc++ is specified.
2057 @item -Wreorder @r{(C++ and Objective-C++ only)}
2059 @opindex Wno-reorder
2060 @cindex reordering, warning
2061 @cindex warning for reordering of member initializers
2062 Warn when the order of member initializers given in the code does not
2063 match the order in which they must be executed. For instance:
2069 A(): j (0), i (1) @{ @}
2073 The compiler will rearrange the member initializers for @samp{i}
2074 and @samp{j} to match the declaration order of the members, emitting
2075 a warning to that effect. This warning is enabled by @option{-Wall}.
2078 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2081 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2084 Warn about violations of the following style guidelines from Scott Meyers'
2085 @cite{Effective C++} book:
2089 Item 11: Define a copy constructor and an assignment operator for classes
2090 with dynamically allocated memory.
2093 Item 12: Prefer initialization to assignment in constructors.
2096 Item 14: Make destructors virtual in base classes.
2099 Item 15: Have @code{operator=} return a reference to @code{*this}.
2102 Item 23: Don't try to return a reference when you must return an object.
2106 Also warn about violations of the following style guidelines from
2107 Scott Meyers' @cite{More Effective C++} book:
2111 Item 6: Distinguish between prefix and postfix forms of increment and
2112 decrement operators.
2115 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2119 When selecting this option, be aware that the standard library
2120 headers do not obey all of these guidelines; use @samp{grep -v}
2121 to filter out those warnings.
2123 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2124 @opindex Wstrict-null-sentinel
2125 @opindex Wno-strict-null-sentinel
2126 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2127 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2128 to @code{__null}. Although it is a null pointer constant not a null pointer,
2129 it is guaranteed to of the same size as a pointer. But this use is
2130 not portable across different compilers.
2132 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2133 @opindex Wno-non-template-friend
2134 @opindex Wnon-template-friend
2135 Disable warnings when non-templatized friend functions are declared
2136 within a template. Since the advent of explicit template specification
2137 support in G++, if the name of the friend is an unqualified-id (i.e.,
2138 @samp{friend foo(int)}), the C++ language specification demands that the
2139 friend declare or define an ordinary, nontemplate function. (Section
2140 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2141 could be interpreted as a particular specialization of a templatized
2142 function. Because this non-conforming behavior is no longer the default
2143 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2144 check existing code for potential trouble spots and is on by default.
2145 This new compiler behavior can be turned off with
2146 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2147 but disables the helpful warning.
2149 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2150 @opindex Wold-style-cast
2151 @opindex Wno-old-style-cast
2152 Warn if an old-style (C-style) cast to a non-void type is used within
2153 a C++ program. The new-style casts (@samp{dynamic_cast},
2154 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2155 less vulnerable to unintended effects and much easier to search for.
2157 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2158 @opindex Woverloaded-virtual
2159 @opindex Wno-overloaded-virtual
2160 @cindex overloaded virtual fn, warning
2161 @cindex warning for overloaded virtual fn
2162 Warn when a function declaration hides virtual functions from a
2163 base class. For example, in:
2170 struct B: public A @{
2175 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2183 will fail to compile.
2185 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
2186 @opindex Wno-pmf-conversions
2187 @opindex Wpmf-conversions
2188 Disable the diagnostic for converting a bound pointer to member function
2191 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
2192 @opindex Wsign-promo
2193 @opindex Wno-sign-promo
2194 Warn when overload resolution chooses a promotion from unsigned or
2195 enumerated type to a signed type, over a conversion to an unsigned type of
2196 the same size. Previous versions of G++ would try to preserve
2197 unsignedness, but the standard mandates the current behavior.
2202 A& operator = (int);
2212 In this example, G++ will synthesize a default @samp{A& operator =
2213 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2216 @node Objective-C and Objective-C++ Dialect Options
2217 @section Options Controlling Objective-C and Objective-C++ Dialects
2219 @cindex compiler options, Objective-C and Objective-C++
2220 @cindex Objective-C and Objective-C++ options, command line
2221 @cindex options, Objective-C and Objective-C++
2222 (NOTE: This manual does not describe the Objective-C and Objective-C++
2223 languages themselves. See @xref{Standards,,Language Standards
2224 Supported by GCC}, for references.)
2226 This section describes the command-line options that are only meaningful
2227 for Objective-C and Objective-C++ programs, but you can also use most of
2228 the language-independent GNU compiler options.
2229 For example, you might compile a file @code{some_class.m} like this:
2232 gcc -g -fgnu-runtime -O -c some_class.m
2236 In this example, @option{-fgnu-runtime} is an option meant only for
2237 Objective-C and Objective-C++ programs; you can use the other options with
2238 any language supported by GCC@.
2240 Note that since Objective-C is an extension of the C language, Objective-C
2241 compilations may also use options specific to the C front-end (e.g.,
2242 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2243 C++-specific options (e.g., @option{-Wabi}).
2245 Here is a list of options that are @emph{only} for compiling Objective-C
2246 and Objective-C++ programs:
2249 @item -fconstant-string-class=@var{class-name}
2250 @opindex fconstant-string-class
2251 Use @var{class-name} as the name of the class to instantiate for each
2252 literal string specified with the syntax @code{@@"@dots{}"}. The default
2253 class name is @code{NXConstantString} if the GNU runtime is being used, and
2254 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2255 @option{-fconstant-cfstrings} option, if also present, will override the
2256 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2257 to be laid out as constant CoreFoundation strings.
2260 @opindex fgnu-runtime
2261 Generate object code compatible with the standard GNU Objective-C
2262 runtime. This is the default for most types of systems.
2264 @item -fnext-runtime
2265 @opindex fnext-runtime
2266 Generate output compatible with the NeXT runtime. This is the default
2267 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2268 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2271 @item -fno-nil-receivers
2272 @opindex fno-nil-receivers
2273 Assume that all Objective-C message dispatches (e.g.,
2274 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2275 is not @code{nil}. This allows for more efficient entry points in the runtime
2276 to be used. Currently, this option is only available in conjunction with
2277 the NeXT runtime on Mac OS X 10.3 and later.
2279 @item -fobjc-call-cxx-cdtors
2280 @opindex fobjc-call-cxx-cdtors
2281 For each Objective-C class, check if any of its instance variables is a
2282 C++ object with a non-trivial default constructor. If so, synthesize a
2283 special @code{- (id) .cxx_construct} instance method that will run
2284 non-trivial default constructors on any such instance variables, in order,
2285 and then return @code{self}. Similarly, check if any instance variable
2286 is a C++ object with a non-trivial destructor, and if so, synthesize a
2287 special @code{- (void) .cxx_destruct} method that will run
2288 all such default destructors, in reverse order.
2290 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2291 thusly generated will only operate on instance variables declared in the
2292 current Objective-C class, and not those inherited from superclasses. It
2293 is the responsibility of the Objective-C runtime to invoke all such methods
2294 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2295 will be invoked by the runtime immediately after a new object
2296 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2297 be invoked immediately before the runtime deallocates an object instance.
2299 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2300 support for invoking the @code{- (id) .cxx_construct} and
2301 @code{- (void) .cxx_destruct} methods.
2303 @item -fobjc-direct-dispatch
2304 @opindex fobjc-direct-dispatch
2305 Allow fast jumps to the message dispatcher. On Darwin this is
2306 accomplished via the comm page.
2308 @item -fobjc-exceptions
2309 @opindex fobjc-exceptions
2310 Enable syntactic support for structured exception handling in Objective-C,
2311 similar to what is offered by C++ and Java. This option is
2312 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2321 @@catch (AnObjCClass *exc) @{
2328 @@catch (AnotherClass *exc) @{
2331 @@catch (id allOthers) @{
2341 The @code{@@throw} statement may appear anywhere in an Objective-C or
2342 Objective-C++ program; when used inside of a @code{@@catch} block, the
2343 @code{@@throw} may appear without an argument (as shown above), in which case
2344 the object caught by the @code{@@catch} will be rethrown.
2346 Note that only (pointers to) Objective-C objects may be thrown and
2347 caught using this scheme. When an object is thrown, it will be caught
2348 by the nearest @code{@@catch} clause capable of handling objects of that type,
2349 analogously to how @code{catch} blocks work in C++ and Java. A
2350 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2351 any and all Objective-C exceptions not caught by previous @code{@@catch}
2354 The @code{@@finally} clause, if present, will be executed upon exit from the
2355 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2356 regardless of whether any exceptions are thrown, caught or rethrown
2357 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2358 of the @code{finally} clause in Java.
2360 There are several caveats to using the new exception mechanism:
2364 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2365 idioms provided by the @code{NSException} class, the new
2366 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2367 systems, due to additional functionality needed in the (NeXT) Objective-C
2371 As mentioned above, the new exceptions do not support handling
2372 types other than Objective-C objects. Furthermore, when used from
2373 Objective-C++, the Objective-C exception model does not interoperate with C++
2374 exceptions at this time. This means you cannot @code{@@throw} an exception
2375 from Objective-C and @code{catch} it in C++, or vice versa
2376 (i.e., @code{throw @dots{} @@catch}).
2379 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2380 blocks for thread-safe execution:
2383 @@synchronized (ObjCClass *guard) @{
2388 Upon entering the @code{@@synchronized} block, a thread of execution shall
2389 first check whether a lock has been placed on the corresponding @code{guard}
2390 object by another thread. If it has, the current thread shall wait until
2391 the other thread relinquishes its lock. Once @code{guard} becomes available,
2392 the current thread will place its own lock on it, execute the code contained in
2393 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2394 making @code{guard} available to other threads).
2396 Unlike Java, Objective-C does not allow for entire methods to be marked
2397 @code{@@synchronized}. Note that throwing exceptions out of
2398 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2399 to be unlocked properly.
2403 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2405 @item -freplace-objc-classes
2406 @opindex freplace-objc-classes
2407 Emit a special marker instructing @command{ld(1)} not to statically link in
2408 the resulting object file, and allow @command{dyld(1)} to load it in at
2409 run time instead. This is used in conjunction with the Fix-and-Continue
2410 debugging mode, where the object file in question may be recompiled and
2411 dynamically reloaded in the course of program execution, without the need
2412 to restart the program itself. Currently, Fix-and-Continue functionality
2413 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2418 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2419 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2420 compile time) with static class references that get initialized at load time,
2421 which improves run-time performance. Specifying the @option{-fzero-link} flag
2422 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2423 to be retained. This is useful in Zero-Link debugging mode, since it allows
2424 for individual class implementations to be modified during program execution.
2428 Dump interface declarations for all classes seen in the source file to a
2429 file named @file{@var{sourcename}.decl}.
2431 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
2432 @opindex Wassign-intercept
2433 @opindex Wno-assign-intercept
2434 Warn whenever an Objective-C assignment is being intercepted by the
2437 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
2438 @opindex Wno-protocol
2440 If a class is declared to implement a protocol, a warning is issued for
2441 every method in the protocol that is not implemented by the class. The
2442 default behavior is to issue a warning for every method not explicitly
2443 implemented in the class, even if a method implementation is inherited
2444 from the superclass. If you use the @option{-Wno-protocol} option, then
2445 methods inherited from the superclass are considered to be implemented,
2446 and no warning is issued for them.
2448 @item -Wselector @r{(Objective-C and Objective-C++ only)}
2450 @opindex Wno-selector
2451 Warn if multiple methods of different types for the same selector are
2452 found during compilation. The check is performed on the list of methods
2453 in the final stage of compilation. Additionally, a check is performed
2454 for each selector appearing in a @code{@@selector(@dots{})}
2455 expression, and a corresponding method for that selector has been found
2456 during compilation. Because these checks scan the method table only at
2457 the end of compilation, these warnings are not produced if the final
2458 stage of compilation is not reached, for example because an error is
2459 found during compilation, or because the @option{-fsyntax-only} option is
2462 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
2463 @opindex Wstrict-selector-match
2464 @opindex Wno-strict-selector-match
2465 Warn if multiple methods with differing argument and/or return types are
2466 found for a given selector when attempting to send a message using this
2467 selector to a receiver of type @code{id} or @code{Class}. When this flag
2468 is off (which is the default behavior), the compiler will omit such warnings
2469 if any differences found are confined to types which share the same size
2472 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
2473 @opindex Wundeclared-selector
2474 @opindex Wno-undeclared-selector
2475 Warn if a @code{@@selector(@dots{})} expression referring to an
2476 undeclared selector is found. A selector is considered undeclared if no
2477 method with that name has been declared before the
2478 @code{@@selector(@dots{})} expression, either explicitly in an
2479 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2480 an @code{@@implementation} section. This option always performs its
2481 checks as soon as a @code{@@selector(@dots{})} expression is found,
2482 while @option{-Wselector} only performs its checks in the final stage of
2483 compilation. This also enforces the coding style convention
2484 that methods and selectors must be declared before being used.
2486 @item -print-objc-runtime-info
2487 @opindex print-objc-runtime-info
2488 Generate C header describing the largest structure that is passed by
2493 @node Language Independent Options
2494 @section Options to Control Diagnostic Messages Formatting
2495 @cindex options to control diagnostics formatting
2496 @cindex diagnostic messages
2497 @cindex message formatting
2499 Traditionally, diagnostic messages have been formatted irrespective of
2500 the output device's aspect (e.g.@: its width, @dots{}). The options described
2501 below can be used to control the diagnostic messages formatting
2502 algorithm, e.g.@: how many characters per line, how often source location
2503 information should be reported. Right now, only the C++ front end can
2504 honor these options. However it is expected, in the near future, that
2505 the remaining front ends would be able to digest them correctly.
2508 @item -fmessage-length=@var{n}
2509 @opindex fmessage-length
2510 Try to format error messages so that they fit on lines of about @var{n}
2511 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2512 the front ends supported by GCC@. If @var{n} is zero, then no
2513 line-wrapping will be done; each error message will appear on a single
2516 @opindex fdiagnostics-show-location
2517 @item -fdiagnostics-show-location=once
2518 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2519 reporter to emit @emph{once} source location information; that is, in
2520 case the message is too long to fit on a single physical line and has to
2521 be wrapped, the source location won't be emitted (as prefix) again,
2522 over and over, in subsequent continuation lines. This is the default
2525 @item -fdiagnostics-show-location=every-line
2526 Only meaningful in line-wrapping mode. Instructs the diagnostic
2527 messages reporter to emit the same source location information (as
2528 prefix) for physical lines that result from the process of breaking
2529 a message which is too long to fit on a single line.
2531 @item -fdiagnostics-show-option
2532 @opindex fdiagnostics-show-option
2533 This option instructs the diagnostic machinery to add text to each
2534 diagnostic emitted, which indicates which command line option directly
2535 controls that diagnostic, when such an option is known to the
2536 diagnostic machinery.
2538 @item -Wcoverage-mismatch
2539 @opindex Wcoverage-mismatch
2540 Warn if feedback profiles do not match when using the
2541 @option{-fprofile-use} option.
2542 If a source file was changed between @option{-fprofile-gen} and
2543 @option{-fprofile-use}, the files with the profile feedback can fail
2544 to match the source file and GCC can not use the profile feedback
2545 information. By default, GCC emits an error message in this case.
2546 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2547 error. GCC does not use appropriate feedback profiles, so using this
2548 option can result in poorly optimized code. This option is useful
2549 only in the case of very minor changes such as bug fixes to an
2554 @node Warning Options
2555 @section Options to Request or Suppress Warnings
2556 @cindex options to control warnings
2557 @cindex warning messages
2558 @cindex messages, warning
2559 @cindex suppressing warnings
2561 Warnings are diagnostic messages that report constructions which
2562 are not inherently erroneous but which are risky or suggest there
2563 may have been an error.
2565 The following language-independent options do not enable specific
2566 warnings but control the kinds of diagnostics produced by GCC.
2569 @cindex syntax checking
2571 @opindex fsyntax-only
2572 Check the code for syntax errors, but don't do anything beyond that.
2576 Inhibit all warning messages.
2581 Make all warnings into errors.
2586 Make the specified warning into an error. The specifier for a warning
2587 is appended, for example @option{-Werror=switch} turns the warnings
2588 controlled by @option{-Wswitch} into errors. This switch takes a
2589 negative form, to be used to negate @option{-Werror} for specific
2590 warnings, for example @option{-Wno-error=switch} makes
2591 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
2592 is in effect. You can use the @option{-fdiagnostics-show-option}
2593 option to have each controllable warning amended with the option which
2594 controls it, to determine what to use with this option.
2596 Note that specifying @option{-Werror=}@var{foo} automatically implies
2597 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
2600 @item -Wfatal-errors
2601 @opindex Wfatal-errors
2602 @opindex Wno-fatal-errors
2603 This option causes the compiler to abort compilation on the first error
2604 occurred rather than trying to keep going and printing further error
2609 You can request many specific warnings with options beginning
2610 @samp{-W}, for example @option{-Wimplicit} to request warnings on
2611 implicit declarations. Each of these specific warning options also
2612 has a negative form beginning @samp{-Wno-} to turn off warnings; for
2613 example, @option{-Wno-implicit}. This manual lists only one of the
2614 two forms, whichever is not the default. For further,
2615 language-specific options also refer to @ref{C++ Dialect Options} and
2616 @ref{Objective-C and Objective-C++ Dialect Options}.
2621 Issue all the warnings demanded by strict ISO C and ISO C++;
2622 reject all programs that use forbidden extensions, and some other
2623 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2624 version of the ISO C standard specified by any @option{-std} option used.
2626 Valid ISO C and ISO C++ programs should compile properly with or without
2627 this option (though a rare few will require @option{-ansi} or a
2628 @option{-std} option specifying the required version of ISO C)@. However,
2629 without this option, certain GNU extensions and traditional C and C++
2630 features are supported as well. With this option, they are rejected.
2632 @option{-pedantic} does not cause warning messages for use of the
2633 alternate keywords whose names begin and end with @samp{__}. Pedantic
2634 warnings are also disabled in the expression that follows
2635 @code{__extension__}. However, only system header files should use
2636 these escape routes; application programs should avoid them.
2637 @xref{Alternate Keywords}.
2639 Some users try to use @option{-pedantic} to check programs for strict ISO
2640 C conformance. They soon find that it does not do quite what they want:
2641 it finds some non-ISO practices, but not all---only those for which
2642 ISO C @emph{requires} a diagnostic, and some others for which
2643 diagnostics have been added.
2645 A feature to report any failure to conform to ISO C might be useful in
2646 some instances, but would require considerable additional work and would
2647 be quite different from @option{-pedantic}. We don't have plans to
2648 support such a feature in the near future.
2650 Where the standard specified with @option{-std} represents a GNU
2651 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2652 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2653 extended dialect is based. Warnings from @option{-pedantic} are given
2654 where they are required by the base standard. (It would not make sense
2655 for such warnings to be given only for features not in the specified GNU
2656 C dialect, since by definition the GNU dialects of C include all
2657 features the compiler supports with the given option, and there would be
2658 nothing to warn about.)
2660 @item -pedantic-errors
2661 @opindex pedantic-errors
2662 Like @option{-pedantic}, except that errors are produced rather than
2668 This enables all the warnings about constructions that some users
2669 consider questionable, and that are easy to avoid (or modify to
2670 prevent the warning), even in conjunction with macros. This also
2671 enables some language-specific warnings described in @ref{C++ Dialect
2672 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
2674 @option{-Wall} turns on the following warning flags:
2676 @gccoptlist{-Waddress @gol
2677 -Warray-bounds @r{(only with} @option{-O2}@r{)} @gol
2679 -Wchar-subscripts @gol
2681 -Wimplicit-function-declaration @gol
2684 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
2685 -Wmissing-braces @gol
2691 -Wsequence-point @gol
2692 -Wsign-compare @r{(only in C++)} @gol
2693 -Wstrict-aliasing @gol
2694 -Wstrict-overflow=1 @gol
2697 -Wuninitialized @r{(only with} @option{-O1} @r{and above)} @gol
2698 -Wunknown-pragmas @gol
2699 -Wunused-function @gol
2702 -Wunused-variable @gol
2703 -Wvolatile-register-var @gol
2706 Note that some warning flags are not implied by @option{-Wall}. Some of
2707 them warn about constructions that users generally do not consider
2708 questionable, but which occasionally you might wish to check for;
2709 others warn about constructions that are necessary or hard to avoid in
2710 some cases, and there is no simple way to modify the code to suppress
2711 the warning. Some of them are enabled by @option{-Wextra} but many of
2712 them must be enabled individually.
2718 This enables some extra warning flags that are not enabled by
2719 @option{-Wall}. (This option used to be called @option{-W}. The older
2720 name is still supported, but the newer name is more descriptive.)
2722 @gccoptlist{-Wclobbered @gol
2724 -Wignored-qualifiers @gol
2725 -Wmissing-field-initializers @gol
2726 -Wmissing-parameter-type @r{(C only)} @gol
2727 -Wold-style-declaration @r{(C only)} @gol
2728 -Woverride-init @gol
2731 -Wuninitialized @r{(only with} @option{-O1} @r{and above)} @gol
2732 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2735 The option @option{-Wextra} also prints warning messages for the
2741 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2742 @samp{>}, or @samp{>=}.
2745 (C++ only) An enumerator and a non-enumerator both appear in a
2746 conditional expression.
2749 (C++ only) A non-static reference or non-static @samp{const} member
2750 appears in a class without constructors.
2753 (C++ only) Ambiguous virtual bases.
2756 (C++ only) Subscripting an array which has been declared @samp{register}.
2759 (C++ only) Taking the address of a variable which has been declared
2763 (C++ only) A base class is not initialized in a derived class' copy
2771 Inhibit warning messages about the use of @samp{#import}.
2773 @item -Wchar-subscripts
2774 @opindex Wchar-subscripts
2775 @opindex Wno-char-subscripts
2776 Warn if an array subscript has type @code{char}. This is a common cause
2777 of error, as programmers often forget that this type is signed on some
2779 This warning is enabled by @option{-Wall}.
2783 @opindex Wno-comment
2784 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2785 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2786 This warning is enabled by @option{-Wall}.
2791 @opindex ffreestanding
2792 @opindex fno-builtin
2793 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2794 the arguments supplied have types appropriate to the format string
2795 specified, and that the conversions specified in the format string make
2796 sense. This includes standard functions, and others specified by format
2797 attributes (@pxref{Function Attributes}), in the @code{printf},
2798 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2799 not in the C standard) families (or other target-specific families).
2800 Which functions are checked without format attributes having been
2801 specified depends on the standard version selected, and such checks of
2802 functions without the attribute specified are disabled by
2803 @option{-ffreestanding} or @option{-fno-builtin}.
2805 The formats are checked against the format features supported by GNU
2806 libc version 2.2. These include all ISO C90 and C99 features, as well
2807 as features from the Single Unix Specification and some BSD and GNU
2808 extensions. Other library implementations may not support all these
2809 features; GCC does not support warning about features that go beyond a
2810 particular library's limitations. However, if @option{-pedantic} is used
2811 with @option{-Wformat}, warnings will be given about format features not
2812 in the selected standard version (but not for @code{strfmon} formats,
2813 since those are not in any version of the C standard). @xref{C Dialect
2814 Options,,Options Controlling C Dialect}.
2816 Since @option{-Wformat} also checks for null format arguments for
2817 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2819 @option{-Wformat} is included in @option{-Wall}. For more control over some
2820 aspects of format checking, the options @option{-Wformat-y2k},
2821 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2822 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2823 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2826 @opindex Wformat-y2k
2827 @opindex Wno-format-y2k
2828 If @option{-Wformat} is specified, also warn about @code{strftime}
2829 formats which may yield only a two-digit year.
2831 @item -Wno-format-contains-nul
2832 @opindex Wno-format-contains-nul
2833 @opindex Wformat-contains-nul
2834 If @option{-Wformat} is specified, do not warn about format strings that
2837 @item -Wno-format-extra-args
2838 @opindex Wno-format-extra-args
2839 @opindex Wformat-extra-args
2840 If @option{-Wformat} is specified, do not warn about excess arguments to a
2841 @code{printf} or @code{scanf} format function. The C standard specifies
2842 that such arguments are ignored.
2844 Where the unused arguments lie between used arguments that are
2845 specified with @samp{$} operand number specifications, normally
2846 warnings are still given, since the implementation could not know what
2847 type to pass to @code{va_arg} to skip the unused arguments. However,
2848 in the case of @code{scanf} formats, this option will suppress the
2849 warning if the unused arguments are all pointers, since the Single
2850 Unix Specification says that such unused arguments are allowed.
2852 @item -Wno-format-zero-length @r{(C and Objective-C only)}
2853 @opindex Wno-format-zero-length
2854 @opindex Wformat-zero-length
2855 If @option{-Wformat} is specified, do not warn about zero-length formats.
2856 The C standard specifies that zero-length formats are allowed.
2858 @item -Wformat-nonliteral
2859 @opindex Wformat-nonliteral
2860 @opindex Wno-format-nonliteral
2861 If @option{-Wformat} is specified, also warn if the format string is not a
2862 string literal and so cannot be checked, unless the format function
2863 takes its format arguments as a @code{va_list}.
2865 @item -Wformat-security
2866 @opindex Wformat-security
2867 @opindex Wno-format-security
2868 If @option{-Wformat} is specified, also warn about uses of format
2869 functions that represent possible security problems. At present, this
2870 warns about calls to @code{printf} and @code{scanf} functions where the
2871 format string is not a string literal and there are no format arguments,
2872 as in @code{printf (foo);}. This may be a security hole if the format
2873 string came from untrusted input and contains @samp{%n}. (This is
2874 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2875 in future warnings may be added to @option{-Wformat-security} that are not
2876 included in @option{-Wformat-nonliteral}.)
2880 @opindex Wno-format=2
2881 Enable @option{-Wformat} plus format checks not included in
2882 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2883 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2885 @item -Wnonnull @r{(C and Objective-C only)}
2887 @opindex Wno-nonnull
2888 Warn about passing a null pointer for arguments marked as
2889 requiring a non-null value by the @code{nonnull} function attribute.
2891 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2892 can be disabled with the @option{-Wno-nonnull} option.
2894 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2896 @opindex Wno-init-self
2897 Warn about uninitialized variables which are initialized with themselves.
2898 Note this option can only be used with the @option{-Wuninitialized} option,
2899 which in turn only works with @option{-O1} and above.
2901 For example, GCC will warn about @code{i} being uninitialized in the
2902 following snippet only when @option{-Winit-self} has been specified:
2913 @item -Wimplicit-int @r{(C and Objective-C only)}
2914 @opindex Wimplicit-int
2915 @opindex Wno-implicit-int
2916 Warn when a declaration does not specify a type.
2917 This warning is enabled by @option{-Wall}.
2919 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
2920 @opindex Wimplicit-function-declaration
2921 @opindex Wno-implicit-function-declaration
2922 Give a warning whenever a function is used before being declared. In
2923 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
2924 enabled by default and it is made into an error by
2925 @option{-pedantic-errors}. This warning is also enabled by
2930 @opindex Wno-implicit
2931 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2932 This warning is enabled by @option{-Wall}.
2934 @item -Wignored-qualifiers @r{(C and C++ only)}
2935 @opindex Wignored-qualifiers
2936 @opindex Wno-ignored-qualifiers
2937 Warn if the return type of a function has a type qualifier
2938 such as @code{const}. For ISO C such a type qualifier has no effect,
2939 since the value returned by a function is not an lvalue.
2940 For C++, the warning is only emitted for scalar types or @code{void}.
2941 ISO C prohibits qualified @code{void} return types on function
2942 definitions, so such return types always receive a warning
2943 even without this option.
2945 This warning is also enabled by @option{-Wextra}.
2950 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2951 function with external linkage, returning int, taking either zero
2952 arguments, two, or three arguments of appropriate types.
2953 This warning is enabled by @option{-Wall}.
2955 @item -Wmissing-braces
2956 @opindex Wmissing-braces
2957 @opindex Wno-missing-braces
2958 Warn if an aggregate or union initializer is not fully bracketed. In
2959 the following example, the initializer for @samp{a} is not fully
2960 bracketed, but that for @samp{b} is fully bracketed.
2963 int a[2][2] = @{ 0, 1, 2, 3 @};
2964 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2967 This warning is enabled by @option{-Wall}.
2969 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2970 @opindex Wmissing-include-dirs
2971 @opindex Wno-missing-include-dirs
2972 Warn if a user-supplied include directory does not exist.
2975 @opindex Wparentheses
2976 @opindex Wno-parentheses
2977 Warn if parentheses are omitted in certain contexts, such
2978 as when there is an assignment in a context where a truth value
2979 is expected, or when operators are nested whose precedence people
2980 often get confused about.
2982 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2983 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2984 interpretation from that of ordinary mathematical notation.
2986 Also warn about constructions where there may be confusion to which
2987 @code{if} statement an @code{else} branch belongs. Here is an example of
3002 In C/C++, every @code{else} branch belongs to the innermost possible
3003 @code{if} statement, which in this example is @code{if (b)}. This is
3004 often not what the programmer expected, as illustrated in the above
3005 example by indentation the programmer chose. When there is the
3006 potential for this confusion, GCC will issue a warning when this flag
3007 is specified. To eliminate the warning, add explicit braces around
3008 the innermost @code{if} statement so there is no way the @code{else}
3009 could belong to the enclosing @code{if}. The resulting code would
3026 This warning is enabled by @option{-Wall}.
3028 @item -Wsequence-point
3029 @opindex Wsequence-point
3030 @opindex Wno-sequence-point
3031 Warn about code that may have undefined semantics because of violations
3032 of sequence point rules in the C and C++ standards.
3034 The C and C++ standards defines the order in which expressions in a C/C++
3035 program are evaluated in terms of @dfn{sequence points}, which represent
3036 a partial ordering between the execution of parts of the program: those
3037 executed before the sequence point, and those executed after it. These
3038 occur after the evaluation of a full expression (one which is not part
3039 of a larger expression), after the evaluation of the first operand of a
3040 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
3041 function is called (but after the evaluation of its arguments and the
3042 expression denoting the called function), and in certain other places.
3043 Other than as expressed by the sequence point rules, the order of
3044 evaluation of subexpressions of an expression is not specified. All
3045 these rules describe only a partial order rather than a total order,
3046 since, for example, if two functions are called within one expression
3047 with no sequence point between them, the order in which the functions
3048 are called is not specified. However, the standards committee have
3049 ruled that function calls do not overlap.
3051 It is not specified when between sequence points modifications to the
3052 values of objects take effect. Programs whose behavior depends on this
3053 have undefined behavior; the C and C++ standards specify that ``Between
3054 the previous and next sequence point an object shall have its stored
3055 value modified at most once by the evaluation of an expression.
3056 Furthermore, the prior value shall be read only to determine the value
3057 to be stored.''. If a program breaks these rules, the results on any
3058 particular implementation are entirely unpredictable.
3060 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
3061 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
3062 diagnosed by this option, and it may give an occasional false positive
3063 result, but in general it has been found fairly effective at detecting
3064 this sort of problem in programs.
3066 The standard is worded confusingly, therefore there is some debate
3067 over the precise meaning of the sequence point rules in subtle cases.
3068 Links to discussions of the problem, including proposed formal
3069 definitions, may be found on the GCC readings page, at
3070 @w{@uref{http://gcc.gnu.org/readings.html}}.
3072 This warning is enabled by @option{-Wall} for C and C++.
3075 @opindex Wreturn-type
3076 @opindex Wno-return-type
3077 Warn whenever a function is defined with a return-type that defaults
3078 to @code{int}. Also warn about any @code{return} statement with no
3079 return-value in a function whose return-type is not @code{void}
3080 (falling off the end of the function body is considered returning
3081 without a value), and about a @code{return} statement with a
3082 expression in a function whose return-type is @code{void}.
3084 For C++, a function without return type always produces a diagnostic
3085 message, even when @option{-Wno-return-type} is specified. The only
3086 exceptions are @samp{main} and functions defined in system headers.
3088 This warning is enabled by @option{-Wall}.
3093 Warn whenever a @code{switch} statement has an index of enumerated type
3094 and lacks a @code{case} for one or more of the named codes of that
3095 enumeration. (The presence of a @code{default} label prevents this
3096 warning.) @code{case} labels outside the enumeration range also
3097 provoke warnings when this option is used.
3098 This warning is enabled by @option{-Wall}.
3100 @item -Wswitch-default
3101 @opindex Wswitch-default
3102 @opindex Wno-switch-default
3103 Warn whenever a @code{switch} statement does not have a @code{default}
3107 @opindex Wswitch-enum
3108 @opindex Wno-switch-enum
3109 Warn whenever a @code{switch} statement has an index of enumerated type
3110 and lacks a @code{case} for one or more of the named codes of that
3111 enumeration. @code{case} labels outside the enumeration range also
3112 provoke warnings when this option is used.
3116 @opindex Wno-trigraphs
3117 Warn if any trigraphs are encountered that might change the meaning of
3118 the program (trigraphs within comments are not warned about).
3119 This warning is enabled by @option{-Wall}.
3121 @item -Wunused-function
3122 @opindex Wunused-function
3123 @opindex Wno-unused-function
3124 Warn whenever a static function is declared but not defined or a
3125 non-inline static function is unused.
3126 This warning is enabled by @option{-Wall}.
3128 @item -Wunused-label
3129 @opindex Wunused-label
3130 @opindex Wno-unused-label
3131 Warn whenever a label is declared but not used.
3132 This warning is enabled by @option{-Wall}.
3134 To suppress this warning use the @samp{unused} attribute
3135 (@pxref{Variable Attributes}).
3137 @item -Wunused-parameter
3138 @opindex Wunused-parameter
3139 @opindex Wno-unused-parameter
3140 Warn whenever a function parameter is unused aside from its declaration.
3142 To suppress this warning use the @samp{unused} attribute
3143 (@pxref{Variable Attributes}).
3145 @item -Wunused-variable
3146 @opindex Wunused-variable
3147 @opindex Wno-unused-variable
3148 Warn whenever a local variable or non-constant static variable is unused
3149 aside from its declaration.
3150 This warning is enabled by @option{-Wall}.
3152 To suppress this warning use the @samp{unused} attribute
3153 (@pxref{Variable Attributes}).
3155 @item -Wunused-value
3156 @opindex Wunused-value
3157 @opindex Wno-unused-value
3158 Warn whenever a statement computes a result that is explicitly not
3159 used. To suppress this warning cast the unused expression to
3160 @samp{void}. This includes an expression-statement or the left-hand
3161 side of a comma expression that contains no side effects. For example,
3162 an expression such as @samp{x[i,j]} will cause a warning, while
3163 @samp{x[(void)i,j]} will not.
3165 This warning is enabled by @option{-Wall}.
3170 All the above @option{-Wunused} options combined.
3172 In order to get a warning about an unused function parameter, you must
3173 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
3174 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
3176 @item -Wuninitialized
3177 @opindex Wuninitialized
3178 @opindex Wno-uninitialized
3179 Warn if an automatic variable is used without first being initialized or
3180 if a variable may be clobbered by a @code{setjmp} call.
3182 These warnings are possible only in optimizing compilation,
3183 because they require data flow information that is computed only
3184 when optimizing. If you do not specify @option{-O}, you will not get
3185 these warnings. Instead, GCC will issue a warning about @option{-Wuninitialized}
3186 requiring @option{-O}.
3188 If you want to warn about code which uses the uninitialized value of the
3189 variable in its own initializer, use the @option{-Winit-self} option.
3191 These warnings occur for individual uninitialized or clobbered
3192 elements of structure, union or array variables as well as for
3193 variables which are uninitialized or clobbered as a whole. They do
3194 not occur for variables or elements declared @code{volatile}. Because
3195 these warnings depend on optimization, the exact variables or elements
3196 for which there are warnings will depend on the precise optimization
3197 options and version of GCC used.
3199 Note that there may be no warning about a variable that is used only
3200 to compute a value that itself is never used, because such
3201 computations may be deleted by data flow analysis before the warnings
3204 These warnings are made optional because GCC is not smart
3205 enough to see all the reasons why the code might be correct
3206 despite appearing to have an error. Here is one example of how
3227 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3228 always initialized, but GCC doesn't know this. Here is
3229 another common case:
3234 if (change_y) save_y = y, y = new_y;
3236 if (change_y) y = save_y;
3241 This has no bug because @code{save_y} is used only if it is set.
3243 @cindex @code{longjmp} warnings
3244 This option also warns when a non-volatile automatic variable might be
3245 changed by a call to @code{longjmp}. These warnings as well are possible
3246 only in optimizing compilation.
3248 The compiler sees only the calls to @code{setjmp}. It cannot know
3249 where @code{longjmp} will be called; in fact, a signal handler could
3250 call it at any point in the code. As a result, you may get a warning
3251 even when there is in fact no problem because @code{longjmp} cannot
3252 in fact be called at the place which would cause a problem.
3254 Some spurious warnings can be avoided if you declare all the functions
3255 you use that never return as @code{noreturn}. @xref{Function
3258 This warning is enabled by @option{-Wall} or @option{-Wextra} in
3259 optimizing compilations (@option{-O1} and above).
3261 @item -Wunknown-pragmas
3262 @opindex Wunknown-pragmas
3263 @opindex Wno-unknown-pragmas
3264 @cindex warning for unknown pragmas
3265 @cindex unknown pragmas, warning
3266 @cindex pragmas, warning of unknown
3267 Warn when a #pragma directive is encountered which is not understood by
3268 GCC@. If this command line option is used, warnings will even be issued
3269 for unknown pragmas in system header files. This is not the case if
3270 the warnings were only enabled by the @option{-Wall} command line option.
3273 @opindex Wno-pragmas
3275 Do not warn about misuses of pragmas, such as incorrect parameters,
3276 invalid syntax, or conflicts between pragmas. See also
3277 @samp{-Wunknown-pragmas}.
3279 @item -Wstrict-aliasing
3280 @opindex Wstrict-aliasing
3281 @opindex Wno-strict-aliasing
3282 This option is only active when @option{-fstrict-aliasing} is active.
3283 It warns about code which might break the strict aliasing rules that the
3284 compiler is using for optimization. The warning does not catch all
3285 cases, but does attempt to catch the more common pitfalls. It is
3286 included in @option{-Wall}.
3287 It is equivalent to @option{-Wstrict-aliasing=3}
3289 @item -Wstrict-aliasing=n
3290 @opindex Wstrict-aliasing=n
3291 @opindex Wno-strict-aliasing=n
3292 This option is only active when @option{-fstrict-aliasing} is active.
3293 It warns about code which might break the strict aliasing rules that the
3294 compiler is using for optimization.
3295 Higher levels correspond to higher accuracy (fewer false positives).
3296 Higher levels also correspond to more effort, similar to the way -O works.
3297 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3300 Level 1: Most aggressive, quick, least accurate.
3301 Possibly useful when higher levels
3302 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3303 false negatives. However, it has many false positives.
3304 Warns for all pointer conversions between possibly incompatible types,
3305 even if never dereferenced. Runs in the frontend only.
3307 Level 2: Aggressive, quick, not too precise.
3308 May still have many false positives (not as many as level 1 though),
3309 and few false negatives (but possibly more than level 1).
3310 Unlike level 1, it only warns when an address is taken. Warns about
3311 incomplete types. Runs in the frontend only.
3313 Level 3 (default for @option{-Wstrict-aliasing}):
3314 Should have very few false positives and few false
3315 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3316 Takes care of the common punn+dereference pattern in the frontend:
3317 @code{*(int*)&some_float}.
3318 If optimization is enabled, it also runs in the backend, where it deals
3319 with multiple statement cases using flow-sensitive points-to information.
3320 Only warns when the converted pointer is dereferenced.
3321 Does not warn about incomplete types.
3323 @item -Wstrict-overflow
3324 @itemx -Wstrict-overflow=@var{n}
3325 @opindex Wstrict-overflow
3326 @opindex Wno-strict-overflow
3327 This option is only active when @option{-fstrict-overflow} is active.
3328 It warns about cases where the compiler optimizes based on the
3329 assumption that signed overflow does not occur. Note that it does not
3330 warn about all cases where the code might overflow: it only warns
3331 about cases where the compiler implements some optimization. Thus
3332 this warning depends on the optimization level.
3334 An optimization which assumes that signed overflow does not occur is
3335 perfectly safe if the values of the variables involved are such that
3336 overflow never does, in fact, occur. Therefore this warning can
3337 easily give a false positive: a warning about code which is not
3338 actually a problem. To help focus on important issues, several
3339 warning levels are defined. No warnings are issued for the use of
3340 undefined signed overflow when estimating how many iterations a loop
3341 will require, in particular when determining whether a loop will be
3345 @item -Wstrict-overflow=1
3346 Warn about cases which are both questionable and easy to avoid. For
3347 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3348 compiler will simplify this to @code{1}. This level of
3349 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3350 are not, and must be explicitly requested.
3352 @item -Wstrict-overflow=2
3353 Also warn about other cases where a comparison is simplified to a
3354 constant. For example: @code{abs (x) >= 0}. This can only be
3355 simplified when @option{-fstrict-overflow} is in effect, because
3356 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3357 zero. @option{-Wstrict-overflow} (with no level) is the same as
3358 @option{-Wstrict-overflow=2}.
3360 @item -Wstrict-overflow=3
3361 Also warn about other cases where a comparison is simplified. For
3362 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3364 @item -Wstrict-overflow=4
3365 Also warn about other simplifications not covered by the above cases.
3366 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3368 @item -Wstrict-overflow=5
3369 Also warn about cases where the compiler reduces the magnitude of a
3370 constant involved in a comparison. For example: @code{x + 2 > y} will
3371 be simplified to @code{x + 1 >= y}. This is reported only at the
3372 highest warning level because this simplification applies to many
3373 comparisons, so this warning level will give a very large number of
3377 @item -Warray-bounds
3378 @opindex Wno-array-bounds
3379 @opindex Warray-bounds
3380 This option is only active when @option{-ftree-vrp} is active
3381 (default for -O2 and above). It warns about subscripts to arrays
3382 that are always out of bounds. This warning is enabled by @option{-Wall}.
3384 @item -Wno-div-by-zero
3385 @opindex Wno-div-by-zero
3386 @opindex Wdiv-by-zero
3387 Do not warn about compile-time integer division by zero. Floating point
3388 division by zero is not warned about, as it can be a legitimate way of
3389 obtaining infinities and NaNs.
3391 @item -Wsystem-headers
3392 @opindex Wsystem-headers
3393 @opindex Wno-system-headers
3394 @cindex warnings from system headers
3395 @cindex system headers, warnings from
3396 Print warning messages for constructs found in system header files.
3397 Warnings from system headers are normally suppressed, on the assumption
3398 that they usually do not indicate real problems and would only make the
3399 compiler output harder to read. Using this command line option tells
3400 GCC to emit warnings from system headers as if they occurred in user
3401 code. However, note that using @option{-Wall} in conjunction with this
3402 option will @emph{not} warn about unknown pragmas in system
3403 headers---for that, @option{-Wunknown-pragmas} must also be used.
3406 @opindex Wfloat-equal
3407 @opindex Wno-float-equal
3408 Warn if floating point values are used in equality comparisons.
3410 The idea behind this is that sometimes it is convenient (for the
3411 programmer) to consider floating-point values as approximations to
3412 infinitely precise real numbers. If you are doing this, then you need
3413 to compute (by analyzing the code, or in some other way) the maximum or
3414 likely maximum error that the computation introduces, and allow for it
3415 when performing comparisons (and when producing output, but that's a
3416 different problem). In particular, instead of testing for equality, you
3417 would check to see whether the two values have ranges that overlap; and
3418 this is done with the relational operators, so equality comparisons are
3421 @item -Wtraditional @r{(C and Objective-C only)}
3422 @opindex Wtraditional
3423 @opindex Wno-traditional
3424 Warn about certain constructs that behave differently in traditional and
3425 ISO C@. Also warn about ISO C constructs that have no traditional C
3426 equivalent, and/or problematic constructs which should be avoided.
3430 Macro parameters that appear within string literals in the macro body.
3431 In traditional C macro replacement takes place within string literals,
3432 but does not in ISO C@.
3435 In traditional C, some preprocessor directives did not exist.
3436 Traditional preprocessors would only consider a line to be a directive
3437 if the @samp{#} appeared in column 1 on the line. Therefore
3438 @option{-Wtraditional} warns about directives that traditional C
3439 understands but would ignore because the @samp{#} does not appear as the
3440 first character on the line. It also suggests you hide directives like
3441 @samp{#pragma} not understood by traditional C by indenting them. Some
3442 traditional implementations would not recognize @samp{#elif}, so it
3443 suggests avoiding it altogether.
3446 A function-like macro that appears without arguments.
3449 The unary plus operator.
3452 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3453 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3454 constants.) Note, these suffixes appear in macros defined in the system
3455 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3456 Use of these macros in user code might normally lead to spurious
3457 warnings, however GCC's integrated preprocessor has enough context to
3458 avoid warning in these cases.
3461 A function declared external in one block and then used after the end of
3465 A @code{switch} statement has an operand of type @code{long}.
3468 A non-@code{static} function declaration follows a @code{static} one.
3469 This construct is not accepted by some traditional C compilers.
3472 The ISO type of an integer constant has a different width or
3473 signedness from its traditional type. This warning is only issued if
3474 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3475 typically represent bit patterns, are not warned about.
3478 Usage of ISO string concatenation is detected.
3481 Initialization of automatic aggregates.
3484 Identifier conflicts with labels. Traditional C lacks a separate
3485 namespace for labels.
3488 Initialization of unions. If the initializer is zero, the warning is
3489 omitted. This is done under the assumption that the zero initializer in
3490 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3491 initializer warnings and relies on default initialization to zero in the
3495 Conversions by prototypes between fixed/floating point values and vice
3496 versa. The absence of these prototypes when compiling with traditional
3497 C would cause serious problems. This is a subset of the possible
3498 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3501 Use of ISO C style function definitions. This warning intentionally is
3502 @emph{not} issued for prototype declarations or variadic functions
3503 because these ISO C features will appear in your code when using
3504 libiberty's traditional C compatibility macros, @code{PARAMS} and
3505 @code{VPARAMS}. This warning is also bypassed for nested functions
3506 because that feature is already a GCC extension and thus not relevant to
3507 traditional C compatibility.
3510 @item -Wtraditional-conversion @r{(C and Objective-C only)}
3511 @opindex Wtraditional-conversion
3512 @opindex Wno-traditional-conversion
3513 Warn if a prototype causes a type conversion that is different from what
3514 would happen to the same argument in the absence of a prototype. This
3515 includes conversions of fixed point to floating and vice versa, and
3516 conversions changing the width or signedness of a fixed point argument
3517 except when the same as the default promotion.
3519 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
3520 @opindex Wdeclaration-after-statement
3521 @opindex Wno-declaration-after-statement
3522 Warn when a declaration is found after a statement in a block. This
3523 construct, known from C++, was introduced with ISO C99 and is by default
3524 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3525 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3530 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3532 @item -Wno-endif-labels
3533 @opindex Wno-endif-labels
3534 @opindex Wendif-labels
3535 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3540 Warn whenever a local variable shadows another local variable, parameter or
3541 global variable or whenever a built-in function is shadowed.
3543 @item -Wlarger-than=@var{len}
3544 @opindex Wlarger-than=@var{len}
3545 @opindex Wlarger-than-@var{len}
3546 Warn whenever an object of larger than @var{len} bytes is defined.
3548 @item -Wframe-larger-than=@var{len}
3549 @opindex Wframe-larger-than
3550 Warn if the size of a function frame is larger than @var{len} bytes.
3551 The computation done to determine the stack frame size is approximate
3552 and not conservative.
3553 The actual requirements may be somewhat greater than @var{len}
3554 even if you do not get a warning. In addition, any space allocated
3555 via @code{alloca}, variable-length arrays, or related constructs
3556 is not included by the compiler when determining
3557 whether or not to issue a warning.
3559 @item -Wunsafe-loop-optimizations
3560 @opindex Wunsafe-loop-optimizations
3561 @opindex Wno-unsafe-loop-optimizations
3562 Warn if the loop cannot be optimized because the compiler could not
3563 assume anything on the bounds of the loop indices. With
3564 @option{-funsafe-loop-optimizations} warn if the compiler made
3567 @item -Wpointer-arith
3568 @opindex Wpointer-arith
3569 @opindex Wno-pointer-arith
3570 Warn about anything that depends on the ``size of'' a function type or
3571 of @code{void}. GNU C assigns these types a size of 1, for
3572 convenience in calculations with @code{void *} pointers and pointers
3573 to functions. In C++, warn also when an arithmetic operation involves
3574 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3577 @opindex Wtype-limits
3578 @opindex Wno-type-limits
3579 Warn if a comparison is always true or always false due to the limited
3580 range of the data type, but do not warn for constant expressions. For
3581 example, warn if an unsigned variable is compared against zero with
3582 @samp{<} or @samp{>=}. This warning is also enabled by
3585 @item -Wbad-function-cast @r{(C and Objective-C only)}
3586 @opindex Wbad-function-cast
3587 @opindex Wno-bad-function-cast
3588 Warn whenever a function call is cast to a non-matching type.
3589 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3591 @item -Wc++-compat @r{(C and Objective-C only)}
3592 Warn about ISO C constructs that are outside of the common subset of
3593 ISO C and ISO C++, e.g.@: request for implicit conversion from
3594 @code{void *} to a pointer to non-@code{void} type.
3596 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3597 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3598 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3599 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3603 @opindex Wno-cast-qual
3604 Warn whenever a pointer is cast so as to remove a type qualifier from
3605 the target type. For example, warn if a @code{const char *} is cast
3606 to an ordinary @code{char *}.
3609 @opindex Wcast-align
3610 @opindex Wno-cast-align
3611 Warn whenever a pointer is cast such that the required alignment of the
3612 target is increased. For example, warn if a @code{char *} is cast to
3613 an @code{int *} on machines where integers can only be accessed at
3614 two- or four-byte boundaries.
3616 @item -Wwrite-strings
3617 @opindex Wwrite-strings
3618 @opindex Wno-write-strings
3619 When compiling C, give string constants the type @code{const
3620 char[@var{length}]} so that
3621 copying the address of one into a non-@code{const} @code{char *}
3622 pointer will get a warning; when compiling C++, warn about the
3623 deprecated conversion from string literals to @code{char *}. This
3624 warning, by default, is enabled for C++ programs.
3625 These warnings will help you find at
3626 compile time code that can try to write into a string constant, but
3627 only if you have been very careful about using @code{const} in
3628 declarations and prototypes. Otherwise, it will just be a nuisance;
3629 this is why we did not make @option{-Wall} request these warnings.
3633 @opindex Wno-clobbered
3634 Warn for variables that might be changed by @samp{longjmp} or
3635 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3638 @opindex Wconversion
3639 @opindex Wno-conversion
3640 Warn for implicit conversions that may alter a value. This includes
3641 conversions between real and integer, like @code{abs (x)} when
3642 @code{x} is @code{double}; conversions between signed and unsigned,
3643 like @code{unsigned ui = -1}; and conversions to smaller types, like
3644 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3645 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3646 changed by the conversion like in @code{abs (2.0)}. Warnings about
3647 conversions between signed and unsigned integers can be disabled by
3648 using @option{-Wno-sign-conversion}.
3650 For C++, also warn for conversions between @code{NULL} and non-pointer
3651 types; confusing overload resolution for user-defined conversions; and
3652 conversions that will never use a type conversion operator:
3653 conversions to @code{void}, the same type, a base class or a reference
3654 to them. Warnings about conversions between signed and unsigned
3655 integers are disabled by default in C++ unless
3656 @option{-Wsign-conversion} is explicitly enabled.
3659 @opindex Wempty-body
3660 @opindex Wno-empty-body
3661 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3662 while} statement. Additionally, in C++, warn when an empty body occurs
3663 in a @samp{while} or @samp{for} statement with no whitespacing before
3664 the semicolon. This warning is also enabled by @option{-Wextra}.
3666 @item -Wsign-compare
3667 @opindex Wsign-compare
3668 @opindex Wno-sign-compare
3669 @cindex warning for comparison of signed and unsigned values
3670 @cindex comparison of signed and unsigned values, warning
3671 @cindex signed and unsigned values, comparison warning
3672 Warn when a comparison between signed and unsigned values could produce
3673 an incorrect result when the signed value is converted to unsigned.
3674 This warning is also enabled by @option{-Wextra}; to get the other warnings
3675 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3677 @item -Wsign-conversion
3678 @opindex Wsign-conversion
3679 @opindex Wno-sign-conversion
3680 Warn for implicit conversions that may change the sign of an integer
3681 value, like assigning a signed integer expression to an unsigned
3682 integer variable. An explicit cast silences the warning. In C, this
3683 option is enabled also by @option{-Wconversion}.
3687 @opindex Wno-address
3688 Warn about suspicious uses of memory addresses. These include using
3689 the address of a function in a conditional expression, such as
3690 @code{void func(void); if (func)}, and comparisons against the memory
3691 address of a string literal, such as @code{if (x == "abc")}. Such
3692 uses typically indicate a programmer error: the address of a function
3693 always evaluates to true, so their use in a conditional usually
3694 indicate that the programmer forgot the parentheses in a function
3695 call; and comparisons against string literals result in unspecified
3696 behavior and are not portable in C, so they usually indicate that the
3697 programmer intended to use @code{strcmp}. This warning is enabled by
3701 @opindex Wlogical-op
3702 @opindex Wno-logical-op
3703 Warn about suspicious uses of logical operators in expressions.
3704 This includes using logical operators in contexts where a
3705 bit-wise operator is likely to be expected.
3707 @item -Waggregate-return
3708 @opindex Waggregate-return
3709 @opindex Wno-aggregate-return
3710 Warn if any functions that return structures or unions are defined or
3711 called. (In languages where you can return an array, this also elicits
3714 @item -Wno-attributes
3715 @opindex Wno-attributes
3716 @opindex Wattributes
3717 Do not warn if an unexpected @code{__attribute__} is used, such as
3718 unrecognized attributes, function attributes applied to variables,
3719 etc. This will not stop errors for incorrect use of supported
3722 @item -Wstrict-prototypes @r{(C and Objective-C only)}
3723 @opindex Wstrict-prototypes
3724 @opindex Wno-strict-prototypes
3725 Warn if a function is declared or defined without specifying the
3726 argument types. (An old-style function definition is permitted without
3727 a warning if preceded by a declaration which specifies the argument
3730 @item -Wold-style-declaration @r{(C and Objective-C only)}
3731 @opindex Wold-style-declaration
3732 @opindex Wno-old-style-declaration
3733 Warn for obsolescent usages, according to the C Standard, in a
3734 declaration. For example, warn if storage-class specifiers like
3735 @code{static} are not the first things in a declaration. This warning
3736 is also enabled by @option{-Wextra}.
3738 @item -Wold-style-definition @r{(C and Objective-C only)}
3739 @opindex Wold-style-definition
3740 @opindex Wno-old-style-definition
3741 Warn if an old-style function definition is used. A warning is given
3742 even if there is a previous prototype.
3744 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
3745 @opindex Wmissing-parameter-type
3746 @opindex Wno-missing-parameter-type
3747 A function parameter is declared without a type specifier in K&R-style
3754 This warning is also enabled by @option{-Wextra}.
3756 @item -Wmissing-prototypes @r{(C and Objective-C only)}
3757 @opindex Wmissing-prototypes
3758 @opindex Wno-missing-prototypes
3759 Warn if a global function is defined without a previous prototype
3760 declaration. This warning is issued even if the definition itself
3761 provides a prototype. The aim is to detect global functions that fail
3762 to be declared in header files.
3764 @item -Wmissing-declarations
3765 @opindex Wmissing-declarations
3766 @opindex Wno-missing-declarations
3767 Warn if a global function is defined without a previous declaration.
3768 Do so even if the definition itself provides a prototype.
3769 Use this option to detect global functions that are not declared in
3770 header files. In C++, no warnings are issued for function templates,
3771 or for inline functions, or for functions in anonymous namespaces.
3773 @item -Wmissing-field-initializers
3774 @opindex Wmissing-field-initializers
3775 @opindex Wno-missing-field-initializers
3779 Warn if a structure's initializer has some fields missing. For
3780 example, the following code would cause such a warning, because
3781 @code{x.h} is implicitly zero:
3784 struct s @{ int f, g, h; @};
3785 struct s x = @{ 3, 4 @};
3788 This option does not warn about designated initializers, so the following
3789 modification would not trigger a warning:
3792 struct s @{ int f, g, h; @};
3793 struct s x = @{ .f = 3, .g = 4 @};
3796 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3797 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3799 @item -Wmissing-noreturn
3800 @opindex Wmissing-noreturn
3801 @opindex Wno-missing-noreturn
3802 Warn about functions which might be candidates for attribute @code{noreturn}.
3803 Note these are only possible candidates, not absolute ones. Care should
3804 be taken to manually verify functions actually do not ever return before
3805 adding the @code{noreturn} attribute, otherwise subtle code generation
3806 bugs could be introduced. You will not get a warning for @code{main} in
3807 hosted C environments.
3809 @item -Wmissing-format-attribute
3810 @opindex Wmissing-format-attribute
3811 @opindex Wno-missing-format-attribute
3814 Warn about function pointers which might be candidates for @code{format}
3815 attributes. Note these are only possible candidates, not absolute ones.
3816 GCC will guess that function pointers with @code{format} attributes that
3817 are used in assignment, initialization, parameter passing or return
3818 statements should have a corresponding @code{format} attribute in the
3819 resulting type. I.e.@: the left-hand side of the assignment or
3820 initialization, the type of the parameter variable, or the return type
3821 of the containing function respectively should also have a @code{format}
3822 attribute to avoid the warning.
3824 GCC will also warn about function definitions which might be
3825 candidates for @code{format} attributes. Again, these are only
3826 possible candidates. GCC will guess that @code{format} attributes
3827 might be appropriate for any function that calls a function like
3828 @code{vprintf} or @code{vscanf}, but this might not always be the
3829 case, and some functions for which @code{format} attributes are
3830 appropriate may not be detected.
3832 @item -Wno-multichar
3833 @opindex Wno-multichar
3835 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3836 Usually they indicate a typo in the user's code, as they have
3837 implementation-defined values, and should not be used in portable code.
3839 @item -Wnormalized=<none|id|nfc|nfkc>
3840 @opindex Wnormalized=
3843 @cindex character set, input normalization
3844 In ISO C and ISO C++, two identifiers are different if they are
3845 different sequences of characters. However, sometimes when characters
3846 outside the basic ASCII character set are used, you can have two
3847 different character sequences that look the same. To avoid confusion,
3848 the ISO 10646 standard sets out some @dfn{normalization rules} which
3849 when applied ensure that two sequences that look the same are turned into
3850 the same sequence. GCC can warn you if you are using identifiers which
3851 have not been normalized; this option controls that warning.
3853 There are four levels of warning that GCC supports. The default is
3854 @option{-Wnormalized=nfc}, which warns about any identifier which is
3855 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3856 recommended form for most uses.
3858 Unfortunately, there are some characters which ISO C and ISO C++ allow
3859 in identifiers that when turned into NFC aren't allowable as
3860 identifiers. That is, there's no way to use these symbols in portable
3861 ISO C or C++ and have all your identifiers in NFC@.
3862 @option{-Wnormalized=id} suppresses the warning for these characters.
3863 It is hoped that future versions of the standards involved will correct
3864 this, which is why this option is not the default.
3866 You can switch the warning off for all characters by writing
3867 @option{-Wnormalized=none}. You would only want to do this if you
3868 were using some other normalization scheme (like ``D''), because
3869 otherwise you can easily create bugs that are literally impossible to see.
3871 Some characters in ISO 10646 have distinct meanings but look identical
3872 in some fonts or display methodologies, especially once formatting has
3873 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3874 LETTER N'', will display just like a regular @code{n} which has been
3875 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3876 normalization scheme to convert all these into a standard form as
3877 well, and GCC will warn if your code is not in NFKC if you use
3878 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3879 about every identifier that contains the letter O because it might be
3880 confused with the digit 0, and so is not the default, but may be
3881 useful as a local coding convention if the programming environment is
3882 unable to be fixed to display these characters distinctly.
3884 @item -Wno-deprecated
3885 @opindex Wno-deprecated
3886 @opindex Wdeprecated
3887 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
3889 @item -Wno-deprecated-declarations
3890 @opindex Wno-deprecated-declarations
3891 @opindex Wdeprecated-declarations
3892 Do not warn about uses of functions (@pxref{Function Attributes}),
3893 variables (@pxref{Variable Attributes}), and types (@pxref{Type
3894 Attributes}) marked as deprecated by using the @code{deprecated}
3898 @opindex Wno-overflow
3900 Do not warn about compile-time overflow in constant expressions.
3902 @item -Woverride-init @r{(C and Objective-C only)}
3903 @opindex Woverride-init
3904 @opindex Wno-override-init
3908 Warn if an initialized field without side effects is overridden when
3909 using designated initializers (@pxref{Designated Inits, , Designated
3912 This warning is included in @option{-Wextra}. To get other
3913 @option{-Wextra} warnings without this one, use @samp{-Wextra
3914 -Wno-override-init}.
3919 Warn if a structure is given the packed attribute, but the packed
3920 attribute has no effect on the layout or size of the structure.
3921 Such structures may be mis-aligned for little benefit. For
3922 instance, in this code, the variable @code{f.x} in @code{struct bar}
3923 will be misaligned even though @code{struct bar} does not itself
3924 have the packed attribute:
3931 @} __attribute__((packed));
3942 Warn if padding is included in a structure, either to align an element
3943 of the structure or to align the whole structure. Sometimes when this
3944 happens it is possible to rearrange the fields of the structure to
3945 reduce the padding and so make the structure smaller.
3947 @item -Wredundant-decls
3948 @opindex Wredundant-decls
3949 @opindex Wno-redundant-decls
3950 Warn if anything is declared more than once in the same scope, even in
3951 cases where multiple declaration is valid and changes nothing.
3953 @item -Wnested-externs @r{(C and Objective-C only)}
3954 @opindex Wnested-externs
3955 @opindex Wno-nested-externs
3956 Warn if an @code{extern} declaration is encountered within a function.
3958 @item -Wunreachable-code
3959 @opindex Wunreachable-code
3960 @opindex Wno-unreachable-code
3961 Warn if the compiler detects that code will never be executed.
3963 This option is intended to warn when the compiler detects that at
3964 least a whole line of source code will never be executed, because
3965 some condition is never satisfied or because it is after a
3966 procedure that never returns.
3968 It is possible for this option to produce a warning even though there
3969 are circumstances under which part of the affected line can be executed,
3970 so care should be taken when removing apparently-unreachable code.
3972 For instance, when a function is inlined, a warning may mean that the
3973 line is unreachable in only one inlined copy of the function.
3975 This option is not made part of @option{-Wall} because in a debugging
3976 version of a program there is often substantial code which checks
3977 correct functioning of the program and is, hopefully, unreachable
3978 because the program does work. Another common use of unreachable
3979 code is to provide behavior which is selectable at compile-time.
3984 Warn if a function can not be inlined and it was declared as inline.
3985 Even with this option, the compiler will not warn about failures to
3986 inline functions declared in system headers.
3988 The compiler uses a variety of heuristics to determine whether or not
3989 to inline a function. For example, the compiler takes into account
3990 the size of the function being inlined and the amount of inlining
3991 that has already been done in the current function. Therefore,
3992 seemingly insignificant changes in the source program can cause the
3993 warnings produced by @option{-Winline} to appear or disappear.
3995 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
3996 @opindex Wno-invalid-offsetof
3997 @opindex Winvalid-offsetof
3998 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3999 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
4000 to a non-POD type is undefined. In existing C++ implementations,
4001 however, @samp{offsetof} typically gives meaningful results even when
4002 applied to certain kinds of non-POD types. (Such as a simple
4003 @samp{struct} that fails to be a POD type only by virtue of having a
4004 constructor.) This flag is for users who are aware that they are
4005 writing nonportable code and who have deliberately chosen to ignore the
4008 The restrictions on @samp{offsetof} may be relaxed in a future version
4009 of the C++ standard.
4011 @item -Wno-int-to-pointer-cast @r{(C and Objective-C only)}
4012 @opindex Wno-int-to-pointer-cast
4013 @opindex Wint-to-pointer-cast
4014 Suppress warnings from casts to pointer type of an integer of a
4017 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
4018 @opindex Wno-pointer-to-int-cast
4019 @opindex Wpointer-to-int-cast
4020 Suppress warnings from casts from a pointer to an integer type of a
4024 @opindex Winvalid-pch
4025 @opindex Wno-invalid-pch
4026 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
4027 the search path but can't be used.
4031 @opindex Wno-long-long
4032 Warn if @samp{long long} type is used. This is default. To inhibit
4033 the warning messages, use @option{-Wno-long-long}. Flags
4034 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
4035 only when @option{-pedantic} flag is used.
4037 @item -Wvariadic-macros
4038 @opindex Wvariadic-macros
4039 @opindex Wno-variadic-macros
4040 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
4041 alternate syntax when in pedantic ISO C99 mode. This is default.
4042 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
4047 Warn if variable length array is used in the code.
4048 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
4049 the variable length array.
4051 @item -Wvolatile-register-var
4052 @opindex Wvolatile-register-var
4053 @opindex Wno-volatile-register-var
4054 Warn if a register variable is declared volatile. The volatile
4055 modifier does not inhibit all optimizations that may eliminate reads
4056 and/or writes to register variables. This warning is enabled by
4059 @item -Wdisabled-optimization
4060 @opindex Wdisabled-optimization
4061 @opindex Wno-disabled-optimization
4062 Warn if a requested optimization pass is disabled. This warning does
4063 not generally indicate that there is anything wrong with your code; it
4064 merely indicates that GCC's optimizers were unable to handle the code
4065 effectively. Often, the problem is that your code is too big or too
4066 complex; GCC will refuse to optimize programs when the optimization
4067 itself is likely to take inordinate amounts of time.
4069 @item -Wpointer-sign @r{(C and Objective-C only)}
4070 @opindex Wpointer-sign
4071 @opindex Wno-pointer-sign
4072 Warn for pointer argument passing or assignment with different signedness.
4073 This option is only supported for C and Objective-C@. It is implied by
4074 @option{-Wall} and by @option{-pedantic}, which can be disabled with
4075 @option{-Wno-pointer-sign}.
4077 @item -Wstack-protector
4078 @opindex Wstack-protector
4079 @opindex Wno-stack-protector
4080 This option is only active when @option{-fstack-protector} is active. It
4081 warns about functions that will not be protected against stack smashing.
4084 @opindex Wno-mudflap
4085 Suppress warnings about constructs that cannot be instrumented by
4088 @item -Woverlength-strings
4089 @opindex Woverlength-strings
4090 @opindex Wno-overlength-strings
4091 Warn about string constants which are longer than the ``minimum
4092 maximum'' length specified in the C standard. Modern compilers
4093 generally allow string constants which are much longer than the
4094 standard's minimum limit, but very portable programs should avoid
4095 using longer strings.
4097 The limit applies @emph{after} string constant concatenation, and does
4098 not count the trailing NUL@. In C89, the limit was 509 characters; in
4099 C99, it was raised to 4095. C++98 does not specify a normative
4100 minimum maximum, so we do not diagnose overlength strings in C++@.
4102 This option is implied by @option{-pedantic}, and can be disabled with
4103 @option{-Wno-overlength-strings}.
4106 @node Debugging Options
4107 @section Options for Debugging Your Program or GCC
4108 @cindex options, debugging
4109 @cindex debugging information options
4111 GCC has various special options that are used for debugging
4112 either your program or GCC:
4117 Produce debugging information in the operating system's native format
4118 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
4121 On most systems that use stabs format, @option{-g} enables use of extra
4122 debugging information that only GDB can use; this extra information
4123 makes debugging work better in GDB but will probably make other debuggers
4125 refuse to read the program. If you want to control for certain whether
4126 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
4127 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
4129 GCC allows you to use @option{-g} with
4130 @option{-O}. The shortcuts taken by optimized code may occasionally
4131 produce surprising results: some variables you declared may not exist
4132 at all; flow of control may briefly move where you did not expect it;
4133 some statements may not be executed because they compute constant
4134 results or their values were already at hand; some statements may
4135 execute in different places because they were moved out of loops.
4137 Nevertheless it proves possible to debug optimized output. This makes
4138 it reasonable to use the optimizer for programs that might have bugs.
4140 The following options are useful when GCC is generated with the
4141 capability for more than one debugging format.
4145 Produce debugging information for use by GDB@. This means to use the
4146 most expressive format available (DWARF 2, stabs, or the native format
4147 if neither of those are supported), including GDB extensions if at all
4152 Produce debugging information in stabs format (if that is supported),
4153 without GDB extensions. This is the format used by DBX on most BSD
4154 systems. On MIPS, Alpha and System V Release 4 systems this option
4155 produces stabs debugging output which is not understood by DBX or SDB@.
4156 On System V Release 4 systems this option requires the GNU assembler.
4158 @item -feliminate-unused-debug-symbols
4159 @opindex feliminate-unused-debug-symbols
4160 Produce debugging information in stabs format (if that is supported),
4161 for only symbols that are actually used.
4163 @item -femit-class-debug-always
4164 Instead of emitting debugging information for a C++ class in only one
4165 object file, emit it in all object files using the class. This option
4166 should be used only with debuggers that are unable to handle the way GCC
4167 normally emits debugging information for classes because using this
4168 option will increase the size of debugging information by as much as a
4173 Produce debugging information in stabs format (if that is supported),
4174 using GNU extensions understood only by the GNU debugger (GDB)@. The
4175 use of these extensions is likely to make other debuggers crash or
4176 refuse to read the program.
4180 Produce debugging information in COFF format (if that is supported).
4181 This is the format used by SDB on most System V systems prior to
4186 Produce debugging information in XCOFF format (if that is supported).
4187 This is the format used by the DBX debugger on IBM RS/6000 systems.
4191 Produce debugging information in XCOFF format (if that is supported),
4192 using GNU extensions understood only by the GNU debugger (GDB)@. The
4193 use of these extensions is likely to make other debuggers crash or
4194 refuse to read the program, and may cause assemblers other than the GNU
4195 assembler (GAS) to fail with an error.
4199 Produce debugging information in DWARF version 2 format (if that is
4200 supported). This is the format used by DBX on IRIX 6. With this
4201 option, GCC uses features of DWARF version 3 when they are useful;
4202 version 3 is upward compatible with version 2, but may still cause
4203 problems for older debuggers.
4207 Produce debugging information in VMS debug format (if that is
4208 supported). This is the format used by DEBUG on VMS systems.
4211 @itemx -ggdb@var{level}
4212 @itemx -gstabs@var{level}
4213 @itemx -gcoff@var{level}
4214 @itemx -gxcoff@var{level}
4215 @itemx -gvms@var{level}
4216 Request debugging information and also use @var{level} to specify how
4217 much information. The default level is 2.
4219 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4222 Level 1 produces minimal information, enough for making backtraces in
4223 parts of the program that you don't plan to debug. This includes
4224 descriptions of functions and external variables, but no information
4225 about local variables and no line numbers.
4227 Level 3 includes extra information, such as all the macro definitions
4228 present in the program. Some debuggers support macro expansion when
4229 you use @option{-g3}.
4231 @option{-gdwarf-2} does not accept a concatenated debug level, because
4232 GCC used to support an option @option{-gdwarf} that meant to generate
4233 debug information in version 1 of the DWARF format (which is very
4234 different from version 2), and it would have been too confusing. That
4235 debug format is long obsolete, but the option cannot be changed now.
4236 Instead use an additional @option{-g@var{level}} option to change the
4237 debug level for DWARF2.
4239 @item -feliminate-dwarf2-dups
4240 @opindex feliminate-dwarf2-dups
4241 Compress DWARF2 debugging information by eliminating duplicated
4242 information about each symbol. This option only makes sense when
4243 generating DWARF2 debugging information with @option{-gdwarf-2}.
4245 @item -femit-struct-debug-baseonly
4246 Emit debug information for struct-like types
4247 only when the base name of the compilation source file
4248 matches the base name of file in which the struct was defined.
4250 This option substantially reduces the size of debugging information,
4251 but at significant potential loss in type information to the debugger.
4252 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4253 See @option{-femit-struct-debug-detailed} for more detailed control.
4255 This option works only with DWARF 2.
4257 @item -femit-struct-debug-reduced
4258 Emit debug information for struct-like types
4259 only when the base name of the compilation source file
4260 matches the base name of file in which the type was defined,
4261 unless the struct is a template or defined in a system header.
4263 This option significantly reduces the size of debugging information,
4264 with some potential loss in type information to the debugger.
4265 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4266 See @option{-femit-struct-debug-detailed} for more detailed control.
4268 This option works only with DWARF 2.
4270 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4271 Specify the struct-like types
4272 for which the compiler will generate debug information.
4273 The intent is to reduce duplicate struct debug information
4274 between different object files within the same program.
4276 This option is a detailed version of
4277 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4278 which will serve for most needs.
4280 A specification has the syntax
4281 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4283 The optional first word limits the specification to
4284 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4285 A struct type is used directly when it is the type of a variable, member.
4286 Indirect uses arise through pointers to structs.
4287 That is, when use of an incomplete struct would be legal, the use is indirect.
4289 @samp{struct one direct; struct two * indirect;}.
4291 The optional second word limits the specification to
4292 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4293 Generic structs are a bit complicated to explain.
4294 For C++, these are non-explicit specializations of template classes,
4295 or non-template classes within the above.
4296 Other programming languages have generics,
4297 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4299 The third word specifies the source files for those
4300 structs for which the compiler will emit debug information.
4301 The values @samp{none} and @samp{any} have the normal meaning.
4302 The value @samp{base} means that
4303 the base of name of the file in which the type declaration appears
4304 must match the base of the name of the main compilation file.
4305 In practice, this means that
4306 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4307 but types declared in other header will not.
4308 The value @samp{sys} means those types satisfying @samp{base}
4309 or declared in system or compiler headers.
4311 You may need to experiment to determine the best settings for your application.
4313 The default is @samp{-femit-struct-debug-detailed=all}.
4315 This option works only with DWARF 2.
4317 @item -fno-merge-debug-strings
4318 @opindex fmerge-debug-strings
4319 @opindex fno-merge-debug-strings
4320 Direct the linker to merge together strings which are identical in
4321 different object files. This is not supported by all assemblers or
4322 linker. This decreases the size of the debug information in the
4323 output file at the cost of increasing link processing time. This is
4326 @item -fdebug-prefix-map=@var{old}=@var{new}
4327 @opindex fdebug-prefix-map
4328 When compiling files in directory @file{@var{old}}, record debugging
4329 information describing them as in @file{@var{new}} instead.
4331 @cindex @command{prof}
4334 Generate extra code to write profile information suitable for the
4335 analysis program @command{prof}. You must use this option when compiling
4336 the source files you want data about, and you must also use it when
4339 @cindex @command{gprof}
4342 Generate extra code to write profile information suitable for the
4343 analysis program @command{gprof}. You must use this option when compiling
4344 the source files you want data about, and you must also use it when
4349 Makes the compiler print out each function name as it is compiled, and
4350 print some statistics about each pass when it finishes.
4353 @opindex ftime-report
4354 Makes the compiler print some statistics about the time consumed by each
4355 pass when it finishes.
4358 @opindex fmem-report
4359 Makes the compiler print some statistics about permanent memory
4360 allocation when it finishes.
4362 @item -fpre-ipa-mem-report
4363 @opindex fpre-ipa-mem-report
4364 @item -fpost-ipa-mem-report
4365 @opindex fpost-ipa-mem-report
4366 Makes the compiler print some statistics about permanent memory
4367 allocation before or after interprocedural optimization.
4369 @item -fprofile-arcs
4370 @opindex fprofile-arcs
4371 Add code so that program flow @dfn{arcs} are instrumented. During
4372 execution the program records how many times each branch and call is
4373 executed and how many times it is taken or returns. When the compiled
4374 program exits it saves this data to a file called
4375 @file{@var{auxname}.gcda} for each source file. The data may be used for
4376 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4377 test coverage analysis (@option{-ftest-coverage}). Each object file's
4378 @var{auxname} is generated from the name of the output file, if
4379 explicitly specified and it is not the final executable, otherwise it is
4380 the basename of the source file. In both cases any suffix is removed
4381 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4382 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4383 @xref{Cross-profiling}.
4385 @cindex @command{gcov}
4389 This option is used to compile and link code instrumented for coverage
4390 analysis. The option is a synonym for @option{-fprofile-arcs}
4391 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4392 linking). See the documentation for those options for more details.
4397 Compile the source files with @option{-fprofile-arcs} plus optimization
4398 and code generation options. For test coverage analysis, use the
4399 additional @option{-ftest-coverage} option. You do not need to profile
4400 every source file in a program.
4403 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4404 (the latter implies the former).
4407 Run the program on a representative workload to generate the arc profile
4408 information. This may be repeated any number of times. You can run
4409 concurrent instances of your program, and provided that the file system
4410 supports locking, the data files will be correctly updated. Also
4411 @code{fork} calls are detected and correctly handled (double counting
4415 For profile-directed optimizations, compile the source files again with
4416 the same optimization and code generation options plus
4417 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4418 Control Optimization}).
4421 For test coverage analysis, use @command{gcov} to produce human readable
4422 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4423 @command{gcov} documentation for further information.
4427 With @option{-fprofile-arcs}, for each function of your program GCC
4428 creates a program flow graph, then finds a spanning tree for the graph.
4429 Only arcs that are not on the spanning tree have to be instrumented: the
4430 compiler adds code to count the number of times that these arcs are
4431 executed. When an arc is the only exit or only entrance to a block, the
4432 instrumentation code can be added to the block; otherwise, a new basic
4433 block must be created to hold the instrumentation code.
4436 @item -ftest-coverage
4437 @opindex ftest-coverage
4438 Produce a notes file that the @command{gcov} code-coverage utility
4439 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4440 show program coverage. Each source file's note file is called
4441 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4442 above for a description of @var{auxname} and instructions on how to
4443 generate test coverage data. Coverage data will match the source files
4444 more closely, if you do not optimize.
4446 @item -fdbg-cnt-list
4447 @opindex fdbg-cnt-list
4448 Print the name and the counter upperbound for all debug counters.
4450 @item -fdbg-cnt=@var{counter-value-list}
4452 Set the internal debug counter upperbound. @var{counter-value-list}
4453 is a comma-separated list of @var{name}:@var{value} pairs
4454 which sets the upperbound of each debug counter @var{name} to @var{value}.
4455 All debug counters have the initial upperbound of @var{UINT_MAX},
4456 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4457 e.g. With -fdbg-cnt=dce:10,tail_call:0
4458 dbg_cnt(dce) will return true only for first 10 invocations
4459 and dbg_cnt(tail_call) will return false always.
4461 @item -d@var{letters}
4462 @itemx -fdump-rtl-@var{pass}
4464 Says to make debugging dumps during compilation at times specified by
4465 @var{letters}. This is used for debugging the RTL-based passes of the
4466 compiler. The file names for most of the dumps are made by appending a
4467 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
4468 from the name of the output file, if explicitly specified and it is not
4469 an executable, otherwise it is the basename of the source file. These
4470 switches may have different effects when @option{-E} is used for
4473 Most debug dumps can be enabled either passing a letter to the @option{-d}
4474 option, or with a long @option{-fdump-rtl} switch; here are the possible
4475 letters for use in @var{letters} and @var{pass}, and their meanings:
4480 Annotate the assembler output with miscellaneous debugging information.
4482 @item -fdump-rtl-bbro
4483 @opindex fdump-rtl-bbro
4484 Dump after block reordering, to @file{@var{file}.148r.bbro}.
4486 @item -fdump-rtl-combine
4487 @opindex fdump-rtl-combine
4488 Dump after the RTL instruction combination pass, to the file
4489 @file{@var{file}.129r.combine}.
4491 @item -fdump-rtl-ce1
4492 @itemx -fdump-rtl-ce2
4493 @opindex fdump-rtl-ce1
4494 @opindex fdump-rtl-ce2
4495 @option{-fdump-rtl-ce1} enable dumping after the
4496 first if conversion, to the file @file{@var{file}.117r.ce1}.
4497 @option{-fdump-rtl-ce2} enable dumping after the second if
4498 conversion, to the file @file{@var{file}.130r.ce2}.
4500 @item -fdump-rtl-btl
4501 @itemx -fdump-rtl-dbr
4502 @opindex fdump-rtl-btl
4503 @opindex fdump-rtl-dbr
4504 @option{-fdump-rtl-btl} enable dumping after branch
4505 target load optimization, to @file{@var{file}.31.btl}.
4506 @option{-fdump-rtl-dbr} enable dumping after delayed branch
4507 scheduling, to @file{@var{file}.36.dbr}.
4511 Dump all macro definitions, at the end of preprocessing, in addition to
4514 @item -fdump-rtl-ce3
4515 @opindex fdump-rtl-ce3
4516 Dump after the third if conversion, to @file{@var{file}.146r.ce3}.
4518 @item -fdump-rtl-cfg
4519 @itemx -fdump-rtl-life
4520 @opindex fdump-rtl-cfg
4521 @opindex fdump-rtl-life
4522 @option{-fdump-rtl-cfg} enable dumping after control
4523 and data flow analysis, to @file{@var{file}.116r.cfg}.
4524 @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
4525 to @file{@var{file}.128r.life1} and @file{@var{file}.135r.life2}.
4527 @item -fdump-rtl-greg
4528 @opindex fdump-rtl-greg
4529 Dump after global register allocation, to @file{@var{file}.139r.greg}.
4531 @item -fdump-rtl-gcse
4532 @itemx -fdump-rtl-bypass
4533 @opindex fdump-rtl-gcse
4534 @opindex fdump-rtl-bypass
4535 @option{-fdump-rtl-gcse} enable dumping after GCSE, to
4536 @file{@var{file}.114r.gcse}. @option{-fdump-rtl-bypass}
4537 enable dumping after jump bypassing and control flow optimizations, to
4538 @file{@var{file}.115r.bypass}.
4541 @opindex fdump-rtl-eh
4542 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
4544 @item -fdump-rtl-sibling
4545 @opindex fdump-rtl-sibling
4546 Dump after sibling call optimizations, to @file{@var{file}.106r.sibling}.
4548 @item -fdump-rtl-jump
4549 @opindex fdump-rtl-jump
4550 Dump after the first jump optimization, to @file{@var{file}.112r.jump}.
4552 @item -fdump-rtl-stack
4553 @opindex fdump-rtl-stack
4554 Dump after conversion from GCC's "flat register file" registers to the
4555 x87's stack-like registers, to @file{@var{file}.152r.stack}.
4557 @item -fdump-rtl-lreg
4558 @opindex fdump-rtl-lreg
4559 Dump after local register allocation, to @file{@var{file}.138r.lreg}.
4561 @item -fdump-rtl-loop2
4562 @opindex fdump-rtl-loop2
4563 @option{-dL} and @option{-fdump-rtl-loop2} enable dumping after the
4564 loop optimization pass, to @file{@var{file}.119r.loop2},
4565 @file{@var{file}.120r.loop2_init},
4566 @file{@var{file}.121r.loop2_invariant}, and
4567 @file{@var{file}.125r.loop2_done}.
4569 @item -fdump-rtl-sms
4570 @opindex fdump-rtl-sms
4571 Dump after modulo scheduling, to @file{@var{file}.136r.sms}.
4573 @item -fdump-rtl-mach
4574 @opindex fdump-rtl-mach
4575 Dump after performing the machine dependent reorganization pass, to
4576 @file{@var{file}.155r.mach} if that pass exists.
4578 @item -fdump-rtl-rnreg
4579 @opindex fdump-rtl-rnreg
4580 Dump after register renumbering, to @file{@var{file}.147r.rnreg}.
4582 @item -fdump-rtl-regmove
4583 @opindex fdump-rtl-regmove
4584 Dump after the register move pass, to @file{@var{file}.132r.regmove}.
4586 @item -fdump-rtl-postreload
4587 @opindex fdump-rtl-postreload
4588 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
4590 @item -fdump-rtl-expand
4591 @opindex fdump-rtl-expand
4592 Dump after RTL generation, to @file{@var{file}.104r.expand}.
4594 @item -fdump-rtl-sched2
4595 @opindex fdump-rtl-sched2
4596 Dump after the second scheduling pass, to @file{@var{file}.149r.sched2}.
4598 @item -fdump-rtl-cse
4599 @opindex fdump-rtl-cse
4600 Dump after CSE (including the jump optimization that sometimes follows
4601 CSE), to @file{@var{file}.113r.cse}.
4603 @item -fdump-rtl-sched1
4604 @opindex fdump-rtl-sched1
4605 Dump after the first scheduling pass, to @file{@var{file}.136r.sched1}.
4607 @item -fdump-rtl-cse2
4608 @opindex fdump-rtl-cse2
4609 Dump after the second CSE pass (including the jump optimization that
4610 sometimes follows CSE), to @file{@var{file}.127r.cse2}.
4612 @item -fdump-rtl-tracer
4613 @opindex fdump-rtl-tracer
4614 Dump after running tracer, to @file{@var{file}.118r.tracer}.
4616 @item -fdump-rtl-vpt
4617 @itemx -fdump-rtl-vartrack
4618 @opindex fdump-rtl-vpt
4619 @opindex fdump-rtl-vartrack
4620 @option{-fdump-rtl-vpt} enable dumping after the value
4621 profile transformations, to @file{@var{file}.10.vpt}.
4622 @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
4623 to @file{@var{file}.154r.vartrack}.
4625 @item -fdump-rtl-flow2
4626 @opindex fdump-rtl-flow2
4627 Dump after the second flow pass, to @file{@var{file}.142r.flow2}.
4629 @item -fdump-rtl-peephole2
4630 @opindex fdump-rtl-peephole2
4631 Dump after the peephole pass, to @file{@var{file}.145r.peephole2}.
4633 @item -fdump-rtl-web
4634 @opindex fdump-rtl-web
4635 Dump after live range splitting, to @file{@var{file}.126r.web}.
4637 @item -fdump-rtl-all
4638 @opindex fdump-rtl-all
4639 Produce all the dumps listed above.
4643 Produce a core dump whenever an error occurs.
4647 Print statistics on memory usage, at the end of the run, to
4652 Annotate the assembler output with a comment indicating which
4653 pattern and alternative was used. The length of each instruction is
4658 Dump the RTL in the assembler output as a comment before each instruction.
4659 Also turns on @option{-dp} annotation.
4663 For each of the other indicated dump files (either with @option{-d} or
4664 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
4665 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
4669 Just generate RTL for a function instead of compiling it. Usually used
4670 with @samp{r} (@option{-fdump-rtl-expand}).
4674 Dump debugging information during parsing, to standard error.
4678 @opindex fdump-noaddr
4679 When doing debugging dumps (see @option{-d} option above), suppress
4680 address output. This makes it more feasible to use diff on debugging
4681 dumps for compiler invocations with different compiler binaries and/or
4682 different text / bss / data / heap / stack / dso start locations.
4684 @item -fdump-unnumbered
4685 @opindex fdump-unnumbered
4686 When doing debugging dumps (see @option{-d} option above), suppress instruction
4687 numbers and address output. This makes it more feasible to
4688 use diff on debugging dumps for compiler invocations with different
4689 options, in particular with and without @option{-g}.
4691 @item -fdump-translation-unit @r{(C++ only)}
4692 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
4693 @opindex fdump-translation-unit
4694 Dump a representation of the tree structure for the entire translation
4695 unit to a file. The file name is made by appending @file{.tu} to the
4696 source file name. If the @samp{-@var{options}} form is used, @var{options}
4697 controls the details of the dump as described for the
4698 @option{-fdump-tree} options.
4700 @item -fdump-class-hierarchy @r{(C++ only)}
4701 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
4702 @opindex fdump-class-hierarchy
4703 Dump a representation of each class's hierarchy and virtual function
4704 table layout to a file. The file name is made by appending @file{.class}
4705 to the source file name. If the @samp{-@var{options}} form is used,
4706 @var{options} controls the details of the dump as described for the
4707 @option{-fdump-tree} options.
4709 @item -fdump-ipa-@var{switch}
4711 Control the dumping at various stages of inter-procedural analysis
4712 language tree to a file. The file name is generated by appending a switch
4713 specific suffix to the source file name. The following dumps are possible:
4717 Enables all inter-procedural analysis dumps.
4720 Dumps information about call-graph optimization, unused function removal,
4721 and inlining decisions.
4724 Dump after function inlining.
4728 @item -fdump-statistics-@var{option}
4729 @opindex -fdump-statistics
4730 Enable and control dumping of pass statistics in a separate file. The
4731 file name is generated by appending a suffix ending in @samp{.statistics}
4732 to the source file name. If the @samp{-@var{option}} form is used,
4733 @samp{-stats} will cause counters to be summed over the whole compilation unit
4734 while @samp{-details} will dump every event as the passes generate them.
4735 The default with no option is to sum counters for each function compiled.
4737 @item -fdump-tree-@var{switch}
4738 @itemx -fdump-tree-@var{switch}-@var{options}
4740 Control the dumping at various stages of processing the intermediate
4741 language tree to a file. The file name is generated by appending a switch
4742 specific suffix to the source file name. If the @samp{-@var{options}}
4743 form is used, @var{options} is a list of @samp{-} separated options that
4744 control the details of the dump. Not all options are applicable to all
4745 dumps, those which are not meaningful will be ignored. The following
4746 options are available
4750 Print the address of each node. Usually this is not meaningful as it
4751 changes according to the environment and source file. Its primary use
4752 is for tying up a dump file with a debug environment.
4754 Inhibit dumping of members of a scope or body of a function merely
4755 because that scope has been reached. Only dump such items when they
4756 are directly reachable by some other path. When dumping pretty-printed
4757 trees, this option inhibits dumping the bodies of control structures.
4759 Print a raw representation of the tree. By default, trees are
4760 pretty-printed into a C-like representation.
4762 Enable more detailed dumps (not honored by every dump option).
4764 Enable dumping various statistics about the pass (not honored by every dump
4767 Enable showing basic block boundaries (disabled in raw dumps).
4769 Enable showing virtual operands for every statement.
4771 Enable showing line numbers for statements.
4773 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4775 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4778 The following tree dumps are possible:
4782 Dump before any tree based optimization, to @file{@var{file}.original}.
4785 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4788 @opindex fdump-tree-gimple
4789 Dump each function before and after the gimplification pass to a file. The
4790 file name is made by appending @file{.gimple} to the source file name.
4793 @opindex fdump-tree-cfg
4794 Dump the control flow graph of each function to a file. The file name is
4795 made by appending @file{.cfg} to the source file name.
4798 @opindex fdump-tree-vcg
4799 Dump the control flow graph of each function to a file in VCG format. The
4800 file name is made by appending @file{.vcg} to the source file name. Note
4801 that if the file contains more than one function, the generated file cannot
4802 be used directly by VCG@. You will need to cut and paste each function's
4803 graph into its own separate file first.
4806 @opindex fdump-tree-ch
4807 Dump each function after copying loop headers. The file name is made by
4808 appending @file{.ch} to the source file name.
4811 @opindex fdump-tree-ssa
4812 Dump SSA related information to a file. The file name is made by appending
4813 @file{.ssa} to the source file name.
4816 @opindex fdump-tree-alias
4817 Dump aliasing information for each function. The file name is made by
4818 appending @file{.alias} to the source file name.
4821 @opindex fdump-tree-ccp
4822 Dump each function after CCP@. The file name is made by appending
4823 @file{.ccp} to the source file name.
4826 @opindex fdump-tree-storeccp
4827 Dump each function after STORE-CCP@. The file name is made by appending
4828 @file{.storeccp} to the source file name.
4831 @opindex fdump-tree-pre
4832 Dump trees after partial redundancy elimination. The file name is made
4833 by appending @file{.pre} to the source file name.
4836 @opindex fdump-tree-fre
4837 Dump trees after full redundancy elimination. The file name is made
4838 by appending @file{.fre} to the source file name.
4841 @opindex fdump-tree-copyprop
4842 Dump trees after copy propagation. The file name is made
4843 by appending @file{.copyprop} to the source file name.
4845 @item store_copyprop
4846 @opindex fdump-tree-store_copyprop
4847 Dump trees after store copy-propagation. The file name is made
4848 by appending @file{.store_copyprop} to the source file name.
4851 @opindex fdump-tree-dce
4852 Dump each function after dead code elimination. The file name is made by
4853 appending @file{.dce} to the source file name.
4856 @opindex fdump-tree-mudflap
4857 Dump each function after adding mudflap instrumentation. The file name is
4858 made by appending @file{.mudflap} to the source file name.
4861 @opindex fdump-tree-sra
4862 Dump each function after performing scalar replacement of aggregates. The
4863 file name is made by appending @file{.sra} to the source file name.
4866 @opindex fdump-tree-sink
4867 Dump each function after performing code sinking. The file name is made
4868 by appending @file{.sink} to the source file name.
4871 @opindex fdump-tree-dom
4872 Dump each function after applying dominator tree optimizations. The file
4873 name is made by appending @file{.dom} to the source file name.
4876 @opindex fdump-tree-dse
4877 Dump each function after applying dead store elimination. The file
4878 name is made by appending @file{.dse} to the source file name.
4881 @opindex fdump-tree-phiopt
4882 Dump each function after optimizing PHI nodes into straightline code. The file
4883 name is made by appending @file{.phiopt} to the source file name.
4886 @opindex fdump-tree-forwprop
4887 Dump each function after forward propagating single use variables. The file
4888 name is made by appending @file{.forwprop} to the source file name.
4891 @opindex fdump-tree-copyrename
4892 Dump each function after applying the copy rename optimization. The file
4893 name is made by appending @file{.copyrename} to the source file name.
4896 @opindex fdump-tree-nrv
4897 Dump each function after applying the named return value optimization on
4898 generic trees. The file name is made by appending @file{.nrv} to the source
4902 @opindex fdump-tree-vect
4903 Dump each function after applying vectorization of loops. The file name is
4904 made by appending @file{.vect} to the source file name.
4907 @opindex fdump-tree-vrp
4908 Dump each function after Value Range Propagation (VRP). The file name
4909 is made by appending @file{.vrp} to the source file name.
4912 @opindex fdump-tree-all
4913 Enable all the available tree dumps with the flags provided in this option.
4916 @item -ftree-vectorizer-verbose=@var{n}
4917 @opindex ftree-vectorizer-verbose
4918 This option controls the amount of debugging output the vectorizer prints.
4919 This information is written to standard error, unless
4920 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
4921 in which case it is output to the usual dump listing file, @file{.vect}.
4922 For @var{n}=0 no diagnostic information is reported.
4923 If @var{n}=1 the vectorizer reports each loop that got vectorized,
4924 and the total number of loops that got vectorized.
4925 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
4926 the first analysis phase (vect_analyze_loop_form) - i.e.@: countable,
4927 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
4928 level that @option{-fdump-tree-vect-stats} uses.
4929 Higher verbosity levels mean either more information dumped for each
4930 reported loop, or same amount of information reported for more loops:
4931 If @var{n}=3, alignment related information is added to the reports.
4932 If @var{n}=4, data-references related information (e.g.@: memory dependences,
4933 memory access-patterns) is added to the reports.
4934 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
4935 that did not pass the first analysis phase (i.e., may not be countable, or
4936 may have complicated control-flow).
4937 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
4938 For @var{n}=7, all the information the vectorizer generates during its
4939 analysis and transformation is reported. This is the same verbosity level
4940 that @option{-fdump-tree-vect-details} uses.
4942 @item -frandom-seed=@var{string}
4943 @opindex frandom-string
4944 This option provides a seed that GCC uses when it would otherwise use
4945 random numbers. It is used to generate certain symbol names
4946 that have to be different in every compiled file. It is also used to
4947 place unique stamps in coverage data files and the object files that
4948 produce them. You can use the @option{-frandom-seed} option to produce
4949 reproducibly identical object files.
4951 The @var{string} should be different for every file you compile.
4953 @item -fsched-verbose=@var{n}
4954 @opindex fsched-verbose
4955 On targets that use instruction scheduling, this option controls the
4956 amount of debugging output the scheduler prints. This information is
4957 written to standard error, unless @option{-dS} or @option{-dR} is
4958 specified, in which case it is output to the usual dump
4959 listing file, @file{.sched} or @file{.sched2} respectively. However
4960 for @var{n} greater than nine, the output is always printed to standard
4963 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4964 same information as @option{-dRS}. For @var{n} greater than one, it
4965 also output basic block probabilities, detailed ready list information
4966 and unit/insn info. For @var{n} greater than two, it includes RTL
4967 at abort point, control-flow and regions info. And for @var{n} over
4968 four, @option{-fsched-verbose} also includes dependence info.
4972 Store the usual ``temporary'' intermediate files permanently; place them
4973 in the current directory and name them based on the source file. Thus,
4974 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4975 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4976 preprocessed @file{foo.i} output file even though the compiler now
4977 normally uses an integrated preprocessor.
4979 When used in combination with the @option{-x} command line option,
4980 @option{-save-temps} is sensible enough to avoid over writing an
4981 input source file with the same extension as an intermediate file.
4982 The corresponding intermediate file may be obtained by renaming the
4983 source file before using @option{-save-temps}.
4987 Report the CPU time taken by each subprocess in the compilation
4988 sequence. For C source files, this is the compiler proper and assembler
4989 (plus the linker if linking is done). The output looks like this:
4996 The first number on each line is the ``user time'', that is time spent
4997 executing the program itself. The second number is ``system time'',
4998 time spent executing operating system routines on behalf of the program.
4999 Both numbers are in seconds.
5001 @item -fvar-tracking
5002 @opindex fvar-tracking
5003 Run variable tracking pass. It computes where variables are stored at each
5004 position in code. Better debugging information is then generated
5005 (if the debugging information format supports this information).
5007 It is enabled by default when compiling with optimization (@option{-Os},
5008 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
5009 the debug info format supports it.
5011 @item -print-file-name=@var{library}
5012 @opindex print-file-name
5013 Print the full absolute name of the library file @var{library} that
5014 would be used when linking---and don't do anything else. With this
5015 option, GCC does not compile or link anything; it just prints the
5018 @item -print-multi-directory
5019 @opindex print-multi-directory
5020 Print the directory name corresponding to the multilib selected by any
5021 other switches present in the command line. This directory is supposed
5022 to exist in @env{GCC_EXEC_PREFIX}.
5024 @item -print-multi-lib
5025 @opindex print-multi-lib
5026 Print the mapping from multilib directory names to compiler switches
5027 that enable them. The directory name is separated from the switches by
5028 @samp{;}, and each switch starts with an @samp{@@} instead of the
5029 @samp{-}, without spaces between multiple switches. This is supposed to
5030 ease shell-processing.
5032 @item -print-prog-name=@var{program}
5033 @opindex print-prog-name
5034 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
5036 @item -print-libgcc-file-name
5037 @opindex print-libgcc-file-name
5038 Same as @option{-print-file-name=libgcc.a}.
5040 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
5041 but you do want to link with @file{libgcc.a}. You can do
5044 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
5047 @item -print-search-dirs
5048 @opindex print-search-dirs
5049 Print the name of the configured installation directory and a list of
5050 program and library directories @command{gcc} will search---and don't do anything else.
5052 This is useful when @command{gcc} prints the error message
5053 @samp{installation problem, cannot exec cpp0: No such file or directory}.
5054 To resolve this you either need to put @file{cpp0} and the other compiler
5055 components where @command{gcc} expects to find them, or you can set the environment
5056 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
5057 Don't forget the trailing @samp{/}.
5058 @xref{Environment Variables}.
5060 @item -print-sysroot-headers-suffix
5061 @opindex print-sysroot-headers-suffix
5062 Print the suffix added to the target sysroot when searching for
5063 headers, or give an error if the compiler is not configured with such
5064 a suffix---and don't do anything else.
5067 @opindex dumpmachine
5068 Print the compiler's target machine (for example,
5069 @samp{i686-pc-linux-gnu})---and don't do anything else.
5072 @opindex dumpversion
5073 Print the compiler version (for example, @samp{3.0})---and don't do
5078 Print the compiler's built-in specs---and don't do anything else. (This
5079 is used when GCC itself is being built.) @xref{Spec Files}.
5081 @item -feliminate-unused-debug-types
5082 @opindex feliminate-unused-debug-types
5083 Normally, when producing DWARF2 output, GCC will emit debugging
5084 information for all types declared in a compilation
5085 unit, regardless of whether or not they are actually used
5086 in that compilation unit. Sometimes this is useful, such as
5087 if, in the debugger, you want to cast a value to a type that is
5088 not actually used in your program (but is declared). More often,
5089 however, this results in a significant amount of wasted space.
5090 With this option, GCC will avoid producing debug symbol output
5091 for types that are nowhere used in the source file being compiled.
5094 @node Optimize Options
5095 @section Options That Control Optimization
5096 @cindex optimize options
5097 @cindex options, optimization
5099 These options control various sorts of optimizations.
5101 Without any optimization option, the compiler's goal is to reduce the
5102 cost of compilation and to make debugging produce the expected
5103 results. Statements are independent: if you stop the program with a
5104 breakpoint between statements, you can then assign a new value to any
5105 variable or change the program counter to any other statement in the
5106 function and get exactly the results you would expect from the source
5109 Turning on optimization flags makes the compiler attempt to improve
5110 the performance and/or code size at the expense of compilation time
5111 and possibly the ability to debug the program.
5113 The compiler performs optimization based on the knowledge it has of
5114 the program. Optimization levels @option{-O} and above, in
5115 particular, enable @emph{unit-at-a-time} mode, which allows the
5116 compiler to consider information gained from later functions in
5117 the file when compiling a function. Compiling multiple files at
5118 once to a single output file in @emph{unit-at-a-time} mode allows
5119 the compiler to use information gained from all of the files when
5120 compiling each of them.
5122 Not all optimizations are controlled directly by a flag. Only
5123 optimizations that have a flag are listed.
5130 Optimize. Optimizing compilation takes somewhat more time, and a lot
5131 more memory for a large function.
5133 With @option{-O}, the compiler tries to reduce code size and execution
5134 time, without performing any optimizations that take a great deal of
5137 @option{-O} turns on the following optimization flags:
5140 -fcprop-registers @gol
5143 -fdelayed-branch @gol
5145 -fguess-branch-probability @gol
5146 -fif-conversion2 @gol
5147 -fif-conversion @gol
5148 -finline-small-functions @gol
5149 -fipa-pure-const @gol
5150 -fipa-reference @gol
5152 -fsplit-wide-types @gol
5155 -ftree-copyrename @gol
5157 -ftree-dominator-opts @gol
5164 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5165 where doing so does not interfere with debugging.
5169 Optimize even more. GCC performs nearly all supported optimizations
5170 that do not involve a space-speed tradeoff. The compiler does not
5171 perform loop unrolling or function inlining when you specify @option{-O2}.
5172 As compared to @option{-O}, this option increases both compilation time
5173 and the performance of the generated code.
5175 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5176 also turns on the following optimization flags:
5177 @gccoptlist{-fthread-jumps @gol
5178 -falign-functions -falign-jumps @gol
5179 -falign-loops -falign-labels @gol
5182 -fcse-follow-jumps -fcse-skip-blocks @gol
5183 -fdelete-null-pointer-checks @gol
5184 -fexpensive-optimizations @gol
5185 -fgcse -fgcse-lm @gol
5186 -foptimize-sibling-calls @gol
5189 -freorder-blocks -freorder-functions @gol
5190 -frerun-cse-after-loop @gol
5191 -fsched-interblock -fsched-spec @gol
5192 -fschedule-insns -fschedule-insns2 @gol
5193 -fstrict-aliasing -fstrict-overflow @gol
5197 Please note the warning under @option{-fgcse} about
5198 invoking @option{-O2} on programs that use computed gotos.
5202 Optimize yet more. @option{-O3} turns on all optimizations specified by
5203 @option{-O2} and also turns on the @option{-finline-functions},
5204 @option{-funswitch-loops}, @option{-fpredictive-commoning},
5205 @option{-fgcse-after-reload} and @option{-ftree-vectorize} options.
5209 Reduce compilation time and make debugging produce the expected
5210 results. This is the default.
5214 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5215 do not typically increase code size. It also performs further
5216 optimizations designed to reduce code size.
5218 @option{-Os} disables the following optimization flags:
5219 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5220 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5221 -fprefetch-loop-arrays -ftree-vect-loop-version}
5223 If you use multiple @option{-O} options, with or without level numbers,
5224 the last such option is the one that is effective.
5227 Options of the form @option{-f@var{flag}} specify machine-independent
5228 flags. Most flags have both positive and negative forms; the negative
5229 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5230 below, only one of the forms is listed---the one you typically will
5231 use. You can figure out the other form by either removing @samp{no-}
5234 The following options control specific optimizations. They are either
5235 activated by @option{-O} options or are related to ones that are. You
5236 can use the following flags in the rare cases when ``fine-tuning'' of
5237 optimizations to be performed is desired.
5240 @item -fno-default-inline
5241 @opindex fno-default-inline
5242 Do not make member functions inline by default merely because they are
5243 defined inside the class scope (C++ only). Otherwise, when you specify
5244 @w{@option{-O}}, member functions defined inside class scope are compiled
5245 inline by default; i.e., you don't need to add @samp{inline} in front of
5246 the member function name.
5248 @item -fno-defer-pop
5249 @opindex fno-defer-pop
5250 Always pop the arguments to each function call as soon as that function
5251 returns. For machines which must pop arguments after a function call,
5252 the compiler normally lets arguments accumulate on the stack for several
5253 function calls and pops them all at once.
5255 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5257 @item -fforward-propagate
5258 @opindex fforward-propagate
5259 Perform a forward propagation pass on RTL@. The pass tries to combine two
5260 instructions and checks if the result can be simplified. If loop unrolling
5261 is active, two passes are performed and the second is scheduled after
5264 This option is enabled by default at optimization levels @option{-O2},
5265 @option{-O3}, @option{-Os}.
5267 @item -fomit-frame-pointer
5268 @opindex fomit-frame-pointer
5269 Don't keep the frame pointer in a register for functions that
5270 don't need one. This avoids the instructions to save, set up and
5271 restore frame pointers; it also makes an extra register available
5272 in many functions. @strong{It also makes debugging impossible on
5275 On some machines, such as the VAX, this flag has no effect, because
5276 the standard calling sequence automatically handles the frame pointer
5277 and nothing is saved by pretending it doesn't exist. The
5278 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5279 whether a target machine supports this flag. @xref{Registers,,Register
5280 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5282 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5284 @item -foptimize-sibling-calls
5285 @opindex foptimize-sibling-calls
5286 Optimize sibling and tail recursive calls.
5288 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5292 Don't pay attention to the @code{inline} keyword. Normally this option
5293 is used to keep the compiler from expanding any functions inline.
5294 Note that if you are not optimizing, no functions can be expanded inline.
5296 @item -finline-small-functions
5297 @opindex finline-small-functions
5298 Integrate functions into their callers when their body is smaller than expected
5299 function call code (so overall size of program gets smaller). The compiler
5300 heuristically decides which functions are simple enough to be worth integrating
5303 Enabled at level @option{-O2}.
5305 @item -finline-functions
5306 @opindex finline-functions
5307 Integrate all simple functions into their callers. The compiler
5308 heuristically decides which functions are simple enough to be worth
5309 integrating in this way.
5311 If all calls to a given function are integrated, and the function is
5312 declared @code{static}, then the function is normally not output as
5313 assembler code in its own right.
5315 Enabled at level @option{-O3}.
5317 @item -finline-functions-called-once
5318 @opindex finline-functions-called-once
5319 Consider all @code{static} functions called once for inlining into their
5320 caller even if they are not marked @code{inline}. If a call to a given
5321 function is integrated, then the function is not output as assembler code
5324 Enabled if @option{-funit-at-a-time} is enabled.
5326 @item -fearly-inlining
5327 @opindex fearly-inlining
5328 Inline functions marked by @code{always_inline} and functions whose body seems
5329 smaller than the function call overhead early before doing
5330 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5331 makes profiling significantly cheaper and usually inlining faster on programs
5332 having large chains of nested wrapper functions.
5336 @item -finline-limit=@var{n}
5337 @opindex finline-limit
5338 By default, GCC limits the size of functions that can be inlined. This flag
5339 allows coarse control of this limit. @var{n} is the size of functions that
5340 can be inlined in number of pseudo instructions.
5342 Inlining is actually controlled by a number of parameters, which may be
5343 specified individually by using @option{--param @var{name}=@var{value}}.
5344 The @option{-finline-limit=@var{n}} option sets some of these parameters
5348 @item max-inline-insns-single
5349 is set to @var{n}/2.
5350 @item max-inline-insns-auto
5351 is set to @var{n}/2.
5354 See below for a documentation of the individual
5355 parameters controlling inlining and for the defaults of these parameters.
5357 @emph{Note:} there may be no value to @option{-finline-limit} that results
5358 in default behavior.
5360 @emph{Note:} pseudo instruction represents, in this particular context, an
5361 abstract measurement of function's size. In no way does it represent a count
5362 of assembly instructions and as such its exact meaning might change from one
5363 release to an another.
5365 @item -fkeep-inline-functions
5366 @opindex fkeep-inline-functions
5367 In C, emit @code{static} functions that are declared @code{inline}
5368 into the object file, even if the function has been inlined into all
5369 of its callers. This switch does not affect functions using the
5370 @code{extern inline} extension in GNU C89@. In C++, emit any and all
5371 inline functions into the object file.
5373 @item -fkeep-static-consts
5374 @opindex fkeep-static-consts
5375 Emit variables declared @code{static const} when optimization isn't turned
5376 on, even if the variables aren't referenced.
5378 GCC enables this option by default. If you want to force the compiler to
5379 check if the variable was referenced, regardless of whether or not
5380 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5382 @item -fmerge-constants
5383 @opindex fmerge-constants
5384 Attempt to merge identical constants (string constants and floating point
5385 constants) across compilation units.
5387 This option is the default for optimized compilation if the assembler and
5388 linker support it. Use @option{-fno-merge-constants} to inhibit this
5391 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5393 @item -fmerge-all-constants
5394 @opindex fmerge-all-constants
5395 Attempt to merge identical constants and identical variables.
5397 This option implies @option{-fmerge-constants}. In addition to
5398 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5399 arrays or initialized constant variables with integral or floating point
5400 types. Languages like C or C++ require each non-automatic variable to
5401 have distinct location, so using this option will result in non-conforming
5404 @item -fmodulo-sched
5405 @opindex fmodulo-sched
5406 Perform swing modulo scheduling immediately before the first scheduling
5407 pass. This pass looks at innermost loops and reorders their
5408 instructions by overlapping different iterations.
5410 @item -fmodulo-sched-allow-regmoves
5411 @opindex fmodulo-sched-allow-regmoves
5412 Perform more aggressive SMS based modulo scheduling with register moves
5413 allowed. By setting this flag certain anti-dependences edges will be
5414 deleted which will trigger the generation of reg-moves based on the
5415 life-range analysis. This option is effective only with
5416 @option{-fmodulo-sched} enabled.
5418 @item -fno-branch-count-reg
5419 @opindex fno-branch-count-reg
5420 Do not use ``decrement and branch'' instructions on a count register,
5421 but instead generate a sequence of instructions that decrement a
5422 register, compare it against zero, then branch based upon the result.
5423 This option is only meaningful on architectures that support such
5424 instructions, which include x86, PowerPC, IA-64 and S/390.
5426 The default is @option{-fbranch-count-reg}.
5428 @item -fno-function-cse
5429 @opindex fno-function-cse
5430 Do not put function addresses in registers; make each instruction that
5431 calls a constant function contain the function's address explicitly.
5433 This option results in less efficient code, but some strange hacks
5434 that alter the assembler output may be confused by the optimizations
5435 performed when this option is not used.
5437 The default is @option{-ffunction-cse}
5439 @item -fno-zero-initialized-in-bss
5440 @opindex fno-zero-initialized-in-bss
5441 If the target supports a BSS section, GCC by default puts variables that
5442 are initialized to zero into BSS@. This can save space in the resulting
5445 This option turns off this behavior because some programs explicitly
5446 rely on variables going to the data section. E.g., so that the
5447 resulting executable can find the beginning of that section and/or make
5448 assumptions based on that.
5450 The default is @option{-fzero-initialized-in-bss}.
5452 @item -fmudflap -fmudflapth -fmudflapir
5456 @cindex bounds checking
5458 For front-ends that support it (C and C++), instrument all risky
5459 pointer/array dereferencing operations, some standard library
5460 string/heap functions, and some other associated constructs with
5461 range/validity tests. Modules so instrumented should be immune to
5462 buffer overflows, invalid heap use, and some other classes of C/C++
5463 programming errors. The instrumentation relies on a separate runtime
5464 library (@file{libmudflap}), which will be linked into a program if
5465 @option{-fmudflap} is given at link time. Run-time behavior of the
5466 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
5467 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
5470 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
5471 link if your program is multi-threaded. Use @option{-fmudflapir}, in
5472 addition to @option{-fmudflap} or @option{-fmudflapth}, if
5473 instrumentation should ignore pointer reads. This produces less
5474 instrumentation (and therefore faster execution) and still provides
5475 some protection against outright memory corrupting writes, but allows
5476 erroneously read data to propagate within a program.
5478 @item -fthread-jumps
5479 @opindex fthread-jumps
5480 Perform optimizations where we check to see if a jump branches to a
5481 location where another comparison subsumed by the first is found. If
5482 so, the first branch is redirected to either the destination of the
5483 second branch or a point immediately following it, depending on whether
5484 the condition is known to be true or false.
5486 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5488 @item -fsplit-wide-types
5489 @opindex fsplit-wide-types
5490 When using a type that occupies multiple registers, such as @code{long
5491 long} on a 32-bit system, split the registers apart and allocate them
5492 independently. This normally generates better code for those types,
5493 but may make debugging more difficult.
5495 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
5498 @item -fcse-follow-jumps
5499 @opindex fcse-follow-jumps
5500 In common subexpression elimination (CSE), scan through jump instructions
5501 when the target of the jump is not reached by any other path. For
5502 example, when CSE encounters an @code{if} statement with an
5503 @code{else} clause, CSE will follow the jump when the condition
5506 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5508 @item -fcse-skip-blocks
5509 @opindex fcse-skip-blocks
5510 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
5511 follow jumps which conditionally skip over blocks. When CSE
5512 encounters a simple @code{if} statement with no else clause,
5513 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
5514 body of the @code{if}.
5516 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5518 @item -frerun-cse-after-loop
5519 @opindex frerun-cse-after-loop
5520 Re-run common subexpression elimination after loop optimizations has been
5523 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5527 Perform a global common subexpression elimination pass.
5528 This pass also performs global constant and copy propagation.
5530 @emph{Note:} When compiling a program using computed gotos, a GCC
5531 extension, you may get better runtime performance if you disable
5532 the global common subexpression elimination pass by adding
5533 @option{-fno-gcse} to the command line.
5535 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5539 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
5540 attempt to move loads which are only killed by stores into themselves. This
5541 allows a loop containing a load/store sequence to be changed to a load outside
5542 the loop, and a copy/store within the loop.
5544 Enabled by default when gcse is enabled.
5548 When @option{-fgcse-sm} is enabled, a store motion pass is run after
5549 global common subexpression elimination. This pass will attempt to move
5550 stores out of loops. When used in conjunction with @option{-fgcse-lm},
5551 loops containing a load/store sequence can be changed to a load before
5552 the loop and a store after the loop.
5554 Not enabled at any optimization level.
5558 When @option{-fgcse-las} is enabled, the global common subexpression
5559 elimination pass eliminates redundant loads that come after stores to the
5560 same memory location (both partial and full redundancies).
5562 Not enabled at any optimization level.
5564 @item -fgcse-after-reload
5565 @opindex fgcse-after-reload
5566 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
5567 pass is performed after reload. The purpose of this pass is to cleanup
5570 @item -funsafe-loop-optimizations
5571 @opindex funsafe-loop-optimizations
5572 If given, the loop optimizer will assume that loop indices do not
5573 overflow, and that the loops with nontrivial exit condition are not
5574 infinite. This enables a wider range of loop optimizations even if
5575 the loop optimizer itself cannot prove that these assumptions are valid.
5576 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
5577 if it finds this kind of loop.
5579 @item -fcrossjumping
5580 @opindex fcrossjumping
5581 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
5582 resulting code may or may not perform better than without cross-jumping.
5584 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5586 @item -fauto-inc-dec
5587 @opindex fauto-inc-dec
5588 Combine increments or decrements of addresses with memory accesses.
5589 This pass is always skipped on architectures that do not have
5590 instructions to support this. Enabled by default at @option{-O} and
5591 higher on architectures that support this.
5595 Perform dead code elimination (DCE) on RTL@.
5596 Enabled by default at @option{-O} and higher.
5600 Perform dead store elimination (DSE) on RTL@.
5601 Enabled by default at @option{-O} and higher.
5603 @item -fif-conversion
5604 @opindex fif-conversion
5605 Attempt to transform conditional jumps into branch-less equivalents. This
5606 include use of conditional moves, min, max, set flags and abs instructions, and
5607 some tricks doable by standard arithmetics. The use of conditional execution
5608 on chips where it is available is controlled by @code{if-conversion2}.
5610 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5612 @item -fif-conversion2
5613 @opindex fif-conversion2
5614 Use conditional execution (where available) to transform conditional jumps into
5615 branch-less equivalents.
5617 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5619 @item -fdelete-null-pointer-checks
5620 @opindex fdelete-null-pointer-checks
5621 Use global dataflow analysis to identify and eliminate useless checks
5622 for null pointers. The compiler assumes that dereferencing a null
5623 pointer would have halted the program. If a pointer is checked after
5624 it has already been dereferenced, it cannot be null.
5626 In some environments, this assumption is not true, and programs can
5627 safely dereference null pointers. Use
5628 @option{-fno-delete-null-pointer-checks} to disable this optimization
5629 for programs which depend on that behavior.
5631 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5633 @item -fexpensive-optimizations
5634 @opindex fexpensive-optimizations
5635 Perform a number of minor optimizations that are relatively expensive.
5637 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5639 @item -foptimize-register-move
5641 @opindex foptimize-register-move
5643 Attempt to reassign register numbers in move instructions and as
5644 operands of other simple instructions in order to maximize the amount of
5645 register tying. This is especially helpful on machines with two-operand
5648 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
5651 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5653 @item -fdelayed-branch
5654 @opindex fdelayed-branch
5655 If supported for the target machine, attempt to reorder instructions
5656 to exploit instruction slots available after delayed branch
5659 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5661 @item -fschedule-insns
5662 @opindex fschedule-insns
5663 If supported for the target machine, attempt to reorder instructions to
5664 eliminate execution stalls due to required data being unavailable. This
5665 helps machines that have slow floating point or memory load instructions
5666 by allowing other instructions to be issued until the result of the load
5667 or floating point instruction is required.
5669 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5671 @item -fschedule-insns2
5672 @opindex fschedule-insns2
5673 Similar to @option{-fschedule-insns}, but requests an additional pass of
5674 instruction scheduling after register allocation has been done. This is
5675 especially useful on machines with a relatively small number of
5676 registers and where memory load instructions take more than one cycle.
5678 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5680 @item -fno-sched-interblock
5681 @opindex fno-sched-interblock
5682 Don't schedule instructions across basic blocks. This is normally
5683 enabled by default when scheduling before register allocation, i.e.@:
5684 with @option{-fschedule-insns} or at @option{-O2} or higher.
5686 @item -fno-sched-spec
5687 @opindex fno-sched-spec
5688 Don't allow speculative motion of non-load instructions. This is normally
5689 enabled by default when scheduling before register allocation, i.e.@:
5690 with @option{-fschedule-insns} or at @option{-O2} or higher.
5692 @item -fsched-spec-load
5693 @opindex fsched-spec-load
5694 Allow speculative motion of some load instructions. This only makes
5695 sense when scheduling before register allocation, i.e.@: with
5696 @option{-fschedule-insns} or at @option{-O2} or higher.
5698 @item -fsched-spec-load-dangerous
5699 @opindex fsched-spec-load-dangerous
5700 Allow speculative motion of more load instructions. This only makes
5701 sense when scheduling before register allocation, i.e.@: with
5702 @option{-fschedule-insns} or at @option{-O2} or higher.
5704 @item -fsched-stalled-insns
5705 @itemx -fsched-stalled-insns=@var{n}
5706 @opindex fsched-stalled-insns
5707 Define how many insns (if any) can be moved prematurely from the queue
5708 of stalled insns into the ready list, during the second scheduling pass.
5709 @option{-fno-sched-stalled-insns} means that no insns will be moved
5710 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
5711 on how many queued insns can be moved prematurely.
5712 @option{-fsched-stalled-insns} without a value is equivalent to
5713 @option{-fsched-stalled-insns=1}.
5715 @item -fsched-stalled-insns-dep
5716 @itemx -fsched-stalled-insns-dep=@var{n}
5717 @opindex fsched-stalled-insns-dep
5718 Define how many insn groups (cycles) will be examined for a dependency
5719 on a stalled insn that is candidate for premature removal from the queue
5720 of stalled insns. This has an effect only during the second scheduling pass,
5721 and only if @option{-fsched-stalled-insns} is used.
5722 @option{-fno-sched-stalled-insns-dep} is equivalent to
5723 @option{-fsched-stalled-insns-dep=0}.
5724 @option{-fsched-stalled-insns-dep} without a value is equivalent to
5725 @option{-fsched-stalled-insns-dep=1}.
5727 @item -fsched2-use-superblocks
5728 @opindex fsched2-use-superblocks
5729 When scheduling after register allocation, do use superblock scheduling
5730 algorithm. Superblock scheduling allows motion across basic block boundaries
5731 resulting on faster schedules. This option is experimental, as not all machine
5732 descriptions used by GCC model the CPU closely enough to avoid unreliable
5733 results from the algorithm.
5735 This only makes sense when scheduling after register allocation, i.e.@: with
5736 @option{-fschedule-insns2} or at @option{-O2} or higher.
5738 @item -fsched2-use-traces
5739 @opindex fsched2-use-traces
5740 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
5741 allocation and additionally perform code duplication in order to increase the
5742 size of superblocks using tracer pass. See @option{-ftracer} for details on
5745 This mode should produce faster but significantly longer programs. Also
5746 without @option{-fbranch-probabilities} the traces constructed may not
5747 match the reality and hurt the performance. This only makes
5748 sense when scheduling after register allocation, i.e.@: with
5749 @option{-fschedule-insns2} or at @option{-O2} or higher.
5753 Eliminate redundant sign extension instructions and move the non-redundant
5754 ones to optimal placement using lazy code motion (LCM).
5756 @item -freschedule-modulo-scheduled-loops
5757 @opindex freschedule-modulo-scheduled-loops
5758 The modulo scheduling comes before the traditional scheduling, if a loop
5759 was modulo scheduled we may want to prevent the later scheduling passes
5760 from changing its schedule, we use this option to control that.
5762 @item -fcaller-saves
5763 @opindex fcaller-saves
5764 Enable values to be allocated in registers that will be clobbered by
5765 function calls, by emitting extra instructions to save and restore the
5766 registers around such calls. Such allocation is done only when it
5767 seems to result in better code than would otherwise be produced.
5769 This option is always enabled by default on certain machines, usually
5770 those which have no call-preserved registers to use instead.
5772 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5774 @item -ftree-reassoc
5775 @opindex ftree-reassoc
5776 Perform reassociation on trees. This flag is enabled by default
5777 at @option{-O} and higher.
5781 Perform partial redundancy elimination (PRE) on trees. This flag is
5782 enabled by default at @option{-O2} and @option{-O3}.
5786 Perform full redundancy elimination (FRE) on trees. The difference
5787 between FRE and PRE is that FRE only considers expressions
5788 that are computed on all paths leading to the redundant computation.
5789 This analysis is faster than PRE, though it exposes fewer redundancies.
5790 This flag is enabled by default at @option{-O} and higher.
5792 @item -ftree-copy-prop
5793 @opindex ftree-copy-prop
5794 Perform copy propagation on trees. This pass eliminates unnecessary
5795 copy operations. This flag is enabled by default at @option{-O} and
5798 @item -fipa-pure-const
5799 @opindex fipa-pure-const
5800 Discover which functions are pure or constant.
5801 Enabled by default at @option{-O} and higher.
5803 @item -fipa-reference
5804 @opindex fipa-reference
5805 Discover which static variables do not escape cannot escape the
5807 Enabled by default at @option{-O} and higher.
5809 @item -fipa-struct-reorg
5810 @opindex fipa-struct-reorg
5811 Perform structure reorganization optimization, that change C-like structures
5812 layout in order to better utilize spatial locality. This transformation is
5813 affective for programs containing arrays of structures. Available in two
5814 compilation modes: profile-based (enabled with @option{-fprofile-generate})
5815 or static (which uses built-in heuristics). Require @option{-fipa-type-escape}
5816 to provide the safety of this transformation. It works only in whole program
5817 mode, so it requires @option{-fwhole-program} and @option{-combine} to be
5818 enabled. Structures considered @samp{cold} by this transformation are not
5819 affected (see @option{--param struct-reorg-cold-struct-ratio=@var{value}}).
5821 With this flag, the program debug info reflects a new structure layout.
5825 Perform interprocedural pointer analysis.
5829 Perform interprocedural constant propagation.
5830 This optimization analyzes the program to determine when values passed
5831 to functions are constants and then optimizes accordingly.
5832 This optimization can substantially increase performance
5833 if the application has constants passed to functions, but
5834 because this optimization can create multiple copies of functions,
5835 it may significantly increase code size.
5837 @item -fipa-matrix-reorg
5838 @opindex fipa-matrix-reorg
5839 Perform matrix flattening and transposing.
5840 Matrix flattening tries to replace a m-dimensional matrix
5841 with its equivalent n-dimensional matrix, where n < m.
5842 This reduces the level of indirection needed for accessing the elements
5843 of the matrix. The second optimization is matrix transposing that
5844 attemps to change the order of the matrix's dimensions in order to
5845 improve cache locality.
5846 Both optimizations need fwhole-program flag.
5847 Transposing is enabled only if profiling information is avaliable.
5852 Perform forward store motion on trees. This flag is
5853 enabled by default at @option{-O} and higher.
5857 Perform sparse conditional constant propagation (CCP) on trees. This
5858 pass only operates on local scalar variables and is enabled by default
5859 at @option{-O} and higher.
5861 @item -ftree-store-ccp
5862 @opindex ftree-store-ccp
5863 Perform sparse conditional constant propagation (CCP) on trees. This
5864 pass operates on both local scalar variables and memory stores and
5865 loads (global variables, structures, arrays, etc). This flag is
5866 enabled by default at @option{-O2} and higher.
5870 Perform dead code elimination (DCE) on trees. This flag is enabled by
5871 default at @option{-O} and higher.
5873 @item -ftree-dominator-opts
5874 @opindex ftree-dominator-opts
5875 Perform a variety of simple scalar cleanups (constant/copy
5876 propagation, redundancy elimination, range propagation and expression
5877 simplification) based on a dominator tree traversal. This also
5878 performs jump threading (to reduce jumps to jumps). This flag is
5879 enabled by default at @option{-O} and higher.
5883 Perform dead store elimination (DSE) on trees. A dead store is a store into
5884 a memory location which will later be overwritten by another store without
5885 any intervening loads. In this case the earlier store can be deleted. This
5886 flag is enabled by default at @option{-O} and higher.
5890 Perform loop header copying on trees. This is beneficial since it increases
5891 effectiveness of code motion optimizations. It also saves one jump. This flag
5892 is enabled by default at @option{-O} and higher. It is not enabled
5893 for @option{-Os}, since it usually increases code size.
5895 @item -ftree-loop-optimize
5896 @opindex ftree-loop-optimize
5897 Perform loop optimizations on trees. This flag is enabled by default
5898 at @option{-O} and higher.
5900 @item -ftree-loop-linear
5901 @opindex ftree-loop-linear
5902 Perform linear loop transformations on tree. This flag can improve cache
5903 performance and allow further loop optimizations to take place.
5905 @item -fcheck-data-deps
5906 @opindex fcheck-data-deps
5907 Compare the results of several data dependence analyzers. This option
5908 is used for debugging the data dependence analyzers.
5910 @item -ftree-loop-distribution
5911 Perform loop distribution. This flag can improve cache performance on
5912 big loop bodies and allow further loop optimizations, like
5913 parallelization or vectorization, to take place. For example, the loop
5930 @item -ftree-loop-im
5931 @opindex ftree-loop-im
5932 Perform loop invariant motion on trees. This pass moves only invariants that
5933 would be hard to handle at RTL level (function calls, operations that expand to
5934 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5935 operands of conditions that are invariant out of the loop, so that we can use
5936 just trivial invariantness analysis in loop unswitching. The pass also includes
5939 @item -ftree-loop-ivcanon
5940 @opindex ftree-loop-ivcanon
5941 Create a canonical counter for number of iterations in the loop for that
5942 determining number of iterations requires complicated analysis. Later
5943 optimizations then may determine the number easily. Useful especially
5944 in connection with unrolling.
5948 Perform induction variable optimizations (strength reduction, induction
5949 variable merging and induction variable elimination) on trees.
5951 @item -ftree-parallelize-loops=n
5952 @opindex ftree-parallelize-loops
5953 Parallelize loops, i.e., split their iteration space to run in n threads.
5954 This is only possible for loops whose iterations are independent
5955 and can be arbitrarily reordered. The optimization is only
5956 profitable on multiprocessor machines, for loops that are CPU-intensive,
5957 rather than constrained e.g.@: by memory bandwidth. This option
5958 implies @option{-pthread}, and thus is only supported on targets
5959 that have support for @option{-pthread}.
5963 Perform scalar replacement of aggregates. This pass replaces structure
5964 references with scalars to prevent committing structures to memory too
5965 early. This flag is enabled by default at @option{-O} and higher.
5967 @item -ftree-copyrename
5968 @opindex ftree-copyrename
5969 Perform copy renaming on trees. This pass attempts to rename compiler
5970 temporaries to other variables at copy locations, usually resulting in
5971 variable names which more closely resemble the original variables. This flag
5972 is enabled by default at @option{-O} and higher.
5976 Perform temporary expression replacement during the SSA->normal phase. Single
5977 use/single def temporaries are replaced at their use location with their
5978 defining expression. This results in non-GIMPLE code, but gives the expanders
5979 much more complex trees to work on resulting in better RTL generation. This is
5980 enabled by default at @option{-O} and higher.
5982 @item -ftree-vectorize
5983 @opindex ftree-vectorize
5984 Perform loop vectorization on trees. This flag is enabled by default at
5987 @item -ftree-vect-loop-version
5988 @opindex ftree-vect-loop-version
5989 Perform loop versioning when doing loop vectorization on trees. When a loop
5990 appears to be vectorizable except that data alignment or data dependence cannot
5991 be determined at compile time then vectorized and non-vectorized versions of
5992 the loop are generated along with runtime checks for alignment or dependence
5993 to control which version is executed. This option is enabled by default
5994 except at level @option{-Os} where it is disabled.
5996 @item -fvect-cost-model
5997 @opindex fvect-cost-model
5998 Enable cost model for vectorization.
6002 Perform Value Range Propagation on trees. This is similar to the
6003 constant propagation pass, but instead of values, ranges of values are
6004 propagated. This allows the optimizers to remove unnecessary range
6005 checks like array bound checks and null pointer checks. This is
6006 enabled by default at @option{-O2} and higher. Null pointer check
6007 elimination is only done if @option{-fdelete-null-pointer-checks} is
6012 Perform tail duplication to enlarge superblock size. This transformation
6013 simplifies the control flow of the function allowing other optimizations to do
6016 @item -funroll-loops
6017 @opindex funroll-loops
6018 Unroll loops whose number of iterations can be determined at compile
6019 time or upon entry to the loop. @option{-funroll-loops} implies
6020 @option{-frerun-cse-after-loop}. This option makes code larger,
6021 and may or may not make it run faster.
6023 @item -funroll-all-loops
6024 @opindex funroll-all-loops
6025 Unroll all loops, even if their number of iterations is uncertain when
6026 the loop is entered. This usually makes programs run more slowly.
6027 @option{-funroll-all-loops} implies the same options as
6028 @option{-funroll-loops},
6030 @item -fsplit-ivs-in-unroller
6031 @opindex fsplit-ivs-in-unroller
6032 Enables expressing of values of induction variables in later iterations
6033 of the unrolled loop using the value in the first iteration. This breaks
6034 long dependency chains, thus improving efficiency of the scheduling passes.
6036 Combination of @option{-fweb} and CSE is often sufficient to obtain the
6037 same effect. However in cases the loop body is more complicated than
6038 a single basic block, this is not reliable. It also does not work at all
6039 on some of the architectures due to restrictions in the CSE pass.
6041 This optimization is enabled by default.
6043 @item -fvariable-expansion-in-unroller
6044 @opindex fvariable-expansion-in-unroller
6045 With this option, the compiler will create multiple copies of some
6046 local variables when unrolling a loop which can result in superior code.
6048 @item -fpredictive-commoning
6049 @opindex fpredictive-commoning
6050 Perform predictive commoning optimization, i.e., reusing computations
6051 (especially memory loads and stores) performed in previous
6052 iterations of loops.
6054 This option is enabled at level @option{-O3}.
6056 @item -fprefetch-loop-arrays
6057 @opindex fprefetch-loop-arrays
6058 If supported by the target machine, generate instructions to prefetch
6059 memory to improve the performance of loops that access large arrays.
6061 This option may generate better or worse code; results are highly
6062 dependent on the structure of loops within the source code.
6064 Disabled at level @option{-Os}.
6067 @itemx -fno-peephole2
6068 @opindex fno-peephole
6069 @opindex fno-peephole2
6070 Disable any machine-specific peephole optimizations. The difference
6071 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
6072 are implemented in the compiler; some targets use one, some use the
6073 other, a few use both.
6075 @option{-fpeephole} is enabled by default.
6076 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6078 @item -fno-guess-branch-probability
6079 @opindex fno-guess-branch-probability
6080 Do not guess branch probabilities using heuristics.
6082 GCC will use heuristics to guess branch probabilities if they are
6083 not provided by profiling feedback (@option{-fprofile-arcs}). These
6084 heuristics are based on the control flow graph. If some branch probabilities
6085 are specified by @samp{__builtin_expect}, then the heuristics will be
6086 used to guess branch probabilities for the rest of the control flow graph,
6087 taking the @samp{__builtin_expect} info into account. The interactions
6088 between the heuristics and @samp{__builtin_expect} can be complex, and in
6089 some cases, it may be useful to disable the heuristics so that the effects
6090 of @samp{__builtin_expect} are easier to understand.
6092 The default is @option{-fguess-branch-probability} at levels
6093 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6095 @item -freorder-blocks
6096 @opindex freorder-blocks
6097 Reorder basic blocks in the compiled function in order to reduce number of
6098 taken branches and improve code locality.
6100 Enabled at levels @option{-O2}, @option{-O3}.
6102 @item -freorder-blocks-and-partition
6103 @opindex freorder-blocks-and-partition
6104 In addition to reordering basic blocks in the compiled function, in order
6105 to reduce number of taken branches, partitions hot and cold basic blocks
6106 into separate sections of the assembly and .o files, to improve
6107 paging and cache locality performance.
6109 This optimization is automatically turned off in the presence of
6110 exception handling, for linkonce sections, for functions with a user-defined
6111 section attribute and on any architecture that does not support named
6114 @item -freorder-functions
6115 @opindex freorder-functions
6116 Reorder functions in the object file in order to
6117 improve code locality. This is implemented by using special
6118 subsections @code{.text.hot} for most frequently executed functions and
6119 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
6120 the linker so object file format must support named sections and linker must
6121 place them in a reasonable way.
6123 Also profile feedback must be available in to make this option effective. See
6124 @option{-fprofile-arcs} for details.
6126 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6128 @item -fstrict-aliasing
6129 @opindex fstrict-aliasing
6130 Allows the compiler to assume the strictest aliasing rules applicable to
6131 the language being compiled. For C (and C++), this activates
6132 optimizations based on the type of expressions. In particular, an
6133 object of one type is assumed never to reside at the same address as an
6134 object of a different type, unless the types are almost the same. For
6135 example, an @code{unsigned int} can alias an @code{int}, but not a
6136 @code{void*} or a @code{double}. A character type may alias any other
6139 Pay special attention to code like this:
6152 The practice of reading from a different union member than the one most
6153 recently written to (called ``type-punning'') is common. Even with
6154 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
6155 is accessed through the union type. So, the code above will work as
6156 expected. However, this code might not:
6167 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6169 @item -fstrict-overflow
6170 @opindex fstrict-overflow
6171 Allow the compiler to assume strict signed overflow rules, depending
6172 on the language being compiled. For C (and C++) this means that
6173 overflow when doing arithmetic with signed numbers is undefined, which
6174 means that the compiler may assume that it will not happen. This
6175 permits various optimizations. For example, the compiler will assume
6176 that an expression like @code{i + 10 > i} will always be true for
6177 signed @code{i}. This assumption is only valid if signed overflow is
6178 undefined, as the expression is false if @code{i + 10} overflows when
6179 using twos complement arithmetic. When this option is in effect any
6180 attempt to determine whether an operation on signed numbers will
6181 overflow must be written carefully to not actually involve overflow.
6183 This option also allows the compiler to assume strict pointer
6184 semantics: given a pointer to an object, if adding an offset to that
6185 pointer does not produce a pointer to the same object, the addition is
6186 undefined. This permits the compiler to conclude that @code{p + u >
6187 p} is always true for a pointer @code{p} and unsigned integer
6188 @code{u}. This assumption is only valid because pointer wraparound is
6189 undefined, as the expression is false if @code{p + u} overflows using
6190 twos complement arithmetic.
6192 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
6193 that integer signed overflow is fully defined: it wraps. When
6194 @option{-fwrapv} is used, there is no difference between
6195 @option{-fstrict-overflow} and @option{-fno-strict-overflow} for
6196 integers. With @option{-fwrapv} certain types of overflow are
6197 permitted. For example, if the compiler gets an overflow when doing
6198 arithmetic on constants, the overflowed value can still be used with
6199 @option{-fwrapv}, but not otherwise.
6201 The @option{-fstrict-overflow} option is enabled at levels
6202 @option{-O2}, @option{-O3}, @option{-Os}.
6204 @item -falign-functions
6205 @itemx -falign-functions=@var{n}
6206 @opindex falign-functions
6207 Align the start of functions to the next power-of-two greater than
6208 @var{n}, skipping up to @var{n} bytes. For instance,
6209 @option{-falign-functions=32} aligns functions to the next 32-byte
6210 boundary, but @option{-falign-functions=24} would align to the next
6211 32-byte boundary only if this can be done by skipping 23 bytes or less.
6213 @option{-fno-align-functions} and @option{-falign-functions=1} are
6214 equivalent and mean that functions will not be aligned.
6216 Some assemblers only support this flag when @var{n} is a power of two;
6217 in that case, it is rounded up.
6219 If @var{n} is not specified or is zero, use a machine-dependent default.
6221 Enabled at levels @option{-O2}, @option{-O3}.
6223 @item -falign-labels
6224 @itemx -falign-labels=@var{n}
6225 @opindex falign-labels
6226 Align all branch targets to a power-of-two boundary, skipping up to
6227 @var{n} bytes like @option{-falign-functions}. This option can easily
6228 make code slower, because it must insert dummy operations for when the
6229 branch target is reached in the usual flow of the code.
6231 @option{-fno-align-labels} and @option{-falign-labels=1} are
6232 equivalent and mean that labels will not be aligned.
6234 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
6235 are greater than this value, then their values are used instead.
6237 If @var{n} is not specified or is zero, use a machine-dependent default
6238 which is very likely to be @samp{1}, meaning no alignment.
6240 Enabled at levels @option{-O2}, @option{-O3}.
6243 @itemx -falign-loops=@var{n}
6244 @opindex falign-loops
6245 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
6246 like @option{-falign-functions}. The hope is that the loop will be
6247 executed many times, which will make up for any execution of the dummy
6250 @option{-fno-align-loops} and @option{-falign-loops=1} are
6251 equivalent and mean that loops will not be aligned.
6253 If @var{n} is not specified or is zero, use a machine-dependent default.
6255 Enabled at levels @option{-O2}, @option{-O3}.
6258 @itemx -falign-jumps=@var{n}
6259 @opindex falign-jumps
6260 Align branch targets to a power-of-two boundary, for branch targets
6261 where the targets can only be reached by jumping, skipping up to @var{n}
6262 bytes like @option{-falign-functions}. In this case, no dummy operations
6265 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
6266 equivalent and mean that loops will not be aligned.
6268 If @var{n} is not specified or is zero, use a machine-dependent default.
6270 Enabled at levels @option{-O2}, @option{-O3}.
6272 @item -funit-at-a-time
6273 @opindex funit-at-a-time
6274 Parse the whole compilation unit before starting to produce code.
6275 This allows some extra optimizations to take place but consumes
6276 more memory (in general). There are some compatibility issues
6277 with @emph{unit-at-a-time} mode:
6280 enabling @emph{unit-at-a-time} mode may change the order
6281 in which functions, variables, and top-level @code{asm} statements
6282 are emitted, and will likely break code relying on some particular
6283 ordering. The majority of such top-level @code{asm} statements,
6284 though, can be replaced by @code{section} attributes. The
6285 @option{fno-toplevel-reorder} option may be used to keep the ordering
6286 used in the input file, at the cost of some optimizations.
6289 @emph{unit-at-a-time} mode removes unreferenced static variables
6290 and functions. This may result in undefined references
6291 when an @code{asm} statement refers directly to variables or functions
6292 that are otherwise unused. In that case either the variable/function
6293 shall be listed as an operand of the @code{asm} statement operand or,
6294 in the case of top-level @code{asm} statements the attribute @code{used}
6295 shall be used on the declaration.
6298 Static functions now can use non-standard passing conventions that
6299 may break @code{asm} statements calling functions directly. Again,
6300 attribute @code{used} will prevent this behavior.
6303 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
6304 but this scheme may not be supported by future releases of GCC@.
6306 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6308 @item -fno-toplevel-reorder
6309 @opindex fno-toplevel-reorder
6310 Do not reorder top-level functions, variables, and @code{asm}
6311 statements. Output them in the same order that they appear in the
6312 input file. When this option is used, unreferenced static variables
6313 will not be removed. This option is intended to support existing code
6314 which relies on a particular ordering. For new code, it is better to
6319 Constructs webs as commonly used for register allocation purposes and assign
6320 each web individual pseudo register. This allows the register allocation pass
6321 to operate on pseudos directly, but also strengthens several other optimization
6322 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
6323 however, make debugging impossible, since variables will no longer stay in a
6326 Enabled by default with @option{-funroll-loops}.
6328 @item -fwhole-program
6329 @opindex fwhole-program
6330 Assume that the current compilation unit represents whole program being
6331 compiled. All public functions and variables with the exception of @code{main}
6332 and those merged by attribute @code{externally_visible} become static functions
6333 and in a affect gets more aggressively optimized by interprocedural optimizers.
6334 While this option is equivalent to proper use of @code{static} keyword for
6335 programs consisting of single file, in combination with option
6336 @option{--combine} this flag can be used to compile most of smaller scale C
6337 programs since the functions and variables become local for the whole combined
6338 compilation unit, not for the single source file itself.
6340 This option is not supported for Fortran programs.
6342 @item -fcprop-registers
6343 @opindex fcprop-registers
6344 After register allocation and post-register allocation instruction splitting,
6345 we perform a copy-propagation pass to try to reduce scheduling dependencies
6346 and occasionally eliminate the copy.
6348 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6350 @item -fprofile-dir=@var{path}
6351 @opindex fprofile-dir
6353 Set the directory to search the profile data files in to @var{path}.
6354 This option affects only the profile data generated by
6355 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
6356 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
6357 and its related options.
6358 By default, GCC will use the current directory as @var{path}
6359 thus the profile data file will appear in the same directory as the object file.
6361 @item -fprofile-generate
6362 @itemx -fprofile-generate=@var{path}
6363 @opindex fprofile-generate
6365 Enable options usually used for instrumenting application to produce
6366 profile useful for later recompilation with profile feedback based
6367 optimization. You must use @option{-fprofile-generate} both when
6368 compiling and when linking your program.
6370 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
6372 If @var{path} is specified, GCC will look at the @var{path} to find
6373 the profile feeedback data files. See @option{-fprofile-dir}.
6376 @itemx -fprofile-use=@var{path}
6377 @opindex fprofile-use
6378 Enable profile feedback directed optimizations, and optimizations
6379 generally profitable only with profile feedback available.
6381 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
6382 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
6384 By default, GCC emits an error message if the feedback profiles do not
6385 match the source code. This error can be turned into a warning by using
6386 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
6389 If @var{path} is specified, GCC will look at the @var{path} to find
6390 the profile feedback data files. See @option{-fprofile-dir}.
6393 The following options control compiler behavior regarding floating
6394 point arithmetic. These options trade off between speed and
6395 correctness. All must be specifically enabled.
6399 @opindex ffloat-store
6400 Do not store floating point variables in registers, and inhibit other
6401 options that might change whether a floating point value is taken from a
6404 @cindex floating point precision
6405 This option prevents undesirable excess precision on machines such as
6406 the 68000 where the floating registers (of the 68881) keep more
6407 precision than a @code{double} is supposed to have. Similarly for the
6408 x86 architecture. For most programs, the excess precision does only
6409 good, but a few programs rely on the precise definition of IEEE floating
6410 point. Use @option{-ffloat-store} for such programs, after modifying
6411 them to store all pertinent intermediate computations into variables.
6415 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
6416 @option{-ffinite-math-only}, @option{-fno-rounding-math},
6417 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
6419 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
6421 This option is not turned on by any @option{-O} option since
6422 it can result in incorrect output for programs which depend on
6423 an exact implementation of IEEE or ISO rules/specifications for
6424 math functions. It may, however, yield faster code for programs
6425 that do not require the guarantees of these specifications.
6427 @item -fno-math-errno
6428 @opindex fno-math-errno
6429 Do not set ERRNO after calling math functions that are executed
6430 with a single instruction, e.g., sqrt. A program that relies on
6431 IEEE exceptions for math error handling may want to use this flag
6432 for speed while maintaining IEEE arithmetic compatibility.
6434 This option is not turned on by any @option{-O} option since
6435 it can result in incorrect output for programs which depend on
6436 an exact implementation of IEEE or ISO rules/specifications for
6437 math functions. It may, however, yield faster code for programs
6438 that do not require the guarantees of these specifications.
6440 The default is @option{-fmath-errno}.
6442 On Darwin systems, the math library never sets @code{errno}. There is
6443 therefore no reason for the compiler to consider the possibility that
6444 it might, and @option{-fno-math-errno} is the default.
6446 @item -funsafe-math-optimizations
6447 @opindex funsafe-math-optimizations
6449 Allow optimizations for floating-point arithmetic that (a) assume
6450 that arguments and results are valid and (b) may violate IEEE or
6451 ANSI standards. When used at link-time, it may include libraries
6452 or startup files that change the default FPU control word or other
6453 similar optimizations.
6455 This option is not turned on by any @option{-O} option since
6456 it can result in incorrect output for programs which depend on
6457 an exact implementation of IEEE or ISO rules/specifications for
6458 math functions. It may, however, yield faster code for programs
6459 that do not require the guarantees of these specifications.
6460 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
6461 @option{-fassociative-math} and @option{-freciprocal-math}.
6463 The default is @option{-fno-unsafe-math-optimizations}.
6465 @item -fassociative-math
6466 @opindex fassociative-math
6468 Allow re-association of operands in series of floating-point operations.
6469 This violates the ISO C and C++ language standard by possibly changing
6470 computation result. NOTE: re-ordering may change the sign of zero as
6471 well as ignore NaNs and inhibit or create underflow or overflow (and
6472 thus cannot be used on a code which relies on rounding behavior like
6473 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
6474 and thus may not be used when ordered comparisons are required.
6475 This option requires that both @option{-fno-signed-zeros} and
6476 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
6477 much sense with @option{-frounding-math}.
6479 The default is @option{-fno-associative-math}.
6481 @item -freciprocal-math
6482 @opindex freciprocal-math
6484 Allow the reciprocal of a value to be used instead of dividing by
6485 the value if this enables optimizations. For example @code{x / y}
6486 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
6487 is subject to common subexpression elimination. Note that this loses
6488 precision and increases the number of flops operating on the value.
6490 The default is @option{-fno-reciprocal-math}.
6492 @item -ffinite-math-only
6493 @opindex ffinite-math-only
6494 Allow optimizations for floating-point arithmetic that assume
6495 that arguments and results are not NaNs or +-Infs.
6497 This option is not turned on by any @option{-O} option since
6498 it can result in incorrect output for programs which depend on
6499 an exact implementation of IEEE or ISO rules/specifications for
6500 math functions. It may, however, yield faster code for programs
6501 that do not require the guarantees of these specifications.
6503 The default is @option{-fno-finite-math-only}.
6505 @item -fno-signed-zeros
6506 @opindex fno-signed-zeros
6507 Allow optimizations for floating point arithmetic that ignore the
6508 signedness of zero. IEEE arithmetic specifies the behavior of
6509 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
6510 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
6511 This option implies that the sign of a zero result isn't significant.
6513 The default is @option{-fsigned-zeros}.
6515 @item -fno-trapping-math
6516 @opindex fno-trapping-math
6517 Compile code assuming that floating-point operations cannot generate
6518 user-visible traps. These traps include division by zero, overflow,
6519 underflow, inexact result and invalid operation. This option requires
6520 that @option{-fno-signaling-nans} be in effect. Setting this option may
6521 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
6523 This option should never be turned on by any @option{-O} option since
6524 it can result in incorrect output for programs which depend on
6525 an exact implementation of IEEE or ISO rules/specifications for
6528 The default is @option{-ftrapping-math}.
6530 @item -frounding-math
6531 @opindex frounding-math
6532 Disable transformations and optimizations that assume default floating
6533 point rounding behavior. This is round-to-zero for all floating point
6534 to integer conversions, and round-to-nearest for all other arithmetic
6535 truncations. This option should be specified for programs that change
6536 the FP rounding mode dynamically, or that may be executed with a
6537 non-default rounding mode. This option disables constant folding of
6538 floating point expressions at compile-time (which may be affected by
6539 rounding mode) and arithmetic transformations that are unsafe in the
6540 presence of sign-dependent rounding modes.
6542 The default is @option{-fno-rounding-math}.
6544 This option is experimental and does not currently guarantee to
6545 disable all GCC optimizations that are affected by rounding mode.
6546 Future versions of GCC may provide finer control of this setting
6547 using C99's @code{FENV_ACCESS} pragma. This command line option
6548 will be used to specify the default state for @code{FENV_ACCESS}.
6550 @item -frtl-abstract-sequences
6551 @opindex frtl-abstract-sequences
6552 It is a size optimization method. This option is to find identical
6553 sequences of code, which can be turned into pseudo-procedures and
6554 then replace all occurrences with calls to the newly created
6555 subroutine. It is kind of an opposite of @option{-finline-functions}.
6556 This optimization runs at RTL level.
6558 @item -fsignaling-nans
6559 @opindex fsignaling-nans
6560 Compile code assuming that IEEE signaling NaNs may generate user-visible
6561 traps during floating-point operations. Setting this option disables
6562 optimizations that may change the number of exceptions visible with
6563 signaling NaNs. This option implies @option{-ftrapping-math}.
6565 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
6568 The default is @option{-fno-signaling-nans}.
6570 This option is experimental and does not currently guarantee to
6571 disable all GCC optimizations that affect signaling NaN behavior.
6573 @item -fsingle-precision-constant
6574 @opindex fsingle-precision-constant
6575 Treat floating point constant as single precision constant instead of
6576 implicitly converting it to double precision constant.
6578 @item -fcx-limited-range
6579 @opindex fcx-limited-range
6580 When enabled, this option states that a range reduction step is not
6581 needed when performing complex division. Also, there is no checking
6582 whether the result of a complex multiplication or division is @code{NaN
6583 + I*NaN}, with an attempt to rescue the situation in that case. The
6584 default is @option{-fno-cx-limited-range}, but is enabled by
6585 @option{-ffast-math}.
6587 This option controls the default setting of the ISO C99
6588 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
6591 @item -fcx-fortran-rules
6592 @opindex fcx-fortran-rules
6593 Complex multiplication and division follow Fortran rules. Range
6594 reduction is done as part of complex division, but there is no checking
6595 whether the result of a complex multiplication or division is @code{NaN
6596 + I*NaN}, with an attempt to rescue the situation in that case.
6598 The default is @option{-fno-cx-fortran-rules}.
6602 The following options control optimizations that may improve
6603 performance, but are not enabled by any @option{-O} options. This
6604 section includes experimental options that may produce broken code.
6607 @item -fbranch-probabilities
6608 @opindex fbranch-probabilities
6609 After running a program compiled with @option{-fprofile-arcs}
6610 (@pxref{Debugging Options,, Options for Debugging Your Program or
6611 @command{gcc}}), you can compile it a second time using
6612 @option{-fbranch-probabilities}, to improve optimizations based on
6613 the number of times each branch was taken. When the program
6614 compiled with @option{-fprofile-arcs} exits it saves arc execution
6615 counts to a file called @file{@var{sourcename}.gcda} for each source
6616 file. The information in this data file is very dependent on the
6617 structure of the generated code, so you must use the same source code
6618 and the same optimization options for both compilations.
6620 With @option{-fbranch-probabilities}, GCC puts a
6621 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
6622 These can be used to improve optimization. Currently, they are only
6623 used in one place: in @file{reorg.c}, instead of guessing which path a
6624 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
6625 exactly determine which path is taken more often.
6627 @item -fprofile-values
6628 @opindex fprofile-values
6629 If combined with @option{-fprofile-arcs}, it adds code so that some
6630 data about values of expressions in the program is gathered.
6632 With @option{-fbranch-probabilities}, it reads back the data gathered
6633 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
6634 notes to instructions for their later usage in optimizations.
6636 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
6640 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
6641 a code to gather information about values of expressions.
6643 With @option{-fbranch-probabilities}, it reads back the data gathered
6644 and actually performs the optimizations based on them.
6645 Currently the optimizations include specialization of division operation
6646 using the knowledge about the value of the denominator.
6648 @item -frename-registers
6649 @opindex frename-registers
6650 Attempt to avoid false dependencies in scheduled code by making use
6651 of registers left over after register allocation. This optimization
6652 will most benefit processors with lots of registers. Depending on the
6653 debug information format adopted by the target, however, it can
6654 make debugging impossible, since variables will no longer stay in
6655 a ``home register''.
6657 Enabled by default with @option{-funroll-loops}.
6661 Perform tail duplication to enlarge superblock size. This transformation
6662 simplifies the control flow of the function allowing other optimizations to do
6665 Enabled with @option{-fprofile-use}.
6667 @item -funroll-loops
6668 @opindex funroll-loops
6669 Unroll loops whose number of iterations can be determined at compile time or
6670 upon entry to the loop. @option{-funroll-loops} implies
6671 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
6672 It also turns on complete loop peeling (i.e.@: complete removal of loops with
6673 small constant number of iterations). This option makes code larger, and may
6674 or may not make it run faster.
6676 Enabled with @option{-fprofile-use}.
6678 @item -funroll-all-loops
6679 @opindex funroll-all-loops
6680 Unroll all loops, even if their number of iterations is uncertain when
6681 the loop is entered. This usually makes programs run more slowly.
6682 @option{-funroll-all-loops} implies the same options as
6683 @option{-funroll-loops}.
6686 @opindex fpeel-loops
6687 Peels the loops for that there is enough information that they do not
6688 roll much (from profile feedback). It also turns on complete loop peeling
6689 (i.e.@: complete removal of loops with small constant number of iterations).
6691 Enabled with @option{-fprofile-use}.
6693 @item -fmove-loop-invariants
6694 @opindex fmove-loop-invariants
6695 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
6696 at level @option{-O1}
6698 @item -funswitch-loops
6699 @opindex funswitch-loops
6700 Move branches with loop invariant conditions out of the loop, with duplicates
6701 of the loop on both branches (modified according to result of the condition).
6703 @item -ffunction-sections
6704 @itemx -fdata-sections
6705 @opindex ffunction-sections
6706 @opindex fdata-sections
6707 Place each function or data item into its own section in the output
6708 file if the target supports arbitrary sections. The name of the
6709 function or the name of the data item determines the section's name
6712 Use these options on systems where the linker can perform optimizations
6713 to improve locality of reference in the instruction space. Most systems
6714 using the ELF object format and SPARC processors running Solaris 2 have
6715 linkers with such optimizations. AIX may have these optimizations in
6718 Only use these options when there are significant benefits from doing
6719 so. When you specify these options, the assembler and linker will
6720 create larger object and executable files and will also be slower.
6721 You will not be able to use @code{gprof} on all systems if you
6722 specify this option and you may have problems with debugging if
6723 you specify both this option and @option{-g}.
6725 @item -fbranch-target-load-optimize
6726 @opindex fbranch-target-load-optimize
6727 Perform branch target register load optimization before prologue / epilogue
6729 The use of target registers can typically be exposed only during reload,
6730 thus hoisting loads out of loops and doing inter-block scheduling needs
6731 a separate optimization pass.
6733 @item -fbranch-target-load-optimize2
6734 @opindex fbranch-target-load-optimize2
6735 Perform branch target register load optimization after prologue / epilogue
6738 @item -fbtr-bb-exclusive
6739 @opindex fbtr-bb-exclusive
6740 When performing branch target register load optimization, don't reuse
6741 branch target registers in within any basic block.
6743 @item -fstack-protector
6744 @opindex fstack-protector
6745 Emit extra code to check for buffer overflows, such as stack smashing
6746 attacks. This is done by adding a guard variable to functions with
6747 vulnerable objects. This includes functions that call alloca, and
6748 functions with buffers larger than 8 bytes. The guards are initialized
6749 when a function is entered and then checked when the function exits.
6750 If a guard check fails, an error message is printed and the program exits.
6752 @item -fstack-protector-all
6753 @opindex fstack-protector-all
6754 Like @option{-fstack-protector} except that all functions are protected.
6756 @item -fsection-anchors
6757 @opindex fsection-anchors
6758 Try to reduce the number of symbolic address calculations by using
6759 shared ``anchor'' symbols to address nearby objects. This transformation
6760 can help to reduce the number of GOT entries and GOT accesses on some
6763 For example, the implementation of the following function @code{foo}:
6767 int foo (void) @{ return a + b + c; @}
6770 would usually calculate the addresses of all three variables, but if you
6771 compile it with @option{-fsection-anchors}, it will access the variables
6772 from a common anchor point instead. The effect is similar to the
6773 following pseudocode (which isn't valid C):
6778 register int *xr = &x;
6779 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
6783 Not all targets support this option.
6785 @item --param @var{name}=@var{value}
6787 In some places, GCC uses various constants to control the amount of
6788 optimization that is done. For example, GCC will not inline functions
6789 that contain more that a certain number of instructions. You can
6790 control some of these constants on the command-line using the
6791 @option{--param} option.
6793 The names of specific parameters, and the meaning of the values, are
6794 tied to the internals of the compiler, and are subject to change
6795 without notice in future releases.
6797 In each case, the @var{value} is an integer. The allowable choices for
6798 @var{name} are given in the following table:
6801 @item sra-max-structure-size
6802 The maximum structure size, in bytes, at which the scalar replacement
6803 of aggregates (SRA) optimization will perform block copies. The
6804 default value, 0, implies that GCC will select the most appropriate
6807 @item sra-field-structure-ratio
6808 The threshold ratio (as a percentage) between instantiated fields and
6809 the complete structure size. We say that if the ratio of the number
6810 of bytes in instantiated fields to the number of bytes in the complete
6811 structure exceeds this parameter, then block copies are not used. The
6814 @item struct-reorg-cold-struct-ratio
6815 The threshold ratio (as a percentage) between a structure frequency
6816 and the frequency of the hottest structure in the program. This parameter
6817 is used by struct-reorg optimization enabled by @option{-fipa-struct-reorg}.
6818 We say that if the ratio of a structure frequency, calculated by profiling,
6819 to the hottest structure frequency in the program is less than this
6820 parameter, then structure reorganization is not applied to this structure.
6823 @item max-crossjump-edges
6824 The maximum number of incoming edges to consider for crossjumping.
6825 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
6826 the number of edges incoming to each block. Increasing values mean
6827 more aggressive optimization, making the compile time increase with
6828 probably small improvement in executable size.
6830 @item min-crossjump-insns
6831 The minimum number of instructions which must be matched at the end
6832 of two blocks before crossjumping will be performed on them. This
6833 value is ignored in the case where all instructions in the block being
6834 crossjumped from are matched. The default value is 5.
6836 @item max-grow-copy-bb-insns
6837 The maximum code size expansion factor when copying basic blocks
6838 instead of jumping. The expansion is relative to a jump instruction.
6839 The default value is 8.
6841 @item max-goto-duplication-insns
6842 The maximum number of instructions to duplicate to a block that jumps
6843 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
6844 passes, GCC factors computed gotos early in the compilation process,
6845 and unfactors them as late as possible. Only computed jumps at the
6846 end of a basic blocks with no more than max-goto-duplication-insns are
6847 unfactored. The default value is 8.
6849 @item max-delay-slot-insn-search
6850 The maximum number of instructions to consider when looking for an
6851 instruction to fill a delay slot. If more than this arbitrary number of
6852 instructions is searched, the time savings from filling the delay slot
6853 will be minimal so stop searching. Increasing values mean more
6854 aggressive optimization, making the compile time increase with probably
6855 small improvement in executable run time.
6857 @item max-delay-slot-live-search
6858 When trying to fill delay slots, the maximum number of instructions to
6859 consider when searching for a block with valid live register
6860 information. Increasing this arbitrarily chosen value means more
6861 aggressive optimization, increasing the compile time. This parameter
6862 should be removed when the delay slot code is rewritten to maintain the
6865 @item max-gcse-memory
6866 The approximate maximum amount of memory that will be allocated in
6867 order to perform the global common subexpression elimination
6868 optimization. If more memory than specified is required, the
6869 optimization will not be done.
6871 @item max-gcse-passes
6872 The maximum number of passes of GCSE to run. The default is 1.
6874 @item max-pending-list-length
6875 The maximum number of pending dependencies scheduling will allow
6876 before flushing the current state and starting over. Large functions
6877 with few branches or calls can create excessively large lists which
6878 needlessly consume memory and resources.
6880 @item max-inline-insns-single
6881 Several parameters control the tree inliner used in gcc.
6882 This number sets the maximum number of instructions (counted in GCC's
6883 internal representation) in a single function that the tree inliner
6884 will consider for inlining. This only affects functions declared
6885 inline and methods implemented in a class declaration (C++).
6886 The default value is 450.
6888 @item max-inline-insns-auto
6889 When you use @option{-finline-functions} (included in @option{-O3}),
6890 a lot of functions that would otherwise not be considered for inlining
6891 by the compiler will be investigated. To those functions, a different
6892 (more restrictive) limit compared to functions declared inline can
6894 The default value is 90.
6896 @item large-function-insns
6897 The limit specifying really large functions. For functions larger than this
6898 limit after inlining inlining is constrained by
6899 @option{--param large-function-growth}. This parameter is useful primarily
6900 to avoid extreme compilation time caused by non-linear algorithms used by the
6902 This parameter is ignored when @option{-funit-at-a-time} is not used.
6903 The default value is 2700.
6905 @item large-function-growth
6906 Specifies maximal growth of large function caused by inlining in percents.
6907 This parameter is ignored when @option{-funit-at-a-time} is not used.
6908 The default value is 100 which limits large function growth to 2.0 times
6911 @item large-unit-insns
6912 The limit specifying large translation unit. Growth caused by inlining of
6913 units larger than this limit is limited by @option{--param inline-unit-growth}.
6914 For small units this might be too tight (consider unit consisting of function A
6915 that is inline and B that just calls A three time. If B is small relative to
6916 A, the growth of unit is 300\% and yet such inlining is very sane. For very
6917 large units consisting of small inlineable functions however the overall unit
6918 growth limit is needed to avoid exponential explosion of code size. Thus for
6919 smaller units, the size is increased to @option{--param large-unit-insns}
6920 before applying @option{--param inline-unit-growth}. The default is 10000
6922 @item inline-unit-growth
6923 Specifies maximal overall growth of the compilation unit caused by inlining.
6924 This parameter is ignored when @option{-funit-at-a-time} is not used.
6925 The default value is 30 which limits unit growth to 1.3 times the original
6928 @item large-stack-frame
6929 The limit specifying large stack frames. While inlining the algorithm is trying
6930 to not grow past this limit too much. Default value is 256 bytes.
6932 @item large-stack-frame-growth
6933 Specifies maximal growth of large stack frames caused by inlining in percents.
6934 The default value is 1000 which limits large stack frame growth to 11 times
6937 @item max-inline-insns-recursive
6938 @itemx max-inline-insns-recursive-auto
6939 Specifies maximum number of instructions out-of-line copy of self recursive inline
6940 function can grow into by performing recursive inlining.
6942 For functions declared inline @option{--param max-inline-insns-recursive} is
6943 taken into account. For function not declared inline, recursive inlining
6944 happens only when @option{-finline-functions} (included in @option{-O3}) is
6945 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
6946 default value is 450.
6948 @item max-inline-recursive-depth
6949 @itemx max-inline-recursive-depth-auto
6950 Specifies maximum recursion depth used by the recursive inlining.
6952 For functions declared inline @option{--param max-inline-recursive-depth} is
6953 taken into account. For function not declared inline, recursive inlining
6954 happens only when @option{-finline-functions} (included in @option{-O3}) is
6955 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
6958 @item min-inline-recursive-probability
6959 Recursive inlining is profitable only for function having deep recursion
6960 in average and can hurt for function having little recursion depth by
6961 increasing the prologue size or complexity of function body to other
6964 When profile feedback is available (see @option{-fprofile-generate}) the actual
6965 recursion depth can be guessed from probability that function will recurse via
6966 given call expression. This parameter limits inlining only to call expression
6967 whose probability exceeds given threshold (in percents). The default value is
6970 @item inline-call-cost
6971 Specify cost of call instruction relative to simple arithmetics operations
6972 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
6973 functions and at the same time increases size of leaf function that is believed to
6974 reduce function size by being inlined. In effect it increases amount of
6975 inlining for code having large abstraction penalty (many functions that just
6976 pass the arguments to other functions) and decrease inlining for code with low
6977 abstraction penalty. The default value is 12.
6979 @item min-vect-loop-bound
6980 The minimum number of iterations under which a loop will not get vectorized
6981 when @option{-ftree-vectorize} is used. The number of iterations after
6982 vectorization needs to be greater than the value specified by this option
6983 to allow vectorization. The default value is 0.
6985 @item max-unrolled-insns
6986 The maximum number of instructions that a loop should have if that loop
6987 is unrolled, and if the loop is unrolled, it determines how many times
6988 the loop code is unrolled.
6990 @item max-average-unrolled-insns
6991 The maximum number of instructions biased by probabilities of their execution
6992 that a loop should have if that loop is unrolled, and if the loop is unrolled,
6993 it determines how many times the loop code is unrolled.
6995 @item max-unroll-times
6996 The maximum number of unrollings of a single loop.
6998 @item max-peeled-insns
6999 The maximum number of instructions that a loop should have if that loop
7000 is peeled, and if the loop is peeled, it determines how many times
7001 the loop code is peeled.
7003 @item max-peel-times
7004 The maximum number of peelings of a single loop.
7006 @item max-completely-peeled-insns
7007 The maximum number of insns of a completely peeled loop.
7009 @item max-completely-peel-times
7010 The maximum number of iterations of a loop to be suitable for complete peeling.
7012 @item max-unswitch-insns
7013 The maximum number of insns of an unswitched loop.
7015 @item max-unswitch-level
7016 The maximum number of branches unswitched in a single loop.
7019 The minimum cost of an expensive expression in the loop invariant motion.
7021 @item iv-consider-all-candidates-bound
7022 Bound on number of candidates for induction variables below that
7023 all candidates are considered for each use in induction variable
7024 optimizations. Only the most relevant candidates are considered
7025 if there are more candidates, to avoid quadratic time complexity.
7027 @item iv-max-considered-uses
7028 The induction variable optimizations give up on loops that contain more
7029 induction variable uses.
7031 @item iv-always-prune-cand-set-bound
7032 If number of candidates in the set is smaller than this value,
7033 we always try to remove unnecessary ivs from the set during its
7034 optimization when a new iv is added to the set.
7036 @item scev-max-expr-size
7037 Bound on size of expressions used in the scalar evolutions analyzer.
7038 Large expressions slow the analyzer.
7040 @item omega-max-vars
7041 The maximum number of variables in an Omega constraint system.
7042 The default value is 128.
7044 @item omega-max-geqs
7045 The maximum number of inequalities in an Omega constraint system.
7046 The default value is 256.
7049 The maximum number of equalities in an Omega constraint system.
7050 The default value is 128.
7052 @item omega-max-wild-cards
7053 The maximum number of wildcard variables that the Omega solver will
7054 be able to insert. The default value is 18.
7056 @item omega-hash-table-size
7057 The size of the hash table in the Omega solver. The default value is
7060 @item omega-max-keys
7061 The maximal number of keys used by the Omega solver. The default
7064 @item omega-eliminate-redundant-constraints
7065 When set to 1, use expensive methods to eliminate all redundant
7066 constraints. The default value is 0.
7068 @item vect-max-version-for-alignment-checks
7069 The maximum number of runtime checks that can be performed when
7070 doing loop versioning for alignment in the vectorizer. See option
7071 ftree-vect-loop-version for more information.
7073 @item vect-max-version-for-alias-checks
7074 The maximum number of runtime checks that can be performed when
7075 doing loop versioning for alias in the vectorizer. See option
7076 ftree-vect-loop-version for more information.
7078 @item max-iterations-to-track
7080 The maximum number of iterations of a loop the brute force algorithm
7081 for analysis of # of iterations of the loop tries to evaluate.
7083 @item hot-bb-count-fraction
7084 Select fraction of the maximal count of repetitions of basic block in program
7085 given basic block needs to have to be considered hot.
7087 @item hot-bb-frequency-fraction
7088 Select fraction of the maximal frequency of executions of basic block in
7089 function given basic block needs to have to be considered hot
7091 @item max-predicted-iterations
7092 The maximum number of loop iterations we predict statically. This is useful
7093 in cases where function contain single loop with known bound and other loop
7094 with unknown. We predict the known number of iterations correctly, while
7095 the unknown number of iterations average to roughly 10. This means that the
7096 loop without bounds would appear artificially cold relative to the other one.
7098 @item align-threshold
7100 Select fraction of the maximal frequency of executions of basic block in
7101 function given basic block will get aligned.
7103 @item align-loop-iterations
7105 A loop expected to iterate at lest the selected number of iterations will get
7108 @item tracer-dynamic-coverage
7109 @itemx tracer-dynamic-coverage-feedback
7111 This value is used to limit superblock formation once the given percentage of
7112 executed instructions is covered. This limits unnecessary code size
7115 The @option{tracer-dynamic-coverage-feedback} is used only when profile
7116 feedback is available. The real profiles (as opposed to statically estimated
7117 ones) are much less balanced allowing the threshold to be larger value.
7119 @item tracer-max-code-growth
7120 Stop tail duplication once code growth has reached given percentage. This is
7121 rather hokey argument, as most of the duplicates will be eliminated later in
7122 cross jumping, so it may be set to much higher values than is the desired code
7125 @item tracer-min-branch-ratio
7127 Stop reverse growth when the reverse probability of best edge is less than this
7128 threshold (in percent).
7130 @item tracer-min-branch-ratio
7131 @itemx tracer-min-branch-ratio-feedback
7133 Stop forward growth if the best edge do have probability lower than this
7136 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
7137 compilation for profile feedback and one for compilation without. The value
7138 for compilation with profile feedback needs to be more conservative (higher) in
7139 order to make tracer effective.
7141 @item max-cse-path-length
7143 Maximum number of basic blocks on path that cse considers. The default is 10.
7146 The maximum instructions CSE process before flushing. The default is 1000.
7148 @item max-aliased-vops
7150 Maximum number of virtual operands per function allowed to represent
7151 aliases before triggering the alias partitioning heuristic. Alias
7152 partitioning reduces compile times and memory consumption needed for
7153 aliasing at the expense of precision loss in alias information. The
7154 default value for this parameter is 100 for -O1, 500 for -O2 and 1000
7157 Notice that if a function contains more memory statements than the
7158 value of this parameter, it is not really possible to achieve this
7159 reduction. In this case, the compiler will use the number of memory
7160 statements as the value for @option{max-aliased-vops}.
7162 @item avg-aliased-vops
7164 Average number of virtual operands per statement allowed to represent
7165 aliases before triggering the alias partitioning heuristic. This
7166 works in conjunction with @option{max-aliased-vops}. If a function
7167 contains more than @option{max-aliased-vops} virtual operators, then
7168 memory symbols will be grouped into memory partitions until either the
7169 total number of virtual operators is below @option{max-aliased-vops}
7170 or the average number of virtual operators per memory statement is
7171 below @option{avg-aliased-vops}. The default value for this parameter
7172 is 1 for -O1 and -O2, and 3 for -O3.
7174 @item ggc-min-expand
7176 GCC uses a garbage collector to manage its own memory allocation. This
7177 parameter specifies the minimum percentage by which the garbage
7178 collector's heap should be allowed to expand between collections.
7179 Tuning this may improve compilation speed; it has no effect on code
7182 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
7183 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
7184 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
7185 GCC is not able to calculate RAM on a particular platform, the lower
7186 bound of 30% is used. Setting this parameter and
7187 @option{ggc-min-heapsize} to zero causes a full collection to occur at
7188 every opportunity. This is extremely slow, but can be useful for
7191 @item ggc-min-heapsize
7193 Minimum size of the garbage collector's heap before it begins bothering
7194 to collect garbage. The first collection occurs after the heap expands
7195 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
7196 tuning this may improve compilation speed, and has no effect on code
7199 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
7200 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
7201 with a lower bound of 4096 (four megabytes) and an upper bound of
7202 131072 (128 megabytes). If GCC is not able to calculate RAM on a
7203 particular platform, the lower bound is used. Setting this parameter
7204 very large effectively disables garbage collection. Setting this
7205 parameter and @option{ggc-min-expand} to zero causes a full collection
7206 to occur at every opportunity.
7208 @item max-reload-search-insns
7209 The maximum number of instruction reload should look backward for equivalent
7210 register. Increasing values mean more aggressive optimization, making the
7211 compile time increase with probably slightly better performance. The default
7214 @item max-cselib-memory-locations
7215 The maximum number of memory locations cselib should take into account.
7216 Increasing values mean more aggressive optimization, making the compile time
7217 increase with probably slightly better performance. The default value is 500.
7219 @item reorder-blocks-duplicate
7220 @itemx reorder-blocks-duplicate-feedback
7222 Used by basic block reordering pass to decide whether to use unconditional
7223 branch or duplicate the code on its destination. Code is duplicated when its
7224 estimated size is smaller than this value multiplied by the estimated size of
7225 unconditional jump in the hot spots of the program.
7227 The @option{reorder-block-duplicate-feedback} is used only when profile
7228 feedback is available and may be set to higher values than
7229 @option{reorder-block-duplicate} since information about the hot spots is more
7232 @item max-sched-ready-insns
7233 The maximum number of instructions ready to be issued the scheduler should
7234 consider at any given time during the first scheduling pass. Increasing
7235 values mean more thorough searches, making the compilation time increase
7236 with probably little benefit. The default value is 100.
7238 @item max-sched-region-blocks
7239 The maximum number of blocks in a region to be considered for
7240 interblock scheduling. The default value is 10.
7242 @item max-sched-region-insns
7243 The maximum number of insns in a region to be considered for
7244 interblock scheduling. The default value is 100.
7247 The minimum probability (in percents) of reaching a source block
7248 for interblock speculative scheduling. The default value is 40.
7250 @item max-sched-extend-regions-iters
7251 The maximum number of iterations through CFG to extend regions.
7252 0 - disable region extension,
7253 N - do at most N iterations.
7254 The default value is 0.
7256 @item max-sched-insn-conflict-delay
7257 The maximum conflict delay for an insn to be considered for speculative motion.
7258 The default value is 3.
7260 @item sched-spec-prob-cutoff
7261 The minimal probability of speculation success (in percents), so that
7262 speculative insn will be scheduled.
7263 The default value is 40.
7265 @item max-last-value-rtl
7267 The maximum size measured as number of RTLs that can be recorded in an expression
7268 in combiner for a pseudo register as last known value of that register. The default
7271 @item integer-share-limit
7272 Small integer constants can use a shared data structure, reducing the
7273 compiler's memory usage and increasing its speed. This sets the maximum
7274 value of a shared integer constant. The default value is 256.
7276 @item min-virtual-mappings
7277 Specifies the minimum number of virtual mappings in the incremental
7278 SSA updater that should be registered to trigger the virtual mappings
7279 heuristic defined by virtual-mappings-ratio. The default value is
7282 @item virtual-mappings-ratio
7283 If the number of virtual mappings is virtual-mappings-ratio bigger
7284 than the number of virtual symbols to be updated, then the incremental
7285 SSA updater switches to a full update for those symbols. The default
7288 @item ssp-buffer-size
7289 The minimum size of buffers (i.e.@: arrays) that will receive stack smashing
7290 protection when @option{-fstack-protection} is used.
7292 @item max-jump-thread-duplication-stmts
7293 Maximum number of statements allowed in a block that needs to be
7294 duplicated when threading jumps.
7296 @item max-fields-for-field-sensitive
7297 Maximum number of fields in a structure we will treat in
7298 a field sensitive manner during pointer analysis.
7300 @item prefetch-latency
7301 Estimate on average number of instructions that are executed before
7302 prefetch finishes. The distance we prefetch ahead is proportional
7303 to this constant. Increasing this number may also lead to less
7304 streams being prefetched (see @option{simultaneous-prefetches}).
7306 @item simultaneous-prefetches
7307 Maximum number of prefetches that can run at the same time.
7309 @item l1-cache-line-size
7310 The size of cache line in L1 cache, in bytes.
7313 The size of L1 cache, in kilobytes.
7316 The size of L2 cache, in kilobytes.
7318 @item use-canonical-types
7319 Whether the compiler should use the ``canonical'' type system. By
7320 default, this should always be 1, which uses a more efficient internal
7321 mechanism for comparing types in C++ and Objective-C++. However, if
7322 bugs in the canonical type system are causing compilation failures,
7323 set this value to 0 to disable canonical types.
7325 @item max-partial-antic-length
7326 Maximum length of the partial antic set computed during the tree
7327 partial redundancy elimination optimization (@option{-ftree-pre}) when
7328 optimizing at @option{-O3} and above. For some sorts of source code
7329 the enhanced partial redundancy elimination optimization can run away,
7330 consuming all of the memory available on the host machine. This
7331 parameter sets a limit on the length of the sets that are computed,
7332 which prevents the runaway behaviour. Setting a value of 0 for
7333 this paramter will allow an unlimited set length.
7335 @item sccvn-max-scc-size
7336 Maximum size of a strongly connected component (SCC) during SCCVN
7337 processing. If this limit is hit, SCCVN processing for the whole
7338 function will not be done and optimizations depending on it will
7339 be disabled. The default maximum SCC size is 10000.
7344 @node Preprocessor Options
7345 @section Options Controlling the Preprocessor
7346 @cindex preprocessor options
7347 @cindex options, preprocessor
7349 These options control the C preprocessor, which is run on each C source
7350 file before actual compilation.
7352 If you use the @option{-E} option, nothing is done except preprocessing.
7353 Some of these options make sense only together with @option{-E} because
7354 they cause the preprocessor output to be unsuitable for actual
7359 You can use @option{-Wp,@var{option}} to bypass the compiler driver
7360 and pass @var{option} directly through to the preprocessor. If
7361 @var{option} contains commas, it is split into multiple options at the
7362 commas. However, many options are modified, translated or interpreted
7363 by the compiler driver before being passed to the preprocessor, and
7364 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
7365 interface is undocumented and subject to change, so whenever possible
7366 you should avoid using @option{-Wp} and let the driver handle the
7369 @item -Xpreprocessor @var{option}
7370 @opindex preprocessor
7371 Pass @var{option} as an option to the preprocessor. You can use this to
7372 supply system-specific preprocessor options which GCC does not know how to
7375 If you want to pass an option that takes an argument, you must use
7376 @option{-Xpreprocessor} twice, once for the option and once for the argument.
7379 @include cppopts.texi
7381 @node Assembler Options
7382 @section Passing Options to the Assembler
7384 @c prevent bad page break with this line
7385 You can pass options to the assembler.
7388 @item -Wa,@var{option}
7390 Pass @var{option} as an option to the assembler. If @var{option}
7391 contains commas, it is split into multiple options at the commas.
7393 @item -Xassembler @var{option}
7395 Pass @var{option} as an option to the assembler. You can use this to
7396 supply system-specific assembler options which GCC does not know how to
7399 If you want to pass an option that takes an argument, you must use
7400 @option{-Xassembler} twice, once for the option and once for the argument.
7405 @section Options for Linking
7406 @cindex link options
7407 @cindex options, linking
7409 These options come into play when the compiler links object files into
7410 an executable output file. They are meaningless if the compiler is
7411 not doing a link step.
7415 @item @var{object-file-name}
7416 A file name that does not end in a special recognized suffix is
7417 considered to name an object file or library. (Object files are
7418 distinguished from libraries by the linker according to the file
7419 contents.) If linking is done, these object files are used as input
7428 If any of these options is used, then the linker is not run, and
7429 object file names should not be used as arguments. @xref{Overall
7433 @item -l@var{library}
7434 @itemx -l @var{library}
7436 Search the library named @var{library} when linking. (The second
7437 alternative with the library as a separate argument is only for
7438 POSIX compliance and is not recommended.)
7440 It makes a difference where in the command you write this option; the
7441 linker searches and processes libraries and object files in the order they
7442 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
7443 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
7444 to functions in @samp{z}, those functions may not be loaded.
7446 The linker searches a standard list of directories for the library,
7447 which is actually a file named @file{lib@var{library}.a}. The linker
7448 then uses this file as if it had been specified precisely by name.
7450 The directories searched include several standard system directories
7451 plus any that you specify with @option{-L}.
7453 Normally the files found this way are library files---archive files
7454 whose members are object files. The linker handles an archive file by
7455 scanning through it for members which define symbols that have so far
7456 been referenced but not defined. But if the file that is found is an
7457 ordinary object file, it is linked in the usual fashion. The only
7458 difference between using an @option{-l} option and specifying a file name
7459 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
7460 and searches several directories.
7464 You need this special case of the @option{-l} option in order to
7465 link an Objective-C or Objective-C++ program.
7468 @opindex nostartfiles
7469 Do not use the standard system startup files when linking.
7470 The standard system libraries are used normally, unless @option{-nostdlib}
7471 or @option{-nodefaultlibs} is used.
7473 @item -nodefaultlibs
7474 @opindex nodefaultlibs
7475 Do not use the standard system libraries when linking.
7476 Only the libraries you specify will be passed to the linker.
7477 The standard startup files are used normally, unless @option{-nostartfiles}
7478 is used. The compiler may generate calls to @code{memcmp},
7479 @code{memset}, @code{memcpy} and @code{memmove}.
7480 These entries are usually resolved by entries in
7481 libc. These entry points should be supplied through some other
7482 mechanism when this option is specified.
7486 Do not use the standard system startup files or libraries when linking.
7487 No startup files and only the libraries you specify will be passed to
7488 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
7489 @code{memcpy} and @code{memmove}.
7490 These entries are usually resolved by entries in
7491 libc. These entry points should be supplied through some other
7492 mechanism when this option is specified.
7494 @cindex @option{-lgcc}, use with @option{-nostdlib}
7495 @cindex @option{-nostdlib} and unresolved references
7496 @cindex unresolved references and @option{-nostdlib}
7497 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
7498 @cindex @option{-nodefaultlibs} and unresolved references
7499 @cindex unresolved references and @option{-nodefaultlibs}
7500 One of the standard libraries bypassed by @option{-nostdlib} and
7501 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
7502 that GCC uses to overcome shortcomings of particular machines, or special
7503 needs for some languages.
7504 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
7505 Collection (GCC) Internals},
7506 for more discussion of @file{libgcc.a}.)
7507 In most cases, you need @file{libgcc.a} even when you want to avoid
7508 other standard libraries. In other words, when you specify @option{-nostdlib}
7509 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
7510 This ensures that you have no unresolved references to internal GCC
7511 library subroutines. (For example, @samp{__main}, used to ensure C++
7512 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
7513 GNU Compiler Collection (GCC) Internals}.)
7517 Produce a position independent executable on targets which support it.
7518 For predictable results, you must also specify the same set of options
7519 that were used to generate code (@option{-fpie}, @option{-fPIE},
7520 or model suboptions) when you specify this option.
7524 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
7525 that support it. This instructs the linker to add all symbols, not
7526 only used ones, to the dynamic symbol table. This option is needed
7527 for some uses of @code{dlopen} or to allow obtaining backtraces
7528 from within a program.
7532 Remove all symbol table and relocation information from the executable.
7536 On systems that support dynamic linking, this prevents linking with the shared
7537 libraries. On other systems, this option has no effect.
7541 Produce a shared object which can then be linked with other objects to
7542 form an executable. Not all systems support this option. For predictable
7543 results, you must also specify the same set of options that were used to
7544 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
7545 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
7546 needs to build supplementary stub code for constructors to work. On
7547 multi-libbed systems, @samp{gcc -shared} must select the correct support
7548 libraries to link against. Failing to supply the correct flags may lead
7549 to subtle defects. Supplying them in cases where they are not necessary
7552 @item -shared-libgcc
7553 @itemx -static-libgcc
7554 @opindex shared-libgcc
7555 @opindex static-libgcc
7556 On systems that provide @file{libgcc} as a shared library, these options
7557 force the use of either the shared or static version respectively.
7558 If no shared version of @file{libgcc} was built when the compiler was
7559 configured, these options have no effect.
7561 There are several situations in which an application should use the
7562 shared @file{libgcc} instead of the static version. The most common
7563 of these is when the application wishes to throw and catch exceptions
7564 across different shared libraries. In that case, each of the libraries
7565 as well as the application itself should use the shared @file{libgcc}.
7567 Therefore, the G++ and GCJ drivers automatically add
7568 @option{-shared-libgcc} whenever you build a shared library or a main
7569 executable, because C++ and Java programs typically use exceptions, so
7570 this is the right thing to do.
7572 If, instead, you use the GCC driver to create shared libraries, you may
7573 find that they will not always be linked with the shared @file{libgcc}.
7574 If GCC finds, at its configuration time, that you have a non-GNU linker
7575 or a GNU linker that does not support option @option{--eh-frame-hdr},
7576 it will link the shared version of @file{libgcc} into shared libraries
7577 by default. Otherwise, it will take advantage of the linker and optimize
7578 away the linking with the shared version of @file{libgcc}, linking with
7579 the static version of libgcc by default. This allows exceptions to
7580 propagate through such shared libraries, without incurring relocation
7581 costs at library load time.
7583 However, if a library or main executable is supposed to throw or catch
7584 exceptions, you must link it using the G++ or GCJ driver, as appropriate
7585 for the languages used in the program, or using the option
7586 @option{-shared-libgcc}, such that it is linked with the shared
7591 Bind references to global symbols when building a shared object. Warn
7592 about any unresolved references (unless overridden by the link editor
7593 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
7596 @item -Xlinker @var{option}
7598 Pass @var{option} as an option to the linker. You can use this to
7599 supply system-specific linker options which GCC does not know how to
7602 If you want to pass an option that takes an argument, you must use
7603 @option{-Xlinker} twice, once for the option and once for the argument.
7604 For example, to pass @option{-assert definitions}, you must write
7605 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
7606 @option{-Xlinker "-assert definitions"}, because this passes the entire
7607 string as a single argument, which is not what the linker expects.
7609 @item -Wl,@var{option}
7611 Pass @var{option} as an option to the linker. If @var{option} contains
7612 commas, it is split into multiple options at the commas.
7614 @item -u @var{symbol}
7616 Pretend the symbol @var{symbol} is undefined, to force linking of
7617 library modules to define it. You can use @option{-u} multiple times with
7618 different symbols to force loading of additional library modules.
7621 @node Directory Options
7622 @section Options for Directory Search
7623 @cindex directory options
7624 @cindex options, directory search
7627 These options specify directories to search for header files, for
7628 libraries and for parts of the compiler:
7633 Add the directory @var{dir} to the head of the list of directories to be
7634 searched for header files. This can be used to override a system header
7635 file, substituting your own version, since these directories are
7636 searched before the system header file directories. However, you should
7637 not use this option to add directories that contain vendor-supplied
7638 system header files (use @option{-isystem} for that). If you use more than
7639 one @option{-I} option, the directories are scanned in left-to-right
7640 order; the standard system directories come after.
7642 If a standard system include directory, or a directory specified with
7643 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
7644 option will be ignored. The directory will still be searched but as a
7645 system directory at its normal position in the system include chain.
7646 This is to ensure that GCC's procedure to fix buggy system headers and
7647 the ordering for the include_next directive are not inadvertently changed.
7648 If you really need to change the search order for system directories,
7649 use the @option{-nostdinc} and/or @option{-isystem} options.
7651 @item -iquote@var{dir}
7653 Add the directory @var{dir} to the head of the list of directories to
7654 be searched for header files only for the case of @samp{#include
7655 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
7656 otherwise just like @option{-I}.
7660 Add directory @var{dir} to the list of directories to be searched
7663 @item -B@var{prefix}
7665 This option specifies where to find the executables, libraries,
7666 include files, and data files of the compiler itself.
7668 The compiler driver program runs one or more of the subprograms
7669 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
7670 @var{prefix} as a prefix for each program it tries to run, both with and
7671 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
7673 For each subprogram to be run, the compiler driver first tries the
7674 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
7675 was not specified, the driver tries two standard prefixes, which are
7676 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
7677 those results in a file name that is found, the unmodified program
7678 name is searched for using the directories specified in your
7679 @env{PATH} environment variable.
7681 The compiler will check to see if the path provided by the @option{-B}
7682 refers to a directory, and if necessary it will add a directory
7683 separator character at the end of the path.
7685 @option{-B} prefixes that effectively specify directory names also apply
7686 to libraries in the linker, because the compiler translates these
7687 options into @option{-L} options for the linker. They also apply to
7688 includes files in the preprocessor, because the compiler translates these
7689 options into @option{-isystem} options for the preprocessor. In this case,
7690 the compiler appends @samp{include} to the prefix.
7692 The run-time support file @file{libgcc.a} can also be searched for using
7693 the @option{-B} prefix, if needed. If it is not found there, the two
7694 standard prefixes above are tried, and that is all. The file is left
7695 out of the link if it is not found by those means.
7697 Another way to specify a prefix much like the @option{-B} prefix is to use
7698 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
7701 As a special kludge, if the path provided by @option{-B} is
7702 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
7703 9, then it will be replaced by @file{[dir/]include}. This is to help
7704 with boot-strapping the compiler.
7706 @item -specs=@var{file}
7708 Process @var{file} after the compiler reads in the standard @file{specs}
7709 file, in order to override the defaults that the @file{gcc} driver
7710 program uses when determining what switches to pass to @file{cc1},
7711 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
7712 @option{-specs=@var{file}} can be specified on the command line, and they
7713 are processed in order, from left to right.
7715 @item --sysroot=@var{dir}
7717 Use @var{dir} as the logical root directory for headers and libraries.
7718 For example, if the compiler would normally search for headers in
7719 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
7720 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
7722 If you use both this option and the @option{-isysroot} option, then
7723 the @option{--sysroot} option will apply to libraries, but the
7724 @option{-isysroot} option will apply to header files.
7726 The GNU linker (beginning with version 2.16) has the necessary support
7727 for this option. If your linker does not support this option, the
7728 header file aspect of @option{--sysroot} will still work, but the
7729 library aspect will not.
7733 This option has been deprecated. Please use @option{-iquote} instead for
7734 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
7735 Any directories you specify with @option{-I} options before the @option{-I-}
7736 option are searched only for the case of @samp{#include "@var{file}"};
7737 they are not searched for @samp{#include <@var{file}>}.
7739 If additional directories are specified with @option{-I} options after
7740 the @option{-I-}, these directories are searched for all @samp{#include}
7741 directives. (Ordinarily @emph{all} @option{-I} directories are used
7744 In addition, the @option{-I-} option inhibits the use of the current
7745 directory (where the current input file came from) as the first search
7746 directory for @samp{#include "@var{file}"}. There is no way to
7747 override this effect of @option{-I-}. With @option{-I.} you can specify
7748 searching the directory which was current when the compiler was
7749 invoked. That is not exactly the same as what the preprocessor does
7750 by default, but it is often satisfactory.
7752 @option{-I-} does not inhibit the use of the standard system directories
7753 for header files. Thus, @option{-I-} and @option{-nostdinc} are
7760 @section Specifying subprocesses and the switches to pass to them
7763 @command{gcc} is a driver program. It performs its job by invoking a
7764 sequence of other programs to do the work of compiling, assembling and
7765 linking. GCC interprets its command-line parameters and uses these to
7766 deduce which programs it should invoke, and which command-line options
7767 it ought to place on their command lines. This behavior is controlled
7768 by @dfn{spec strings}. In most cases there is one spec string for each
7769 program that GCC can invoke, but a few programs have multiple spec
7770 strings to control their behavior. The spec strings built into GCC can
7771 be overridden by using the @option{-specs=} command-line switch to specify
7774 @dfn{Spec files} are plaintext files that are used to construct spec
7775 strings. They consist of a sequence of directives separated by blank
7776 lines. The type of directive is determined by the first non-whitespace
7777 character on the line and it can be one of the following:
7780 @item %@var{command}
7781 Issues a @var{command} to the spec file processor. The commands that can
7785 @item %include <@var{file}>
7787 Search for @var{file} and insert its text at the current point in the
7790 @item %include_noerr <@var{file}>
7791 @cindex %include_noerr
7792 Just like @samp{%include}, but do not generate an error message if the include
7793 file cannot be found.
7795 @item %rename @var{old_name} @var{new_name}
7797 Rename the spec string @var{old_name} to @var{new_name}.
7801 @item *[@var{spec_name}]:
7802 This tells the compiler to create, override or delete the named spec
7803 string. All lines after this directive up to the next directive or
7804 blank line are considered to be the text for the spec string. If this
7805 results in an empty string then the spec will be deleted. (Or, if the
7806 spec did not exist, then nothing will happened.) Otherwise, if the spec
7807 does not currently exist a new spec will be created. If the spec does
7808 exist then its contents will be overridden by the text of this
7809 directive, unless the first character of that text is the @samp{+}
7810 character, in which case the text will be appended to the spec.
7812 @item [@var{suffix}]:
7813 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
7814 and up to the next directive or blank line are considered to make up the
7815 spec string for the indicated suffix. When the compiler encounters an
7816 input file with the named suffix, it will processes the spec string in
7817 order to work out how to compile that file. For example:
7824 This says that any input file whose name ends in @samp{.ZZ} should be
7825 passed to the program @samp{z-compile}, which should be invoked with the
7826 command-line switch @option{-input} and with the result of performing the
7827 @samp{%i} substitution. (See below.)
7829 As an alternative to providing a spec string, the text that follows a
7830 suffix directive can be one of the following:
7833 @item @@@var{language}
7834 This says that the suffix is an alias for a known @var{language}. This is
7835 similar to using the @option{-x} command-line switch to GCC to specify a
7836 language explicitly. For example:
7843 Says that .ZZ files are, in fact, C++ source files.
7846 This causes an error messages saying:
7849 @var{name} compiler not installed on this system.
7853 GCC already has an extensive list of suffixes built into it.
7854 This directive will add an entry to the end of the list of suffixes, but
7855 since the list is searched from the end backwards, it is effectively
7856 possible to override earlier entries using this technique.
7860 GCC has the following spec strings built into it. Spec files can
7861 override these strings or create their own. Note that individual
7862 targets can also add their own spec strings to this list.
7865 asm Options to pass to the assembler
7866 asm_final Options to pass to the assembler post-processor
7867 cpp Options to pass to the C preprocessor
7868 cc1 Options to pass to the C compiler
7869 cc1plus Options to pass to the C++ compiler
7870 endfile Object files to include at the end of the link
7871 link Options to pass to the linker
7872 lib Libraries to include on the command line to the linker
7873 libgcc Decides which GCC support library to pass to the linker
7874 linker Sets the name of the linker
7875 predefines Defines to be passed to the C preprocessor
7876 signed_char Defines to pass to CPP to say whether @code{char} is signed
7878 startfile Object files to include at the start of the link
7881 Here is a small example of a spec file:
7887 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
7890 This example renames the spec called @samp{lib} to @samp{old_lib} and
7891 then overrides the previous definition of @samp{lib} with a new one.
7892 The new definition adds in some extra command-line options before
7893 including the text of the old definition.
7895 @dfn{Spec strings} are a list of command-line options to be passed to their
7896 corresponding program. In addition, the spec strings can contain
7897 @samp{%}-prefixed sequences to substitute variable text or to
7898 conditionally insert text into the command line. Using these constructs
7899 it is possible to generate quite complex command lines.
7901 Here is a table of all defined @samp{%}-sequences for spec
7902 strings. Note that spaces are not generated automatically around the
7903 results of expanding these sequences. Therefore you can concatenate them
7904 together or combine them with constant text in a single argument.
7908 Substitute one @samp{%} into the program name or argument.
7911 Substitute the name of the input file being processed.
7914 Substitute the basename of the input file being processed.
7915 This is the substring up to (and not including) the last period
7916 and not including the directory.
7919 This is the same as @samp{%b}, but include the file suffix (text after
7923 Marks the argument containing or following the @samp{%d} as a
7924 temporary file name, so that that file will be deleted if GCC exits
7925 successfully. Unlike @samp{%g}, this contributes no text to the
7928 @item %g@var{suffix}
7929 Substitute a file name that has suffix @var{suffix} and is chosen
7930 once per compilation, and mark the argument in the same way as
7931 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
7932 name is now chosen in a way that is hard to predict even when previously
7933 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
7934 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
7935 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
7936 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
7937 was simply substituted with a file name chosen once per compilation,
7938 without regard to any appended suffix (which was therefore treated
7939 just like ordinary text), making such attacks more likely to succeed.
7941 @item %u@var{suffix}
7942 Like @samp{%g}, but generates a new temporary file name even if
7943 @samp{%u@var{suffix}} was already seen.
7945 @item %U@var{suffix}
7946 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
7947 new one if there is no such last file name. In the absence of any
7948 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
7949 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
7950 would involve the generation of two distinct file names, one
7951 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
7952 simply substituted with a file name chosen for the previous @samp{%u},
7953 without regard to any appended suffix.
7955 @item %j@var{suffix}
7956 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
7957 writable, and if save-temps is off; otherwise, substitute the name
7958 of a temporary file, just like @samp{%u}. This temporary file is not
7959 meant for communication between processes, but rather as a junk
7962 @item %|@var{suffix}
7963 @itemx %m@var{suffix}
7964 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
7965 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
7966 all. These are the two most common ways to instruct a program that it
7967 should read from standard input or write to standard output. If you
7968 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
7969 construct: see for example @file{f/lang-specs.h}.
7971 @item %.@var{SUFFIX}
7972 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
7973 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
7974 terminated by the next space or %.
7977 Marks the argument containing or following the @samp{%w} as the
7978 designated output file of this compilation. This puts the argument
7979 into the sequence of arguments that @samp{%o} will substitute later.
7982 Substitutes the names of all the output files, with spaces
7983 automatically placed around them. You should write spaces
7984 around the @samp{%o} as well or the results are undefined.
7985 @samp{%o} is for use in the specs for running the linker.
7986 Input files whose names have no recognized suffix are not compiled
7987 at all, but they are included among the output files, so they will
7991 Substitutes the suffix for object files. Note that this is
7992 handled specially when it immediately follows @samp{%g, %u, or %U},
7993 because of the need for those to form complete file names. The
7994 handling is such that @samp{%O} is treated exactly as if it had already
7995 been substituted, except that @samp{%g, %u, and %U} do not currently
7996 support additional @var{suffix} characters following @samp{%O} as they would
7997 following, for example, @samp{.o}.
8000 Substitutes the standard macro predefinitions for the
8001 current target machine. Use this when running @code{cpp}.
8004 Like @samp{%p}, but puts @samp{__} before and after the name of each
8005 predefined macro, except for macros that start with @samp{__} or with
8006 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
8010 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
8011 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
8012 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
8013 and @option{-imultilib} as necessary.
8016 Current argument is the name of a library or startup file of some sort.
8017 Search for that file in a standard list of directories and substitute
8018 the full name found.
8021 Print @var{str} as an error message. @var{str} is terminated by a newline.
8022 Use this when inconsistent options are detected.
8025 Substitute the contents of spec string @var{name} at this point.
8028 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
8030 @item %x@{@var{option}@}
8031 Accumulate an option for @samp{%X}.
8034 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
8038 Output the accumulated assembler options specified by @option{-Wa}.
8041 Output the accumulated preprocessor options specified by @option{-Wp}.
8044 Process the @code{asm} spec. This is used to compute the
8045 switches to be passed to the assembler.
8048 Process the @code{asm_final} spec. This is a spec string for
8049 passing switches to an assembler post-processor, if such a program is
8053 Process the @code{link} spec. This is the spec for computing the
8054 command line passed to the linker. Typically it will make use of the
8055 @samp{%L %G %S %D and %E} sequences.
8058 Dump out a @option{-L} option for each directory that GCC believes might
8059 contain startup files. If the target supports multilibs then the
8060 current multilib directory will be prepended to each of these paths.
8063 Process the @code{lib} spec. This is a spec string for deciding which
8064 libraries should be included on the command line to the linker.
8067 Process the @code{libgcc} spec. This is a spec string for deciding
8068 which GCC support library should be included on the command line to the linker.
8071 Process the @code{startfile} spec. This is a spec for deciding which
8072 object files should be the first ones passed to the linker. Typically
8073 this might be a file named @file{crt0.o}.
8076 Process the @code{endfile} spec. This is a spec string that specifies
8077 the last object files that will be passed to the linker.
8080 Process the @code{cpp} spec. This is used to construct the arguments
8081 to be passed to the C preprocessor.
8084 Process the @code{cc1} spec. This is used to construct the options to be
8085 passed to the actual C compiler (@samp{cc1}).
8088 Process the @code{cc1plus} spec. This is used to construct the options to be
8089 passed to the actual C++ compiler (@samp{cc1plus}).
8092 Substitute the variable part of a matched option. See below.
8093 Note that each comma in the substituted string is replaced by
8097 Remove all occurrences of @code{-S} from the command line. Note---this
8098 command is position dependent. @samp{%} commands in the spec string
8099 before this one will see @code{-S}, @samp{%} commands in the spec string
8100 after this one will not.
8102 @item %:@var{function}(@var{args})
8103 Call the named function @var{function}, passing it @var{args}.
8104 @var{args} is first processed as a nested spec string, then split
8105 into an argument vector in the usual fashion. The function returns
8106 a string which is processed as if it had appeared literally as part
8107 of the current spec.
8109 The following built-in spec functions are provided:
8113 The @code{getenv} spec function takes two arguments: an environment
8114 variable name and a string. If the environment variable is not
8115 defined, a fatal error is issued. Otherwise, the return value is the
8116 value of the environment variable concatenated with the string. For
8117 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
8120 %:getenv(TOPDIR /include)
8123 expands to @file{/path/to/top/include}.
8125 @item @code{if-exists}
8126 The @code{if-exists} spec function takes one argument, an absolute
8127 pathname to a file. If the file exists, @code{if-exists} returns the
8128 pathname. Here is a small example of its usage:
8132 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
8135 @item @code{if-exists-else}
8136 The @code{if-exists-else} spec function is similar to the @code{if-exists}
8137 spec function, except that it takes two arguments. The first argument is
8138 an absolute pathname to a file. If the file exists, @code{if-exists-else}
8139 returns the pathname. If it does not exist, it returns the second argument.
8140 This way, @code{if-exists-else} can be used to select one file or another,
8141 based on the existence of the first. Here is a small example of its usage:
8145 crt0%O%s %:if-exists(crti%O%s) \
8146 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
8149 @item @code{replace-outfile}
8150 The @code{replace-outfile} spec function takes two arguments. It looks for the
8151 first argument in the outfiles array and replaces it with the second argument. Here
8152 is a small example of its usage:
8155 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
8158 @item @code{print-asm-header}
8159 The @code{print-asm-header} function takes no arguments and simply
8160 prints a banner like:
8166 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
8169 It is used to separate compiler options from assembler options
8170 in the @option{--target-help} output.
8174 Substitutes the @code{-S} switch, if that switch was given to GCC@.
8175 If that switch was not specified, this substitutes nothing. Note that
8176 the leading dash is omitted when specifying this option, and it is
8177 automatically inserted if the substitution is performed. Thus the spec
8178 string @samp{%@{foo@}} would match the command-line option @option{-foo}
8179 and would output the command line option @option{-foo}.
8181 @item %W@{@code{S}@}
8182 Like %@{@code{S}@} but mark last argument supplied within as a file to be
8185 @item %@{@code{S}*@}
8186 Substitutes all the switches specified to GCC whose names start
8187 with @code{-S}, but which also take an argument. This is used for
8188 switches like @option{-o}, @option{-D}, @option{-I}, etc.
8189 GCC considers @option{-o foo} as being
8190 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
8191 text, including the space. Thus two arguments would be generated.
8193 @item %@{@code{S}*&@code{T}*@}
8194 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
8195 (the order of @code{S} and @code{T} in the spec is not significant).
8196 There can be any number of ampersand-separated variables; for each the
8197 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
8199 @item %@{@code{S}:@code{X}@}
8200 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
8202 @item %@{!@code{S}:@code{X}@}
8203 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
8205 @item %@{@code{S}*:@code{X}@}
8206 Substitutes @code{X} if one or more switches whose names start with
8207 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
8208 once, no matter how many such switches appeared. However, if @code{%*}
8209 appears somewhere in @code{X}, then @code{X} will be substituted once
8210 for each matching switch, with the @code{%*} replaced by the part of
8211 that switch that matched the @code{*}.
8213 @item %@{.@code{S}:@code{X}@}
8214 Substitutes @code{X}, if processing a file with suffix @code{S}.
8216 @item %@{!.@code{S}:@code{X}@}
8217 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
8219 @item %@{,@code{S}:@code{X}@}
8220 Substitutes @code{X}, if processing a file for language @code{S}.
8222 @item %@{!,@code{S}:@code{X}@}
8223 Substitutes @code{X}, if not processing a file for language @code{S}.
8225 @item %@{@code{S}|@code{P}:@code{X}@}
8226 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
8227 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
8228 @code{*} sequences as well, although they have a stronger binding than
8229 the @samp{|}. If @code{%*} appears in @code{X}, all of the
8230 alternatives must be starred, and only the first matching alternative
8233 For example, a spec string like this:
8236 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
8239 will output the following command-line options from the following input
8240 command-line options:
8245 -d fred.c -foo -baz -boggle
8246 -d jim.d -bar -baz -boggle
8249 @item %@{S:X; T:Y; :D@}
8251 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
8252 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
8253 be as many clauses as you need. This may be combined with @code{.},
8254 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
8259 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
8260 construct may contain other nested @samp{%} constructs or spaces, or
8261 even newlines. They are processed as usual, as described above.
8262 Trailing white space in @code{X} is ignored. White space may also
8263 appear anywhere on the left side of the colon in these constructs,
8264 except between @code{.} or @code{*} and the corresponding word.
8266 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
8267 handled specifically in these constructs. If another value of
8268 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
8269 @option{-W} switch is found later in the command line, the earlier
8270 switch value is ignored, except with @{@code{S}*@} where @code{S} is
8271 just one letter, which passes all matching options.
8273 The character @samp{|} at the beginning of the predicate text is used to
8274 indicate that a command should be piped to the following command, but
8275 only if @option{-pipe} is specified.
8277 It is built into GCC which switches take arguments and which do not.
8278 (You might think it would be useful to generalize this to allow each
8279 compiler's spec to say which switches take arguments. But this cannot
8280 be done in a consistent fashion. GCC cannot even decide which input
8281 files have been specified without knowing which switches take arguments,
8282 and it must know which input files to compile in order to tell which
8285 GCC also knows implicitly that arguments starting in @option{-l} are to be
8286 treated as compiler output files, and passed to the linker in their
8287 proper position among the other output files.
8289 @c man begin OPTIONS
8291 @node Target Options
8292 @section Specifying Target Machine and Compiler Version
8293 @cindex target options
8294 @cindex cross compiling
8295 @cindex specifying machine version
8296 @cindex specifying compiler version and target machine
8297 @cindex compiler version, specifying
8298 @cindex target machine, specifying
8300 The usual way to run GCC is to run the executable called @file{gcc}, or
8301 @file{<machine>-gcc} when cross-compiling, or
8302 @file{<machine>-gcc-<version>} to run a version other than the one that
8303 was installed last. Sometimes this is inconvenient, so GCC provides
8304 options that will switch to another cross-compiler or version.
8307 @item -b @var{machine}
8309 The argument @var{machine} specifies the target machine for compilation.
8311 The value to use for @var{machine} is the same as was specified as the
8312 machine type when configuring GCC as a cross-compiler. For
8313 example, if a cross-compiler was configured with @samp{configure
8314 arm-elf}, meaning to compile for an arm processor with elf binaries,
8315 then you would specify @option{-b arm-elf} to run that cross compiler.
8316 Because there are other options beginning with @option{-b}, the
8317 configuration must contain a hyphen.
8319 @item -V @var{version}
8321 The argument @var{version} specifies which version of GCC to run.
8322 This is useful when multiple versions are installed. For example,
8323 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
8326 The @option{-V} and @option{-b} options work by running the
8327 @file{<machine>-gcc-<version>} executable, so there's no real reason to
8328 use them if you can just run that directly.
8330 @node Submodel Options
8331 @section Hardware Models and Configurations
8332 @cindex submodel options
8333 @cindex specifying hardware config
8334 @cindex hardware models and configurations, specifying
8335 @cindex machine dependent options
8337 Earlier we discussed the standard option @option{-b} which chooses among
8338 different installed compilers for completely different target
8339 machines, such as VAX vs.@: 68000 vs.@: 80386.
8341 In addition, each of these target machine types can have its own
8342 special options, starting with @samp{-m}, to choose among various
8343 hardware models or configurations---for example, 68010 vs 68020,
8344 floating coprocessor or none. A single installed version of the
8345 compiler can compile for any model or configuration, according to the
8348 Some configurations of the compiler also support additional special
8349 options, usually for compatibility with other compilers on the same
8352 @c This list is ordered alphanumerically by subsection name.
8353 @c It should be the same order and spelling as these options are listed
8354 @c in Machine Dependent Options
8360 * Blackfin Options::
8364 * DEC Alpha Options::
8365 * DEC Alpha/VMS Options::
8367 * GNU/Linux Options::
8370 * i386 and x86-64 Options::
8383 * RS/6000 and PowerPC Options::
8384 * S/390 and zSeries Options::
8389 * System V Options::
8394 * Xstormy16 Options::
8400 @subsection ARC Options
8403 These options are defined for ARC implementations:
8408 Compile code for little endian mode. This is the default.
8412 Compile code for big endian mode.
8415 @opindex mmangle-cpu
8416 Prepend the name of the cpu to all public symbol names.
8417 In multiple-processor systems, there are many ARC variants with different
8418 instruction and register set characteristics. This flag prevents code
8419 compiled for one cpu to be linked with code compiled for another.
8420 No facility exists for handling variants that are ``almost identical''.
8421 This is an all or nothing option.
8423 @item -mcpu=@var{cpu}
8425 Compile code for ARC variant @var{cpu}.
8426 Which variants are supported depend on the configuration.
8427 All variants support @option{-mcpu=base}, this is the default.
8429 @item -mtext=@var{text-section}
8430 @itemx -mdata=@var{data-section}
8431 @itemx -mrodata=@var{readonly-data-section}
8435 Put functions, data, and readonly data in @var{text-section},
8436 @var{data-section}, and @var{readonly-data-section} respectively
8437 by default. This can be overridden with the @code{section} attribute.
8438 @xref{Variable Attributes}.
8443 @subsection ARM Options
8446 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
8450 @item -mabi=@var{name}
8452 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
8453 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
8456 @opindex mapcs-frame
8457 Generate a stack frame that is compliant with the ARM Procedure Call
8458 Standard for all functions, even if this is not strictly necessary for
8459 correct execution of the code. Specifying @option{-fomit-frame-pointer}
8460 with this option will cause the stack frames not to be generated for
8461 leaf functions. The default is @option{-mno-apcs-frame}.
8465 This is a synonym for @option{-mapcs-frame}.
8468 @c not currently implemented
8469 @item -mapcs-stack-check
8470 @opindex mapcs-stack-check
8471 Generate code to check the amount of stack space available upon entry to
8472 every function (that actually uses some stack space). If there is
8473 insufficient space available then either the function
8474 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
8475 called, depending upon the amount of stack space required. The run time
8476 system is required to provide these functions. The default is
8477 @option{-mno-apcs-stack-check}, since this produces smaller code.
8479 @c not currently implemented
8481 @opindex mapcs-float
8482 Pass floating point arguments using the float point registers. This is
8483 one of the variants of the APCS@. This option is recommended if the
8484 target hardware has a floating point unit or if a lot of floating point
8485 arithmetic is going to be performed by the code. The default is
8486 @option{-mno-apcs-float}, since integer only code is slightly increased in
8487 size if @option{-mapcs-float} is used.
8489 @c not currently implemented
8490 @item -mapcs-reentrant
8491 @opindex mapcs-reentrant
8492 Generate reentrant, position independent code. The default is
8493 @option{-mno-apcs-reentrant}.
8496 @item -mthumb-interwork
8497 @opindex mthumb-interwork
8498 Generate code which supports calling between the ARM and Thumb
8499 instruction sets. Without this option the two instruction sets cannot
8500 be reliably used inside one program. The default is
8501 @option{-mno-thumb-interwork}, since slightly larger code is generated
8502 when @option{-mthumb-interwork} is specified.
8504 @item -mno-sched-prolog
8505 @opindex mno-sched-prolog
8506 Prevent the reordering of instructions in the function prolog, or the
8507 merging of those instruction with the instructions in the function's
8508 body. This means that all functions will start with a recognizable set
8509 of instructions (or in fact one of a choice from a small set of
8510 different function prologues), and this information can be used to
8511 locate the start if functions inside an executable piece of code. The
8512 default is @option{-msched-prolog}.
8515 @opindex mhard-float
8516 Generate output containing floating point instructions. This is the
8520 @opindex msoft-float
8521 Generate output containing library calls for floating point.
8522 @strong{Warning:} the requisite libraries are not available for all ARM
8523 targets. Normally the facilities of the machine's usual C compiler are
8524 used, but this cannot be done directly in cross-compilation. You must make
8525 your own arrangements to provide suitable library functions for
8528 @option{-msoft-float} changes the calling convention in the output file;
8529 therefore, it is only useful if you compile @emph{all} of a program with
8530 this option. In particular, you need to compile @file{libgcc.a}, the
8531 library that comes with GCC, with @option{-msoft-float} in order for
8534 @item -mfloat-abi=@var{name}
8536 Specifies which ABI to use for floating point values. Permissible values
8537 are: @samp{soft}, @samp{softfp} and @samp{hard}.
8539 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
8540 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
8541 of floating point instructions, but still uses the soft-float calling
8544 @item -mlittle-endian
8545 @opindex mlittle-endian
8546 Generate code for a processor running in little-endian mode. This is
8547 the default for all standard configurations.
8550 @opindex mbig-endian
8551 Generate code for a processor running in big-endian mode; the default is
8552 to compile code for a little-endian processor.
8554 @item -mwords-little-endian
8555 @opindex mwords-little-endian
8556 This option only applies when generating code for big-endian processors.
8557 Generate code for a little-endian word order but a big-endian byte
8558 order. That is, a byte order of the form @samp{32107654}. Note: this
8559 option should only be used if you require compatibility with code for
8560 big-endian ARM processors generated by versions of the compiler prior to
8563 @item -mcpu=@var{name}
8565 This specifies the name of the target ARM processor. GCC uses this name
8566 to determine what kind of instructions it can emit when generating
8567 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
8568 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
8569 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
8570 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
8571 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
8572 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
8573 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
8574 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
8575 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
8576 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
8577 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
8578 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
8579 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
8580 @samp{arm1156t2-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
8581 @samp{cortex-a8}, @samp{cortex-r4}, @samp{cortex-m3}, @samp{cortex-m1},
8582 @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
8584 @item -mtune=@var{name}
8586 This option is very similar to the @option{-mcpu=} option, except that
8587 instead of specifying the actual target processor type, and hence
8588 restricting which instructions can be used, it specifies that GCC should
8589 tune the performance of the code as if the target were of the type
8590 specified in this option, but still choosing the instructions that it
8591 will generate based on the cpu specified by a @option{-mcpu=} option.
8592 For some ARM implementations better performance can be obtained by using
8595 @item -march=@var{name}
8597 This specifies the name of the target ARM architecture. GCC uses this
8598 name to determine what kind of instructions it can emit when generating
8599 assembly code. This option can be used in conjunction with or instead
8600 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
8601 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
8602 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
8603 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv6-m},
8604 @samp{armv7}, @samp{armv7-a}, @samp{armv7-r}, @samp{armv7-m},
8605 @samp{iwmmxt}, @samp{ep9312}.
8607 @item -mfpu=@var{name}
8608 @itemx -mfpe=@var{number}
8609 @itemx -mfp=@var{number}
8613 This specifies what floating point hardware (or hardware emulation) is
8614 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
8615 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
8616 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
8617 with older versions of GCC@.
8619 If @option{-msoft-float} is specified this specifies the format of
8620 floating point values.
8622 @item -mstructure-size-boundary=@var{n}
8623 @opindex mstructure-size-boundary
8624 The size of all structures and unions will be rounded up to a multiple
8625 of the number of bits set by this option. Permissible values are 8, 32
8626 and 64. The default value varies for different toolchains. For the COFF
8627 targeted toolchain the default value is 8. A value of 64 is only allowed
8628 if the underlying ABI supports it.
8630 Specifying the larger number can produce faster, more efficient code, but
8631 can also increase the size of the program. Different values are potentially
8632 incompatible. Code compiled with one value cannot necessarily expect to
8633 work with code or libraries compiled with another value, if they exchange
8634 information using structures or unions.
8636 @item -mabort-on-noreturn
8637 @opindex mabort-on-noreturn
8638 Generate a call to the function @code{abort} at the end of a
8639 @code{noreturn} function. It will be executed if the function tries to
8643 @itemx -mno-long-calls
8644 @opindex mlong-calls
8645 @opindex mno-long-calls
8646 Tells the compiler to perform function calls by first loading the
8647 address of the function into a register and then performing a subroutine
8648 call on this register. This switch is needed if the target function
8649 will lie outside of the 64 megabyte addressing range of the offset based
8650 version of subroutine call instruction.
8652 Even if this switch is enabled, not all function calls will be turned
8653 into long calls. The heuristic is that static functions, functions
8654 which have the @samp{short-call} attribute, functions that are inside
8655 the scope of a @samp{#pragma no_long_calls} directive and functions whose
8656 definitions have already been compiled within the current compilation
8657 unit, will not be turned into long calls. The exception to this rule is
8658 that weak function definitions, functions with the @samp{long-call}
8659 attribute or the @samp{section} attribute, and functions that are within
8660 the scope of a @samp{#pragma long_calls} directive, will always be
8661 turned into long calls.
8663 This feature is not enabled by default. Specifying
8664 @option{-mno-long-calls} will restore the default behavior, as will
8665 placing the function calls within the scope of a @samp{#pragma
8666 long_calls_off} directive. Note these switches have no effect on how
8667 the compiler generates code to handle function calls via function
8670 @item -mnop-fun-dllimport
8671 @opindex mnop-fun-dllimport
8672 Disable support for the @code{dllimport} attribute.
8674 @item -msingle-pic-base
8675 @opindex msingle-pic-base
8676 Treat the register used for PIC addressing as read-only, rather than
8677 loading it in the prologue for each function. The run-time system is
8678 responsible for initializing this register with an appropriate value
8679 before execution begins.
8681 @item -mpic-register=@var{reg}
8682 @opindex mpic-register
8683 Specify the register to be used for PIC addressing. The default is R10
8684 unless stack-checking is enabled, when R9 is used.
8686 @item -mcirrus-fix-invalid-insns
8687 @opindex mcirrus-fix-invalid-insns
8688 @opindex mno-cirrus-fix-invalid-insns
8689 Insert NOPs into the instruction stream to in order to work around
8690 problems with invalid Maverick instruction combinations. This option
8691 is only valid if the @option{-mcpu=ep9312} option has been used to
8692 enable generation of instructions for the Cirrus Maverick floating
8693 point co-processor. This option is not enabled by default, since the
8694 problem is only present in older Maverick implementations. The default
8695 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
8698 @item -mpoke-function-name
8699 @opindex mpoke-function-name
8700 Write the name of each function into the text section, directly
8701 preceding the function prologue. The generated code is similar to this:
8705 .ascii "arm_poke_function_name", 0
8708 .word 0xff000000 + (t1 - t0)
8709 arm_poke_function_name
8711 stmfd sp!, @{fp, ip, lr, pc@}
8715 When performing a stack backtrace, code can inspect the value of
8716 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
8717 location @code{pc - 12} and the top 8 bits are set, then we know that
8718 there is a function name embedded immediately preceding this location
8719 and has length @code{((pc[-3]) & 0xff000000)}.
8723 Generate code for the Thumb instruction set. The default is to
8724 use the 32-bit ARM instruction set.
8725 This option automatically enables either 16-bit Thumb-1 or
8726 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
8727 and @option{-march=@var{name}} options.
8730 @opindex mtpcs-frame
8731 Generate a stack frame that is compliant with the Thumb Procedure Call
8732 Standard for all non-leaf functions. (A leaf function is one that does
8733 not call any other functions.) The default is @option{-mno-tpcs-frame}.
8735 @item -mtpcs-leaf-frame
8736 @opindex mtpcs-leaf-frame
8737 Generate a stack frame that is compliant with the Thumb Procedure Call
8738 Standard for all leaf functions. (A leaf function is one that does
8739 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
8741 @item -mcallee-super-interworking
8742 @opindex mcallee-super-interworking
8743 Gives all externally visible functions in the file being compiled an ARM
8744 instruction set header which switches to Thumb mode before executing the
8745 rest of the function. This allows these functions to be called from
8746 non-interworking code.
8748 @item -mcaller-super-interworking
8749 @opindex mcaller-super-interworking
8750 Allows calls via function pointers (including virtual functions) to
8751 execute correctly regardless of whether the target code has been
8752 compiled for interworking or not. There is a small overhead in the cost
8753 of executing a function pointer if this option is enabled.
8755 @item -mtp=@var{name}
8757 Specify the access model for the thread local storage pointer. The valid
8758 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
8759 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
8760 (supported in the arm6k architecture), and @option{auto}, which uses the
8761 best available method for the selected processor. The default setting is
8767 @subsection AVR Options
8770 These options are defined for AVR implementations:
8773 @item -mmcu=@var{mcu}
8775 Specify ATMEL AVR instruction set or MCU type.
8777 Instruction set avr1 is for the minimal AVR core, not supported by the C
8778 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8779 attiny11, attiny12, attiny15, attiny28).
8781 Instruction set avr2 (default) is for the classic AVR core with up to
8782 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8783 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8784 at90c8534, at90s8535).
8786 Instruction set avr3 is for the classic AVR core with up to 128K program
8787 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8789 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8790 memory space (MCU types: atmega8, atmega83, atmega85).
8792 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8793 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8794 atmega64, atmega128, at43usb355, at94k).
8798 Output instruction sizes to the asm file.
8800 @item -minit-stack=@var{N}
8801 @opindex minit-stack
8802 Specify the initial stack address, which may be a symbol or numeric value,
8803 @samp{__stack} is the default.
8805 @item -mno-interrupts
8806 @opindex mno-interrupts
8807 Generated code is not compatible with hardware interrupts.
8808 Code size will be smaller.
8810 @item -mcall-prologues
8811 @opindex mcall-prologues
8812 Functions prologues/epilogues expanded as call to appropriate
8813 subroutines. Code size will be smaller.
8815 @item -mno-tablejump
8816 @opindex mno-tablejump
8817 Do not generate tablejump insns which sometimes increase code size.
8820 @opindex mtiny-stack
8821 Change only the low 8 bits of the stack pointer.
8825 Assume int to be 8 bit integer. This affects the sizes of all types: A
8826 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
8827 and long long will be 4 bytes. Please note that this option does not
8828 comply to the C standards, but it will provide you with smaller code
8832 @node Blackfin Options
8833 @subsection Blackfin Options
8834 @cindex Blackfin Options
8837 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
8839 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
8840 can be one of @samp{bf522}, @samp{bf523}, @samp{bf524},
8841 @samp{bf525}, @samp{bf526}, @samp{bf527},
8842 @samp{bf531}, @samp{bf532}, @samp{bf533}, @samp{bf534},
8843 @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
8844 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
8846 The optional @var{sirevision} specifies the silicon revision of the target
8847 Blackfin processor. Any workarounds available for the targeted silicon revision
8848 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
8849 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
8850 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
8851 hexadecimal digits representing the major and minor numbers in the silicon
8852 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
8853 is not defined. If @var{sirevision} is @samp{any}, the
8854 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
8855 If this optional @var{sirevision} is not used, GCC assumes the latest known
8856 silicon revision of the targeted Blackfin processor.
8858 Support for @samp{bf561} is incomplete. For @samp{bf561},
8859 Only the processor macro is defined.
8860 Without this option, @samp{bf532} is used as the processor by default.
8861 The corresponding predefined processor macros for @var{cpu} is to
8862 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
8863 provided by libgloss to be linked in if @option{-msim} is not given.
8867 Specifies that the program will be run on the simulator. This causes
8868 the simulator BSP provided by libgloss to be linked in. This option
8869 has effect only for @samp{bfin-elf} toolchain.
8870 Certain other options, such as @option{-mid-shared-library} and
8871 @option{-mfdpic}, imply @option{-msim}.
8873 @item -momit-leaf-frame-pointer
8874 @opindex momit-leaf-frame-pointer
8875 Don't keep the frame pointer in a register for leaf functions. This
8876 avoids the instructions to save, set up and restore frame pointers and
8877 makes an extra register available in leaf functions. The option
8878 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8879 which might make debugging harder.
8881 @item -mspecld-anomaly
8882 @opindex mspecld-anomaly
8883 When enabled, the compiler will ensure that the generated code does not
8884 contain speculative loads after jump instructions. If this option is used,
8885 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
8887 @item -mno-specld-anomaly
8888 @opindex mno-specld-anomaly
8889 Don't generate extra code to prevent speculative loads from occurring.
8891 @item -mcsync-anomaly
8892 @opindex mcsync-anomaly
8893 When enabled, the compiler will ensure that the generated code does not
8894 contain CSYNC or SSYNC instructions too soon after conditional branches.
8895 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
8897 @item -mno-csync-anomaly
8898 @opindex mno-csync-anomaly
8899 Don't generate extra code to prevent CSYNC or SSYNC instructions from
8900 occurring too soon after a conditional branch.
8904 When enabled, the compiler is free to take advantage of the knowledge that
8905 the entire program fits into the low 64k of memory.
8908 @opindex mno-low-64k
8909 Assume that the program is arbitrarily large. This is the default.
8911 @item -mstack-check-l1
8912 @opindex mstack-check-l1
8913 Do stack checking using information placed into L1 scratchpad memory by the
8916 @item -mid-shared-library
8917 @opindex mid-shared-library
8918 Generate code that supports shared libraries via the library ID method.
8919 This allows for execute in place and shared libraries in an environment
8920 without virtual memory management. This option implies @option{-fPIC}.
8921 With a @samp{bfin-elf} target, this option implies @option{-msim}.
8923 @item -mno-id-shared-library
8924 @opindex mno-id-shared-library
8925 Generate code that doesn't assume ID based shared libraries are being used.
8926 This is the default.
8928 @item -mleaf-id-shared-library
8929 @opindex mleaf-id-shared-library
8930 Generate code that supports shared libraries via the library ID method,
8931 but assumes that this library or executable won't link against any other
8932 ID shared libraries. That allows the compiler to use faster code for jumps
8935 @item -mno-leaf-id-shared-library
8936 @opindex mno-leaf-id-shared-library
8937 Do not assume that the code being compiled won't link against any ID shared
8938 libraries. Slower code will be generated for jump and call insns.
8940 @item -mshared-library-id=n
8941 @opindex mshared-library-id
8942 Specified the identification number of the ID based shared library being
8943 compiled. Specifying a value of 0 will generate more compact code, specifying
8944 other values will force the allocation of that number to the current
8945 library but is no more space or time efficient than omitting this option.
8949 Generate code that allows the data segment to be located in a different
8950 area of memory from the text segment. This allows for execute in place in
8951 an environment without virtual memory management by eliminating relocations
8952 against the text section.
8955 @opindex mno-sep-data
8956 Generate code that assumes that the data segment follows the text segment.
8957 This is the default.
8960 @itemx -mno-long-calls
8961 @opindex mlong-calls
8962 @opindex mno-long-calls
8963 Tells the compiler to perform function calls by first loading the
8964 address of the function into a register and then performing a subroutine
8965 call on this register. This switch is needed if the target function
8966 will lie outside of the 24 bit addressing range of the offset based
8967 version of subroutine call instruction.
8969 This feature is not enabled by default. Specifying
8970 @option{-mno-long-calls} will restore the default behavior. Note these
8971 switches have no effect on how the compiler generates code to handle
8972 function calls via function pointers.
8976 Link with the fast floating-point library. This library relaxes some of
8977 the IEEE floating-point standard's rules for checking inputs against
8978 Not-a-Number (NAN), in the interest of performance.
8981 @opindex minline-plt
8982 Enable inlining of PLT entries in function calls to functions that are
8983 not known to bind locally. It has no effect without @option{-mfdpic}.
8987 @subsection CRIS Options
8988 @cindex CRIS Options
8990 These options are defined specifically for the CRIS ports.
8993 @item -march=@var{architecture-type}
8994 @itemx -mcpu=@var{architecture-type}
8997 Generate code for the specified architecture. The choices for
8998 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
8999 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
9000 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9003 @item -mtune=@var{architecture-type}
9005 Tune to @var{architecture-type} everything applicable about the generated
9006 code, except for the ABI and the set of available instructions. The
9007 choices for @var{architecture-type} are the same as for
9008 @option{-march=@var{architecture-type}}.
9010 @item -mmax-stack-frame=@var{n}
9011 @opindex mmax-stack-frame
9012 Warn when the stack frame of a function exceeds @var{n} bytes.
9014 @item -melinux-stacksize=@var{n}
9015 @opindex melinux-stacksize
9016 Only available with the @samp{cris-axis-aout} target. Arranges for
9017 indications in the program to the kernel loader that the stack of the
9018 program should be set to @var{n} bytes.
9024 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9025 @option{-march=v3} and @option{-march=v8} respectively.
9027 @item -mmul-bug-workaround
9028 @itemx -mno-mul-bug-workaround
9029 @opindex mmul-bug-workaround
9030 @opindex mno-mul-bug-workaround
9031 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
9032 models where it applies. This option is active by default.
9036 Enable CRIS-specific verbose debug-related information in the assembly
9037 code. This option also has the effect to turn off the @samp{#NO_APP}
9038 formatted-code indicator to the assembler at the beginning of the
9043 Do not use condition-code results from previous instruction; always emit
9044 compare and test instructions before use of condition codes.
9046 @item -mno-side-effects
9047 @opindex mno-side-effects
9048 Do not emit instructions with side-effects in addressing modes other than
9052 @itemx -mno-stack-align
9054 @itemx -mno-data-align
9055 @itemx -mconst-align
9056 @itemx -mno-const-align
9057 @opindex mstack-align
9058 @opindex mno-stack-align
9059 @opindex mdata-align
9060 @opindex mno-data-align
9061 @opindex mconst-align
9062 @opindex mno-const-align
9063 These options (no-options) arranges (eliminate arrangements) for the
9064 stack-frame, individual data and constants to be aligned for the maximum
9065 single data access size for the chosen CPU model. The default is to
9066 arrange for 32-bit alignment. ABI details such as structure layout are
9067 not affected by these options.
9075 Similar to the stack- data- and const-align options above, these options
9076 arrange for stack-frame, writable data and constants to all be 32-bit,
9077 16-bit or 8-bit aligned. The default is 32-bit alignment.
9079 @item -mno-prologue-epilogue
9080 @itemx -mprologue-epilogue
9081 @opindex mno-prologue-epilogue
9082 @opindex mprologue-epilogue
9083 With @option{-mno-prologue-epilogue}, the normal function prologue and
9084 epilogue that sets up the stack-frame are omitted and no return
9085 instructions or return sequences are generated in the code. Use this
9086 option only together with visual inspection of the compiled code: no
9087 warnings or errors are generated when call-saved registers must be saved,
9088 or storage for local variable needs to be allocated.
9094 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9095 instruction sequences that load addresses for functions from the PLT part
9096 of the GOT rather than (traditional on other architectures) calls to the
9097 PLT@. The default is @option{-mgotplt}.
9101 Legacy no-op option only recognized with the cris-axis-aout target.
9105 Legacy no-op option only recognized with the cris-axis-elf and
9106 cris-axis-linux-gnu targets.
9110 Only recognized with the cris-axis-aout target, where it selects a
9111 GNU/linux-like multilib, include files and instruction set for
9116 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9120 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9121 to link with input-output functions from a simulator library. Code,
9122 initialized data and zero-initialized data are allocated consecutively.
9126 Like @option{-sim}, but pass linker options to locate initialized data at
9127 0x40000000 and zero-initialized data at 0x80000000.
9131 @subsection CRX Options
9134 These options are defined specifically for the CRX ports.
9140 Enable the use of multiply-accumulate instructions. Disabled by default.
9144 Push instructions will be used to pass outgoing arguments when functions
9145 are called. Enabled by default.
9148 @node Darwin Options
9149 @subsection Darwin Options
9150 @cindex Darwin options
9152 These options are defined for all architectures running the Darwin operating
9155 FSF GCC on Darwin does not create ``fat'' object files; it will create
9156 an object file for the single architecture that it was built to
9157 target. Apple's GCC on Darwin does create ``fat'' files if multiple
9158 @option{-arch} options are used; it does so by running the compiler or
9159 linker multiple times and joining the results together with
9162 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
9163 @samp{i686}) is determined by the flags that specify the ISA
9164 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
9165 @option{-force_cpusubtype_ALL} option can be used to override this.
9167 The Darwin tools vary in their behavior when presented with an ISA
9168 mismatch. The assembler, @file{as}, will only permit instructions to
9169 be used that are valid for the subtype of the file it is generating,
9170 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
9171 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
9172 and print an error if asked to create a shared library with a less
9173 restrictive subtype than its input files (for instance, trying to put
9174 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
9175 for executables, @file{ld}, will quietly give the executable the most
9176 restrictive subtype of any of its input files.
9181 Add the framework directory @var{dir} to the head of the list of
9182 directories to be searched for header files. These directories are
9183 interleaved with those specified by @option{-I} options and are
9184 scanned in a left-to-right order.
9186 A framework directory is a directory with frameworks in it. A
9187 framework is a directory with a @samp{"Headers"} and/or
9188 @samp{"PrivateHeaders"} directory contained directly in it that ends
9189 in @samp{".framework"}. The name of a framework is the name of this
9190 directory excluding the @samp{".framework"}. Headers associated with
9191 the framework are found in one of those two directories, with
9192 @samp{"Headers"} being searched first. A subframework is a framework
9193 directory that is in a framework's @samp{"Frameworks"} directory.
9194 Includes of subframework headers can only appear in a header of a
9195 framework that contains the subframework, or in a sibling subframework
9196 header. Two subframeworks are siblings if they occur in the same
9197 framework. A subframework should not have the same name as a
9198 framework, a warning will be issued if this is violated. Currently a
9199 subframework cannot have subframeworks, in the future, the mechanism
9200 may be extended to support this. The standard frameworks can be found
9201 in @samp{"/System/Library/Frameworks"} and
9202 @samp{"/Library/Frameworks"}. An example include looks like
9203 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
9204 the name of the framework and header.h is found in the
9205 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
9207 @item -iframework@var{dir}
9209 Like @option{-F} except the directory is a treated as a system
9210 directory. The main difference between this @option{-iframework} and
9211 @option{-F} is that with @option{-iframework} the compiler does not
9212 warn about constructs contained within header files found via
9213 @var{dir}. This option is valid only for the C family of languages.
9217 Emit debugging information for symbols that are used. For STABS
9218 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
9219 This is by default ON@.
9223 Emit debugging information for all symbols and types.
9225 @item -mmacosx-version-min=@var{version}
9226 The earliest version of MacOS X that this executable will run on
9227 is @var{version}. Typical values of @var{version} include @code{10.1},
9228 @code{10.2}, and @code{10.3.9}.
9230 If the compiler was built to use the system's headers by default,
9231 then the default for this option is the system version on which the
9232 compiler is running, otherwise the default is to make choices which
9233 are compatible with as many systems and code bases as possible.
9237 Enable kernel development mode. The @option{-mkernel} option sets
9238 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
9239 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
9240 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
9241 applicable. This mode also sets @option{-mno-altivec},
9242 @option{-msoft-float}, @option{-fno-builtin} and
9243 @option{-mlong-branch} for PowerPC targets.
9245 @item -mone-byte-bool
9246 @opindex mone-byte-bool
9247 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
9248 By default @samp{sizeof(bool)} is @samp{4} when compiling for
9249 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
9250 option has no effect on x86.
9252 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
9253 to generate code that is not binary compatible with code generated
9254 without that switch. Using this switch may require recompiling all
9255 other modules in a program, including system libraries. Use this
9256 switch to conform to a non-default data model.
9258 @item -mfix-and-continue
9259 @itemx -ffix-and-continue
9260 @itemx -findirect-data
9261 @opindex mfix-and-continue
9262 @opindex ffix-and-continue
9263 @opindex findirect-data
9264 Generate code suitable for fast turn around development. Needed to
9265 enable gdb to dynamically load @code{.o} files into already running
9266 programs. @option{-findirect-data} and @option{-ffix-and-continue}
9267 are provided for backwards compatibility.
9271 Loads all members of static archive libraries.
9272 See man ld(1) for more information.
9274 @item -arch_errors_fatal
9275 @opindex arch_errors_fatal
9276 Cause the errors having to do with files that have the wrong architecture
9280 @opindex bind_at_load
9281 Causes the output file to be marked such that the dynamic linker will
9282 bind all undefined references when the file is loaded or launched.
9286 Produce a Mach-o bundle format file.
9287 See man ld(1) for more information.
9289 @item -bundle_loader @var{executable}
9290 @opindex bundle_loader
9291 This option specifies the @var{executable} that will be loading the build
9292 output file being linked. See man ld(1) for more information.
9296 When passed this option, GCC will produce a dynamic library instead of
9297 an executable when linking, using the Darwin @file{libtool} command.
9299 @item -force_cpusubtype_ALL
9300 @opindex force_cpusubtype_ALL
9301 This causes GCC's output file to have the @var{ALL} subtype, instead of
9302 one controlled by the @option{-mcpu} or @option{-march} option.
9304 @item -allowable_client @var{client_name}
9306 @itemx -compatibility_version
9307 @itemx -current_version
9309 @itemx -dependency-file
9311 @itemx -dylinker_install_name
9313 @itemx -exported_symbols_list
9315 @itemx -flat_namespace
9316 @itemx -force_flat_namespace
9317 @itemx -headerpad_max_install_names
9320 @itemx -install_name
9321 @itemx -keep_private_externs
9322 @itemx -multi_module
9323 @itemx -multiply_defined
9324 @itemx -multiply_defined_unused
9326 @itemx -no_dead_strip_inits_and_terms
9327 @itemx -nofixprebinding
9330 @itemx -noseglinkedit
9331 @itemx -pagezero_size
9333 @itemx -prebind_all_twolevel_modules
9334 @itemx -private_bundle
9335 @itemx -read_only_relocs
9337 @itemx -sectobjectsymbols
9341 @itemx -sectobjectsymbols
9344 @itemx -segs_read_only_addr
9345 @itemx -segs_read_write_addr
9346 @itemx -seg_addr_table
9347 @itemx -seg_addr_table_filename
9350 @itemx -segs_read_only_addr
9351 @itemx -segs_read_write_addr
9352 @itemx -single_module
9355 @itemx -sub_umbrella
9356 @itemx -twolevel_namespace
9359 @itemx -unexported_symbols_list
9360 @itemx -weak_reference_mismatches
9362 @opindex allowable_client
9363 @opindex client_name
9364 @opindex compatibility_version
9365 @opindex current_version
9367 @opindex dependency-file
9369 @opindex dylinker_install_name
9371 @opindex exported_symbols_list
9373 @opindex flat_namespace
9374 @opindex force_flat_namespace
9375 @opindex headerpad_max_install_names
9378 @opindex install_name
9379 @opindex keep_private_externs
9380 @opindex multi_module
9381 @opindex multiply_defined
9382 @opindex multiply_defined_unused
9384 @opindex no_dead_strip_inits_and_terms
9385 @opindex nofixprebinding
9386 @opindex nomultidefs
9388 @opindex noseglinkedit
9389 @opindex pagezero_size
9391 @opindex prebind_all_twolevel_modules
9392 @opindex private_bundle
9393 @opindex read_only_relocs
9395 @opindex sectobjectsymbols
9399 @opindex sectobjectsymbols
9402 @opindex segs_read_only_addr
9403 @opindex segs_read_write_addr
9404 @opindex seg_addr_table
9405 @opindex seg_addr_table_filename
9406 @opindex seglinkedit
9408 @opindex segs_read_only_addr
9409 @opindex segs_read_write_addr
9410 @opindex single_module
9412 @opindex sub_library
9413 @opindex sub_umbrella
9414 @opindex twolevel_namespace
9417 @opindex unexported_symbols_list
9418 @opindex weak_reference_mismatches
9419 @opindex whatsloaded
9420 These options are passed to the Darwin linker. The Darwin linker man page
9421 describes them in detail.
9424 @node DEC Alpha Options
9425 @subsection DEC Alpha Options
9427 These @samp{-m} options are defined for the DEC Alpha implementations:
9430 @item -mno-soft-float
9432 @opindex mno-soft-float
9433 @opindex msoft-float
9434 Use (do not use) the hardware floating-point instructions for
9435 floating-point operations. When @option{-msoft-float} is specified,
9436 functions in @file{libgcc.a} will be used to perform floating-point
9437 operations. Unless they are replaced by routines that emulate the
9438 floating-point operations, or compiled in such a way as to call such
9439 emulations routines, these routines will issue floating-point
9440 operations. If you are compiling for an Alpha without floating-point
9441 operations, you must ensure that the library is built so as not to call
9444 Note that Alpha implementations without floating-point operations are
9445 required to have floating-point registers.
9450 @opindex mno-fp-regs
9451 Generate code that uses (does not use) the floating-point register set.
9452 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
9453 register set is not used, floating point operands are passed in integer
9454 registers as if they were integers and floating-point results are passed
9455 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
9456 so any function with a floating-point argument or return value called by code
9457 compiled with @option{-mno-fp-regs} must also be compiled with that
9460 A typical use of this option is building a kernel that does not use,
9461 and hence need not save and restore, any floating-point registers.
9465 The Alpha architecture implements floating-point hardware optimized for
9466 maximum performance. It is mostly compliant with the IEEE floating
9467 point standard. However, for full compliance, software assistance is
9468 required. This option generates code fully IEEE compliant code
9469 @emph{except} that the @var{inexact-flag} is not maintained (see below).
9470 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
9471 defined during compilation. The resulting code is less efficient but is
9472 able to correctly support denormalized numbers and exceptional IEEE
9473 values such as not-a-number and plus/minus infinity. Other Alpha
9474 compilers call this option @option{-ieee_with_no_inexact}.
9476 @item -mieee-with-inexact
9477 @opindex mieee-with-inexact
9478 This is like @option{-mieee} except the generated code also maintains
9479 the IEEE @var{inexact-flag}. Turning on this option causes the
9480 generated code to implement fully-compliant IEEE math. In addition to
9481 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
9482 macro. On some Alpha implementations the resulting code may execute
9483 significantly slower than the code generated by default. Since there is
9484 very little code that depends on the @var{inexact-flag}, you should
9485 normally not specify this option. Other Alpha compilers call this
9486 option @option{-ieee_with_inexact}.
9488 @item -mfp-trap-mode=@var{trap-mode}
9489 @opindex mfp-trap-mode
9490 This option controls what floating-point related traps are enabled.
9491 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9492 The trap mode can be set to one of four values:
9496 This is the default (normal) setting. The only traps that are enabled
9497 are the ones that cannot be disabled in software (e.g., division by zero
9501 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9505 Like @samp{u}, but the instructions are marked to be safe for software
9506 completion (see Alpha architecture manual for details).
9509 Like @samp{su}, but inexact traps are enabled as well.
9512 @item -mfp-rounding-mode=@var{rounding-mode}
9513 @opindex mfp-rounding-mode
9514 Selects the IEEE rounding mode. Other Alpha compilers call this option
9515 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9520 Normal IEEE rounding mode. Floating point numbers are rounded towards
9521 the nearest machine number or towards the even machine number in case
9525 Round towards minus infinity.
9528 Chopped rounding mode. Floating point numbers are rounded towards zero.
9531 Dynamic rounding mode. A field in the floating point control register
9532 (@var{fpcr}, see Alpha architecture reference manual) controls the
9533 rounding mode in effect. The C library initializes this register for
9534 rounding towards plus infinity. Thus, unless your program modifies the
9535 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9538 @item -mtrap-precision=@var{trap-precision}
9539 @opindex mtrap-precision
9540 In the Alpha architecture, floating point traps are imprecise. This
9541 means without software assistance it is impossible to recover from a
9542 floating trap and program execution normally needs to be terminated.
9543 GCC can generate code that can assist operating system trap handlers
9544 in determining the exact location that caused a floating point trap.
9545 Depending on the requirements of an application, different levels of
9546 precisions can be selected:
9550 Program precision. This option is the default and means a trap handler
9551 can only identify which program caused a floating point exception.
9554 Function precision. The trap handler can determine the function that
9555 caused a floating point exception.
9558 Instruction precision. The trap handler can determine the exact
9559 instruction that caused a floating point exception.
9562 Other Alpha compilers provide the equivalent options called
9563 @option{-scope_safe} and @option{-resumption_safe}.
9565 @item -mieee-conformant
9566 @opindex mieee-conformant
9567 This option marks the generated code as IEEE conformant. You must not
9568 use this option unless you also specify @option{-mtrap-precision=i} and either
9569 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9570 is to emit the line @samp{.eflag 48} in the function prologue of the
9571 generated assembly file. Under DEC Unix, this has the effect that
9572 IEEE-conformant math library routines will be linked in.
9574 @item -mbuild-constants
9575 @opindex mbuild-constants
9576 Normally GCC examines a 32- or 64-bit integer constant to
9577 see if it can construct it from smaller constants in two or three
9578 instructions. If it cannot, it will output the constant as a literal and
9579 generate code to load it from the data segment at runtime.
9581 Use this option to require GCC to construct @emph{all} integer constants
9582 using code, even if it takes more instructions (the maximum is six).
9584 You would typically use this option to build a shared library dynamic
9585 loader. Itself a shared library, it must relocate itself in memory
9586 before it can find the variables and constants in its own data segment.
9592 Select whether to generate code to be assembled by the vendor-supplied
9593 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9611 Indicate whether GCC should generate code to use the optional BWX,
9612 CIX, FIX and MAX instruction sets. The default is to use the instruction
9613 sets supported by the CPU type specified via @option{-mcpu=} option or that
9614 of the CPU on which GCC was built if none was specified.
9619 @opindex mfloat-ieee
9620 Generate code that uses (does not use) VAX F and G floating point
9621 arithmetic instead of IEEE single and double precision.
9623 @item -mexplicit-relocs
9624 @itemx -mno-explicit-relocs
9625 @opindex mexplicit-relocs
9626 @opindex mno-explicit-relocs
9627 Older Alpha assemblers provided no way to generate symbol relocations
9628 except via assembler macros. Use of these macros does not allow
9629 optimal instruction scheduling. GNU binutils as of version 2.12
9630 supports a new syntax that allows the compiler to explicitly mark
9631 which relocations should apply to which instructions. This option
9632 is mostly useful for debugging, as GCC detects the capabilities of
9633 the assembler when it is built and sets the default accordingly.
9637 @opindex msmall-data
9638 @opindex mlarge-data
9639 When @option{-mexplicit-relocs} is in effect, static data is
9640 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9641 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9642 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9643 16-bit relocations off of the @code{$gp} register. This limits the
9644 size of the small data area to 64KB, but allows the variables to be
9645 directly accessed via a single instruction.
9647 The default is @option{-mlarge-data}. With this option the data area
9648 is limited to just below 2GB@. Programs that require more than 2GB of
9649 data must use @code{malloc} or @code{mmap} to allocate the data in the
9650 heap instead of in the program's data segment.
9652 When generating code for shared libraries, @option{-fpic} implies
9653 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9657 @opindex msmall-text
9658 @opindex mlarge-text
9659 When @option{-msmall-text} is used, the compiler assumes that the
9660 code of the entire program (or shared library) fits in 4MB, and is
9661 thus reachable with a branch instruction. When @option{-msmall-data}
9662 is used, the compiler can assume that all local symbols share the
9663 same @code{$gp} value, and thus reduce the number of instructions
9664 required for a function call from 4 to 1.
9666 The default is @option{-mlarge-text}.
9668 @item -mcpu=@var{cpu_type}
9670 Set the instruction set and instruction scheduling parameters for
9671 machine type @var{cpu_type}. You can specify either the @samp{EV}
9672 style name or the corresponding chip number. GCC supports scheduling
9673 parameters for the EV4, EV5 and EV6 family of processors and will
9674 choose the default values for the instruction set from the processor
9675 you specify. If you do not specify a processor type, GCC will default
9676 to the processor on which the compiler was built.
9678 Supported values for @var{cpu_type} are
9684 Schedules as an EV4 and has no instruction set extensions.
9688 Schedules as an EV5 and has no instruction set extensions.
9692 Schedules as an EV5 and supports the BWX extension.
9697 Schedules as an EV5 and supports the BWX and MAX extensions.
9701 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9705 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9708 @item -mtune=@var{cpu_type}
9710 Set only the instruction scheduling parameters for machine type
9711 @var{cpu_type}. The instruction set is not changed.
9713 @item -mmemory-latency=@var{time}
9714 @opindex mmemory-latency
9715 Sets the latency the scheduler should assume for typical memory
9716 references as seen by the application. This number is highly
9717 dependent on the memory access patterns used by the application
9718 and the size of the external cache on the machine.
9720 Valid options for @var{time} are
9724 A decimal number representing clock cycles.
9730 The compiler contains estimates of the number of clock cycles for
9731 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9732 (also called Dcache, Scache, and Bcache), as well as to main memory.
9733 Note that L3 is only valid for EV5.
9738 @node DEC Alpha/VMS Options
9739 @subsection DEC Alpha/VMS Options
9741 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9744 @item -mvms-return-codes
9745 @opindex mvms-return-codes
9746 Return VMS condition codes from main. The default is to return POSIX
9747 style condition (e.g.@: error) codes.
9751 @subsection FRV Options
9758 Only use the first 32 general purpose registers.
9763 Use all 64 general purpose registers.
9768 Use only the first 32 floating point registers.
9773 Use all 64 floating point registers
9776 @opindex mhard-float
9778 Use hardware instructions for floating point operations.
9781 @opindex msoft-float
9783 Use library routines for floating point operations.
9788 Dynamically allocate condition code registers.
9793 Do not try to dynamically allocate condition code registers, only
9794 use @code{icc0} and @code{fcc0}.
9799 Change ABI to use double word insns.
9804 Do not use double word instructions.
9809 Use floating point double instructions.
9814 Do not use floating point double instructions.
9819 Use media instructions.
9824 Do not use media instructions.
9829 Use multiply and add/subtract instructions.
9834 Do not use multiply and add/subtract instructions.
9839 Select the FDPIC ABI, that uses function descriptors to represent
9840 pointers to functions. Without any PIC/PIE-related options, it
9841 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
9842 assumes GOT entries and small data are within a 12-bit range from the
9843 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
9844 are computed with 32 bits.
9845 With a @samp{bfin-elf} target, this option implies @option{-msim}.
9848 @opindex minline-plt
9850 Enable inlining of PLT entries in function calls to functions that are
9851 not known to bind locally. It has no effect without @option{-mfdpic}.
9852 It's enabled by default if optimizing for speed and compiling for
9853 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
9854 optimization option such as @option{-O3} or above is present in the
9860 Assume a large TLS segment when generating thread-local code.
9865 Do not assume a large TLS segment when generating thread-local code.
9870 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
9871 that is known to be in read-only sections. It's enabled by default,
9872 except for @option{-fpic} or @option{-fpie}: even though it may help
9873 make the global offset table smaller, it trades 1 instruction for 4.
9874 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
9875 one of which may be shared by multiple symbols, and it avoids the need
9876 for a GOT entry for the referenced symbol, so it's more likely to be a
9877 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
9879 @item -multilib-library-pic
9880 @opindex multilib-library-pic
9882 Link with the (library, not FD) pic libraries. It's implied by
9883 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
9884 @option{-fpic} without @option{-mfdpic}. You should never have to use
9890 Follow the EABI requirement of always creating a frame pointer whenever
9891 a stack frame is allocated. This option is enabled by default and can
9892 be disabled with @option{-mno-linked-fp}.
9895 @opindex mlong-calls
9897 Use indirect addressing to call functions outside the current
9898 compilation unit. This allows the functions to be placed anywhere
9899 within the 32-bit address space.
9901 @item -malign-labels
9902 @opindex malign-labels
9904 Try to align labels to an 8-byte boundary by inserting nops into the
9905 previous packet. This option only has an effect when VLIW packing
9906 is enabled. It doesn't create new packets; it merely adds nops to
9910 @opindex mlibrary-pic
9912 Generate position-independent EABI code.
9917 Use only the first four media accumulator registers.
9922 Use all eight media accumulator registers.
9927 Pack VLIW instructions.
9932 Do not pack VLIW instructions.
9937 Do not mark ABI switches in e_flags.
9942 Enable the use of conditional-move instructions (default).
9944 This switch is mainly for debugging the compiler and will likely be removed
9945 in a future version.
9947 @item -mno-cond-move
9948 @opindex mno-cond-move
9950 Disable the use of conditional-move instructions.
9952 This switch is mainly for debugging the compiler and will likely be removed
9953 in a future version.
9958 Enable the use of conditional set instructions (default).
9960 This switch is mainly for debugging the compiler and will likely be removed
9961 in a future version.
9966 Disable the use of conditional set instructions.
9968 This switch is mainly for debugging the compiler and will likely be removed
9969 in a future version.
9974 Enable the use of conditional execution (default).
9976 This switch is mainly for debugging the compiler and will likely be removed
9977 in a future version.
9979 @item -mno-cond-exec
9980 @opindex mno-cond-exec
9982 Disable the use of conditional execution.
9984 This switch is mainly for debugging the compiler and will likely be removed
9985 in a future version.
9988 @opindex mvliw-branch
9990 Run a pass to pack branches into VLIW instructions (default).
9992 This switch is mainly for debugging the compiler and will likely be removed
9993 in a future version.
9995 @item -mno-vliw-branch
9996 @opindex mno-vliw-branch
9998 Do not run a pass to pack branches into VLIW instructions.
10000 This switch is mainly for debugging the compiler and will likely be removed
10001 in a future version.
10003 @item -mmulti-cond-exec
10004 @opindex mmulti-cond-exec
10006 Enable optimization of @code{&&} and @code{||} in conditional execution
10009 This switch is mainly for debugging the compiler and will likely be removed
10010 in a future version.
10012 @item -mno-multi-cond-exec
10013 @opindex mno-multi-cond-exec
10015 Disable optimization of @code{&&} and @code{||} in conditional execution.
10017 This switch is mainly for debugging the compiler and will likely be removed
10018 in a future version.
10020 @item -mnested-cond-exec
10021 @opindex mnested-cond-exec
10023 Enable nested conditional execution optimizations (default).
10025 This switch is mainly for debugging the compiler and will likely be removed
10026 in a future version.
10028 @item -mno-nested-cond-exec
10029 @opindex mno-nested-cond-exec
10031 Disable nested conditional execution optimizations.
10033 This switch is mainly for debugging the compiler and will likely be removed
10034 in a future version.
10036 @item -moptimize-membar
10037 @opindex moptimize-membar
10039 This switch removes redundant @code{membar} instructions from the
10040 compiler generated code. It is enabled by default.
10042 @item -mno-optimize-membar
10043 @opindex mno-optimize-membar
10045 This switch disables the automatic removal of redundant @code{membar}
10046 instructions from the generated code.
10048 @item -mtomcat-stats
10049 @opindex mtomcat-stats
10051 Cause gas to print out tomcat statistics.
10053 @item -mcpu=@var{cpu}
10056 Select the processor type for which to generate code. Possible values are
10057 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
10058 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
10062 @node GNU/Linux Options
10063 @subsection GNU/Linux Options
10065 These @samp{-m} options are defined for GNU/Linux targets:
10070 Use the GNU C library instead of uClibc. This is the default except
10071 on @samp{*-*-linux-*uclibc*} targets.
10075 Use uClibc instead of the GNU C library. This is the default on
10076 @samp{*-*-linux-*uclibc*} targets.
10079 @node H8/300 Options
10080 @subsection H8/300 Options
10082 These @samp{-m} options are defined for the H8/300 implementations:
10087 Shorten some address references at link time, when possible; uses the
10088 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
10089 ld, Using ld}, for a fuller description.
10093 Generate code for the H8/300H@.
10097 Generate code for the H8S@.
10101 Generate code for the H8S and H8/300H in the normal mode. This switch
10102 must be used either with @option{-mh} or @option{-ms}.
10106 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
10110 Make @code{int} data 32 bits by default.
10113 @opindex malign-300
10114 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
10115 The default for the H8/300H and H8S is to align longs and floats on 4
10117 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
10118 This option has no effect on the H8/300.
10122 @subsection HPPA Options
10123 @cindex HPPA Options
10125 These @samp{-m} options are defined for the HPPA family of computers:
10128 @item -march=@var{architecture-type}
10130 Generate code for the specified architecture. The choices for
10131 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
10132 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
10133 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
10134 architecture option for your machine. Code compiled for lower numbered
10135 architectures will run on higher numbered architectures, but not the
10138 @item -mpa-risc-1-0
10139 @itemx -mpa-risc-1-1
10140 @itemx -mpa-risc-2-0
10141 @opindex mpa-risc-1-0
10142 @opindex mpa-risc-1-1
10143 @opindex mpa-risc-2-0
10144 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
10147 @opindex mbig-switch
10148 Generate code suitable for big switch tables. Use this option only if
10149 the assembler/linker complain about out of range branches within a switch
10152 @item -mjump-in-delay
10153 @opindex mjump-in-delay
10154 Fill delay slots of function calls with unconditional jump instructions
10155 by modifying the return pointer for the function call to be the target
10156 of the conditional jump.
10158 @item -mdisable-fpregs
10159 @opindex mdisable-fpregs
10160 Prevent floating point registers from being used in any manner. This is
10161 necessary for compiling kernels which perform lazy context switching of
10162 floating point registers. If you use this option and attempt to perform
10163 floating point operations, the compiler will abort.
10165 @item -mdisable-indexing
10166 @opindex mdisable-indexing
10167 Prevent the compiler from using indexing address modes. This avoids some
10168 rather obscure problems when compiling MIG generated code under MACH@.
10170 @item -mno-space-regs
10171 @opindex mno-space-regs
10172 Generate code that assumes the target has no space registers. This allows
10173 GCC to generate faster indirect calls and use unscaled index address modes.
10175 Such code is suitable for level 0 PA systems and kernels.
10177 @item -mfast-indirect-calls
10178 @opindex mfast-indirect-calls
10179 Generate code that assumes calls never cross space boundaries. This
10180 allows GCC to emit code which performs faster indirect calls.
10182 This option will not work in the presence of shared libraries or nested
10185 @item -mfixed-range=@var{register-range}
10186 @opindex mfixed-range
10187 Generate code treating the given register range as fixed registers.
10188 A fixed register is one that the register allocator can not use. This is
10189 useful when compiling kernel code. A register range is specified as
10190 two registers separated by a dash. Multiple register ranges can be
10191 specified separated by a comma.
10193 @item -mlong-load-store
10194 @opindex mlong-load-store
10195 Generate 3-instruction load and store sequences as sometimes required by
10196 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
10199 @item -mportable-runtime
10200 @opindex mportable-runtime
10201 Use the portable calling conventions proposed by HP for ELF systems.
10205 Enable the use of assembler directives only GAS understands.
10207 @item -mschedule=@var{cpu-type}
10209 Schedule code according to the constraints for the machine type
10210 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
10211 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
10212 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
10213 proper scheduling option for your machine. The default scheduling is
10217 @opindex mlinker-opt
10218 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
10219 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
10220 linkers in which they give bogus error messages when linking some programs.
10223 @opindex msoft-float
10224 Generate output containing library calls for floating point.
10225 @strong{Warning:} the requisite libraries are not available for all HPPA
10226 targets. Normally the facilities of the machine's usual C compiler are
10227 used, but this cannot be done directly in cross-compilation. You must make
10228 your own arrangements to provide suitable library functions for
10229 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
10230 does provide software floating point support.
10232 @option{-msoft-float} changes the calling convention in the output file;
10233 therefore, it is only useful if you compile @emph{all} of a program with
10234 this option. In particular, you need to compile @file{libgcc.a}, the
10235 library that comes with GCC, with @option{-msoft-float} in order for
10240 Generate the predefine, @code{_SIO}, for server IO@. The default is
10241 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
10242 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
10243 options are available under HP-UX and HI-UX@.
10247 Use GNU ld specific options. This passes @option{-shared} to ld when
10248 building a shared library. It is the default when GCC is configured,
10249 explicitly or implicitly, with the GNU linker. This option does not
10250 have any affect on which ld is called, it only changes what parameters
10251 are passed to that ld. The ld that is called is determined by the
10252 @option{--with-ld} configure option, GCC's program search path, and
10253 finally by the user's @env{PATH}. The linker used by GCC can be printed
10254 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
10255 on the 64 bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
10259 Use HP ld specific options. This passes @option{-b} to ld when building
10260 a shared library and passes @option{+Accept TypeMismatch} to ld on all
10261 links. It is the default when GCC is configured, explicitly or
10262 implicitly, with the HP linker. This option does not have any affect on
10263 which ld is called, it only changes what parameters are passed to that
10264 ld. The ld that is called is determined by the @option{--with-ld}
10265 configure option, GCC's program search path, and finally by the user's
10266 @env{PATH}. The linker used by GCC can be printed using @samp{which
10267 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
10268 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
10271 @opindex mno-long-calls
10272 Generate code that uses long call sequences. This ensures that a call
10273 is always able to reach linker generated stubs. The default is to generate
10274 long calls only when the distance from the call site to the beginning
10275 of the function or translation unit, as the case may be, exceeds a
10276 predefined limit set by the branch type being used. The limits for
10277 normal calls are 7,600,000 and 240,000 bytes, respectively for the
10278 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
10281 Distances are measured from the beginning of functions when using the
10282 @option{-ffunction-sections} option, or when using the @option{-mgas}
10283 and @option{-mno-portable-runtime} options together under HP-UX with
10286 It is normally not desirable to use this option as it will degrade
10287 performance. However, it may be useful in large applications,
10288 particularly when partial linking is used to build the application.
10290 The types of long calls used depends on the capabilities of the
10291 assembler and linker, and the type of code being generated. The
10292 impact on systems that support long absolute calls, and long pic
10293 symbol-difference or pc-relative calls should be relatively small.
10294 However, an indirect call is used on 32-bit ELF systems in pic code
10295 and it is quite long.
10297 @item -munix=@var{unix-std}
10299 Generate compiler predefines and select a startfile for the specified
10300 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
10301 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
10302 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
10303 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
10304 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
10307 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
10308 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
10309 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
10310 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
10311 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
10312 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
10314 It is @emph{important} to note that this option changes the interfaces
10315 for various library routines. It also affects the operational behavior
10316 of the C library. Thus, @emph{extreme} care is needed in using this
10319 Library code that is intended to operate with more than one UNIX
10320 standard must test, set and restore the variable @var{__xpg4_extended_mask}
10321 as appropriate. Most GNU software doesn't provide this capability.
10325 Suppress the generation of link options to search libdld.sl when the
10326 @option{-static} option is specified on HP-UX 10 and later.
10330 The HP-UX implementation of setlocale in libc has a dependency on
10331 libdld.sl. There isn't an archive version of libdld.sl. Thus,
10332 when the @option{-static} option is specified, special link options
10333 are needed to resolve this dependency.
10335 On HP-UX 10 and later, the GCC driver adds the necessary options to
10336 link with libdld.sl when the @option{-static} option is specified.
10337 This causes the resulting binary to be dynamic. On the 64-bit port,
10338 the linkers generate dynamic binaries by default in any case. The
10339 @option{-nolibdld} option can be used to prevent the GCC driver from
10340 adding these link options.
10344 Add support for multithreading with the @dfn{dce thread} library
10345 under HP-UX@. This option sets flags for both the preprocessor and
10349 @node i386 and x86-64 Options
10350 @subsection Intel 386 and AMD x86-64 Options
10351 @cindex i386 Options
10352 @cindex x86-64 Options
10353 @cindex Intel 386 Options
10354 @cindex AMD x86-64 Options
10356 These @samp{-m} options are defined for the i386 and x86-64 family of
10360 @item -mtune=@var{cpu-type}
10362 Tune to @var{cpu-type} everything applicable about the generated code, except
10363 for the ABI and the set of available instructions. The choices for
10364 @var{cpu-type} are:
10367 Produce code optimized for the most common IA32/AMD64/EM64T processors.
10368 If you know the CPU on which your code will run, then you should use
10369 the corresponding @option{-mtune} option instead of
10370 @option{-mtune=generic}. But, if you do not know exactly what CPU users
10371 of your application will have, then you should use this option.
10373 As new processors are deployed in the marketplace, the behavior of this
10374 option will change. Therefore, if you upgrade to a newer version of
10375 GCC, the code generated option will change to reflect the processors
10376 that were most common when that version of GCC was released.
10378 There is no @option{-march=generic} option because @option{-march}
10379 indicates the instruction set the compiler can use, and there is no
10380 generic instruction set applicable to all processors. In contrast,
10381 @option{-mtune} indicates the processor (or, in this case, collection of
10382 processors) for which the code is optimized.
10384 This selects the CPU to tune for at compilation time by determining
10385 the processor type of the compiling machine. Using @option{-mtune=native}
10386 will produce code optimized for the local machine under the constraints
10387 of the selected instruction set. Using @option{-march=native} will
10388 enable all instruction subsets supported by the local machine (hence
10389 the result might not run on different machines).
10391 Original Intel's i386 CPU@.
10393 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
10394 @item i586, pentium
10395 Intel Pentium CPU with no MMX support.
10397 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
10399 Intel PentiumPro CPU@.
10401 Same as @code{generic}, but when used as @code{march} option, PentiumPro
10402 instruction set will be used, so the code will run on all i686 family chips.
10404 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
10405 @item pentium3, pentium3m
10406 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
10409 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
10410 support. Used by Centrino notebooks.
10411 @item pentium4, pentium4m
10412 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
10414 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
10417 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
10418 SSE2 and SSE3 instruction set support.
10420 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
10421 instruction set support.
10423 AMD K6 CPU with MMX instruction set support.
10425 Improved versions of AMD K6 CPU with MMX and 3dNOW!@: instruction set support.
10426 @item athlon, athlon-tbird
10427 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and SSE prefetch instructions
10429 @item athlon-4, athlon-xp, athlon-mp
10430 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and full SSE
10431 instruction set support.
10432 @item k8, opteron, athlon64, athlon-fx
10433 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
10434 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW!@: and 64-bit instruction set extensions.)
10435 @item k8-sse3, opteron-sse3, athlon64-sse3
10436 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
10437 @item amdfam10, barcelona
10438 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
10439 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3dNOW!, enhanced 3dNOW!, ABM and 64-bit
10440 instruction set extensions.)
10442 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
10445 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!@:
10446 instruction set support.
10448 Via C3 CPU with MMX and 3dNOW!@: instruction set support. (No scheduling is
10449 implemented for this chip.)
10451 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
10452 implemented for this chip.)
10454 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
10457 While picking a specific @var{cpu-type} will schedule things appropriately
10458 for that particular chip, the compiler will not generate any code that
10459 does not run on the i386 without the @option{-march=@var{cpu-type}} option
10462 @item -march=@var{cpu-type}
10464 Generate instructions for the machine type @var{cpu-type}. The choices
10465 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
10466 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
10468 @item -mcpu=@var{cpu-type}
10470 A deprecated synonym for @option{-mtune}.
10472 @item -mfpmath=@var{unit}
10474 Generate floating point arithmetics for selected unit @var{unit}. The choices
10475 for @var{unit} are:
10479 Use the standard 387 floating point coprocessor present majority of chips and
10480 emulated otherwise. Code compiled with this option will run almost everywhere.
10481 The temporary results are computed in 80bit precision instead of precision
10482 specified by the type resulting in slightly different results compared to most
10483 of other chips. See @option{-ffloat-store} for more detailed description.
10485 This is the default choice for i386 compiler.
10488 Use scalar floating point instructions present in the SSE instruction set.
10489 This instruction set is supported by Pentium3 and newer chips, in the AMD line
10490 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
10491 instruction set supports only single precision arithmetics, thus the double and
10492 extended precision arithmetics is still done using 387. Later version, present
10493 only in Pentium4 and the future AMD x86-64 chips supports double precision
10496 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
10497 or @option{-msse2} switches to enable SSE extensions and make this option
10498 effective. For the x86-64 compiler, these extensions are enabled by default.
10500 The resulting code should be considerably faster in the majority of cases and avoid
10501 the numerical instability problems of 387 code, but may break some existing
10502 code that expects temporaries to be 80bit.
10504 This is the default choice for the x86-64 compiler.
10507 Attempt to utilize both instruction sets at once. This effectively double the
10508 amount of available registers and on chips with separate execution units for
10509 387 and SSE the execution resources too. Use this option with care, as it is
10510 still experimental, because the GCC register allocator does not model separate
10511 functional units well resulting in instable performance.
10514 @item -masm=@var{dialect}
10515 @opindex masm=@var{dialect}
10516 Output asm instructions using selected @var{dialect}. Supported
10517 choices are @samp{intel} or @samp{att} (the default one). Darwin does
10518 not support @samp{intel}.
10521 @itemx -mno-ieee-fp
10523 @opindex mno-ieee-fp
10524 Control whether or not the compiler uses IEEE floating point
10525 comparisons. These handle correctly the case where the result of a
10526 comparison is unordered.
10529 @opindex msoft-float
10530 Generate output containing library calls for floating point.
10531 @strong{Warning:} the requisite libraries are not part of GCC@.
10532 Normally the facilities of the machine's usual C compiler are used, but
10533 this can't be done directly in cross-compilation. You must make your
10534 own arrangements to provide suitable library functions for
10537 On machines where a function returns floating point results in the 80387
10538 register stack, some floating point opcodes may be emitted even if
10539 @option{-msoft-float} is used.
10541 @item -mno-fp-ret-in-387
10542 @opindex mno-fp-ret-in-387
10543 Do not use the FPU registers for return values of functions.
10545 The usual calling convention has functions return values of types
10546 @code{float} and @code{double} in an FPU register, even if there
10547 is no FPU@. The idea is that the operating system should emulate
10550 The option @option{-mno-fp-ret-in-387} causes such values to be returned
10551 in ordinary CPU registers instead.
10553 @item -mno-fancy-math-387
10554 @opindex mno-fancy-math-387
10555 Some 387 emulators do not support the @code{sin}, @code{cos} and
10556 @code{sqrt} instructions for the 387. Specify this option to avoid
10557 generating those instructions. This option is the default on FreeBSD,
10558 OpenBSD and NetBSD@. This option is overridden when @option{-march}
10559 indicates that the target cpu will always have an FPU and so the
10560 instruction will not need emulation. As of revision 2.6.1, these
10561 instructions are not generated unless you also use the
10562 @option{-funsafe-math-optimizations} switch.
10564 @item -malign-double
10565 @itemx -mno-align-double
10566 @opindex malign-double
10567 @opindex mno-align-double
10568 Control whether GCC aligns @code{double}, @code{long double}, and
10569 @code{long long} variables on a two word boundary or a one word
10570 boundary. Aligning @code{double} variables on a two word boundary will
10571 produce code that runs somewhat faster on a @samp{Pentium} at the
10572 expense of more memory.
10574 On x86-64, @option{-malign-double} is enabled by default.
10576 @strong{Warning:} if you use the @option{-malign-double} switch,
10577 structures containing the above types will be aligned differently than
10578 the published application binary interface specifications for the 386
10579 and will not be binary compatible with structures in code compiled
10580 without that switch.
10582 @item -m96bit-long-double
10583 @itemx -m128bit-long-double
10584 @opindex m96bit-long-double
10585 @opindex m128bit-long-double
10586 These switches control the size of @code{long double} type. The i386
10587 application binary interface specifies the size to be 96 bits,
10588 so @option{-m96bit-long-double} is the default in 32 bit mode.
10590 Modern architectures (Pentium and newer) would prefer @code{long double}
10591 to be aligned to an 8 or 16 byte boundary. In arrays or structures
10592 conforming to the ABI, this would not be possible. So specifying a
10593 @option{-m128bit-long-double} will align @code{long double}
10594 to a 16 byte boundary by padding the @code{long double} with an additional
10597 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
10598 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
10600 Notice that neither of these options enable any extra precision over the x87
10601 standard of 80 bits for a @code{long double}.
10603 @strong{Warning:} if you override the default value for your target ABI, the
10604 structures and arrays containing @code{long double} variables will change
10605 their size as well as function calling convention for function taking
10606 @code{long double} will be modified. Hence they will not be binary
10607 compatible with arrays or structures in code compiled without that switch.
10609 @item -mmlarge-data-threshold=@var{number}
10610 @opindex mlarge-data-threshold=@var{number}
10611 When @option{-mcmodel=medium} is specified, the data greater than
10612 @var{threshold} are placed in large data section. This value must be the
10613 same across all object linked into the binary and defaults to 65535.
10617 Use a different function-calling convention, in which functions that
10618 take a fixed number of arguments return with the @code{ret} @var{num}
10619 instruction, which pops their arguments while returning. This saves one
10620 instruction in the caller since there is no need to pop the arguments
10623 You can specify that an individual function is called with this calling
10624 sequence with the function attribute @samp{stdcall}. You can also
10625 override the @option{-mrtd} option by using the function attribute
10626 @samp{cdecl}. @xref{Function Attributes}.
10628 @strong{Warning:} this calling convention is incompatible with the one
10629 normally used on Unix, so you cannot use it if you need to call
10630 libraries compiled with the Unix compiler.
10632 Also, you must provide function prototypes for all functions that
10633 take variable numbers of arguments (including @code{printf});
10634 otherwise incorrect code will be generated for calls to those
10637 In addition, seriously incorrect code will result if you call a
10638 function with too many arguments. (Normally, extra arguments are
10639 harmlessly ignored.)
10641 @item -mregparm=@var{num}
10643 Control how many registers are used to pass integer arguments. By
10644 default, no registers are used to pass arguments, and at most 3
10645 registers can be used. You can control this behavior for a specific
10646 function by using the function attribute @samp{regparm}.
10647 @xref{Function Attributes}.
10649 @strong{Warning:} if you use this switch, and
10650 @var{num} is nonzero, then you must build all modules with the same
10651 value, including any libraries. This includes the system libraries and
10655 @opindex msseregparm
10656 Use SSE register passing conventions for float and double arguments
10657 and return values. You can control this behavior for a specific
10658 function by using the function attribute @samp{sseregparm}.
10659 @xref{Function Attributes}.
10661 @strong{Warning:} if you use this switch then you must build all
10662 modules with the same value, including any libraries. This includes
10663 the system libraries and startup modules.
10672 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
10673 is specified, the significands of results of floating-point operations are
10674 rounded to 24 bits (single precision); @option{-mpc64} rounds the the
10675 significands of results of floating-point operations to 53 bits (double
10676 precision) and @option{-mpc80} rounds the significands of results of
10677 floating-point operations to 64 bits (extended double precision), which is
10678 the default. When this option is used, floating-point operations in higher
10679 precisions are not available to the programmer without setting the FPU
10680 control word explicitly.
10682 Setting the rounding of floating-point operations to less than the default
10683 80 bits can speed some programs by 2% or more. Note that some mathematical
10684 libraries assume that extended precision (80 bit) floating-point operations
10685 are enabled by default; routines in such libraries could suffer significant
10686 loss of accuracy, typically through so-called "catastrophic cancellation",
10687 when this option is used to set the precision to less than extended precision.
10689 @item -mstackrealign
10690 @opindex mstackrealign
10691 Realign the stack at entry. On the Intel x86, the
10692 @option{-mstackrealign} option will generate an alternate prologue and
10693 epilogue that realigns the runtime stack. This supports mixing legacy
10694 codes that keep a 4-byte aligned stack with modern codes that keep a
10695 16-byte stack for SSE compatibility. The alternate prologue and
10696 epilogue are slower and bigger than the regular ones, and the
10697 alternate prologue requires an extra scratch register; this lowers the
10698 number of registers available if used in conjunction with the
10699 @code{regparm} attribute. The @option{-mstackrealign} option is
10700 incompatible with the nested function prologue; this is considered a
10701 hard error. See also the attribute @code{force_align_arg_pointer},
10702 applicable to individual functions.
10704 @item -mpreferred-stack-boundary=@var{num}
10705 @opindex mpreferred-stack-boundary
10706 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
10707 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
10708 the default is 4 (16 bytes or 128 bits).
10710 On Pentium and PentiumPro, @code{double} and @code{long double} values
10711 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
10712 suffer significant run time performance penalties. On Pentium III, the
10713 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
10714 properly if it is not 16 byte aligned.
10716 To ensure proper alignment of this values on the stack, the stack boundary
10717 must be as aligned as that required by any value stored on the stack.
10718 Further, every function must be generated such that it keeps the stack
10719 aligned. Thus calling a function compiled with a higher preferred
10720 stack boundary from a function compiled with a lower preferred stack
10721 boundary will most likely misalign the stack. It is recommended that
10722 libraries that use callbacks always use the default setting.
10724 This extra alignment does consume extra stack space, and generally
10725 increases code size. Code that is sensitive to stack space usage, such
10726 as embedded systems and operating system kernels, may want to reduce the
10727 preferred alignment to @option{-mpreferred-stack-boundary=2}.
10765 These switches enable or disable the use of instructions in the MMX,
10766 SSE, SSE2, SSE3, SSSE3, SSE4.1, AES, PCLMUL, SSE4A, SSE5, ABM or
10767 3DNow!@: extended instruction sets.
10768 These extensions are also available as built-in functions: see
10769 @ref{X86 Built-in Functions}, for details of the functions enabled and
10770 disabled by these switches.
10772 To have SSE/SSE2 instructions generated automatically from floating-point
10773 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
10775 These options will enable GCC to use these extended instructions in
10776 generated code, even without @option{-mfpmath=sse}. Applications which
10777 perform runtime CPU detection must compile separate files for each
10778 supported architecture, using the appropriate flags. In particular,
10779 the file containing the CPU detection code should be compiled without
10784 This option will enable GCC to use CMPXCHG16B instruction in generated code.
10785 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
10786 data types. This is useful for high resolution counters that could be updated
10787 by multiple processors (or cores). This instruction is generated as part of
10788 atomic built-in functions: see @ref{Atomic Builtins} for details.
10792 This option will enable GCC to use SAHF instruction in generated 64-bit code.
10793 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
10794 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
10795 SAHF are load and store instructions, respectively, for certain status flags.
10796 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
10797 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
10801 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
10802 vectorized variants RCPPS and RSQRTPS) with an additional Newton-Rhapson step
10803 to increase precision instead of DIVSS and SQRTSS (and their vectorized
10804 variants) for single precision floating point arguments. These instructions
10805 are generated only when @option{-funsafe-math-optimizations} is enabled
10806 together with @option{-finite-math-only} and @option{-fno-trapping-math}.
10807 Note that while the throughput of the sequence is higher than the throughput
10808 of the non-reciprocal instruction, the precision of the sequence can be
10809 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
10811 @item -mveclibabi=@var{type}
10812 @opindex mveclibabi
10813 Specifies the ABI type to use for vectorizing intrinsics using an
10814 external library. Supported types are @code{svml} for the Intel short
10815 vector math library and @code{acml} for the AMD math core library style
10816 of interfacing. GCC will currently emit calls to @code{vmldExp2},
10817 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
10818 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
10819 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
10820 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
10821 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
10822 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
10823 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
10824 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
10825 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
10826 function type when @option{-mveclibabi=svml} is used and @code{__vrd2_sin},
10827 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
10828 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
10829 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
10830 @code{__vrs4_log10f} and @code{__vrs4_powf} for corresponding function type
10831 when @option{-mveclibabi=acml} is used. Both @option{-ftree-vectorize} and
10832 @option{-funsafe-math-optimizations} have to be enabled. A SVML or ACML ABI
10833 compatible library will have to be specified at link time.
10836 @itemx -mno-push-args
10837 @opindex mpush-args
10838 @opindex mno-push-args
10839 Use PUSH operations to store outgoing parameters. This method is shorter
10840 and usually equally fast as method using SUB/MOV operations and is enabled
10841 by default. In some cases disabling it may improve performance because of
10842 improved scheduling and reduced dependencies.
10844 @item -maccumulate-outgoing-args
10845 @opindex maccumulate-outgoing-args
10846 If enabled, the maximum amount of space required for outgoing arguments will be
10847 computed in the function prologue. This is faster on most modern CPUs
10848 because of reduced dependencies, improved scheduling and reduced stack usage
10849 when preferred stack boundary is not equal to 2. The drawback is a notable
10850 increase in code size. This switch implies @option{-mno-push-args}.
10854 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
10855 on thread-safe exception handling must compile and link all code with the
10856 @option{-mthreads} option. When compiling, @option{-mthreads} defines
10857 @option{-D_MT}; when linking, it links in a special thread helper library
10858 @option{-lmingwthrd} which cleans up per thread exception handling data.
10860 @item -mno-align-stringops
10861 @opindex mno-align-stringops
10862 Do not align destination of inlined string operations. This switch reduces
10863 code size and improves performance in case the destination is already aligned,
10864 but GCC doesn't know about it.
10866 @item -minline-all-stringops
10867 @opindex minline-all-stringops
10868 By default GCC inlines string operations only when destination is known to be
10869 aligned at least to 4 byte boundary. This enables more inlining, increase code
10870 size, but may improve performance of code that depends on fast memcpy, strlen
10871 and memset for short lengths.
10873 @item -minline-stringops-dynamically
10874 @opindex minline-stringops-dynamically
10875 For string operation of unknown size, inline runtime checks so for small
10876 blocks inline code is used, while for large blocks library call is used.
10878 @item -mstringop-strategy=@var{alg}
10879 @opindex mstringop-strategy=@var{alg}
10880 Overwrite internal decision heuristic about particular algorithm to inline
10881 string operation with. The allowed values are @code{rep_byte},
10882 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
10883 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
10884 expanding inline loop, @code{libcall} for always expanding library call.
10886 @item -momit-leaf-frame-pointer
10887 @opindex momit-leaf-frame-pointer
10888 Don't keep the frame pointer in a register for leaf functions. This
10889 avoids the instructions to save, set up and restore frame pointers and
10890 makes an extra register available in leaf functions. The option
10891 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10892 which might make debugging harder.
10894 @item -mtls-direct-seg-refs
10895 @itemx -mno-tls-direct-seg-refs
10896 @opindex mtls-direct-seg-refs
10897 Controls whether TLS variables may be accessed with offsets from the
10898 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
10899 or whether the thread base pointer must be added. Whether or not this
10900 is legal depends on the operating system, and whether it maps the
10901 segment to cover the entire TLS area.
10903 For systems that use GNU libc, the default is on.
10906 @itemx -mno-fused-madd
10907 @opindex mfused-madd
10908 Enable automatic generation of fused floating point multiply-add instructions
10909 if the ISA supports such instructions. The -mfused-madd option is on by
10910 default. The fused multiply-add instructions have a different
10911 rounding behavior compared to executing a multiply followed by an add.
10914 These @samp{-m} switches are supported in addition to the above
10915 on AMD x86-64 processors in 64-bit environments.
10922 Generate code for a 32-bit or 64-bit environment.
10923 The 32-bit environment sets int, long and pointer to 32 bits and
10924 generates code that runs on any i386 system.
10925 The 64-bit environment sets int to 32 bits and long and pointer
10926 to 64 bits and generates code for AMD's x86-64 architecture. For
10927 darwin only the -m64 option turns off the @option{-fno-pic} and
10928 @option{-mdynamic-no-pic} options.
10930 @item -mno-red-zone
10931 @opindex no-red-zone
10932 Do not use a so called red zone for x86-64 code. The red zone is mandated
10933 by the x86-64 ABI, it is a 128-byte area beyond the location of the
10934 stack pointer that will not be modified by signal or interrupt handlers
10935 and therefore can be used for temporary data without adjusting the stack
10936 pointer. The flag @option{-mno-red-zone} disables this red zone.
10938 @item -mcmodel=small
10939 @opindex mcmodel=small
10940 Generate code for the small code model: the program and its symbols must
10941 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
10942 Programs can be statically or dynamically linked. This is the default
10945 @item -mcmodel=kernel
10946 @opindex mcmodel=kernel
10947 Generate code for the kernel code model. The kernel runs in the
10948 negative 2 GB of the address space.
10949 This model has to be used for Linux kernel code.
10951 @item -mcmodel=medium
10952 @opindex mcmodel=medium
10953 Generate code for the medium model: The program is linked in the lower 2
10954 GB of the address space but symbols can be located anywhere in the
10955 address space. Programs can be statically or dynamically linked, but
10956 building of shared libraries are not supported with the medium model.
10958 @item -mcmodel=large
10959 @opindex mcmodel=large
10960 Generate code for the large model: This model makes no assumptions
10961 about addresses and sizes of sections.
10964 @node IA-64 Options
10965 @subsection IA-64 Options
10966 @cindex IA-64 Options
10968 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10972 @opindex mbig-endian
10973 Generate code for a big endian target. This is the default for HP-UX@.
10975 @item -mlittle-endian
10976 @opindex mlittle-endian
10977 Generate code for a little endian target. This is the default for AIX5
10983 @opindex mno-gnu-as
10984 Generate (or don't) code for the GNU assembler. This is the default.
10985 @c Also, this is the default if the configure option @option{--with-gnu-as}
10991 @opindex mno-gnu-ld
10992 Generate (or don't) code for the GNU linker. This is the default.
10993 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10998 Generate code that does not use a global pointer register. The result
10999 is not position independent code, and violates the IA-64 ABI@.
11001 @item -mvolatile-asm-stop
11002 @itemx -mno-volatile-asm-stop
11003 @opindex mvolatile-asm-stop
11004 @opindex mno-volatile-asm-stop
11005 Generate (or don't) a stop bit immediately before and after volatile asm
11008 @item -mregister-names
11009 @itemx -mno-register-names
11010 @opindex mregister-names
11011 @opindex mno-register-names
11012 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
11013 the stacked registers. This may make assembler output more readable.
11019 Disable (or enable) optimizations that use the small data section. This may
11020 be useful for working around optimizer bugs.
11022 @item -mconstant-gp
11023 @opindex mconstant-gp
11024 Generate code that uses a single constant global pointer value. This is
11025 useful when compiling kernel code.
11029 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
11030 This is useful when compiling firmware code.
11032 @item -minline-float-divide-min-latency
11033 @opindex minline-float-divide-min-latency
11034 Generate code for inline divides of floating point values
11035 using the minimum latency algorithm.
11037 @item -minline-float-divide-max-throughput
11038 @opindex minline-float-divide-max-throughput
11039 Generate code for inline divides of floating point values
11040 using the maximum throughput algorithm.
11042 @item -minline-int-divide-min-latency
11043 @opindex minline-int-divide-min-latency
11044 Generate code for inline divides of integer values
11045 using the minimum latency algorithm.
11047 @item -minline-int-divide-max-throughput
11048 @opindex minline-int-divide-max-throughput
11049 Generate code for inline divides of integer values
11050 using the maximum throughput algorithm.
11052 @item -minline-sqrt-min-latency
11053 @opindex minline-sqrt-min-latency
11054 Generate code for inline square roots
11055 using the minimum latency algorithm.
11057 @item -minline-sqrt-max-throughput
11058 @opindex minline-sqrt-max-throughput
11059 Generate code for inline square roots
11060 using the maximum throughput algorithm.
11062 @item -mno-dwarf2-asm
11063 @itemx -mdwarf2-asm
11064 @opindex mno-dwarf2-asm
11065 @opindex mdwarf2-asm
11066 Don't (or do) generate assembler code for the DWARF2 line number debugging
11067 info. This may be useful when not using the GNU assembler.
11069 @item -mearly-stop-bits
11070 @itemx -mno-early-stop-bits
11071 @opindex mearly-stop-bits
11072 @opindex mno-early-stop-bits
11073 Allow stop bits to be placed earlier than immediately preceding the
11074 instruction that triggered the stop bit. This can improve instruction
11075 scheduling, but does not always do so.
11077 @item -mfixed-range=@var{register-range}
11078 @opindex mfixed-range
11079 Generate code treating the given register range as fixed registers.
11080 A fixed register is one that the register allocator can not use. This is
11081 useful when compiling kernel code. A register range is specified as
11082 two registers separated by a dash. Multiple register ranges can be
11083 specified separated by a comma.
11085 @item -mtls-size=@var{tls-size}
11087 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
11090 @item -mtune=@var{cpu-type}
11092 Tune the instruction scheduling for a particular CPU, Valid values are
11093 itanium, itanium1, merced, itanium2, and mckinley.
11099 Add support for multithreading using the POSIX threads library. This
11100 option sets flags for both the preprocessor and linker. It does
11101 not affect the thread safety of object code produced by the compiler or
11102 that of libraries supplied with it. These are HP-UX specific flags.
11108 Generate code for a 32-bit or 64-bit environment.
11109 The 32-bit environment sets int, long and pointer to 32 bits.
11110 The 64-bit environment sets int to 32 bits and long and pointer
11111 to 64 bits. These are HP-UX specific flags.
11113 @item -mno-sched-br-data-spec
11114 @itemx -msched-br-data-spec
11115 @opindex mno-sched-br-data-spec
11116 @opindex msched-br-data-spec
11117 (Dis/En)able data speculative scheduling before reload.
11118 This will result in generation of the ld.a instructions and
11119 the corresponding check instructions (ld.c / chk.a).
11120 The default is 'disable'.
11122 @item -msched-ar-data-spec
11123 @itemx -mno-sched-ar-data-spec
11124 @opindex msched-ar-data-spec
11125 @opindex mno-sched-ar-data-spec
11126 (En/Dis)able data speculative scheduling after reload.
11127 This will result in generation of the ld.a instructions and
11128 the corresponding check instructions (ld.c / chk.a).
11129 The default is 'enable'.
11131 @item -mno-sched-control-spec
11132 @itemx -msched-control-spec
11133 @opindex mno-sched-control-spec
11134 @opindex msched-control-spec
11135 (Dis/En)able control speculative scheduling. This feature is
11136 available only during region scheduling (i.e.@: before reload).
11137 This will result in generation of the ld.s instructions and
11138 the corresponding check instructions chk.s .
11139 The default is 'disable'.
11141 @item -msched-br-in-data-spec
11142 @itemx -mno-sched-br-in-data-spec
11143 @opindex msched-br-in-data-spec
11144 @opindex mno-sched-br-in-data-spec
11145 (En/Dis)able speculative scheduling of the instructions that
11146 are dependent on the data speculative loads before reload.
11147 This is effective only with @option{-msched-br-data-spec} enabled.
11148 The default is 'enable'.
11150 @item -msched-ar-in-data-spec
11151 @itemx -mno-sched-ar-in-data-spec
11152 @opindex msched-ar-in-data-spec
11153 @opindex mno-sched-ar-in-data-spec
11154 (En/Dis)able speculative scheduling of the instructions that
11155 are dependent on the data speculative loads after reload.
11156 This is effective only with @option{-msched-ar-data-spec} enabled.
11157 The default is 'enable'.
11159 @item -msched-in-control-spec
11160 @itemx -mno-sched-in-control-spec
11161 @opindex msched-in-control-spec
11162 @opindex mno-sched-in-control-spec
11163 (En/Dis)able speculative scheduling of the instructions that
11164 are dependent on the control speculative loads.
11165 This is effective only with @option{-msched-control-spec} enabled.
11166 The default is 'enable'.
11169 @itemx -mno-sched-ldc
11170 @opindex msched-ldc
11171 @opindex mno-sched-ldc
11172 (En/Dis)able use of simple data speculation checks ld.c .
11173 If disabled, only chk.a instructions will be emitted to check
11174 data speculative loads.
11175 The default is 'enable'.
11177 @item -mno-sched-control-ldc
11178 @itemx -msched-control-ldc
11179 @opindex mno-sched-control-ldc
11180 @opindex msched-control-ldc
11181 (Dis/En)able use of ld.c instructions to check control speculative loads.
11182 If enabled, in case of control speculative load with no speculatively
11183 scheduled dependent instructions this load will be emitted as ld.sa and
11184 ld.c will be used to check it.
11185 The default is 'disable'.
11187 @item -mno-sched-spec-verbose
11188 @itemx -msched-spec-verbose
11189 @opindex mno-sched-spec-verbose
11190 @opindex msched-spec-verbose
11191 (Dis/En)able printing of the information about speculative motions.
11193 @item -mno-sched-prefer-non-data-spec-insns
11194 @itemx -msched-prefer-non-data-spec-insns
11195 @opindex mno-sched-prefer-non-data-spec-insns
11196 @opindex msched-prefer-non-data-spec-insns
11197 If enabled, data speculative instructions will be chosen for schedule
11198 only if there are no other choices at the moment. This will make
11199 the use of the data speculation much more conservative.
11200 The default is 'disable'.
11202 @item -mno-sched-prefer-non-control-spec-insns
11203 @itemx -msched-prefer-non-control-spec-insns
11204 @opindex mno-sched-prefer-non-control-spec-insns
11205 @opindex msched-prefer-non-control-spec-insns
11206 If enabled, control speculative instructions will be chosen for schedule
11207 only if there are no other choices at the moment. This will make
11208 the use of the control speculation much more conservative.
11209 The default is 'disable'.
11211 @item -mno-sched-count-spec-in-critical-path
11212 @itemx -msched-count-spec-in-critical-path
11213 @opindex mno-sched-count-spec-in-critical-path
11214 @opindex msched-count-spec-in-critical-path
11215 If enabled, speculative dependencies will be considered during
11216 computation of the instructions priorities. This will make the use of the
11217 speculation a bit more conservative.
11218 The default is 'disable'.
11223 @subsection M32C Options
11224 @cindex M32C options
11227 @item -mcpu=@var{name}
11229 Select the CPU for which code is generated. @var{name} may be one of
11230 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
11231 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
11232 the M32C/80 series.
11236 Specifies that the program will be run on the simulator. This causes
11237 an alternate runtime library to be linked in which supports, for
11238 example, file I/O@. You must not use this option when generating
11239 programs that will run on real hardware; you must provide your own
11240 runtime library for whatever I/O functions are needed.
11242 @item -memregs=@var{number}
11244 Specifies the number of memory-based pseudo-registers GCC will use
11245 during code generation. These pseudo-registers will be used like real
11246 registers, so there is a tradeoff between GCC's ability to fit the
11247 code into available registers, and the performance penalty of using
11248 memory instead of registers. Note that all modules in a program must
11249 be compiled with the same value for this option. Because of that, you
11250 must not use this option with the default runtime libraries gcc
11255 @node M32R/D Options
11256 @subsection M32R/D Options
11257 @cindex M32R/D options
11259 These @option{-m} options are defined for Renesas M32R/D architectures:
11264 Generate code for the M32R/2@.
11268 Generate code for the M32R/X@.
11272 Generate code for the M32R@. This is the default.
11274 @item -mmodel=small
11275 @opindex mmodel=small
11276 Assume all objects live in the lower 16MB of memory (so that their addresses
11277 can be loaded with the @code{ld24} instruction), and assume all subroutines
11278 are reachable with the @code{bl} instruction.
11279 This is the default.
11281 The addressability of a particular object can be set with the
11282 @code{model} attribute.
11284 @item -mmodel=medium
11285 @opindex mmodel=medium
11286 Assume objects may be anywhere in the 32-bit address space (the compiler
11287 will generate @code{seth/add3} instructions to load their addresses), and
11288 assume all subroutines are reachable with the @code{bl} instruction.
11290 @item -mmodel=large
11291 @opindex mmodel=large
11292 Assume objects may be anywhere in the 32-bit address space (the compiler
11293 will generate @code{seth/add3} instructions to load their addresses), and
11294 assume subroutines may not be reachable with the @code{bl} instruction
11295 (the compiler will generate the much slower @code{seth/add3/jl}
11296 instruction sequence).
11299 @opindex msdata=none
11300 Disable use of the small data area. Variables will be put into
11301 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
11302 @code{section} attribute has been specified).
11303 This is the default.
11305 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
11306 Objects may be explicitly put in the small data area with the
11307 @code{section} attribute using one of these sections.
11309 @item -msdata=sdata
11310 @opindex msdata=sdata
11311 Put small global and static data in the small data area, but do not
11312 generate special code to reference them.
11315 @opindex msdata=use
11316 Put small global and static data in the small data area, and generate
11317 special instructions to reference them.
11321 @cindex smaller data references
11322 Put global and static objects less than or equal to @var{num} bytes
11323 into the small data or bss sections instead of the normal data or bss
11324 sections. The default value of @var{num} is 8.
11325 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
11326 for this option to have any effect.
11328 All modules should be compiled with the same @option{-G @var{num}} value.
11329 Compiling with different values of @var{num} may or may not work; if it
11330 doesn't the linker will give an error message---incorrect code will not be
11335 Makes the M32R specific code in the compiler display some statistics
11336 that might help in debugging programs.
11338 @item -malign-loops
11339 @opindex malign-loops
11340 Align all loops to a 32-byte boundary.
11342 @item -mno-align-loops
11343 @opindex mno-align-loops
11344 Do not enforce a 32-byte alignment for loops. This is the default.
11346 @item -missue-rate=@var{number}
11347 @opindex missue-rate=@var{number}
11348 Issue @var{number} instructions per cycle. @var{number} can only be 1
11351 @item -mbranch-cost=@var{number}
11352 @opindex mbranch-cost=@var{number}
11353 @var{number} can only be 1 or 2. If it is 1 then branches will be
11354 preferred over conditional code, if it is 2, then the opposite will
11357 @item -mflush-trap=@var{number}
11358 @opindex mflush-trap=@var{number}
11359 Specifies the trap number to use to flush the cache. The default is
11360 12. Valid numbers are between 0 and 15 inclusive.
11362 @item -mno-flush-trap
11363 @opindex mno-flush-trap
11364 Specifies that the cache cannot be flushed by using a trap.
11366 @item -mflush-func=@var{name}
11367 @opindex mflush-func=@var{name}
11368 Specifies the name of the operating system function to call to flush
11369 the cache. The default is @emph{_flush_cache}, but a function call
11370 will only be used if a trap is not available.
11372 @item -mno-flush-func
11373 @opindex mno-flush-func
11374 Indicates that there is no OS function for flushing the cache.
11378 @node M680x0 Options
11379 @subsection M680x0 Options
11380 @cindex M680x0 options
11382 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
11383 The default settings depend on which architecture was selected when
11384 the compiler was configured; the defaults for the most common choices
11388 @item -march=@var{arch}
11390 Generate code for a specific M680x0 or ColdFire instruction set
11391 architecture. Permissible values of @var{arch} for M680x0
11392 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
11393 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
11394 architectures are selected according to Freescale's ISA classification
11395 and the permissible values are: @samp{isaa}, @samp{isaaplus},
11396 @samp{isab} and @samp{isac}.
11398 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
11399 code for a ColdFire target. The @var{arch} in this macro is one of the
11400 @option{-march} arguments given above.
11402 When used together, @option{-march} and @option{-mtune} select code
11403 that runs on a family of similar processors but that is optimized
11404 for a particular microarchitecture.
11406 @item -mcpu=@var{cpu}
11408 Generate code for a specific M680x0 or ColdFire processor.
11409 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
11410 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
11411 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
11412 below, which also classifies the CPUs into families:
11414 @multitable @columnfractions 0.20 0.80
11415 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
11416 @item @samp{51qe} @tab @samp{51qe}
11417 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
11418 @item @samp{5206e} @tab @samp{5206e}
11419 @item @samp{5208} @tab @samp{5207} @samp{5208}
11420 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
11421 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
11422 @item @samp{5216} @tab @samp{5214} @samp{5216}
11423 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
11424 @item @samp{5225} @tab @samp{5224} @samp{5225}
11425 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
11426 @item @samp{5249} @tab @samp{5249}
11427 @item @samp{5250} @tab @samp{5250}
11428 @item @samp{5271} @tab @samp{5270} @samp{5271}
11429 @item @samp{5272} @tab @samp{5272}
11430 @item @samp{5275} @tab @samp{5274} @samp{5275}
11431 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
11432 @item @samp{5307} @tab @samp{5307}
11433 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
11434 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
11435 @item @samp{5407} @tab @samp{5407}
11436 @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}
11439 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
11440 @var{arch} is compatible with @var{cpu}. Other combinations of
11441 @option{-mcpu} and @option{-march} are rejected.
11443 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
11444 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
11445 where the value of @var{family} is given by the table above.
11447 @item -mtune=@var{tune}
11449 Tune the code for a particular microarchitecture, within the
11450 constraints set by @option{-march} and @option{-mcpu}.
11451 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
11452 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
11453 and @samp{cpu32}. The ColdFire microarchitectures
11454 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
11456 You can also use @option{-mtune=68020-40} for code that needs
11457 to run relatively well on 68020, 68030 and 68040 targets.
11458 @option{-mtune=68020-60} is similar but includes 68060 targets
11459 as well. These two options select the same tuning decisions as
11460 @option{-m68020-40} and @option{-m68020-60} respectively.
11462 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
11463 when tuning for 680x0 architecture @var{arch}. It also defines
11464 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
11465 option is used. If gcc is tuning for a range of architectures,
11466 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
11467 it defines the macros for every architecture in the range.
11469 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
11470 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
11471 of the arguments given above.
11477 Generate output for a 68000. This is the default
11478 when the compiler is configured for 68000-based systems.
11479 It is equivalent to @option{-march=68000}.
11481 Use this option for microcontrollers with a 68000 or EC000 core,
11482 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
11486 Generate output for a 68010. This is the default
11487 when the compiler is configured for 68010-based systems.
11488 It is equivalent to @option{-march=68010}.
11494 Generate output for a 68020. This is the default
11495 when the compiler is configured for 68020-based systems.
11496 It is equivalent to @option{-march=68020}.
11500 Generate output for a 68030. This is the default when the compiler is
11501 configured for 68030-based systems. It is equivalent to
11502 @option{-march=68030}.
11506 Generate output for a 68040. This is the default when the compiler is
11507 configured for 68040-based systems. It is equivalent to
11508 @option{-march=68040}.
11510 This option inhibits the use of 68881/68882 instructions that have to be
11511 emulated by software on the 68040. Use this option if your 68040 does not
11512 have code to emulate those instructions.
11516 Generate output for a 68060. This is the default when the compiler is
11517 configured for 68060-based systems. It is equivalent to
11518 @option{-march=68060}.
11520 This option inhibits the use of 68020 and 68881/68882 instructions that
11521 have to be emulated by software on the 68060. Use this option if your 68060
11522 does not have code to emulate those instructions.
11526 Generate output for a CPU32. This is the default
11527 when the compiler is configured for CPU32-based systems.
11528 It is equivalent to @option{-march=cpu32}.
11530 Use this option for microcontrollers with a
11531 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
11532 68336, 68340, 68341, 68349 and 68360.
11536 Generate output for a 520X ColdFire CPU@. This is the default
11537 when the compiler is configured for 520X-based systems.
11538 It is equivalent to @option{-mcpu=5206}, and is now deprecated
11539 in favor of that option.
11541 Use this option for microcontroller with a 5200 core, including
11542 the MCF5202, MCF5203, MCF5204 and MCF5206.
11546 Generate output for a 5206e ColdFire CPU@. The option is now
11547 deprecated in favor of the equivalent @option{-mcpu=5206e}.
11551 Generate output for a member of the ColdFire 528X family.
11552 The option is now deprecated in favor of the equivalent
11553 @option{-mcpu=528x}.
11557 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
11558 in favor of the equivalent @option{-mcpu=5307}.
11562 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
11563 in favor of the equivalent @option{-mcpu=5407}.
11567 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
11568 This includes use of hardware floating point instructions.
11569 The option is equivalent to @option{-mcpu=547x}, and is now
11570 deprecated in favor of that option.
11574 Generate output for a 68040, without using any of the new instructions.
11575 This results in code which can run relatively efficiently on either a
11576 68020/68881 or a 68030 or a 68040. The generated code does use the
11577 68881 instructions that are emulated on the 68040.
11579 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
11583 Generate output for a 68060, without using any of the new instructions.
11584 This results in code which can run relatively efficiently on either a
11585 68020/68881 or a 68030 or a 68040. The generated code does use the
11586 68881 instructions that are emulated on the 68060.
11588 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
11592 @opindex mhard-float
11594 Generate floating-point instructions. This is the default for 68020
11595 and above, and for ColdFire devices that have an FPU@. It defines the
11596 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
11597 on ColdFire targets.
11600 @opindex msoft-float
11601 Do not generate floating-point instructions; use library calls instead.
11602 This is the default for 68000, 68010, and 68832 targets. It is also
11603 the default for ColdFire devices that have no FPU.
11609 Generate (do not generate) ColdFire hardware divide and remainder
11610 instructions. If @option{-march} is used without @option{-mcpu},
11611 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
11612 architectures. Otherwise, the default is taken from the target CPU
11613 (either the default CPU, or the one specified by @option{-mcpu}). For
11614 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
11615 @option{-mcpu=5206e}.
11617 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
11621 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11622 Additionally, parameters passed on the stack are also aligned to a
11623 16-bit boundary even on targets whose API mandates promotion to 32-bit.
11627 Do not consider type @code{int} to be 16 bits wide. This is the default.
11630 @itemx -mno-bitfield
11631 @opindex mnobitfield
11632 @opindex mno-bitfield
11633 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
11634 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
11638 Do use the bit-field instructions. The @option{-m68020} option implies
11639 @option{-mbitfield}. This is the default if you use a configuration
11640 designed for a 68020.
11644 Use a different function-calling convention, in which functions
11645 that take a fixed number of arguments return with the @code{rtd}
11646 instruction, which pops their arguments while returning. This
11647 saves one instruction in the caller since there is no need to pop
11648 the arguments there.
11650 This calling convention is incompatible with the one normally
11651 used on Unix, so you cannot use it if you need to call libraries
11652 compiled with the Unix compiler.
11654 Also, you must provide function prototypes for all functions that
11655 take variable numbers of arguments (including @code{printf});
11656 otherwise incorrect code will be generated for calls to those
11659 In addition, seriously incorrect code will result if you call a
11660 function with too many arguments. (Normally, extra arguments are
11661 harmlessly ignored.)
11663 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
11664 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
11668 Do not use the calling conventions selected by @option{-mrtd}.
11669 This is the default.
11672 @itemx -mno-align-int
11673 @opindex malign-int
11674 @opindex mno-align-int
11675 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
11676 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
11677 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
11678 Aligning variables on 32-bit boundaries produces code that runs somewhat
11679 faster on processors with 32-bit busses at the expense of more memory.
11681 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
11682 align structures containing the above types differently than
11683 most published application binary interface specifications for the m68k.
11687 Use the pc-relative addressing mode of the 68000 directly, instead of
11688 using a global offset table. At present, this option implies @option{-fpic},
11689 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
11690 not presently supported with @option{-mpcrel}, though this could be supported for
11691 68020 and higher processors.
11693 @item -mno-strict-align
11694 @itemx -mstrict-align
11695 @opindex mno-strict-align
11696 @opindex mstrict-align
11697 Do not (do) assume that unaligned memory references will be handled by
11701 Generate code that allows the data segment to be located in a different
11702 area of memory from the text segment. This allows for execute in place in
11703 an environment without virtual memory management. This option implies
11706 @item -mno-sep-data
11707 Generate code that assumes that the data segment follows the text segment.
11708 This is the default.
11710 @item -mid-shared-library
11711 Generate code that supports shared libraries via the library ID method.
11712 This allows for execute in place and shared libraries in an environment
11713 without virtual memory management. This option implies @option{-fPIC}.
11715 @item -mno-id-shared-library
11716 Generate code that doesn't assume ID based shared libraries are being used.
11717 This is the default.
11719 @item -mshared-library-id=n
11720 Specified the identification number of the ID based shared library being
11721 compiled. Specifying a value of 0 will generate more compact code, specifying
11722 other values will force the allocation of that number to the current
11723 library but is no more space or time efficient than omitting this option.
11727 @node M68hc1x Options
11728 @subsection M68hc1x Options
11729 @cindex M68hc1x options
11731 These are the @samp{-m} options defined for the 68hc11 and 68hc12
11732 microcontrollers. The default values for these options depends on
11733 which style of microcontroller was selected when the compiler was configured;
11734 the defaults for the most common choices are given below.
11741 Generate output for a 68HC11. This is the default
11742 when the compiler is configured for 68HC11-based systems.
11748 Generate output for a 68HC12. This is the default
11749 when the compiler is configured for 68HC12-based systems.
11755 Generate output for a 68HCS12.
11757 @item -mauto-incdec
11758 @opindex mauto-incdec
11759 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
11766 Enable the use of 68HC12 min and max instructions.
11769 @itemx -mno-long-calls
11770 @opindex mlong-calls
11771 @opindex mno-long-calls
11772 Treat all calls as being far away (near). If calls are assumed to be
11773 far away, the compiler will use the @code{call} instruction to
11774 call a function and the @code{rtc} instruction for returning.
11778 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11780 @item -msoft-reg-count=@var{count}
11781 @opindex msoft-reg-count
11782 Specify the number of pseudo-soft registers which are used for the
11783 code generation. The maximum number is 32. Using more pseudo-soft
11784 register may or may not result in better code depending on the program.
11785 The default is 4 for 68HC11 and 2 for 68HC12.
11789 @node MCore Options
11790 @subsection MCore Options
11791 @cindex MCore options
11793 These are the @samp{-m} options defined for the Motorola M*Core
11799 @itemx -mno-hardlit
11801 @opindex mno-hardlit
11802 Inline constants into the code stream if it can be done in two
11803 instructions or less.
11809 Use the divide instruction. (Enabled by default).
11811 @item -mrelax-immediate
11812 @itemx -mno-relax-immediate
11813 @opindex mrelax-immediate
11814 @opindex mno-relax-immediate
11815 Allow arbitrary sized immediates in bit operations.
11817 @item -mwide-bitfields
11818 @itemx -mno-wide-bitfields
11819 @opindex mwide-bitfields
11820 @opindex mno-wide-bitfields
11821 Always treat bit-fields as int-sized.
11823 @item -m4byte-functions
11824 @itemx -mno-4byte-functions
11825 @opindex m4byte-functions
11826 @opindex mno-4byte-functions
11827 Force all functions to be aligned to a four byte boundary.
11829 @item -mcallgraph-data
11830 @itemx -mno-callgraph-data
11831 @opindex mcallgraph-data
11832 @opindex mno-callgraph-data
11833 Emit callgraph information.
11836 @itemx -mno-slow-bytes
11837 @opindex mslow-bytes
11838 @opindex mno-slow-bytes
11839 Prefer word access when reading byte quantities.
11841 @item -mlittle-endian
11842 @itemx -mbig-endian
11843 @opindex mlittle-endian
11844 @opindex mbig-endian
11845 Generate code for a little endian target.
11851 Generate code for the 210 processor.
11855 @subsection MIPS Options
11856 @cindex MIPS options
11862 Generate big-endian code.
11866 Generate little-endian code. This is the default for @samp{mips*el-*-*}
11869 @item -march=@var{arch}
11871 Generate code that will run on @var{arch}, which can be the name of a
11872 generic MIPS ISA, or the name of a particular processor.
11874 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
11875 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
11876 The processor names are:
11877 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
11878 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
11879 @samp{5kc}, @samp{5kf},
11881 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
11882 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
11883 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
11884 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
11887 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
11888 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
11889 @samp{rm7000}, @samp{rm9000},
11892 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
11893 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
11894 The special value @samp{from-abi} selects the
11895 most compatible architecture for the selected ABI (that is,
11896 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
11898 In processor names, a final @samp{000} can be abbreviated as @samp{k}
11899 (for example, @samp{-march=r2k}). Prefixes are optional, and
11900 @samp{vr} may be written @samp{r}.
11902 Names of the form @samp{@var{n}f2_1} refer to processors with
11903 FPUs clocked at half the rate of the core, names of the form
11904 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
11905 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
11906 processors with FPUs clocked a ratio of 3:2 with respect to the core.
11907 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
11908 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
11909 accepted as synonyms for @samp{@var{n}f1_1}.
11911 GCC defines two macros based on the value of this option. The first
11912 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
11913 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
11914 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
11915 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
11916 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
11918 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
11919 above. In other words, it will have the full prefix and will not
11920 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
11921 the macro names the resolved architecture (either @samp{"mips1"} or
11922 @samp{"mips3"}). It names the default architecture when no
11923 @option{-march} option is given.
11925 @item -mtune=@var{arch}
11927 Optimize for @var{arch}. Among other things, this option controls
11928 the way instructions are scheduled, and the perceived cost of arithmetic
11929 operations. The list of @var{arch} values is the same as for
11932 When this option is not used, GCC will optimize for the processor
11933 specified by @option{-march}. By using @option{-march} and
11934 @option{-mtune} together, it is possible to generate code that will
11935 run on a family of processors, but optimize the code for one
11936 particular member of that family.
11938 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
11939 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
11940 @samp{-march} ones described above.
11944 Equivalent to @samp{-march=mips1}.
11948 Equivalent to @samp{-march=mips2}.
11952 Equivalent to @samp{-march=mips3}.
11956 Equivalent to @samp{-march=mips4}.
11960 Equivalent to @samp{-march=mips32}.
11964 Equivalent to @samp{-march=mips32r2}.
11968 Equivalent to @samp{-march=mips64}.
11973 @opindex mno-mips16
11974 Generate (do not generate) MIPS16 code. If GCC is targetting a
11975 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
11977 MIPS16 code generation can also be controlled on a per-function basis
11978 by means of @code{mips16} and @code{nomips16} attributes.
11979 @xref{Function Attributes}, for more information.
11981 @item -mflip-mips16
11982 @opindex mflip-mips16
11983 Generate MIPS16 code on alternating functions. This option is provided
11984 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
11985 not intended for ordinary use in compiling user code.
11987 @item -minterlink-mips16
11988 @itemx -mno-interlink-mips16
11989 @opindex minterlink-mips16
11990 @opindex mno-interlink-mips16
11991 Require (do not require) that non-MIPS16 code be link-compatible with
11994 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
11995 it must either use a call or an indirect jump. @option{-minterlink-mips16}
11996 therefore disables direct jumps unless GCC knows that the target of the
11997 jump is not MIPS16.
12009 Generate code for the given ABI@.
12011 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
12012 generates 64-bit code when you select a 64-bit architecture, but you
12013 can use @option{-mgp32} to get 32-bit code instead.
12015 For information about the O64 ABI, see
12016 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
12018 GCC supports a variant of the o32 ABI in which floating-point registers
12019 are 64 rather than 32 bits wide. You can select this combination with
12020 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
12021 and @samp{mfhc1} instructions and is therefore only supported for
12022 MIPS32R2 processors.
12024 The register assignments for arguments and return values remain the
12025 same, but each scalar value is passed in a single 64-bit register
12026 rather than a pair of 32-bit registers. For example, scalar
12027 floating-point values are returned in @samp{$f0} only, not a
12028 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
12029 remains the same, but all 64 bits are saved.
12032 @itemx -mno-abicalls
12034 @opindex mno-abicalls
12035 Generate (do not generate) code that is suitable for SVR4-style
12036 dynamic objects. @option{-mabicalls} is the default for SVR4-based
12041 Generate (do not generate) code that is fully position-independent,
12042 and that can therefore be linked into shared libraries. This option
12043 only affects @option{-mabicalls}.
12045 All @option{-mabicalls} code has traditionally been position-independent,
12046 regardless of options like @option{-fPIC} and @option{-fpic}. However,
12047 as an extension, the GNU toolchain allows executables to use absolute
12048 accesses for locally-binding symbols. It can also use shorter GP
12049 initialization sequences and generate direct calls to locally-defined
12050 functions. This mode is selected by @option{-mno-shared}.
12052 @option{-mno-shared} depends on binutils 2.16 or higher and generates
12053 objects that can only be linked by the GNU linker. However, the option
12054 does not affect the ABI of the final executable; it only affects the ABI
12055 of relocatable objects. Using @option{-mno-shared} will generally make
12056 executables both smaller and quicker.
12058 @option{-mshared} is the default.
12064 Lift (do not lift) the usual restrictions on the size of the global
12067 GCC normally uses a single instruction to load values from the GOT@.
12068 While this is relatively efficient, it will only work if the GOT
12069 is smaller than about 64k. Anything larger will cause the linker
12070 to report an error such as:
12072 @cindex relocation truncated to fit (MIPS)
12074 relocation truncated to fit: R_MIPS_GOT16 foobar
12077 If this happens, you should recompile your code with @option{-mxgot}.
12078 It should then work with very large GOTs, although it will also be
12079 less efficient, since it will take three instructions to fetch the
12080 value of a global symbol.
12082 Note that some linkers can create multiple GOTs. If you have such a
12083 linker, you should only need to use @option{-mxgot} when a single object
12084 file accesses more than 64k's worth of GOT entries. Very few do.
12086 These options have no effect unless GCC is generating position
12091 Assume that general-purpose registers are 32 bits wide.
12095 Assume that general-purpose registers are 64 bits wide.
12099 Assume that floating-point registers are 32 bits wide.
12103 Assume that floating-point registers are 64 bits wide.
12106 @opindex mhard-float
12107 Use floating-point coprocessor instructions.
12110 @opindex msoft-float
12111 Do not use floating-point coprocessor instructions. Implement
12112 floating-point calculations using library calls instead.
12114 @item -msingle-float
12115 @opindex msingle-float
12116 Assume that the floating-point coprocessor only supports single-precision
12119 @item -mdouble-float
12120 @opindex mdouble-float
12121 Assume that the floating-point coprocessor supports double-precision
12122 operations. This is the default.
12128 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
12129 implement atomic memory built-in functions. When neither option is
12130 specified, GCC will use the instructions if the target architecture
12133 @option{-mllsc} is useful if the runtime environment can emulate the
12134 instructions and @option{-mno-llsc} can be useful when compiling for
12135 nonstandard ISAs. You can make either option the default by
12136 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
12137 respectively. @option{--with-llsc} is the default for some
12138 configurations; see the installation documentation for details.
12144 Use (do not use) revision 1 of the MIPS DSP ASE@.
12145 @xref{MIPS DSP Built-in Functions}. This option defines the
12146 preprocessor macro @samp{__mips_dsp}. It also defines
12147 @samp{__mips_dsp_rev} to 1.
12153 Use (do not use) revision 2 of the MIPS DSP ASE@.
12154 @xref{MIPS DSP Built-in Functions}. This option defines the
12155 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
12156 It also defines @samp{__mips_dsp_rev} to 2.
12159 @itemx -mno-smartmips
12160 @opindex msmartmips
12161 @opindex mno-smartmips
12162 Use (do not use) the MIPS SmartMIPS ASE.
12164 @item -mpaired-single
12165 @itemx -mno-paired-single
12166 @opindex mpaired-single
12167 @opindex mno-paired-single
12168 Use (do not use) paired-single floating-point instructions.
12169 @xref{MIPS Paired-Single Support}. This option requires
12170 hardware floating-point support to be enabled.
12176 Use (do not use) MIPS Digital Media Extension instructions.
12177 This option can only be used when generating 64-bit code and requires
12178 hardware floating-point support to be enabled.
12183 @opindex mno-mips3d
12184 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
12185 The option @option{-mips3d} implies @option{-mpaired-single}.
12191 Use (do not use) MT Multithreading instructions.
12195 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
12196 an explanation of the default and the way that the pointer size is
12201 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
12203 The default size of @code{int}s, @code{long}s and pointers depends on
12204 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
12205 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
12206 32-bit @code{long}s. Pointers are the same size as @code{long}s,
12207 or the same size as integer registers, whichever is smaller.
12213 Assume (do not assume) that all symbols have 32-bit values, regardless
12214 of the selected ABI@. This option is useful in combination with
12215 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
12216 to generate shorter and faster references to symbolic addresses.
12220 Put definitions of externally-visible data in a small data section
12221 if that data is no bigger than @var{num} bytes. GCC can then access
12222 the data more efficiently; see @option{-mgpopt} for details.
12224 The default @option{-G} option depends on the configuration.
12226 @item -mlocal-sdata
12227 @itemx -mno-local-sdata
12228 @opindex mlocal-sdata
12229 @opindex mno-local-sdata
12230 Extend (do not extend) the @option{-G} behavior to local data too,
12231 such as to static variables in C@. @option{-mlocal-sdata} is the
12232 default for all configurations.
12234 If the linker complains that an application is using too much small data,
12235 you might want to try rebuilding the less performance-critical parts with
12236 @option{-mno-local-sdata}. You might also want to build large
12237 libraries with @option{-mno-local-sdata}, so that the libraries leave
12238 more room for the main program.
12240 @item -mextern-sdata
12241 @itemx -mno-extern-sdata
12242 @opindex mextern-sdata
12243 @opindex mno-extern-sdata
12244 Assume (do not assume) that externally-defined data will be in
12245 a small data section if that data is within the @option{-G} limit.
12246 @option{-mextern-sdata} is the default for all configurations.
12248 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
12249 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
12250 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
12251 is placed in a small data section. If @var{Var} is defined by another
12252 module, you must either compile that module with a high-enough
12253 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
12254 definition. If @var{Var} is common, you must link the application
12255 with a high-enough @option{-G} setting.
12257 The easiest way of satisfying these restrictions is to compile
12258 and link every module with the same @option{-G} option. However,
12259 you may wish to build a library that supports several different
12260 small data limits. You can do this by compiling the library with
12261 the highest supported @option{-G} setting and additionally using
12262 @option{-mno-extern-sdata} to stop the library from making assumptions
12263 about externally-defined data.
12269 Use (do not use) GP-relative accesses for symbols that are known to be
12270 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
12271 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
12274 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
12275 might not hold the value of @code{_gp}. For example, if the code is
12276 part of a library that might be used in a boot monitor, programs that
12277 call boot monitor routines will pass an unknown value in @code{$gp}.
12278 (In such situations, the boot monitor itself would usually be compiled
12279 with @option{-G0}.)
12281 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
12282 @option{-mno-extern-sdata}.
12284 @item -membedded-data
12285 @itemx -mno-embedded-data
12286 @opindex membedded-data
12287 @opindex mno-embedded-data
12288 Allocate variables to the read-only data section first if possible, then
12289 next in the small data section if possible, otherwise in data. This gives
12290 slightly slower code than the default, but reduces the amount of RAM required
12291 when executing, and thus may be preferred for some embedded systems.
12293 @item -muninit-const-in-rodata
12294 @itemx -mno-uninit-const-in-rodata
12295 @opindex muninit-const-in-rodata
12296 @opindex mno-uninit-const-in-rodata
12297 Put uninitialized @code{const} variables in the read-only data section.
12298 This option is only meaningful in conjunction with @option{-membedded-data}.
12300 @item -mcode-readable=@var{setting}
12301 @opindex mcode-readable
12302 Specify whether GCC may generate code that reads from executable sections.
12303 There are three possible settings:
12306 @item -mcode-readable=yes
12307 Instructions may freely access executable sections. This is the
12310 @item -mcode-readable=pcrel
12311 MIPS16 PC-relative load instructions can access executable sections,
12312 but other instructions must not do so. This option is useful on 4KSc
12313 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
12314 It is also useful on processors that can be configured to have a dual
12315 instruction/data SRAM interface and that, like the M4K, automatically
12316 redirect PC-relative loads to the instruction RAM.
12318 @item -mcode-readable=no
12319 Instructions must not access executable sections. This option can be
12320 useful on targets that are configured to have a dual instruction/data
12321 SRAM interface but that (unlike the M4K) do not automatically redirect
12322 PC-relative loads to the instruction RAM.
12325 @item -msplit-addresses
12326 @itemx -mno-split-addresses
12327 @opindex msplit-addresses
12328 @opindex mno-split-addresses
12329 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
12330 relocation operators. This option has been superseded by
12331 @option{-mexplicit-relocs} but is retained for backwards compatibility.
12333 @item -mexplicit-relocs
12334 @itemx -mno-explicit-relocs
12335 @opindex mexplicit-relocs
12336 @opindex mno-explicit-relocs
12337 Use (do not use) assembler relocation operators when dealing with symbolic
12338 addresses. The alternative, selected by @option{-mno-explicit-relocs},
12339 is to use assembler macros instead.
12341 @option{-mexplicit-relocs} is the default if GCC was configured
12342 to use an assembler that supports relocation operators.
12344 @item -mcheck-zero-division
12345 @itemx -mno-check-zero-division
12346 @opindex mcheck-zero-division
12347 @opindex mno-check-zero-division
12348 Trap (do not trap) on integer division by zero.
12350 The default is @option{-mcheck-zero-division}.
12352 @item -mdivide-traps
12353 @itemx -mdivide-breaks
12354 @opindex mdivide-traps
12355 @opindex mdivide-breaks
12356 MIPS systems check for division by zero by generating either a
12357 conditional trap or a break instruction. Using traps results in
12358 smaller code, but is only supported on MIPS II and later. Also, some
12359 versions of the Linux kernel have a bug that prevents trap from
12360 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
12361 allow conditional traps on architectures that support them and
12362 @option{-mdivide-breaks} to force the use of breaks.
12364 The default is usually @option{-mdivide-traps}, but this can be
12365 overridden at configure time using @option{--with-divide=breaks}.
12366 Divide-by-zero checks can be completely disabled using
12367 @option{-mno-check-zero-division}.
12372 @opindex mno-memcpy
12373 Force (do not force) the use of @code{memcpy()} for non-trivial block
12374 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
12375 most constant-sized copies.
12378 @itemx -mno-long-calls
12379 @opindex mlong-calls
12380 @opindex mno-long-calls
12381 Disable (do not disable) use of the @code{jal} instruction. Calling
12382 functions using @code{jal} is more efficient but requires the caller
12383 and callee to be in the same 256 megabyte segment.
12385 This option has no effect on abicalls code. The default is
12386 @option{-mno-long-calls}.
12392 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
12393 instructions, as provided by the R4650 ISA@.
12396 @itemx -mno-fused-madd
12397 @opindex mfused-madd
12398 @opindex mno-fused-madd
12399 Enable (disable) use of the floating point multiply-accumulate
12400 instructions, when they are available. The default is
12401 @option{-mfused-madd}.
12403 When multiply-accumulate instructions are used, the intermediate
12404 product is calculated to infinite precision and is not subject to
12405 the FCSR Flush to Zero bit. This may be undesirable in some
12410 Tell the MIPS assembler to not run its preprocessor over user
12411 assembler files (with a @samp{.s} suffix) when assembling them.
12414 @itemx -mno-fix-r4000
12415 @opindex mfix-r4000
12416 @opindex mno-fix-r4000
12417 Work around certain R4000 CPU errata:
12420 A double-word or a variable shift may give an incorrect result if executed
12421 immediately after starting an integer division.
12423 A double-word or a variable shift may give an incorrect result if executed
12424 while an integer multiplication is in progress.
12426 An integer division may give an incorrect result if started in a delay slot
12427 of a taken branch or a jump.
12431 @itemx -mno-fix-r4400
12432 @opindex mfix-r4400
12433 @opindex mno-fix-r4400
12434 Work around certain R4400 CPU errata:
12437 A double-word or a variable shift may give an incorrect result if executed
12438 immediately after starting an integer division.
12442 @itemx -mno-fix-vr4120
12443 @opindex mfix-vr4120
12444 Work around certain VR4120 errata:
12447 @code{dmultu} does not always produce the correct result.
12449 @code{div} and @code{ddiv} do not always produce the correct result if one
12450 of the operands is negative.
12452 The workarounds for the division errata rely on special functions in
12453 @file{libgcc.a}. At present, these functions are only provided by
12454 the @code{mips64vr*-elf} configurations.
12456 Other VR4120 errata require a nop to be inserted between certain pairs of
12457 instructions. These errata are handled by the assembler, not by GCC itself.
12460 @opindex mfix-vr4130
12461 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
12462 workarounds are implemented by the assembler rather than by GCC,
12463 although GCC will avoid using @code{mflo} and @code{mfhi} if the
12464 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
12465 instructions are available instead.
12468 @itemx -mno-fix-sb1
12470 Work around certain SB-1 CPU core errata.
12471 (This flag currently works around the SB-1 revision 2
12472 ``F1'' and ``F2'' floating point errata.)
12474 @item -mflush-func=@var{func}
12475 @itemx -mno-flush-func
12476 @opindex mflush-func
12477 Specifies the function to call to flush the I and D caches, or to not
12478 call any such function. If called, the function must take the same
12479 arguments as the common @code{_flush_func()}, that is, the address of the
12480 memory range for which the cache is being flushed, the size of the
12481 memory range, and the number 3 (to flush both caches). The default
12482 depends on the target GCC was configured for, but commonly is either
12483 @samp{_flush_func} or @samp{__cpu_flush}.
12485 @item mbranch-cost=@var{num}
12486 @opindex mbranch-cost
12487 Set the cost of branches to roughly @var{num} ``simple'' instructions.
12488 This cost is only a heuristic and is not guaranteed to produce
12489 consistent results across releases. A zero cost redundantly selects
12490 the default, which is based on the @option{-mtune} setting.
12492 @item -mbranch-likely
12493 @itemx -mno-branch-likely
12494 @opindex mbranch-likely
12495 @opindex mno-branch-likely
12496 Enable or disable use of Branch Likely instructions, regardless of the
12497 default for the selected architecture. By default, Branch Likely
12498 instructions may be generated if they are supported by the selected
12499 architecture. An exception is for the MIPS32 and MIPS64 architectures
12500 and processors which implement those architectures; for those, Branch
12501 Likely instructions will not be generated by default because the MIPS32
12502 and MIPS64 architectures specifically deprecate their use.
12504 @item -mfp-exceptions
12505 @itemx -mno-fp-exceptions
12506 @opindex mfp-exceptions
12507 Specifies whether FP exceptions are enabled. This affects how we schedule
12508 FP instructions for some processors. The default is that FP exceptions are
12511 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
12512 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
12515 @item -mvr4130-align
12516 @itemx -mno-vr4130-align
12517 @opindex mvr4130-align
12518 The VR4130 pipeline is two-way superscalar, but can only issue two
12519 instructions together if the first one is 8-byte aligned. When this
12520 option is enabled, GCC will align pairs of instructions that it
12521 thinks should execute in parallel.
12523 This option only has an effect when optimizing for the VR4130.
12524 It normally makes code faster, but at the expense of making it bigger.
12525 It is enabled by default at optimization level @option{-O3}.
12529 @subsection MMIX Options
12530 @cindex MMIX Options
12532 These options are defined for the MMIX:
12536 @itemx -mno-libfuncs
12538 @opindex mno-libfuncs
12539 Specify that intrinsic library functions are being compiled, passing all
12540 values in registers, no matter the size.
12543 @itemx -mno-epsilon
12545 @opindex mno-epsilon
12546 Generate floating-point comparison instructions that compare with respect
12547 to the @code{rE} epsilon register.
12549 @item -mabi=mmixware
12551 @opindex mabi-mmixware
12553 Generate code that passes function parameters and return values that (in
12554 the called function) are seen as registers @code{$0} and up, as opposed to
12555 the GNU ABI which uses global registers @code{$231} and up.
12557 @item -mzero-extend
12558 @itemx -mno-zero-extend
12559 @opindex mzero-extend
12560 @opindex mno-zero-extend
12561 When reading data from memory in sizes shorter than 64 bits, use (do not
12562 use) zero-extending load instructions by default, rather than
12563 sign-extending ones.
12566 @itemx -mno-knuthdiv
12568 @opindex mno-knuthdiv
12569 Make the result of a division yielding a remainder have the same sign as
12570 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
12571 remainder follows the sign of the dividend. Both methods are
12572 arithmetically valid, the latter being almost exclusively used.
12574 @item -mtoplevel-symbols
12575 @itemx -mno-toplevel-symbols
12576 @opindex mtoplevel-symbols
12577 @opindex mno-toplevel-symbols
12578 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
12579 code can be used with the @code{PREFIX} assembly directive.
12583 Generate an executable in the ELF format, rather than the default
12584 @samp{mmo} format used by the @command{mmix} simulator.
12586 @item -mbranch-predict
12587 @itemx -mno-branch-predict
12588 @opindex mbranch-predict
12589 @opindex mno-branch-predict
12590 Use (do not use) the probable-branch instructions, when static branch
12591 prediction indicates a probable branch.
12593 @item -mbase-addresses
12594 @itemx -mno-base-addresses
12595 @opindex mbase-addresses
12596 @opindex mno-base-addresses
12597 Generate (do not generate) code that uses @emph{base addresses}. Using a
12598 base address automatically generates a request (handled by the assembler
12599 and the linker) for a constant to be set up in a global register. The
12600 register is used for one or more base address requests within the range 0
12601 to 255 from the value held in the register. The generally leads to short
12602 and fast code, but the number of different data items that can be
12603 addressed is limited. This means that a program that uses lots of static
12604 data may require @option{-mno-base-addresses}.
12606 @item -msingle-exit
12607 @itemx -mno-single-exit
12608 @opindex msingle-exit
12609 @opindex mno-single-exit
12610 Force (do not force) generated code to have a single exit point in each
12614 @node MN10300 Options
12615 @subsection MN10300 Options
12616 @cindex MN10300 options
12618 These @option{-m} options are defined for Matsushita MN10300 architectures:
12623 Generate code to avoid bugs in the multiply instructions for the MN10300
12624 processors. This is the default.
12626 @item -mno-mult-bug
12627 @opindex mno-mult-bug
12628 Do not generate code to avoid bugs in the multiply instructions for the
12629 MN10300 processors.
12633 Generate code which uses features specific to the AM33 processor.
12637 Do not generate code which uses features specific to the AM33 processor. This
12640 @item -mreturn-pointer-on-d0
12641 @opindex mreturn-pointer-on-d0
12642 When generating a function which returns a pointer, return the pointer
12643 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
12644 only in a0, and attempts to call such functions without a prototype
12645 would result in errors. Note that this option is on by default; use
12646 @option{-mno-return-pointer-on-d0} to disable it.
12650 Do not link in the C run-time initialization object file.
12654 Indicate to the linker that it should perform a relaxation optimization pass
12655 to shorten branches, calls and absolute memory addresses. This option only
12656 has an effect when used on the command line for the final link step.
12658 This option makes symbolic debugging impossible.
12662 @subsection MT Options
12665 These @option{-m} options are defined for Morpho MT architectures:
12669 @item -march=@var{cpu-type}
12671 Generate code that will run on @var{cpu-type}, which is the name of a system
12672 representing a certain processor type. Possible values for
12673 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
12674 @samp{ms1-16-003} and @samp{ms2}.
12676 When this option is not used, the default is @option{-march=ms1-16-002}.
12680 Use byte loads and stores when generating code.
12684 Do not use byte loads and stores when generating code.
12688 Use simulator runtime
12692 Do not link in the C run-time initialization object file
12693 @file{crti.o}. Other run-time initialization and termination files
12694 such as @file{startup.o} and @file{exit.o} are still included on the
12695 linker command line.
12699 @node PDP-11 Options
12700 @subsection PDP-11 Options
12701 @cindex PDP-11 Options
12703 These options are defined for the PDP-11:
12708 Use hardware FPP floating point. This is the default. (FIS floating
12709 point on the PDP-11/40 is not supported.)
12712 @opindex msoft-float
12713 Do not use hardware floating point.
12717 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
12721 Return floating-point results in memory. This is the default.
12725 Generate code for a PDP-11/40.
12729 Generate code for a PDP-11/45. This is the default.
12733 Generate code for a PDP-11/10.
12735 @item -mbcopy-builtin
12736 @opindex bcopy-builtin
12737 Use inline @code{movmemhi} patterns for copying memory. This is the
12742 Do not use inline @code{movmemhi} patterns for copying memory.
12748 Use 16-bit @code{int}. This is the default.
12754 Use 32-bit @code{int}.
12757 @itemx -mno-float32
12759 @opindex mno-float32
12760 Use 64-bit @code{float}. This is the default.
12763 @itemx -mno-float64
12765 @opindex mno-float64
12766 Use 32-bit @code{float}.
12770 Use @code{abshi2} pattern. This is the default.
12774 Do not use @code{abshi2} pattern.
12776 @item -mbranch-expensive
12777 @opindex mbranch-expensive
12778 Pretend that branches are expensive. This is for experimenting with
12779 code generation only.
12781 @item -mbranch-cheap
12782 @opindex mbranch-cheap
12783 Do not pretend that branches are expensive. This is the default.
12787 Generate code for a system with split I&D@.
12791 Generate code for a system without split I&D@. This is the default.
12795 Use Unix assembler syntax. This is the default when configured for
12796 @samp{pdp11-*-bsd}.
12800 Use DEC assembler syntax. This is the default when configured for any
12801 PDP-11 target other than @samp{pdp11-*-bsd}.
12804 @node PowerPC Options
12805 @subsection PowerPC Options
12806 @cindex PowerPC options
12808 These are listed under @xref{RS/6000 and PowerPC Options}.
12810 @node RS/6000 and PowerPC Options
12811 @subsection IBM RS/6000 and PowerPC Options
12812 @cindex RS/6000 and PowerPC Options
12813 @cindex IBM RS/6000 and PowerPC Options
12815 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
12822 @itemx -mno-powerpc
12823 @itemx -mpowerpc-gpopt
12824 @itemx -mno-powerpc-gpopt
12825 @itemx -mpowerpc-gfxopt
12826 @itemx -mno-powerpc-gfxopt
12828 @itemx -mno-powerpc64
12832 @itemx -mno-popcntb
12840 @itemx -mno-hard-dfp
12844 @opindex mno-power2
12846 @opindex mno-powerpc
12847 @opindex mpowerpc-gpopt
12848 @opindex mno-powerpc-gpopt
12849 @opindex mpowerpc-gfxopt
12850 @opindex mno-powerpc-gfxopt
12851 @opindex mpowerpc64
12852 @opindex mno-powerpc64
12856 @opindex mno-popcntb
12862 @opindex mno-mfpgpr
12864 @opindex mno-hard-dfp
12865 GCC supports two related instruction set architectures for the
12866 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
12867 instructions supported by the @samp{rios} chip set used in the original
12868 RS/6000 systems and the @dfn{PowerPC} instruction set is the
12869 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
12870 the IBM 4xx, 6xx, and follow-on microprocessors.
12872 Neither architecture is a subset of the other. However there is a
12873 large common subset of instructions supported by both. An MQ
12874 register is included in processors supporting the POWER architecture.
12876 You use these options to specify which instructions are available on the
12877 processor you are using. The default value of these options is
12878 determined when configuring GCC@. Specifying the
12879 @option{-mcpu=@var{cpu_type}} overrides the specification of these
12880 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
12881 rather than the options listed above.
12883 The @option{-mpower} option allows GCC to generate instructions that
12884 are found only in the POWER architecture and to use the MQ register.
12885 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
12886 to generate instructions that are present in the POWER2 architecture but
12887 not the original POWER architecture.
12889 The @option{-mpowerpc} option allows GCC to generate instructions that
12890 are found only in the 32-bit subset of the PowerPC architecture.
12891 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
12892 GCC to use the optional PowerPC architecture instructions in the
12893 General Purpose group, including floating-point square root. Specifying
12894 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
12895 use the optional PowerPC architecture instructions in the Graphics
12896 group, including floating-point select.
12898 The @option{-mmfcrf} option allows GCC to generate the move from
12899 condition register field instruction implemented on the POWER4
12900 processor and other processors that support the PowerPC V2.01
12902 The @option{-mpopcntb} option allows GCC to generate the popcount and
12903 double precision FP reciprocal estimate instruction implemented on the
12904 POWER5 processor and other processors that support the PowerPC V2.02
12906 The @option{-mfprnd} option allows GCC to generate the FP round to
12907 integer instructions implemented on the POWER5+ processor and other
12908 processors that support the PowerPC V2.03 architecture.
12909 The @option{-mcmpb} option allows GCC to generate the compare bytes
12910 instruction implemented on the POWER6 processor and other processors
12911 that support the PowerPC V2.05 architecture.
12912 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
12913 general purpose register instructions implemented on the POWER6X
12914 processor and other processors that support the extended PowerPC V2.05
12916 The @option{-mhard-dfp} option allows GCC to generate the decimal floating
12917 point instructions implemented on some POWER processors.
12919 The @option{-mpowerpc64} option allows GCC to generate the additional
12920 64-bit instructions that are found in the full PowerPC64 architecture
12921 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
12922 @option{-mno-powerpc64}.
12924 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
12925 will use only the instructions in the common subset of both
12926 architectures plus some special AIX common-mode calls, and will not use
12927 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
12928 permits GCC to use any instruction from either architecture and to
12929 allow use of the MQ register; specify this for the Motorola MPC601.
12931 @item -mnew-mnemonics
12932 @itemx -mold-mnemonics
12933 @opindex mnew-mnemonics
12934 @opindex mold-mnemonics
12935 Select which mnemonics to use in the generated assembler code. With
12936 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
12937 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
12938 assembler mnemonics defined for the POWER architecture. Instructions
12939 defined in only one architecture have only one mnemonic; GCC uses that
12940 mnemonic irrespective of which of these options is specified.
12942 GCC defaults to the mnemonics appropriate for the architecture in
12943 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
12944 value of these option. Unless you are building a cross-compiler, you
12945 should normally not specify either @option{-mnew-mnemonics} or
12946 @option{-mold-mnemonics}, but should instead accept the default.
12948 @item -mcpu=@var{cpu_type}
12950 Set architecture type, register usage, choice of mnemonics, and
12951 instruction scheduling parameters for machine type @var{cpu_type}.
12952 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
12953 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
12954 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
12955 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
12956 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
12957 @samp{860}, @samp{970}, @samp{8540}, @samp{e300c2}, @samp{e300c3},
12958 @samp{ec603e}, @samp{G3}, @samp{G4}, @samp{G5}, @samp{power},
12959 @samp{power2}, @samp{power3}, @samp{power4}, @samp{power5},
12960 @samp{power5+}, @samp{power6}, @samp{power6x}, @samp{common},
12961 @samp{powerpc}, @samp{powerpc64}, @samp{rios}, @samp{rios1},
12962 @samp{rios2}, @samp{rsc}, and @samp{rs64}.
12964 @option{-mcpu=common} selects a completely generic processor. Code
12965 generated under this option will run on any POWER or PowerPC processor.
12966 GCC will use only the instructions in the common subset of both
12967 architectures, and will not use the MQ register. GCC assumes a generic
12968 processor model for scheduling purposes.
12970 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
12971 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
12972 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
12973 types, with an appropriate, generic processor model assumed for
12974 scheduling purposes.
12976 The other options specify a specific processor. Code generated under
12977 those options will run best on that processor, and may not run at all on
12980 The @option{-mcpu} options automatically enable or disable the
12983 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
12984 -mnew-mnemonics -mpopcntb -mpower -mpower2 -mpowerpc64 @gol
12985 -mpowerpc-gpopt -mpowerpc-gfxopt -mstring -mmulhw -mdlmzb -mmfpgpr}
12987 The particular options set for any particular CPU will vary between
12988 compiler versions, depending on what setting seems to produce optimal
12989 code for that CPU; it doesn't necessarily reflect the actual hardware's
12990 capabilities. If you wish to set an individual option to a particular
12991 value, you may specify it after the @option{-mcpu} option, like
12992 @samp{-mcpu=970 -mno-altivec}.
12994 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
12995 not enabled or disabled by the @option{-mcpu} option at present because
12996 AIX does not have full support for these options. You may still
12997 enable or disable them individually if you're sure it'll work in your
13000 @item -mtune=@var{cpu_type}
13002 Set the instruction scheduling parameters for machine type
13003 @var{cpu_type}, but do not set the architecture type, register usage, or
13004 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
13005 values for @var{cpu_type} are used for @option{-mtune} as for
13006 @option{-mcpu}. If both are specified, the code generated will use the
13007 architecture, registers, and mnemonics set by @option{-mcpu}, but the
13008 scheduling parameters set by @option{-mtune}.
13014 Generate code to compute division as reciprocal estimate and iterative
13015 refinement, creating opportunities for increased throughput. This
13016 feature requires: optional PowerPC Graphics instruction set for single
13017 precision and FRE instruction for double precision, assuming divides
13018 cannot generate user-visible traps, and the domain values not include
13019 Infinities, denormals or zero denominator.
13022 @itemx -mno-altivec
13024 @opindex mno-altivec
13025 Generate code that uses (does not use) AltiVec instructions, and also
13026 enable the use of built-in functions that allow more direct access to
13027 the AltiVec instruction set. You may also need to set
13028 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
13034 @opindex mno-vrsave
13035 Generate VRSAVE instructions when generating AltiVec code.
13038 @opindex msecure-plt
13039 Generate code that allows ld and ld.so to build executables and shared
13040 libraries with non-exec .plt and .got sections. This is a PowerPC
13041 32-bit SYSV ABI option.
13045 Generate code that uses a BSS .plt section that ld.so fills in, and
13046 requires .plt and .got sections that are both writable and executable.
13047 This is a PowerPC 32-bit SYSV ABI option.
13053 This switch enables or disables the generation of ISEL instructions.
13055 @item -misel=@var{yes/no}
13056 This switch has been deprecated. Use @option{-misel} and
13057 @option{-mno-isel} instead.
13063 This switch enables or disables the generation of SPE simd
13069 @opindex mno-paired
13070 This switch enables or disables the generation of PAIRED simd
13073 @item -mspe=@var{yes/no}
13074 This option has been deprecated. Use @option{-mspe} and
13075 @option{-mno-spe} instead.
13077 @item -mfloat-gprs=@var{yes/single/double/no}
13078 @itemx -mfloat-gprs
13079 @opindex mfloat-gprs
13080 This switch enables or disables the generation of floating point
13081 operations on the general purpose registers for architectures that
13084 The argument @var{yes} or @var{single} enables the use of
13085 single-precision floating point operations.
13087 The argument @var{double} enables the use of single and
13088 double-precision floating point operations.
13090 The argument @var{no} disables floating point operations on the
13091 general purpose registers.
13093 This option is currently only available on the MPC854x.
13099 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
13100 targets (including GNU/Linux). The 32-bit environment sets int, long
13101 and pointer to 32 bits and generates code that runs on any PowerPC
13102 variant. The 64-bit environment sets int to 32 bits and long and
13103 pointer to 64 bits, and generates code for PowerPC64, as for
13104 @option{-mpowerpc64}.
13107 @itemx -mno-fp-in-toc
13108 @itemx -mno-sum-in-toc
13109 @itemx -mminimal-toc
13111 @opindex mno-fp-in-toc
13112 @opindex mno-sum-in-toc
13113 @opindex mminimal-toc
13114 Modify generation of the TOC (Table Of Contents), which is created for
13115 every executable file. The @option{-mfull-toc} option is selected by
13116 default. In that case, GCC will allocate at least one TOC entry for
13117 each unique non-automatic variable reference in your program. GCC
13118 will also place floating-point constants in the TOC@. However, only
13119 16,384 entries are available in the TOC@.
13121 If you receive a linker error message that saying you have overflowed
13122 the available TOC space, you can reduce the amount of TOC space used
13123 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
13124 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
13125 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
13126 generate code to calculate the sum of an address and a constant at
13127 run-time instead of putting that sum into the TOC@. You may specify one
13128 or both of these options. Each causes GCC to produce very slightly
13129 slower and larger code at the expense of conserving TOC space.
13131 If you still run out of space in the TOC even when you specify both of
13132 these options, specify @option{-mminimal-toc} instead. This option causes
13133 GCC to make only one TOC entry for every file. When you specify this
13134 option, GCC will produce code that is slower and larger but which
13135 uses extremely little TOC space. You may wish to use this option
13136 only on files that contain less frequently executed code.
13142 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
13143 @code{long} type, and the infrastructure needed to support them.
13144 Specifying @option{-maix64} implies @option{-mpowerpc64} and
13145 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
13146 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
13149 @itemx -mno-xl-compat
13150 @opindex mxl-compat
13151 @opindex mno-xl-compat
13152 Produce code that conforms more closely to IBM XL compiler semantics
13153 when using AIX-compatible ABI@. Pass floating-point arguments to
13154 prototyped functions beyond the register save area (RSA) on the stack
13155 in addition to argument FPRs. Do not assume that most significant
13156 double in 128-bit long double value is properly rounded when comparing
13157 values and converting to double. Use XL symbol names for long double
13160 The AIX calling convention was extended but not initially documented to
13161 handle an obscure K&R C case of calling a function that takes the
13162 address of its arguments with fewer arguments than declared. IBM XL
13163 compilers access floating point arguments which do not fit in the
13164 RSA from the stack when a subroutine is compiled without
13165 optimization. Because always storing floating-point arguments on the
13166 stack is inefficient and rarely needed, this option is not enabled by
13167 default and only is necessary when calling subroutines compiled by IBM
13168 XL compilers without optimization.
13172 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
13173 application written to use message passing with special startup code to
13174 enable the application to run. The system must have PE installed in the
13175 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
13176 must be overridden with the @option{-specs=} option to specify the
13177 appropriate directory location. The Parallel Environment does not
13178 support threads, so the @option{-mpe} option and the @option{-pthread}
13179 option are incompatible.
13181 @item -malign-natural
13182 @itemx -malign-power
13183 @opindex malign-natural
13184 @opindex malign-power
13185 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
13186 @option{-malign-natural} overrides the ABI-defined alignment of larger
13187 types, such as floating-point doubles, on their natural size-based boundary.
13188 The option @option{-malign-power} instructs GCC to follow the ABI-specified
13189 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
13191 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
13195 @itemx -mhard-float
13196 @opindex msoft-float
13197 @opindex mhard-float
13198 Generate code that does not use (uses) the floating-point register set.
13199 Software floating point emulation is provided if you use the
13200 @option{-msoft-float} option, and pass the option to GCC when linking.
13203 @itemx -mno-multiple
13205 @opindex mno-multiple
13206 Generate code that uses (does not use) the load multiple word
13207 instructions and the store multiple word instructions. These
13208 instructions are generated by default on POWER systems, and not
13209 generated on PowerPC systems. Do not use @option{-mmultiple} on little
13210 endian PowerPC systems, since those instructions do not work when the
13211 processor is in little endian mode. The exceptions are PPC740 and
13212 PPC750 which permit the instructions usage in little endian mode.
13217 @opindex mno-string
13218 Generate code that uses (does not use) the load string instructions
13219 and the store string word instructions to save multiple registers and
13220 do small block moves. These instructions are generated by default on
13221 POWER systems, and not generated on PowerPC systems. Do not use
13222 @option{-mstring} on little endian PowerPC systems, since those
13223 instructions do not work when the processor is in little endian mode.
13224 The exceptions are PPC740 and PPC750 which permit the instructions
13225 usage in little endian mode.
13230 @opindex mno-update
13231 Generate code that uses (does not use) the load or store instructions
13232 that update the base register to the address of the calculated memory
13233 location. These instructions are generated by default. If you use
13234 @option{-mno-update}, there is a small window between the time that the
13235 stack pointer is updated and the address of the previous frame is
13236 stored, which means code that walks the stack frame across interrupts or
13237 signals may get corrupted data.
13240 @itemx -mno-fused-madd
13241 @opindex mfused-madd
13242 @opindex mno-fused-madd
13243 Generate code that uses (does not use) the floating point multiply and
13244 accumulate instructions. These instructions are generated by default if
13245 hardware floating is used.
13251 Generate code that uses (does not use) the half-word multiply and
13252 multiply-accumulate instructions on the IBM 405 and 440 processors.
13253 These instructions are generated by default when targetting those
13260 Generate code that uses (does not use) the string-search @samp{dlmzb}
13261 instruction on the IBM 405 and 440 processors. This instruction is
13262 generated by default when targetting those processors.
13264 @item -mno-bit-align
13266 @opindex mno-bit-align
13267 @opindex mbit-align
13268 On System V.4 and embedded PowerPC systems do not (do) force structures
13269 and unions that contain bit-fields to be aligned to the base type of the
13272 For example, by default a structure containing nothing but 8
13273 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
13274 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
13275 the structure would be aligned to a 1 byte boundary and be one byte in
13278 @item -mno-strict-align
13279 @itemx -mstrict-align
13280 @opindex mno-strict-align
13281 @opindex mstrict-align
13282 On System V.4 and embedded PowerPC systems do not (do) assume that
13283 unaligned memory references will be handled by the system.
13285 @item -mrelocatable
13286 @itemx -mno-relocatable
13287 @opindex mrelocatable
13288 @opindex mno-relocatable
13289 On embedded PowerPC systems generate code that allows (does not allow)
13290 the program to be relocated to a different address at runtime. If you
13291 use @option{-mrelocatable} on any module, all objects linked together must
13292 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
13294 @item -mrelocatable-lib
13295 @itemx -mno-relocatable-lib
13296 @opindex mrelocatable-lib
13297 @opindex mno-relocatable-lib
13298 On embedded PowerPC systems generate code that allows (does not allow)
13299 the program to be relocated to a different address at runtime. Modules
13300 compiled with @option{-mrelocatable-lib} can be linked with either modules
13301 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
13302 with modules compiled with the @option{-mrelocatable} options.
13308 On System V.4 and embedded PowerPC systems do not (do) assume that
13309 register 2 contains a pointer to a global area pointing to the addresses
13310 used in the program.
13313 @itemx -mlittle-endian
13315 @opindex mlittle-endian
13316 On System V.4 and embedded PowerPC systems compile code for the
13317 processor in little endian mode. The @option{-mlittle-endian} option is
13318 the same as @option{-mlittle}.
13321 @itemx -mbig-endian
13323 @opindex mbig-endian
13324 On System V.4 and embedded PowerPC systems compile code for the
13325 processor in big endian mode. The @option{-mbig-endian} option is
13326 the same as @option{-mbig}.
13328 @item -mdynamic-no-pic
13329 @opindex mdynamic-no-pic
13330 On Darwin and Mac OS X systems, compile code so that it is not
13331 relocatable, but that its external references are relocatable. The
13332 resulting code is suitable for applications, but not shared
13335 @item -mprioritize-restricted-insns=@var{priority}
13336 @opindex mprioritize-restricted-insns
13337 This option controls the priority that is assigned to
13338 dispatch-slot restricted instructions during the second scheduling
13339 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
13340 @var{no/highest/second-highest} priority to dispatch slot restricted
13343 @item -msched-costly-dep=@var{dependence_type}
13344 @opindex msched-costly-dep
13345 This option controls which dependences are considered costly
13346 by the target during instruction scheduling. The argument
13347 @var{dependence_type} takes one of the following values:
13348 @var{no}: no dependence is costly,
13349 @var{all}: all dependences are costly,
13350 @var{true_store_to_load}: a true dependence from store to load is costly,
13351 @var{store_to_load}: any dependence from store to load is costly,
13352 @var{number}: any dependence which latency >= @var{number} is costly.
13354 @item -minsert-sched-nops=@var{scheme}
13355 @opindex minsert-sched-nops
13356 This option controls which nop insertion scheme will be used during
13357 the second scheduling pass. The argument @var{scheme} takes one of the
13359 @var{no}: Don't insert nops.
13360 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
13361 according to the scheduler's grouping.
13362 @var{regroup_exact}: Insert nops to force costly dependent insns into
13363 separate groups. Insert exactly as many nops as needed to force an insn
13364 to a new group, according to the estimated processor grouping.
13365 @var{number}: Insert nops to force costly dependent insns into
13366 separate groups. Insert @var{number} nops to force an insn to a new group.
13369 @opindex mcall-sysv
13370 On System V.4 and embedded PowerPC systems compile code using calling
13371 conventions that adheres to the March 1995 draft of the System V
13372 Application Binary Interface, PowerPC processor supplement. This is the
13373 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
13375 @item -mcall-sysv-eabi
13376 @opindex mcall-sysv-eabi
13377 Specify both @option{-mcall-sysv} and @option{-meabi} options.
13379 @item -mcall-sysv-noeabi
13380 @opindex mcall-sysv-noeabi
13381 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
13383 @item -mcall-solaris
13384 @opindex mcall-solaris
13385 On System V.4 and embedded PowerPC systems compile code for the Solaris
13389 @opindex mcall-linux
13390 On System V.4 and embedded PowerPC systems compile code for the
13391 Linux-based GNU system.
13395 On System V.4 and embedded PowerPC systems compile code for the
13396 Hurd-based GNU system.
13398 @item -mcall-netbsd
13399 @opindex mcall-netbsd
13400 On System V.4 and embedded PowerPC systems compile code for the
13401 NetBSD operating system.
13403 @item -maix-struct-return
13404 @opindex maix-struct-return
13405 Return all structures in memory (as specified by the AIX ABI)@.
13407 @item -msvr4-struct-return
13408 @opindex msvr4-struct-return
13409 Return structures smaller than 8 bytes in registers (as specified by the
13412 @item -mabi=@var{abi-type}
13414 Extend the current ABI with a particular extension, or remove such extension.
13415 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
13416 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
13420 Extend the current ABI with SPE ABI extensions. This does not change
13421 the default ABI, instead it adds the SPE ABI extensions to the current
13425 @opindex mabi=no-spe
13426 Disable Booke SPE ABI extensions for the current ABI@.
13428 @item -mabi=ibmlongdouble
13429 @opindex mabi=ibmlongdouble
13430 Change the current ABI to use IBM extended precision long double.
13431 This is a PowerPC 32-bit SYSV ABI option.
13433 @item -mabi=ieeelongdouble
13434 @opindex mabi=ieeelongdouble
13435 Change the current ABI to use IEEE extended precision long double.
13436 This is a PowerPC 32-bit Linux ABI option.
13439 @itemx -mno-prototype
13440 @opindex mprototype
13441 @opindex mno-prototype
13442 On System V.4 and embedded PowerPC systems assume that all calls to
13443 variable argument functions are properly prototyped. Otherwise, the
13444 compiler must insert an instruction before every non prototyped call to
13445 set or clear bit 6 of the condition code register (@var{CR}) to
13446 indicate whether floating point values were passed in the floating point
13447 registers in case the function takes a variable arguments. With
13448 @option{-mprototype}, only calls to prototyped variable argument functions
13449 will set or clear the bit.
13453 On embedded PowerPC systems, assume that the startup module is called
13454 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
13455 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
13460 On embedded PowerPC systems, assume that the startup module is called
13461 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
13466 On embedded PowerPC systems, assume that the startup module is called
13467 @file{crt0.o} and the standard C libraries are @file{libads.a} and
13470 @item -myellowknife
13471 @opindex myellowknife
13472 On embedded PowerPC systems, assume that the startup module is called
13473 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
13478 On System V.4 and embedded PowerPC systems, specify that you are
13479 compiling for a VxWorks system.
13483 Specify that you are compiling for the WindISS simulation environment.
13487 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
13488 header to indicate that @samp{eabi} extended relocations are used.
13494 On System V.4 and embedded PowerPC systems do (do not) adhere to the
13495 Embedded Applications Binary Interface (eabi) which is a set of
13496 modifications to the System V.4 specifications. Selecting @option{-meabi}
13497 means that the stack is aligned to an 8 byte boundary, a function
13498 @code{__eabi} is called to from @code{main} to set up the eabi
13499 environment, and the @option{-msdata} option can use both @code{r2} and
13500 @code{r13} to point to two separate small data areas. Selecting
13501 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
13502 do not call an initialization function from @code{main}, and the
13503 @option{-msdata} option will only use @code{r13} to point to a single
13504 small data area. The @option{-meabi} option is on by default if you
13505 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
13508 @opindex msdata=eabi
13509 On System V.4 and embedded PowerPC systems, put small initialized
13510 @code{const} global and static data in the @samp{.sdata2} section, which
13511 is pointed to by register @code{r2}. Put small initialized
13512 non-@code{const} global and static data in the @samp{.sdata} section,
13513 which is pointed to by register @code{r13}. Put small uninitialized
13514 global and static data in the @samp{.sbss} section, which is adjacent to
13515 the @samp{.sdata} section. The @option{-msdata=eabi} option is
13516 incompatible with the @option{-mrelocatable} option. The
13517 @option{-msdata=eabi} option also sets the @option{-memb} option.
13520 @opindex msdata=sysv
13521 On System V.4 and embedded PowerPC systems, put small global and static
13522 data in the @samp{.sdata} section, which is pointed to by register
13523 @code{r13}. Put small uninitialized global and static data in the
13524 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
13525 The @option{-msdata=sysv} option is incompatible with the
13526 @option{-mrelocatable} option.
13528 @item -msdata=default
13530 @opindex msdata=default
13532 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
13533 compile code the same as @option{-msdata=eabi}, otherwise compile code the
13534 same as @option{-msdata=sysv}.
13537 @opindex msdata-data
13538 On System V.4 and embedded PowerPC systems, put small global
13539 data in the @samp{.sdata} section. Put small uninitialized global
13540 data in the @samp{.sbss} section. Do not use register @code{r13}
13541 to address small data however. This is the default behavior unless
13542 other @option{-msdata} options are used.
13546 @opindex msdata=none
13548 On embedded PowerPC systems, put all initialized global and static data
13549 in the @samp{.data} section, and all uninitialized data in the
13550 @samp{.bss} section.
13554 @cindex smaller data references (PowerPC)
13555 @cindex .sdata/.sdata2 references (PowerPC)
13556 On embedded PowerPC systems, put global and static items less than or
13557 equal to @var{num} bytes into the small data or bss sections instead of
13558 the normal data or bss section. By default, @var{num} is 8. The
13559 @option{-G @var{num}} switch is also passed to the linker.
13560 All modules should be compiled with the same @option{-G @var{num}} value.
13563 @itemx -mno-regnames
13565 @opindex mno-regnames
13566 On System V.4 and embedded PowerPC systems do (do not) emit register
13567 names in the assembly language output using symbolic forms.
13570 @itemx -mno-longcall
13572 @opindex mno-longcall
13573 By default assume that all calls are far away so that a longer more
13574 expensive calling sequence is required. This is required for calls
13575 further than 32 megabytes (33,554,432 bytes) from the current location.
13576 A short call will be generated if the compiler knows
13577 the call cannot be that far away. This setting can be overridden by
13578 the @code{shortcall} function attribute, or by @code{#pragma
13581 Some linkers are capable of detecting out-of-range calls and generating
13582 glue code on the fly. On these systems, long calls are unnecessary and
13583 generate slower code. As of this writing, the AIX linker can do this,
13584 as can the GNU linker for PowerPC/64. It is planned to add this feature
13585 to the GNU linker for 32-bit PowerPC systems as well.
13587 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
13588 callee, L42'', plus a ``branch island'' (glue code). The two target
13589 addresses represent the callee and the ``branch island''. The
13590 Darwin/PPC linker will prefer the first address and generate a ``bl
13591 callee'' if the PPC ``bl'' instruction will reach the callee directly;
13592 otherwise, the linker will generate ``bl L42'' to call the ``branch
13593 island''. The ``branch island'' is appended to the body of the
13594 calling function; it computes the full 32-bit address of the callee
13597 On Mach-O (Darwin) systems, this option directs the compiler emit to
13598 the glue for every direct call, and the Darwin linker decides whether
13599 to use or discard it.
13601 In the future, we may cause GCC to ignore all longcall specifications
13602 when the linker is known to generate glue.
13606 Adds support for multithreading with the @dfn{pthreads} library.
13607 This option sets flags for both the preprocessor and linker.
13611 @node S/390 and zSeries Options
13612 @subsection S/390 and zSeries Options
13613 @cindex S/390 and zSeries Options
13615 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
13619 @itemx -msoft-float
13620 @opindex mhard-float
13621 @opindex msoft-float
13622 Use (do not use) the hardware floating-point instructions and registers
13623 for floating-point operations. When @option{-msoft-float} is specified,
13624 functions in @file{libgcc.a} will be used to perform floating-point
13625 operations. When @option{-mhard-float} is specified, the compiler
13626 generates IEEE floating-point instructions. This is the default.
13628 @item -mlong-double-64
13629 @itemx -mlong-double-128
13630 @opindex mlong-double-64
13631 @opindex mlong-double-128
13632 These switches control the size of @code{long double} type. A size
13633 of 64bit makes the @code{long double} type equivalent to the @code{double}
13634 type. This is the default.
13637 @itemx -mno-backchain
13638 @opindex mbackchain
13639 @opindex mno-backchain
13640 Store (do not store) the address of the caller's frame as backchain pointer
13641 into the callee's stack frame.
13642 A backchain may be needed to allow debugging using tools that do not understand
13643 DWARF-2 call frame information.
13644 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
13645 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
13646 the backchain is placed into the topmost word of the 96/160 byte register
13649 In general, code compiled with @option{-mbackchain} is call-compatible with
13650 code compiled with @option{-mmo-backchain}; however, use of the backchain
13651 for debugging purposes usually requires that the whole binary is built with
13652 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
13653 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13654 to build a linux kernel use @option{-msoft-float}.
13656 The default is to not maintain the backchain.
13658 @item -mpacked-stack
13659 @itemx -mno-packed-stack
13660 @opindex mpacked-stack
13661 @opindex mno-packed-stack
13662 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
13663 specified, the compiler uses the all fields of the 96/160 byte register save
13664 area only for their default purpose; unused fields still take up stack space.
13665 When @option{-mpacked-stack} is specified, register save slots are densely
13666 packed at the top of the register save area; unused space is reused for other
13667 purposes, allowing for more efficient use of the available stack space.
13668 However, when @option{-mbackchain} is also in effect, the topmost word of
13669 the save area is always used to store the backchain, and the return address
13670 register is always saved two words below the backchain.
13672 As long as the stack frame backchain is not used, code generated with
13673 @option{-mpacked-stack} is call-compatible with code generated with
13674 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
13675 S/390 or zSeries generated code that uses the stack frame backchain at run
13676 time, not just for debugging purposes. Such code is not call-compatible
13677 with code compiled with @option{-mpacked-stack}. Also, note that the
13678 combination of @option{-mbackchain},
13679 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13680 to build a linux kernel use @option{-msoft-float}.
13682 The default is to not use the packed stack layout.
13685 @itemx -mno-small-exec
13686 @opindex msmall-exec
13687 @opindex mno-small-exec
13688 Generate (or do not generate) code using the @code{bras} instruction
13689 to do subroutine calls.
13690 This only works reliably if the total executable size does not
13691 exceed 64k. The default is to use the @code{basr} instruction instead,
13692 which does not have this limitation.
13698 When @option{-m31} is specified, generate code compliant to the
13699 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
13700 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
13701 particular to generate 64-bit instructions. For the @samp{s390}
13702 targets, the default is @option{-m31}, while the @samp{s390x}
13703 targets default to @option{-m64}.
13709 When @option{-mzarch} is specified, generate code using the
13710 instructions available on z/Architecture.
13711 When @option{-mesa} is specified, generate code using the
13712 instructions available on ESA/390. Note that @option{-mesa} is
13713 not possible with @option{-m64}.
13714 When generating code compliant to the GNU/Linux for S/390 ABI,
13715 the default is @option{-mesa}. When generating code compliant
13716 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
13722 Generate (or do not generate) code using the @code{mvcle} instruction
13723 to perform block moves. When @option{-mno-mvcle} is specified,
13724 use a @code{mvc} loop instead. This is the default unless optimizing for
13731 Print (or do not print) additional debug information when compiling.
13732 The default is to not print debug information.
13734 @item -march=@var{cpu-type}
13736 Generate code that will run on @var{cpu-type}, which is the name of a system
13737 representing a certain processor type. Possible values for
13738 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
13739 When generating code using the instructions available on z/Architecture,
13740 the default is @option{-march=z900}. Otherwise, the default is
13741 @option{-march=g5}.
13743 @item -mtune=@var{cpu-type}
13745 Tune to @var{cpu-type} everything applicable about the generated code,
13746 except for the ABI and the set of available instructions.
13747 The list of @var{cpu-type} values is the same as for @option{-march}.
13748 The default is the value used for @option{-march}.
13751 @itemx -mno-tpf-trace
13752 @opindex mtpf-trace
13753 @opindex mno-tpf-trace
13754 Generate code that adds (does not add) in TPF OS specific branches to trace
13755 routines in the operating system. This option is off by default, even
13756 when compiling for the TPF OS@.
13759 @itemx -mno-fused-madd
13760 @opindex mfused-madd
13761 @opindex mno-fused-madd
13762 Generate code that uses (does not use) the floating point multiply and
13763 accumulate instructions. These instructions are generated by default if
13764 hardware floating point is used.
13766 @item -mwarn-framesize=@var{framesize}
13767 @opindex mwarn-framesize
13768 Emit a warning if the current function exceeds the given frame size. Because
13769 this is a compile time check it doesn't need to be a real problem when the program
13770 runs. It is intended to identify functions which most probably cause
13771 a stack overflow. It is useful to be used in an environment with limited stack
13772 size e.g.@: the linux kernel.
13774 @item -mwarn-dynamicstack
13775 @opindex mwarn-dynamicstack
13776 Emit a warning if the function calls alloca or uses dynamically
13777 sized arrays. This is generally a bad idea with a limited stack size.
13779 @item -mstack-guard=@var{stack-guard}
13780 @itemx -mstack-size=@var{stack-size}
13781 @opindex mstack-guard
13782 @opindex mstack-size
13783 If these options are provided the s390 back end emits additional instructions in
13784 the function prologue which trigger a trap if the stack size is @var{stack-guard}
13785 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
13786 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
13787 the frame size of the compiled function is chosen.
13788 These options are intended to be used to help debugging stack overflow problems.
13789 The additionally emitted code causes only little overhead and hence can also be
13790 used in production like systems without greater performance degradation. The given
13791 values have to be exact powers of 2 and @var{stack-size} has to be greater than
13792 @var{stack-guard} without exceeding 64k.
13793 In order to be efficient the extra code makes the assumption that the stack starts
13794 at an address aligned to the value given by @var{stack-size}.
13795 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
13798 @node Score Options
13799 @subsection Score Options
13800 @cindex Score Options
13802 These options are defined for Score implementations:
13807 Compile code for big endian mode. This is the default.
13811 Compile code for little endian mode.
13815 Disable generate bcnz instruction.
13819 Enable generate unaligned load and store instruction.
13823 Enable the use of multiply-accumulate instructions. Disabled by default.
13827 Specify the SCORE5 as the target architecture.
13831 Specify the SCORE5U of the target architecture.
13835 Specify the SCORE7 as the target architecture. This is the default.
13839 Specify the SCORE7D as the target architecture.
13843 @subsection SH Options
13845 These @samp{-m} options are defined for the SH implementations:
13850 Generate code for the SH1.
13854 Generate code for the SH2.
13857 Generate code for the SH2e.
13861 Generate code for the SH3.
13865 Generate code for the SH3e.
13869 Generate code for the SH4 without a floating-point unit.
13871 @item -m4-single-only
13872 @opindex m4-single-only
13873 Generate code for the SH4 with a floating-point unit that only
13874 supports single-precision arithmetic.
13878 Generate code for the SH4 assuming the floating-point unit is in
13879 single-precision mode by default.
13883 Generate code for the SH4.
13887 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
13888 floating-point unit is not used.
13890 @item -m4a-single-only
13891 @opindex m4a-single-only
13892 Generate code for the SH4a, in such a way that no double-precision
13893 floating point operations are used.
13896 @opindex m4a-single
13897 Generate code for the SH4a assuming the floating-point unit is in
13898 single-precision mode by default.
13902 Generate code for the SH4a.
13906 Same as @option{-m4a-nofpu}, except that it implicitly passes
13907 @option{-dsp} to the assembler. GCC doesn't generate any DSP
13908 instructions at the moment.
13912 Compile code for the processor in big endian mode.
13916 Compile code for the processor in little endian mode.
13920 Align doubles at 64-bit boundaries. Note that this changes the calling
13921 conventions, and thus some functions from the standard C library will
13922 not work unless you recompile it first with @option{-mdalign}.
13926 Shorten some address references at link time, when possible; uses the
13927 linker option @option{-relax}.
13931 Use 32-bit offsets in @code{switch} tables. The default is to use
13936 Enable the use of bit manipulation instructions on SH2A.
13940 Enable the use of the instruction @code{fmovd}.
13944 Comply with the calling conventions defined by Renesas.
13948 Comply with the calling conventions defined by Renesas.
13952 Comply with the calling conventions defined for GCC before the Renesas
13953 conventions were available. This option is the default for all
13954 targets of the SH toolchain except for @samp{sh-symbianelf}.
13957 @opindex mnomacsave
13958 Mark the @code{MAC} register as call-clobbered, even if
13959 @option{-mhitachi} is given.
13963 Increase IEEE-compliance of floating-point code.
13964 At the moment, this is equivalent to @option{-fno-finite-math-only}.
13965 When generating 16 bit SH opcodes, getting IEEE-conforming results for
13966 comparisons of NANs / infinities incurs extra overhead in every
13967 floating point comparison, therefore the default is set to
13968 @option{-ffinite-math-only}.
13970 @item -minline-ic_invalidate
13971 @opindex minline-ic_invalidate
13972 Inline code to invalidate instruction cache entries after setting up
13973 nested function trampolines.
13974 This option has no effect if -musermode is in effect and the selected
13975 code generation option (e.g. -m4) does not allow the use of the icbi
13977 If the selected code generation option does not allow the use of the icbi
13978 instruction, and -musermode is not in effect, the inlined code will
13979 manipulate the instruction cache address array directly with an associative
13980 write. This not only requires privileged mode, but it will also
13981 fail if the cache line had been mapped via the TLB and has become unmapped.
13985 Dump instruction size and location in the assembly code.
13988 @opindex mpadstruct
13989 This option is deprecated. It pads structures to multiple of 4 bytes,
13990 which is incompatible with the SH ABI@.
13994 Optimize for space instead of speed. Implied by @option{-Os}.
13997 @opindex mprefergot
13998 When generating position-independent code, emit function calls using
13999 the Global Offset Table instead of the Procedure Linkage Table.
14003 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
14004 if the inlined code would not work in user mode.
14005 This is the default when the target is @code{sh-*-linux*}.
14007 @item -multcost=@var{number}
14008 @opindex multcost=@var{number}
14009 Set the cost to assume for a multiply insn.
14011 @item -mdiv=@var{strategy}
14012 @opindex mdiv=@var{strategy}
14013 Set the division strategy to use for SHmedia code. @var{strategy} must be
14014 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
14015 inv:call2, inv:fp .
14016 "fp" performs the operation in floating point. This has a very high latency,
14017 but needs only a few instructions, so it might be a good choice if
14018 your code has enough easily exploitable ILP to allow the compiler to
14019 schedule the floating point instructions together with other instructions.
14020 Division by zero causes a floating point exception.
14021 "inv" uses integer operations to calculate the inverse of the divisor,
14022 and then multiplies the dividend with the inverse. This strategy allows
14023 cse and hoisting of the inverse calculation. Division by zero calculates
14024 an unspecified result, but does not trap.
14025 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
14026 have been found, or if the entire operation has been hoisted to the same
14027 place, the last stages of the inverse calculation are intertwined with the
14028 final multiply to reduce the overall latency, at the expense of using a few
14029 more instructions, and thus offering fewer scheduling opportunities with
14031 "call" calls a library function that usually implements the inv:minlat
14033 This gives high code density for m5-*media-nofpu compilations.
14034 "call2" uses a different entry point of the same library function, where it
14035 assumes that a pointer to a lookup table has already been set up, which
14036 exposes the pointer load to cse / code hoisting optimizations.
14037 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
14038 code generation, but if the code stays unoptimized, revert to the "call",
14039 "call2", or "fp" strategies, respectively. Note that the
14040 potentially-trapping side effect of division by zero is carried by a
14041 separate instruction, so it is possible that all the integer instructions
14042 are hoisted out, but the marker for the side effect stays where it is.
14043 A recombination to fp operations or a call is not possible in that case.
14044 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
14045 that the inverse calculation was nor separated from the multiply, they speed
14046 up division where the dividend fits into 20 bits (plus sign where applicable),
14047 by inserting a test to skip a number of operations in this case; this test
14048 slows down the case of larger dividends. inv20u assumes the case of a such
14049 a small dividend to be unlikely, and inv20l assumes it to be likely.
14051 @item -mdivsi3_libfunc=@var{name}
14052 @opindex mdivsi3_libfunc=@var{name}
14053 Set the name of the library function used for 32 bit signed division to
14054 @var{name}. This only affect the name used in the call and inv:call
14055 division strategies, and the compiler will still expect the same
14056 sets of input/output/clobbered registers as if this option was not present.
14058 @item -madjust-unroll
14059 @opindex madjust-unroll
14060 Throttle unrolling to avoid thrashing target registers.
14061 This option only has an effect if the gcc code base supports the
14062 TARGET_ADJUST_UNROLL_MAX target hook.
14064 @item -mindexed-addressing
14065 @opindex mindexed-addressing
14066 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
14067 This is only safe if the hardware and/or OS implement 32 bit wrap-around
14068 semantics for the indexed addressing mode. The architecture allows the
14069 implementation of processors with 64 bit MMU, which the OS could use to
14070 get 32 bit addressing, but since no current hardware implementation supports
14071 this or any other way to make the indexed addressing mode safe to use in
14072 the 32 bit ABI, the default is -mno-indexed-addressing.
14074 @item -mgettrcost=@var{number}
14075 @opindex mgettrcost=@var{number}
14076 Set the cost assumed for the gettr instruction to @var{number}.
14077 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
14081 Assume pt* instructions won't trap. This will generally generate better
14082 scheduled code, but is unsafe on current hardware. The current architecture
14083 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
14084 This has the unintentional effect of making it unsafe to schedule ptabs /
14085 ptrel before a branch, or hoist it out of a loop. For example,
14086 __do_global_ctors, a part of libgcc that runs constructors at program
14087 startup, calls functions in a list which is delimited by @minus{}1. With the
14088 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
14089 That means that all the constructors will be run a bit quicker, but when
14090 the loop comes to the end of the list, the program crashes because ptabs
14091 loads @minus{}1 into a target register. Since this option is unsafe for any
14092 hardware implementing the current architecture specification, the default
14093 is -mno-pt-fixed. Unless the user specifies a specific cost with
14094 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
14095 this deters register allocation using target registers for storing
14098 @item -minvalid-symbols
14099 @opindex minvalid-symbols
14100 Assume symbols might be invalid. Ordinary function symbols generated by
14101 the compiler will always be valid to load with movi/shori/ptabs or
14102 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
14103 to generate symbols that will cause ptabs / ptrel to trap.
14104 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
14105 It will then prevent cross-basic-block cse, hoisting and most scheduling
14106 of symbol loads. The default is @option{-mno-invalid-symbols}.
14109 @node SPARC Options
14110 @subsection SPARC Options
14111 @cindex SPARC options
14113 These @samp{-m} options are supported on the SPARC:
14116 @item -mno-app-regs
14118 @opindex mno-app-regs
14120 Specify @option{-mapp-regs} to generate output using the global registers
14121 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
14124 To be fully SVR4 ABI compliant at the cost of some performance loss,
14125 specify @option{-mno-app-regs}. You should compile libraries and system
14126 software with this option.
14129 @itemx -mhard-float
14131 @opindex mhard-float
14132 Generate output containing floating point instructions. This is the
14136 @itemx -msoft-float
14138 @opindex msoft-float
14139 Generate output containing library calls for floating point.
14140 @strong{Warning:} the requisite libraries are not available for all SPARC
14141 targets. Normally the facilities of the machine's usual C compiler are
14142 used, but this cannot be done directly in cross-compilation. You must make
14143 your own arrangements to provide suitable library functions for
14144 cross-compilation. The embedded targets @samp{sparc-*-aout} and
14145 @samp{sparclite-*-*} do provide software floating point support.
14147 @option{-msoft-float} changes the calling convention in the output file;
14148 therefore, it is only useful if you compile @emph{all} of a program with
14149 this option. In particular, you need to compile @file{libgcc.a}, the
14150 library that comes with GCC, with @option{-msoft-float} in order for
14153 @item -mhard-quad-float
14154 @opindex mhard-quad-float
14155 Generate output containing quad-word (long double) floating point
14158 @item -msoft-quad-float
14159 @opindex msoft-quad-float
14160 Generate output containing library calls for quad-word (long double)
14161 floating point instructions. The functions called are those specified
14162 in the SPARC ABI@. This is the default.
14164 As of this writing, there are no SPARC implementations that have hardware
14165 support for the quad-word floating point instructions. They all invoke
14166 a trap handler for one of these instructions, and then the trap handler
14167 emulates the effect of the instruction. Because of the trap handler overhead,
14168 this is much slower than calling the ABI library routines. Thus the
14169 @option{-msoft-quad-float} option is the default.
14171 @item -mno-unaligned-doubles
14172 @itemx -munaligned-doubles
14173 @opindex mno-unaligned-doubles
14174 @opindex munaligned-doubles
14175 Assume that doubles have 8 byte alignment. This is the default.
14177 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
14178 alignment only if they are contained in another type, or if they have an
14179 absolute address. Otherwise, it assumes they have 4 byte alignment.
14180 Specifying this option avoids some rare compatibility problems with code
14181 generated by other compilers. It is not the default because it results
14182 in a performance loss, especially for floating point code.
14184 @item -mno-faster-structs
14185 @itemx -mfaster-structs
14186 @opindex mno-faster-structs
14187 @opindex mfaster-structs
14188 With @option{-mfaster-structs}, the compiler assumes that structures
14189 should have 8 byte alignment. This enables the use of pairs of
14190 @code{ldd} and @code{std} instructions for copies in structure
14191 assignment, in place of twice as many @code{ld} and @code{st} pairs.
14192 However, the use of this changed alignment directly violates the SPARC
14193 ABI@. Thus, it's intended only for use on targets where the developer
14194 acknowledges that their resulting code will not be directly in line with
14195 the rules of the ABI@.
14197 @item -mimpure-text
14198 @opindex mimpure-text
14199 @option{-mimpure-text}, used in addition to @option{-shared}, tells
14200 the compiler to not pass @option{-z text} to the linker when linking a
14201 shared object. Using this option, you can link position-dependent
14202 code into a shared object.
14204 @option{-mimpure-text} suppresses the ``relocations remain against
14205 allocatable but non-writable sections'' linker error message.
14206 However, the necessary relocations will trigger copy-on-write, and the
14207 shared object is not actually shared across processes. Instead of
14208 using @option{-mimpure-text}, you should compile all source code with
14209 @option{-fpic} or @option{-fPIC}.
14211 This option is only available on SunOS and Solaris.
14213 @item -mcpu=@var{cpu_type}
14215 Set the instruction set, register set, and instruction scheduling parameters
14216 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
14217 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
14218 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
14219 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
14220 @samp{ultrasparc3}, @samp{niagara} and @samp{niagara2}.
14222 Default instruction scheduling parameters are used for values that select
14223 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
14224 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
14226 Here is a list of each supported architecture and their supported
14231 v8: supersparc, hypersparc
14232 sparclite: f930, f934, sparclite86x
14234 v9: ultrasparc, ultrasparc3, niagara, niagara2
14237 By default (unless configured otherwise), GCC generates code for the V7
14238 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
14239 additionally optimizes it for the Cypress CY7C602 chip, as used in the
14240 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
14241 SPARCStation 1, 2, IPX etc.
14243 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
14244 architecture. The only difference from V7 code is that the compiler emits
14245 the integer multiply and integer divide instructions which exist in SPARC-V8
14246 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
14247 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
14250 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
14251 the SPARC architecture. This adds the integer multiply, integer divide step
14252 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
14253 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
14254 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
14255 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
14256 MB86934 chip, which is the more recent SPARClite with FPU@.
14258 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
14259 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
14260 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
14261 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
14262 optimizes it for the TEMIC SPARClet chip.
14264 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
14265 architecture. This adds 64-bit integer and floating-point move instructions,
14266 3 additional floating-point condition code registers and conditional move
14267 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
14268 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
14269 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
14270 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
14271 @option{-mcpu=niagara}, the compiler additionally optimizes it for
14272 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
14273 additionally optimizes it for Sun UltraSPARC T2 chips.
14275 @item -mtune=@var{cpu_type}
14277 Set the instruction scheduling parameters for machine type
14278 @var{cpu_type}, but do not set the instruction set or register set that the
14279 option @option{-mcpu=@var{cpu_type}} would.
14281 The same values for @option{-mcpu=@var{cpu_type}} can be used for
14282 @option{-mtune=@var{cpu_type}}, but the only useful values are those
14283 that select a particular cpu implementation. Those are @samp{cypress},
14284 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
14285 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
14286 @samp{ultrasparc3}, @samp{niagara}, and @samp{niagara2}.
14291 @opindex mno-v8plus
14292 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
14293 difference from the V8 ABI is that the global and out registers are
14294 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
14295 mode for all SPARC-V9 processors.
14301 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
14302 Visual Instruction Set extensions. The default is @option{-mno-vis}.
14305 These @samp{-m} options are supported in addition to the above
14306 on SPARC-V9 processors in 64-bit environments:
14309 @item -mlittle-endian
14310 @opindex mlittle-endian
14311 Generate code for a processor running in little-endian mode. It is only
14312 available for a few configurations and most notably not on Solaris and Linux.
14318 Generate code for a 32-bit or 64-bit environment.
14319 The 32-bit environment sets int, long and pointer to 32 bits.
14320 The 64-bit environment sets int to 32 bits and long and pointer
14323 @item -mcmodel=medlow
14324 @opindex mcmodel=medlow
14325 Generate code for the Medium/Low code model: 64-bit addresses, programs
14326 must be linked in the low 32 bits of memory. Programs can be statically
14327 or dynamically linked.
14329 @item -mcmodel=medmid
14330 @opindex mcmodel=medmid
14331 Generate code for the Medium/Middle code model: 64-bit addresses, programs
14332 must be linked in the low 44 bits of memory, the text and data segments must
14333 be less than 2GB in size and the data segment must be located within 2GB of
14336 @item -mcmodel=medany
14337 @opindex mcmodel=medany
14338 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
14339 may be linked anywhere in memory, the text and data segments must be less
14340 than 2GB in size and the data segment must be located within 2GB of the
14343 @item -mcmodel=embmedany
14344 @opindex mcmodel=embmedany
14345 Generate code for the Medium/Anywhere code model for embedded systems:
14346 64-bit addresses, the text and data segments must be less than 2GB in
14347 size, both starting anywhere in memory (determined at link time). The
14348 global register %g4 points to the base of the data segment. Programs
14349 are statically linked and PIC is not supported.
14352 @itemx -mno-stack-bias
14353 @opindex mstack-bias
14354 @opindex mno-stack-bias
14355 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
14356 frame pointer if present, are offset by @minus{}2047 which must be added back
14357 when making stack frame references. This is the default in 64-bit mode.
14358 Otherwise, assume no such offset is present.
14361 These switches are supported in addition to the above on Solaris:
14366 Add support for multithreading using the Solaris threads library. This
14367 option sets flags for both the preprocessor and linker. This option does
14368 not affect the thread safety of object code produced by the compiler or
14369 that of libraries supplied with it.
14373 Add support for multithreading using the POSIX threads library. This
14374 option sets flags for both the preprocessor and linker. This option does
14375 not affect the thread safety of object code produced by the compiler or
14376 that of libraries supplied with it.
14380 This is a synonym for @option{-pthreads}.
14384 @subsection SPU Options
14385 @cindex SPU options
14387 These @samp{-m} options are supported on the SPU:
14391 @itemx -merror-reloc
14392 @opindex mwarn-reloc
14393 @opindex merror-reloc
14395 The loader for SPU does not handle dynamic relocations. By default, GCC
14396 will give an error when it generates code that requires a dynamic
14397 relocation. @option{-mno-error-reloc} disables the error,
14398 @option{-mwarn-reloc} will generate a warning instead.
14401 @itemx -munsafe-dma
14403 @opindex munsafe-dma
14405 Instructions which initiate or test completion of DMA must not be
14406 reordered with respect to loads and stores of the memory which is being
14407 accessed. Users typically address this problem using the volatile
14408 keyword, but that can lead to inefficient code in places where the
14409 memory is known to not change. Rather than mark the memory as volatile
14410 we treat the DMA instructions as potentially effecting all memory. With
14411 @option{-munsafe-dma} users must use the volatile keyword to protect
14414 @item -mbranch-hints
14415 @opindex mbranch-hints
14417 By default, GCC will generate a branch hint instruction to avoid
14418 pipeline stalls for always taken or probably taken branches. A hint
14419 will not be generated closer than 8 instructions away from its branch.
14420 There is little reason to disable them, except for debugging purposes,
14421 or to make an object a little bit smaller.
14425 @opindex msmall-mem
14426 @opindex mlarge-mem
14428 By default, GCC generates code assuming that addresses are never larger
14429 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
14430 a full 32 bit address.
14435 By default, GCC links against startup code that assumes the SPU-style
14436 main function interface (which has an unconventional parameter list).
14437 With @option{-mstdmain}, GCC will link your program against startup
14438 code that assumes a C99-style interface to @code{main}, including a
14439 local copy of @code{argv} strings.
14441 @item -mfixed-range=@var{register-range}
14442 @opindex mfixed-range
14443 Generate code treating the given register range as fixed registers.
14444 A fixed register is one that the register allocator can not use. This is
14445 useful when compiling kernel code. A register range is specified as
14446 two registers separated by a dash. Multiple register ranges can be
14447 specified separated by a comma.
14451 @node System V Options
14452 @subsection Options for System V
14454 These additional options are available on System V Release 4 for
14455 compatibility with other compilers on those systems:
14460 Create a shared object.
14461 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
14465 Identify the versions of each tool used by the compiler, in a
14466 @code{.ident} assembler directive in the output.
14470 Refrain from adding @code{.ident} directives to the output file (this is
14473 @item -YP,@var{dirs}
14475 Search the directories @var{dirs}, and no others, for libraries
14476 specified with @option{-l}.
14478 @item -Ym,@var{dir}
14480 Look in the directory @var{dir} to find the M4 preprocessor.
14481 The assembler uses this option.
14482 @c This is supposed to go with a -Yd for predefined M4 macro files, but
14483 @c the generic assembler that comes with Solaris takes just -Ym.
14487 @subsection V850 Options
14488 @cindex V850 Options
14490 These @samp{-m} options are defined for V850 implementations:
14494 @itemx -mno-long-calls
14495 @opindex mlong-calls
14496 @opindex mno-long-calls
14497 Treat all calls as being far away (near). If calls are assumed to be
14498 far away, the compiler will always load the functions address up into a
14499 register, and call indirect through the pointer.
14505 Do not optimize (do optimize) basic blocks that use the same index
14506 pointer 4 or more times to copy pointer into the @code{ep} register, and
14507 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
14508 option is on by default if you optimize.
14510 @item -mno-prolog-function
14511 @itemx -mprolog-function
14512 @opindex mno-prolog-function
14513 @opindex mprolog-function
14514 Do not use (do use) external functions to save and restore registers
14515 at the prologue and epilogue of a function. The external functions
14516 are slower, but use less code space if more than one function saves
14517 the same number of registers. The @option{-mprolog-function} option
14518 is on by default if you optimize.
14522 Try to make the code as small as possible. At present, this just turns
14523 on the @option{-mep} and @option{-mprolog-function} options.
14525 @item -mtda=@var{n}
14527 Put static or global variables whose size is @var{n} bytes or less into
14528 the tiny data area that register @code{ep} points to. The tiny data
14529 area can hold up to 256 bytes in total (128 bytes for byte references).
14531 @item -msda=@var{n}
14533 Put static or global variables whose size is @var{n} bytes or less into
14534 the small data area that register @code{gp} points to. The small data
14535 area can hold up to 64 kilobytes.
14537 @item -mzda=@var{n}
14539 Put static or global variables whose size is @var{n} bytes or less into
14540 the first 32 kilobytes of memory.
14544 Specify that the target processor is the V850.
14547 @opindex mbig-switch
14548 Generate code suitable for big switch tables. Use this option only if
14549 the assembler/linker complain about out of range branches within a switch
14554 This option will cause r2 and r5 to be used in the code generated by
14555 the compiler. This setting is the default.
14557 @item -mno-app-regs
14558 @opindex mno-app-regs
14559 This option will cause r2 and r5 to be treated as fixed registers.
14563 Specify that the target processor is the V850E1. The preprocessor
14564 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
14565 this option is used.
14569 Specify that the target processor is the V850E@. The preprocessor
14570 constant @samp{__v850e__} will be defined if this option is used.
14572 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
14573 are defined then a default target processor will be chosen and the
14574 relevant @samp{__v850*__} preprocessor constant will be defined.
14576 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
14577 defined, regardless of which processor variant is the target.
14579 @item -mdisable-callt
14580 @opindex mdisable-callt
14581 This option will suppress generation of the CALLT instruction for the
14582 v850e and v850e1 flavors of the v850 architecture. The default is
14583 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
14588 @subsection VAX Options
14589 @cindex VAX options
14591 These @samp{-m} options are defined for the VAX:
14596 Do not output certain jump instructions (@code{aobleq} and so on)
14597 that the Unix assembler for the VAX cannot handle across long
14602 Do output those jump instructions, on the assumption that you
14603 will assemble with the GNU assembler.
14607 Output code for g-format floating point numbers instead of d-format.
14610 @node VxWorks Options
14611 @subsection VxWorks Options
14612 @cindex VxWorks Options
14614 The options in this section are defined for all VxWorks targets.
14615 Options specific to the target hardware are listed with the other
14616 options for that target.
14621 GCC can generate code for both VxWorks kernels and real time processes
14622 (RTPs). This option switches from the former to the latter. It also
14623 defines the preprocessor macro @code{__RTP__}.
14626 @opindex non-static
14627 Link an RTP executable against shared libraries rather than static
14628 libraries. The options @option{-static} and @option{-shared} can
14629 also be used for RTPs (@pxref{Link Options}); @option{-static}
14636 These options are passed down to the linker. They are defined for
14637 compatibility with Diab.
14640 @opindex Xbind-lazy
14641 Enable lazy binding of function calls. This option is equivalent to
14642 @option{-Wl,-z,now} and is defined for compatibility with Diab.
14646 Disable lazy binding of function calls. This option is the default and
14647 is defined for compatibility with Diab.
14650 @node x86-64 Options
14651 @subsection x86-64 Options
14652 @cindex x86-64 options
14654 These are listed under @xref{i386 and x86-64 Options}.
14656 @node Xstormy16 Options
14657 @subsection Xstormy16 Options
14658 @cindex Xstormy16 Options
14660 These options are defined for Xstormy16:
14665 Choose startup files and linker script suitable for the simulator.
14668 @node Xtensa Options
14669 @subsection Xtensa Options
14670 @cindex Xtensa Options
14672 These options are supported for Xtensa targets:
14676 @itemx -mno-const16
14678 @opindex mno-const16
14679 Enable or disable use of @code{CONST16} instructions for loading
14680 constant values. The @code{CONST16} instruction is currently not a
14681 standard option from Tensilica. When enabled, @code{CONST16}
14682 instructions are always used in place of the standard @code{L32R}
14683 instructions. The use of @code{CONST16} is enabled by default only if
14684 the @code{L32R} instruction is not available.
14687 @itemx -mno-fused-madd
14688 @opindex mfused-madd
14689 @opindex mno-fused-madd
14690 Enable or disable use of fused multiply/add and multiply/subtract
14691 instructions in the floating-point option. This has no effect if the
14692 floating-point option is not also enabled. Disabling fused multiply/add
14693 and multiply/subtract instructions forces the compiler to use separate
14694 instructions for the multiply and add/subtract operations. This may be
14695 desirable in some cases where strict IEEE 754-compliant results are
14696 required: the fused multiply add/subtract instructions do not round the
14697 intermediate result, thereby producing results with @emph{more} bits of
14698 precision than specified by the IEEE standard. Disabling fused multiply
14699 add/subtract instructions also ensures that the program output is not
14700 sensitive to the compiler's ability to combine multiply and add/subtract
14703 @item -mserialize-volatile
14704 @itemx -mno-serialize-volatile
14705 @opindex mserialize-volatile
14706 @opindex mno-serialize-volatile
14707 When this option is enabled, GCC inserts @code{MEMW} instructions before
14708 @code{volatile} memory references to guarantee sequential consistency.
14709 The default is @option{-mserialize-volatile}. Use
14710 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
14712 @item -mtext-section-literals
14713 @itemx -mno-text-section-literals
14714 @opindex mtext-section-literals
14715 @opindex mno-text-section-literals
14716 Control the treatment of literal pools. The default is
14717 @option{-mno-text-section-literals}, which places literals in a separate
14718 section in the output file. This allows the literal pool to be placed
14719 in a data RAM/ROM, and it also allows the linker to combine literal
14720 pools from separate object files to remove redundant literals and
14721 improve code size. With @option{-mtext-section-literals}, the literals
14722 are interspersed in the text section in order to keep them as close as
14723 possible to their references. This may be necessary for large assembly
14726 @item -mtarget-align
14727 @itemx -mno-target-align
14728 @opindex mtarget-align
14729 @opindex mno-target-align
14730 When this option is enabled, GCC instructs the assembler to
14731 automatically align instructions to reduce branch penalties at the
14732 expense of some code density. The assembler attempts to widen density
14733 instructions to align branch targets and the instructions following call
14734 instructions. If there are not enough preceding safe density
14735 instructions to align a target, no widening will be performed. The
14736 default is @option{-mtarget-align}. These options do not affect the
14737 treatment of auto-aligned instructions like @code{LOOP}, which the
14738 assembler will always align, either by widening density instructions or
14739 by inserting no-op instructions.
14742 @itemx -mno-longcalls
14743 @opindex mlongcalls
14744 @opindex mno-longcalls
14745 When this option is enabled, GCC instructs the assembler to translate
14746 direct calls to indirect calls unless it can determine that the target
14747 of a direct call is in the range allowed by the call instruction. This
14748 translation typically occurs for calls to functions in other source
14749 files. Specifically, the assembler translates a direct @code{CALL}
14750 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
14751 The default is @option{-mno-longcalls}. This option should be used in
14752 programs where the call target can potentially be out of range. This
14753 option is implemented in the assembler, not the compiler, so the
14754 assembly code generated by GCC will still show direct call
14755 instructions---look at the disassembled object code to see the actual
14756 instructions. Note that the assembler will use an indirect call for
14757 every cross-file call, not just those that really will be out of range.
14760 @node zSeries Options
14761 @subsection zSeries Options
14762 @cindex zSeries options
14764 These are listed under @xref{S/390 and zSeries Options}.
14766 @node Code Gen Options
14767 @section Options for Code Generation Conventions
14768 @cindex code generation conventions
14769 @cindex options, code generation
14770 @cindex run-time options
14772 These machine-independent options control the interface conventions
14773 used in code generation.
14775 Most of them have both positive and negative forms; the negative form
14776 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
14777 one of the forms is listed---the one which is not the default. You
14778 can figure out the other form by either removing @samp{no-} or adding
14782 @item -fbounds-check
14783 @opindex fbounds-check
14784 For front-ends that support it, generate additional code to check that
14785 indices used to access arrays are within the declared range. This is
14786 currently only supported by the Java and Fortran front-ends, where
14787 this option defaults to true and false respectively.
14791 This option generates traps for signed overflow on addition, subtraction,
14792 multiplication operations.
14796 This option instructs the compiler to assume that signed arithmetic
14797 overflow of addition, subtraction and multiplication wraps around
14798 using twos-complement representation. This flag enables some optimizations
14799 and disables others. This option is enabled by default for the Java
14800 front-end, as required by the Java language specification.
14803 @opindex fexceptions
14804 Enable exception handling. Generates extra code needed to propagate
14805 exceptions. For some targets, this implies GCC will generate frame
14806 unwind information for all functions, which can produce significant data
14807 size overhead, although it does not affect execution. If you do not
14808 specify this option, GCC will enable it by default for languages like
14809 C++ which normally require exception handling, and disable it for
14810 languages like C that do not normally require it. However, you may need
14811 to enable this option when compiling C code that needs to interoperate
14812 properly with exception handlers written in C++. You may also wish to
14813 disable this option if you are compiling older C++ programs that don't
14814 use exception handling.
14816 @item -fnon-call-exceptions
14817 @opindex fnon-call-exceptions
14818 Generate code that allows trapping instructions to throw exceptions.
14819 Note that this requires platform-specific runtime support that does
14820 not exist everywhere. Moreover, it only allows @emph{trapping}
14821 instructions to throw exceptions, i.e.@: memory references or floating
14822 point instructions. It does not allow exceptions to be thrown from
14823 arbitrary signal handlers such as @code{SIGALRM}.
14825 @item -funwind-tables
14826 @opindex funwind-tables
14827 Similar to @option{-fexceptions}, except that it will just generate any needed
14828 static data, but will not affect the generated code in any other way.
14829 You will normally not enable this option; instead, a language processor
14830 that needs this handling would enable it on your behalf.
14832 @item -fasynchronous-unwind-tables
14833 @opindex fasynchronous-unwind-tables
14834 Generate unwind table in dwarf2 format, if supported by target machine. The
14835 table is exact at each instruction boundary, so it can be used for stack
14836 unwinding from asynchronous events (such as debugger or garbage collector).
14838 @item -fpcc-struct-return
14839 @opindex fpcc-struct-return
14840 Return ``short'' @code{struct} and @code{union} values in memory like
14841 longer ones, rather than in registers. This convention is less
14842 efficient, but it has the advantage of allowing intercallability between
14843 GCC-compiled files and files compiled with other compilers, particularly
14844 the Portable C Compiler (pcc).
14846 The precise convention for returning structures in memory depends
14847 on the target configuration macros.
14849 Short structures and unions are those whose size and alignment match
14850 that of some integer type.
14852 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
14853 switch is not binary compatible with code compiled with the
14854 @option{-freg-struct-return} switch.
14855 Use it to conform to a non-default application binary interface.
14857 @item -freg-struct-return
14858 @opindex freg-struct-return
14859 Return @code{struct} and @code{union} values in registers when possible.
14860 This is more efficient for small structures than
14861 @option{-fpcc-struct-return}.
14863 If you specify neither @option{-fpcc-struct-return} nor
14864 @option{-freg-struct-return}, GCC defaults to whichever convention is
14865 standard for the target. If there is no standard convention, GCC
14866 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
14867 the principal compiler. In those cases, we can choose the standard, and
14868 we chose the more efficient register return alternative.
14870 @strong{Warning:} code compiled with the @option{-freg-struct-return}
14871 switch is not binary compatible with code compiled with the
14872 @option{-fpcc-struct-return} switch.
14873 Use it to conform to a non-default application binary interface.
14875 @item -fshort-enums
14876 @opindex fshort-enums
14877 Allocate to an @code{enum} type only as many bytes as it needs for the
14878 declared range of possible values. Specifically, the @code{enum} type
14879 will be equivalent to the smallest integer type which has enough room.
14881 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
14882 code that is not binary compatible with code generated without that switch.
14883 Use it to conform to a non-default application binary interface.
14885 @item -fshort-double
14886 @opindex fshort-double
14887 Use the same size for @code{double} as for @code{float}.
14889 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
14890 code that is not binary compatible with code generated without that switch.
14891 Use it to conform to a non-default application binary interface.
14893 @item -fshort-wchar
14894 @opindex fshort-wchar
14895 Override the underlying type for @samp{wchar_t} to be @samp{short
14896 unsigned int} instead of the default for the target. This option is
14897 useful for building programs to run under WINE@.
14899 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
14900 code that is not binary compatible with code generated without that switch.
14901 Use it to conform to a non-default application binary interface.
14904 @opindex fno-common
14905 In C, allocate even uninitialized global variables in the data section of the
14906 object file, rather than generating them as common blocks. This has the
14907 effect that if the same variable is declared (without @code{extern}) in
14908 two different compilations, you will get an error when you link them.
14909 The only reason this might be useful is if you wish to verify that the
14910 program will work on other systems which always work this way.
14914 Ignore the @samp{#ident} directive.
14916 @item -finhibit-size-directive
14917 @opindex finhibit-size-directive
14918 Don't output a @code{.size} assembler directive, or anything else that
14919 would cause trouble if the function is split in the middle, and the
14920 two halves are placed at locations far apart in memory. This option is
14921 used when compiling @file{crtstuff.c}; you should not need to use it
14924 @item -fverbose-asm
14925 @opindex fverbose-asm
14926 Put extra commentary information in the generated assembly code to
14927 make it more readable. This option is generally only of use to those
14928 who actually need to read the generated assembly code (perhaps while
14929 debugging the compiler itself).
14931 @option{-fno-verbose-asm}, the default, causes the
14932 extra information to be omitted and is useful when comparing two assembler
14935 @item -frecord-gcc-switches
14936 @opindex frecord-gcc-switches
14937 This switch causes the command line that was used to invoke the
14938 compiler to be recorded into the object file that is being created.
14939 This switch is only implemented on some targets and the exact format
14940 of the recording is target and binary file format dependent, but it
14941 usually takes the form of a section containing ASCII text. This
14942 switch is related to the @option{-fverbose-asm} switch, but that
14943 switch only records information in the assembler output file as
14944 comments, so it never reaches the object file.
14948 @cindex global offset table
14950 Generate position-independent code (PIC) suitable for use in a shared
14951 library, if supported for the target machine. Such code accesses all
14952 constant addresses through a global offset table (GOT)@. The dynamic
14953 loader resolves the GOT entries when the program starts (the dynamic
14954 loader is not part of GCC; it is part of the operating system). If
14955 the GOT size for the linked executable exceeds a machine-specific
14956 maximum size, you get an error message from the linker indicating that
14957 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
14958 instead. (These maximums are 8k on the SPARC and 32k
14959 on the m68k and RS/6000. The 386 has no such limit.)
14961 Position-independent code requires special support, and therefore works
14962 only on certain machines. For the 386, GCC supports PIC for System V
14963 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
14964 position-independent.
14966 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14971 If supported for the target machine, emit position-independent code,
14972 suitable for dynamic linking and avoiding any limit on the size of the
14973 global offset table. This option makes a difference on the m68k,
14974 PowerPC and SPARC@.
14976 Position-independent code requires special support, and therefore works
14977 only on certain machines.
14979 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14986 These options are similar to @option{-fpic} and @option{-fPIC}, but
14987 generated position independent code can be only linked into executables.
14988 Usually these options are used when @option{-pie} GCC option will be
14989 used during linking.
14991 @option{-fpie} and @option{-fPIE} both define the macros
14992 @code{__pie__} and @code{__PIE__}. The macros have the value 1
14993 for @option{-fpie} and 2 for @option{-fPIE}.
14995 @item -fno-jump-tables
14996 @opindex fno-jump-tables
14997 Do not use jump tables for switch statements even where it would be
14998 more efficient than other code generation strategies. This option is
14999 of use in conjunction with @option{-fpic} or @option{-fPIC} for
15000 building code which forms part of a dynamic linker and cannot
15001 reference the address of a jump table. On some targets, jump tables
15002 do not require a GOT and this option is not needed.
15004 @item -ffixed-@var{reg}
15006 Treat the register named @var{reg} as a fixed register; generated code
15007 should never refer to it (except perhaps as a stack pointer, frame
15008 pointer or in some other fixed role).
15010 @var{reg} must be the name of a register. The register names accepted
15011 are machine-specific and are defined in the @code{REGISTER_NAMES}
15012 macro in the machine description macro file.
15014 This flag does not have a negative form, because it specifies a
15017 @item -fcall-used-@var{reg}
15018 @opindex fcall-used
15019 Treat the register named @var{reg} as an allocable register that is
15020 clobbered by function calls. It may be allocated for temporaries or
15021 variables that do not live across a call. Functions compiled this way
15022 will not save and restore the register @var{reg}.
15024 It is an error to used this flag with the frame pointer or stack pointer.
15025 Use of this flag for other registers that have fixed pervasive roles in
15026 the machine's execution model will produce disastrous results.
15028 This flag does not have a negative form, because it specifies a
15031 @item -fcall-saved-@var{reg}
15032 @opindex fcall-saved
15033 Treat the register named @var{reg} as an allocable register saved by
15034 functions. It may be allocated even for temporaries or variables that
15035 live across a call. Functions compiled this way will save and restore
15036 the register @var{reg} if they use it.
15038 It is an error to used this flag with the frame pointer or stack pointer.
15039 Use of this flag for other registers that have fixed pervasive roles in
15040 the machine's execution model will produce disastrous results.
15042 A different sort of disaster will result from the use of this flag for
15043 a register in which function values may be returned.
15045 This flag does not have a negative form, because it specifies a
15048 @item -fpack-struct[=@var{n}]
15049 @opindex fpack-struct
15050 Without a value specified, pack all structure members together without
15051 holes. When a value is specified (which must be a small power of two), pack
15052 structure members according to this value, representing the maximum
15053 alignment (that is, objects with default alignment requirements larger than
15054 this will be output potentially unaligned at the next fitting location.
15056 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
15057 code that is not binary compatible with code generated without that switch.
15058 Additionally, it makes the code suboptimal.
15059 Use it to conform to a non-default application binary interface.
15061 @item -finstrument-functions
15062 @opindex finstrument-functions
15063 Generate instrumentation calls for entry and exit to functions. Just
15064 after function entry and just before function exit, the following
15065 profiling functions will be called with the address of the current
15066 function and its call site. (On some platforms,
15067 @code{__builtin_return_address} does not work beyond the current
15068 function, so the call site information may not be available to the
15069 profiling functions otherwise.)
15072 void __cyg_profile_func_enter (void *this_fn,
15074 void __cyg_profile_func_exit (void *this_fn,
15078 The first argument is the address of the start of the current function,
15079 which may be looked up exactly in the symbol table.
15081 This instrumentation is also done for functions expanded inline in other
15082 functions. The profiling calls will indicate where, conceptually, the
15083 inline function is entered and exited. This means that addressable
15084 versions of such functions must be available. If all your uses of a
15085 function are expanded inline, this may mean an additional expansion of
15086 code size. If you use @samp{extern inline} in your C code, an
15087 addressable version of such functions must be provided. (This is
15088 normally the case anyways, but if you get lucky and the optimizer always
15089 expands the functions inline, you might have gotten away without
15090 providing static copies.)
15092 A function may be given the attribute @code{no_instrument_function}, in
15093 which case this instrumentation will not be done. This can be used, for
15094 example, for the profiling functions listed above, high-priority
15095 interrupt routines, and any functions from which the profiling functions
15096 cannot safely be called (perhaps signal handlers, if the profiling
15097 routines generate output or allocate memory).
15099 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
15100 @opindex finstrument-functions-exclude-file-list
15102 Set the list of functions that are excluded from instrumentation (see
15103 the description of @code{-finstrument-functions}). If the file that
15104 contains a function definition matches with one of @var{file}, then
15105 that function is not instrumented. The match is done on substrings:
15106 if the @var{file} parameter is a substring of the file name, it is
15107 considered to be a match.
15110 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
15111 will exclude any inline function defined in files whose pathnames
15112 contain @code{/bits/stl} or @code{include/sys}.
15114 If, for some reason, you want to include letter @code{','} in one of
15115 @var{sym}, write @code{'\,'}. For example,
15116 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
15117 (note the single quote surrounding the option).
15119 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
15120 @opindex finstrument-functions-exclude-function-list
15122 This is similar to @code{-finstrument-functions-exclude-file-list},
15123 but this option sets the list of function names to be excluded from
15124 instrumentation. The function name to be matched is its user-visible
15125 name, such as @code{vector<int> blah(const vector<int> &)}, not the
15126 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
15127 match is done on substrings: if the @var{sym} parameter is a substring
15128 of the function name, it is considered to be a match.
15130 @item -fstack-check
15131 @opindex fstack-check
15132 Generate code to verify that you do not go beyond the boundary of the
15133 stack. You should specify this flag if you are running in an
15134 environment with multiple threads, but only rarely need to specify it in
15135 a single-threaded environment since stack overflow is automatically
15136 detected on nearly all systems if there is only one stack.
15138 Note that this switch does not actually cause checking to be done; the
15139 operating system must do that. The switch causes generation of code
15140 to ensure that the operating system sees the stack being extended.
15142 @item -fstack-limit-register=@var{reg}
15143 @itemx -fstack-limit-symbol=@var{sym}
15144 @itemx -fno-stack-limit
15145 @opindex fstack-limit-register
15146 @opindex fstack-limit-symbol
15147 @opindex fno-stack-limit
15148 Generate code to ensure that the stack does not grow beyond a certain value,
15149 either the value of a register or the address of a symbol. If the stack
15150 would grow beyond the value, a signal is raised. For most targets,
15151 the signal is raised before the stack overruns the boundary, so
15152 it is possible to catch the signal without taking special precautions.
15154 For instance, if the stack starts at absolute address @samp{0x80000000}
15155 and grows downwards, you can use the flags
15156 @option{-fstack-limit-symbol=__stack_limit} and
15157 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
15158 of 128KB@. Note that this may only work with the GNU linker.
15160 @cindex aliasing of parameters
15161 @cindex parameters, aliased
15162 @item -fargument-alias
15163 @itemx -fargument-noalias
15164 @itemx -fargument-noalias-global
15165 @itemx -fargument-noalias-anything
15166 @opindex fargument-alias
15167 @opindex fargument-noalias
15168 @opindex fargument-noalias-global
15169 @opindex fargument-noalias-anything
15170 Specify the possible relationships among parameters and between
15171 parameters and global data.
15173 @option{-fargument-alias} specifies that arguments (parameters) may
15174 alias each other and may alias global storage.@*
15175 @option{-fargument-noalias} specifies that arguments do not alias
15176 each other, but may alias global storage.@*
15177 @option{-fargument-noalias-global} specifies that arguments do not
15178 alias each other and do not alias global storage.
15179 @option{-fargument-noalias-anything} specifies that arguments do not
15180 alias any other storage.
15182 Each language will automatically use whatever option is required by
15183 the language standard. You should not need to use these options yourself.
15185 @item -fleading-underscore
15186 @opindex fleading-underscore
15187 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
15188 change the way C symbols are represented in the object file. One use
15189 is to help link with legacy assembly code.
15191 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
15192 generate code that is not binary compatible with code generated without that
15193 switch. Use it to conform to a non-default application binary interface.
15194 Not all targets provide complete support for this switch.
15196 @item -ftls-model=@var{model}
15197 @opindex ftls-model
15198 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
15199 The @var{model} argument should be one of @code{global-dynamic},
15200 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
15202 The default without @option{-fpic} is @code{initial-exec}; with
15203 @option{-fpic} the default is @code{global-dynamic}.
15205 @item -fvisibility=@var{default|internal|hidden|protected}
15206 @opindex fvisibility
15207 Set the default ELF image symbol visibility to the specified option---all
15208 symbols will be marked with this unless overridden within the code.
15209 Using this feature can very substantially improve linking and
15210 load times of shared object libraries, produce more optimized
15211 code, provide near-perfect API export and prevent symbol clashes.
15212 It is @strong{strongly} recommended that you use this in any shared objects
15215 Despite the nomenclature, @code{default} always means public ie;
15216 available to be linked against from outside the shared object.
15217 @code{protected} and @code{internal} are pretty useless in real-world
15218 usage so the only other commonly used option will be @code{hidden}.
15219 The default if @option{-fvisibility} isn't specified is
15220 @code{default}, i.e., make every
15221 symbol public---this causes the same behavior as previous versions of
15224 A good explanation of the benefits offered by ensuring ELF
15225 symbols have the correct visibility is given by ``How To Write
15226 Shared Libraries'' by Ulrich Drepper (which can be found at
15227 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
15228 solution made possible by this option to marking things hidden when
15229 the default is public is to make the default hidden and mark things
15230 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
15231 and @code{__attribute__ ((visibility("default")))} instead of
15232 @code{__declspec(dllexport)} you get almost identical semantics with
15233 identical syntax. This is a great boon to those working with
15234 cross-platform projects.
15236 For those adding visibility support to existing code, you may find
15237 @samp{#pragma GCC visibility} of use. This works by you enclosing
15238 the declarations you wish to set visibility for with (for example)
15239 @samp{#pragma GCC visibility push(hidden)} and
15240 @samp{#pragma GCC visibility pop}.
15241 Bear in mind that symbol visibility should be viewed @strong{as
15242 part of the API interface contract} and thus all new code should
15243 always specify visibility when it is not the default ie; declarations
15244 only for use within the local DSO should @strong{always} be marked explicitly
15245 as hidden as so to avoid PLT indirection overheads---making this
15246 abundantly clear also aids readability and self-documentation of the code.
15247 Note that due to ISO C++ specification requirements, operator new and
15248 operator delete must always be of default visibility.
15250 Be aware that headers from outside your project, in particular system
15251 headers and headers from any other library you use, may not be
15252 expecting to be compiled with visibility other than the default. You
15253 may need to explicitly say @samp{#pragma GCC visibility push(default)}
15254 before including any such headers.
15256 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
15257 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
15258 no modifications. However, this means that calls to @samp{extern}
15259 functions with no explicit visibility will use the PLT, so it is more
15260 effective to use @samp{__attribute ((visibility))} and/or
15261 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
15262 declarations should be treated as hidden.
15264 Note that @samp{-fvisibility} does affect C++ vague linkage
15265 entities. This means that, for instance, an exception class that will
15266 be thrown between DSOs must be explicitly marked with default
15267 visibility so that the @samp{type_info} nodes will be unified between
15270 An overview of these techniques, their benefits and how to use them
15271 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
15277 @node Environment Variables
15278 @section Environment Variables Affecting GCC
15279 @cindex environment variables
15281 @c man begin ENVIRONMENT
15282 This section describes several environment variables that affect how GCC
15283 operates. Some of them work by specifying directories or prefixes to use
15284 when searching for various kinds of files. Some are used to specify other
15285 aspects of the compilation environment.
15287 Note that you can also specify places to search using options such as
15288 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
15289 take precedence over places specified using environment variables, which
15290 in turn take precedence over those specified by the configuration of GCC@.
15291 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
15292 GNU Compiler Collection (GCC) Internals}.
15297 @c @itemx LC_COLLATE
15299 @c @itemx LC_MONETARY
15300 @c @itemx LC_NUMERIC
15305 @c @findex LC_COLLATE
15306 @findex LC_MESSAGES
15307 @c @findex LC_MONETARY
15308 @c @findex LC_NUMERIC
15312 These environment variables control the way that GCC uses
15313 localization information that allow GCC to work with different
15314 national conventions. GCC inspects the locale categories
15315 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
15316 so. These locale categories can be set to any value supported by your
15317 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
15318 Kingdom encoded in UTF-8.
15320 The @env{LC_CTYPE} environment variable specifies character
15321 classification. GCC uses it to determine the character boundaries in
15322 a string; this is needed for some multibyte encodings that contain quote
15323 and escape characters that would otherwise be interpreted as a string
15326 The @env{LC_MESSAGES} environment variable specifies the language to
15327 use in diagnostic messages.
15329 If the @env{LC_ALL} environment variable is set, it overrides the value
15330 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
15331 and @env{LC_MESSAGES} default to the value of the @env{LANG}
15332 environment variable. If none of these variables are set, GCC
15333 defaults to traditional C English behavior.
15337 If @env{TMPDIR} is set, it specifies the directory to use for temporary
15338 files. GCC uses temporary files to hold the output of one stage of
15339 compilation which is to be used as input to the next stage: for example,
15340 the output of the preprocessor, which is the input to the compiler
15343 @item GCC_EXEC_PREFIX
15344 @findex GCC_EXEC_PREFIX
15345 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
15346 names of the subprograms executed by the compiler. No slash is added
15347 when this prefix is combined with the name of a subprogram, but you can
15348 specify a prefix that ends with a slash if you wish.
15350 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
15351 an appropriate prefix to use based on the pathname it was invoked with.
15353 If GCC cannot find the subprogram using the specified prefix, it
15354 tries looking in the usual places for the subprogram.
15356 The default value of @env{GCC_EXEC_PREFIX} is
15357 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
15358 the installed compiler. In many cases @var{prefix} is the value
15359 of @code{prefix} when you ran the @file{configure} script.
15361 Other prefixes specified with @option{-B} take precedence over this prefix.
15363 This prefix is also used for finding files such as @file{crt0.o} that are
15366 In addition, the prefix is used in an unusual way in finding the
15367 directories to search for header files. For each of the standard
15368 directories whose name normally begins with @samp{/usr/local/lib/gcc}
15369 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
15370 replacing that beginning with the specified prefix to produce an
15371 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
15372 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
15373 These alternate directories are searched first; the standard directories
15374 come next. If a standard directory begins with the configured
15375 @var{prefix} then the value of @var{prefix} is replaced by
15376 @env{GCC_EXEC_PREFIX} when looking for header files.
15378 @item COMPILER_PATH
15379 @findex COMPILER_PATH
15380 The value of @env{COMPILER_PATH} is a colon-separated list of
15381 directories, much like @env{PATH}. GCC tries the directories thus
15382 specified when searching for subprograms, if it can't find the
15383 subprograms using @env{GCC_EXEC_PREFIX}.
15386 @findex LIBRARY_PATH
15387 The value of @env{LIBRARY_PATH} is a colon-separated list of
15388 directories, much like @env{PATH}. When configured as a native compiler,
15389 GCC tries the directories thus specified when searching for special
15390 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
15391 using GCC also uses these directories when searching for ordinary
15392 libraries for the @option{-l} option (but directories specified with
15393 @option{-L} come first).
15397 @cindex locale definition
15398 This variable is used to pass locale information to the compiler. One way in
15399 which this information is used is to determine the character set to be used
15400 when character literals, string literals and comments are parsed in C and C++.
15401 When the compiler is configured to allow multibyte characters,
15402 the following values for @env{LANG} are recognized:
15406 Recognize JIS characters.
15408 Recognize SJIS characters.
15410 Recognize EUCJP characters.
15413 If @env{LANG} is not defined, or if it has some other value, then the
15414 compiler will use mblen and mbtowc as defined by the default locale to
15415 recognize and translate multibyte characters.
15419 Some additional environments variables affect the behavior of the
15422 @include cppenv.texi
15426 @node Precompiled Headers
15427 @section Using Precompiled Headers
15428 @cindex precompiled headers
15429 @cindex speed of compilation
15431 Often large projects have many header files that are included in every
15432 source file. The time the compiler takes to process these header files
15433 over and over again can account for nearly all of the time required to
15434 build the project. To make builds faster, GCC allows users to
15435 `precompile' a header file; then, if builds can use the precompiled
15436 header file they will be much faster.
15438 To create a precompiled header file, simply compile it as you would any
15439 other file, if necessary using the @option{-x} option to make the driver
15440 treat it as a C or C++ header file. You will probably want to use a
15441 tool like @command{make} to keep the precompiled header up-to-date when
15442 the headers it contains change.
15444 A precompiled header file will be searched for when @code{#include} is
15445 seen in the compilation. As it searches for the included file
15446 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
15447 compiler looks for a precompiled header in each directory just before it
15448 looks for the include file in that directory. The name searched for is
15449 the name specified in the @code{#include} with @samp{.gch} appended. If
15450 the precompiled header file can't be used, it is ignored.
15452 For instance, if you have @code{#include "all.h"}, and you have
15453 @file{all.h.gch} in the same directory as @file{all.h}, then the
15454 precompiled header file will be used if possible, and the original
15455 header will be used otherwise.
15457 Alternatively, you might decide to put the precompiled header file in a
15458 directory and use @option{-I} to ensure that directory is searched
15459 before (or instead of) the directory containing the original header.
15460 Then, if you want to check that the precompiled header file is always
15461 used, you can put a file of the same name as the original header in this
15462 directory containing an @code{#error} command.
15464 This also works with @option{-include}. So yet another way to use
15465 precompiled headers, good for projects not designed with precompiled
15466 header files in mind, is to simply take most of the header files used by
15467 a project, include them from another header file, precompile that header
15468 file, and @option{-include} the precompiled header. If the header files
15469 have guards against multiple inclusion, they will be skipped because
15470 they've already been included (in the precompiled header).
15472 If you need to precompile the same header file for different
15473 languages, targets, or compiler options, you can instead make a
15474 @emph{directory} named like @file{all.h.gch}, and put each precompiled
15475 header in the directory, perhaps using @option{-o}. It doesn't matter
15476 what you call the files in the directory, every precompiled header in
15477 the directory will be considered. The first precompiled header
15478 encountered in the directory that is valid for this compilation will
15479 be used; they're searched in no particular order.
15481 There are many other possibilities, limited only by your imagination,
15482 good sense, and the constraints of your build system.
15484 A precompiled header file can be used only when these conditions apply:
15488 Only one precompiled header can be used in a particular compilation.
15491 A precompiled header can't be used once the first C token is seen. You
15492 can have preprocessor directives before a precompiled header; you can
15493 even include a precompiled header from inside another header, so long as
15494 there are no C tokens before the @code{#include}.
15497 The precompiled header file must be produced for the same language as
15498 the current compilation. You can't use a C precompiled header for a C++
15502 The precompiled header file must have been produced by the same compiler
15503 binary as the current compilation is using.
15506 Any macros defined before the precompiled header is included must
15507 either be defined in the same way as when the precompiled header was
15508 generated, or must not affect the precompiled header, which usually
15509 means that they don't appear in the precompiled header at all.
15511 The @option{-D} option is one way to define a macro before a
15512 precompiled header is included; using a @code{#define} can also do it.
15513 There are also some options that define macros implicitly, like
15514 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
15517 @item If debugging information is output when using the precompiled
15518 header, using @option{-g} or similar, the same kind of debugging information
15519 must have been output when building the precompiled header. However,
15520 a precompiled header built using @option{-g} can be used in a compilation
15521 when no debugging information is being output.
15523 @item The same @option{-m} options must generally be used when building
15524 and using the precompiled header. @xref{Submodel Options},
15525 for any cases where this rule is relaxed.
15527 @item Each of the following options must be the same when building and using
15528 the precompiled header:
15530 @gccoptlist{-fexceptions -funit-at-a-time}
15533 Some other command-line options starting with @option{-f},
15534 @option{-p}, or @option{-O} must be defined in the same way as when
15535 the precompiled header was generated. At present, it's not clear
15536 which options are safe to change and which are not; the safest choice
15537 is to use exactly the same options when generating and using the
15538 precompiled header. The following are known to be safe:
15540 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
15541 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
15542 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
15547 For all of these except the last, the compiler will automatically
15548 ignore the precompiled header if the conditions aren't met. If you
15549 find an option combination that doesn't work and doesn't cause the
15550 precompiled header to be ignored, please consider filing a bug report,
15553 If you do use differing options when generating and using the
15554 precompiled header, the actual behavior will be a mixture of the
15555 behavior for the options. For instance, if you use @option{-g} to
15556 generate the precompiled header but not when using it, you may or may
15557 not get debugging information for routines in the precompiled header.
15559 @node Running Protoize
15560 @section Running Protoize
15562 The program @code{protoize} is an optional part of GCC@. You can use
15563 it to add prototypes to a program, thus converting the program to ISO
15564 C in one respect. The companion program @code{unprotoize} does the
15565 reverse: it removes argument types from any prototypes that are found.
15567 When you run these programs, you must specify a set of source files as
15568 command line arguments. The conversion programs start out by compiling
15569 these files to see what functions they define. The information gathered
15570 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
15572 After scanning comes actual conversion. The specified files are all
15573 eligible to be converted; any files they include (whether sources or
15574 just headers) are eligible as well.
15576 But not all the eligible files are converted. By default,
15577 @code{protoize} and @code{unprotoize} convert only source and header
15578 files in the current directory. You can specify additional directories
15579 whose files should be converted with the @option{-d @var{directory}}
15580 option. You can also specify particular files to exclude with the
15581 @option{-x @var{file}} option. A file is converted if it is eligible, its
15582 directory name matches one of the specified directory names, and its
15583 name within the directory has not been excluded.
15585 Basic conversion with @code{protoize} consists of rewriting most
15586 function definitions and function declarations to specify the types of
15587 the arguments. The only ones not rewritten are those for varargs
15590 @code{protoize} optionally inserts prototype declarations at the
15591 beginning of the source file, to make them available for any calls that
15592 precede the function's definition. Or it can insert prototype
15593 declarations with block scope in the blocks where undeclared functions
15596 Basic conversion with @code{unprotoize} consists of rewriting most
15597 function declarations to remove any argument types, and rewriting
15598 function definitions to the old-style pre-ISO form.
15600 Both conversion programs print a warning for any function declaration or
15601 definition that they can't convert. You can suppress these warnings
15604 The output from @code{protoize} or @code{unprotoize} replaces the
15605 original source file. The original file is renamed to a name ending
15606 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
15607 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
15608 for DOS) file already exists, then the source file is simply discarded.
15610 @code{protoize} and @code{unprotoize} both depend on GCC itself to
15611 scan the program and collect information about the functions it uses.
15612 So neither of these programs will work until GCC is installed.
15614 Here is a table of the options you can use with @code{protoize} and
15615 @code{unprotoize}. Each option works with both programs unless
15619 @item -B @var{directory}
15620 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
15621 usual directory (normally @file{/usr/local/lib}). This file contains
15622 prototype information about standard system functions. This option
15623 applies only to @code{protoize}.
15625 @item -c @var{compilation-options}
15626 Use @var{compilation-options} as the options when running @command{gcc} to
15627 produce the @samp{.X} files. The special option @option{-aux-info} is
15628 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
15630 Note that the compilation options must be given as a single argument to
15631 @code{protoize} or @code{unprotoize}. If you want to specify several
15632 @command{gcc} options, you must quote the entire set of compilation options
15633 to make them a single word in the shell.
15635 There are certain @command{gcc} arguments that you cannot use, because they
15636 would produce the wrong kind of output. These include @option{-g},
15637 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
15638 the @var{compilation-options}, they are ignored.
15641 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
15642 systems) instead of @samp{.c}. This is convenient if you are converting
15643 a C program to C++. This option applies only to @code{protoize}.
15646 Add explicit global declarations. This means inserting explicit
15647 declarations at the beginning of each source file for each function
15648 that is called in the file and was not declared. These declarations
15649 precede the first function definition that contains a call to an
15650 undeclared function. This option applies only to @code{protoize}.
15652 @item -i @var{string}
15653 Indent old-style parameter declarations with the string @var{string}.
15654 This option applies only to @code{protoize}.
15656 @code{unprotoize} converts prototyped function definitions to old-style
15657 function definitions, where the arguments are declared between the
15658 argument list and the initial @samp{@{}. By default, @code{unprotoize}
15659 uses five spaces as the indentation. If you want to indent with just
15660 one space instead, use @option{-i " "}.
15663 Keep the @samp{.X} files. Normally, they are deleted after conversion
15667 Add explicit local declarations. @code{protoize} with @option{-l} inserts
15668 a prototype declaration for each function in each block which calls the
15669 function without any declaration. This option applies only to
15673 Make no real changes. This mode just prints information about the conversions
15674 that would have been done without @option{-n}.
15677 Make no @samp{.save} files. The original files are simply deleted.
15678 Use this option with caution.
15680 @item -p @var{program}
15681 Use the program @var{program} as the compiler. Normally, the name
15682 @file{gcc} is used.
15685 Work quietly. Most warnings are suppressed.
15688 Print the version number, just like @option{-v} for @command{gcc}.
15691 If you need special compiler options to compile one of your program's
15692 source files, then you should generate that file's @samp{.X} file
15693 specially, by running @command{gcc} on that source file with the
15694 appropriate options and the option @option{-aux-info}. Then run
15695 @code{protoize} on the entire set of files. @code{protoize} will use
15696 the existing @samp{.X} file because it is newer than the source file.
15700 gcc -Dfoo=bar file1.c -aux-info file1.X
15705 You need to include the special files along with the rest in the
15706 @code{protoize} command, even though their @samp{.X} files already
15707 exist, because otherwise they won't get converted.
15709 @xref{Protoize Caveats}, for more information on how to use
15710 @code{protoize} successfully.