1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007
3 @c Free Software Foundation, Inc.
4 @c This is part of the GCC manual.
5 @c For copying conditions, see the file gcc.texi.
12 @c man begin COPYRIGHT
13 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
14 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
16 Permission is granted to copy, distribute and/or modify this document
17 under the terms of the GNU Free Documentation License, Version 1.2 or
18 any later version published by the Free Software Foundation; with the
19 Invariant Sections being ``GNU General Public License'' and ``Funding
20 Free Software'', the Front-Cover texts being (a) (see below), and with
21 the Back-Cover Texts being (b) (see below). A copy of the license is
22 included in the gfdl(7) man page.
24 (a) The FSF's Front-Cover Text is:
28 (b) The FSF's Back-Cover Text is:
30 You have freedom to copy and modify this GNU Manual, like GNU
31 software. Copies published by the Free Software Foundation raise
32 funds for GNU development.
34 @c Set file name and title for the man page.
36 @settitle GNU project C and C++ compiler
38 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
39 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
40 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
41 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
42 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
43 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
44 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
46 Only the most useful options are listed here; see below for the
47 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
50 gpl(7), gfdl(7), fsf-funding(7),
51 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
52 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
53 @file{ld}, @file{binutils} and @file{gdb}.
56 For instructions on reporting bugs, see
57 @w{@uref{http://gcc.gnu.org/bugs.html}}.
60 See the Info entry for @command{gcc}, or
61 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
62 for contributors to GCC@.
67 @chapter GCC Command Options
68 @cindex GCC command options
69 @cindex command options
70 @cindex options, GCC command
72 @c man begin DESCRIPTION
73 When you invoke GCC, it normally does preprocessing, compilation,
74 assembly and linking. The ``overall options'' allow you to stop this
75 process at an intermediate stage. For example, the @option{-c} option
76 says not to run the linker. Then the output consists of object files
77 output by the assembler.
79 Other options are passed on to one stage of processing. Some options
80 control the preprocessor and others the compiler itself. Yet other
81 options control the assembler and linker; most of these are not
82 documented here, since you rarely need to use any of them.
84 @cindex C compilation options
85 Most of the command line options that you can use with GCC are useful
86 for C programs; when an option is only useful with another language
87 (usually C++), the explanation says so explicitly. If the description
88 for a particular option does not mention a source language, you can use
89 that option with all supported languages.
91 @cindex C++ compilation options
92 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
93 options for compiling C++ programs.
95 @cindex grouping options
96 @cindex options, grouping
97 The @command{gcc} program accepts options and file names as operands. Many
98 options have multi-letter names; therefore multiple single-letter options
99 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
102 @cindex order of options
103 @cindex options, order
104 You can mix options and other arguments. For the most part, the order
105 you use doesn't matter. Order does matter when you use several options
106 of the same kind; for example, if you specify @option{-L} more than once,
107 the directories are searched in the order specified.
109 Many options have long names starting with @samp{-f} or with
110 @samp{-W}---for example,
111 @option{-fmove-loop-invariants}, @option{-Wformat} and so on. Most of
112 these have both positive and negative forms; the negative form of
113 @option{-ffoo} would be @option{-fno-foo}. This manual documents
114 only one of these two forms, whichever one is not the default.
118 @xref{Option Index}, for an index to GCC's options.
121 * Option Summary:: Brief list of all options, without explanations.
122 * Overall Options:: Controlling the kind of output:
123 an executable, object files, assembler files,
124 or preprocessed source.
125 * Invoking G++:: Compiling C++ programs.
126 * C Dialect Options:: Controlling the variant of C language compiled.
127 * C++ Dialect Options:: Variations on C++.
128 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
130 * Language Independent Options:: Controlling how diagnostics should be
132 * Warning Options:: How picky should the compiler be?
133 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
134 * Optimize Options:: How much optimization?
135 * Preprocessor Options:: Controlling header files and macro definitions.
136 Also, getting dependency information for Make.
137 * Assembler Options:: Passing options to the assembler.
138 * Link Options:: Specifying libraries and so on.
139 * Directory Options:: Where to find header files and libraries.
140 Where to find the compiler executable files.
141 * Spec Files:: How to pass switches to sub-processes.
142 * Target Options:: Running a cross-compiler, or an old version of GCC.
143 * Submodel Options:: Specifying minor hardware or convention variations,
144 such as 68010 vs 68020.
145 * Code Gen Options:: Specifying conventions for function calls, data layout
147 * Environment Variables:: Env vars that affect GCC.
148 * Precompiled Headers:: Compiling a header once, and using it many times.
149 * Running Protoize:: Automatically adding or removing function prototypes.
155 @section Option Summary
157 Here is a summary of all the options, grouped by type. Explanations are
158 in the following sections.
161 @item Overall Options
162 @xref{Overall Options,,Options Controlling the Kind of Output}.
163 @gccoptlist{-c -S -E -o @var{file} -combine -pipe -pass-exit-codes @gol
164 -x @var{language} -v -### --help@r{[}=@var{class}@r{]} --target-help @gol
165 --version @@@var{file}}
167 @item C Language Options
168 @xref{C Dialect Options,,Options Controlling C Dialect}.
169 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
170 -aux-info @var{filename} @gol
171 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
172 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
173 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
174 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
175 -fsigned-bitfields -fsigned-char @gol
176 -funsigned-bitfields -funsigned-char}
178 @item C++ Language Options
179 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
180 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
181 -fconserve-space -ffriend-injection @gol
182 -fno-elide-constructors @gol
183 -fno-enforce-eh-specs @gol
184 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
185 -fno-implicit-templates @gol
186 -fno-implicit-inline-templates @gol
187 -fno-implement-inlines -fms-extensions @gol
188 -fno-nonansi-builtins -fno-operator-names @gol
189 -fno-optional-diags -fpermissive @gol
190 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
191 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
192 -fno-default-inline -fvisibility-inlines-hidden @gol
193 -Wabi -Wctor-dtor-privacy @gol
194 -Wnon-virtual-dtor -Wreorder @gol
195 -Weffc++ -Wno-deprecated -Wstrict-null-sentinel @gol
196 -Wno-non-template-friend -Wold-style-cast @gol
197 -Woverloaded-virtual -Wno-pmf-conversions @gol
200 @item Objective-C and Objective-C++ Language Options
201 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
202 Objective-C and Objective-C++ Dialects}.
203 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
204 -fgnu-runtime -fnext-runtime @gol
205 -fno-nil-receivers @gol
206 -fobjc-call-cxx-cdtors @gol
207 -fobjc-direct-dispatch @gol
208 -fobjc-exceptions @gol
210 -freplace-objc-classes @gol
213 -Wassign-intercept @gol
214 -Wno-protocol -Wselector @gol
215 -Wstrict-selector-match @gol
216 -Wundeclared-selector}
218 @item Language Independent Options
219 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
220 @gccoptlist{-fmessage-length=@var{n} @gol
221 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
222 -fdiagnostics-show-option}
224 @item Warning Options
225 @xref{Warning Options,,Options to Request or Suppress Warnings}.
226 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
227 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
228 -Wno-attributes -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
229 -Wchar-subscripts -Wclobbered -Wcomment @gol
230 -Wconversion -Wcoverage-mismatch -Wno-deprecated-declarations @gol
231 -Wdisabled-optimization -Wno-div-by-zero @gol
232 -Wempty-body -Wno-endif-labels @gol
233 -Werror -Werror=* @gol
234 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
235 -Wno-format-extra-args -Wformat-nonliteral @gol
236 -Wformat-security -Wformat-y2k @gol
237 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
238 -Wimport -Wno-import -Winit-self -Winline @gol
239 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
240 -Winvalid-pch -Wlarger-than-@var{len} -Wunsafe-loop-optimizations @gol
241 -Wlogical-op -Wlong-long @gol
242 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
243 -Wmissing-format-attribute -Wmissing-include-dirs @gol
244 -Wmissing-noreturn @gol
245 -Wno-multichar -Wnonnull -Wno-overflow @gol
246 -Woverlength-strings -Wpacked -Wpadded @gol
247 -Wparentheses -Wpointer-arith -Wno-pointer-to-int-cast @gol
248 -Wredundant-decls @gol
249 -Wreturn-type -Wsequence-point -Wshadow @gol
250 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
251 -Wstrict-aliasing -Wstrict-aliasing=n @gol
252 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
253 -Wswitch -Wswitch-default -Wswitch-enum @gol
254 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
255 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
256 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
257 -Wunused-value -Wunused-variable @gol
258 -Wvariadic-macros -Wvla @gol
259 -Wvolatile-register-var -Wwrite-strings}
261 @item C-only Warning Options
262 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
263 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
264 -Wold-style-declaration -Wold-style-definition @gol
265 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
266 -Wdeclaration-after-statement -Wpointer-sign}
268 @item Debugging Options
269 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
270 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
271 -fdump-noaddr -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
272 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
273 -fdump-ipa-all -fdump-ipa-cgraph @gol
275 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
276 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
277 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
278 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
280 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
281 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
282 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
283 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
284 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
285 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
287 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
288 -fdump-tree-nrv -fdump-tree-vect @gol
289 -fdump-tree-sink @gol
290 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
291 -fdump-tree-salias @gol
292 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
293 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
294 -ftree-vectorizer-verbose=@var{n} @gol
295 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
296 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
297 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
298 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
299 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
300 -ftest-coverage -ftime-report -fvar-tracking @gol
301 -g -g@var{level} -gcoff -gdwarf-2 @gol
302 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
303 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
304 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
305 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
306 -print-multi-directory -print-multi-lib @gol
307 -print-prog-name=@var{program} -print-search-dirs -Q @gol
308 -print-sysroot-headers-suffix @gol
311 @item Optimization Options
312 @xref{Optimize Options,,Options that Control Optimization}.
313 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
314 -falign-labels=@var{n} -falign-loops=@var{n} @gol
315 -fbounds-check -fmudflap -fmudflapth -fmudflapir @gol
316 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
317 -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol
318 -fcaller-saves -fcprop-registers -fcse-follow-jumps @gol
319 -fcse-skip-blocks -fcx-limited-range -fdata-sections @gol
320 -fdelayed-branch -fdelete-null-pointer-checks -fearly-inlining @gol
321 -fexpensive-optimizations -ffast-math -ffloat-store @gol
322 -fforce-addr -fforward-propagate -ffunction-sections @gol
323 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -fgcse-after-reload @gol
324 -fcrossjumping -fif-conversion -fif-conversion2 @gol
325 -finline-functions -finline-functions-called-once @gol
326 -finline-limit=@var{n} -fkeep-inline-functions @gol
327 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
328 -fmodulo-sched -fno-branch-count-reg @gol
329 -fno-default-inline -fno-defer-pop -fmove-loop-invariants @gol
330 -fno-function-cse -fno-guess-branch-probability @gol
331 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
332 -funsafe-math-optimizations -funsafe-loop-optimizations @gol
333 -ffinite-math-only -fno-signed-zeros @gol
334 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
335 -fomit-frame-pointer -foptimize-register-move @gol
336 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
337 -fprofile-generate -fprofile-use @gol
338 -fregmove -frename-registers @gol
339 -freorder-blocks -freorder-blocks-and-partition -freorder-functions @gol
340 -frerun-cse-after-loop @gol
341 -frounding-math -frtl-abstract-sequences @gol
342 -fschedule-insns -fschedule-insns2 @gol
343 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
344 -fsched-spec-load-dangerous @gol
345 -fsched-stalled-insns=@var{n} -fsched-stalled-insns-dep=@var{n} @gol
346 -fsched2-use-superblocks @gol
347 -fsched2-use-traces -fsee -freschedule-modulo-scheduled-loops @gol
348 -fsection-anchors -fsignaling-nans -fsingle-precision-constant @gol
349 -fno-split-wide-types -fstack-protector -fstack-protector-all @gol
350 -fstrict-aliasing -fstrict-overflow -ftracer -fthread-jumps @gol
351 -funroll-all-loops -funroll-loops -fpeel-loops @gol
352 -fsplit-ivs-in-unroller -funswitch-loops @gol
353 -fvariable-expansion-in-unroller @gol
354 -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol
355 -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts @gol
356 -fcheck-data-deps @gol
357 -ftree-dominator-opts -ftree-dse -ftree-copyrename -ftree-sink @gol
358 -ftree-ch -ftree-sra -ftree-ter -ftree-fre -ftree-vectorize @gol
359 -ftree-vect-loop-version -ftree-salias -fipa-pta -fweb @gol
360 -ftree-copy-prop -ftree-store-ccp -ftree-store-copy-prop -fwhole-program @gol
361 --param @var{name}=@var{value}
362 -O -O0 -O1 -O2 -O3 -Os}
364 @item Preprocessor Options
365 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
366 @gccoptlist{-A@var{question}=@var{answer} @gol
367 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
368 -C -dD -dI -dM -dN @gol
369 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
370 -idirafter @var{dir} @gol
371 -include @var{file} -imacros @var{file} @gol
372 -iprefix @var{file} -iwithprefix @var{dir} @gol
373 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
374 -imultilib @var{dir} -isysroot @var{dir} @gol
375 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
376 -P -fworking-directory -remap @gol
377 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
378 -Xpreprocessor @var{option}}
380 @item Assembler Option
381 @xref{Assembler Options,,Passing Options to the Assembler}.
382 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
385 @xref{Link Options,,Options for Linking}.
386 @gccoptlist{@var{object-file-name} -l@var{library} @gol
387 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
388 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
389 -Wl,@var{option} -Xlinker @var{option} @gol
392 @item Directory Options
393 @xref{Directory Options,,Options for Directory Search}.
394 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
395 -specs=@var{file} -I- --sysroot=@var{dir}}
398 @c I wrote this xref this way to avoid overfull hbox. -- rms
399 @xref{Target Options}.
400 @gccoptlist{-V @var{version} -b @var{machine}}
402 @item Machine Dependent Options
403 @xref{Submodel Options,,Hardware Models and Configurations}.
404 @c This list is ordered alphanumerically by subsection name.
405 @c Try and put the significant identifier (CPU or system) first,
406 @c so users have a clue at guessing where the ones they want will be.
409 @gccoptlist{-EB -EL @gol
410 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
411 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
414 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
415 -mabi=@var{name} @gol
416 -mapcs-stack-check -mno-apcs-stack-check @gol
417 -mapcs-float -mno-apcs-float @gol
418 -mapcs-reentrant -mno-apcs-reentrant @gol
419 -msched-prolog -mno-sched-prolog @gol
420 -mlittle-endian -mbig-endian -mwords-little-endian @gol
421 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
422 -mthumb-interwork -mno-thumb-interwork @gol
423 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
424 -mstructure-size-boundary=@var{n} @gol
425 -mabort-on-noreturn @gol
426 -mlong-calls -mno-long-calls @gol
427 -msingle-pic-base -mno-single-pic-base @gol
428 -mpic-register=@var{reg} @gol
429 -mnop-fun-dllimport @gol
430 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
431 -mpoke-function-name @gol
433 -mtpcs-frame -mtpcs-leaf-frame @gol
434 -mcaller-super-interworking -mcallee-super-interworking @gol
438 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
439 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
441 @emph{Blackfin Options}
442 @gccoptlist{-mcpu=@var{cpu} -msim -momit-leaf-frame-pointer @gol
443 -mno-omit-leaf-frame-pointer -mspecld-anomaly -mno-specld-anomaly @gol
444 -mcsync-anomaly -mno-csync-anomaly -mlow-64k -mno-low64k @gol
445 -mstack-check-l1 -mid-shared-library -mno-id-shared-library @gol
446 -mshared-library-id=@var{n} -mleaf-id-shared-library @gol
447 -mno-leaf-id-shared-library -msep-data -mno-sep-data -mlong-calls @gol
451 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
452 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
453 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
454 -mstack-align -mdata-align -mconst-align @gol
455 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
456 -melf -maout -melinux -mlinux -sim -sim2 @gol
457 -mmul-bug-workaround -mno-mul-bug-workaround}
460 @gccoptlist{-mmac -mpush-args}
462 @emph{Darwin Options}
463 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
464 -arch_only -bind_at_load -bundle -bundle_loader @gol
465 -client_name -compatibility_version -current_version @gol
467 -dependency-file -dylib_file -dylinker_install_name @gol
468 -dynamic -dynamiclib -exported_symbols_list @gol
469 -filelist -flat_namespace -force_cpusubtype_ALL @gol
470 -force_flat_namespace -headerpad_max_install_names @gol
472 -image_base -init -install_name -keep_private_externs @gol
473 -multi_module -multiply_defined -multiply_defined_unused @gol
474 -noall_load -no_dead_strip_inits_and_terms @gol
475 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
476 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
477 -private_bundle -read_only_relocs -sectalign @gol
478 -sectobjectsymbols -whyload -seg1addr @gol
479 -sectcreate -sectobjectsymbols -sectorder @gol
480 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
481 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
482 -segprot -segs_read_only_addr -segs_read_write_addr @gol
483 -single_module -static -sub_library -sub_umbrella @gol
484 -twolevel_namespace -umbrella -undefined @gol
485 -unexported_symbols_list -weak_reference_mismatches @gol
486 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
487 -mkernel -mone-byte-bool}
489 @emph{DEC Alpha Options}
490 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
491 -mieee -mieee-with-inexact -mieee-conformant @gol
492 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
493 -mtrap-precision=@var{mode} -mbuild-constants @gol
494 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
495 -mbwx -mmax -mfix -mcix @gol
496 -mfloat-vax -mfloat-ieee @gol
497 -mexplicit-relocs -msmall-data -mlarge-data @gol
498 -msmall-text -mlarge-text @gol
499 -mmemory-latency=@var{time}}
501 @emph{DEC Alpha/VMS Options}
502 @gccoptlist{-mvms-return-codes}
505 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
506 -mhard-float -msoft-float @gol
507 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
508 -mdouble -mno-double @gol
509 -mmedia -mno-media -mmuladd -mno-muladd @gol
510 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
511 -mlinked-fp -mlong-calls -malign-labels @gol
512 -mlibrary-pic -macc-4 -macc-8 @gol
513 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
514 -moptimize-membar -mno-optimize-membar @gol
515 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
516 -mvliw-branch -mno-vliw-branch @gol
517 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
518 -mno-nested-cond-exec -mtomcat-stats @gol
522 @emph{GNU/Linux Options}
523 @gccoptlist{-muclibc}
525 @emph{H8/300 Options}
526 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
529 @gccoptlist{-march=@var{architecture-type} @gol
530 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
531 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
532 -mfixed-range=@var{register-range} @gol
533 -mjump-in-delay -mlinker-opt -mlong-calls @gol
534 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
535 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
536 -mno-jump-in-delay -mno-long-load-store @gol
537 -mno-portable-runtime -mno-soft-float @gol
538 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
539 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
540 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
541 -munix=@var{unix-std} -nolibdld -static -threads}
543 @emph{i386 and x86-64 Options}
544 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
545 -mfpmath=@var{unit} @gol
546 -masm=@var{dialect} -mno-fancy-math-387 @gol
547 -mno-fp-ret-in-387 -msoft-float @gol
548 -mno-wide-multiply -mrtd -malign-double @gol
549 -mpreferred-stack-boundary=@var{num} -mcx16 -msahf @gol
550 -mmmx -msse -msse2 -msse3 -mssse3 -msse4a -m3dnow -mpopcnt -mabm @gol
551 -mthreads -mno-align-stringops -minline-all-stringops @gol
552 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
553 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
554 -mpc32 -mpc64 -mpc80 mstackrealign @gol
555 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
556 -mcmodel=@var{code-model} @gol
557 -m32 -m64 -mlarge-data-threshold=@var{num}}
560 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
561 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
562 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
563 -minline-float-divide-max-throughput @gol
564 -minline-int-divide-min-latency @gol
565 -minline-int-divide-max-throughput @gol
566 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
567 -mno-dwarf2-asm -mearly-stop-bits @gol
568 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
569 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64 @gol
570 -mno-sched-br-data-spec -msched-ar-data-spec -mno-sched-control-spec @gol
571 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
572 -msched-ldc -mno-sched-control-ldc -mno-sched-spec-verbose @gol
573 -mno-sched-prefer-non-data-spec-insns @gol
574 -mno-sched-prefer-non-control-spec-insns @gol
575 -mno-sched-count-spec-in-critical-path}
577 @emph{M32R/D Options}
578 @gccoptlist{-m32r2 -m32rx -m32r @gol
580 -malign-loops -mno-align-loops @gol
581 -missue-rate=@var{number} @gol
582 -mbranch-cost=@var{number} @gol
583 -mmodel=@var{code-size-model-type} @gol
584 -msdata=@var{sdata-type} @gol
585 -mno-flush-func -mflush-func=@var{name} @gol
586 -mno-flush-trap -mflush-trap=@var{number} @gol
590 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
592 @emph{M680x0 Options}
593 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
594 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
595 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
596 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
597 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
598 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
599 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
600 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
602 @emph{M68hc1x Options}
603 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
604 -mauto-incdec -minmax -mlong-calls -mshort @gol
605 -msoft-reg-count=@var{count}}
608 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
609 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
610 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
611 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
612 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
615 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
616 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
617 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
618 -mshared -mno-shared -mxgot -mno-xgot -mgp32 -mgp64 @gol
619 -mfp32 -mfp64 -mhard-float -msoft-float @gol
620 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
621 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
622 -mips3d -mno-mips3d -mmt -mno-mt @gol
623 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
624 -G@var{num} -membedded-data -mno-embedded-data @gol
625 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
626 -msplit-addresses -mno-split-addresses @gol
627 -mexplicit-relocs -mno-explicit-relocs @gol
628 -mcheck-zero-division -mno-check-zero-division @gol
629 -mdivide-traps -mdivide-breaks @gol
630 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
631 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
632 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
633 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 -mno-fix-vr4130 @gol
634 -mfix-sb1 -mno-fix-sb1 @gol
635 -mflush-func=@var{func} -mno-flush-func @gol
636 -mbranch-likely -mno-branch-likely @gol
637 -mfp-exceptions -mno-fp-exceptions @gol
638 -mvr4130-align -mno-vr4130-align}
641 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
642 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
643 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
644 -mno-base-addresses -msingle-exit -mno-single-exit}
646 @emph{MN10300 Options}
647 @gccoptlist{-mmult-bug -mno-mult-bug @gol
648 -mam33 -mno-am33 @gol
649 -mam33-2 -mno-am33-2 @gol
650 -mreturn-pointer-on-d0 @gol
654 @gccoptlist{-mno-crt0 -mbacc -msim @gol
655 -march=@var{cpu-type} }
657 @emph{PDP-11 Options}
658 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
659 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
660 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
661 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
662 -mbranch-expensive -mbranch-cheap @gol
663 -msplit -mno-split -munix-asm -mdec-asm}
665 @emph{PowerPC Options}
666 See RS/6000 and PowerPC Options.
668 @emph{RS/6000 and PowerPC Options}
669 @gccoptlist{-mcpu=@var{cpu-type} @gol
670 -mtune=@var{cpu-type} @gol
671 -mpower -mno-power -mpower2 -mno-power2 @gol
672 -mpowerpc -mpowerpc64 -mno-powerpc @gol
673 -maltivec -mno-altivec @gol
674 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
675 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
676 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
677 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mdfp -mno-dfp @gol
678 -mnew-mnemonics -mold-mnemonics @gol
679 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
680 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
681 -malign-power -malign-natural @gol
682 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
683 -mstring -mno-string -mupdate -mno-update @gol
684 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
685 -mstrict-align -mno-strict-align -mrelocatable @gol
686 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
687 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
688 -mdynamic-no-pic -maltivec -mswdiv @gol
689 -mprioritize-restricted-insns=@var{priority} @gol
690 -msched-costly-dep=@var{dependence_type} @gol
691 -minsert-sched-nops=@var{scheme} @gol
692 -mcall-sysv -mcall-netbsd @gol
693 -maix-struct-return -msvr4-struct-return @gol
694 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
695 -misel -mno-isel @gol
696 -misel=yes -misel=no @gol
698 -mspe=yes -mspe=no @gol
699 -mvrsave -mno-vrsave @gol
700 -mmulhw -mno-mulhw @gol
701 -mdlmzb -mno-dlmzb @gol
702 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
703 -mprototype -mno-prototype @gol
704 -msim -mmvme -mads -myellowknife -memb -msdata @gol
705 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
707 @emph{S/390 and zSeries Options}
708 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
709 -mhard-float -msoft-float -mlong-double-64 -mlong-double-128 @gol
710 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
711 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
712 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
713 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
714 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
717 @gccoptlist{-meb -mel @gol
721 -mscore5 -mscore5u -mscore7 -mscore7d}
724 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
725 -m4-nofpu -m4-single-only -m4-single -m4 @gol
726 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
727 -m5-64media -m5-64media-nofpu @gol
728 -m5-32media -m5-32media-nofpu @gol
729 -m5-compact -m5-compact-nofpu @gol
730 -mb -ml -mdalign -mrelax @gol
731 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
732 -mieee -misize -minline-ic_invalidate -mpadstruct -mspace @gol
733 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
734 -mdivsi3_libfunc=@var{name} @gol
735 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
739 @gccoptlist{-mcpu=@var{cpu-type} @gol
740 -mtune=@var{cpu-type} @gol
741 -mcmodel=@var{code-model} @gol
742 -m32 -m64 -mapp-regs -mno-app-regs @gol
743 -mfaster-structs -mno-faster-structs @gol
744 -mfpu -mno-fpu -mhard-float -msoft-float @gol
745 -mhard-quad-float -msoft-quad-float @gol
746 -mimpure-text -mno-impure-text -mlittle-endian @gol
747 -mstack-bias -mno-stack-bias @gol
748 -munaligned-doubles -mno-unaligned-doubles @gol
749 -mv8plus -mno-v8plus -mvis -mno-vis
750 -threads -pthreads -pthread}
753 @gccoptlist{-mwarn-reloc -merror-reloc @gol
754 -msafe-dma -munsafe-dma @gol
756 -msmall-mem -mlarge-mem -mstdmain @gol
757 -mfixed-range=@var{register-range}}
759 @emph{System V Options}
760 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
762 @emph{TMS320C3x/C4x Options}
763 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
764 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
765 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
766 -mparallel-insns -mparallel-mpy -mpreserve-float}
769 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
770 -mprolog-function -mno-prolog-function -mspace @gol
771 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
772 -mapp-regs -mno-app-regs @gol
773 -mdisable-callt -mno-disable-callt @gol
779 @gccoptlist{-mg -mgnu -munix}
781 @emph{VxWorks Options}
782 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
783 -Xbind-lazy -Xbind-now}
785 @emph{x86-64 Options}
786 See i386 and x86-64 Options.
788 @emph{Xstormy16 Options}
791 @emph{Xtensa Options}
792 @gccoptlist{-mconst16 -mno-const16 @gol
793 -mfused-madd -mno-fused-madd @gol
794 -mtext-section-literals -mno-text-section-literals @gol
795 -mtarget-align -mno-target-align @gol
796 -mlongcalls -mno-longcalls}
798 @emph{zSeries Options}
799 See S/390 and zSeries Options.
801 @item Code Generation Options
802 @xref{Code Gen Options,,Options for Code Generation Conventions}.
803 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
804 -ffixed-@var{reg} -fexceptions @gol
805 -fnon-call-exceptions -funwind-tables @gol
806 -fasynchronous-unwind-tables @gol
807 -finhibit-size-directive -finstrument-functions @gol
808 -fno-common -fno-ident @gol
809 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
810 -fno-jump-tables @gol
811 -frecord-gcc-switches @gol
812 -freg-struct-return -fshort-enums @gol
813 -fshort-double -fshort-wchar @gol
814 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
815 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
816 -fargument-alias -fargument-noalias @gol
817 -fargument-noalias-global -fargument-noalias-anything
818 -fleading-underscore -ftls-model=@var{model} @gol
819 -ftrapv -fwrapv -fbounds-check @gol
824 * Overall Options:: Controlling the kind of output:
825 an executable, object files, assembler files,
826 or preprocessed source.
827 * C Dialect Options:: Controlling the variant of C language compiled.
828 * C++ Dialect Options:: Variations on C++.
829 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
831 * Language Independent Options:: Controlling how diagnostics should be
833 * Warning Options:: How picky should the compiler be?
834 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
835 * Optimize Options:: How much optimization?
836 * Preprocessor Options:: Controlling header files and macro definitions.
837 Also, getting dependency information for Make.
838 * Assembler Options:: Passing options to the assembler.
839 * Link Options:: Specifying libraries and so on.
840 * Directory Options:: Where to find header files and libraries.
841 Where to find the compiler executable files.
842 * Spec Files:: How to pass switches to sub-processes.
843 * Target Options:: Running a cross-compiler, or an old version of GCC.
846 @node Overall Options
847 @section Options Controlling the Kind of Output
849 Compilation can involve up to four stages: preprocessing, compilation
850 proper, assembly and linking, always in that order. GCC is capable of
851 preprocessing and compiling several files either into several
852 assembler input files, or into one assembler input file; then each
853 assembler input file produces an object file, and linking combines all
854 the object files (those newly compiled, and those specified as input)
855 into an executable file.
857 @cindex file name suffix
858 For any given input file, the file name suffix determines what kind of
863 C source code which must be preprocessed.
866 C source code which should not be preprocessed.
869 C++ source code which should not be preprocessed.
872 Objective-C source code. Note that you must link with the @file{libobjc}
873 library to make an Objective-C program work.
876 Objective-C source code which should not be preprocessed.
880 Objective-C++ source code. Note that you must link with the @file{libobjc}
881 library to make an Objective-C++ program work. Note that @samp{.M} refers
882 to a literal capital M@.
885 Objective-C++ source code which should not be preprocessed.
888 C, C++, Objective-C or Objective-C++ header file to be turned into a
893 @itemx @var{file}.cxx
894 @itemx @var{file}.cpp
895 @itemx @var{file}.CPP
896 @itemx @var{file}.c++
898 C++ source code which must be preprocessed. Note that in @samp{.cxx},
899 the last two letters must both be literally @samp{x}. Likewise,
900 @samp{.C} refers to a literal capital C@.
904 Objective-C++ source code which must be preprocessed.
907 Objective-C++ source code which should not be preprocessed.
911 C++ header file to be turned into a precompiled header.
914 @itemx @var{file}.for
915 @itemx @var{file}.FOR
916 Fixed form Fortran source code which should not be preprocessed.
919 @itemx @var{file}.fpp
920 @itemx @var{file}.FPP
921 Fixed form Fortran source code which must be preprocessed (with the traditional
925 @itemx @var{file}.f95
926 Free form Fortran source code which should not be preprocessed.
929 @itemx @var{file}.F95
930 Free form Fortran source code which must be preprocessed (with the
931 traditional preprocessor).
933 @c FIXME: Descriptions of Java file types.
940 Ada source code file which contains a library unit declaration (a
941 declaration of a package, subprogram, or generic, or a generic
942 instantiation), or a library unit renaming declaration (a package,
943 generic, or subprogram renaming declaration). Such files are also
946 @itemx @var{file}.adb
947 Ada source code file containing a library unit body (a subprogram or
948 package body). Such files are also called @dfn{bodies}.
950 @c GCC also knows about some suffixes for languages not yet included:
961 Assembler code which must be preprocessed.
964 An object file to be fed straight into linking.
965 Any file name with no recognized suffix is treated this way.
969 You can specify the input language explicitly with the @option{-x} option:
972 @item -x @var{language}
973 Specify explicitly the @var{language} for the following input files
974 (rather than letting the compiler choose a default based on the file
975 name suffix). This option applies to all following input files until
976 the next @option{-x} option. Possible values for @var{language} are:
978 c c-header c-cpp-output
979 c++ c++-header c++-cpp-output
980 objective-c objective-c-header objective-c-cpp-output
981 objective-c++ objective-c++-header objective-c++-cpp-output
982 assembler assembler-with-cpp
990 Turn off any specification of a language, so that subsequent files are
991 handled according to their file name suffixes (as they are if @option{-x}
992 has not been used at all).
994 @item -pass-exit-codes
995 @opindex pass-exit-codes
996 Normally the @command{gcc} program will exit with the code of 1 if any
997 phase of the compiler returns a non-success return code. If you specify
998 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
999 numerically highest error produced by any phase that returned an error
1000 indication. The C, C++, and Fortran frontends return 4, if an internal
1001 compiler error is encountered.
1004 If you only want some of the stages of compilation, you can use
1005 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1006 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1007 @command{gcc} is to stop. Note that some combinations (for example,
1008 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1013 Compile or assemble the source files, but do not link. The linking
1014 stage simply is not done. The ultimate output is in the form of an
1015 object file for each source file.
1017 By default, the object file name for a source file is made by replacing
1018 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1020 Unrecognized input files, not requiring compilation or assembly, are
1025 Stop after the stage of compilation proper; do not assemble. The output
1026 is in the form of an assembler code file for each non-assembler input
1029 By default, the assembler file name for a source file is made by
1030 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1032 Input files that don't require compilation are ignored.
1036 Stop after the preprocessing stage; do not run the compiler proper. The
1037 output is in the form of preprocessed source code, which is sent to the
1040 Input files which don't require preprocessing are ignored.
1042 @cindex output file option
1045 Place output in file @var{file}. This applies regardless to whatever
1046 sort of output is being produced, whether it be an executable file,
1047 an object file, an assembler file or preprocessed C code.
1049 If @option{-o} is not specified, the default is to put an executable
1050 file in @file{a.out}, the object file for
1051 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1052 assembler file in @file{@var{source}.s}, a precompiled header file in
1053 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1058 Print (on standard error output) the commands executed to run the stages
1059 of compilation. Also print the version number of the compiler driver
1060 program and of the preprocessor and the compiler proper.
1064 Like @option{-v} except the commands are not executed and all command
1065 arguments are quoted. This is useful for shell scripts to capture the
1066 driver-generated command lines.
1070 Use pipes rather than temporary files for communication between the
1071 various stages of compilation. This fails to work on some systems where
1072 the assembler is unable to read from a pipe; but the GNU assembler has
1077 If you are compiling multiple source files, this option tells the driver
1078 to pass all the source files to the compiler at once (for those
1079 languages for which the compiler can handle this). This will allow
1080 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1081 language for which this is supported is C@. If you pass source files for
1082 multiple languages to the driver, using this option, the driver will invoke
1083 the compiler(s) that support IMA once each, passing each compiler all the
1084 source files appropriate for it. For those languages that do not support
1085 IMA this option will be ignored, and the compiler will be invoked once for
1086 each source file in that language. If you use this option in conjunction
1087 with @option{-save-temps}, the compiler will generate multiple
1089 (one for each source file), but only one (combined) @file{.o} or
1094 Print (on the standard output) a description of the command line options
1095 understood by @command{gcc}. If the @option{-v} option is also specified
1096 then @option{--help} will also be passed on to the various processes
1097 invoked by @command{gcc}, so that they can display the command line options
1098 they accept. If the @option{-Wextra} option has also been specified
1099 (prior to the @option{--help} option), then command line options which
1100 have no documentation associated with them will also be displayed.
1103 @opindex target-help
1104 Print (on the standard output) a description of target-specific command
1105 line options for each tool.
1107 @item --help=@var{class}@r{[},@var{qualifier}@r{]}
1108 Print (on the standard output) a description of the command line
1109 options understood by the compiler that fit into a specific class.
1110 The class can be one of @samp{optimizers}, @samp{warnings}, @samp{target},
1111 @samp{params}, or @var{language}:
1114 @item @samp{optimizers}
1115 This will display all of the optimization options supported by the
1118 @item @samp{warnings}
1119 This will display all of the options controlling warning messages
1120 produced by the compiler.
1123 This will display target-specific options. Unlike the
1124 @option{--target-help} option however, target-specific options of the
1125 linker and assembler will not be displayed. This is because those
1126 tools do not currently support the extended @option{--help=} syntax.
1129 This will display the values recognized by the @option{--param}
1132 @item @var{language}
1133 This will display the options supported for @var{language}, where
1134 @var{language} is the name of one of the languages supported in this
1138 This will display the options that are common to all languages.
1141 It is possible to further refine the output of the @option{--help=}
1142 option by adding a comma separated list of qualifiers after the
1143 class. These can be any from the following list:
1146 @item @samp{undocumented}
1147 Display only those options which are undocumented.
1150 Display options which take an argument that appears after an equal
1151 sign in the same continuous piece of text, such as:
1152 @samp{--help=target}.
1154 @item @samp{separate}
1155 Display options which take an argument that appears as a separate word
1156 following the original option, such as: @samp{-o output-file}.
1159 Thus for example to display all the undocumented target-specific
1160 switches supported by the compiler the following can be used:
1163 --help=target,undocumented
1166 The sense of a qualifier can be inverted by prefixing it with the
1167 @var{^} character, so for example to display all binary warning
1168 options (i.e. ones that are either on or off and that do not take an
1169 argument), which have a description the following can be used:
1172 --help=warnings,^joined,^undocumented
1175 A class can also be used as a qualifier, although this usually
1176 restricts the output by so much that there is nothing to display. One
1177 case where it does work however is when one of the classes is
1178 @var{target}. So for example to display all the target-specific
1179 optimization options the following can be used:
1182 --help=target,optimizers
1185 The @option{--help=} option can be repeated on the command line. Each
1186 successive use will display its requested class of options, skipping
1187 those that have already been displayed.
1189 If the @option{-Q} option appears on the command line before the
1190 @option{--help=} option, then the descriptive text displayed by
1191 @option{--help=} is changed. Instead of describing the displayed
1192 options, an indication is given as to whether the option is enabled,
1193 disabled or set to a specific value (assuming that the compiler
1194 knows this at the point where the @option{--help=} option is used).
1196 Here is a truncated example from the ARM port of @command{gcc}:
1199 % gcc -Q -mabi=2 --help=target -c
1200 The following options are target specific:
1202 -mabort-on-noreturn [disabled]
1206 The output is sensitive to the effects of previous command line
1207 options, so for example it is possible to find out which optimizations
1208 are enabled at @option{-O2} by using:
1211 -O2 --help=optimizers
1214 Alternatively you can discover which binary optimizations are enabled
1215 by @option{-O3} by using:
1218 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1219 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1220 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1225 Display the version number and copyrights of the invoked GCC@.
1227 @include @value{srcdir}/../libiberty/at-file.texi
1231 @section Compiling C++ Programs
1233 @cindex suffixes for C++ source
1234 @cindex C++ source file suffixes
1235 C++ source files conventionally use one of the suffixes @samp{.C},
1236 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1237 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
1238 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1239 files with these names and compiles them as C++ programs even if you
1240 call the compiler the same way as for compiling C programs (usually
1241 with the name @command{gcc}).
1245 However, the use of @command{gcc} does not add the C++ library.
1246 @command{g++} is a program that calls GCC and treats @samp{.c},
1247 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1248 files unless @option{-x} is used, and automatically specifies linking
1249 against the C++ library. This program is also useful when
1250 precompiling a C header file with a @samp{.h} extension for use in C++
1251 compilations. On many systems, @command{g++} is also installed with
1252 the name @command{c++}.
1254 @cindex invoking @command{g++}
1255 When you compile C++ programs, you may specify many of the same
1256 command-line options that you use for compiling programs in any
1257 language; or command-line options meaningful for C and related
1258 languages; or options that are meaningful only for C++ programs.
1259 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1260 explanations of options for languages related to C@.
1261 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1262 explanations of options that are meaningful only for C++ programs.
1264 @node C Dialect Options
1265 @section Options Controlling C Dialect
1266 @cindex dialect options
1267 @cindex language dialect options
1268 @cindex options, dialect
1270 The following options control the dialect of C (or languages derived
1271 from C, such as C++, Objective-C and Objective-C++) that the compiler
1275 @cindex ANSI support
1279 In C mode, support all ISO C90 programs. In C++ mode,
1280 remove GNU extensions that conflict with ISO C++.
1282 This turns off certain features of GCC that are incompatible with ISO
1283 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1284 such as the @code{asm} and @code{typeof} keywords, and
1285 predefined macros such as @code{unix} and @code{vax} that identify the
1286 type of system you are using. It also enables the undesirable and
1287 rarely used ISO trigraph feature. For the C compiler,
1288 it disables recognition of C++ style @samp{//} comments as well as
1289 the @code{inline} keyword.
1291 The alternate keywords @code{__asm__}, @code{__extension__},
1292 @code{__inline__} and @code{__typeof__} continue to work despite
1293 @option{-ansi}. You would not want to use them in an ISO C program, of
1294 course, but it is useful to put them in header files that might be included
1295 in compilations done with @option{-ansi}. Alternate predefined macros
1296 such as @code{__unix__} and @code{__vax__} are also available, with or
1297 without @option{-ansi}.
1299 The @option{-ansi} option does not cause non-ISO programs to be
1300 rejected gratuitously. For that, @option{-pedantic} is required in
1301 addition to @option{-ansi}. @xref{Warning Options}.
1303 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1304 option is used. Some header files may notice this macro and refrain
1305 from declaring certain functions or defining certain macros that the
1306 ISO standard doesn't call for; this is to avoid interfering with any
1307 programs that might use these names for other things.
1309 Functions which would normally be built in but do not have semantics
1310 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1311 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1312 built-in functions provided by GCC}, for details of the functions
1317 Determine the language standard. This option is currently only
1318 supported when compiling C or C++. A value for this option must be
1319 provided; possible values are
1324 ISO C90 (same as @option{-ansi}).
1326 @item iso9899:199409
1327 ISO C90 as modified in amendment 1.
1333 ISO C99. Note that this standard is not yet fully supported; see
1334 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1335 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1338 Default, ISO C90 plus GNU extensions (including some C99 features).
1342 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1343 this will become the default. The name @samp{gnu9x} is deprecated.
1346 The 1998 ISO C++ standard plus amendments.
1349 The same as @option{-std=c++98} plus GNU extensions. This is the
1350 default for C++ code.
1353 The working draft of the upcoming ISO C++0x standard. This option
1354 enables experimental features that are likely to be included in
1355 C++0x. The working draft is constantly changing, and any feature that is
1356 enabled by this flag may be removed from future versions of GCC if it is
1357 not part of the C++0x standard.
1360 The same as @option{-std=c++0x} plus GNU extensions. As with
1361 @option{-std=c++0x}, this option enables experimental features that may
1362 be removed in future versions of GCC.
1365 Even when this option is not specified, you can still use some of the
1366 features of newer standards in so far as they do not conflict with
1367 previous C standards. For example, you may use @code{__restrict__} even
1368 when @option{-std=c99} is not specified.
1370 The @option{-std} options specifying some version of ISO C have the same
1371 effects as @option{-ansi}, except that features that were not in ISO C90
1372 but are in the specified version (for example, @samp{//} comments and
1373 the @code{inline} keyword in ISO C99) are not disabled.
1375 @xref{Standards,,Language Standards Supported by GCC}, for details of
1376 these standard versions.
1378 @item -fgnu89-inline
1379 @opindex fgnu89-inline
1380 The option @option{-fgnu89-inline} tells GCC to use the traditional
1381 GNU semantics for @code{inline} functions when in C99 mode.
1382 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1383 is accepted and ignored by GCC versions 4.1.3 up to but not including
1384 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1385 C99 mode. Using this option is roughly equivalent to adding the
1386 @code{gnu_inline} function attribute to all inline functions
1387 (@pxref{Function Attributes}).
1389 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1390 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1391 specifies the default behavior). This option was first supported in
1392 GCC 4.3. This option is not supported in C89 or gnu89 mode.
1394 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1395 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1396 in effect for @code{inline} functions. @xref{Common Predefined
1397 Macros,,,cpp,The C Preprocessor}.
1399 @item -aux-info @var{filename}
1401 Output to the given filename prototyped declarations for all functions
1402 declared and/or defined in a translation unit, including those in header
1403 files. This option is silently ignored in any language other than C@.
1405 Besides declarations, the file indicates, in comments, the origin of
1406 each declaration (source file and line), whether the declaration was
1407 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1408 @samp{O} for old, respectively, in the first character after the line
1409 number and the colon), and whether it came from a declaration or a
1410 definition (@samp{C} or @samp{F}, respectively, in the following
1411 character). In the case of function definitions, a K&R-style list of
1412 arguments followed by their declarations is also provided, inside
1413 comments, after the declaration.
1417 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1418 keyword, so that code can use these words as identifiers. You can use
1419 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1420 instead. @option{-ansi} implies @option{-fno-asm}.
1422 In C++, this switch only affects the @code{typeof} keyword, since
1423 @code{asm} and @code{inline} are standard keywords. You may want to
1424 use the @option{-fno-gnu-keywords} flag instead, which has the same
1425 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1426 switch only affects the @code{asm} and @code{typeof} keywords, since
1427 @code{inline} is a standard keyword in ISO C99.
1430 @itemx -fno-builtin-@var{function}
1431 @opindex fno-builtin
1432 @cindex built-in functions
1433 Don't recognize built-in functions that do not begin with
1434 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1435 functions provided by GCC}, for details of the functions affected,
1436 including those which are not built-in functions when @option{-ansi} or
1437 @option{-std} options for strict ISO C conformance are used because they
1438 do not have an ISO standard meaning.
1440 GCC normally generates special code to handle certain built-in functions
1441 more efficiently; for instance, calls to @code{alloca} may become single
1442 instructions that adjust the stack directly, and calls to @code{memcpy}
1443 may become inline copy loops. The resulting code is often both smaller
1444 and faster, but since the function calls no longer appear as such, you
1445 cannot set a breakpoint on those calls, nor can you change the behavior
1446 of the functions by linking with a different library. In addition,
1447 when a function is recognized as a built-in function, GCC may use
1448 information about that function to warn about problems with calls to
1449 that function, or to generate more efficient code, even if the
1450 resulting code still contains calls to that function. For example,
1451 warnings are given with @option{-Wformat} for bad calls to
1452 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1453 known not to modify global memory.
1455 With the @option{-fno-builtin-@var{function}} option
1456 only the built-in function @var{function} is
1457 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1458 function is named this is not built-in in this version of GCC, this
1459 option is ignored. There is no corresponding
1460 @option{-fbuiltin-@var{function}} option; if you wish to enable
1461 built-in functions selectively when using @option{-fno-builtin} or
1462 @option{-ffreestanding}, you may define macros such as:
1465 #define abs(n) __builtin_abs ((n))
1466 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1471 @cindex hosted environment
1473 Assert that compilation takes place in a hosted environment. This implies
1474 @option{-fbuiltin}. A hosted environment is one in which the
1475 entire standard library is available, and in which @code{main} has a return
1476 type of @code{int}. Examples are nearly everything except a kernel.
1477 This is equivalent to @option{-fno-freestanding}.
1479 @item -ffreestanding
1480 @opindex ffreestanding
1481 @cindex hosted environment
1483 Assert that compilation takes place in a freestanding environment. This
1484 implies @option{-fno-builtin}. A freestanding environment
1485 is one in which the standard library may not exist, and program startup may
1486 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1487 This is equivalent to @option{-fno-hosted}.
1489 @xref{Standards,,Language Standards Supported by GCC}, for details of
1490 freestanding and hosted environments.
1494 @cindex openmp parallel
1495 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1496 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1497 compiler generates parallel code according to the OpenMP Application
1498 Program Interface v2.5 @w{@uref{http://www.openmp.org/}}.
1500 @item -fms-extensions
1501 @opindex fms-extensions
1502 Accept some non-standard constructs used in Microsoft header files.
1504 Some cases of unnamed fields in structures and unions are only
1505 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1506 fields within structs/unions}, for details.
1510 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1511 options for strict ISO C conformance) implies @option{-trigraphs}.
1513 @item -no-integrated-cpp
1514 @opindex no-integrated-cpp
1515 Performs a compilation in two passes: preprocessing and compiling. This
1516 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1517 @option{-B} option. The user supplied compilation step can then add in
1518 an additional preprocessing step after normal preprocessing but before
1519 compiling. The default is to use the integrated cpp (internal cpp)
1521 The semantics of this option will change if "cc1", "cc1plus", and
1522 "cc1obj" are merged.
1524 @cindex traditional C language
1525 @cindex C language, traditional
1527 @itemx -traditional-cpp
1528 @opindex traditional-cpp
1529 @opindex traditional
1530 Formerly, these options caused GCC to attempt to emulate a pre-standard
1531 C compiler. They are now only supported with the @option{-E} switch.
1532 The preprocessor continues to support a pre-standard mode. See the GNU
1533 CPP manual for details.
1535 @item -fcond-mismatch
1536 @opindex fcond-mismatch
1537 Allow conditional expressions with mismatched types in the second and
1538 third arguments. The value of such an expression is void. This option
1539 is not supported for C++.
1541 @item -flax-vector-conversions
1542 @opindex flax-vector-conversions
1543 Allow implicit conversions between vectors with differing numbers of
1544 elements and/or incompatible element types. This option should not be
1547 @item -funsigned-char
1548 @opindex funsigned-char
1549 Let the type @code{char} be unsigned, like @code{unsigned char}.
1551 Each kind of machine has a default for what @code{char} should
1552 be. It is either like @code{unsigned char} by default or like
1553 @code{signed char} by default.
1555 Ideally, a portable program should always use @code{signed char} or
1556 @code{unsigned char} when it depends on the signedness of an object.
1557 But many programs have been written to use plain @code{char} and
1558 expect it to be signed, or expect it to be unsigned, depending on the
1559 machines they were written for. This option, and its inverse, let you
1560 make such a program work with the opposite default.
1562 The type @code{char} is always a distinct type from each of
1563 @code{signed char} or @code{unsigned char}, even though its behavior
1564 is always just like one of those two.
1567 @opindex fsigned-char
1568 Let the type @code{char} be signed, like @code{signed char}.
1570 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1571 the negative form of @option{-funsigned-char}. Likewise, the option
1572 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1574 @item -fsigned-bitfields
1575 @itemx -funsigned-bitfields
1576 @itemx -fno-signed-bitfields
1577 @itemx -fno-unsigned-bitfields
1578 @opindex fsigned-bitfields
1579 @opindex funsigned-bitfields
1580 @opindex fno-signed-bitfields
1581 @opindex fno-unsigned-bitfields
1582 These options control whether a bit-field is signed or unsigned, when the
1583 declaration does not use either @code{signed} or @code{unsigned}. By
1584 default, such a bit-field is signed, because this is consistent: the
1585 basic integer types such as @code{int} are signed types.
1588 @node C++ Dialect Options
1589 @section Options Controlling C++ Dialect
1591 @cindex compiler options, C++
1592 @cindex C++ options, command line
1593 @cindex options, C++
1594 This section describes the command-line options that are only meaningful
1595 for C++ programs; but you can also use most of the GNU compiler options
1596 regardless of what language your program is in. For example, you
1597 might compile a file @code{firstClass.C} like this:
1600 g++ -g -frepo -O -c firstClass.C
1604 In this example, only @option{-frepo} is an option meant
1605 only for C++ programs; you can use the other options with any
1606 language supported by GCC@.
1608 Here is a list of options that are @emph{only} for compiling C++ programs:
1612 @item -fabi-version=@var{n}
1613 @opindex fabi-version
1614 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1615 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1616 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1617 the version that conforms most closely to the C++ ABI specification.
1618 Therefore, the ABI obtained using version 0 will change as ABI bugs
1621 The default is version 2.
1623 @item -fno-access-control
1624 @opindex fno-access-control
1625 Turn off all access checking. This switch is mainly useful for working
1626 around bugs in the access control code.
1630 Check that the pointer returned by @code{operator new} is non-null
1631 before attempting to modify the storage allocated. This check is
1632 normally unnecessary because the C++ standard specifies that
1633 @code{operator new} will only return @code{0} if it is declared
1634 @samp{throw()}, in which case the compiler will always check the
1635 return value even without this option. In all other cases, when
1636 @code{operator new} has a non-empty exception specification, memory
1637 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1638 @samp{new (nothrow)}.
1640 @item -fconserve-space
1641 @opindex fconserve-space
1642 Put uninitialized or runtime-initialized global variables into the
1643 common segment, as C does. This saves space in the executable at the
1644 cost of not diagnosing duplicate definitions. If you compile with this
1645 flag and your program mysteriously crashes after @code{main()} has
1646 completed, you may have an object that is being destroyed twice because
1647 two definitions were merged.
1649 This option is no longer useful on most targets, now that support has
1650 been added for putting variables into BSS without making them common.
1652 @item -ffriend-injection
1653 @opindex ffriend-injection
1654 Inject friend functions into the enclosing namespace, so that they are
1655 visible outside the scope of the class in which they are declared.
1656 Friend functions were documented to work this way in the old Annotated
1657 C++ Reference Manual, and versions of G++ before 4.1 always worked
1658 that way. However, in ISO C++ a friend function which is not declared
1659 in an enclosing scope can only be found using argument dependent
1660 lookup. This option causes friends to be injected as they were in
1663 This option is for compatibility, and may be removed in a future
1666 @item -fno-elide-constructors
1667 @opindex fno-elide-constructors
1668 The C++ standard allows an implementation to omit creating a temporary
1669 which is only used to initialize another object of the same type.
1670 Specifying this option disables that optimization, and forces G++ to
1671 call the copy constructor in all cases.
1673 @item -fno-enforce-eh-specs
1674 @opindex fno-enforce-eh-specs
1675 Don't generate code to check for violation of exception specifications
1676 at runtime. This option violates the C++ standard, but may be useful
1677 for reducing code size in production builds, much like defining
1678 @samp{NDEBUG}. This does not give user code permission to throw
1679 exceptions in violation of the exception specifications; the compiler
1680 will still optimize based on the specifications, so throwing an
1681 unexpected exception will result in undefined behavior.
1684 @itemx -fno-for-scope
1686 @opindex fno-for-scope
1687 If @option{-ffor-scope} is specified, the scope of variables declared in
1688 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1689 as specified by the C++ standard.
1690 If @option{-fno-for-scope} is specified, the scope of variables declared in
1691 a @i{for-init-statement} extends to the end of the enclosing scope,
1692 as was the case in old versions of G++, and other (traditional)
1693 implementations of C++.
1695 The default if neither flag is given to follow the standard,
1696 but to allow and give a warning for old-style code that would
1697 otherwise be invalid, or have different behavior.
1699 @item -fno-gnu-keywords
1700 @opindex fno-gnu-keywords
1701 Do not recognize @code{typeof} as a keyword, so that code can use this
1702 word as an identifier. You can use the keyword @code{__typeof__} instead.
1703 @option{-ansi} implies @option{-fno-gnu-keywords}.
1705 @item -fno-implicit-templates
1706 @opindex fno-implicit-templates
1707 Never emit code for non-inline templates which are instantiated
1708 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1709 @xref{Template Instantiation}, for more information.
1711 @item -fno-implicit-inline-templates
1712 @opindex fno-implicit-inline-templates
1713 Don't emit code for implicit instantiations of inline templates, either.
1714 The default is to handle inlines differently so that compiles with and
1715 without optimization will need the same set of explicit instantiations.
1717 @item -fno-implement-inlines
1718 @opindex fno-implement-inlines
1719 To save space, do not emit out-of-line copies of inline functions
1720 controlled by @samp{#pragma implementation}. This will cause linker
1721 errors if these functions are not inlined everywhere they are called.
1723 @item -fms-extensions
1724 @opindex fms-extensions
1725 Disable pedantic warnings about constructs used in MFC, such as implicit
1726 int and getting a pointer to member function via non-standard syntax.
1728 @item -fno-nonansi-builtins
1729 @opindex fno-nonansi-builtins
1730 Disable built-in declarations of functions that are not mandated by
1731 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1732 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1734 @item -fno-operator-names
1735 @opindex fno-operator-names
1736 Do not treat the operator name keywords @code{and}, @code{bitand},
1737 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1738 synonyms as keywords.
1740 @item -fno-optional-diags
1741 @opindex fno-optional-diags
1742 Disable diagnostics that the standard says a compiler does not need to
1743 issue. Currently, the only such diagnostic issued by G++ is the one for
1744 a name having multiple meanings within a class.
1747 @opindex fpermissive
1748 Downgrade some diagnostics about nonconformant code from errors to
1749 warnings. Thus, using @option{-fpermissive} will allow some
1750 nonconforming code to compile.
1754 Enable automatic template instantiation at link time. This option also
1755 implies @option{-fno-implicit-templates}. @xref{Template
1756 Instantiation}, for more information.
1760 Disable generation of information about every class with virtual
1761 functions for use by the C++ runtime type identification features
1762 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1763 of the language, you can save some space by using this flag. Note that
1764 exception handling uses the same information, but it will generate it as
1765 needed. The @samp{dynamic_cast} operator can still be used for casts that
1766 do not require runtime type information, i.e. casts to @code{void *} or to
1767 unambiguous base classes.
1771 Emit statistics about front-end processing at the end of the compilation.
1772 This information is generally only useful to the G++ development team.
1774 @item -ftemplate-depth-@var{n}
1775 @opindex ftemplate-depth
1776 Set the maximum instantiation depth for template classes to @var{n}.
1777 A limit on the template instantiation depth is needed to detect
1778 endless recursions during template class instantiation. ANSI/ISO C++
1779 conforming programs must not rely on a maximum depth greater than 17.
1781 @item -fno-threadsafe-statics
1782 @opindex fno-threadsafe-statics
1783 Do not emit the extra code to use the routines specified in the C++
1784 ABI for thread-safe initialization of local statics. You can use this
1785 option to reduce code size slightly in code that doesn't need to be
1788 @item -fuse-cxa-atexit
1789 @opindex fuse-cxa-atexit
1790 Register destructors for objects with static storage duration with the
1791 @code{__cxa_atexit} function rather than the @code{atexit} function.
1792 This option is required for fully standards-compliant handling of static
1793 destructors, but will only work if your C library supports
1794 @code{__cxa_atexit}.
1796 @item -fno-use-cxa-get-exception-ptr
1797 @opindex fno-use-cxa-get-exception-ptr
1798 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1799 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1800 if the runtime routine is not available.
1802 @item -fvisibility-inlines-hidden
1803 @opindex fvisibility-inlines-hidden
1804 This switch declares that the user does not attempt to compare
1805 pointers to inline methods where the addresses of the two functions
1806 were taken in different shared objects.
1808 The effect of this is that GCC may, effectively, mark inline methods with
1809 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1810 appear in the export table of a DSO and do not require a PLT indirection
1811 when used within the DSO@. Enabling this option can have a dramatic effect
1812 on load and link times of a DSO as it massively reduces the size of the
1813 dynamic export table when the library makes heavy use of templates.
1815 The behavior of this switch is not quite the same as marking the
1816 methods as hidden directly, because it does not affect static variables
1817 local to the function or cause the compiler to deduce that
1818 the function is defined in only one shared object.
1820 You may mark a method as having a visibility explicitly to negate the
1821 effect of the switch for that method. For example, if you do want to
1822 compare pointers to a particular inline method, you might mark it as
1823 having default visibility. Marking the enclosing class with explicit
1824 visibility will have no effect.
1826 Explicitly instantiated inline methods are unaffected by this option
1827 as their linkage might otherwise cross a shared library boundary.
1828 @xref{Template Instantiation}.
1832 Do not use weak symbol support, even if it is provided by the linker.
1833 By default, G++ will use weak symbols if they are available. This
1834 option exists only for testing, and should not be used by end-users;
1835 it will result in inferior code and has no benefits. This option may
1836 be removed in a future release of G++.
1840 Do not search for header files in the standard directories specific to
1841 C++, but do still search the other standard directories. (This option
1842 is used when building the C++ library.)
1845 In addition, these optimization, warning, and code generation options
1846 have meanings only for C++ programs:
1849 @item -fno-default-inline
1850 @opindex fno-default-inline
1851 Do not assume @samp{inline} for functions defined inside a class scope.
1852 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1853 functions will have linkage like inline functions; they just won't be
1856 @item -Wabi @r{(C++ only)}
1858 Warn when G++ generates code that is probably not compatible with the
1859 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1860 all such cases, there are probably some cases that are not warned about,
1861 even though G++ is generating incompatible code. There may also be
1862 cases where warnings are emitted even though the code that is generated
1865 You should rewrite your code to avoid these warnings if you are
1866 concerned about the fact that code generated by G++ may not be binary
1867 compatible with code generated by other compilers.
1869 The known incompatibilities at this point include:
1874 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1875 pack data into the same byte as a base class. For example:
1878 struct A @{ virtual void f(); int f1 : 1; @};
1879 struct B : public A @{ int f2 : 1; @};
1883 In this case, G++ will place @code{B::f2} into the same byte
1884 as@code{A::f1}; other compilers will not. You can avoid this problem
1885 by explicitly padding @code{A} so that its size is a multiple of the
1886 byte size on your platform; that will cause G++ and other compilers to
1887 layout @code{B} identically.
1890 Incorrect handling of tail-padding for virtual bases. G++ does not use
1891 tail padding when laying out virtual bases. For example:
1894 struct A @{ virtual void f(); char c1; @};
1895 struct B @{ B(); char c2; @};
1896 struct C : public A, public virtual B @{@};
1900 In this case, G++ will not place @code{B} into the tail-padding for
1901 @code{A}; other compilers will. You can avoid this problem by
1902 explicitly padding @code{A} so that its size is a multiple of its
1903 alignment (ignoring virtual base classes); that will cause G++ and other
1904 compilers to layout @code{C} identically.
1907 Incorrect handling of bit-fields with declared widths greater than that
1908 of their underlying types, when the bit-fields appear in a union. For
1912 union U @{ int i : 4096; @};
1916 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1917 union too small by the number of bits in an @code{int}.
1920 Empty classes can be placed at incorrect offsets. For example:
1930 struct C : public B, public A @{@};
1934 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1935 it should be placed at offset zero. G++ mistakenly believes that the
1936 @code{A} data member of @code{B} is already at offset zero.
1939 Names of template functions whose types involve @code{typename} or
1940 template template parameters can be mangled incorrectly.
1943 template <typename Q>
1944 void f(typename Q::X) @{@}
1946 template <template <typename> class Q>
1947 void f(typename Q<int>::X) @{@}
1951 Instantiations of these templates may be mangled incorrectly.
1955 @item -Wctor-dtor-privacy @r{(C++ only)}
1956 @opindex Wctor-dtor-privacy
1957 Warn when a class seems unusable because all the constructors or
1958 destructors in that class are private, and it has neither friends nor
1959 public static member functions.
1961 @item -Wnon-virtual-dtor @r{(C++ only)}
1962 @opindex Wnon-virtual-dtor
1963 Warn when a class appears to be polymorphic, thereby requiring a virtual
1964 destructor, yet it declares a non-virtual one. This warning is also
1965 enabled if -Weffc++ is specified.
1967 @item -Wreorder @r{(C++ only)}
1969 @cindex reordering, warning
1970 @cindex warning for reordering of member initializers
1971 Warn when the order of member initializers given in the code does not
1972 match the order in which they must be executed. For instance:
1978 A(): j (0), i (1) @{ @}
1982 The compiler will rearrange the member initializers for @samp{i}
1983 and @samp{j} to match the declaration order of the members, emitting
1984 a warning to that effect. This warning is enabled by @option{-Wall}.
1987 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1990 @item -Weffc++ @r{(C++ only)}
1992 Warn about violations of the following style guidelines from Scott Meyers'
1993 @cite{Effective C++} book:
1997 Item 11: Define a copy constructor and an assignment operator for classes
1998 with dynamically allocated memory.
2001 Item 12: Prefer initialization to assignment in constructors.
2004 Item 14: Make destructors virtual in base classes.
2007 Item 15: Have @code{operator=} return a reference to @code{*this}.
2010 Item 23: Don't try to return a reference when you must return an object.
2014 Also warn about violations of the following style guidelines from
2015 Scott Meyers' @cite{More Effective C++} book:
2019 Item 6: Distinguish between prefix and postfix forms of increment and
2020 decrement operators.
2023 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2027 When selecting this option, be aware that the standard library
2028 headers do not obey all of these guidelines; use @samp{grep -v}
2029 to filter out those warnings.
2031 @item -Wno-deprecated @r{(C++ only)}
2032 @opindex Wno-deprecated
2033 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
2035 @item -Wstrict-null-sentinel @r{(C++ only)}
2036 @opindex Wstrict-null-sentinel
2037 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2038 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2039 to @code{__null}. Although it is a null pointer constant not a null pointer,
2040 it is guaranteed to of the same size as a pointer. But this use is
2041 not portable across different compilers.
2043 @item -Wno-non-template-friend @r{(C++ only)}
2044 @opindex Wno-non-template-friend
2045 Disable warnings when non-templatized friend functions are declared
2046 within a template. Since the advent of explicit template specification
2047 support in G++, if the name of the friend is an unqualified-id (i.e.,
2048 @samp{friend foo(int)}), the C++ language specification demands that the
2049 friend declare or define an ordinary, nontemplate function. (Section
2050 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2051 could be interpreted as a particular specialization of a templatized
2052 function. Because this non-conforming behavior is no longer the default
2053 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2054 check existing code for potential trouble spots and is on by default.
2055 This new compiler behavior can be turned off with
2056 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2057 but disables the helpful warning.
2059 @item -Wold-style-cast @r{(C++ only)}
2060 @opindex Wold-style-cast
2061 Warn if an old-style (C-style) cast to a non-void type is used within
2062 a C++ program. The new-style casts (@samp{dynamic_cast},
2063 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2064 less vulnerable to unintended effects and much easier to search for.
2066 @item -Woverloaded-virtual @r{(C++ only)}
2067 @opindex Woverloaded-virtual
2068 @cindex overloaded virtual fn, warning
2069 @cindex warning for overloaded virtual fn
2070 Warn when a function declaration hides virtual functions from a
2071 base class. For example, in:
2078 struct B: public A @{
2083 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2091 will fail to compile.
2093 @item -Wno-pmf-conversions @r{(C++ only)}
2094 @opindex Wno-pmf-conversions
2095 Disable the diagnostic for converting a bound pointer to member function
2098 @item -Wsign-promo @r{(C++ only)}
2099 @opindex Wsign-promo
2100 Warn when overload resolution chooses a promotion from unsigned or
2101 enumerated type to a signed type, over a conversion to an unsigned type of
2102 the same size. Previous versions of G++ would try to preserve
2103 unsignedness, but the standard mandates the current behavior.
2108 A& operator = (int);
2118 In this example, G++ will synthesize a default @samp{A& operator =
2119 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2122 @node Objective-C and Objective-C++ Dialect Options
2123 @section Options Controlling Objective-C and Objective-C++ Dialects
2125 @cindex compiler options, Objective-C and Objective-C++
2126 @cindex Objective-C and Objective-C++ options, command line
2127 @cindex options, Objective-C and Objective-C++
2128 (NOTE: This manual does not describe the Objective-C and Objective-C++
2129 languages themselves. See @xref{Standards,,Language Standards
2130 Supported by GCC}, for references.)
2132 This section describes the command-line options that are only meaningful
2133 for Objective-C and Objective-C++ programs, but you can also use most of
2134 the language-independent GNU compiler options.
2135 For example, you might compile a file @code{some_class.m} like this:
2138 gcc -g -fgnu-runtime -O -c some_class.m
2142 In this example, @option{-fgnu-runtime} is an option meant only for
2143 Objective-C and Objective-C++ programs; you can use the other options with
2144 any language supported by GCC@.
2146 Note that since Objective-C is an extension of the C language, Objective-C
2147 compilations may also use options specific to the C front-end (e.g.,
2148 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2149 C++-specific options (e.g., @option{-Wabi}).
2151 Here is a list of options that are @emph{only} for compiling Objective-C
2152 and Objective-C++ programs:
2155 @item -fconstant-string-class=@var{class-name}
2156 @opindex fconstant-string-class
2157 Use @var{class-name} as the name of the class to instantiate for each
2158 literal string specified with the syntax @code{@@"@dots{}"}. The default
2159 class name is @code{NXConstantString} if the GNU runtime is being used, and
2160 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2161 @option{-fconstant-cfstrings} option, if also present, will override the
2162 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2163 to be laid out as constant CoreFoundation strings.
2166 @opindex fgnu-runtime
2167 Generate object code compatible with the standard GNU Objective-C
2168 runtime. This is the default for most types of systems.
2170 @item -fnext-runtime
2171 @opindex fnext-runtime
2172 Generate output compatible with the NeXT runtime. This is the default
2173 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2174 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2177 @item -fno-nil-receivers
2178 @opindex fno-nil-receivers
2179 Assume that all Objective-C message dispatches (e.g.,
2180 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2181 is not @code{nil}. This allows for more efficient entry points in the runtime
2182 to be used. Currently, this option is only available in conjunction with
2183 the NeXT runtime on Mac OS X 10.3 and later.
2185 @item -fobjc-call-cxx-cdtors
2186 @opindex fobjc-call-cxx-cdtors
2187 For each Objective-C class, check if any of its instance variables is a
2188 C++ object with a non-trivial default constructor. If so, synthesize a
2189 special @code{- (id) .cxx_construct} instance method that will run
2190 non-trivial default constructors on any such instance variables, in order,
2191 and then return @code{self}. Similarly, check if any instance variable
2192 is a C++ object with a non-trivial destructor, and if so, synthesize a
2193 special @code{- (void) .cxx_destruct} method that will run
2194 all such default destructors, in reverse order.
2196 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2197 thusly generated will only operate on instance variables declared in the
2198 current Objective-C class, and not those inherited from superclasses. It
2199 is the responsibility of the Objective-C runtime to invoke all such methods
2200 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2201 will be invoked by the runtime immediately after a new object
2202 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2203 be invoked immediately before the runtime deallocates an object instance.
2205 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2206 support for invoking the @code{- (id) .cxx_construct} and
2207 @code{- (void) .cxx_destruct} methods.
2209 @item -fobjc-direct-dispatch
2210 @opindex fobjc-direct-dispatch
2211 Allow fast jumps to the message dispatcher. On Darwin this is
2212 accomplished via the comm page.
2214 @item -fobjc-exceptions
2215 @opindex fobjc-exceptions
2216 Enable syntactic support for structured exception handling in Objective-C,
2217 similar to what is offered by C++ and Java. This option is
2218 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2227 @@catch (AnObjCClass *exc) @{
2234 @@catch (AnotherClass *exc) @{
2237 @@catch (id allOthers) @{
2247 The @code{@@throw} statement may appear anywhere in an Objective-C or
2248 Objective-C++ program; when used inside of a @code{@@catch} block, the
2249 @code{@@throw} may appear without an argument (as shown above), in which case
2250 the object caught by the @code{@@catch} will be rethrown.
2252 Note that only (pointers to) Objective-C objects may be thrown and
2253 caught using this scheme. When an object is thrown, it will be caught
2254 by the nearest @code{@@catch} clause capable of handling objects of that type,
2255 analogously to how @code{catch} blocks work in C++ and Java. A
2256 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2257 any and all Objective-C exceptions not caught by previous @code{@@catch}
2260 The @code{@@finally} clause, if present, will be executed upon exit from the
2261 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2262 regardless of whether any exceptions are thrown, caught or rethrown
2263 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2264 of the @code{finally} clause in Java.
2266 There are several caveats to using the new exception mechanism:
2270 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2271 idioms provided by the @code{NSException} class, the new
2272 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2273 systems, due to additional functionality needed in the (NeXT) Objective-C
2277 As mentioned above, the new exceptions do not support handling
2278 types other than Objective-C objects. Furthermore, when used from
2279 Objective-C++, the Objective-C exception model does not interoperate with C++
2280 exceptions at this time. This means you cannot @code{@@throw} an exception
2281 from Objective-C and @code{catch} it in C++, or vice versa
2282 (i.e., @code{throw @dots{} @@catch}).
2285 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2286 blocks for thread-safe execution:
2289 @@synchronized (ObjCClass *guard) @{
2294 Upon entering the @code{@@synchronized} block, a thread of execution shall
2295 first check whether a lock has been placed on the corresponding @code{guard}
2296 object by another thread. If it has, the current thread shall wait until
2297 the other thread relinquishes its lock. Once @code{guard} becomes available,
2298 the current thread will place its own lock on it, execute the code contained in
2299 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2300 making @code{guard} available to other threads).
2302 Unlike Java, Objective-C does not allow for entire methods to be marked
2303 @code{@@synchronized}. Note that throwing exceptions out of
2304 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2305 to be unlocked properly.
2309 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2311 @item -freplace-objc-classes
2312 @opindex freplace-objc-classes
2313 Emit a special marker instructing @command{ld(1)} not to statically link in
2314 the resulting object file, and allow @command{dyld(1)} to load it in at
2315 run time instead. This is used in conjunction with the Fix-and-Continue
2316 debugging mode, where the object file in question may be recompiled and
2317 dynamically reloaded in the course of program execution, without the need
2318 to restart the program itself. Currently, Fix-and-Continue functionality
2319 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2324 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2325 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2326 compile time) with static class references that get initialized at load time,
2327 which improves run-time performance. Specifying the @option{-fzero-link} flag
2328 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2329 to be retained. This is useful in Zero-Link debugging mode, since it allows
2330 for individual class implementations to be modified during program execution.
2334 Dump interface declarations for all classes seen in the source file to a
2335 file named @file{@var{sourcename}.decl}.
2337 @item -Wassign-intercept
2338 @opindex Wassign-intercept
2339 Warn whenever an Objective-C assignment is being intercepted by the
2343 @opindex Wno-protocol
2344 If a class is declared to implement a protocol, a warning is issued for
2345 every method in the protocol that is not implemented by the class. The
2346 default behavior is to issue a warning for every method not explicitly
2347 implemented in the class, even if a method implementation is inherited
2348 from the superclass. If you use the @option{-Wno-protocol} option, then
2349 methods inherited from the superclass are considered to be implemented,
2350 and no warning is issued for them.
2354 Warn if multiple methods of different types for the same selector are
2355 found during compilation. The check is performed on the list of methods
2356 in the final stage of compilation. Additionally, a check is performed
2357 for each selector appearing in a @code{@@selector(@dots{})}
2358 expression, and a corresponding method for that selector has been found
2359 during compilation. Because these checks scan the method table only at
2360 the end of compilation, these warnings are not produced if the final
2361 stage of compilation is not reached, for example because an error is
2362 found during compilation, or because the @option{-fsyntax-only} option is
2365 @item -Wstrict-selector-match
2366 @opindex Wstrict-selector-match
2367 Warn if multiple methods with differing argument and/or return types are
2368 found for a given selector when attempting to send a message using this
2369 selector to a receiver of type @code{id} or @code{Class}. When this flag
2370 is off (which is the default behavior), the compiler will omit such warnings
2371 if any differences found are confined to types which share the same size
2374 @item -Wundeclared-selector
2375 @opindex Wundeclared-selector
2376 Warn if a @code{@@selector(@dots{})} expression referring to an
2377 undeclared selector is found. A selector is considered undeclared if no
2378 method with that name has been declared before the
2379 @code{@@selector(@dots{})} expression, either explicitly in an
2380 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2381 an @code{@@implementation} section. This option always performs its
2382 checks as soon as a @code{@@selector(@dots{})} expression is found,
2383 while @option{-Wselector} only performs its checks in the final stage of
2384 compilation. This also enforces the coding style convention
2385 that methods and selectors must be declared before being used.
2387 @item -print-objc-runtime-info
2388 @opindex print-objc-runtime-info
2389 Generate C header describing the largest structure that is passed by
2394 @node Language Independent Options
2395 @section Options to Control Diagnostic Messages Formatting
2396 @cindex options to control diagnostics formatting
2397 @cindex diagnostic messages
2398 @cindex message formatting
2400 Traditionally, diagnostic messages have been formatted irrespective of
2401 the output device's aspect (e.g.@: its width, @dots{}). The options described
2402 below can be used to control the diagnostic messages formatting
2403 algorithm, e.g.@: how many characters per line, how often source location
2404 information should be reported. Right now, only the C++ front end can
2405 honor these options. However it is expected, in the near future, that
2406 the remaining front ends would be able to digest them correctly.
2409 @item -fmessage-length=@var{n}
2410 @opindex fmessage-length
2411 Try to format error messages so that they fit on lines of about @var{n}
2412 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2413 the front ends supported by GCC@. If @var{n} is zero, then no
2414 line-wrapping will be done; each error message will appear on a single
2417 @opindex fdiagnostics-show-location
2418 @item -fdiagnostics-show-location=once
2419 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2420 reporter to emit @emph{once} source location information; that is, in
2421 case the message is too long to fit on a single physical line and has to
2422 be wrapped, the source location won't be emitted (as prefix) again,
2423 over and over, in subsequent continuation lines. This is the default
2426 @item -fdiagnostics-show-location=every-line
2427 Only meaningful in line-wrapping mode. Instructs the diagnostic
2428 messages reporter to emit the same source location information (as
2429 prefix) for physical lines that result from the process of breaking
2430 a message which is too long to fit on a single line.
2432 @item -fdiagnostics-show-option
2433 @opindex fdiagnostics-show-option
2434 This option instructs the diagnostic machinery to add text to each
2435 diagnostic emitted, which indicates which command line option directly
2436 controls that diagnostic, when such an option is known to the
2437 diagnostic machinery.
2439 @item -Wcoverage-mismatch
2440 @opindex Wcoverage-mismatch
2441 Warn if feedback profiles do not match when using the
2442 @option{-fprofile-use} option.
2443 If a source file was changed between @option{-fprofile-gen} and
2444 @option{-fprofile-use}, the files with the profile feedback can fail
2445 to match the source file and GCC can not use the profile feedback
2446 information. By default, GCC emits an error message in this case.
2447 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2448 error. GCC does not use appropriate feedback profiles, so using this
2449 option can result in poorly optimized code. This option is useful
2450 only in the case of very minor changes such as bug fixes to an
2455 @node Warning Options
2456 @section Options to Request or Suppress Warnings
2457 @cindex options to control warnings
2458 @cindex warning messages
2459 @cindex messages, warning
2460 @cindex suppressing warnings
2462 Warnings are diagnostic messages that report constructions which
2463 are not inherently erroneous but which are risky or suggest there
2464 may have been an error.
2466 You can request many specific warnings with options beginning @samp{-W},
2467 for example @option{-Wimplicit} to request warnings on implicit
2468 declarations. Each of these specific warning options also has a
2469 negative form beginning @samp{-Wno-} to turn off warnings;
2470 for example, @option{-Wno-implicit}. This manual lists only one of the
2471 two forms, whichever is not the default.
2473 The following options control the amount and kinds of warnings produced
2474 by GCC; for further, language-specific options also refer to
2475 @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect
2479 @cindex syntax checking
2481 @opindex fsyntax-only
2482 Check the code for syntax errors, but don't do anything beyond that.
2486 Issue all the warnings demanded by strict ISO C and ISO C++;
2487 reject all programs that use forbidden extensions, and some other
2488 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2489 version of the ISO C standard specified by any @option{-std} option used.
2491 Valid ISO C and ISO C++ programs should compile properly with or without
2492 this option (though a rare few will require @option{-ansi} or a
2493 @option{-std} option specifying the required version of ISO C)@. However,
2494 without this option, certain GNU extensions and traditional C and C++
2495 features are supported as well. With this option, they are rejected.
2497 @option{-pedantic} does not cause warning messages for use of the
2498 alternate keywords whose names begin and end with @samp{__}. Pedantic
2499 warnings are also disabled in the expression that follows
2500 @code{__extension__}. However, only system header files should use
2501 these escape routes; application programs should avoid them.
2502 @xref{Alternate Keywords}.
2504 Some users try to use @option{-pedantic} to check programs for strict ISO
2505 C conformance. They soon find that it does not do quite what they want:
2506 it finds some non-ISO practices, but not all---only those for which
2507 ISO C @emph{requires} a diagnostic, and some others for which
2508 diagnostics have been added.
2510 A feature to report any failure to conform to ISO C might be useful in
2511 some instances, but would require considerable additional work and would
2512 be quite different from @option{-pedantic}. We don't have plans to
2513 support such a feature in the near future.
2515 Where the standard specified with @option{-std} represents a GNU
2516 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2517 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2518 extended dialect is based. Warnings from @option{-pedantic} are given
2519 where they are required by the base standard. (It would not make sense
2520 for such warnings to be given only for features not in the specified GNU
2521 C dialect, since by definition the GNU dialects of C include all
2522 features the compiler supports with the given option, and there would be
2523 nothing to warn about.)
2525 @item -pedantic-errors
2526 @opindex pedantic-errors
2527 Like @option{-pedantic}, except that errors are produced rather than
2532 Inhibit all warning messages.
2536 Inhibit warning messages about the use of @samp{#import}.
2538 @item -Wchar-subscripts
2539 @opindex Wchar-subscripts
2540 Warn if an array subscript has type @code{char}. This is a common cause
2541 of error, as programmers often forget that this type is signed on some
2543 This warning is enabled by @option{-Wall}.
2547 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2548 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2549 This warning is enabled by @option{-Wall}.
2551 @item -Wfatal-errors
2552 @opindex Wfatal-errors
2553 This option causes the compiler to abort compilation on the first error
2554 occurred rather than trying to keep going and printing further error
2559 @opindex ffreestanding
2560 @opindex fno-builtin
2561 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2562 the arguments supplied have types appropriate to the format string
2563 specified, and that the conversions specified in the format string make
2564 sense. This includes standard functions, and others specified by format
2565 attributes (@pxref{Function Attributes}), in the @code{printf},
2566 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2567 not in the C standard) families (or other target-specific families).
2568 Which functions are checked without format attributes having been
2569 specified depends on the standard version selected, and such checks of
2570 functions without the attribute specified are disabled by
2571 @option{-ffreestanding} or @option{-fno-builtin}.
2573 The formats are checked against the format features supported by GNU
2574 libc version 2.2. These include all ISO C90 and C99 features, as well
2575 as features from the Single Unix Specification and some BSD and GNU
2576 extensions. Other library implementations may not support all these
2577 features; GCC does not support warning about features that go beyond a
2578 particular library's limitations. However, if @option{-pedantic} is used
2579 with @option{-Wformat}, warnings will be given about format features not
2580 in the selected standard version (but not for @code{strfmon} formats,
2581 since those are not in any version of the C standard). @xref{C Dialect
2582 Options,,Options Controlling C Dialect}.
2584 Since @option{-Wformat} also checks for null format arguments for
2585 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2587 @option{-Wformat} is included in @option{-Wall}. For more control over some
2588 aspects of format checking, the options @option{-Wformat-y2k},
2589 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2590 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2591 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2594 @opindex Wformat-y2k
2595 If @option{-Wformat} is specified, also warn about @code{strftime}
2596 formats which may yield only a two-digit year.
2598 @item -Wno-format-extra-args
2599 @opindex Wno-format-extra-args
2600 If @option{-Wformat} is specified, do not warn about excess arguments to a
2601 @code{printf} or @code{scanf} format function. The C standard specifies
2602 that such arguments are ignored.
2604 Where the unused arguments lie between used arguments that are
2605 specified with @samp{$} operand number specifications, normally
2606 warnings are still given, since the implementation could not know what
2607 type to pass to @code{va_arg} to skip the unused arguments. However,
2608 in the case of @code{scanf} formats, this option will suppress the
2609 warning if the unused arguments are all pointers, since the Single
2610 Unix Specification says that such unused arguments are allowed.
2612 @item -Wno-format-zero-length
2613 @opindex Wno-format-zero-length
2614 If @option{-Wformat} is specified, do not warn about zero-length formats.
2615 The C standard specifies that zero-length formats are allowed.
2617 @item -Wformat-nonliteral
2618 @opindex Wformat-nonliteral
2619 If @option{-Wformat} is specified, also warn if the format string is not a
2620 string literal and so cannot be checked, unless the format function
2621 takes its format arguments as a @code{va_list}.
2623 @item -Wformat-security
2624 @opindex Wformat-security
2625 If @option{-Wformat} is specified, also warn about uses of format
2626 functions that represent possible security problems. At present, this
2627 warns about calls to @code{printf} and @code{scanf} functions where the
2628 format string is not a string literal and there are no format arguments,
2629 as in @code{printf (foo);}. This may be a security hole if the format
2630 string came from untrusted input and contains @samp{%n}. (This is
2631 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2632 in future warnings may be added to @option{-Wformat-security} that are not
2633 included in @option{-Wformat-nonliteral}.)
2637 Enable @option{-Wformat} plus format checks not included in
2638 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2639 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2643 Warn about passing a null pointer for arguments marked as
2644 requiring a non-null value by the @code{nonnull} function attribute.
2646 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2647 can be disabled with the @option{-Wno-nonnull} option.
2649 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2651 Warn about uninitialized variables which are initialized with themselves.
2652 Note this option can only be used with the @option{-Wuninitialized} option,
2653 which in turn only works with @option{-O1} and above.
2655 For example, GCC will warn about @code{i} being uninitialized in the
2656 following snippet only when @option{-Winit-self} has been specified:
2667 @item -Wimplicit-int
2668 @opindex Wimplicit-int
2669 Warn when a declaration does not specify a type.
2670 This warning is enabled by @option{-Wall}.
2672 @item -Wimplicit-function-declaration
2673 @opindex Wimplicit-function-declaration
2674 @opindex Wno-implicit-function-declaration
2675 Give a warning whenever a function is used before being declared. In
2676 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
2677 enabled by default and it is made into an error by
2678 @option{-pedantic-errors}. This warning is also enabled by
2683 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2684 This warning is enabled by @option{-Wall}.
2688 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2689 function with external linkage, returning int, taking either zero
2690 arguments, two, or three arguments of appropriate types.
2691 This warning is enabled by @option{-Wall}.
2693 @item -Wmissing-braces
2694 @opindex Wmissing-braces
2695 Warn if an aggregate or union initializer is not fully bracketed. In
2696 the following example, the initializer for @samp{a} is not fully
2697 bracketed, but that for @samp{b} is fully bracketed.
2700 int a[2][2] = @{ 0, 1, 2, 3 @};
2701 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2704 This warning is enabled by @option{-Wall}.
2706 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2707 @opindex Wmissing-include-dirs
2708 Warn if a user-supplied include directory does not exist.
2711 @opindex Wparentheses
2712 Warn if parentheses are omitted in certain contexts, such
2713 as when there is an assignment in a context where a truth value
2714 is expected, or when operators are nested whose precedence people
2715 often get confused about.
2717 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2718 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2719 interpretation from that of ordinary mathematical notation.
2721 Also warn about constructions where there may be confusion to which
2722 @code{if} statement an @code{else} branch belongs. Here is an example of
2737 In C/C++, every @code{else} branch belongs to the innermost possible
2738 @code{if} statement, which in this example is @code{if (b)}. This is
2739 often not what the programmer expected, as illustrated in the above
2740 example by indentation the programmer chose. When there is the
2741 potential for this confusion, GCC will issue a warning when this flag
2742 is specified. To eliminate the warning, add explicit braces around
2743 the innermost @code{if} statement so there is no way the @code{else}
2744 could belong to the enclosing @code{if}. The resulting code would
2761 This warning is enabled by @option{-Wall}.
2763 @item -Wsequence-point
2764 @opindex Wsequence-point
2765 Warn about code that may have undefined semantics because of violations
2766 of sequence point rules in the C and C++ standards.
2768 The C and C++ standards defines the order in which expressions in a C/C++
2769 program are evaluated in terms of @dfn{sequence points}, which represent
2770 a partial ordering between the execution of parts of the program: those
2771 executed before the sequence point, and those executed after it. These
2772 occur after the evaluation of a full expression (one which is not part
2773 of a larger expression), after the evaluation of the first operand of a
2774 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2775 function is called (but after the evaluation of its arguments and the
2776 expression denoting the called function), and in certain other places.
2777 Other than as expressed by the sequence point rules, the order of
2778 evaluation of subexpressions of an expression is not specified. All
2779 these rules describe only a partial order rather than a total order,
2780 since, for example, if two functions are called within one expression
2781 with no sequence point between them, the order in which the functions
2782 are called is not specified. However, the standards committee have
2783 ruled that function calls do not overlap.
2785 It is not specified when between sequence points modifications to the
2786 values of objects take effect. Programs whose behavior depends on this
2787 have undefined behavior; the C and C++ standards specify that ``Between
2788 the previous and next sequence point an object shall have its stored
2789 value modified at most once by the evaluation of an expression.
2790 Furthermore, the prior value shall be read only to determine the value
2791 to be stored.''. If a program breaks these rules, the results on any
2792 particular implementation are entirely unpredictable.
2794 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2795 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2796 diagnosed by this option, and it may give an occasional false positive
2797 result, but in general it has been found fairly effective at detecting
2798 this sort of problem in programs.
2800 The standard is worded confusingly, therefore there is some debate
2801 over the precise meaning of the sequence point rules in subtle cases.
2802 Links to discussions of the problem, including proposed formal
2803 definitions, may be found on the GCC readings page, at
2804 @w{@uref{http://gcc.gnu.org/readings.html}}.
2806 This warning is enabled by @option{-Wall} for C and C++.
2809 @opindex Wreturn-type
2810 @opindex Wno-return-type
2811 Warn whenever a function is defined with a return-type that defaults
2812 to @code{int}. Also warn about any @code{return} statement with no
2813 return-value in a function whose return-type is not @code{void}
2814 (falling off the end of the function body is considered returning
2815 without a value), and about a @code{return} statement with a
2816 expression in a function whose return-type is @code{void}.
2818 Also warn if the return type of a function has a type qualifier
2819 such as @code{const}. For ISO C such a type qualifier has no effect,
2820 since the value returned by a function is not an lvalue.
2821 For C++, the warning is only emitted for scalar types or @code{void}.
2822 ISO C prohibits qualified @code{void} return types on function
2823 definitions, so such return types always receive a warning
2824 even without this option.
2826 For C++, a function without return type always produces a diagnostic
2827 message, even when @option{-Wno-return-type} is specified. The only
2828 exceptions are @samp{main} and functions defined in system headers.
2830 This warning is enabled by @option{-Wall}.
2834 Warn whenever a @code{switch} statement has an index of enumerated type
2835 and lacks a @code{case} for one or more of the named codes of that
2836 enumeration. (The presence of a @code{default} label prevents this
2837 warning.) @code{case} labels outside the enumeration range also
2838 provoke warnings when this option is used.
2839 This warning is enabled by @option{-Wall}.
2841 @item -Wswitch-default
2842 @opindex Wswitch-switch
2843 Warn whenever a @code{switch} statement does not have a @code{default}
2847 @opindex Wswitch-enum
2848 Warn whenever a @code{switch} statement has an index of enumerated type
2849 and lacks a @code{case} for one or more of the named codes of that
2850 enumeration. @code{case} labels outside the enumeration range also
2851 provoke warnings when this option is used.
2855 Warn if any trigraphs are encountered that might change the meaning of
2856 the program (trigraphs within comments are not warned about).
2857 This warning is enabled by @option{-Wall}.
2859 @item -Wunused-function
2860 @opindex Wunused-function
2861 Warn whenever a static function is declared but not defined or a
2862 non-inline static function is unused.
2863 This warning is enabled by @option{-Wall}.
2865 @item -Wunused-label
2866 @opindex Wunused-label
2867 Warn whenever a label is declared but not used.
2868 This warning is enabled by @option{-Wall}.
2870 To suppress this warning use the @samp{unused} attribute
2871 (@pxref{Variable Attributes}).
2873 @item -Wunused-parameter
2874 @opindex Wunused-parameter
2875 Warn whenever a function parameter is unused aside from its declaration.
2877 To suppress this warning use the @samp{unused} attribute
2878 (@pxref{Variable Attributes}).
2880 @item -Wunused-variable
2881 @opindex Wunused-variable
2882 Warn whenever a local variable or non-constant static variable is unused
2883 aside from its declaration.
2884 This warning is enabled by @option{-Wall}.
2886 To suppress this warning use the @samp{unused} attribute
2887 (@pxref{Variable Attributes}).
2889 @item -Wunused-value
2890 @opindex Wunused-value
2891 Warn whenever a statement computes a result that is explicitly not
2892 used. To suppress this warning cast the unused expression to
2893 @samp{void}. This includes an expression-statement or the left-hand
2894 side of a comma expression that contains no side effects. For example,
2895 an expression such as @samp{x[i,j]} will cause a warning, while
2896 @samp{x[(void)i,j]} will not.
2898 This warning is enabled by @option{-Wall}.
2902 All the above @option{-Wunused} options combined.
2904 In order to get a warning about an unused function parameter, you must
2905 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2906 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2908 @item -Wuninitialized
2909 @opindex Wuninitialized
2910 Warn if an automatic variable is used without first being initialized or
2911 if a variable may be clobbered by a @code{setjmp} call.
2913 These warnings are possible only in optimizing compilation,
2914 because they require data flow information that is computed only
2915 when optimizing. If you do not specify @option{-O}, you will not get
2916 these warnings. Instead, GCC will issue a warning about @option{-Wuninitialized}
2917 requiring @option{-O}.
2919 If you want to warn about code which uses the uninitialized value of the
2920 variable in its own initializer, use the @option{-Winit-self} option.
2922 These warnings occur for individual uninitialized or clobbered
2923 elements of structure, union or array variables as well as for
2924 variables which are uninitialized or clobbered as a whole. They do
2925 not occur for variables or elements declared @code{volatile}. Because
2926 these warnings depend on optimization, the exact variables or elements
2927 for which there are warnings will depend on the precise optimization
2928 options and version of GCC used.
2930 Note that there may be no warning about a variable that is used only
2931 to compute a value that itself is never used, because such
2932 computations may be deleted by data flow analysis before the warnings
2935 These warnings are made optional because GCC is not smart
2936 enough to see all the reasons why the code might be correct
2937 despite appearing to have an error. Here is one example of how
2958 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2959 always initialized, but GCC doesn't know this. Here is
2960 another common case:
2965 if (change_y) save_y = y, y = new_y;
2967 if (change_y) y = save_y;
2972 This has no bug because @code{save_y} is used only if it is set.
2974 @cindex @code{longjmp} warnings
2975 This option also warns when a non-volatile automatic variable might be
2976 changed by a call to @code{longjmp}. These warnings as well are possible
2977 only in optimizing compilation.
2979 The compiler sees only the calls to @code{setjmp}. It cannot know
2980 where @code{longjmp} will be called; in fact, a signal handler could
2981 call it at any point in the code. As a result, you may get a warning
2982 even when there is in fact no problem because @code{longjmp} cannot
2983 in fact be called at the place which would cause a problem.
2985 Some spurious warnings can be avoided if you declare all the functions
2986 you use that never return as @code{noreturn}. @xref{Function
2989 This warning is enabled by @option{-Wall}.
2991 @item -Wunknown-pragmas
2992 @opindex Wunknown-pragmas
2993 @cindex warning for unknown pragmas
2994 @cindex unknown pragmas, warning
2995 @cindex pragmas, warning of unknown
2996 Warn when a #pragma directive is encountered which is not understood by
2997 GCC@. If this command line option is used, warnings will even be issued
2998 for unknown pragmas in system header files. This is not the case if
2999 the warnings were only enabled by the @option{-Wall} command line option.
3002 @opindex Wno-pragmas
3004 Do not warn about misuses of pragmas, such as incorrect parameters,
3005 invalid syntax, or conflicts between pragmas. See also
3006 @samp{-Wunknown-pragmas}.
3008 @item -Wstrict-aliasing
3009 @opindex Wstrict-aliasing
3010 This option is only active when @option{-fstrict-aliasing} is active.
3011 It warns about code which might break the strict aliasing rules that the
3012 compiler is using for optimization. The warning does not catch all
3013 cases, but does attempt to catch the more common pitfalls. It is
3014 included in @option{-Wall}.
3015 It is equivalent to -Wstrict-aliasing=3
3017 @item -Wstrict-aliasing=n
3018 @opindex Wstrict-aliasing=n
3019 This option is only active when @option{-fstrict-aliasing} is active.
3020 It warns about code which might break the strict aliasing rules that the
3021 compiler is using for optimization.
3022 Higher levels correspond to higher accuracy (fewer false positives).
3023 Higher levels also correspond to more effort, similar to the way -O works.
3024 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3027 Level 1: Most aggressive, quick, least accurate.
3028 Possibly useful when higher levels
3029 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3030 false negatives. However, it has many false positives.
3031 Warns for all pointer conversions between possibly incompatible types,
3032 even if never dereferenced. Runs in the frontend only.
3034 Level 2: Aggressive, quick, not too precise.
3035 May still have many false positives (not as many as level 1 though),
3036 and few false negatives (but possibly more than level 1).
3037 Unlike level 1, it only warns when an address is taken. Warns about
3038 incomplete types. Runs in the frontend only.
3040 Level 3 (default for @option{-Wstrict-aliasing}):
3041 Should have very few false positives and few false
3042 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3043 Takes care of the common punn+dereference pattern in the frontend:
3044 @code{*(int*)&some_float}.
3045 If optimization is enabled, it also runs in the backend, where it deals
3046 with multiple statement cases using flow-sensitive points-to information.
3047 Only warns when the converted pointer is dereferenced.
3048 Does not warn about incomplete types.
3050 @item -Wstrict-overflow
3051 @item -Wstrict-overflow=@var{n}
3052 @opindex Wstrict-overflow
3053 This option is only active when @option{-fstrict-overflow} is active.
3054 It warns about cases where the compiler optimizes based on the
3055 assumption that signed overflow does not occur. Note that it does not
3056 warn about all cases where the code might overflow: it only warns
3057 about cases where the compiler implements some optimization. Thus
3058 this warning depends on the optimization level.
3060 An optimization which assumes that signed overflow does not occur is
3061 perfectly safe if the values of the variables involved are such that
3062 overflow never does, in fact, occur. Therefore this warning can
3063 easily give a false positive: a warning about code which is not
3064 actually a problem. To help focus on important issues, several
3065 warning levels are defined. No warnings are issued for the use of
3066 undefined signed overflow when estimating how many iterations a loop
3067 will require, in particular when determining whether a loop will be
3071 @item -Wstrict-overflow=1
3072 Warn about cases which are both questionable and easy to avoid. For
3073 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3074 compiler will simplify this to @code{1}. This level of
3075 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3076 are not, and must be explicitly requested.
3078 @item -Wstrict-overflow=2
3079 Also warn about other cases where a comparison is simplified to a
3080 constant. For example: @code{abs (x) >= 0}. This can only be
3081 simplified when @option{-fstrict-overflow} is in effect, because
3082 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3083 zero. @option{-Wstrict-overflow} (with no level) is the same as
3084 @option{-Wstrict-overflow=2}.
3086 @item -Wstrict-overflow=3
3087 Also warn about other cases where a comparison is simplified. For
3088 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3090 @item -Wstrict-overflow=4
3091 Also warn about other simplifications not covered by the above cases.
3092 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3094 @item -Wstrict-overflow=5
3095 Also warn about cases where the compiler reduces the magnitude of a
3096 constant involved in a comparison. For example: @code{x + 2 > y} will
3097 be simplified to @code{x + 1 >= y}. This is reported only at the
3098 highest warning level because this simplification applies to many
3099 comparisons, so this warning level will give a very large number of
3103 @item -Warray-bounds
3104 @opindex Wno-array-bounds
3105 @opindex Warray-bounds
3106 This option is only active when @option{-ftree-vrp} is active
3107 (default for -O2 and above). It warns about subscripts to arrays
3108 that are always out of bounds. This warning is enabled by @option{-Wall}.
3112 All of the above @samp{-W} options combined. This enables all the
3113 warnings about constructions that some users consider questionable, and
3114 that are easy to avoid (or modify to prevent the warning), even in
3115 conjunction with macros. This also enables some language-specific
3116 warnings described in @ref{C++ Dialect Options} and
3117 @ref{Objective-C and Objective-C++ Dialect Options}.
3120 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
3121 Some of them warn about constructions that users generally do not
3122 consider questionable, but which occasionally you might wish to check
3123 for; others warn about constructions that are necessary or hard to avoid
3124 in some cases, and there is no simple way to modify the code to suppress
3131 (This option used to be called @option{-W}. The older name is still
3132 supported, but the newer name is more descriptive.) Print extra warning
3133 messages for these events:
3137 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
3140 Storage-class specifiers like @code{static} are not the first things
3141 in a declaration. According to the C Standard, this usage is
3142 obsolescent. This warning can be independently controlled by
3143 @option{-Wold-style-declaration}.
3146 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
3150 A comparison between signed and unsigned values could produce an
3151 incorrect result when the signed value is converted to unsigned.
3152 (But don't warn if @option{-Wno-sign-compare} is also specified.)
3155 An aggregate has an initializer which does not initialize all members.
3156 This warning can be independently controlled by
3157 @option{-Wmissing-field-initializers}.
3160 An initialized field without side effects is overridden when using
3161 designated initializers (@pxref{Designated Inits, , Designated
3162 Initializers}). This warning can be independently controlled by
3163 @option{-Woverride-init}.
3166 A function parameter is declared without a type specifier in K&R-style
3167 functions. This warning can be independently controlled by
3168 @option{-Wmissing-parameter-type}.
3171 An empty body occurs in an @samp{if}, @samp{else} or
3172 @samp{do while} statement. This warning can be independently
3173 controlled by @option{-Wempty-body}.
3175 @item @r{(C++ only)}
3176 An empty body occurs in a @samp{while} or @samp{for} statement with no
3177 whitespacing before the semicolon. This warning can be independently
3178 controlled by @option{-Wempty-body}.
3181 A pointer is compared against integer zero with @samp{<}, @samp{<=},
3182 @samp{>}, or @samp{>=}.
3185 A variable might be changed by @samp{longjmp} or @samp{vfork}.
3186 This warning can be independently controlled by @option{-Wclobbered}.
3188 @item @r{(C++ only)}
3189 An enumerator and a non-enumerator both appear in a conditional expression.
3191 @item @r{(C++ only)}
3192 A non-static reference or non-static @samp{const} member appears in a
3193 class without constructors.
3195 @item @r{(C++ only)}
3196 Ambiguous virtual bases.
3198 @item @r{(C++ only)}
3199 Subscripting an array which has been declared @samp{register}.
3201 @item @r{(C++ only)}
3202 Taking the address of a variable which has been declared @samp{register}.
3204 @item @r{(C++ only)}
3205 A base class is not initialized in a derived class' copy constructor.
3208 @item -Wno-div-by-zero
3209 @opindex Wno-div-by-zero
3210 @opindex Wdiv-by-zero
3211 Do not warn about compile-time integer division by zero. Floating point
3212 division by zero is not warned about, as it can be a legitimate way of
3213 obtaining infinities and NaNs.
3215 @item -Wsystem-headers
3216 @opindex Wsystem-headers
3217 @cindex warnings from system headers
3218 @cindex system headers, warnings from
3219 Print warning messages for constructs found in system header files.
3220 Warnings from system headers are normally suppressed, on the assumption
3221 that they usually do not indicate real problems and would only make the
3222 compiler output harder to read. Using this command line option tells
3223 GCC to emit warnings from system headers as if they occurred in user
3224 code. However, note that using @option{-Wall} in conjunction with this
3225 option will @emph{not} warn about unknown pragmas in system
3226 headers---for that, @option{-Wunknown-pragmas} must also be used.
3229 @opindex Wfloat-equal
3230 Warn if floating point values are used in equality comparisons.
3232 The idea behind this is that sometimes it is convenient (for the
3233 programmer) to consider floating-point values as approximations to
3234 infinitely precise real numbers. If you are doing this, then you need
3235 to compute (by analyzing the code, or in some other way) the maximum or
3236 likely maximum error that the computation introduces, and allow for it
3237 when performing comparisons (and when producing output, but that's a
3238 different problem). In particular, instead of testing for equality, you
3239 would check to see whether the two values have ranges that overlap; and
3240 this is done with the relational operators, so equality comparisons are
3243 @item -Wtraditional @r{(C only)}
3244 @opindex Wtraditional
3245 Warn about certain constructs that behave differently in traditional and
3246 ISO C@. Also warn about ISO C constructs that have no traditional C
3247 equivalent, and/or problematic constructs which should be avoided.
3251 Macro parameters that appear within string literals in the macro body.
3252 In traditional C macro replacement takes place within string literals,
3253 but does not in ISO C@.
3256 In traditional C, some preprocessor directives did not exist.
3257 Traditional preprocessors would only consider a line to be a directive
3258 if the @samp{#} appeared in column 1 on the line. Therefore
3259 @option{-Wtraditional} warns about directives that traditional C
3260 understands but would ignore because the @samp{#} does not appear as the
3261 first character on the line. It also suggests you hide directives like
3262 @samp{#pragma} not understood by traditional C by indenting them. Some
3263 traditional implementations would not recognize @samp{#elif}, so it
3264 suggests avoiding it altogether.
3267 A function-like macro that appears without arguments.
3270 The unary plus operator.
3273 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3274 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3275 constants.) Note, these suffixes appear in macros defined in the system
3276 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3277 Use of these macros in user code might normally lead to spurious
3278 warnings, however GCC's integrated preprocessor has enough context to
3279 avoid warning in these cases.
3282 A function declared external in one block and then used after the end of
3286 A @code{switch} statement has an operand of type @code{long}.
3289 A non-@code{static} function declaration follows a @code{static} one.
3290 This construct is not accepted by some traditional C compilers.
3293 The ISO type of an integer constant has a different width or
3294 signedness from its traditional type. This warning is only issued if
3295 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3296 typically represent bit patterns, are not warned about.
3299 Usage of ISO string concatenation is detected.
3302 Initialization of automatic aggregates.
3305 Identifier conflicts with labels. Traditional C lacks a separate
3306 namespace for labels.
3309 Initialization of unions. If the initializer is zero, the warning is
3310 omitted. This is done under the assumption that the zero initializer in
3311 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3312 initializer warnings and relies on default initialization to zero in the
3316 Conversions by prototypes between fixed/floating point values and vice
3317 versa. The absence of these prototypes when compiling with traditional
3318 C would cause serious problems. This is a subset of the possible
3319 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3322 Use of ISO C style function definitions. This warning intentionally is
3323 @emph{not} issued for prototype declarations or variadic functions
3324 because these ISO C features will appear in your code when using
3325 libiberty's traditional C compatibility macros, @code{PARAMS} and
3326 @code{VPARAMS}. This warning is also bypassed for nested functions
3327 because that feature is already a GCC extension and thus not relevant to
3328 traditional C compatibility.
3331 @item -Wtraditional-conversion @r{(C only)}
3332 @opindex Wtraditional-conversion
3333 Warn if a prototype causes a type conversion that is different from what
3334 would happen to the same argument in the absence of a prototype. This
3335 includes conversions of fixed point to floating and vice versa, and
3336 conversions changing the width or signedness of a fixed point argument
3337 except when the same as the default promotion.
3339 @item -Wdeclaration-after-statement @r{(C only)}
3340 @opindex Wdeclaration-after-statement
3341 Warn when a declaration is found after a statement in a block. This
3342 construct, known from C++, was introduced with ISO C99 and is by default
3343 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3344 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3348 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3350 @item -Wno-endif-labels
3351 @opindex Wno-endif-labels
3352 @opindex Wendif-labels
3353 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3357 Warn whenever a local variable shadows another local variable, parameter or
3358 global variable or whenever a built-in function is shadowed.
3360 @item -Wlarger-than-@var{len}
3361 @opindex Wlarger-than
3362 Warn whenever an object of larger than @var{len} bytes is defined.
3364 @item -Wunsafe-loop-optimizations
3365 @opindex Wunsafe-loop-optimizations
3366 Warn if the loop cannot be optimized because the compiler could not
3367 assume anything on the bounds of the loop indices. With
3368 @option{-funsafe-loop-optimizations} warn if the compiler made
3371 @item -Wpointer-arith
3372 @opindex Wpointer-arith
3373 Warn about anything that depends on the ``size of'' a function type or
3374 of @code{void}. GNU C assigns these types a size of 1, for
3375 convenience in calculations with @code{void *} pointers and pointers
3376 to functions. In C++, warn also when an arithmetic operation involves
3377 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3379 @item -Wbad-function-cast @r{(C only)}
3380 @opindex Wbad-function-cast
3381 Warn whenever a function call is cast to a non-matching type.
3382 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3385 Warn about ISO C constructs that are outside of the common subset of
3386 ISO C and ISO C++, e.g.@: request for implicit conversion from
3387 @code{void *} to a pointer to non-@code{void} type.
3389 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3390 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3391 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3392 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3396 Warn whenever a pointer is cast so as to remove a type qualifier from
3397 the target type. For example, warn if a @code{const char *} is cast
3398 to an ordinary @code{char *}.
3401 @opindex Wcast-align
3402 Warn whenever a pointer is cast such that the required alignment of the
3403 target is increased. For example, warn if a @code{char *} is cast to
3404 an @code{int *} on machines where integers can only be accessed at
3405 two- or four-byte boundaries.
3407 @item -Wwrite-strings
3408 @opindex Wwrite-strings
3409 When compiling C, give string constants the type @code{const
3410 char[@var{length}]} so that
3411 copying the address of one into a non-@code{const} @code{char *}
3412 pointer will get a warning; when compiling C++, warn about the
3413 deprecated conversion from string literals to @code{char *}. This
3414 warning, by default, is enabled for C++ programs.
3415 These warnings will help you find at
3416 compile time code that can try to write into a string constant, but
3417 only if you have been very careful about using @code{const} in
3418 declarations and prototypes. Otherwise, it will just be a nuisance;
3419 this is why we did not make @option{-Wall} request these warnings.
3423 Warn for variables that might be changed by @samp{longjmp} or
3424 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3427 @opindex Wconversion
3428 @opindex Wno-conversion
3429 Warn for implicit conversions that may alter a value. This includes
3430 conversions between real and integer, like @code{abs (x)} when
3431 @code{x} is @code{double}; conversions between signed and unsigned,
3432 like @code{unsigned ui = -1}; and conversions to smaller types, like
3433 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3434 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3435 changed by the conversion like in @code{abs (2.0)}. Warnings about
3436 conversions between signed and unsigned integers can be disabled by
3437 using @option{-Wno-sign-conversion}.
3439 For C++, also warn for conversions between @code{NULL} and non-pointer
3440 types; confusing overload resolution for user-defined conversions; and
3441 conversions that will never use a type conversion operator:
3442 conversions to @code{void}, the same type, a base class or a reference
3443 to them. Warnings about conversions between signed and unsigned
3444 integers are disabled by default in C++ unless
3445 @option{-Wsign-conversion} is explicitly enabled.
3448 @opindex Wempty-body
3449 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3450 while} statement. Additionally, in C++, warn when an empty body occurs
3451 in a @samp{while} or @samp{for} statement with no whitespacing before
3452 the semicolon. This warning is also enabled by @option{-Wextra}.
3454 @item -Wsign-compare
3455 @opindex Wsign-compare
3456 @cindex warning for comparison of signed and unsigned values
3457 @cindex comparison of signed and unsigned values, warning
3458 @cindex signed and unsigned values, comparison warning
3459 Warn when a comparison between signed and unsigned values could produce
3460 an incorrect result when the signed value is converted to unsigned.
3461 This warning is also enabled by @option{-Wextra}; to get the other warnings
3462 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3464 @item -Wsign-conversion
3465 @opindex Wsign-conversion
3466 @opindex Wno-sign-conversion
3467 Warn for implicit conversions that may change the sign of an integer
3468 value, like assigning a signed integer expression to an unsigned
3469 integer variable. An explicit cast silences the warning. In C, this
3470 option is enabled also by @option{-Wconversion}.
3474 @opindex Wno-address
3475 Warn about suspicious uses of memory addresses. These include using
3476 the address of a function in a conditional expression, such as
3477 @code{void func(void); if (func)}, and comparisons against the memory
3478 address of a string literal, such as @code{if (x == "abc")}. Such
3479 uses typically indicate a programmer error: the address of a function
3480 always evaluates to true, so their use in a conditional usually
3481 indicate that the programmer forgot the parentheses in a function
3482 call; and comparisons against string literals result in unspecified
3483 behavior and are not portable in C, so they usually indicate that the
3484 programmer intended to use @code{strcmp}. This warning is enabled by
3488 @opindex Wlogical-op
3489 @opindex Wno-logical-op
3490 Warn about suspicious uses of logical operators in expressions.
3491 This includes using logical operators in contexts where a
3492 bit-wise operator is likely to be expected.
3494 @item -Waggregate-return
3495 @opindex Waggregate-return
3496 Warn if any functions that return structures or unions are defined or
3497 called. (In languages where you can return an array, this also elicits
3500 @item -Wno-attributes
3501 @opindex Wno-attributes
3502 @opindex Wattributes
3503 Do not warn if an unexpected @code{__attribute__} is used, such as
3504 unrecognized attributes, function attributes applied to variables,
3505 etc. This will not stop errors for incorrect use of supported
3508 @item -Wstrict-prototypes @r{(C only)}
3509 @opindex Wstrict-prototypes
3510 Warn if a function is declared or defined without specifying the
3511 argument types. (An old-style function definition is permitted without
3512 a warning if preceded by a declaration which specifies the argument
3515 @item -Wold-style-declaration @r{(C only)}
3516 @opindex Wold-style-declaration
3517 Warn for obsolescent usages, according to the C Standard, in a
3518 declaration. For example, warn if storage-class specifiers like
3519 @code{static} are not the first things in a declaration. This warning
3520 is also enabled by @option{-Wextra}.
3522 @item -Wold-style-definition @r{(C only)}
3523 @opindex Wold-style-definition
3524 Warn if an old-style function definition is used. A warning is given
3525 even if there is a previous prototype.
3527 @item -Wmissing-parameter-type @r{(C only)}
3528 @opindex Wmissing-parameter-type
3529 A function parameter is declared without a type specifier in K&R-style
3536 This warning is also enabled by @option{-Wextra}.
3538 @item -Wmissing-prototypes @r{(C only)}
3539 @opindex Wmissing-prototypes
3540 Warn if a global function is defined without a previous prototype
3541 declaration. This warning is issued even if the definition itself
3542 provides a prototype. The aim is to detect global functions that fail
3543 to be declared in header files.
3545 @item -Wmissing-declarations @r{(C and C++ only)}
3546 @opindex Wmissing-declarations
3547 Warn if a global function is defined without a previous declaration.
3548 Do so even if the definition itself provides a prototype.
3549 Use this option to detect global functions that are not declared in
3550 header files. In C++, no warnings are issued for function templates,
3551 or for inline functions, or for functions in anonymous namespaces.
3553 @item -Wmissing-field-initializers
3554 @opindex Wmissing-field-initializers
3557 Warn if a structure's initializer has some fields missing. For
3558 example, the following code would cause such a warning, because
3559 @code{x.h} is implicitly zero:
3562 struct s @{ int f, g, h; @};
3563 struct s x = @{ 3, 4 @};
3566 This option does not warn about designated initializers, so the following
3567 modification would not trigger a warning:
3570 struct s @{ int f, g, h; @};
3571 struct s x = @{ .f = 3, .g = 4 @};
3574 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3575 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3577 @item -Wmissing-noreturn
3578 @opindex Wmissing-noreturn
3579 Warn about functions which might be candidates for attribute @code{noreturn}.
3580 Note these are only possible candidates, not absolute ones. Care should
3581 be taken to manually verify functions actually do not ever return before
3582 adding the @code{noreturn} attribute, otherwise subtle code generation
3583 bugs could be introduced. You will not get a warning for @code{main} in
3584 hosted C environments.
3586 @item -Wmissing-format-attribute
3587 @opindex Wmissing-format-attribute
3589 Warn about function pointers which might be candidates for @code{format}
3590 attributes. Note these are only possible candidates, not absolute ones.
3591 GCC will guess that function pointers with @code{format} attributes that
3592 are used in assignment, initialization, parameter passing or return
3593 statements should have a corresponding @code{format} attribute in the
3594 resulting type. I.e.@: the left-hand side of the assignment or
3595 initialization, the type of the parameter variable, or the return type
3596 of the containing function respectively should also have a @code{format}
3597 attribute to avoid the warning.
3599 GCC will also warn about function definitions which might be
3600 candidates for @code{format} attributes. Again, these are only
3601 possible candidates. GCC will guess that @code{format} attributes
3602 might be appropriate for any function that calls a function like
3603 @code{vprintf} or @code{vscanf}, but this might not always be the
3604 case, and some functions for which @code{format} attributes are
3605 appropriate may not be detected.
3607 @item -Wno-multichar
3608 @opindex Wno-multichar
3610 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3611 Usually they indicate a typo in the user's code, as they have
3612 implementation-defined values, and should not be used in portable code.
3614 @item -Wnormalized=<none|id|nfc|nfkc>
3615 @opindex Wnormalized
3618 @cindex character set, input normalization
3619 In ISO C and ISO C++, two identifiers are different if they are
3620 different sequences of characters. However, sometimes when characters
3621 outside the basic ASCII character set are used, you can have two
3622 different character sequences that look the same. To avoid confusion,
3623 the ISO 10646 standard sets out some @dfn{normalization rules} which
3624 when applied ensure that two sequences that look the same are turned into
3625 the same sequence. GCC can warn you if you are using identifiers which
3626 have not been normalized; this option controls that warning.
3628 There are four levels of warning that GCC supports. The default is
3629 @option{-Wnormalized=nfc}, which warns about any identifier which is
3630 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3631 recommended form for most uses.
3633 Unfortunately, there are some characters which ISO C and ISO C++ allow
3634 in identifiers that when turned into NFC aren't allowable as
3635 identifiers. That is, there's no way to use these symbols in portable
3636 ISO C or C++ and have all your identifiers in NFC.
3637 @option{-Wnormalized=id} suppresses the warning for these characters.
3638 It is hoped that future versions of the standards involved will correct
3639 this, which is why this option is not the default.
3641 You can switch the warning off for all characters by writing
3642 @option{-Wnormalized=none}. You would only want to do this if you
3643 were using some other normalization scheme (like ``D''), because
3644 otherwise you can easily create bugs that are literally impossible to see.
3646 Some characters in ISO 10646 have distinct meanings but look identical
3647 in some fonts or display methodologies, especially once formatting has
3648 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3649 LETTER N'', will display just like a regular @code{n} which has been
3650 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3651 normalization scheme to convert all these into a standard form as
3652 well, and GCC will warn if your code is not in NFKC if you use
3653 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3654 about every identifier that contains the letter O because it might be
3655 confused with the digit 0, and so is not the default, but may be
3656 useful as a local coding convention if the programming environment is
3657 unable to be fixed to display these characters distinctly.
3659 @item -Wno-deprecated-declarations
3660 @opindex Wno-deprecated-declarations
3661 Do not warn about uses of functions (@pxref{Function Attributes}),
3662 variables (@pxref{Variable Attributes}), and types (@pxref{Type
3663 Attributes}) marked as deprecated by using the @code{deprecated}
3667 @opindex Wno-overflow
3668 Do not warn about compile-time overflow in constant expressions.
3670 @item -Woverride-init
3671 @opindex Woverride-init
3674 Warn if an initialized field without side effects is overridden when
3675 using designated initializers (@pxref{Designated Inits, , Designated
3678 This warning is included in @option{-Wextra}. To get other
3679 @option{-Wextra} warnings without this one, use @samp{-Wextra
3680 -Wno-override-init}.
3684 Warn if a structure is given the packed attribute, but the packed
3685 attribute has no effect on the layout or size of the structure.
3686 Such structures may be mis-aligned for little benefit. For
3687 instance, in this code, the variable @code{f.x} in @code{struct bar}
3688 will be misaligned even though @code{struct bar} does not itself
3689 have the packed attribute:
3696 @} __attribute__((packed));
3706 Warn if padding is included in a structure, either to align an element
3707 of the structure or to align the whole structure. Sometimes when this
3708 happens it is possible to rearrange the fields of the structure to
3709 reduce the padding and so make the structure smaller.
3711 @item -Wredundant-decls
3712 @opindex Wredundant-decls
3713 Warn if anything is declared more than once in the same scope, even in
3714 cases where multiple declaration is valid and changes nothing.
3716 @item -Wnested-externs @r{(C only)}
3717 @opindex Wnested-externs
3718 Warn if an @code{extern} declaration is encountered within a function.
3720 @item -Wunreachable-code
3721 @opindex Wunreachable-code
3722 Warn if the compiler detects that code will never be executed.
3724 This option is intended to warn when the compiler detects that at
3725 least a whole line of source code will never be executed, because
3726 some condition is never satisfied or because it is after a
3727 procedure that never returns.
3729 It is possible for this option to produce a warning even though there
3730 are circumstances under which part of the affected line can be executed,
3731 so care should be taken when removing apparently-unreachable code.
3733 For instance, when a function is inlined, a warning may mean that the
3734 line is unreachable in only one inlined copy of the function.
3736 This option is not made part of @option{-Wall} because in a debugging
3737 version of a program there is often substantial code which checks
3738 correct functioning of the program and is, hopefully, unreachable
3739 because the program does work. Another common use of unreachable
3740 code is to provide behavior which is selectable at compile-time.
3744 Warn if a function can not be inlined and it was declared as inline.
3745 Even with this option, the compiler will not warn about failures to
3746 inline functions declared in system headers.
3748 The compiler uses a variety of heuristics to determine whether or not
3749 to inline a function. For example, the compiler takes into account
3750 the size of the function being inlined and the amount of inlining
3751 that has already been done in the current function. Therefore,
3752 seemingly insignificant changes in the source program can cause the
3753 warnings produced by @option{-Winline} to appear or disappear.
3755 @item -Wno-invalid-offsetof @r{(C++ only)}
3756 @opindex Wno-invalid-offsetof
3757 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3758 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3759 to a non-POD type is undefined. In existing C++ implementations,
3760 however, @samp{offsetof} typically gives meaningful results even when
3761 applied to certain kinds of non-POD types. (Such as a simple
3762 @samp{struct} that fails to be a POD type only by virtue of having a
3763 constructor.) This flag is for users who are aware that they are
3764 writing nonportable code and who have deliberately chosen to ignore the
3767 The restrictions on @samp{offsetof} may be relaxed in a future version
3768 of the C++ standard.
3770 @item -Wno-int-to-pointer-cast @r{(C only)}
3771 @opindex Wno-int-to-pointer-cast
3772 Suppress warnings from casts to pointer type of an integer of a
3775 @item -Wno-pointer-to-int-cast @r{(C only)}
3776 @opindex Wno-pointer-to-int-cast
3777 Suppress warnings from casts from a pointer to an integer type of a
3781 @opindex Winvalid-pch
3782 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3783 the search path but can't be used.
3787 @opindex Wno-long-long
3788 Warn if @samp{long long} type is used. This is default. To inhibit
3789 the warning messages, use @option{-Wno-long-long}. Flags
3790 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3791 only when @option{-pedantic} flag is used.
3793 @item -Wvariadic-macros
3794 @opindex Wvariadic-macros
3795 @opindex Wno-variadic-macros
3796 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3797 alternate syntax when in pedantic ISO C99 mode. This is default.
3798 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3803 Warn if variable length array is used in the code.
3804 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
3805 the variable length array.
3807 @item -Wvolatile-register-var
3808 @opindex Wvolatile-register-var
3809 @opindex Wno-volatile-register-var
3810 Warn if a register variable is declared volatile. The volatile
3811 modifier does not inhibit all optimizations that may eliminate reads
3812 and/or writes to register variables.
3814 @item -Wdisabled-optimization
3815 @opindex Wdisabled-optimization
3816 Warn if a requested optimization pass is disabled. This warning does
3817 not generally indicate that there is anything wrong with your code; it
3818 merely indicates that GCC's optimizers were unable to handle the code
3819 effectively. Often, the problem is that your code is too big or too
3820 complex; GCC will refuse to optimize programs when the optimization
3821 itself is likely to take inordinate amounts of time.
3823 @item -Wpointer-sign
3824 @opindex Wpointer-sign
3825 @opindex Wno-pointer-sign
3826 Warn for pointer argument passing or assignment with different signedness.
3827 This option is only supported for C and Objective-C@. It is implied by
3828 @option{-Wall} and by @option{-pedantic}, which can be disabled with
3829 @option{-Wno-pointer-sign}.
3833 Make all warnings into errors.
3837 Make the specified warning into an errors. The specifier for a
3838 warning is appended, for example @option{-Werror=switch} turns the
3839 warnings controlled by @option{-Wswitch} into errors. This switch
3840 takes a negative form, to be used to negate @option{-Werror} for
3841 specific warnings, for example @option{-Wno-error=switch} makes
3842 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
3843 is in effect. You can use the @option{-fdiagnostics-show-option}
3844 option to have each controllable warning amended with the option which
3845 controls it, to determine what to use with this option.
3847 Note that specifying @option{-Werror=}@var{foo} automatically implies
3848 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
3851 @item -Wstack-protector
3852 @opindex Wstack-protector
3853 This option is only active when @option{-fstack-protector} is active. It
3854 warns about functions that will not be protected against stack smashing.
3856 @item -Woverlength-strings
3857 @opindex Woverlength-strings
3858 Warn about string constants which are longer than the ``minimum
3859 maximum'' length specified in the C standard. Modern compilers
3860 generally allow string constants which are much longer than the
3861 standard's minimum limit, but very portable programs should avoid
3862 using longer strings.
3864 The limit applies @emph{after} string constant concatenation, and does
3865 not count the trailing NUL@. In C89, the limit was 509 characters; in
3866 C99, it was raised to 4095. C++98 does not specify a normative
3867 minimum maximum, so we do not diagnose overlength strings in C++@.
3869 This option is implied by @option{-pedantic}, and can be disabled with
3870 @option{-Wno-overlength-strings}.
3873 @node Debugging Options
3874 @section Options for Debugging Your Program or GCC
3875 @cindex options, debugging
3876 @cindex debugging information options
3878 GCC has various special options that are used for debugging
3879 either your program or GCC:
3884 Produce debugging information in the operating system's native format
3885 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
3888 On most systems that use stabs format, @option{-g} enables use of extra
3889 debugging information that only GDB can use; this extra information
3890 makes debugging work better in GDB but will probably make other debuggers
3892 refuse to read the program. If you want to control for certain whether
3893 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3894 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3896 GCC allows you to use @option{-g} with
3897 @option{-O}. The shortcuts taken by optimized code may occasionally
3898 produce surprising results: some variables you declared may not exist
3899 at all; flow of control may briefly move where you did not expect it;
3900 some statements may not be executed because they compute constant
3901 results or their values were already at hand; some statements may
3902 execute in different places because they were moved out of loops.
3904 Nevertheless it proves possible to debug optimized output. This makes
3905 it reasonable to use the optimizer for programs that might have bugs.
3907 The following options are useful when GCC is generated with the
3908 capability for more than one debugging format.
3912 Produce debugging information for use by GDB@. This means to use the
3913 most expressive format available (DWARF 2, stabs, or the native format
3914 if neither of those are supported), including GDB extensions if at all
3919 Produce debugging information in stabs format (if that is supported),
3920 without GDB extensions. This is the format used by DBX on most BSD
3921 systems. On MIPS, Alpha and System V Release 4 systems this option
3922 produces stabs debugging output which is not understood by DBX or SDB@.
3923 On System V Release 4 systems this option requires the GNU assembler.
3925 @item -feliminate-unused-debug-symbols
3926 @opindex feliminate-unused-debug-symbols
3927 Produce debugging information in stabs format (if that is supported),
3928 for only symbols that are actually used.
3930 @item -femit-class-debug-always
3931 Instead of emitting debugging information for a C++ class in only one
3932 object file, emit it in all object files using the class. This option
3933 should be used only with debuggers that are unable to handle the way GCC
3934 normally emits debugging information for classes because using this
3935 option will increase the size of debugging information by as much as a
3940 Produce debugging information in stabs format (if that is supported),
3941 using GNU extensions understood only by the GNU debugger (GDB)@. The
3942 use of these extensions is likely to make other debuggers crash or
3943 refuse to read the program.
3947 Produce debugging information in COFF format (if that is supported).
3948 This is the format used by SDB on most System V systems prior to
3953 Produce debugging information in XCOFF format (if that is supported).
3954 This is the format used by the DBX debugger on IBM RS/6000 systems.
3958 Produce debugging information in XCOFF format (if that is supported),
3959 using GNU extensions understood only by the GNU debugger (GDB)@. The
3960 use of these extensions is likely to make other debuggers crash or
3961 refuse to read the program, and may cause assemblers other than the GNU
3962 assembler (GAS) to fail with an error.
3966 Produce debugging information in DWARF version 2 format (if that is
3967 supported). This is the format used by DBX on IRIX 6. With this
3968 option, GCC uses features of DWARF version 3 when they are useful;
3969 version 3 is upward compatible with version 2, but may still cause
3970 problems for older debuggers.
3974 Produce debugging information in VMS debug format (if that is
3975 supported). This is the format used by DEBUG on VMS systems.
3978 @itemx -ggdb@var{level}
3979 @itemx -gstabs@var{level}
3980 @itemx -gcoff@var{level}
3981 @itemx -gxcoff@var{level}
3982 @itemx -gvms@var{level}
3983 Request debugging information and also use @var{level} to specify how
3984 much information. The default level is 2.
3986 Level 0 produces no debug information at all. Thus, @option{-g0} negates
3989 Level 1 produces minimal information, enough for making backtraces in
3990 parts of the program that you don't plan to debug. This includes
3991 descriptions of functions and external variables, but no information
3992 about local variables and no line numbers.
3994 Level 3 includes extra information, such as all the macro definitions
3995 present in the program. Some debuggers support macro expansion when
3996 you use @option{-g3}.
3998 @option{-gdwarf-2} does not accept a concatenated debug level, because
3999 GCC used to support an option @option{-gdwarf} that meant to generate
4000 debug information in version 1 of the DWARF format (which is very
4001 different from version 2), and it would have been too confusing. That
4002 debug format is long obsolete, but the option cannot be changed now.
4003 Instead use an additional @option{-g@var{level}} option to change the
4004 debug level for DWARF2.
4006 @item -feliminate-dwarf2-dups
4007 @opindex feliminate-dwarf2-dups
4008 Compress DWARF2 debugging information by eliminating duplicated
4009 information about each symbol. This option only makes sense when
4010 generating DWARF2 debugging information with @option{-gdwarf-2}.
4012 @item -femit-struct-debug-baseonly
4013 Emit debug information for struct-like types
4014 only when the base name of the compilation source file
4015 matches the base name of file in which the struct was defined.
4017 This option substantially reduces the size of debugging information,
4018 but at significant potential loss in type information to the debugger.
4019 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4020 See @option{-femit-struct-debug-detailed} for more detailed control.
4022 This option works only with DWARF 2.
4024 @item -femit-struct-debug-reduced
4025 Emit debug information for struct-like types
4026 only when the base name of the compilation source file
4027 matches the base name of file in which the type was defined,
4028 unless the struct is a template or defined in a system header.
4030 This option significantly reduces the size of debugging information,
4031 with some potential loss in type information to the debugger.
4032 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4033 See @option{-femit-struct-debug-detailed} for more detailed control.
4035 This option works only with DWARF 2.
4037 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4038 Specify the struct-like types
4039 for which the compiler will generate debug information.
4040 The intent is to reduce duplicate struct debug information
4041 between different object files within the same program.
4043 This option is a detailed version of
4044 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4045 which will serve for most needs.
4047 A specification has the syntax
4048 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4050 The optional first word limits the specification to
4051 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4052 A struct type is used directly when it is the type of a variable, member.
4053 Indirect uses arise through pointers to structs.
4054 That is, when use of an incomplete struct would be legal, the use is indirect.
4056 @samp{struct one direct; struct two * indirect;}.
4058 The optional second word limits the specification to
4059 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4060 Generic structs are a bit complicated to explain.
4061 For C++, these are non-explicit specializations of template classes,
4062 or non-template classes within the above.
4063 Other programming languages have generics,
4064 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4066 The third word specifies the source files for those
4067 structs for which the compiler will emit debug information.
4068 The values @samp{none} and @samp{any} have the normal meaning.
4069 The value @samp{base} means that
4070 the base of name of the file in which the type declaration appears
4071 must match the base of the name of the main compilation file.
4072 In practice, this means that
4073 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4074 but types declared in other header will not.
4075 The value @samp{sys} means those types satisfying @samp{base}
4076 or declared in system or compiler headers.
4078 You may need to experiment to determine the best settings for your application.
4080 The default is @samp{-femit-struct-debug-detailed=all}.
4082 This option works only with DWARF 2.
4084 @cindex @command{prof}
4087 Generate extra code to write profile information suitable for the
4088 analysis program @command{prof}. You must use this option when compiling
4089 the source files you want data about, and you must also use it when
4092 @cindex @command{gprof}
4095 Generate extra code to write profile information suitable for the
4096 analysis program @command{gprof}. You must use this option when compiling
4097 the source files you want data about, and you must also use it when
4102 Makes the compiler print out each function name as it is compiled, and
4103 print some statistics about each pass when it finishes.
4106 @opindex ftime-report
4107 Makes the compiler print some statistics about the time consumed by each
4108 pass when it finishes.
4111 @opindex fmem-report
4112 Makes the compiler print some statistics about permanent memory
4113 allocation when it finishes.
4115 @item -fpre-ipa-mem-report
4116 @opindex fpre-ipa-mem-report
4117 @item -fpost-ipa-mem-report
4118 @opindex fpost-ipa-mem-report
4119 Makes the compiler print some statistics about permanent memory
4120 allocation before or after interprocedural optimization.
4122 @item -fprofile-arcs
4123 @opindex fprofile-arcs
4124 Add code so that program flow @dfn{arcs} are instrumented. During
4125 execution the program records how many times each branch and call is
4126 executed and how many times it is taken or returns. When the compiled
4127 program exits it saves this data to a file called
4128 @file{@var{auxname}.gcda} for each source file. The data may be used for
4129 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4130 test coverage analysis (@option{-ftest-coverage}). Each object file's
4131 @var{auxname} is generated from the name of the output file, if
4132 explicitly specified and it is not the final executable, otherwise it is
4133 the basename of the source file. In both cases any suffix is removed
4134 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4135 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4136 @xref{Cross-profiling}.
4138 @cindex @command{gcov}
4142 This option is used to compile and link code instrumented for coverage
4143 analysis. The option is a synonym for @option{-fprofile-arcs}
4144 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4145 linking). See the documentation for those options for more details.
4150 Compile the source files with @option{-fprofile-arcs} plus optimization
4151 and code generation options. For test coverage analysis, use the
4152 additional @option{-ftest-coverage} option. You do not need to profile
4153 every source file in a program.
4156 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4157 (the latter implies the former).
4160 Run the program on a representative workload to generate the arc profile
4161 information. This may be repeated any number of times. You can run
4162 concurrent instances of your program, and provided that the file system
4163 supports locking, the data files will be correctly updated. Also
4164 @code{fork} calls are detected and correctly handled (double counting
4168 For profile-directed optimizations, compile the source files again with
4169 the same optimization and code generation options plus
4170 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4171 Control Optimization}).
4174 For test coverage analysis, use @command{gcov} to produce human readable
4175 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4176 @command{gcov} documentation for further information.
4180 With @option{-fprofile-arcs}, for each function of your program GCC
4181 creates a program flow graph, then finds a spanning tree for the graph.
4182 Only arcs that are not on the spanning tree have to be instrumented: the
4183 compiler adds code to count the number of times that these arcs are
4184 executed. When an arc is the only exit or only entrance to a block, the
4185 instrumentation code can be added to the block; otherwise, a new basic
4186 block must be created to hold the instrumentation code.
4189 @item -ftest-coverage
4190 @opindex ftest-coverage
4191 Produce a notes file that the @command{gcov} code-coverage utility
4192 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4193 show program coverage. Each source file's note file is called
4194 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4195 above for a description of @var{auxname} and instructions on how to
4196 generate test coverage data. Coverage data will match the source files
4197 more closely, if you do not optimize.
4199 @item -d@var{letters}
4200 @item -fdump-rtl-@var{pass}
4202 Says to make debugging dumps during compilation at times specified by
4203 @var{letters}. This is used for debugging the RTL-based passes of the
4204 compiler. The file names for most of the dumps are made by appending a
4205 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
4206 from the name of the output file, if explicitly specified and it is not
4207 an executable, otherwise it is the basename of the source file.
4209 Most debug dumps can be enabled either passing a letter to the @option{-d}
4210 option, or with a long @option{-fdump-rtl} switch; here are the possible
4211 letters for use in @var{letters} and @var{pass}, and their meanings:
4216 Annotate the assembler output with miscellaneous debugging information.
4219 @itemx -fdump-rtl-bbro
4221 @opindex fdump-rtl-bbro
4222 Dump after block reordering, to @file{@var{file}.148r.bbro}.
4225 @itemx -fdump-rtl-combine
4227 @opindex fdump-rtl-combine
4228 Dump after the RTL instruction combination pass, to the file
4229 @file{@var{file}.129r.combine}.
4232 @itemx -fdump-rtl-ce1
4233 @itemx -fdump-rtl-ce2
4235 @opindex fdump-rtl-ce1
4236 @opindex fdump-rtl-ce2
4237 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
4238 first if conversion, to the file @file{@var{file}.117r.ce1}. @option{-dC}
4239 and @option{-fdump-rtl-ce2} enable dumping after the second if
4240 conversion, to the file @file{@var{file}.130r.ce2}.
4243 @itemx -fdump-rtl-btl
4244 @itemx -fdump-rtl-dbr
4246 @opindex fdump-rtl-btl
4247 @opindex fdump-rtl-dbr
4248 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
4249 target load optimization, to @file{@var{file}.31.btl}. @option{-dd}
4250 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
4251 scheduling, to @file{@var{file}.36.dbr}.
4255 Dump all macro definitions, at the end of preprocessing, in addition to
4259 @itemx -fdump-rtl-ce3
4261 @opindex fdump-rtl-ce3
4262 Dump after the third if conversion, to @file{@var{file}.146r.ce3}.
4265 @itemx -fdump-rtl-cfg
4266 @itemx -fdump-rtl-life
4268 @opindex fdump-rtl-cfg
4269 @opindex fdump-rtl-life
4270 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
4271 and data flow analysis, to @file{@var{file}.116r.cfg}. @option{-df}
4272 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
4273 to @file{@var{file}.128r.life1} and @file{@var{file}.135r.life2}.
4276 @itemx -fdump-rtl-greg
4278 @opindex fdump-rtl-greg
4279 Dump after global register allocation, to @file{@var{file}.139r.greg}.
4282 @itemx -fdump-rtl-gcse
4283 @itemx -fdump-rtl-bypass
4285 @opindex fdump-rtl-gcse
4286 @opindex fdump-rtl-bypass
4287 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
4288 @file{@var{file}.114r.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
4289 enable dumping after jump bypassing and control flow optimizations, to
4290 @file{@var{file}.115r.bypass}.
4293 @itemx -fdump-rtl-eh
4295 @opindex fdump-rtl-eh
4296 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
4299 @itemx -fdump-rtl-sibling
4301 @opindex fdump-rtl-sibling
4302 Dump after sibling call optimizations, to @file{@var{file}.106r.sibling}.
4305 @itemx -fdump-rtl-jump
4307 @opindex fdump-rtl-jump
4308 Dump after the first jump optimization, to @file{@var{file}.112r.jump}.
4311 @itemx -fdump-rtl-stack
4313 @opindex fdump-rtl-stack
4314 Dump after conversion from GCC's "flat register file" registers to the
4315 x87's stack-like registers, to @file{@var{file}.152r.stack}.
4318 @itemx -fdump-rtl-lreg
4320 @opindex fdump-rtl-lreg
4321 Dump after local register allocation, to @file{@var{file}.138r.lreg}.
4324 @itemx -fdump-rtl-loop2
4326 @opindex fdump-rtl-loop2
4327 @option{-dL} and @option{-fdump-rtl-loop2} enable dumping after the
4328 loop optimization pass, to @file{@var{file}.119r.loop2},
4329 @file{@var{file}.120r.loop2_init},
4330 @file{@var{file}.121r.loop2_invariant}, and
4331 @file{@var{file}.125r.loop2_done}.
4334 @itemx -fdump-rtl-sms
4336 @opindex fdump-rtl-sms
4337 Dump after modulo scheduling, to @file{@var{file}.136r.sms}.
4340 @itemx -fdump-rtl-mach
4342 @opindex fdump-rtl-mach
4343 Dump after performing the machine dependent reorganization pass, to
4344 @file{@var{file}.155r.mach}.
4347 @itemx -fdump-rtl-rnreg
4349 @opindex fdump-rtl-rnreg
4350 Dump after register renumbering, to @file{@var{file}.147r.rnreg}.
4353 @itemx -fdump-rtl-regmove
4355 @opindex fdump-rtl-regmove
4356 Dump after the register move pass, to @file{@var{file}.132r.regmove}.
4359 @itemx -fdump-rtl-postreload
4361 @opindex fdump-rtl-postreload
4362 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
4365 @itemx -fdump-rtl-expand
4367 @opindex fdump-rtl-expand
4368 Dump after RTL generation, to @file{@var{file}.104r.expand}.
4371 @itemx -fdump-rtl-sched2
4373 @opindex fdump-rtl-sched2
4374 Dump after the second scheduling pass, to @file{@var{file}.149r.sched2}.
4377 @itemx -fdump-rtl-cse
4379 @opindex fdump-rtl-cse
4380 Dump after CSE (including the jump optimization that sometimes follows
4381 CSE), to @file{@var{file}.113r.cse}.
4384 @itemx -fdump-rtl-sched1
4386 @opindex fdump-rtl-sched1
4387 Dump after the first scheduling pass, to @file{@var{file}.136r.sched1}.
4390 @itemx -fdump-rtl-cse2
4392 @opindex fdump-rtl-cse2
4393 Dump after the second CSE pass (including the jump optimization that
4394 sometimes follows CSE), to @file{@var{file}.127r.cse2}.
4397 @itemx -fdump-rtl-tracer
4399 @opindex fdump-rtl-tracer
4400 Dump after running tracer, to @file{@var{file}.118r.tracer}.
4403 @itemx -fdump-rtl-vpt
4404 @itemx -fdump-rtl-vartrack
4406 @opindex fdump-rtl-vpt
4407 @opindex fdump-rtl-vartrack
4408 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
4409 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
4410 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
4411 to @file{@var{file}.154r.vartrack}.
4414 @itemx -fdump-rtl-flow2
4416 @opindex fdump-rtl-flow2
4417 Dump after the second flow pass, to @file{@var{file}.142r.flow2}.
4420 @itemx -fdump-rtl-peephole2
4422 @opindex fdump-rtl-peephole2
4423 Dump after the peephole pass, to @file{@var{file}.145r.peephole2}.
4426 @itemx -fdump-rtl-web
4428 @opindex fdump-rtl-web
4429 Dump after live range splitting, to @file{@var{file}.126r.web}.
4432 @itemx -fdump-rtl-all
4434 @opindex fdump-rtl-all
4435 Produce all the dumps listed above.
4439 Produce a core dump whenever an error occurs.
4443 Print statistics on memory usage, at the end of the run, to
4448 Annotate the assembler output with a comment indicating which
4449 pattern and alternative was used. The length of each instruction is
4454 Dump the RTL in the assembler output as a comment before each instruction.
4455 Also turns on @option{-dp} annotation.
4459 For each of the other indicated dump files (either with @option{-d} or
4460 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
4461 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
4465 Just generate RTL for a function instead of compiling it. Usually used
4466 with @samp{r} (@option{-fdump-rtl-expand}).
4470 Dump debugging information during parsing, to standard error.
4474 @opindex fdump-noaddr
4475 When doing debugging dumps (see @option{-d} option above), suppress
4476 address output. This makes it more feasible to use diff on debugging
4477 dumps for compiler invocations with different compiler binaries and/or
4478 different text / bss / data / heap / stack / dso start locations.
4480 @item -fdump-unnumbered
4481 @opindex fdump-unnumbered
4482 When doing debugging dumps (see @option{-d} option above), suppress instruction
4483 numbers, line number note and address output. This makes it more feasible to
4484 use diff on debugging dumps for compiler invocations with different
4485 options, in particular with and without @option{-g}.
4487 @item -fdump-translation-unit @r{(C++ only)}
4488 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
4489 @opindex fdump-translation-unit
4490 Dump a representation of the tree structure for the entire translation
4491 unit to a file. The file name is made by appending @file{.tu} to the
4492 source file name. If the @samp{-@var{options}} form is used, @var{options}
4493 controls the details of the dump as described for the
4494 @option{-fdump-tree} options.
4496 @item -fdump-class-hierarchy @r{(C++ only)}
4497 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
4498 @opindex fdump-class-hierarchy
4499 Dump a representation of each class's hierarchy and virtual function
4500 table layout to a file. The file name is made by appending @file{.class}
4501 to the source file name. If the @samp{-@var{options}} form is used,
4502 @var{options} controls the details of the dump as described for the
4503 @option{-fdump-tree} options.
4505 @item -fdump-ipa-@var{switch}
4507 Control the dumping at various stages of inter-procedural analysis
4508 language tree to a file. The file name is generated by appending a switch
4509 specific suffix to the source file name. The following dumps are possible:
4513 Enables all inter-procedural analysis dumps; currently the only produced
4514 dump is the @samp{cgraph} dump.
4517 Dumps information about call-graph optimization, unused function removal,
4518 and inlining decisions.
4521 @item -fdump-tree-@var{switch}
4522 @itemx -fdump-tree-@var{switch}-@var{options}
4524 Control the dumping at various stages of processing the intermediate
4525 language tree to a file. The file name is generated by appending a switch
4526 specific suffix to the source file name. If the @samp{-@var{options}}
4527 form is used, @var{options} is a list of @samp{-} separated options that
4528 control the details of the dump. Not all options are applicable to all
4529 dumps, those which are not meaningful will be ignored. The following
4530 options are available
4534 Print the address of each node. Usually this is not meaningful as it
4535 changes according to the environment and source file. Its primary use
4536 is for tying up a dump file with a debug environment.
4538 Inhibit dumping of members of a scope or body of a function merely
4539 because that scope has been reached. Only dump such items when they
4540 are directly reachable by some other path. When dumping pretty-printed
4541 trees, this option inhibits dumping the bodies of control structures.
4543 Print a raw representation of the tree. By default, trees are
4544 pretty-printed into a C-like representation.
4546 Enable more detailed dumps (not honored by every dump option).
4548 Enable dumping various statistics about the pass (not honored by every dump
4551 Enable showing basic block boundaries (disabled in raw dumps).
4553 Enable showing virtual operands for every statement.
4555 Enable showing line numbers for statements.
4557 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4559 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4562 The following tree dumps are possible:
4566 Dump before any tree based optimization, to @file{@var{file}.original}.
4569 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4572 Dump after function inlining, to @file{@var{file}.inlined}.
4575 @opindex fdump-tree-gimple
4576 Dump each function before and after the gimplification pass to a file. The
4577 file name is made by appending @file{.gimple} to the source file name.
4580 @opindex fdump-tree-cfg
4581 Dump the control flow graph of each function to a file. The file name is
4582 made by appending @file{.cfg} to the source file name.
4585 @opindex fdump-tree-vcg
4586 Dump the control flow graph of each function to a file in VCG format. The
4587 file name is made by appending @file{.vcg} to the source file name. Note
4588 that if the file contains more than one function, the generated file cannot
4589 be used directly by VCG@. You will need to cut and paste each function's
4590 graph into its own separate file first.
4593 @opindex fdump-tree-ch
4594 Dump each function after copying loop headers. The file name is made by
4595 appending @file{.ch} to the source file name.
4598 @opindex fdump-tree-ssa
4599 Dump SSA related information to a file. The file name is made by appending
4600 @file{.ssa} to the source file name.
4603 @opindex fdump-tree-salias
4604 Dump structure aliasing variable information to a file. This file name
4605 is made by appending @file{.salias} to the source file name.
4608 @opindex fdump-tree-alias
4609 Dump aliasing information for each function. The file name is made by
4610 appending @file{.alias} to the source file name.
4613 @opindex fdump-tree-ccp
4614 Dump each function after CCP@. The file name is made by appending
4615 @file{.ccp} to the source file name.
4618 @opindex fdump-tree-storeccp
4619 Dump each function after STORE-CCP. The file name is made by appending
4620 @file{.storeccp} to the source file name.
4623 @opindex fdump-tree-pre
4624 Dump trees after partial redundancy elimination. The file name is made
4625 by appending @file{.pre} to the source file name.
4628 @opindex fdump-tree-fre
4629 Dump trees after full redundancy elimination. The file name is made
4630 by appending @file{.fre} to the source file name.
4633 @opindex fdump-tree-copyprop
4634 Dump trees after copy propagation. The file name is made
4635 by appending @file{.copyprop} to the source file name.
4637 @item store_copyprop
4638 @opindex fdump-tree-store_copyprop
4639 Dump trees after store copy-propagation. The file name is made
4640 by appending @file{.store_copyprop} to the source file name.
4643 @opindex fdump-tree-dce
4644 Dump each function after dead code elimination. The file name is made by
4645 appending @file{.dce} to the source file name.
4648 @opindex fdump-tree-mudflap
4649 Dump each function after adding mudflap instrumentation. The file name is
4650 made by appending @file{.mudflap} to the source file name.
4653 @opindex fdump-tree-sra
4654 Dump each function after performing scalar replacement of aggregates. The
4655 file name is made by appending @file{.sra} to the source file name.
4658 @opindex fdump-tree-sink
4659 Dump each function after performing code sinking. The file name is made
4660 by appending @file{.sink} to the source file name.
4663 @opindex fdump-tree-dom
4664 Dump each function after applying dominator tree optimizations. The file
4665 name is made by appending @file{.dom} to the source file name.
4668 @opindex fdump-tree-dse
4669 Dump each function after applying dead store elimination. The file
4670 name is made by appending @file{.dse} to the source file name.
4673 @opindex fdump-tree-phiopt
4674 Dump each function after optimizing PHI nodes into straightline code. The file
4675 name is made by appending @file{.phiopt} to the source file name.
4678 @opindex fdump-tree-forwprop
4679 Dump each function after forward propagating single use variables. The file
4680 name is made by appending @file{.forwprop} to the source file name.
4683 @opindex fdump-tree-copyrename
4684 Dump each function after applying the copy rename optimization. The file
4685 name is made by appending @file{.copyrename} to the source file name.
4688 @opindex fdump-tree-nrv
4689 Dump each function after applying the named return value optimization on
4690 generic trees. The file name is made by appending @file{.nrv} to the source
4694 @opindex fdump-tree-vect
4695 Dump each function after applying vectorization of loops. The file name is
4696 made by appending @file{.vect} to the source file name.
4699 @opindex fdump-tree-vrp
4700 Dump each function after Value Range Propagation (VRP). The file name
4701 is made by appending @file{.vrp} to the source file name.
4704 @opindex fdump-tree-all
4705 Enable all the available tree dumps with the flags provided in this option.
4708 @item -ftree-vectorizer-verbose=@var{n}
4709 @opindex ftree-vectorizer-verbose
4710 This option controls the amount of debugging output the vectorizer prints.
4711 This information is written to standard error, unless
4712 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
4713 in which case it is output to the usual dump listing file, @file{.vect}.
4714 For @var{n}=0 no diagnostic information is reported.
4715 If @var{n}=1 the vectorizer reports each loop that got vectorized,
4716 and the total number of loops that got vectorized.
4717 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
4718 the first analysis phase (vect_analyze_loop_form) - i.e. countable,
4719 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
4720 level that @option{-fdump-tree-vect-stats} uses.
4721 Higher verbosity levels mean either more information dumped for each
4722 reported loop, or same amount of information reported for more loops:
4723 If @var{n}=3, alignment related information is added to the reports.
4724 If @var{n}=4, data-references related information (e.g. memory dependences,
4725 memory access-patterns) is added to the reports.
4726 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
4727 that did not pass the first analysis phase (i.e. may not be countable, or
4728 may have complicated control-flow).
4729 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
4730 For @var{n}=7, all the information the vectorizer generates during its
4731 analysis and transformation is reported. This is the same verbosity level
4732 that @option{-fdump-tree-vect-details} uses.
4734 @item -frandom-seed=@var{string}
4735 @opindex frandom-string
4736 This option provides a seed that GCC uses when it would otherwise use
4737 random numbers. It is used to generate certain symbol names
4738 that have to be different in every compiled file. It is also used to
4739 place unique stamps in coverage data files and the object files that
4740 produce them. You can use the @option{-frandom-seed} option to produce
4741 reproducibly identical object files.
4743 The @var{string} should be different for every file you compile.
4745 @item -fsched-verbose=@var{n}
4746 @opindex fsched-verbose
4747 On targets that use instruction scheduling, this option controls the
4748 amount of debugging output the scheduler prints. This information is
4749 written to standard error, unless @option{-dS} or @option{-dR} is
4750 specified, in which case it is output to the usual dump
4751 listing file, @file{.sched} or @file{.sched2} respectively. However
4752 for @var{n} greater than nine, the output is always printed to standard
4755 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4756 same information as @option{-dRS}. For @var{n} greater than one, it
4757 also output basic block probabilities, detailed ready list information
4758 and unit/insn info. For @var{n} greater than two, it includes RTL
4759 at abort point, control-flow and regions info. And for @var{n} over
4760 four, @option{-fsched-verbose} also includes dependence info.
4764 Store the usual ``temporary'' intermediate files permanently; place them
4765 in the current directory and name them based on the source file. Thus,
4766 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4767 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4768 preprocessed @file{foo.i} output file even though the compiler now
4769 normally uses an integrated preprocessor.
4771 When used in combination with the @option{-x} command line option,
4772 @option{-save-temps} is sensible enough to avoid over writing an
4773 input source file with the same extension as an intermediate file.
4774 The corresponding intermediate file may be obtained by renaming the
4775 source file before using @option{-save-temps}.
4779 Report the CPU time taken by each subprocess in the compilation
4780 sequence. For C source files, this is the compiler proper and assembler
4781 (plus the linker if linking is done). The output looks like this:
4788 The first number on each line is the ``user time'', that is time spent
4789 executing the program itself. The second number is ``system time'',
4790 time spent executing operating system routines on behalf of the program.
4791 Both numbers are in seconds.
4793 @item -fvar-tracking
4794 @opindex fvar-tracking
4795 Run variable tracking pass. It computes where variables are stored at each
4796 position in code. Better debugging information is then generated
4797 (if the debugging information format supports this information).
4799 It is enabled by default when compiling with optimization (@option{-Os},
4800 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
4801 the debug info format supports it.
4803 @item -print-file-name=@var{library}
4804 @opindex print-file-name
4805 Print the full absolute name of the library file @var{library} that
4806 would be used when linking---and don't do anything else. With this
4807 option, GCC does not compile or link anything; it just prints the
4810 @item -print-multi-directory
4811 @opindex print-multi-directory
4812 Print the directory name corresponding to the multilib selected by any
4813 other switches present in the command line. This directory is supposed
4814 to exist in @env{GCC_EXEC_PREFIX}.
4816 @item -print-multi-lib
4817 @opindex print-multi-lib
4818 Print the mapping from multilib directory names to compiler switches
4819 that enable them. The directory name is separated from the switches by
4820 @samp{;}, and each switch starts with an @samp{@@} instead of the
4821 @samp{-}, without spaces between multiple switches. This is supposed to
4822 ease shell-processing.
4824 @item -print-prog-name=@var{program}
4825 @opindex print-prog-name
4826 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
4828 @item -print-libgcc-file-name
4829 @opindex print-libgcc-file-name
4830 Same as @option{-print-file-name=libgcc.a}.
4832 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
4833 but you do want to link with @file{libgcc.a}. You can do
4836 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
4839 @item -print-search-dirs
4840 @opindex print-search-dirs
4841 Print the name of the configured installation directory and a list of
4842 program and library directories @command{gcc} will search---and don't do anything else.
4844 This is useful when @command{gcc} prints the error message
4845 @samp{installation problem, cannot exec cpp0: No such file or directory}.
4846 To resolve this you either need to put @file{cpp0} and the other compiler
4847 components where @command{gcc} expects to find them, or you can set the environment
4848 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
4849 Don't forget the trailing @samp{/}.
4850 @xref{Environment Variables}.
4852 @item -print-sysroot-headers-suffix
4853 @opindex print-sysroot-headers-suffix
4854 Print the suffix added to the target sysroot when searching for
4855 headers, or give an error if the compiler is not configured with such
4856 a suffix---and don't do anything else.
4859 @opindex dumpmachine
4860 Print the compiler's target machine (for example,
4861 @samp{i686-pc-linux-gnu})---and don't do anything else.
4864 @opindex dumpversion
4865 Print the compiler version (for example, @samp{3.0})---and don't do
4870 Print the compiler's built-in specs---and don't do anything else. (This
4871 is used when GCC itself is being built.) @xref{Spec Files}.
4873 @item -feliminate-unused-debug-types
4874 @opindex feliminate-unused-debug-types
4875 Normally, when producing DWARF2 output, GCC will emit debugging
4876 information for all types declared in a compilation
4877 unit, regardless of whether or not they are actually used
4878 in that compilation unit. Sometimes this is useful, such as
4879 if, in the debugger, you want to cast a value to a type that is
4880 not actually used in your program (but is declared). More often,
4881 however, this results in a significant amount of wasted space.
4882 With this option, GCC will avoid producing debug symbol output
4883 for types that are nowhere used in the source file being compiled.
4886 @node Optimize Options
4887 @section Options That Control Optimization
4888 @cindex optimize options
4889 @cindex options, optimization
4891 These options control various sorts of optimizations.
4893 Without any optimization option, the compiler's goal is to reduce the
4894 cost of compilation and to make debugging produce the expected
4895 results. Statements are independent: if you stop the program with a
4896 breakpoint between statements, you can then assign a new value to any
4897 variable or change the program counter to any other statement in the
4898 function and get exactly the results you would expect from the source
4901 Turning on optimization flags makes the compiler attempt to improve
4902 the performance and/or code size at the expense of compilation time
4903 and possibly the ability to debug the program.
4905 The compiler performs optimization based on the knowledge it has of
4906 the program. Optimization levels @option{-O} and above, in
4907 particular, enable @emph{unit-at-a-time} mode, which allows the
4908 compiler to consider information gained from later functions in
4909 the file when compiling a function. Compiling multiple files at
4910 once to a single output file in @emph{unit-at-a-time} mode allows
4911 the compiler to use information gained from all of the files when
4912 compiling each of them.
4914 Not all optimizations are controlled directly by a flag. Only
4915 optimizations that have a flag are listed.
4922 Optimize. Optimizing compilation takes somewhat more time, and a lot
4923 more memory for a large function.
4925 With @option{-O}, the compiler tries to reduce code size and execution
4926 time, without performing any optimizations that take a great deal of
4929 @option{-O} turns on the following optimization flags:
4930 @gccoptlist{-fdefer-pop @gol
4931 -fdelayed-branch @gol
4932 -fguess-branch-probability @gol
4933 -fcprop-registers @gol
4934 -fif-conversion @gol
4935 -fif-conversion2 @gol
4936 -fsplit-wide-types @gol
4939 -ftree-dominator-opts @gol
4943 -ftree-copyrename @gol
4946 -funit-at-a-time @gol
4949 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
4950 where doing so does not interfere with debugging.
4954 Optimize even more. GCC performs nearly all supported optimizations
4955 that do not involve a space-speed tradeoff. The compiler does not
4956 perform loop unrolling or function inlining when you specify @option{-O2}.
4957 As compared to @option{-O}, this option increases both compilation time
4958 and the performance of the generated code.
4960 @option{-O2} turns on all optimization flags specified by @option{-O}. It
4961 also turns on the following optimization flags:
4962 @gccoptlist{-fthread-jumps @gol
4964 -foptimize-sibling-calls @gol
4965 -fcse-follow-jumps -fcse-skip-blocks @gol
4966 -fgcse -fgcse-lm @gol
4967 -fexpensive-optimizations @gol
4968 -frerun-cse-after-loop @gol
4971 -fschedule-insns -fschedule-insns2 @gol
4972 -fsched-interblock -fsched-spec @gol
4974 -fstrict-aliasing -fstrict-overflow @gol
4975 -fdelete-null-pointer-checks @gol
4976 -freorder-blocks -freorder-functions @gol
4977 -falign-functions -falign-jumps @gol
4978 -falign-loops -falign-labels @gol
4982 Please note the warning under @option{-fgcse} about
4983 invoking @option{-O2} on programs that use computed gotos.
4987 Optimize yet more. @option{-O3} turns on all optimizations specified by
4988 @option{-O2} and also turns on the @option{-finline-functions},
4989 @option{-funswitch-loops} and @option{-fgcse-after-reload} options.
4993 Reduce compilation time and make debugging produce the expected
4994 results. This is the default.
4998 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
4999 do not typically increase code size. It also performs further
5000 optimizations designed to reduce code size.
5002 @option{-Os} disables the following optimization flags:
5003 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5004 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5005 -fprefetch-loop-arrays -ftree-vect-loop-version}
5007 If you use multiple @option{-O} options, with or without level numbers,
5008 the last such option is the one that is effective.
5011 Options of the form @option{-f@var{flag}} specify machine-independent
5012 flags. Most flags have both positive and negative forms; the negative
5013 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5014 below, only one of the forms is listed---the one you typically will
5015 use. You can figure out the other form by either removing @samp{no-}
5018 The following options control specific optimizations. They are either
5019 activated by @option{-O} options or are related to ones that are. You
5020 can use the following flags in the rare cases when ``fine-tuning'' of
5021 optimizations to be performed is desired.
5024 @item -fno-default-inline
5025 @opindex fno-default-inline
5026 Do not make member functions inline by default merely because they are
5027 defined inside the class scope (C++ only). Otherwise, when you specify
5028 @w{@option{-O}}, member functions defined inside class scope are compiled
5029 inline by default; i.e., you don't need to add @samp{inline} in front of
5030 the member function name.
5032 @item -fno-defer-pop
5033 @opindex fno-defer-pop
5034 Always pop the arguments to each function call as soon as that function
5035 returns. For machines which must pop arguments after a function call,
5036 the compiler normally lets arguments accumulate on the stack for several
5037 function calls and pops them all at once.
5039 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5042 @opindex fforce-addr
5043 Force memory address constants to be copied into registers before
5044 doing arithmetic on them.
5046 @item -fforward-propagate
5047 @opindex fforward-propagate
5048 Perform a forward propagation pass on RTL. The pass tries to combine two
5049 instructions and checks if the result can be simplified. If loop unrolling
5050 is active, two passes are performed and the second is scheduled after
5053 This option is enabled by default at optimization levels @option{-O2},
5054 @option{-O3}, @option{-Os}.
5056 @item -fomit-frame-pointer
5057 @opindex fomit-frame-pointer
5058 Don't keep the frame pointer in a register for functions that
5059 don't need one. This avoids the instructions to save, set up and
5060 restore frame pointers; it also makes an extra register available
5061 in many functions. @strong{It also makes debugging impossible on
5064 On some machines, such as the VAX, this flag has no effect, because
5065 the standard calling sequence automatically handles the frame pointer
5066 and nothing is saved by pretending it doesn't exist. The
5067 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5068 whether a target machine supports this flag. @xref{Registers,,Register
5069 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5071 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5073 @item -foptimize-sibling-calls
5074 @opindex foptimize-sibling-calls
5075 Optimize sibling and tail recursive calls.
5077 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5081 Don't pay attention to the @code{inline} keyword. Normally this option
5082 is used to keep the compiler from expanding any functions inline.
5083 Note that if you are not optimizing, no functions can be expanded inline.
5085 @item -finline-functions
5086 @opindex finline-functions
5087 Integrate all simple functions into their callers. The compiler
5088 heuristically decides which functions are simple enough to be worth
5089 integrating in this way.
5091 If all calls to a given function are integrated, and the function is
5092 declared @code{static}, then the function is normally not output as
5093 assembler code in its own right.
5095 Enabled at level @option{-O3}.
5097 @item -finline-functions-called-once
5098 @opindex finline-functions-called-once
5099 Consider all @code{static} functions called once for inlining into their
5100 caller even if they are not marked @code{inline}. If a call to a given
5101 function is integrated, then the function is not output as assembler code
5104 Enabled if @option{-funit-at-a-time} is enabled.
5106 @item -fearly-inlining
5107 @opindex fearly-inlining
5108 Inline functions marked by @code{always_inline} and functions whose body seems
5109 smaller than the function call overhead early before doing
5110 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5111 makes profiling significantly cheaper and usually inlining faster on programs
5112 having large chains of nested wrapper functions.
5116 @item -finline-limit=@var{n}
5117 @opindex finline-limit
5118 By default, GCC limits the size of functions that can be inlined. This flag
5119 allows the control of this limit for functions that are explicitly marked as
5120 inline (i.e., marked with the inline keyword or defined within the class
5121 definition in c++). @var{n} is the size of functions that can be inlined in
5122 number of pseudo instructions (not counting parameter handling). The default
5123 value of @var{n} is 600.
5124 Increasing this value can result in more inlined code at
5125 the cost of compilation time and memory consumption. Decreasing usually makes
5126 the compilation faster and less code will be inlined (which presumably
5127 means slower programs). This option is particularly useful for programs that
5128 use inlining heavily such as those based on recursive templates with C++.
5130 Inlining is actually controlled by a number of parameters, which may be
5131 specified individually by using @option{--param @var{name}=@var{value}}.
5132 The @option{-finline-limit=@var{n}} option sets some of these parameters
5136 @item max-inline-insns-single
5137 is set to @var{n}/2.
5138 @item max-inline-insns-auto
5139 is set to @var{n}/2.
5140 @item min-inline-insns
5141 is set to 130 or @var{n}/4, whichever is smaller.
5142 @item max-inline-insns-rtl
5146 See below for a documentation of the individual
5147 parameters controlling inlining.
5149 @emph{Note:} pseudo instruction represents, in this particular context, an
5150 abstract measurement of function's size. In no way does it represent a count
5151 of assembly instructions and as such its exact meaning might change from one
5152 release to an another.
5154 @item -fkeep-inline-functions
5155 @opindex fkeep-inline-functions
5156 In C, emit @code{static} functions that are declared @code{inline}
5157 into the object file, even if the function has been inlined into all
5158 of its callers. This switch does not affect functions using the
5159 @code{extern inline} extension in GNU C89@. In C++, emit any and all
5160 inline functions into the object file.
5162 @item -fkeep-static-consts
5163 @opindex fkeep-static-consts
5164 Emit variables declared @code{static const} when optimization isn't turned
5165 on, even if the variables aren't referenced.
5167 GCC enables this option by default. If you want to force the compiler to
5168 check if the variable was referenced, regardless of whether or not
5169 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5171 @item -fmerge-constants
5172 Attempt to merge identical constants (string constants and floating point
5173 constants) across compilation units.
5175 This option is the default for optimized compilation if the assembler and
5176 linker support it. Use @option{-fno-merge-constants} to inhibit this
5179 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5181 @item -fmerge-all-constants
5182 Attempt to merge identical constants and identical variables.
5184 This option implies @option{-fmerge-constants}. In addition to
5185 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5186 arrays or initialized constant variables with integral or floating point
5187 types. Languages like C or C++ require each non-automatic variable to
5188 have distinct location, so using this option will result in non-conforming
5191 @item -fmodulo-sched
5192 @opindex fmodulo-sched
5193 Perform swing modulo scheduling immediately before the first scheduling
5194 pass. This pass looks at innermost loops and reorders their
5195 instructions by overlapping different iterations.
5197 @item -fno-branch-count-reg
5198 @opindex fno-branch-count-reg
5199 Do not use ``decrement and branch'' instructions on a count register,
5200 but instead generate a sequence of instructions that decrement a
5201 register, compare it against zero, then branch based upon the result.
5202 This option is only meaningful on architectures that support such
5203 instructions, which include x86, PowerPC, IA-64 and S/390.
5205 The default is @option{-fbranch-count-reg}.
5207 @item -fno-function-cse
5208 @opindex fno-function-cse
5209 Do not put function addresses in registers; make each instruction that
5210 calls a constant function contain the function's address explicitly.
5212 This option results in less efficient code, but some strange hacks
5213 that alter the assembler output may be confused by the optimizations
5214 performed when this option is not used.
5216 The default is @option{-ffunction-cse}
5218 @item -fno-zero-initialized-in-bss
5219 @opindex fno-zero-initialized-in-bss
5220 If the target supports a BSS section, GCC by default puts variables that
5221 are initialized to zero into BSS@. This can save space in the resulting
5224 This option turns off this behavior because some programs explicitly
5225 rely on variables going to the data section. E.g., so that the
5226 resulting executable can find the beginning of that section and/or make
5227 assumptions based on that.
5229 The default is @option{-fzero-initialized-in-bss}.
5231 @item -fbounds-check
5232 @opindex fbounds-check
5233 For front-ends that support it, generate additional code to check that
5234 indices used to access arrays are within the declared range. This is
5235 currently only supported by the Java and Fortran front-ends, where
5236 this option defaults to true and false respectively.
5238 @item -fmudflap -fmudflapth -fmudflapir
5242 @cindex bounds checking
5244 For front-ends that support it (C and C++), instrument all risky
5245 pointer/array dereferencing operations, some standard library
5246 string/heap functions, and some other associated constructs with
5247 range/validity tests. Modules so instrumented should be immune to
5248 buffer overflows, invalid heap use, and some other classes of C/C++
5249 programming errors. The instrumentation relies on a separate runtime
5250 library (@file{libmudflap}), which will be linked into a program if
5251 @option{-fmudflap} is given at link time. Run-time behavior of the
5252 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
5253 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
5256 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
5257 link if your program is multi-threaded. Use @option{-fmudflapir}, in
5258 addition to @option{-fmudflap} or @option{-fmudflapth}, if
5259 instrumentation should ignore pointer reads. This produces less
5260 instrumentation (and therefore faster execution) and still provides
5261 some protection against outright memory corrupting writes, but allows
5262 erroneously read data to propagate within a program.
5264 @item -fthread-jumps
5265 @opindex fthread-jumps
5266 Perform optimizations where we check to see if a jump branches to a
5267 location where another comparison subsumed by the first is found. If
5268 so, the first branch is redirected to either the destination of the
5269 second branch or a point immediately following it, depending on whether
5270 the condition is known to be true or false.
5272 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5274 @item -fsplit-wide-types
5275 @opindex fsplit-wide-types
5276 When using a type that occupies multiple registers, such as @code{long
5277 long} on a 32-bit system, split the registers apart and allocate them
5278 independently. This normally generates better code for those types,
5279 but may make debugging more difficult.
5281 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
5284 @item -fcse-follow-jumps
5285 @opindex fcse-follow-jumps
5286 In common subexpression elimination, scan through jump instructions
5287 when the target of the jump is not reached by any other path. For
5288 example, when CSE encounters an @code{if} statement with an
5289 @code{else} clause, CSE will follow the jump when the condition
5292 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5294 @item -fcse-skip-blocks
5295 @opindex fcse-skip-blocks
5296 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
5297 follow jumps which conditionally skip over blocks. When CSE
5298 encounters a simple @code{if} statement with no else clause,
5299 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
5300 body of the @code{if}.
5302 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5304 @item -frerun-cse-after-loop
5305 @opindex frerun-cse-after-loop
5306 Re-run common subexpression elimination after loop optimizations has been
5309 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5313 Perform a global common subexpression elimination pass.
5314 This pass also performs global constant and copy propagation.
5316 @emph{Note:} When compiling a program using computed gotos, a GCC
5317 extension, you may get better runtime performance if you disable
5318 the global common subexpression elimination pass by adding
5319 @option{-fno-gcse} to the command line.
5321 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5325 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
5326 attempt to move loads which are only killed by stores into themselves. This
5327 allows a loop containing a load/store sequence to be changed to a load outside
5328 the loop, and a copy/store within the loop.
5330 Enabled by default when gcse is enabled.
5334 When @option{-fgcse-sm} is enabled, a store motion pass is run after
5335 global common subexpression elimination. This pass will attempt to move
5336 stores out of loops. When used in conjunction with @option{-fgcse-lm},
5337 loops containing a load/store sequence can be changed to a load before
5338 the loop and a store after the loop.
5340 Not enabled at any optimization level.
5344 When @option{-fgcse-las} is enabled, the global common subexpression
5345 elimination pass eliminates redundant loads that come after stores to the
5346 same memory location (both partial and full redundancies).
5348 Not enabled at any optimization level.
5350 @item -fgcse-after-reload
5351 @opindex fgcse-after-reload
5352 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
5353 pass is performed after reload. The purpose of this pass is to cleanup
5356 @item -funsafe-loop-optimizations
5357 @opindex funsafe-loop-optimizations
5358 If given, the loop optimizer will assume that loop indices do not
5359 overflow, and that the loops with nontrivial exit condition are not
5360 infinite. This enables a wider range of loop optimizations even if
5361 the loop optimizer itself cannot prove that these assumptions are valid.
5362 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
5363 if it finds this kind of loop.
5365 @item -fcrossjumping
5366 @opindex crossjumping
5367 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
5368 resulting code may or may not perform better than without cross-jumping.
5370 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5372 @item -fif-conversion
5373 @opindex if-conversion
5374 Attempt to transform conditional jumps into branch-less equivalents. This
5375 include use of conditional moves, min, max, set flags and abs instructions, and
5376 some tricks doable by standard arithmetics. The use of conditional execution
5377 on chips where it is available is controlled by @code{if-conversion2}.
5379 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5381 @item -fif-conversion2
5382 @opindex if-conversion2
5383 Use conditional execution (where available) to transform conditional jumps into
5384 branch-less equivalents.
5386 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5388 @item -fdelete-null-pointer-checks
5389 @opindex fdelete-null-pointer-checks
5390 Use global dataflow analysis to identify and eliminate useless checks
5391 for null pointers. The compiler assumes that dereferencing a null
5392 pointer would have halted the program. If a pointer is checked after
5393 it has already been dereferenced, it cannot be null.
5395 In some environments, this assumption is not true, and programs can
5396 safely dereference null pointers. Use
5397 @option{-fno-delete-null-pointer-checks} to disable this optimization
5398 for programs which depend on that behavior.
5400 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5402 @item -fexpensive-optimizations
5403 @opindex fexpensive-optimizations
5404 Perform a number of minor optimizations that are relatively expensive.
5406 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5408 @item -foptimize-register-move
5410 @opindex foptimize-register-move
5412 Attempt to reassign register numbers in move instructions and as
5413 operands of other simple instructions in order to maximize the amount of
5414 register tying. This is especially helpful on machines with two-operand
5417 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
5420 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5422 @item -fdelayed-branch
5423 @opindex fdelayed-branch
5424 If supported for the target machine, attempt to reorder instructions
5425 to exploit instruction slots available after delayed branch
5428 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5430 @item -fschedule-insns
5431 @opindex fschedule-insns
5432 If supported for the target machine, attempt to reorder instructions to
5433 eliminate execution stalls due to required data being unavailable. This
5434 helps machines that have slow floating point or memory load instructions
5435 by allowing other instructions to be issued until the result of the load
5436 or floating point instruction is required.
5438 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5440 @item -fschedule-insns2
5441 @opindex fschedule-insns2
5442 Similar to @option{-fschedule-insns}, but requests an additional pass of
5443 instruction scheduling after register allocation has been done. This is
5444 especially useful on machines with a relatively small number of
5445 registers and where memory load instructions take more than one cycle.
5447 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5449 @item -fno-sched-interblock
5450 @opindex fno-sched-interblock
5451 Don't schedule instructions across basic blocks. This is normally
5452 enabled by default when scheduling before register allocation, i.e.@:
5453 with @option{-fschedule-insns} or at @option{-O2} or higher.
5455 @item -fno-sched-spec
5456 @opindex fno-sched-spec
5457 Don't allow speculative motion of non-load instructions. This is normally
5458 enabled by default when scheduling before register allocation, i.e.@:
5459 with @option{-fschedule-insns} or at @option{-O2} or higher.
5461 @item -fsched-spec-load
5462 @opindex fsched-spec-load
5463 Allow speculative motion of some load instructions. This only makes
5464 sense when scheduling before register allocation, i.e.@: with
5465 @option{-fschedule-insns} or at @option{-O2} or higher.
5467 @item -fsched-spec-load-dangerous
5468 @opindex fsched-spec-load-dangerous
5469 Allow speculative motion of more load instructions. This only makes
5470 sense when scheduling before register allocation, i.e.@: with
5471 @option{-fschedule-insns} or at @option{-O2} or higher.
5473 @item -fsched-stalled-insns=@var{n}
5474 @opindex fsched-stalled-insns
5475 Define how many insns (if any) can be moved prematurely from the queue
5476 of stalled insns into the ready list, during the second scheduling pass.
5478 @item -fsched-stalled-insns-dep=@var{n}
5479 @opindex fsched-stalled-insns-dep
5480 Define how many insn groups (cycles) will be examined for a dependency
5481 on a stalled insn that is candidate for premature removal from the queue
5482 of stalled insns. Has an effect only during the second scheduling pass,
5483 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
5485 @item -fsched2-use-superblocks
5486 @opindex fsched2-use-superblocks
5487 When scheduling after register allocation, do use superblock scheduling
5488 algorithm. Superblock scheduling allows motion across basic block boundaries
5489 resulting on faster schedules. This option is experimental, as not all machine
5490 descriptions used by GCC model the CPU closely enough to avoid unreliable
5491 results from the algorithm.
5493 This only makes sense when scheduling after register allocation, i.e.@: with
5494 @option{-fschedule-insns2} or at @option{-O2} or higher.
5496 @item -fsched2-use-traces
5497 @opindex fsched2-use-traces
5498 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
5499 allocation and additionally perform code duplication in order to increase the
5500 size of superblocks using tracer pass. See @option{-ftracer} for details on
5503 This mode should produce faster but significantly longer programs. Also
5504 without @option{-fbranch-probabilities} the traces constructed may not
5505 match the reality and hurt the performance. This only makes
5506 sense when scheduling after register allocation, i.e.@: with
5507 @option{-fschedule-insns2} or at @option{-O2} or higher.
5511 Eliminates redundant extension instructions and move the non redundant
5512 ones to optimal placement using LCM.
5514 @item -freschedule-modulo-scheduled-loops
5515 @opindex fscheduling-in-modulo-scheduled-loops
5516 The modulo scheduling comes before the traditional scheduling, if a loop was modulo scheduled
5517 we may want to prevent the later scheduling passes from changing its schedule, we use this
5518 option to control that.
5520 @item -fcaller-saves
5521 @opindex fcaller-saves
5522 Enable values to be allocated in registers that will be clobbered by
5523 function calls, by emitting extra instructions to save and restore the
5524 registers around such calls. Such allocation is done only when it
5525 seems to result in better code than would otherwise be produced.
5527 This option is always enabled by default on certain machines, usually
5528 those which have no call-preserved registers to use instead.
5530 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5533 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
5534 enabled by default at @option{-O2} and @option{-O3}.
5537 Perform Full Redundancy Elimination (FRE) on trees. The difference
5538 between FRE and PRE is that FRE only considers expressions
5539 that are computed on all paths leading to the redundant computation.
5540 This analysis is faster than PRE, though it exposes fewer redundancies.
5541 This flag is enabled by default at @option{-O} and higher.
5543 @item -ftree-copy-prop
5544 Perform copy propagation on trees. This pass eliminates unnecessary
5545 copy operations. This flag is enabled by default at @option{-O} and
5548 @item -ftree-store-copy-prop
5549 Perform copy propagation of memory loads and stores. This pass
5550 eliminates unnecessary copy operations in memory references
5551 (structures, global variables, arrays, etc). This flag is enabled by
5552 default at @option{-O2} and higher.
5555 Perform structural alias analysis on trees. This flag
5556 is enabled by default at @option{-O} and higher.
5559 Perform interprocedural pointer analysis.
5562 Perform forward store motion on trees. This flag is
5563 enabled by default at @option{-O} and higher.
5566 Perform sparse conditional constant propagation (CCP) on trees. This
5567 pass only operates on local scalar variables and is enabled by default
5568 at @option{-O} and higher.
5570 @item -ftree-store-ccp
5571 Perform sparse conditional constant propagation (CCP) on trees. This
5572 pass operates on both local scalar variables and memory stores and
5573 loads (global variables, structures, arrays, etc). This flag is
5574 enabled by default at @option{-O2} and higher.
5577 Perform dead code elimination (DCE) on trees. This flag is enabled by
5578 default at @option{-O} and higher.
5580 @item -ftree-dominator-opts
5581 Perform a variety of simple scalar cleanups (constant/copy
5582 propagation, redundancy elimination, range propagation and expression
5583 simplification) based on a dominator tree traversal. This also
5584 performs jump threading (to reduce jumps to jumps). This flag is
5585 enabled by default at @option{-O} and higher.
5588 Perform loop header copying on trees. This is beneficial since it increases
5589 effectiveness of code motion optimizations. It also saves one jump. This flag
5590 is enabled by default at @option{-O} and higher. It is not enabled
5591 for @option{-Os}, since it usually increases code size.
5593 @item -ftree-loop-optimize
5594 Perform loop optimizations on trees. This flag is enabled by default
5595 at @option{-O} and higher.
5597 @item -ftree-loop-linear
5598 Perform linear loop transformations on tree. This flag can improve cache
5599 performance and allow further loop optimizations to take place.
5601 @item -fcheck-data-deps
5602 Compare the results of several data dependence analyzers. This option
5603 is used for debugging the data dependence analyzers.
5605 @item -ftree-loop-im
5606 Perform loop invariant motion on trees. This pass moves only invariants that
5607 would be hard to handle at RTL level (function calls, operations that expand to
5608 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5609 operands of conditions that are invariant out of the loop, so that we can use
5610 just trivial invariantness analysis in loop unswitching. The pass also includes
5613 @item -ftree-loop-ivcanon
5614 Create a canonical counter for number of iterations in the loop for that
5615 determining number of iterations requires complicated analysis. Later
5616 optimizations then may determine the number easily. Useful especially
5617 in connection with unrolling.
5620 Perform induction variable optimizations (strength reduction, induction
5621 variable merging and induction variable elimination) on trees.
5624 Perform scalar replacement of aggregates. This pass replaces structure
5625 references with scalars to prevent committing structures to memory too
5626 early. This flag is enabled by default at @option{-O} and higher.
5628 @item -ftree-copyrename
5629 Perform copy renaming on trees. This pass attempts to rename compiler
5630 temporaries to other variables at copy locations, usually resulting in
5631 variable names which more closely resemble the original variables. This flag
5632 is enabled by default at @option{-O} and higher.
5635 Perform temporary expression replacement during the SSA->normal phase. Single
5636 use/single def temporaries are replaced at their use location with their
5637 defining expression. This results in non-GIMPLE code, but gives the expanders
5638 much more complex trees to work on resulting in better RTL generation. This is
5639 enabled by default at @option{-O} and higher.
5641 @item -ftree-vectorize
5642 Perform loop vectorization on trees.
5644 @item -ftree-vect-loop-version
5645 @opindex ftree-vect-loop-version
5646 Perform loop versioning when doing loop vectorization on trees. When a loop
5647 appears to be vectorizable except that data alignment or data dependence cannot
5648 be determined at compile time then vectorized and non-vectorized versions of
5649 the loop are generated along with runtime checks for alignment or dependence
5650 to control which version is executed. This option is enabled by default
5651 except at level @option{-Os} where it is disabled.
5654 Perform Value Range Propagation on trees. This is similar to the
5655 constant propagation pass, but instead of values, ranges of values are
5656 propagated. This allows the optimizers to remove unnecessary range
5657 checks like array bound checks and null pointer checks. This is
5658 enabled by default at @option{-O2} and higher. Null pointer check
5659 elimination is only done if @option{-fdelete-null-pointer-checks} is
5664 Perform tail duplication to enlarge superblock size. This transformation
5665 simplifies the control flow of the function allowing other optimizations to do
5668 @item -funroll-loops
5669 @opindex funroll-loops
5670 Unroll loops whose number of iterations can be determined at compile
5671 time or upon entry to the loop. @option{-funroll-loops} implies
5672 @option{-frerun-cse-after-loop}. This option makes code larger,
5673 and may or may not make it run faster.
5675 @item -funroll-all-loops
5676 @opindex funroll-all-loops
5677 Unroll all loops, even if their number of iterations is uncertain when
5678 the loop is entered. This usually makes programs run more slowly.
5679 @option{-funroll-all-loops} implies the same options as
5680 @option{-funroll-loops},
5682 @item -fsplit-ivs-in-unroller
5683 @opindex fsplit-ivs-in-unroller
5684 Enables expressing of values of induction variables in later iterations
5685 of the unrolled loop using the value in the first iteration. This breaks
5686 long dependency chains, thus improving efficiency of the scheduling passes.
5688 Combination of @option{-fweb} and CSE is often sufficient to obtain the
5689 same effect. However in cases the loop body is more complicated than
5690 a single basic block, this is not reliable. It also does not work at all
5691 on some of the architectures due to restrictions in the CSE pass.
5693 This optimization is enabled by default.
5695 @item -fvariable-expansion-in-unroller
5696 @opindex fvariable-expansion-in-unroller
5697 With this option, the compiler will create multiple copies of some
5698 local variables when unrolling a loop which can result in superior code.
5700 @item -fprefetch-loop-arrays
5701 @opindex fprefetch-loop-arrays
5702 If supported by the target machine, generate instructions to prefetch
5703 memory to improve the performance of loops that access large arrays.
5705 This option may generate better or worse code; results are highly
5706 dependent on the structure of loops within the source code.
5708 Disabled at level @option{-Os}.
5711 @itemx -fno-peephole2
5712 @opindex fno-peephole
5713 @opindex fno-peephole2
5714 Disable any machine-specific peephole optimizations. The difference
5715 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
5716 are implemented in the compiler; some targets use one, some use the
5717 other, a few use both.
5719 @option{-fpeephole} is enabled by default.
5720 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5722 @item -fno-guess-branch-probability
5723 @opindex fno-guess-branch-probability
5724 Do not guess branch probabilities using heuristics.
5726 GCC will use heuristics to guess branch probabilities if they are
5727 not provided by profiling feedback (@option{-fprofile-arcs}). These
5728 heuristics are based on the control flow graph. If some branch probabilities
5729 are specified by @samp{__builtin_expect}, then the heuristics will be
5730 used to guess branch probabilities for the rest of the control flow graph,
5731 taking the @samp{__builtin_expect} info into account. The interactions
5732 between the heuristics and @samp{__builtin_expect} can be complex, and in
5733 some cases, it may be useful to disable the heuristics so that the effects
5734 of @samp{__builtin_expect} are easier to understand.
5736 The default is @option{-fguess-branch-probability} at levels
5737 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5739 @item -freorder-blocks
5740 @opindex freorder-blocks
5741 Reorder basic blocks in the compiled function in order to reduce number of
5742 taken branches and improve code locality.
5744 Enabled at levels @option{-O2}, @option{-O3}.
5746 @item -freorder-blocks-and-partition
5747 @opindex freorder-blocks-and-partition
5748 In addition to reordering basic blocks in the compiled function, in order
5749 to reduce number of taken branches, partitions hot and cold basic blocks
5750 into separate sections of the assembly and .o files, to improve
5751 paging and cache locality performance.
5753 This optimization is automatically turned off in the presence of
5754 exception handling, for linkonce sections, for functions with a user-defined
5755 section attribute and on any architecture that does not support named
5758 @item -freorder-functions
5759 @opindex freorder-functions
5760 Reorder functions in the object file in order to
5761 improve code locality. This is implemented by using special
5762 subsections @code{.text.hot} for most frequently executed functions and
5763 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
5764 the linker so object file format must support named sections and linker must
5765 place them in a reasonable way.
5767 Also profile feedback must be available in to make this option effective. See
5768 @option{-fprofile-arcs} for details.
5770 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5772 @item -fstrict-aliasing
5773 @opindex fstrict-aliasing
5774 Allows the compiler to assume the strictest aliasing rules applicable to
5775 the language being compiled. For C (and C++), this activates
5776 optimizations based on the type of expressions. In particular, an
5777 object of one type is assumed never to reside at the same address as an
5778 object of a different type, unless the types are almost the same. For
5779 example, an @code{unsigned int} can alias an @code{int}, but not a
5780 @code{void*} or a @code{double}. A character type may alias any other
5783 Pay special attention to code like this:
5796 The practice of reading from a different union member than the one most
5797 recently written to (called ``type-punning'') is common. Even with
5798 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
5799 is accessed through the union type. So, the code above will work as
5800 expected. However, this code might not:
5811 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5813 @item -fstrict-overflow
5814 @opindex fstrict-overflow
5815 Allow the compiler to assume strict signed overflow rules, depending
5816 on the language being compiled. For C (and C++) this means that
5817 overflow when doing arithmetic with signed numbers is undefined, which
5818 means that the compiler may assume that it will not happen. This
5819 permits various optimizations. For example, the compiler will assume
5820 that an expression like @code{i + 10 > i} will always be true for
5821 signed @code{i}. This assumption is only valid if signed overflow is
5822 undefined, as the expression is false if @code{i + 10} overflows when
5823 using twos complement arithmetic. When this option is in effect any
5824 attempt to determine whether an operation on signed numbers will
5825 overflow must be written carefully to not actually involve overflow.
5827 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
5828 that signed overflow is fully defined: it wraps. When
5829 @option{-fwrapv} is used, there is no difference between
5830 @option{-fstrict-overflow} and @option{-fno-strict-overflow}. With
5831 @option{-fwrapv} certain types of overflow are permitted. For
5832 example, if the compiler gets an overflow when doing arithmetic on
5833 constants, the overflowed value can still be used with
5834 @option{-fwrapv}, but not otherwise.
5836 The @option{-fstrict-overflow} option is enabled at levels
5837 @option{-O2}, @option{-O3}, @option{-Os}.
5839 @item -falign-functions
5840 @itemx -falign-functions=@var{n}
5841 @opindex falign-functions
5842 Align the start of functions to the next power-of-two greater than
5843 @var{n}, skipping up to @var{n} bytes. For instance,
5844 @option{-falign-functions=32} aligns functions to the next 32-byte
5845 boundary, but @option{-falign-functions=24} would align to the next
5846 32-byte boundary only if this can be done by skipping 23 bytes or less.
5848 @option{-fno-align-functions} and @option{-falign-functions=1} are
5849 equivalent and mean that functions will not be aligned.
5851 Some assemblers only support this flag when @var{n} is a power of two;
5852 in that case, it is rounded up.
5854 If @var{n} is not specified or is zero, use a machine-dependent default.
5856 Enabled at levels @option{-O2}, @option{-O3}.
5858 @item -falign-labels
5859 @itemx -falign-labels=@var{n}
5860 @opindex falign-labels
5861 Align all branch targets to a power-of-two boundary, skipping up to
5862 @var{n} bytes like @option{-falign-functions}. This option can easily
5863 make code slower, because it must insert dummy operations for when the
5864 branch target is reached in the usual flow of the code.
5866 @option{-fno-align-labels} and @option{-falign-labels=1} are
5867 equivalent and mean that labels will not be aligned.
5869 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
5870 are greater than this value, then their values are used instead.
5872 If @var{n} is not specified or is zero, use a machine-dependent default
5873 which is very likely to be @samp{1}, meaning no alignment.
5875 Enabled at levels @option{-O2}, @option{-O3}.
5878 @itemx -falign-loops=@var{n}
5879 @opindex falign-loops
5880 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
5881 like @option{-falign-functions}. The hope is that the loop will be
5882 executed many times, which will make up for any execution of the dummy
5885 @option{-fno-align-loops} and @option{-falign-loops=1} are
5886 equivalent and mean that loops will not be aligned.
5888 If @var{n} is not specified or is zero, use a machine-dependent default.
5890 Enabled at levels @option{-O2}, @option{-O3}.
5893 @itemx -falign-jumps=@var{n}
5894 @opindex falign-jumps
5895 Align branch targets to a power-of-two boundary, for branch targets
5896 where the targets can only be reached by jumping, skipping up to @var{n}
5897 bytes like @option{-falign-functions}. In this case, no dummy operations
5900 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
5901 equivalent and mean that loops will not be aligned.
5903 If @var{n} is not specified or is zero, use a machine-dependent default.
5905 Enabled at levels @option{-O2}, @option{-O3}.
5907 @item -funit-at-a-time
5908 @opindex funit-at-a-time
5909 Parse the whole compilation unit before starting to produce code.
5910 This allows some extra optimizations to take place but consumes
5911 more memory (in general). There are some compatibility issues
5912 with @emph{unit-at-a-time} mode:
5915 enabling @emph{unit-at-a-time} mode may change the order
5916 in which functions, variables, and top-level @code{asm} statements
5917 are emitted, and will likely break code relying on some particular
5918 ordering. The majority of such top-level @code{asm} statements,
5919 though, can be replaced by @code{section} attributes. The
5920 @option{fno-toplevel-reorder} option may be used to keep the ordering
5921 used in the input file, at the cost of some optimizations.
5924 @emph{unit-at-a-time} mode removes unreferenced static variables
5925 and functions. This may result in undefined references
5926 when an @code{asm} statement refers directly to variables or functions
5927 that are otherwise unused. In that case either the variable/function
5928 shall be listed as an operand of the @code{asm} statement operand or,
5929 in the case of top-level @code{asm} statements the attribute @code{used}
5930 shall be used on the declaration.
5933 Static functions now can use non-standard passing conventions that
5934 may break @code{asm} statements calling functions directly. Again,
5935 attribute @code{used} will prevent this behavior.
5938 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
5939 but this scheme may not be supported by future releases of GCC@.
5941 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5943 @item -fno-toplevel-reorder
5944 Do not reorder top-level functions, variables, and @code{asm}
5945 statements. Output them in the same order that they appear in the
5946 input file. When this option is used, unreferenced static variables
5947 will not be removed. This option is intended to support existing code
5948 which relies on a particular ordering. For new code, it is better to
5953 Constructs webs as commonly used for register allocation purposes and assign
5954 each web individual pseudo register. This allows the register allocation pass
5955 to operate on pseudos directly, but also strengthens several other optimization
5956 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
5957 however, make debugging impossible, since variables will no longer stay in a
5960 Enabled by default with @option{-funroll-loops}.
5962 @item -fwhole-program
5963 @opindex fwhole-program
5964 Assume that the current compilation unit represents whole program being
5965 compiled. All public functions and variables with the exception of @code{main}
5966 and those merged by attribute @code{externally_visible} become static functions
5967 and in a affect gets more aggressively optimized by interprocedural optimizers.
5968 While this option is equivalent to proper use of @code{static} keyword for
5969 programs consisting of single file, in combination with option
5970 @option{--combine} this flag can be used to compile most of smaller scale C
5971 programs since the functions and variables become local for the whole combined
5972 compilation unit, not for the single source file itself.
5975 @item -fno-cprop-registers
5976 @opindex fno-cprop-registers
5977 After register allocation and post-register allocation instruction splitting,
5978 we perform a copy-propagation pass to try to reduce scheduling dependencies
5979 and occasionally eliminate the copy.
5981 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5983 @item -fprofile-generate
5984 @opindex fprofile-generate
5986 Enable options usually used for instrumenting application to produce
5987 profile useful for later recompilation with profile feedback based
5988 optimization. You must use @option{-fprofile-generate} both when
5989 compiling and when linking your program.
5991 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
5994 @opindex fprofile-use
5995 Enable profile feedback directed optimizations, and optimizations
5996 generally profitable only with profile feedback available.
5998 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
5999 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
6001 By default, GCC emits an error message if the feedback profiles do not
6002 match the source code. This error can be turned into a warning by using
6003 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
6007 The following options control compiler behavior regarding floating
6008 point arithmetic. These options trade off between speed and
6009 correctness. All must be specifically enabled.
6013 @opindex ffloat-store
6014 Do not store floating point variables in registers, and inhibit other
6015 options that might change whether a floating point value is taken from a
6018 @cindex floating point precision
6019 This option prevents undesirable excess precision on machines such as
6020 the 68000 where the floating registers (of the 68881) keep more
6021 precision than a @code{double} is supposed to have. Similarly for the
6022 x86 architecture. For most programs, the excess precision does only
6023 good, but a few programs rely on the precise definition of IEEE floating
6024 point. Use @option{-ffloat-store} for such programs, after modifying
6025 them to store all pertinent intermediate computations into variables.
6029 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
6030 @option{-fno-trapping-math}, @option{-ffinite-math-only},
6031 @option{-fno-rounding-math}, @option{-fno-signaling-nans},
6032 @option{-fno-signed-zeros} and @option{fcx-limited-range}.
6034 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
6036 This option is not turned on by any @option{-O} option since
6037 it can result in incorrect output for programs which depend on
6038 an exact implementation of IEEE or ISO rules/specifications for
6039 math functions. It may, however, yield faster code for programs
6040 that do not require the guarantees of these specifications.
6042 @item -fno-math-errno
6043 @opindex fno-math-errno
6044 Do not set ERRNO after calling math functions that are executed
6045 with a single instruction, e.g., sqrt. A program that relies on
6046 IEEE exceptions for math error handling may want to use this flag
6047 for speed while maintaining IEEE arithmetic compatibility.
6049 This option is not turned on by any @option{-O} option since
6050 it can result in incorrect output for programs which depend on
6051 an exact implementation of IEEE or ISO rules/specifications for
6052 math functions. It may, however, yield faster code for programs
6053 that do not require the guarantees of these specifications.
6055 The default is @option{-fmath-errno}.
6057 On Darwin systems, the math library never sets @code{errno}. There is
6058 therefore no reason for the compiler to consider the possibility that
6059 it might, and @option{-fno-math-errno} is the default.
6061 @item -funsafe-math-optimizations
6062 @opindex funsafe-math-optimizations
6063 Allow optimizations for floating-point arithmetic that (a) assume
6064 that arguments and results are valid and (b) may violate IEEE or
6065 ANSI standards. When used at link-time, it may include libraries
6066 or startup files that change the default FPU control word or other
6067 similar optimizations.
6069 This option is not turned on by any @option{-O} option since
6070 it can result in incorrect output for programs which depend on
6071 an exact implementation of IEEE or ISO rules/specifications for
6072 math functions. It may, however, yield faster code for programs
6073 that do not require the guarantees of these specifications.
6075 The default is @option{-fno-unsafe-math-optimizations}.
6077 @item -ffinite-math-only
6078 @opindex ffinite-math-only
6079 Allow optimizations for floating-point arithmetic that assume
6080 that arguments and results are not NaNs or +-Infs.
6082 This option is not turned on by any @option{-O} option since
6083 it can result in incorrect output for programs which depend on
6084 an exact implementation of IEEE or ISO rules/specifications for
6085 math functions. It may, however, yield faster code for programs
6086 that do not require the guarantees of these specifications.
6088 The default is @option{-fno-finite-math-only}.
6090 @item -fno-signed-zeros
6091 @opindex fno-signed-zeros
6092 Allow optimizations for floating point arithmetic that ignore the
6093 signedness of zero. IEEE arithmetic specifies the behavior of
6094 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
6095 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
6096 This option implies that the sign of a zero result isn't significant.
6098 The default is @option{-fsigned-zeros}.
6100 @item -fno-trapping-math
6101 @opindex fno-trapping-math
6102 Compile code assuming that floating-point operations cannot generate
6103 user-visible traps. These traps include division by zero, overflow,
6104 underflow, inexact result and invalid operation. This option implies
6105 @option{-fno-signaling-nans}. Setting this option may allow faster
6106 code if one relies on ``non-stop'' IEEE arithmetic, for example.
6108 This option should never be turned on by any @option{-O} option since
6109 it can result in incorrect output for programs which depend on
6110 an exact implementation of IEEE or ISO rules/specifications for
6113 The default is @option{-ftrapping-math}.
6115 @item -frounding-math
6116 @opindex frounding-math
6117 Disable transformations and optimizations that assume default floating
6118 point rounding behavior. This is round-to-zero for all floating point
6119 to integer conversions, and round-to-nearest for all other arithmetic
6120 truncations. This option should be specified for programs that change
6121 the FP rounding mode dynamically, or that may be executed with a
6122 non-default rounding mode. This option disables constant folding of
6123 floating point expressions at compile-time (which may be affected by
6124 rounding mode) and arithmetic transformations that are unsafe in the
6125 presence of sign-dependent rounding modes.
6127 The default is @option{-fno-rounding-math}.
6129 This option is experimental and does not currently guarantee to
6130 disable all GCC optimizations that are affected by rounding mode.
6131 Future versions of GCC may provide finer control of this setting
6132 using C99's @code{FENV_ACCESS} pragma. This command line option
6133 will be used to specify the default state for @code{FENV_ACCESS}.
6135 @item -frtl-abstract-sequences
6136 @opindex frtl-abstract-sequences
6137 It is a size optimization method. This option is to find identical
6138 sequences of code, which can be turned into pseudo-procedures and
6139 then replace all occurrences with calls to the newly created
6140 subroutine. It is kind of an opposite of @option{-finline-functions}.
6141 This optimization runs at RTL level.
6143 @item -fsignaling-nans
6144 @opindex fsignaling-nans
6145 Compile code assuming that IEEE signaling NaNs may generate user-visible
6146 traps during floating-point operations. Setting this option disables
6147 optimizations that may change the number of exceptions visible with
6148 signaling NaNs. This option implies @option{-ftrapping-math}.
6150 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
6153 The default is @option{-fno-signaling-nans}.
6155 This option is experimental and does not currently guarantee to
6156 disable all GCC optimizations that affect signaling NaN behavior.
6158 @item -fsingle-precision-constant
6159 @opindex fsingle-precision-constant
6160 Treat floating point constant as single precision constant instead of
6161 implicitly converting it to double precision constant.
6163 @item -fcx-limited-range
6164 @itemx -fno-cx-limited-range
6165 @opindex fcx-limited-range
6166 @opindex fno-cx-limited-range
6167 When enabled, this option states that a range reduction step is not
6168 needed when performing complex division. The default is
6169 @option{-fno-cx-limited-range}, but is enabled by @option{-ffast-math}.
6171 This option controls the default setting of the ISO C99
6172 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
6177 The following options control optimizations that may improve
6178 performance, but are not enabled by any @option{-O} options. This
6179 section includes experimental options that may produce broken code.
6182 @item -fbranch-probabilities
6183 @opindex fbranch-probabilities
6184 After running a program compiled with @option{-fprofile-arcs}
6185 (@pxref{Debugging Options,, Options for Debugging Your Program or
6186 @command{gcc}}), you can compile it a second time using
6187 @option{-fbranch-probabilities}, to improve optimizations based on
6188 the number of times each branch was taken. When the program
6189 compiled with @option{-fprofile-arcs} exits it saves arc execution
6190 counts to a file called @file{@var{sourcename}.gcda} for each source
6191 file. The information in this data file is very dependent on the
6192 structure of the generated code, so you must use the same source code
6193 and the same optimization options for both compilations.
6195 With @option{-fbranch-probabilities}, GCC puts a
6196 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
6197 These can be used to improve optimization. Currently, they are only
6198 used in one place: in @file{reorg.c}, instead of guessing which path a
6199 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
6200 exactly determine which path is taken more often.
6202 @item -fprofile-values
6203 @opindex fprofile-values
6204 If combined with @option{-fprofile-arcs}, it adds code so that some
6205 data about values of expressions in the program is gathered.
6207 With @option{-fbranch-probabilities}, it reads back the data gathered
6208 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
6209 notes to instructions for their later usage in optimizations.
6211 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
6215 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
6216 a code to gather information about values of expressions.
6218 With @option{-fbranch-probabilities}, it reads back the data gathered
6219 and actually performs the optimizations based on them.
6220 Currently the optimizations include specialization of division operation
6221 using the knowledge about the value of the denominator.
6223 @item -frename-registers
6224 @opindex frename-registers
6225 Attempt to avoid false dependencies in scheduled code by making use
6226 of registers left over after register allocation. This optimization
6227 will most benefit processors with lots of registers. Depending on the
6228 debug information format adopted by the target, however, it can
6229 make debugging impossible, since variables will no longer stay in
6230 a ``home register''.
6232 Enabled by default with @option{-funroll-loops}.
6236 Perform tail duplication to enlarge superblock size. This transformation
6237 simplifies the control flow of the function allowing other optimizations to do
6240 Enabled with @option{-fprofile-use}.
6242 @item -funroll-loops
6243 @opindex funroll-loops
6244 Unroll loops whose number of iterations can be determined at compile time or
6245 upon entry to the loop. @option{-funroll-loops} implies
6246 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
6247 It also turns on complete loop peeling (i.e.@: complete removal of loops with
6248 small constant number of iterations). This option makes code larger, and may
6249 or may not make it run faster.
6251 Enabled with @option{-fprofile-use}.
6253 @item -funroll-all-loops
6254 @opindex funroll-all-loops
6255 Unroll all loops, even if their number of iterations is uncertain when
6256 the loop is entered. This usually makes programs run more slowly.
6257 @option{-funroll-all-loops} implies the same options as
6258 @option{-funroll-loops}.
6261 @opindex fpeel-loops
6262 Peels the loops for that there is enough information that they do not
6263 roll much (from profile feedback). It also turns on complete loop peeling
6264 (i.e.@: complete removal of loops with small constant number of iterations).
6266 Enabled with @option{-fprofile-use}.
6268 @item -fmove-loop-invariants
6269 @opindex fmove-loop-invariants
6270 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
6271 at level @option{-O1}
6273 @item -funswitch-loops
6274 @opindex funswitch-loops
6275 Move branches with loop invariant conditions out of the loop, with duplicates
6276 of the loop on both branches (modified according to result of the condition).
6278 @item -ffunction-sections
6279 @itemx -fdata-sections
6280 @opindex ffunction-sections
6281 @opindex fdata-sections
6282 Place each function or data item into its own section in the output
6283 file if the target supports arbitrary sections. The name of the
6284 function or the name of the data item determines the section's name
6287 Use these options on systems where the linker can perform optimizations
6288 to improve locality of reference in the instruction space. Most systems
6289 using the ELF object format and SPARC processors running Solaris 2 have
6290 linkers with such optimizations. AIX may have these optimizations in
6293 Only use these options when there are significant benefits from doing
6294 so. When you specify these options, the assembler and linker will
6295 create larger object and executable files and will also be slower.
6296 You will not be able to use @code{gprof} on all systems if you
6297 specify this option and you may have problems with debugging if
6298 you specify both this option and @option{-g}.
6300 @item -fbranch-target-load-optimize
6301 @opindex fbranch-target-load-optimize
6302 Perform branch target register load optimization before prologue / epilogue
6304 The use of target registers can typically be exposed only during reload,
6305 thus hoisting loads out of loops and doing inter-block scheduling needs
6306 a separate optimization pass.
6308 @item -fbranch-target-load-optimize2
6309 @opindex fbranch-target-load-optimize2
6310 Perform branch target register load optimization after prologue / epilogue
6313 @item -fbtr-bb-exclusive
6314 @opindex fbtr-bb-exclusive
6315 When performing branch target register load optimization, don't reuse
6316 branch target registers in within any basic block.
6318 @item -fstack-protector
6319 Emit extra code to check for buffer overflows, such as stack smashing
6320 attacks. This is done by adding a guard variable to functions with
6321 vulnerable objects. This includes functions that call alloca, and
6322 functions with buffers larger than 8 bytes. The guards are initialized
6323 when a function is entered and then checked when the function exits.
6324 If a guard check fails, an error message is printed and the program exits.
6326 @item -fstack-protector-all
6327 Like @option{-fstack-protector} except that all functions are protected.
6329 @item -fsection-anchors
6330 @opindex fsection-anchors
6331 Try to reduce the number of symbolic address calculations by using
6332 shared ``anchor'' symbols to address nearby objects. This transformation
6333 can help to reduce the number of GOT entries and GOT accesses on some
6336 For example, the implementation of the following function @code{foo}:
6340 int foo (void) @{ return a + b + c; @}
6343 would usually calculate the addresses of all three variables, but if you
6344 compile it with @option{-fsection-anchors}, it will access the variables
6345 from a common anchor point instead. The effect is similar to the
6346 following pseudocode (which isn't valid C):
6351 register int *xr = &x;
6352 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
6356 Not all targets support this option.
6358 @item --param @var{name}=@var{value}
6360 In some places, GCC uses various constants to control the amount of
6361 optimization that is done. For example, GCC will not inline functions
6362 that contain more that a certain number of instructions. You can
6363 control some of these constants on the command-line using the
6364 @option{--param} option.
6366 The names of specific parameters, and the meaning of the values, are
6367 tied to the internals of the compiler, and are subject to change
6368 without notice in future releases.
6370 In each case, the @var{value} is an integer. The allowable choices for
6371 @var{name} are given in the following table:
6374 @item salias-max-implicit-fields
6375 The maximum number of fields in a variable without direct
6376 structure accesses for which structure aliasing will consider trying
6377 to track each field. The default is 5
6379 @item salias-max-array-elements
6380 The maximum number of elements an array can have and its elements
6381 still be tracked individually by structure aliasing. The default is 4
6383 @item sra-max-structure-size
6384 The maximum structure size, in bytes, at which the scalar replacement
6385 of aggregates (SRA) optimization will perform block copies. The
6386 default value, 0, implies that GCC will select the most appropriate
6389 @item sra-field-structure-ratio
6390 The threshold ratio (as a percentage) between instantiated fields and
6391 the complete structure size. We say that if the ratio of the number
6392 of bytes in instantiated fields to the number of bytes in the complete
6393 structure exceeds this parameter, then block copies are not used. The
6396 @item max-crossjump-edges
6397 The maximum number of incoming edges to consider for crossjumping.
6398 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
6399 the number of edges incoming to each block. Increasing values mean
6400 more aggressive optimization, making the compile time increase with
6401 probably small improvement in executable size.
6403 @item min-crossjump-insns
6404 The minimum number of instructions which must be matched at the end
6405 of two blocks before crossjumping will be performed on them. This
6406 value is ignored in the case where all instructions in the block being
6407 crossjumped from are matched. The default value is 5.
6409 @item max-grow-copy-bb-insns
6410 The maximum code size expansion factor when copying basic blocks
6411 instead of jumping. The expansion is relative to a jump instruction.
6412 The default value is 8.
6414 @item max-goto-duplication-insns
6415 The maximum number of instructions to duplicate to a block that jumps
6416 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
6417 passes, GCC factors computed gotos early in the compilation process,
6418 and unfactors them as late as possible. Only computed jumps at the
6419 end of a basic blocks with no more than max-goto-duplication-insns are
6420 unfactored. The default value is 8.
6422 @item max-delay-slot-insn-search
6423 The maximum number of instructions to consider when looking for an
6424 instruction to fill a delay slot. If more than this arbitrary number of
6425 instructions is searched, the time savings from filling the delay slot
6426 will be minimal so stop searching. Increasing values mean more
6427 aggressive optimization, making the compile time increase with probably
6428 small improvement in executable run time.
6430 @item max-delay-slot-live-search
6431 When trying to fill delay slots, the maximum number of instructions to
6432 consider when searching for a block with valid live register
6433 information. Increasing this arbitrarily chosen value means more
6434 aggressive optimization, increasing the compile time. This parameter
6435 should be removed when the delay slot code is rewritten to maintain the
6438 @item max-gcse-memory
6439 The approximate maximum amount of memory that will be allocated in
6440 order to perform the global common subexpression elimination
6441 optimization. If more memory than specified is required, the
6442 optimization will not be done.
6444 @item max-gcse-passes
6445 The maximum number of passes of GCSE to run. The default is 1.
6447 @item max-pending-list-length
6448 The maximum number of pending dependencies scheduling will allow
6449 before flushing the current state and starting over. Large functions
6450 with few branches or calls can create excessively large lists which
6451 needlessly consume memory and resources.
6453 @item max-inline-insns-single
6454 Several parameters control the tree inliner used in gcc.
6455 This number sets the maximum number of instructions (counted in GCC's
6456 internal representation) in a single function that the tree inliner
6457 will consider for inlining. This only affects functions declared
6458 inline and methods implemented in a class declaration (C++).
6459 The default value is 450.
6461 @item max-inline-insns-auto
6462 When you use @option{-finline-functions} (included in @option{-O3}),
6463 a lot of functions that would otherwise not be considered for inlining
6464 by the compiler will be investigated. To those functions, a different
6465 (more restrictive) limit compared to functions declared inline can
6467 The default value is 90.
6469 @item large-function-insns
6470 The limit specifying really large functions. For functions larger than this
6471 limit after inlining inlining is constrained by
6472 @option{--param large-function-growth}. This parameter is useful primarily
6473 to avoid extreme compilation time caused by non-linear algorithms used by the
6475 This parameter is ignored when @option{-funit-at-a-time} is not used.
6476 The default value is 2700.
6478 @item large-function-growth
6479 Specifies maximal growth of large function caused by inlining in percents.
6480 This parameter is ignored when @option{-funit-at-a-time} is not used.
6481 The default value is 100 which limits large function growth to 2.0 times
6484 @item large-unit-insns
6485 The limit specifying large translation unit. Growth caused by inlining of
6486 units larger than this limit is limited by @option{--param inline-unit-growth}.
6487 For small units this might be too tight (consider unit consisting of function A
6488 that is inline and B that just calls A three time. If B is small relative to
6489 A, the growth of unit is 300\% and yet such inlining is very sane. For very
6490 large units consisting of small inlininable functions however the overall unit
6491 growth limit is needed to avoid exponential explosion of code size. Thus for
6492 smaller units, the size is increased to @option{--param large-unit-insns}
6493 before applying @option{--param inline-unit-growth}. The default is 10000
6495 @item inline-unit-growth
6496 Specifies maximal overall growth of the compilation unit caused by inlining.
6497 This parameter is ignored when @option{-funit-at-a-time} is not used.
6498 The default value is 30 which limits unit growth to 1.3 times the original
6501 @item large-stack-frame
6502 The limit specifying large stack frames. While inlining the algorithm is trying
6503 to not grow past this limit too much. Default value is 256 bytes.
6505 @item large-stack-frame-growth
6506 Specifies maximal growth of large stack frames caused by inlining in percents.
6507 The default value is 1000 which limits large stack frame growth to 11 times
6510 @item max-inline-insns-recursive
6511 @itemx max-inline-insns-recursive-auto
6512 Specifies maximum number of instructions out-of-line copy of self recursive inline
6513 function can grow into by performing recursive inlining.
6515 For functions declared inline @option{--param max-inline-insns-recursive} is
6516 taken into account. For function not declared inline, recursive inlining
6517 happens only when @option{-finline-functions} (included in @option{-O3}) is
6518 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
6519 default value is 450.
6521 @item max-inline-recursive-depth
6522 @itemx max-inline-recursive-depth-auto
6523 Specifies maximum recursion depth used by the recursive inlining.
6525 For functions declared inline @option{--param max-inline-recursive-depth} is
6526 taken into account. For function not declared inline, recursive inlining
6527 happens only when @option{-finline-functions} (included in @option{-O3}) is
6528 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
6529 default value is 450.
6531 @item min-inline-recursive-probability
6532 Recursive inlining is profitable only for function having deep recursion
6533 in average and can hurt for function having little recursion depth by
6534 increasing the prologue size or complexity of function body to other
6537 When profile feedback is available (see @option{-fprofile-generate}) the actual
6538 recursion depth can be guessed from probability that function will recurse via
6539 given call expression. This parameter limits inlining only to call expression
6540 whose probability exceeds given threshold (in percents). The default value is
6543 @item inline-call-cost
6544 Specify cost of call instruction relative to simple arithmetics operations
6545 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
6546 functions and at the same time increases size of leaf function that is believed to
6547 reduce function size by being inlined. In effect it increases amount of
6548 inlining for code having large abstraction penalty (many functions that just
6549 pass the arguments to other functions) and decrease inlining for code with low
6550 abstraction penalty. The default value is 16.
6552 @item min-vect-loop-bound
6553 The minimum number of iterations under which a loop will not get vectorized
6554 when @option{-ftree-vectorize} is used. The number of iterations after
6555 vectorization needs to be greater than the value specified by this option
6556 to allow vectorization. The default value is 0.
6558 @item max-unrolled-insns
6559 The maximum number of instructions that a loop should have if that loop
6560 is unrolled, and if the loop is unrolled, it determines how many times
6561 the loop code is unrolled.
6563 @item max-average-unrolled-insns
6564 The maximum number of instructions biased by probabilities of their execution
6565 that a loop should have if that loop is unrolled, and if the loop is unrolled,
6566 it determines how many times the loop code is unrolled.
6568 @item max-unroll-times
6569 The maximum number of unrollings of a single loop.
6571 @item max-peeled-insns
6572 The maximum number of instructions that a loop should have if that loop
6573 is peeled, and if the loop is peeled, it determines how many times
6574 the loop code is peeled.
6576 @item max-peel-times
6577 The maximum number of peelings of a single loop.
6579 @item max-completely-peeled-insns
6580 The maximum number of insns of a completely peeled loop.
6582 @item max-completely-peel-times
6583 The maximum number of iterations of a loop to be suitable for complete peeling.
6585 @item max-unswitch-insns
6586 The maximum number of insns of an unswitched loop.
6588 @item max-unswitch-level
6589 The maximum number of branches unswitched in a single loop.
6592 The minimum cost of an expensive expression in the loop invariant motion.
6594 @item iv-consider-all-candidates-bound
6595 Bound on number of candidates for induction variables below that
6596 all candidates are considered for each use in induction variable
6597 optimizations. Only the most relevant candidates are considered
6598 if there are more candidates, to avoid quadratic time complexity.
6600 @item iv-max-considered-uses
6601 The induction variable optimizations give up on loops that contain more
6602 induction variable uses.
6604 @item iv-always-prune-cand-set-bound
6605 If number of candidates in the set is smaller than this value,
6606 we always try to remove unnecessary ivs from the set during its
6607 optimization when a new iv is added to the set.
6609 @item scev-max-expr-size
6610 Bound on size of expressions used in the scalar evolutions analyzer.
6611 Large expressions slow the analyzer.
6613 @item omega-max-vars
6614 The maximum number of variables in an Omega constraint system.
6615 The default value is 128.
6617 @item omega-max-geqs
6618 The maximum number of inequalities in an Omega constraint system.
6619 The default value is 256.
6622 The maximum number of equalities in an Omega constraint system.
6623 The default value is 128.
6625 @item omega-max-wild-cards
6626 The maximum number of wildcard variables that the Omega solver will
6627 be able to insert. The default value is 18.
6629 @item omega-hash-table-size
6630 The size of the hash table in the Omega solver. The default value is
6633 @item omega-max-keys
6634 The maximal number of keys used by the Omega solver. The default
6637 @item omega-eliminate-redundant-constraints
6638 When set to 1, use expensive methods to eliminate all redundant
6639 constraints. The default value is 0.
6641 @item vect-max-version-checks
6642 The maximum number of runtime checks that can be performed when doing
6643 loop versioning in the vectorizer. See option ftree-vect-loop-version
6644 for more information.
6646 @item max-iterations-to-track
6648 The maximum number of iterations of a loop the brute force algorithm
6649 for analysis of # of iterations of the loop tries to evaluate.
6651 @item hot-bb-count-fraction
6652 Select fraction of the maximal count of repetitions of basic block in program
6653 given basic block needs to have to be considered hot.
6655 @item hot-bb-frequency-fraction
6656 Select fraction of the maximal frequency of executions of basic block in
6657 function given basic block needs to have to be considered hot
6659 @item max-predicted-iterations
6660 The maximum number of loop iterations we predict statically. This is useful
6661 in cases where function contain single loop with known bound and other loop
6662 with unknown. We predict the known number of iterations correctly, while
6663 the unknown number of iterations average to roughly 10. This means that the
6664 loop without bounds would appear artificially cold relative to the other one.
6666 @item tracer-dynamic-coverage
6667 @itemx tracer-dynamic-coverage-feedback
6669 This value is used to limit superblock formation once the given percentage of
6670 executed instructions is covered. This limits unnecessary code size
6673 The @option{tracer-dynamic-coverage-feedback} is used only when profile
6674 feedback is available. The real profiles (as opposed to statically estimated
6675 ones) are much less balanced allowing the threshold to be larger value.
6677 @item tracer-max-code-growth
6678 Stop tail duplication once code growth has reached given percentage. This is
6679 rather hokey argument, as most of the duplicates will be eliminated later in
6680 cross jumping, so it may be set to much higher values than is the desired code
6683 @item tracer-min-branch-ratio
6685 Stop reverse growth when the reverse probability of best edge is less than this
6686 threshold (in percent).
6688 @item tracer-min-branch-ratio
6689 @itemx tracer-min-branch-ratio-feedback
6691 Stop forward growth if the best edge do have probability lower than this
6694 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
6695 compilation for profile feedback and one for compilation without. The value
6696 for compilation with profile feedback needs to be more conservative (higher) in
6697 order to make tracer effective.
6699 @item max-cse-path-length
6701 Maximum number of basic blocks on path that cse considers. The default is 10.
6704 The maximum instructions CSE process before flushing. The default is 1000.
6706 @item max-aliased-vops
6708 Maximum number of virtual operands per function allowed to represent
6709 aliases before triggering the alias partitioning heuristic. Alias
6710 partitioning reduces compile times and memory consumption needed for
6711 aliasing at the expense of precision loss in alias information. The
6712 default value for this parameter is 100 for -O1, 500 for -O2 and 1000
6715 Notice that if a function contains more memory statements than the
6716 value of this parameter, it is not really possible to achieve this
6717 reduction. In this case, the compiler will use the number of memory
6718 statements as the value for @option{max-aliased-vops}.
6720 @item avg-aliased-vops
6722 Average number of virtual operands per statement allowed to represent
6723 aliases before triggering the alias partitioning heuristic. This
6724 works in conjunction with @option{max-aliased-vops}. If a function
6725 contains more than @option{max-aliased-vops} virtual operators, then
6726 memory symbols will be grouped into memory partitions until either the
6727 total number of virtual operators is below @option{max-aliased-vops}
6728 or the average number of virtual operators per memory statement is
6729 below @option{avg-aliased-vops}. The default value for this parameter
6730 is 1 for -O1 and -O2, and 3 for -O3.
6732 @item ggc-min-expand
6734 GCC uses a garbage collector to manage its own memory allocation. This
6735 parameter specifies the minimum percentage by which the garbage
6736 collector's heap should be allowed to expand between collections.
6737 Tuning this may improve compilation speed; it has no effect on code
6740 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
6741 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
6742 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
6743 GCC is not able to calculate RAM on a particular platform, the lower
6744 bound of 30% is used. Setting this parameter and
6745 @option{ggc-min-heapsize} to zero causes a full collection to occur at
6746 every opportunity. This is extremely slow, but can be useful for
6749 @item ggc-min-heapsize
6751 Minimum size of the garbage collector's heap before it begins bothering
6752 to collect garbage. The first collection occurs after the heap expands
6753 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
6754 tuning this may improve compilation speed, and has no effect on code
6757 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
6758 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
6759 with a lower bound of 4096 (four megabytes) and an upper bound of
6760 131072 (128 megabytes). If GCC is not able to calculate RAM on a
6761 particular platform, the lower bound is used. Setting this parameter
6762 very large effectively disables garbage collection. Setting this
6763 parameter and @option{ggc-min-expand} to zero causes a full collection
6764 to occur at every opportunity.
6766 @item max-reload-search-insns
6767 The maximum number of instruction reload should look backward for equivalent
6768 register. Increasing values mean more aggressive optimization, making the
6769 compile time increase with probably slightly better performance. The default
6772 @item max-cselib-memory-locations
6773 The maximum number of memory locations cselib should take into account.
6774 Increasing values mean more aggressive optimization, making the compile time
6775 increase with probably slightly better performance. The default value is 500.
6777 @item max-flow-memory-locations
6778 Similar as @option{max-cselib-memory-locations} but for dataflow liveness.
6779 The default value is 100.
6781 @item reorder-blocks-duplicate
6782 @itemx reorder-blocks-duplicate-feedback
6784 Used by basic block reordering pass to decide whether to use unconditional
6785 branch or duplicate the code on its destination. Code is duplicated when its
6786 estimated size is smaller than this value multiplied by the estimated size of
6787 unconditional jump in the hot spots of the program.
6789 The @option{reorder-block-duplicate-feedback} is used only when profile
6790 feedback is available and may be set to higher values than
6791 @option{reorder-block-duplicate} since information about the hot spots is more
6794 @item max-sched-ready-insns
6795 The maximum number of instructions ready to be issued the scheduler should
6796 consider at any given time during the first scheduling pass. Increasing
6797 values mean more thorough searches, making the compilation time increase
6798 with probably little benefit. The default value is 100.
6800 @item max-sched-region-blocks
6801 The maximum number of blocks in a region to be considered for
6802 interblock scheduling. The default value is 10.
6804 @item max-sched-region-insns
6805 The maximum number of insns in a region to be considered for
6806 interblock scheduling. The default value is 100.
6809 The minimum probability (in percents) of reaching a source block
6810 for interblock speculative scheduling. The default value is 40.
6812 @item max-sched-extend-regions-iters
6813 The maximum number of iterations through CFG to extend regions.
6814 0 - disable region extension,
6815 N - do at most N iterations.
6816 The default value is 0.
6818 @item max-sched-insn-conflict-delay
6819 The maximum conflict delay for an insn to be considered for speculative motion.
6820 The default value is 3.
6822 @item sched-spec-prob-cutoff
6823 The minimal probability of speculation success (in percents), so that
6824 speculative insn will be scheduled.
6825 The default value is 40.
6827 @item max-last-value-rtl
6829 The maximum size measured as number of RTLs that can be recorded in an expression
6830 in combiner for a pseudo register as last known value of that register. The default
6833 @item integer-share-limit
6834 Small integer constants can use a shared data structure, reducing the
6835 compiler's memory usage and increasing its speed. This sets the maximum
6836 value of a shared integer constant's. The default value is 256.
6838 @item min-virtual-mappings
6839 Specifies the minimum number of virtual mappings in the incremental
6840 SSA updater that should be registered to trigger the virtual mappings
6841 heuristic defined by virtual-mappings-ratio. The default value is
6844 @item virtual-mappings-ratio
6845 If the number of virtual mappings is virtual-mappings-ratio bigger
6846 than the number of virtual symbols to be updated, then the incremental
6847 SSA updater switches to a full update for those symbols. The default
6850 @item ssp-buffer-size
6851 The minimum size of buffers (i.e. arrays) that will receive stack smashing
6852 protection when @option{-fstack-protection} is used.
6854 @item max-jump-thread-duplication-stmts
6855 Maximum number of statements allowed in a block that needs to be
6856 duplicated when threading jumps.
6858 @item max-fields-for-field-sensitive
6859 Maximum number of fields in a structure we will treat in
6860 a field sensitive manner during pointer analysis.
6862 @item prefetch-latency
6863 Estimate on average number of instructions that are executed before
6864 prefetch finishes. The distance we prefetch ahead is proportional
6865 to this constant. Increasing this number may also lead to less
6866 streams being prefetched (see @option{simultaneous-prefetches}).
6868 @item simultaneous-prefetches
6869 Maximum number of prefetches that can run at the same time.
6871 @item l1-cache-line-size
6872 The size of cache line in L1 cache, in bytes.
6875 The number of cache lines in L1 cache.
6877 @item verify-canonical-types
6878 Whether the compiler should verify the ``canonical'' types used for
6879 type equality comparisons within the C++ and Objective-C++ front
6880 ends. Set to 1 (the default when GCC is configured with
6881 --enable-checking) to enable verification, 0 to disable verification
6882 (the default when GCC is configured with --disable-checking).
6887 @node Preprocessor Options
6888 @section Options Controlling the Preprocessor
6889 @cindex preprocessor options
6890 @cindex options, preprocessor
6892 These options control the C preprocessor, which is run on each C source
6893 file before actual compilation.
6895 If you use the @option{-E} option, nothing is done except preprocessing.
6896 Some of these options make sense only together with @option{-E} because
6897 they cause the preprocessor output to be unsuitable for actual
6902 You can use @option{-Wp,@var{option}} to bypass the compiler driver
6903 and pass @var{option} directly through to the preprocessor. If
6904 @var{option} contains commas, it is split into multiple options at the
6905 commas. However, many options are modified, translated or interpreted
6906 by the compiler driver before being passed to the preprocessor, and
6907 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
6908 interface is undocumented and subject to change, so whenever possible
6909 you should avoid using @option{-Wp} and let the driver handle the
6912 @item -Xpreprocessor @var{option}
6913 @opindex preprocessor
6914 Pass @var{option} as an option to the preprocessor. You can use this to
6915 supply system-specific preprocessor options which GCC does not know how to
6918 If you want to pass an option that takes an argument, you must use
6919 @option{-Xpreprocessor} twice, once for the option and once for the argument.
6922 @include cppopts.texi
6924 @node Assembler Options
6925 @section Passing Options to the Assembler
6927 @c prevent bad page break with this line
6928 You can pass options to the assembler.
6931 @item -Wa,@var{option}
6933 Pass @var{option} as an option to the assembler. If @var{option}
6934 contains commas, it is split into multiple options at the commas.
6936 @item -Xassembler @var{option}
6938 Pass @var{option} as an option to the assembler. You can use this to
6939 supply system-specific assembler options which GCC does not know how to
6942 If you want to pass an option that takes an argument, you must use
6943 @option{-Xassembler} twice, once for the option and once for the argument.
6948 @section Options for Linking
6949 @cindex link options
6950 @cindex options, linking
6952 These options come into play when the compiler links object files into
6953 an executable output file. They are meaningless if the compiler is
6954 not doing a link step.
6958 @item @var{object-file-name}
6959 A file name that does not end in a special recognized suffix is
6960 considered to name an object file or library. (Object files are
6961 distinguished from libraries by the linker according to the file
6962 contents.) If linking is done, these object files are used as input
6971 If any of these options is used, then the linker is not run, and
6972 object file names should not be used as arguments. @xref{Overall
6976 @item -l@var{library}
6977 @itemx -l @var{library}
6979 Search the library named @var{library} when linking. (The second
6980 alternative with the library as a separate argument is only for
6981 POSIX compliance and is not recommended.)
6983 It makes a difference where in the command you write this option; the
6984 linker searches and processes libraries and object files in the order they
6985 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
6986 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
6987 to functions in @samp{z}, those functions may not be loaded.
6989 The linker searches a standard list of directories for the library,
6990 which is actually a file named @file{lib@var{library}.a}. The linker
6991 then uses this file as if it had been specified precisely by name.
6993 The directories searched include several standard system directories
6994 plus any that you specify with @option{-L}.
6996 Normally the files found this way are library files---archive files
6997 whose members are object files. The linker handles an archive file by
6998 scanning through it for members which define symbols that have so far
6999 been referenced but not defined. But if the file that is found is an
7000 ordinary object file, it is linked in the usual fashion. The only
7001 difference between using an @option{-l} option and specifying a file name
7002 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
7003 and searches several directories.
7007 You need this special case of the @option{-l} option in order to
7008 link an Objective-C or Objective-C++ program.
7011 @opindex nostartfiles
7012 Do not use the standard system startup files when linking.
7013 The standard system libraries are used normally, unless @option{-nostdlib}
7014 or @option{-nodefaultlibs} is used.
7016 @item -nodefaultlibs
7017 @opindex nodefaultlibs
7018 Do not use the standard system libraries when linking.
7019 Only the libraries you specify will be passed to the linker.
7020 The standard startup files are used normally, unless @option{-nostartfiles}
7021 is used. The compiler may generate calls to @code{memcmp},
7022 @code{memset}, @code{memcpy} and @code{memmove}.
7023 These entries are usually resolved by entries in
7024 libc. These entry points should be supplied through some other
7025 mechanism when this option is specified.
7029 Do not use the standard system startup files or libraries when linking.
7030 No startup files and only the libraries you specify will be passed to
7031 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
7032 @code{memcpy} and @code{memmove}.
7033 These entries are usually resolved by entries in
7034 libc. These entry points should be supplied through some other
7035 mechanism when this option is specified.
7037 @cindex @option{-lgcc}, use with @option{-nostdlib}
7038 @cindex @option{-nostdlib} and unresolved references
7039 @cindex unresolved references and @option{-nostdlib}
7040 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
7041 @cindex @option{-nodefaultlibs} and unresolved references
7042 @cindex unresolved references and @option{-nodefaultlibs}
7043 One of the standard libraries bypassed by @option{-nostdlib} and
7044 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
7045 that GCC uses to overcome shortcomings of particular machines, or special
7046 needs for some languages.
7047 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
7048 Collection (GCC) Internals},
7049 for more discussion of @file{libgcc.a}.)
7050 In most cases, you need @file{libgcc.a} even when you want to avoid
7051 other standard libraries. In other words, when you specify @option{-nostdlib}
7052 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
7053 This ensures that you have no unresolved references to internal GCC
7054 library subroutines. (For example, @samp{__main}, used to ensure C++
7055 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
7056 GNU Compiler Collection (GCC) Internals}.)
7060 Produce a position independent executable on targets which support it.
7061 For predictable results, you must also specify the same set of options
7062 that were used to generate code (@option{-fpie}, @option{-fPIE},
7063 or model suboptions) when you specify this option.
7067 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
7068 that support it. This instructs the linker to add all symbols, not
7069 only used ones, to the dynamic symbol table. This option is needed
7070 for some uses of @code{dlopen} or to allow obtaining backtraces
7071 from within a program.
7075 Remove all symbol table and relocation information from the executable.
7079 On systems that support dynamic linking, this prevents linking with the shared
7080 libraries. On other systems, this option has no effect.
7084 Produce a shared object which can then be linked with other objects to
7085 form an executable. Not all systems support this option. For predictable
7086 results, you must also specify the same set of options that were used to
7087 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
7088 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
7089 needs to build supplementary stub code for constructors to work. On
7090 multi-libbed systems, @samp{gcc -shared} must select the correct support
7091 libraries to link against. Failing to supply the correct flags may lead
7092 to subtle defects. Supplying them in cases where they are not necessary
7095 @item -shared-libgcc
7096 @itemx -static-libgcc
7097 @opindex shared-libgcc
7098 @opindex static-libgcc
7099 On systems that provide @file{libgcc} as a shared library, these options
7100 force the use of either the shared or static version respectively.
7101 If no shared version of @file{libgcc} was built when the compiler was
7102 configured, these options have no effect.
7104 There are several situations in which an application should use the
7105 shared @file{libgcc} instead of the static version. The most common
7106 of these is when the application wishes to throw and catch exceptions
7107 across different shared libraries. In that case, each of the libraries
7108 as well as the application itself should use the shared @file{libgcc}.
7110 Therefore, the G++ and GCJ drivers automatically add
7111 @option{-shared-libgcc} whenever you build a shared library or a main
7112 executable, because C++ and Java programs typically use exceptions, so
7113 this is the right thing to do.
7115 If, instead, you use the GCC driver to create shared libraries, you may
7116 find that they will not always be linked with the shared @file{libgcc}.
7117 If GCC finds, at its configuration time, that you have a non-GNU linker
7118 or a GNU linker that does not support option @option{--eh-frame-hdr},
7119 it will link the shared version of @file{libgcc} into shared libraries
7120 by default. Otherwise, it will take advantage of the linker and optimize
7121 away the linking with the shared version of @file{libgcc}, linking with
7122 the static version of libgcc by default. This allows exceptions to
7123 propagate through such shared libraries, without incurring relocation
7124 costs at library load time.
7126 However, if a library or main executable is supposed to throw or catch
7127 exceptions, you must link it using the G++ or GCJ driver, as appropriate
7128 for the languages used in the program, or using the option
7129 @option{-shared-libgcc}, such that it is linked with the shared
7134 Bind references to global symbols when building a shared object. Warn
7135 about any unresolved references (unless overridden by the link editor
7136 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
7139 @item -Xlinker @var{option}
7141 Pass @var{option} as an option to the linker. You can use this to
7142 supply system-specific linker options which GCC does not know how to
7145 If you want to pass an option that takes an argument, you must use
7146 @option{-Xlinker} twice, once for the option and once for the argument.
7147 For example, to pass @option{-assert definitions}, you must write
7148 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
7149 @option{-Xlinker "-assert definitions"}, because this passes the entire
7150 string as a single argument, which is not what the linker expects.
7152 @item -Wl,@var{option}
7154 Pass @var{option} as an option to the linker. If @var{option} contains
7155 commas, it is split into multiple options at the commas.
7157 @item -u @var{symbol}
7159 Pretend the symbol @var{symbol} is undefined, to force linking of
7160 library modules to define it. You can use @option{-u} multiple times with
7161 different symbols to force loading of additional library modules.
7164 @node Directory Options
7165 @section Options for Directory Search
7166 @cindex directory options
7167 @cindex options, directory search
7170 These options specify directories to search for header files, for
7171 libraries and for parts of the compiler:
7176 Add the directory @var{dir} to the head of the list of directories to be
7177 searched for header files. This can be used to override a system header
7178 file, substituting your own version, since these directories are
7179 searched before the system header file directories. However, you should
7180 not use this option to add directories that contain vendor-supplied
7181 system header files (use @option{-isystem} for that). If you use more than
7182 one @option{-I} option, the directories are scanned in left-to-right
7183 order; the standard system directories come after.
7185 If a standard system include directory, or a directory specified with
7186 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
7187 option will be ignored. The directory will still be searched but as a
7188 system directory at its normal position in the system include chain.
7189 This is to ensure that GCC's procedure to fix buggy system headers and
7190 the ordering for the include_next directive are not inadvertently changed.
7191 If you really need to change the search order for system directories,
7192 use the @option{-nostdinc} and/or @option{-isystem} options.
7194 @item -iquote@var{dir}
7196 Add the directory @var{dir} to the head of the list of directories to
7197 be searched for header files only for the case of @samp{#include
7198 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
7199 otherwise just like @option{-I}.
7203 Add directory @var{dir} to the list of directories to be searched
7206 @item -B@var{prefix}
7208 This option specifies where to find the executables, libraries,
7209 include files, and data files of the compiler itself.
7211 The compiler driver program runs one or more of the subprograms
7212 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
7213 @var{prefix} as a prefix for each program it tries to run, both with and
7214 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
7216 For each subprogram to be run, the compiler driver first tries the
7217 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
7218 was not specified, the driver tries two standard prefixes, which are
7219 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
7220 those results in a file name that is found, the unmodified program
7221 name is searched for using the directories specified in your
7222 @env{PATH} environment variable.
7224 The compiler will check to see if the path provided by the @option{-B}
7225 refers to a directory, and if necessary it will add a directory
7226 separator character at the end of the path.
7228 @option{-B} prefixes that effectively specify directory names also apply
7229 to libraries in the linker, because the compiler translates these
7230 options into @option{-L} options for the linker. They also apply to
7231 includes files in the preprocessor, because the compiler translates these
7232 options into @option{-isystem} options for the preprocessor. In this case,
7233 the compiler appends @samp{include} to the prefix.
7235 The run-time support file @file{libgcc.a} can also be searched for using
7236 the @option{-B} prefix, if needed. If it is not found there, the two
7237 standard prefixes above are tried, and that is all. The file is left
7238 out of the link if it is not found by those means.
7240 Another way to specify a prefix much like the @option{-B} prefix is to use
7241 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
7244 As a special kludge, if the path provided by @option{-B} is
7245 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
7246 9, then it will be replaced by @file{[dir/]include}. This is to help
7247 with boot-strapping the compiler.
7249 @item -specs=@var{file}
7251 Process @var{file} after the compiler reads in the standard @file{specs}
7252 file, in order to override the defaults that the @file{gcc} driver
7253 program uses when determining what switches to pass to @file{cc1},
7254 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
7255 @option{-specs=@var{file}} can be specified on the command line, and they
7256 are processed in order, from left to right.
7258 @item --sysroot=@var{dir}
7260 Use @var{dir} as the logical root directory for headers and libraries.
7261 For example, if the compiler would normally search for headers in
7262 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
7263 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
7265 If you use both this option and the @option{-isysroot} option, then
7266 the @option{--sysroot} option will apply to libraries, but the
7267 @option{-isysroot} option will apply to header files.
7269 The GNU linker (beginning with version 2.16) has the necessary support
7270 for this option. If your linker does not support this option, the
7271 header file aspect of @option{--sysroot} will still work, but the
7272 library aspect will not.
7276 This option has been deprecated. Please use @option{-iquote} instead for
7277 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
7278 Any directories you specify with @option{-I} options before the @option{-I-}
7279 option are searched only for the case of @samp{#include "@var{file}"};
7280 they are not searched for @samp{#include <@var{file}>}.
7282 If additional directories are specified with @option{-I} options after
7283 the @option{-I-}, these directories are searched for all @samp{#include}
7284 directives. (Ordinarily @emph{all} @option{-I} directories are used
7287 In addition, the @option{-I-} option inhibits the use of the current
7288 directory (where the current input file came from) as the first search
7289 directory for @samp{#include "@var{file}"}. There is no way to
7290 override this effect of @option{-I-}. With @option{-I.} you can specify
7291 searching the directory which was current when the compiler was
7292 invoked. That is not exactly the same as what the preprocessor does
7293 by default, but it is often satisfactory.
7295 @option{-I-} does not inhibit the use of the standard system directories
7296 for header files. Thus, @option{-I-} and @option{-nostdinc} are
7303 @section Specifying subprocesses and the switches to pass to them
7306 @command{gcc} is a driver program. It performs its job by invoking a
7307 sequence of other programs to do the work of compiling, assembling and
7308 linking. GCC interprets its command-line parameters and uses these to
7309 deduce which programs it should invoke, and which command-line options
7310 it ought to place on their command lines. This behavior is controlled
7311 by @dfn{spec strings}. In most cases there is one spec string for each
7312 program that GCC can invoke, but a few programs have multiple spec
7313 strings to control their behavior. The spec strings built into GCC can
7314 be overridden by using the @option{-specs=} command-line switch to specify
7317 @dfn{Spec files} are plaintext files that are used to construct spec
7318 strings. They consist of a sequence of directives separated by blank
7319 lines. The type of directive is determined by the first non-whitespace
7320 character on the line and it can be one of the following:
7323 @item %@var{command}
7324 Issues a @var{command} to the spec file processor. The commands that can
7328 @item %include <@var{file}>
7330 Search for @var{file} and insert its text at the current point in the
7333 @item %include_noerr <@var{file}>
7334 @cindex %include_noerr
7335 Just like @samp{%include}, but do not generate an error message if the include
7336 file cannot be found.
7338 @item %rename @var{old_name} @var{new_name}
7340 Rename the spec string @var{old_name} to @var{new_name}.
7344 @item *[@var{spec_name}]:
7345 This tells the compiler to create, override or delete the named spec
7346 string. All lines after this directive up to the next directive or
7347 blank line are considered to be the text for the spec string. If this
7348 results in an empty string then the spec will be deleted. (Or, if the
7349 spec did not exist, then nothing will happened.) Otherwise, if the spec
7350 does not currently exist a new spec will be created. If the spec does
7351 exist then its contents will be overridden by the text of this
7352 directive, unless the first character of that text is the @samp{+}
7353 character, in which case the text will be appended to the spec.
7355 @item [@var{suffix}]:
7356 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
7357 and up to the next directive or blank line are considered to make up the
7358 spec string for the indicated suffix. When the compiler encounters an
7359 input file with the named suffix, it will processes the spec string in
7360 order to work out how to compile that file. For example:
7367 This says that any input file whose name ends in @samp{.ZZ} should be
7368 passed to the program @samp{z-compile}, which should be invoked with the
7369 command-line switch @option{-input} and with the result of performing the
7370 @samp{%i} substitution. (See below.)
7372 As an alternative to providing a spec string, the text that follows a
7373 suffix directive can be one of the following:
7376 @item @@@var{language}
7377 This says that the suffix is an alias for a known @var{language}. This is
7378 similar to using the @option{-x} command-line switch to GCC to specify a
7379 language explicitly. For example:
7386 Says that .ZZ files are, in fact, C++ source files.
7389 This causes an error messages saying:
7392 @var{name} compiler not installed on this system.
7396 GCC already has an extensive list of suffixes built into it.
7397 This directive will add an entry to the end of the list of suffixes, but
7398 since the list is searched from the end backwards, it is effectively
7399 possible to override earlier entries using this technique.
7403 GCC has the following spec strings built into it. Spec files can
7404 override these strings or create their own. Note that individual
7405 targets can also add their own spec strings to this list.
7408 asm Options to pass to the assembler
7409 asm_final Options to pass to the assembler post-processor
7410 cpp Options to pass to the C preprocessor
7411 cc1 Options to pass to the C compiler
7412 cc1plus Options to pass to the C++ compiler
7413 endfile Object files to include at the end of the link
7414 link Options to pass to the linker
7415 lib Libraries to include on the command line to the linker
7416 libgcc Decides which GCC support library to pass to the linker
7417 linker Sets the name of the linker
7418 predefines Defines to be passed to the C preprocessor
7419 signed_char Defines to pass to CPP to say whether @code{char} is signed
7421 startfile Object files to include at the start of the link
7424 Here is a small example of a spec file:
7430 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
7433 This example renames the spec called @samp{lib} to @samp{old_lib} and
7434 then overrides the previous definition of @samp{lib} with a new one.
7435 The new definition adds in some extra command-line options before
7436 including the text of the old definition.
7438 @dfn{Spec strings} are a list of command-line options to be passed to their
7439 corresponding program. In addition, the spec strings can contain
7440 @samp{%}-prefixed sequences to substitute variable text or to
7441 conditionally insert text into the command line. Using these constructs
7442 it is possible to generate quite complex command lines.
7444 Here is a table of all defined @samp{%}-sequences for spec
7445 strings. Note that spaces are not generated automatically around the
7446 results of expanding these sequences. Therefore you can concatenate them
7447 together or combine them with constant text in a single argument.
7451 Substitute one @samp{%} into the program name or argument.
7454 Substitute the name of the input file being processed.
7457 Substitute the basename of the input file being processed.
7458 This is the substring up to (and not including) the last period
7459 and not including the directory.
7462 This is the same as @samp{%b}, but include the file suffix (text after
7466 Marks the argument containing or following the @samp{%d} as a
7467 temporary file name, so that that file will be deleted if GCC exits
7468 successfully. Unlike @samp{%g}, this contributes no text to the
7471 @item %g@var{suffix}
7472 Substitute a file name that has suffix @var{suffix} and is chosen
7473 once per compilation, and mark the argument in the same way as
7474 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
7475 name is now chosen in a way that is hard to predict even when previously
7476 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
7477 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
7478 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
7479 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
7480 was simply substituted with a file name chosen once per compilation,
7481 without regard to any appended suffix (which was therefore treated
7482 just like ordinary text), making such attacks more likely to succeed.
7484 @item %u@var{suffix}
7485 Like @samp{%g}, but generates a new temporary file name even if
7486 @samp{%u@var{suffix}} was already seen.
7488 @item %U@var{suffix}
7489 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
7490 new one if there is no such last file name. In the absence of any
7491 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
7492 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
7493 would involve the generation of two distinct file names, one
7494 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
7495 simply substituted with a file name chosen for the previous @samp{%u},
7496 without regard to any appended suffix.
7498 @item %j@var{suffix}
7499 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
7500 writable, and if save-temps is off; otherwise, substitute the name
7501 of a temporary file, just like @samp{%u}. This temporary file is not
7502 meant for communication between processes, but rather as a junk
7505 @item %|@var{suffix}
7506 @itemx %m@var{suffix}
7507 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
7508 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
7509 all. These are the two most common ways to instruct a program that it
7510 should read from standard input or write to standard output. If you
7511 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
7512 construct: see for example @file{f/lang-specs.h}.
7514 @item %.@var{SUFFIX}
7515 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
7516 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
7517 terminated by the next space or %.
7520 Marks the argument containing or following the @samp{%w} as the
7521 designated output file of this compilation. This puts the argument
7522 into the sequence of arguments that @samp{%o} will substitute later.
7525 Substitutes the names of all the output files, with spaces
7526 automatically placed around them. You should write spaces
7527 around the @samp{%o} as well or the results are undefined.
7528 @samp{%o} is for use in the specs for running the linker.
7529 Input files whose names have no recognized suffix are not compiled
7530 at all, but they are included among the output files, so they will
7534 Substitutes the suffix for object files. Note that this is
7535 handled specially when it immediately follows @samp{%g, %u, or %U},
7536 because of the need for those to form complete file names. The
7537 handling is such that @samp{%O} is treated exactly as if it had already
7538 been substituted, except that @samp{%g, %u, and %U} do not currently
7539 support additional @var{suffix} characters following @samp{%O} as they would
7540 following, for example, @samp{.o}.
7543 Substitutes the standard macro predefinitions for the
7544 current target machine. Use this when running @code{cpp}.
7547 Like @samp{%p}, but puts @samp{__} before and after the name of each
7548 predefined macro, except for macros that start with @samp{__} or with
7549 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
7553 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
7554 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
7555 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
7556 and @option{-imultilib} as necessary.
7559 Current argument is the name of a library or startup file of some sort.
7560 Search for that file in a standard list of directories and substitute
7561 the full name found.
7564 Print @var{str} as an error message. @var{str} is terminated by a newline.
7565 Use this when inconsistent options are detected.
7568 Substitute the contents of spec string @var{name} at this point.
7571 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
7573 @item %x@{@var{option}@}
7574 Accumulate an option for @samp{%X}.
7577 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
7581 Output the accumulated assembler options specified by @option{-Wa}.
7584 Output the accumulated preprocessor options specified by @option{-Wp}.
7587 Process the @code{asm} spec. This is used to compute the
7588 switches to be passed to the assembler.
7591 Process the @code{asm_final} spec. This is a spec string for
7592 passing switches to an assembler post-processor, if such a program is
7596 Process the @code{link} spec. This is the spec for computing the
7597 command line passed to the linker. Typically it will make use of the
7598 @samp{%L %G %S %D and %E} sequences.
7601 Dump out a @option{-L} option for each directory that GCC believes might
7602 contain startup files. If the target supports multilibs then the
7603 current multilib directory will be prepended to each of these paths.
7606 Process the @code{lib} spec. This is a spec string for deciding which
7607 libraries should be included on the command line to the linker.
7610 Process the @code{libgcc} spec. This is a spec string for deciding
7611 which GCC support library should be included on the command line to the linker.
7614 Process the @code{startfile} spec. This is a spec for deciding which
7615 object files should be the first ones passed to the linker. Typically
7616 this might be a file named @file{crt0.o}.
7619 Process the @code{endfile} spec. This is a spec string that specifies
7620 the last object files that will be passed to the linker.
7623 Process the @code{cpp} spec. This is used to construct the arguments
7624 to be passed to the C preprocessor.
7627 Process the @code{cc1} spec. This is used to construct the options to be
7628 passed to the actual C compiler (@samp{cc1}).
7631 Process the @code{cc1plus} spec. This is used to construct the options to be
7632 passed to the actual C++ compiler (@samp{cc1plus}).
7635 Substitute the variable part of a matched option. See below.
7636 Note that each comma in the substituted string is replaced by
7640 Remove all occurrences of @code{-S} from the command line. Note---this
7641 command is position dependent. @samp{%} commands in the spec string
7642 before this one will see @code{-S}, @samp{%} commands in the spec string
7643 after this one will not.
7645 @item %:@var{function}(@var{args})
7646 Call the named function @var{function}, passing it @var{args}.
7647 @var{args} is first processed as a nested spec string, then split
7648 into an argument vector in the usual fashion. The function returns
7649 a string which is processed as if it had appeared literally as part
7650 of the current spec.
7652 The following built-in spec functions are provided:
7656 The @code{getenv} spec function takes two arguments: an environment
7657 variable name and a string. If the environment variable is not
7658 defined, a fatal error is issued. Otherwise, the return value is the
7659 value of the environment variable concatenated with the string. For
7660 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
7663 %:getenv(TOPDIR /include)
7666 expands to @file{/path/to/top/include}.
7668 @item @code{if-exists}
7669 The @code{if-exists} spec function takes one argument, an absolute
7670 pathname to a file. If the file exists, @code{if-exists} returns the
7671 pathname. Here is a small example of its usage:
7675 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
7678 @item @code{if-exists-else}
7679 The @code{if-exists-else} spec function is similar to the @code{if-exists}
7680 spec function, except that it takes two arguments. The first argument is
7681 an absolute pathname to a file. If the file exists, @code{if-exists-else}
7682 returns the pathname. If it does not exist, it returns the second argument.
7683 This way, @code{if-exists-else} can be used to select one file or another,
7684 based on the existence of the first. Here is a small example of its usage:
7688 crt0%O%s %:if-exists(crti%O%s) \
7689 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
7692 @item @code{replace-outfile}
7693 The @code{replace-outfile} spec function takes two arguments. It looks for the
7694 first argument in the outfiles array and replaces it with the second argument. Here
7695 is a small example of its usage:
7698 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
7701 @item @code{print-asm-header}
7702 The @code{print-asm-header} function takes no arguments and simply
7703 prints a banner like:
7709 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
7712 It is used to separate compiler options from assembler options
7713 in the @option{--target-help} output.
7717 Substitutes the @code{-S} switch, if that switch was given to GCC@.
7718 If that switch was not specified, this substitutes nothing. Note that
7719 the leading dash is omitted when specifying this option, and it is
7720 automatically inserted if the substitution is performed. Thus the spec
7721 string @samp{%@{foo@}} would match the command-line option @option{-foo}
7722 and would output the command line option @option{-foo}.
7724 @item %W@{@code{S}@}
7725 Like %@{@code{S}@} but mark last argument supplied within as a file to be
7728 @item %@{@code{S}*@}
7729 Substitutes all the switches specified to GCC whose names start
7730 with @code{-S}, but which also take an argument. This is used for
7731 switches like @option{-o}, @option{-D}, @option{-I}, etc.
7732 GCC considers @option{-o foo} as being
7733 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
7734 text, including the space. Thus two arguments would be generated.
7736 @item %@{@code{S}*&@code{T}*@}
7737 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
7738 (the order of @code{S} and @code{T} in the spec is not significant).
7739 There can be any number of ampersand-separated variables; for each the
7740 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
7742 @item %@{@code{S}:@code{X}@}
7743 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
7745 @item %@{!@code{S}:@code{X}@}
7746 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
7748 @item %@{@code{S}*:@code{X}@}
7749 Substitutes @code{X} if one or more switches whose names start with
7750 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
7751 once, no matter how many such switches appeared. However, if @code{%*}
7752 appears somewhere in @code{X}, then @code{X} will be substituted once
7753 for each matching switch, with the @code{%*} replaced by the part of
7754 that switch that matched the @code{*}.
7756 @item %@{.@code{S}:@code{X}@}
7757 Substitutes @code{X}, if processing a file with suffix @code{S}.
7759 @item %@{!.@code{S}:@code{X}@}
7760 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
7762 @item %@{,@code{S}:@code{X}@}
7763 Substitutes @code{X}, if processing a file for language @code{S}.
7765 @item %@{!,@code{S}:@code{X}@}
7766 Substitutes @code{X}, if not processing a file for language @code{S}.
7768 @item %@{@code{S}|@code{P}:@code{X}@}
7769 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
7770 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
7771 @code{*} sequences as well, although they have a stronger binding than
7772 the @samp{|}. If @code{%*} appears in @code{X}, all of the
7773 alternatives must be starred, and only the first matching alternative
7776 For example, a spec string like this:
7779 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
7782 will output the following command-line options from the following input
7783 command-line options:
7788 -d fred.c -foo -baz -boggle
7789 -d jim.d -bar -baz -boggle
7792 @item %@{S:X; T:Y; :D@}
7794 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
7795 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
7796 be as many clauses as you need. This may be combined with @code{.},
7797 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
7802 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
7803 construct may contain other nested @samp{%} constructs or spaces, or
7804 even newlines. They are processed as usual, as described above.
7805 Trailing white space in @code{X} is ignored. White space may also
7806 appear anywhere on the left side of the colon in these constructs,
7807 except between @code{.} or @code{*} and the corresponding word.
7809 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
7810 handled specifically in these constructs. If another value of
7811 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
7812 @option{-W} switch is found later in the command line, the earlier
7813 switch value is ignored, except with @{@code{S}*@} where @code{S} is
7814 just one letter, which passes all matching options.
7816 The character @samp{|} at the beginning of the predicate text is used to
7817 indicate that a command should be piped to the following command, but
7818 only if @option{-pipe} is specified.
7820 It is built into GCC which switches take arguments and which do not.
7821 (You might think it would be useful to generalize this to allow each
7822 compiler's spec to say which switches take arguments. But this cannot
7823 be done in a consistent fashion. GCC cannot even decide which input
7824 files have been specified without knowing which switches take arguments,
7825 and it must know which input files to compile in order to tell which
7828 GCC also knows implicitly that arguments starting in @option{-l} are to be
7829 treated as compiler output files, and passed to the linker in their
7830 proper position among the other output files.
7832 @c man begin OPTIONS
7834 @node Target Options
7835 @section Specifying Target Machine and Compiler Version
7836 @cindex target options
7837 @cindex cross compiling
7838 @cindex specifying machine version
7839 @cindex specifying compiler version and target machine
7840 @cindex compiler version, specifying
7841 @cindex target machine, specifying
7843 The usual way to run GCC is to run the executable called @file{gcc}, or
7844 @file{<machine>-gcc} when cross-compiling, or
7845 @file{<machine>-gcc-<version>} to run a version other than the one that
7846 was installed last. Sometimes this is inconvenient, so GCC provides
7847 options that will switch to another cross-compiler or version.
7850 @item -b @var{machine}
7852 The argument @var{machine} specifies the target machine for compilation.
7854 The value to use for @var{machine} is the same as was specified as the
7855 machine type when configuring GCC as a cross-compiler. For
7856 example, if a cross-compiler was configured with @samp{configure
7857 arm-elf}, meaning to compile for an arm processor with elf binaries,
7858 then you would specify @option{-b arm-elf} to run that cross compiler.
7859 Because there are other options beginning with @option{-b}, the
7860 configuration must contain a hyphen.
7862 @item -V @var{version}
7864 The argument @var{version} specifies which version of GCC to run.
7865 This is useful when multiple versions are installed. For example,
7866 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
7869 The @option{-V} and @option{-b} options work by running the
7870 @file{<machine>-gcc-<version>} executable, so there's no real reason to
7871 use them if you can just run that directly.
7873 @node Submodel Options
7874 @section Hardware Models and Configurations
7875 @cindex submodel options
7876 @cindex specifying hardware config
7877 @cindex hardware models and configurations, specifying
7878 @cindex machine dependent options
7880 Earlier we discussed the standard option @option{-b} which chooses among
7881 different installed compilers for completely different target
7882 machines, such as VAX vs.@: 68000 vs.@: 80386.
7884 In addition, each of these target machine types can have its own
7885 special options, starting with @samp{-m}, to choose among various
7886 hardware models or configurations---for example, 68010 vs 68020,
7887 floating coprocessor or none. A single installed version of the
7888 compiler can compile for any model or configuration, according to the
7891 Some configurations of the compiler also support additional special
7892 options, usually for compatibility with other compilers on the same
7895 @c This list is ordered alphanumerically by subsection name.
7896 @c It should be the same order and spelling as these options are listed
7897 @c in Machine Dependent Options
7903 * Blackfin Options::
7907 * DEC Alpha Options::
7908 * DEC Alpha/VMS Options::
7910 * GNU/Linux Options::
7913 * i386 and x86-64 Options::
7926 * RS/6000 and PowerPC Options::
7927 * S/390 and zSeries Options::
7932 * System V Options::
7933 * TMS320C3x/C4x Options::
7938 * Xstormy16 Options::
7944 @subsection ARC Options
7947 These options are defined for ARC implementations:
7952 Compile code for little endian mode. This is the default.
7956 Compile code for big endian mode.
7959 @opindex mmangle-cpu
7960 Prepend the name of the cpu to all public symbol names.
7961 In multiple-processor systems, there are many ARC variants with different
7962 instruction and register set characteristics. This flag prevents code
7963 compiled for one cpu to be linked with code compiled for another.
7964 No facility exists for handling variants that are ``almost identical''.
7965 This is an all or nothing option.
7967 @item -mcpu=@var{cpu}
7969 Compile code for ARC variant @var{cpu}.
7970 Which variants are supported depend on the configuration.
7971 All variants support @option{-mcpu=base}, this is the default.
7973 @item -mtext=@var{text-section}
7974 @itemx -mdata=@var{data-section}
7975 @itemx -mrodata=@var{readonly-data-section}
7979 Put functions, data, and readonly data in @var{text-section},
7980 @var{data-section}, and @var{readonly-data-section} respectively
7981 by default. This can be overridden with the @code{section} attribute.
7982 @xref{Variable Attributes}.
7987 @subsection ARM Options
7990 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
7994 @item -mabi=@var{name}
7996 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
7997 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
8000 @opindex mapcs-frame
8001 Generate a stack frame that is compliant with the ARM Procedure Call
8002 Standard for all functions, even if this is not strictly necessary for
8003 correct execution of the code. Specifying @option{-fomit-frame-pointer}
8004 with this option will cause the stack frames not to be generated for
8005 leaf functions. The default is @option{-mno-apcs-frame}.
8009 This is a synonym for @option{-mapcs-frame}.
8012 @c not currently implemented
8013 @item -mapcs-stack-check
8014 @opindex mapcs-stack-check
8015 Generate code to check the amount of stack space available upon entry to
8016 every function (that actually uses some stack space). If there is
8017 insufficient space available then either the function
8018 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
8019 called, depending upon the amount of stack space required. The run time
8020 system is required to provide these functions. The default is
8021 @option{-mno-apcs-stack-check}, since this produces smaller code.
8023 @c not currently implemented
8025 @opindex mapcs-float
8026 Pass floating point arguments using the float point registers. This is
8027 one of the variants of the APCS@. This option is recommended if the
8028 target hardware has a floating point unit or if a lot of floating point
8029 arithmetic is going to be performed by the code. The default is
8030 @option{-mno-apcs-float}, since integer only code is slightly increased in
8031 size if @option{-mapcs-float} is used.
8033 @c not currently implemented
8034 @item -mapcs-reentrant
8035 @opindex mapcs-reentrant
8036 Generate reentrant, position independent code. The default is
8037 @option{-mno-apcs-reentrant}.
8040 @item -mthumb-interwork
8041 @opindex mthumb-interwork
8042 Generate code which supports calling between the ARM and Thumb
8043 instruction sets. Without this option the two instruction sets cannot
8044 be reliably used inside one program. The default is
8045 @option{-mno-thumb-interwork}, since slightly larger code is generated
8046 when @option{-mthumb-interwork} is specified.
8048 @item -mno-sched-prolog
8049 @opindex mno-sched-prolog
8050 Prevent the reordering of instructions in the function prolog, or the
8051 merging of those instruction with the instructions in the function's
8052 body. This means that all functions will start with a recognizable set
8053 of instructions (or in fact one of a choice from a small set of
8054 different function prologues), and this information can be used to
8055 locate the start if functions inside an executable piece of code. The
8056 default is @option{-msched-prolog}.
8059 @opindex mhard-float
8060 Generate output containing floating point instructions. This is the
8064 @opindex msoft-float
8065 Generate output containing library calls for floating point.
8066 @strong{Warning:} the requisite libraries are not available for all ARM
8067 targets. Normally the facilities of the machine's usual C compiler are
8068 used, but this cannot be done directly in cross-compilation. You must make
8069 your own arrangements to provide suitable library functions for
8072 @option{-msoft-float} changes the calling convention in the output file;
8073 therefore, it is only useful if you compile @emph{all} of a program with
8074 this option. In particular, you need to compile @file{libgcc.a}, the
8075 library that comes with GCC, with @option{-msoft-float} in order for
8078 @item -mfloat-abi=@var{name}
8080 Specifies which ABI to use for floating point values. Permissible values
8081 are: @samp{soft}, @samp{softfp} and @samp{hard}.
8083 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
8084 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
8085 of floating point instructions, but still uses the soft-float calling
8088 @item -mlittle-endian
8089 @opindex mlittle-endian
8090 Generate code for a processor running in little-endian mode. This is
8091 the default for all standard configurations.
8094 @opindex mbig-endian
8095 Generate code for a processor running in big-endian mode; the default is
8096 to compile code for a little-endian processor.
8098 @item -mwords-little-endian
8099 @opindex mwords-little-endian
8100 This option only applies when generating code for big-endian processors.
8101 Generate code for a little-endian word order but a big-endian byte
8102 order. That is, a byte order of the form @samp{32107654}. Note: this
8103 option should only be used if you require compatibility with code for
8104 big-endian ARM processors generated by versions of the compiler prior to
8107 @item -mcpu=@var{name}
8109 This specifies the name of the target ARM processor. GCC uses this name
8110 to determine what kind of instructions it can emit when generating
8111 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
8112 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
8113 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
8114 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
8115 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
8116 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
8117 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
8118 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
8119 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
8120 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
8121 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
8122 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
8123 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
8124 @samp{arm1156t2-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
8125 @samp{cortex-a8}, @samp{cortex-r4}, @samp{cortex-m3},
8126 @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
8128 @itemx -mtune=@var{name}
8130 This option is very similar to the @option{-mcpu=} option, except that
8131 instead of specifying the actual target processor type, and hence
8132 restricting which instructions can be used, it specifies that GCC should
8133 tune the performance of the code as if the target were of the type
8134 specified in this option, but still choosing the instructions that it
8135 will generate based on the cpu specified by a @option{-mcpu=} option.
8136 For some ARM implementations better performance can be obtained by using
8139 @item -march=@var{name}
8141 This specifies the name of the target ARM architecture. GCC uses this
8142 name to determine what kind of instructions it can emit when generating
8143 assembly code. This option can be used in conjunction with or instead
8144 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
8145 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
8146 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
8147 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv7}, @samp{armv7-a},
8148 @samp{armv7-r}, @samp{armv7-m}, @samp{iwmmxt}, @samp{ep9312}.
8150 @item -mfpu=@var{name}
8151 @itemx -mfpe=@var{number}
8152 @itemx -mfp=@var{number}
8156 This specifies what floating point hardware (or hardware emulation) is
8157 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
8158 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
8159 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
8160 with older versions of GCC@.
8162 If @option{-msoft-float} is specified this specifies the format of
8163 floating point values.
8165 @item -mstructure-size-boundary=@var{n}
8166 @opindex mstructure-size-boundary
8167 The size of all structures and unions will be rounded up to a multiple
8168 of the number of bits set by this option. Permissible values are 8, 32
8169 and 64. The default value varies for different toolchains. For the COFF
8170 targeted toolchain the default value is 8. A value of 64 is only allowed
8171 if the underlying ABI supports it.
8173 Specifying the larger number can produce faster, more efficient code, but
8174 can also increase the size of the program. Different values are potentially
8175 incompatible. Code compiled with one value cannot necessarily expect to
8176 work with code or libraries compiled with another value, if they exchange
8177 information using structures or unions.
8179 @item -mabort-on-noreturn
8180 @opindex mabort-on-noreturn
8181 Generate a call to the function @code{abort} at the end of a
8182 @code{noreturn} function. It will be executed if the function tries to
8186 @itemx -mno-long-calls
8187 @opindex mlong-calls
8188 @opindex mno-long-calls
8189 Tells the compiler to perform function calls by first loading the
8190 address of the function into a register and then performing a subroutine
8191 call on this register. This switch is needed if the target function
8192 will lie outside of the 64 megabyte addressing range of the offset based
8193 version of subroutine call instruction.
8195 Even if this switch is enabled, not all function calls will be turned
8196 into long calls. The heuristic is that static functions, functions
8197 which have the @samp{short-call} attribute, functions that are inside
8198 the scope of a @samp{#pragma no_long_calls} directive and functions whose
8199 definitions have already been compiled within the current compilation
8200 unit, will not be turned into long calls. The exception to this rule is
8201 that weak function definitions, functions with the @samp{long-call}
8202 attribute or the @samp{section} attribute, and functions that are within
8203 the scope of a @samp{#pragma long_calls} directive, will always be
8204 turned into long calls.
8206 This feature is not enabled by default. Specifying
8207 @option{-mno-long-calls} will restore the default behavior, as will
8208 placing the function calls within the scope of a @samp{#pragma
8209 long_calls_off} directive. Note these switches have no effect on how
8210 the compiler generates code to handle function calls via function
8213 @item -mnop-fun-dllimport
8214 @opindex mnop-fun-dllimport
8215 Disable support for the @code{dllimport} attribute.
8217 @item -msingle-pic-base
8218 @opindex msingle-pic-base
8219 Treat the register used for PIC addressing as read-only, rather than
8220 loading it in the prologue for each function. The run-time system is
8221 responsible for initializing this register with an appropriate value
8222 before execution begins.
8224 @item -mpic-register=@var{reg}
8225 @opindex mpic-register
8226 Specify the register to be used for PIC addressing. The default is R10
8227 unless stack-checking is enabled, when R9 is used.
8229 @item -mcirrus-fix-invalid-insns
8230 @opindex mcirrus-fix-invalid-insns
8231 @opindex mno-cirrus-fix-invalid-insns
8232 Insert NOPs into the instruction stream to in order to work around
8233 problems with invalid Maverick instruction combinations. This option
8234 is only valid if the @option{-mcpu=ep9312} option has been used to
8235 enable generation of instructions for the Cirrus Maverick floating
8236 point co-processor. This option is not enabled by default, since the
8237 problem is only present in older Maverick implementations. The default
8238 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
8241 @item -mpoke-function-name
8242 @opindex mpoke-function-name
8243 Write the name of each function into the text section, directly
8244 preceding the function prologue. The generated code is similar to this:
8248 .ascii "arm_poke_function_name", 0
8251 .word 0xff000000 + (t1 - t0)
8252 arm_poke_function_name
8254 stmfd sp!, @{fp, ip, lr, pc@}
8258 When performing a stack backtrace, code can inspect the value of
8259 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
8260 location @code{pc - 12} and the top 8 bits are set, then we know that
8261 there is a function name embedded immediately preceding this location
8262 and has length @code{((pc[-3]) & 0xff000000)}.
8266 Generate code for the Thumb instruction set. The default is to
8267 use the 32-bit ARM instruction set.
8268 This option automatically enables either 16-bit Thumb-1 or
8269 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
8270 and @option{-march=@var{name}} options.
8273 @opindex mtpcs-frame
8274 Generate a stack frame that is compliant with the Thumb Procedure Call
8275 Standard for all non-leaf functions. (A leaf function is one that does
8276 not call any other functions.) The default is @option{-mno-tpcs-frame}.
8278 @item -mtpcs-leaf-frame
8279 @opindex mtpcs-leaf-frame
8280 Generate a stack frame that is compliant with the Thumb Procedure Call
8281 Standard for all leaf functions. (A leaf function is one that does
8282 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
8284 @item -mcallee-super-interworking
8285 @opindex mcallee-super-interworking
8286 Gives all externally visible functions in the file being compiled an ARM
8287 instruction set header which switches to Thumb mode before executing the
8288 rest of the function. This allows these functions to be called from
8289 non-interworking code.
8291 @item -mcaller-super-interworking
8292 @opindex mcaller-super-interworking
8293 Allows calls via function pointers (including virtual functions) to
8294 execute correctly regardless of whether the target code has been
8295 compiled for interworking or not. There is a small overhead in the cost
8296 of executing a function pointer if this option is enabled.
8298 @item -mtp=@var{name}
8300 Specify the access model for the thread local storage pointer. The valid
8301 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
8302 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
8303 (supported in the arm6k architecture), and @option{auto}, which uses the
8304 best available method for the selected processor. The default setting is
8310 @subsection AVR Options
8313 These options are defined for AVR implementations:
8316 @item -mmcu=@var{mcu}
8318 Specify ATMEL AVR instruction set or MCU type.
8320 Instruction set avr1 is for the minimal AVR core, not supported by the C
8321 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8322 attiny11, attiny12, attiny15, attiny28).
8324 Instruction set avr2 (default) is for the classic AVR core with up to
8325 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8326 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8327 at90c8534, at90s8535).
8329 Instruction set avr3 is for the classic AVR core with up to 128K program
8330 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8332 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8333 memory space (MCU types: atmega8, atmega83, atmega85).
8335 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8336 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8337 atmega64, atmega128, at43usb355, at94k).
8341 Output instruction sizes to the asm file.
8343 @item -minit-stack=@var{N}
8344 @opindex minit-stack
8345 Specify the initial stack address, which may be a symbol or numeric value,
8346 @samp{__stack} is the default.
8348 @item -mno-interrupts
8349 @opindex mno-interrupts
8350 Generated code is not compatible with hardware interrupts.
8351 Code size will be smaller.
8353 @item -mcall-prologues
8354 @opindex mcall-prologues
8355 Functions prologues/epilogues expanded as call to appropriate
8356 subroutines. Code size will be smaller.
8358 @item -mno-tablejump
8359 @opindex mno-tablejump
8360 Do not generate tablejump insns which sometimes increase code size.
8363 @opindex mtiny-stack
8364 Change only the low 8 bits of the stack pointer.
8368 Assume int to be 8 bit integer. This affects the sizes of all types: A
8369 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
8370 and long long will be 4 bytes. Please note that this option does not
8371 comply to the C standards, but it will provide you with smaller code
8375 @node Blackfin Options
8376 @subsection Blackfin Options
8377 @cindex Blackfin Options
8380 @item -mcpu=@var{cpu}
8382 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
8383 can be one of @samp{bf531}, @samp{bf532}, @samp{bf533},
8384 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf561}.
8385 Without this option, @samp{bf532} is used as the processor by default.
8386 The corresponding predefined processor macros for @var{cpu} is to
8387 be defined. For the @samp{bfin-elf} toolchain, this causes the hardware
8388 BSP provided by libgloss to be linked in if @samp{-msim} is not given.
8389 Support for @samp{bf561} is incomplete; only the processor macro is defined.
8393 Specifies that the program will be run on the simulator. This causes
8394 the simulator BSP provided by libgloss to be linked in. This option
8395 has effect only for @samp{bfin-elf} toolchain.
8397 @item -momit-leaf-frame-pointer
8398 @opindex momit-leaf-frame-pointer
8399 Don't keep the frame pointer in a register for leaf functions. This
8400 avoids the instructions to save, set up and restore frame pointers and
8401 makes an extra register available in leaf functions. The option
8402 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8403 which might make debugging harder.
8405 @item -mspecld-anomaly
8406 @opindex mspecld-anomaly
8407 When enabled, the compiler will ensure that the generated code does not
8408 contain speculative loads after jump instructions. This option is enabled
8411 @item -mno-specld-anomaly
8412 @opindex mno-specld-anomaly
8413 Don't generate extra code to prevent speculative loads from occurring.
8415 @item -mcsync-anomaly
8416 @opindex mcsync-anomaly
8417 When enabled, the compiler will ensure that the generated code does not
8418 contain CSYNC or SSYNC instructions too soon after conditional branches.
8419 This option is enabled by default.
8421 @item -mno-csync-anomaly
8422 @opindex mno-csync-anomaly
8423 Don't generate extra code to prevent CSYNC or SSYNC instructions from
8424 occurring too soon after a conditional branch.
8428 When enabled, the compiler is free to take advantage of the knowledge that
8429 the entire program fits into the low 64k of memory.
8432 @opindex mno-low-64k
8433 Assume that the program is arbitrarily large. This is the default.
8435 @item -mstack-check-l1
8436 @opindex mstack-check-l1
8437 Do stack checking using information placed into L1 scratchpad memory by the
8440 @item -mid-shared-library
8441 @opindex mid-shared-library
8442 Generate code that supports shared libraries via the library ID method.
8443 This allows for execute in place and shared libraries in an environment
8444 without virtual memory management. This option implies @option{-fPIC}.
8446 @item -mno-id-shared-library
8447 @opindex mno-id-shared-library
8448 Generate code that doesn't assume ID based shared libraries are being used.
8449 This is the default.
8451 @item -mleaf-id-shared-library
8452 @opindex mleaf-id-shared-library
8453 Generate code that supports shared libraries via the library ID method,
8454 but assumes that this library or executable won't link against any other
8455 ID shared libraries. That allows the compiler to use faster code for jumps
8458 @item -mno-leaf-id-shared-library
8459 @opindex mno-leaf-id-shared-library
8460 Do not assume that the code being compiled won't link against any ID shared
8461 libraries. Slower code will be generated for jump and call insns.
8463 @item -mshared-library-id=n
8464 @opindex mshared-library-id
8465 Specified the identification number of the ID based shared library being
8466 compiled. Specifying a value of 0 will generate more compact code, specifying
8467 other values will force the allocation of that number to the current
8468 library but is no more space or time efficient than omitting this option.
8472 Generate code that allows the data segment to be located in a different
8473 area of memory from the text segment. This allows for execute in place in
8474 an environment without virtual memory management by eliminating relocations
8475 against the text section.
8478 @opindex mno-sep-data
8479 Generate code that assumes that the data segment follows the text segment.
8480 This is the default.
8483 @itemx -mno-long-calls
8484 @opindex mlong-calls
8485 @opindex mno-long-calls
8486 Tells the compiler to perform function calls by first loading the
8487 address of the function into a register and then performing a subroutine
8488 call on this register. This switch is needed if the target function
8489 will lie outside of the 24 bit addressing range of the offset based
8490 version of subroutine call instruction.
8492 This feature is not enabled by default. Specifying
8493 @option{-mno-long-calls} will restore the default behavior. Note these
8494 switches have no effect on how the compiler generates code to handle
8495 function calls via function pointers.
8499 @subsection CRIS Options
8500 @cindex CRIS Options
8502 These options are defined specifically for the CRIS ports.
8505 @item -march=@var{architecture-type}
8506 @itemx -mcpu=@var{architecture-type}
8509 Generate code for the specified architecture. The choices for
8510 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
8511 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
8512 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
8515 @item -mtune=@var{architecture-type}
8517 Tune to @var{architecture-type} everything applicable about the generated
8518 code, except for the ABI and the set of available instructions. The
8519 choices for @var{architecture-type} are the same as for
8520 @option{-march=@var{architecture-type}}.
8522 @item -mmax-stack-frame=@var{n}
8523 @opindex mmax-stack-frame
8524 Warn when the stack frame of a function exceeds @var{n} bytes.
8526 @item -melinux-stacksize=@var{n}
8527 @opindex melinux-stacksize
8528 Only available with the @samp{cris-axis-aout} target. Arranges for
8529 indications in the program to the kernel loader that the stack of the
8530 program should be set to @var{n} bytes.
8536 The options @option{-metrax4} and @option{-metrax100} are synonyms for
8537 @option{-march=v3} and @option{-march=v8} respectively.
8539 @item -mmul-bug-workaround
8540 @itemx -mno-mul-bug-workaround
8541 @opindex mmul-bug-workaround
8542 @opindex mno-mul-bug-workaround
8543 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
8544 models where it applies. This option is active by default.
8548 Enable CRIS-specific verbose debug-related information in the assembly
8549 code. This option also has the effect to turn off the @samp{#NO_APP}
8550 formatted-code indicator to the assembler at the beginning of the
8555 Do not use condition-code results from previous instruction; always emit
8556 compare and test instructions before use of condition codes.
8558 @item -mno-side-effects
8559 @opindex mno-side-effects
8560 Do not emit instructions with side-effects in addressing modes other than
8564 @itemx -mno-stack-align
8566 @itemx -mno-data-align
8567 @itemx -mconst-align
8568 @itemx -mno-const-align
8569 @opindex mstack-align
8570 @opindex mno-stack-align
8571 @opindex mdata-align
8572 @opindex mno-data-align
8573 @opindex mconst-align
8574 @opindex mno-const-align
8575 These options (no-options) arranges (eliminate arrangements) for the
8576 stack-frame, individual data and constants to be aligned for the maximum
8577 single data access size for the chosen CPU model. The default is to
8578 arrange for 32-bit alignment. ABI details such as structure layout are
8579 not affected by these options.
8587 Similar to the stack- data- and const-align options above, these options
8588 arrange for stack-frame, writable data and constants to all be 32-bit,
8589 16-bit or 8-bit aligned. The default is 32-bit alignment.
8591 @item -mno-prologue-epilogue
8592 @itemx -mprologue-epilogue
8593 @opindex mno-prologue-epilogue
8594 @opindex mprologue-epilogue
8595 With @option{-mno-prologue-epilogue}, the normal function prologue and
8596 epilogue that sets up the stack-frame are omitted and no return
8597 instructions or return sequences are generated in the code. Use this
8598 option only together with visual inspection of the compiled code: no
8599 warnings or errors are generated when call-saved registers must be saved,
8600 or storage for local variable needs to be allocated.
8606 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
8607 instruction sequences that load addresses for functions from the PLT part
8608 of the GOT rather than (traditional on other architectures) calls to the
8609 PLT@. The default is @option{-mgotplt}.
8613 Legacy no-op option only recognized with the cris-axis-aout target.
8617 Legacy no-op option only recognized with the cris-axis-elf and
8618 cris-axis-linux-gnu targets.
8622 Only recognized with the cris-axis-aout target, where it selects a
8623 GNU/linux-like multilib, include files and instruction set for
8628 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
8632 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
8633 to link with input-output functions from a simulator library. Code,
8634 initialized data and zero-initialized data are allocated consecutively.
8638 Like @option{-sim}, but pass linker options to locate initialized data at
8639 0x40000000 and zero-initialized data at 0x80000000.
8643 @subsection CRX Options
8646 These options are defined specifically for the CRX ports.
8652 Enable the use of multiply-accumulate instructions. Disabled by default.
8656 Push instructions will be used to pass outgoing arguments when functions
8657 are called. Enabled by default.
8660 @node Darwin Options
8661 @subsection Darwin Options
8662 @cindex Darwin options
8664 These options are defined for all architectures running the Darwin operating
8667 FSF GCC on Darwin does not create ``fat'' object files; it will create
8668 an object file for the single architecture that it was built to
8669 target. Apple's GCC on Darwin does create ``fat'' files if multiple
8670 @option{-arch} options are used; it does so by running the compiler or
8671 linker multiple times and joining the results together with
8674 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
8675 @samp{i686}) is determined by the flags that specify the ISA
8676 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
8677 @option{-force_cpusubtype_ALL} option can be used to override this.
8679 The Darwin tools vary in their behavior when presented with an ISA
8680 mismatch. The assembler, @file{as}, will only permit instructions to
8681 be used that are valid for the subtype of the file it is generating,
8682 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
8683 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
8684 and print an error if asked to create a shared library with a less
8685 restrictive subtype than its input files (for instance, trying to put
8686 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
8687 for executables, @file{ld}, will quietly give the executable the most
8688 restrictive subtype of any of its input files.
8693 Add the framework directory @var{dir} to the head of the list of
8694 directories to be searched for header files. These directories are
8695 interleaved with those specified by @option{-I} options and are
8696 scanned in a left-to-right order.
8698 A framework directory is a directory with frameworks in it. A
8699 framework is a directory with a @samp{"Headers"} and/or
8700 @samp{"PrivateHeaders"} directory contained directly in it that ends
8701 in @samp{".framework"}. The name of a framework is the name of this
8702 directory excluding the @samp{".framework"}. Headers associated with
8703 the framework are found in one of those two directories, with
8704 @samp{"Headers"} being searched first. A subframework is a framework
8705 directory that is in a framework's @samp{"Frameworks"} directory.
8706 Includes of subframework headers can only appear in a header of a
8707 framework that contains the subframework, or in a sibling subframework
8708 header. Two subframeworks are siblings if they occur in the same
8709 framework. A subframework should not have the same name as a
8710 framework, a warning will be issued if this is violated. Currently a
8711 subframework cannot have subframeworks, in the future, the mechanism
8712 may be extended to support this. The standard frameworks can be found
8713 in @samp{"/System/Library/Frameworks"} and
8714 @samp{"/Library/Frameworks"}. An example include looks like
8715 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
8716 the name of the framework and header.h is found in the
8717 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
8719 @item -iframework@var{dir}
8721 Like @option{-F} except the directory is a treated as a system
8722 directory. The main difference between this @option{-iframework} and
8723 @option{-F} is that with @option{-iframework} the compiler does not
8724 warn about constructs contained within header files found via
8725 @var{dir}. This option is valid only for the C family of languages.
8729 Emit debugging information for symbols that are used. For STABS
8730 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
8731 This is by default ON@.
8735 Emit debugging information for all symbols and types.
8737 @item -mmacosx-version-min=@var{version}
8738 The earliest version of MacOS X that this executable will run on
8739 is @var{version}. Typical values of @var{version} include @code{10.1},
8740 @code{10.2}, and @code{10.3.9}.
8742 The default for this option is to make choices that seem to be most
8747 Enable kernel development mode. The @option{-mkernel} option sets
8748 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
8749 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
8750 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
8751 applicable. This mode also sets @option{-mno-altivec},
8752 @option{-msoft-float}, @option{-fno-builtin} and
8753 @option{-mlong-branch} for PowerPC targets.
8755 @item -mone-byte-bool
8756 @opindex mone-byte-bool
8757 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
8758 By default @samp{sizeof(bool)} is @samp{4} when compiling for
8759 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
8760 option has no effect on x86.
8762 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
8763 to generate code that is not binary compatible with code generated
8764 without that switch. Using this switch may require recompiling all
8765 other modules in a program, including system libraries. Use this
8766 switch to conform to a non-default data model.
8768 @item -mfix-and-continue
8769 @itemx -ffix-and-continue
8770 @itemx -findirect-data
8771 @opindex mfix-and-continue
8772 @opindex ffix-and-continue
8773 @opindex findirect-data
8774 Generate code suitable for fast turn around development. Needed to
8775 enable gdb to dynamically load @code{.o} files into already running
8776 programs. @option{-findirect-data} and @option{-ffix-and-continue}
8777 are provided for backwards compatibility.
8781 Loads all members of static archive libraries.
8782 See man ld(1) for more information.
8784 @item -arch_errors_fatal
8785 @opindex arch_errors_fatal
8786 Cause the errors having to do with files that have the wrong architecture
8790 @opindex bind_at_load
8791 Causes the output file to be marked such that the dynamic linker will
8792 bind all undefined references when the file is loaded or launched.
8796 Produce a Mach-o bundle format file.
8797 See man ld(1) for more information.
8799 @item -bundle_loader @var{executable}
8800 @opindex bundle_loader
8801 This option specifies the @var{executable} that will be loading the build
8802 output file being linked. See man ld(1) for more information.
8806 When passed this option, GCC will produce a dynamic library instead of
8807 an executable when linking, using the Darwin @file{libtool} command.
8809 @item -force_cpusubtype_ALL
8810 @opindex force_cpusubtype_ALL
8811 This causes GCC's output file to have the @var{ALL} subtype, instead of
8812 one controlled by the @option{-mcpu} or @option{-march} option.
8814 @item -allowable_client @var{client_name}
8816 @itemx -compatibility_version
8817 @itemx -current_version
8819 @itemx -dependency-file
8821 @itemx -dylinker_install_name
8823 @itemx -exported_symbols_list
8825 @itemx -flat_namespace
8826 @itemx -force_flat_namespace
8827 @itemx -headerpad_max_install_names
8830 @itemx -install_name
8831 @itemx -keep_private_externs
8832 @itemx -multi_module
8833 @itemx -multiply_defined
8834 @itemx -multiply_defined_unused
8836 @itemx -no_dead_strip_inits_and_terms
8837 @itemx -nofixprebinding
8840 @itemx -noseglinkedit
8841 @itemx -pagezero_size
8843 @itemx -prebind_all_twolevel_modules
8844 @itemx -private_bundle
8845 @itemx -read_only_relocs
8847 @itemx -sectobjectsymbols
8851 @itemx -sectobjectsymbols
8854 @itemx -segs_read_only_addr
8855 @itemx -segs_read_write_addr
8856 @itemx -seg_addr_table
8857 @itemx -seg_addr_table_filename
8860 @itemx -segs_read_only_addr
8861 @itemx -segs_read_write_addr
8862 @itemx -single_module
8865 @itemx -sub_umbrella
8866 @itemx -twolevel_namespace
8869 @itemx -unexported_symbols_list
8870 @itemx -weak_reference_mismatches
8873 @opindex allowable_client
8874 @opindex client_name
8875 @opindex compatibility_version
8876 @opindex current_version
8878 @opindex dependency-file
8880 @opindex dylinker_install_name
8882 @opindex exported_symbols_list
8884 @opindex flat_namespace
8885 @opindex force_flat_namespace
8886 @opindex headerpad_max_install_names
8889 @opindex install_name
8890 @opindex keep_private_externs
8891 @opindex multi_module
8892 @opindex multiply_defined
8893 @opindex multiply_defined_unused
8895 @opindex no_dead_strip_inits_and_terms
8896 @opindex nofixprebinding
8897 @opindex nomultidefs
8899 @opindex noseglinkedit
8900 @opindex pagezero_size
8902 @opindex prebind_all_twolevel_modules
8903 @opindex private_bundle
8904 @opindex read_only_relocs
8906 @opindex sectobjectsymbols
8910 @opindex sectobjectsymbols
8913 @opindex segs_read_only_addr
8914 @opindex segs_read_write_addr
8915 @opindex seg_addr_table
8916 @opindex seg_addr_table_filename
8917 @opindex seglinkedit
8919 @opindex segs_read_only_addr
8920 @opindex segs_read_write_addr
8921 @opindex single_module
8923 @opindex sub_library
8924 @opindex sub_umbrella
8925 @opindex twolevel_namespace
8928 @opindex unexported_symbols_list
8929 @opindex weak_reference_mismatches
8930 @opindex whatsloaded
8932 These options are passed to the Darwin linker. The Darwin linker man page
8933 describes them in detail.
8936 @node DEC Alpha Options
8937 @subsection DEC Alpha Options
8939 These @samp{-m} options are defined for the DEC Alpha implementations:
8942 @item -mno-soft-float
8944 @opindex mno-soft-float
8945 @opindex msoft-float
8946 Use (do not use) the hardware floating-point instructions for
8947 floating-point operations. When @option{-msoft-float} is specified,
8948 functions in @file{libgcc.a} will be used to perform floating-point
8949 operations. Unless they are replaced by routines that emulate the
8950 floating-point operations, or compiled in such a way as to call such
8951 emulations routines, these routines will issue floating-point
8952 operations. If you are compiling for an Alpha without floating-point
8953 operations, you must ensure that the library is built so as not to call
8956 Note that Alpha implementations without floating-point operations are
8957 required to have floating-point registers.
8962 @opindex mno-fp-regs
8963 Generate code that uses (does not use) the floating-point register set.
8964 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8965 register set is not used, floating point operands are passed in integer
8966 registers as if they were integers and floating-point results are passed
8967 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8968 so any function with a floating-point argument or return value called by code
8969 compiled with @option{-mno-fp-regs} must also be compiled with that
8972 A typical use of this option is building a kernel that does not use,
8973 and hence need not save and restore, any floating-point registers.
8977 The Alpha architecture implements floating-point hardware optimized for
8978 maximum performance. It is mostly compliant with the IEEE floating
8979 point standard. However, for full compliance, software assistance is
8980 required. This option generates code fully IEEE compliant code
8981 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8982 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8983 defined during compilation. The resulting code is less efficient but is
8984 able to correctly support denormalized numbers and exceptional IEEE
8985 values such as not-a-number and plus/minus infinity. Other Alpha
8986 compilers call this option @option{-ieee_with_no_inexact}.
8988 @item -mieee-with-inexact
8989 @opindex mieee-with-inexact
8990 This is like @option{-mieee} except the generated code also maintains
8991 the IEEE @var{inexact-flag}. Turning on this option causes the
8992 generated code to implement fully-compliant IEEE math. In addition to
8993 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8994 macro. On some Alpha implementations the resulting code may execute
8995 significantly slower than the code generated by default. Since there is
8996 very little code that depends on the @var{inexact-flag}, you should
8997 normally not specify this option. Other Alpha compilers call this
8998 option @option{-ieee_with_inexact}.
9000 @item -mfp-trap-mode=@var{trap-mode}
9001 @opindex mfp-trap-mode
9002 This option controls what floating-point related traps are enabled.
9003 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9004 The trap mode can be set to one of four values:
9008 This is the default (normal) setting. The only traps that are enabled
9009 are the ones that cannot be disabled in software (e.g., division by zero
9013 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9017 Like @samp{u}, but the instructions are marked to be safe for software
9018 completion (see Alpha architecture manual for details).
9021 Like @samp{su}, but inexact traps are enabled as well.
9024 @item -mfp-rounding-mode=@var{rounding-mode}
9025 @opindex mfp-rounding-mode
9026 Selects the IEEE rounding mode. Other Alpha compilers call this option
9027 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9032 Normal IEEE rounding mode. Floating point numbers are rounded towards
9033 the nearest machine number or towards the even machine number in case
9037 Round towards minus infinity.
9040 Chopped rounding mode. Floating point numbers are rounded towards zero.
9043 Dynamic rounding mode. A field in the floating point control register
9044 (@var{fpcr}, see Alpha architecture reference manual) controls the
9045 rounding mode in effect. The C library initializes this register for
9046 rounding towards plus infinity. Thus, unless your program modifies the
9047 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9050 @item -mtrap-precision=@var{trap-precision}
9051 @opindex mtrap-precision
9052 In the Alpha architecture, floating point traps are imprecise. This
9053 means without software assistance it is impossible to recover from a
9054 floating trap and program execution normally needs to be terminated.
9055 GCC can generate code that can assist operating system trap handlers
9056 in determining the exact location that caused a floating point trap.
9057 Depending on the requirements of an application, different levels of
9058 precisions can be selected:
9062 Program precision. This option is the default and means a trap handler
9063 can only identify which program caused a floating point exception.
9066 Function precision. The trap handler can determine the function that
9067 caused a floating point exception.
9070 Instruction precision. The trap handler can determine the exact
9071 instruction that caused a floating point exception.
9074 Other Alpha compilers provide the equivalent options called
9075 @option{-scope_safe} and @option{-resumption_safe}.
9077 @item -mieee-conformant
9078 @opindex mieee-conformant
9079 This option marks the generated code as IEEE conformant. You must not
9080 use this option unless you also specify @option{-mtrap-precision=i} and either
9081 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9082 is to emit the line @samp{.eflag 48} in the function prologue of the
9083 generated assembly file. Under DEC Unix, this has the effect that
9084 IEEE-conformant math library routines will be linked in.
9086 @item -mbuild-constants
9087 @opindex mbuild-constants
9088 Normally GCC examines a 32- or 64-bit integer constant to
9089 see if it can construct it from smaller constants in two or three
9090 instructions. If it cannot, it will output the constant as a literal and
9091 generate code to load it from the data segment at runtime.
9093 Use this option to require GCC to construct @emph{all} integer constants
9094 using code, even if it takes more instructions (the maximum is six).
9096 You would typically use this option to build a shared library dynamic
9097 loader. Itself a shared library, it must relocate itself in memory
9098 before it can find the variables and constants in its own data segment.
9104 Select whether to generate code to be assembled by the vendor-supplied
9105 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9123 Indicate whether GCC should generate code to use the optional BWX,
9124 CIX, FIX and MAX instruction sets. The default is to use the instruction
9125 sets supported by the CPU type specified via @option{-mcpu=} option or that
9126 of the CPU on which GCC was built if none was specified.
9131 @opindex mfloat-ieee
9132 Generate code that uses (does not use) VAX F and G floating point
9133 arithmetic instead of IEEE single and double precision.
9135 @item -mexplicit-relocs
9136 @itemx -mno-explicit-relocs
9137 @opindex mexplicit-relocs
9138 @opindex mno-explicit-relocs
9139 Older Alpha assemblers provided no way to generate symbol relocations
9140 except via assembler macros. Use of these macros does not allow
9141 optimal instruction scheduling. GNU binutils as of version 2.12
9142 supports a new syntax that allows the compiler to explicitly mark
9143 which relocations should apply to which instructions. This option
9144 is mostly useful for debugging, as GCC detects the capabilities of
9145 the assembler when it is built and sets the default accordingly.
9149 @opindex msmall-data
9150 @opindex mlarge-data
9151 When @option{-mexplicit-relocs} is in effect, static data is
9152 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9153 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9154 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9155 16-bit relocations off of the @code{$gp} register. This limits the
9156 size of the small data area to 64KB, but allows the variables to be
9157 directly accessed via a single instruction.
9159 The default is @option{-mlarge-data}. With this option the data area
9160 is limited to just below 2GB@. Programs that require more than 2GB of
9161 data must use @code{malloc} or @code{mmap} to allocate the data in the
9162 heap instead of in the program's data segment.
9164 When generating code for shared libraries, @option{-fpic} implies
9165 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9169 @opindex msmall-text
9170 @opindex mlarge-text
9171 When @option{-msmall-text} is used, the compiler assumes that the
9172 code of the entire program (or shared library) fits in 4MB, and is
9173 thus reachable with a branch instruction. When @option{-msmall-data}
9174 is used, the compiler can assume that all local symbols share the
9175 same @code{$gp} value, and thus reduce the number of instructions
9176 required for a function call from 4 to 1.
9178 The default is @option{-mlarge-text}.
9180 @item -mcpu=@var{cpu_type}
9182 Set the instruction set and instruction scheduling parameters for
9183 machine type @var{cpu_type}. You can specify either the @samp{EV}
9184 style name or the corresponding chip number. GCC supports scheduling
9185 parameters for the EV4, EV5 and EV6 family of processors and will
9186 choose the default values for the instruction set from the processor
9187 you specify. If you do not specify a processor type, GCC will default
9188 to the processor on which the compiler was built.
9190 Supported values for @var{cpu_type} are
9196 Schedules as an EV4 and has no instruction set extensions.
9200 Schedules as an EV5 and has no instruction set extensions.
9204 Schedules as an EV5 and supports the BWX extension.
9209 Schedules as an EV5 and supports the BWX and MAX extensions.
9213 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9217 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9220 @item -mtune=@var{cpu_type}
9222 Set only the instruction scheduling parameters for machine type
9223 @var{cpu_type}. The instruction set is not changed.
9225 @item -mmemory-latency=@var{time}
9226 @opindex mmemory-latency
9227 Sets the latency the scheduler should assume for typical memory
9228 references as seen by the application. This number is highly
9229 dependent on the memory access patterns used by the application
9230 and the size of the external cache on the machine.
9232 Valid options for @var{time} are
9236 A decimal number representing clock cycles.
9242 The compiler contains estimates of the number of clock cycles for
9243 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9244 (also called Dcache, Scache, and Bcache), as well as to main memory.
9245 Note that L3 is only valid for EV5.
9250 @node DEC Alpha/VMS Options
9251 @subsection DEC Alpha/VMS Options
9253 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9256 @item -mvms-return-codes
9257 @opindex mvms-return-codes
9258 Return VMS condition codes from main. The default is to return POSIX
9259 style condition (e.g.@ error) codes.
9263 @subsection FRV Options
9270 Only use the first 32 general purpose registers.
9275 Use all 64 general purpose registers.
9280 Use only the first 32 floating point registers.
9285 Use all 64 floating point registers
9288 @opindex mhard-float
9290 Use hardware instructions for floating point operations.
9293 @opindex msoft-float
9295 Use library routines for floating point operations.
9300 Dynamically allocate condition code registers.
9305 Do not try to dynamically allocate condition code registers, only
9306 use @code{icc0} and @code{fcc0}.
9311 Change ABI to use double word insns.
9316 Do not use double word instructions.
9321 Use floating point double instructions.
9326 Do not use floating point double instructions.
9331 Use media instructions.
9336 Do not use media instructions.
9341 Use multiply and add/subtract instructions.
9346 Do not use multiply and add/subtract instructions.
9351 Select the FDPIC ABI, that uses function descriptors to represent
9352 pointers to functions. Without any PIC/PIE-related options, it
9353 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
9354 assumes GOT entries and small data are within a 12-bit range from the
9355 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
9356 are computed with 32 bits.
9359 @opindex minline-plt
9361 Enable inlining of PLT entries in function calls to functions that are
9362 not known to bind locally. It has no effect without @option{-mfdpic}.
9363 It's enabled by default if optimizing for speed and compiling for
9364 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
9365 optimization option such as @option{-O3} or above is present in the
9371 Assume a large TLS segment when generating thread-local code.
9376 Do not assume a large TLS segment when generating thread-local code.
9381 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
9382 that is known to be in read-only sections. It's enabled by default,
9383 except for @option{-fpic} or @option{-fpie}: even though it may help
9384 make the global offset table smaller, it trades 1 instruction for 4.
9385 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
9386 one of which may be shared by multiple symbols, and it avoids the need
9387 for a GOT entry for the referenced symbol, so it's more likely to be a
9388 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
9390 @item -multilib-library-pic
9391 @opindex multilib-library-pic
9393 Link with the (library, not FD) pic libraries. It's implied by
9394 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
9395 @option{-fpic} without @option{-mfdpic}. You should never have to use
9401 Follow the EABI requirement of always creating a frame pointer whenever
9402 a stack frame is allocated. This option is enabled by default and can
9403 be disabled with @option{-mno-linked-fp}.
9406 @opindex mlong-calls
9408 Use indirect addressing to call functions outside the current
9409 compilation unit. This allows the functions to be placed anywhere
9410 within the 32-bit address space.
9412 @item -malign-labels
9413 @opindex malign-labels
9415 Try to align labels to an 8-byte boundary by inserting nops into the
9416 previous packet. This option only has an effect when VLIW packing
9417 is enabled. It doesn't create new packets; it merely adds nops to
9421 @opindex mlibrary-pic
9423 Generate position-independent EABI code.
9428 Use only the first four media accumulator registers.
9433 Use all eight media accumulator registers.
9438 Pack VLIW instructions.
9443 Do not pack VLIW instructions.
9448 Do not mark ABI switches in e_flags.
9453 Enable the use of conditional-move instructions (default).
9455 This switch is mainly for debugging the compiler and will likely be removed
9456 in a future version.
9458 @item -mno-cond-move
9459 @opindex mno-cond-move
9461 Disable the use of conditional-move instructions.
9463 This switch is mainly for debugging the compiler and will likely be removed
9464 in a future version.
9469 Enable the use of conditional set instructions (default).
9471 This switch is mainly for debugging the compiler and will likely be removed
9472 in a future version.
9477 Disable the use of conditional set instructions.
9479 This switch is mainly for debugging the compiler and will likely be removed
9480 in a future version.
9485 Enable the use of conditional execution (default).
9487 This switch is mainly for debugging the compiler and will likely be removed
9488 in a future version.
9490 @item -mno-cond-exec
9491 @opindex mno-cond-exec
9493 Disable the use of conditional execution.
9495 This switch is mainly for debugging the compiler and will likely be removed
9496 in a future version.
9499 @opindex mvliw-branch
9501 Run a pass to pack branches into VLIW instructions (default).
9503 This switch is mainly for debugging the compiler and will likely be removed
9504 in a future version.
9506 @item -mno-vliw-branch
9507 @opindex mno-vliw-branch
9509 Do not run a pass to pack branches into VLIW instructions.
9511 This switch is mainly for debugging the compiler and will likely be removed
9512 in a future version.
9514 @item -mmulti-cond-exec
9515 @opindex mmulti-cond-exec
9517 Enable optimization of @code{&&} and @code{||} in conditional execution
9520 This switch is mainly for debugging the compiler and will likely be removed
9521 in a future version.
9523 @item -mno-multi-cond-exec
9524 @opindex mno-multi-cond-exec
9526 Disable optimization of @code{&&} and @code{||} in conditional execution.
9528 This switch is mainly for debugging the compiler and will likely be removed
9529 in a future version.
9531 @item -mnested-cond-exec
9532 @opindex mnested-cond-exec
9534 Enable nested conditional execution optimizations (default).
9536 This switch is mainly for debugging the compiler and will likely be removed
9537 in a future version.
9539 @item -mno-nested-cond-exec
9540 @opindex mno-nested-cond-exec
9542 Disable nested conditional execution optimizations.
9544 This switch is mainly for debugging the compiler and will likely be removed
9545 in a future version.
9547 @item -moptimize-membar
9548 @opindex moptimize-membar
9550 This switch removes redundant @code{membar} instructions from the
9551 compiler generated code. It is enabled by default.
9553 @item -mno-optimize-membar
9554 @opindex mno-optimize-membar
9556 This switch disables the automatic removal of redundant @code{membar}
9557 instructions from the generated code.
9559 @item -mtomcat-stats
9560 @opindex mtomcat-stats
9562 Cause gas to print out tomcat statistics.
9564 @item -mcpu=@var{cpu}
9567 Select the processor type for which to generate code. Possible values are
9568 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
9569 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
9573 @node GNU/Linux Options
9574 @subsection GNU/Linux Options
9576 These @samp{-m} options are defined for GNU/Linux targets:
9581 Use the GNU C library instead of uClibc. This is the default except
9582 on @samp{*-*-linux-*uclibc*} targets.
9586 Use uClibc instead of the GNU C library. This is the default on
9587 @samp{*-*-linux-*uclibc*} targets.
9590 @node H8/300 Options
9591 @subsection H8/300 Options
9593 These @samp{-m} options are defined for the H8/300 implementations:
9598 Shorten some address references at link time, when possible; uses the
9599 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9600 ld, Using ld}, for a fuller description.
9604 Generate code for the H8/300H@.
9608 Generate code for the H8S@.
9612 Generate code for the H8S and H8/300H in the normal mode. This switch
9613 must be used either with @option{-mh} or @option{-ms}.
9617 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9621 Make @code{int} data 32 bits by default.
9625 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9626 The default for the H8/300H and H8S is to align longs and floats on 4
9628 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9629 This option has no effect on the H8/300.
9633 @subsection HPPA Options
9634 @cindex HPPA Options
9636 These @samp{-m} options are defined for the HPPA family of computers:
9639 @item -march=@var{architecture-type}
9641 Generate code for the specified architecture. The choices for
9642 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
9643 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
9644 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
9645 architecture option for your machine. Code compiled for lower numbered
9646 architectures will run on higher numbered architectures, but not the
9650 @itemx -mpa-risc-1-1
9651 @itemx -mpa-risc-2-0
9652 @opindex mpa-risc-1-0
9653 @opindex mpa-risc-1-1
9654 @opindex mpa-risc-2-0
9655 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
9658 @opindex mbig-switch
9659 Generate code suitable for big switch tables. Use this option only if
9660 the assembler/linker complain about out of range branches within a switch
9663 @item -mjump-in-delay
9664 @opindex mjump-in-delay
9665 Fill delay slots of function calls with unconditional jump instructions
9666 by modifying the return pointer for the function call to be the target
9667 of the conditional jump.
9669 @item -mdisable-fpregs
9670 @opindex mdisable-fpregs
9671 Prevent floating point registers from being used in any manner. This is
9672 necessary for compiling kernels which perform lazy context switching of
9673 floating point registers. If you use this option and attempt to perform
9674 floating point operations, the compiler will abort.
9676 @item -mdisable-indexing
9677 @opindex mdisable-indexing
9678 Prevent the compiler from using indexing address modes. This avoids some
9679 rather obscure problems when compiling MIG generated code under MACH@.
9681 @item -mno-space-regs
9682 @opindex mno-space-regs
9683 Generate code that assumes the target has no space registers. This allows
9684 GCC to generate faster indirect calls and use unscaled index address modes.
9686 Such code is suitable for level 0 PA systems and kernels.
9688 @item -mfast-indirect-calls
9689 @opindex mfast-indirect-calls
9690 Generate code that assumes calls never cross space boundaries. This
9691 allows GCC to emit code which performs faster indirect calls.
9693 This option will not work in the presence of shared libraries or nested
9696 @item -mfixed-range=@var{register-range}
9697 @opindex mfixed-range
9698 Generate code treating the given register range as fixed registers.
9699 A fixed register is one that the register allocator can not use. This is
9700 useful when compiling kernel code. A register range is specified as
9701 two registers separated by a dash. Multiple register ranges can be
9702 specified separated by a comma.
9704 @item -mlong-load-store
9705 @opindex mlong-load-store
9706 Generate 3-instruction load and store sequences as sometimes required by
9707 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
9710 @item -mportable-runtime
9711 @opindex mportable-runtime
9712 Use the portable calling conventions proposed by HP for ELF systems.
9716 Enable the use of assembler directives only GAS understands.
9718 @item -mschedule=@var{cpu-type}
9720 Schedule code according to the constraints for the machine type
9721 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
9722 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
9723 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
9724 proper scheduling option for your machine. The default scheduling is
9728 @opindex mlinker-opt
9729 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
9730 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
9731 linkers in which they give bogus error messages when linking some programs.
9734 @opindex msoft-float
9735 Generate output containing library calls for floating point.
9736 @strong{Warning:} the requisite libraries are not available for all HPPA
9737 targets. Normally the facilities of the machine's usual C compiler are
9738 used, but this cannot be done directly in cross-compilation. You must make
9739 your own arrangements to provide suitable library functions for
9740 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
9741 does provide software floating point support.
9743 @option{-msoft-float} changes the calling convention in the output file;
9744 therefore, it is only useful if you compile @emph{all} of a program with
9745 this option. In particular, you need to compile @file{libgcc.a}, the
9746 library that comes with GCC, with @option{-msoft-float} in order for
9751 Generate the predefine, @code{_SIO}, for server IO@. The default is
9752 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
9753 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
9754 options are available under HP-UX and HI-UX@.
9758 Use GNU ld specific options. This passes @option{-shared} to ld when
9759 building a shared library. It is the default when GCC is configured,
9760 explicitly or implicitly, with the GNU linker. This option does not
9761 have any affect on which ld is called, it only changes what parameters
9762 are passed to that ld. The ld that is called is determined by the
9763 @option{--with-ld} configure option, GCC's program search path, and
9764 finally by the user's @env{PATH}. The linker used by GCC can be printed
9765 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
9766 on the 64 bit HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9770 Use HP ld specific options. This passes @option{-b} to ld when building
9771 a shared library and passes @option{+Accept TypeMismatch} to ld on all
9772 links. It is the default when GCC is configured, explicitly or
9773 implicitly, with the HP linker. This option does not have any affect on
9774 which ld is called, it only changes what parameters are passed to that
9775 ld. The ld that is called is determined by the @option{--with-ld}
9776 configure option, GCC's program search path, and finally by the user's
9777 @env{PATH}. The linker used by GCC can be printed using @samp{which
9778 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
9779 HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9782 @opindex mno-long-calls
9783 Generate code that uses long call sequences. This ensures that a call
9784 is always able to reach linker generated stubs. The default is to generate
9785 long calls only when the distance from the call site to the beginning
9786 of the function or translation unit, as the case may be, exceeds a
9787 predefined limit set by the branch type being used. The limits for
9788 normal calls are 7,600,000 and 240,000 bytes, respectively for the
9789 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
9792 Distances are measured from the beginning of functions when using the
9793 @option{-ffunction-sections} option, or when using the @option{-mgas}
9794 and @option{-mno-portable-runtime} options together under HP-UX with
9797 It is normally not desirable to use this option as it will degrade
9798 performance. However, it may be useful in large applications,
9799 particularly when partial linking is used to build the application.
9801 The types of long calls used depends on the capabilities of the
9802 assembler and linker, and the type of code being generated. The
9803 impact on systems that support long absolute calls, and long pic
9804 symbol-difference or pc-relative calls should be relatively small.
9805 However, an indirect call is used on 32-bit ELF systems in pic code
9806 and it is quite long.
9808 @item -munix=@var{unix-std}
9810 Generate compiler predefines and select a startfile for the specified
9811 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
9812 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
9813 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
9814 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
9815 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
9818 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
9819 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
9820 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
9821 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
9822 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
9823 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
9825 It is @emph{important} to note that this option changes the interfaces
9826 for various library routines. It also affects the operational behavior
9827 of the C library. Thus, @emph{extreme} care is needed in using this
9830 Library code that is intended to operate with more than one UNIX
9831 standard must test, set and restore the variable @var{__xpg4_extended_mask}
9832 as appropriate. Most GNU software doesn't provide this capability.
9836 Suppress the generation of link options to search libdld.sl when the
9837 @option{-static} option is specified on HP-UX 10 and later.
9841 The HP-UX implementation of setlocale in libc has a dependency on
9842 libdld.sl. There isn't an archive version of libdld.sl. Thus,
9843 when the @option{-static} option is specified, special link options
9844 are needed to resolve this dependency.
9846 On HP-UX 10 and later, the GCC driver adds the necessary options to
9847 link with libdld.sl when the @option{-static} option is specified.
9848 This causes the resulting binary to be dynamic. On the 64-bit port,
9849 the linkers generate dynamic binaries by default in any case. The
9850 @option{-nolibdld} option can be used to prevent the GCC driver from
9851 adding these link options.
9855 Add support for multithreading with the @dfn{dce thread} library
9856 under HP-UX@. This option sets flags for both the preprocessor and
9860 @node i386 and x86-64 Options
9861 @subsection Intel 386 and AMD x86-64 Options
9862 @cindex i386 Options
9863 @cindex x86-64 Options
9864 @cindex Intel 386 Options
9865 @cindex AMD x86-64 Options
9867 These @samp{-m} options are defined for the i386 and x86-64 family of
9871 @item -mtune=@var{cpu-type}
9873 Tune to @var{cpu-type} everything applicable about the generated code, except
9874 for the ABI and the set of available instructions. The choices for
9878 Produce code optimized for the most common IA32/AMD64/EM64T processors.
9879 If you know the CPU on which your code will run, then you should use
9880 the corresponding @option{-mtune} option instead of
9881 @option{-mtune=generic}. But, if you do not know exactly what CPU users
9882 of your application will have, then you should use this option.
9884 As new processors are deployed in the marketplace, the behavior of this
9885 option will change. Therefore, if you upgrade to a newer version of
9886 GCC, the code generated option will change to reflect the processors
9887 that were most common when that version of GCC was released.
9889 There is no @option{-march=generic} option because @option{-march}
9890 indicates the instruction set the compiler can use, and there is no
9891 generic instruction set applicable to all processors. In contrast,
9892 @option{-mtune} indicates the processor (or, in this case, collection of
9893 processors) for which the code is optimized.
9895 This selects the CPU to tune for at compilation time by determining
9896 the processor type of the compiling machine. Using @option{-mtune=native}
9897 will produce code optimized for the local machine under the constraints
9898 of the selected instruction set. Using @option{-march=native} will
9899 enable all instruction subsets supported by the local machine (hence
9900 the result might not run on different machines).
9902 Original Intel's i386 CPU@.
9904 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
9906 Intel Pentium CPU with no MMX support.
9908 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
9910 Intel PentiumPro CPU@.
9912 Same as @code{generic}, but when used as @code{march} option, PentiumPro
9913 instruction set will be used, so the code will run on all i686 family chips.
9915 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
9916 @item pentium3, pentium3m
9917 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
9920 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
9921 support. Used by Centrino notebooks.
9922 @item pentium4, pentium4m
9923 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
9925 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
9928 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
9929 SSE2 and SSE3 instruction set support.
9931 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
9932 instruction set support.
9934 AMD K6 CPU with MMX instruction set support.
9936 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
9937 @item athlon, athlon-tbird
9938 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
9940 @item athlon-4, athlon-xp, athlon-mp
9941 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
9942 instruction set support.
9943 @item k8, opteron, athlon64, athlon-fx
9944 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
9945 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
9946 @item k8-sse3, opteron-sse3, athlon64-sse3
9947 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
9948 @item amdfam10, barcelona
9949 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
9950 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3dNOW!, enhanced 3dNOW!, ABM and 64-bit
9951 instruction set extensions.)
9953 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
9956 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
9957 instruction set support.
9959 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
9960 implemented for this chip.)
9962 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
9963 implemented for this chip.)
9965 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
9968 While picking a specific @var{cpu-type} will schedule things appropriately
9969 for that particular chip, the compiler will not generate any code that
9970 does not run on the i386 without the @option{-march=@var{cpu-type}} option
9973 @item -march=@var{cpu-type}
9975 Generate instructions for the machine type @var{cpu-type}. The choices
9976 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
9977 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
9979 @item -mcpu=@var{cpu-type}
9981 A deprecated synonym for @option{-mtune}.
9983 @item -mfpmath=@var{unit}
9985 Generate floating point arithmetics for selected unit @var{unit}. The choices
9990 Use the standard 387 floating point coprocessor present majority of chips and
9991 emulated otherwise. Code compiled with this option will run almost everywhere.
9992 The temporary results are computed in 80bit precision instead of precision
9993 specified by the type resulting in slightly different results compared to most
9994 of other chips. See @option{-ffloat-store} for more detailed description.
9996 This is the default choice for i386 compiler.
9999 Use scalar floating point instructions present in the SSE instruction set.
10000 This instruction set is supported by Pentium3 and newer chips, in the AMD line
10001 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
10002 instruction set supports only single precision arithmetics, thus the double and
10003 extended precision arithmetics is still done using 387. Later version, present
10004 only in Pentium4 and the future AMD x86-64 chips supports double precision
10007 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
10008 or @option{-msse2} switches to enable SSE extensions and make this option
10009 effective. For the x86-64 compiler, these extensions are enabled by default.
10011 The resulting code should be considerably faster in the majority of cases and avoid
10012 the numerical instability problems of 387 code, but may break some existing
10013 code that expects temporaries to be 80bit.
10015 This is the default choice for the x86-64 compiler.
10018 Attempt to utilize both instruction sets at once. This effectively double the
10019 amount of available registers and on chips with separate execution units for
10020 387 and SSE the execution resources too. Use this option with care, as it is
10021 still experimental, because the GCC register allocator does not model separate
10022 functional units well resulting in instable performance.
10025 @item -masm=@var{dialect}
10026 @opindex masm=@var{dialect}
10027 Output asm instructions using selected @var{dialect}. Supported
10028 choices are @samp{intel} or @samp{att} (the default one). Darwin does
10029 not support @samp{intel}.
10032 @itemx -mno-ieee-fp
10034 @opindex mno-ieee-fp
10035 Control whether or not the compiler uses IEEE floating point
10036 comparisons. These handle correctly the case where the result of a
10037 comparison is unordered.
10040 @opindex msoft-float
10041 Generate output containing library calls for floating point.
10042 @strong{Warning:} the requisite libraries are not part of GCC@.
10043 Normally the facilities of the machine's usual C compiler are used, but
10044 this can't be done directly in cross-compilation. You must make your
10045 own arrangements to provide suitable library functions for
10048 On machines where a function returns floating point results in the 80387
10049 register stack, some floating point opcodes may be emitted even if
10050 @option{-msoft-float} is used.
10052 @item -mno-fp-ret-in-387
10053 @opindex mno-fp-ret-in-387
10054 Do not use the FPU registers for return values of functions.
10056 The usual calling convention has functions return values of types
10057 @code{float} and @code{double} in an FPU register, even if there
10058 is no FPU@. The idea is that the operating system should emulate
10061 The option @option{-mno-fp-ret-in-387} causes such values to be returned
10062 in ordinary CPU registers instead.
10064 @item -mno-fancy-math-387
10065 @opindex mno-fancy-math-387
10066 Some 387 emulators do not support the @code{sin}, @code{cos} and
10067 @code{sqrt} instructions for the 387. Specify this option to avoid
10068 generating those instructions. This option is the default on FreeBSD,
10069 OpenBSD and NetBSD@. This option is overridden when @option{-march}
10070 indicates that the target cpu will always have an FPU and so the
10071 instruction will not need emulation. As of revision 2.6.1, these
10072 instructions are not generated unless you also use the
10073 @option{-funsafe-math-optimizations} switch.
10075 @item -malign-double
10076 @itemx -mno-align-double
10077 @opindex malign-double
10078 @opindex mno-align-double
10079 Control whether GCC aligns @code{double}, @code{long double}, and
10080 @code{long long} variables on a two word boundary or a one word
10081 boundary. Aligning @code{double} variables on a two word boundary will
10082 produce code that runs somewhat faster on a @samp{Pentium} at the
10083 expense of more memory.
10085 On x86-64, @option{-malign-double} is enabled by default.
10087 @strong{Warning:} if you use the @option{-malign-double} switch,
10088 structures containing the above types will be aligned differently than
10089 the published application binary interface specifications for the 386
10090 and will not be binary compatible with structures in code compiled
10091 without that switch.
10093 @item -m96bit-long-double
10094 @itemx -m128bit-long-double
10095 @opindex m96bit-long-double
10096 @opindex m128bit-long-double
10097 These switches control the size of @code{long double} type. The i386
10098 application binary interface specifies the size to be 96 bits,
10099 so @option{-m96bit-long-double} is the default in 32 bit mode.
10101 Modern architectures (Pentium and newer) would prefer @code{long double}
10102 to be aligned to an 8 or 16 byte boundary. In arrays or structures
10103 conforming to the ABI, this would not be possible. So specifying a
10104 @option{-m128bit-long-double} will align @code{long double}
10105 to a 16 byte boundary by padding the @code{long double} with an additional
10108 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
10109 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
10111 Notice that neither of these options enable any extra precision over the x87
10112 standard of 80 bits for a @code{long double}.
10114 @strong{Warning:} if you override the default value for your target ABI, the
10115 structures and arrays containing @code{long double} variables will change
10116 their size as well as function calling convention for function taking
10117 @code{long double} will be modified. Hence they will not be binary
10118 compatible with arrays or structures in code compiled without that switch.
10120 @item -mmlarge-data-threshold=@var{number}
10121 @opindex mlarge-data-threshold=@var{number}
10122 When @option{-mcmodel=medium} is specified, the data greater than
10123 @var{threshold} are placed in large data section. This value must be the
10124 same across all object linked into the binary and defaults to 65535.
10128 Use a different function-calling convention, in which functions that
10129 take a fixed number of arguments return with the @code{ret} @var{num}
10130 instruction, which pops their arguments while returning. This saves one
10131 instruction in the caller since there is no need to pop the arguments
10134 You can specify that an individual function is called with this calling
10135 sequence with the function attribute @samp{stdcall}. You can also
10136 override the @option{-mrtd} option by using the function attribute
10137 @samp{cdecl}. @xref{Function Attributes}.
10139 @strong{Warning:} this calling convention is incompatible with the one
10140 normally used on Unix, so you cannot use it if you need to call
10141 libraries compiled with the Unix compiler.
10143 Also, you must provide function prototypes for all functions that
10144 take variable numbers of arguments (including @code{printf});
10145 otherwise incorrect code will be generated for calls to those
10148 In addition, seriously incorrect code will result if you call a
10149 function with too many arguments. (Normally, extra arguments are
10150 harmlessly ignored.)
10152 @item -mregparm=@var{num}
10154 Control how many registers are used to pass integer arguments. By
10155 default, no registers are used to pass arguments, and at most 3
10156 registers can be used. You can control this behavior for a specific
10157 function by using the function attribute @samp{regparm}.
10158 @xref{Function Attributes}.
10160 @strong{Warning:} if you use this switch, and
10161 @var{num} is nonzero, then you must build all modules with the same
10162 value, including any libraries. This includes the system libraries and
10166 @opindex msseregparm
10167 Use SSE register passing conventions for float and double arguments
10168 and return values. You can control this behavior for a specific
10169 function by using the function attribute @samp{sseregparm}.
10170 @xref{Function Attributes}.
10172 @strong{Warning:} if you use this switch then you must build all
10173 modules with the same value, including any libraries. This includes
10174 the system libraries and startup modules.
10183 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
10184 is specified, the significands of results of floating-point operations are
10185 rounded to 24 bits (single precision); @option{-mpc64} rounds the the
10186 significands of results of floating-point operations to 53 bits (double
10187 precision) and @option{-mpc80} rounds the significands of results of
10188 floating-point operations to 64 bits (extended double precision), which is
10189 the default. When this option is used, floating-point operations in higher
10190 precisions are not available to the programmer without setting the FPU
10191 control word explicitly.
10193 Setting the rounding of floating-point operations to less than the default
10194 80 bits can speed some programs by 2% or more. Note that some mathematical
10195 libraries assume that extended precision (80 bit) floating-point operations
10196 are enabled by default; routines in such libraries could suffer significant
10197 loss of accuracy, typically through so-called "catastrophic cancellation",
10198 when this option is used to set the precision to less than extended precision.
10200 @item -mstackrealign
10201 @opindex mstackrealign
10202 Realign the stack at entry. On the Intel x86, the
10203 @option{-mstackrealign} option will generate an alternate prologue and
10204 epilogue that realigns the runtime stack. This supports mixing legacy
10205 codes that keep a 4-byte aligned stack with modern codes that keep a
10206 16-byte stack for SSE compatibility. The alternate prologue and
10207 epilogue are slower and bigger than the regular ones, and the
10208 alternate prologue requires an extra scratch register; this lowers the
10209 number of registers available if used in conjunction with the
10210 @code{regparm} attribute. The @option{-mstackrealign} option is
10211 incompatible with the nested function prologue; this is considered a
10212 hard error. See also the attribute @code{force_align_arg_pointer},
10213 applicable to individual functions.
10215 @item -mpreferred-stack-boundary=@var{num}
10216 @opindex mpreferred-stack-boundary
10217 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
10218 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
10219 the default is 4 (16 bytes or 128 bits).
10221 On Pentium and PentiumPro, @code{double} and @code{long double} values
10222 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
10223 suffer significant run time performance penalties. On Pentium III, the
10224 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
10225 properly if it is not 16 byte aligned.
10227 To ensure proper alignment of this values on the stack, the stack boundary
10228 must be as aligned as that required by any value stored on the stack.
10229 Further, every function must be generated such that it keeps the stack
10230 aligned. Thus calling a function compiled with a higher preferred
10231 stack boundary from a function compiled with a lower preferred stack
10232 boundary will most likely misalign the stack. It is recommended that
10233 libraries that use callbacks always use the default setting.
10235 This extra alignment does consume extra stack space, and generally
10236 increases code size. Code that is sensitive to stack space usage, such
10237 as embedded systems and operating system kernels, may want to reduce the
10238 preferred alignment to @option{-mpreferred-stack-boundary=2}.
10264 These switches enable or disable the use of instructions in the MMX,
10265 SSE, SSE2, SSE3, SSSE3, SSE4A, ABM or 3DNow! extended instruction sets.
10266 These extensions are also available as built-in functions: see
10267 @ref{X86 Built-in Functions}, for details of the functions enabled and
10268 disabled by these switches.
10270 To have SSE/SSE2 instructions generated automatically from floating-point
10271 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
10273 These options will enable GCC to use these extended instructions in
10274 generated code, even without @option{-mfpmath=sse}. Applications which
10275 perform runtime CPU detection must compile separate files for each
10276 supported architecture, using the appropriate flags. In particular,
10277 the file containing the CPU detection code should be compiled without
10282 This option will enable GCC to use CMPXCHG16B instruction in generated code.
10283 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
10284 data types. This is useful for high resolution counters that could be updated
10285 by multiple processors (or cores). This instruction is generated as part of
10286 atomic built-in functions: see @ref{Atomic Builtins} for details.
10290 This option will enable GCC to use SAHF instruction in generated 64-bit code.
10291 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
10292 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
10293 SAHF are load and store instructions, respectively, for certain status flags.
10294 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
10295 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
10298 @itemx -mno-push-args
10299 @opindex mpush-args
10300 @opindex mno-push-args
10301 Use PUSH operations to store outgoing parameters. This method is shorter
10302 and usually equally fast as method using SUB/MOV operations and is enabled
10303 by default. In some cases disabling it may improve performance because of
10304 improved scheduling and reduced dependencies.
10306 @item -maccumulate-outgoing-args
10307 @opindex maccumulate-outgoing-args
10308 If enabled, the maximum amount of space required for outgoing arguments will be
10309 computed in the function prologue. This is faster on most modern CPUs
10310 because of reduced dependencies, improved scheduling and reduced stack usage
10311 when preferred stack boundary is not equal to 2. The drawback is a notable
10312 increase in code size. This switch implies @option{-mno-push-args}.
10316 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
10317 on thread-safe exception handling must compile and link all code with the
10318 @option{-mthreads} option. When compiling, @option{-mthreads} defines
10319 @option{-D_MT}; when linking, it links in a special thread helper library
10320 @option{-lmingwthrd} which cleans up per thread exception handling data.
10322 @item -mno-align-stringops
10323 @opindex mno-align-stringops
10324 Do not align destination of inlined string operations. This switch reduces
10325 code size and improves performance in case the destination is already aligned,
10326 but GCC doesn't know about it.
10328 @item -minline-all-stringops
10329 @opindex minline-all-stringops
10330 By default GCC inlines string operations only when destination is known to be
10331 aligned at least to 4 byte boundary. This enables more inlining, increase code
10332 size, but may improve performance of code that depends on fast memcpy, strlen
10333 and memset for short lengths.
10335 @item -minline-stringops-dynamically
10336 @opindex minline-stringops-dynamically
10337 For string operation of unknown size, inline runtime checks so for small
10338 blocks inline code is used, while for large blocks library call is used.
10340 @item -mstringop-strategy=@var{alg}
10341 @opindex mstringop-strategy=@var{alg}
10342 Overwrite internal decision heuristic about particular algorithm to inline
10343 string operation with. The allowed values are @code{rep_byte},
10344 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
10345 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
10346 expanding inline loop, @code{libcall} for always expanding library call.
10348 @item -momit-leaf-frame-pointer
10349 @opindex momit-leaf-frame-pointer
10350 Don't keep the frame pointer in a register for leaf functions. This
10351 avoids the instructions to save, set up and restore frame pointers and
10352 makes an extra register available in leaf functions. The option
10353 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10354 which might make debugging harder.
10356 @item -mtls-direct-seg-refs
10357 @itemx -mno-tls-direct-seg-refs
10358 @opindex mtls-direct-seg-refs
10359 Controls whether TLS variables may be accessed with offsets from the
10360 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
10361 or whether the thread base pointer must be added. Whether or not this
10362 is legal depends on the operating system, and whether it maps the
10363 segment to cover the entire TLS area.
10365 For systems that use GNU libc, the default is on.
10368 These @samp{-m} switches are supported in addition to the above
10369 on AMD x86-64 processors in 64-bit environments.
10376 Generate code for a 32-bit or 64-bit environment.
10377 The 32-bit environment sets int, long and pointer to 32 bits and
10378 generates code that runs on any i386 system.
10379 The 64-bit environment sets int to 32 bits and long and pointer
10380 to 64 bits and generates code for AMD's x86-64 architecture. For
10381 darwin only the -m64 option turns off the @option{-fno-pic} and
10382 @option{-mdynamic-no-pic} options.
10384 @item -mno-red-zone
10385 @opindex no-red-zone
10386 Do not use a so called red zone for x86-64 code. The red zone is mandated
10387 by the x86-64 ABI, it is a 128-byte area beyond the location of the
10388 stack pointer that will not be modified by signal or interrupt handlers
10389 and therefore can be used for temporary data without adjusting the stack
10390 pointer. The flag @option{-mno-red-zone} disables this red zone.
10392 @item -mcmodel=small
10393 @opindex mcmodel=small
10394 Generate code for the small code model: the program and its symbols must
10395 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
10396 Programs can be statically or dynamically linked. This is the default
10399 @item -mcmodel=kernel
10400 @opindex mcmodel=kernel
10401 Generate code for the kernel code model. The kernel runs in the
10402 negative 2 GB of the address space.
10403 This model has to be used for Linux kernel code.
10405 @item -mcmodel=medium
10406 @opindex mcmodel=medium
10407 Generate code for the medium model: The program is linked in the lower 2
10408 GB of the address space but symbols can be located anywhere in the
10409 address space. Programs can be statically or dynamically linked, but
10410 building of shared libraries are not supported with the medium model.
10412 @item -mcmodel=large
10413 @opindex mcmodel=large
10414 Generate code for the large model: This model makes no assumptions
10415 about addresses and sizes of sections.
10418 @node IA-64 Options
10419 @subsection IA-64 Options
10420 @cindex IA-64 Options
10422 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10426 @opindex mbig-endian
10427 Generate code for a big endian target. This is the default for HP-UX@.
10429 @item -mlittle-endian
10430 @opindex mlittle-endian
10431 Generate code for a little endian target. This is the default for AIX5
10437 @opindex mno-gnu-as
10438 Generate (or don't) code for the GNU assembler. This is the default.
10439 @c Also, this is the default if the configure option @option{--with-gnu-as}
10445 @opindex mno-gnu-ld
10446 Generate (or don't) code for the GNU linker. This is the default.
10447 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10452 Generate code that does not use a global pointer register. The result
10453 is not position independent code, and violates the IA-64 ABI@.
10455 @item -mvolatile-asm-stop
10456 @itemx -mno-volatile-asm-stop
10457 @opindex mvolatile-asm-stop
10458 @opindex mno-volatile-asm-stop
10459 Generate (or don't) a stop bit immediately before and after volatile asm
10462 @item -mregister-names
10463 @itemx -mno-register-names
10464 @opindex mregister-names
10465 @opindex mno-register-names
10466 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
10467 the stacked registers. This may make assembler output more readable.
10473 Disable (or enable) optimizations that use the small data section. This may
10474 be useful for working around optimizer bugs.
10476 @item -mconstant-gp
10477 @opindex mconstant-gp
10478 Generate code that uses a single constant global pointer value. This is
10479 useful when compiling kernel code.
10483 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
10484 This is useful when compiling firmware code.
10486 @item -minline-float-divide-min-latency
10487 @opindex minline-float-divide-min-latency
10488 Generate code for inline divides of floating point values
10489 using the minimum latency algorithm.
10491 @item -minline-float-divide-max-throughput
10492 @opindex minline-float-divide-max-throughput
10493 Generate code for inline divides of floating point values
10494 using the maximum throughput algorithm.
10496 @item -minline-int-divide-min-latency
10497 @opindex minline-int-divide-min-latency
10498 Generate code for inline divides of integer values
10499 using the minimum latency algorithm.
10501 @item -minline-int-divide-max-throughput
10502 @opindex minline-int-divide-max-throughput
10503 Generate code for inline divides of integer values
10504 using the maximum throughput algorithm.
10506 @item -minline-sqrt-min-latency
10507 @opindex minline-sqrt-min-latency
10508 Generate code for inline square roots
10509 using the minimum latency algorithm.
10511 @item -minline-sqrt-max-throughput
10512 @opindex minline-sqrt-max-throughput
10513 Generate code for inline square roots
10514 using the maximum throughput algorithm.
10516 @item -mno-dwarf2-asm
10517 @itemx -mdwarf2-asm
10518 @opindex mno-dwarf2-asm
10519 @opindex mdwarf2-asm
10520 Don't (or do) generate assembler code for the DWARF2 line number debugging
10521 info. This may be useful when not using the GNU assembler.
10523 @item -mearly-stop-bits
10524 @itemx -mno-early-stop-bits
10525 @opindex mearly-stop-bits
10526 @opindex mno-early-stop-bits
10527 Allow stop bits to be placed earlier than immediately preceding the
10528 instruction that triggered the stop bit. This can improve instruction
10529 scheduling, but does not always do so.
10531 @item -mfixed-range=@var{register-range}
10532 @opindex mfixed-range
10533 Generate code treating the given register range as fixed registers.
10534 A fixed register is one that the register allocator can not use. This is
10535 useful when compiling kernel code. A register range is specified as
10536 two registers separated by a dash. Multiple register ranges can be
10537 specified separated by a comma.
10539 @item -mtls-size=@var{tls-size}
10541 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
10544 @item -mtune=@var{cpu-type}
10546 Tune the instruction scheduling for a particular CPU, Valid values are
10547 itanium, itanium1, merced, itanium2, and mckinley.
10553 Add support for multithreading using the POSIX threads library. This
10554 option sets flags for both the preprocessor and linker. It does
10555 not affect the thread safety of object code produced by the compiler or
10556 that of libraries supplied with it. These are HP-UX specific flags.
10562 Generate code for a 32-bit or 64-bit environment.
10563 The 32-bit environment sets int, long and pointer to 32 bits.
10564 The 64-bit environment sets int to 32 bits and long and pointer
10565 to 64 bits. These are HP-UX specific flags.
10567 @item -mno-sched-br-data-spec
10568 @itemx -msched-br-data-spec
10569 @opindex mno-sched-br-data-spec
10570 @opindex msched-br-data-spec
10571 (Dis/En)able data speculative scheduling before reload.
10572 This will result in generation of the ld.a instructions and
10573 the corresponding check instructions (ld.c / chk.a).
10574 The default is 'disable'.
10576 @item -msched-ar-data-spec
10577 @itemx -mno-sched-ar-data-spec
10578 @opindex msched-ar-data-spec
10579 @opindex mno-sched-ar-data-spec
10580 (En/Dis)able data speculative scheduling after reload.
10581 This will result in generation of the ld.a instructions and
10582 the corresponding check instructions (ld.c / chk.a).
10583 The default is 'enable'.
10585 @item -mno-sched-control-spec
10586 @itemx -msched-control-spec
10587 @opindex mno-sched-control-spec
10588 @opindex msched-control-spec
10589 (Dis/En)able control speculative scheduling. This feature is
10590 available only during region scheduling (i.e. before reload).
10591 This will result in generation of the ld.s instructions and
10592 the corresponding check instructions chk.s .
10593 The default is 'disable'.
10595 @item -msched-br-in-data-spec
10596 @itemx -mno-sched-br-in-data-spec
10597 @opindex msched-br-in-data-spec
10598 @opindex mno-sched-br-in-data-spec
10599 (En/Dis)able speculative scheduling of the instructions that
10600 are dependent on the data speculative loads before reload.
10601 This is effective only with @option{-msched-br-data-spec} enabled.
10602 The default is 'enable'.
10604 @item -msched-ar-in-data-spec
10605 @itemx -mno-sched-ar-in-data-spec
10606 @opindex msched-ar-in-data-spec
10607 @opindex mno-sched-ar-in-data-spec
10608 (En/Dis)able speculative scheduling of the instructions that
10609 are dependent on the data speculative loads after reload.
10610 This is effective only with @option{-msched-ar-data-spec} enabled.
10611 The default is 'enable'.
10613 @item -msched-in-control-spec
10614 @itemx -mno-sched-in-control-spec
10615 @opindex msched-in-control-spec
10616 @opindex mno-sched-in-control-spec
10617 (En/Dis)able speculative scheduling of the instructions that
10618 are dependent on the control speculative loads.
10619 This is effective only with @option{-msched-control-spec} enabled.
10620 The default is 'enable'.
10623 @itemx -mno-sched-ldc
10624 @opindex msched-ldc
10625 @opindex mno-sched-ldc
10626 (En/Dis)able use of simple data speculation checks ld.c .
10627 If disabled, only chk.a instructions will be emitted to check
10628 data speculative loads.
10629 The default is 'enable'.
10631 @item -mno-sched-control-ldc
10632 @itemx -msched-control-ldc
10633 @opindex mno-sched-control-ldc
10634 @opindex msched-control-ldc
10635 (Dis/En)able use of ld.c instructions to check control speculative loads.
10636 If enabled, in case of control speculative load with no speculatively
10637 scheduled dependent instructions this load will be emitted as ld.sa and
10638 ld.c will be used to check it.
10639 The default is 'disable'.
10641 @item -mno-sched-spec-verbose
10642 @itemx -msched-spec-verbose
10643 @opindex mno-sched-spec-verbose
10644 @opindex msched-spec-verbose
10645 (Dis/En)able printing of the information about speculative motions.
10647 @item -mno-sched-prefer-non-data-spec-insns
10648 @itemx -msched-prefer-non-data-spec-insns
10649 @opindex mno-sched-prefer-non-data-spec-insns
10650 @opindex msched-prefer-non-data-spec-insns
10651 If enabled, data speculative instructions will be chosen for schedule
10652 only if there are no other choices at the moment. This will make
10653 the use of the data speculation much more conservative.
10654 The default is 'disable'.
10656 @item -mno-sched-prefer-non-control-spec-insns
10657 @itemx -msched-prefer-non-control-spec-insns
10658 @opindex mno-sched-prefer-non-control-spec-insns
10659 @opindex msched-prefer-non-control-spec-insns
10660 If enabled, control speculative instructions will be chosen for schedule
10661 only if there are no other choices at the moment. This will make
10662 the use of the control speculation much more conservative.
10663 The default is 'disable'.
10665 @item -mno-sched-count-spec-in-critical-path
10666 @itemx -msched-count-spec-in-critical-path
10667 @opindex mno-sched-count-spec-in-critical-path
10668 @opindex msched-count-spec-in-critical-path
10669 If enabled, speculative dependencies will be considered during
10670 computation of the instructions priorities. This will make the use of the
10671 speculation a bit more conservative.
10672 The default is 'disable'.
10677 @subsection M32C Options
10678 @cindex M32C options
10681 @item -mcpu=@var{name}
10683 Select the CPU for which code is generated. @var{name} may be one of
10684 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
10685 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
10686 the M32C/80 series.
10690 Specifies that the program will be run on the simulator. This causes
10691 an alternate runtime library to be linked in which supports, for
10692 example, file I/O. You must not use this option when generating
10693 programs that will run on real hardware; you must provide your own
10694 runtime library for whatever I/O functions are needed.
10696 @item -memregs=@var{number}
10698 Specifies the number of memory-based pseudo-registers GCC will use
10699 during code generation. These pseudo-registers will be used like real
10700 registers, so there is a tradeoff between GCC's ability to fit the
10701 code into available registers, and the performance penalty of using
10702 memory instead of registers. Note that all modules in a program must
10703 be compiled with the same value for this option. Because of that, you
10704 must not use this option with the default runtime libraries gcc
10709 @node M32R/D Options
10710 @subsection M32R/D Options
10711 @cindex M32R/D options
10713 These @option{-m} options are defined for Renesas M32R/D architectures:
10718 Generate code for the M32R/2@.
10722 Generate code for the M32R/X@.
10726 Generate code for the M32R@. This is the default.
10728 @item -mmodel=small
10729 @opindex mmodel=small
10730 Assume all objects live in the lower 16MB of memory (so that their addresses
10731 can be loaded with the @code{ld24} instruction), and assume all subroutines
10732 are reachable with the @code{bl} instruction.
10733 This is the default.
10735 The addressability of a particular object can be set with the
10736 @code{model} attribute.
10738 @item -mmodel=medium
10739 @opindex mmodel=medium
10740 Assume objects may be anywhere in the 32-bit address space (the compiler
10741 will generate @code{seth/add3} instructions to load their addresses), and
10742 assume all subroutines are reachable with the @code{bl} instruction.
10744 @item -mmodel=large
10745 @opindex mmodel=large
10746 Assume objects may be anywhere in the 32-bit address space (the compiler
10747 will generate @code{seth/add3} instructions to load their addresses), and
10748 assume subroutines may not be reachable with the @code{bl} instruction
10749 (the compiler will generate the much slower @code{seth/add3/jl}
10750 instruction sequence).
10753 @opindex msdata=none
10754 Disable use of the small data area. Variables will be put into
10755 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
10756 @code{section} attribute has been specified).
10757 This is the default.
10759 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
10760 Objects may be explicitly put in the small data area with the
10761 @code{section} attribute using one of these sections.
10763 @item -msdata=sdata
10764 @opindex msdata=sdata
10765 Put small global and static data in the small data area, but do not
10766 generate special code to reference them.
10769 @opindex msdata=use
10770 Put small global and static data in the small data area, and generate
10771 special instructions to reference them.
10775 @cindex smaller data references
10776 Put global and static objects less than or equal to @var{num} bytes
10777 into the small data or bss sections instead of the normal data or bss
10778 sections. The default value of @var{num} is 8.
10779 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
10780 for this option to have any effect.
10782 All modules should be compiled with the same @option{-G @var{num}} value.
10783 Compiling with different values of @var{num} may or may not work; if it
10784 doesn't the linker will give an error message---incorrect code will not be
10789 Makes the M32R specific code in the compiler display some statistics
10790 that might help in debugging programs.
10792 @item -malign-loops
10793 @opindex malign-loops
10794 Align all loops to a 32-byte boundary.
10796 @item -mno-align-loops
10797 @opindex mno-align-loops
10798 Do not enforce a 32-byte alignment for loops. This is the default.
10800 @item -missue-rate=@var{number}
10801 @opindex missue-rate=@var{number}
10802 Issue @var{number} instructions per cycle. @var{number} can only be 1
10805 @item -mbranch-cost=@var{number}
10806 @opindex mbranch-cost=@var{number}
10807 @var{number} can only be 1 or 2. If it is 1 then branches will be
10808 preferred over conditional code, if it is 2, then the opposite will
10811 @item -mflush-trap=@var{number}
10812 @opindex mflush-trap=@var{number}
10813 Specifies the trap number to use to flush the cache. The default is
10814 12. Valid numbers are between 0 and 15 inclusive.
10816 @item -mno-flush-trap
10817 @opindex mno-flush-trap
10818 Specifies that the cache cannot be flushed by using a trap.
10820 @item -mflush-func=@var{name}
10821 @opindex mflush-func=@var{name}
10822 Specifies the name of the operating system function to call to flush
10823 the cache. The default is @emph{_flush_cache}, but a function call
10824 will only be used if a trap is not available.
10826 @item -mno-flush-func
10827 @opindex mno-flush-func
10828 Indicates that there is no OS function for flushing the cache.
10832 @node M680x0 Options
10833 @subsection M680x0 Options
10834 @cindex M680x0 options
10836 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
10837 The default settings depend on which architecture was selected when
10838 the compiler was configured; the defaults for the most common choices
10842 @item -march=@var{arch}
10844 Generate code for a specific M680x0 or ColdFire instruction set
10845 architecture. Permissible values of @var{arch} for M680x0
10846 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
10847 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
10848 architectures are selected according to Freescale's ISA classification
10849 and the permissible values are: @samp{isaa}, @samp{isaaplus},
10850 @samp{isab} and @samp{isac}.
10852 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
10853 code for a ColdFire target. The @var{arch} in this macro is one of the
10854 @option{-march} arguments given above.
10856 When used together, @option{-march} and @option{-mtune} select code
10857 that runs on a family of similar processors but that is optimized
10858 for a particular microarchitecture.
10860 @item -mcpu=@var{cpu}
10862 Generate code for a specific M680x0 or ColdFire processor.
10863 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
10864 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
10865 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
10866 below, which also classifies the CPUs into families:
10868 @multitable @columnfractions 0.20 0.80
10869 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
10870 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
10871 @item @samp{5206e} @tab @samp{5206e}
10872 @item @samp{5208} @tab @samp{5207} @samp{5208}
10873 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
10874 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
10875 @item @samp{5216} @tab @samp{5214} @samp{5216}
10876 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
10877 @item @samp{5225} @tab @samp{5224} @samp{5225}
10878 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
10879 @item @samp{5249} @tab @samp{5249}
10880 @item @samp{5250} @tab @samp{5250}
10881 @item @samp{5271} @tab @samp{5270} @samp{5271}
10882 @item @samp{5272} @tab @samp{5272}
10883 @item @samp{5275} @tab @samp{5274} @samp{5275}
10884 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
10885 @item @samp{5307} @tab @samp{5307}
10886 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
10887 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
10888 @item @samp{5407} @tab @samp{5407}
10889 @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}
10892 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
10893 @var{arch} is compatible with @var{cpu}. Other combinations of
10894 @option{-mcpu} and @option{-march} are rejected.
10896 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
10897 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
10898 where the value of @var{family} is given by the table above.
10900 @item -mtune=@var{tune}
10902 Tune the code for a particular microarchitecture, within the
10903 constraints set by @option{-march} and @option{-mcpu}.
10904 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
10905 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
10906 and @samp{cpu32}. The ColdFire microarchitectures
10907 are: @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
10909 You can also use @option{-mtune=68020-40} for code that needs
10910 to run relatively well on 68020, 68030 and 68040 targets.
10911 @option{-mtune=68020-60} is similar but includes 68060 targets
10912 as well. These two options select the same tuning decisions as
10913 @option{-m68020-40} and @option{-m68020-60} respectively.
10915 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
10916 when tuning for 680x0 architecture @var{arch}. It also defines
10917 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
10918 option is used. If gcc is tuning for a range of architectures,
10919 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
10920 it defines the macros for every architecture in the range.
10922 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
10923 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
10924 of the arguments given above.
10930 Generate output for a 68000. This is the default
10931 when the compiler is configured for 68000-based systems.
10932 It is equivalent to @option{-march=68000}.
10934 Use this option for microcontrollers with a 68000 or EC000 core,
10935 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
10939 Generate output for a 68010. This is the default
10940 when the compiler is configured for 68010-based systems.
10941 It is equivalent to @option{-march=68010}.
10947 Generate output for a 68020. This is the default
10948 when the compiler is configured for 68020-based systems.
10949 It is equivalent to @option{-march=68020}.
10953 Generate output for a 68030. This is the default when the compiler is
10954 configured for 68030-based systems. It is equivalent to
10955 @option{-march=68030}.
10959 Generate output for a 68040. This is the default when the compiler is
10960 configured for 68040-based systems. It is equivalent to
10961 @option{-march=68040}.
10963 This option inhibits the use of 68881/68882 instructions that have to be
10964 emulated by software on the 68040. Use this option if your 68040 does not
10965 have code to emulate those instructions.
10969 Generate output for a 68060. This is the default when the compiler is
10970 configured for 68060-based systems. It is equivalent to
10971 @option{-march=68060}.
10973 This option inhibits the use of 68020 and 68881/68882 instructions that
10974 have to be emulated by software on the 68060. Use this option if your 68060
10975 does not have code to emulate those instructions.
10979 Generate output for a CPU32. This is the default
10980 when the compiler is configured for CPU32-based systems.
10981 It is equivalent to @option{-march=cpu32}.
10983 Use this option for microcontrollers with a
10984 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
10985 68336, 68340, 68341, 68349 and 68360.
10989 Generate output for a 520X ColdFire CPU. This is the default
10990 when the compiler is configured for 520X-based systems.
10991 It is equivalent to @option{-mcpu=5206}, and is now deprecated
10992 in favor of that option.
10994 Use this option for microcontroller with a 5200 core, including
10995 the MCF5202, MCF5203, MCF5204 and MCF5206.
10999 Generate output for a 5206e ColdFire CPU. The option is now
11000 deprecated in favor of the equivalent @option{-mcpu=5206e}.
11004 Generate output for a member of the ColdFire 528X family.
11005 The option is now deprecated in favor of the equivalent
11006 @option{-mcpu=528x}.
11010 Generate output for a ColdFire 5307 CPU. The option is now deprecated
11011 in favor of the equivalent @option{-mcpu=5307}.
11015 Generate output for a ColdFire 5407 CPU. The option is now deprecated
11016 in favor of the equivalent @option{-mcpu=5407}.
11020 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
11021 This includes use of hardware floating point instructions.
11022 The option is equivalent to @option{-mcpu=547x}, and is now
11023 deprecated in favor of that option.
11027 Generate output for a 68040, without using any of the new instructions.
11028 This results in code which can run relatively efficiently on either a
11029 68020/68881 or a 68030 or a 68040. The generated code does use the
11030 68881 instructions that are emulated on the 68040.
11032 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
11036 Generate output for a 68060, without using any of the new instructions.
11037 This results in code which can run relatively efficiently on either a
11038 68020/68881 or a 68030 or a 68040. The generated code does use the
11039 68881 instructions that are emulated on the 68060.
11041 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
11045 @opindex mhard-float
11047 Generate floating-point instructions. This is the default for 68020
11048 and above, and for ColdFire devices that have an FPU. It defines the
11049 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
11050 on ColdFire targets.
11053 @opindex msoft-float
11054 Do not generate floating-point instructions; use library calls instead.
11055 This is the default for 68000, 68010, and 68832 targets. It is also
11056 the default for ColdFire devices that have no FPU.
11062 Generate (do not generate) ColdFire hardware divide and remainder
11063 instructions. If @option{-march} is used without @option{-mcpu},
11064 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
11065 architectures. Otherwise, the default is taken from the target CPU
11066 (either the default CPU, or the one specified by @option{-mcpu}). For
11067 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
11068 @option{-mcpu=5206e}.
11070 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
11074 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11075 Additionally, parameters passed on the stack are also aligned to a
11076 16-bit boundary even on targets whose API mandates promotion to 32-bit.
11080 Do not consider type @code{int} to be 16 bits wide. This is the default.
11083 @itemx -mno-bitfield
11084 @opindex mnobitfield
11085 @opindex mno-bitfield
11086 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
11087 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
11091 Do use the bit-field instructions. The @option{-m68020} option implies
11092 @option{-mbitfield}. This is the default if you use a configuration
11093 designed for a 68020.
11097 Use a different function-calling convention, in which functions
11098 that take a fixed number of arguments return with the @code{rtd}
11099 instruction, which pops their arguments while returning. This
11100 saves one instruction in the caller since there is no need to pop
11101 the arguments there.
11103 This calling convention is incompatible with the one normally
11104 used on Unix, so you cannot use it if you need to call libraries
11105 compiled with the Unix compiler.
11107 Also, you must provide function prototypes for all functions that
11108 take variable numbers of arguments (including @code{printf});
11109 otherwise incorrect code will be generated for calls to those
11112 In addition, seriously incorrect code will result if you call a
11113 function with too many arguments. (Normally, extra arguments are
11114 harmlessly ignored.)
11116 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
11117 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
11121 Do not use the calling conventions selected by @option{-mrtd}.
11122 This is the default.
11125 @itemx -mno-align-int
11126 @opindex malign-int
11127 @opindex mno-align-int
11128 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
11129 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
11130 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
11131 Aligning variables on 32-bit boundaries produces code that runs somewhat
11132 faster on processors with 32-bit busses at the expense of more memory.
11134 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
11135 align structures containing the above types differently than
11136 most published application binary interface specifications for the m68k.
11140 Use the pc-relative addressing mode of the 68000 directly, instead of
11141 using a global offset table. At present, this option implies @option{-fpic},
11142 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
11143 not presently supported with @option{-mpcrel}, though this could be supported for
11144 68020 and higher processors.
11146 @item -mno-strict-align
11147 @itemx -mstrict-align
11148 @opindex mno-strict-align
11149 @opindex mstrict-align
11150 Do not (do) assume that unaligned memory references will be handled by
11154 Generate code that allows the data segment to be located in a different
11155 area of memory from the text segment. This allows for execute in place in
11156 an environment without virtual memory management. This option implies
11159 @item -mno-sep-data
11160 Generate code that assumes that the data segment follows the text segment.
11161 This is the default.
11163 @item -mid-shared-library
11164 Generate code that supports shared libraries via the library ID method.
11165 This allows for execute in place and shared libraries in an environment
11166 without virtual memory management. This option implies @option{-fPIC}.
11168 @item -mno-id-shared-library
11169 Generate code that doesn't assume ID based shared libraries are being used.
11170 This is the default.
11172 @item -mshared-library-id=n
11173 Specified the identification number of the ID based shared library being
11174 compiled. Specifying a value of 0 will generate more compact code, specifying
11175 other values will force the allocation of that number to the current
11176 library but is no more space or time efficient than omitting this option.
11180 @node M68hc1x Options
11181 @subsection M68hc1x Options
11182 @cindex M68hc1x options
11184 These are the @samp{-m} options defined for the 68hc11 and 68hc12
11185 microcontrollers. The default values for these options depends on
11186 which style of microcontroller was selected when the compiler was configured;
11187 the defaults for the most common choices are given below.
11194 Generate output for a 68HC11. This is the default
11195 when the compiler is configured for 68HC11-based systems.
11201 Generate output for a 68HC12. This is the default
11202 when the compiler is configured for 68HC12-based systems.
11208 Generate output for a 68HCS12.
11210 @item -mauto-incdec
11211 @opindex mauto-incdec
11212 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
11219 Enable the use of 68HC12 min and max instructions.
11222 @itemx -mno-long-calls
11223 @opindex mlong-calls
11224 @opindex mno-long-calls
11225 Treat all calls as being far away (near). If calls are assumed to be
11226 far away, the compiler will use the @code{call} instruction to
11227 call a function and the @code{rtc} instruction for returning.
11231 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11233 @item -msoft-reg-count=@var{count}
11234 @opindex msoft-reg-count
11235 Specify the number of pseudo-soft registers which are used for the
11236 code generation. The maximum number is 32. Using more pseudo-soft
11237 register may or may not result in better code depending on the program.
11238 The default is 4 for 68HC11 and 2 for 68HC12.
11242 @node MCore Options
11243 @subsection MCore Options
11244 @cindex MCore options
11246 These are the @samp{-m} options defined for the Motorola M*Core
11252 @itemx -mno-hardlit
11254 @opindex mno-hardlit
11255 Inline constants into the code stream if it can be done in two
11256 instructions or less.
11262 Use the divide instruction. (Enabled by default).
11264 @item -mrelax-immediate
11265 @itemx -mno-relax-immediate
11266 @opindex mrelax-immediate
11267 @opindex mno-relax-immediate
11268 Allow arbitrary sized immediates in bit operations.
11270 @item -mwide-bitfields
11271 @itemx -mno-wide-bitfields
11272 @opindex mwide-bitfields
11273 @opindex mno-wide-bitfields
11274 Always treat bit-fields as int-sized.
11276 @item -m4byte-functions
11277 @itemx -mno-4byte-functions
11278 @opindex m4byte-functions
11279 @opindex mno-4byte-functions
11280 Force all functions to be aligned to a four byte boundary.
11282 @item -mcallgraph-data
11283 @itemx -mno-callgraph-data
11284 @opindex mcallgraph-data
11285 @opindex mno-callgraph-data
11286 Emit callgraph information.
11289 @itemx -mno-slow-bytes
11290 @opindex mslow-bytes
11291 @opindex mno-slow-bytes
11292 Prefer word access when reading byte quantities.
11294 @item -mlittle-endian
11295 @itemx -mbig-endian
11296 @opindex mlittle-endian
11297 @opindex mbig-endian
11298 Generate code for a little endian target.
11304 Generate code for the 210 processor.
11308 @subsection MIPS Options
11309 @cindex MIPS options
11315 Generate big-endian code.
11319 Generate little-endian code. This is the default for @samp{mips*el-*-*}
11322 @item -march=@var{arch}
11324 Generate code that will run on @var{arch}, which can be the name of a
11325 generic MIPS ISA, or the name of a particular processor.
11327 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
11328 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
11329 The processor names are:
11330 @samp{4kc}, @samp{4km}, @samp{4kp},
11331 @samp{4kec}, @samp{4kem}, @samp{4kep},
11332 @samp{5kc}, @samp{5kf},
11334 @samp{24kc}, @samp{24kf}, @samp{24kx},
11335 @samp{24kec}, @samp{24kef}, @samp{24kex},
11336 @samp{34kc}, @samp{34kf}, @samp{34kx},
11337 @samp{74kc}, @samp{74kf}, @samp{74kx},
11340 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
11341 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
11342 @samp{rm7000}, @samp{rm9000},
11345 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
11346 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
11347 The special value @samp{from-abi} selects the
11348 most compatible architecture for the selected ABI (that is,
11349 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
11351 In processor names, a final @samp{000} can be abbreviated as @samp{k}
11352 (for example, @samp{-march=r2k}). Prefixes are optional, and
11353 @samp{vr} may be written @samp{r}.
11355 GCC defines two macros based on the value of this option. The first
11356 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
11357 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
11358 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
11359 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
11360 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
11362 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
11363 above. In other words, it will have the full prefix and will not
11364 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
11365 the macro names the resolved architecture (either @samp{"mips1"} or
11366 @samp{"mips3"}). It names the default architecture when no
11367 @option{-march} option is given.
11369 @item -mtune=@var{arch}
11371 Optimize for @var{arch}. Among other things, this option controls
11372 the way instructions are scheduled, and the perceived cost of arithmetic
11373 operations. The list of @var{arch} values is the same as for
11376 When this option is not used, GCC will optimize for the processor
11377 specified by @option{-march}. By using @option{-march} and
11378 @option{-mtune} together, it is possible to generate code that will
11379 run on a family of processors, but optimize the code for one
11380 particular member of that family.
11382 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
11383 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
11384 @samp{-march} ones described above.
11388 Equivalent to @samp{-march=mips1}.
11392 Equivalent to @samp{-march=mips2}.
11396 Equivalent to @samp{-march=mips3}.
11400 Equivalent to @samp{-march=mips4}.
11404 Equivalent to @samp{-march=mips32}.
11408 Equivalent to @samp{-march=mips32r2}.
11412 Equivalent to @samp{-march=mips64}.
11417 @opindex mno-mips16
11418 Generate (do not generate) MIPS16 code. If GCC is targetting a
11419 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
11431 Generate code for the given ABI@.
11433 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
11434 generates 64-bit code when you select a 64-bit architecture, but you
11435 can use @option{-mgp32} to get 32-bit code instead.
11437 For information about the O64 ABI, see
11438 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
11440 GCC supports a variant of the o32 ABI in which floating-point registers
11441 are 64 rather than 32 bits wide. You can select this combination with
11442 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
11443 and @samp{mfhc1} instructions and is therefore only supported for
11444 MIPS32R2 processors.
11446 The register assignments for arguments and return values remain the
11447 same, but each scalar value is passed in a single 64-bit register
11448 rather than a pair of 32-bit registers. For example, scalar
11449 floating-point values are returned in @samp{$f0} only, not a
11450 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
11451 remains the same, but all 64 bits are saved.
11454 @itemx -mno-abicalls
11456 @opindex mno-abicalls
11457 Generate (do not generate) code that is suitable for SVR4-style
11458 dynamic objects. @option{-mabicalls} is the default for SVR4-based
11463 Generate (do not generate) code that is fully position-independent,
11464 and that can therefore be linked into shared libraries. This option
11465 only affects @option{-mabicalls}.
11467 All @option{-mabicalls} code has traditionally been position-independent,
11468 regardless of options like @option{-fPIC} and @option{-fpic}. However,
11469 as an extension, the GNU toolchain allows executables to use absolute
11470 accesses for locally-binding symbols. It can also use shorter GP
11471 initialization sequences and generate direct calls to locally-defined
11472 functions. This mode is selected by @option{-mno-shared}.
11474 @option{-mno-shared} depends on binutils 2.16 or higher and generates
11475 objects that can only be linked by the GNU linker. However, the option
11476 does not affect the ABI of the final executable; it only affects the ABI
11477 of relocatable objects. Using @option{-mno-shared} will generally make
11478 executables both smaller and quicker.
11480 @option{-mshared} is the default.
11486 Lift (do not lift) the usual restrictions on the size of the global
11489 GCC normally uses a single instruction to load values from the GOT@.
11490 While this is relatively efficient, it will only work if the GOT
11491 is smaller than about 64k. Anything larger will cause the linker
11492 to report an error such as:
11494 @cindex relocation truncated to fit (MIPS)
11496 relocation truncated to fit: R_MIPS_GOT16 foobar
11499 If this happens, you should recompile your code with @option{-mxgot}.
11500 It should then work with very large GOTs, although it will also be
11501 less efficient, since it will take three instructions to fetch the
11502 value of a global symbol.
11504 Note that some linkers can create multiple GOTs. If you have such a
11505 linker, you should only need to use @option{-mxgot} when a single object
11506 file accesses more than 64k's worth of GOT entries. Very few do.
11508 These options have no effect unless GCC is generating position
11513 Assume that general-purpose registers are 32 bits wide.
11517 Assume that general-purpose registers are 64 bits wide.
11521 Assume that floating-point registers are 32 bits wide.
11525 Assume that floating-point registers are 64 bits wide.
11528 @opindex mhard-float
11529 Use floating-point coprocessor instructions.
11532 @opindex msoft-float
11533 Do not use floating-point coprocessor instructions. Implement
11534 floating-point calculations using library calls instead.
11536 @item -msingle-float
11537 @opindex msingle-float
11538 Assume that the floating-point coprocessor only supports single-precision
11541 @item -mdouble-float
11542 @opindex mdouble-float
11543 Assume that the floating-point coprocessor supports double-precision
11544 operations. This is the default.
11550 Use (do not use) the MIPS DSP ASE. @xref{MIPS DSP Built-in Functions}.
11556 Use (do not use) the MIPS DSP ASE REV 2. @xref{MIPS DSP Built-in Functions}.
11557 The option @option{-mdspr2} implies @option{-mdsp}.
11559 @item -mpaired-single
11560 @itemx -mno-paired-single
11561 @opindex mpaired-single
11562 @opindex mno-paired-single
11563 Use (do not use) paired-single floating-point instructions.
11564 @xref{MIPS Paired-Single Support}. This option can only be used
11565 when generating 64-bit code and requires hardware floating-point
11566 support to be enabled.
11572 Use (do not use) MIPS Digital Media Extension instructions.
11573 This option can only be used when generating 64-bit code and requires
11574 hardware floating-point support to be enabled.
11579 @opindex mno-mips3d
11580 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
11581 The option @option{-mips3d} implies @option{-mpaired-single}.
11587 Use (do not use) MT Multithreading instructions.
11591 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
11592 an explanation of the default and the way that the pointer size is
11597 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
11599 The default size of @code{int}s, @code{long}s and pointers depends on
11600 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
11601 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
11602 32-bit @code{long}s. Pointers are the same size as @code{long}s,
11603 or the same size as integer registers, whichever is smaller.
11609 Assume (do not assume) that all symbols have 32-bit values, regardless
11610 of the selected ABI@. This option is useful in combination with
11611 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
11612 to generate shorter and faster references to symbolic addresses.
11616 @cindex smaller data references (MIPS)
11617 @cindex gp-relative references (MIPS)
11618 Put global and static items less than or equal to @var{num} bytes into
11619 the small data or bss section instead of the normal data or bss section.
11620 This allows the data to be accessed using a single instruction.
11622 All modules should be compiled with the same @option{-G @var{num}}
11625 @item -membedded-data
11626 @itemx -mno-embedded-data
11627 @opindex membedded-data
11628 @opindex mno-embedded-data
11629 Allocate variables to the read-only data section first if possible, then
11630 next in the small data section if possible, otherwise in data. This gives
11631 slightly slower code than the default, but reduces the amount of RAM required
11632 when executing, and thus may be preferred for some embedded systems.
11634 @item -muninit-const-in-rodata
11635 @itemx -mno-uninit-const-in-rodata
11636 @opindex muninit-const-in-rodata
11637 @opindex mno-uninit-const-in-rodata
11638 Put uninitialized @code{const} variables in the read-only data section.
11639 This option is only meaningful in conjunction with @option{-membedded-data}.
11641 @item -msplit-addresses
11642 @itemx -mno-split-addresses
11643 @opindex msplit-addresses
11644 @opindex mno-split-addresses
11645 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
11646 relocation operators. This option has been superseded by
11647 @option{-mexplicit-relocs} but is retained for backwards compatibility.
11649 @item -mexplicit-relocs
11650 @itemx -mno-explicit-relocs
11651 @opindex mexplicit-relocs
11652 @opindex mno-explicit-relocs
11653 Use (do not use) assembler relocation operators when dealing with symbolic
11654 addresses. The alternative, selected by @option{-mno-explicit-relocs},
11655 is to use assembler macros instead.
11657 @option{-mexplicit-relocs} is the default if GCC was configured
11658 to use an assembler that supports relocation operators.
11660 @item -mcheck-zero-division
11661 @itemx -mno-check-zero-division
11662 @opindex mcheck-zero-division
11663 @opindex mno-check-zero-division
11664 Trap (do not trap) on integer division by zero.
11666 The default is @option{-mcheck-zero-division}.
11668 @item -mdivide-traps
11669 @itemx -mdivide-breaks
11670 @opindex mdivide-traps
11671 @opindex mdivide-breaks
11672 MIPS systems check for division by zero by generating either a
11673 conditional trap or a break instruction. Using traps results in
11674 smaller code, but is only supported on MIPS II and later. Also, some
11675 versions of the Linux kernel have a bug that prevents trap from
11676 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
11677 allow conditional traps on architectures that support them and
11678 @option{-mdivide-breaks} to force the use of breaks.
11680 The default is usually @option{-mdivide-traps}, but this can be
11681 overridden at configure time using @option{--with-divide=breaks}.
11682 Divide-by-zero checks can be completely disabled using
11683 @option{-mno-check-zero-division}.
11688 @opindex mno-memcpy
11689 Force (do not force) the use of @code{memcpy()} for non-trivial block
11690 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
11691 most constant-sized copies.
11694 @itemx -mno-long-calls
11695 @opindex mlong-calls
11696 @opindex mno-long-calls
11697 Disable (do not disable) use of the @code{jal} instruction. Calling
11698 functions using @code{jal} is more efficient but requires the caller
11699 and callee to be in the same 256 megabyte segment.
11701 This option has no effect on abicalls code. The default is
11702 @option{-mno-long-calls}.
11708 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
11709 instructions, as provided by the R4650 ISA@.
11712 @itemx -mno-fused-madd
11713 @opindex mfused-madd
11714 @opindex mno-fused-madd
11715 Enable (disable) use of the floating point multiply-accumulate
11716 instructions, when they are available. The default is
11717 @option{-mfused-madd}.
11719 When multiply-accumulate instructions are used, the intermediate
11720 product is calculated to infinite precision and is not subject to
11721 the FCSR Flush to Zero bit. This may be undesirable in some
11726 Tell the MIPS assembler to not run its preprocessor over user
11727 assembler files (with a @samp{.s} suffix) when assembling them.
11730 @itemx -mno-fix-r4000
11731 @opindex mfix-r4000
11732 @opindex mno-fix-r4000
11733 Work around certain R4000 CPU errata:
11736 A double-word or a variable shift may give an incorrect result if executed
11737 immediately after starting an integer division.
11739 A double-word or a variable shift may give an incorrect result if executed
11740 while an integer multiplication is in progress.
11742 An integer division may give an incorrect result if started in a delay slot
11743 of a taken branch or a jump.
11747 @itemx -mno-fix-r4400
11748 @opindex mfix-r4400
11749 @opindex mno-fix-r4400
11750 Work around certain R4400 CPU errata:
11753 A double-word or a variable shift may give an incorrect result if executed
11754 immediately after starting an integer division.
11758 @itemx -mno-fix-vr4120
11759 @opindex mfix-vr4120
11760 Work around certain VR4120 errata:
11763 @code{dmultu} does not always produce the correct result.
11765 @code{div} and @code{ddiv} do not always produce the correct result if one
11766 of the operands is negative.
11768 The workarounds for the division errata rely on special functions in
11769 @file{libgcc.a}. At present, these functions are only provided by
11770 the @code{mips64vr*-elf} configurations.
11772 Other VR4120 errata require a nop to be inserted between certain pairs of
11773 instructions. These errata are handled by the assembler, not by GCC itself.
11776 @opindex mfix-vr4130
11777 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
11778 workarounds are implemented by the assembler rather than by GCC,
11779 although GCC will avoid using @code{mflo} and @code{mfhi} if the
11780 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
11781 instructions are available instead.
11784 @itemx -mno-fix-sb1
11786 Work around certain SB-1 CPU core errata.
11787 (This flag currently works around the SB-1 revision 2
11788 ``F1'' and ``F2'' floating point errata.)
11790 @item -mflush-func=@var{func}
11791 @itemx -mno-flush-func
11792 @opindex mflush-func
11793 Specifies the function to call to flush the I and D caches, or to not
11794 call any such function. If called, the function must take the same
11795 arguments as the common @code{_flush_func()}, that is, the address of the
11796 memory range for which the cache is being flushed, the size of the
11797 memory range, and the number 3 (to flush both caches). The default
11798 depends on the target GCC was configured for, but commonly is either
11799 @samp{_flush_func} or @samp{__cpu_flush}.
11801 @item -mbranch-likely
11802 @itemx -mno-branch-likely
11803 @opindex mbranch-likely
11804 @opindex mno-branch-likely
11805 Enable or disable use of Branch Likely instructions, regardless of the
11806 default for the selected architecture. By default, Branch Likely
11807 instructions may be generated if they are supported by the selected
11808 architecture. An exception is for the MIPS32 and MIPS64 architectures
11809 and processors which implement those architectures; for those, Branch
11810 Likely instructions will not be generated by default because the MIPS32
11811 and MIPS64 architectures specifically deprecate their use.
11813 @item -mfp-exceptions
11814 @itemx -mno-fp-exceptions
11815 @opindex mfp-exceptions
11816 Specifies whether FP exceptions are enabled. This affects how we schedule
11817 FP instructions for some processors. The default is that FP exceptions are
11820 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
11821 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
11824 @item -mvr4130-align
11825 @itemx -mno-vr4130-align
11826 @opindex mvr4130-align
11827 The VR4130 pipeline is two-way superscalar, but can only issue two
11828 instructions together if the first one is 8-byte aligned. When this
11829 option is enabled, GCC will align pairs of instructions that it
11830 thinks should execute in parallel.
11832 This option only has an effect when optimizing for the VR4130.
11833 It normally makes code faster, but at the expense of making it bigger.
11834 It is enabled by default at optimization level @option{-O3}.
11838 @subsection MMIX Options
11839 @cindex MMIX Options
11841 These options are defined for the MMIX:
11845 @itemx -mno-libfuncs
11847 @opindex mno-libfuncs
11848 Specify that intrinsic library functions are being compiled, passing all
11849 values in registers, no matter the size.
11852 @itemx -mno-epsilon
11854 @opindex mno-epsilon
11855 Generate floating-point comparison instructions that compare with respect
11856 to the @code{rE} epsilon register.
11858 @item -mabi=mmixware
11860 @opindex mabi-mmixware
11862 Generate code that passes function parameters and return values that (in
11863 the called function) are seen as registers @code{$0} and up, as opposed to
11864 the GNU ABI which uses global registers @code{$231} and up.
11866 @item -mzero-extend
11867 @itemx -mno-zero-extend
11868 @opindex mzero-extend
11869 @opindex mno-zero-extend
11870 When reading data from memory in sizes shorter than 64 bits, use (do not
11871 use) zero-extending load instructions by default, rather than
11872 sign-extending ones.
11875 @itemx -mno-knuthdiv
11877 @opindex mno-knuthdiv
11878 Make the result of a division yielding a remainder have the same sign as
11879 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
11880 remainder follows the sign of the dividend. Both methods are
11881 arithmetically valid, the latter being almost exclusively used.
11883 @item -mtoplevel-symbols
11884 @itemx -mno-toplevel-symbols
11885 @opindex mtoplevel-symbols
11886 @opindex mno-toplevel-symbols
11887 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
11888 code can be used with the @code{PREFIX} assembly directive.
11892 Generate an executable in the ELF format, rather than the default
11893 @samp{mmo} format used by the @command{mmix} simulator.
11895 @item -mbranch-predict
11896 @itemx -mno-branch-predict
11897 @opindex mbranch-predict
11898 @opindex mno-branch-predict
11899 Use (do not use) the probable-branch instructions, when static branch
11900 prediction indicates a probable branch.
11902 @item -mbase-addresses
11903 @itemx -mno-base-addresses
11904 @opindex mbase-addresses
11905 @opindex mno-base-addresses
11906 Generate (do not generate) code that uses @emph{base addresses}. Using a
11907 base address automatically generates a request (handled by the assembler
11908 and the linker) for a constant to be set up in a global register. The
11909 register is used for one or more base address requests within the range 0
11910 to 255 from the value held in the register. The generally leads to short
11911 and fast code, but the number of different data items that can be
11912 addressed is limited. This means that a program that uses lots of static
11913 data may require @option{-mno-base-addresses}.
11915 @item -msingle-exit
11916 @itemx -mno-single-exit
11917 @opindex msingle-exit
11918 @opindex mno-single-exit
11919 Force (do not force) generated code to have a single exit point in each
11923 @node MN10300 Options
11924 @subsection MN10300 Options
11925 @cindex MN10300 options
11927 These @option{-m} options are defined for Matsushita MN10300 architectures:
11932 Generate code to avoid bugs in the multiply instructions for the MN10300
11933 processors. This is the default.
11935 @item -mno-mult-bug
11936 @opindex mno-mult-bug
11937 Do not generate code to avoid bugs in the multiply instructions for the
11938 MN10300 processors.
11942 Generate code which uses features specific to the AM33 processor.
11946 Do not generate code which uses features specific to the AM33 processor. This
11949 @item -mreturn-pointer-on-d0
11950 @opindex mreturn-pointer-on-d0
11951 When generating a function which returns a pointer, return the pointer
11952 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
11953 only in a0, and attempts to call such functions without a prototype
11954 would result in errors. Note that this option is on by default; use
11955 @option{-mno-return-pointer-on-d0} to disable it.
11959 Do not link in the C run-time initialization object file.
11963 Indicate to the linker that it should perform a relaxation optimization pass
11964 to shorten branches, calls and absolute memory addresses. This option only
11965 has an effect when used on the command line for the final link step.
11967 This option makes symbolic debugging impossible.
11971 @subsection MT Options
11974 These @option{-m} options are defined for Morpho MT architectures:
11978 @item -march=@var{cpu-type}
11980 Generate code that will run on @var{cpu-type}, which is the name of a system
11981 representing a certain processor type. Possible values for
11982 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
11983 @samp{ms1-16-003} and @samp{ms2}.
11985 When this option is not used, the default is @option{-march=ms1-16-002}.
11989 Use byte loads and stores when generating code.
11993 Do not use byte loads and stores when generating code.
11997 Use simulator runtime
12001 Do not link in the C run-time initialization object file
12002 @file{crti.o}. Other run-time initialization and termination files
12003 such as @file{startup.o} and @file{exit.o} are still included on the
12004 linker command line.
12008 @node PDP-11 Options
12009 @subsection PDP-11 Options
12010 @cindex PDP-11 Options
12012 These options are defined for the PDP-11:
12017 Use hardware FPP floating point. This is the default. (FIS floating
12018 point on the PDP-11/40 is not supported.)
12021 @opindex msoft-float
12022 Do not use hardware floating point.
12026 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
12030 Return floating-point results in memory. This is the default.
12034 Generate code for a PDP-11/40.
12038 Generate code for a PDP-11/45. This is the default.
12042 Generate code for a PDP-11/10.
12044 @item -mbcopy-builtin
12045 @opindex bcopy-builtin
12046 Use inline @code{movmemhi} patterns for copying memory. This is the
12051 Do not use inline @code{movmemhi} patterns for copying memory.
12057 Use 16-bit @code{int}. This is the default.
12063 Use 32-bit @code{int}.
12066 @itemx -mno-float32
12068 @opindex mno-float32
12069 Use 64-bit @code{float}. This is the default.
12072 @itemx -mno-float64
12074 @opindex mno-float64
12075 Use 32-bit @code{float}.
12079 Use @code{abshi2} pattern. This is the default.
12083 Do not use @code{abshi2} pattern.
12085 @item -mbranch-expensive
12086 @opindex mbranch-expensive
12087 Pretend that branches are expensive. This is for experimenting with
12088 code generation only.
12090 @item -mbranch-cheap
12091 @opindex mbranch-cheap
12092 Do not pretend that branches are expensive. This is the default.
12096 Generate code for a system with split I&D@.
12100 Generate code for a system without split I&D@. This is the default.
12104 Use Unix assembler syntax. This is the default when configured for
12105 @samp{pdp11-*-bsd}.
12109 Use DEC assembler syntax. This is the default when configured for any
12110 PDP-11 target other than @samp{pdp11-*-bsd}.
12113 @node PowerPC Options
12114 @subsection PowerPC Options
12115 @cindex PowerPC options
12117 These are listed under @xref{RS/6000 and PowerPC Options}.
12119 @node RS/6000 and PowerPC Options
12120 @subsection IBM RS/6000 and PowerPC Options
12121 @cindex RS/6000 and PowerPC Options
12122 @cindex IBM RS/6000 and PowerPC Options
12124 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
12131 @itemx -mno-powerpc
12132 @itemx -mpowerpc-gpopt
12133 @itemx -mno-powerpc-gpopt
12134 @itemx -mpowerpc-gfxopt
12135 @itemx -mno-powerpc-gfxopt
12137 @itemx -mno-powerpc64
12141 @itemx -mno-popcntb
12153 @opindex mno-power2
12155 @opindex mno-powerpc
12156 @opindex mpowerpc-gpopt
12157 @opindex mno-powerpc-gpopt
12158 @opindex mpowerpc-gfxopt
12159 @opindex mno-powerpc-gfxopt
12160 @opindex mpowerpc64
12161 @opindex mno-powerpc64
12165 @opindex mno-popcntb
12171 @opindex mno-mfpgpr
12174 GCC supports two related instruction set architectures for the
12175 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
12176 instructions supported by the @samp{rios} chip set used in the original
12177 RS/6000 systems and the @dfn{PowerPC} instruction set is the
12178 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
12179 the IBM 4xx, 6xx, and follow-on microprocessors.
12181 Neither architecture is a subset of the other. However there is a
12182 large common subset of instructions supported by both. An MQ
12183 register is included in processors supporting the POWER architecture.
12185 You use these options to specify which instructions are available on the
12186 processor you are using. The default value of these options is
12187 determined when configuring GCC@. Specifying the
12188 @option{-mcpu=@var{cpu_type}} overrides the specification of these
12189 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
12190 rather than the options listed above.
12192 The @option{-mpower} option allows GCC to generate instructions that
12193 are found only in the POWER architecture and to use the MQ register.
12194 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
12195 to generate instructions that are present in the POWER2 architecture but
12196 not the original POWER architecture.
12198 The @option{-mpowerpc} option allows GCC to generate instructions that
12199 are found only in the 32-bit subset of the PowerPC architecture.
12200 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
12201 GCC to use the optional PowerPC architecture instructions in the
12202 General Purpose group, including floating-point square root. Specifying
12203 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
12204 use the optional PowerPC architecture instructions in the Graphics
12205 group, including floating-point select.
12207 The @option{-mmfcrf} option allows GCC to generate the move from
12208 condition register field instruction implemented on the POWER4
12209 processor and other processors that support the PowerPC V2.01
12211 The @option{-mpopcntb} option allows GCC to generate the popcount and
12212 double precision FP reciprocal estimate instruction implemented on the
12213 POWER5 processor and other processors that support the PowerPC V2.02
12215 The @option{-mfprnd} option allows GCC to generate the FP round to
12216 integer instructions implemented on the POWER5+ processor and other
12217 processors that support the PowerPC V2.03 architecture.
12218 The @option{-mcmpb} option allows GCC to generate the compare bytes
12219 instruction implemented on the POWER6 processor and other processors
12220 that support the PowerPC V2.05 architecture.
12221 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
12222 general purpose register instructions implemented on the POWER6X
12223 processor and other processors that support the extended PowerPC V2.05
12225 The @option{-mdfp} option allows GCC to generate the decimal floating
12226 point instructions implemented on some POWER processors.
12228 The @option{-mpowerpc64} option allows GCC to generate the additional
12229 64-bit instructions that are found in the full PowerPC64 architecture
12230 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
12231 @option{-mno-powerpc64}.
12233 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
12234 will use only the instructions in the common subset of both
12235 architectures plus some special AIX common-mode calls, and will not use
12236 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
12237 permits GCC to use any instruction from either architecture and to
12238 allow use of the MQ register; specify this for the Motorola MPC601.
12240 @item -mnew-mnemonics
12241 @itemx -mold-mnemonics
12242 @opindex mnew-mnemonics
12243 @opindex mold-mnemonics
12244 Select which mnemonics to use in the generated assembler code. With
12245 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
12246 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
12247 assembler mnemonics defined for the POWER architecture. Instructions
12248 defined in only one architecture have only one mnemonic; GCC uses that
12249 mnemonic irrespective of which of these options is specified.
12251 GCC defaults to the mnemonics appropriate for the architecture in
12252 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
12253 value of these option. Unless you are building a cross-compiler, you
12254 should normally not specify either @option{-mnew-mnemonics} or
12255 @option{-mold-mnemonics}, but should instead accept the default.
12257 @item -mcpu=@var{cpu_type}
12259 Set architecture type, register usage, choice of mnemonics, and
12260 instruction scheduling parameters for machine type @var{cpu_type}.
12261 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
12262 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
12263 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
12264 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
12265 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
12266 @samp{860}, @samp{970}, @samp{8540}, @samp{ec603e}, @samp{G3},
12267 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
12268 @samp{power4}, @samp{power5}, @samp{power5+}, @samp{power6},
12269 @samp{power6x}, @samp{common}, @samp{powerpc}, @samp{powerpc64},
12270 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
12272 @option{-mcpu=common} selects a completely generic processor. Code
12273 generated under this option will run on any POWER or PowerPC processor.
12274 GCC will use only the instructions in the common subset of both
12275 architectures, and will not use the MQ register. GCC assumes a generic
12276 processor model for scheduling purposes.
12278 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
12279 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
12280 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
12281 types, with an appropriate, generic processor model assumed for
12282 scheduling purposes.
12284 The other options specify a specific processor. Code generated under
12285 those options will run best on that processor, and may not run at all on
12288 The @option{-mcpu} options automatically enable or disable the
12291 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
12292 -mnew-mnemonics -mpopcntb -mpower -mpower2 -mpowerpc64 @gol
12293 -mpowerpc-gpopt -mpowerpc-gfxopt -mstring -mmulhw -mdlmzb -mmfpgpr}
12295 The particular options set for any particular CPU will vary between
12296 compiler versions, depending on what setting seems to produce optimal
12297 code for that CPU; it doesn't necessarily reflect the actual hardware's
12298 capabilities. If you wish to set an individual option to a particular
12299 value, you may specify it after the @option{-mcpu} option, like
12300 @samp{-mcpu=970 -mno-altivec}.
12302 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
12303 not enabled or disabled by the @option{-mcpu} option at present because
12304 AIX does not have full support for these options. You may still
12305 enable or disable them individually if you're sure it'll work in your
12308 @item -mtune=@var{cpu_type}
12310 Set the instruction scheduling parameters for machine type
12311 @var{cpu_type}, but do not set the architecture type, register usage, or
12312 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
12313 values for @var{cpu_type} are used for @option{-mtune} as for
12314 @option{-mcpu}. If both are specified, the code generated will use the
12315 architecture, registers, and mnemonics set by @option{-mcpu}, but the
12316 scheduling parameters set by @option{-mtune}.
12322 Generate code to compute division as reciprocal estimate and iterative
12323 refinement, creating opportunities for increased throughput. This
12324 feature requires: optional PowerPC Graphics instruction set for single
12325 precision and FRE instruction for double precision, assuming divides
12326 cannot generate user-visible traps, and the domain values not include
12327 Infinities, denormals or zero denominator.
12330 @itemx -mno-altivec
12332 @opindex mno-altivec
12333 Generate code that uses (does not use) AltiVec instructions, and also
12334 enable the use of built-in functions that allow more direct access to
12335 the AltiVec instruction set. You may also need to set
12336 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
12342 @opindex mno-vrsave
12343 Generate VRSAVE instructions when generating AltiVec code.
12346 @opindex msecure-plt
12347 Generate code that allows ld and ld.so to build executables and shared
12348 libraries with non-exec .plt and .got sections. This is a PowerPC
12349 32-bit SYSV ABI option.
12353 Generate code that uses a BSS .plt section that ld.so fills in, and
12354 requires .plt and .got sections that are both writable and executable.
12355 This is a PowerPC 32-bit SYSV ABI option.
12361 This switch enables or disables the generation of ISEL instructions.
12363 @item -misel=@var{yes/no}
12364 This switch has been deprecated. Use @option{-misel} and
12365 @option{-mno-isel} instead.
12371 This switch enables or disables the generation of SPE simd
12374 @item -mspe=@var{yes/no}
12375 This option has been deprecated. Use @option{-mspe} and
12376 @option{-mno-spe} instead.
12378 @item -mfloat-gprs=@var{yes/single/double/no}
12379 @itemx -mfloat-gprs
12380 @opindex mfloat-gprs
12381 This switch enables or disables the generation of floating point
12382 operations on the general purpose registers for architectures that
12385 The argument @var{yes} or @var{single} enables the use of
12386 single-precision floating point operations.
12388 The argument @var{double} enables the use of single and
12389 double-precision floating point operations.
12391 The argument @var{no} disables floating point operations on the
12392 general purpose registers.
12394 This option is currently only available on the MPC854x.
12400 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
12401 targets (including GNU/Linux). The 32-bit environment sets int, long
12402 and pointer to 32 bits and generates code that runs on any PowerPC
12403 variant. The 64-bit environment sets int to 32 bits and long and
12404 pointer to 64 bits, and generates code for PowerPC64, as for
12405 @option{-mpowerpc64}.
12408 @itemx -mno-fp-in-toc
12409 @itemx -mno-sum-in-toc
12410 @itemx -mminimal-toc
12412 @opindex mno-fp-in-toc
12413 @opindex mno-sum-in-toc
12414 @opindex mminimal-toc
12415 Modify generation of the TOC (Table Of Contents), which is created for
12416 every executable file. The @option{-mfull-toc} option is selected by
12417 default. In that case, GCC will allocate at least one TOC entry for
12418 each unique non-automatic variable reference in your program. GCC
12419 will also place floating-point constants in the TOC@. However, only
12420 16,384 entries are available in the TOC@.
12422 If you receive a linker error message that saying you have overflowed
12423 the available TOC space, you can reduce the amount of TOC space used
12424 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
12425 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
12426 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
12427 generate code to calculate the sum of an address and a constant at
12428 run-time instead of putting that sum into the TOC@. You may specify one
12429 or both of these options. Each causes GCC to produce very slightly
12430 slower and larger code at the expense of conserving TOC space.
12432 If you still run out of space in the TOC even when you specify both of
12433 these options, specify @option{-mminimal-toc} instead. This option causes
12434 GCC to make only one TOC entry for every file. When you specify this
12435 option, GCC will produce code that is slower and larger but which
12436 uses extremely little TOC space. You may wish to use this option
12437 only on files that contain less frequently executed code.
12443 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
12444 @code{long} type, and the infrastructure needed to support them.
12445 Specifying @option{-maix64} implies @option{-mpowerpc64} and
12446 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
12447 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
12450 @itemx -mno-xl-compat
12451 @opindex mxl-compat
12452 @opindex mno-xl-compat
12453 Produce code that conforms more closely to IBM XL compiler semantics
12454 when using AIX-compatible ABI. Pass floating-point arguments to
12455 prototyped functions beyond the register save area (RSA) on the stack
12456 in addition to argument FPRs. Do not assume that most significant
12457 double in 128-bit long double value is properly rounded when comparing
12458 values and converting to double. Use XL symbol names for long double
12461 The AIX calling convention was extended but not initially documented to
12462 handle an obscure K&R C case of calling a function that takes the
12463 address of its arguments with fewer arguments than declared. IBM XL
12464 compilers access floating point arguments which do not fit in the
12465 RSA from the stack when a subroutine is compiled without
12466 optimization. Because always storing floating-point arguments on the
12467 stack is inefficient and rarely needed, this option is not enabled by
12468 default and only is necessary when calling subroutines compiled by IBM
12469 XL compilers without optimization.
12473 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
12474 application written to use message passing with special startup code to
12475 enable the application to run. The system must have PE installed in the
12476 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
12477 must be overridden with the @option{-specs=} option to specify the
12478 appropriate directory location. The Parallel Environment does not
12479 support threads, so the @option{-mpe} option and the @option{-pthread}
12480 option are incompatible.
12482 @item -malign-natural
12483 @itemx -malign-power
12484 @opindex malign-natural
12485 @opindex malign-power
12486 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
12487 @option{-malign-natural} overrides the ABI-defined alignment of larger
12488 types, such as floating-point doubles, on their natural size-based boundary.
12489 The option @option{-malign-power} instructs GCC to follow the ABI-specified
12490 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
12492 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
12496 @itemx -mhard-float
12497 @opindex msoft-float
12498 @opindex mhard-float
12499 Generate code that does not use (uses) the floating-point register set.
12500 Software floating point emulation is provided if you use the
12501 @option{-msoft-float} option, and pass the option to GCC when linking.
12504 @itemx -mno-multiple
12506 @opindex mno-multiple
12507 Generate code that uses (does not use) the load multiple word
12508 instructions and the store multiple word instructions. These
12509 instructions are generated by default on POWER systems, and not
12510 generated on PowerPC systems. Do not use @option{-mmultiple} on little
12511 endian PowerPC systems, since those instructions do not work when the
12512 processor is in little endian mode. The exceptions are PPC740 and
12513 PPC750 which permit the instructions usage in little endian mode.
12518 @opindex mno-string
12519 Generate code that uses (does not use) the load string instructions
12520 and the store string word instructions to save multiple registers and
12521 do small block moves. These instructions are generated by default on
12522 POWER systems, and not generated on PowerPC systems. Do not use
12523 @option{-mstring} on little endian PowerPC systems, since those
12524 instructions do not work when the processor is in little endian mode.
12525 The exceptions are PPC740 and PPC750 which permit the instructions
12526 usage in little endian mode.
12531 @opindex mno-update
12532 Generate code that uses (does not use) the load or store instructions
12533 that update the base register to the address of the calculated memory
12534 location. These instructions are generated by default. If you use
12535 @option{-mno-update}, there is a small window between the time that the
12536 stack pointer is updated and the address of the previous frame is
12537 stored, which means code that walks the stack frame across interrupts or
12538 signals may get corrupted data.
12541 @itemx -mno-fused-madd
12542 @opindex mfused-madd
12543 @opindex mno-fused-madd
12544 Generate code that uses (does not use) the floating point multiply and
12545 accumulate instructions. These instructions are generated by default if
12546 hardware floating is used.
12552 Generate code that uses (does not use) the half-word multiply and
12553 multiply-accumulate instructions on the IBM 405 and 440 processors.
12554 These instructions are generated by default when targetting those
12561 Generate code that uses (does not use) the string-search @samp{dlmzb}
12562 instruction on the IBM 405 and 440 processors. This instruction is
12563 generated by default when targetting those processors.
12565 @item -mno-bit-align
12567 @opindex mno-bit-align
12568 @opindex mbit-align
12569 On System V.4 and embedded PowerPC systems do not (do) force structures
12570 and unions that contain bit-fields to be aligned to the base type of the
12573 For example, by default a structure containing nothing but 8
12574 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
12575 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
12576 the structure would be aligned to a 1 byte boundary and be one byte in
12579 @item -mno-strict-align
12580 @itemx -mstrict-align
12581 @opindex mno-strict-align
12582 @opindex mstrict-align
12583 On System V.4 and embedded PowerPC systems do not (do) assume that
12584 unaligned memory references will be handled by the system.
12586 @item -mrelocatable
12587 @itemx -mno-relocatable
12588 @opindex mrelocatable
12589 @opindex mno-relocatable
12590 On embedded PowerPC systems generate code that allows (does not allow)
12591 the program to be relocated to a different address at runtime. If you
12592 use @option{-mrelocatable} on any module, all objects linked together must
12593 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
12595 @item -mrelocatable-lib
12596 @itemx -mno-relocatable-lib
12597 @opindex mrelocatable-lib
12598 @opindex mno-relocatable-lib
12599 On embedded PowerPC systems generate code that allows (does not allow)
12600 the program to be relocated to a different address at runtime. Modules
12601 compiled with @option{-mrelocatable-lib} can be linked with either modules
12602 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
12603 with modules compiled with the @option{-mrelocatable} options.
12609 On System V.4 and embedded PowerPC systems do not (do) assume that
12610 register 2 contains a pointer to a global area pointing to the addresses
12611 used in the program.
12614 @itemx -mlittle-endian
12616 @opindex mlittle-endian
12617 On System V.4 and embedded PowerPC systems compile code for the
12618 processor in little endian mode. The @option{-mlittle-endian} option is
12619 the same as @option{-mlittle}.
12622 @itemx -mbig-endian
12624 @opindex mbig-endian
12625 On System V.4 and embedded PowerPC systems compile code for the
12626 processor in big endian mode. The @option{-mbig-endian} option is
12627 the same as @option{-mbig}.
12629 @item -mdynamic-no-pic
12630 @opindex mdynamic-no-pic
12631 On Darwin and Mac OS X systems, compile code so that it is not
12632 relocatable, but that its external references are relocatable. The
12633 resulting code is suitable for applications, but not shared
12636 @item -mprioritize-restricted-insns=@var{priority}
12637 @opindex mprioritize-restricted-insns
12638 This option controls the priority that is assigned to
12639 dispatch-slot restricted instructions during the second scheduling
12640 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
12641 @var{no/highest/second-highest} priority to dispatch slot restricted
12644 @item -msched-costly-dep=@var{dependence_type}
12645 @opindex msched-costly-dep
12646 This option controls which dependences are considered costly
12647 by the target during instruction scheduling. The argument
12648 @var{dependence_type} takes one of the following values:
12649 @var{no}: no dependence is costly,
12650 @var{all}: all dependences are costly,
12651 @var{true_store_to_load}: a true dependence from store to load is costly,
12652 @var{store_to_load}: any dependence from store to load is costly,
12653 @var{number}: any dependence which latency >= @var{number} is costly.
12655 @item -minsert-sched-nops=@var{scheme}
12656 @opindex minsert-sched-nops
12657 This option controls which nop insertion scheme will be used during
12658 the second scheduling pass. The argument @var{scheme} takes one of the
12660 @var{no}: Don't insert nops.
12661 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
12662 according to the scheduler's grouping.
12663 @var{regroup_exact}: Insert nops to force costly dependent insns into
12664 separate groups. Insert exactly as many nops as needed to force an insn
12665 to a new group, according to the estimated processor grouping.
12666 @var{number}: Insert nops to force costly dependent insns into
12667 separate groups. Insert @var{number} nops to force an insn to a new group.
12670 @opindex mcall-sysv
12671 On System V.4 and embedded PowerPC systems compile code using calling
12672 conventions that adheres to the March 1995 draft of the System V
12673 Application Binary Interface, PowerPC processor supplement. This is the
12674 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
12676 @item -mcall-sysv-eabi
12677 @opindex mcall-sysv-eabi
12678 Specify both @option{-mcall-sysv} and @option{-meabi} options.
12680 @item -mcall-sysv-noeabi
12681 @opindex mcall-sysv-noeabi
12682 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
12684 @item -mcall-solaris
12685 @opindex mcall-solaris
12686 On System V.4 and embedded PowerPC systems compile code for the Solaris
12690 @opindex mcall-linux
12691 On System V.4 and embedded PowerPC systems compile code for the
12692 Linux-based GNU system.
12696 On System V.4 and embedded PowerPC systems compile code for the
12697 Hurd-based GNU system.
12699 @item -mcall-netbsd
12700 @opindex mcall-netbsd
12701 On System V.4 and embedded PowerPC systems compile code for the
12702 NetBSD operating system.
12704 @item -maix-struct-return
12705 @opindex maix-struct-return
12706 Return all structures in memory (as specified by the AIX ABI)@.
12708 @item -msvr4-struct-return
12709 @opindex msvr4-struct-return
12710 Return structures smaller than 8 bytes in registers (as specified by the
12713 @item -mabi=@var{abi-type}
12715 Extend the current ABI with a particular extension, or remove such extension.
12716 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
12717 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
12721 Extend the current ABI with SPE ABI extensions. This does not change
12722 the default ABI, instead it adds the SPE ABI extensions to the current
12726 @opindex mabi=no-spe
12727 Disable Booke SPE ABI extensions for the current ABI@.
12729 @item -mabi=ibmlongdouble
12730 @opindex mabi=ibmlongdouble
12731 Change the current ABI to use IBM extended precision long double.
12732 This is a PowerPC 32-bit SYSV ABI option.
12734 @item -mabi=ieeelongdouble
12735 @opindex mabi=ieeelongdouble
12736 Change the current ABI to use IEEE extended precision long double.
12737 This is a PowerPC 32-bit Linux ABI option.
12740 @itemx -mno-prototype
12741 @opindex mprototype
12742 @opindex mno-prototype
12743 On System V.4 and embedded PowerPC systems assume that all calls to
12744 variable argument functions are properly prototyped. Otherwise, the
12745 compiler must insert an instruction before every non prototyped call to
12746 set or clear bit 6 of the condition code register (@var{CR}) to
12747 indicate whether floating point values were passed in the floating point
12748 registers in case the function takes a variable arguments. With
12749 @option{-mprototype}, only calls to prototyped variable argument functions
12750 will set or clear the bit.
12754 On embedded PowerPC systems, assume that the startup module is called
12755 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
12756 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
12761 On embedded PowerPC systems, assume that the startup module is called
12762 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
12767 On embedded PowerPC systems, assume that the startup module is called
12768 @file{crt0.o} and the standard C libraries are @file{libads.a} and
12771 @item -myellowknife
12772 @opindex myellowknife
12773 On embedded PowerPC systems, assume that the startup module is called
12774 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
12779 On System V.4 and embedded PowerPC systems, specify that you are
12780 compiling for a VxWorks system.
12784 Specify that you are compiling for the WindISS simulation environment.
12788 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
12789 header to indicate that @samp{eabi} extended relocations are used.
12795 On System V.4 and embedded PowerPC systems do (do not) adhere to the
12796 Embedded Applications Binary Interface (eabi) which is a set of
12797 modifications to the System V.4 specifications. Selecting @option{-meabi}
12798 means that the stack is aligned to an 8 byte boundary, a function
12799 @code{__eabi} is called to from @code{main} to set up the eabi
12800 environment, and the @option{-msdata} option can use both @code{r2} and
12801 @code{r13} to point to two separate small data areas. Selecting
12802 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
12803 do not call an initialization function from @code{main}, and the
12804 @option{-msdata} option will only use @code{r13} to point to a single
12805 small data area. The @option{-meabi} option is on by default if you
12806 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
12809 @opindex msdata=eabi
12810 On System V.4 and embedded PowerPC systems, put small initialized
12811 @code{const} global and static data in the @samp{.sdata2} section, which
12812 is pointed to by register @code{r2}. Put small initialized
12813 non-@code{const} global and static data in the @samp{.sdata} section,
12814 which is pointed to by register @code{r13}. Put small uninitialized
12815 global and static data in the @samp{.sbss} section, which is adjacent to
12816 the @samp{.sdata} section. The @option{-msdata=eabi} option is
12817 incompatible with the @option{-mrelocatable} option. The
12818 @option{-msdata=eabi} option also sets the @option{-memb} option.
12821 @opindex msdata=sysv
12822 On System V.4 and embedded PowerPC systems, put small global and static
12823 data in the @samp{.sdata} section, which is pointed to by register
12824 @code{r13}. Put small uninitialized global and static data in the
12825 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
12826 The @option{-msdata=sysv} option is incompatible with the
12827 @option{-mrelocatable} option.
12829 @item -msdata=default
12831 @opindex msdata=default
12833 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
12834 compile code the same as @option{-msdata=eabi}, otherwise compile code the
12835 same as @option{-msdata=sysv}.
12838 @opindex msdata-data
12839 On System V.4 and embedded PowerPC systems, put small global
12840 data in the @samp{.sdata} section. Put small uninitialized global
12841 data in the @samp{.sbss} section. Do not use register @code{r13}
12842 to address small data however. This is the default behavior unless
12843 other @option{-msdata} options are used.
12847 @opindex msdata=none
12849 On embedded PowerPC systems, put all initialized global and static data
12850 in the @samp{.data} section, and all uninitialized data in the
12851 @samp{.bss} section.
12855 @cindex smaller data references (PowerPC)
12856 @cindex .sdata/.sdata2 references (PowerPC)
12857 On embedded PowerPC systems, put global and static items less than or
12858 equal to @var{num} bytes into the small data or bss sections instead of
12859 the normal data or bss section. By default, @var{num} is 8. The
12860 @option{-G @var{num}} switch is also passed to the linker.
12861 All modules should be compiled with the same @option{-G @var{num}} value.
12864 @itemx -mno-regnames
12866 @opindex mno-regnames
12867 On System V.4 and embedded PowerPC systems do (do not) emit register
12868 names in the assembly language output using symbolic forms.
12871 @itemx -mno-longcall
12873 @opindex mno-longcall
12874 By default assume that all calls are far away so that a longer more
12875 expensive calling sequence is required. This is required for calls
12876 further than 32 megabytes (33,554,432 bytes) from the current location.
12877 A short call will be generated if the compiler knows
12878 the call cannot be that far away. This setting can be overridden by
12879 the @code{shortcall} function attribute, or by @code{#pragma
12882 Some linkers are capable of detecting out-of-range calls and generating
12883 glue code on the fly. On these systems, long calls are unnecessary and
12884 generate slower code. As of this writing, the AIX linker can do this,
12885 as can the GNU linker for PowerPC/64. It is planned to add this feature
12886 to the GNU linker for 32-bit PowerPC systems as well.
12888 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
12889 callee, L42'', plus a ``branch island'' (glue code). The two target
12890 addresses represent the callee and the ``branch island''. The
12891 Darwin/PPC linker will prefer the first address and generate a ``bl
12892 callee'' if the PPC ``bl'' instruction will reach the callee directly;
12893 otherwise, the linker will generate ``bl L42'' to call the ``branch
12894 island''. The ``branch island'' is appended to the body of the
12895 calling function; it computes the full 32-bit address of the callee
12898 On Mach-O (Darwin) systems, this option directs the compiler emit to
12899 the glue for every direct call, and the Darwin linker decides whether
12900 to use or discard it.
12902 In the future, we may cause GCC to ignore all longcall specifications
12903 when the linker is known to generate glue.
12907 Adds support for multithreading with the @dfn{pthreads} library.
12908 This option sets flags for both the preprocessor and linker.
12912 @node S/390 and zSeries Options
12913 @subsection S/390 and zSeries Options
12914 @cindex S/390 and zSeries Options
12916 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
12920 @itemx -msoft-float
12921 @opindex mhard-float
12922 @opindex msoft-float
12923 Use (do not use) the hardware floating-point instructions and registers
12924 for floating-point operations. When @option{-msoft-float} is specified,
12925 functions in @file{libgcc.a} will be used to perform floating-point
12926 operations. When @option{-mhard-float} is specified, the compiler
12927 generates IEEE floating-point instructions. This is the default.
12929 @item -mlong-double-64
12930 @itemx -mlong-double-128
12931 @opindex mlong-double-64
12932 @opindex mlong-double-128
12933 These switches control the size of @code{long double} type. A size
12934 of 64bit makes the @code{long double} type equivalent to the @code{double}
12935 type. This is the default.
12938 @itemx -mno-backchain
12939 @opindex mbackchain
12940 @opindex mno-backchain
12941 Store (do not store) the address of the caller's frame as backchain pointer
12942 into the callee's stack frame.
12943 A backchain may be needed to allow debugging using tools that do not understand
12944 DWARF-2 call frame information.
12945 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
12946 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
12947 the backchain is placed into the topmost word of the 96/160 byte register
12950 In general, code compiled with @option{-mbackchain} is call-compatible with
12951 code compiled with @option{-mmo-backchain}; however, use of the backchain
12952 for debugging purposes usually requires that the whole binary is built with
12953 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
12954 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
12955 to build a linux kernel use @option{-msoft-float}.
12957 The default is to not maintain the backchain.
12959 @item -mpacked-stack
12960 @item -mno-packed-stack
12961 @opindex mpacked-stack
12962 @opindex mno-packed-stack
12963 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
12964 specified, the compiler uses the all fields of the 96/160 byte register save
12965 area only for their default purpose; unused fields still take up stack space.
12966 When @option{-mpacked-stack} is specified, register save slots are densely
12967 packed at the top of the register save area; unused space is reused for other
12968 purposes, allowing for more efficient use of the available stack space.
12969 However, when @option{-mbackchain} is also in effect, the topmost word of
12970 the save area is always used to store the backchain, and the return address
12971 register is always saved two words below the backchain.
12973 As long as the stack frame backchain is not used, code generated with
12974 @option{-mpacked-stack} is call-compatible with code generated with
12975 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
12976 S/390 or zSeries generated code that uses the stack frame backchain at run
12977 time, not just for debugging purposes. Such code is not call-compatible
12978 with code compiled with @option{-mpacked-stack}. Also, note that the
12979 combination of @option{-mbackchain},
12980 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
12981 to build a linux kernel use @option{-msoft-float}.
12983 The default is to not use the packed stack layout.
12986 @itemx -mno-small-exec
12987 @opindex msmall-exec
12988 @opindex mno-small-exec
12989 Generate (or do not generate) code using the @code{bras} instruction
12990 to do subroutine calls.
12991 This only works reliably if the total executable size does not
12992 exceed 64k. The default is to use the @code{basr} instruction instead,
12993 which does not have this limitation.
12999 When @option{-m31} is specified, generate code compliant to the
13000 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
13001 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
13002 particular to generate 64-bit instructions. For the @samp{s390}
13003 targets, the default is @option{-m31}, while the @samp{s390x}
13004 targets default to @option{-m64}.
13010 When @option{-mzarch} is specified, generate code using the
13011 instructions available on z/Architecture.
13012 When @option{-mesa} is specified, generate code using the
13013 instructions available on ESA/390. Note that @option{-mesa} is
13014 not possible with @option{-m64}.
13015 When generating code compliant to the GNU/Linux for S/390 ABI,
13016 the default is @option{-mesa}. When generating code compliant
13017 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
13023 Generate (or do not generate) code using the @code{mvcle} instruction
13024 to perform block moves. When @option{-mno-mvcle} is specified,
13025 use a @code{mvc} loop instead. This is the default unless optimizing for
13032 Print (or do not print) additional debug information when compiling.
13033 The default is to not print debug information.
13035 @item -march=@var{cpu-type}
13037 Generate code that will run on @var{cpu-type}, which is the name of a system
13038 representing a certain processor type. Possible values for
13039 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
13040 When generating code using the instructions available on z/Architecture,
13041 the default is @option{-march=z900}. Otherwise, the default is
13042 @option{-march=g5}.
13044 @item -mtune=@var{cpu-type}
13046 Tune to @var{cpu-type} everything applicable about the generated code,
13047 except for the ABI and the set of available instructions.
13048 The list of @var{cpu-type} values is the same as for @option{-march}.
13049 The default is the value used for @option{-march}.
13052 @itemx -mno-tpf-trace
13053 @opindex mtpf-trace
13054 @opindex mno-tpf-trace
13055 Generate code that adds (does not add) in TPF OS specific branches to trace
13056 routines in the operating system. This option is off by default, even
13057 when compiling for the TPF OS@.
13060 @itemx -mno-fused-madd
13061 @opindex mfused-madd
13062 @opindex mno-fused-madd
13063 Generate code that uses (does not use) the floating point multiply and
13064 accumulate instructions. These instructions are generated by default if
13065 hardware floating point is used.
13067 @item -mwarn-framesize=@var{framesize}
13068 @opindex mwarn-framesize
13069 Emit a warning if the current function exceeds the given frame size. Because
13070 this is a compile time check it doesn't need to be a real problem when the program
13071 runs. It is intended to identify functions which most probably cause
13072 a stack overflow. It is useful to be used in an environment with limited stack
13073 size e.g.@: the linux kernel.
13075 @item -mwarn-dynamicstack
13076 @opindex mwarn-dynamicstack
13077 Emit a warning if the function calls alloca or uses dynamically
13078 sized arrays. This is generally a bad idea with a limited stack size.
13080 @item -mstack-guard=@var{stack-guard}
13081 @item -mstack-size=@var{stack-size}
13082 @opindex mstack-guard
13083 @opindex mstack-size
13084 If these options are provided the s390 back end emits additional instructions in
13085 the function prologue which trigger a trap if the stack size is @var{stack-guard}
13086 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
13087 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
13088 the frame size of the compiled function is chosen.
13089 These options are intended to be used to help debugging stack overflow problems.
13090 The additionally emitted code causes only little overhead and hence can also be
13091 used in production like systems without greater performance degradation. The given
13092 values have to be exact powers of 2 and @var{stack-size} has to be greater than
13093 @var{stack-guard} without exceeding 64k.
13094 In order to be efficient the extra code makes the assumption that the stack starts
13095 at an address aligned to the value given by @var{stack-size}.
13096 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
13099 @node Score Options
13100 @subsection Score Options
13101 @cindex Score Options
13103 These options are defined for Score implementations:
13108 Compile code for big endian mode. This is the default.
13112 Compile code for little endian mode.
13116 Disable generate bcnz instruction.
13120 Enable generate unaligned load and store instruction.
13124 Enable the use of multiply-accumulate instructions. Disabled by default.
13128 Specify the SCORE5 as the target architecture.
13132 Specify the SCORE5U of the target architecture.
13136 Specify the SCORE7 as the target architecture. This is the default.
13140 Specify the SCORE7D as the target architecture.
13144 @subsection SH Options
13146 These @samp{-m} options are defined for the SH implementations:
13151 Generate code for the SH1.
13155 Generate code for the SH2.
13158 Generate code for the SH2e.
13162 Generate code for the SH3.
13166 Generate code for the SH3e.
13170 Generate code for the SH4 without a floating-point unit.
13172 @item -m4-single-only
13173 @opindex m4-single-only
13174 Generate code for the SH4 with a floating-point unit that only
13175 supports single-precision arithmetic.
13179 Generate code for the SH4 assuming the floating-point unit is in
13180 single-precision mode by default.
13184 Generate code for the SH4.
13188 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
13189 floating-point unit is not used.
13191 @item -m4a-single-only
13192 @opindex m4a-single-only
13193 Generate code for the SH4a, in such a way that no double-precision
13194 floating point operations are used.
13197 @opindex m4a-single
13198 Generate code for the SH4a assuming the floating-point unit is in
13199 single-precision mode by default.
13203 Generate code for the SH4a.
13207 Same as @option{-m4a-nofpu}, except that it implicitly passes
13208 @option{-dsp} to the assembler. GCC doesn't generate any DSP
13209 instructions at the moment.
13213 Compile code for the processor in big endian mode.
13217 Compile code for the processor in little endian mode.
13221 Align doubles at 64-bit boundaries. Note that this changes the calling
13222 conventions, and thus some functions from the standard C library will
13223 not work unless you recompile it first with @option{-mdalign}.
13227 Shorten some address references at link time, when possible; uses the
13228 linker option @option{-relax}.
13232 Use 32-bit offsets in @code{switch} tables. The default is to use
13237 Enable the use of the instruction @code{fmovd}.
13241 Comply with the calling conventions defined by Renesas.
13245 Comply with the calling conventions defined by Renesas.
13249 Comply with the calling conventions defined for GCC before the Renesas
13250 conventions were available. This option is the default for all
13251 targets of the SH toolchain except for @samp{sh-symbianelf}.
13254 @opindex mnomacsave
13255 Mark the @code{MAC} register as call-clobbered, even if
13256 @option{-mhitachi} is given.
13260 Increase IEEE-compliance of floating-point code.
13261 At the moment, this is equivalent to @option{-fno-finite-math-only}.
13262 When generating 16 bit SH opcodes, getting IEEE-conforming results for
13263 comparisons of NANs / infinities incurs extra overhead in every
13264 floating point comparison, therefore the default is set to
13265 @option{-ffinite-math-only}.
13267 @item -minline-ic_invalidate
13268 @opindex minline-ic_invalidate
13269 Inline code to invalidate instruction cache entries after setting up
13270 nested function trampolines.
13271 This option has no effect if -musermode is in effect and the selected
13272 code generation option (e.g. -m4) does not allow the use of the icbi
13274 If the selected code generation option does not allow the use of the icbi
13275 instruction, and -musermode is not in effect, the inlined code will
13276 manipulate the instruction cache address array directly with an associative
13277 write. This not only requires privileged mode, but it will also
13278 fail if the cache line had been mapped via the TLB and has become unmapped.
13282 Dump instruction size and location in the assembly code.
13285 @opindex mpadstruct
13286 This option is deprecated. It pads structures to multiple of 4 bytes,
13287 which is incompatible with the SH ABI@.
13291 Optimize for space instead of speed. Implied by @option{-Os}.
13294 @opindex mprefergot
13295 When generating position-independent code, emit function calls using
13296 the Global Offset Table instead of the Procedure Linkage Table.
13300 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
13301 if the inlined code would not work in user mode.
13302 This is the default when the target is @code{sh-*-linux*}.
13304 @item -multcost=@var{number}
13305 @opindex multcost=@var{number}
13306 Set the cost to assume for a multiply insn.
13308 @item -mdiv=@var{strategy}
13309 @opindex mdiv=@var{strategy}
13310 Set the division strategy to use for SHmedia code. @var{strategy} must be
13311 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
13312 inv:call2, inv:fp .
13313 "fp" performs the operation in floating point. This has a very high latency,
13314 but needs only a few instructions, so it might be a good choice if
13315 your code has enough easily exploitable ILP to allow the compiler to
13316 schedule the floating point instructions together with other instructions.
13317 Division by zero causes a floating point exception.
13318 "inv" uses integer operations to calculate the inverse of the divisor,
13319 and then multiplies the dividend with the inverse. This strategy allows
13320 cse and hoisting of the inverse calculation. Division by zero calculates
13321 an unspecified result, but does not trap.
13322 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
13323 have been found, or if the entire operation has been hoisted to the same
13324 place, the last stages of the inverse calculation are intertwined with the
13325 final multiply to reduce the overall latency, at the expense of using a few
13326 more instructions, and thus offering fewer scheduling opportunities with
13328 "call" calls a library function that usually implements the inv:minlat
13330 This gives high code density for m5-*media-nofpu compilations.
13331 "call2" uses a different entry point of the same library function, where it
13332 assumes that a pointer to a lookup table has already been set up, which
13333 exposes the pointer load to cse / code hoisting optimizations.
13334 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
13335 code generation, but if the code stays unoptimized, revert to the "call",
13336 "call2", or "fp" strategies, respectively. Note that the
13337 potentially-trapping side effect of division by zero is carried by a
13338 separate instruction, so it is possible that all the integer instructions
13339 are hoisted out, but the marker for the side effect stays where it is.
13340 A recombination to fp operations or a call is not possible in that case.
13341 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
13342 that the inverse calculation was nor separated from the multiply, they speed
13343 up division where the dividend fits into 20 bits (plus sign where applicable),
13344 by inserting a test to skip a number of operations in this case; this test
13345 slows down the case of larger dividends. inv20u assumes the case of a such
13346 a small dividend to be unlikely, and inv20l assumes it to be likely.
13348 @item -mdivsi3_libfunc=@var{name}
13349 @opindex mdivsi3_libfunc=@var{name}
13350 Set the name of the library function used for 32 bit signed division to
13351 @var{name}. This only affect the name used in the call and inv:call
13352 division strategies, and the compiler will still expect the same
13353 sets of input/output/clobbered registers as if this option was not present.
13355 @item -madjust-unroll
13356 @opindex madjust-unroll
13357 Throttle unrolling to avoid thrashing target registers.
13358 This option only has an effect if the gcc code base supports the
13359 TARGET_ADJUST_UNROLL_MAX target hook.
13361 @item -mindexed-addressing
13362 @opindex mindexed-addressing
13363 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
13364 This is only safe if the hardware and/or OS implement 32 bit wrap-around
13365 semantics for the indexed addressing mode. The architecture allows the
13366 implementation of processors with 64 bit MMU, which the OS could use to
13367 get 32 bit addressing, but since no current hardware implementation supports
13368 this or any other way to make the indexed addressing mode safe to use in
13369 the 32 bit ABI, the default is -mno-indexed-addressing.
13371 @item -mgettrcost=@var{number}
13372 @opindex mgettrcost=@var{number}
13373 Set the cost assumed for the gettr instruction to @var{number}.
13374 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
13378 Assume pt* instructions won't trap. This will generally generate better
13379 scheduled code, but is unsafe on current hardware. The current architecture
13380 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
13381 This has the unintentional effect of making it unsafe to schedule ptabs /
13382 ptrel before a branch, or hoist it out of a loop. For example,
13383 __do_global_ctors, a part of libgcc that runs constructors at program
13384 startup, calls functions in a list which is delimited by @minus{}1. With the
13385 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
13386 That means that all the constructors will be run a bit quicker, but when
13387 the loop comes to the end of the list, the program crashes because ptabs
13388 loads @minus{}1 into a target register. Since this option is unsafe for any
13389 hardware implementing the current architecture specification, the default
13390 is -mno-pt-fixed. Unless the user specifies a specific cost with
13391 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
13392 this deters register allocation using target registers for storing
13395 @item -minvalid-symbols
13396 @opindex minvalid-symbols
13397 Assume symbols might be invalid. Ordinary function symbols generated by
13398 the compiler will always be valid to load with movi/shori/ptabs or
13399 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
13400 to generate symbols that will cause ptabs / ptrel to trap.
13401 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
13402 It will then prevent cross-basic-block cse, hoisting and most scheduling
13403 of symbol loads. The default is @option{-mno-invalid-symbols}.
13406 @node SPARC Options
13407 @subsection SPARC Options
13408 @cindex SPARC options
13410 These @samp{-m} options are supported on the SPARC:
13413 @item -mno-app-regs
13415 @opindex mno-app-regs
13417 Specify @option{-mapp-regs} to generate output using the global registers
13418 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
13421 To be fully SVR4 ABI compliant at the cost of some performance loss,
13422 specify @option{-mno-app-regs}. You should compile libraries and system
13423 software with this option.
13426 @itemx -mhard-float
13428 @opindex mhard-float
13429 Generate output containing floating point instructions. This is the
13433 @itemx -msoft-float
13435 @opindex msoft-float
13436 Generate output containing library calls for floating point.
13437 @strong{Warning:} the requisite libraries are not available for all SPARC
13438 targets. Normally the facilities of the machine's usual C compiler are
13439 used, but this cannot be done directly in cross-compilation. You must make
13440 your own arrangements to provide suitable library functions for
13441 cross-compilation. The embedded targets @samp{sparc-*-aout} and
13442 @samp{sparclite-*-*} do provide software floating point support.
13444 @option{-msoft-float} changes the calling convention in the output file;
13445 therefore, it is only useful if you compile @emph{all} of a program with
13446 this option. In particular, you need to compile @file{libgcc.a}, the
13447 library that comes with GCC, with @option{-msoft-float} in order for
13450 @item -mhard-quad-float
13451 @opindex mhard-quad-float
13452 Generate output containing quad-word (long double) floating point
13455 @item -msoft-quad-float
13456 @opindex msoft-quad-float
13457 Generate output containing library calls for quad-word (long double)
13458 floating point instructions. The functions called are those specified
13459 in the SPARC ABI@. This is the default.
13461 As of this writing, there are no SPARC implementations that have hardware
13462 support for the quad-word floating point instructions. They all invoke
13463 a trap handler for one of these instructions, and then the trap handler
13464 emulates the effect of the instruction. Because of the trap handler overhead,
13465 this is much slower than calling the ABI library routines. Thus the
13466 @option{-msoft-quad-float} option is the default.
13468 @item -mno-unaligned-doubles
13469 @itemx -munaligned-doubles
13470 @opindex mno-unaligned-doubles
13471 @opindex munaligned-doubles
13472 Assume that doubles have 8 byte alignment. This is the default.
13474 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
13475 alignment only if they are contained in another type, or if they have an
13476 absolute address. Otherwise, it assumes they have 4 byte alignment.
13477 Specifying this option avoids some rare compatibility problems with code
13478 generated by other compilers. It is not the default because it results
13479 in a performance loss, especially for floating point code.
13481 @item -mno-faster-structs
13482 @itemx -mfaster-structs
13483 @opindex mno-faster-structs
13484 @opindex mfaster-structs
13485 With @option{-mfaster-structs}, the compiler assumes that structures
13486 should have 8 byte alignment. This enables the use of pairs of
13487 @code{ldd} and @code{std} instructions for copies in structure
13488 assignment, in place of twice as many @code{ld} and @code{st} pairs.
13489 However, the use of this changed alignment directly violates the SPARC
13490 ABI@. Thus, it's intended only for use on targets where the developer
13491 acknowledges that their resulting code will not be directly in line with
13492 the rules of the ABI@.
13494 @item -mimpure-text
13495 @opindex mimpure-text
13496 @option{-mimpure-text}, used in addition to @option{-shared}, tells
13497 the compiler to not pass @option{-z text} to the linker when linking a
13498 shared object. Using this option, you can link position-dependent
13499 code into a shared object.
13501 @option{-mimpure-text} suppresses the ``relocations remain against
13502 allocatable but non-writable sections'' linker error message.
13503 However, the necessary relocations will trigger copy-on-write, and the
13504 shared object is not actually shared across processes. Instead of
13505 using @option{-mimpure-text}, you should compile all source code with
13506 @option{-fpic} or @option{-fPIC}.
13508 This option is only available on SunOS and Solaris.
13510 @item -mcpu=@var{cpu_type}
13512 Set the instruction set, register set, and instruction scheduling parameters
13513 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
13514 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
13515 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
13516 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
13517 @samp{ultrasparc3}, and @samp{niagara}.
13519 Default instruction scheduling parameters are used for values that select
13520 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
13521 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
13523 Here is a list of each supported architecture and their supported
13528 v8: supersparc, hypersparc
13529 sparclite: f930, f934, sparclite86x
13531 v9: ultrasparc, ultrasparc3, niagara
13534 By default (unless configured otherwise), GCC generates code for the V7
13535 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
13536 additionally optimizes it for the Cypress CY7C602 chip, as used in the
13537 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
13538 SPARCStation 1, 2, IPX etc.
13540 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
13541 architecture. The only difference from V7 code is that the compiler emits
13542 the integer multiply and integer divide instructions which exist in SPARC-V8
13543 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
13544 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
13547 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
13548 the SPARC architecture. This adds the integer multiply, integer divide step
13549 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
13550 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
13551 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
13552 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
13553 MB86934 chip, which is the more recent SPARClite with FPU@.
13555 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
13556 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
13557 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
13558 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
13559 optimizes it for the TEMIC SPARClet chip.
13561 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
13562 architecture. This adds 64-bit integer and floating-point move instructions,
13563 3 additional floating-point condition code registers and conditional move
13564 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
13565 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
13566 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
13567 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
13568 @option{-mcpu=niagara}, the compiler additionally optimizes it for
13569 Sun UltraSPARC T1 chips.
13571 @item -mtune=@var{cpu_type}
13573 Set the instruction scheduling parameters for machine type
13574 @var{cpu_type}, but do not set the instruction set or register set that the
13575 option @option{-mcpu=@var{cpu_type}} would.
13577 The same values for @option{-mcpu=@var{cpu_type}} can be used for
13578 @option{-mtune=@var{cpu_type}}, but the only useful values are those
13579 that select a particular cpu implementation. Those are @samp{cypress},
13580 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
13581 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
13582 @samp{ultrasparc3}, and @samp{niagara}.
13587 @opindex mno-v8plus
13588 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
13589 difference from the V8 ABI is that the global and out registers are
13590 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
13591 mode for all SPARC-V9 processors.
13597 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
13598 Visual Instruction Set extensions. The default is @option{-mno-vis}.
13601 These @samp{-m} options are supported in addition to the above
13602 on SPARC-V9 processors in 64-bit environments:
13605 @item -mlittle-endian
13606 @opindex mlittle-endian
13607 Generate code for a processor running in little-endian mode. It is only
13608 available for a few configurations and most notably not on Solaris and Linux.
13614 Generate code for a 32-bit or 64-bit environment.
13615 The 32-bit environment sets int, long and pointer to 32 bits.
13616 The 64-bit environment sets int to 32 bits and long and pointer
13619 @item -mcmodel=medlow
13620 @opindex mcmodel=medlow
13621 Generate code for the Medium/Low code model: 64-bit addresses, programs
13622 must be linked in the low 32 bits of memory. Programs can be statically
13623 or dynamically linked.
13625 @item -mcmodel=medmid
13626 @opindex mcmodel=medmid
13627 Generate code for the Medium/Middle code model: 64-bit addresses, programs
13628 must be linked in the low 44 bits of memory, the text and data segments must
13629 be less than 2GB in size and the data segment must be located within 2GB of
13632 @item -mcmodel=medany
13633 @opindex mcmodel=medany
13634 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
13635 may be linked anywhere in memory, the text and data segments must be less
13636 than 2GB in size and the data segment must be located within 2GB of the
13639 @item -mcmodel=embmedany
13640 @opindex mcmodel=embmedany
13641 Generate code for the Medium/Anywhere code model for embedded systems:
13642 64-bit addresses, the text and data segments must be less than 2GB in
13643 size, both starting anywhere in memory (determined at link time). The
13644 global register %g4 points to the base of the data segment. Programs
13645 are statically linked and PIC is not supported.
13648 @itemx -mno-stack-bias
13649 @opindex mstack-bias
13650 @opindex mno-stack-bias
13651 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
13652 frame pointer if present, are offset by @minus{}2047 which must be added back
13653 when making stack frame references. This is the default in 64-bit mode.
13654 Otherwise, assume no such offset is present.
13657 These switches are supported in addition to the above on Solaris:
13662 Add support for multithreading using the Solaris threads library. This
13663 option sets flags for both the preprocessor and linker. This option does
13664 not affect the thread safety of object code produced by the compiler or
13665 that of libraries supplied with it.
13669 Add support for multithreading using the POSIX threads library. This
13670 option sets flags for both the preprocessor and linker. This option does
13671 not affect the thread safety of object code produced by the compiler or
13672 that of libraries supplied with it.
13676 This is a synonym for @option{-pthreads}.
13680 @subsection SPU Options
13681 @cindex SPU options
13683 These @samp{-m} options are supported on the SPU:
13687 @itemx -merror-reloc
13688 @opindex mwarn-reloc
13689 @opindex merror-reloc
13691 The loader for SPU does not handle dynamic relocations. By default, GCC
13692 will give an error when it generates code that requires a dynamic
13693 relocation. @option{-mno-error-reloc} disables the error,
13694 @option{-mwarn-reloc} will generate a warning instead.
13697 @itemx -munsafe-dma
13699 @opindex munsafe-dma
13701 Instructions which initiate or test completion of DMA must not be
13702 reordered with respect to loads and stores of the memory which is being
13703 accessed. Users typically address this problem using the volatile
13704 keyword, but that can lead to inefficient code in places where the
13705 memory is known to not change. Rather than mark the memory as volatile
13706 we treat the DMA instructions as potentially effecting all memory. With
13707 @option{-munsafe-dma} users must use the volatile keyword to protect
13710 @item -mbranch-hints
13711 @opindex mbranch-hints
13713 By default, GCC will generate a branch hint instruction to avoid
13714 pipeline stalls for always taken or probably taken branches. A hint
13715 will not be generated closer than 8 instructions away from its branch.
13716 There is little reason to disable them, except for debugging purposes,
13717 or to make an object a little bit smaller.
13721 @opindex msmall-mem
13722 @opindex mlarge-mem
13724 By default, GCC generates code assuming that addresses are never larger
13725 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
13726 a full 32 bit address.
13731 By default, GCC links against startup code that assumes the SPU-style
13732 main function interface (which has an unconventional parameter list).
13733 With @option{-mstdmain}, GCC will link your program against startup
13734 code that assumes a C99-style interface to @code{main}, including a
13735 local copy of @code{argv} strings.
13737 @item -mfixed-range=@var{register-range}
13738 @opindex mfixed-range
13739 Generate code treating the given register range as fixed registers.
13740 A fixed register is one that the register allocator can not use. This is
13741 useful when compiling kernel code. A register range is specified as
13742 two registers separated by a dash. Multiple register ranges can be
13743 specified separated by a comma.
13747 @node System V Options
13748 @subsection Options for System V
13750 These additional options are available on System V Release 4 for
13751 compatibility with other compilers on those systems:
13756 Create a shared object.
13757 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
13761 Identify the versions of each tool used by the compiler, in a
13762 @code{.ident} assembler directive in the output.
13766 Refrain from adding @code{.ident} directives to the output file (this is
13769 @item -YP,@var{dirs}
13771 Search the directories @var{dirs}, and no others, for libraries
13772 specified with @option{-l}.
13774 @item -Ym,@var{dir}
13776 Look in the directory @var{dir} to find the M4 preprocessor.
13777 The assembler uses this option.
13778 @c This is supposed to go with a -Yd for predefined M4 macro files, but
13779 @c the generic assembler that comes with Solaris takes just -Ym.
13782 @node TMS320C3x/C4x Options
13783 @subsection TMS320C3x/C4x Options
13784 @cindex TMS320C3x/C4x Options
13786 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
13790 @item -mcpu=@var{cpu_type}
13792 Set the instruction set, register set, and instruction scheduling
13793 parameters for machine type @var{cpu_type}. Supported values for
13794 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
13795 @samp{c44}. The default is @samp{c40} to generate code for the
13800 @itemx -msmall-memory
13802 @opindex mbig-memory
13804 @opindex msmall-memory
13806 Generates code for the big or small memory model. The small memory
13807 model assumed that all data fits into one 64K word page. At run-time
13808 the data page (DP) register must be set to point to the 64K page
13809 containing the .bss and .data program sections. The big memory model is
13810 the default and requires reloading of the DP register for every direct
13817 Allow (disallow) allocation of general integer operands into the block
13818 count register BK@.
13824 Enable (disable) generation of code using decrement and branch,
13825 DBcond(D), instructions. This is enabled by default for the C4x. To be
13826 on the safe side, this is disabled for the C3x, since the maximum
13827 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
13828 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
13829 that it can utilize the decrement and branch instruction, but will give
13830 up if there is more than one memory reference in the loop. Thus a loop
13831 where the loop counter is decremented can generate slightly more
13832 efficient code, in cases where the RPTB instruction cannot be utilized.
13834 @item -mdp-isr-reload
13836 @opindex mdp-isr-reload
13838 Force the DP register to be saved on entry to an interrupt service
13839 routine (ISR), reloaded to point to the data section, and restored on
13840 exit from the ISR@. This should not be required unless someone has
13841 violated the small memory model by modifying the DP register, say within
13848 For the C3x use the 24-bit MPYI instruction for integer multiplies
13849 instead of a library call to guarantee 32-bit results. Note that if one
13850 of the operands is a constant, then the multiplication will be performed
13851 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
13852 then squaring operations are performed inline instead of a library call.
13855 @itemx -mno-fast-fix
13857 @opindex mno-fast-fix
13858 The C3x/C4x FIX instruction to convert a floating point value to an
13859 integer value chooses the nearest integer less than or equal to the
13860 floating point value rather than to the nearest integer. Thus if the
13861 floating point number is negative, the result will be incorrectly
13862 truncated an additional code is necessary to detect and correct this
13863 case. This option can be used to disable generation of the additional
13864 code required to correct the result.
13870 Enable (disable) generation of repeat block sequences using the RPTB
13871 instruction for zero overhead looping. The RPTB construct is only used
13872 for innermost loops that do not call functions or jump across the loop
13873 boundaries. There is no advantage having nested RPTB loops due to the
13874 overhead required to save and restore the RC, RS, and RE registers.
13875 This is enabled by default with @option{-O2}.
13877 @item -mrpts=@var{count}
13881 Enable (disable) the use of the single instruction repeat instruction
13882 RPTS@. If a repeat block contains a single instruction, and the loop
13883 count can be guaranteed to be less than the value @var{count}, GCC will
13884 emit a RPTS instruction instead of a RPTB@. If no value is specified,
13885 then a RPTS will be emitted even if the loop count cannot be determined
13886 at compile time. Note that the repeated instruction following RPTS does
13887 not have to be reloaded from memory each iteration, thus freeing up the
13888 CPU buses for operands. However, since interrupts are blocked by this
13889 instruction, it is disabled by default.
13891 @item -mloop-unsigned
13892 @itemx -mno-loop-unsigned
13893 @opindex mloop-unsigned
13894 @opindex mno-loop-unsigned
13895 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
13896 is @math{2^{31} + 1} since these instructions test if the iteration count is
13897 negative to terminate the loop. If the iteration count is unsigned
13898 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
13899 exceeded. This switch allows an unsigned iteration count.
13903 Try to emit an assembler syntax that the TI assembler (asm30) is happy
13904 with. This also enforces compatibility with the API employed by the TI
13905 C3x C compiler. For example, long doubles are passed as structures
13906 rather than in floating point registers.
13912 Generate code that uses registers (stack) for passing arguments to functions.
13913 By default, arguments are passed in registers where possible rather
13914 than by pushing arguments on to the stack.
13916 @item -mparallel-insns
13917 @itemx -mno-parallel-insns
13918 @opindex mparallel-insns
13919 @opindex mno-parallel-insns
13920 Allow the generation of parallel instructions. This is enabled by
13921 default with @option{-O2}.
13923 @item -mparallel-mpy
13924 @itemx -mno-parallel-mpy
13925 @opindex mparallel-mpy
13926 @opindex mno-parallel-mpy
13927 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
13928 provided @option{-mparallel-insns} is also specified. These instructions have
13929 tight register constraints which can pessimize the code generation
13930 of large functions.
13935 @subsection V850 Options
13936 @cindex V850 Options
13938 These @samp{-m} options are defined for V850 implementations:
13942 @itemx -mno-long-calls
13943 @opindex mlong-calls
13944 @opindex mno-long-calls
13945 Treat all calls as being far away (near). If calls are assumed to be
13946 far away, the compiler will always load the functions address up into a
13947 register, and call indirect through the pointer.
13953 Do not optimize (do optimize) basic blocks that use the same index
13954 pointer 4 or more times to copy pointer into the @code{ep} register, and
13955 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
13956 option is on by default if you optimize.
13958 @item -mno-prolog-function
13959 @itemx -mprolog-function
13960 @opindex mno-prolog-function
13961 @opindex mprolog-function
13962 Do not use (do use) external functions to save and restore registers
13963 at the prologue and epilogue of a function. The external functions
13964 are slower, but use less code space if more than one function saves
13965 the same number of registers. The @option{-mprolog-function} option
13966 is on by default if you optimize.
13970 Try to make the code as small as possible. At present, this just turns
13971 on the @option{-mep} and @option{-mprolog-function} options.
13973 @item -mtda=@var{n}
13975 Put static or global variables whose size is @var{n} bytes or less into
13976 the tiny data area that register @code{ep} points to. The tiny data
13977 area can hold up to 256 bytes in total (128 bytes for byte references).
13979 @item -msda=@var{n}
13981 Put static or global variables whose size is @var{n} bytes or less into
13982 the small data area that register @code{gp} points to. The small data
13983 area can hold up to 64 kilobytes.
13985 @item -mzda=@var{n}
13987 Put static or global variables whose size is @var{n} bytes or less into
13988 the first 32 kilobytes of memory.
13992 Specify that the target processor is the V850.
13995 @opindex mbig-switch
13996 Generate code suitable for big switch tables. Use this option only if
13997 the assembler/linker complain about out of range branches within a switch
14002 This option will cause r2 and r5 to be used in the code generated by
14003 the compiler. This setting is the default.
14005 @item -mno-app-regs
14006 @opindex mno-app-regs
14007 This option will cause r2 and r5 to be treated as fixed registers.
14011 Specify that the target processor is the V850E1. The preprocessor
14012 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
14013 this option is used.
14017 Specify that the target processor is the V850E@. The preprocessor
14018 constant @samp{__v850e__} will be defined if this option is used.
14020 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
14021 are defined then a default target processor will be chosen and the
14022 relevant @samp{__v850*__} preprocessor constant will be defined.
14024 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
14025 defined, regardless of which processor variant is the target.
14027 @item -mdisable-callt
14028 @opindex mdisable-callt
14029 This option will suppress generation of the CALLT instruction for the
14030 v850e and v850e1 flavors of the v850 architecture. The default is
14031 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
14036 @subsection VAX Options
14037 @cindex VAX options
14039 These @samp{-m} options are defined for the VAX:
14044 Do not output certain jump instructions (@code{aobleq} and so on)
14045 that the Unix assembler for the VAX cannot handle across long
14050 Do output those jump instructions, on the assumption that you
14051 will assemble with the GNU assembler.
14055 Output code for g-format floating point numbers instead of d-format.
14058 @node VxWorks Options
14059 @subsection VxWorks Options
14060 @cindex VxWorks Options
14062 The options in this section are defined for all VxWorks targets.
14063 Options specific to the target hardware are listed with the other
14064 options for that target.
14069 GCC can generate code for both VxWorks kernels and real time processes
14070 (RTPs). This option switches from the former to the latter. It also
14071 defines the preprocessor macro @code{__RTP__}.
14074 @opindex non-static
14075 Link an RTP executable against shared libraries rather than static
14076 libraries. The options @option{-static} and @option{-shared} can
14077 also be used for RTPs (@pxref{Link Options}); @option{-static}
14084 These options are passed down to the linker. They are defined for
14085 compatibility with Diab.
14088 @opindex Xbind-lazy
14089 Enable lazy binding of function calls. This option is equivalent to
14090 @option{-Wl,-z,now} and is defined for compatibility with Diab.
14094 Disable lazy binding of function calls. This option is the default and
14095 is defined for compatibility with Diab.
14098 @node x86-64 Options
14099 @subsection x86-64 Options
14100 @cindex x86-64 options
14102 These are listed under @xref{i386 and x86-64 Options}.
14104 @node Xstormy16 Options
14105 @subsection Xstormy16 Options
14106 @cindex Xstormy16 Options
14108 These options are defined for Xstormy16:
14113 Choose startup files and linker script suitable for the simulator.
14116 @node Xtensa Options
14117 @subsection Xtensa Options
14118 @cindex Xtensa Options
14120 These options are supported for Xtensa targets:
14124 @itemx -mno-const16
14126 @opindex mno-const16
14127 Enable or disable use of @code{CONST16} instructions for loading
14128 constant values. The @code{CONST16} instruction is currently not a
14129 standard option from Tensilica. When enabled, @code{CONST16}
14130 instructions are always used in place of the standard @code{L32R}
14131 instructions. The use of @code{CONST16} is enabled by default only if
14132 the @code{L32R} instruction is not available.
14135 @itemx -mno-fused-madd
14136 @opindex mfused-madd
14137 @opindex mno-fused-madd
14138 Enable or disable use of fused multiply/add and multiply/subtract
14139 instructions in the floating-point option. This has no effect if the
14140 floating-point option is not also enabled. Disabling fused multiply/add
14141 and multiply/subtract instructions forces the compiler to use separate
14142 instructions for the multiply and add/subtract operations. This may be
14143 desirable in some cases where strict IEEE 754-compliant results are
14144 required: the fused multiply add/subtract instructions do not round the
14145 intermediate result, thereby producing results with @emph{more} bits of
14146 precision than specified by the IEEE standard. Disabling fused multiply
14147 add/subtract instructions also ensures that the program output is not
14148 sensitive to the compiler's ability to combine multiply and add/subtract
14151 @item -mtext-section-literals
14152 @itemx -mno-text-section-literals
14153 @opindex mtext-section-literals
14154 @opindex mno-text-section-literals
14155 Control the treatment of literal pools. The default is
14156 @option{-mno-text-section-literals}, which places literals in a separate
14157 section in the output file. This allows the literal pool to be placed
14158 in a data RAM/ROM, and it also allows the linker to combine literal
14159 pools from separate object files to remove redundant literals and
14160 improve code size. With @option{-mtext-section-literals}, the literals
14161 are interspersed in the text section in order to keep them as close as
14162 possible to their references. This may be necessary for large assembly
14165 @item -mtarget-align
14166 @itemx -mno-target-align
14167 @opindex mtarget-align
14168 @opindex mno-target-align
14169 When this option is enabled, GCC instructs the assembler to
14170 automatically align instructions to reduce branch penalties at the
14171 expense of some code density. The assembler attempts to widen density
14172 instructions to align branch targets and the instructions following call
14173 instructions. If there are not enough preceding safe density
14174 instructions to align a target, no widening will be performed. The
14175 default is @option{-mtarget-align}. These options do not affect the
14176 treatment of auto-aligned instructions like @code{LOOP}, which the
14177 assembler will always align, either by widening density instructions or
14178 by inserting no-op instructions.
14181 @itemx -mno-longcalls
14182 @opindex mlongcalls
14183 @opindex mno-longcalls
14184 When this option is enabled, GCC instructs the assembler to translate
14185 direct calls to indirect calls unless it can determine that the target
14186 of a direct call is in the range allowed by the call instruction. This
14187 translation typically occurs for calls to functions in other source
14188 files. Specifically, the assembler translates a direct @code{CALL}
14189 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
14190 The default is @option{-mno-longcalls}. This option should be used in
14191 programs where the call target can potentially be out of range. This
14192 option is implemented in the assembler, not the compiler, so the
14193 assembly code generated by GCC will still show direct call
14194 instructions---look at the disassembled object code to see the actual
14195 instructions. Note that the assembler will use an indirect call for
14196 every cross-file call, not just those that really will be out of range.
14199 @node zSeries Options
14200 @subsection zSeries Options
14201 @cindex zSeries options
14203 These are listed under @xref{S/390 and zSeries Options}.
14205 @node Code Gen Options
14206 @section Options for Code Generation Conventions
14207 @cindex code generation conventions
14208 @cindex options, code generation
14209 @cindex run-time options
14211 These machine-independent options control the interface conventions
14212 used in code generation.
14214 Most of them have both positive and negative forms; the negative form
14215 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
14216 one of the forms is listed---the one which is not the default. You
14217 can figure out the other form by either removing @samp{no-} or adding
14221 @item -fbounds-check
14222 @opindex fbounds-check
14223 For front-ends that support it, generate additional code to check that
14224 indices used to access arrays are within the declared range. This is
14225 currently only supported by the Java and Fortran front-ends, where
14226 this option defaults to true and false respectively.
14230 This option generates traps for signed overflow on addition, subtraction,
14231 multiplication operations.
14235 This option instructs the compiler to assume that signed arithmetic
14236 overflow of addition, subtraction and multiplication wraps around
14237 using twos-complement representation. This flag enables some optimizations
14238 and disables others. This option is enabled by default for the Java
14239 front-end, as required by the Java language specification.
14242 @opindex fexceptions
14243 Enable exception handling. Generates extra code needed to propagate
14244 exceptions. For some targets, this implies GCC will generate frame
14245 unwind information for all functions, which can produce significant data
14246 size overhead, although it does not affect execution. If you do not
14247 specify this option, GCC will enable it by default for languages like
14248 C++ which normally require exception handling, and disable it for
14249 languages like C that do not normally require it. However, you may need
14250 to enable this option when compiling C code that needs to interoperate
14251 properly with exception handlers written in C++. You may also wish to
14252 disable this option if you are compiling older C++ programs that don't
14253 use exception handling.
14255 @item -fnon-call-exceptions
14256 @opindex fnon-call-exceptions
14257 Generate code that allows trapping instructions to throw exceptions.
14258 Note that this requires platform-specific runtime support that does
14259 not exist everywhere. Moreover, it only allows @emph{trapping}
14260 instructions to throw exceptions, i.e.@: memory references or floating
14261 point instructions. It does not allow exceptions to be thrown from
14262 arbitrary signal handlers such as @code{SIGALRM}.
14264 @item -funwind-tables
14265 @opindex funwind-tables
14266 Similar to @option{-fexceptions}, except that it will just generate any needed
14267 static data, but will not affect the generated code in any other way.
14268 You will normally not enable this option; instead, a language processor
14269 that needs this handling would enable it on your behalf.
14271 @item -fasynchronous-unwind-tables
14272 @opindex fasynchronous-unwind-tables
14273 Generate unwind table in dwarf2 format, if supported by target machine. The
14274 table is exact at each instruction boundary, so it can be used for stack
14275 unwinding from asynchronous events (such as debugger or garbage collector).
14277 @item -fpcc-struct-return
14278 @opindex fpcc-struct-return
14279 Return ``short'' @code{struct} and @code{union} values in memory like
14280 longer ones, rather than in registers. This convention is less
14281 efficient, but it has the advantage of allowing intercallability between
14282 GCC-compiled files and files compiled with other compilers, particularly
14283 the Portable C Compiler (pcc).
14285 The precise convention for returning structures in memory depends
14286 on the target configuration macros.
14288 Short structures and unions are those whose size and alignment match
14289 that of some integer type.
14291 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
14292 switch is not binary compatible with code compiled with the
14293 @option{-freg-struct-return} switch.
14294 Use it to conform to a non-default application binary interface.
14296 @item -freg-struct-return
14297 @opindex freg-struct-return
14298 Return @code{struct} and @code{union} values in registers when possible.
14299 This is more efficient for small structures than
14300 @option{-fpcc-struct-return}.
14302 If you specify neither @option{-fpcc-struct-return} nor
14303 @option{-freg-struct-return}, GCC defaults to whichever convention is
14304 standard for the target. If there is no standard convention, GCC
14305 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
14306 the principal compiler. In those cases, we can choose the standard, and
14307 we chose the more efficient register return alternative.
14309 @strong{Warning:} code compiled with the @option{-freg-struct-return}
14310 switch is not binary compatible with code compiled with the
14311 @option{-fpcc-struct-return} switch.
14312 Use it to conform to a non-default application binary interface.
14314 @item -fshort-enums
14315 @opindex fshort-enums
14316 Allocate to an @code{enum} type only as many bytes as it needs for the
14317 declared range of possible values. Specifically, the @code{enum} type
14318 will be equivalent to the smallest integer type which has enough room.
14320 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
14321 code that is not binary compatible with code generated without that switch.
14322 Use it to conform to a non-default application binary interface.
14324 @item -fshort-double
14325 @opindex fshort-double
14326 Use the same size for @code{double} as for @code{float}.
14328 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
14329 code that is not binary compatible with code generated without that switch.
14330 Use it to conform to a non-default application binary interface.
14332 @item -fshort-wchar
14333 @opindex fshort-wchar
14334 Override the underlying type for @samp{wchar_t} to be @samp{short
14335 unsigned int} instead of the default for the target. This option is
14336 useful for building programs to run under WINE@.
14338 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
14339 code that is not binary compatible with code generated without that switch.
14340 Use it to conform to a non-default application binary interface.
14343 @opindex fno-common
14344 In C, allocate even uninitialized global variables in the data section of the
14345 object file, rather than generating them as common blocks. This has the
14346 effect that if the same variable is declared (without @code{extern}) in
14347 two different compilations, you will get an error when you link them.
14348 The only reason this might be useful is if you wish to verify that the
14349 program will work on other systems which always work this way.
14353 Ignore the @samp{#ident} directive.
14355 @item -finhibit-size-directive
14356 @opindex finhibit-size-directive
14357 Don't output a @code{.size} assembler directive, or anything else that
14358 would cause trouble if the function is split in the middle, and the
14359 two halves are placed at locations far apart in memory. This option is
14360 used when compiling @file{crtstuff.c}; you should not need to use it
14363 @item -fverbose-asm
14364 @opindex fverbose-asm
14365 Put extra commentary information in the generated assembly code to
14366 make it more readable. This option is generally only of use to those
14367 who actually need to read the generated assembly code (perhaps while
14368 debugging the compiler itself).
14370 @option{-fno-verbose-asm}, the default, causes the
14371 extra information to be omitted and is useful when comparing two assembler
14374 @item -frecord-gcc-switches
14375 @opindex frecord-gcc-switches
14376 This switch causes the command line that was used to invoke the
14377 compiler to be recorded into the object file that is being created.
14378 This switch is only implemented on some targets and the exact format
14379 of the recording is target and binary file format dependent, but it
14380 usually takes the form of a section containing ASCII text. This
14381 switch is related to the @option{-fverbose-asm} switch, but that
14382 switch only records information in the assembler output file as
14383 comments, so it never reaches the object file.
14387 @cindex global offset table
14389 Generate position-independent code (PIC) suitable for use in a shared
14390 library, if supported for the target machine. Such code accesses all
14391 constant addresses through a global offset table (GOT)@. The dynamic
14392 loader resolves the GOT entries when the program starts (the dynamic
14393 loader is not part of GCC; it is part of the operating system). If
14394 the GOT size for the linked executable exceeds a machine-specific
14395 maximum size, you get an error message from the linker indicating that
14396 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
14397 instead. (These maximums are 8k on the SPARC and 32k
14398 on the m68k and RS/6000. The 386 has no such limit.)
14400 Position-independent code requires special support, and therefore works
14401 only on certain machines. For the 386, GCC supports PIC for System V
14402 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
14403 position-independent.
14405 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14410 If supported for the target machine, emit position-independent code,
14411 suitable for dynamic linking and avoiding any limit on the size of the
14412 global offset table. This option makes a difference on the m68k,
14413 PowerPC and SPARC@.
14415 Position-independent code requires special support, and therefore works
14416 only on certain machines.
14418 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14425 These options are similar to @option{-fpic} and @option{-fPIC}, but
14426 generated position independent code can be only linked into executables.
14427 Usually these options are used when @option{-pie} GCC option will be
14428 used during linking.
14430 @option{-fpie} and @option{-fPIE} both define the macros
14431 @code{__pie__} and @code{__PIE__}. The macros have the value 1
14432 for @option{-fpie} and 2 for @option{-fPIE}.
14434 @item -fno-jump-tables
14435 @opindex fno-jump-tables
14436 Do not use jump tables for switch statements even where it would be
14437 more efficient than other code generation strategies. This option is
14438 of use in conjunction with @option{-fpic} or @option{-fPIC} for
14439 building code which forms part of a dynamic linker and cannot
14440 reference the address of a jump table. On some targets, jump tables
14441 do not require a GOT and this option is not needed.
14443 @item -ffixed-@var{reg}
14445 Treat the register named @var{reg} as a fixed register; generated code
14446 should never refer to it (except perhaps as a stack pointer, frame
14447 pointer or in some other fixed role).
14449 @var{reg} must be the name of a register. The register names accepted
14450 are machine-specific and are defined in the @code{REGISTER_NAMES}
14451 macro in the machine description macro file.
14453 This flag does not have a negative form, because it specifies a
14456 @item -fcall-used-@var{reg}
14457 @opindex fcall-used
14458 Treat the register named @var{reg} as an allocable register that is
14459 clobbered by function calls. It may be allocated for temporaries or
14460 variables that do not live across a call. Functions compiled this way
14461 will not save and restore the register @var{reg}.
14463 It is an error to used this flag with the frame pointer or stack pointer.
14464 Use of this flag for other registers that have fixed pervasive roles in
14465 the machine's execution model will produce disastrous results.
14467 This flag does not have a negative form, because it specifies a
14470 @item -fcall-saved-@var{reg}
14471 @opindex fcall-saved
14472 Treat the register named @var{reg} as an allocable register saved by
14473 functions. It may be allocated even for temporaries or variables that
14474 live across a call. Functions compiled this way will save and restore
14475 the register @var{reg} if they use it.
14477 It is an error to used this flag with the frame pointer or stack pointer.
14478 Use of this flag for other registers that have fixed pervasive roles in
14479 the machine's execution model will produce disastrous results.
14481 A different sort of disaster will result from the use of this flag for
14482 a register in which function values may be returned.
14484 This flag does not have a negative form, because it specifies a
14487 @item -fpack-struct[=@var{n}]
14488 @opindex fpack-struct
14489 Without a value specified, pack all structure members together without
14490 holes. When a value is specified (which must be a small power of two), pack
14491 structure members according to this value, representing the maximum
14492 alignment (that is, objects with default alignment requirements larger than
14493 this will be output potentially unaligned at the next fitting location.
14495 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
14496 code that is not binary compatible with code generated without that switch.
14497 Additionally, it makes the code suboptimal.
14498 Use it to conform to a non-default application binary interface.
14500 @item -finstrument-functions
14501 @opindex finstrument-functions
14502 Generate instrumentation calls for entry and exit to functions. Just
14503 after function entry and just before function exit, the following
14504 profiling functions will be called with the address of the current
14505 function and its call site. (On some platforms,
14506 @code{__builtin_return_address} does not work beyond the current
14507 function, so the call site information may not be available to the
14508 profiling functions otherwise.)
14511 void __cyg_profile_func_enter (void *this_fn,
14513 void __cyg_profile_func_exit (void *this_fn,
14517 The first argument is the address of the start of the current function,
14518 which may be looked up exactly in the symbol table.
14520 This instrumentation is also done for functions expanded inline in other
14521 functions. The profiling calls will indicate where, conceptually, the
14522 inline function is entered and exited. This means that addressable
14523 versions of such functions must be available. If all your uses of a
14524 function are expanded inline, this may mean an additional expansion of
14525 code size. If you use @samp{extern inline} in your C code, an
14526 addressable version of such functions must be provided. (This is
14527 normally the case anyways, but if you get lucky and the optimizer always
14528 expands the functions inline, you might have gotten away without
14529 providing static copies.)
14531 A function may be given the attribute @code{no_instrument_function}, in
14532 which case this instrumentation will not be done. This can be used, for
14533 example, for the profiling functions listed above, high-priority
14534 interrupt routines, and any functions from which the profiling functions
14535 cannot safely be called (perhaps signal handlers, if the profiling
14536 routines generate output or allocate memory).
14538 @item -fstack-check
14539 @opindex fstack-check
14540 Generate code to verify that you do not go beyond the boundary of the
14541 stack. You should specify this flag if you are running in an
14542 environment with multiple threads, but only rarely need to specify it in
14543 a single-threaded environment since stack overflow is automatically
14544 detected on nearly all systems if there is only one stack.
14546 Note that this switch does not actually cause checking to be done; the
14547 operating system must do that. The switch causes generation of code
14548 to ensure that the operating system sees the stack being extended.
14550 @item -fstack-limit-register=@var{reg}
14551 @itemx -fstack-limit-symbol=@var{sym}
14552 @itemx -fno-stack-limit
14553 @opindex fstack-limit-register
14554 @opindex fstack-limit-symbol
14555 @opindex fno-stack-limit
14556 Generate code to ensure that the stack does not grow beyond a certain value,
14557 either the value of a register or the address of a symbol. If the stack
14558 would grow beyond the value, a signal is raised. For most targets,
14559 the signal is raised before the stack overruns the boundary, so
14560 it is possible to catch the signal without taking special precautions.
14562 For instance, if the stack starts at absolute address @samp{0x80000000}
14563 and grows downwards, you can use the flags
14564 @option{-fstack-limit-symbol=__stack_limit} and
14565 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
14566 of 128KB@. Note that this may only work with the GNU linker.
14568 @cindex aliasing of parameters
14569 @cindex parameters, aliased
14570 @item -fargument-alias
14571 @itemx -fargument-noalias
14572 @itemx -fargument-noalias-global
14573 @itemx -fargument-noalias-anything
14574 @opindex fargument-alias
14575 @opindex fargument-noalias
14576 @opindex fargument-noalias-global
14577 @opindex fargument-noalias-anything
14578 Specify the possible relationships among parameters and between
14579 parameters and global data.
14581 @option{-fargument-alias} specifies that arguments (parameters) may
14582 alias each other and may alias global storage.@*
14583 @option{-fargument-noalias} specifies that arguments do not alias
14584 each other, but may alias global storage.@*
14585 @option{-fargument-noalias-global} specifies that arguments do not
14586 alias each other and do not alias global storage.
14587 @option{-fargument-noalias-anything} specifies that arguments do not
14588 alias any other storage.
14590 Each language will automatically use whatever option is required by
14591 the language standard. You should not need to use these options yourself.
14593 @item -fleading-underscore
14594 @opindex fleading-underscore
14595 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
14596 change the way C symbols are represented in the object file. One use
14597 is to help link with legacy assembly code.
14599 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
14600 generate code that is not binary compatible with code generated without that
14601 switch. Use it to conform to a non-default application binary interface.
14602 Not all targets provide complete support for this switch.
14604 @item -ftls-model=@var{model}
14605 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
14606 The @var{model} argument should be one of @code{global-dynamic},
14607 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
14609 The default without @option{-fpic} is @code{initial-exec}; with
14610 @option{-fpic} the default is @code{global-dynamic}.
14612 @item -fvisibility=@var{default|internal|hidden|protected}
14613 @opindex fvisibility
14614 Set the default ELF image symbol visibility to the specified option---all
14615 symbols will be marked with this unless overridden within the code.
14616 Using this feature can very substantially improve linking and
14617 load times of shared object libraries, produce more optimized
14618 code, provide near-perfect API export and prevent symbol clashes.
14619 It is @strong{strongly} recommended that you use this in any shared objects
14622 Despite the nomenclature, @code{default} always means public ie;
14623 available to be linked against from outside the shared object.
14624 @code{protected} and @code{internal} are pretty useless in real-world
14625 usage so the only other commonly used option will be @code{hidden}.
14626 The default if @option{-fvisibility} isn't specified is
14627 @code{default}, i.e., make every
14628 symbol public---this causes the same behavior as previous versions of
14631 A good explanation of the benefits offered by ensuring ELF
14632 symbols have the correct visibility is given by ``How To Write
14633 Shared Libraries'' by Ulrich Drepper (which can be found at
14634 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
14635 solution made possible by this option to marking things hidden when
14636 the default is public is to make the default hidden and mark things
14637 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
14638 and @code{__attribute__ ((visibility("default")))} instead of
14639 @code{__declspec(dllexport)} you get almost identical semantics with
14640 identical syntax. This is a great boon to those working with
14641 cross-platform projects.
14643 For those adding visibility support to existing code, you may find
14644 @samp{#pragma GCC visibility} of use. This works by you enclosing
14645 the declarations you wish to set visibility for with (for example)
14646 @samp{#pragma GCC visibility push(hidden)} and
14647 @samp{#pragma GCC visibility pop}.
14648 Bear in mind that symbol visibility should be viewed @strong{as
14649 part of the API interface contract} and thus all new code should
14650 always specify visibility when it is not the default ie; declarations
14651 only for use within the local DSO should @strong{always} be marked explicitly
14652 as hidden as so to avoid PLT indirection overheads---making this
14653 abundantly clear also aids readability and self-documentation of the code.
14654 Note that due to ISO C++ specification requirements, operator new and
14655 operator delete must always be of default visibility.
14657 Be aware that headers from outside your project, in particular system
14658 headers and headers from any other library you use, may not be
14659 expecting to be compiled with visibility other than the default. You
14660 may need to explicitly say @samp{#pragma GCC visibility push(default)}
14661 before including any such headers.
14663 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
14664 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
14665 no modifications. However, this means that calls to @samp{extern}
14666 functions with no explicit visibility will use the PLT, so it is more
14667 effective to use @samp{__attribute ((visibility))} and/or
14668 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
14669 declarations should be treated as hidden.
14671 Note that @samp{-fvisibility} does affect C++ vague linkage
14672 entities. This means that, for instance, an exception class that will
14673 be thrown between DSOs must be explicitly marked with default
14674 visibility so that the @samp{type_info} nodes will be unified between
14677 An overview of these techniques, their benefits and how to use them
14678 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
14684 @node Environment Variables
14685 @section Environment Variables Affecting GCC
14686 @cindex environment variables
14688 @c man begin ENVIRONMENT
14689 This section describes several environment variables that affect how GCC
14690 operates. Some of them work by specifying directories or prefixes to use
14691 when searching for various kinds of files. Some are used to specify other
14692 aspects of the compilation environment.
14694 Note that you can also specify places to search using options such as
14695 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
14696 take precedence over places specified using environment variables, which
14697 in turn take precedence over those specified by the configuration of GCC@.
14698 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
14699 GNU Compiler Collection (GCC) Internals}.
14704 @c @itemx LC_COLLATE
14706 @c @itemx LC_MONETARY
14707 @c @itemx LC_NUMERIC
14712 @c @findex LC_COLLATE
14713 @findex LC_MESSAGES
14714 @c @findex LC_MONETARY
14715 @c @findex LC_NUMERIC
14719 These environment variables control the way that GCC uses
14720 localization information that allow GCC to work with different
14721 national conventions. GCC inspects the locale categories
14722 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
14723 so. These locale categories can be set to any value supported by your
14724 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
14725 Kingdom encoded in UTF-8.
14727 The @env{LC_CTYPE} environment variable specifies character
14728 classification. GCC uses it to determine the character boundaries in
14729 a string; this is needed for some multibyte encodings that contain quote
14730 and escape characters that would otherwise be interpreted as a string
14733 The @env{LC_MESSAGES} environment variable specifies the language to
14734 use in diagnostic messages.
14736 If the @env{LC_ALL} environment variable is set, it overrides the value
14737 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
14738 and @env{LC_MESSAGES} default to the value of the @env{LANG}
14739 environment variable. If none of these variables are set, GCC
14740 defaults to traditional C English behavior.
14744 If @env{TMPDIR} is set, it specifies the directory to use for temporary
14745 files. GCC uses temporary files to hold the output of one stage of
14746 compilation which is to be used as input to the next stage: for example,
14747 the output of the preprocessor, which is the input to the compiler
14750 @item GCC_EXEC_PREFIX
14751 @findex GCC_EXEC_PREFIX
14752 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
14753 names of the subprograms executed by the compiler. No slash is added
14754 when this prefix is combined with the name of a subprogram, but you can
14755 specify a prefix that ends with a slash if you wish.
14757 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
14758 an appropriate prefix to use based on the pathname it was invoked with.
14760 If GCC cannot find the subprogram using the specified prefix, it
14761 tries looking in the usual places for the subprogram.
14763 The default value of @env{GCC_EXEC_PREFIX} is
14764 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
14765 the installed compiler. In many cases @var{prefix} is the value
14766 of @code{prefix} when you ran the @file{configure} script.
14768 Other prefixes specified with @option{-B} take precedence over this prefix.
14770 This prefix is also used for finding files such as @file{crt0.o} that are
14773 In addition, the prefix is used in an unusual way in finding the
14774 directories to search for header files. For each of the standard
14775 directories whose name normally begins with @samp{/usr/local/lib/gcc}
14776 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
14777 replacing that beginning with the specified prefix to produce an
14778 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
14779 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
14780 These alternate directories are searched first; the standard directories
14781 come next. If a standard directory begins with the configured
14782 @var{prefix} then the value of @var{prefix} is replaced by
14783 @env{GCC_EXEC_PREFIX} when looking for header files.
14785 @item COMPILER_PATH
14786 @findex COMPILER_PATH
14787 The value of @env{COMPILER_PATH} is a colon-separated list of
14788 directories, much like @env{PATH}. GCC tries the directories thus
14789 specified when searching for subprograms, if it can't find the
14790 subprograms using @env{GCC_EXEC_PREFIX}.
14793 @findex LIBRARY_PATH
14794 The value of @env{LIBRARY_PATH} is a colon-separated list of
14795 directories, much like @env{PATH}. When configured as a native compiler,
14796 GCC tries the directories thus specified when searching for special
14797 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
14798 using GCC also uses these directories when searching for ordinary
14799 libraries for the @option{-l} option (but directories specified with
14800 @option{-L} come first).
14804 @cindex locale definition
14805 This variable is used to pass locale information to the compiler. One way in
14806 which this information is used is to determine the character set to be used
14807 when character literals, string literals and comments are parsed in C and C++.
14808 When the compiler is configured to allow multibyte characters,
14809 the following values for @env{LANG} are recognized:
14813 Recognize JIS characters.
14815 Recognize SJIS characters.
14817 Recognize EUCJP characters.
14820 If @env{LANG} is not defined, or if it has some other value, then the
14821 compiler will use mblen and mbtowc as defined by the default locale to
14822 recognize and translate multibyte characters.
14826 Some additional environments variables affect the behavior of the
14829 @include cppenv.texi
14833 @node Precompiled Headers
14834 @section Using Precompiled Headers
14835 @cindex precompiled headers
14836 @cindex speed of compilation
14838 Often large projects have many header files that are included in every
14839 source file. The time the compiler takes to process these header files
14840 over and over again can account for nearly all of the time required to
14841 build the project. To make builds faster, GCC allows users to
14842 `precompile' a header file; then, if builds can use the precompiled
14843 header file they will be much faster.
14845 To create a precompiled header file, simply compile it as you would any
14846 other file, if necessary using the @option{-x} option to make the driver
14847 treat it as a C or C++ header file. You will probably want to use a
14848 tool like @command{make} to keep the precompiled header up-to-date when
14849 the headers it contains change.
14851 A precompiled header file will be searched for when @code{#include} is
14852 seen in the compilation. As it searches for the included file
14853 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
14854 compiler looks for a precompiled header in each directory just before it
14855 looks for the include file in that directory. The name searched for is
14856 the name specified in the @code{#include} with @samp{.gch} appended. If
14857 the precompiled header file can't be used, it is ignored.
14859 For instance, if you have @code{#include "all.h"}, and you have
14860 @file{all.h.gch} in the same directory as @file{all.h}, then the
14861 precompiled header file will be used if possible, and the original
14862 header will be used otherwise.
14864 Alternatively, you might decide to put the precompiled header file in a
14865 directory and use @option{-I} to ensure that directory is searched
14866 before (or instead of) the directory containing the original header.
14867 Then, if you want to check that the precompiled header file is always
14868 used, you can put a file of the same name as the original header in this
14869 directory containing an @code{#error} command.
14871 This also works with @option{-include}. So yet another way to use
14872 precompiled headers, good for projects not designed with precompiled
14873 header files in mind, is to simply take most of the header files used by
14874 a project, include them from another header file, precompile that header
14875 file, and @option{-include} the precompiled header. If the header files
14876 have guards against multiple inclusion, they will be skipped because
14877 they've already been included (in the precompiled header).
14879 If you need to precompile the same header file for different
14880 languages, targets, or compiler options, you can instead make a
14881 @emph{directory} named like @file{all.h.gch}, and put each precompiled
14882 header in the directory, perhaps using @option{-o}. It doesn't matter
14883 what you call the files in the directory, every precompiled header in
14884 the directory will be considered. The first precompiled header
14885 encountered in the directory that is valid for this compilation will
14886 be used; they're searched in no particular order.
14888 There are many other possibilities, limited only by your imagination,
14889 good sense, and the constraints of your build system.
14891 A precompiled header file can be used only when these conditions apply:
14895 Only one precompiled header can be used in a particular compilation.
14898 A precompiled header can't be used once the first C token is seen. You
14899 can have preprocessor directives before a precompiled header; you can
14900 even include a precompiled header from inside another header, so long as
14901 there are no C tokens before the @code{#include}.
14904 The precompiled header file must be produced for the same language as
14905 the current compilation. You can't use a C precompiled header for a C++
14909 The precompiled header file must have been produced by the same compiler
14910 binary as the current compilation is using.
14913 Any macros defined before the precompiled header is included must
14914 either be defined in the same way as when the precompiled header was
14915 generated, or must not affect the precompiled header, which usually
14916 means that they don't appear in the precompiled header at all.
14918 The @option{-D} option is one way to define a macro before a
14919 precompiled header is included; using a @code{#define} can also do it.
14920 There are also some options that define macros implicitly, like
14921 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
14924 @item If debugging information is output when using the precompiled
14925 header, using @option{-g} or similar, the same kind of debugging information
14926 must have been output when building the precompiled header. However,
14927 a precompiled header built using @option{-g} can be used in a compilation
14928 when no debugging information is being output.
14930 @item The same @option{-m} options must generally be used when building
14931 and using the precompiled header. @xref{Submodel Options},
14932 for any cases where this rule is relaxed.
14934 @item Each of the following options must be the same when building and using
14935 the precompiled header:
14937 @gccoptlist{-fexceptions -funit-at-a-time}
14940 Some other command-line options starting with @option{-f},
14941 @option{-p}, or @option{-O} must be defined in the same way as when
14942 the precompiled header was generated. At present, it's not clear
14943 which options are safe to change and which are not; the safest choice
14944 is to use exactly the same options when generating and using the
14945 precompiled header. The following are known to be safe:
14947 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
14948 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
14949 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
14954 For all of these except the last, the compiler will automatically
14955 ignore the precompiled header if the conditions aren't met. If you
14956 find an option combination that doesn't work and doesn't cause the
14957 precompiled header to be ignored, please consider filing a bug report,
14960 If you do use differing options when generating and using the
14961 precompiled header, the actual behavior will be a mixture of the
14962 behavior for the options. For instance, if you use @option{-g} to
14963 generate the precompiled header but not when using it, you may or may
14964 not get debugging information for routines in the precompiled header.
14966 @node Running Protoize
14967 @section Running Protoize
14969 The program @code{protoize} is an optional part of GCC@. You can use
14970 it to add prototypes to a program, thus converting the program to ISO
14971 C in one respect. The companion program @code{unprotoize} does the
14972 reverse: it removes argument types from any prototypes that are found.
14974 When you run these programs, you must specify a set of source files as
14975 command line arguments. The conversion programs start out by compiling
14976 these files to see what functions they define. The information gathered
14977 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
14979 After scanning comes actual conversion. The specified files are all
14980 eligible to be converted; any files they include (whether sources or
14981 just headers) are eligible as well.
14983 But not all the eligible files are converted. By default,
14984 @code{protoize} and @code{unprotoize} convert only source and header
14985 files in the current directory. You can specify additional directories
14986 whose files should be converted with the @option{-d @var{directory}}
14987 option. You can also specify particular files to exclude with the
14988 @option{-x @var{file}} option. A file is converted if it is eligible, its
14989 directory name matches one of the specified directory names, and its
14990 name within the directory has not been excluded.
14992 Basic conversion with @code{protoize} consists of rewriting most
14993 function definitions and function declarations to specify the types of
14994 the arguments. The only ones not rewritten are those for varargs
14997 @code{protoize} optionally inserts prototype declarations at the
14998 beginning of the source file, to make them available for any calls that
14999 precede the function's definition. Or it can insert prototype
15000 declarations with block scope in the blocks where undeclared functions
15003 Basic conversion with @code{unprotoize} consists of rewriting most
15004 function declarations to remove any argument types, and rewriting
15005 function definitions to the old-style pre-ISO form.
15007 Both conversion programs print a warning for any function declaration or
15008 definition that they can't convert. You can suppress these warnings
15011 The output from @code{protoize} or @code{unprotoize} replaces the
15012 original source file. The original file is renamed to a name ending
15013 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
15014 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
15015 for DOS) file already exists, then the source file is simply discarded.
15017 @code{protoize} and @code{unprotoize} both depend on GCC itself to
15018 scan the program and collect information about the functions it uses.
15019 So neither of these programs will work until GCC is installed.
15021 Here is a table of the options you can use with @code{protoize} and
15022 @code{unprotoize}. Each option works with both programs unless
15026 @item -B @var{directory}
15027 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
15028 usual directory (normally @file{/usr/local/lib}). This file contains
15029 prototype information about standard system functions. This option
15030 applies only to @code{protoize}.
15032 @item -c @var{compilation-options}
15033 Use @var{compilation-options} as the options when running @command{gcc} to
15034 produce the @samp{.X} files. The special option @option{-aux-info} is
15035 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
15037 Note that the compilation options must be given as a single argument to
15038 @code{protoize} or @code{unprotoize}. If you want to specify several
15039 @command{gcc} options, you must quote the entire set of compilation options
15040 to make them a single word in the shell.
15042 There are certain @command{gcc} arguments that you cannot use, because they
15043 would produce the wrong kind of output. These include @option{-g},
15044 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
15045 the @var{compilation-options}, they are ignored.
15048 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
15049 systems) instead of @samp{.c}. This is convenient if you are converting
15050 a C program to C++. This option applies only to @code{protoize}.
15053 Add explicit global declarations. This means inserting explicit
15054 declarations at the beginning of each source file for each function
15055 that is called in the file and was not declared. These declarations
15056 precede the first function definition that contains a call to an
15057 undeclared function. This option applies only to @code{protoize}.
15059 @item -i @var{string}
15060 Indent old-style parameter declarations with the string @var{string}.
15061 This option applies only to @code{protoize}.
15063 @code{unprotoize} converts prototyped function definitions to old-style
15064 function definitions, where the arguments are declared between the
15065 argument list and the initial @samp{@{}. By default, @code{unprotoize}
15066 uses five spaces as the indentation. If you want to indent with just
15067 one space instead, use @option{-i " "}.
15070 Keep the @samp{.X} files. Normally, they are deleted after conversion
15074 Add explicit local declarations. @code{protoize} with @option{-l} inserts
15075 a prototype declaration for each function in each block which calls the
15076 function without any declaration. This option applies only to
15080 Make no real changes. This mode just prints information about the conversions
15081 that would have been done without @option{-n}.
15084 Make no @samp{.save} files. The original files are simply deleted.
15085 Use this option with caution.
15087 @item -p @var{program}
15088 Use the program @var{program} as the compiler. Normally, the name
15089 @file{gcc} is used.
15092 Work quietly. Most warnings are suppressed.
15095 Print the version number, just like @option{-v} for @command{gcc}.
15098 If you need special compiler options to compile one of your program's
15099 source files, then you should generate that file's @samp{.X} file
15100 specially, by running @command{gcc} on that source file with the
15101 appropriate options and the option @option{-aux-info}. Then run
15102 @code{protoize} on the entire set of files. @code{protoize} will use
15103 the existing @samp{.X} file because it is newer than the source file.
15107 gcc -Dfoo=bar file1.c -aux-info file1.X
15112 You need to include the special files along with the rest in the
15113 @code{protoize} command, even though their @samp{.X} files already
15114 exist, because otherwise they won't get converted.
15116 @xref{Protoize Caveats}, for more information on how to use
15117 @code{protoize} successfully.