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 --target-help --version @@@var{file}}
166 @item C Language Options
167 @xref{C Dialect Options,,Options Controlling C Dialect}.
168 @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol
169 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
170 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
171 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
172 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
173 -fsigned-bitfields -fsigned-char @gol
174 -funsigned-bitfields -funsigned-char}
176 @item C++ Language Options
177 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
178 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
179 -fconserve-space -ffriend-injection @gol
180 -fno-elide-constructors @gol
181 -fno-enforce-eh-specs @gol
182 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
183 -fno-implicit-templates @gol
184 -fno-implicit-inline-templates @gol
185 -fno-implement-inlines -fms-extensions @gol
186 -fno-nonansi-builtins -fno-operator-names @gol
187 -fno-optional-diags -fpermissive @gol
188 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
189 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
190 -fno-default-inline -fvisibility-inlines-hidden @gol
191 -Wabi -Wctor-dtor-privacy @gol
192 -Wnon-virtual-dtor -Wreorder @gol
193 -Weffc++ -Wno-deprecated -Wstrict-null-sentinel @gol
194 -Wno-non-template-friend -Wold-style-cast @gol
195 -Woverloaded-virtual -Wno-pmf-conversions @gol
198 @item Objective-C and Objective-C++ Language Options
199 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
200 Objective-C and Objective-C++ Dialects}.
201 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
202 -fgnu-runtime -fnext-runtime @gol
203 -fno-nil-receivers @gol
204 -fobjc-call-cxx-cdtors @gol
205 -fobjc-direct-dispatch @gol
206 -fobjc-exceptions @gol
208 -freplace-objc-classes @gol
211 -Wassign-intercept @gol
212 -Wno-protocol -Wselector @gol
213 -Wstrict-selector-match @gol
214 -Wundeclared-selector}
216 @item Language Independent Options
217 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
218 @gccoptlist{-fmessage-length=@var{n} @gol
219 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
220 -fdiagnostics-show-option}
222 @item Warning Options
223 @xref{Warning Options,,Options to Request or Suppress Warnings}.
224 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
225 -w -Wextra -Wall -Waggregate-return -Walways-true -Warray-bounds @gol
226 -Wno-attributes -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
227 -Wchar-subscripts -Wclobbered -Wcomment @gol
228 -Wconversion -Wcoverage-mismatch -Wno-deprecated-declarations @gol
229 -Wdisabled-optimization -Wno-div-by-zero @gol
230 -Wempty-body -Wno-endif-labels @gol
231 -Werror -Werror-* -Werror-implicit-function-declaration @gol
232 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
233 -Wno-format-extra-args -Wformat-nonliteral @gol
234 -Wformat-security -Wformat-y2k @gol
235 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
236 -Wimport -Wno-import -Winit-self -Winline @gol
237 -Wno-int-to-pointer-cast @gol
238 -Wno-invalid-offsetof -Winvalid-pch @gol
239 -Wlarger-than-@var{len} -Wunsafe-loop-optimizations -Wlong-long @gol
240 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
241 -Wmissing-format-attribute -Wmissing-include-dirs @gol
242 -Wmissing-noreturn @gol
243 -Wno-multichar -Wnonnull -Wno-overflow @gol
244 -Woverlength-strings -Wpacked -Wpadded @gol
245 -Wparentheses -Wpointer-arith -Wno-pointer-to-int-cast @gol
246 -Wredundant-decls @gol
247 -Wreturn-type -Wsequence-point -Wshadow @gol
248 -Wsign-compare -Wstack-protector @gol
249 -Wstrict-aliasing -Wstrict-aliasing=2 @gol
250 -Wstring-literal-comparison @gol
251 -Wswitch -Wswitch-default -Wswitch-enum @gol
252 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
253 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
254 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
255 -Wunused-value -Wunused-variable -Wvariadic-macros @gol
256 -Wvolatile-register-var -Wwrite-strings}
258 @item C-only Warning Options
259 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
260 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
261 -Wold-style-declaration -Wold-style-definition @gol
262 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
263 -Wdeclaration-after-statement -Wpointer-sign}
265 @item Debugging Options
266 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
267 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
268 -fdump-noaddr -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
269 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
270 -fdump-ipa-all -fdump-ipa-cgraph @gol
272 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
273 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
274 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
275 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
277 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
278 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
279 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
280 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
281 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
282 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
283 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
284 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
285 -fdump-tree-nrv -fdump-tree-vect @gol
286 -fdump-tree-sink @gol
287 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
288 -fdump-tree-salias @gol
289 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
290 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
291 -ftree-vectorizer-verbose=@var{n} @gol
292 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
293 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
294 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
295 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
296 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
297 -ftest-coverage -ftime-report -fvar-tracking @gol
298 -g -g@var{level} -gcoff -gdwarf-2 @gol
299 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
300 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
301 -print-multi-directory -print-multi-lib @gol
302 -print-prog-name=@var{program} -print-search-dirs -Q @gol
305 @item Optimization Options
306 @xref{Optimize Options,,Options that Control Optimization}.
307 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
308 -falign-labels=@var{n} -falign-loops=@var{n} @gol
309 -fbounds-check -fmudflap -fmudflapth -fmudflapir @gol
310 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
311 -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol
312 -fcaller-saves -fcprop-registers -fcse-follow-jumps @gol
313 -fcse-skip-blocks -fcx-limited-range -fdata-sections @gol
314 -fdelayed-branch -fdelete-null-pointer-checks -fearly-inlining @gol
315 -fexpensive-optimizations -ffast-math -ffloat-store @gol
316 -fforce-addr -fforward-propagate -ffunction-sections @gol
317 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -fgcse-after-reload @gol
318 -fcrossjumping -fif-conversion -fif-conversion2 @gol
319 -finline-functions -finline-functions-called-once @gol
320 -finline-limit=@var{n} -fkeep-inline-functions @gol
321 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
322 -fmodulo-sched -fno-branch-count-reg @gol
323 -fno-default-inline -fno-defer-pop -fmove-loop-invariants @gol
324 -fno-function-cse -fno-guess-branch-probability @gol
325 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
326 -funsafe-math-optimizations -funsafe-loop-optimizations @gol
327 -ffinite-math-only -fno-signed-zeros @gol
328 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
329 -fomit-frame-pointer -foptimize-register-move @gol
330 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
331 -fprofile-generate -fprofile-use @gol
332 -fregmove -frename-registers @gol
333 -freorder-blocks -freorder-blocks-and-partition -freorder-functions @gol
334 -frerun-cse-after-loop @gol
335 -frounding-math -frtl-abstract-sequences @gol
336 -fschedule-insns -fschedule-insns2 @gol
337 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
338 -fsched-spec-load-dangerous @gol
339 -fsched-stalled-insns=@var{n} -fsched-stalled-insns-dep=@var{n} @gol
340 -fsched2-use-superblocks @gol
341 -fsched2-use-traces -fsee -freschedule-modulo-scheduled-loops @gol
342 -fsection-anchors -fsignaling-nans -fsingle-precision-constant @gol
343 -fstack-protector -fstack-protector-all @gol
344 -fstrict-aliasing -fstrict-overflow -ftracer -fthread-jumps @gol
345 -funroll-all-loops -funroll-loops -fpeel-loops @gol
346 -fsplit-ivs-in-unroller -funswitch-loops @gol
347 -fvariable-expansion-in-unroller @gol
348 -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol
349 -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts @gol
350 -ftree-dominator-opts -ftree-dse -ftree-copyrename -ftree-sink @gol
351 -ftree-ch -ftree-sra -ftree-ter -ftree-lrs -ftree-fre -ftree-vectorize @gol
352 -ftree-vect-loop-version -ftree-salias -fipa-pta -fweb @gol
353 -ftree-copy-prop -ftree-store-ccp -ftree-store-copy-prop -fwhole-program @gol
354 --param @var{name}=@var{value}
355 -O -O0 -O1 -O2 -O3 -Os}
357 @item Preprocessor Options
358 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
359 @gccoptlist{-A@var{question}=@var{answer} @gol
360 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
361 -C -dD -dI -dM -dN @gol
362 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
363 -idirafter @var{dir} @gol
364 -include @var{file} -imacros @var{file} @gol
365 -iprefix @var{file} -iwithprefix @var{dir} @gol
366 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
367 -imultilib @var{dir} -isysroot @var{dir} @gol
368 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
369 -P -fworking-directory -remap @gol
370 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
371 -Xpreprocessor @var{option}}
373 @item Assembler Option
374 @xref{Assembler Options,,Passing Options to the Assembler}.
375 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
378 @xref{Link Options,,Options for Linking}.
379 @gccoptlist{@var{object-file-name} -l@var{library} @gol
380 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
381 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
382 -Wl,@var{option} -Xlinker @var{option} @gol
385 @item Directory Options
386 @xref{Directory Options,,Options for Directory Search}.
387 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
388 -specs=@var{file} -I- --sysroot=@var{dir}}
391 @c I wrote this xref this way to avoid overfull hbox. -- rms
392 @xref{Target Options}.
393 @gccoptlist{-V @var{version} -b @var{machine}}
395 @item Machine Dependent Options
396 @xref{Submodel Options,,Hardware Models and Configurations}.
397 @c This list is ordered alphanumerically by subsection name.
398 @c Try and put the significant identifier (CPU or system) first,
399 @c so users have a clue at guessing where the ones they want will be.
402 @gccoptlist{-EB -EL @gol
403 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
404 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
407 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
408 -mabi=@var{name} @gol
409 -mapcs-stack-check -mno-apcs-stack-check @gol
410 -mapcs-float -mno-apcs-float @gol
411 -mapcs-reentrant -mno-apcs-reentrant @gol
412 -msched-prolog -mno-sched-prolog @gol
413 -mlittle-endian -mbig-endian -mwords-little-endian @gol
414 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
415 -mthumb-interwork -mno-thumb-interwork @gol
416 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
417 -mstructure-size-boundary=@var{n} @gol
418 -mabort-on-noreturn @gol
419 -mlong-calls -mno-long-calls @gol
420 -msingle-pic-base -mno-single-pic-base @gol
421 -mpic-register=@var{reg} @gol
422 -mnop-fun-dllimport @gol
423 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
424 -mpoke-function-name @gol
426 -mtpcs-frame -mtpcs-leaf-frame @gol
427 -mcaller-super-interworking -mcallee-super-interworking @gol
431 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
432 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
434 @emph{Blackfin Options}
435 @gccoptlist{-momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
436 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
437 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
438 -mno-id-shared-library -mshared-library-id=@var{n} @gol
439 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
440 -msep-data -mno-sep-data -mlong-calls -mno-long-calls}
443 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
444 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
445 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
446 -mstack-align -mdata-align -mconst-align @gol
447 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
448 -melf -maout -melinux -mlinux -sim -sim2 @gol
449 -mmul-bug-workaround -mno-mul-bug-workaround}
452 @gccoptlist{-mmac -mpush-args}
454 @emph{Darwin Options}
455 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
456 -arch_only -bind_at_load -bundle -bundle_loader @gol
457 -client_name -compatibility_version -current_version @gol
459 -dependency-file -dylib_file -dylinker_install_name @gol
460 -dynamic -dynamiclib -exported_symbols_list @gol
461 -filelist -flat_namespace -force_cpusubtype_ALL @gol
462 -force_flat_namespace -headerpad_max_install_names @gol
463 -image_base -init -install_name -keep_private_externs @gol
464 -multi_module -multiply_defined -multiply_defined_unused @gol
465 -noall_load -no_dead_strip_inits_and_terms @gol
466 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
467 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
468 -private_bundle -read_only_relocs -sectalign @gol
469 -sectobjectsymbols -whyload -seg1addr @gol
470 -sectcreate -sectobjectsymbols -sectorder @gol
471 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
472 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
473 -segprot -segs_read_only_addr -segs_read_write_addr @gol
474 -single_module -static -sub_library -sub_umbrella @gol
475 -twolevel_namespace -umbrella -undefined @gol
476 -unexported_symbols_list -weak_reference_mismatches @gol
477 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
478 -mkernel -mone-byte-bool}
480 @emph{DEC Alpha Options}
481 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
482 -mieee -mieee-with-inexact -mieee-conformant @gol
483 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
484 -mtrap-precision=@var{mode} -mbuild-constants @gol
485 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
486 -mbwx -mmax -mfix -mcix @gol
487 -mfloat-vax -mfloat-ieee @gol
488 -mexplicit-relocs -msmall-data -mlarge-data @gol
489 -msmall-text -mlarge-text @gol
490 -mmemory-latency=@var{time}}
492 @emph{DEC Alpha/VMS Options}
493 @gccoptlist{-mvms-return-codes}
496 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
497 -mhard-float -msoft-float @gol
498 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
499 -mdouble -mno-double @gol
500 -mmedia -mno-media -mmuladd -mno-muladd @gol
501 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
502 -mlinked-fp -mlong-calls -malign-labels @gol
503 -mlibrary-pic -macc-4 -macc-8 @gol
504 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
505 -moptimize-membar -mno-optimize-membar @gol
506 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
507 -mvliw-branch -mno-vliw-branch @gol
508 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
509 -mno-nested-cond-exec -mtomcat-stats @gol
513 @emph{GNU/Linux Options}
514 @gccoptlist{-muclibc}
516 @emph{H8/300 Options}
517 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
520 @gccoptlist{-march=@var{architecture-type} @gol
521 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
522 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
523 -mfixed-range=@var{register-range} @gol
524 -mjump-in-delay -mlinker-opt -mlong-calls @gol
525 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
526 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
527 -mno-jump-in-delay -mno-long-load-store @gol
528 -mno-portable-runtime -mno-soft-float @gol
529 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
530 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
531 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
532 -munix=@var{unix-std} -nolibdld -static -threads}
534 @emph{i386 and x86-64 Options}
535 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
536 -mfpmath=@var{unit} @gol
537 -masm=@var{dialect} -mno-fancy-math-387 @gol
538 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
539 -mno-wide-multiply -mrtd -malign-double @gol
540 -mpreferred-stack-boundary=@var{num} @gol
541 -mmmx -msse -msse2 -msse3 -mssse3 -m3dnow @gol
542 -mthreads -mno-align-stringops -minline-all-stringops @gol
543 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
544 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
546 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
547 -mcmodel=@var{code-model} @gol
548 -m32 -m64 -mlarge-data-threshold=@var{num}}
551 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
552 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
553 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
554 -minline-float-divide-max-throughput @gol
555 -minline-int-divide-min-latency @gol
556 -minline-int-divide-max-throughput @gol
557 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
558 -mno-dwarf2-asm -mearly-stop-bits @gol
559 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
560 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64 @gol
561 -mno-sched-br-data-spec -msched-ar-data-spec -mno-sched-control-spec @gol
562 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
563 -msched-ldc -mno-sched-control-ldc -mno-sched-spec-verbose @gol
564 -mno-sched-prefer-non-data-spec-insns @gol
565 -mno-sched-prefer-non-control-spec-insns @gol
566 -mno-sched-count-spec-in-critical-path}
568 @emph{M32R/D Options}
569 @gccoptlist{-m32r2 -m32rx -m32r @gol
571 -malign-loops -mno-align-loops @gol
572 -missue-rate=@var{number} @gol
573 -mbranch-cost=@var{number} @gol
574 -mmodel=@var{code-size-model-type} @gol
575 -msdata=@var{sdata-type} @gol
576 -mno-flush-func -mflush-func=@var{name} @gol
577 -mno-flush-trap -mflush-trap=@var{number} @gol
581 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
583 @emph{M680x0 Options}
584 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
585 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
586 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
587 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
588 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
589 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
590 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
591 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
593 @emph{M68hc1x Options}
594 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
595 -mauto-incdec -minmax -mlong-calls -mshort @gol
596 -msoft-reg-count=@var{count}}
599 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
600 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
601 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
602 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
603 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
606 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
607 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
608 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
609 -mshared -mno-shared -mxgot -mno-xgot -mgp32 -mgp64 @gol
610 -mfp32 -mfp64 -mhard-float -msoft-float @gol
611 -msingle-float -mdouble-float -mdsp -mpaired-single -mips3d @gol
612 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
613 -G@var{num} -membedded-data -mno-embedded-data @gol
614 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
615 -msplit-addresses -mno-split-addresses @gol
616 -mexplicit-relocs -mno-explicit-relocs @gol
617 -mcheck-zero-division -mno-check-zero-division @gol
618 -mdivide-traps -mdivide-breaks @gol
619 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
620 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
621 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
622 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 @gol
623 -mfix-sb1 -mno-fix-sb1 @gol
624 -mflush-func=@var{func} -mno-flush-func @gol
625 -mbranch-likely -mno-branch-likely @gol
626 -mfp-exceptions -mno-fp-exceptions @gol
627 -mvr4130-align -mno-vr4130-align}
630 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
631 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
632 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
633 -mno-base-addresses -msingle-exit -mno-single-exit}
635 @emph{MN10300 Options}
636 @gccoptlist{-mmult-bug -mno-mult-bug @gol
637 -mam33 -mno-am33 @gol
638 -mam33-2 -mno-am33-2 @gol
639 -mreturn-pointer-on-d0 @gol
643 @gccoptlist{-mno-crt0 -mbacc -msim @gol
644 -march=@var{cpu-type} }
646 @emph{PDP-11 Options}
647 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
648 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
649 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
650 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
651 -mbranch-expensive -mbranch-cheap @gol
652 -msplit -mno-split -munix-asm -mdec-asm}
654 @emph{PowerPC Options}
655 See RS/6000 and PowerPC Options.
657 @emph{RS/6000 and PowerPC Options}
658 @gccoptlist{-mcpu=@var{cpu-type} @gol
659 -mtune=@var{cpu-type} @gol
660 -mpower -mno-power -mpower2 -mno-power2 @gol
661 -mpowerpc -mpowerpc64 -mno-powerpc @gol
662 -maltivec -mno-altivec @gol
663 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
664 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
665 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
666 -mmfpgpr -mno-mfpgpr @gol
667 -mnew-mnemonics -mold-mnemonics @gol
668 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
669 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
670 -malign-power -malign-natural @gol
671 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
672 -mstring -mno-string -mupdate -mno-update @gol
673 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
674 -mstrict-align -mno-strict-align -mrelocatable @gol
675 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
676 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
677 -mdynamic-no-pic -maltivec -mswdiv @gol
678 -mprioritize-restricted-insns=@var{priority} @gol
679 -msched-costly-dep=@var{dependence_type} @gol
680 -minsert-sched-nops=@var{scheme} @gol
681 -mcall-sysv -mcall-netbsd @gol
682 -maix-struct-return -msvr4-struct-return @gol
683 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
684 -misel -mno-isel @gol
685 -misel=yes -misel=no @gol
687 -mspe=yes -mspe=no @gol
688 -mvrsave -mno-vrsave @gol
689 -mmulhw -mno-mulhw @gol
690 -mdlmzb -mno-dlmzb @gol
691 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
692 -mprototype -mno-prototype @gol
693 -msim -mmvme -mads -myellowknife -memb -msdata @gol
694 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
696 @emph{S/390 and zSeries Options}
697 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
698 -mhard-float -msoft-float -mlong-double-64 -mlong-double-128 @gol
699 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
700 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
701 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
702 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
703 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
706 @gccoptlist{-mel -mel @gol
711 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
712 -m4-nofpu -m4-single-only -m4-single -m4 @gol
713 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
714 -m5-64media -m5-64media-nofpu @gol
715 -m5-32media -m5-32media-nofpu @gol
716 -m5-compact -m5-compact-nofpu @gol
717 -mb -ml -mdalign -mrelax @gol
718 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
719 -mieee -misize -minline-ic_invalidate -mpadstruct -mspace @gol
720 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
721 -mdivsi3_libfunc=@var{name} @gol
722 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
726 @gccoptlist{-mcpu=@var{cpu-type} @gol
727 -mtune=@var{cpu-type} @gol
728 -mcmodel=@var{code-model} @gol
729 -m32 -m64 -mapp-regs -mno-app-regs @gol
730 -mfaster-structs -mno-faster-structs @gol
731 -mfpu -mno-fpu -mhard-float -msoft-float @gol
732 -mhard-quad-float -msoft-quad-float @gol
733 -mimpure-text -mno-impure-text -mlittle-endian @gol
734 -mstack-bias -mno-stack-bias @gol
735 -munaligned-doubles -mno-unaligned-doubles @gol
736 -mv8plus -mno-v8plus -mvis -mno-vis
737 -threads -pthreads -pthread}
740 @gccoptlist{-mwarn-reloc -merror-reloc @gol
741 -msafe-dma -munsafe-dma @gol
743 -msmall-mem -mlarge-mem -mstdmain @gol
744 -mfixed-range=@var{register-range}}
746 @emph{System V Options}
747 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
749 @emph{TMS320C3x/C4x Options}
750 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
751 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
752 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
753 -mparallel-insns -mparallel-mpy -mpreserve-float}
756 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
757 -mprolog-function -mno-prolog-function -mspace @gol
758 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
759 -mapp-regs -mno-app-regs @gol
760 -mdisable-callt -mno-disable-callt @gol
766 @gccoptlist{-mg -mgnu -munix}
768 @emph{x86-64 Options}
769 See i386 and x86-64 Options.
771 @emph{Xstormy16 Options}
774 @emph{Xtensa Options}
775 @gccoptlist{-mconst16 -mno-const16 @gol
776 -mfused-madd -mno-fused-madd @gol
777 -mtext-section-literals -mno-text-section-literals @gol
778 -mtarget-align -mno-target-align @gol
779 -mlongcalls -mno-longcalls}
781 @emph{zSeries Options}
782 See S/390 and zSeries Options.
784 @item Code Generation Options
785 @xref{Code Gen Options,,Options for Code Generation Conventions}.
786 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
787 -ffixed-@var{reg} -fexceptions @gol
788 -fnon-call-exceptions -funwind-tables @gol
789 -fasynchronous-unwind-tables @gol
790 -finhibit-size-directive -finstrument-functions @gol
791 -fno-common -fno-ident @gol
792 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
793 -fno-jump-tables @gol
794 -frecord-gcc-switches @gol
795 -freg-struct-return -fshort-enums @gol
796 -fshort-double -fshort-wchar @gol
797 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
798 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
799 -fargument-alias -fargument-noalias @gol
800 -fargument-noalias-global -fargument-noalias-anything
801 -fleading-underscore -ftls-model=@var{model} @gol
802 -ftrapv -fwrapv -fbounds-check @gol
807 * Overall Options:: Controlling the kind of output:
808 an executable, object files, assembler files,
809 or preprocessed source.
810 * C Dialect Options:: Controlling the variant of C language compiled.
811 * C++ Dialect Options:: Variations on C++.
812 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
814 * Language Independent Options:: Controlling how diagnostics should be
816 * Warning Options:: How picky should the compiler be?
817 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
818 * Optimize Options:: How much optimization?
819 * Preprocessor Options:: Controlling header files and macro definitions.
820 Also, getting dependency information for Make.
821 * Assembler Options:: Passing options to the assembler.
822 * Link Options:: Specifying libraries and so on.
823 * Directory Options:: Where to find header files and libraries.
824 Where to find the compiler executable files.
825 * Spec Files:: How to pass switches to sub-processes.
826 * Target Options:: Running a cross-compiler, or an old version of GCC.
829 @node Overall Options
830 @section Options Controlling the Kind of Output
832 Compilation can involve up to four stages: preprocessing, compilation
833 proper, assembly and linking, always in that order. GCC is capable of
834 preprocessing and compiling several files either into several
835 assembler input files, or into one assembler input file; then each
836 assembler input file produces an object file, and linking combines all
837 the object files (those newly compiled, and those specified as input)
838 into an executable file.
840 @cindex file name suffix
841 For any given input file, the file name suffix determines what kind of
846 C source code which must be preprocessed.
849 C source code which should not be preprocessed.
852 C++ source code which should not be preprocessed.
855 Objective-C source code. Note that you must link with the @file{libobjc}
856 library to make an Objective-C program work.
859 Objective-C source code which should not be preprocessed.
863 Objective-C++ source code. Note that you must link with the @file{libobjc}
864 library to make an Objective-C++ program work. Note that @samp{.M} refers
865 to a literal capital M@.
868 Objective-C++ source code which should not be preprocessed.
871 C, C++, Objective-C or Objective-C++ header file to be turned into a
876 @itemx @var{file}.cxx
877 @itemx @var{file}.cpp
878 @itemx @var{file}.CPP
879 @itemx @var{file}.c++
881 C++ source code which must be preprocessed. Note that in @samp{.cxx},
882 the last two letters must both be literally @samp{x}. Likewise,
883 @samp{.C} refers to a literal capital C@.
887 Objective-C++ source code which must be preprocessed.
890 Objective-C++ source code which should not be preprocessed.
894 C++ header file to be turned into a precompiled header.
897 @itemx @var{file}.for
898 @itemx @var{file}.FOR
899 Fixed form Fortran source code which should not be preprocessed.
902 @itemx @var{file}.fpp
903 @itemx @var{file}.FPP
904 Fixed form Fortran source code which must be preprocessed (with the traditional
908 @itemx @var{file}.f95
909 Free form Fortran source code which should not be preprocessed.
912 @itemx @var{file}.F95
913 Free form Fortran source code which must be preprocessed (with the
914 traditional preprocessor).
916 @c FIXME: Descriptions of Java file types.
923 Ada source code file which contains a library unit declaration (a
924 declaration of a package, subprogram, or generic, or a generic
925 instantiation), or a library unit renaming declaration (a package,
926 generic, or subprogram renaming declaration). Such files are also
929 @itemx @var{file}.adb
930 Ada source code file containing a library unit body (a subprogram or
931 package body). Such files are also called @dfn{bodies}.
933 @c GCC also knows about some suffixes for languages not yet included:
944 Assembler code which must be preprocessed.
947 An object file to be fed straight into linking.
948 Any file name with no recognized suffix is treated this way.
952 You can specify the input language explicitly with the @option{-x} option:
955 @item -x @var{language}
956 Specify explicitly the @var{language} for the following input files
957 (rather than letting the compiler choose a default based on the file
958 name suffix). This option applies to all following input files until
959 the next @option{-x} option. Possible values for @var{language} are:
961 c c-header c-cpp-output
962 c++ c++-header c++-cpp-output
963 objective-c objective-c-header objective-c-cpp-output
964 objective-c++ objective-c++-header objective-c++-cpp-output
965 assembler assembler-with-cpp
973 Turn off any specification of a language, so that subsequent files are
974 handled according to their file name suffixes (as they are if @option{-x}
975 has not been used at all).
977 @item -pass-exit-codes
978 @opindex pass-exit-codes
979 Normally the @command{gcc} program will exit with the code of 1 if any
980 phase of the compiler returns a non-success return code. If you specify
981 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
982 numerically highest error produced by any phase that returned an error
983 indication. The C, C++, and Fortran frontends return 4, if an internal
984 compiler error is encountered.
987 If you only want some of the stages of compilation, you can use
988 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
989 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
990 @command{gcc} is to stop. Note that some combinations (for example,
991 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
996 Compile or assemble the source files, but do not link. The linking
997 stage simply is not done. The ultimate output is in the form of an
998 object file for each source file.
1000 By default, the object file name for a source file is made by replacing
1001 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1003 Unrecognized input files, not requiring compilation or assembly, are
1008 Stop after the stage of compilation proper; do not assemble. The output
1009 is in the form of an assembler code file for each non-assembler input
1012 By default, the assembler file name for a source file is made by
1013 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1015 Input files that don't require compilation are ignored.
1019 Stop after the preprocessing stage; do not run the compiler proper. The
1020 output is in the form of preprocessed source code, which is sent to the
1023 Input files which don't require preprocessing are ignored.
1025 @cindex output file option
1028 Place output in file @var{file}. This applies regardless to whatever
1029 sort of output is being produced, whether it be an executable file,
1030 an object file, an assembler file or preprocessed C code.
1032 If @option{-o} is not specified, the default is to put an executable
1033 file in @file{a.out}, the object file for
1034 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1035 assembler file in @file{@var{source}.s}, a precompiled header file in
1036 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1041 Print (on standard error output) the commands executed to run the stages
1042 of compilation. Also print the version number of the compiler driver
1043 program and of the preprocessor and the compiler proper.
1047 Like @option{-v} except the commands are not executed and all command
1048 arguments are quoted. This is useful for shell scripts to capture the
1049 driver-generated command lines.
1053 Use pipes rather than temporary files for communication between the
1054 various stages of compilation. This fails to work on some systems where
1055 the assembler is unable to read from a pipe; but the GNU assembler has
1060 If you are compiling multiple source files, this option tells the driver
1061 to pass all the source files to the compiler at once (for those
1062 languages for which the compiler can handle this). This will allow
1063 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1064 language for which this is supported is C@. If you pass source files for
1065 multiple languages to the driver, using this option, the driver will invoke
1066 the compiler(s) that support IMA once each, passing each compiler all the
1067 source files appropriate for it. For those languages that do not support
1068 IMA this option will be ignored, and the compiler will be invoked once for
1069 each source file in that language. If you use this option in conjunction
1070 with @option{-save-temps}, the compiler will generate multiple
1072 (one for each source file), but only one (combined) @file{.o} or
1077 Print (on the standard output) a description of the command line options
1078 understood by @command{gcc}. If the @option{-v} option is also specified
1079 then @option{--help} will also be passed on to the various processes
1080 invoked by @command{gcc}, so that they can display the command line options
1081 they accept. If the @option{-Wextra} option is also specified then command
1082 line options which have no documentation associated with them will also
1086 @opindex target-help
1087 Print (on the standard output) a description of target specific command
1088 line options for each tool.
1092 Display the version number and copyrights of the invoked GCC@.
1094 @include @value{srcdir}/../libiberty/at-file.texi
1098 @section Compiling C++ Programs
1100 @cindex suffixes for C++ source
1101 @cindex C++ source file suffixes
1102 C++ source files conventionally use one of the suffixes @samp{.C},
1103 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1104 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
1105 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1106 files with these names and compiles them as C++ programs even if you
1107 call the compiler the same way as for compiling C programs (usually
1108 with the name @command{gcc}).
1112 However, the use of @command{gcc} does not add the C++ library.
1113 @command{g++} is a program that calls GCC and treats @samp{.c},
1114 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1115 files unless @option{-x} is used, and automatically specifies linking
1116 against the C++ library. This program is also useful when
1117 precompiling a C header file with a @samp{.h} extension for use in C++
1118 compilations. On many systems, @command{g++} is also installed with
1119 the name @command{c++}.
1121 @cindex invoking @command{g++}
1122 When you compile C++ programs, you may specify many of the same
1123 command-line options that you use for compiling programs in any
1124 language; or command-line options meaningful for C and related
1125 languages; or options that are meaningful only for C++ programs.
1126 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1127 explanations of options for languages related to C@.
1128 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1129 explanations of options that are meaningful only for C++ programs.
1131 @node C Dialect Options
1132 @section Options Controlling C Dialect
1133 @cindex dialect options
1134 @cindex language dialect options
1135 @cindex options, dialect
1137 The following options control the dialect of C (or languages derived
1138 from C, such as C++, Objective-C and Objective-C++) that the compiler
1142 @cindex ANSI support
1146 In C mode, support all ISO C90 programs. In C++ mode,
1147 remove GNU extensions that conflict with ISO C++.
1149 This turns off certain features of GCC that are incompatible with ISO
1150 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1151 such as the @code{asm} and @code{typeof} keywords, and
1152 predefined macros such as @code{unix} and @code{vax} that identify the
1153 type of system you are using. It also enables the undesirable and
1154 rarely used ISO trigraph feature. For the C compiler,
1155 it disables recognition of C++ style @samp{//} comments as well as
1156 the @code{inline} keyword.
1158 The alternate keywords @code{__asm__}, @code{__extension__},
1159 @code{__inline__} and @code{__typeof__} continue to work despite
1160 @option{-ansi}. You would not want to use them in an ISO C program, of
1161 course, but it is useful to put them in header files that might be included
1162 in compilations done with @option{-ansi}. Alternate predefined macros
1163 such as @code{__unix__} and @code{__vax__} are also available, with or
1164 without @option{-ansi}.
1166 The @option{-ansi} option does not cause non-ISO programs to be
1167 rejected gratuitously. For that, @option{-pedantic} is required in
1168 addition to @option{-ansi}. @xref{Warning Options}.
1170 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1171 option is used. Some header files may notice this macro and refrain
1172 from declaring certain functions or defining certain macros that the
1173 ISO standard doesn't call for; this is to avoid interfering with any
1174 programs that might use these names for other things.
1176 Functions which would normally be built in but do not have semantics
1177 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1178 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1179 built-in functions provided by GCC}, for details of the functions
1184 Determine the language standard. This option is currently only
1185 supported when compiling C or C++. A value for this option must be
1186 provided; possible values are
1191 ISO C90 (same as @option{-ansi}).
1193 @item iso9899:199409
1194 ISO C90 as modified in amendment 1.
1200 ISO C99. Note that this standard is not yet fully supported; see
1201 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1202 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1205 Default, ISO C90 plus GNU extensions (including some C99 features).
1209 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1210 this will become the default. The name @samp{gnu9x} is deprecated.
1213 The 1998 ISO C++ standard plus amendments.
1216 The same as @option{-std=c++98} plus GNU extensions. This is the
1217 default for C++ code.
1220 The working draft of the upcoming ISO C++0x standard. This option
1221 enables experimental features that are likely to be included in
1222 C++0x. The working draft is constantly changing, and any feature that is
1223 enabled by this flag may be removed from future versions of GCC if it is
1224 not part of the C++0x standard.
1227 The same as @option{-std=c++0x} plus GNU extensions. As with
1228 @option{-std=c++0x}, this option enables experimental features that may
1229 be removed in future versions of GCC.
1232 Even when this option is not specified, you can still use some of the
1233 features of newer standards in so far as they do not conflict with
1234 previous C standards. For example, you may use @code{__restrict__} even
1235 when @option{-std=c99} is not specified.
1237 The @option{-std} options specifying some version of ISO C have the same
1238 effects as @option{-ansi}, except that features that were not in ISO C90
1239 but are in the specified version (for example, @samp{//} comments and
1240 the @code{inline} keyword in ISO C99) are not disabled.
1242 @xref{Standards,,Language Standards Supported by GCC}, for details of
1243 these standard versions.
1245 @item -aux-info @var{filename}
1247 Output to the given filename prototyped declarations for all functions
1248 declared and/or defined in a translation unit, including those in header
1249 files. This option is silently ignored in any language other than C@.
1251 Besides declarations, the file indicates, in comments, the origin of
1252 each declaration (source file and line), whether the declaration was
1253 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1254 @samp{O} for old, respectively, in the first character after the line
1255 number and the colon), and whether it came from a declaration or a
1256 definition (@samp{C} or @samp{F}, respectively, in the following
1257 character). In the case of function definitions, a K&R-style list of
1258 arguments followed by their declarations is also provided, inside
1259 comments, after the declaration.
1263 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1264 keyword, so that code can use these words as identifiers. You can use
1265 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1266 instead. @option{-ansi} implies @option{-fno-asm}.
1268 In C++, this switch only affects the @code{typeof} keyword, since
1269 @code{asm} and @code{inline} are standard keywords. You may want to
1270 use the @option{-fno-gnu-keywords} flag instead, which has the same
1271 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1272 switch only affects the @code{asm} and @code{typeof} keywords, since
1273 @code{inline} is a standard keyword in ISO C99.
1276 @itemx -fno-builtin-@var{function}
1277 @opindex fno-builtin
1278 @cindex built-in functions
1279 Don't recognize built-in functions that do not begin with
1280 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1281 functions provided by GCC}, for details of the functions affected,
1282 including those which are not built-in functions when @option{-ansi} or
1283 @option{-std} options for strict ISO C conformance are used because they
1284 do not have an ISO standard meaning.
1286 GCC normally generates special code to handle certain built-in functions
1287 more efficiently; for instance, calls to @code{alloca} may become single
1288 instructions that adjust the stack directly, and calls to @code{memcpy}
1289 may become inline copy loops. The resulting code is often both smaller
1290 and faster, but since the function calls no longer appear as such, you
1291 cannot set a breakpoint on those calls, nor can you change the behavior
1292 of the functions by linking with a different library. In addition,
1293 when a function is recognized as a built-in function, GCC may use
1294 information about that function to warn about problems with calls to
1295 that function, or to generate more efficient code, even if the
1296 resulting code still contains calls to that function. For example,
1297 warnings are given with @option{-Wformat} for bad calls to
1298 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1299 known not to modify global memory.
1301 With the @option{-fno-builtin-@var{function}} option
1302 only the built-in function @var{function} is
1303 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1304 function is named this is not built-in in this version of GCC, this
1305 option is ignored. There is no corresponding
1306 @option{-fbuiltin-@var{function}} option; if you wish to enable
1307 built-in functions selectively when using @option{-fno-builtin} or
1308 @option{-ffreestanding}, you may define macros such as:
1311 #define abs(n) __builtin_abs ((n))
1312 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1317 @cindex hosted environment
1319 Assert that compilation takes place in a hosted environment. This implies
1320 @option{-fbuiltin}. A hosted environment is one in which the
1321 entire standard library is available, and in which @code{main} has a return
1322 type of @code{int}. Examples are nearly everything except a kernel.
1323 This is equivalent to @option{-fno-freestanding}.
1325 @item -ffreestanding
1326 @opindex ffreestanding
1327 @cindex hosted environment
1329 Assert that compilation takes place in a freestanding environment. This
1330 implies @option{-fno-builtin}. A freestanding environment
1331 is one in which the standard library may not exist, and program startup may
1332 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1333 This is equivalent to @option{-fno-hosted}.
1335 @xref{Standards,,Language Standards Supported by GCC}, for details of
1336 freestanding and hosted environments.
1340 @cindex openmp parallel
1341 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1342 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1343 compiler generates parallel code according to the OpenMP Application
1344 Program Interface v2.5 @w{@uref{http://www.openmp.org/}}.
1346 @item -fms-extensions
1347 @opindex fms-extensions
1348 Accept some non-standard constructs used in Microsoft header files.
1350 Some cases of unnamed fields in structures and unions are only
1351 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1352 fields within structs/unions}, for details.
1356 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1357 options for strict ISO C conformance) implies @option{-trigraphs}.
1359 @item -no-integrated-cpp
1360 @opindex no-integrated-cpp
1361 Performs a compilation in two passes: preprocessing and compiling. This
1362 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1363 @option{-B} option. The user supplied compilation step can then add in
1364 an additional preprocessing step after normal preprocessing but before
1365 compiling. The default is to use the integrated cpp (internal cpp)
1367 The semantics of this option will change if "cc1", "cc1plus", and
1368 "cc1obj" are merged.
1370 @cindex traditional C language
1371 @cindex C language, traditional
1373 @itemx -traditional-cpp
1374 @opindex traditional-cpp
1375 @opindex traditional
1376 Formerly, these options caused GCC to attempt to emulate a pre-standard
1377 C compiler. They are now only supported with the @option{-E} switch.
1378 The preprocessor continues to support a pre-standard mode. See the GNU
1379 CPP manual for details.
1381 @item -fcond-mismatch
1382 @opindex fcond-mismatch
1383 Allow conditional expressions with mismatched types in the second and
1384 third arguments. The value of such an expression is void. This option
1385 is not supported for C++.
1387 @item -flax-vector-conversions
1388 @opindex flax-vector-conversions
1389 Allow implicit conversions between vectors with differing numbers of
1390 elements and/or incompatible element types. This option should not be
1393 @item -funsigned-char
1394 @opindex funsigned-char
1395 Let the type @code{char} be unsigned, like @code{unsigned char}.
1397 Each kind of machine has a default for what @code{char} should
1398 be. It is either like @code{unsigned char} by default or like
1399 @code{signed char} by default.
1401 Ideally, a portable program should always use @code{signed char} or
1402 @code{unsigned char} when it depends on the signedness of an object.
1403 But many programs have been written to use plain @code{char} and
1404 expect it to be signed, or expect it to be unsigned, depending on the
1405 machines they were written for. This option, and its inverse, let you
1406 make such a program work with the opposite default.
1408 The type @code{char} is always a distinct type from each of
1409 @code{signed char} or @code{unsigned char}, even though its behavior
1410 is always just like one of those two.
1413 @opindex fsigned-char
1414 Let the type @code{char} be signed, like @code{signed char}.
1416 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1417 the negative form of @option{-funsigned-char}. Likewise, the option
1418 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1420 @item -fsigned-bitfields
1421 @itemx -funsigned-bitfields
1422 @itemx -fno-signed-bitfields
1423 @itemx -fno-unsigned-bitfields
1424 @opindex fsigned-bitfields
1425 @opindex funsigned-bitfields
1426 @opindex fno-signed-bitfields
1427 @opindex fno-unsigned-bitfields
1428 These options control whether a bit-field is signed or unsigned, when the
1429 declaration does not use either @code{signed} or @code{unsigned}. By
1430 default, such a bit-field is signed, because this is consistent: the
1431 basic integer types such as @code{int} are signed types.
1434 @node C++ Dialect Options
1435 @section Options Controlling C++ Dialect
1437 @cindex compiler options, C++
1438 @cindex C++ options, command line
1439 @cindex options, C++
1440 This section describes the command-line options that are only meaningful
1441 for C++ programs; but you can also use most of the GNU compiler options
1442 regardless of what language your program is in. For example, you
1443 might compile a file @code{firstClass.C} like this:
1446 g++ -g -frepo -O -c firstClass.C
1450 In this example, only @option{-frepo} is an option meant
1451 only for C++ programs; you can use the other options with any
1452 language supported by GCC@.
1454 Here is a list of options that are @emph{only} for compiling C++ programs:
1458 @item -fabi-version=@var{n}
1459 @opindex fabi-version
1460 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1461 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1462 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1463 the version that conforms most closely to the C++ ABI specification.
1464 Therefore, the ABI obtained using version 0 will change as ABI bugs
1467 The default is version 2.
1469 @item -fno-access-control
1470 @opindex fno-access-control
1471 Turn off all access checking. This switch is mainly useful for working
1472 around bugs in the access control code.
1476 Check that the pointer returned by @code{operator new} is non-null
1477 before attempting to modify the storage allocated. This check is
1478 normally unnecessary because the C++ standard specifies that
1479 @code{operator new} will only return @code{0} if it is declared
1480 @samp{throw()}, in which case the compiler will always check the
1481 return value even without this option. In all other cases, when
1482 @code{operator new} has a non-empty exception specification, memory
1483 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1484 @samp{new (nothrow)}.
1486 @item -fconserve-space
1487 @opindex fconserve-space
1488 Put uninitialized or runtime-initialized global variables into the
1489 common segment, as C does. This saves space in the executable at the
1490 cost of not diagnosing duplicate definitions. If you compile with this
1491 flag and your program mysteriously crashes after @code{main()} has
1492 completed, you may have an object that is being destroyed twice because
1493 two definitions were merged.
1495 This option is no longer useful on most targets, now that support has
1496 been added for putting variables into BSS without making them common.
1498 @item -ffriend-injection
1499 @opindex ffriend-injection
1500 Inject friend functions into the enclosing namespace, so that they are
1501 visible outside the scope of the class in which they are declared.
1502 Friend functions were documented to work this way in the old Annotated
1503 C++ Reference Manual, and versions of G++ before 4.1 always worked
1504 that way. However, in ISO C++ a friend function which is not declared
1505 in an enclosing scope can only be found using argument dependent
1506 lookup. This option causes friends to be injected as they were in
1509 This option is for compatibility, and may be removed in a future
1512 @item -fno-elide-constructors
1513 @opindex fno-elide-constructors
1514 The C++ standard allows an implementation to omit creating a temporary
1515 which is only used to initialize another object of the same type.
1516 Specifying this option disables that optimization, and forces G++ to
1517 call the copy constructor in all cases.
1519 @item -fno-enforce-eh-specs
1520 @opindex fno-enforce-eh-specs
1521 Don't generate code to check for violation of exception specifications
1522 at runtime. This option violates the C++ standard, but may be useful
1523 for reducing code size in production builds, much like defining
1524 @samp{NDEBUG}. This does not give user code permission to throw
1525 exceptions in violation of the exception specifications; the compiler
1526 will still optimize based on the specifications, so throwing an
1527 unexpected exception will result in undefined behavior.
1530 @itemx -fno-for-scope
1532 @opindex fno-for-scope
1533 If @option{-ffor-scope} is specified, the scope of variables declared in
1534 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1535 as specified by the C++ standard.
1536 If @option{-fno-for-scope} is specified, the scope of variables declared in
1537 a @i{for-init-statement} extends to the end of the enclosing scope,
1538 as was the case in old versions of G++, and other (traditional)
1539 implementations of C++.
1541 The default if neither flag is given to follow the standard,
1542 but to allow and give a warning for old-style code that would
1543 otherwise be invalid, or have different behavior.
1545 @item -fno-gnu-keywords
1546 @opindex fno-gnu-keywords
1547 Do not recognize @code{typeof} as a keyword, so that code can use this
1548 word as an identifier. You can use the keyword @code{__typeof__} instead.
1549 @option{-ansi} implies @option{-fno-gnu-keywords}.
1551 @item -fno-implicit-templates
1552 @opindex fno-implicit-templates
1553 Never emit code for non-inline templates which are instantiated
1554 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1555 @xref{Template Instantiation}, for more information.
1557 @item -fno-implicit-inline-templates
1558 @opindex fno-implicit-inline-templates
1559 Don't emit code for implicit instantiations of inline templates, either.
1560 The default is to handle inlines differently so that compiles with and
1561 without optimization will need the same set of explicit instantiations.
1563 @item -fno-implement-inlines
1564 @opindex fno-implement-inlines
1565 To save space, do not emit out-of-line copies of inline functions
1566 controlled by @samp{#pragma implementation}. This will cause linker
1567 errors if these functions are not inlined everywhere they are called.
1569 @item -fms-extensions
1570 @opindex fms-extensions
1571 Disable pedantic warnings about constructs used in MFC, such as implicit
1572 int and getting a pointer to member function via non-standard syntax.
1574 @item -fno-nonansi-builtins
1575 @opindex fno-nonansi-builtins
1576 Disable built-in declarations of functions that are not mandated by
1577 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1578 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1580 @item -fno-operator-names
1581 @opindex fno-operator-names
1582 Do not treat the operator name keywords @code{and}, @code{bitand},
1583 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1584 synonyms as keywords.
1586 @item -fno-optional-diags
1587 @opindex fno-optional-diags
1588 Disable diagnostics that the standard says a compiler does not need to
1589 issue. Currently, the only such diagnostic issued by G++ is the one for
1590 a name having multiple meanings within a class.
1593 @opindex fpermissive
1594 Downgrade some diagnostics about nonconformant code from errors to
1595 warnings. Thus, using @option{-fpermissive} will allow some
1596 nonconforming code to compile.
1600 Enable automatic template instantiation at link time. This option also
1601 implies @option{-fno-implicit-templates}. @xref{Template
1602 Instantiation}, for more information.
1606 Disable generation of information about every class with virtual
1607 functions for use by the C++ runtime type identification features
1608 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1609 of the language, you can save some space by using this flag. Note that
1610 exception handling uses the same information, but it will generate it as
1611 needed. The @samp{dynamic_cast} operator can still be used for casts that
1612 do not require runtime type information, i.e. casts to @code{void *} or to
1613 unambiguous base classes.
1617 Emit statistics about front-end processing at the end of the compilation.
1618 This information is generally only useful to the G++ development team.
1620 @item -ftemplate-depth-@var{n}
1621 @opindex ftemplate-depth
1622 Set the maximum instantiation depth for template classes to @var{n}.
1623 A limit on the template instantiation depth is needed to detect
1624 endless recursions during template class instantiation. ANSI/ISO C++
1625 conforming programs must not rely on a maximum depth greater than 17.
1627 @item -fno-threadsafe-statics
1628 @opindex fno-threadsafe-statics
1629 Do not emit the extra code to use the routines specified in the C++
1630 ABI for thread-safe initialization of local statics. You can use this
1631 option to reduce code size slightly in code that doesn't need to be
1634 @item -fuse-cxa-atexit
1635 @opindex fuse-cxa-atexit
1636 Register destructors for objects with static storage duration with the
1637 @code{__cxa_atexit} function rather than the @code{atexit} function.
1638 This option is required for fully standards-compliant handling of static
1639 destructors, but will only work if your C library supports
1640 @code{__cxa_atexit}.
1642 @item -fno-use-cxa-get-exception-ptr
1643 @opindex fno-use-cxa-get-exception-ptr
1644 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1645 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1646 if the runtime routine is not available.
1648 @item -fvisibility-inlines-hidden
1649 @opindex fvisibility-inlines-hidden
1650 This switch declares that the user does not attempt to compare
1651 pointers to inline methods where the addresses of the two functions
1652 were taken in different shared objects.
1654 The effect of this is that GCC may, effectively, mark inline methods with
1655 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1656 appear in the export table of a DSO and do not require a PLT indirection
1657 when used within the DSO@. Enabling this option can have a dramatic effect
1658 on load and link times of a DSO as it massively reduces the size of the
1659 dynamic export table when the library makes heavy use of templates.
1661 The behaviour of this switch is not quite the same as marking the
1662 methods as hidden directly, because it does not affect static variables
1663 local to the function or cause the compiler to deduce that
1664 the function is defined in only one shared object.
1666 You may mark a method as having a visibility explicitly to negate the
1667 effect of the switch for that method. For example, if you do want to
1668 compare pointers to a particular inline method, you might mark it as
1669 having default visibility. Marking the enclosing class with explicit
1670 visibility will have no effect.
1672 Explicitly instantiated inline methods are unaffected by this option
1673 as their linkage might otherwise cross a shared library boundary.
1674 @xref{Template Instantiation}.
1678 Do not use weak symbol support, even if it is provided by the linker.
1679 By default, G++ will use weak symbols if they are available. This
1680 option exists only for testing, and should not be used by end-users;
1681 it will result in inferior code and has no benefits. This option may
1682 be removed in a future release of G++.
1686 Do not search for header files in the standard directories specific to
1687 C++, but do still search the other standard directories. (This option
1688 is used when building the C++ library.)
1691 In addition, these optimization, warning, and code generation options
1692 have meanings only for C++ programs:
1695 @item -fno-default-inline
1696 @opindex fno-default-inline
1697 Do not assume @samp{inline} for functions defined inside a class scope.
1698 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1699 functions will have linkage like inline functions; they just won't be
1702 @item -Wabi @r{(C++ only)}
1704 Warn when G++ generates code that is probably not compatible with the
1705 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1706 all such cases, there are probably some cases that are not warned about,
1707 even though G++ is generating incompatible code. There may also be
1708 cases where warnings are emitted even though the code that is generated
1711 You should rewrite your code to avoid these warnings if you are
1712 concerned about the fact that code generated by G++ may not be binary
1713 compatible with code generated by other compilers.
1715 The known incompatibilities at this point include:
1720 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1721 pack data into the same byte as a base class. For example:
1724 struct A @{ virtual void f(); int f1 : 1; @};
1725 struct B : public A @{ int f2 : 1; @};
1729 In this case, G++ will place @code{B::f2} into the same byte
1730 as@code{A::f1}; other compilers will not. You can avoid this problem
1731 by explicitly padding @code{A} so that its size is a multiple of the
1732 byte size on your platform; that will cause G++ and other compilers to
1733 layout @code{B} identically.
1736 Incorrect handling of tail-padding for virtual bases. G++ does not use
1737 tail padding when laying out virtual bases. For example:
1740 struct A @{ virtual void f(); char c1; @};
1741 struct B @{ B(); char c2; @};
1742 struct C : public A, public virtual B @{@};
1746 In this case, G++ will not place @code{B} into the tail-padding for
1747 @code{A}; other compilers will. You can avoid this problem by
1748 explicitly padding @code{A} so that its size is a multiple of its
1749 alignment (ignoring virtual base classes); that will cause G++ and other
1750 compilers to layout @code{C} identically.
1753 Incorrect handling of bit-fields with declared widths greater than that
1754 of their underlying types, when the bit-fields appear in a union. For
1758 union U @{ int i : 4096; @};
1762 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1763 union too small by the number of bits in an @code{int}.
1766 Empty classes can be placed at incorrect offsets. For example:
1776 struct C : public B, public A @{@};
1780 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1781 it should be placed at offset zero. G++ mistakenly believes that the
1782 @code{A} data member of @code{B} is already at offset zero.
1785 Names of template functions whose types involve @code{typename} or
1786 template template parameters can be mangled incorrectly.
1789 template <typename Q>
1790 void f(typename Q::X) @{@}
1792 template <template <typename> class Q>
1793 void f(typename Q<int>::X) @{@}
1797 Instantiations of these templates may be mangled incorrectly.
1801 @item -Wctor-dtor-privacy @r{(C++ only)}
1802 @opindex Wctor-dtor-privacy
1803 Warn when a class seems unusable because all the constructors or
1804 destructors in that class are private, and it has neither friends nor
1805 public static member functions.
1807 @item -Wnon-virtual-dtor @r{(C++ only)}
1808 @opindex Wnon-virtual-dtor
1809 Warn when a class appears to be polymorphic, thereby requiring a virtual
1810 destructor, yet it declares a non-virtual one. This warning is also
1811 enabled if -Weffc++ is specified.
1813 @item -Wreorder @r{(C++ only)}
1815 @cindex reordering, warning
1816 @cindex warning for reordering of member initializers
1817 Warn when the order of member initializers given in the code does not
1818 match the order in which they must be executed. For instance:
1824 A(): j (0), i (1) @{ @}
1828 The compiler will rearrange the member initializers for @samp{i}
1829 and @samp{j} to match the declaration order of the members, emitting
1830 a warning to that effect. This warning is enabled by @option{-Wall}.
1833 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1836 @item -Weffc++ @r{(C++ only)}
1838 Warn about violations of the following style guidelines from Scott Meyers'
1839 @cite{Effective C++} book:
1843 Item 11: Define a copy constructor and an assignment operator for classes
1844 with dynamically allocated memory.
1847 Item 12: Prefer initialization to assignment in constructors.
1850 Item 14: Make destructors virtual in base classes.
1853 Item 15: Have @code{operator=} return a reference to @code{*this}.
1856 Item 23: Don't try to return a reference when you must return an object.
1860 Also warn about violations of the following style guidelines from
1861 Scott Meyers' @cite{More Effective C++} book:
1865 Item 6: Distinguish between prefix and postfix forms of increment and
1866 decrement operators.
1869 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1873 When selecting this option, be aware that the standard library
1874 headers do not obey all of these guidelines; use @samp{grep -v}
1875 to filter out those warnings.
1877 @item -Wno-deprecated @r{(C++ only)}
1878 @opindex Wno-deprecated
1879 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1881 @item -Wstrict-null-sentinel @r{(C++ only)}
1882 @opindex Wstrict-null-sentinel
1883 Warn also about the use of an uncasted @code{NULL} as sentinel. When
1884 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
1885 to @code{__null}. Although it is a null pointer constant not a null pointer,
1886 it is guaranteed to of the same size as a pointer. But this use is
1887 not portable across different compilers.
1889 @item -Wno-non-template-friend @r{(C++ only)}
1890 @opindex Wno-non-template-friend
1891 Disable warnings when non-templatized friend functions are declared
1892 within a template. Since the advent of explicit template specification
1893 support in G++, if the name of the friend is an unqualified-id (i.e.,
1894 @samp{friend foo(int)}), the C++ language specification demands that the
1895 friend declare or define an ordinary, nontemplate function. (Section
1896 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1897 could be interpreted as a particular specialization of a templatized
1898 function. Because this non-conforming behavior is no longer the default
1899 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1900 check existing code for potential trouble spots and is on by default.
1901 This new compiler behavior can be turned off with
1902 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1903 but disables the helpful warning.
1905 @item -Wold-style-cast @r{(C++ only)}
1906 @opindex Wold-style-cast
1907 Warn if an old-style (C-style) cast to a non-void type is used within
1908 a C++ program. The new-style casts (@samp{dynamic_cast},
1909 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
1910 less vulnerable to unintended effects and much easier to search for.
1912 @item -Woverloaded-virtual @r{(C++ only)}
1913 @opindex Woverloaded-virtual
1914 @cindex overloaded virtual fn, warning
1915 @cindex warning for overloaded virtual fn
1916 Warn when a function declaration hides virtual functions from a
1917 base class. For example, in:
1924 struct B: public A @{
1929 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1937 will fail to compile.
1939 @item -Wno-pmf-conversions @r{(C++ only)}
1940 @opindex Wno-pmf-conversions
1941 Disable the diagnostic for converting a bound pointer to member function
1944 @item -Wsign-promo @r{(C++ only)}
1945 @opindex Wsign-promo
1946 Warn when overload resolution chooses a promotion from unsigned or
1947 enumerated type to a signed type, over a conversion to an unsigned type of
1948 the same size. Previous versions of G++ would try to preserve
1949 unsignedness, but the standard mandates the current behavior.
1954 A& operator = (int);
1964 In this example, G++ will synthesize a default @samp{A& operator =
1965 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1968 @node Objective-C and Objective-C++ Dialect Options
1969 @section Options Controlling Objective-C and Objective-C++ Dialects
1971 @cindex compiler options, Objective-C and Objective-C++
1972 @cindex Objective-C and Objective-C++ options, command line
1973 @cindex options, Objective-C and Objective-C++
1974 (NOTE: This manual does not describe the Objective-C and Objective-C++
1975 languages themselves. See @xref{Standards,,Language Standards
1976 Supported by GCC}, for references.)
1978 This section describes the command-line options that are only meaningful
1979 for Objective-C and Objective-C++ programs, but you can also use most of
1980 the language-independent GNU compiler options.
1981 For example, you might compile a file @code{some_class.m} like this:
1984 gcc -g -fgnu-runtime -O -c some_class.m
1988 In this example, @option{-fgnu-runtime} is an option meant only for
1989 Objective-C and Objective-C++ programs; you can use the other options with
1990 any language supported by GCC@.
1992 Note that since Objective-C is an extension of the C language, Objective-C
1993 compilations may also use options specific to the C front-end (e.g.,
1994 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
1995 C++-specific options (e.g., @option{-Wabi}).
1997 Here is a list of options that are @emph{only} for compiling Objective-C
1998 and Objective-C++ programs:
2001 @item -fconstant-string-class=@var{class-name}
2002 @opindex fconstant-string-class
2003 Use @var{class-name} as the name of the class to instantiate for each
2004 literal string specified with the syntax @code{@@"@dots{}"}. The default
2005 class name is @code{NXConstantString} if the GNU runtime is being used, and
2006 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2007 @option{-fconstant-cfstrings} option, if also present, will override the
2008 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2009 to be laid out as constant CoreFoundation strings.
2012 @opindex fgnu-runtime
2013 Generate object code compatible with the standard GNU Objective-C
2014 runtime. This is the default for most types of systems.
2016 @item -fnext-runtime
2017 @opindex fnext-runtime
2018 Generate output compatible with the NeXT runtime. This is the default
2019 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2020 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2023 @item -fno-nil-receivers
2024 @opindex fno-nil-receivers
2025 Assume that all Objective-C message dispatches (e.g.,
2026 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2027 is not @code{nil}. This allows for more efficient entry points in the runtime
2028 to be used. Currently, this option is only available in conjunction with
2029 the NeXT runtime on Mac OS X 10.3 and later.
2031 @item -fobjc-call-cxx-cdtors
2032 @opindex fobjc-call-cxx-cdtors
2033 For each Objective-C class, check if any of its instance variables is a
2034 C++ object with a non-trivial default constructor. If so, synthesize a
2035 special @code{- (id) .cxx_construct} instance method that will run
2036 non-trivial default constructors on any such instance variables, in order,
2037 and then return @code{self}. Similarly, check if any instance variable
2038 is a C++ object with a non-trivial destructor, and if so, synthesize a
2039 special @code{- (void) .cxx_destruct} method that will run
2040 all such default destructors, in reverse order.
2042 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2043 thusly generated will only operate on instance variables declared in the
2044 current Objective-C class, and not those inherited from superclasses. It
2045 is the responsibility of the Objective-C runtime to invoke all such methods
2046 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2047 will be invoked by the runtime immediately after a new object
2048 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2049 be invoked immediately before the runtime deallocates an object instance.
2051 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2052 support for invoking the @code{- (id) .cxx_construct} and
2053 @code{- (void) .cxx_destruct} methods.
2055 @item -fobjc-direct-dispatch
2056 @opindex fobjc-direct-dispatch
2057 Allow fast jumps to the message dispatcher. On Darwin this is
2058 accomplished via the comm page.
2060 @item -fobjc-exceptions
2061 @opindex fobjc-exceptions
2062 Enable syntactic support for structured exception handling in Objective-C,
2063 similar to what is offered by C++ and Java. This option is
2064 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2073 @@catch (AnObjCClass *exc) @{
2080 @@catch (AnotherClass *exc) @{
2083 @@catch (id allOthers) @{
2093 The @code{@@throw} statement may appear anywhere in an Objective-C or
2094 Objective-C++ program; when used inside of a @code{@@catch} block, the
2095 @code{@@throw} may appear without an argument (as shown above), in which case
2096 the object caught by the @code{@@catch} will be rethrown.
2098 Note that only (pointers to) Objective-C objects may be thrown and
2099 caught using this scheme. When an object is thrown, it will be caught
2100 by the nearest @code{@@catch} clause capable of handling objects of that type,
2101 analogously to how @code{catch} blocks work in C++ and Java. A
2102 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2103 any and all Objective-C exceptions not caught by previous @code{@@catch}
2106 The @code{@@finally} clause, if present, will be executed upon exit from the
2107 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2108 regardless of whether any exceptions are thrown, caught or rethrown
2109 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2110 of the @code{finally} clause in Java.
2112 There are several caveats to using the new exception mechanism:
2116 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2117 idioms provided by the @code{NSException} class, the new
2118 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2119 systems, due to additional functionality needed in the (NeXT) Objective-C
2123 As mentioned above, the new exceptions do not support handling
2124 types other than Objective-C objects. Furthermore, when used from
2125 Objective-C++, the Objective-C exception model does not interoperate with C++
2126 exceptions at this time. This means you cannot @code{@@throw} an exception
2127 from Objective-C and @code{catch} it in C++, or vice versa
2128 (i.e., @code{throw @dots{} @@catch}).
2131 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2132 blocks for thread-safe execution:
2135 @@synchronized (ObjCClass *guard) @{
2140 Upon entering the @code{@@synchronized} block, a thread of execution shall
2141 first check whether a lock has been placed on the corresponding @code{guard}
2142 object by another thread. If it has, the current thread shall wait until
2143 the other thread relinquishes its lock. Once @code{guard} becomes available,
2144 the current thread will place its own lock on it, execute the code contained in
2145 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2146 making @code{guard} available to other threads).
2148 Unlike Java, Objective-C does not allow for entire methods to be marked
2149 @code{@@synchronized}. Note that throwing exceptions out of
2150 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2151 to be unlocked properly.
2155 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2157 @item -freplace-objc-classes
2158 @opindex freplace-objc-classes
2159 Emit a special marker instructing @command{ld(1)} not to statically link in
2160 the resulting object file, and allow @command{dyld(1)} to load it in at
2161 run time instead. This is used in conjunction with the Fix-and-Continue
2162 debugging mode, where the object file in question may be recompiled and
2163 dynamically reloaded in the course of program execution, without the need
2164 to restart the program itself. Currently, Fix-and-Continue functionality
2165 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2170 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2171 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2172 compile time) with static class references that get initialized at load time,
2173 which improves run-time performance. Specifying the @option{-fzero-link} flag
2174 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2175 to be retained. This is useful in Zero-Link debugging mode, since it allows
2176 for individual class implementations to be modified during program execution.
2180 Dump interface declarations for all classes seen in the source file to a
2181 file named @file{@var{sourcename}.decl}.
2183 @item -Wassign-intercept
2184 @opindex Wassign-intercept
2185 Warn whenever an Objective-C assignment is being intercepted by the
2189 @opindex Wno-protocol
2190 If a class is declared to implement a protocol, a warning is issued for
2191 every method in the protocol that is not implemented by the class. The
2192 default behavior is to issue a warning for every method not explicitly
2193 implemented in the class, even if a method implementation is inherited
2194 from the superclass. If you use the @option{-Wno-protocol} option, then
2195 methods inherited from the superclass are considered to be implemented,
2196 and no warning is issued for them.
2200 Warn if multiple methods of different types for the same selector are
2201 found during compilation. The check is performed on the list of methods
2202 in the final stage of compilation. Additionally, a check is performed
2203 for each selector appearing in a @code{@@selector(@dots{})}
2204 expression, and a corresponding method for that selector has been found
2205 during compilation. Because these checks scan the method table only at
2206 the end of compilation, these warnings are not produced if the final
2207 stage of compilation is not reached, for example because an error is
2208 found during compilation, or because the @option{-fsyntax-only} option is
2211 @item -Wstrict-selector-match
2212 @opindex Wstrict-selector-match
2213 Warn if multiple methods with differing argument and/or return types are
2214 found for a given selector when attempting to send a message using this
2215 selector to a receiver of type @code{id} or @code{Class}. When this flag
2216 is off (which is the default behavior), the compiler will omit such warnings
2217 if any differences found are confined to types which share the same size
2220 @item -Wundeclared-selector
2221 @opindex Wundeclared-selector
2222 Warn if a @code{@@selector(@dots{})} expression referring to an
2223 undeclared selector is found. A selector is considered undeclared if no
2224 method with that name has been declared before the
2225 @code{@@selector(@dots{})} expression, either explicitly in an
2226 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2227 an @code{@@implementation} section. This option always performs its
2228 checks as soon as a @code{@@selector(@dots{})} expression is found,
2229 while @option{-Wselector} only performs its checks in the final stage of
2230 compilation. This also enforces the coding style convention
2231 that methods and selectors must be declared before being used.
2233 @item -print-objc-runtime-info
2234 @opindex print-objc-runtime-info
2235 Generate C header describing the largest structure that is passed by
2240 @node Language Independent Options
2241 @section Options to Control Diagnostic Messages Formatting
2242 @cindex options to control diagnostics formatting
2243 @cindex diagnostic messages
2244 @cindex message formatting
2246 Traditionally, diagnostic messages have been formatted irrespective of
2247 the output device's aspect (e.g.@: its width, @dots{}). The options described
2248 below can be used to control the diagnostic messages formatting
2249 algorithm, e.g.@: how many characters per line, how often source location
2250 information should be reported. Right now, only the C++ front end can
2251 honor these options. However it is expected, in the near future, that
2252 the remaining front ends would be able to digest them correctly.
2255 @item -fmessage-length=@var{n}
2256 @opindex fmessage-length
2257 Try to format error messages so that they fit on lines of about @var{n}
2258 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2259 the front ends supported by GCC@. If @var{n} is zero, then no
2260 line-wrapping will be done; each error message will appear on a single
2263 @opindex fdiagnostics-show-location
2264 @item -fdiagnostics-show-location=once
2265 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2266 reporter to emit @emph{once} source location information; that is, in
2267 case the message is too long to fit on a single physical line and has to
2268 be wrapped, the source location won't be emitted (as prefix) again,
2269 over and over, in subsequent continuation lines. This is the default
2272 @item -fdiagnostics-show-location=every-line
2273 Only meaningful in line-wrapping mode. Instructs the diagnostic
2274 messages reporter to emit the same source location information (as
2275 prefix) for physical lines that result from the process of breaking
2276 a message which is too long to fit on a single line.
2278 @item -fdiagnostics-show-option
2279 @opindex fdiagnostics-show-option
2280 This option instructs the diagnostic machinery to add text to each
2281 diagnostic emitted, which indicates which command line option directly
2282 controls that diagnostic, when such an option is known to the
2283 diagnostic machinery.
2285 @item -Wcoverage-mismatch
2286 @opindex Wcoverage-mismatch
2287 Warn if feedback profiles do not match when using the
2288 @option{-fprofile-use} option.
2289 If a source file was changed between @option{-fprofile-gen} and
2290 @option{-fprofile-use}, the files with the profile feedback can fail
2291 to match the source file and GCC can not use the profile feedback
2292 information. By default, GCC emits an error message in this case.
2293 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2294 error. GCC does not use appropriate feedback profiles, so using this
2295 option can result in poorly optimized code. This option is useful
2296 only in the case of very minor changes such as bugfixes to an
2301 @node Warning Options
2302 @section Options to Request or Suppress Warnings
2303 @cindex options to control warnings
2304 @cindex warning messages
2305 @cindex messages, warning
2306 @cindex suppressing warnings
2308 Warnings are diagnostic messages that report constructions which
2309 are not inherently erroneous but which are risky or suggest there
2310 may have been an error.
2312 You can request many specific warnings with options beginning @samp{-W},
2313 for example @option{-Wimplicit} to request warnings on implicit
2314 declarations. Each of these specific warning options also has a
2315 negative form beginning @samp{-Wno-} to turn off warnings;
2316 for example, @option{-Wno-implicit}. This manual lists only one of the
2317 two forms, whichever is not the default.
2319 The following options control the amount and kinds of warnings produced
2320 by GCC; for further, language-specific options also refer to
2321 @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect
2325 @cindex syntax checking
2327 @opindex fsyntax-only
2328 Check the code for syntax errors, but don't do anything beyond that.
2332 Issue all the warnings demanded by strict ISO C and ISO C++;
2333 reject all programs that use forbidden extensions, and some other
2334 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2335 version of the ISO C standard specified by any @option{-std} option used.
2337 Valid ISO C and ISO C++ programs should compile properly with or without
2338 this option (though a rare few will require @option{-ansi} or a
2339 @option{-std} option specifying the required version of ISO C)@. However,
2340 without this option, certain GNU extensions and traditional C and C++
2341 features are supported as well. With this option, they are rejected.
2343 @option{-pedantic} does not cause warning messages for use of the
2344 alternate keywords whose names begin and end with @samp{__}. Pedantic
2345 warnings are also disabled in the expression that follows
2346 @code{__extension__}. However, only system header files should use
2347 these escape routes; application programs should avoid them.
2348 @xref{Alternate Keywords}.
2350 Some users try to use @option{-pedantic} to check programs for strict ISO
2351 C conformance. They soon find that it does not do quite what they want:
2352 it finds some non-ISO practices, but not all---only those for which
2353 ISO C @emph{requires} a diagnostic, and some others for which
2354 diagnostics have been added.
2356 A feature to report any failure to conform to ISO C might be useful in
2357 some instances, but would require considerable additional work and would
2358 be quite different from @option{-pedantic}. We don't have plans to
2359 support such a feature in the near future.
2361 Where the standard specified with @option{-std} represents a GNU
2362 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2363 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2364 extended dialect is based. Warnings from @option{-pedantic} are given
2365 where they are required by the base standard. (It would not make sense
2366 for such warnings to be given only for features not in the specified GNU
2367 C dialect, since by definition the GNU dialects of C include all
2368 features the compiler supports with the given option, and there would be
2369 nothing to warn about.)
2371 @item -pedantic-errors
2372 @opindex pedantic-errors
2373 Like @option{-pedantic}, except that errors are produced rather than
2378 Inhibit all warning messages.
2382 Inhibit warning messages about the use of @samp{#import}.
2384 @item -Wchar-subscripts
2385 @opindex Wchar-subscripts
2386 Warn if an array subscript has type @code{char}. This is a common cause
2387 of error, as programmers often forget that this type is signed on some
2389 This warning is enabled by @option{-Wall}.
2393 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2394 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2395 This warning is enabled by @option{-Wall}.
2397 @item -Wfatal-errors
2398 @opindex Wfatal-errors
2399 This option causes the compiler to abort compilation on the first error
2400 occurred rather than trying to keep going and printing further error
2405 @opindex ffreestanding
2406 @opindex fno-builtin
2407 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2408 the arguments supplied have types appropriate to the format string
2409 specified, and that the conversions specified in the format string make
2410 sense. This includes standard functions, and others specified by format
2411 attributes (@pxref{Function Attributes}), in the @code{printf},
2412 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2413 not in the C standard) families (or other target-specific families).
2414 Which functions are checked without format attributes having been
2415 specified depends on the standard version selected, and such checks of
2416 functions without the attribute specified are disabled by
2417 @option{-ffreestanding} or @option{-fno-builtin}.
2419 The formats are checked against the format features supported by GNU
2420 libc version 2.2. These include all ISO C90 and C99 features, as well
2421 as features from the Single Unix Specification and some BSD and GNU
2422 extensions. Other library implementations may not support all these
2423 features; GCC does not support warning about features that go beyond a
2424 particular library's limitations. However, if @option{-pedantic} is used
2425 with @option{-Wformat}, warnings will be given about format features not
2426 in the selected standard version (but not for @code{strfmon} formats,
2427 since those are not in any version of the C standard). @xref{C Dialect
2428 Options,,Options Controlling C Dialect}.
2430 Since @option{-Wformat} also checks for null format arguments for
2431 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2433 @option{-Wformat} is included in @option{-Wall}. For more control over some
2434 aspects of format checking, the options @option{-Wformat-y2k},
2435 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2436 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2437 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2440 @opindex Wformat-y2k
2441 If @option{-Wformat} is specified, also warn about @code{strftime}
2442 formats which may yield only a two-digit year.
2444 @item -Wno-format-extra-args
2445 @opindex Wno-format-extra-args
2446 If @option{-Wformat} is specified, do not warn about excess arguments to a
2447 @code{printf} or @code{scanf} format function. The C standard specifies
2448 that such arguments are ignored.
2450 Where the unused arguments lie between used arguments that are
2451 specified with @samp{$} operand number specifications, normally
2452 warnings are still given, since the implementation could not know what
2453 type to pass to @code{va_arg} to skip the unused arguments. However,
2454 in the case of @code{scanf} formats, this option will suppress the
2455 warning if the unused arguments are all pointers, since the Single
2456 Unix Specification says that such unused arguments are allowed.
2458 @item -Wno-format-zero-length
2459 @opindex Wno-format-zero-length
2460 If @option{-Wformat} is specified, do not warn about zero-length formats.
2461 The C standard specifies that zero-length formats are allowed.
2463 @item -Wformat-nonliteral
2464 @opindex Wformat-nonliteral
2465 If @option{-Wformat} is specified, also warn if the format string is not a
2466 string literal and so cannot be checked, unless the format function
2467 takes its format arguments as a @code{va_list}.
2469 @item -Wformat-security
2470 @opindex Wformat-security
2471 If @option{-Wformat} is specified, also warn about uses of format
2472 functions that represent possible security problems. At present, this
2473 warns about calls to @code{printf} and @code{scanf} functions where the
2474 format string is not a string literal and there are no format arguments,
2475 as in @code{printf (foo);}. This may be a security hole if the format
2476 string came from untrusted input and contains @samp{%n}. (This is
2477 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2478 in future warnings may be added to @option{-Wformat-security} that are not
2479 included in @option{-Wformat-nonliteral}.)
2483 Enable @option{-Wformat} plus format checks not included in
2484 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2485 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2489 Warn about passing a null pointer for arguments marked as
2490 requiring a non-null value by the @code{nonnull} function attribute.
2492 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2493 can be disabled with the @option{-Wno-nonnull} option.
2495 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2497 Warn about uninitialized variables which are initialized with themselves.
2498 Note this option can only be used with the @option{-Wuninitialized} option,
2499 which in turn only works with @option{-O1} and above.
2501 For example, GCC will warn about @code{i} being uninitialized in the
2502 following snippet only when @option{-Winit-self} has been specified:
2513 @item -Wimplicit-int
2514 @opindex Wimplicit-int
2515 Warn when a declaration does not specify a type.
2516 This warning is enabled by @option{-Wall}.
2518 @item -Wimplicit-function-declaration
2519 @itemx -Werror-implicit-function-declaration
2520 @opindex Wimplicit-function-declaration
2521 @opindex Werror-implicit-function-declaration
2522 Give a warning (or error) whenever a function is used before being
2523 declared. The form @option{-Wno-error-implicit-function-declaration}
2525 This warning is enabled by @option{-Wall} (as a warning, not an error).
2529 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2530 This warning is enabled by @option{-Wall}.
2534 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2535 function with external linkage, returning int, taking either zero
2536 arguments, two, or three arguments of appropriate types.
2537 This warning is enabled by @option{-Wall}.
2539 @item -Wmissing-braces
2540 @opindex Wmissing-braces
2541 Warn if an aggregate or union initializer is not fully bracketed. In
2542 the following example, the initializer for @samp{a} is not fully
2543 bracketed, but that for @samp{b} is fully bracketed.
2546 int a[2][2] = @{ 0, 1, 2, 3 @};
2547 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2550 This warning is enabled by @option{-Wall}.
2552 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2553 @opindex Wmissing-include-dirs
2554 Warn if a user-supplied include directory does not exist.
2557 @opindex Wparentheses
2558 Warn if parentheses are omitted in certain contexts, such
2559 as when there is an assignment in a context where a truth value
2560 is expected, or when operators are nested whose precedence people
2561 often get confused about.
2563 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2564 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2565 interpretation from that of ordinary mathematical notation.
2567 Also warn about constructions where there may be confusion to which
2568 @code{if} statement an @code{else} branch belongs. Here is an example of
2583 In C/C++, every @code{else} branch belongs to the innermost possible
2584 @code{if} statement, which in this example is @code{if (b)}. This is
2585 often not what the programmer expected, as illustrated in the above
2586 example by indentation the programmer chose. When there is the
2587 potential for this confusion, GCC will issue a warning when this flag
2588 is specified. To eliminate the warning, add explicit braces around
2589 the innermost @code{if} statement so there is no way the @code{else}
2590 could belong to the enclosing @code{if}. The resulting code would
2607 This warning is enabled by @option{-Wall}.
2609 @item -Wsequence-point
2610 @opindex Wsequence-point
2611 Warn about code that may have undefined semantics because of violations
2612 of sequence point rules in the C and C++ standards.
2614 The C and C++ standards defines the order in which expressions in a C/C++
2615 program are evaluated in terms of @dfn{sequence points}, which represent
2616 a partial ordering between the execution of parts of the program: those
2617 executed before the sequence point, and those executed after it. These
2618 occur after the evaluation of a full expression (one which is not part
2619 of a larger expression), after the evaluation of the first operand of a
2620 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2621 function is called (but after the evaluation of its arguments and the
2622 expression denoting the called function), and in certain other places.
2623 Other than as expressed by the sequence point rules, the order of
2624 evaluation of subexpressions of an expression is not specified. All
2625 these rules describe only a partial order rather than a total order,
2626 since, for example, if two functions are called within one expression
2627 with no sequence point between them, the order in which the functions
2628 are called is not specified. However, the standards committee have
2629 ruled that function calls do not overlap.
2631 It is not specified when between sequence points modifications to the
2632 values of objects take effect. Programs whose behavior depends on this
2633 have undefined behavior; the C and C++ standards specify that ``Between
2634 the previous and next sequence point an object shall have its stored
2635 value modified at most once by the evaluation of an expression.
2636 Furthermore, the prior value shall be read only to determine the value
2637 to be stored.''. If a program breaks these rules, the results on any
2638 particular implementation are entirely unpredictable.
2640 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2641 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2642 diagnosed by this option, and it may give an occasional false positive
2643 result, but in general it has been found fairly effective at detecting
2644 this sort of problem in programs.
2646 The standard is worded confusingly, therefore there is some debate
2647 over the precise meaning of the sequence point rules in subtle cases.
2648 Links to discussions of the problem, including proposed formal
2649 definitions, may be found on the GCC readings page, at
2650 @w{@uref{http://gcc.gnu.org/readings.html}}.
2652 This warning is enabled by @option{-Wall} for C and C++.
2655 @opindex Wreturn-type
2656 Warn whenever a function is defined with a return-type that defaults to
2657 @code{int}. Also warn about any @code{return} statement with no
2658 return-value in a function whose return-type is not @code{void}.
2660 For C, also warn if the return type of a function has a type qualifier
2661 such as @code{const}. Such a type qualifier has no effect, since the
2662 value returned by a function is not an lvalue. ISO C prohibits
2663 qualified @code{void} return types on function definitions, so such
2664 return types always receive a warning even without this option.
2666 For C++, a function without return type always produces a diagnostic
2667 message, even when @option{-Wno-return-type} is specified. The only
2668 exceptions are @samp{main} and functions defined in system headers.
2670 This warning is enabled by @option{-Wall}.
2674 Warn whenever a @code{switch} statement has an index of enumerated type
2675 and lacks a @code{case} for one or more of the named codes of that
2676 enumeration. (The presence of a @code{default} label prevents this
2677 warning.) @code{case} labels outside the enumeration range also
2678 provoke warnings when this option is used.
2679 This warning is enabled by @option{-Wall}.
2681 @item -Wswitch-default
2682 @opindex Wswitch-switch
2683 Warn whenever a @code{switch} statement does not have a @code{default}
2687 @opindex Wswitch-enum
2688 Warn whenever a @code{switch} statement has an index of enumerated type
2689 and lacks a @code{case} for one or more of the named codes of that
2690 enumeration. @code{case} labels outside the enumeration range also
2691 provoke warnings when this option is used.
2695 Warn if any trigraphs are encountered that might change the meaning of
2696 the program (trigraphs within comments are not warned about).
2697 This warning is enabled by @option{-Wall}.
2699 @item -Wunused-function
2700 @opindex Wunused-function
2701 Warn whenever a static function is declared but not defined or a
2702 non-inline static function is unused.
2703 This warning is enabled by @option{-Wall}.
2705 @item -Wunused-label
2706 @opindex Wunused-label
2707 Warn whenever a label is declared but not used.
2708 This warning is enabled by @option{-Wall}.
2710 To suppress this warning use the @samp{unused} attribute
2711 (@pxref{Variable Attributes}).
2713 @item -Wunused-parameter
2714 @opindex Wunused-parameter
2715 Warn whenever a function parameter is unused aside from its declaration.
2717 To suppress this warning use the @samp{unused} attribute
2718 (@pxref{Variable Attributes}).
2720 @item -Wunused-variable
2721 @opindex Wunused-variable
2722 Warn whenever a local variable or non-constant static variable is unused
2723 aside from its declaration.
2724 This warning is enabled by @option{-Wall}.
2726 To suppress this warning use the @samp{unused} attribute
2727 (@pxref{Variable Attributes}).
2729 @item -Wunused-value
2730 @opindex Wunused-value
2731 Warn whenever a statement computes a result that is explicitly not used.
2732 This warning is enabled by @option{-Wall}.
2734 To suppress this warning cast the expression to @samp{void}.
2738 All the above @option{-Wunused} options combined.
2740 In order to get a warning about an unused function parameter, you must
2741 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2742 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2744 @item -Wuninitialized
2745 @opindex Wuninitialized
2746 Warn if an automatic variable is used without first being initialized or
2747 if a variable may be clobbered by a @code{setjmp} call.
2749 These warnings are possible only in optimizing compilation,
2750 because they require data flow information that is computed only
2751 when optimizing. If you do not specify @option{-O}, you will not get
2752 these warnings. Instead, GCC will issue a warning about @option{-Wuninitialized}
2753 requiring @option{-O}.
2755 If you want to warn about code which uses the uninitialized value of the
2756 variable in its own initializer, use the @option{-Winit-self} option.
2758 These warnings occur for individual uninitialized or clobbered
2759 elements of structure, union or array variables as well as for
2760 variables which are uninitialized or clobbered as a whole. They do
2761 not occur for variables or elements declared @code{volatile}. Because
2762 these warnings depend on optimization, the exact variables or elements
2763 for which there are warnings will depend on the precise optimization
2764 options and version of GCC used.
2766 Note that there may be no warning about a variable that is used only
2767 to compute a value that itself is never used, because such
2768 computations may be deleted by data flow analysis before the warnings
2771 These warnings are made optional because GCC is not smart
2772 enough to see all the reasons why the code might be correct
2773 despite appearing to have an error. Here is one example of how
2794 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2795 always initialized, but GCC doesn't know this. Here is
2796 another common case:
2801 if (change_y) save_y = y, y = new_y;
2803 if (change_y) y = save_y;
2808 This has no bug because @code{save_y} is used only if it is set.
2810 @cindex @code{longjmp} warnings
2811 This option also warns when a non-volatile automatic variable might be
2812 changed by a call to @code{longjmp}. These warnings as well are possible
2813 only in optimizing compilation.
2815 The compiler sees only the calls to @code{setjmp}. It cannot know
2816 where @code{longjmp} will be called; in fact, a signal handler could
2817 call it at any point in the code. As a result, you may get a warning
2818 even when there is in fact no problem because @code{longjmp} cannot
2819 in fact be called at the place which would cause a problem.
2821 Some spurious warnings can be avoided if you declare all the functions
2822 you use that never return as @code{noreturn}. @xref{Function
2825 This warning is enabled by @option{-Wall}.
2827 @item -Wunknown-pragmas
2828 @opindex Wunknown-pragmas
2829 @cindex warning for unknown pragmas
2830 @cindex unknown pragmas, warning
2831 @cindex pragmas, warning of unknown
2832 Warn when a #pragma directive is encountered which is not understood by
2833 GCC@. If this command line option is used, warnings will even be issued
2834 for unknown pragmas in system header files. This is not the case if
2835 the warnings were only enabled by the @option{-Wall} command line option.
2838 @opindex Wno-pragmas
2840 Do not warn about misuses of pragmas, such as incorrect parameters,
2841 invalid syntax, or conflicts between pragmas. See also
2842 @samp{-Wunknown-pragmas}.
2844 @item -Wstrict-aliasing
2845 @opindex Wstrict-aliasing
2846 This option is only active when @option{-fstrict-aliasing} is active.
2847 It warns about code which might break the strict aliasing rules that the
2848 compiler is using for optimization. The warning does not catch all
2849 cases, but does attempt to catch the more common pitfalls. It is
2850 included in @option{-Wall}.
2852 @item -Wstrict-aliasing=2
2853 @opindex Wstrict-aliasing=2
2854 This option is only active when @option{-fstrict-aliasing} is active.
2855 It warns about code which might break the strict aliasing rules that the
2856 compiler is using for optimization. This warning catches more cases than
2857 @option{-Wstrict-aliasing}, but it will also give a warning for some ambiguous
2858 cases that are safe.
2860 @item -Warray-bounds
2861 @opindex Wno-array-bounds
2862 @opindex Warray-bounds
2863 This option is only active when @option{-ftree-vrp} is active
2864 (default for -O2 and above). It warns about subscripts to arrays
2865 that are always out of bounds. This warning is enabled by @option{-Wall}.
2869 All of the above @samp{-W} options combined. This enables all the
2870 warnings about constructions that some users consider questionable, and
2871 that are easy to avoid (or modify to prevent the warning), even in
2872 conjunction with macros. This also enables some language-specific
2873 warnings described in @ref{C++ Dialect Options} and
2874 @ref{Objective-C and Objective-C++ Dialect Options}.
2877 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2878 Some of them warn about constructions that users generally do not
2879 consider questionable, but which occasionally you might wish to check
2880 for; others warn about constructions that are necessary or hard to avoid
2881 in some cases, and there is no simple way to modify the code to suppress
2888 (This option used to be called @option{-W}. The older name is still
2889 supported, but the newer name is more descriptive.) Print extra warning
2890 messages for these events:
2894 A function can return either with or without a value. (Falling
2895 off the end of the function body is considered returning without
2896 a value.) For example, this function would evoke such a
2910 An expression-statement or the left-hand side of a comma expression
2911 contains no side effects.
2912 To suppress the warning, cast the unused expression to void.
2913 For example, an expression such as @samp{x[i,j]} will cause a warning,
2914 but @samp{x[(void)i,j]} will not.
2917 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2920 Storage-class specifiers like @code{static} are not the first things
2921 in a declaration. According to the C Standard, this usage is
2922 obsolescent. This warning can be independently controlled by
2923 @option{-Wold-style-declaration}.
2926 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2930 A comparison between signed and unsigned values could produce an
2931 incorrect result when the signed value is converted to unsigned.
2932 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2935 An aggregate has an initializer which does not initialize all members.
2936 This warning can be independently controlled by
2937 @option{-Wmissing-field-initializers}.
2940 An initialized field without side effects is overridden when using
2941 designated initializers (@pxref{Designated Inits, , Designated
2942 Initializers}). This warning can be independently controlled by
2943 @option{-Woverride-init}.
2946 A function parameter is declared without a type specifier in K&R-style
2947 functions. This warning can be independently controlled by
2948 @option{-Wmissing-parameter-type}.
2951 An empty body occurs in an @samp{if} or @samp{else} statement. This
2952 warning can be independently controlled by @option{-Wempty-body}.
2955 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2956 @samp{>}, or @samp{>=}.
2959 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2960 This warning can be independently controlled by @option{-Wclobbered}.
2963 Any of several floating-point events that often indicate errors, such as
2964 overflow, underflow, loss of precision, etc.
2966 @item @r{(C++ only)}
2967 An enumerator and a non-enumerator both appear in a conditional expression.
2969 @item @r{(C++ only)}
2970 A non-static reference or non-static @samp{const} member appears in a
2971 class without constructors.
2973 @item @r{(C++ only)}
2974 Ambiguous virtual bases.
2976 @item @r{(C++ only)}
2977 Subscripting an array which has been declared @samp{register}.
2979 @item @r{(C++ only)}
2980 Taking the address of a variable which has been declared @samp{register}.
2982 @item @r{(C++ only)}
2983 A base class is not initialized in a derived class' copy constructor.
2986 @item -Wno-div-by-zero
2987 @opindex Wno-div-by-zero
2988 @opindex Wdiv-by-zero
2989 Do not warn about compile-time integer division by zero. Floating point
2990 division by zero is not warned about, as it can be a legitimate way of
2991 obtaining infinities and NaNs.
2993 @item -Wsystem-headers
2994 @opindex Wsystem-headers
2995 @cindex warnings from system headers
2996 @cindex system headers, warnings from
2997 Print warning messages for constructs found in system header files.
2998 Warnings from system headers are normally suppressed, on the assumption
2999 that they usually do not indicate real problems and would only make the
3000 compiler output harder to read. Using this command line option tells
3001 GCC to emit warnings from system headers as if they occurred in user
3002 code. However, note that using @option{-Wall} in conjunction with this
3003 option will @emph{not} warn about unknown pragmas in system
3004 headers---for that, @option{-Wunknown-pragmas} must also be used.
3007 @opindex Wfloat-equal
3008 Warn if floating point values are used in equality comparisons.
3010 The idea behind this is that sometimes it is convenient (for the
3011 programmer) to consider floating-point values as approximations to
3012 infinitely precise real numbers. If you are doing this, then you need
3013 to compute (by analyzing the code, or in some other way) the maximum or
3014 likely maximum error that the computation introduces, and allow for it
3015 when performing comparisons (and when producing output, but that's a
3016 different problem). In particular, instead of testing for equality, you
3017 would check to see whether the two values have ranges that overlap; and
3018 this is done with the relational operators, so equality comparisons are
3021 @item -Wtraditional @r{(C only)}
3022 @opindex Wtraditional
3023 Warn about certain constructs that behave differently in traditional and
3024 ISO C@. Also warn about ISO C constructs that have no traditional C
3025 equivalent, and/or problematic constructs which should be avoided.
3029 Macro parameters that appear within string literals in the macro body.
3030 In traditional C macro replacement takes place within string literals,
3031 but does not in ISO C@.
3034 In traditional C, some preprocessor directives did not exist.
3035 Traditional preprocessors would only consider a line to be a directive
3036 if the @samp{#} appeared in column 1 on the line. Therefore
3037 @option{-Wtraditional} warns about directives that traditional C
3038 understands but would ignore because the @samp{#} does not appear as the
3039 first character on the line. It also suggests you hide directives like
3040 @samp{#pragma} not understood by traditional C by indenting them. Some
3041 traditional implementations would not recognize @samp{#elif}, so it
3042 suggests avoiding it altogether.
3045 A function-like macro that appears without arguments.
3048 The unary plus operator.
3051 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3052 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3053 constants.) Note, these suffixes appear in macros defined in the system
3054 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3055 Use of these macros in user code might normally lead to spurious
3056 warnings, however GCC's integrated preprocessor has enough context to
3057 avoid warning in these cases.
3060 A function declared external in one block and then used after the end of
3064 A @code{switch} statement has an operand of type @code{long}.
3067 A non-@code{static} function declaration follows a @code{static} one.
3068 This construct is not accepted by some traditional C compilers.
3071 The ISO type of an integer constant has a different width or
3072 signedness from its traditional type. This warning is only issued if
3073 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3074 typically represent bit patterns, are not warned about.
3077 Usage of ISO string concatenation is detected.
3080 Initialization of automatic aggregates.
3083 Identifier conflicts with labels. Traditional C lacks a separate
3084 namespace for labels.
3087 Initialization of unions. If the initializer is zero, the warning is
3088 omitted. This is done under the assumption that the zero initializer in
3089 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3090 initializer warnings and relies on default initialization to zero in the
3094 Conversions by prototypes between fixed/floating point values and vice
3095 versa. The absence of these prototypes when compiling with traditional
3096 C would cause serious problems. This is a subset of the possible
3097 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3100 Use of ISO C style function definitions. This warning intentionally is
3101 @emph{not} issued for prototype declarations or variadic functions
3102 because these ISO C features will appear in your code when using
3103 libiberty's traditional C compatibility macros, @code{PARAMS} and
3104 @code{VPARAMS}. This warning is also bypassed for nested functions
3105 because that feature is already a GCC extension and thus not relevant to
3106 traditional C compatibility.
3109 @item -Wtraditional-conversion @r{(C only)}
3110 @opindex Wtraditional-conversion
3111 Warn if a prototype causes a type conversion that is different from what
3112 would happen to the same argument in the absence of a prototype. This
3113 includes conversions of fixed point to floating and vice versa, and
3114 conversions changing the width or signedness of a fixed point argument
3115 except when the same as the default promotion.
3117 @item -Wdeclaration-after-statement @r{(C only)}
3118 @opindex Wdeclaration-after-statement
3119 Warn when a declaration is found after a statement in a block. This
3120 construct, known from C++, was introduced with ISO C99 and is by default
3121 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3122 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3126 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3128 @item -Wno-endif-labels
3129 @opindex Wno-endif-labels
3130 @opindex Wendif-labels
3131 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3135 Warn whenever a local variable shadows another local variable, parameter or
3136 global variable or whenever a built-in function is shadowed.
3138 @item -Wlarger-than-@var{len}
3139 @opindex Wlarger-than
3140 Warn whenever an object of larger than @var{len} bytes is defined.
3142 @item -Wunsafe-loop-optimizations
3143 @opindex Wunsafe-loop-optimizations
3144 Warn if the loop cannot be optimized because the compiler could not
3145 assume anything on the bounds of the loop indices. With
3146 @option{-funsafe-loop-optimizations} warn if the compiler made
3149 @item -Wpointer-arith
3150 @opindex Wpointer-arith
3151 Warn about anything that depends on the ``size of'' a function type or
3152 of @code{void}. GNU C assigns these types a size of 1, for
3153 convenience in calculations with @code{void *} pointers and pointers
3156 @item -Wbad-function-cast @r{(C only)}
3157 @opindex Wbad-function-cast
3158 Warn whenever a function call is cast to a non-matching type.
3159 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3162 Warn about ISO C constructs that are outside of the common subset of
3163 ISO C and ISO C++, e.g.@: request for implicit conversion from
3164 @code{void *} to a pointer to non-@code{void} type.
3166 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3167 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3168 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3169 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3173 Warn whenever a pointer is cast so as to remove a type qualifier from
3174 the target type. For example, warn if a @code{const char *} is cast
3175 to an ordinary @code{char *}.
3178 @opindex Wcast-align
3179 Warn whenever a pointer is cast such that the required alignment of the
3180 target is increased. For example, warn if a @code{char *} is cast to
3181 an @code{int *} on machines where integers can only be accessed at
3182 two- or four-byte boundaries.
3184 @item -Wwrite-strings
3185 @opindex Wwrite-strings
3186 When compiling C, give string constants the type @code{const
3187 char[@var{length}]} so that
3188 copying the address of one into a non-@code{const} @code{char *}
3189 pointer will get a warning; when compiling C++, warn about the
3190 deprecated conversion from string literals to @code{char *}. This
3191 warning, by default, is enabled for C++ programs.
3192 These warnings will help you find at
3193 compile time code that can try to write into a string constant, but
3194 only if you have been very careful about using @code{const} in
3195 declarations and prototypes. Otherwise, it will just be a nuisance;
3196 this is why we did not make @option{-Wall} request these warnings.
3200 Warn for variables that might be changed by @samp{longjmp} or
3201 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3204 @opindex Wconversion
3205 Warn for implicit conversions that may alter a value. This includes
3206 conversions between real and integer, like @code{abs (x)} when
3207 @code{x} is @code{double}; conversions between signed and unsigned,
3208 like @code{unsigned ui = -1}; and conversions to smaller types, like
3209 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3210 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3211 changed by the conversion like in @code{abs (2.0)}.
3214 @opindex Wempty-body
3215 An empty body occurs in an @samp{if} or @samp{else} statement.
3216 This warning is also enabled by @option{-Wextra}.
3218 @item -Wsign-compare
3219 @opindex Wsign-compare
3220 @cindex warning for comparison of signed and unsigned values
3221 @cindex comparison of signed and unsigned values, warning
3222 @cindex signed and unsigned values, comparison warning
3223 Warn when a comparison between signed and unsigned values could produce
3224 an incorrect result when the signed value is converted to unsigned.
3225 This warning is also enabled by @option{-Wextra}; to get the other warnings
3226 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3228 @item -Waggregate-return
3229 @opindex Waggregate-return
3230 Warn if any functions that return structures or unions are defined or
3231 called. (In languages where you can return an array, this also elicits
3235 @opindex Walways-true
3236 Warn about comparisons which are always true such as testing if
3237 unsigned values are greater than or equal to zero. This warning is
3238 enabled by @option{-Wall}.
3240 @item -Wno-attributes
3241 @opindex Wno-attributes
3242 @opindex Wattributes
3243 Do not warn if an unexpected @code{__attribute__} is used, such as
3244 unrecognized attributes, function attributes applied to variables,
3245 etc. This will not stop errors for incorrect use of supported
3248 @item -Wstrict-prototypes @r{(C only)}
3249 @opindex Wstrict-prototypes
3250 Warn if a function is declared or defined without specifying the
3251 argument types. (An old-style function definition is permitted without
3252 a warning if preceded by a declaration which specifies the argument
3255 @item -Wold-style-declaration @r{(C only)}
3256 @opindex Wold-style-declaration
3257 Warn for obsolescent usages, according to the C Standard, in a
3258 declaration. For example, warn if storage-class specifiers like
3259 @code{static} are not the first things in a declaration. This warning
3260 is also enabled by @option{-Wextra}.
3262 @item -Wold-style-definition @r{(C only)}
3263 @opindex Wold-style-definition
3264 Warn if an old-style function definition is used. A warning is given
3265 even if there is a previous prototype.
3267 @item -Wmissing-parameter-type @r{(C only)}
3268 @opindex Wmissing-parameter-type
3269 A function parameter is declared without a type specifier in K&R-style
3276 This warning is also enabled by @option{-Wextra}.
3278 @item -Wmissing-prototypes @r{(C only)}
3279 @opindex Wmissing-prototypes
3280 Warn if a global function is defined without a previous prototype
3281 declaration. This warning is issued even if the definition itself
3282 provides a prototype. The aim is to detect global functions that fail
3283 to be declared in header files.
3285 @item -Wmissing-declarations @r{(C and C++ only)}
3286 @opindex Wmissing-declarations
3287 Warn if a global function is defined without a previous declaration.
3288 Do so even if the definition itself provides a prototype.
3289 Use this option to detect global functions that are not declared in
3290 header files. In C++, no warnings are issued for function templates,
3291 or for inline functions, or for functions in anonymous namespaces.
3293 @item -Wmissing-field-initializers
3294 @opindex Wmissing-field-initializers
3297 Warn if a structure's initializer has some fields missing. For
3298 example, the following code would cause such a warning, because
3299 @code{x.h} is implicitly zero:
3302 struct s @{ int f, g, h; @};
3303 struct s x = @{ 3, 4 @};
3306 This option does not warn about designated initializers, so the following
3307 modification would not trigger a warning:
3310 struct s @{ int f, g, h; @};
3311 struct s x = @{ .f = 3, .g = 4 @};
3314 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3315 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3317 @item -Wmissing-noreturn
3318 @opindex Wmissing-noreturn
3319 Warn about functions which might be candidates for attribute @code{noreturn}.
3320 Note these are only possible candidates, not absolute ones. Care should
3321 be taken to manually verify functions actually do not ever return before
3322 adding the @code{noreturn} attribute, otherwise subtle code generation
3323 bugs could be introduced. You will not get a warning for @code{main} in
3324 hosted C environments.
3326 @item -Wmissing-format-attribute
3327 @opindex Wmissing-format-attribute
3329 Warn about function pointers which might be candidates for @code{format}
3330 attributes. Note these are only possible candidates, not absolute ones.
3331 GCC will guess that function pointers with @code{format} attributes that
3332 are used in assignment, initialization, parameter passing or return
3333 statements should have a corresponding @code{format} attribute in the
3334 resulting type. I.e.@: the left-hand side of the assignment or
3335 initialization, the type of the parameter variable, or the return type
3336 of the containing function respectively should also have a @code{format}
3337 attribute to avoid the warning.
3339 GCC will also warn about function definitions which might be
3340 candidates for @code{format} attributes. Again, these are only
3341 possible candidates. GCC will guess that @code{format} attributes
3342 might be appropriate for any function that calls a function like
3343 @code{vprintf} or @code{vscanf}, but this might not always be the
3344 case, and some functions for which @code{format} attributes are
3345 appropriate may not be detected.
3347 @item -Wno-multichar
3348 @opindex Wno-multichar
3350 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3351 Usually they indicate a typo in the user's code, as they have
3352 implementation-defined values, and should not be used in portable code.
3354 @item -Wnormalized=<none|id|nfc|nfkc>
3355 @opindex Wnormalized
3358 @cindex character set, input normalization
3359 In ISO C and ISO C++, two identifiers are different if they are
3360 different sequences of characters. However, sometimes when characters
3361 outside the basic ASCII character set are used, you can have two
3362 different character sequences that look the same. To avoid confusion,
3363 the ISO 10646 standard sets out some @dfn{normalization rules} which
3364 when applied ensure that two sequences that look the same are turned into
3365 the same sequence. GCC can warn you if you are using identifiers which
3366 have not been normalized; this option controls that warning.
3368 There are four levels of warning that GCC supports. The default is
3369 @option{-Wnormalized=nfc}, which warns about any identifier which is
3370 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3371 recommended form for most uses.
3373 Unfortunately, there are some characters which ISO C and ISO C++ allow
3374 in identifiers that when turned into NFC aren't allowable as
3375 identifiers. That is, there's no way to use these symbols in portable
3376 ISO C or C++ and have all your identifiers in NFC.
3377 @option{-Wnormalized=id} suppresses the warning for these characters.
3378 It is hoped that future versions of the standards involved will correct
3379 this, which is why this option is not the default.
3381 You can switch the warning off for all characters by writing
3382 @option{-Wnormalized=none}. You would only want to do this if you
3383 were using some other normalization scheme (like ``D''), because
3384 otherwise you can easily create bugs that are literally impossible to see.
3386 Some characters in ISO 10646 have distinct meanings but look identical
3387 in some fonts or display methodologies, especially once formatting has
3388 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3389 LETTER N'', will display just like a regular @code{n} which has been
3390 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3391 normalization scheme to convert all these into a standard form as
3392 well, and GCC will warn if your code is not in NFKC if you use
3393 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3394 about every identifier that contains the letter O because it might be
3395 confused with the digit 0, and so is not the default, but may be
3396 useful as a local coding convention if the programming environment is
3397 unable to be fixed to display these characters distinctly.
3399 @item -Wno-deprecated-declarations
3400 @opindex Wno-deprecated-declarations
3401 Do not warn about uses of functions (@pxref{Function Attributes}),
3402 variables (@pxref{Variable Attributes}), and types (@pxref{Type
3403 Attributes}) marked as deprecated by using the @code{deprecated}
3407 @opindex Wno-overflow
3408 Do not warn about compile-time overflow in constant expressions.
3410 @item -Woverride-init
3411 @opindex Woverride-init
3414 Warn if an initialized field without side effects is overridden when
3415 using designated initializers (@pxref{Designated Inits, , Designated
3418 This warning is included in @option{-Wextra}. To get other
3419 @option{-Wextra} warnings without this one, use @samp{-Wextra
3420 -Wno-override-init}.
3424 Warn if a structure is given the packed attribute, but the packed
3425 attribute has no effect on the layout or size of the structure.
3426 Such structures may be mis-aligned for little benefit. For
3427 instance, in this code, the variable @code{f.x} in @code{struct bar}
3428 will be misaligned even though @code{struct bar} does not itself
3429 have the packed attribute:
3436 @} __attribute__((packed));
3446 Warn if padding is included in a structure, either to align an element
3447 of the structure or to align the whole structure. Sometimes when this
3448 happens it is possible to rearrange the fields of the structure to
3449 reduce the padding and so make the structure smaller.
3451 @item -Wredundant-decls
3452 @opindex Wredundant-decls
3453 Warn if anything is declared more than once in the same scope, even in
3454 cases where multiple declaration is valid and changes nothing.
3456 @item -Wnested-externs @r{(C only)}
3457 @opindex Wnested-externs
3458 Warn if an @code{extern} declaration is encountered within a function.
3460 @item -Wunreachable-code
3461 @opindex Wunreachable-code
3462 Warn if the compiler detects that code will never be executed.
3464 This option is intended to warn when the compiler detects that at
3465 least a whole line of source code will never be executed, because
3466 some condition is never satisfied or because it is after a
3467 procedure that never returns.
3469 It is possible for this option to produce a warning even though there
3470 are circumstances under which part of the affected line can be executed,
3471 so care should be taken when removing apparently-unreachable code.
3473 For instance, when a function is inlined, a warning may mean that the
3474 line is unreachable in only one inlined copy of the function.
3476 This option is not made part of @option{-Wall} because in a debugging
3477 version of a program there is often substantial code which checks
3478 correct functioning of the program and is, hopefully, unreachable
3479 because the program does work. Another common use of unreachable
3480 code is to provide behavior which is selectable at compile-time.
3484 Warn if a function can not be inlined and it was declared as inline.
3485 Even with this option, the compiler will not warn about failures to
3486 inline functions declared in system headers.
3488 The compiler uses a variety of heuristics to determine whether or not
3489 to inline a function. For example, the compiler takes into account
3490 the size of the function being inlined and the amount of inlining
3491 that has already been done in the current function. Therefore,
3492 seemingly insignificant changes in the source program can cause the
3493 warnings produced by @option{-Winline} to appear or disappear.
3495 @item -Wno-invalid-offsetof @r{(C++ only)}
3496 @opindex Wno-invalid-offsetof
3497 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3498 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3499 to a non-POD type is undefined. In existing C++ implementations,
3500 however, @samp{offsetof} typically gives meaningful results even when
3501 applied to certain kinds of non-POD types. (Such as a simple
3502 @samp{struct} that fails to be a POD type only by virtue of having a
3503 constructor.) This flag is for users who are aware that they are
3504 writing nonportable code and who have deliberately chosen to ignore the
3507 The restrictions on @samp{offsetof} may be relaxed in a future version
3508 of the C++ standard.
3510 @item -Wno-int-to-pointer-cast @r{(C only)}
3511 @opindex Wno-int-to-pointer-cast
3512 Suppress warnings from casts to pointer type of an integer of a
3515 @item -Wno-pointer-to-int-cast @r{(C only)}
3516 @opindex Wno-pointer-to-int-cast
3517 Suppress warnings from casts from a pointer to an integer type of a
3521 @opindex Winvalid-pch
3522 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3523 the search path but can't be used.
3527 @opindex Wno-long-long
3528 Warn if @samp{long long} type is used. This is default. To inhibit
3529 the warning messages, use @option{-Wno-long-long}. Flags
3530 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3531 only when @option{-pedantic} flag is used.
3533 @item -Wvariadic-macros
3534 @opindex Wvariadic-macros
3535 @opindex Wno-variadic-macros
3536 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3537 alternate syntax when in pedantic ISO C99 mode. This is default.
3538 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3540 @item -Wvolatile-register-var
3541 @opindex Wvolatile-register-var
3542 @opindex Wno-volatile-register-var
3543 Warn if a register variable is declared volatile. The volatile
3544 modifier does not inhibit all optimizations that may eliminate reads
3545 and/or writes to register variables.
3547 @item -Wdisabled-optimization
3548 @opindex Wdisabled-optimization
3549 Warn if a requested optimization pass is disabled. This warning does
3550 not generally indicate that there is anything wrong with your code; it
3551 merely indicates that GCC's optimizers were unable to handle the code
3552 effectively. Often, the problem is that your code is too big or too
3553 complex; GCC will refuse to optimize programs when the optimization
3554 itself is likely to take inordinate amounts of time.
3556 @item -Wpointer-sign
3557 @opindex Wpointer-sign
3558 @opindex Wno-pointer-sign
3559 Warn for pointer argument passing or assignment with different signedness.
3560 This option is only supported for C and Objective-C@. It is implied by
3561 @option{-Wall} and by @option{-pedantic}, which can be disabled with
3562 @option{-Wno-pointer-sign}.
3566 Make all warnings into errors.
3570 Make the specified warning into an errors. The specifier for a
3571 warning is appended, for example @option{-Werror=switch} turns the
3572 warnings controlled by @option{-Wswitch} into errors. This switch
3573 takes a negative form, to be used to negate @option{-Werror} for
3574 specific warnings, for example @option{-Wno-error=switch} makes
3575 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
3576 is in effect. You can use the @option{-fdiagnostics-show-option}
3577 option to have each controllable warning amended with the option which
3578 controls it, to determine what to use with this option.
3580 Note that specifying @option{-Werror=}@var{foo} automatically implies
3581 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
3584 @item -Wstack-protector
3585 @opindex Wstack-protector
3586 This option is only active when @option{-fstack-protector} is active. It
3587 warns about functions that will not be protected against stack smashing.
3589 @item -Wstring-literal-comparison
3590 @opindex Wstring-literal-comparison
3591 Warn about suspicious comparisons to string literal constants. In C,
3592 direct comparisons against the memory address of a string literal, such
3593 as @code{if (x == "abc")}, typically indicate a programmer error, and
3594 even when intentional, result in unspecified behavior and are not portable.
3595 Usually these warnings alert that the programmer intended to use
3596 @code{strcmp}. This warning is enabled by @option{-Wall}.
3598 @item -Woverlength-strings
3599 @opindex Woverlength-strings
3600 Warn about string constants which are longer than the ``minimum
3601 maximum'' length specified in the C standard. Modern compilers
3602 generally allow string constants which are much longer than the
3603 standard's minimum limit, but very portable programs should avoid
3604 using longer strings.
3606 The limit applies @emph{after} string constant concatenation, and does
3607 not count the trailing NUL@. In C89, the limit was 509 characters; in
3608 C99, it was raised to 4095. C++98 does not specify a normative
3609 minimum maximum, so we do not diagnose overlength strings in C++@.
3611 This option is implied by @option{-pedantic}, and can be disabled with
3612 @option{-Wno-overlength-strings}.
3615 @node Debugging Options
3616 @section Options for Debugging Your Program or GCC
3617 @cindex options, debugging
3618 @cindex debugging information options
3620 GCC has various special options that are used for debugging
3621 either your program or GCC:
3626 Produce debugging information in the operating system's native format
3627 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
3630 On most systems that use stabs format, @option{-g} enables use of extra
3631 debugging information that only GDB can use; this extra information
3632 makes debugging work better in GDB but will probably make other debuggers
3634 refuse to read the program. If you want to control for certain whether
3635 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3636 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3638 GCC allows you to use @option{-g} with
3639 @option{-O}. The shortcuts taken by optimized code may occasionally
3640 produce surprising results: some variables you declared may not exist
3641 at all; flow of control may briefly move where you did not expect it;
3642 some statements may not be executed because they compute constant
3643 results or their values were already at hand; some statements may
3644 execute in different places because they were moved out of loops.
3646 Nevertheless it proves possible to debug optimized output. This makes
3647 it reasonable to use the optimizer for programs that might have bugs.
3649 The following options are useful when GCC is generated with the
3650 capability for more than one debugging format.
3654 Produce debugging information for use by GDB@. This means to use the
3655 most expressive format available (DWARF 2, stabs, or the native format
3656 if neither of those are supported), including GDB extensions if at all
3661 Produce debugging information in stabs format (if that is supported),
3662 without GDB extensions. This is the format used by DBX on most BSD
3663 systems. On MIPS, Alpha and System V Release 4 systems this option
3664 produces stabs debugging output which is not understood by DBX or SDB@.
3665 On System V Release 4 systems this option requires the GNU assembler.
3667 @item -feliminate-unused-debug-symbols
3668 @opindex feliminate-unused-debug-symbols
3669 Produce debugging information in stabs format (if that is supported),
3670 for only symbols that are actually used.
3672 @item -femit-class-debug-always
3673 Instead of emitting debugging information for a C++ class in only one
3674 object file, emit it in all object files using the class. This option
3675 should be used only with debuggers that are unable to handle the way GCC
3676 normally emits debugging information for classes because using this
3677 option will increase the size of debugging information by as much as a
3682 Produce debugging information in stabs format (if that is supported),
3683 using GNU extensions understood only by the GNU debugger (GDB)@. The
3684 use of these extensions is likely to make other debuggers crash or
3685 refuse to read the program.
3689 Produce debugging information in COFF format (if that is supported).
3690 This is the format used by SDB on most System V systems prior to
3695 Produce debugging information in XCOFF format (if that is supported).
3696 This is the format used by the DBX debugger on IBM RS/6000 systems.
3700 Produce debugging information in XCOFF format (if that is supported),
3701 using GNU extensions understood only by the GNU debugger (GDB)@. The
3702 use of these extensions is likely to make other debuggers crash or
3703 refuse to read the program, and may cause assemblers other than the GNU
3704 assembler (GAS) to fail with an error.
3708 Produce debugging information in DWARF version 2 format (if that is
3709 supported). This is the format used by DBX on IRIX 6. With this
3710 option, GCC uses features of DWARF version 3 when they are useful;
3711 version 3 is upward compatible with version 2, but may still cause
3712 problems for older debuggers.
3716 Produce debugging information in VMS debug format (if that is
3717 supported). This is the format used by DEBUG on VMS systems.
3720 @itemx -ggdb@var{level}
3721 @itemx -gstabs@var{level}
3722 @itemx -gcoff@var{level}
3723 @itemx -gxcoff@var{level}
3724 @itemx -gvms@var{level}
3725 Request debugging information and also use @var{level} to specify how
3726 much information. The default level is 2.
3728 Level 0 produces no debug information at all. Thus, @option{-g0} negates
3731 Level 1 produces minimal information, enough for making backtraces in
3732 parts of the program that you don't plan to debug. This includes
3733 descriptions of functions and external variables, but no information
3734 about local variables and no line numbers.
3736 Level 3 includes extra information, such as all the macro definitions
3737 present in the program. Some debuggers support macro expansion when
3738 you use @option{-g3}.
3740 @option{-gdwarf-2} does not accept a concatenated debug level, because
3741 GCC used to support an option @option{-gdwarf} that meant to generate
3742 debug information in version 1 of the DWARF format (which is very
3743 different from version 2), and it would have been too confusing. That
3744 debug format is long obsolete, but the option cannot be changed now.
3745 Instead use an additional @option{-g@var{level}} option to change the
3746 debug level for DWARF2.
3748 @item -feliminate-dwarf2-dups
3749 @opindex feliminate-dwarf2-dups
3750 Compress DWARF2 debugging information by eliminating duplicated
3751 information about each symbol. This option only makes sense when
3752 generating DWARF2 debugging information with @option{-gdwarf-2}.
3754 @cindex @command{prof}
3757 Generate extra code to write profile information suitable for the
3758 analysis program @command{prof}. You must use this option when compiling
3759 the source files you want data about, and you must also use it when
3762 @cindex @command{gprof}
3765 Generate extra code to write profile information suitable for the
3766 analysis program @command{gprof}. You must use this option when compiling
3767 the source files you want data about, and you must also use it when
3772 Makes the compiler print out each function name as it is compiled, and
3773 print some statistics about each pass when it finishes.
3776 @opindex ftime-report
3777 Makes the compiler print some statistics about the time consumed by each
3778 pass when it finishes.
3781 @opindex fmem-report
3782 Makes the compiler print some statistics about permanent memory
3783 allocation when it finishes.
3785 @item -fpre-ipa-mem-report
3786 @opindex fpre-ipa-mem-report
3787 @item -fpost-ipa-mem-report
3788 @opindex fpost-ipa-mem-report
3789 Makes the compiler print some statistics about permanent memory
3790 allocation before or after interprocedural optimization.
3792 @item -fprofile-arcs
3793 @opindex fprofile-arcs
3794 Add code so that program flow @dfn{arcs} are instrumented. During
3795 execution the program records how many times each branch and call is
3796 executed and how many times it is taken or returns. When the compiled
3797 program exits it saves this data to a file called
3798 @file{@var{auxname}.gcda} for each source file. The data may be used for
3799 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3800 test coverage analysis (@option{-ftest-coverage}). Each object file's
3801 @var{auxname} is generated from the name of the output file, if
3802 explicitly specified and it is not the final executable, otherwise it is
3803 the basename of the source file. In both cases any suffix is removed
3804 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
3805 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3806 @xref{Cross-profiling}.
3808 @cindex @command{gcov}
3812 This option is used to compile and link code instrumented for coverage
3813 analysis. The option is a synonym for @option{-fprofile-arcs}
3814 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
3815 linking). See the documentation for those options for more details.
3820 Compile the source files with @option{-fprofile-arcs} plus optimization
3821 and code generation options. For test coverage analysis, use the
3822 additional @option{-ftest-coverage} option. You do not need to profile
3823 every source file in a program.
3826 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3827 (the latter implies the former).
3830 Run the program on a representative workload to generate the arc profile
3831 information. This may be repeated any number of times. You can run
3832 concurrent instances of your program, and provided that the file system
3833 supports locking, the data files will be correctly updated. Also
3834 @code{fork} calls are detected and correctly handled (double counting
3838 For profile-directed optimizations, compile the source files again with
3839 the same optimization and code generation options plus
3840 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3841 Control Optimization}).
3844 For test coverage analysis, use @command{gcov} to produce human readable
3845 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3846 @command{gcov} documentation for further information.
3850 With @option{-fprofile-arcs}, for each function of your program GCC
3851 creates a program flow graph, then finds a spanning tree for the graph.
3852 Only arcs that are not on the spanning tree have to be instrumented: the
3853 compiler adds code to count the number of times that these arcs are
3854 executed. When an arc is the only exit or only entrance to a block, the
3855 instrumentation code can be added to the block; otherwise, a new basic
3856 block must be created to hold the instrumentation code.
3859 @item -ftest-coverage
3860 @opindex ftest-coverage
3861 Produce a notes file that the @command{gcov} code-coverage utility
3862 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3863 show program coverage. Each source file's note file is called
3864 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3865 above for a description of @var{auxname} and instructions on how to
3866 generate test coverage data. Coverage data will match the source files
3867 more closely, if you do not optimize.
3869 @item -d@var{letters}
3870 @item -fdump-rtl-@var{pass}
3872 Says to make debugging dumps during compilation at times specified by
3873 @var{letters}. This is used for debugging the RTL-based passes of the
3874 compiler. The file names for most of the dumps are made by appending a
3875 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
3876 from the name of the output file, if explicitly specified and it is not
3877 an executable, otherwise it is the basename of the source file.
3879 Most debug dumps can be enabled either passing a letter to the @option{-d}
3880 option, or with a long @option{-fdump-rtl} switch; here are the possible
3881 letters for use in @var{letters} and @var{pass}, and their meanings:
3886 Annotate the assembler output with miscellaneous debugging information.
3889 @itemx -fdump-rtl-bbro
3891 @opindex fdump-rtl-bbro
3892 Dump after block reordering, to @file{@var{file}.148r.bbro}.
3895 @itemx -fdump-rtl-combine
3897 @opindex fdump-rtl-combine
3898 Dump after the RTL instruction combination pass, to the file
3899 @file{@var{file}.129r.combine}.
3902 @itemx -fdump-rtl-ce1
3903 @itemx -fdump-rtl-ce2
3905 @opindex fdump-rtl-ce1
3906 @opindex fdump-rtl-ce2
3907 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
3908 first if conversion, to the file @file{@var{file}.117r.ce1}. @option{-dC}
3909 and @option{-fdump-rtl-ce2} enable dumping after the second if
3910 conversion, to the file @file{@var{file}.130r.ce2}.
3913 @itemx -fdump-rtl-btl
3914 @itemx -fdump-rtl-dbr
3916 @opindex fdump-rtl-btl
3917 @opindex fdump-rtl-dbr
3918 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
3919 target load optimization, to @file{@var{file}.31.btl}. @option{-dd}
3920 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
3921 scheduling, to @file{@var{file}.36.dbr}.
3925 Dump all macro definitions, at the end of preprocessing, in addition to
3929 @itemx -fdump-rtl-ce3
3931 @opindex fdump-rtl-ce3
3932 Dump after the third if conversion, to @file{@var{file}.146r.ce3}.
3935 @itemx -fdump-rtl-cfg
3936 @itemx -fdump-rtl-life
3938 @opindex fdump-rtl-cfg
3939 @opindex fdump-rtl-life
3940 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
3941 and data flow analysis, to @file{@var{file}.116r.cfg}. @option{-df}
3942 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
3943 to @file{@var{file}.128r.life1} and @file{@var{file}.135r.life2}.
3946 @itemx -fdump-rtl-greg
3948 @opindex fdump-rtl-greg
3949 Dump after global register allocation, to @file{@var{file}.139r.greg}.
3952 @itemx -fdump-rtl-gcse
3953 @itemx -fdump-rtl-bypass
3955 @opindex fdump-rtl-gcse
3956 @opindex fdump-rtl-bypass
3957 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
3958 @file{@var{file}.114r.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
3959 enable dumping after jump bypassing and control flow optimizations, to
3960 @file{@var{file}.115r.bypass}.
3963 @itemx -fdump-rtl-eh
3965 @opindex fdump-rtl-eh
3966 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
3969 @itemx -fdump-rtl-sibling
3971 @opindex fdump-rtl-sibling
3972 Dump after sibling call optimizations, to @file{@var{file}.106r.sibling}.
3975 @itemx -fdump-rtl-jump
3977 @opindex fdump-rtl-jump
3978 Dump after the first jump optimization, to @file{@var{file}.112r.jump}.
3981 @itemx -fdump-rtl-stack
3983 @opindex fdump-rtl-stack
3984 Dump after conversion from GCC's "flat register file" registers to the
3985 x87's stack-like registers, to @file{@var{file}.152r.stack}.
3988 @itemx -fdump-rtl-lreg
3990 @opindex fdump-rtl-lreg
3991 Dump after local register allocation, to @file{@var{file}.138r.lreg}.
3994 @itemx -fdump-rtl-loop2
3996 @opindex fdump-rtl-loop2
3997 @option{-dL} and @option{-fdump-rtl-loop2} enable dumping after the
3998 loop optimization pass, to @file{@var{file}.119r.loop2},
3999 @file{@var{file}.120r.loop2_init},
4000 @file{@var{file}.121r.loop2_invariant}, and
4001 @file{@var{file}.125r.loop2_done}.
4004 @itemx -fdump-rtl-sms
4006 @opindex fdump-rtl-sms
4007 Dump after modulo scheduling, to @file{@var{file}.136r.sms}.
4010 @itemx -fdump-rtl-mach
4012 @opindex fdump-rtl-mach
4013 Dump after performing the machine dependent reorganization pass, to
4014 @file{@var{file}.155r.mach}.
4017 @itemx -fdump-rtl-rnreg
4019 @opindex fdump-rtl-rnreg
4020 Dump after register renumbering, to @file{@var{file}.147r.rnreg}.
4023 @itemx -fdump-rtl-regmove
4025 @opindex fdump-rtl-regmove
4026 Dump after the register move pass, to @file{@var{file}.132r.regmove}.
4029 @itemx -fdump-rtl-postreload
4031 @opindex fdump-rtl-postreload
4032 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
4035 @itemx -fdump-rtl-expand
4037 @opindex fdump-rtl-expand
4038 Dump after RTL generation, to @file{@var{file}.104r.expand}.
4041 @itemx -fdump-rtl-sched2
4043 @opindex fdump-rtl-sched2
4044 Dump after the second scheduling pass, to @file{@var{file}.149r.sched2}.
4047 @itemx -fdump-rtl-cse
4049 @opindex fdump-rtl-cse
4050 Dump after CSE (including the jump optimization that sometimes follows
4051 CSE), to @file{@var{file}.113r.cse}.
4054 @itemx -fdump-rtl-sched1
4056 @opindex fdump-rtl-sched1
4057 Dump after the first scheduling pass, to @file{@var{file}.136r.sched1}.
4060 @itemx -fdump-rtl-cse2
4062 @opindex fdump-rtl-cse2
4063 Dump after the second CSE pass (including the jump optimization that
4064 sometimes follows CSE), to @file{@var{file}.127r.cse2}.
4067 @itemx -fdump-rtl-tracer
4069 @opindex fdump-rtl-tracer
4070 Dump after running tracer, to @file{@var{file}.118r.tracer}.
4073 @itemx -fdump-rtl-vpt
4074 @itemx -fdump-rtl-vartrack
4076 @opindex fdump-rtl-vpt
4077 @opindex fdump-rtl-vartrack
4078 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
4079 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
4080 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
4081 to @file{@var{file}.154r.vartrack}.
4084 @itemx -fdump-rtl-flow2
4086 @opindex fdump-rtl-flow2
4087 Dump after the second flow pass, to @file{@var{file}.142r.flow2}.
4090 @itemx -fdump-rtl-peephole2
4092 @opindex fdump-rtl-peephole2
4093 Dump after the peephole pass, to @file{@var{file}.145r.peephole2}.
4096 @itemx -fdump-rtl-web
4098 @opindex fdump-rtl-web
4099 Dump after live range splitting, to @file{@var{file}.126r.web}.
4102 @itemx -fdump-rtl-all
4104 @opindex fdump-rtl-all
4105 Produce all the dumps listed above.
4109 Produce a core dump whenever an error occurs.
4113 Print statistics on memory usage, at the end of the run, to
4118 Annotate the assembler output with a comment indicating which
4119 pattern and alternative was used. The length of each instruction is
4124 Dump the RTL in the assembler output as a comment before each instruction.
4125 Also turns on @option{-dp} annotation.
4129 For each of the other indicated dump files (either with @option{-d} or
4130 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
4131 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
4135 Just generate RTL for a function instead of compiling it. Usually used
4136 with @samp{r} (@option{-fdump-rtl-expand}).
4140 Dump debugging information during parsing, to standard error.
4144 @opindex fdump-noaddr
4145 When doing debugging dumps (see @option{-d} option above), suppress
4146 address output. This makes it more feasible to use diff on debugging
4147 dumps for compiler invocations with different compiler binaries and/or
4148 different text / bss / data / heap / stack / dso start locations.
4150 @item -fdump-unnumbered
4151 @opindex fdump-unnumbered
4152 When doing debugging dumps (see @option{-d} option above), suppress instruction
4153 numbers, line number note and address output. This makes it more feasible to
4154 use diff on debugging dumps for compiler invocations with different
4155 options, in particular with and without @option{-g}.
4157 @item -fdump-translation-unit @r{(C++ only)}
4158 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
4159 @opindex fdump-translation-unit
4160 Dump a representation of the tree structure for the entire translation
4161 unit to a file. The file name is made by appending @file{.tu} to the
4162 source file name. If the @samp{-@var{options}} form is used, @var{options}
4163 controls the details of the dump as described for the
4164 @option{-fdump-tree} options.
4166 @item -fdump-class-hierarchy @r{(C++ only)}
4167 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
4168 @opindex fdump-class-hierarchy
4169 Dump a representation of each class's hierarchy and virtual function
4170 table layout to a file. The file name is made by appending @file{.class}
4171 to the source file name. If the @samp{-@var{options}} form is used,
4172 @var{options} controls the details of the dump as described for the
4173 @option{-fdump-tree} options.
4175 @item -fdump-ipa-@var{switch}
4177 Control the dumping at various stages of inter-procedural analysis
4178 language tree to a file. The file name is generated by appending a switch
4179 specific suffix to the source file name. The following dumps are possible:
4183 Enables all inter-procedural analysis dumps; currently the only produced
4184 dump is the @samp{cgraph} dump.
4187 Dumps information about call-graph optimization, unused function removal,
4188 and inlining decisions.
4191 @item -fdump-tree-@var{switch}
4192 @itemx -fdump-tree-@var{switch}-@var{options}
4194 Control the dumping at various stages of processing the intermediate
4195 language tree to a file. The file name is generated by appending a switch
4196 specific suffix to the source file name. If the @samp{-@var{options}}
4197 form is used, @var{options} is a list of @samp{-} separated options that
4198 control the details of the dump. Not all options are applicable to all
4199 dumps, those which are not meaningful will be ignored. The following
4200 options are available
4204 Print the address of each node. Usually this is not meaningful as it
4205 changes according to the environment and source file. Its primary use
4206 is for tying up a dump file with a debug environment.
4208 Inhibit dumping of members of a scope or body of a function merely
4209 because that scope has been reached. Only dump such items when they
4210 are directly reachable by some other path. When dumping pretty-printed
4211 trees, this option inhibits dumping the bodies of control structures.
4213 Print a raw representation of the tree. By default, trees are
4214 pretty-printed into a C-like representation.
4216 Enable more detailed dumps (not honored by every dump option).
4218 Enable dumping various statistics about the pass (not honored by every dump
4221 Enable showing basic block boundaries (disabled in raw dumps).
4223 Enable showing virtual operands for every statement.
4225 Enable showing line numbers for statements.
4227 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4229 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4232 The following tree dumps are possible:
4236 Dump before any tree based optimization, to @file{@var{file}.original}.
4239 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4242 Dump after function inlining, to @file{@var{file}.inlined}.
4245 @opindex fdump-tree-gimple
4246 Dump each function before and after the gimplification pass to a file. The
4247 file name is made by appending @file{.gimple} to the source file name.
4250 @opindex fdump-tree-cfg
4251 Dump the control flow graph of each function to a file. The file name is
4252 made by appending @file{.cfg} to the source file name.
4255 @opindex fdump-tree-vcg
4256 Dump the control flow graph of each function to a file in VCG format. The
4257 file name is made by appending @file{.vcg} to the source file name. Note
4258 that if the file contains more than one function, the generated file cannot
4259 be used directly by VCG@. You will need to cut and paste each function's
4260 graph into its own separate file first.
4263 @opindex fdump-tree-ch
4264 Dump each function after copying loop headers. The file name is made by
4265 appending @file{.ch} to the source file name.
4268 @opindex fdump-tree-ssa
4269 Dump SSA related information to a file. The file name is made by appending
4270 @file{.ssa} to the source file name.
4273 @opindex fdump-tree-salias
4274 Dump structure aliasing variable information to a file. This file name
4275 is made by appending @file{.salias} to the source file name.
4278 @opindex fdump-tree-alias
4279 Dump aliasing information for each function. The file name is made by
4280 appending @file{.alias} to the source file name.
4283 @opindex fdump-tree-ccp
4284 Dump each function after CCP@. The file name is made by appending
4285 @file{.ccp} to the source file name.
4288 @opindex fdump-tree-storeccp
4289 Dump each function after STORE-CCP. The file name is made by appending
4290 @file{.storeccp} to the source file name.
4293 @opindex fdump-tree-pre
4294 Dump trees after partial redundancy elimination. The file name is made
4295 by appending @file{.pre} to the source file name.
4298 @opindex fdump-tree-fre
4299 Dump trees after full redundancy elimination. The file name is made
4300 by appending @file{.fre} to the source file name.
4303 @opindex fdump-tree-copyprop
4304 Dump trees after copy propagation. The file name is made
4305 by appending @file{.copyprop} to the source file name.
4307 @item store_copyprop
4308 @opindex fdump-tree-store_copyprop
4309 Dump trees after store copy-propagation. The file name is made
4310 by appending @file{.store_copyprop} to the source file name.
4313 @opindex fdump-tree-dce
4314 Dump each function after dead code elimination. The file name is made by
4315 appending @file{.dce} to the source file name.
4318 @opindex fdump-tree-mudflap
4319 Dump each function after adding mudflap instrumentation. The file name is
4320 made by appending @file{.mudflap} to the source file name.
4323 @opindex fdump-tree-sra
4324 Dump each function after performing scalar replacement of aggregates. The
4325 file name is made by appending @file{.sra} to the source file name.
4328 @opindex fdump-tree-sink
4329 Dump each function after performing code sinking. The file name is made
4330 by appending @file{.sink} to the source file name.
4333 @opindex fdump-tree-dom
4334 Dump each function after applying dominator tree optimizations. The file
4335 name is made by appending @file{.dom} to the source file name.
4338 @opindex fdump-tree-dse
4339 Dump each function after applying dead store elimination. The file
4340 name is made by appending @file{.dse} to the source file name.
4343 @opindex fdump-tree-phiopt
4344 Dump each function after optimizing PHI nodes into straightline code. The file
4345 name is made by appending @file{.phiopt} to the source file name.
4348 @opindex fdump-tree-forwprop
4349 Dump each function after forward propagating single use variables. The file
4350 name is made by appending @file{.forwprop} to the source file name.
4353 @opindex fdump-tree-copyrename
4354 Dump each function after applying the copy rename optimization. The file
4355 name is made by appending @file{.copyrename} to the source file name.
4358 @opindex fdump-tree-nrv
4359 Dump each function after applying the named return value optimization on
4360 generic trees. The file name is made by appending @file{.nrv} to the source
4364 @opindex fdump-tree-vect
4365 Dump each function after applying vectorization of loops. The file name is
4366 made by appending @file{.vect} to the source file name.
4369 @opindex fdump-tree-vrp
4370 Dump each function after Value Range Propagation (VRP). The file name
4371 is made by appending @file{.vrp} to the source file name.
4374 @opindex fdump-tree-all
4375 Enable all the available tree dumps with the flags provided in this option.
4378 @item -ftree-vectorizer-verbose=@var{n}
4379 @opindex ftree-vectorizer-verbose
4380 This option controls the amount of debugging output the vectorizer prints.
4381 This information is written to standard error, unless
4382 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
4383 in which case it is output to the usual dump listing file, @file{.vect}.
4384 For @var{n}=0 no diagnostic information is reported.
4385 If @var{n}=1 the vectorizer reports each loop that got vectorized,
4386 and the total number of loops that got vectorized.
4387 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
4388 the first analysis phase (vect_analyze_loop_form) - i.e. countable,
4389 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
4390 level that @option{-fdump-tree-vect-stats} uses.
4391 Higher verbosity levels mean either more information dumped for each
4392 reported loop, or same amount of information reported for more loops:
4393 If @var{n}=3, alignment related information is added to the reports.
4394 If @var{n}=4, data-references related information (e.g. memory dependences,
4395 memory access-patterns) is added to the reports.
4396 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
4397 that did not pass the first analysis phase (i.e. may not be countable, or
4398 may have complicated control-flow).
4399 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
4400 For @var{n}=7, all the information the vectorizer generates during its
4401 analysis and transformation is reported. This is the same verbosity level
4402 that @option{-fdump-tree-vect-details} uses.
4404 @item -frandom-seed=@var{string}
4405 @opindex frandom-string
4406 This option provides a seed that GCC uses when it would otherwise use
4407 random numbers. It is used to generate certain symbol names
4408 that have to be different in every compiled file. It is also used to
4409 place unique stamps in coverage data files and the object files that
4410 produce them. You can use the @option{-frandom-seed} option to produce
4411 reproducibly identical object files.
4413 The @var{string} should be different for every file you compile.
4415 @item -fsched-verbose=@var{n}
4416 @opindex fsched-verbose
4417 On targets that use instruction scheduling, this option controls the
4418 amount of debugging output the scheduler prints. This information is
4419 written to standard error, unless @option{-dS} or @option{-dR} is
4420 specified, in which case it is output to the usual dump
4421 listing file, @file{.sched} or @file{.sched2} respectively. However
4422 for @var{n} greater than nine, the output is always printed to standard
4425 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4426 same information as @option{-dRS}. For @var{n} greater than one, it
4427 also output basic block probabilities, detailed ready list information
4428 and unit/insn info. For @var{n} greater than two, it includes RTL
4429 at abort point, control-flow and regions info. And for @var{n} over
4430 four, @option{-fsched-verbose} also includes dependence info.
4434 Store the usual ``temporary'' intermediate files permanently; place them
4435 in the current directory and name them based on the source file. Thus,
4436 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4437 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4438 preprocessed @file{foo.i} output file even though the compiler now
4439 normally uses an integrated preprocessor.
4441 When used in combination with the @option{-x} command line option,
4442 @option{-save-temps} is sensible enough to avoid over writing an
4443 input source file with the same extension as an intermediate file.
4444 The corresponding intermediate file may be obtained by renaming the
4445 source file before using @option{-save-temps}.
4449 Report the CPU time taken by each subprocess in the compilation
4450 sequence. For C source files, this is the compiler proper and assembler
4451 (plus the linker if linking is done). The output looks like this:
4458 The first number on each line is the ``user time'', that is time spent
4459 executing the program itself. The second number is ``system time'',
4460 time spent executing operating system routines on behalf of the program.
4461 Both numbers are in seconds.
4463 @item -fvar-tracking
4464 @opindex fvar-tracking
4465 Run variable tracking pass. It computes where variables are stored at each
4466 position in code. Better debugging information is then generated
4467 (if the debugging information format supports this information).
4469 It is enabled by default when compiling with optimization (@option{-Os},
4470 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
4471 the debug info format supports it.
4473 @item -print-file-name=@var{library}
4474 @opindex print-file-name
4475 Print the full absolute name of the library file @var{library} that
4476 would be used when linking---and don't do anything else. With this
4477 option, GCC does not compile or link anything; it just prints the
4480 @item -print-multi-directory
4481 @opindex print-multi-directory
4482 Print the directory name corresponding to the multilib selected by any
4483 other switches present in the command line. This directory is supposed
4484 to exist in @env{GCC_EXEC_PREFIX}.
4486 @item -print-multi-lib
4487 @opindex print-multi-lib
4488 Print the mapping from multilib directory names to compiler switches
4489 that enable them. The directory name is separated from the switches by
4490 @samp{;}, and each switch starts with an @samp{@@} instead of the
4491 @samp{-}, without spaces between multiple switches. This is supposed to
4492 ease shell-processing.
4494 @item -print-prog-name=@var{program}
4495 @opindex print-prog-name
4496 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
4498 @item -print-libgcc-file-name
4499 @opindex print-libgcc-file-name
4500 Same as @option{-print-file-name=libgcc.a}.
4502 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
4503 but you do want to link with @file{libgcc.a}. You can do
4506 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
4509 @item -print-search-dirs
4510 @opindex print-search-dirs
4511 Print the name of the configured installation directory and a list of
4512 program and library directories @command{gcc} will search---and don't do anything else.
4514 This is useful when @command{gcc} prints the error message
4515 @samp{installation problem, cannot exec cpp0: No such file or directory}.
4516 To resolve this you either need to put @file{cpp0} and the other compiler
4517 components where @command{gcc} expects to find them, or you can set the environment
4518 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
4519 Don't forget the trailing @samp{/}.
4520 @xref{Environment Variables}.
4523 @opindex dumpmachine
4524 Print the compiler's target machine (for example,
4525 @samp{i686-pc-linux-gnu})---and don't do anything else.
4528 @opindex dumpversion
4529 Print the compiler version (for example, @samp{3.0})---and don't do
4534 Print the compiler's built-in specs---and don't do anything else. (This
4535 is used when GCC itself is being built.) @xref{Spec Files}.
4537 @item -feliminate-unused-debug-types
4538 @opindex feliminate-unused-debug-types
4539 Normally, when producing DWARF2 output, GCC will emit debugging
4540 information for all types declared in a compilation
4541 unit, regardless of whether or not they are actually used
4542 in that compilation unit. Sometimes this is useful, such as
4543 if, in the debugger, you want to cast a value to a type that is
4544 not actually used in your program (but is declared). More often,
4545 however, this results in a significant amount of wasted space.
4546 With this option, GCC will avoid producing debug symbol output
4547 for types that are nowhere used in the source file being compiled.
4550 @node Optimize Options
4551 @section Options That Control Optimization
4552 @cindex optimize options
4553 @cindex options, optimization
4555 These options control various sorts of optimizations.
4557 Without any optimization option, the compiler's goal is to reduce the
4558 cost of compilation and to make debugging produce the expected
4559 results. Statements are independent: if you stop the program with a
4560 breakpoint between statements, you can then assign a new value to any
4561 variable or change the program counter to any other statement in the
4562 function and get exactly the results you would expect from the source
4565 Turning on optimization flags makes the compiler attempt to improve
4566 the performance and/or code size at the expense of compilation time
4567 and possibly the ability to debug the program.
4569 The compiler performs optimization based on the knowledge it has of
4570 the program. Optimization levels @option{-O} and above, in
4571 particular, enable @emph{unit-at-a-time} mode, which allows the
4572 compiler to consider information gained from later functions in
4573 the file when compiling a function. Compiling multiple files at
4574 once to a single output file in @emph{unit-at-a-time} mode allows
4575 the compiler to use information gained from all of the files when
4576 compiling each of them.
4578 Not all optimizations are controlled directly by a flag. Only
4579 optimizations that have a flag are listed.
4586 Optimize. Optimizing compilation takes somewhat more time, and a lot
4587 more memory for a large function.
4589 With @option{-O}, the compiler tries to reduce code size and execution
4590 time, without performing any optimizations that take a great deal of
4593 @option{-O} turns on the following optimization flags:
4594 @gccoptlist{-fdefer-pop @gol
4595 -fdelayed-branch @gol
4596 -fguess-branch-probability @gol
4597 -fcprop-registers @gol
4598 -fif-conversion @gol
4599 -fif-conversion2 @gol
4602 -ftree-dominator-opts @gol
4607 -ftree-copyrename @gol
4610 -funit-at-a-time @gol
4613 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
4614 where doing so does not interfere with debugging.
4618 Optimize even more. GCC performs nearly all supported optimizations
4619 that do not involve a space-speed tradeoff. The compiler does not
4620 perform loop unrolling or function inlining when you specify @option{-O2}.
4621 As compared to @option{-O}, this option increases both compilation time
4622 and the performance of the generated code.
4624 @option{-O2} turns on all optimization flags specified by @option{-O}. It
4625 also turns on the following optimization flags:
4626 @gccoptlist{-fthread-jumps @gol
4628 -foptimize-sibling-calls @gol
4629 -fcse-follow-jumps -fcse-skip-blocks @gol
4630 -fgcse -fgcse-lm @gol
4631 -fexpensive-optimizations @gol
4632 -frerun-cse-after-loop @gol
4635 -fschedule-insns -fschedule-insns2 @gol
4636 -fsched-interblock -fsched-spec @gol
4638 -fstrict-aliasing -fstrict-overflow @gol
4639 -fdelete-null-pointer-checks @gol
4640 -freorder-blocks -freorder-functions @gol
4641 -falign-functions -falign-jumps @gol
4642 -falign-loops -falign-labels @gol
4646 Please note the warning under @option{-fgcse} about
4647 invoking @option{-O2} on programs that use computed gotos.
4651 Optimize yet more. @option{-O3} turns on all optimizations specified by
4652 @option{-O2} and also turns on the @option{-finline-functions},
4653 @option{-funswitch-loops} and @option{-fgcse-after-reload} options.
4657 Reduce compilation time and make debugging produce the expected
4658 results. This is the default.
4662 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
4663 do not typically increase code size. It also performs further
4664 optimizations designed to reduce code size.
4666 @option{-Os} disables the following optimization flags:
4667 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
4668 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
4669 -fprefetch-loop-arrays -ftree-vect-loop-version}
4671 If you use multiple @option{-O} options, with or without level numbers,
4672 the last such option is the one that is effective.
4675 Options of the form @option{-f@var{flag}} specify machine-independent
4676 flags. Most flags have both positive and negative forms; the negative
4677 form of @option{-ffoo} would be @option{-fno-foo}. In the table
4678 below, only one of the forms is listed---the one you typically will
4679 use. You can figure out the other form by either removing @samp{no-}
4682 The following options control specific optimizations. They are either
4683 activated by @option{-O} options or are related to ones that are. You
4684 can use the following flags in the rare cases when ``fine-tuning'' of
4685 optimizations to be performed is desired.
4688 @item -fno-default-inline
4689 @opindex fno-default-inline
4690 Do not make member functions inline by default merely because they are
4691 defined inside the class scope (C++ only). Otherwise, when you specify
4692 @w{@option{-O}}, member functions defined inside class scope are compiled
4693 inline by default; i.e., you don't need to add @samp{inline} in front of
4694 the member function name.
4696 @item -fno-defer-pop
4697 @opindex fno-defer-pop
4698 Always pop the arguments to each function call as soon as that function
4699 returns. For machines which must pop arguments after a function call,
4700 the compiler normally lets arguments accumulate on the stack for several
4701 function calls and pops them all at once.
4703 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4706 @opindex fforce-addr
4707 Force memory address constants to be copied into registers before
4708 doing arithmetic on them.
4710 @item -fforward-propagate
4711 @opindex fforward-propagate
4712 Perform a forward propagation pass on RTL. The pass tries to combine two
4713 instructions and checks if the result can be simplified. If loop unrolling
4714 is active, two passes are performed and the second is scheduled after
4717 This option is enabled by default at optimization levels @option{-O2},
4718 @option{-O3}, @option{-Os}.
4720 @item -fomit-frame-pointer
4721 @opindex fomit-frame-pointer
4722 Don't keep the frame pointer in a register for functions that
4723 don't need one. This avoids the instructions to save, set up and
4724 restore frame pointers; it also makes an extra register available
4725 in many functions. @strong{It also makes debugging impossible on
4728 On some machines, such as the VAX, this flag has no effect, because
4729 the standard calling sequence automatically handles the frame pointer
4730 and nothing is saved by pretending it doesn't exist. The
4731 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
4732 whether a target machine supports this flag. @xref{Registers,,Register
4733 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
4735 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4737 @item -foptimize-sibling-calls
4738 @opindex foptimize-sibling-calls
4739 Optimize sibling and tail recursive calls.
4741 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4745 Don't pay attention to the @code{inline} keyword. Normally this option
4746 is used to keep the compiler from expanding any functions inline.
4747 Note that if you are not optimizing, no functions can be expanded inline.
4749 @item -finline-functions
4750 @opindex finline-functions
4751 Integrate all simple functions into their callers. The compiler
4752 heuristically decides which functions are simple enough to be worth
4753 integrating in this way.
4755 If all calls to a given function are integrated, and the function is
4756 declared @code{static}, then the function is normally not output as
4757 assembler code in its own right.
4759 Enabled at level @option{-O3}.
4761 @item -finline-functions-called-once
4762 @opindex finline-functions-called-once
4763 Consider all @code{static} functions called once for inlining into their
4764 caller even if they are not marked @code{inline}. If a call to a given
4765 function is integrated, then the function is not output as assembler code
4768 Enabled if @option{-funit-at-a-time} is enabled.
4770 @item -fearly-inlining
4771 @opindex fearly-inlining
4772 Inline functions marked by @code{always_inline} and functions whose body seems
4773 smaller than the function call overhead early before doing
4774 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
4775 makes profiling significantly cheaper and usually inlining faster on programs
4776 having large chains of nested wrapper functions.
4780 @item -finline-limit=@var{n}
4781 @opindex finline-limit
4782 By default, GCC limits the size of functions that can be inlined. This flag
4783 allows the control of this limit for functions that are explicitly marked as
4784 inline (i.e., marked with the inline keyword or defined within the class
4785 definition in c++). @var{n} is the size of functions that can be inlined in
4786 number of pseudo instructions (not counting parameter handling). The default
4787 value of @var{n} is 600.
4788 Increasing this value can result in more inlined code at
4789 the cost of compilation time and memory consumption. Decreasing usually makes
4790 the compilation faster and less code will be inlined (which presumably
4791 means slower programs). This option is particularly useful for programs that
4792 use inlining heavily such as those based on recursive templates with C++.
4794 Inlining is actually controlled by a number of parameters, which may be
4795 specified individually by using @option{--param @var{name}=@var{value}}.
4796 The @option{-finline-limit=@var{n}} option sets some of these parameters
4800 @item max-inline-insns-single
4801 is set to @var{n}/2.
4802 @item max-inline-insns-auto
4803 is set to @var{n}/2.
4804 @item min-inline-insns
4805 is set to 130 or @var{n}/4, whichever is smaller.
4806 @item max-inline-insns-rtl
4810 See below for a documentation of the individual
4811 parameters controlling inlining.
4813 @emph{Note:} pseudo instruction represents, in this particular context, an
4814 abstract measurement of function's size. In no way does it represent a count
4815 of assembly instructions and as such its exact meaning might change from one
4816 release to an another.
4818 @item -fkeep-inline-functions
4819 @opindex fkeep-inline-functions
4820 In C, emit @code{static} functions that are declared @code{inline}
4821 into the object file, even if the function has been inlined into all
4822 of its callers. This switch does not affect functions using the
4823 @code{extern inline} extension in GNU C89@. In C++, emit any and all
4824 inline functions into the object file.
4826 @item -fkeep-static-consts
4827 @opindex fkeep-static-consts
4828 Emit variables declared @code{static const} when optimization isn't turned
4829 on, even if the variables aren't referenced.
4831 GCC enables this option by default. If you want to force the compiler to
4832 check if the variable was referenced, regardless of whether or not
4833 optimization is turned on, use the @option{-fno-keep-static-consts} option.
4835 @item -fmerge-constants
4836 Attempt to merge identical constants (string constants and floating point
4837 constants) across compilation units.
4839 This option is the default for optimized compilation if the assembler and
4840 linker support it. Use @option{-fno-merge-constants} to inhibit this
4843 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4845 @item -fmerge-all-constants
4846 Attempt to merge identical constants and identical variables.
4848 This option implies @option{-fmerge-constants}. In addition to
4849 @option{-fmerge-constants} this considers e.g.@: even constant initialized
4850 arrays or initialized constant variables with integral or floating point
4851 types. Languages like C or C++ require each non-automatic variable to
4852 have distinct location, so using this option will result in non-conforming
4855 @item -fmodulo-sched
4856 @opindex fmodulo-sched
4857 Perform swing modulo scheduling immediately before the first scheduling
4858 pass. This pass looks at innermost loops and reorders their
4859 instructions by overlapping different iterations.
4861 @item -fno-branch-count-reg
4862 @opindex fno-branch-count-reg
4863 Do not use ``decrement and branch'' instructions on a count register,
4864 but instead generate a sequence of instructions that decrement a
4865 register, compare it against zero, then branch based upon the result.
4866 This option is only meaningful on architectures that support such
4867 instructions, which include x86, PowerPC, IA-64 and S/390.
4869 The default is @option{-fbranch-count-reg}.
4871 @item -fno-function-cse
4872 @opindex fno-function-cse
4873 Do not put function addresses in registers; make each instruction that
4874 calls a constant function contain the function's address explicitly.
4876 This option results in less efficient code, but some strange hacks
4877 that alter the assembler output may be confused by the optimizations
4878 performed when this option is not used.
4880 The default is @option{-ffunction-cse}
4882 @item -fno-zero-initialized-in-bss
4883 @opindex fno-zero-initialized-in-bss
4884 If the target supports a BSS section, GCC by default puts variables that
4885 are initialized to zero into BSS@. This can save space in the resulting
4888 This option turns off this behavior because some programs explicitly
4889 rely on variables going to the data section. E.g., so that the
4890 resulting executable can find the beginning of that section and/or make
4891 assumptions based on that.
4893 The default is @option{-fzero-initialized-in-bss}.
4895 @item -fbounds-check
4896 @opindex fbounds-check
4897 For front-ends that support it, generate additional code to check that
4898 indices used to access arrays are within the declared range. This is
4899 currently only supported by the Java and Fortran front-ends, where
4900 this option defaults to true and false respectively.
4902 @item -fmudflap -fmudflapth -fmudflapir
4906 @cindex bounds checking
4908 For front-ends that support it (C and C++), instrument all risky
4909 pointer/array dereferencing operations, some standard library
4910 string/heap functions, and some other associated constructs with
4911 range/validity tests. Modules so instrumented should be immune to
4912 buffer overflows, invalid heap use, and some other classes of C/C++
4913 programming errors. The instrumentation relies on a separate runtime
4914 library (@file{libmudflap}), which will be linked into a program if
4915 @option{-fmudflap} is given at link time. Run-time behavior of the
4916 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
4917 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
4920 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
4921 link if your program is multi-threaded. Use @option{-fmudflapir}, in
4922 addition to @option{-fmudflap} or @option{-fmudflapth}, if
4923 instrumentation should ignore pointer reads. This produces less
4924 instrumentation (and therefore faster execution) and still provides
4925 some protection against outright memory corrupting writes, but allows
4926 erroneously read data to propagate within a program.
4928 @item -fthread-jumps
4929 @opindex fthread-jumps
4930 Perform optimizations where we check to see if a jump branches to a
4931 location where another comparison subsumed by the first is found. If
4932 so, the first branch is redirected to either the destination of the
4933 second branch or a point immediately following it, depending on whether
4934 the condition is known to be true or false.
4936 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4938 @item -fcse-follow-jumps
4939 @opindex fcse-follow-jumps
4940 In common subexpression elimination, scan through jump instructions
4941 when the target of the jump is not reached by any other path. For
4942 example, when CSE encounters an @code{if} statement with an
4943 @code{else} clause, CSE will follow the jump when the condition
4946 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4948 @item -fcse-skip-blocks
4949 @opindex fcse-skip-blocks
4950 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
4951 follow jumps which conditionally skip over blocks. When CSE
4952 encounters a simple @code{if} statement with no else clause,
4953 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
4954 body of the @code{if}.
4956 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4958 @item -frerun-cse-after-loop
4959 @opindex frerun-cse-after-loop
4960 Re-run common subexpression elimination after loop optimizations has been
4963 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4967 Perform a global common subexpression elimination pass.
4968 This pass also performs global constant and copy propagation.
4970 @emph{Note:} When compiling a program using computed gotos, a GCC
4971 extension, you may get better runtime performance if you disable
4972 the global common subexpression elimination pass by adding
4973 @option{-fno-gcse} to the command line.
4975 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4979 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
4980 attempt to move loads which are only killed by stores into themselves. This
4981 allows a loop containing a load/store sequence to be changed to a load outside
4982 the loop, and a copy/store within the loop.
4984 Enabled by default when gcse is enabled.
4988 When @option{-fgcse-sm} is enabled, a store motion pass is run after
4989 global common subexpression elimination. This pass will attempt to move
4990 stores out of loops. When used in conjunction with @option{-fgcse-lm},
4991 loops containing a load/store sequence can be changed to a load before
4992 the loop and a store after the loop.
4994 Not enabled at any optimization level.
4998 When @option{-fgcse-las} is enabled, the global common subexpression
4999 elimination pass eliminates redundant loads that come after stores to the
5000 same memory location (both partial and full redundancies).
5002 Not enabled at any optimization level.
5004 @item -fgcse-after-reload
5005 @opindex fgcse-after-reload
5006 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
5007 pass is performed after reload. The purpose of this pass is to cleanup
5010 @item -funsafe-loop-optimizations
5011 @opindex funsafe-loop-optimizations
5012 If given, the loop optimizer will assume that loop indices do not
5013 overflow, and that the loops with nontrivial exit condition are not
5014 infinite. This enables a wider range of loop optimizations even if
5015 the loop optimizer itself cannot prove that these assumptions are valid.
5016 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
5017 if it finds this kind of loop.
5019 @item -fcrossjumping
5020 @opindex crossjumping
5021 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
5022 resulting code may or may not perform better than without cross-jumping.
5024 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5026 @item -fif-conversion
5027 @opindex if-conversion
5028 Attempt to transform conditional jumps into branch-less equivalents. This
5029 include use of conditional moves, min, max, set flags and abs instructions, and
5030 some tricks doable by standard arithmetics. The use of conditional execution
5031 on chips where it is available is controlled by @code{if-conversion2}.
5033 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5035 @item -fif-conversion2
5036 @opindex if-conversion2
5037 Use conditional execution (where available) to transform conditional jumps into
5038 branch-less equivalents.
5040 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5042 @item -fdelete-null-pointer-checks
5043 @opindex fdelete-null-pointer-checks
5044 Use global dataflow analysis to identify and eliminate useless checks
5045 for null pointers. The compiler assumes that dereferencing a null
5046 pointer would have halted the program. If a pointer is checked after
5047 it has already been dereferenced, it cannot be null.
5049 In some environments, this assumption is not true, and programs can
5050 safely dereference null pointers. Use
5051 @option{-fno-delete-null-pointer-checks} to disable this optimization
5052 for programs which depend on that behavior.
5054 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5056 @item -fexpensive-optimizations
5057 @opindex fexpensive-optimizations
5058 Perform a number of minor optimizations that are relatively expensive.
5060 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5062 @item -foptimize-register-move
5064 @opindex foptimize-register-move
5066 Attempt to reassign register numbers in move instructions and as
5067 operands of other simple instructions in order to maximize the amount of
5068 register tying. This is especially helpful on machines with two-operand
5071 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
5074 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5076 @item -fdelayed-branch
5077 @opindex fdelayed-branch
5078 If supported for the target machine, attempt to reorder instructions
5079 to exploit instruction slots available after delayed branch
5082 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5084 @item -fschedule-insns
5085 @opindex fschedule-insns
5086 If supported for the target machine, attempt to reorder instructions to
5087 eliminate execution stalls due to required data being unavailable. This
5088 helps machines that have slow floating point or memory load instructions
5089 by allowing other instructions to be issued until the result of the load
5090 or floating point instruction is required.
5092 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5094 @item -fschedule-insns2
5095 @opindex fschedule-insns2
5096 Similar to @option{-fschedule-insns}, but requests an additional pass of
5097 instruction scheduling after register allocation has been done. This is
5098 especially useful on machines with a relatively small number of
5099 registers and where memory load instructions take more than one cycle.
5101 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5103 @item -fno-sched-interblock
5104 @opindex fno-sched-interblock
5105 Don't schedule instructions across basic blocks. This is normally
5106 enabled by default when scheduling before register allocation, i.e.@:
5107 with @option{-fschedule-insns} or at @option{-O2} or higher.
5109 @item -fno-sched-spec
5110 @opindex fno-sched-spec
5111 Don't allow speculative motion of non-load instructions. This is normally
5112 enabled by default when scheduling before register allocation, i.e.@:
5113 with @option{-fschedule-insns} or at @option{-O2} or higher.
5115 @item -fsched-spec-load
5116 @opindex fsched-spec-load
5117 Allow speculative motion of some load instructions. This only makes
5118 sense when scheduling before register allocation, i.e.@: with
5119 @option{-fschedule-insns} or at @option{-O2} or higher.
5121 @item -fsched-spec-load-dangerous
5122 @opindex fsched-spec-load-dangerous
5123 Allow speculative motion of more load instructions. This only makes
5124 sense when scheduling before register allocation, i.e.@: with
5125 @option{-fschedule-insns} or at @option{-O2} or higher.
5127 @item -fsched-stalled-insns=@var{n}
5128 @opindex fsched-stalled-insns
5129 Define how many insns (if any) can be moved prematurely from the queue
5130 of stalled insns into the ready list, during the second scheduling pass.
5132 @item -fsched-stalled-insns-dep=@var{n}
5133 @opindex fsched-stalled-insns-dep
5134 Define how many insn groups (cycles) will be examined for a dependency
5135 on a stalled insn that is candidate for premature removal from the queue
5136 of stalled insns. Has an effect only during the second scheduling pass,
5137 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
5139 @item -fsched2-use-superblocks
5140 @opindex fsched2-use-superblocks
5141 When scheduling after register allocation, do use superblock scheduling
5142 algorithm. Superblock scheduling allows motion across basic block boundaries
5143 resulting on faster schedules. This option is experimental, as not all machine
5144 descriptions used by GCC model the CPU closely enough to avoid unreliable
5145 results from the algorithm.
5147 This only makes sense when scheduling after register allocation, i.e.@: with
5148 @option{-fschedule-insns2} or at @option{-O2} or higher.
5150 @item -fsched2-use-traces
5151 @opindex fsched2-use-traces
5152 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
5153 allocation and additionally perform code duplication in order to increase the
5154 size of superblocks using tracer pass. See @option{-ftracer} for details on
5157 This mode should produce faster but significantly longer programs. Also
5158 without @option{-fbranch-probabilities} the traces constructed may not
5159 match the reality and hurt the performance. This only makes
5160 sense when scheduling after register allocation, i.e.@: with
5161 @option{-fschedule-insns2} or at @option{-O2} or higher.
5165 Eliminates redundant extension instructions and move the non redundant
5166 ones to optimal placement using LCM.
5168 @item -freschedule-modulo-scheduled-loops
5169 @opindex fscheduling-in-modulo-scheduled-loops
5170 The modulo scheduling comes before the traditional scheduling, if a loop was modulo scheduled
5171 we may want to prevent the later scheduling passes from changing its schedule, we use this
5172 option to control that.
5174 @item -fcaller-saves
5175 @opindex fcaller-saves
5176 Enable values to be allocated in registers that will be clobbered by
5177 function calls, by emitting extra instructions to save and restore the
5178 registers around such calls. Such allocation is done only when it
5179 seems to result in better code than would otherwise be produced.
5181 This option is always enabled by default on certain machines, usually
5182 those which have no call-preserved registers to use instead.
5184 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5187 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
5188 enabled by default at @option{-O2} and @option{-O3}.
5191 Perform Full Redundancy Elimination (FRE) on trees. The difference
5192 between FRE and PRE is that FRE only considers expressions
5193 that are computed on all paths leading to the redundant computation.
5194 This analysis faster than PRE, though it exposes fewer redundancies.
5195 This flag is enabled by default at @option{-O} and higher.
5197 @item -ftree-copy-prop
5198 Perform copy propagation on trees. This pass eliminates unnecessary
5199 copy operations. This flag is enabled by default at @option{-O} and
5202 @item -ftree-store-copy-prop
5203 Perform copy propagation of memory loads and stores. This pass
5204 eliminates unnecessary copy operations in memory references
5205 (structures, global variables, arrays, etc). This flag is enabled by
5206 default at @option{-O2} and higher.
5209 Perform structural alias analysis on trees. This flag
5210 is enabled by default at @option{-O} and higher.
5213 Perform interprocedural pointer analysis.
5216 Perform forward store motion on trees. This flag is
5217 enabled by default at @option{-O} and higher.
5220 Perform sparse conditional constant propagation (CCP) on trees. This
5221 pass only operates on local scalar variables and is enabled by default
5222 at @option{-O} and higher.
5224 @item -ftree-store-ccp
5225 Perform sparse conditional constant propagation (CCP) on trees. This
5226 pass operates on both local scalar variables and memory stores and
5227 loads (global variables, structures, arrays, etc). This flag is
5228 enabled by default at @option{-O2} and higher.
5231 Perform dead code elimination (DCE) on trees. This flag is enabled by
5232 default at @option{-O} and higher.
5234 @item -ftree-dominator-opts
5235 Perform a variety of simple scalar cleanups (constant/copy
5236 propagation, redundancy elimination, range propagation and expression
5237 simplification) based on a dominator tree traversal. This also
5238 performs jump threading (to reduce jumps to jumps). This flag is
5239 enabled by default at @option{-O} and higher.
5242 Perform loop header copying on trees. This is beneficial since it increases
5243 effectiveness of code motion optimizations. It also saves one jump. This flag
5244 is enabled by default at @option{-O} and higher. It is not enabled
5245 for @option{-Os}, since it usually increases code size.
5247 @item -ftree-loop-optimize
5248 Perform loop optimizations on trees. This flag is enabled by default
5249 at @option{-O} and higher.
5251 @item -ftree-loop-linear
5252 Perform linear loop transformations on tree. This flag can improve cache
5253 performance and allow further loop optimizations to take place.
5255 @item -ftree-loop-im
5256 Perform loop invariant motion on trees. This pass moves only invariants that
5257 would be hard to handle at RTL level (function calls, operations that expand to
5258 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5259 operands of conditions that are invariant out of the loop, so that we can use
5260 just trivial invariantness analysis in loop unswitching. The pass also includes
5263 @item -ftree-loop-ivcanon
5264 Create a canonical counter for number of iterations in the loop for that
5265 determining number of iterations requires complicated analysis. Later
5266 optimizations then may determine the number easily. Useful especially
5267 in connection with unrolling.
5270 Perform induction variable optimizations (strength reduction, induction
5271 variable merging and induction variable elimination) on trees.
5274 Perform scalar replacement of aggregates. This pass replaces structure
5275 references with scalars to prevent committing structures to memory too
5276 early. This flag is enabled by default at @option{-O} and higher.
5278 @item -ftree-copyrename
5279 Perform copy renaming on trees. This pass attempts to rename compiler
5280 temporaries to other variables at copy locations, usually resulting in
5281 variable names which more closely resemble the original variables. This flag
5282 is enabled by default at @option{-O} and higher.
5285 Perform temporary expression replacement during the SSA->normal phase. Single
5286 use/single def temporaries are replaced at their use location with their
5287 defining expression. This results in non-GIMPLE code, but gives the expanders
5288 much more complex trees to work on resulting in better RTL generation. This is
5289 enabled by default at @option{-O} and higher.
5292 Perform live range splitting during the SSA->normal phase. Distinct live
5293 ranges of a variable are split into unique variables, allowing for better
5294 optimization later. This is enabled by default at @option{-O} and higher.
5296 @item -ftree-vectorize
5297 Perform loop vectorization on trees.
5299 @item -ftree-vect-loop-version
5300 @opindex ftree-vect-loop-version
5301 Perform loop versioning when doing loop vectorization on trees. When a loop
5302 appears to be vectorizable except that data alignment or data dependence cannot
5303 be determined at compile time then vectorized and non-vectorized versions of
5304 the loop are generated along with runtime checks for alignment or dependence
5305 to control which version is executed. This option is enabled by default
5306 except at level @option{-Os} where it is disabled.
5309 Perform Value Range Propagation on trees. This is similar to the
5310 constant propagation pass, but instead of values, ranges of values are
5311 propagated. This allows the optimizers to remove unnecessary range
5312 checks like array bound checks and null pointer checks. This is
5313 enabled by default at @option{-O2} and higher. Null pointer check
5314 elimination is only done if @option{-fdelete-null-pointer-checks} is
5319 Perform tail duplication to enlarge superblock size. This transformation
5320 simplifies the control flow of the function allowing other optimizations to do
5323 @item -funroll-loops
5324 @opindex funroll-loops
5325 Unroll loops whose number of iterations can be determined at compile
5326 time or upon entry to the loop. @option{-funroll-loops} implies
5327 @option{-frerun-cse-after-loop}. This option makes code larger,
5328 and may or may not make it run faster.
5330 @item -funroll-all-loops
5331 @opindex funroll-all-loops
5332 Unroll all loops, even if their number of iterations is uncertain when
5333 the loop is entered. This usually makes programs run more slowly.
5334 @option{-funroll-all-loops} implies the same options as
5335 @option{-funroll-loops},
5337 @item -fsplit-ivs-in-unroller
5338 @opindex -fsplit-ivs-in-unroller
5339 Enables expressing of values of induction variables in later iterations
5340 of the unrolled loop using the value in the first iteration. This breaks
5341 long dependency chains, thus improving efficiency of the scheduling passes.
5343 Combination of @option{-fweb} and CSE is often sufficient to obtain the
5344 same effect. However in cases the loop body is more complicated than
5345 a single basic block, this is not reliable. It also does not work at all
5346 on some of the architectures due to restrictions in the CSE pass.
5348 This optimization is enabled by default.
5350 @item -fvariable-expansion-in-unroller
5351 @opindex -fvariable-expansion-in-unroller
5352 With this option, the compiler will create multiple copies of some
5353 local variables when unrolling a loop which can result in superior code.
5355 @item -fprefetch-loop-arrays
5356 @opindex fprefetch-loop-arrays
5357 If supported by the target machine, generate instructions to prefetch
5358 memory to improve the performance of loops that access large arrays.
5360 This option may generate better or worse code; results are highly
5361 dependent on the structure of loops within the source code.
5363 Disabled at level @option{-Os}.
5366 @itemx -fno-peephole2
5367 @opindex fno-peephole
5368 @opindex fno-peephole2
5369 Disable any machine-specific peephole optimizations. The difference
5370 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
5371 are implemented in the compiler; some targets use one, some use the
5372 other, a few use both.
5374 @option{-fpeephole} is enabled by default.
5375 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5377 @item -fno-guess-branch-probability
5378 @opindex fno-guess-branch-probability
5379 Do not guess branch probabilities using heuristics.
5381 GCC will use heuristics to guess branch probabilities if they are
5382 not provided by profiling feedback (@option{-fprofile-arcs}). These
5383 heuristics are based on the control flow graph. If some branch probabilities
5384 are specified by @samp{__builtin_expect}, then the heuristics will be
5385 used to guess branch probabilities for the rest of the control flow graph,
5386 taking the @samp{__builtin_expect} info into account. The interactions
5387 between the heuristics and @samp{__builtin_expect} can be complex, and in
5388 some cases, it may be useful to disable the heuristics so that the effects
5389 of @samp{__builtin_expect} are easier to understand.
5391 The default is @option{-fguess-branch-probability} at levels
5392 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5394 @item -freorder-blocks
5395 @opindex freorder-blocks
5396 Reorder basic blocks in the compiled function in order to reduce number of
5397 taken branches and improve code locality.
5399 Enabled at levels @option{-O2}, @option{-O3}.
5401 @item -freorder-blocks-and-partition
5402 @opindex freorder-blocks-and-partition
5403 In addition to reordering basic blocks in the compiled function, in order
5404 to reduce number of taken branches, partitions hot and cold basic blocks
5405 into separate sections of the assembly and .o files, to improve
5406 paging and cache locality performance.
5408 This optimization is automatically turned off in the presence of
5409 exception handling, for linkonce sections, for functions with a user-defined
5410 section attribute and on any architecture that does not support named
5413 @item -freorder-functions
5414 @opindex freorder-functions
5415 Reorder functions in the object file in order to
5416 improve code locality. This is implemented by using special
5417 subsections @code{.text.hot} for most frequently executed functions and
5418 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
5419 the linker so object file format must support named sections and linker must
5420 place them in a reasonable way.
5422 Also profile feedback must be available in to make this option effective. See
5423 @option{-fprofile-arcs} for details.
5425 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5427 @item -fstrict-aliasing
5428 @opindex fstrict-aliasing
5429 Allows the compiler to assume the strictest aliasing rules applicable to
5430 the language being compiled. For C (and C++), this activates
5431 optimizations based on the type of expressions. In particular, an
5432 object of one type is assumed never to reside at the same address as an
5433 object of a different type, unless the types are almost the same. For
5434 example, an @code{unsigned int} can alias an @code{int}, but not a
5435 @code{void*} or a @code{double}. A character type may alias any other
5438 Pay special attention to code like this:
5451 The practice of reading from a different union member than the one most
5452 recently written to (called ``type-punning'') is common. Even with
5453 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
5454 is accessed through the union type. So, the code above will work as
5455 expected. However, this code might not:
5466 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5468 @item -fstrict-overflow
5469 @opindex fstrict-overflow
5470 Allow the compiler to assume strict signed overflow rules, depending
5471 on the language being compiled. For C (and C++) this means that
5472 overflow when doing arithmetic with signed numbers is undefined, which
5473 means that the compiler may assume that it will not happen. This
5474 permits various optimizations. For example, the compiler will assume
5475 that an expression like @code{i + 10 > i} will always be true for
5476 signed @code{i}. This assumption is only valid if signed overflow is
5477 undefined, as the expression is false if @code{i + 10} overflows when
5478 using twos complement arithmetic. When this option is in effect any
5479 attempt to determine whether an operation on signed numbers will
5480 overflow must be written carefully to not actually involve overflow.
5482 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
5483 that signed overflow is fully defined: it wraps. When
5484 @option{-fwrapv} is used, there is no difference between
5485 @option{-fstrict-overflow} and @option{-fno-strict-overflow}. With
5486 @option{-fwrapv} certain types of overflow are permitted. For
5487 example, if the compiler gets an overflow when doing arithmetic on
5488 constants, the overflowed value can still be used with
5489 @option{-fwrapv}, but not otherwise.
5491 The @option{-fstrict-overflow} option is enabled at levels
5492 @option{-O2}, @option{-O3}, @option{-Os}.
5494 @item -falign-functions
5495 @itemx -falign-functions=@var{n}
5496 @opindex falign-functions
5497 Align the start of functions to the next power-of-two greater than
5498 @var{n}, skipping up to @var{n} bytes. For instance,
5499 @option{-falign-functions=32} aligns functions to the next 32-byte
5500 boundary, but @option{-falign-functions=24} would align to the next
5501 32-byte boundary only if this can be done by skipping 23 bytes or less.
5503 @option{-fno-align-functions} and @option{-falign-functions=1} are
5504 equivalent and mean that functions will not be aligned.
5506 Some assemblers only support this flag when @var{n} is a power of two;
5507 in that case, it is rounded up.
5509 If @var{n} is not specified or is zero, use a machine-dependent default.
5511 Enabled at levels @option{-O2}, @option{-O3}.
5513 @item -falign-labels
5514 @itemx -falign-labels=@var{n}
5515 @opindex falign-labels
5516 Align all branch targets to a power-of-two boundary, skipping up to
5517 @var{n} bytes like @option{-falign-functions}. This option can easily
5518 make code slower, because it must insert dummy operations for when the
5519 branch target is reached in the usual flow of the code.
5521 @option{-fno-align-labels} and @option{-falign-labels=1} are
5522 equivalent and mean that labels will not be aligned.
5524 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
5525 are greater than this value, then their values are used instead.
5527 If @var{n} is not specified or is zero, use a machine-dependent default
5528 which is very likely to be @samp{1}, meaning no alignment.
5530 Enabled at levels @option{-O2}, @option{-O3}.
5533 @itemx -falign-loops=@var{n}
5534 @opindex falign-loops
5535 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
5536 like @option{-falign-functions}. The hope is that the loop will be
5537 executed many times, which will make up for any execution of the dummy
5540 @option{-fno-align-loops} and @option{-falign-loops=1} are
5541 equivalent and mean that loops will not be aligned.
5543 If @var{n} is not specified or is zero, use a machine-dependent default.
5545 Enabled at levels @option{-O2}, @option{-O3}.
5548 @itemx -falign-jumps=@var{n}
5549 @opindex falign-jumps
5550 Align branch targets to a power-of-two boundary, for branch targets
5551 where the targets can only be reached by jumping, skipping up to @var{n}
5552 bytes like @option{-falign-functions}. In this case, no dummy operations
5555 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
5556 equivalent and mean that loops will not be aligned.
5558 If @var{n} is not specified or is zero, use a machine-dependent default.
5560 Enabled at levels @option{-O2}, @option{-O3}.
5562 @item -funit-at-a-time
5563 @opindex funit-at-a-time
5564 Parse the whole compilation unit before starting to produce code.
5565 This allows some extra optimizations to take place but consumes
5566 more memory (in general). There are some compatibility issues
5567 with @emph{unit-at-a-time} mode:
5570 enabling @emph{unit-at-a-time} mode may change the order
5571 in which functions, variables, and top-level @code{asm} statements
5572 are emitted, and will likely break code relying on some particular
5573 ordering. The majority of such top-level @code{asm} statements,
5574 though, can be replaced by @code{section} attributes. The
5575 @option{fno-toplevel-reorder} option may be used to keep the ordering
5576 used in the input file, at the cost of some optimizations.
5579 @emph{unit-at-a-time} mode removes unreferenced static variables
5580 and functions. This may result in undefined references
5581 when an @code{asm} statement refers directly to variables or functions
5582 that are otherwise unused. In that case either the variable/function
5583 shall be listed as an operand of the @code{asm} statement operand or,
5584 in the case of top-level @code{asm} statements the attribute @code{used}
5585 shall be used on the declaration.
5588 Static functions now can use non-standard passing conventions that
5589 may break @code{asm} statements calling functions directly. Again,
5590 attribute @code{used} will prevent this behavior.
5593 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
5594 but this scheme may not be supported by future releases of GCC@.
5596 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5598 @item -fno-toplevel-reorder
5599 Do not reorder top-level functions, variables, and @code{asm}
5600 statements. Output them in the same order that they appear in the
5601 input file. When this option is used, unreferenced static variables
5602 will not be removed. This option is intended to support existing code
5603 which relies on a particular ordering. For new code, it is better to
5608 Constructs webs as commonly used for register allocation purposes and assign
5609 each web individual pseudo register. This allows the register allocation pass
5610 to operate on pseudos directly, but also strengthens several other optimization
5611 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
5612 however, make debugging impossible, since variables will no longer stay in a
5615 Enabled by default with @option{-funroll-loops}.
5617 @item -fwhole-program
5618 @opindex fwhole-program
5619 Assume that the current compilation unit represents whole program being
5620 compiled. All public functions and variables with the exception of @code{main}
5621 and those merged by attribute @code{externally_visible} become static functions
5622 and in a affect gets more aggressively optimized by interprocedural optimizers.
5623 While this option is equivalent to proper use of @code{static} keyword for
5624 programs consisting of single file, in combination with option
5625 @option{--combine} this flag can be used to compile most of smaller scale C
5626 programs since the functions and variables become local for the whole combined
5627 compilation unit, not for the single source file itself.
5630 @item -fno-cprop-registers
5631 @opindex fno-cprop-registers
5632 After register allocation and post-register allocation instruction splitting,
5633 we perform a copy-propagation pass to try to reduce scheduling dependencies
5634 and occasionally eliminate the copy.
5636 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5638 @item -fprofile-generate
5639 @opindex fprofile-generate
5641 Enable options usually used for instrumenting application to produce
5642 profile useful for later recompilation with profile feedback based
5643 optimization. You must use @option{-fprofile-generate} both when
5644 compiling and when linking your program.
5646 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
5649 @opindex fprofile-use
5650 Enable profile feedback directed optimizations, and optimizations
5651 generally profitable only with profile feedback available.
5653 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
5654 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
5656 By default, GCC emits an error message if the feedback profiles do not
5657 match the source code. This error can be turned into a warning by using
5658 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
5662 The following options control compiler behavior regarding floating
5663 point arithmetic. These options trade off between speed and
5664 correctness. All must be specifically enabled.
5668 @opindex ffloat-store
5669 Do not store floating point variables in registers, and inhibit other
5670 options that might change whether a floating point value is taken from a
5673 @cindex floating point precision
5674 This option prevents undesirable excess precision on machines such as
5675 the 68000 where the floating registers (of the 68881) keep more
5676 precision than a @code{double} is supposed to have. Similarly for the
5677 x86 architecture. For most programs, the excess precision does only
5678 good, but a few programs rely on the precise definition of IEEE floating
5679 point. Use @option{-ffloat-store} for such programs, after modifying
5680 them to store all pertinent intermediate computations into variables.
5684 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
5685 @option{-fno-trapping-math}, @option{-ffinite-math-only},
5686 @option{-fno-rounding-math}, @option{-fno-signaling-nans},
5687 @option{-fno-signed-zeros} and @option{fcx-limited-range}.
5689 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
5691 This option should never be turned on by any @option{-O} option since
5692 it can result in incorrect output for programs which depend on
5693 an exact implementation of IEEE or ISO rules/specifications for
5696 @item -fno-math-errno
5697 @opindex fno-math-errno
5698 Do not set ERRNO after calling math functions that are executed
5699 with a single instruction, e.g., sqrt. A program that relies on
5700 IEEE exceptions for math error handling may want to use this flag
5701 for speed while maintaining IEEE arithmetic compatibility.
5703 This option should never be turned on by any @option{-O} option since
5704 it can result in incorrect output for programs which depend on
5705 an exact implementation of IEEE or ISO rules/specifications for
5708 The default is @option{-fmath-errno}.
5710 On Darwin systems, the math library never sets @code{errno}. There is
5711 therefore no reason for the compiler to consider the possibility that
5712 it might, and @option{-fno-math-errno} is the default.
5714 @item -funsafe-math-optimizations
5715 @opindex funsafe-math-optimizations
5716 Allow optimizations for floating-point arithmetic that (a) assume
5717 that arguments and results are valid and (b) may violate IEEE or
5718 ANSI standards. When used at link-time, it may include libraries
5719 or startup files that change the default FPU control word or other
5720 similar optimizations.
5722 This option should never be turned on by any @option{-O} option since
5723 it can result in incorrect output for programs which depend on
5724 an exact implementation of IEEE or ISO rules/specifications for
5727 The default is @option{-fno-unsafe-math-optimizations}.
5729 @item -ffinite-math-only
5730 @opindex ffinite-math-only
5731 Allow optimizations for floating-point arithmetic that assume
5732 that arguments and results are not NaNs or +-Infs.
5734 This option should never be turned on by any @option{-O} option since
5735 it can result in incorrect output for programs which depend on
5736 an exact implementation of IEEE or ISO rules/specifications.
5738 The default is @option{-fno-finite-math-only}.
5740 @item -fno-signed-zeros
5741 @opindex fno-signed-zeros
5742 Allow optimizations for floating point arithmetic that ignore the
5743 signedness of zero. IEEE arithmetic specifies the behavior of
5744 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
5745 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
5746 This option implies that the sign of a zero result isn't significant.
5748 The default is @option{-fsigned-zeros}.
5750 @item -fno-trapping-math
5751 @opindex fno-trapping-math
5752 Compile code assuming that floating-point operations cannot generate
5753 user-visible traps. These traps include division by zero, overflow,
5754 underflow, inexact result and invalid operation. This option implies
5755 @option{-fno-signaling-nans}. Setting this option may allow faster
5756 code if one relies on ``non-stop'' IEEE arithmetic, for example.
5758 This option should never be turned on by any @option{-O} option since
5759 it can result in incorrect output for programs which depend on
5760 an exact implementation of IEEE or ISO rules/specifications for
5763 The default is @option{-ftrapping-math}.
5765 @item -frounding-math
5766 @opindex frounding-math
5767 Disable transformations and optimizations that assume default floating
5768 point rounding behavior. This is round-to-zero for all floating point
5769 to integer conversions, and round-to-nearest for all other arithmetic
5770 truncations. This option should be specified for programs that change
5771 the FP rounding mode dynamically, or that may be executed with a
5772 non-default rounding mode. This option disables constant folding of
5773 floating point expressions at compile-time (which may be affected by
5774 rounding mode) and arithmetic transformations that are unsafe in the
5775 presence of sign-dependent rounding modes.
5777 The default is @option{-fno-rounding-math}.
5779 This option is experimental and does not currently guarantee to
5780 disable all GCC optimizations that are affected by rounding mode.
5781 Future versions of GCC may provide finer control of this setting
5782 using C99's @code{FENV_ACCESS} pragma. This command line option
5783 will be used to specify the default state for @code{FENV_ACCESS}.
5785 @item -frtl-abstract-sequences
5786 @opindex frtl-abstract-sequences
5787 It is a size optimization method. This option is to find identical
5788 sequences of code, which can be turned into pseudo-procedures and
5789 then replace all occurrences with calls to the newly created
5790 subroutine. It is kind of an opposite of @option{-finline-functions}.
5791 This optimization runs at RTL level.
5793 @item -fsignaling-nans
5794 @opindex fsignaling-nans
5795 Compile code assuming that IEEE signaling NaNs may generate user-visible
5796 traps during floating-point operations. Setting this option disables
5797 optimizations that may change the number of exceptions visible with
5798 signaling NaNs. This option implies @option{-ftrapping-math}.
5800 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
5803 The default is @option{-fno-signaling-nans}.
5805 This option is experimental and does not currently guarantee to
5806 disable all GCC optimizations that affect signaling NaN behavior.
5808 @item -fsingle-precision-constant
5809 @opindex fsingle-precision-constant
5810 Treat floating point constant as single precision constant instead of
5811 implicitly converting it to double precision constant.
5813 @item -fcx-limited-range
5814 @itemx -fno-cx-limited-range
5815 @opindex fcx-limited-range
5816 @opindex fno-cx-limited-range
5817 When enabled, this option states that a range reduction step is not
5818 needed when performing complex division. The default is
5819 @option{-fno-cx-limited-range}, but is enabled by @option{-ffast-math}.
5821 This option controls the default setting of the ISO C99
5822 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
5827 The following options control optimizations that may improve
5828 performance, but are not enabled by any @option{-O} options. This
5829 section includes experimental options that may produce broken code.
5832 @item -fbranch-probabilities
5833 @opindex fbranch-probabilities
5834 After running a program compiled with @option{-fprofile-arcs}
5835 (@pxref{Debugging Options,, Options for Debugging Your Program or
5836 @command{gcc}}), you can compile it a second time using
5837 @option{-fbranch-probabilities}, to improve optimizations based on
5838 the number of times each branch was taken. When the program
5839 compiled with @option{-fprofile-arcs} exits it saves arc execution
5840 counts to a file called @file{@var{sourcename}.gcda} for each source
5841 file. The information in this data file is very dependent on the
5842 structure of the generated code, so you must use the same source code
5843 and the same optimization options for both compilations.
5845 With @option{-fbranch-probabilities}, GCC puts a
5846 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
5847 These can be used to improve optimization. Currently, they are only
5848 used in one place: in @file{reorg.c}, instead of guessing which path a
5849 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
5850 exactly determine which path is taken more often.
5852 @item -fprofile-values
5853 @opindex fprofile-values
5854 If combined with @option{-fprofile-arcs}, it adds code so that some
5855 data about values of expressions in the program is gathered.
5857 With @option{-fbranch-probabilities}, it reads back the data gathered
5858 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
5859 notes to instructions for their later usage in optimizations.
5861 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
5865 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
5866 a code to gather information about values of expressions.
5868 With @option{-fbranch-probabilities}, it reads back the data gathered
5869 and actually performs the optimizations based on them.
5870 Currently the optimizations include specialization of division operation
5871 using the knowledge about the value of the denominator.
5873 @item -frename-registers
5874 @opindex frename-registers
5875 Attempt to avoid false dependencies in scheduled code by making use
5876 of registers left over after register allocation. This optimization
5877 will most benefit processors with lots of registers. Depending on the
5878 debug information format adopted by the target, however, it can
5879 make debugging impossible, since variables will no longer stay in
5880 a ``home register''.
5882 Enabled by default with @option{-funroll-loops}.
5886 Perform tail duplication to enlarge superblock size. This transformation
5887 simplifies the control flow of the function allowing other optimizations to do
5890 Enabled with @option{-fprofile-use}.
5892 @item -funroll-loops
5893 @opindex funroll-loops
5894 Unroll loops whose number of iterations can be determined at compile time or
5895 upon entry to the loop. @option{-funroll-loops} implies
5896 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
5897 It also turns on complete loop peeling (i.e.@: complete removal of loops with
5898 small constant number of iterations). This option makes code larger, and may
5899 or may not make it run faster.
5901 Enabled with @option{-fprofile-use}.
5903 @item -funroll-all-loops
5904 @opindex funroll-all-loops
5905 Unroll all loops, even if their number of iterations is uncertain when
5906 the loop is entered. This usually makes programs run more slowly.
5907 @option{-funroll-all-loops} implies the same options as
5908 @option{-funroll-loops}.
5911 @opindex fpeel-loops
5912 Peels the loops for that there is enough information that they do not
5913 roll much (from profile feedback). It also turns on complete loop peeling
5914 (i.e.@: complete removal of loops with small constant number of iterations).
5916 Enabled with @option{-fprofile-use}.
5918 @item -fmove-loop-invariants
5919 @opindex fmove-loop-invariants
5920 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
5921 at level @option{-O1}
5923 @item -funswitch-loops
5924 @opindex funswitch-loops
5925 Move branches with loop invariant conditions out of the loop, with duplicates
5926 of the loop on both branches (modified according to result of the condition).
5928 @item -ffunction-sections
5929 @itemx -fdata-sections
5930 @opindex ffunction-sections
5931 @opindex fdata-sections
5932 Place each function or data item into its own section in the output
5933 file if the target supports arbitrary sections. The name of the
5934 function or the name of the data item determines the section's name
5937 Use these options on systems where the linker can perform optimizations
5938 to improve locality of reference in the instruction space. Most systems
5939 using the ELF object format and SPARC processors running Solaris 2 have
5940 linkers with such optimizations. AIX may have these optimizations in
5943 Only use these options when there are significant benefits from doing
5944 so. When you specify these options, the assembler and linker will
5945 create larger object and executable files and will also be slower.
5946 You will not be able to use @code{gprof} on all systems if you
5947 specify this option and you may have problems with debugging if
5948 you specify both this option and @option{-g}.
5950 @item -fbranch-target-load-optimize
5951 @opindex fbranch-target-load-optimize
5952 Perform branch target register load optimization before prologue / epilogue
5954 The use of target registers can typically be exposed only during reload,
5955 thus hoisting loads out of loops and doing inter-block scheduling needs
5956 a separate optimization pass.
5958 @item -fbranch-target-load-optimize2
5959 @opindex fbranch-target-load-optimize2
5960 Perform branch target register load optimization after prologue / epilogue
5963 @item -fbtr-bb-exclusive
5964 @opindex fbtr-bb-exclusive
5965 When performing branch target register load optimization, don't reuse
5966 branch target registers in within any basic block.
5968 @item -fstack-protector
5969 Emit extra code to check for buffer overflows, such as stack smashing
5970 attacks. This is done by adding a guard variable to functions with
5971 vulnerable objects. This includes functions that call alloca, and
5972 functions with buffers larger than 8 bytes. The guards are initialized
5973 when a function is entered and then checked when the function exits.
5974 If a guard check fails, an error message is printed and the program exits.
5976 @item -fstack-protector-all
5977 Like @option{-fstack-protector} except that all functions are protected.
5979 @item -fsection-anchors
5980 @opindex fsection-anchors
5981 Try to reduce the number of symbolic address calculations by using
5982 shared ``anchor'' symbols to address nearby objects. This transformation
5983 can help to reduce the number of GOT entries and GOT accesses on some
5986 For example, the implementation of the following function @code{foo}:
5990 int foo (void) @{ return a + b + c; @}
5993 would usually calculate the addresses of all three variables, but if you
5994 compile it with @option{-fsection-anchors}, it will access the variables
5995 from a common anchor point instead. The effect is similar to the
5996 following pseudocode (which isn't valid C):
6001 register int *xr = &x;
6002 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
6006 Not all targets support this option.
6008 @item --param @var{name}=@var{value}
6010 In some places, GCC uses various constants to control the amount of
6011 optimization that is done. For example, GCC will not inline functions
6012 that contain more that a certain number of instructions. You can
6013 control some of these constants on the command-line using the
6014 @option{--param} option.
6016 The names of specific parameters, and the meaning of the values, are
6017 tied to the internals of the compiler, and are subject to change
6018 without notice in future releases.
6020 In each case, the @var{value} is an integer. The allowable choices for
6021 @var{name} are given in the following table:
6024 @item salias-max-implicit-fields
6025 The maximum number of fields in a variable without direct
6026 structure accesses for which structure aliasing will consider trying
6027 to track each field. The default is 5
6029 @item salias-max-array-elements
6030 The maximum number of elements an array can have and its elements
6031 still be tracked individually by structure aliasing. The default is 4
6033 @item sra-max-structure-size
6034 The maximum structure size, in bytes, at which the scalar replacement
6035 of aggregates (SRA) optimization will perform block copies. The
6036 default value, 0, implies that GCC will select the most appropriate
6039 @item sra-field-structure-ratio
6040 The threshold ratio (as a percentage) between instantiated fields and
6041 the complete structure size. We say that if the ratio of the number
6042 of bytes in instantiated fields to the number of bytes in the complete
6043 structure exceeds this parameter, then block copies are not used. The
6046 @item max-crossjump-edges
6047 The maximum number of incoming edges to consider for crossjumping.
6048 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
6049 the number of edges incoming to each block. Increasing values mean
6050 more aggressive optimization, making the compile time increase with
6051 probably small improvement in executable size.
6053 @item min-crossjump-insns
6054 The minimum number of instructions which must be matched at the end
6055 of two blocks before crossjumping will be performed on them. This
6056 value is ignored in the case where all instructions in the block being
6057 crossjumped from are matched. The default value is 5.
6059 @item max-grow-copy-bb-insns
6060 The maximum code size expansion factor when copying basic blocks
6061 instead of jumping. The expansion is relative to a jump instruction.
6062 The default value is 8.
6064 @item max-goto-duplication-insns
6065 The maximum number of instructions to duplicate to a block that jumps
6066 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
6067 passes, GCC factors computed gotos early in the compilation process,
6068 and unfactors them as late as possible. Only computed jumps at the
6069 end of a basic blocks with no more than max-goto-duplication-insns are
6070 unfactored. The default value is 8.
6072 @item max-delay-slot-insn-search
6073 The maximum number of instructions to consider when looking for an
6074 instruction to fill a delay slot. If more than this arbitrary number of
6075 instructions is searched, the time savings from filling the delay slot
6076 will be minimal so stop searching. Increasing values mean more
6077 aggressive optimization, making the compile time increase with probably
6078 small improvement in executable run time.
6080 @item max-delay-slot-live-search
6081 When trying to fill delay slots, the maximum number of instructions to
6082 consider when searching for a block with valid live register
6083 information. Increasing this arbitrarily chosen value means more
6084 aggressive optimization, increasing the compile time. This parameter
6085 should be removed when the delay slot code is rewritten to maintain the
6088 @item max-gcse-memory
6089 The approximate maximum amount of memory that will be allocated in
6090 order to perform the global common subexpression elimination
6091 optimization. If more memory than specified is required, the
6092 optimization will not be done.
6094 @item max-gcse-passes
6095 The maximum number of passes of GCSE to run. The default is 1.
6097 @item max-pending-list-length
6098 The maximum number of pending dependencies scheduling will allow
6099 before flushing the current state and starting over. Large functions
6100 with few branches or calls can create excessively large lists which
6101 needlessly consume memory and resources.
6103 @item max-inline-insns-single
6104 Several parameters control the tree inliner used in gcc.
6105 This number sets the maximum number of instructions (counted in GCC's
6106 internal representation) in a single function that the tree inliner
6107 will consider for inlining. This only affects functions declared
6108 inline and methods implemented in a class declaration (C++).
6109 The default value is 450.
6111 @item max-inline-insns-auto
6112 When you use @option{-finline-functions} (included in @option{-O3}),
6113 a lot of functions that would otherwise not be considered for inlining
6114 by the compiler will be investigated. To those functions, a different
6115 (more restrictive) limit compared to functions declared inline can
6117 The default value is 90.
6119 @item large-function-insns
6120 The limit specifying really large functions. For functions larger than this
6121 limit after inlining inlining is constrained by
6122 @option{--param large-function-growth}. This parameter is useful primarily
6123 to avoid extreme compilation time caused by non-linear algorithms used by the
6125 This parameter is ignored when @option{-funit-at-a-time} is not used.
6126 The default value is 2700.
6128 @item large-function-growth
6129 Specifies maximal growth of large function caused by inlining in percents.
6130 This parameter is ignored when @option{-funit-at-a-time} is not used.
6131 The default value is 100 which limits large function growth to 2.0 times
6134 @item large-unit-insns
6135 The limit specifying large translation unit. Growth caused by inlining of
6136 units larger than this limit is limited by @option{--param inline-unit-growth}.
6137 For small units this might be too tight (consider unit consisting of function A
6138 that is inline and B that just calls A three time. If B is small relative to
6139 A, the growth of unit is 300\% and yet such inlining is very sane. For very
6140 large units consisting of small inlininable functions however the overall unit
6141 growth limit is needed to avoid exponential explosion of code size. Thus for
6142 smaller units, the size is increased to @option{--param large-unit-insns}
6143 before applying @option{--param inline-unit-growth}. The default is 10000
6145 @item inline-unit-growth
6146 Specifies maximal overall growth of the compilation unit caused by inlining.
6147 This parameter is ignored when @option{-funit-at-a-time} is not used.
6148 The default value is 60 which limits unit growth to 1.6 times the original
6151 @item large-stack-frame
6152 The limit specifying large stack frames. While inlining the algorithm is trying
6153 to not grow past this limit too much. Default value is 256 bytes.
6155 @item large-stack-frame-growth
6156 Specifies maximal growth of large stack frames caused by inlining in percents.
6157 The default value is 1000 which limits large stack frame growth to 11 times
6160 @item max-inline-insns-recursive
6161 @itemx max-inline-insns-recursive-auto
6162 Specifies maximum number of instructions out-of-line copy of self recursive inline
6163 function can grow into by performing recursive inlining.
6165 For functions declared inline @option{--param max-inline-insns-recursive} is
6166 taken into account. For function not declared inline, recursive inlining
6167 happens only when @option{-finline-functions} (included in @option{-O3}) is
6168 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
6169 default value is 450.
6171 @item max-inline-recursive-depth
6172 @itemx max-inline-recursive-depth-auto
6173 Specifies maximum recursion depth used by the recursive inlining.
6175 For functions declared inline @option{--param max-inline-recursive-depth} is
6176 taken into account. For function not declared inline, recursive inlining
6177 happens only when @option{-finline-functions} (included in @option{-O3}) is
6178 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
6179 default value is 450.
6181 @item min-inline-recursive-probability
6182 Recursive inlining is profitable only for function having deep recursion
6183 in average and can hurt for function having little recursion depth by
6184 increasing the prologue size or complexity of function body to other
6187 When profile feedback is available (see @option{-fprofile-generate}) the actual
6188 recursion depth can be guessed from probability that function will recurse via
6189 given call expression. This parameter limits inlining only to call expression
6190 whose probability exceeds given threshold (in percents). The default value is
6193 @item inline-call-cost
6194 Specify cost of call instruction relative to simple arithmetics operations
6195 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
6196 functions and at the same time increases size of leaf function that is believed to
6197 reduce function size by being inlined. In effect it increases amount of
6198 inlining for code having large abstraction penalty (many functions that just
6199 pass the arguments to other functions) and decrease inlining for code with low
6200 abstraction penalty. The default value is 16.
6202 @item min-vect-loop-bound
6203 The minimum number of iterations under which a loop will not get vectorized
6204 when @option{-ftree-vectorize} is used. The number of iterations after
6205 vectorization needs to be greater than the value specified by this option
6206 to allow vectorization. The default value is 0.
6208 @item max-unrolled-insns
6209 The maximum number of instructions that a loop should have if that loop
6210 is unrolled, and if the loop is unrolled, it determines how many times
6211 the loop code is unrolled.
6213 @item max-average-unrolled-insns
6214 The maximum number of instructions biased by probabilities of their execution
6215 that a loop should have if that loop is unrolled, and if the loop is unrolled,
6216 it determines how many times the loop code is unrolled.
6218 @item max-unroll-times
6219 The maximum number of unrollings of a single loop.
6221 @item max-peeled-insns
6222 The maximum number of instructions that a loop should have if that loop
6223 is peeled, and if the loop is peeled, it determines how many times
6224 the loop code is peeled.
6226 @item max-peel-times
6227 The maximum number of peelings of a single loop.
6229 @item max-completely-peeled-insns
6230 The maximum number of insns of a completely peeled loop.
6232 @item max-completely-peel-times
6233 The maximum number of iterations of a loop to be suitable for complete peeling.
6235 @item max-unswitch-insns
6236 The maximum number of insns of an unswitched loop.
6238 @item max-unswitch-level
6239 The maximum number of branches unswitched in a single loop.
6242 The minimum cost of an expensive expression in the loop invariant motion.
6244 @item iv-consider-all-candidates-bound
6245 Bound on number of candidates for induction variables below that
6246 all candidates are considered for each use in induction variable
6247 optimizations. Only the most relevant candidates are considered
6248 if there are more candidates, to avoid quadratic time complexity.
6250 @item iv-max-considered-uses
6251 The induction variable optimizations give up on loops that contain more
6252 induction variable uses.
6254 @item iv-always-prune-cand-set-bound
6255 If number of candidates in the set is smaller than this value,
6256 we always try to remove unnecessary ivs from the set during its
6257 optimization when a new iv is added to the set.
6259 @item scev-max-expr-size
6260 Bound on size of expressions used in the scalar evolutions analyzer.
6261 Large expressions slow the analyzer.
6263 @item vect-max-version-checks
6264 The maximum number of runtime checks that can be performed when doing
6265 loop versioning in the vectorizer. See option ftree-vect-loop-version
6266 for more information.
6268 @item max-iterations-to-track
6270 The maximum number of iterations of a loop the brute force algorithm
6271 for analysis of # of iterations of the loop tries to evaluate.
6273 @item hot-bb-count-fraction
6274 Select fraction of the maximal count of repetitions of basic block in program
6275 given basic block needs to have to be considered hot.
6277 @item hot-bb-frequency-fraction
6278 Select fraction of the maximal frequency of executions of basic block in
6279 function given basic block needs to have to be considered hot
6281 @item max-predicted-iterations
6282 The maximum number of loop iterations we predict statically. This is useful
6283 in cases where function contain single loop with known bound and other loop
6284 with unknown. We predict the known number of iterations correctly, while
6285 the unknown number of iterations average to roughly 10. This means that the
6286 loop without bounds would appear artificially cold relative to the other one.
6288 @item tracer-dynamic-coverage
6289 @itemx tracer-dynamic-coverage-feedback
6291 This value is used to limit superblock formation once the given percentage of
6292 executed instructions is covered. This limits unnecessary code size
6295 The @option{tracer-dynamic-coverage-feedback} is used only when profile
6296 feedback is available. The real profiles (as opposed to statically estimated
6297 ones) are much less balanced allowing the threshold to be larger value.
6299 @item tracer-max-code-growth
6300 Stop tail duplication once code growth has reached given percentage. This is
6301 rather hokey argument, as most of the duplicates will be eliminated later in
6302 cross jumping, so it may be set to much higher values than is the desired code
6305 @item tracer-min-branch-ratio
6307 Stop reverse growth when the reverse probability of best edge is less than this
6308 threshold (in percent).
6310 @item tracer-min-branch-ratio
6311 @itemx tracer-min-branch-ratio-feedback
6313 Stop forward growth if the best edge do have probability lower than this
6316 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
6317 compilation for profile feedback and one for compilation without. The value
6318 for compilation with profile feedback needs to be more conservative (higher) in
6319 order to make tracer effective.
6321 @item max-cse-path-length
6323 Maximum number of basic blocks on path that cse considers. The default is 10.
6326 The maximum instructions CSE process before flushing. The default is 1000.
6328 @item max-aliased-vops
6330 Maximum number of virtual operands per statement allowed to represent
6331 aliases before triggering the alias grouping heuristic. Alias
6332 grouping reduces compile times and memory consumption needed for
6333 aliasing at the expense of precision loss in alias information.
6335 @item ggc-min-expand
6337 GCC uses a garbage collector to manage its own memory allocation. This
6338 parameter specifies the minimum percentage by which the garbage
6339 collector's heap should be allowed to expand between collections.
6340 Tuning this may improve compilation speed; it has no effect on code
6343 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
6344 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
6345 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
6346 GCC is not able to calculate RAM on a particular platform, the lower
6347 bound of 30% is used. Setting this parameter and
6348 @option{ggc-min-heapsize} to zero causes a full collection to occur at
6349 every opportunity. This is extremely slow, but can be useful for
6352 @item ggc-min-heapsize
6354 Minimum size of the garbage collector's heap before it begins bothering
6355 to collect garbage. The first collection occurs after the heap expands
6356 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
6357 tuning this may improve compilation speed, and has no effect on code
6360 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
6361 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
6362 with a lower bound of 4096 (four megabytes) and an upper bound of
6363 131072 (128 megabytes). If GCC is not able to calculate RAM on a
6364 particular platform, the lower bound is used. Setting this parameter
6365 very large effectively disables garbage collection. Setting this
6366 parameter and @option{ggc-min-expand} to zero causes a full collection
6367 to occur at every opportunity.
6369 @item max-reload-search-insns
6370 The maximum number of instruction reload should look backward for equivalent
6371 register. Increasing values mean more aggressive optimization, making the
6372 compile time increase with probably slightly better performance. The default
6375 @item max-cselib-memory-locations
6376 The maximum number of memory locations cselib should take into account.
6377 Increasing values mean more aggressive optimization, making the compile time
6378 increase with probably slightly better performance. The default value is 500.
6380 @item max-flow-memory-locations
6381 Similar as @option{max-cselib-memory-locations} but for dataflow liveness.
6382 The default value is 100.
6384 @item reorder-blocks-duplicate
6385 @itemx reorder-blocks-duplicate-feedback
6387 Used by basic block reordering pass to decide whether to use unconditional
6388 branch or duplicate the code on its destination. Code is duplicated when its
6389 estimated size is smaller than this value multiplied by the estimated size of
6390 unconditional jump in the hot spots of the program.
6392 The @option{reorder-block-duplicate-feedback} is used only when profile
6393 feedback is available and may be set to higher values than
6394 @option{reorder-block-duplicate} since information about the hot spots is more
6397 @item max-sched-ready-insns
6398 The maximum number of instructions ready to be issued the scheduler should
6399 consider at any given time during the first scheduling pass. Increasing
6400 values mean more thorough searches, making the compilation time increase
6401 with probably little benefit. The default value is 100.
6403 @item max-sched-region-blocks
6404 The maximum number of blocks in a region to be considered for
6405 interblock scheduling. The default value is 10.
6407 @item max-sched-region-insns
6408 The maximum number of insns in a region to be considered for
6409 interblock scheduling. The default value is 100.
6412 The minimum probability (in percents) of reaching a source block
6413 for interblock speculative scheduling. The default value is 40.
6415 @item max-sched-extend-regions-iters
6416 The maximum number of iterations through CFG to extend regions.
6417 0 - disable region extension,
6418 N - do at most N iterations.
6419 The default value is 0.
6421 @item max-sched-insn-conflict-delay
6422 The maximum conflict delay for an insn to be considered for speculative motion.
6423 The default value is 3.
6425 @item sched-spec-prob-cutoff
6426 The minimal probability of speculation success (in percents), so that
6427 speculative insn will be scheduled.
6428 The default value is 40.
6430 @item max-last-value-rtl
6432 The maximum size measured as number of RTLs that can be recorded in an expression
6433 in combiner for a pseudo register as last known value of that register. The default
6436 @item integer-share-limit
6437 Small integer constants can use a shared data structure, reducing the
6438 compiler's memory usage and increasing its speed. This sets the maximum
6439 value of a shared integer constant's. The default value is 256.
6441 @item min-virtual-mappings
6442 Specifies the minimum number of virtual mappings in the incremental
6443 SSA updater that should be registered to trigger the virtual mappings
6444 heuristic defined by virtual-mappings-ratio. The default value is
6447 @item virtual-mappings-ratio
6448 If the number of virtual mappings is virtual-mappings-ratio bigger
6449 than the number of virtual symbols to be updated, then the incremental
6450 SSA updater switches to a full update for those symbols. The default
6453 @item ssp-buffer-size
6454 The minimum size of buffers (i.e. arrays) that will receive stack smashing
6455 protection when @option{-fstack-protection} is used.
6457 @item max-jump-thread-duplication-stmts
6458 Maximum number of statements allowed in a block that needs to be
6459 duplicated when threading jumps.
6461 @item max-fields-for-field-sensitive
6462 Maximum number of fields in a structure we will treat in
6463 a field sensitive manner during pointer analysis.
6465 @item prefetch-latency
6466 Estimate on average number of instructions that are executed before
6467 prefetch finishes. The distance we prefetch ahead is proportional
6468 to this constant. Increasing this number may also lead to less
6469 streams being prefetched (see @option{simultaneous-prefetches}).
6471 @item simultaneous-prefetches
6472 Maximum number of prefetches that can run at the same time.
6474 @item l1-cache-line-size
6475 The size of cache line in L1 cache, in bytes.
6478 The number of cache lines in L1 cache.
6480 @item verify-canonical-types
6481 Whether the compiler should verify the ``canonical'' types used for
6482 type equality comparisons within the C++ and Objective-C++ front
6483 ends. Set to 1 (the default when GCC is configured with
6484 --enable-checking) to enable verification, 0 to disable verification
6485 (the default when GCC is configured with --disable-checking).
6490 @node Preprocessor Options
6491 @section Options Controlling the Preprocessor
6492 @cindex preprocessor options
6493 @cindex options, preprocessor
6495 These options control the C preprocessor, which is run on each C source
6496 file before actual compilation.
6498 If you use the @option{-E} option, nothing is done except preprocessing.
6499 Some of these options make sense only together with @option{-E} because
6500 they cause the preprocessor output to be unsuitable for actual
6505 You can use @option{-Wp,@var{option}} to bypass the compiler driver
6506 and pass @var{option} directly through to the preprocessor. If
6507 @var{option} contains commas, it is split into multiple options at the
6508 commas. However, many options are modified, translated or interpreted
6509 by the compiler driver before being passed to the preprocessor, and
6510 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
6511 interface is undocumented and subject to change, so whenever possible
6512 you should avoid using @option{-Wp} and let the driver handle the
6515 @item -Xpreprocessor @var{option}
6516 @opindex preprocessor
6517 Pass @var{option} as an option to the preprocessor. You can use this to
6518 supply system-specific preprocessor options which GCC does not know how to
6521 If you want to pass an option that takes an argument, you must use
6522 @option{-Xpreprocessor} twice, once for the option and once for the argument.
6525 @include cppopts.texi
6527 @node Assembler Options
6528 @section Passing Options to the Assembler
6530 @c prevent bad page break with this line
6531 You can pass options to the assembler.
6534 @item -Wa,@var{option}
6536 Pass @var{option} as an option to the assembler. If @var{option}
6537 contains commas, it is split into multiple options at the commas.
6539 @item -Xassembler @var{option}
6541 Pass @var{option} as an option to the assembler. You can use this to
6542 supply system-specific assembler options which GCC does not know how to
6545 If you want to pass an option that takes an argument, you must use
6546 @option{-Xassembler} twice, once for the option and once for the argument.
6551 @section Options for Linking
6552 @cindex link options
6553 @cindex options, linking
6555 These options come into play when the compiler links object files into
6556 an executable output file. They are meaningless if the compiler is
6557 not doing a link step.
6561 @item @var{object-file-name}
6562 A file name that does not end in a special recognized suffix is
6563 considered to name an object file or library. (Object files are
6564 distinguished from libraries by the linker according to the file
6565 contents.) If linking is done, these object files are used as input
6574 If any of these options is used, then the linker is not run, and
6575 object file names should not be used as arguments. @xref{Overall
6579 @item -l@var{library}
6580 @itemx -l @var{library}
6582 Search the library named @var{library} when linking. (The second
6583 alternative with the library as a separate argument is only for
6584 POSIX compliance and is not recommended.)
6586 It makes a difference where in the command you write this option; the
6587 linker searches and processes libraries and object files in the order they
6588 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
6589 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
6590 to functions in @samp{z}, those functions may not be loaded.
6592 The linker searches a standard list of directories for the library,
6593 which is actually a file named @file{lib@var{library}.a}. The linker
6594 then uses this file as if it had been specified precisely by name.
6596 The directories searched include several standard system directories
6597 plus any that you specify with @option{-L}.
6599 Normally the files found this way are library files---archive files
6600 whose members are object files. The linker handles an archive file by
6601 scanning through it for members which define symbols that have so far
6602 been referenced but not defined. But if the file that is found is an
6603 ordinary object file, it is linked in the usual fashion. The only
6604 difference between using an @option{-l} option and specifying a file name
6605 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
6606 and searches several directories.
6610 You need this special case of the @option{-l} option in order to
6611 link an Objective-C or Objective-C++ program.
6614 @opindex nostartfiles
6615 Do not use the standard system startup files when linking.
6616 The standard system libraries are used normally, unless @option{-nostdlib}
6617 or @option{-nodefaultlibs} is used.
6619 @item -nodefaultlibs
6620 @opindex nodefaultlibs
6621 Do not use the standard system libraries when linking.
6622 Only the libraries you specify will be passed to the linker.
6623 The standard startup files are used normally, unless @option{-nostartfiles}
6624 is used. The compiler may generate calls to @code{memcmp},
6625 @code{memset}, @code{memcpy} and @code{memmove}.
6626 These entries are usually resolved by entries in
6627 libc. These entry points should be supplied through some other
6628 mechanism when this option is specified.
6632 Do not use the standard system startup files or libraries when linking.
6633 No startup files and only the libraries you specify will be passed to
6634 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
6635 @code{memcpy} and @code{memmove}.
6636 These entries are usually resolved by entries in
6637 libc. These entry points should be supplied through some other
6638 mechanism when this option is specified.
6640 @cindex @option{-lgcc}, use with @option{-nostdlib}
6641 @cindex @option{-nostdlib} and unresolved references
6642 @cindex unresolved references and @option{-nostdlib}
6643 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
6644 @cindex @option{-nodefaultlibs} and unresolved references
6645 @cindex unresolved references and @option{-nodefaultlibs}
6646 One of the standard libraries bypassed by @option{-nostdlib} and
6647 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
6648 that GCC uses to overcome shortcomings of particular machines, or special
6649 needs for some languages.
6650 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
6651 Collection (GCC) Internals},
6652 for more discussion of @file{libgcc.a}.)
6653 In most cases, you need @file{libgcc.a} even when you want to avoid
6654 other standard libraries. In other words, when you specify @option{-nostdlib}
6655 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
6656 This ensures that you have no unresolved references to internal GCC
6657 library subroutines. (For example, @samp{__main}, used to ensure C++
6658 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
6659 GNU Compiler Collection (GCC) Internals}.)
6663 Produce a position independent executable on targets which support it.
6664 For predictable results, you must also specify the same set of options
6665 that were used to generate code (@option{-fpie}, @option{-fPIE},
6666 or model suboptions) when you specify this option.
6670 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
6671 that support it. This instructs the linker to add all symbols, not
6672 only used ones, to the dynamic symbol table. This option is needed
6673 for some uses of @code{dlopen} or to allow obtaining backtraces
6674 from within a program.
6678 Remove all symbol table and relocation information from the executable.
6682 On systems that support dynamic linking, this prevents linking with the shared
6683 libraries. On other systems, this option has no effect.
6687 Produce a shared object which can then be linked with other objects to
6688 form an executable. Not all systems support this option. For predictable
6689 results, you must also specify the same set of options that were used to
6690 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
6691 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
6692 needs to build supplementary stub code for constructors to work. On
6693 multi-libbed systems, @samp{gcc -shared} must select the correct support
6694 libraries to link against. Failing to supply the correct flags may lead
6695 to subtle defects. Supplying them in cases where they are not necessary
6698 @item -shared-libgcc
6699 @itemx -static-libgcc
6700 @opindex shared-libgcc
6701 @opindex static-libgcc
6702 On systems that provide @file{libgcc} as a shared library, these options
6703 force the use of either the shared or static version respectively.
6704 If no shared version of @file{libgcc} was built when the compiler was
6705 configured, these options have no effect.
6707 There are several situations in which an application should use the
6708 shared @file{libgcc} instead of the static version. The most common
6709 of these is when the application wishes to throw and catch exceptions
6710 across different shared libraries. In that case, each of the libraries
6711 as well as the application itself should use the shared @file{libgcc}.
6713 Therefore, the G++ and GCJ drivers automatically add
6714 @option{-shared-libgcc} whenever you build a shared library or a main
6715 executable, because C++ and Java programs typically use exceptions, so
6716 this is the right thing to do.
6718 If, instead, you use the GCC driver to create shared libraries, you may
6719 find that they will not always be linked with the shared @file{libgcc}.
6720 If GCC finds, at its configuration time, that you have a non-GNU linker
6721 or a GNU linker that does not support option @option{--eh-frame-hdr},
6722 it will link the shared version of @file{libgcc} into shared libraries
6723 by default. Otherwise, it will take advantage of the linker and optimize
6724 away the linking with the shared version of @file{libgcc}, linking with
6725 the static version of libgcc by default. This allows exceptions to
6726 propagate through such shared libraries, without incurring relocation
6727 costs at library load time.
6729 However, if a library or main executable is supposed to throw or catch
6730 exceptions, you must link it using the G++ or GCJ driver, as appropriate
6731 for the languages used in the program, or using the option
6732 @option{-shared-libgcc}, such that it is linked with the shared
6737 Bind references to global symbols when building a shared object. Warn
6738 about any unresolved references (unless overridden by the link editor
6739 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
6742 @item -Xlinker @var{option}
6744 Pass @var{option} as an option to the linker. You can use this to
6745 supply system-specific linker options which GCC does not know how to
6748 If you want to pass an option that takes an argument, you must use
6749 @option{-Xlinker} twice, once for the option and once for the argument.
6750 For example, to pass @option{-assert definitions}, you must write
6751 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
6752 @option{-Xlinker "-assert definitions"}, because this passes the entire
6753 string as a single argument, which is not what the linker expects.
6755 @item -Wl,@var{option}
6757 Pass @var{option} as an option to the linker. If @var{option} contains
6758 commas, it is split into multiple options at the commas.
6760 @item -u @var{symbol}
6762 Pretend the symbol @var{symbol} is undefined, to force linking of
6763 library modules to define it. You can use @option{-u} multiple times with
6764 different symbols to force loading of additional library modules.
6767 @node Directory Options
6768 @section Options for Directory Search
6769 @cindex directory options
6770 @cindex options, directory search
6773 These options specify directories to search for header files, for
6774 libraries and for parts of the compiler:
6779 Add the directory @var{dir} to the head of the list of directories to be
6780 searched for header files. This can be used to override a system header
6781 file, substituting your own version, since these directories are
6782 searched before the system header file directories. However, you should
6783 not use this option to add directories that contain vendor-supplied
6784 system header files (use @option{-isystem} for that). If you use more than
6785 one @option{-I} option, the directories are scanned in left-to-right
6786 order; the standard system directories come after.
6788 If a standard system include directory, or a directory specified with
6789 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
6790 option will be ignored. The directory will still be searched but as a
6791 system directory at its normal position in the system include chain.
6792 This is to ensure that GCC's procedure to fix buggy system headers and
6793 the ordering for the include_next directive are not inadvertently changed.
6794 If you really need to change the search order for system directories,
6795 use the @option{-nostdinc} and/or @option{-isystem} options.
6797 @item -iquote@var{dir}
6799 Add the directory @var{dir} to the head of the list of directories to
6800 be searched for header files only for the case of @samp{#include
6801 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
6802 otherwise just like @option{-I}.
6806 Add directory @var{dir} to the list of directories to be searched
6809 @item -B@var{prefix}
6811 This option specifies where to find the executables, libraries,
6812 include files, and data files of the compiler itself.
6814 The compiler driver program runs one or more of the subprograms
6815 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
6816 @var{prefix} as a prefix for each program it tries to run, both with and
6817 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
6819 For each subprogram to be run, the compiler driver first tries the
6820 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
6821 was not specified, the driver tries two standard prefixes, which are
6822 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
6823 those results in a file name that is found, the unmodified program
6824 name is searched for using the directories specified in your
6825 @env{PATH} environment variable.
6827 The compiler will check to see if the path provided by the @option{-B}
6828 refers to a directory, and if necessary it will add a directory
6829 separator character at the end of the path.
6831 @option{-B} prefixes that effectively specify directory names also apply
6832 to libraries in the linker, because the compiler translates these
6833 options into @option{-L} options for the linker. They also apply to
6834 includes files in the preprocessor, because the compiler translates these
6835 options into @option{-isystem} options for the preprocessor. In this case,
6836 the compiler appends @samp{include} to the prefix.
6838 The run-time support file @file{libgcc.a} can also be searched for using
6839 the @option{-B} prefix, if needed. If it is not found there, the two
6840 standard prefixes above are tried, and that is all. The file is left
6841 out of the link if it is not found by those means.
6843 Another way to specify a prefix much like the @option{-B} prefix is to use
6844 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
6847 As a special kludge, if the path provided by @option{-B} is
6848 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
6849 9, then it will be replaced by @file{[dir/]include}. This is to help
6850 with boot-strapping the compiler.
6852 @item -specs=@var{file}
6854 Process @var{file} after the compiler reads in the standard @file{specs}
6855 file, in order to override the defaults that the @file{gcc} driver
6856 program uses when determining what switches to pass to @file{cc1},
6857 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
6858 @option{-specs=@var{file}} can be specified on the command line, and they
6859 are processed in order, from left to right.
6861 @item --sysroot=@var{dir}
6863 Use @var{dir} as the logical root directory for headers and libraries.
6864 For example, if the compiler would normally search for headers in
6865 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
6866 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
6868 If you use both this option and the @option{-isysroot} option, then
6869 the @option{--sysroot} option will apply to libraries, but the
6870 @option{-isysroot} option will apply to header files.
6872 The GNU linker (beginning with version 2.16) has the necessary support
6873 for this option. If your linker does not support this option, the
6874 header file aspect of @option{--sysroot} will still work, but the
6875 library aspect will not.
6879 This option has been deprecated. Please use @option{-iquote} instead for
6880 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
6881 Any directories you specify with @option{-I} options before the @option{-I-}
6882 option are searched only for the case of @samp{#include "@var{file}"};
6883 they are not searched for @samp{#include <@var{file}>}.
6885 If additional directories are specified with @option{-I} options after
6886 the @option{-I-}, these directories are searched for all @samp{#include}
6887 directives. (Ordinarily @emph{all} @option{-I} directories are used
6890 In addition, the @option{-I-} option inhibits the use of the current
6891 directory (where the current input file came from) as the first search
6892 directory for @samp{#include "@var{file}"}. There is no way to
6893 override this effect of @option{-I-}. With @option{-I.} you can specify
6894 searching the directory which was current when the compiler was
6895 invoked. That is not exactly the same as what the preprocessor does
6896 by default, but it is often satisfactory.
6898 @option{-I-} does not inhibit the use of the standard system directories
6899 for header files. Thus, @option{-I-} and @option{-nostdinc} are
6906 @section Specifying subprocesses and the switches to pass to them
6909 @command{gcc} is a driver program. It performs its job by invoking a
6910 sequence of other programs to do the work of compiling, assembling and
6911 linking. GCC interprets its command-line parameters and uses these to
6912 deduce which programs it should invoke, and which command-line options
6913 it ought to place on their command lines. This behavior is controlled
6914 by @dfn{spec strings}. In most cases there is one spec string for each
6915 program that GCC can invoke, but a few programs have multiple spec
6916 strings to control their behavior. The spec strings built into GCC can
6917 be overridden by using the @option{-specs=} command-line switch to specify
6920 @dfn{Spec files} are plaintext files that are used to construct spec
6921 strings. They consist of a sequence of directives separated by blank
6922 lines. The type of directive is determined by the first non-whitespace
6923 character on the line and it can be one of the following:
6926 @item %@var{command}
6927 Issues a @var{command} to the spec file processor. The commands that can
6931 @item %include <@var{file}>
6933 Search for @var{file} and insert its text at the current point in the
6936 @item %include_noerr <@var{file}>
6937 @cindex %include_noerr
6938 Just like @samp{%include}, but do not generate an error message if the include
6939 file cannot be found.
6941 @item %rename @var{old_name} @var{new_name}
6943 Rename the spec string @var{old_name} to @var{new_name}.
6947 @item *[@var{spec_name}]:
6948 This tells the compiler to create, override or delete the named spec
6949 string. All lines after this directive up to the next directive or
6950 blank line are considered to be the text for the spec string. If this
6951 results in an empty string then the spec will be deleted. (Or, if the
6952 spec did not exist, then nothing will happened.) Otherwise, if the spec
6953 does not currently exist a new spec will be created. If the spec does
6954 exist then its contents will be overridden by the text of this
6955 directive, unless the first character of that text is the @samp{+}
6956 character, in which case the text will be appended to the spec.
6958 @item [@var{suffix}]:
6959 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
6960 and up to the next directive or blank line are considered to make up the
6961 spec string for the indicated suffix. When the compiler encounters an
6962 input file with the named suffix, it will processes the spec string in
6963 order to work out how to compile that file. For example:
6970 This says that any input file whose name ends in @samp{.ZZ} should be
6971 passed to the program @samp{z-compile}, which should be invoked with the
6972 command-line switch @option{-input} and with the result of performing the
6973 @samp{%i} substitution. (See below.)
6975 As an alternative to providing a spec string, the text that follows a
6976 suffix directive can be one of the following:
6979 @item @@@var{language}
6980 This says that the suffix is an alias for a known @var{language}. This is
6981 similar to using the @option{-x} command-line switch to GCC to specify a
6982 language explicitly. For example:
6989 Says that .ZZ files are, in fact, C++ source files.
6992 This causes an error messages saying:
6995 @var{name} compiler not installed on this system.
6999 GCC already has an extensive list of suffixes built into it.
7000 This directive will add an entry to the end of the list of suffixes, but
7001 since the list is searched from the end backwards, it is effectively
7002 possible to override earlier entries using this technique.
7006 GCC has the following spec strings built into it. Spec files can
7007 override these strings or create their own. Note that individual
7008 targets can also add their own spec strings to this list.
7011 asm Options to pass to the assembler
7012 asm_final Options to pass to the assembler post-processor
7013 cpp Options to pass to the C preprocessor
7014 cc1 Options to pass to the C compiler
7015 cc1plus Options to pass to the C++ compiler
7016 endfile Object files to include at the end of the link
7017 link Options to pass to the linker
7018 lib Libraries to include on the command line to the linker
7019 libgcc Decides which GCC support library to pass to the linker
7020 linker Sets the name of the linker
7021 predefines Defines to be passed to the C preprocessor
7022 signed_char Defines to pass to CPP to say whether @code{char} is signed
7024 startfile Object files to include at the start of the link
7027 Here is a small example of a spec file:
7033 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
7036 This example renames the spec called @samp{lib} to @samp{old_lib} and
7037 then overrides the previous definition of @samp{lib} with a new one.
7038 The new definition adds in some extra command-line options before
7039 including the text of the old definition.
7041 @dfn{Spec strings} are a list of command-line options to be passed to their
7042 corresponding program. In addition, the spec strings can contain
7043 @samp{%}-prefixed sequences to substitute variable text or to
7044 conditionally insert text into the command line. Using these constructs
7045 it is possible to generate quite complex command lines.
7047 Here is a table of all defined @samp{%}-sequences for spec
7048 strings. Note that spaces are not generated automatically around the
7049 results of expanding these sequences. Therefore you can concatenate them
7050 together or combine them with constant text in a single argument.
7054 Substitute one @samp{%} into the program name or argument.
7057 Substitute the name of the input file being processed.
7060 Substitute the basename of the input file being processed.
7061 This is the substring up to (and not including) the last period
7062 and not including the directory.
7065 This is the same as @samp{%b}, but include the file suffix (text after
7069 Marks the argument containing or following the @samp{%d} as a
7070 temporary file name, so that that file will be deleted if GCC exits
7071 successfully. Unlike @samp{%g}, this contributes no text to the
7074 @item %g@var{suffix}
7075 Substitute a file name that has suffix @var{suffix} and is chosen
7076 once per compilation, and mark the argument in the same way as
7077 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
7078 name is now chosen in a way that is hard to predict even when previously
7079 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
7080 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
7081 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
7082 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
7083 was simply substituted with a file name chosen once per compilation,
7084 without regard to any appended suffix (which was therefore treated
7085 just like ordinary text), making such attacks more likely to succeed.
7087 @item %u@var{suffix}
7088 Like @samp{%g}, but generates a new temporary file name even if
7089 @samp{%u@var{suffix}} was already seen.
7091 @item %U@var{suffix}
7092 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
7093 new one if there is no such last file name. In the absence of any
7094 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
7095 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
7096 would involve the generation of two distinct file names, one
7097 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
7098 simply substituted with a file name chosen for the previous @samp{%u},
7099 without regard to any appended suffix.
7101 @item %j@var{suffix}
7102 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
7103 writable, and if save-temps is off; otherwise, substitute the name
7104 of a temporary file, just like @samp{%u}. This temporary file is not
7105 meant for communication between processes, but rather as a junk
7108 @item %|@var{suffix}
7109 @itemx %m@var{suffix}
7110 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
7111 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
7112 all. These are the two most common ways to instruct a program that it
7113 should read from standard input or write to standard output. If you
7114 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
7115 construct: see for example @file{f/lang-specs.h}.
7117 @item %.@var{SUFFIX}
7118 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
7119 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
7120 terminated by the next space or %.
7123 Marks the argument containing or following the @samp{%w} as the
7124 designated output file of this compilation. This puts the argument
7125 into the sequence of arguments that @samp{%o} will substitute later.
7128 Substitutes the names of all the output files, with spaces
7129 automatically placed around them. You should write spaces
7130 around the @samp{%o} as well or the results are undefined.
7131 @samp{%o} is for use in the specs for running the linker.
7132 Input files whose names have no recognized suffix are not compiled
7133 at all, but they are included among the output files, so they will
7137 Substitutes the suffix for object files. Note that this is
7138 handled specially when it immediately follows @samp{%g, %u, or %U},
7139 because of the need for those to form complete file names. The
7140 handling is such that @samp{%O} is treated exactly as if it had already
7141 been substituted, except that @samp{%g, %u, and %U} do not currently
7142 support additional @var{suffix} characters following @samp{%O} as they would
7143 following, for example, @samp{.o}.
7146 Substitutes the standard macro predefinitions for the
7147 current target machine. Use this when running @code{cpp}.
7150 Like @samp{%p}, but puts @samp{__} before and after the name of each
7151 predefined macro, except for macros that start with @samp{__} or with
7152 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
7156 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
7157 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
7158 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
7159 and @option{-imultilib} as necessary.
7162 Current argument is the name of a library or startup file of some sort.
7163 Search for that file in a standard list of directories and substitute
7164 the full name found.
7167 Print @var{str} as an error message. @var{str} is terminated by a newline.
7168 Use this when inconsistent options are detected.
7171 Substitute the contents of spec string @var{name} at this point.
7174 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
7176 @item %x@{@var{option}@}
7177 Accumulate an option for @samp{%X}.
7180 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
7184 Output the accumulated assembler options specified by @option{-Wa}.
7187 Output the accumulated preprocessor options specified by @option{-Wp}.
7190 Process the @code{asm} spec. This is used to compute the
7191 switches to be passed to the assembler.
7194 Process the @code{asm_final} spec. This is a spec string for
7195 passing switches to an assembler post-processor, if such a program is
7199 Process the @code{link} spec. This is the spec for computing the
7200 command line passed to the linker. Typically it will make use of the
7201 @samp{%L %G %S %D and %E} sequences.
7204 Dump out a @option{-L} option for each directory that GCC believes might
7205 contain startup files. If the target supports multilibs then the
7206 current multilib directory will be prepended to each of these paths.
7209 Process the @code{lib} spec. This is a spec string for deciding which
7210 libraries should be included on the command line to the linker.
7213 Process the @code{libgcc} spec. This is a spec string for deciding
7214 which GCC support library should be included on the command line to the linker.
7217 Process the @code{startfile} spec. This is a spec for deciding which
7218 object files should be the first ones passed to the linker. Typically
7219 this might be a file named @file{crt0.o}.
7222 Process the @code{endfile} spec. This is a spec string that specifies
7223 the last object files that will be passed to the linker.
7226 Process the @code{cpp} spec. This is used to construct the arguments
7227 to be passed to the C preprocessor.
7230 Process the @code{cc1} spec. This is used to construct the options to be
7231 passed to the actual C compiler (@samp{cc1}).
7234 Process the @code{cc1plus} spec. This is used to construct the options to be
7235 passed to the actual C++ compiler (@samp{cc1plus}).
7238 Substitute the variable part of a matched option. See below.
7239 Note that each comma in the substituted string is replaced by
7243 Remove all occurrences of @code{-S} from the command line. Note---this
7244 command is position dependent. @samp{%} commands in the spec string
7245 before this one will see @code{-S}, @samp{%} commands in the spec string
7246 after this one will not.
7248 @item %:@var{function}(@var{args})
7249 Call the named function @var{function}, passing it @var{args}.
7250 @var{args} is first processed as a nested spec string, then split
7251 into an argument vector in the usual fashion. The function returns
7252 a string which is processed as if it had appeared literally as part
7253 of the current spec.
7255 The following built-in spec functions are provided:
7258 @item @code{if-exists}
7259 The @code{if-exists} spec function takes one argument, an absolute
7260 pathname to a file. If the file exists, @code{if-exists} returns the
7261 pathname. Here is a small example of its usage:
7265 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
7268 @item @code{if-exists-else}
7269 The @code{if-exists-else} spec function is similar to the @code{if-exists}
7270 spec function, except that it takes two arguments. The first argument is
7271 an absolute pathname to a file. If the file exists, @code{if-exists-else}
7272 returns the pathname. If it does not exist, it returns the second argument.
7273 This way, @code{if-exists-else} can be used to select one file or another,
7274 based on the existence of the first. Here is a small example of its usage:
7278 crt0%O%s %:if-exists(crti%O%s) \
7279 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
7282 @item @code{replace-outfile}
7283 The @code{replace-outfile} spec function takes two arguments. It looks for the
7284 first argument in the outfiles array and replaces it with the second argument. Here
7285 is a small example of its usage:
7288 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
7294 Substitutes the @code{-S} switch, if that switch was given to GCC@.
7295 If that switch was not specified, this substitutes nothing. Note that
7296 the leading dash is omitted when specifying this option, and it is
7297 automatically inserted if the substitution is performed. Thus the spec
7298 string @samp{%@{foo@}} would match the command-line option @option{-foo}
7299 and would output the command line option @option{-foo}.
7301 @item %W@{@code{S}@}
7302 Like %@{@code{S}@} but mark last argument supplied within as a file to be
7305 @item %@{@code{S}*@}
7306 Substitutes all the switches specified to GCC whose names start
7307 with @code{-S}, but which also take an argument. This is used for
7308 switches like @option{-o}, @option{-D}, @option{-I}, etc.
7309 GCC considers @option{-o foo} as being
7310 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
7311 text, including the space. Thus two arguments would be generated.
7313 @item %@{@code{S}*&@code{T}*@}
7314 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
7315 (the order of @code{S} and @code{T} in the spec is not significant).
7316 There can be any number of ampersand-separated variables; for each the
7317 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
7319 @item %@{@code{S}:@code{X}@}
7320 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
7322 @item %@{!@code{S}:@code{X}@}
7323 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
7325 @item %@{@code{S}*:@code{X}@}
7326 Substitutes @code{X} if one or more switches whose names start with
7327 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
7328 once, no matter how many such switches appeared. However, if @code{%*}
7329 appears somewhere in @code{X}, then @code{X} will be substituted once
7330 for each matching switch, with the @code{%*} replaced by the part of
7331 that switch that matched the @code{*}.
7333 @item %@{.@code{S}:@code{X}@}
7334 Substitutes @code{X}, if processing a file with suffix @code{S}.
7336 @item %@{!.@code{S}:@code{X}@}
7337 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
7339 @item %@{@code{S}|@code{P}:@code{X}@}
7340 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
7341 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
7342 although they have a stronger binding than the @samp{|}. If @code{%*}
7343 appears in @code{X}, all of the alternatives must be starred, and only
7344 the first matching alternative is substituted.
7346 For example, a spec string like this:
7349 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
7352 will output the following command-line options from the following input
7353 command-line options:
7358 -d fred.c -foo -baz -boggle
7359 -d jim.d -bar -baz -boggle
7362 @item %@{S:X; T:Y; :D@}
7364 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
7365 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
7366 be as many clauses as you need. This may be combined with @code{.},
7367 @code{!}, @code{|}, and @code{*} as needed.
7372 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
7373 construct may contain other nested @samp{%} constructs or spaces, or
7374 even newlines. They are processed as usual, as described above.
7375 Trailing white space in @code{X} is ignored. White space may also
7376 appear anywhere on the left side of the colon in these constructs,
7377 except between @code{.} or @code{*} and the corresponding word.
7379 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
7380 handled specifically in these constructs. If another value of
7381 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
7382 @option{-W} switch is found later in the command line, the earlier
7383 switch value is ignored, except with @{@code{S}*@} where @code{S} is
7384 just one letter, which passes all matching options.
7386 The character @samp{|} at the beginning of the predicate text is used to
7387 indicate that a command should be piped to the following command, but
7388 only if @option{-pipe} is specified.
7390 It is built into GCC which switches take arguments and which do not.
7391 (You might think it would be useful to generalize this to allow each
7392 compiler's spec to say which switches take arguments. But this cannot
7393 be done in a consistent fashion. GCC cannot even decide which input
7394 files have been specified without knowing which switches take arguments,
7395 and it must know which input files to compile in order to tell which
7398 GCC also knows implicitly that arguments starting in @option{-l} are to be
7399 treated as compiler output files, and passed to the linker in their
7400 proper position among the other output files.
7402 @c man begin OPTIONS
7404 @node Target Options
7405 @section Specifying Target Machine and Compiler Version
7406 @cindex target options
7407 @cindex cross compiling
7408 @cindex specifying machine version
7409 @cindex specifying compiler version and target machine
7410 @cindex compiler version, specifying
7411 @cindex target machine, specifying
7413 The usual way to run GCC is to run the executable called @file{gcc}, or
7414 @file{<machine>-gcc} when cross-compiling, or
7415 @file{<machine>-gcc-<version>} to run a version other than the one that
7416 was installed last. Sometimes this is inconvenient, so GCC provides
7417 options that will switch to another cross-compiler or version.
7420 @item -b @var{machine}
7422 The argument @var{machine} specifies the target machine for compilation.
7424 The value to use for @var{machine} is the same as was specified as the
7425 machine type when configuring GCC as a cross-compiler. For
7426 example, if a cross-compiler was configured with @samp{configure
7427 arm-elf}, meaning to compile for an arm processor with elf binaries,
7428 then you would specify @option{-b arm-elf} to run that cross compiler.
7429 Because there are other options beginning with @option{-b}, the
7430 configuration must contain a hyphen.
7432 @item -V @var{version}
7434 The argument @var{version} specifies which version of GCC to run.
7435 This is useful when multiple versions are installed. For example,
7436 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
7439 The @option{-V} and @option{-b} options work by running the
7440 @file{<machine>-gcc-<version>} executable, so there's no real reason to
7441 use them if you can just run that directly.
7443 @node Submodel Options
7444 @section Hardware Models and Configurations
7445 @cindex submodel options
7446 @cindex specifying hardware config
7447 @cindex hardware models and configurations, specifying
7448 @cindex machine dependent options
7450 Earlier we discussed the standard option @option{-b} which chooses among
7451 different installed compilers for completely different target
7452 machines, such as VAX vs.@: 68000 vs.@: 80386.
7454 In addition, each of these target machine types can have its own
7455 special options, starting with @samp{-m}, to choose among various
7456 hardware models or configurations---for example, 68010 vs 68020,
7457 floating coprocessor or none. A single installed version of the
7458 compiler can compile for any model or configuration, according to the
7461 Some configurations of the compiler also support additional special
7462 options, usually for compatibility with other compilers on the same
7465 @c This list is ordered alphanumerically by subsection name.
7466 @c It should be the same order and spelling as these options are listed
7467 @c in Machine Dependent Options
7473 * Blackfin Options::
7477 * DEC Alpha Options::
7478 * DEC Alpha/VMS Options::
7480 * GNU/Linux Options::
7483 * i386 and x86-64 Options::
7496 * RS/6000 and PowerPC Options::
7497 * S/390 and zSeries Options::
7502 * System V Options::
7503 * TMS320C3x/C4x Options::
7507 * Xstormy16 Options::
7513 @subsection ARC Options
7516 These options are defined for ARC implementations:
7521 Compile code for little endian mode. This is the default.
7525 Compile code for big endian mode.
7528 @opindex mmangle-cpu
7529 Prepend the name of the cpu to all public symbol names.
7530 In multiple-processor systems, there are many ARC variants with different
7531 instruction and register set characteristics. This flag prevents code
7532 compiled for one cpu to be linked with code compiled for another.
7533 No facility exists for handling variants that are ``almost identical''.
7534 This is an all or nothing option.
7536 @item -mcpu=@var{cpu}
7538 Compile code for ARC variant @var{cpu}.
7539 Which variants are supported depend on the configuration.
7540 All variants support @option{-mcpu=base}, this is the default.
7542 @item -mtext=@var{text-section}
7543 @itemx -mdata=@var{data-section}
7544 @itemx -mrodata=@var{readonly-data-section}
7548 Put functions, data, and readonly data in @var{text-section},
7549 @var{data-section}, and @var{readonly-data-section} respectively
7550 by default. This can be overridden with the @code{section} attribute.
7551 @xref{Variable Attributes}.
7556 @subsection ARM Options
7559 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
7563 @item -mabi=@var{name}
7565 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
7566 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
7569 @opindex mapcs-frame
7570 Generate a stack frame that is compliant with the ARM Procedure Call
7571 Standard for all functions, even if this is not strictly necessary for
7572 correct execution of the code. Specifying @option{-fomit-frame-pointer}
7573 with this option will cause the stack frames not to be generated for
7574 leaf functions. The default is @option{-mno-apcs-frame}.
7578 This is a synonym for @option{-mapcs-frame}.
7581 @c not currently implemented
7582 @item -mapcs-stack-check
7583 @opindex mapcs-stack-check
7584 Generate code to check the amount of stack space available upon entry to
7585 every function (that actually uses some stack space). If there is
7586 insufficient space available then either the function
7587 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
7588 called, depending upon the amount of stack space required. The run time
7589 system is required to provide these functions. The default is
7590 @option{-mno-apcs-stack-check}, since this produces smaller code.
7592 @c not currently implemented
7594 @opindex mapcs-float
7595 Pass floating point arguments using the float point registers. This is
7596 one of the variants of the APCS@. This option is recommended if the
7597 target hardware has a floating point unit or if a lot of floating point
7598 arithmetic is going to be performed by the code. The default is
7599 @option{-mno-apcs-float}, since integer only code is slightly increased in
7600 size if @option{-mapcs-float} is used.
7602 @c not currently implemented
7603 @item -mapcs-reentrant
7604 @opindex mapcs-reentrant
7605 Generate reentrant, position independent code. The default is
7606 @option{-mno-apcs-reentrant}.
7609 @item -mthumb-interwork
7610 @opindex mthumb-interwork
7611 Generate code which supports calling between the ARM and Thumb
7612 instruction sets. Without this option the two instruction sets cannot
7613 be reliably used inside one program. The default is
7614 @option{-mno-thumb-interwork}, since slightly larger code is generated
7615 when @option{-mthumb-interwork} is specified.
7617 @item -mno-sched-prolog
7618 @opindex mno-sched-prolog
7619 Prevent the reordering of instructions in the function prolog, or the
7620 merging of those instruction with the instructions in the function's
7621 body. This means that all functions will start with a recognizable set
7622 of instructions (or in fact one of a choice from a small set of
7623 different function prologues), and this information can be used to
7624 locate the start if functions inside an executable piece of code. The
7625 default is @option{-msched-prolog}.
7628 @opindex mhard-float
7629 Generate output containing floating point instructions. This is the
7633 @opindex msoft-float
7634 Generate output containing library calls for floating point.
7635 @strong{Warning:} the requisite libraries are not available for all ARM
7636 targets. Normally the facilities of the machine's usual C compiler are
7637 used, but this cannot be done directly in cross-compilation. You must make
7638 your own arrangements to provide suitable library functions for
7641 @option{-msoft-float} changes the calling convention in the output file;
7642 therefore, it is only useful if you compile @emph{all} of a program with
7643 this option. In particular, you need to compile @file{libgcc.a}, the
7644 library that comes with GCC, with @option{-msoft-float} in order for
7647 @item -mfloat-abi=@var{name}
7649 Specifies which ABI to use for floating point values. Permissible values
7650 are: @samp{soft}, @samp{softfp} and @samp{hard}.
7652 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
7653 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
7654 of floating point instructions, but still uses the soft-float calling
7657 @item -mlittle-endian
7658 @opindex mlittle-endian
7659 Generate code for a processor running in little-endian mode. This is
7660 the default for all standard configurations.
7663 @opindex mbig-endian
7664 Generate code for a processor running in big-endian mode; the default is
7665 to compile code for a little-endian processor.
7667 @item -mwords-little-endian
7668 @opindex mwords-little-endian
7669 This option only applies when generating code for big-endian processors.
7670 Generate code for a little-endian word order but a big-endian byte
7671 order. That is, a byte order of the form @samp{32107654}. Note: this
7672 option should only be used if you require compatibility with code for
7673 big-endian ARM processors generated by versions of the compiler prior to
7676 @item -mcpu=@var{name}
7678 This specifies the name of the target ARM processor. GCC uses this name
7679 to determine what kind of instructions it can emit when generating
7680 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
7681 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
7682 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
7683 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
7684 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
7685 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
7686 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
7687 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
7688 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
7689 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
7690 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
7691 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
7692 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
7693 @samp{arm1156t2-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
7694 @samp{cortex-a8}, @samp{cortex-r4}, @samp{cortex-m3},
7695 @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
7697 @itemx -mtune=@var{name}
7699 This option is very similar to the @option{-mcpu=} option, except that
7700 instead of specifying the actual target processor type, and hence
7701 restricting which instructions can be used, it specifies that GCC should
7702 tune the performance of the code as if the target were of the type
7703 specified in this option, but still choosing the instructions that it
7704 will generate based on the cpu specified by a @option{-mcpu=} option.
7705 For some ARM implementations better performance can be obtained by using
7708 @item -march=@var{name}
7710 This specifies the name of the target ARM architecture. GCC uses this
7711 name to determine what kind of instructions it can emit when generating
7712 assembly code. This option can be used in conjunction with or instead
7713 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
7714 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
7715 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
7716 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv7}, @samp{armv7-a},
7717 @samp{armv7-r}, @samp{armv7-m}, @samp{iwmmxt}, @samp{ep9312}.
7719 @item -mfpu=@var{name}
7720 @itemx -mfpe=@var{number}
7721 @itemx -mfp=@var{number}
7725 This specifies what floating point hardware (or hardware emulation) is
7726 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
7727 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
7728 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
7729 with older versions of GCC@.
7731 If @option{-msoft-float} is specified this specifies the format of
7732 floating point values.
7734 @item -mstructure-size-boundary=@var{n}
7735 @opindex mstructure-size-boundary
7736 The size of all structures and unions will be rounded up to a multiple
7737 of the number of bits set by this option. Permissible values are 8, 32
7738 and 64. The default value varies for different toolchains. For the COFF
7739 targeted toolchain the default value is 8. A value of 64 is only allowed
7740 if the underlying ABI supports it.
7742 Specifying the larger number can produce faster, more efficient code, but
7743 can also increase the size of the program. Different values are potentially
7744 incompatible. Code compiled with one value cannot necessarily expect to
7745 work with code or libraries compiled with another value, if they exchange
7746 information using structures or unions.
7748 @item -mabort-on-noreturn
7749 @opindex mabort-on-noreturn
7750 Generate a call to the function @code{abort} at the end of a
7751 @code{noreturn} function. It will be executed if the function tries to
7755 @itemx -mno-long-calls
7756 @opindex mlong-calls
7757 @opindex mno-long-calls
7758 Tells the compiler to perform function calls by first loading the
7759 address of the function into a register and then performing a subroutine
7760 call on this register. This switch is needed if the target function
7761 will lie outside of the 64 megabyte addressing range of the offset based
7762 version of subroutine call instruction.
7764 Even if this switch is enabled, not all function calls will be turned
7765 into long calls. The heuristic is that static functions, functions
7766 which have the @samp{short-call} attribute, functions that are inside
7767 the scope of a @samp{#pragma no_long_calls} directive and functions whose
7768 definitions have already been compiled within the current compilation
7769 unit, will not be turned into long calls. The exception to this rule is
7770 that weak function definitions, functions with the @samp{long-call}
7771 attribute or the @samp{section} attribute, and functions that are within
7772 the scope of a @samp{#pragma long_calls} directive, will always be
7773 turned into long calls.
7775 This feature is not enabled by default. Specifying
7776 @option{-mno-long-calls} will restore the default behavior, as will
7777 placing the function calls within the scope of a @samp{#pragma
7778 long_calls_off} directive. Note these switches have no effect on how
7779 the compiler generates code to handle function calls via function
7782 @item -mnop-fun-dllimport
7783 @opindex mnop-fun-dllimport
7784 Disable support for the @code{dllimport} attribute.
7786 @item -msingle-pic-base
7787 @opindex msingle-pic-base
7788 Treat the register used for PIC addressing as read-only, rather than
7789 loading it in the prologue for each function. The run-time system is
7790 responsible for initializing this register with an appropriate value
7791 before execution begins.
7793 @item -mpic-register=@var{reg}
7794 @opindex mpic-register
7795 Specify the register to be used for PIC addressing. The default is R10
7796 unless stack-checking is enabled, when R9 is used.
7798 @item -mcirrus-fix-invalid-insns
7799 @opindex mcirrus-fix-invalid-insns
7800 @opindex mno-cirrus-fix-invalid-insns
7801 Insert NOPs into the instruction stream to in order to work around
7802 problems with invalid Maverick instruction combinations. This option
7803 is only valid if the @option{-mcpu=ep9312} option has been used to
7804 enable generation of instructions for the Cirrus Maverick floating
7805 point co-processor. This option is not enabled by default, since the
7806 problem is only present in older Maverick implementations. The default
7807 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
7810 @item -mpoke-function-name
7811 @opindex mpoke-function-name
7812 Write the name of each function into the text section, directly
7813 preceding the function prologue. The generated code is similar to this:
7817 .ascii "arm_poke_function_name", 0
7820 .word 0xff000000 + (t1 - t0)
7821 arm_poke_function_name
7823 stmfd sp!, @{fp, ip, lr, pc@}
7827 When performing a stack backtrace, code can inspect the value of
7828 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
7829 location @code{pc - 12} and the top 8 bits are set, then we know that
7830 there is a function name embedded immediately preceding this location
7831 and has length @code{((pc[-3]) & 0xff000000)}.
7835 Generate code for the Thumb instruction set. The default is to
7836 use the 32-bit ARM instruction set.
7837 This option automatically enables either 16-bit Thumb-1 or
7838 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
7839 and @option{-march=@var{name}} options.
7842 @opindex mtpcs-frame
7843 Generate a stack frame that is compliant with the Thumb Procedure Call
7844 Standard for all non-leaf functions. (A leaf function is one that does
7845 not call any other functions.) The default is @option{-mno-tpcs-frame}.
7847 @item -mtpcs-leaf-frame
7848 @opindex mtpcs-leaf-frame
7849 Generate a stack frame that is compliant with the Thumb Procedure Call
7850 Standard for all leaf functions. (A leaf function is one that does
7851 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
7853 @item -mcallee-super-interworking
7854 @opindex mcallee-super-interworking
7855 Gives all externally visible functions in the file being compiled an ARM
7856 instruction set header which switches to Thumb mode before executing the
7857 rest of the function. This allows these functions to be called from
7858 non-interworking code.
7860 @item -mcaller-super-interworking
7861 @opindex mcaller-super-interworking
7862 Allows calls via function pointers (including virtual functions) to
7863 execute correctly regardless of whether the target code has been
7864 compiled for interworking or not. There is a small overhead in the cost
7865 of executing a function pointer if this option is enabled.
7867 @item -mtp=@var{name}
7869 Specify the access model for the thread local storage pointer. The valid
7870 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
7871 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
7872 (supported in the arm6k architecture), and @option{auto}, which uses the
7873 best available method for the selected processor. The default setting is
7879 @subsection AVR Options
7882 These options are defined for AVR implementations:
7885 @item -mmcu=@var{mcu}
7887 Specify ATMEL AVR instruction set or MCU type.
7889 Instruction set avr1 is for the minimal AVR core, not supported by the C
7890 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
7891 attiny11, attiny12, attiny15, attiny28).
7893 Instruction set avr2 (default) is for the classic AVR core with up to
7894 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
7895 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
7896 at90c8534, at90s8535).
7898 Instruction set avr3 is for the classic AVR core with up to 128K program
7899 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
7901 Instruction set avr4 is for the enhanced AVR core with up to 8K program
7902 memory space (MCU types: atmega8, atmega83, atmega85).
7904 Instruction set avr5 is for the enhanced AVR core with up to 128K program
7905 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
7906 atmega64, atmega128, at43usb355, at94k).
7910 Output instruction sizes to the asm file.
7912 @item -minit-stack=@var{N}
7913 @opindex minit-stack
7914 Specify the initial stack address, which may be a symbol or numeric value,
7915 @samp{__stack} is the default.
7917 @item -mno-interrupts
7918 @opindex mno-interrupts
7919 Generated code is not compatible with hardware interrupts.
7920 Code size will be smaller.
7922 @item -mcall-prologues
7923 @opindex mcall-prologues
7924 Functions prologues/epilogues expanded as call to appropriate
7925 subroutines. Code size will be smaller.
7927 @item -mno-tablejump
7928 @opindex mno-tablejump
7929 Do not generate tablejump insns which sometimes increase code size.
7932 @opindex mtiny-stack
7933 Change only the low 8 bits of the stack pointer.
7937 Assume int to be 8 bit integer. This affects the sizes of all types: A
7938 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
7939 and long long will be 4 bytes. Please note that this option does not
7940 comply to the C standards, but it will provide you with smaller code
7944 @node Blackfin Options
7945 @subsection Blackfin Options
7946 @cindex Blackfin Options
7949 @item -momit-leaf-frame-pointer
7950 @opindex momit-leaf-frame-pointer
7951 Don't keep the frame pointer in a register for leaf functions. This
7952 avoids the instructions to save, set up and restore frame pointers and
7953 makes an extra register available in leaf functions. The option
7954 @option{-fomit-frame-pointer} removes the frame pointer for all functions
7955 which might make debugging harder.
7957 @item -mspecld-anomaly
7958 @opindex mspecld-anomaly
7959 When enabled, the compiler will ensure that the generated code does not
7960 contain speculative loads after jump instructions. This option is enabled
7963 @item -mno-specld-anomaly
7964 @opindex mno-specld-anomaly
7965 Don't generate extra code to prevent speculative loads from occurring.
7967 @item -mcsync-anomaly
7968 @opindex mcsync-anomaly
7969 When enabled, the compiler will ensure that the generated code does not
7970 contain CSYNC or SSYNC instructions too soon after conditional branches.
7971 This option is enabled by default.
7973 @item -mno-csync-anomaly
7974 @opindex mno-csync-anomaly
7975 Don't generate extra code to prevent CSYNC or SSYNC instructions from
7976 occurring too soon after a conditional branch.
7980 When enabled, the compiler is free to take advantage of the knowledge that
7981 the entire program fits into the low 64k of memory.
7984 @opindex mno-low-64k
7985 Assume that the program is arbitrarily large. This is the default.
7987 @item -mstack-check-l1
7988 @opindex mstack-check-l1
7989 Do stack checking using information placed into L1 scratchpad memory by the
7992 @item -mid-shared-library
7993 @opindex mid-shared-library
7994 Generate code that supports shared libraries via the library ID method.
7995 This allows for execute in place and shared libraries in an environment
7996 without virtual memory management. This option implies @option{-fPIC}.
7998 @item -mno-id-shared-library
7999 @opindex mno-id-shared-library
8000 Generate code that doesn't assume ID based shared libraries are being used.
8001 This is the default.
8003 @item -mleaf-id-shared-library
8004 @opindex mleaf-id-shared-library
8005 Generate code that supports shared libraries via the library ID method,
8006 but assumes that this library or executable won't link against any other
8007 ID shared libraries. That allows the compiler to use faster code for jumps
8010 @item -mno-leaf-id-shared-library
8011 @opindex mno-leaf-id-shared-library
8012 Do not assume that the code being compiled won't link against any ID shared
8013 libraries. Slower code will be generated for jump and call insns.
8015 @item -mshared-library-id=n
8016 @opindex mshared-library-id
8017 Specified the identification number of the ID based shared library being
8018 compiled. Specifying a value of 0 will generate more compact code, specifying
8019 other values will force the allocation of that number to the current
8020 library but is no more space or time efficient than omitting this option.
8024 Generate code that allows the data segment to be located in a different
8025 area of memory from the text segment. This allows for execute in place in
8026 an environment without virtual memory management by eliminating relocations
8027 against the text section.
8030 @opindex mno-sep-data
8031 Generate code that assumes that the data segment follows the text segment.
8032 This is the default.
8035 @itemx -mno-long-calls
8036 @opindex mlong-calls
8037 @opindex mno-long-calls
8038 Tells the compiler to perform function calls by first loading the
8039 address of the function into a register and then performing a subroutine
8040 call on this register. This switch is needed if the target function
8041 will lie outside of the 24 bit addressing range of the offset based
8042 version of subroutine call instruction.
8044 This feature is not enabled by default. Specifying
8045 @option{-mno-long-calls} will restore the default behavior. Note these
8046 switches have no effect on how the compiler generates code to handle
8047 function calls via function pointers.
8051 @subsection CRIS Options
8052 @cindex CRIS Options
8054 These options are defined specifically for the CRIS ports.
8057 @item -march=@var{architecture-type}
8058 @itemx -mcpu=@var{architecture-type}
8061 Generate code for the specified architecture. The choices for
8062 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
8063 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
8064 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
8067 @item -mtune=@var{architecture-type}
8069 Tune to @var{architecture-type} everything applicable about the generated
8070 code, except for the ABI and the set of available instructions. The
8071 choices for @var{architecture-type} are the same as for
8072 @option{-march=@var{architecture-type}}.
8074 @item -mmax-stack-frame=@var{n}
8075 @opindex mmax-stack-frame
8076 Warn when the stack frame of a function exceeds @var{n} bytes.
8078 @item -melinux-stacksize=@var{n}
8079 @opindex melinux-stacksize
8080 Only available with the @samp{cris-axis-aout} target. Arranges for
8081 indications in the program to the kernel loader that the stack of the
8082 program should be set to @var{n} bytes.
8088 The options @option{-metrax4} and @option{-metrax100} are synonyms for
8089 @option{-march=v3} and @option{-march=v8} respectively.
8091 @item -mmul-bug-workaround
8092 @itemx -mno-mul-bug-workaround
8093 @opindex mmul-bug-workaround
8094 @opindex mno-mul-bug-workaround
8095 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
8096 models where it applies. This option is active by default.
8100 Enable CRIS-specific verbose debug-related information in the assembly
8101 code. This option also has the effect to turn off the @samp{#NO_APP}
8102 formatted-code indicator to the assembler at the beginning of the
8107 Do not use condition-code results from previous instruction; always emit
8108 compare and test instructions before use of condition codes.
8110 @item -mno-side-effects
8111 @opindex mno-side-effects
8112 Do not emit instructions with side-effects in addressing modes other than
8116 @itemx -mno-stack-align
8118 @itemx -mno-data-align
8119 @itemx -mconst-align
8120 @itemx -mno-const-align
8121 @opindex mstack-align
8122 @opindex mno-stack-align
8123 @opindex mdata-align
8124 @opindex mno-data-align
8125 @opindex mconst-align
8126 @opindex mno-const-align
8127 These options (no-options) arranges (eliminate arrangements) for the
8128 stack-frame, individual data and constants to be aligned for the maximum
8129 single data access size for the chosen CPU model. The default is to
8130 arrange for 32-bit alignment. ABI details such as structure layout are
8131 not affected by these options.
8139 Similar to the stack- data- and const-align options above, these options
8140 arrange for stack-frame, writable data and constants to all be 32-bit,
8141 16-bit or 8-bit aligned. The default is 32-bit alignment.
8143 @item -mno-prologue-epilogue
8144 @itemx -mprologue-epilogue
8145 @opindex mno-prologue-epilogue
8146 @opindex mprologue-epilogue
8147 With @option{-mno-prologue-epilogue}, the normal function prologue and
8148 epilogue that sets up the stack-frame are omitted and no return
8149 instructions or return sequences are generated in the code. Use this
8150 option only together with visual inspection of the compiled code: no
8151 warnings or errors are generated when call-saved registers must be saved,
8152 or storage for local variable needs to be allocated.
8158 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
8159 instruction sequences that load addresses for functions from the PLT part
8160 of the GOT rather than (traditional on other architectures) calls to the
8161 PLT@. The default is @option{-mgotplt}.
8165 Legacy no-op option only recognized with the cris-axis-aout target.
8169 Legacy no-op option only recognized with the cris-axis-elf and
8170 cris-axis-linux-gnu targets.
8174 Only recognized with the cris-axis-aout target, where it selects a
8175 GNU/linux-like multilib, include files and instruction set for
8180 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
8184 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
8185 to link with input-output functions from a simulator library. Code,
8186 initialized data and zero-initialized data are allocated consecutively.
8190 Like @option{-sim}, but pass linker options to locate initialized data at
8191 0x40000000 and zero-initialized data at 0x80000000.
8195 @subsection CRX Options
8198 These options are defined specifically for the CRX ports.
8204 Enable the use of multiply-accumulate instructions. Disabled by default.
8208 Push instructions will be used to pass outgoing arguments when functions
8209 are called. Enabled by default.
8212 @node Darwin Options
8213 @subsection Darwin Options
8214 @cindex Darwin options
8216 These options are defined for all architectures running the Darwin operating
8219 FSF GCC on Darwin does not create ``fat'' object files; it will create
8220 an object file for the single architecture that it was built to
8221 target. Apple's GCC on Darwin does create ``fat'' files if multiple
8222 @option{-arch} options are used; it does so by running the compiler or
8223 linker multiple times and joining the results together with
8226 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
8227 @samp{i686}) is determined by the flags that specify the ISA
8228 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
8229 @option{-force_cpusubtype_ALL} option can be used to override this.
8231 The Darwin tools vary in their behavior when presented with an ISA
8232 mismatch. The assembler, @file{as}, will only permit instructions to
8233 be used that are valid for the subtype of the file it is generating,
8234 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
8235 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
8236 and print an error if asked to create a shared library with a less
8237 restrictive subtype than its input files (for instance, trying to put
8238 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
8239 for executables, @file{ld}, will quietly give the executable the most
8240 restrictive subtype of any of its input files.
8245 Add the framework directory @var{dir} to the head of the list of
8246 directories to be searched for header files. These directories are
8247 interleaved with those specified by @option{-I} options and are
8248 scanned in a left-to-right order.
8250 A framework directory is a directory with frameworks in it. A
8251 framework is a directory with a @samp{"Headers"} and/or
8252 @samp{"PrivateHeaders"} directory contained directly in it that ends
8253 in @samp{".framework"}. The name of a framework is the name of this
8254 directory excluding the @samp{".framework"}. Headers associated with
8255 the framework are found in one of those two directories, with
8256 @samp{"Headers"} being searched first. A subframework is a framework
8257 directory that is in a framework's @samp{"Frameworks"} directory.
8258 Includes of subframework headers can only appear in a header of a
8259 framework that contains the subframework, or in a sibling subframework
8260 header. Two subframeworks are siblings if they occur in the same
8261 framework. A subframework should not have the same name as a
8262 framework, a warning will be issued if this is violated. Currently a
8263 subframework cannot have subframeworks, in the future, the mechanism
8264 may be extended to support this. The standard frameworks can be found
8265 in @samp{"/System/Library/Frameworks"} and
8266 @samp{"/Library/Frameworks"}. An example include looks like
8267 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
8268 the name of the framework and header.h is found in the
8269 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
8273 Emit debugging information for symbols that are used. For STABS
8274 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
8275 This is by default ON@.
8279 Emit debugging information for all symbols and types.
8281 @item -mmacosx-version-min=@var{version}
8282 The earliest version of MacOS X that this executable will run on
8283 is @var{version}. Typical values of @var{version} include @code{10.1},
8284 @code{10.2}, and @code{10.3.9}.
8286 The default for this option is to make choices that seem to be most
8291 Enable kernel development mode. The @option{-mkernel} option sets
8292 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
8293 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
8294 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
8295 applicable. This mode also sets @option{-mno-altivec},
8296 @option{-msoft-float}, @option{-fno-builtin} and
8297 @option{-mlong-branch} for PowerPC targets.
8299 @item -mone-byte-bool
8300 @opindex -mone-byte-bool
8301 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
8302 By default @samp{sizeof(bool)} is @samp{4} when compiling for
8303 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
8304 option has no effect on x86.
8306 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
8307 to generate code that is not binary compatible with code generated
8308 without that switch. Using this switch may require recompiling all
8309 other modules in a program, including system libraries. Use this
8310 switch to conform to a non-default data model.
8312 @item -mfix-and-continue
8313 @itemx -ffix-and-continue
8314 @itemx -findirect-data
8315 @opindex mfix-and-continue
8316 @opindex ffix-and-continue
8317 @opindex findirect-data
8318 Generate code suitable for fast turn around development. Needed to
8319 enable gdb to dynamically load @code{.o} files into already running
8320 programs. @option{-findirect-data} and @option{-ffix-and-continue}
8321 are provided for backwards compatibility.
8325 Loads all members of static archive libraries.
8326 See man ld(1) for more information.
8328 @item -arch_errors_fatal
8329 @opindex arch_errors_fatal
8330 Cause the errors having to do with files that have the wrong architecture
8334 @opindex bind_at_load
8335 Causes the output file to be marked such that the dynamic linker will
8336 bind all undefined references when the file is loaded or launched.
8340 Produce a Mach-o bundle format file.
8341 See man ld(1) for more information.
8343 @item -bundle_loader @var{executable}
8344 @opindex bundle_loader
8345 This option specifies the @var{executable} that will be loading the build
8346 output file being linked. See man ld(1) for more information.
8349 @opindex -dynamiclib
8350 When passed this option, GCC will produce a dynamic library instead of
8351 an executable when linking, using the Darwin @file{libtool} command.
8353 @item -force_cpusubtype_ALL
8354 @opindex -force_cpusubtype_ALL
8355 This causes GCC's output file to have the @var{ALL} subtype, instead of
8356 one controlled by the @option{-mcpu} or @option{-march} option.
8358 @item -allowable_client @var{client_name}
8360 @itemx -compatibility_version
8361 @itemx -current_version
8363 @itemx -dependency-file
8365 @itemx -dylinker_install_name
8367 @itemx -exported_symbols_list
8369 @itemx -flat_namespace
8370 @itemx -force_flat_namespace
8371 @itemx -headerpad_max_install_names
8374 @itemx -install_name
8375 @itemx -keep_private_externs
8376 @itemx -multi_module
8377 @itemx -multiply_defined
8378 @itemx -multiply_defined_unused
8380 @itemx -no_dead_strip_inits_and_terms
8381 @itemx -nofixprebinding
8384 @itemx -noseglinkedit
8385 @itemx -pagezero_size
8387 @itemx -prebind_all_twolevel_modules
8388 @itemx -private_bundle
8389 @itemx -read_only_relocs
8391 @itemx -sectobjectsymbols
8395 @itemx -sectobjectsymbols
8398 @itemx -segs_read_only_addr
8399 @itemx -segs_read_write_addr
8400 @itemx -seg_addr_table
8401 @itemx -seg_addr_table_filename
8404 @itemx -segs_read_only_addr
8405 @itemx -segs_read_write_addr
8406 @itemx -single_module
8409 @itemx -sub_umbrella
8410 @itemx -twolevel_namespace
8413 @itemx -unexported_symbols_list
8414 @itemx -weak_reference_mismatches
8417 @opindex allowable_client
8418 @opindex client_name
8419 @opindex compatibility_version
8420 @opindex current_version
8422 @opindex dependency-file
8424 @opindex dylinker_install_name
8426 @opindex exported_symbols_list
8428 @opindex flat_namespace
8429 @opindex force_flat_namespace
8430 @opindex headerpad_max_install_names
8433 @opindex install_name
8434 @opindex keep_private_externs
8435 @opindex multi_module
8436 @opindex multiply_defined
8437 @opindex multiply_defined_unused
8439 @opindex no_dead_strip_inits_and_terms
8440 @opindex nofixprebinding
8441 @opindex nomultidefs
8443 @opindex noseglinkedit
8444 @opindex pagezero_size
8446 @opindex prebind_all_twolevel_modules
8447 @opindex private_bundle
8448 @opindex read_only_relocs
8450 @opindex sectobjectsymbols
8454 @opindex sectobjectsymbols
8457 @opindex segs_read_only_addr
8458 @opindex segs_read_write_addr
8459 @opindex seg_addr_table
8460 @opindex seg_addr_table_filename
8461 @opindex seglinkedit
8463 @opindex segs_read_only_addr
8464 @opindex segs_read_write_addr
8465 @opindex single_module
8467 @opindex sub_library
8468 @opindex sub_umbrella
8469 @opindex twolevel_namespace
8472 @opindex unexported_symbols_list
8473 @opindex weak_reference_mismatches
8474 @opindex whatsloaded
8476 These options are passed to the Darwin linker. The Darwin linker man page
8477 describes them in detail.
8480 @node DEC Alpha Options
8481 @subsection DEC Alpha Options
8483 These @samp{-m} options are defined for the DEC Alpha implementations:
8486 @item -mno-soft-float
8488 @opindex mno-soft-float
8489 @opindex msoft-float
8490 Use (do not use) the hardware floating-point instructions for
8491 floating-point operations. When @option{-msoft-float} is specified,
8492 functions in @file{libgcc.a} will be used to perform floating-point
8493 operations. Unless they are replaced by routines that emulate the
8494 floating-point operations, or compiled in such a way as to call such
8495 emulations routines, these routines will issue floating-point
8496 operations. If you are compiling for an Alpha without floating-point
8497 operations, you must ensure that the library is built so as not to call
8500 Note that Alpha implementations without floating-point operations are
8501 required to have floating-point registers.
8506 @opindex mno-fp-regs
8507 Generate code that uses (does not use) the floating-point register set.
8508 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8509 register set is not used, floating point operands are passed in integer
8510 registers as if they were integers and floating-point results are passed
8511 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8512 so any function with a floating-point argument or return value called by code
8513 compiled with @option{-mno-fp-regs} must also be compiled with that
8516 A typical use of this option is building a kernel that does not use,
8517 and hence need not save and restore, any floating-point registers.
8521 The Alpha architecture implements floating-point hardware optimized for
8522 maximum performance. It is mostly compliant with the IEEE floating
8523 point standard. However, for full compliance, software assistance is
8524 required. This option generates code fully IEEE compliant code
8525 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8526 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8527 defined during compilation. The resulting code is less efficient but is
8528 able to correctly support denormalized numbers and exceptional IEEE
8529 values such as not-a-number and plus/minus infinity. Other Alpha
8530 compilers call this option @option{-ieee_with_no_inexact}.
8532 @item -mieee-with-inexact
8533 @opindex mieee-with-inexact
8534 This is like @option{-mieee} except the generated code also maintains
8535 the IEEE @var{inexact-flag}. Turning on this option causes the
8536 generated code to implement fully-compliant IEEE math. In addition to
8537 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8538 macro. On some Alpha implementations the resulting code may execute
8539 significantly slower than the code generated by default. Since there is
8540 very little code that depends on the @var{inexact-flag}, you should
8541 normally not specify this option. Other Alpha compilers call this
8542 option @option{-ieee_with_inexact}.
8544 @item -mfp-trap-mode=@var{trap-mode}
8545 @opindex mfp-trap-mode
8546 This option controls what floating-point related traps are enabled.
8547 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8548 The trap mode can be set to one of four values:
8552 This is the default (normal) setting. The only traps that are enabled
8553 are the ones that cannot be disabled in software (e.g., division by zero
8557 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8561 Like @samp{u}, but the instructions are marked to be safe for software
8562 completion (see Alpha architecture manual for details).
8565 Like @samp{su}, but inexact traps are enabled as well.
8568 @item -mfp-rounding-mode=@var{rounding-mode}
8569 @opindex mfp-rounding-mode
8570 Selects the IEEE rounding mode. Other Alpha compilers call this option
8571 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8576 Normal IEEE rounding mode. Floating point numbers are rounded towards
8577 the nearest machine number or towards the even machine number in case
8581 Round towards minus infinity.
8584 Chopped rounding mode. Floating point numbers are rounded towards zero.
8587 Dynamic rounding mode. A field in the floating point control register
8588 (@var{fpcr}, see Alpha architecture reference manual) controls the
8589 rounding mode in effect. The C library initializes this register for
8590 rounding towards plus infinity. Thus, unless your program modifies the
8591 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8594 @item -mtrap-precision=@var{trap-precision}
8595 @opindex mtrap-precision
8596 In the Alpha architecture, floating point traps are imprecise. This
8597 means without software assistance it is impossible to recover from a
8598 floating trap and program execution normally needs to be terminated.
8599 GCC can generate code that can assist operating system trap handlers
8600 in determining the exact location that caused a floating point trap.
8601 Depending on the requirements of an application, different levels of
8602 precisions can be selected:
8606 Program precision. This option is the default and means a trap handler
8607 can only identify which program caused a floating point exception.
8610 Function precision. The trap handler can determine the function that
8611 caused a floating point exception.
8614 Instruction precision. The trap handler can determine the exact
8615 instruction that caused a floating point exception.
8618 Other Alpha compilers provide the equivalent options called
8619 @option{-scope_safe} and @option{-resumption_safe}.
8621 @item -mieee-conformant
8622 @opindex mieee-conformant
8623 This option marks the generated code as IEEE conformant. You must not
8624 use this option unless you also specify @option{-mtrap-precision=i} and either
8625 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8626 is to emit the line @samp{.eflag 48} in the function prologue of the
8627 generated assembly file. Under DEC Unix, this has the effect that
8628 IEEE-conformant math library routines will be linked in.
8630 @item -mbuild-constants
8631 @opindex mbuild-constants
8632 Normally GCC examines a 32- or 64-bit integer constant to
8633 see if it can construct it from smaller constants in two or three
8634 instructions. If it cannot, it will output the constant as a literal and
8635 generate code to load it from the data segment at runtime.
8637 Use this option to require GCC to construct @emph{all} integer constants
8638 using code, even if it takes more instructions (the maximum is six).
8640 You would typically use this option to build a shared library dynamic
8641 loader. Itself a shared library, it must relocate itself in memory
8642 before it can find the variables and constants in its own data segment.
8648 Select whether to generate code to be assembled by the vendor-supplied
8649 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8667 Indicate whether GCC should generate code to use the optional BWX,
8668 CIX, FIX and MAX instruction sets. The default is to use the instruction
8669 sets supported by the CPU type specified via @option{-mcpu=} option or that
8670 of the CPU on which GCC was built if none was specified.
8675 @opindex mfloat-ieee
8676 Generate code that uses (does not use) VAX F and G floating point
8677 arithmetic instead of IEEE single and double precision.
8679 @item -mexplicit-relocs
8680 @itemx -mno-explicit-relocs
8681 @opindex mexplicit-relocs
8682 @opindex mno-explicit-relocs
8683 Older Alpha assemblers provided no way to generate symbol relocations
8684 except via assembler macros. Use of these macros does not allow
8685 optimal instruction scheduling. GNU binutils as of version 2.12
8686 supports a new syntax that allows the compiler to explicitly mark
8687 which relocations should apply to which instructions. This option
8688 is mostly useful for debugging, as GCC detects the capabilities of
8689 the assembler when it is built and sets the default accordingly.
8693 @opindex msmall-data
8694 @opindex mlarge-data
8695 When @option{-mexplicit-relocs} is in effect, static data is
8696 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8697 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8698 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8699 16-bit relocations off of the @code{$gp} register. This limits the
8700 size of the small data area to 64KB, but allows the variables to be
8701 directly accessed via a single instruction.
8703 The default is @option{-mlarge-data}. With this option the data area
8704 is limited to just below 2GB@. Programs that require more than 2GB of
8705 data must use @code{malloc} or @code{mmap} to allocate the data in the
8706 heap instead of in the program's data segment.
8708 When generating code for shared libraries, @option{-fpic} implies
8709 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8713 @opindex msmall-text
8714 @opindex mlarge-text
8715 When @option{-msmall-text} is used, the compiler assumes that the
8716 code of the entire program (or shared library) fits in 4MB, and is
8717 thus reachable with a branch instruction. When @option{-msmall-data}
8718 is used, the compiler can assume that all local symbols share the
8719 same @code{$gp} value, and thus reduce the number of instructions
8720 required for a function call from 4 to 1.
8722 The default is @option{-mlarge-text}.
8724 @item -mcpu=@var{cpu_type}
8726 Set the instruction set and instruction scheduling parameters for
8727 machine type @var{cpu_type}. You can specify either the @samp{EV}
8728 style name or the corresponding chip number. GCC supports scheduling
8729 parameters for the EV4, EV5 and EV6 family of processors and will
8730 choose the default values for the instruction set from the processor
8731 you specify. If you do not specify a processor type, GCC will default
8732 to the processor on which the compiler was built.
8734 Supported values for @var{cpu_type} are
8740 Schedules as an EV4 and has no instruction set extensions.
8744 Schedules as an EV5 and has no instruction set extensions.
8748 Schedules as an EV5 and supports the BWX extension.
8753 Schedules as an EV5 and supports the BWX and MAX extensions.
8757 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8761 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8764 @item -mtune=@var{cpu_type}
8766 Set only the instruction scheduling parameters for machine type
8767 @var{cpu_type}. The instruction set is not changed.
8769 @item -mmemory-latency=@var{time}
8770 @opindex mmemory-latency
8771 Sets the latency the scheduler should assume for typical memory
8772 references as seen by the application. This number is highly
8773 dependent on the memory access patterns used by the application
8774 and the size of the external cache on the machine.
8776 Valid options for @var{time} are
8780 A decimal number representing clock cycles.
8786 The compiler contains estimates of the number of clock cycles for
8787 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8788 (also called Dcache, Scache, and Bcache), as well as to main memory.
8789 Note that L3 is only valid for EV5.
8794 @node DEC Alpha/VMS Options
8795 @subsection DEC Alpha/VMS Options
8797 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8800 @item -mvms-return-codes
8801 @opindex mvms-return-codes
8802 Return VMS condition codes from main. The default is to return POSIX
8803 style condition (e.g.@ error) codes.
8807 @subsection FRV Options
8814 Only use the first 32 general purpose registers.
8819 Use all 64 general purpose registers.
8824 Use only the first 32 floating point registers.
8829 Use all 64 floating point registers
8832 @opindex mhard-float
8834 Use hardware instructions for floating point operations.
8837 @opindex msoft-float
8839 Use library routines for floating point operations.
8844 Dynamically allocate condition code registers.
8849 Do not try to dynamically allocate condition code registers, only
8850 use @code{icc0} and @code{fcc0}.
8855 Change ABI to use double word insns.
8860 Do not use double word instructions.
8865 Use floating point double instructions.
8870 Do not use floating point double instructions.
8875 Use media instructions.
8880 Do not use media instructions.
8885 Use multiply and add/subtract instructions.
8890 Do not use multiply and add/subtract instructions.
8895 Select the FDPIC ABI, that uses function descriptors to represent
8896 pointers to functions. Without any PIC/PIE-related options, it
8897 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
8898 assumes GOT entries and small data are within a 12-bit range from the
8899 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
8900 are computed with 32 bits.
8903 @opindex minline-plt
8905 Enable inlining of PLT entries in function calls to functions that are
8906 not known to bind locally. It has no effect without @option{-mfdpic}.
8907 It's enabled by default if optimizing for speed and compiling for
8908 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
8909 optimization option such as @option{-O3} or above is present in the
8915 Assume a large TLS segment when generating thread-local code.
8920 Do not assume a large TLS segment when generating thread-local code.
8925 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
8926 that is known to be in read-only sections. It's enabled by default,
8927 except for @option{-fpic} or @option{-fpie}: even though it may help
8928 make the global offset table smaller, it trades 1 instruction for 4.
8929 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
8930 one of which may be shared by multiple symbols, and it avoids the need
8931 for a GOT entry for the referenced symbol, so it's more likely to be a
8932 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
8934 @item -multilib-library-pic
8935 @opindex multilib-library-pic
8937 Link with the (library, not FD) pic libraries. It's implied by
8938 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
8939 @option{-fpic} without @option{-mfdpic}. You should never have to use
8945 Follow the EABI requirement of always creating a frame pointer whenever
8946 a stack frame is allocated. This option is enabled by default and can
8947 be disabled with @option{-mno-linked-fp}.
8950 @opindex mlong-calls
8952 Use indirect addressing to call functions outside the current
8953 compilation unit. This allows the functions to be placed anywhere
8954 within the 32-bit address space.
8956 @item -malign-labels
8957 @opindex malign-labels
8959 Try to align labels to an 8-byte boundary by inserting nops into the
8960 previous packet. This option only has an effect when VLIW packing
8961 is enabled. It doesn't create new packets; it merely adds nops to
8965 @opindex mlibrary-pic
8967 Generate position-independent EABI code.
8972 Use only the first four media accumulator registers.
8977 Use all eight media accumulator registers.
8982 Pack VLIW instructions.
8987 Do not pack VLIW instructions.
8992 Do not mark ABI switches in e_flags.
8997 Enable the use of conditional-move instructions (default).
8999 This switch is mainly for debugging the compiler and will likely be removed
9000 in a future version.
9002 @item -mno-cond-move
9003 @opindex mno-cond-move
9005 Disable the use of conditional-move instructions.
9007 This switch is mainly for debugging the compiler and will likely be removed
9008 in a future version.
9013 Enable the use of conditional set instructions (default).
9015 This switch is mainly for debugging the compiler and will likely be removed
9016 in a future version.
9021 Disable the use of conditional set instructions.
9023 This switch is mainly for debugging the compiler and will likely be removed
9024 in a future version.
9029 Enable the use of conditional execution (default).
9031 This switch is mainly for debugging the compiler and will likely be removed
9032 in a future version.
9034 @item -mno-cond-exec
9035 @opindex mno-cond-exec
9037 Disable the use of conditional execution.
9039 This switch is mainly for debugging the compiler and will likely be removed
9040 in a future version.
9043 @opindex mvliw-branch
9045 Run a pass to pack branches into VLIW instructions (default).
9047 This switch is mainly for debugging the compiler and will likely be removed
9048 in a future version.
9050 @item -mno-vliw-branch
9051 @opindex mno-vliw-branch
9053 Do not run a pass to pack branches into VLIW instructions.
9055 This switch is mainly for debugging the compiler and will likely be removed
9056 in a future version.
9058 @item -mmulti-cond-exec
9059 @opindex mmulti-cond-exec
9061 Enable optimization of @code{&&} and @code{||} in conditional execution
9064 This switch is mainly for debugging the compiler and will likely be removed
9065 in a future version.
9067 @item -mno-multi-cond-exec
9068 @opindex mno-multi-cond-exec
9070 Disable optimization of @code{&&} and @code{||} in conditional execution.
9072 This switch is mainly for debugging the compiler and will likely be removed
9073 in a future version.
9075 @item -mnested-cond-exec
9076 @opindex mnested-cond-exec
9078 Enable nested conditional execution optimizations (default).
9080 This switch is mainly for debugging the compiler and will likely be removed
9081 in a future version.
9083 @item -mno-nested-cond-exec
9084 @opindex mno-nested-cond-exec
9086 Disable nested conditional execution optimizations.
9088 This switch is mainly for debugging the compiler and will likely be removed
9089 in a future version.
9091 @item -moptimize-membar
9092 @opindex moptimize-membar
9094 This switch removes redundant @code{membar} instructions from the
9095 compiler generated code. It is enabled by default.
9097 @item -mno-optimize-membar
9098 @opindex mno-optimize-membar
9100 This switch disables the automatic removal of redundant @code{membar}
9101 instructions from the generated code.
9103 @item -mtomcat-stats
9104 @opindex mtomcat-stats
9106 Cause gas to print out tomcat statistics.
9108 @item -mcpu=@var{cpu}
9111 Select the processor type for which to generate code. Possible values are
9112 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
9113 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
9117 @node GNU/Linux Options
9118 @subsection GNU/Linux Options
9120 These @samp{-m} options are defined for GNU/Linux targets:
9125 Use the GNU C library instead of uClibc. This is the default except
9126 on @samp{*-*-linux-*uclibc*} targets.
9130 Use uClibc instead of the GNU C library. This is the default on
9131 @samp{*-*-linux-*uclibc*} targets.
9134 @node H8/300 Options
9135 @subsection H8/300 Options
9137 These @samp{-m} options are defined for the H8/300 implementations:
9142 Shorten some address references at link time, when possible; uses the
9143 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9144 ld, Using ld}, for a fuller description.
9148 Generate code for the H8/300H@.
9152 Generate code for the H8S@.
9156 Generate code for the H8S and H8/300H in the normal mode. This switch
9157 must be used either with @option{-mh} or @option{-ms}.
9161 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9165 Make @code{int} data 32 bits by default.
9169 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9170 The default for the H8/300H and H8S is to align longs and floats on 4
9172 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9173 This option has no effect on the H8/300.
9177 @subsection HPPA Options
9178 @cindex HPPA Options
9180 These @samp{-m} options are defined for the HPPA family of computers:
9183 @item -march=@var{architecture-type}
9185 Generate code for the specified architecture. The choices for
9186 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
9187 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
9188 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
9189 architecture option for your machine. Code compiled for lower numbered
9190 architectures will run on higher numbered architectures, but not the
9194 @itemx -mpa-risc-1-1
9195 @itemx -mpa-risc-2-0
9196 @opindex mpa-risc-1-0
9197 @opindex mpa-risc-1-1
9198 @opindex mpa-risc-2-0
9199 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
9202 @opindex mbig-switch
9203 Generate code suitable for big switch tables. Use this option only if
9204 the assembler/linker complain about out of range branches within a switch
9207 @item -mjump-in-delay
9208 @opindex mjump-in-delay
9209 Fill delay slots of function calls with unconditional jump instructions
9210 by modifying the return pointer for the function call to be the target
9211 of the conditional jump.
9213 @item -mdisable-fpregs
9214 @opindex mdisable-fpregs
9215 Prevent floating point registers from being used in any manner. This is
9216 necessary for compiling kernels which perform lazy context switching of
9217 floating point registers. If you use this option and attempt to perform
9218 floating point operations, the compiler will abort.
9220 @item -mdisable-indexing
9221 @opindex mdisable-indexing
9222 Prevent the compiler from using indexing address modes. This avoids some
9223 rather obscure problems when compiling MIG generated code under MACH@.
9225 @item -mno-space-regs
9226 @opindex mno-space-regs
9227 Generate code that assumes the target has no space registers. This allows
9228 GCC to generate faster indirect calls and use unscaled index address modes.
9230 Such code is suitable for level 0 PA systems and kernels.
9232 @item -mfast-indirect-calls
9233 @opindex mfast-indirect-calls
9234 Generate code that assumes calls never cross space boundaries. This
9235 allows GCC to emit code which performs faster indirect calls.
9237 This option will not work in the presence of shared libraries or nested
9240 @item -mfixed-range=@var{register-range}
9241 @opindex mfixed-range
9242 Generate code treating the given register range as fixed registers.
9243 A fixed register is one that the register allocator can not use. This is
9244 useful when compiling kernel code. A register range is specified as
9245 two registers separated by a dash. Multiple register ranges can be
9246 specified separated by a comma.
9248 @item -mlong-load-store
9249 @opindex mlong-load-store
9250 Generate 3-instruction load and store sequences as sometimes required by
9251 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
9254 @item -mportable-runtime
9255 @opindex mportable-runtime
9256 Use the portable calling conventions proposed by HP for ELF systems.
9260 Enable the use of assembler directives only GAS understands.
9262 @item -mschedule=@var{cpu-type}
9264 Schedule code according to the constraints for the machine type
9265 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
9266 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
9267 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
9268 proper scheduling option for your machine. The default scheduling is
9272 @opindex mlinker-opt
9273 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
9274 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
9275 linkers in which they give bogus error messages when linking some programs.
9278 @opindex msoft-float
9279 Generate output containing library calls for floating point.
9280 @strong{Warning:} the requisite libraries are not available for all HPPA
9281 targets. Normally the facilities of the machine's usual C compiler are
9282 used, but this cannot be done directly in cross-compilation. You must make
9283 your own arrangements to provide suitable library functions for
9284 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
9285 does provide software floating point support.
9287 @option{-msoft-float} changes the calling convention in the output file;
9288 therefore, it is only useful if you compile @emph{all} of a program with
9289 this option. In particular, you need to compile @file{libgcc.a}, the
9290 library that comes with GCC, with @option{-msoft-float} in order for
9295 Generate the predefine, @code{_SIO}, for server IO@. The default is
9296 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
9297 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
9298 options are available under HP-UX and HI-UX@.
9302 Use GNU ld specific options. This passes @option{-shared} to ld when
9303 building a shared library. It is the default when GCC is configured,
9304 explicitly or implicitly, with the GNU linker. This option does not
9305 have any affect on which ld is called, it only changes what parameters
9306 are passed to that ld. The ld that is called is determined by the
9307 @option{--with-ld} configure option, GCC's program search path, and
9308 finally by the user's @env{PATH}. The linker used by GCC can be printed
9309 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
9310 on the 64 bit HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9314 Use HP ld specific options. This passes @option{-b} to ld when building
9315 a shared library and passes @option{+Accept TypeMismatch} to ld on all
9316 links. It is the default when GCC is configured, explicitly or
9317 implicitly, with the HP linker. This option does not have any affect on
9318 which ld is called, it only changes what parameters are passed to that
9319 ld. The ld that is called is determined by the @option{--with-ld}
9320 configure option, GCC's program search path, and finally by the user's
9321 @env{PATH}. The linker used by GCC can be printed using @samp{which
9322 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
9323 HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}.
9326 @opindex mno-long-calls
9327 Generate code that uses long call sequences. This ensures that a call
9328 is always able to reach linker generated stubs. The default is to generate
9329 long calls only when the distance from the call site to the beginning
9330 of the function or translation unit, as the case may be, exceeds a
9331 predefined limit set by the branch type being used. The limits for
9332 normal calls are 7,600,000 and 240,000 bytes, respectively for the
9333 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
9336 Distances are measured from the beginning of functions when using the
9337 @option{-ffunction-sections} option, or when using the @option{-mgas}
9338 and @option{-mno-portable-runtime} options together under HP-UX with
9341 It is normally not desirable to use this option as it will degrade
9342 performance. However, it may be useful in large applications,
9343 particularly when partial linking is used to build the application.
9345 The types of long calls used depends on the capabilities of the
9346 assembler and linker, and the type of code being generated. The
9347 impact on systems that support long absolute calls, and long pic
9348 symbol-difference or pc-relative calls should be relatively small.
9349 However, an indirect call is used on 32-bit ELF systems in pic code
9350 and it is quite long.
9352 @item -munix=@var{unix-std}
9354 Generate compiler predefines and select a startfile for the specified
9355 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
9356 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
9357 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
9358 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
9359 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
9362 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
9363 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
9364 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
9365 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
9366 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
9367 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
9369 It is @emph{important} to note that this option changes the interfaces
9370 for various library routines. It also affects the operational behavior
9371 of the C library. Thus, @emph{extreme} care is needed in using this
9374 Library code that is intended to operate with more than one UNIX
9375 standard must test, set and restore the variable @var{__xpg4_extended_mask}
9376 as appropriate. Most GNU software doesn't provide this capability.
9380 Suppress the generation of link options to search libdld.sl when the
9381 @option{-static} option is specified on HP-UX 10 and later.
9385 The HP-UX implementation of setlocale in libc has a dependency on
9386 libdld.sl. There isn't an archive version of libdld.sl. Thus,
9387 when the @option{-static} option is specified, special link options
9388 are needed to resolve this dependency.
9390 On HP-UX 10 and later, the GCC driver adds the necessary options to
9391 link with libdld.sl when the @option{-static} option is specified.
9392 This causes the resulting binary to be dynamic. On the 64-bit port,
9393 the linkers generate dynamic binaries by default in any case. The
9394 @option{-nolibdld} option can be used to prevent the GCC driver from
9395 adding these link options.
9399 Add support for multithreading with the @dfn{dce thread} library
9400 under HP-UX@. This option sets flags for both the preprocessor and
9404 @node i386 and x86-64 Options
9405 @subsection Intel 386 and AMD x86-64 Options
9406 @cindex i386 Options
9407 @cindex x86-64 Options
9408 @cindex Intel 386 Options
9409 @cindex AMD x86-64 Options
9411 These @samp{-m} options are defined for the i386 and x86-64 family of
9415 @item -mtune=@var{cpu-type}
9417 Tune to @var{cpu-type} everything applicable about the generated code, except
9418 for the ABI and the set of available instructions. The choices for
9422 Produce code optimized for the most common IA32/AMD64/EM64T processors.
9423 If you know the CPU on which your code will run, then you should use
9424 the corresponding @option{-mtune} option instead of
9425 @option{-mtune=generic}. But, if you do not know exactly what CPU users
9426 of your application will have, then you should use this option.
9428 As new processors are deployed in the marketplace, the behavior of this
9429 option will change. Therefore, if you upgrade to a newer version of
9430 GCC, the code generated option will change to reflect the processors
9431 that were most common when that version of GCC was released.
9433 There is no @option{-march=generic} option because @option{-march}
9434 indicates the instruction set the compiler can use, and there is no
9435 generic instruction set applicable to all processors. In contrast,
9436 @option{-mtune} indicates the processor (or, in this case, collection of
9437 processors) for which the code is optimized.
9439 This selects the CPU to tune for at compilation time by determining
9440 the processor type of the compiling machine. Using @option{-mtune=native}
9441 will produce code optimized for the local machine under the constraints
9442 of the selected instruction set. Using @option{-march=native} will
9443 enable all instruction subsets supported by the local machine (hence
9444 the result might not run on different machines).
9446 Original Intel's i386 CPU@.
9448 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
9450 Intel Pentium CPU with no MMX support.
9452 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
9454 Intel PentiumPro CPU@.
9456 Same as @code{generic}, but when used as @code{march} option, PentiumPro
9457 instruction set will be used, so the code will run on all i686 family chips.
9459 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
9460 @item pentium3, pentium3m
9461 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
9464 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
9465 support. Used by Centrino notebooks.
9466 @item pentium4, pentium4m
9467 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
9469 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
9472 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
9473 SSE2 and SSE3 instruction set support.
9475 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
9476 instruction set support.
9478 AMD K6 CPU with MMX instruction set support.
9480 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
9481 @item athlon, athlon-tbird
9482 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
9484 @item athlon-4, athlon-xp, athlon-mp
9485 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
9486 instruction set support.
9487 @item k8, opteron, athlon64, athlon-fx
9488 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
9489 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
9491 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
9494 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
9495 instruction set support.
9497 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
9498 implemented for this chip.)
9500 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
9501 implemented for this chip.)
9503 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
9506 While picking a specific @var{cpu-type} will schedule things appropriately
9507 for that particular chip, the compiler will not generate any code that
9508 does not run on the i386 without the @option{-march=@var{cpu-type}} option
9511 @item -march=@var{cpu-type}
9513 Generate instructions for the machine type @var{cpu-type}. The choices
9514 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
9515 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
9517 @item -mcpu=@var{cpu-type}
9519 A deprecated synonym for @option{-mtune}.
9528 @opindex mpentiumpro
9529 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
9530 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
9531 These synonyms are deprecated.
9533 @item -mfpmath=@var{unit}
9535 Generate floating point arithmetics for selected unit @var{unit}. The choices
9540 Use the standard 387 floating point coprocessor present majority of chips and
9541 emulated otherwise. Code compiled with this option will run almost everywhere.
9542 The temporary results are computed in 80bit precision instead of precision
9543 specified by the type resulting in slightly different results compared to most
9544 of other chips. See @option{-ffloat-store} for more detailed description.
9546 This is the default choice for i386 compiler.
9549 Use scalar floating point instructions present in the SSE instruction set.
9550 This instruction set is supported by Pentium3 and newer chips, in the AMD line
9551 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
9552 instruction set supports only single precision arithmetics, thus the double and
9553 extended precision arithmetics is still done using 387. Later version, present
9554 only in Pentium4 and the future AMD x86-64 chips supports double precision
9557 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
9558 or @option{-msse2} switches to enable SSE extensions and make this option
9559 effective. For the x86-64 compiler, these extensions are enabled by default.
9561 The resulting code should be considerably faster in the majority of cases and avoid
9562 the numerical instability problems of 387 code, but may break some existing
9563 code that expects temporaries to be 80bit.
9565 This is the default choice for the x86-64 compiler.
9568 Attempt to utilize both instruction sets at once. This effectively double the
9569 amount of available registers and on chips with separate execution units for
9570 387 and SSE the execution resources too. Use this option with care, as it is
9571 still experimental, because the GCC register allocator does not model separate
9572 functional units well resulting in instable performance.
9575 @item -masm=@var{dialect}
9576 @opindex masm=@var{dialect}
9577 Output asm instructions using selected @var{dialect}. Supported
9578 choices are @samp{intel} or @samp{att} (the default one). Darwin does
9579 not support @samp{intel}.
9584 @opindex mno-ieee-fp
9585 Control whether or not the compiler uses IEEE floating point
9586 comparisons. These handle correctly the case where the result of a
9587 comparison is unordered.
9590 @opindex msoft-float
9591 Generate output containing library calls for floating point.
9592 @strong{Warning:} the requisite libraries are not part of GCC@.
9593 Normally the facilities of the machine's usual C compiler are used, but
9594 this can't be done directly in cross-compilation. You must make your
9595 own arrangements to provide suitable library functions for
9598 On machines where a function returns floating point results in the 80387
9599 register stack, some floating point opcodes may be emitted even if
9600 @option{-msoft-float} is used.
9602 @item -mno-fp-ret-in-387
9603 @opindex mno-fp-ret-in-387
9604 Do not use the FPU registers for return values of functions.
9606 The usual calling convention has functions return values of types
9607 @code{float} and @code{double} in an FPU register, even if there
9608 is no FPU@. The idea is that the operating system should emulate
9611 The option @option{-mno-fp-ret-in-387} causes such values to be returned
9612 in ordinary CPU registers instead.
9614 @item -mno-fancy-math-387
9615 @opindex mno-fancy-math-387
9616 Some 387 emulators do not support the @code{sin}, @code{cos} and
9617 @code{sqrt} instructions for the 387. Specify this option to avoid
9618 generating those instructions. This option is the default on FreeBSD,
9619 OpenBSD and NetBSD@. This option is overridden when @option{-march}
9620 indicates that the target cpu will always have an FPU and so the
9621 instruction will not need emulation. As of revision 2.6.1, these
9622 instructions are not generated unless you also use the
9623 @option{-funsafe-math-optimizations} switch.
9625 @item -malign-double
9626 @itemx -mno-align-double
9627 @opindex malign-double
9628 @opindex mno-align-double
9629 Control whether GCC aligns @code{double}, @code{long double}, and
9630 @code{long long} variables on a two word boundary or a one word
9631 boundary. Aligning @code{double} variables on a two word boundary will
9632 produce code that runs somewhat faster on a @samp{Pentium} at the
9633 expense of more memory.
9635 On x86-64, @option{-malign-double} is enabled by default.
9637 @strong{Warning:} if you use the @option{-malign-double} switch,
9638 structures containing the above types will be aligned differently than
9639 the published application binary interface specifications for the 386
9640 and will not be binary compatible with structures in code compiled
9641 without that switch.
9643 @item -m96bit-long-double
9644 @itemx -m128bit-long-double
9645 @opindex m96bit-long-double
9646 @opindex m128bit-long-double
9647 These switches control the size of @code{long double} type. The i386
9648 application binary interface specifies the size to be 96 bits,
9649 so @option{-m96bit-long-double} is the default in 32 bit mode.
9651 Modern architectures (Pentium and newer) would prefer @code{long double}
9652 to be aligned to an 8 or 16 byte boundary. In arrays or structures
9653 conforming to the ABI, this would not be possible. So specifying a
9654 @option{-m128bit-long-double} will align @code{long double}
9655 to a 16 byte boundary by padding the @code{long double} with an additional
9658 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
9659 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
9661 Notice that neither of these options enable any extra precision over the x87
9662 standard of 80 bits for a @code{long double}.
9664 @strong{Warning:} if you override the default value for your target ABI, the
9665 structures and arrays containing @code{long double} variables will change
9666 their size as well as function calling convention for function taking
9667 @code{long double} will be modified. Hence they will not be binary
9668 compatible with arrays or structures in code compiled without that switch.
9670 @item -mmlarge-data-threshold=@var{number}
9671 @opindex mlarge-data-threshold=@var{number}
9672 When @option{-mcmodel=medium} is specified, the data greater than
9673 @var{threshold} are placed in large data section. This value must be the
9674 same across all object linked into the binary and defaults to 65535.
9677 @itemx -mno-svr3-shlib
9678 @opindex msvr3-shlib
9679 @opindex mno-svr3-shlib
9680 Control whether GCC places uninitialized local variables into the
9681 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
9682 into @code{bss}. These options are meaningful only on System V Release 3.
9686 Use a different function-calling convention, in which functions that
9687 take a fixed number of arguments return with the @code{ret} @var{num}
9688 instruction, which pops their arguments while returning. This saves one
9689 instruction in the caller since there is no need to pop the arguments
9692 You can specify that an individual function is called with this calling
9693 sequence with the function attribute @samp{stdcall}. You can also
9694 override the @option{-mrtd} option by using the function attribute
9695 @samp{cdecl}. @xref{Function Attributes}.
9697 @strong{Warning:} this calling convention is incompatible with the one
9698 normally used on Unix, so you cannot use it if you need to call
9699 libraries compiled with the Unix compiler.
9701 Also, you must provide function prototypes for all functions that
9702 take variable numbers of arguments (including @code{printf});
9703 otherwise incorrect code will be generated for calls to those
9706 In addition, seriously incorrect code will result if you call a
9707 function with too many arguments. (Normally, extra arguments are
9708 harmlessly ignored.)
9710 @item -mregparm=@var{num}
9712 Control how many registers are used to pass integer arguments. By
9713 default, no registers are used to pass arguments, and at most 3
9714 registers can be used. You can control this behavior for a specific
9715 function by using the function attribute @samp{regparm}.
9716 @xref{Function Attributes}.
9718 @strong{Warning:} if you use this switch, and
9719 @var{num} is nonzero, then you must build all modules with the same
9720 value, including any libraries. This includes the system libraries and
9724 @opindex msseregparm
9725 Use SSE register passing conventions for float and double arguments
9726 and return values. You can control this behavior for a specific
9727 function by using the function attribute @samp{sseregparm}.
9728 @xref{Function Attributes}.
9730 @strong{Warning:} if you use this switch then you must build all
9731 modules with the same value, including any libraries. This includes
9732 the system libraries and startup modules.
9734 @item -mstackrealign
9735 @opindex mstackrealign
9736 Realign the stack at entry. On the Intel x86, the
9737 @option{-mstackrealign} option will generate an alternate prologue and
9738 epilogue that realigns the runtime stack. This supports mixing legacy
9739 codes that keep a 4-byte aligned stack with modern codes that keep a
9740 16-byte stack for SSE compatibility. The alternate prologue and
9741 epilogue are slower and bigger than the regular ones, and the
9742 alternate prologue requires an extra scratch register; this lowers the
9743 number of registers available if used in conjunction with the
9744 @code{regparm} attribute. The @option{-mstackrealign} option is
9745 incompatible with the nested function prologue; this is considered a
9746 hard error. See also the attribute @code{force_align_arg_pointer},
9747 applicable to individual functions.
9749 @item -mpreferred-stack-boundary=@var{num}
9750 @opindex mpreferred-stack-boundary
9751 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
9752 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
9753 the default is 4 (16 bytes or 128 bits).
9755 On Pentium and PentiumPro, @code{double} and @code{long double} values
9756 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
9757 suffer significant run time performance penalties. On Pentium III, the
9758 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
9759 properly if it is not 16 byte aligned.
9761 To ensure proper alignment of this values on the stack, the stack boundary
9762 must be as aligned as that required by any value stored on the stack.
9763 Further, every function must be generated such that it keeps the stack
9764 aligned. Thus calling a function compiled with a higher preferred
9765 stack boundary from a function compiled with a lower preferred stack
9766 boundary will most likely misalign the stack. It is recommended that
9767 libraries that use callbacks always use the default setting.
9769 This extra alignment does consume extra stack space, and generally
9770 increases code size. Code that is sensitive to stack space usage, such
9771 as embedded systems and operating system kernels, may want to reduce the
9772 preferred alignment to @option{-mpreferred-stack-boundary=2}.
9792 These switches enable or disable the use of instructions in the MMX,
9793 SSE, SSE2, SSE3, SSSE3 or 3DNow! extended instruction sets.
9794 These extensions are also available as built-in functions: see
9795 @ref{X86 Built-in Functions}, for details of the functions enabled and
9796 disabled by these switches.
9798 To have SSE/SSE2 instructions generated automatically from floating-point
9799 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
9801 These options will enable GCC to use these extended instructions in
9802 generated code, even without @option{-mfpmath=sse}. Applications which
9803 perform runtime CPU detection must compile separate files for each
9804 supported architecture, using the appropriate flags. In particular,
9805 the file containing the CPU detection code should be compiled without
9809 @itemx -mno-push-args
9811 @opindex mno-push-args
9812 Use PUSH operations to store outgoing parameters. This method is shorter
9813 and usually equally fast as method using SUB/MOV operations and is enabled
9814 by default. In some cases disabling it may improve performance because of
9815 improved scheduling and reduced dependencies.
9817 @item -maccumulate-outgoing-args
9818 @opindex maccumulate-outgoing-args
9819 If enabled, the maximum amount of space required for outgoing arguments will be
9820 computed in the function prologue. This is faster on most modern CPUs
9821 because of reduced dependencies, improved scheduling and reduced stack usage
9822 when preferred stack boundary is not equal to 2. The drawback is a notable
9823 increase in code size. This switch implies @option{-mno-push-args}.
9827 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
9828 on thread-safe exception handling must compile and link all code with the
9829 @option{-mthreads} option. When compiling, @option{-mthreads} defines
9830 @option{-D_MT}; when linking, it links in a special thread helper library
9831 @option{-lmingwthrd} which cleans up per thread exception handling data.
9833 @item -mno-align-stringops
9834 @opindex mno-align-stringops
9835 Do not align destination of inlined string operations. This switch reduces
9836 code size and improves performance in case the destination is already aligned,
9837 but GCC doesn't know about it.
9839 @item -minline-all-stringops
9840 @opindex minline-all-stringops
9841 By default GCC inlines string operations only when destination is known to be
9842 aligned at least to 4 byte boundary. This enables more inlining, increase code
9843 size, but may improve performance of code that depends on fast memcpy, strlen
9844 and memset for short lengths.
9846 @item -minline-stringops-dynamically
9847 @opindex minline-stringops-dynamically
9848 For string operation of unknown size, inline runtime checks so for small
9849 blocks inline code is used, while for large blocks library call is used.
9851 @item -mstringop-strategy=@var{alg}
9852 @opindex mstringop-strategy=@var{alg}
9853 Overwrite internal decision heuristic about particular algorithm to inline
9854 string operation with. The allowed values are @code{rep_byte},
9855 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
9856 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
9857 expanding inline loop, @code{libcall} for always expanding library call.
9859 @item -momit-leaf-frame-pointer
9860 @opindex momit-leaf-frame-pointer
9861 Don't keep the frame pointer in a register for leaf functions. This
9862 avoids the instructions to save, set up and restore frame pointers and
9863 makes an extra register available in leaf functions. The option
9864 @option{-fomit-frame-pointer} removes the frame pointer for all functions
9865 which might make debugging harder.
9867 @item -mtls-direct-seg-refs
9868 @itemx -mno-tls-direct-seg-refs
9869 @opindex mtls-direct-seg-refs
9870 Controls whether TLS variables may be accessed with offsets from the
9871 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
9872 or whether the thread base pointer must be added. Whether or not this
9873 is legal depends on the operating system, and whether it maps the
9874 segment to cover the entire TLS area.
9876 For systems that use GNU libc, the default is on.
9879 These @samp{-m} switches are supported in addition to the above
9880 on AMD x86-64 processors in 64-bit environments.
9887 Generate code for a 32-bit or 64-bit environment.
9888 The 32-bit environment sets int, long and pointer to 32 bits and
9889 generates code that runs on any i386 system.
9890 The 64-bit environment sets int to 32 bits and long and pointer
9891 to 64 bits and generates code for AMD's x86-64 architecture.
9894 @opindex no-red-zone
9895 Do not use a so called red zone for x86-64 code. The red zone is mandated
9896 by the x86-64 ABI, it is a 128-byte area beyond the location of the
9897 stack pointer that will not be modified by signal or interrupt handlers
9898 and therefore can be used for temporary data without adjusting the stack
9899 pointer. The flag @option{-mno-red-zone} disables this red zone.
9901 @item -mcmodel=small
9902 @opindex mcmodel=small
9903 Generate code for the small code model: the program and its symbols must
9904 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
9905 Programs can be statically or dynamically linked. This is the default
9908 @item -mcmodel=kernel
9909 @opindex mcmodel=kernel
9910 Generate code for the kernel code model. The kernel runs in the
9911 negative 2 GB of the address space.
9912 This model has to be used for Linux kernel code.
9914 @item -mcmodel=medium
9915 @opindex mcmodel=medium
9916 Generate code for the medium model: The program is linked in the lower 2
9917 GB of the address space but symbols can be located anywhere in the
9918 address space. Programs can be statically or dynamically linked, but
9919 building of shared libraries are not supported with the medium model.
9921 @item -mcmodel=large
9922 @opindex mcmodel=large
9923 Generate code for the large model: This model makes no assumptions
9924 about addresses and sizes of sections. Currently GCC does not implement
9929 @subsection IA-64 Options
9930 @cindex IA-64 Options
9932 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9936 @opindex mbig-endian
9937 Generate code for a big endian target. This is the default for HP-UX@.
9939 @item -mlittle-endian
9940 @opindex mlittle-endian
9941 Generate code for a little endian target. This is the default for AIX5
9948 Generate (or don't) code for the GNU assembler. This is the default.
9949 @c Also, this is the default if the configure option @option{--with-gnu-as}
9956 Generate (or don't) code for the GNU linker. This is the default.
9957 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9962 Generate code that does not use a global pointer register. The result
9963 is not position independent code, and violates the IA-64 ABI@.
9965 @item -mvolatile-asm-stop
9966 @itemx -mno-volatile-asm-stop
9967 @opindex mvolatile-asm-stop
9968 @opindex mno-volatile-asm-stop
9969 Generate (or don't) a stop bit immediately before and after volatile asm
9972 @item -mregister-names
9973 @itemx -mno-register-names
9974 @opindex mregister-names
9975 @opindex mno-register-names
9976 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9977 the stacked registers. This may make assembler output more readable.
9983 Disable (or enable) optimizations that use the small data section. This may
9984 be useful for working around optimizer bugs.
9987 @opindex mconstant-gp
9988 Generate code that uses a single constant global pointer value. This is
9989 useful when compiling kernel code.
9993 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9994 This is useful when compiling firmware code.
9996 @item -minline-float-divide-min-latency
9997 @opindex minline-float-divide-min-latency
9998 Generate code for inline divides of floating point values
9999 using the minimum latency algorithm.
10001 @item -minline-float-divide-max-throughput
10002 @opindex minline-float-divide-max-throughput
10003 Generate code for inline divides of floating point values
10004 using the maximum throughput algorithm.
10006 @item -minline-int-divide-min-latency
10007 @opindex minline-int-divide-min-latency
10008 Generate code for inline divides of integer values
10009 using the minimum latency algorithm.
10011 @item -minline-int-divide-max-throughput
10012 @opindex minline-int-divide-max-throughput
10013 Generate code for inline divides of integer values
10014 using the maximum throughput algorithm.
10016 @item -minline-sqrt-min-latency
10017 @opindex minline-sqrt-min-latency
10018 Generate code for inline square roots
10019 using the minimum latency algorithm.
10021 @item -minline-sqrt-max-throughput
10022 @opindex minline-sqrt-max-throughput
10023 Generate code for inline square roots
10024 using the maximum throughput algorithm.
10026 @item -mno-dwarf2-asm
10027 @itemx -mdwarf2-asm
10028 @opindex mno-dwarf2-asm
10029 @opindex mdwarf2-asm
10030 Don't (or do) generate assembler code for the DWARF2 line number debugging
10031 info. This may be useful when not using the GNU assembler.
10033 @item -mearly-stop-bits
10034 @itemx -mno-early-stop-bits
10035 @opindex mearly-stop-bits
10036 @opindex mno-early-stop-bits
10037 Allow stop bits to be placed earlier than immediately preceding the
10038 instruction that triggered the stop bit. This can improve instruction
10039 scheduling, but does not always do so.
10041 @item -mfixed-range=@var{register-range}
10042 @opindex mfixed-range
10043 Generate code treating the given register range as fixed registers.
10044 A fixed register is one that the register allocator can not use. This is
10045 useful when compiling kernel code. A register range is specified as
10046 two registers separated by a dash. Multiple register ranges can be
10047 specified separated by a comma.
10049 @item -mtls-size=@var{tls-size}
10051 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
10054 @item -mtune=@var{cpu-type}
10056 Tune the instruction scheduling for a particular CPU, Valid values are
10057 itanium, itanium1, merced, itanium2, and mckinley.
10063 Add support for multithreading using the POSIX threads library. This
10064 option sets flags for both the preprocessor and linker. It does
10065 not affect the thread safety of object code produced by the compiler or
10066 that of libraries supplied with it. These are HP-UX specific flags.
10072 Generate code for a 32-bit or 64-bit environment.
10073 The 32-bit environment sets int, long and pointer to 32 bits.
10074 The 64-bit environment sets int to 32 bits and long and pointer
10075 to 64 bits. These are HP-UX specific flags.
10077 @item -mno-sched-br-data-spec
10078 @itemx -msched-br-data-spec
10079 @opindex -mno-sched-br-data-spec
10080 @opindex -msched-br-data-spec
10081 (Dis/En)able data speculative scheduling before reload.
10082 This will result in generation of the ld.a instructions and
10083 the corresponding check instructions (ld.c / chk.a).
10084 The default is 'disable'.
10086 @item -msched-ar-data-spec
10087 @itemx -mno-sched-ar-data-spec
10088 @opindex -msched-ar-data-spec
10089 @opindex -mno-sched-ar-data-spec
10090 (En/Dis)able data speculative scheduling after reload.
10091 This will result in generation of the ld.a instructions and
10092 the corresponding check instructions (ld.c / chk.a).
10093 The default is 'enable'.
10095 @item -mno-sched-control-spec
10096 @itemx -msched-control-spec
10097 @opindex -mno-sched-control-spec
10098 @opindex -msched-control-spec
10099 (Dis/En)able control speculative scheduling. This feature is
10100 available only during region scheduling (i.e. before reload).
10101 This will result in generation of the ld.s instructions and
10102 the corresponding check instructions chk.s .
10103 The default is 'disable'.
10105 @item -msched-br-in-data-spec
10106 @itemx -mno-sched-br-in-data-spec
10107 @opindex -msched-br-in-data-spec
10108 @opindex -mno-sched-br-in-data-spec
10109 (En/Dis)able speculative scheduling of the instructions that
10110 are dependent on the data speculative loads before reload.
10111 This is effective only with @option{-msched-br-data-spec} enabled.
10112 The default is 'enable'.
10114 @item -msched-ar-in-data-spec
10115 @itemx -mno-sched-ar-in-data-spec
10116 @opindex -msched-ar-in-data-spec
10117 @opindex -mno-sched-ar-in-data-spec
10118 (En/Dis)able speculative scheduling of the instructions that
10119 are dependent on the data speculative loads after reload.
10120 This is effective only with @option{-msched-ar-data-spec} enabled.
10121 The default is 'enable'.
10123 @item -msched-in-control-spec
10124 @itemx -mno-sched-in-control-spec
10125 @opindex -msched-in-control-spec
10126 @opindex -mno-sched-in-control-spec
10127 (En/Dis)able speculative scheduling of the instructions that
10128 are dependent on the control speculative loads.
10129 This is effective only with @option{-msched-control-spec} enabled.
10130 The default is 'enable'.
10133 @itemx -mno-sched-ldc
10134 @opindex -msched-ldc
10135 @opindex -mno-sched-ldc
10136 (En/Dis)able use of simple data speculation checks ld.c .
10137 If disabled, only chk.a instructions will be emitted to check
10138 data speculative loads.
10139 The default is 'enable'.
10141 @item -mno-sched-control-ldc
10142 @itemx -msched-control-ldc
10143 @opindex -mno-sched-control-ldc
10144 @opindex -msched-control-ldc
10145 (Dis/En)able use of ld.c instructions to check control speculative loads.
10146 If enabled, in case of control speculative load with no speculatively
10147 scheduled dependent instructions this load will be emitted as ld.sa and
10148 ld.c will be used to check it.
10149 The default is 'disable'.
10151 @item -mno-sched-spec-verbose
10152 @itemx -msched-spec-verbose
10153 @opindex -mno-sched-spec-verbose
10154 @opindex -msched-spec-verbose
10155 (Dis/En)able printing of the information about speculative motions.
10157 @item -mno-sched-prefer-non-data-spec-insns
10158 @itemx -msched-prefer-non-data-spec-insns
10159 @opindex -mno-sched-prefer-non-data-spec-insns
10160 @opindex -msched-prefer-non-data-spec-insns
10161 If enabled, data speculative instructions will be chosen for schedule
10162 only if there are no other choices at the moment. This will make
10163 the use of the data speculation much more conservative.
10164 The default is 'disable'.
10166 @item -mno-sched-prefer-non-control-spec-insns
10167 @itemx -msched-prefer-non-control-spec-insns
10168 @opindex -mno-sched-prefer-non-control-spec-insns
10169 @opindex -msched-prefer-non-control-spec-insns
10170 If enabled, control speculative instructions will be chosen for schedule
10171 only if there are no other choices at the moment. This will make
10172 the use of the control speculation much more conservative.
10173 The default is 'disable'.
10175 @item -mno-sched-count-spec-in-critical-path
10176 @itemx -msched-count-spec-in-critical-path
10177 @opindex -mno-sched-count-spec-in-critical-path
10178 @opindex -msched-count-spec-in-critical-path
10179 If enabled, speculative dependencies will be considered during
10180 computation of the instructions priorities. This will make the use of the
10181 speculation a bit more conservative.
10182 The default is 'disable'.
10187 @subsection M32C Options
10188 @cindex M32C options
10191 @item -mcpu=@var{name}
10193 Select the CPU for which code is generated. @var{name} may be one of
10194 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
10195 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
10196 the M32C/80 series.
10200 Specifies that the program will be run on the simulator. This causes
10201 an alternate runtime library to be linked in which supports, for
10202 example, file I/O. You must not use this option when generating
10203 programs that will run on real hardware; you must provide your own
10204 runtime library for whatever I/O functions are needed.
10206 @item -memregs=@var{number}
10208 Specifies the number of memory-based pseudo-registers GCC will use
10209 during code generation. These pseudo-registers will be used like real
10210 registers, so there is a tradeoff between GCC's ability to fit the
10211 code into available registers, and the performance penalty of using
10212 memory instead of registers. Note that all modules in a program must
10213 be compiled with the same value for this option. Because of that, you
10214 must not use this option with the default runtime libraries gcc
10219 @node M32R/D Options
10220 @subsection M32R/D Options
10221 @cindex M32R/D options
10223 These @option{-m} options are defined for Renesas M32R/D architectures:
10228 Generate code for the M32R/2@.
10232 Generate code for the M32R/X@.
10236 Generate code for the M32R@. This is the default.
10238 @item -mmodel=small
10239 @opindex mmodel=small
10240 Assume all objects live in the lower 16MB of memory (so that their addresses
10241 can be loaded with the @code{ld24} instruction), and assume all subroutines
10242 are reachable with the @code{bl} instruction.
10243 This is the default.
10245 The addressability of a particular object can be set with the
10246 @code{model} attribute.
10248 @item -mmodel=medium
10249 @opindex mmodel=medium
10250 Assume objects may be anywhere in the 32-bit address space (the compiler
10251 will generate @code{seth/add3} instructions to load their addresses), and
10252 assume all subroutines are reachable with the @code{bl} instruction.
10254 @item -mmodel=large
10255 @opindex mmodel=large
10256 Assume objects may be anywhere in the 32-bit address space (the compiler
10257 will generate @code{seth/add3} instructions to load their addresses), and
10258 assume subroutines may not be reachable with the @code{bl} instruction
10259 (the compiler will generate the much slower @code{seth/add3/jl}
10260 instruction sequence).
10263 @opindex msdata=none
10264 Disable use of the small data area. Variables will be put into
10265 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
10266 @code{section} attribute has been specified).
10267 This is the default.
10269 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
10270 Objects may be explicitly put in the small data area with the
10271 @code{section} attribute using one of these sections.
10273 @item -msdata=sdata
10274 @opindex msdata=sdata
10275 Put small global and static data in the small data area, but do not
10276 generate special code to reference them.
10279 @opindex msdata=use
10280 Put small global and static data in the small data area, and generate
10281 special instructions to reference them.
10285 @cindex smaller data references
10286 Put global and static objects less than or equal to @var{num} bytes
10287 into the small data or bss sections instead of the normal data or bss
10288 sections. The default value of @var{num} is 8.
10289 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
10290 for this option to have any effect.
10292 All modules should be compiled with the same @option{-G @var{num}} value.
10293 Compiling with different values of @var{num} may or may not work; if it
10294 doesn't the linker will give an error message---incorrect code will not be
10299 Makes the M32R specific code in the compiler display some statistics
10300 that might help in debugging programs.
10302 @item -malign-loops
10303 @opindex malign-loops
10304 Align all loops to a 32-byte boundary.
10306 @item -mno-align-loops
10307 @opindex mno-align-loops
10308 Do not enforce a 32-byte alignment for loops. This is the default.
10310 @item -missue-rate=@var{number}
10311 @opindex missue-rate=@var{number}
10312 Issue @var{number} instructions per cycle. @var{number} can only be 1
10315 @item -mbranch-cost=@var{number}
10316 @opindex mbranch-cost=@var{number}
10317 @var{number} can only be 1 or 2. If it is 1 then branches will be
10318 preferred over conditional code, if it is 2, then the opposite will
10321 @item -mflush-trap=@var{number}
10322 @opindex mflush-trap=@var{number}
10323 Specifies the trap number to use to flush the cache. The default is
10324 12. Valid numbers are between 0 and 15 inclusive.
10326 @item -mno-flush-trap
10327 @opindex mno-flush-trap
10328 Specifies that the cache cannot be flushed by using a trap.
10330 @item -mflush-func=@var{name}
10331 @opindex mflush-func=@var{name}
10332 Specifies the name of the operating system function to call to flush
10333 the cache. The default is @emph{_flush_cache}, but a function call
10334 will only be used if a trap is not available.
10336 @item -mno-flush-func
10337 @opindex mno-flush-func
10338 Indicates that there is no OS function for flushing the cache.
10342 @node M680x0 Options
10343 @subsection M680x0 Options
10344 @cindex M680x0 options
10346 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
10347 The default settings depend on which architecture was selected when
10348 the compiler was configured; the defaults for the most common choices
10352 @item -march=@var{arch}
10354 Generate code for a specific M680x0 or ColdFire instruction set
10355 architecture. Permissible values of @var{arch} for M680x0
10356 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
10357 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
10358 architectures are selected according to Freescale's ISA classification
10359 and the permissible values are: @samp{isaa}, @samp{isaaplus},
10360 @samp{isab} and @samp{isac}.
10362 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
10363 code for a ColdFire target. The @var{arch} in this macro is one of the
10364 @option{-march} arguments given above.
10366 When used together, @option{-march} and @option{-mtune} select code
10367 that runs on a family of similar processors but that is optimized
10368 for a particular microarchitecture.
10370 @item -mcpu=@var{cpu}
10372 Generate code for a specific M680x0 or ColdFire processor.
10373 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
10374 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
10375 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
10376 below, which also classifies the CPUs into families:
10378 @multitable @columnfractions 0.20 0.80
10379 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
10380 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
10381 @item @samp{5206e} @tab @samp{5206e}
10382 @item @samp{5208} @tab @samp{5207} @samp{5208}
10383 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
10384 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
10385 @item @samp{5216} @tab @samp{5214} @samp{5216}
10386 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
10387 @item @samp{5225} @tab @samp{5224} @samp{5225}
10388 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
10389 @item @samp{5249} @tab @samp{5249}
10390 @item @samp{5250} @tab @samp{5250}
10391 @item @samp{5271} @tab @samp{5270} @samp{5271}
10392 @item @samp{5272} @tab @samp{5272}
10393 @item @samp{5275} @tab @samp{5274} @samp{5275}
10394 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
10395 @item @samp{5307} @tab @samp{5307}
10396 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
10397 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
10398 @item @samp{5407} @tab @samp{5407}
10399 @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}
10402 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
10403 @var{arch} is compatible with @var{cpu}. Other combinations of
10404 @option{-mcpu} and @option{-march} are rejected.
10406 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
10407 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
10408 where the value of @var{family} is given by the table above.
10410 @item -mtune=@var{tune}
10412 Tune the code for a particular microarchitecture, within the
10413 constraints set by @option{-march} and @option{-mcpu}.
10414 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
10415 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
10416 and @samp{cpu32}. The ColdFire microarchitectures
10417 are: @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
10419 You can also use @option{-mtune=68020-40} for code that needs
10420 to run relatively well on 68020, 68030 and 68040 targets.
10421 @option{-mtune=68020-60} is similar but includes 68060 targets
10422 as well. These two options select the same tuning decisions as
10423 @option{-m68020-40} and @option{-m68020-60} respectively.
10425 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
10426 when tuning for 680x0 architecture @var{arch}. It also defines
10427 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
10428 option is used. If gcc is tuning for a range of architectures,
10429 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
10430 it defines the macros for every architecture in the range.
10432 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
10433 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
10434 of the arguments given above.
10440 Generate output for a 68000. This is the default
10441 when the compiler is configured for 68000-based systems.
10442 It is equivalent to @option{-march=68000}.
10444 Use this option for microcontrollers with a 68000 or EC000 core,
10445 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
10449 Generate output for a 68010. This is the default
10450 when the compiler is configured for 68010-based systems.
10451 It is equivalent to @option{-march=68010}.
10457 Generate output for a 68020. This is the default
10458 when the compiler is configured for 68020-based systems.
10459 It is equivalent to @option{-march=68020}.
10463 Generate output for a 68030. This is the default when the compiler is
10464 configured for 68030-based systems. It is equivalent to
10465 @option{-march=68030}.
10469 Generate output for a 68040. This is the default when the compiler is
10470 configured for 68040-based systems. It is equivalent to
10471 @option{-march=68040}.
10473 This option inhibits the use of 68881/68882 instructions that have to be
10474 emulated by software on the 68040. Use this option if your 68040 does not
10475 have code to emulate those instructions.
10479 Generate output for a 68060. This is the default when the compiler is
10480 configured for 68060-based systems. It is equivalent to
10481 @option{-march=68060}.
10483 This option inhibits the use of 68020 and 68881/68882 instructions that
10484 have to be emulated by software on the 68060. Use this option if your 68060
10485 does not have code to emulate those instructions.
10489 Generate output for a CPU32. This is the default
10490 when the compiler is configured for CPU32-based systems.
10491 It is equivalent to @option{-march=cpu32}.
10493 Use this option for microcontrollers with a
10494 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
10495 68336, 68340, 68341, 68349 and 68360.
10499 Generate output for a 520X ColdFire CPU. This is the default
10500 when the compiler is configured for 520X-based systems.
10501 It is equivalent to @option{-mcpu=5206}, and is now deprecated
10502 in favor of that option.
10504 Use this option for microcontroller with a 5200 core, including
10505 the MCF5202, MCF5203, MCF5204 and MCF5206.
10509 Generate output for a 5206e ColdFire CPU. The option is now
10510 deprecated in favor of the equivalent @option{-mcpu=5206e}.
10514 Generate output for a member of the ColdFire 528X family.
10515 The option is now deprecated in favor of the equivalent
10516 @option{-mcpu=528x}.
10520 Generate output for a ColdFire 5307 CPU. The option is now deprecated
10521 in favor of the equivalent @option{-mcpu=5307}.
10525 Generate output for a ColdFire 5407 CPU. The option is now deprecated
10526 in favor of the equivalent @option{-mcpu=5407}.
10530 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
10531 This includes use of hardware floating point instructions.
10532 The option is equivalent to @option{-mcpu=547x}, and is now
10533 deprecated in favor of that option.
10537 Generate output for a 68040, without using any of the new instructions.
10538 This results in code which can run relatively efficiently on either a
10539 68020/68881 or a 68030 or a 68040. The generated code does use the
10540 68881 instructions that are emulated on the 68040.
10542 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
10546 Generate output for a 68060, without using any of the new instructions.
10547 This results in code which can run relatively efficiently on either a
10548 68020/68881 or a 68030 or a 68040. The generated code does use the
10549 68881 instructions that are emulated on the 68060.
10551 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
10555 @opindex mhard-float
10557 Generate floating-point instructions. This is the default for 68020
10558 and above, and for ColdFire devices that have an FPU. It defines the
10559 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
10560 on ColdFire targets.
10563 @opindex msoft-float
10564 Do not generate floating-point instructions; use library calls instead.
10565 This is the default for 68000, 68010, and 68832 targets. It is also
10566 the default for ColdFire devices that have no FPU.
10572 Generate (do not generate) ColdFire hardware divide and remainder
10573 instructions. If @option{-march} is used without @option{-mcpu},
10574 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
10575 architectures. Otherwise, the default is taken from the target CPU
10576 (either the default CPU, or the one specified by @option{-mcpu}). For
10577 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
10578 @option{-mcpu=5206e}.
10580 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
10584 Consider type @code{int} to be 16 bits wide, like @code{short int}.
10585 Additionally, parameters passed on the stack are also aligned to a
10586 16-bit boundary even on targets whose API mandates promotion to 32-bit.
10589 @opindex -mno-short
10590 Do not consider type @code{int} to be 16 bits wide. This is the default.
10593 @itemx -mno-bitfield
10594 @opindex mnobitfield
10595 @opindex mno-bitfield
10596 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
10597 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
10601 Do use the bit-field instructions. The @option{-m68020} option implies
10602 @option{-mbitfield}. This is the default if you use a configuration
10603 designed for a 68020.
10607 Use a different function-calling convention, in which functions
10608 that take a fixed number of arguments return with the @code{rtd}
10609 instruction, which pops their arguments while returning. This
10610 saves one instruction in the caller since there is no need to pop
10611 the arguments there.
10613 This calling convention is incompatible with the one normally
10614 used on Unix, so you cannot use it if you need to call libraries
10615 compiled with the Unix compiler.
10617 Also, you must provide function prototypes for all functions that
10618 take variable numbers of arguments (including @code{printf});
10619 otherwise incorrect code will be generated for calls to those
10622 In addition, seriously incorrect code will result if you call a
10623 function with too many arguments. (Normally, extra arguments are
10624 harmlessly ignored.)
10626 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
10627 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
10631 Do not use the calling conventions selected by @option{-mrtd}.
10632 This is the default.
10635 @itemx -mno-align-int
10636 @opindex malign-int
10637 @opindex mno-align-int
10638 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
10639 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
10640 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
10641 Aligning variables on 32-bit boundaries produces code that runs somewhat
10642 faster on processors with 32-bit busses at the expense of more memory.
10644 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
10645 align structures containing the above types differently than
10646 most published application binary interface specifications for the m68k.
10650 Use the pc-relative addressing mode of the 68000 directly, instead of
10651 using a global offset table. At present, this option implies @option{-fpic},
10652 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
10653 not presently supported with @option{-mpcrel}, though this could be supported for
10654 68020 and higher processors.
10656 @item -mno-strict-align
10657 @itemx -mstrict-align
10658 @opindex mno-strict-align
10659 @opindex mstrict-align
10660 Do not (do) assume that unaligned memory references will be handled by
10664 Generate code that allows the data segment to be located in a different
10665 area of memory from the text segment. This allows for execute in place in
10666 an environment without virtual memory management. This option implies
10669 @item -mno-sep-data
10670 Generate code that assumes that the data segment follows the text segment.
10671 This is the default.
10673 @item -mid-shared-library
10674 Generate code that supports shared libraries via the library ID method.
10675 This allows for execute in place and shared libraries in an environment
10676 without virtual memory management. This option implies @option{-fPIC}.
10678 @item -mno-id-shared-library
10679 Generate code that doesn't assume ID based shared libraries are being used.
10680 This is the default.
10682 @item -mshared-library-id=n
10683 Specified the identification number of the ID based shared library being
10684 compiled. Specifying a value of 0 will generate more compact code, specifying
10685 other values will force the allocation of that number to the current
10686 library but is no more space or time efficient than omitting this option.
10690 @node M68hc1x Options
10691 @subsection M68hc1x Options
10692 @cindex M68hc1x options
10694 These are the @samp{-m} options defined for the 68hc11 and 68hc12
10695 microcontrollers. The default values for these options depends on
10696 which style of microcontroller was selected when the compiler was configured;
10697 the defaults for the most common choices are given below.
10704 Generate output for a 68HC11. This is the default
10705 when the compiler is configured for 68HC11-based systems.
10711 Generate output for a 68HC12. This is the default
10712 when the compiler is configured for 68HC12-based systems.
10718 Generate output for a 68HCS12.
10720 @item -mauto-incdec
10721 @opindex mauto-incdec
10722 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
10729 Enable the use of 68HC12 min and max instructions.
10732 @itemx -mno-long-calls
10733 @opindex mlong-calls
10734 @opindex mno-long-calls
10735 Treat all calls as being far away (near). If calls are assumed to be
10736 far away, the compiler will use the @code{call} instruction to
10737 call a function and the @code{rtc} instruction for returning.
10741 Consider type @code{int} to be 16 bits wide, like @code{short int}.
10743 @item -msoft-reg-count=@var{count}
10744 @opindex msoft-reg-count
10745 Specify the number of pseudo-soft registers which are used for the
10746 code generation. The maximum number is 32. Using more pseudo-soft
10747 register may or may not result in better code depending on the program.
10748 The default is 4 for 68HC11 and 2 for 68HC12.
10752 @node MCore Options
10753 @subsection MCore Options
10754 @cindex MCore options
10756 These are the @samp{-m} options defined for the Motorola M*Core
10762 @itemx -mno-hardlit
10764 @opindex mno-hardlit
10765 Inline constants into the code stream if it can be done in two
10766 instructions or less.
10772 Use the divide instruction. (Enabled by default).
10774 @item -mrelax-immediate
10775 @itemx -mno-relax-immediate
10776 @opindex mrelax-immediate
10777 @opindex mno-relax-immediate
10778 Allow arbitrary sized immediates in bit operations.
10780 @item -mwide-bitfields
10781 @itemx -mno-wide-bitfields
10782 @opindex mwide-bitfields
10783 @opindex mno-wide-bitfields
10784 Always treat bit-fields as int-sized.
10786 @item -m4byte-functions
10787 @itemx -mno-4byte-functions
10788 @opindex m4byte-functions
10789 @opindex mno-4byte-functions
10790 Force all functions to be aligned to a four byte boundary.
10792 @item -mcallgraph-data
10793 @itemx -mno-callgraph-data
10794 @opindex mcallgraph-data
10795 @opindex mno-callgraph-data
10796 Emit callgraph information.
10799 @itemx -mno-slow-bytes
10800 @opindex mslow-bytes
10801 @opindex mno-slow-bytes
10802 Prefer word access when reading byte quantities.
10804 @item -mlittle-endian
10805 @itemx -mbig-endian
10806 @opindex mlittle-endian
10807 @opindex mbig-endian
10808 Generate code for a little endian target.
10814 Generate code for the 210 processor.
10818 @subsection MIPS Options
10819 @cindex MIPS options
10825 Generate big-endian code.
10829 Generate little-endian code. This is the default for @samp{mips*el-*-*}
10832 @item -march=@var{arch}
10834 Generate code that will run on @var{arch}, which can be the name of a
10835 generic MIPS ISA, or the name of a particular processor.
10837 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
10838 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
10839 The processor names are:
10840 @samp{4kc}, @samp{4km}, @samp{4kp},
10841 @samp{4kec}, @samp{4kem}, @samp{4kep},
10842 @samp{5kc}, @samp{5kf},
10844 @samp{24kc}, @samp{24kf}, @samp{24kx},
10845 @samp{24kec}, @samp{24kef}, @samp{24kex},
10846 @samp{34kc}, @samp{34kf}, @samp{34kx},
10849 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
10850 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
10851 @samp{rm7000}, @samp{rm9000},
10854 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
10855 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
10856 The special value @samp{from-abi} selects the
10857 most compatible architecture for the selected ABI (that is,
10858 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
10860 In processor names, a final @samp{000} can be abbreviated as @samp{k}
10861 (for example, @samp{-march=r2k}). Prefixes are optional, and
10862 @samp{vr} may be written @samp{r}.
10864 GCC defines two macros based on the value of this option. The first
10865 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
10866 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
10867 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
10868 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
10869 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
10871 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
10872 above. In other words, it will have the full prefix and will not
10873 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
10874 the macro names the resolved architecture (either @samp{"mips1"} or
10875 @samp{"mips3"}). It names the default architecture when no
10876 @option{-march} option is given.
10878 @item -mtune=@var{arch}
10880 Optimize for @var{arch}. Among other things, this option controls
10881 the way instructions are scheduled, and the perceived cost of arithmetic
10882 operations. The list of @var{arch} values is the same as for
10885 When this option is not used, GCC will optimize for the processor
10886 specified by @option{-march}. By using @option{-march} and
10887 @option{-mtune} together, it is possible to generate code that will
10888 run on a family of processors, but optimize the code for one
10889 particular member of that family.
10891 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
10892 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
10893 @samp{-march} ones described above.
10897 Equivalent to @samp{-march=mips1}.
10901 Equivalent to @samp{-march=mips2}.
10905 Equivalent to @samp{-march=mips3}.
10909 Equivalent to @samp{-march=mips4}.
10913 Equivalent to @samp{-march=mips32}.
10917 Equivalent to @samp{-march=mips32r2}.
10921 Equivalent to @samp{-march=mips64}.
10926 @opindex mno-mips16
10927 Generate (do not generate) MIPS16 code. If GCC is targetting a
10928 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
10940 Generate code for the given ABI@.
10942 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
10943 generates 64-bit code when you select a 64-bit architecture, but you
10944 can use @option{-mgp32} to get 32-bit code instead.
10946 For information about the O64 ABI, see
10947 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
10949 GCC supports a variant of the o32 ABI in which floating-point registers
10950 are 64 rather than 32 bits wide. You can select this combination with
10951 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
10952 and @samp{mfhc1} instructions and is therefore only supported for
10953 MIPS32R2 processors.
10955 The register assignments for arguments and return values remain the
10956 same, but each scalar value is passed in a single 64-bit register
10957 rather than a pair of 32-bit registers. For example, scalar
10958 floating-point values are returned in @samp{$f0} only, not a
10959 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
10960 remains the same, but all 64 bits are saved.
10963 @itemx -mno-abicalls
10965 @opindex mno-abicalls
10966 Generate (do not generate) code that is suitable for SVR4-style
10967 dynamic objects. @option{-mabicalls} is the default for SVR4-based
10972 Generate (do not generate) code that is fully position-independent,
10973 and that can therefore be linked into shared libraries. This option
10974 only affects @option{-mabicalls}.
10976 All @option{-mabicalls} code has traditionally been position-independent,
10977 regardless of options like @option{-fPIC} and @option{-fpic}. However,
10978 as an extension, the GNU toolchain allows executables to use absolute
10979 accesses for locally-binding symbols. It can also use shorter GP
10980 initialization sequences and generate direct calls to locally-defined
10981 functions. This mode is selected by @option{-mno-shared}.
10983 @option{-mno-shared} depends on binutils 2.16 or higher and generates
10984 objects that can only be linked by the GNU linker. However, the option
10985 does not affect the ABI of the final executable; it only affects the ABI
10986 of relocatable objects. Using @option{-mno-shared} will generally make
10987 executables both smaller and quicker.
10989 @option{-mshared} is the default.
10995 Lift (do not lift) the usual restrictions on the size of the global
10998 GCC normally uses a single instruction to load values from the GOT@.
10999 While this is relatively efficient, it will only work if the GOT
11000 is smaller than about 64k. Anything larger will cause the linker
11001 to report an error such as:
11003 @cindex relocation truncated to fit (MIPS)
11005 relocation truncated to fit: R_MIPS_GOT16 foobar
11008 If this happens, you should recompile your code with @option{-mxgot}.
11009 It should then work with very large GOTs, although it will also be
11010 less efficient, since it will take three instructions to fetch the
11011 value of a global symbol.
11013 Note that some linkers can create multiple GOTs. If you have such a
11014 linker, you should only need to use @option{-mxgot} when a single object
11015 file accesses more than 64k's worth of GOT entries. Very few do.
11017 These options have no effect unless GCC is generating position
11022 Assume that general-purpose registers are 32 bits wide.
11026 Assume that general-purpose registers are 64 bits wide.
11030 Assume that floating-point registers are 32 bits wide.
11034 Assume that floating-point registers are 64 bits wide.
11037 @opindex mhard-float
11038 Use floating-point coprocessor instructions.
11041 @opindex msoft-float
11042 Do not use floating-point coprocessor instructions. Implement
11043 floating-point calculations using library calls instead.
11045 @item -msingle-float
11046 @opindex msingle-float
11047 Assume that the floating-point coprocessor only supports single-precision
11050 @itemx -mdouble-float
11051 @opindex mdouble-float
11052 Assume that the floating-point coprocessor supports double-precision
11053 operations. This is the default.
11059 Use (do not use) the MIPS DSP ASE. @xref{MIPS DSP Built-in Functions}.
11061 @itemx -mpaired-single
11062 @itemx -mno-paired-single
11063 @opindex mpaired-single
11064 @opindex mno-paired-single
11065 Use (do not use) paired-single floating-point instructions.
11066 @xref{MIPS Paired-Single Support}. This option can only be used
11067 when generating 64-bit code and requires hardware floating-point
11068 support to be enabled.
11073 @opindex mno-mips3d
11074 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
11075 The option @option{-mips3d} implies @option{-mpaired-single}.
11079 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
11080 an explanation of the default and the way that the pointer size is
11085 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
11087 The default size of @code{int}s, @code{long}s and pointers depends on
11088 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
11089 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
11090 32-bit @code{long}s. Pointers are the same size as @code{long}s,
11091 or the same size as integer registers, whichever is smaller.
11097 Assume (do not assume) that all symbols have 32-bit values, regardless
11098 of the selected ABI@. This option is useful in combination with
11099 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
11100 to generate shorter and faster references to symbolic addresses.
11104 @cindex smaller data references (MIPS)
11105 @cindex gp-relative references (MIPS)
11106 Put global and static items less than or equal to @var{num} bytes into
11107 the small data or bss section instead of the normal data or bss section.
11108 This allows the data to be accessed using a single instruction.
11110 All modules should be compiled with the same @option{-G @var{num}}
11113 @item -membedded-data
11114 @itemx -mno-embedded-data
11115 @opindex membedded-data
11116 @opindex mno-embedded-data
11117 Allocate variables to the read-only data section first if possible, then
11118 next in the small data section if possible, otherwise in data. This gives
11119 slightly slower code than the default, but reduces the amount of RAM required
11120 when executing, and thus may be preferred for some embedded systems.
11122 @item -muninit-const-in-rodata
11123 @itemx -mno-uninit-const-in-rodata
11124 @opindex muninit-const-in-rodata
11125 @opindex mno-uninit-const-in-rodata
11126 Put uninitialized @code{const} variables in the read-only data section.
11127 This option is only meaningful in conjunction with @option{-membedded-data}.
11129 @item -msplit-addresses
11130 @itemx -mno-split-addresses
11131 @opindex msplit-addresses
11132 @opindex mno-split-addresses
11133 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
11134 relocation operators. This option has been superseded by
11135 @option{-mexplicit-relocs} but is retained for backwards compatibility.
11137 @item -mexplicit-relocs
11138 @itemx -mno-explicit-relocs
11139 @opindex mexplicit-relocs
11140 @opindex mno-explicit-relocs
11141 Use (do not use) assembler relocation operators when dealing with symbolic
11142 addresses. The alternative, selected by @option{-mno-explicit-relocs},
11143 is to use assembler macros instead.
11145 @option{-mexplicit-relocs} is the default if GCC was configured
11146 to use an assembler that supports relocation operators.
11148 @item -mcheck-zero-division
11149 @itemx -mno-check-zero-division
11150 @opindex mcheck-zero-division
11151 @opindex mno-check-zero-division
11152 Trap (do not trap) on integer division by zero.
11154 The default is @option{-mcheck-zero-division}.
11156 @item -mdivide-traps
11157 @itemx -mdivide-breaks
11158 @opindex mdivide-traps
11159 @opindex mdivide-breaks
11160 MIPS systems check for division by zero by generating either a
11161 conditional trap or a break instruction. Using traps results in
11162 smaller code, but is only supported on MIPS II and later. Also, some
11163 versions of the Linux kernel have a bug that prevents trap from
11164 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
11165 allow conditional traps on architectures that support them and
11166 @option{-mdivide-breaks} to force the use of breaks.
11168 The default is usually @option{-mdivide-traps}, but this can be
11169 overridden at configure time using @option{--with-divide=breaks}.
11170 Divide-by-zero checks can be completely disabled using
11171 @option{-mno-check-zero-division}.
11176 @opindex mno-memcpy
11177 Force (do not force) the use of @code{memcpy()} for non-trivial block
11178 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
11179 most constant-sized copies.
11182 @itemx -mno-long-calls
11183 @opindex mlong-calls
11184 @opindex mno-long-calls
11185 Disable (do not disable) use of the @code{jal} instruction. Calling
11186 functions using @code{jal} is more efficient but requires the caller
11187 and callee to be in the same 256 megabyte segment.
11189 This option has no effect on abicalls code. The default is
11190 @option{-mno-long-calls}.
11196 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
11197 instructions, as provided by the R4650 ISA@.
11200 @itemx -mno-fused-madd
11201 @opindex mfused-madd
11202 @opindex mno-fused-madd
11203 Enable (disable) use of the floating point multiply-accumulate
11204 instructions, when they are available. The default is
11205 @option{-mfused-madd}.
11207 When multiply-accumulate instructions are used, the intermediate
11208 product is calculated to infinite precision and is not subject to
11209 the FCSR Flush to Zero bit. This may be undesirable in some
11214 Tell the MIPS assembler to not run its preprocessor over user
11215 assembler files (with a @samp{.s} suffix) when assembling them.
11218 @itemx -mno-fix-r4000
11219 @opindex mfix-r4000
11220 @opindex mno-fix-r4000
11221 Work around certain R4000 CPU errata:
11224 A double-word or a variable shift may give an incorrect result if executed
11225 immediately after starting an integer division.
11227 A double-word or a variable shift may give an incorrect result if executed
11228 while an integer multiplication is in progress.
11230 An integer division may give an incorrect result if started in a delay slot
11231 of a taken branch or a jump.
11235 @itemx -mno-fix-r4400
11236 @opindex mfix-r4400
11237 @opindex mno-fix-r4400
11238 Work around certain R4400 CPU errata:
11241 A double-word or a variable shift may give an incorrect result if executed
11242 immediately after starting an integer division.
11246 @itemx -mno-fix-vr4120
11247 @opindex mfix-vr4120
11248 Work around certain VR4120 errata:
11251 @code{dmultu} does not always produce the correct result.
11253 @code{div} and @code{ddiv} do not always produce the correct result if one
11254 of the operands is negative.
11256 The workarounds for the division errata rely on special functions in
11257 @file{libgcc.a}. At present, these functions are only provided by
11258 the @code{mips64vr*-elf} configurations.
11260 Other VR4120 errata require a nop to be inserted between certain pairs of
11261 instructions. These errata are handled by the assembler, not by GCC itself.
11264 @opindex mfix-vr4130
11265 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
11266 workarounds are implemented by the assembler rather than by GCC,
11267 although GCC will avoid using @code{mflo} and @code{mfhi} if the
11268 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
11269 instructions are available instead.
11272 @itemx -mno-fix-sb1
11274 Work around certain SB-1 CPU core errata.
11275 (This flag currently works around the SB-1 revision 2
11276 ``F1'' and ``F2'' floating point errata.)
11278 @item -mflush-func=@var{func}
11279 @itemx -mno-flush-func
11280 @opindex mflush-func
11281 Specifies the function to call to flush the I and D caches, or to not
11282 call any such function. If called, the function must take the same
11283 arguments as the common @code{_flush_func()}, that is, the address of the
11284 memory range for which the cache is being flushed, the size of the
11285 memory range, and the number 3 (to flush both caches). The default
11286 depends on the target GCC was configured for, but commonly is either
11287 @samp{_flush_func} or @samp{__cpu_flush}.
11289 @item -mbranch-likely
11290 @itemx -mno-branch-likely
11291 @opindex mbranch-likely
11292 @opindex mno-branch-likely
11293 Enable or disable use of Branch Likely instructions, regardless of the
11294 default for the selected architecture. By default, Branch Likely
11295 instructions may be generated if they are supported by the selected
11296 architecture. An exception is for the MIPS32 and MIPS64 architectures
11297 and processors which implement those architectures; for those, Branch
11298 Likely instructions will not be generated by default because the MIPS32
11299 and MIPS64 architectures specifically deprecate their use.
11301 @item -mfp-exceptions
11302 @itemx -mno-fp-exceptions
11303 @opindex mfp-exceptions
11304 Specifies whether FP exceptions are enabled. This affects how we schedule
11305 FP instructions for some processors. The default is that FP exceptions are
11308 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
11309 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
11312 @item -mvr4130-align
11313 @itemx -mno-vr4130-align
11314 @opindex mvr4130-align
11315 The VR4130 pipeline is two-way superscalar, but can only issue two
11316 instructions together if the first one is 8-byte aligned. When this
11317 option is enabled, GCC will align pairs of instructions that it
11318 thinks should execute in parallel.
11320 This option only has an effect when optimizing for the VR4130.
11321 It normally makes code faster, but at the expense of making it bigger.
11322 It is enabled by default at optimization level @option{-O3}.
11326 @subsection MMIX Options
11327 @cindex MMIX Options
11329 These options are defined for the MMIX:
11333 @itemx -mno-libfuncs
11335 @opindex mno-libfuncs
11336 Specify that intrinsic library functions are being compiled, passing all
11337 values in registers, no matter the size.
11340 @itemx -mno-epsilon
11342 @opindex mno-epsilon
11343 Generate floating-point comparison instructions that compare with respect
11344 to the @code{rE} epsilon register.
11346 @item -mabi=mmixware
11348 @opindex mabi-mmixware
11350 Generate code that passes function parameters and return values that (in
11351 the called function) are seen as registers @code{$0} and up, as opposed to
11352 the GNU ABI which uses global registers @code{$231} and up.
11354 @item -mzero-extend
11355 @itemx -mno-zero-extend
11356 @opindex mzero-extend
11357 @opindex mno-zero-extend
11358 When reading data from memory in sizes shorter than 64 bits, use (do not
11359 use) zero-extending load instructions by default, rather than
11360 sign-extending ones.
11363 @itemx -mno-knuthdiv
11365 @opindex mno-knuthdiv
11366 Make the result of a division yielding a remainder have the same sign as
11367 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
11368 remainder follows the sign of the dividend. Both methods are
11369 arithmetically valid, the latter being almost exclusively used.
11371 @item -mtoplevel-symbols
11372 @itemx -mno-toplevel-symbols
11373 @opindex mtoplevel-symbols
11374 @opindex mno-toplevel-symbols
11375 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
11376 code can be used with the @code{PREFIX} assembly directive.
11380 Generate an executable in the ELF format, rather than the default
11381 @samp{mmo} format used by the @command{mmix} simulator.
11383 @item -mbranch-predict
11384 @itemx -mno-branch-predict
11385 @opindex mbranch-predict
11386 @opindex mno-branch-predict
11387 Use (do not use) the probable-branch instructions, when static branch
11388 prediction indicates a probable branch.
11390 @item -mbase-addresses
11391 @itemx -mno-base-addresses
11392 @opindex mbase-addresses
11393 @opindex mno-base-addresses
11394 Generate (do not generate) code that uses @emph{base addresses}. Using a
11395 base address automatically generates a request (handled by the assembler
11396 and the linker) for a constant to be set up in a global register. The
11397 register is used for one or more base address requests within the range 0
11398 to 255 from the value held in the register. The generally leads to short
11399 and fast code, but the number of different data items that can be
11400 addressed is limited. This means that a program that uses lots of static
11401 data may require @option{-mno-base-addresses}.
11403 @item -msingle-exit
11404 @itemx -mno-single-exit
11405 @opindex msingle-exit
11406 @opindex mno-single-exit
11407 Force (do not force) generated code to have a single exit point in each
11411 @node MN10300 Options
11412 @subsection MN10300 Options
11413 @cindex MN10300 options
11415 These @option{-m} options are defined for Matsushita MN10300 architectures:
11420 Generate code to avoid bugs in the multiply instructions for the MN10300
11421 processors. This is the default.
11423 @item -mno-mult-bug
11424 @opindex mno-mult-bug
11425 Do not generate code to avoid bugs in the multiply instructions for the
11426 MN10300 processors.
11430 Generate code which uses features specific to the AM33 processor.
11434 Do not generate code which uses features specific to the AM33 processor. This
11437 @item -mreturn-pointer-on-d0
11438 @opindex mreturn-pointer-on-d0
11439 When generating a function which returns a pointer, return the pointer
11440 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
11441 only in a0, and attempts to call such functions without a prototype
11442 would result in errors. Note that this option is on by default; use
11443 @option{-mno-return-pointer-on-d0} to disable it.
11447 Do not link in the C run-time initialization object file.
11451 Indicate to the linker that it should perform a relaxation optimization pass
11452 to shorten branches, calls and absolute memory addresses. This option only
11453 has an effect when used on the command line for the final link step.
11455 This option makes symbolic debugging impossible.
11459 @subsection MT Options
11462 These @option{-m} options are defined for Morpho MT architectures:
11466 @item -march=@var{cpu-type}
11468 Generate code that will run on @var{cpu-type}, which is the name of a system
11469 representing a certain processor type. Possible values for
11470 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
11471 @samp{ms1-16-003} and @samp{ms2}.
11473 When this option is not used, the default is @option{-march=ms1-16-002}.
11477 Use byte loads and stores when generating code.
11481 Do not use byte loads and stores when generating code.
11485 Use simulator runtime
11489 Do not link in the C run-time initialization object file
11490 @file{crti.o}. Other run-time initialization and termination files
11491 such as @file{startup.o} and @file{exit.o} are still included on the
11492 linker command line.
11496 @node PDP-11 Options
11497 @subsection PDP-11 Options
11498 @cindex PDP-11 Options
11500 These options are defined for the PDP-11:
11505 Use hardware FPP floating point. This is the default. (FIS floating
11506 point on the PDP-11/40 is not supported.)
11509 @opindex msoft-float
11510 Do not use hardware floating point.
11514 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
11518 Return floating-point results in memory. This is the default.
11522 Generate code for a PDP-11/40.
11526 Generate code for a PDP-11/45. This is the default.
11530 Generate code for a PDP-11/10.
11532 @item -mbcopy-builtin
11533 @opindex bcopy-builtin
11534 Use inline @code{movmemhi} patterns for copying memory. This is the
11539 Do not use inline @code{movmemhi} patterns for copying memory.
11545 Use 16-bit @code{int}. This is the default.
11551 Use 32-bit @code{int}.
11554 @itemx -mno-float32
11556 @opindex mno-float32
11557 Use 64-bit @code{float}. This is the default.
11560 @itemx -mno-float64
11562 @opindex mno-float64
11563 Use 32-bit @code{float}.
11567 Use @code{abshi2} pattern. This is the default.
11571 Do not use @code{abshi2} pattern.
11573 @item -mbranch-expensive
11574 @opindex mbranch-expensive
11575 Pretend that branches are expensive. This is for experimenting with
11576 code generation only.
11578 @item -mbranch-cheap
11579 @opindex mbranch-cheap
11580 Do not pretend that branches are expensive. This is the default.
11584 Generate code for a system with split I&D@.
11588 Generate code for a system without split I&D@. This is the default.
11592 Use Unix assembler syntax. This is the default when configured for
11593 @samp{pdp11-*-bsd}.
11597 Use DEC assembler syntax. This is the default when configured for any
11598 PDP-11 target other than @samp{pdp11-*-bsd}.
11601 @node PowerPC Options
11602 @subsection PowerPC Options
11603 @cindex PowerPC options
11605 These are listed under @xref{RS/6000 and PowerPC Options}.
11607 @node RS/6000 and PowerPC Options
11608 @subsection IBM RS/6000 and PowerPC Options
11609 @cindex RS/6000 and PowerPC Options
11610 @cindex IBM RS/6000 and PowerPC Options
11612 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
11619 @itemx -mno-powerpc
11620 @itemx -mpowerpc-gpopt
11621 @itemx -mno-powerpc-gpopt
11622 @itemx -mpowerpc-gfxopt
11623 @itemx -mno-powerpc-gfxopt
11625 @itemx -mno-powerpc64
11629 @itemx -mno-popcntb
11637 @opindex mno-power2
11639 @opindex mno-powerpc
11640 @opindex mpowerpc-gpopt
11641 @opindex mno-powerpc-gpopt
11642 @opindex mpowerpc-gfxopt
11643 @opindex mno-powerpc-gfxopt
11644 @opindex mpowerpc64
11645 @opindex mno-powerpc64
11649 @opindex mno-popcntb
11653 @opindex mno-mfpgpr
11654 GCC supports two related instruction set architectures for the
11655 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
11656 instructions supported by the @samp{rios} chip set used in the original
11657 RS/6000 systems and the @dfn{PowerPC} instruction set is the
11658 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
11659 the IBM 4xx, 6xx, and follow-on microprocessors.
11661 Neither architecture is a subset of the other. However there is a
11662 large common subset of instructions supported by both. An MQ
11663 register is included in processors supporting the POWER architecture.
11665 You use these options to specify which instructions are available on the
11666 processor you are using. The default value of these options is
11667 determined when configuring GCC@. Specifying the
11668 @option{-mcpu=@var{cpu_type}} overrides the specification of these
11669 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
11670 rather than the options listed above.
11672 The @option{-mpower} option allows GCC to generate instructions that
11673 are found only in the POWER architecture and to use the MQ register.
11674 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
11675 to generate instructions that are present in the POWER2 architecture but
11676 not the original POWER architecture.
11678 The @option{-mpowerpc} option allows GCC to generate instructions that
11679 are found only in the 32-bit subset of the PowerPC architecture.
11680 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
11681 GCC to use the optional PowerPC architecture instructions in the
11682 General Purpose group, including floating-point square root. Specifying
11683 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
11684 use the optional PowerPC architecture instructions in the Graphics
11685 group, including floating-point select.
11687 The @option{-mmfcrf} option allows GCC to generate the move from
11688 condition register field instruction implemented on the POWER4
11689 processor and other processors that support the PowerPC V2.01
11691 The @option{-mpopcntb} option allows GCC to generate the popcount and
11692 double precision FP reciprocal estimate instruction implemented on the
11693 POWER5 processor and other processors that support the PowerPC V2.02
11695 The @option{-mfprnd} option allows GCC to generate the FP round to
11696 integer instructions implemented on the POWER5+ processor and other
11697 processors that support the PowerPC V2.03 architecture.
11698 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
11699 general purpose register instructions implemented on the POWER6X
11700 processor and other processors that support the extended PowerPC V2.05
11703 The @option{-mpowerpc64} option allows GCC to generate the additional
11704 64-bit instructions that are found in the full PowerPC64 architecture
11705 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
11706 @option{-mno-powerpc64}.
11708 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
11709 will use only the instructions in the common subset of both
11710 architectures plus some special AIX common-mode calls, and will not use
11711 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
11712 permits GCC to use any instruction from either architecture and to
11713 allow use of the MQ register; specify this for the Motorola MPC601.
11715 @item -mnew-mnemonics
11716 @itemx -mold-mnemonics
11717 @opindex mnew-mnemonics
11718 @opindex mold-mnemonics
11719 Select which mnemonics to use in the generated assembler code. With
11720 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
11721 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
11722 assembler mnemonics defined for the POWER architecture. Instructions
11723 defined in only one architecture have only one mnemonic; GCC uses that
11724 mnemonic irrespective of which of these options is specified.
11726 GCC defaults to the mnemonics appropriate for the architecture in
11727 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
11728 value of these option. Unless you are building a cross-compiler, you
11729 should normally not specify either @option{-mnew-mnemonics} or
11730 @option{-mold-mnemonics}, but should instead accept the default.
11732 @item -mcpu=@var{cpu_type}
11734 Set architecture type, register usage, choice of mnemonics, and
11735 instruction scheduling parameters for machine type @var{cpu_type}.
11736 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
11737 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
11738 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
11739 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
11740 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
11741 @samp{860}, @samp{970}, @samp{8540}, @samp{ec603e}, @samp{G3},
11742 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
11743 @samp{power4}, @samp{power5}, @samp{power5+}, @samp{power6},
11744 @samp{power6x}, @samp{common}, @samp{powerpc}, @samp{powerpc64},
11745 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
11747 @option{-mcpu=common} selects a completely generic processor. Code
11748 generated under this option will run on any POWER or PowerPC processor.
11749 GCC will use only the instructions in the common subset of both
11750 architectures, and will not use the MQ register. GCC assumes a generic
11751 processor model for scheduling purposes.
11753 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
11754 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
11755 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
11756 types, with an appropriate, generic processor model assumed for
11757 scheduling purposes.
11759 The other options specify a specific processor. Code generated under
11760 those options will run best on that processor, and may not run at all on
11763 The @option{-mcpu} options automatically enable or disable the
11766 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
11767 -mnew-mnemonics -mpopcntb -mpower -mpower2 -mpowerpc64 @gol
11768 -mpowerpc-gpopt -mpowerpc-gfxopt -mstring -mmulhw -mdlmzb -mmfpgpr}
11770 The particular options set for any particular CPU will vary between
11771 compiler versions, depending on what setting seems to produce optimal
11772 code for that CPU; it doesn't necessarily reflect the actual hardware's
11773 capabilities. If you wish to set an individual option to a particular
11774 value, you may specify it after the @option{-mcpu} option, like
11775 @samp{-mcpu=970 -mno-altivec}.
11777 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
11778 not enabled or disabled by the @option{-mcpu} option at present because
11779 AIX does not have full support for these options. You may still
11780 enable or disable them individually if you're sure it'll work in your
11783 @item -mtune=@var{cpu_type}
11785 Set the instruction scheduling parameters for machine type
11786 @var{cpu_type}, but do not set the architecture type, register usage, or
11787 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
11788 values for @var{cpu_type} are used for @option{-mtune} as for
11789 @option{-mcpu}. If both are specified, the code generated will use the
11790 architecture, registers, and mnemonics set by @option{-mcpu}, but the
11791 scheduling parameters set by @option{-mtune}.
11797 Generate code to compute division as reciprocal estimate and iterative
11798 refinement, creating opportunities for increased throughput. This
11799 feature requires: optional PowerPC Graphics instruction set for single
11800 precision and FRE instruction for double precision, assuming divides
11801 cannot generate user-visible traps, and the domain values not include
11802 Infinities, denormals or zero denominator.
11805 @itemx -mno-altivec
11807 @opindex mno-altivec
11808 Generate code that uses (does not use) AltiVec instructions, and also
11809 enable the use of built-in functions that allow more direct access to
11810 the AltiVec instruction set. You may also need to set
11811 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
11817 @opindex mno-vrsave
11818 Generate VRSAVE instructions when generating AltiVec code.
11821 @opindex msecure-plt
11822 Generate code that allows ld and ld.so to build executables and shared
11823 libraries with non-exec .plt and .got sections. This is a PowerPC
11824 32-bit SYSV ABI option.
11828 Generate code that uses a BSS .plt section that ld.so fills in, and
11829 requires .plt and .got sections that are both writable and executable.
11830 This is a PowerPC 32-bit SYSV ABI option.
11836 This switch enables or disables the generation of ISEL instructions.
11838 @item -misel=@var{yes/no}
11839 This switch has been deprecated. Use @option{-misel} and
11840 @option{-mno-isel} instead.
11846 This switch enables or disables the generation of SPE simd
11849 @item -mspe=@var{yes/no}
11850 This option has been deprecated. Use @option{-mspe} and
11851 @option{-mno-spe} instead.
11853 @item -mfloat-gprs=@var{yes/single/double/no}
11854 @itemx -mfloat-gprs
11855 @opindex mfloat-gprs
11856 This switch enables or disables the generation of floating point
11857 operations on the general purpose registers for architectures that
11860 The argument @var{yes} or @var{single} enables the use of
11861 single-precision floating point operations.
11863 The argument @var{double} enables the use of single and
11864 double-precision floating point operations.
11866 The argument @var{no} disables floating point operations on the
11867 general purpose registers.
11869 This option is currently only available on the MPC854x.
11875 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
11876 targets (including GNU/Linux). The 32-bit environment sets int, long
11877 and pointer to 32 bits and generates code that runs on any PowerPC
11878 variant. The 64-bit environment sets int to 32 bits and long and
11879 pointer to 64 bits, and generates code for PowerPC64, as for
11880 @option{-mpowerpc64}.
11883 @itemx -mno-fp-in-toc
11884 @itemx -mno-sum-in-toc
11885 @itemx -mminimal-toc
11887 @opindex mno-fp-in-toc
11888 @opindex mno-sum-in-toc
11889 @opindex mminimal-toc
11890 Modify generation of the TOC (Table Of Contents), which is created for
11891 every executable file. The @option{-mfull-toc} option is selected by
11892 default. In that case, GCC will allocate at least one TOC entry for
11893 each unique non-automatic variable reference in your program. GCC
11894 will also place floating-point constants in the TOC@. However, only
11895 16,384 entries are available in the TOC@.
11897 If you receive a linker error message that saying you have overflowed
11898 the available TOC space, you can reduce the amount of TOC space used
11899 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
11900 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
11901 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
11902 generate code to calculate the sum of an address and a constant at
11903 run-time instead of putting that sum into the TOC@. You may specify one
11904 or both of these options. Each causes GCC to produce very slightly
11905 slower and larger code at the expense of conserving TOC space.
11907 If you still run out of space in the TOC even when you specify both of
11908 these options, specify @option{-mminimal-toc} instead. This option causes
11909 GCC to make only one TOC entry for every file. When you specify this
11910 option, GCC will produce code that is slower and larger but which
11911 uses extremely little TOC space. You may wish to use this option
11912 only on files that contain less frequently executed code.
11918 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
11919 @code{long} type, and the infrastructure needed to support them.
11920 Specifying @option{-maix64} implies @option{-mpowerpc64} and
11921 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
11922 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
11925 @itemx -mno-xl-compat
11926 @opindex mxl-compat
11927 @opindex mno-xl-compat
11928 Produce code that conforms more closely to IBM XL compiler semantics
11929 when using AIX-compatible ABI. Pass floating-point arguments to
11930 prototyped functions beyond the register save area (RSA) on the stack
11931 in addition to argument FPRs. Do not assume that most significant
11932 double in 128-bit long double value is properly rounded when comparing
11933 values and converting to double. Use XL symbol names for long double
11936 The AIX calling convention was extended but not initially documented to
11937 handle an obscure K&R C case of calling a function that takes the
11938 address of its arguments with fewer arguments than declared. IBM XL
11939 compilers access floating point arguments which do not fit in the
11940 RSA from the stack when a subroutine is compiled without
11941 optimization. Because always storing floating-point arguments on the
11942 stack is inefficient and rarely needed, this option is not enabled by
11943 default and only is necessary when calling subroutines compiled by IBM
11944 XL compilers without optimization.
11948 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
11949 application written to use message passing with special startup code to
11950 enable the application to run. The system must have PE installed in the
11951 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
11952 must be overridden with the @option{-specs=} option to specify the
11953 appropriate directory location. The Parallel Environment does not
11954 support threads, so the @option{-mpe} option and the @option{-pthread}
11955 option are incompatible.
11957 @item -malign-natural
11958 @itemx -malign-power
11959 @opindex malign-natural
11960 @opindex malign-power
11961 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
11962 @option{-malign-natural} overrides the ABI-defined alignment of larger
11963 types, such as floating-point doubles, on their natural size-based boundary.
11964 The option @option{-malign-power} instructs GCC to follow the ABI-specified
11965 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
11967 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
11971 @itemx -mhard-float
11972 @opindex msoft-float
11973 @opindex mhard-float
11974 Generate code that does not use (uses) the floating-point register set.
11975 Software floating point emulation is provided if you use the
11976 @option{-msoft-float} option, and pass the option to GCC when linking.
11979 @itemx -mno-multiple
11981 @opindex mno-multiple
11982 Generate code that uses (does not use) the load multiple word
11983 instructions and the store multiple word instructions. These
11984 instructions are generated by default on POWER systems, and not
11985 generated on PowerPC systems. Do not use @option{-mmultiple} on little
11986 endian PowerPC systems, since those instructions do not work when the
11987 processor is in little endian mode. The exceptions are PPC740 and
11988 PPC750 which permit the instructions usage in little endian mode.
11993 @opindex mno-string
11994 Generate code that uses (does not use) the load string instructions
11995 and the store string word instructions to save multiple registers and
11996 do small block moves. These instructions are generated by default on
11997 POWER systems, and not generated on PowerPC systems. Do not use
11998 @option{-mstring} on little endian PowerPC systems, since those
11999 instructions do not work when the processor is in little endian mode.
12000 The exceptions are PPC740 and PPC750 which permit the instructions
12001 usage in little endian mode.
12006 @opindex mno-update
12007 Generate code that uses (does not use) the load or store instructions
12008 that update the base register to the address of the calculated memory
12009 location. These instructions are generated by default. If you use
12010 @option{-mno-update}, there is a small window between the time that the
12011 stack pointer is updated and the address of the previous frame is
12012 stored, which means code that walks the stack frame across interrupts or
12013 signals may get corrupted data.
12016 @itemx -mno-fused-madd
12017 @opindex mfused-madd
12018 @opindex mno-fused-madd
12019 Generate code that uses (does not use) the floating point multiply and
12020 accumulate instructions. These instructions are generated by default if
12021 hardware floating is used.
12027 Generate code that uses (does not use) the half-word multiply and
12028 multiply-accumulate instructions on the IBM 405 and 440 processors.
12029 These instructions are generated by default when targetting those
12036 Generate code that uses (does not use) the string-search @samp{dlmzb}
12037 instruction on the IBM 405 and 440 processors. This instruction is
12038 generated by default when targetting those processors.
12040 @item -mno-bit-align
12042 @opindex mno-bit-align
12043 @opindex mbit-align
12044 On System V.4 and embedded PowerPC systems do not (do) force structures
12045 and unions that contain bit-fields to be aligned to the base type of the
12048 For example, by default a structure containing nothing but 8
12049 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
12050 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
12051 the structure would be aligned to a 1 byte boundary and be one byte in
12054 @item -mno-strict-align
12055 @itemx -mstrict-align
12056 @opindex mno-strict-align
12057 @opindex mstrict-align
12058 On System V.4 and embedded PowerPC systems do not (do) assume that
12059 unaligned memory references will be handled by the system.
12061 @item -mrelocatable
12062 @itemx -mno-relocatable
12063 @opindex mrelocatable
12064 @opindex mno-relocatable
12065 On embedded PowerPC systems generate code that allows (does not allow)
12066 the program to be relocated to a different address at runtime. If you
12067 use @option{-mrelocatable} on any module, all objects linked together must
12068 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
12070 @item -mrelocatable-lib
12071 @itemx -mno-relocatable-lib
12072 @opindex mrelocatable-lib
12073 @opindex mno-relocatable-lib
12074 On embedded PowerPC systems generate code that allows (does not allow)
12075 the program to be relocated to a different address at runtime. Modules
12076 compiled with @option{-mrelocatable-lib} can be linked with either modules
12077 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
12078 with modules compiled with the @option{-mrelocatable} options.
12084 On System V.4 and embedded PowerPC systems do not (do) assume that
12085 register 2 contains a pointer to a global area pointing to the addresses
12086 used in the program.
12089 @itemx -mlittle-endian
12091 @opindex mlittle-endian
12092 On System V.4 and embedded PowerPC systems compile code for the
12093 processor in little endian mode. The @option{-mlittle-endian} option is
12094 the same as @option{-mlittle}.
12097 @itemx -mbig-endian
12099 @opindex mbig-endian
12100 On System V.4 and embedded PowerPC systems compile code for the
12101 processor in big endian mode. The @option{-mbig-endian} option is
12102 the same as @option{-mbig}.
12104 @item -mdynamic-no-pic
12105 @opindex mdynamic-no-pic
12106 On Darwin and Mac OS X systems, compile code so that it is not
12107 relocatable, but that its external references are relocatable. The
12108 resulting code is suitable for applications, but not shared
12111 @item -mprioritize-restricted-insns=@var{priority}
12112 @opindex mprioritize-restricted-insns
12113 This option controls the priority that is assigned to
12114 dispatch-slot restricted instructions during the second scheduling
12115 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
12116 @var{no/highest/second-highest} priority to dispatch slot restricted
12119 @item -msched-costly-dep=@var{dependence_type}
12120 @opindex msched-costly-dep
12121 This option controls which dependences are considered costly
12122 by the target during instruction scheduling. The argument
12123 @var{dependence_type} takes one of the following values:
12124 @var{no}: no dependence is costly,
12125 @var{all}: all dependences are costly,
12126 @var{true_store_to_load}: a true dependence from store to load is costly,
12127 @var{store_to_load}: any dependence from store to load is costly,
12128 @var{number}: any dependence which latency >= @var{number} is costly.
12130 @item -minsert-sched-nops=@var{scheme}
12131 @opindex minsert-sched-nops
12132 This option controls which nop insertion scheme will be used during
12133 the second scheduling pass. The argument @var{scheme} takes one of the
12135 @var{no}: Don't insert nops.
12136 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
12137 according to the scheduler's grouping.
12138 @var{regroup_exact}: Insert nops to force costly dependent insns into
12139 separate groups. Insert exactly as many nops as needed to force an insn
12140 to a new group, according to the estimated processor grouping.
12141 @var{number}: Insert nops to force costly dependent insns into
12142 separate groups. Insert @var{number} nops to force an insn to a new group.
12145 @opindex mcall-sysv
12146 On System V.4 and embedded PowerPC systems compile code using calling
12147 conventions that adheres to the March 1995 draft of the System V
12148 Application Binary Interface, PowerPC processor supplement. This is the
12149 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
12151 @item -mcall-sysv-eabi
12152 @opindex mcall-sysv-eabi
12153 Specify both @option{-mcall-sysv} and @option{-meabi} options.
12155 @item -mcall-sysv-noeabi
12156 @opindex mcall-sysv-noeabi
12157 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
12159 @item -mcall-solaris
12160 @opindex mcall-solaris
12161 On System V.4 and embedded PowerPC systems compile code for the Solaris
12165 @opindex mcall-linux
12166 On System V.4 and embedded PowerPC systems compile code for the
12167 Linux-based GNU system.
12171 On System V.4 and embedded PowerPC systems compile code for the
12172 Hurd-based GNU system.
12174 @item -mcall-netbsd
12175 @opindex mcall-netbsd
12176 On System V.4 and embedded PowerPC systems compile code for the
12177 NetBSD operating system.
12179 @item -maix-struct-return
12180 @opindex maix-struct-return
12181 Return all structures in memory (as specified by the AIX ABI)@.
12183 @item -msvr4-struct-return
12184 @opindex msvr4-struct-return
12185 Return structures smaller than 8 bytes in registers (as specified by the
12188 @item -mabi=@var{abi-type}
12190 Extend the current ABI with a particular extension, or remove such extension.
12191 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
12192 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
12196 Extend the current ABI with SPE ABI extensions. This does not change
12197 the default ABI, instead it adds the SPE ABI extensions to the current
12201 @opindex mabi=no-spe
12202 Disable Booke SPE ABI extensions for the current ABI@.
12204 @item -mabi=ibmlongdouble
12205 @opindex mabi=ibmlongdouble
12206 Change the current ABI to use IBM extended precision long double.
12207 This is a PowerPC 32-bit SYSV ABI option.
12209 @item -mabi=ieeelongdouble
12210 @opindex mabi=ieeelongdouble
12211 Change the current ABI to use IEEE extended precision long double.
12212 This is a PowerPC 32-bit Linux ABI option.
12215 @itemx -mno-prototype
12216 @opindex mprototype
12217 @opindex mno-prototype
12218 On System V.4 and embedded PowerPC systems assume that all calls to
12219 variable argument functions are properly prototyped. Otherwise, the
12220 compiler must insert an instruction before every non prototyped call to
12221 set or clear bit 6 of the condition code register (@var{CR}) to
12222 indicate whether floating point values were passed in the floating point
12223 registers in case the function takes a variable arguments. With
12224 @option{-mprototype}, only calls to prototyped variable argument functions
12225 will set or clear the bit.
12229 On embedded PowerPC systems, assume that the startup module is called
12230 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
12231 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
12236 On embedded PowerPC systems, assume that the startup module is called
12237 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
12242 On embedded PowerPC systems, assume that the startup module is called
12243 @file{crt0.o} and the standard C libraries are @file{libads.a} and
12246 @item -myellowknife
12247 @opindex myellowknife
12248 On embedded PowerPC systems, assume that the startup module is called
12249 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
12254 On System V.4 and embedded PowerPC systems, specify that you are
12255 compiling for a VxWorks system.
12259 Specify that you are compiling for the WindISS simulation environment.
12263 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
12264 header to indicate that @samp{eabi} extended relocations are used.
12270 On System V.4 and embedded PowerPC systems do (do not) adhere to the
12271 Embedded Applications Binary Interface (eabi) which is a set of
12272 modifications to the System V.4 specifications. Selecting @option{-meabi}
12273 means that the stack is aligned to an 8 byte boundary, a function
12274 @code{__eabi} is called to from @code{main} to set up the eabi
12275 environment, and the @option{-msdata} option can use both @code{r2} and
12276 @code{r13} to point to two separate small data areas. Selecting
12277 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
12278 do not call an initialization function from @code{main}, and the
12279 @option{-msdata} option will only use @code{r13} to point to a single
12280 small data area. The @option{-meabi} option is on by default if you
12281 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
12284 @opindex msdata=eabi
12285 On System V.4 and embedded PowerPC systems, put small initialized
12286 @code{const} global and static data in the @samp{.sdata2} section, which
12287 is pointed to by register @code{r2}. Put small initialized
12288 non-@code{const} global and static data in the @samp{.sdata} section,
12289 which is pointed to by register @code{r13}. Put small uninitialized
12290 global and static data in the @samp{.sbss} section, which is adjacent to
12291 the @samp{.sdata} section. The @option{-msdata=eabi} option is
12292 incompatible with the @option{-mrelocatable} option. The
12293 @option{-msdata=eabi} option also sets the @option{-memb} option.
12296 @opindex msdata=sysv
12297 On System V.4 and embedded PowerPC systems, put small global and static
12298 data in the @samp{.sdata} section, which is pointed to by register
12299 @code{r13}. Put small uninitialized global and static data in the
12300 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
12301 The @option{-msdata=sysv} option is incompatible with the
12302 @option{-mrelocatable} option.
12304 @item -msdata=default
12306 @opindex msdata=default
12308 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
12309 compile code the same as @option{-msdata=eabi}, otherwise compile code the
12310 same as @option{-msdata=sysv}.
12313 @opindex msdata-data
12314 On System V.4 and embedded PowerPC systems, put small global
12315 data in the @samp{.sdata} section. Put small uninitialized global
12316 data in the @samp{.sbss} section. Do not use register @code{r13}
12317 to address small data however. This is the default behavior unless
12318 other @option{-msdata} options are used.
12322 @opindex msdata=none
12324 On embedded PowerPC systems, put all initialized global and static data
12325 in the @samp{.data} section, and all uninitialized data in the
12326 @samp{.bss} section.
12330 @cindex smaller data references (PowerPC)
12331 @cindex .sdata/.sdata2 references (PowerPC)
12332 On embedded PowerPC systems, put global and static items less than or
12333 equal to @var{num} bytes into the small data or bss sections instead of
12334 the normal data or bss section. By default, @var{num} is 8. The
12335 @option{-G @var{num}} switch is also passed to the linker.
12336 All modules should be compiled with the same @option{-G @var{num}} value.
12339 @itemx -mno-regnames
12341 @opindex mno-regnames
12342 On System V.4 and embedded PowerPC systems do (do not) emit register
12343 names in the assembly language output using symbolic forms.
12346 @itemx -mno-longcall
12348 @opindex mno-longcall
12349 By default assume that all calls are far away so that a longer more
12350 expensive calling sequence is required. This is required for calls
12351 further than 32 megabytes (33,554,432 bytes) from the current location.
12352 A short call will be generated if the compiler knows
12353 the call cannot be that far away. This setting can be overridden by
12354 the @code{shortcall} function attribute, or by @code{#pragma
12357 Some linkers are capable of detecting out-of-range calls and generating
12358 glue code on the fly. On these systems, long calls are unnecessary and
12359 generate slower code. As of this writing, the AIX linker can do this,
12360 as can the GNU linker for PowerPC/64. It is planned to add this feature
12361 to the GNU linker for 32-bit PowerPC systems as well.
12363 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
12364 callee, L42'', plus a ``branch island'' (glue code). The two target
12365 addresses represent the callee and the ``branch island''. The
12366 Darwin/PPC linker will prefer the first address and generate a ``bl
12367 callee'' if the PPC ``bl'' instruction will reach the callee directly;
12368 otherwise, the linker will generate ``bl L42'' to call the ``branch
12369 island''. The ``branch island'' is appended to the body of the
12370 calling function; it computes the full 32-bit address of the callee
12373 On Mach-O (Darwin) systems, this option directs the compiler emit to
12374 the glue for every direct call, and the Darwin linker decides whether
12375 to use or discard it.
12377 In the future, we may cause GCC to ignore all longcall specifications
12378 when the linker is known to generate glue.
12382 Adds support for multithreading with the @dfn{pthreads} library.
12383 This option sets flags for both the preprocessor and linker.
12387 @node S/390 and zSeries Options
12388 @subsection S/390 and zSeries Options
12389 @cindex S/390 and zSeries Options
12391 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
12395 @itemx -msoft-float
12396 @opindex mhard-float
12397 @opindex msoft-float
12398 Use (do not use) the hardware floating-point instructions and registers
12399 for floating-point operations. When @option{-msoft-float} is specified,
12400 functions in @file{libgcc.a} will be used to perform floating-point
12401 operations. When @option{-mhard-float} is specified, the compiler
12402 generates IEEE floating-point instructions. This is the default.
12404 @item -mlong-double-64
12405 @itemx -mlong-double-128
12406 @opindex mlong-double-64
12407 @opindex mlong-double-128
12408 These switches control the size of @code{long double} type. A size
12409 of 64bit makes the @code{long double} type equivalent to the @code{double}
12410 type. This is the default.
12413 @itemx -mno-backchain
12414 @opindex mbackchain
12415 @opindex mno-backchain
12416 Store (do not store) the address of the caller's frame as backchain pointer
12417 into the callee's stack frame.
12418 A backchain may be needed to allow debugging using tools that do not understand
12419 DWARF-2 call frame information.
12420 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
12421 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
12422 the backchain is placed into the topmost word of the 96/160 byte register
12425 In general, code compiled with @option{-mbackchain} is call-compatible with
12426 code compiled with @option{-mmo-backchain}; however, use of the backchain
12427 for debugging purposes usually requires that the whole binary is built with
12428 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
12429 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
12430 to build a linux kernel use @option{-msoft-float}.
12432 The default is to not maintain the backchain.
12434 @item -mpacked-stack
12435 @item -mno-packed-stack
12436 @opindex mpacked-stack
12437 @opindex mno-packed-stack
12438 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
12439 specified, the compiler uses the all fields of the 96/160 byte register save
12440 area only for their default purpose; unused fields still take up stack space.
12441 When @option{-mpacked-stack} is specified, register save slots are densely
12442 packed at the top of the register save area; unused space is reused for other
12443 purposes, allowing for more efficient use of the available stack space.
12444 However, when @option{-mbackchain} is also in effect, the topmost word of
12445 the save area is always used to store the backchain, and the return address
12446 register is always saved two words below the backchain.
12448 As long as the stack frame backchain is not used, code generated with
12449 @option{-mpacked-stack} is call-compatible with code generated with
12450 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
12451 S/390 or zSeries generated code that uses the stack frame backchain at run
12452 time, not just for debugging purposes. Such code is not call-compatible
12453 with code compiled with @option{-mpacked-stack}. Also, note that the
12454 combination of @option{-mbackchain},
12455 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
12456 to build a linux kernel use @option{-msoft-float}.
12458 The default is to not use the packed stack layout.
12461 @itemx -mno-small-exec
12462 @opindex msmall-exec
12463 @opindex mno-small-exec
12464 Generate (or do not generate) code using the @code{bras} instruction
12465 to do subroutine calls.
12466 This only works reliably if the total executable size does not
12467 exceed 64k. The default is to use the @code{basr} instruction instead,
12468 which does not have this limitation.
12474 When @option{-m31} is specified, generate code compliant to the
12475 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
12476 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
12477 particular to generate 64-bit instructions. For the @samp{s390}
12478 targets, the default is @option{-m31}, while the @samp{s390x}
12479 targets default to @option{-m64}.
12485 When @option{-mzarch} is specified, generate code using the
12486 instructions available on z/Architecture.
12487 When @option{-mesa} is specified, generate code using the
12488 instructions available on ESA/390. Note that @option{-mesa} is
12489 not possible with @option{-m64}.
12490 When generating code compliant to the GNU/Linux for S/390 ABI,
12491 the default is @option{-mesa}. When generating code compliant
12492 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
12498 Generate (or do not generate) code using the @code{mvcle} instruction
12499 to perform block moves. When @option{-mno-mvcle} is specified,
12500 use a @code{mvc} loop instead. This is the default unless optimizing for
12507 Print (or do not print) additional debug information when compiling.
12508 The default is to not print debug information.
12510 @item -march=@var{cpu-type}
12512 Generate code that will run on @var{cpu-type}, which is the name of a system
12513 representing a certain processor type. Possible values for
12514 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
12515 When generating code using the instructions available on z/Architecture,
12516 the default is @option{-march=z900}. Otherwise, the default is
12517 @option{-march=g5}.
12519 @item -mtune=@var{cpu-type}
12521 Tune to @var{cpu-type} everything applicable about the generated code,
12522 except for the ABI and the set of available instructions.
12523 The list of @var{cpu-type} values is the same as for @option{-march}.
12524 The default is the value used for @option{-march}.
12527 @itemx -mno-tpf-trace
12528 @opindex mtpf-trace
12529 @opindex mno-tpf-trace
12530 Generate code that adds (does not add) in TPF OS specific branches to trace
12531 routines in the operating system. This option is off by default, even
12532 when compiling for the TPF OS@.
12535 @itemx -mno-fused-madd
12536 @opindex mfused-madd
12537 @opindex mno-fused-madd
12538 Generate code that uses (does not use) the floating point multiply and
12539 accumulate instructions. These instructions are generated by default if
12540 hardware floating point is used.
12542 @item -mwarn-framesize=@var{framesize}
12543 @opindex mwarn-framesize
12544 Emit a warning if the current function exceeds the given frame size. Because
12545 this is a compile time check it doesn't need to be a real problem when the program
12546 runs. It is intended to identify functions which most probably cause
12547 a stack overflow. It is useful to be used in an environment with limited stack
12548 size e.g.@: the linux kernel.
12550 @item -mwarn-dynamicstack
12551 @opindex mwarn-dynamicstack
12552 Emit a warning if the function calls alloca or uses dynamically
12553 sized arrays. This is generally a bad idea with a limited stack size.
12555 @item -mstack-guard=@var{stack-guard}
12556 @item -mstack-size=@var{stack-size}
12557 @opindex mstack-guard
12558 @opindex mstack-size
12559 These arguments always have to be used in conjunction. If they are present the s390
12560 back end emits additional instructions in the function prologue which trigger a trap
12561 if the stack size is @var{stack-guard} bytes above the @var{stack-size}
12562 (remember that the stack on s390 grows downward). These options are intended to
12563 be used to help debugging stack overflow problems. The additionally emitted code
12564 causes only little overhead and hence can also be used in production like systems
12565 without greater performance degradation. The given values have to be exact
12566 powers of 2 and @var{stack-size} has to be greater than @var{stack-guard} without
12568 In order to be efficient the extra code makes the assumption that the stack starts
12569 at an address aligned to the value given by @var{stack-size}.
12572 @node Score Options
12573 @subsection Score Options
12574 @cindex Score Options
12576 These options are defined for Score implementations:
12581 Compile code for little endian mode.
12585 Compile code for big endian mode. This is the default.
12589 Enable the use of multiply-accumulate instructions. Disabled by default.
12593 Specify the SCORE5U of the target architecture.
12597 Specify the SCORE7 of the target architecture. This is the default.
12601 @subsection SH Options
12603 These @samp{-m} options are defined for the SH implementations:
12608 Generate code for the SH1.
12612 Generate code for the SH2.
12615 Generate code for the SH2e.
12619 Generate code for the SH3.
12623 Generate code for the SH3e.
12627 Generate code for the SH4 without a floating-point unit.
12629 @item -m4-single-only
12630 @opindex m4-single-only
12631 Generate code for the SH4 with a floating-point unit that only
12632 supports single-precision arithmetic.
12636 Generate code for the SH4 assuming the floating-point unit is in
12637 single-precision mode by default.
12641 Generate code for the SH4.
12645 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
12646 floating-point unit is not used.
12648 @item -m4a-single-only
12649 @opindex m4a-single-only
12650 Generate code for the SH4a, in such a way that no double-precision
12651 floating point operations are used.
12654 @opindex m4a-single
12655 Generate code for the SH4a assuming the floating-point unit is in
12656 single-precision mode by default.
12660 Generate code for the SH4a.
12664 Same as @option{-m4a-nofpu}, except that it implicitly passes
12665 @option{-dsp} to the assembler. GCC doesn't generate any DSP
12666 instructions at the moment.
12670 Compile code for the processor in big endian mode.
12674 Compile code for the processor in little endian mode.
12678 Align doubles at 64-bit boundaries. Note that this changes the calling
12679 conventions, and thus some functions from the standard C library will
12680 not work unless you recompile it first with @option{-mdalign}.
12684 Shorten some address references at link time, when possible; uses the
12685 linker option @option{-relax}.
12689 Use 32-bit offsets in @code{switch} tables. The default is to use
12694 Enable the use of the instruction @code{fmovd}.
12698 Comply with the calling conventions defined by Renesas.
12702 Comply with the calling conventions defined by Renesas.
12706 Comply with the calling conventions defined for GCC before the Renesas
12707 conventions were available. This option is the default for all
12708 targets of the SH toolchain except for @samp{sh-symbianelf}.
12711 @opindex mnomacsave
12712 Mark the @code{MAC} register as call-clobbered, even if
12713 @option{-mhitachi} is given.
12717 Increase IEEE-compliance of floating-point code.
12718 At the moment, this is equivalent to @option{-fno-finite-math-only}.
12719 When generating 16 bit SH opcodes, getting IEEE-conforming results for
12720 comparisons of NANs / infinities incurs extra overhead in every
12721 floating point comparison, therefore the default is set to
12722 @option{-ffinite-math-only}.
12724 @item -minline-ic_invalidate
12725 @opindex minline-ic_invalidate
12726 Inline code to invalidate instruction cache entries after setting up
12727 nested function trampolines.
12728 This option has no effect if -musermode is in effect and the selected
12729 code generation option (e.g. -m4) does not allow the use of the icbi
12731 If the selected code generation option does not allow the use of the icbi
12732 instruction, and -musermode is not in effect, the inlined code will
12733 manipulate the instruction cache address array directly with an associative
12734 write. This not only requires privileged mode, but it will also
12735 fail if the cache line had been mapped via the TLB and has become unmapped.
12739 Dump instruction size and location in the assembly code.
12742 @opindex mpadstruct
12743 This option is deprecated. It pads structures to multiple of 4 bytes,
12744 which is incompatible with the SH ABI@.
12748 Optimize for space instead of speed. Implied by @option{-Os}.
12751 @opindex mprefergot
12752 When generating position-independent code, emit function calls using
12753 the Global Offset Table instead of the Procedure Linkage Table.
12757 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
12758 if the inlined code would not work in user mode.
12759 This is the default when the target is @code{sh-*-linux*}.
12761 @item -multcost=@var{number}
12762 @opindex multcost=@var{number}
12763 Set the cost to assume for a multiply insn.
12765 @item -mdiv=@var{strategy}
12766 @opindex mdiv=@var{strategy}
12767 Set the division strategy to use for SHmedia code. @var{strategy} must be
12768 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
12769 inv:call2, inv:fp .
12770 "fp" performs the operation in floating point. This has a very high latency,
12771 but needs only a few instructions, so it might be a good choice if
12772 your code has enough easily exploitable ILP to allow the compiler to
12773 schedule the floating point instructions together with other instructions.
12774 Division by zero causes a floating point exception.
12775 "inv" uses integer operations to calculate the inverse of the divisor,
12776 and then multiplies the dividend with the inverse. This strategy allows
12777 cse and hoisting of the inverse calculation. Division by zero calculates
12778 an unspecified result, but does not trap.
12779 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
12780 have been found, or if the entire operation has been hoisted to the same
12781 place, the last stages of the inverse calculation are intertwined with the
12782 final multiply to reduce the overall latency, at the expense of using a few
12783 more instructions, and thus offering fewer scheduling opportunities with
12785 "call" calls a library function that usually implements the inv:minlat
12787 This gives high code density for m5-*media-nofpu compilations.
12788 "call2" uses a different entry point of the same library function, where it
12789 assumes that a pointer to a lookup table has already been set up, which
12790 exposes the pointer load to cse / code hoisting optimizations.
12791 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
12792 code generation, but if the code stays unoptimized, revert to the "call",
12793 "call2", or "fp" strategies, respectively. Note that the
12794 potentially-trapping side effect of division by zero is carried by a
12795 separate instruction, so it is possible that all the integer instructions
12796 are hoisted out, but the marker for the side effect stays where it is.
12797 A recombination to fp operations or a call is not possible in that case.
12798 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
12799 that the inverse calculation was nor separated from the multiply, they speed
12800 up division where the dividend fits into 20 bits (plus sign where applicable),
12801 by inserting a test to skip a number of operations in this case; this test
12802 slows down the case of larger dividends. inv20u assumes the case of a such
12803 a small dividend to be unlikely, and inv20l assumes it to be likely.
12805 @item -mdivsi3_libfunc=@var{name}
12806 @opindex mdivsi3_libfunc=@var{name}
12807 Set the name of the library function used for 32 bit signed division to
12808 @var{name}. This only affect the name used in the call and inv:call
12809 division strategies, and the compiler will still expect the same
12810 sets of input/output/clobbered registers as if this option was not present.
12812 @item -madjust-unroll
12813 @opindex madjust-unroll
12814 Throttle unrolling to avoid thrashing target registers.
12815 This option only has an effect if the gcc code base supports the
12816 TARGET_ADJUST_UNROLL_MAX target hook.
12818 @item -mindexed-addressing
12819 @opindex mindexed-addressing
12820 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
12821 This is only safe if the hardware and/or OS implement 32 bit wrap-around
12822 semantics for the indexed addressing mode. The architecture allows the
12823 implementation of processors with 64 bit MMU, which the OS could use to
12824 get 32 bit addressing, but since no current hardware implementation supports
12825 this or any other way to make the indexed addressing mode safe to use in
12826 the 32 bit ABI, the default is -mno-indexed-addressing.
12828 @item -mgettrcost=@var{number}
12829 @opindex mgettrcost=@var{number}
12830 Set the cost assumed for the gettr instruction to @var{number}.
12831 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
12835 Assume pt* instructions won't trap. This will generally generate better
12836 scheduled code, but is unsafe on current hardware. The current architecture
12837 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
12838 This has the unintentional effect of making it unsafe to schedule ptabs /
12839 ptrel before a branch, or hoist it out of a loop. For example,
12840 __do_global_ctors, a part of libgcc that runs constructors at program
12841 startup, calls functions in a list which is delimited by @minus{}1. With the
12842 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
12843 That means that all the constructors will be run a bit quicker, but when
12844 the loop comes to the end of the list, the program crashes because ptabs
12845 loads @minus{}1 into a target register. Since this option is unsafe for any
12846 hardware implementing the current architecture specification, the default
12847 is -mno-pt-fixed. Unless the user specifies a specific cost with
12848 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
12849 this deters register allocation using target registers for storing
12852 @item -minvalid-symbols
12853 @opindex minvalid-symbols
12854 Assume symbols might be invalid. Ordinary function symbols generated by
12855 the compiler will always be valid to load with movi/shori/ptabs or
12856 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
12857 to generate symbols that will cause ptabs / ptrel to trap.
12858 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
12859 It will then prevent cross-basic-block cse, hoisting and most scheduling
12860 of symbol loads. The default is @option{-mno-invalid-symbols}.
12863 @node SPARC Options
12864 @subsection SPARC Options
12865 @cindex SPARC options
12867 These @samp{-m} options are supported on the SPARC:
12870 @item -mno-app-regs
12872 @opindex mno-app-regs
12874 Specify @option{-mapp-regs} to generate output using the global registers
12875 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
12878 To be fully SVR4 ABI compliant at the cost of some performance loss,
12879 specify @option{-mno-app-regs}. You should compile libraries and system
12880 software with this option.
12883 @itemx -mhard-float
12885 @opindex mhard-float
12886 Generate output containing floating point instructions. This is the
12890 @itemx -msoft-float
12892 @opindex msoft-float
12893 Generate output containing library calls for floating point.
12894 @strong{Warning:} the requisite libraries are not available for all SPARC
12895 targets. Normally the facilities of the machine's usual C compiler are
12896 used, but this cannot be done directly in cross-compilation. You must make
12897 your own arrangements to provide suitable library functions for
12898 cross-compilation. The embedded targets @samp{sparc-*-aout} and
12899 @samp{sparclite-*-*} do provide software floating point support.
12901 @option{-msoft-float} changes the calling convention in the output file;
12902 therefore, it is only useful if you compile @emph{all} of a program with
12903 this option. In particular, you need to compile @file{libgcc.a}, the
12904 library that comes with GCC, with @option{-msoft-float} in order for
12907 @item -mhard-quad-float
12908 @opindex mhard-quad-float
12909 Generate output containing quad-word (long double) floating point
12912 @item -msoft-quad-float
12913 @opindex msoft-quad-float
12914 Generate output containing library calls for quad-word (long double)
12915 floating point instructions. The functions called are those specified
12916 in the SPARC ABI@. This is the default.
12918 As of this writing, there are no SPARC implementations that have hardware
12919 support for the quad-word floating point instructions. They all invoke
12920 a trap handler for one of these instructions, and then the trap handler
12921 emulates the effect of the instruction. Because of the trap handler overhead,
12922 this is much slower than calling the ABI library routines. Thus the
12923 @option{-msoft-quad-float} option is the default.
12925 @item -mno-unaligned-doubles
12926 @itemx -munaligned-doubles
12927 @opindex mno-unaligned-doubles
12928 @opindex munaligned-doubles
12929 Assume that doubles have 8 byte alignment. This is the default.
12931 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
12932 alignment only if they are contained in another type, or if they have an
12933 absolute address. Otherwise, it assumes they have 4 byte alignment.
12934 Specifying this option avoids some rare compatibility problems with code
12935 generated by other compilers. It is not the default because it results
12936 in a performance loss, especially for floating point code.
12938 @item -mno-faster-structs
12939 @itemx -mfaster-structs
12940 @opindex mno-faster-structs
12941 @opindex mfaster-structs
12942 With @option{-mfaster-structs}, the compiler assumes that structures
12943 should have 8 byte alignment. This enables the use of pairs of
12944 @code{ldd} and @code{std} instructions for copies in structure
12945 assignment, in place of twice as many @code{ld} and @code{st} pairs.
12946 However, the use of this changed alignment directly violates the SPARC
12947 ABI@. Thus, it's intended only for use on targets where the developer
12948 acknowledges that their resulting code will not be directly in line with
12949 the rules of the ABI@.
12951 @item -mimpure-text
12952 @opindex mimpure-text
12953 @option{-mimpure-text}, used in addition to @option{-shared}, tells
12954 the compiler to not pass @option{-z text} to the linker when linking a
12955 shared object. Using this option, you can link position-dependent
12956 code into a shared object.
12958 @option{-mimpure-text} suppresses the ``relocations remain against
12959 allocatable but non-writable sections'' linker error message.
12960 However, the necessary relocations will trigger copy-on-write, and the
12961 shared object is not actually shared across processes. Instead of
12962 using @option{-mimpure-text}, you should compile all source code with
12963 @option{-fpic} or @option{-fPIC}.
12965 This option is only available on SunOS and Solaris.
12967 @item -mcpu=@var{cpu_type}
12969 Set the instruction set, register set, and instruction scheduling parameters
12970 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
12971 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
12972 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
12973 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
12974 @samp{ultrasparc3}, and @samp{niagara}.
12976 Default instruction scheduling parameters are used for values that select
12977 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
12978 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
12980 Here is a list of each supported architecture and their supported
12985 v8: supersparc, hypersparc
12986 sparclite: f930, f934, sparclite86x
12988 v9: ultrasparc, ultrasparc3, niagara
12991 By default (unless configured otherwise), GCC generates code for the V7
12992 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
12993 additionally optimizes it for the Cypress CY7C602 chip, as used in the
12994 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
12995 SPARCStation 1, 2, IPX etc.
12997 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
12998 architecture. The only difference from V7 code is that the compiler emits
12999 the integer multiply and integer divide instructions which exist in SPARC-V8
13000 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
13001 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
13004 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
13005 the SPARC architecture. This adds the integer multiply, integer divide step
13006 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
13007 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
13008 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
13009 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
13010 MB86934 chip, which is the more recent SPARClite with FPU@.
13012 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
13013 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
13014 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
13015 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
13016 optimizes it for the TEMIC SPARClet chip.
13018 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
13019 architecture. This adds 64-bit integer and floating-point move instructions,
13020 3 additional floating-point condition code registers and conditional move
13021 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
13022 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
13023 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
13024 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
13025 @option{-mcpu=niagara}, the compiler additionally optimizes it for
13026 Sun UltraSPARC T1 chips.
13028 @item -mtune=@var{cpu_type}
13030 Set the instruction scheduling parameters for machine type
13031 @var{cpu_type}, but do not set the instruction set or register set that the
13032 option @option{-mcpu=@var{cpu_type}} would.
13034 The same values for @option{-mcpu=@var{cpu_type}} can be used for
13035 @option{-mtune=@var{cpu_type}}, but the only useful values are those
13036 that select a particular cpu implementation. Those are @samp{cypress},
13037 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
13038 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
13039 @samp{ultrasparc3}, and @samp{niagara}.
13044 @opindex mno-v8plus
13045 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
13046 difference from the V8 ABI is that the global and out registers are
13047 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
13048 mode for all SPARC-V9 processors.
13054 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
13055 Visual Instruction Set extensions. The default is @option{-mno-vis}.
13058 These @samp{-m} options are supported in addition to the above
13059 on SPARC-V9 processors in 64-bit environments:
13062 @item -mlittle-endian
13063 @opindex mlittle-endian
13064 Generate code for a processor running in little-endian mode. It is only
13065 available for a few configurations and most notably not on Solaris and Linux.
13071 Generate code for a 32-bit or 64-bit environment.
13072 The 32-bit environment sets int, long and pointer to 32 bits.
13073 The 64-bit environment sets int to 32 bits and long and pointer
13076 @item -mcmodel=medlow
13077 @opindex mcmodel=medlow
13078 Generate code for the Medium/Low code model: 64-bit addresses, programs
13079 must be linked in the low 32 bits of memory. Programs can be statically
13080 or dynamically linked.
13082 @item -mcmodel=medmid
13083 @opindex mcmodel=medmid
13084 Generate code for the Medium/Middle code model: 64-bit addresses, programs
13085 must be linked in the low 44 bits of memory, the text and data segments must
13086 be less than 2GB in size and the data segment must be located within 2GB of
13089 @item -mcmodel=medany
13090 @opindex mcmodel=medany
13091 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
13092 may be linked anywhere in memory, the text and data segments must be less
13093 than 2GB in size and the data segment must be located within 2GB of the
13096 @item -mcmodel=embmedany
13097 @opindex mcmodel=embmedany
13098 Generate code for the Medium/Anywhere code model for embedded systems:
13099 64-bit addresses, the text and data segments must be less than 2GB in
13100 size, both starting anywhere in memory (determined at link time). The
13101 global register %g4 points to the base of the data segment. Programs
13102 are statically linked and PIC is not supported.
13105 @itemx -mno-stack-bias
13106 @opindex mstack-bias
13107 @opindex mno-stack-bias
13108 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
13109 frame pointer if present, are offset by @minus{}2047 which must be added back
13110 when making stack frame references. This is the default in 64-bit mode.
13111 Otherwise, assume no such offset is present.
13114 These switches are supported in addition to the above on Solaris:
13119 Add support for multithreading using the Solaris threads library. This
13120 option sets flags for both the preprocessor and linker. This option does
13121 not affect the thread safety of object code produced by the compiler or
13122 that of libraries supplied with it.
13126 Add support for multithreading using the POSIX threads library. This
13127 option sets flags for both the preprocessor and linker. This option does
13128 not affect the thread safety of object code produced by the compiler or
13129 that of libraries supplied with it.
13133 This is a synonym for @option{-pthreads}.
13137 @subsection SPU Options
13138 @cindex SPU options
13140 These @samp{-m} options are supported on the SPU:
13144 @itemx -merror-reloc
13145 @opindex mwarn-reloc
13146 @opindex merror-reloc
13148 The loader for SPU does not handle dynamic relocations. By default, GCC
13149 will give an error when it generates code that requires a dynamic
13150 relocation. @option{-mno-error-reloc} disables the error,
13151 @option{-mwarn-reloc} will generate a warning instead.
13154 @itemx -munsafe-dma
13156 @opindex munsafe-dma
13158 Instructions which initiate or test completion of DMA must not be
13159 reordered with respect to loads and stores of the memory which is being
13160 accessed. Users typically address this problem using the volatile
13161 keyword, but that can lead to inefficient code in places where the
13162 memory is known to not change. Rather than mark the memory as volatile
13163 we treat the DMA instructions as potentially effecting all memory. With
13164 @option{-munsafe-dma} users must use the volatile keyword to protect
13167 @item -mbranch-hints
13168 @opindex mbranch-hints
13170 By default, GCC will generate a branch hint instruction to avoid
13171 pipeline stalls for always taken or probably taken branches. A hint
13172 will not be generated closer than 8 instructions away from its branch.
13173 There is little reason to disable them, except for debugging purposes,
13174 or to make an object a little bit smaller.
13178 @opindex msmall-mem
13179 @opindex mlarge-mem
13181 By default, GCC generates code assuming that addresses are never larger
13182 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
13183 a full 32 bit address.
13188 By default, GCC links against startup code that assumes the SPU-style
13189 main function interface (which has an unconventional parameter list).
13190 With @option{-mstdmain}, GCC will link your program against startup
13191 code that assumes a C99-style interface to @code{main}, including a
13192 local copy of @code{argv} strings.
13194 @item -mfixed-range=@var{register-range}
13195 @opindex mfixed-range
13196 Generate code treating the given register range as fixed registers.
13197 A fixed register is one that the register allocator can not use. This is
13198 useful when compiling kernel code. A register range is specified as
13199 two registers separated by a dash. Multiple register ranges can be
13200 specified separated by a comma.
13204 @node System V Options
13205 @subsection Options for System V
13207 These additional options are available on System V Release 4 for
13208 compatibility with other compilers on those systems:
13213 Create a shared object.
13214 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
13218 Identify the versions of each tool used by the compiler, in a
13219 @code{.ident} assembler directive in the output.
13223 Refrain from adding @code{.ident} directives to the output file (this is
13226 @item -YP,@var{dirs}
13228 Search the directories @var{dirs}, and no others, for libraries
13229 specified with @option{-l}.
13231 @item -Ym,@var{dir}
13233 Look in the directory @var{dir} to find the M4 preprocessor.
13234 The assembler uses this option.
13235 @c This is supposed to go with a -Yd for predefined M4 macro files, but
13236 @c the generic assembler that comes with Solaris takes just -Ym.
13239 @node TMS320C3x/C4x Options
13240 @subsection TMS320C3x/C4x Options
13241 @cindex TMS320C3x/C4x Options
13243 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
13247 @item -mcpu=@var{cpu_type}
13249 Set the instruction set, register set, and instruction scheduling
13250 parameters for machine type @var{cpu_type}. Supported values for
13251 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
13252 @samp{c44}. The default is @samp{c40} to generate code for the
13257 @itemx -msmall-memory
13259 @opindex mbig-memory
13261 @opindex msmall-memory
13263 Generates code for the big or small memory model. The small memory
13264 model assumed that all data fits into one 64K word page. At run-time
13265 the data page (DP) register must be set to point to the 64K page
13266 containing the .bss and .data program sections. The big memory model is
13267 the default and requires reloading of the DP register for every direct
13274 Allow (disallow) allocation of general integer operands into the block
13275 count register BK@.
13281 Enable (disable) generation of code using decrement and branch,
13282 DBcond(D), instructions. This is enabled by default for the C4x. To be
13283 on the safe side, this is disabled for the C3x, since the maximum
13284 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
13285 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
13286 that it can utilize the decrement and branch instruction, but will give
13287 up if there is more than one memory reference in the loop. Thus a loop
13288 where the loop counter is decremented can generate slightly more
13289 efficient code, in cases where the RPTB instruction cannot be utilized.
13291 @item -mdp-isr-reload
13293 @opindex mdp-isr-reload
13295 Force the DP register to be saved on entry to an interrupt service
13296 routine (ISR), reloaded to point to the data section, and restored on
13297 exit from the ISR@. This should not be required unless someone has
13298 violated the small memory model by modifying the DP register, say within
13305 For the C3x use the 24-bit MPYI instruction for integer multiplies
13306 instead of a library call to guarantee 32-bit results. Note that if one
13307 of the operands is a constant, then the multiplication will be performed
13308 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
13309 then squaring operations are performed inline instead of a library call.
13312 @itemx -mno-fast-fix
13314 @opindex mno-fast-fix
13315 The C3x/C4x FIX instruction to convert a floating point value to an
13316 integer value chooses the nearest integer less than or equal to the
13317 floating point value rather than to the nearest integer. Thus if the
13318 floating point number is negative, the result will be incorrectly
13319 truncated an additional code is necessary to detect and correct this
13320 case. This option can be used to disable generation of the additional
13321 code required to correct the result.
13327 Enable (disable) generation of repeat block sequences using the RPTB
13328 instruction for zero overhead looping. The RPTB construct is only used
13329 for innermost loops that do not call functions or jump across the loop
13330 boundaries. There is no advantage having nested RPTB loops due to the
13331 overhead required to save and restore the RC, RS, and RE registers.
13332 This is enabled by default with @option{-O2}.
13334 @item -mrpts=@var{count}
13338 Enable (disable) the use of the single instruction repeat instruction
13339 RPTS@. If a repeat block contains a single instruction, and the loop
13340 count can be guaranteed to be less than the value @var{count}, GCC will
13341 emit a RPTS instruction instead of a RPTB@. If no value is specified,
13342 then a RPTS will be emitted even if the loop count cannot be determined
13343 at compile time. Note that the repeated instruction following RPTS does
13344 not have to be reloaded from memory each iteration, thus freeing up the
13345 CPU buses for operands. However, since interrupts are blocked by this
13346 instruction, it is disabled by default.
13348 @item -mloop-unsigned
13349 @itemx -mno-loop-unsigned
13350 @opindex mloop-unsigned
13351 @opindex mno-loop-unsigned
13352 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
13353 is @math{2^{31} + 1} since these instructions test if the iteration count is
13354 negative to terminate the loop. If the iteration count is unsigned
13355 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
13356 exceeded. This switch allows an unsigned iteration count.
13360 Try to emit an assembler syntax that the TI assembler (asm30) is happy
13361 with. This also enforces compatibility with the API employed by the TI
13362 C3x C compiler. For example, long doubles are passed as structures
13363 rather than in floating point registers.
13369 Generate code that uses registers (stack) for passing arguments to functions.
13370 By default, arguments are passed in registers where possible rather
13371 than by pushing arguments on to the stack.
13373 @item -mparallel-insns
13374 @itemx -mno-parallel-insns
13375 @opindex mparallel-insns
13376 @opindex mno-parallel-insns
13377 Allow the generation of parallel instructions. This is enabled by
13378 default with @option{-O2}.
13380 @item -mparallel-mpy
13381 @itemx -mno-parallel-mpy
13382 @opindex mparallel-mpy
13383 @opindex mno-parallel-mpy
13384 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
13385 provided @option{-mparallel-insns} is also specified. These instructions have
13386 tight register constraints which can pessimize the code generation
13387 of large functions.
13392 @subsection V850 Options
13393 @cindex V850 Options
13395 These @samp{-m} options are defined for V850 implementations:
13399 @itemx -mno-long-calls
13400 @opindex mlong-calls
13401 @opindex mno-long-calls
13402 Treat all calls as being far away (near). If calls are assumed to be
13403 far away, the compiler will always load the functions address up into a
13404 register, and call indirect through the pointer.
13410 Do not optimize (do optimize) basic blocks that use the same index
13411 pointer 4 or more times to copy pointer into the @code{ep} register, and
13412 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
13413 option is on by default if you optimize.
13415 @item -mno-prolog-function
13416 @itemx -mprolog-function
13417 @opindex mno-prolog-function
13418 @opindex mprolog-function
13419 Do not use (do use) external functions to save and restore registers
13420 at the prologue and epilogue of a function. The external functions
13421 are slower, but use less code space if more than one function saves
13422 the same number of registers. The @option{-mprolog-function} option
13423 is on by default if you optimize.
13427 Try to make the code as small as possible. At present, this just turns
13428 on the @option{-mep} and @option{-mprolog-function} options.
13430 @item -mtda=@var{n}
13432 Put static or global variables whose size is @var{n} bytes or less into
13433 the tiny data area that register @code{ep} points to. The tiny data
13434 area can hold up to 256 bytes in total (128 bytes for byte references).
13436 @item -msda=@var{n}
13438 Put static or global variables whose size is @var{n} bytes or less into
13439 the small data area that register @code{gp} points to. The small data
13440 area can hold up to 64 kilobytes.
13442 @item -mzda=@var{n}
13444 Put static or global variables whose size is @var{n} bytes or less into
13445 the first 32 kilobytes of memory.
13449 Specify that the target processor is the V850.
13452 @opindex mbig-switch
13453 Generate code suitable for big switch tables. Use this option only if
13454 the assembler/linker complain about out of range branches within a switch
13459 This option will cause r2 and r5 to be used in the code generated by
13460 the compiler. This setting is the default.
13462 @item -mno-app-regs
13463 @opindex mno-app-regs
13464 This option will cause r2 and r5 to be treated as fixed registers.
13468 Specify that the target processor is the V850E1. The preprocessor
13469 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
13470 this option is used.
13474 Specify that the target processor is the V850E@. The preprocessor
13475 constant @samp{__v850e__} will be defined if this option is used.
13477 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
13478 are defined then a default target processor will be chosen and the
13479 relevant @samp{__v850*__} preprocessor constant will be defined.
13481 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
13482 defined, regardless of which processor variant is the target.
13484 @item -mdisable-callt
13485 @opindex mdisable-callt
13486 This option will suppress generation of the CALLT instruction for the
13487 v850e and v850e1 flavors of the v850 architecture. The default is
13488 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
13493 @subsection VAX Options
13494 @cindex VAX options
13496 These @samp{-m} options are defined for the VAX:
13501 Do not output certain jump instructions (@code{aobleq} and so on)
13502 that the Unix assembler for the VAX cannot handle across long
13507 Do output those jump instructions, on the assumption that you
13508 will assemble with the GNU assembler.
13512 Output code for g-format floating point numbers instead of d-format.
13515 @node x86-64 Options
13516 @subsection x86-64 Options
13517 @cindex x86-64 options
13519 These are listed under @xref{i386 and x86-64 Options}.
13521 @node Xstormy16 Options
13522 @subsection Xstormy16 Options
13523 @cindex Xstormy16 Options
13525 These options are defined for Xstormy16:
13530 Choose startup files and linker script suitable for the simulator.
13533 @node Xtensa Options
13534 @subsection Xtensa Options
13535 @cindex Xtensa Options
13537 These options are supported for Xtensa targets:
13541 @itemx -mno-const16
13543 @opindex mno-const16
13544 Enable or disable use of @code{CONST16} instructions for loading
13545 constant values. The @code{CONST16} instruction is currently not a
13546 standard option from Tensilica. When enabled, @code{CONST16}
13547 instructions are always used in place of the standard @code{L32R}
13548 instructions. The use of @code{CONST16} is enabled by default only if
13549 the @code{L32R} instruction is not available.
13552 @itemx -mno-fused-madd
13553 @opindex mfused-madd
13554 @opindex mno-fused-madd
13555 Enable or disable use of fused multiply/add and multiply/subtract
13556 instructions in the floating-point option. This has no effect if the
13557 floating-point option is not also enabled. Disabling fused multiply/add
13558 and multiply/subtract instructions forces the compiler to use separate
13559 instructions for the multiply and add/subtract operations. This may be
13560 desirable in some cases where strict IEEE 754-compliant results are
13561 required: the fused multiply add/subtract instructions do not round the
13562 intermediate result, thereby producing results with @emph{more} bits of
13563 precision than specified by the IEEE standard. Disabling fused multiply
13564 add/subtract instructions also ensures that the program output is not
13565 sensitive to the compiler's ability to combine multiply and add/subtract
13568 @item -mtext-section-literals
13569 @itemx -mno-text-section-literals
13570 @opindex mtext-section-literals
13571 @opindex mno-text-section-literals
13572 Control the treatment of literal pools. The default is
13573 @option{-mno-text-section-literals}, which places literals in a separate
13574 section in the output file. This allows the literal pool to be placed
13575 in a data RAM/ROM, and it also allows the linker to combine literal
13576 pools from separate object files to remove redundant literals and
13577 improve code size. With @option{-mtext-section-literals}, the literals
13578 are interspersed in the text section in order to keep them as close as
13579 possible to their references. This may be necessary for large assembly
13582 @item -mtarget-align
13583 @itemx -mno-target-align
13584 @opindex mtarget-align
13585 @opindex mno-target-align
13586 When this option is enabled, GCC instructs the assembler to
13587 automatically align instructions to reduce branch penalties at the
13588 expense of some code density. The assembler attempts to widen density
13589 instructions to align branch targets and the instructions following call
13590 instructions. If there are not enough preceding safe density
13591 instructions to align a target, no widening will be performed. The
13592 default is @option{-mtarget-align}. These options do not affect the
13593 treatment of auto-aligned instructions like @code{LOOP}, which the
13594 assembler will always align, either by widening density instructions or
13595 by inserting no-op instructions.
13598 @itemx -mno-longcalls
13599 @opindex mlongcalls
13600 @opindex mno-longcalls
13601 When this option is enabled, GCC instructs the assembler to translate
13602 direct calls to indirect calls unless it can determine that the target
13603 of a direct call is in the range allowed by the call instruction. This
13604 translation typically occurs for calls to functions in other source
13605 files. Specifically, the assembler translates a direct @code{CALL}
13606 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
13607 The default is @option{-mno-longcalls}. This option should be used in
13608 programs where the call target can potentially be out of range. This
13609 option is implemented in the assembler, not the compiler, so the
13610 assembly code generated by GCC will still show direct call
13611 instructions---look at the disassembled object code to see the actual
13612 instructions. Note that the assembler will use an indirect call for
13613 every cross-file call, not just those that really will be out of range.
13616 @node zSeries Options
13617 @subsection zSeries Options
13618 @cindex zSeries options
13620 These are listed under @xref{S/390 and zSeries Options}.
13622 @node Code Gen Options
13623 @section Options for Code Generation Conventions
13624 @cindex code generation conventions
13625 @cindex options, code generation
13626 @cindex run-time options
13628 These machine-independent options control the interface conventions
13629 used in code generation.
13631 Most of them have both positive and negative forms; the negative form
13632 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
13633 one of the forms is listed---the one which is not the default. You
13634 can figure out the other form by either removing @samp{no-} or adding
13638 @item -fbounds-check
13639 @opindex fbounds-check
13640 For front-ends that support it, generate additional code to check that
13641 indices used to access arrays are within the declared range. This is
13642 currently only supported by the Java and Fortran front-ends, where
13643 this option defaults to true and false respectively.
13647 This option generates traps for signed overflow on addition, subtraction,
13648 multiplication operations.
13652 This option instructs the compiler to assume that signed arithmetic
13653 overflow of addition, subtraction and multiplication wraps around
13654 using twos-complement representation. This flag enables some optimizations
13655 and disables others. This option is enabled by default for the Java
13656 front-end, as required by the Java language specification.
13659 @opindex fexceptions
13660 Enable exception handling. Generates extra code needed to propagate
13661 exceptions. For some targets, this implies GCC will generate frame
13662 unwind information for all functions, which can produce significant data
13663 size overhead, although it does not affect execution. If you do not
13664 specify this option, GCC will enable it by default for languages like
13665 C++ which normally require exception handling, and disable it for
13666 languages like C that do not normally require it. However, you may need
13667 to enable this option when compiling C code that needs to interoperate
13668 properly with exception handlers written in C++. You may also wish to
13669 disable this option if you are compiling older C++ programs that don't
13670 use exception handling.
13672 @item -fnon-call-exceptions
13673 @opindex fnon-call-exceptions
13674 Generate code that allows trapping instructions to throw exceptions.
13675 Note that this requires platform-specific runtime support that does
13676 not exist everywhere. Moreover, it only allows @emph{trapping}
13677 instructions to throw exceptions, i.e.@: memory references or floating
13678 point instructions. It does not allow exceptions to be thrown from
13679 arbitrary signal handlers such as @code{SIGALRM}.
13681 @item -funwind-tables
13682 @opindex funwind-tables
13683 Similar to @option{-fexceptions}, except that it will just generate any needed
13684 static data, but will not affect the generated code in any other way.
13685 You will normally not enable this option; instead, a language processor
13686 that needs this handling would enable it on your behalf.
13688 @item -fasynchronous-unwind-tables
13689 @opindex fasynchronous-unwind-tables
13690 Generate unwind table in dwarf2 format, if supported by target machine. The
13691 table is exact at each instruction boundary, so it can be used for stack
13692 unwinding from asynchronous events (such as debugger or garbage collector).
13694 @item -fpcc-struct-return
13695 @opindex fpcc-struct-return
13696 Return ``short'' @code{struct} and @code{union} values in memory like
13697 longer ones, rather than in registers. This convention is less
13698 efficient, but it has the advantage of allowing intercallability between
13699 GCC-compiled files and files compiled with other compilers, particularly
13700 the Portable C Compiler (pcc).
13702 The precise convention for returning structures in memory depends
13703 on the target configuration macros.
13705 Short structures and unions are those whose size and alignment match
13706 that of some integer type.
13708 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
13709 switch is not binary compatible with code compiled with the
13710 @option{-freg-struct-return} switch.
13711 Use it to conform to a non-default application binary interface.
13713 @item -freg-struct-return
13714 @opindex freg-struct-return
13715 Return @code{struct} and @code{union} values in registers when possible.
13716 This is more efficient for small structures than
13717 @option{-fpcc-struct-return}.
13719 If you specify neither @option{-fpcc-struct-return} nor
13720 @option{-freg-struct-return}, GCC defaults to whichever convention is
13721 standard for the target. If there is no standard convention, GCC
13722 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
13723 the principal compiler. In those cases, we can choose the standard, and
13724 we chose the more efficient register return alternative.
13726 @strong{Warning:} code compiled with the @option{-freg-struct-return}
13727 switch is not binary compatible with code compiled with the
13728 @option{-fpcc-struct-return} switch.
13729 Use it to conform to a non-default application binary interface.
13731 @item -fshort-enums
13732 @opindex fshort-enums
13733 Allocate to an @code{enum} type only as many bytes as it needs for the
13734 declared range of possible values. Specifically, the @code{enum} type
13735 will be equivalent to the smallest integer type which has enough room.
13737 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
13738 code that is not binary compatible with code generated without that switch.
13739 Use it to conform to a non-default application binary interface.
13741 @item -fshort-double
13742 @opindex fshort-double
13743 Use the same size for @code{double} as for @code{float}.
13745 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
13746 code that is not binary compatible with code generated without that switch.
13747 Use it to conform to a non-default application binary interface.
13749 @item -fshort-wchar
13750 @opindex fshort-wchar
13751 Override the underlying type for @samp{wchar_t} to be @samp{short
13752 unsigned int} instead of the default for the target. This option is
13753 useful for building programs to run under WINE@.
13755 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
13756 code that is not binary compatible with code generated without that switch.
13757 Use it to conform to a non-default application binary interface.
13760 @opindex fno-common
13761 In C, allocate even uninitialized global variables in the data section of the
13762 object file, rather than generating them as common blocks. This has the
13763 effect that if the same variable is declared (without @code{extern}) in
13764 two different compilations, you will get an error when you link them.
13765 The only reason this might be useful is if you wish to verify that the
13766 program will work on other systems which always work this way.
13770 Ignore the @samp{#ident} directive.
13772 @item -finhibit-size-directive
13773 @opindex finhibit-size-directive
13774 Don't output a @code{.size} assembler directive, or anything else that
13775 would cause trouble if the function is split in the middle, and the
13776 two halves are placed at locations far apart in memory. This option is
13777 used when compiling @file{crtstuff.c}; you should not need to use it
13780 @item -fverbose-asm
13781 @opindex fverbose-asm
13782 Put extra commentary information in the generated assembly code to
13783 make it more readable. This option is generally only of use to those
13784 who actually need to read the generated assembly code (perhaps while
13785 debugging the compiler itself).
13787 @option{-fno-verbose-asm}, the default, causes the
13788 extra information to be omitted and is useful when comparing two assembler
13791 @item -frecord-gcc-switches
13792 @opindex frecord-gcc-switches
13793 This switch causes the command line that was used to invoke the
13794 compiler to be recorded into the object file that is being created.
13795 This switch is only implemented on some targets and the exact format
13796 of the recording is target and binary file format dependent, but it
13797 usually takes the form of a section containing ASCII text. This
13798 switch is related to the @option{-fverbose-asm} switch, but that
13799 switch only records information in the assembler output file as
13800 comments, so it never reaches the object file.
13804 @cindex global offset table
13806 Generate position-independent code (PIC) suitable for use in a shared
13807 library, if supported for the target machine. Such code accesses all
13808 constant addresses through a global offset table (GOT)@. The dynamic
13809 loader resolves the GOT entries when the program starts (the dynamic
13810 loader is not part of GCC; it is part of the operating system). If
13811 the GOT size for the linked executable exceeds a machine-specific
13812 maximum size, you get an error message from the linker indicating that
13813 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
13814 instead. (These maximums are 8k on the SPARC and 32k
13815 on the m68k and RS/6000. The 386 has no such limit.)
13817 Position-independent code requires special support, and therefore works
13818 only on certain machines. For the 386, GCC supports PIC for System V
13819 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
13820 position-independent.
13822 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
13827 If supported for the target machine, emit position-independent code,
13828 suitable for dynamic linking and avoiding any limit on the size of the
13829 global offset table. This option makes a difference on the m68k,
13830 PowerPC and SPARC@.
13832 Position-independent code requires special support, and therefore works
13833 only on certain machines.
13835 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
13842 These options are similar to @option{-fpic} and @option{-fPIC}, but
13843 generated position independent code can be only linked into executables.
13844 Usually these options are used when @option{-pie} GCC option will be
13845 used during linking.
13847 @item -fno-jump-tables
13848 @opindex fno-jump-tables
13849 Do not use jump tables for switch statements even where it would be
13850 more efficient than other code generation strategies. This option is
13851 of use in conjunction with @option{-fpic} or @option{-fPIC} for
13852 building code which forms part of a dynamic linker and cannot
13853 reference the address of a jump table. On some targets, jump tables
13854 do not require a GOT and this option is not needed.
13856 @item -ffixed-@var{reg}
13858 Treat the register named @var{reg} as a fixed register; generated code
13859 should never refer to it (except perhaps as a stack pointer, frame
13860 pointer or in some other fixed role).
13862 @var{reg} must be the name of a register. The register names accepted
13863 are machine-specific and are defined in the @code{REGISTER_NAMES}
13864 macro in the machine description macro file.
13866 This flag does not have a negative form, because it specifies a
13869 @item -fcall-used-@var{reg}
13870 @opindex fcall-used
13871 Treat the register named @var{reg} as an allocable register that is
13872 clobbered by function calls. It may be allocated for temporaries or
13873 variables that do not live across a call. Functions compiled this way
13874 will not save and restore the register @var{reg}.
13876 It is an error to used this flag with the frame pointer or stack pointer.
13877 Use of this flag for other registers that have fixed pervasive roles in
13878 the machine's execution model will produce disastrous results.
13880 This flag does not have a negative form, because it specifies a
13883 @item -fcall-saved-@var{reg}
13884 @opindex fcall-saved
13885 Treat the register named @var{reg} as an allocable register saved by
13886 functions. It may be allocated even for temporaries or variables that
13887 live across a call. Functions compiled this way will save and restore
13888 the register @var{reg} if they use it.
13890 It is an error to used this flag with the frame pointer or stack pointer.
13891 Use of this flag for other registers that have fixed pervasive roles in
13892 the machine's execution model will produce disastrous results.
13894 A different sort of disaster will result from the use of this flag for
13895 a register in which function values may be returned.
13897 This flag does not have a negative form, because it specifies a
13900 @item -fpack-struct[=@var{n}]
13901 @opindex fpack-struct
13902 Without a value specified, pack all structure members together without
13903 holes. When a value is specified (which must be a small power of two), pack
13904 structure members according to this value, representing the maximum
13905 alignment (that is, objects with default alignment requirements larger than
13906 this will be output potentially unaligned at the next fitting location.
13908 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
13909 code that is not binary compatible with code generated without that switch.
13910 Additionally, it makes the code suboptimal.
13911 Use it to conform to a non-default application binary interface.
13913 @item -finstrument-functions
13914 @opindex finstrument-functions
13915 Generate instrumentation calls for entry and exit to functions. Just
13916 after function entry and just before function exit, the following
13917 profiling functions will be called with the address of the current
13918 function and its call site. (On some platforms,
13919 @code{__builtin_return_address} does not work beyond the current
13920 function, so the call site information may not be available to the
13921 profiling functions otherwise.)
13924 void __cyg_profile_func_enter (void *this_fn,
13926 void __cyg_profile_func_exit (void *this_fn,
13930 The first argument is the address of the start of the current function,
13931 which may be looked up exactly in the symbol table.
13933 This instrumentation is also done for functions expanded inline in other
13934 functions. The profiling calls will indicate where, conceptually, the
13935 inline function is entered and exited. This means that addressable
13936 versions of such functions must be available. If all your uses of a
13937 function are expanded inline, this may mean an additional expansion of
13938 code size. If you use @samp{extern inline} in your C code, an
13939 addressable version of such functions must be provided. (This is
13940 normally the case anyways, but if you get lucky and the optimizer always
13941 expands the functions inline, you might have gotten away without
13942 providing static copies.)
13944 A function may be given the attribute @code{no_instrument_function}, in
13945 which case this instrumentation will not be done. This can be used, for
13946 example, for the profiling functions listed above, high-priority
13947 interrupt routines, and any functions from which the profiling functions
13948 cannot safely be called (perhaps signal handlers, if the profiling
13949 routines generate output or allocate memory).
13951 @item -fstack-check
13952 @opindex fstack-check
13953 Generate code to verify that you do not go beyond the boundary of the
13954 stack. You should specify this flag if you are running in an
13955 environment with multiple threads, but only rarely need to specify it in
13956 a single-threaded environment since stack overflow is automatically
13957 detected on nearly all systems if there is only one stack.
13959 Note that this switch does not actually cause checking to be done; the
13960 operating system must do that. The switch causes generation of code
13961 to ensure that the operating system sees the stack being extended.
13963 @item -fstack-limit-register=@var{reg}
13964 @itemx -fstack-limit-symbol=@var{sym}
13965 @itemx -fno-stack-limit
13966 @opindex fstack-limit-register
13967 @opindex fstack-limit-symbol
13968 @opindex fno-stack-limit
13969 Generate code to ensure that the stack does not grow beyond a certain value,
13970 either the value of a register or the address of a symbol. If the stack
13971 would grow beyond the value, a signal is raised. For most targets,
13972 the signal is raised before the stack overruns the boundary, so
13973 it is possible to catch the signal without taking special precautions.
13975 For instance, if the stack starts at absolute address @samp{0x80000000}
13976 and grows downwards, you can use the flags
13977 @option{-fstack-limit-symbol=__stack_limit} and
13978 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
13979 of 128KB@. Note that this may only work with the GNU linker.
13981 @cindex aliasing of parameters
13982 @cindex parameters, aliased
13983 @item -fargument-alias
13984 @itemx -fargument-noalias
13985 @itemx -fargument-noalias-global
13986 @itemx -fargument-noalias-anything
13987 @opindex fargument-alias
13988 @opindex fargument-noalias
13989 @opindex fargument-noalias-global
13990 @opindex fargument-noalias-anything
13991 Specify the possible relationships among parameters and between
13992 parameters and global data.
13994 @option{-fargument-alias} specifies that arguments (parameters) may
13995 alias each other and may alias global storage.@*
13996 @option{-fargument-noalias} specifies that arguments do not alias
13997 each other, but may alias global storage.@*
13998 @option{-fargument-noalias-global} specifies that arguments do not
13999 alias each other and do not alias global storage.
14000 @option{-fargument-noalias-anything} specifies that arguments do not
14001 alias any other storage.
14003 Each language will automatically use whatever option is required by
14004 the language standard. You should not need to use these options yourself.
14006 @item -fleading-underscore
14007 @opindex fleading-underscore
14008 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
14009 change the way C symbols are represented in the object file. One use
14010 is to help link with legacy assembly code.
14012 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
14013 generate code that is not binary compatible with code generated without that
14014 switch. Use it to conform to a non-default application binary interface.
14015 Not all targets provide complete support for this switch.
14017 @item -ftls-model=@var{model}
14018 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
14019 The @var{model} argument should be one of @code{global-dynamic},
14020 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
14022 The default without @option{-fpic} is @code{initial-exec}; with
14023 @option{-fpic} the default is @code{global-dynamic}.
14025 @item -fvisibility=@var{default|internal|hidden|protected}
14026 @opindex fvisibility
14027 Set the default ELF image symbol visibility to the specified option---all
14028 symbols will be marked with this unless overridden within the code.
14029 Using this feature can very substantially improve linking and
14030 load times of shared object libraries, produce more optimized
14031 code, provide near-perfect API export and prevent symbol clashes.
14032 It is @strong{strongly} recommended that you use this in any shared objects
14035 Despite the nomenclature, @code{default} always means public ie;
14036 available to be linked against from outside the shared object.
14037 @code{protected} and @code{internal} are pretty useless in real-world
14038 usage so the only other commonly used option will be @code{hidden}.
14039 The default if @option{-fvisibility} isn't specified is
14040 @code{default}, i.e., make every
14041 symbol public---this causes the same behavior as previous versions of
14044 A good explanation of the benefits offered by ensuring ELF
14045 symbols have the correct visibility is given by ``How To Write
14046 Shared Libraries'' by Ulrich Drepper (which can be found at
14047 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
14048 solution made possible by this option to marking things hidden when
14049 the default is public is to make the default hidden and mark things
14050 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
14051 and @code{__attribute__ ((visibility("default")))} instead of
14052 @code{__declspec(dllexport)} you get almost identical semantics with
14053 identical syntax. This is a great boon to those working with
14054 cross-platform projects.
14056 For those adding visibility support to existing code, you may find
14057 @samp{#pragma GCC visibility} of use. This works by you enclosing
14058 the declarations you wish to set visibility for with (for example)
14059 @samp{#pragma GCC visibility push(hidden)} and
14060 @samp{#pragma GCC visibility pop}.
14061 Bear in mind that symbol visibility should be viewed @strong{as
14062 part of the API interface contract} and thus all new code should
14063 always specify visibility when it is not the default ie; declarations
14064 only for use within the local DSO should @strong{always} be marked explicitly
14065 as hidden as so to avoid PLT indirection overheads---making this
14066 abundantly clear also aids readability and self-documentation of the code.
14067 Note that due to ISO C++ specification requirements, operator new and
14068 operator delete must always be of default visibility.
14070 Be aware that headers from outside your project, in particular system
14071 headers and headers from any other library you use, may not be
14072 expecting to be compiled with visibility other than the default. You
14073 may need to explicitly say @samp{#pragma GCC visibility push(default)}
14074 before including any such headers.
14076 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
14077 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
14078 no modifications. However, this means that calls to @samp{extern}
14079 functions with no explicit visibility will use the PLT, so it is more
14080 effective to use @samp{__attribute ((visibility))} and/or
14081 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
14082 declarations should be treated as hidden.
14084 Note that @samp{-fvisibility} does affect C++ vague linkage
14085 entities. This means that, for instance, an exception class that will
14086 be thrown between DSOs must be explicitly marked with default
14087 visibility so that the @samp{type_info} nodes will be unified between
14090 An overview of these techniques, their benefits and how to use them
14091 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
14097 @node Environment Variables
14098 @section Environment Variables Affecting GCC
14099 @cindex environment variables
14101 @c man begin ENVIRONMENT
14102 This section describes several environment variables that affect how GCC
14103 operates. Some of them work by specifying directories or prefixes to use
14104 when searching for various kinds of files. Some are used to specify other
14105 aspects of the compilation environment.
14107 Note that you can also specify places to search using options such as
14108 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
14109 take precedence over places specified using environment variables, which
14110 in turn take precedence over those specified by the configuration of GCC@.
14111 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
14112 GNU Compiler Collection (GCC) Internals}.
14117 @c @itemx LC_COLLATE
14119 @c @itemx LC_MONETARY
14120 @c @itemx LC_NUMERIC
14125 @c @findex LC_COLLATE
14126 @findex LC_MESSAGES
14127 @c @findex LC_MONETARY
14128 @c @findex LC_NUMERIC
14132 These environment variables control the way that GCC uses
14133 localization information that allow GCC to work with different
14134 national conventions. GCC inspects the locale categories
14135 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
14136 so. These locale categories can be set to any value supported by your
14137 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
14138 Kingdom encoded in UTF-8.
14140 The @env{LC_CTYPE} environment variable specifies character
14141 classification. GCC uses it to determine the character boundaries in
14142 a string; this is needed for some multibyte encodings that contain quote
14143 and escape characters that would otherwise be interpreted as a string
14146 The @env{LC_MESSAGES} environment variable specifies the language to
14147 use in diagnostic messages.
14149 If the @env{LC_ALL} environment variable is set, it overrides the value
14150 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
14151 and @env{LC_MESSAGES} default to the value of the @env{LANG}
14152 environment variable. If none of these variables are set, GCC
14153 defaults to traditional C English behavior.
14157 If @env{TMPDIR} is set, it specifies the directory to use for temporary
14158 files. GCC uses temporary files to hold the output of one stage of
14159 compilation which is to be used as input to the next stage: for example,
14160 the output of the preprocessor, which is the input to the compiler
14163 @item GCC_EXEC_PREFIX
14164 @findex GCC_EXEC_PREFIX
14165 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
14166 names of the subprograms executed by the compiler. No slash is added
14167 when this prefix is combined with the name of a subprogram, but you can
14168 specify a prefix that ends with a slash if you wish.
14170 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
14171 an appropriate prefix to use based on the pathname it was invoked with.
14173 If GCC cannot find the subprogram using the specified prefix, it
14174 tries looking in the usual places for the subprogram.
14176 The default value of @env{GCC_EXEC_PREFIX} is
14177 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
14178 the installed compiler. In many cases @var{prefix} is the value
14179 of @code{prefix} when you ran the @file{configure} script.
14181 Other prefixes specified with @option{-B} take precedence over this prefix.
14183 This prefix is also used for finding files such as @file{crt0.o} that are
14186 In addition, the prefix is used in an unusual way in finding the
14187 directories to search for header files. For each of the standard
14188 directories whose name normally begins with @samp{/usr/local/lib/gcc}
14189 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
14190 replacing that beginning with the specified prefix to produce an
14191 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
14192 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
14193 These alternate directories are searched first; the standard directories
14194 come next. If a standard directory begins with the configured
14195 @var{prefix} then the value of @var{prefix} is replaced by
14196 @env{GCC_EXEC_PREFIX} when looking for header files.
14198 @item COMPILER_PATH
14199 @findex COMPILER_PATH
14200 The value of @env{COMPILER_PATH} is a colon-separated list of
14201 directories, much like @env{PATH}. GCC tries the directories thus
14202 specified when searching for subprograms, if it can't find the
14203 subprograms using @env{GCC_EXEC_PREFIX}.
14206 @findex LIBRARY_PATH
14207 The value of @env{LIBRARY_PATH} is a colon-separated list of
14208 directories, much like @env{PATH}. When configured as a native compiler,
14209 GCC tries the directories thus specified when searching for special
14210 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
14211 using GCC also uses these directories when searching for ordinary
14212 libraries for the @option{-l} option (but directories specified with
14213 @option{-L} come first).
14217 @cindex locale definition
14218 This variable is used to pass locale information to the compiler. One way in
14219 which this information is used is to determine the character set to be used
14220 when character literals, string literals and comments are parsed in C and C++.
14221 When the compiler is configured to allow multibyte characters,
14222 the following values for @env{LANG} are recognized:
14226 Recognize JIS characters.
14228 Recognize SJIS characters.
14230 Recognize EUCJP characters.
14233 If @env{LANG} is not defined, or if it has some other value, then the
14234 compiler will use mblen and mbtowc as defined by the default locale to
14235 recognize and translate multibyte characters.
14239 Some additional environments variables affect the behavior of the
14242 @include cppenv.texi
14246 @node Precompiled Headers
14247 @section Using Precompiled Headers
14248 @cindex precompiled headers
14249 @cindex speed of compilation
14251 Often large projects have many header files that are included in every
14252 source file. The time the compiler takes to process these header files
14253 over and over again can account for nearly all of the time required to
14254 build the project. To make builds faster, GCC allows users to
14255 `precompile' a header file; then, if builds can use the precompiled
14256 header file they will be much faster.
14258 To create a precompiled header file, simply compile it as you would any
14259 other file, if necessary using the @option{-x} option to make the driver
14260 treat it as a C or C++ header file. You will probably want to use a
14261 tool like @command{make} to keep the precompiled header up-to-date when
14262 the headers it contains change.
14264 A precompiled header file will be searched for when @code{#include} is
14265 seen in the compilation. As it searches for the included file
14266 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
14267 compiler looks for a precompiled header in each directory just before it
14268 looks for the include file in that directory. The name searched for is
14269 the name specified in the @code{#include} with @samp{.gch} appended. If
14270 the precompiled header file can't be used, it is ignored.
14272 For instance, if you have @code{#include "all.h"}, and you have
14273 @file{all.h.gch} in the same directory as @file{all.h}, then the
14274 precompiled header file will be used if possible, and the original
14275 header will be used otherwise.
14277 Alternatively, you might decide to put the precompiled header file in a
14278 directory and use @option{-I} to ensure that directory is searched
14279 before (or instead of) the directory containing the original header.
14280 Then, if you want to check that the precompiled header file is always
14281 used, you can put a file of the same name as the original header in this
14282 directory containing an @code{#error} command.
14284 This also works with @option{-include}. So yet another way to use
14285 precompiled headers, good for projects not designed with precompiled
14286 header files in mind, is to simply take most of the header files used by
14287 a project, include them from another header file, precompile that header
14288 file, and @option{-include} the precompiled header. If the header files
14289 have guards against multiple inclusion, they will be skipped because
14290 they've already been included (in the precompiled header).
14292 If you need to precompile the same header file for different
14293 languages, targets, or compiler options, you can instead make a
14294 @emph{directory} named like @file{all.h.gch}, and put each precompiled
14295 header in the directory, perhaps using @option{-o}. It doesn't matter
14296 what you call the files in the directory, every precompiled header in
14297 the directory will be considered. The first precompiled header
14298 encountered in the directory that is valid for this compilation will
14299 be used; they're searched in no particular order.
14301 There are many other possibilities, limited only by your imagination,
14302 good sense, and the constraints of your build system.
14304 A precompiled header file can be used only when these conditions apply:
14308 Only one precompiled header can be used in a particular compilation.
14311 A precompiled header can't be used once the first C token is seen. You
14312 can have preprocessor directives before a precompiled header; you can
14313 even include a precompiled header from inside another header, so long as
14314 there are no C tokens before the @code{#include}.
14317 The precompiled header file must be produced for the same language as
14318 the current compilation. You can't use a C precompiled header for a C++
14322 The precompiled header file must have been produced by the same compiler
14323 binary as the current compilation is using.
14326 Any macros defined before the precompiled header is included must
14327 either be defined in the same way as when the precompiled header was
14328 generated, or must not affect the precompiled header, which usually
14329 means that they don't appear in the precompiled header at all.
14331 The @option{-D} option is one way to define a macro before a
14332 precompiled header is included; using a @code{#define} can also do it.
14333 There are also some options that define macros implicitly, like
14334 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
14337 @item If debugging information is output when using the precompiled
14338 header, using @option{-g} or similar, the same kind of debugging information
14339 must have been output when building the precompiled header. However,
14340 a precompiled header built using @option{-g} can be used in a compilation
14341 when no debugging information is being output.
14343 @item The same @option{-m} options must generally be used when building
14344 and using the precompiled header. @xref{Submodel Options},
14345 for any cases where this rule is relaxed.
14347 @item Each of the following options must be the same when building and using
14348 the precompiled header:
14350 @gccoptlist{-fexceptions -funit-at-a-time}
14353 Some other command-line options starting with @option{-f},
14354 @option{-p}, or @option{-O} must be defined in the same way as when
14355 the precompiled header was generated. At present, it's not clear
14356 which options are safe to change and which are not; the safest choice
14357 is to use exactly the same options when generating and using the
14358 precompiled header. The following are known to be safe:
14360 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
14361 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
14362 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
14367 For all of these except the last, the compiler will automatically
14368 ignore the precompiled header if the conditions aren't met. If you
14369 find an option combination that doesn't work and doesn't cause the
14370 precompiled header to be ignored, please consider filing a bug report,
14373 If you do use differing options when generating and using the
14374 precompiled header, the actual behavior will be a mixture of the
14375 behavior for the options. For instance, if you use @option{-g} to
14376 generate the precompiled header but not when using it, you may or may
14377 not get debugging information for routines in the precompiled header.
14379 @node Running Protoize
14380 @section Running Protoize
14382 The program @code{protoize} is an optional part of GCC@. You can use
14383 it to add prototypes to a program, thus converting the program to ISO
14384 C in one respect. The companion program @code{unprotoize} does the
14385 reverse: it removes argument types from any prototypes that are found.
14387 When you run these programs, you must specify a set of source files as
14388 command line arguments. The conversion programs start out by compiling
14389 these files to see what functions they define. The information gathered
14390 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
14392 After scanning comes actual conversion. The specified files are all
14393 eligible to be converted; any files they include (whether sources or
14394 just headers) are eligible as well.
14396 But not all the eligible files are converted. By default,
14397 @code{protoize} and @code{unprotoize} convert only source and header
14398 files in the current directory. You can specify additional directories
14399 whose files should be converted with the @option{-d @var{directory}}
14400 option. You can also specify particular files to exclude with the
14401 @option{-x @var{file}} option. A file is converted if it is eligible, its
14402 directory name matches one of the specified directory names, and its
14403 name within the directory has not been excluded.
14405 Basic conversion with @code{protoize} consists of rewriting most
14406 function definitions and function declarations to specify the types of
14407 the arguments. The only ones not rewritten are those for varargs
14410 @code{protoize} optionally inserts prototype declarations at the
14411 beginning of the source file, to make them available for any calls that
14412 precede the function's definition. Or it can insert prototype
14413 declarations with block scope in the blocks where undeclared functions
14416 Basic conversion with @code{unprotoize} consists of rewriting most
14417 function declarations to remove any argument types, and rewriting
14418 function definitions to the old-style pre-ISO form.
14420 Both conversion programs print a warning for any function declaration or
14421 definition that they can't convert. You can suppress these warnings
14424 The output from @code{protoize} or @code{unprotoize} replaces the
14425 original source file. The original file is renamed to a name ending
14426 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
14427 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
14428 for DOS) file already exists, then the source file is simply discarded.
14430 @code{protoize} and @code{unprotoize} both depend on GCC itself to
14431 scan the program and collect information about the functions it uses.
14432 So neither of these programs will work until GCC is installed.
14434 Here is a table of the options you can use with @code{protoize} and
14435 @code{unprotoize}. Each option works with both programs unless
14439 @item -B @var{directory}
14440 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
14441 usual directory (normally @file{/usr/local/lib}). This file contains
14442 prototype information about standard system functions. This option
14443 applies only to @code{protoize}.
14445 @item -c @var{compilation-options}
14446 Use @var{compilation-options} as the options when running @command{gcc} to
14447 produce the @samp{.X} files. The special option @option{-aux-info} is
14448 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
14450 Note that the compilation options must be given as a single argument to
14451 @code{protoize} or @code{unprotoize}. If you want to specify several
14452 @command{gcc} options, you must quote the entire set of compilation options
14453 to make them a single word in the shell.
14455 There are certain @command{gcc} arguments that you cannot use, because they
14456 would produce the wrong kind of output. These include @option{-g},
14457 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
14458 the @var{compilation-options}, they are ignored.
14461 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
14462 systems) instead of @samp{.c}. This is convenient if you are converting
14463 a C program to C++. This option applies only to @code{protoize}.
14466 Add explicit global declarations. This means inserting explicit
14467 declarations at the beginning of each source file for each function
14468 that is called in the file and was not declared. These declarations
14469 precede the first function definition that contains a call to an
14470 undeclared function. This option applies only to @code{protoize}.
14472 @item -i @var{string}
14473 Indent old-style parameter declarations with the string @var{string}.
14474 This option applies only to @code{protoize}.
14476 @code{unprotoize} converts prototyped function definitions to old-style
14477 function definitions, where the arguments are declared between the
14478 argument list and the initial @samp{@{}. By default, @code{unprotoize}
14479 uses five spaces as the indentation. If you want to indent with just
14480 one space instead, use @option{-i " "}.
14483 Keep the @samp{.X} files. Normally, they are deleted after conversion
14487 Add explicit local declarations. @code{protoize} with @option{-l} inserts
14488 a prototype declaration for each function in each block which calls the
14489 function without any declaration. This option applies only to
14493 Make no real changes. This mode just prints information about the conversions
14494 that would have been done without @option{-n}.
14497 Make no @samp{.save} files. The original files are simply deleted.
14498 Use this option with caution.
14500 @item -p @var{program}
14501 Use the program @var{program} as the compiler. Normally, the name
14502 @file{gcc} is used.
14505 Work quietly. Most warnings are suppressed.
14508 Print the version number, just like @option{-v} for @command{gcc}.
14511 If you need special compiler options to compile one of your program's
14512 source files, then you should generate that file's @samp{.X} file
14513 specially, by running @command{gcc} on that source file with the
14514 appropriate options and the option @option{-aux-info}. Then run
14515 @code{protoize} on the entire set of files. @code{protoize} will use
14516 the existing @samp{.X} file because it is newer than the source file.
14520 gcc -Dfoo=bar file1.c -aux-info file1.X
14525 You need to include the special files along with the rest in the
14526 @code{protoize} command, even though their @samp{.X} files already
14527 exist, because otherwise they won't get converted.
14529 @xref{Protoize Caveats}, for more information on how to use
14530 @code{protoize} successfully.