1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
3 @c Free Software Foundation, Inc.
4 @c This is part of the GCC manual.
5 @c For copying conditions, see the file gcc.texi.
12 @c man begin COPYRIGHT
13 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
14 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
15 Free Software Foundation, Inc.
17 Permission is granted to copy, distribute and/or modify this document
18 under the terms of the GNU Free Documentation License, Version 1.2 or
19 any later version published by the Free Software Foundation; with the
20 Invariant Sections being ``GNU General Public License'' and ``Funding
21 Free Software'', the Front-Cover texts being (a) (see below), and with
22 the Back-Cover Texts being (b) (see below). A copy of the license is
23 included in the gfdl(7) man page.
25 (a) The FSF's Front-Cover Text is:
29 (b) The FSF's Back-Cover Text is:
31 You have freedom to copy and modify this GNU Manual, like GNU
32 software. Copies published by the Free Software Foundation raise
33 funds for GNU development.
35 @c Set file name and title for the man page.
37 @settitle GNU project C and C++ compiler
39 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
40 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
41 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
42 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
43 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
44 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
45 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
47 Only the most useful options are listed here; see below for the
48 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
51 gpl(7), gfdl(7), fsf-funding(7),
52 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
53 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
54 @file{ld}, @file{binutils} and @file{gdb}.
57 For instructions on reporting bugs, see
61 See the Info entry for @command{gcc}, or
62 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
63 for contributors to GCC@.
68 @chapter GCC Command Options
69 @cindex GCC command options
70 @cindex command options
71 @cindex options, GCC command
73 @c man begin DESCRIPTION
74 When you invoke GCC, it normally does preprocessing, compilation,
75 assembly and linking. The ``overall options'' allow you to stop this
76 process at an intermediate stage. For example, the @option{-c} option
77 says not to run the linker. Then the output consists of object files
78 output by the assembler.
80 Other options are passed on to one stage of processing. Some options
81 control the preprocessor and others the compiler itself. Yet other
82 options control the assembler and linker; most of these are not
83 documented here, since you rarely need to use any of them.
85 @cindex C compilation options
86 Most of the command line options that you can use with GCC are useful
87 for C programs; when an option is only useful with another language
88 (usually C++), the explanation says so explicitly. If the description
89 for a particular option does not mention a source language, you can use
90 that option with all supported languages.
92 @cindex C++ compilation options
93 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
94 options for compiling C++ programs.
96 @cindex grouping options
97 @cindex options, grouping
98 The @command{gcc} program accepts options and file names as operands. Many
99 options have multi-letter names; therefore multiple single-letter options
100 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
103 @cindex order of options
104 @cindex options, order
105 You can mix options and other arguments. For the most part, the order
106 you use doesn't matter. Order does matter when you use several
107 options of the same kind; for example, if you specify @option{-L} more
108 than once, the directories are searched in the order specified. Also,
109 the placement of the @option{-l} option is significant.
111 Many options have long names starting with @samp{-f} or with
112 @samp{-W}---for example,
113 @option{-fmove-loop-invariants}, @option{-Wformat} and so on. Most of
114 these have both positive and negative forms; the negative form of
115 @option{-ffoo} would be @option{-fno-foo}. This manual documents
116 only one of these two forms, whichever one is not the default.
120 @xref{Option Index}, for an index to GCC's options.
123 * Option Summary:: Brief list of all options, without explanations.
124 * Overall Options:: Controlling the kind of output:
125 an executable, object files, assembler files,
126 or preprocessed source.
127 * Invoking G++:: Compiling C++ programs.
128 * C Dialect Options:: Controlling the variant of C language compiled.
129 * C++ Dialect Options:: Variations on C++.
130 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
132 * Language Independent Options:: Controlling how diagnostics should be
134 * Warning Options:: How picky should the compiler be?
135 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
136 * Optimize Options:: How much optimization?
137 * Preprocessor Options:: Controlling header files and macro definitions.
138 Also, getting dependency information for Make.
139 * Assembler Options:: Passing options to the assembler.
140 * Link Options:: Specifying libraries and so on.
141 * Directory Options:: Where to find header files and libraries.
142 Where to find the compiler executable files.
143 * Spec Files:: How to pass switches to sub-processes.
144 * Target Options:: Running a cross-compiler, or an old version of GCC.
145 * Submodel Options:: Specifying minor hardware or convention variations,
146 such as 68010 vs 68020.
147 * Code Gen Options:: Specifying conventions for function calls, data layout
149 * Environment Variables:: Env vars that affect GCC.
150 * Precompiled Headers:: Compiling a header once, and using it many times.
151 * Running Protoize:: Automatically adding or removing function prototypes.
157 @section Option Summary
159 Here is a summary of all the options, grouped by type. Explanations are
160 in the following sections.
163 @item Overall Options
164 @xref{Overall Options,,Options Controlling the Kind of Output}.
165 @gccoptlist{-c -S -E -o @var{file} -combine -pipe -pass-exit-codes @gol
166 -x @var{language} -v -### --help@r{[}=@var{class}@r{]} --target-help @gol
167 --version @@@var{file}}
169 @item C Language Options
170 @xref{C Dialect Options,,Options Controlling C Dialect}.
171 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
172 -aux-info @var{filename} @gol
173 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
174 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
175 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
176 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
177 -fsigned-bitfields -fsigned-char @gol
178 -funsigned-bitfields -funsigned-char}
180 @item C++ Language Options
181 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
182 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
183 -fconserve-space -ffriend-injection @gol
184 -fno-elide-constructors @gol
185 -fno-enforce-eh-specs @gol
186 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
187 -fno-implicit-templates @gol
188 -fno-implicit-inline-templates @gol
189 -fno-implement-inlines -fms-extensions @gol
190 -fno-nonansi-builtins -fno-operator-names @gol
191 -fno-optional-diags -fpermissive @gol
192 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
193 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
194 -fno-default-inline -fvisibility-inlines-hidden @gol
195 -fvisibility-ms-compat @gol
196 -Wabi -Wctor-dtor-privacy @gol
197 -Wnon-virtual-dtor -Wreorder @gol
198 -Weffc++ -Wno-deprecated -Wstrict-null-sentinel @gol
199 -Wno-non-template-friend -Wold-style-cast @gol
200 -Woverloaded-virtual -Wno-pmf-conversions @gol
203 @item Objective-C and Objective-C++ Language Options
204 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
205 Objective-C and Objective-C++ Dialects}.
206 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
207 -fgnu-runtime -fnext-runtime @gol
208 -fno-nil-receivers @gol
209 -fobjc-call-cxx-cdtors @gol
210 -fobjc-direct-dispatch @gol
211 -fobjc-exceptions @gol
213 -freplace-objc-classes @gol
216 -Wassign-intercept @gol
217 -Wno-protocol -Wselector @gol
218 -Wstrict-selector-match @gol
219 -Wundeclared-selector}
221 @item Language Independent Options
222 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
223 @gccoptlist{-fmessage-length=@var{n} @gol
224 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
225 -fdiagnostics-show-option}
227 @item Warning Options
228 @xref{Warning Options,,Options to Request or Suppress Warnings}.
229 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
230 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
231 -Wno-attributes -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
232 -Wchar-subscripts -Wclobbered -Wcomment @gol
233 -Wconversion -Wcoverage-mismatch -Wno-deprecated-declarations @gol
234 -Wdisabled-optimization -Wno-div-by-zero @gol
235 -Wempty-body -Wno-endif-labels @gol
236 -Werror -Werror=* @gol
237 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
238 -Wno-format-extra-args -Wformat-nonliteral @gol
239 -Wformat-security -Wformat-y2k @gol
240 -Wframe-larger-than=@var{len} -Wignored-qualifiers @gol
241 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
242 -Wimport -Wno-import -Winit-self -Winline @gol
243 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
244 -Winvalid-pch -Wlarger-than=@var{len} -Wunsafe-loop-optimizations @gol
245 -Wlogical-op -Wlong-long @gol
246 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
247 -Wmissing-format-attribute -Wmissing-include-dirs @gol
248 -Wmissing-noreturn @gol
249 -Wno-multichar -Wnonnull -Wno-overflow @gol
250 -Woverlength-strings -Wpacked -Wpadded @gol
251 -Wparentheses -Wpointer-arith -Wno-pointer-to-int-cast @gol
252 -Wredundant-decls @gol
253 -Wreturn-type -Wsequence-point -Wshadow @gol
254 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
255 -Wstrict-aliasing -Wstrict-aliasing=n @gol
256 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
257 -Wswitch -Wswitch-default -Wswitch-enum @gol
258 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
259 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
260 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
261 -Wunused-value -Wunused-variable @gol
262 -Wvariadic-macros -Wvla @gol
263 -Wvolatile-register-var -Wwrite-strings}
265 @item C and Objective-C-only Warning Options
266 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
267 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
268 -Wold-style-declaration -Wold-style-definition @gol
269 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
270 -Wdeclaration-after-statement -Wpointer-sign}
272 @item Debugging Options
273 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
274 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
275 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
276 -fdump-noaddr -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
277 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
278 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
280 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
281 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
282 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
284 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
285 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
287 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
288 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
289 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
290 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
291 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-nrv -fdump-tree-vect @gol
293 -fdump-tree-sink @gol
294 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
295 -fdump-tree-salias @gol
296 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
297 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
298 -ftree-vectorizer-verbose=@var{n} @gol
299 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
300 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
301 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
302 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
303 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
304 -ftest-coverage -ftime-report -fvar-tracking @gol
305 -g -g@var{level} -gcoff -gdwarf-2 @gol
306 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
307 -fno-merge-debug-strings -fdebug-prefix-map=@var{old}=@var{new} @gol
308 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
309 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
310 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
311 -print-multi-directory -print-multi-lib @gol
312 -print-prog-name=@var{program} -print-search-dirs -Q @gol
313 -print-sysroot-headers-suffix @gol
316 @item Optimization Options
317 @xref{Optimize Options,,Options that Control Optimization}.
319 -falign-functions[=@var{n}] -falign-jumps[=@var{n}] @gol
320 -falign-labels[=@var{n}] -falign-loops[=@var{n}] -fassociative-math @gol
321 -fauto-inc-dec -fbranch-probabilities -fbranch-target-load-optimize @gol
322 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fcaller-saves @gol
323 -fcheck-data-deps -fcprop-registers -fcrossjumping -fcse-follow-jumps @gol
324 -fcse-skip-blocks -fcx-fortran-rules -fcx-limited-range @gol
325 -fdata-sections -fdce -fdce @gol
326 -fdelayed-branch -fdelete-null-pointer-checks -fdse -fdse @gol
327 -fearly-inlining -fexpensive-optimizations -ffast-math @gol
328 -ffinite-math-only -ffloat-store -fforward-propagate @gol
329 -ffunction-sections -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol
330 -fgcse-sm -fif-conversion -fif-conversion2 -finline-functions @gol
331 -finline-functions-called-once -finline-limit=@var{n} @gol
332 -finline-small-functions -fipa-cp -fipa-marix-reorg -fipa-pta @gol
333 -fipa-pure-const -fipa-reference -fipa-struct-reorg @gol
334 -fipa-type-escape -fivopts -fkeep-inline-functions -fkeep-static-consts @gol
335 -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol
336 -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol
337 -fmudflapir -fmudflapth -fno-branch-count-reg -fno-default-inline @gol
338 -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
339 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
340 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
341 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
342 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
343 -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
344 -fprofile-generate -fprofile-use -fprofile-values -freciprocal-math @gol
345 -fregmove -frename-registers -freorder-blocks @gol
346 -freorder-blocks-and-partition -freorder-functions @gol
347 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
348 -frounding-math -frtl-abstract-sequences -fsched2-use-superblocks @gol
349 -fsched2-use-traces -fsched-spec-load -fsched-spec-load-dangerous @gol
350 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
351 -fschedule-insns -fschedule-insns2 -fsection-anchors -fsee @gol
352 -fsignaling-nans -fsingle-precision-constant -fsplit-ivs-in-unroller @gol
353 -fsplit-wide-types -fstack-protector -fstack-protector-all @gol
354 -fstrict-aliasing -fstrict-overflow -fthread-jumps -ftracer -ftree-ccp @gol
355 -ftree-ch -ftree-copy-prop -ftree-copyrename -ftree-dce @gol
356 -ftree-dominator-opts -ftree-dse -ftree-fre -ftree-loop-im @gol
357 -ftree-loop-distribution @gol
358 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
359 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-reassoc -ftree-salias @gol
360 -ftree-sink -ftree-sra -ftree-store-ccp -ftree-ter @gol
361 -ftree-vect-loop-version -ftree-vectorize -ftree-vrp -funit-at-a-time @gol
362 -funroll-all-loops -funroll-loops -funsafe-loop-optimizations @gol
363 -funsafe-math-optimizations -funswitch-loops @gol
364 -fvariable-expansion-in-unroller -fvect-cost-model -fvpt -fweb @gol
366 --param @var{name}=@var{value}
367 -O -O0 -O1 -O2 -O3 -Os}
369 @item Preprocessor Options
370 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
371 @gccoptlist{-A@var{question}=@var{answer} @gol
372 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
373 -C -dD -dI -dM -dN @gol
374 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
375 -idirafter @var{dir} @gol
376 -include @var{file} -imacros @var{file} @gol
377 -iprefix @var{file} -iwithprefix @var{dir} @gol
378 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
379 -imultilib @var{dir} -isysroot @var{dir} @gol
380 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
381 -P -fworking-directory -remap @gol
382 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
383 -Xpreprocessor @var{option}}
385 @item Assembler Option
386 @xref{Assembler Options,,Passing Options to the Assembler}.
387 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
390 @xref{Link Options,,Options for Linking}.
391 @gccoptlist{@var{object-file-name} -l@var{library} @gol
392 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
393 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
394 -Wl,@var{option} -Xlinker @var{option} @gol
397 @item Directory Options
398 @xref{Directory Options,,Options for Directory Search}.
399 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
400 -specs=@var{file} -I- --sysroot=@var{dir}}
403 @c I wrote this xref this way to avoid overfull hbox. -- rms
404 @xref{Target Options}.
405 @gccoptlist{-V @var{version} -b @var{machine}}
407 @item Machine Dependent Options
408 @xref{Submodel Options,,Hardware Models and Configurations}.
409 @c This list is ordered alphanumerically by subsection name.
410 @c Try and put the significant identifier (CPU or system) first,
411 @c so users have a clue at guessing where the ones they want will be.
414 @gccoptlist{-EB -EL @gol
415 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
416 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
419 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
420 -mabi=@var{name} @gol
421 -mapcs-stack-check -mno-apcs-stack-check @gol
422 -mapcs-float -mno-apcs-float @gol
423 -mapcs-reentrant -mno-apcs-reentrant @gol
424 -msched-prolog -mno-sched-prolog @gol
425 -mlittle-endian -mbig-endian -mwords-little-endian @gol
426 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
427 -mthumb-interwork -mno-thumb-interwork @gol
428 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
429 -mstructure-size-boundary=@var{n} @gol
430 -mabort-on-noreturn @gol
431 -mlong-calls -mno-long-calls @gol
432 -msingle-pic-base -mno-single-pic-base @gol
433 -mpic-register=@var{reg} @gol
434 -mnop-fun-dllimport @gol
435 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
436 -mpoke-function-name @gol
438 -mtpcs-frame -mtpcs-leaf-frame @gol
439 -mcaller-super-interworking -mcallee-super-interworking @gol
443 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
444 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
446 @emph{Blackfin Options}
447 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
448 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
449 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
450 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
451 -mno-id-shared-library -mshared-library-id=@var{n} @gol
452 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
453 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
454 -mfast-fp -minline-plt}
457 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
458 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
459 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
460 -mstack-align -mdata-align -mconst-align @gol
461 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
462 -melf -maout -melinux -mlinux -sim -sim2 @gol
463 -mmul-bug-workaround -mno-mul-bug-workaround}
466 @gccoptlist{-mmac -mpush-args}
468 @emph{Darwin Options}
469 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
470 -arch_only -bind_at_load -bundle -bundle_loader @gol
471 -client_name -compatibility_version -current_version @gol
473 -dependency-file -dylib_file -dylinker_install_name @gol
474 -dynamic -dynamiclib -exported_symbols_list @gol
475 -filelist -flat_namespace -force_cpusubtype_ALL @gol
476 -force_flat_namespace -headerpad_max_install_names @gol
478 -image_base -init -install_name -keep_private_externs @gol
479 -multi_module -multiply_defined -multiply_defined_unused @gol
480 -noall_load -no_dead_strip_inits_and_terms @gol
481 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
482 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
483 -private_bundle -read_only_relocs -sectalign @gol
484 -sectobjectsymbols -whyload -seg1addr @gol
485 -sectcreate -sectobjectsymbols -sectorder @gol
486 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
487 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
488 -segprot -segs_read_only_addr -segs_read_write_addr @gol
489 -single_module -static -sub_library -sub_umbrella @gol
490 -twolevel_namespace -umbrella -undefined @gol
491 -unexported_symbols_list -weak_reference_mismatches @gol
492 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
493 -mkernel -mone-byte-bool}
495 @emph{DEC Alpha Options}
496 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
497 -mieee -mieee-with-inexact -mieee-conformant @gol
498 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
499 -mtrap-precision=@var{mode} -mbuild-constants @gol
500 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
501 -mbwx -mmax -mfix -mcix @gol
502 -mfloat-vax -mfloat-ieee @gol
503 -mexplicit-relocs -msmall-data -mlarge-data @gol
504 -msmall-text -mlarge-text @gol
505 -mmemory-latency=@var{time}}
507 @emph{DEC Alpha/VMS Options}
508 @gccoptlist{-mvms-return-codes}
511 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
512 -mhard-float -msoft-float @gol
513 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
514 -mdouble -mno-double @gol
515 -mmedia -mno-media -mmuladd -mno-muladd @gol
516 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
517 -mlinked-fp -mlong-calls -malign-labels @gol
518 -mlibrary-pic -macc-4 -macc-8 @gol
519 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
520 -moptimize-membar -mno-optimize-membar @gol
521 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
522 -mvliw-branch -mno-vliw-branch @gol
523 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
524 -mno-nested-cond-exec -mtomcat-stats @gol
528 @emph{GNU/Linux Options}
529 @gccoptlist{-muclibc}
531 @emph{H8/300 Options}
532 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
535 @gccoptlist{-march=@var{architecture-type} @gol
536 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
537 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
538 -mfixed-range=@var{register-range} @gol
539 -mjump-in-delay -mlinker-opt -mlong-calls @gol
540 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
541 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
542 -mno-jump-in-delay -mno-long-load-store @gol
543 -mno-portable-runtime -mno-soft-float @gol
544 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
545 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
546 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
547 -munix=@var{unix-std} -nolibdld -static -threads}
549 @emph{i386 and x86-64 Options}
550 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
551 -mfpmath=@var{unit} @gol
552 -masm=@var{dialect} -mno-fancy-math-387 @gol
553 -mno-fp-ret-in-387 -msoft-float @gol
554 -mno-wide-multiply -mrtd -malign-double @gol
555 -mpreferred-stack-boundary=@var{num} -mcx16 -msahf -mrecip @gol
556 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 @gol
557 -msse4a -m3dnow -mpopcnt -mabm -msse5 @gol
558 -mthreads -mno-align-stringops -minline-all-stringops @gol
559 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
560 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
561 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
562 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
563 -mcmodel=@var{code-model} @gol
564 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
565 -mfused-madd -mno-fused-madd}
568 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
569 -mvolatile-asm-stop -mregister-names -mno-sdata @gol
570 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
571 -minline-float-divide-max-throughput @gol
572 -minline-int-divide-min-latency @gol
573 -minline-int-divide-max-throughput @gol
574 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
575 -mno-dwarf2-asm -mearly-stop-bits @gol
576 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
577 -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64 @gol
578 -mno-sched-br-data-spec -msched-ar-data-spec -mno-sched-control-spec @gol
579 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
580 -msched-ldc -mno-sched-control-ldc -mno-sched-spec-verbose @gol
581 -mno-sched-prefer-non-data-spec-insns @gol
582 -mno-sched-prefer-non-control-spec-insns @gol
583 -mno-sched-count-spec-in-critical-path}
585 @emph{M32R/D Options}
586 @gccoptlist{-m32r2 -m32rx -m32r @gol
588 -malign-loops -mno-align-loops @gol
589 -missue-rate=@var{number} @gol
590 -mbranch-cost=@var{number} @gol
591 -mmodel=@var{code-size-model-type} @gol
592 -msdata=@var{sdata-type} @gol
593 -mno-flush-func -mflush-func=@var{name} @gol
594 -mno-flush-trap -mflush-trap=@var{number} @gol
598 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
600 @emph{M680x0 Options}
601 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
602 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
603 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
604 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
605 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
606 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
607 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
608 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
610 @emph{M68hc1x Options}
611 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
612 -mauto-incdec -minmax -mlong-calls -mshort @gol
613 -msoft-reg-count=@var{count}}
616 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
617 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
618 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
619 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
620 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
623 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
624 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
625 -mips16 -mno-mips16 -mflip-mips16 @gol
626 -minterlink-mips16 -mno-interlink-mips16 @gol
627 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
628 -mshared -mno-shared -mxgot -mno-xgot -mgp32 -mgp64 @gol
629 -mfp32 -mfp64 -mhard-float -msoft-float @gol
630 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
631 -msmartmips -mno-smartmips @gol
632 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
633 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
634 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
635 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
636 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
637 -membedded-data -mno-embedded-data @gol
638 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
639 -mcode-readable=@var{setting} @gol
640 -msplit-addresses -mno-split-addresses @gol
641 -mexplicit-relocs -mno-explicit-relocs @gol
642 -mcheck-zero-division -mno-check-zero-division @gol
643 -mdivide-traps -mdivide-breaks @gol
644 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
645 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
646 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
647 -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 -mno-fix-vr4130 @gol
648 -mfix-sb1 -mno-fix-sb1 @gol
649 -mflush-func=@var{func} -mno-flush-func @gol
650 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
651 -mfp-exceptions -mno-fp-exceptions @gol
652 -mvr4130-align -mno-vr4130-align}
655 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
656 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
657 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
658 -mno-base-addresses -msingle-exit -mno-single-exit}
660 @emph{MN10300 Options}
661 @gccoptlist{-mmult-bug -mno-mult-bug @gol
662 -mam33 -mno-am33 @gol
663 -mam33-2 -mno-am33-2 @gol
664 -mreturn-pointer-on-d0 @gol
668 @gccoptlist{-mno-crt0 -mbacc -msim @gol
669 -march=@var{cpu-type} }
671 @emph{PDP-11 Options}
672 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
673 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
674 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
675 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
676 -mbranch-expensive -mbranch-cheap @gol
677 -msplit -mno-split -munix-asm -mdec-asm}
679 @emph{PowerPC Options}
680 See RS/6000 and PowerPC Options.
682 @emph{RS/6000 and PowerPC Options}
683 @gccoptlist{-mcpu=@var{cpu-type} @gol
684 -mtune=@var{cpu-type} @gol
685 -mpower -mno-power -mpower2 -mno-power2 @gol
686 -mpowerpc -mpowerpc64 -mno-powerpc @gol
687 -maltivec -mno-altivec @gol
688 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
689 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
690 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol
691 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
692 -mnew-mnemonics -mold-mnemonics @gol
693 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
694 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
695 -malign-power -malign-natural @gol
696 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
697 -mstring -mno-string -mupdate -mno-update @gol
698 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
699 -mstrict-align -mno-strict-align -mrelocatable @gol
700 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
701 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
702 -mdynamic-no-pic -maltivec -mswdiv @gol
703 -mprioritize-restricted-insns=@var{priority} @gol
704 -msched-costly-dep=@var{dependence_type} @gol
705 -minsert-sched-nops=@var{scheme} @gol
706 -mcall-sysv -mcall-netbsd @gol
707 -maix-struct-return -msvr4-struct-return @gol
708 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
709 -misel -mno-isel @gol
710 -misel=yes -misel=no @gol
712 -mspe=yes -mspe=no @gol
714 -mvrsave -mno-vrsave @gol
715 -mmulhw -mno-mulhw @gol
716 -mdlmzb -mno-dlmzb @gol
717 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
718 -mprototype -mno-prototype @gol
719 -msim -mmvme -mads -myellowknife -memb -msdata @gol
720 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
722 @emph{S/390 and zSeries Options}
723 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
724 -mhard-float -msoft-float -mlong-double-64 -mlong-double-128 @gol
725 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
726 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
727 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
728 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
729 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
732 @gccoptlist{-meb -mel @gol
736 -mscore5 -mscore5u -mscore7 -mscore7d}
739 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
740 -m4-nofpu -m4-single-only -m4-single -m4 @gol
741 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
742 -m5-64media -m5-64media-nofpu @gol
743 -m5-32media -m5-32media-nofpu @gol
744 -m5-compact -m5-compact-nofpu @gol
745 -mb -ml -mdalign -mrelax @gol
746 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
747 -mieee -misize -minline-ic_invalidate -mpadstruct -mspace @gol
748 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
749 -mdivsi3_libfunc=@var{name} @gol
750 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
754 @gccoptlist{-mcpu=@var{cpu-type} @gol
755 -mtune=@var{cpu-type} @gol
756 -mcmodel=@var{code-model} @gol
757 -m32 -m64 -mapp-regs -mno-app-regs @gol
758 -mfaster-structs -mno-faster-structs @gol
759 -mfpu -mno-fpu -mhard-float -msoft-float @gol
760 -mhard-quad-float -msoft-quad-float @gol
761 -mimpure-text -mno-impure-text -mlittle-endian @gol
762 -mstack-bias -mno-stack-bias @gol
763 -munaligned-doubles -mno-unaligned-doubles @gol
764 -mv8plus -mno-v8plus -mvis -mno-vis
765 -threads -pthreads -pthread}
768 @gccoptlist{-mwarn-reloc -merror-reloc @gol
769 -msafe-dma -munsafe-dma @gol
771 -msmall-mem -mlarge-mem -mstdmain @gol
772 -mfixed-range=@var{register-range}}
774 @emph{System V Options}
775 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
778 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
779 -mprolog-function -mno-prolog-function -mspace @gol
780 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
781 -mapp-regs -mno-app-regs @gol
782 -mdisable-callt -mno-disable-callt @gol
788 @gccoptlist{-mg -mgnu -munix}
790 @emph{VxWorks Options}
791 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
792 -Xbind-lazy -Xbind-now}
794 @emph{x86-64 Options}
795 See i386 and x86-64 Options.
797 @emph{Xstormy16 Options}
800 @emph{Xtensa Options}
801 @gccoptlist{-mconst16 -mno-const16 @gol
802 -mfused-madd -mno-fused-madd @gol
803 -mtext-section-literals -mno-text-section-literals @gol
804 -mtarget-align -mno-target-align @gol
805 -mlongcalls -mno-longcalls}
807 @emph{zSeries Options}
808 See S/390 and zSeries Options.
810 @item Code Generation Options
811 @xref{Code Gen Options,,Options for Code Generation Conventions}.
812 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
813 -ffixed-@var{reg} -fexceptions @gol
814 -fnon-call-exceptions -funwind-tables @gol
815 -fasynchronous-unwind-tables @gol
816 -finhibit-size-directive -finstrument-functions @gol
817 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
818 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
819 -fno-common -fno-ident @gol
820 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
821 -fno-jump-tables @gol
822 -frecord-gcc-switches @gol
823 -freg-struct-return -fshort-enums @gol
824 -fshort-double -fshort-wchar @gol
825 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
826 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
827 -fno-stack-limit -fargument-alias -fargument-noalias @gol
828 -fargument-noalias-global -fargument-noalias-anything @gol
829 -fleading-underscore -ftls-model=@var{model} @gol
830 -ftrapv -fwrapv -fbounds-check @gol
835 * Overall Options:: Controlling the kind of output:
836 an executable, object files, assembler files,
837 or preprocessed source.
838 * C Dialect Options:: Controlling the variant of C language compiled.
839 * C++ Dialect Options:: Variations on C++.
840 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
842 * Language Independent Options:: Controlling how diagnostics should be
844 * Warning Options:: How picky should the compiler be?
845 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
846 * Optimize Options:: How much optimization?
847 * Preprocessor Options:: Controlling header files and macro definitions.
848 Also, getting dependency information for Make.
849 * Assembler Options:: Passing options to the assembler.
850 * Link Options:: Specifying libraries and so on.
851 * Directory Options:: Where to find header files and libraries.
852 Where to find the compiler executable files.
853 * Spec Files:: How to pass switches to sub-processes.
854 * Target Options:: Running a cross-compiler, or an old version of GCC.
857 @node Overall Options
858 @section Options Controlling the Kind of Output
860 Compilation can involve up to four stages: preprocessing, compilation
861 proper, assembly and linking, always in that order. GCC is capable of
862 preprocessing and compiling several files either into several
863 assembler input files, or into one assembler input file; then each
864 assembler input file produces an object file, and linking combines all
865 the object files (those newly compiled, and those specified as input)
866 into an executable file.
868 @cindex file name suffix
869 For any given input file, the file name suffix determines what kind of
874 C source code which must be preprocessed.
877 C source code which should not be preprocessed.
880 C++ source code which should not be preprocessed.
883 Objective-C source code. Note that you must link with the @file{libobjc}
884 library to make an Objective-C program work.
887 Objective-C source code which should not be preprocessed.
891 Objective-C++ source code. Note that you must link with the @file{libobjc}
892 library to make an Objective-C++ program work. Note that @samp{.M} refers
893 to a literal capital M@.
896 Objective-C++ source code which should not be preprocessed.
899 C, C++, Objective-C or Objective-C++ header file to be turned into a
904 @itemx @var{file}.cxx
905 @itemx @var{file}.cpp
906 @itemx @var{file}.CPP
907 @itemx @var{file}.c++
909 C++ source code which must be preprocessed. Note that in @samp{.cxx},
910 the last two letters must both be literally @samp{x}. Likewise,
911 @samp{.C} refers to a literal capital C@.
915 Objective-C++ source code which must be preprocessed.
918 Objective-C++ source code which should not be preprocessed.
923 @itemx @var{file}.hxx
924 @itemx @var{file}.hpp
925 @itemx @var{file}.HPP
926 @itemx @var{file}.h++
927 @itemx @var{file}.tcc
928 C++ header file to be turned into a precompiled header.
931 @itemx @var{file}.for
932 @itemx @var{file}.FOR
933 Fixed form Fortran source code which should not be preprocessed.
936 @itemx @var{file}.fpp
937 @itemx @var{file}.FPP
938 Fixed form Fortran source code which must be preprocessed (with the traditional
942 @itemx @var{file}.f95
943 Free form Fortran source code which should not be preprocessed.
946 @itemx @var{file}.F95
947 Free form Fortran source code which must be preprocessed (with the
948 traditional preprocessor).
950 @c FIXME: Descriptions of Java file types.
957 Ada source code file which contains a library unit declaration (a
958 declaration of a package, subprogram, or generic, or a generic
959 instantiation), or a library unit renaming declaration (a package,
960 generic, or subprogram renaming declaration). Such files are also
963 @itemx @var{file}.adb
964 Ada source code file containing a library unit body (a subprogram or
965 package body). Such files are also called @dfn{bodies}.
967 @c GCC also knows about some suffixes for languages not yet included:
979 Assembler code which must be preprocessed.
982 An object file to be fed straight into linking.
983 Any file name with no recognized suffix is treated this way.
987 You can specify the input language explicitly with the @option{-x} option:
990 @item -x @var{language}
991 Specify explicitly the @var{language} for the following input files
992 (rather than letting the compiler choose a default based on the file
993 name suffix). This option applies to all following input files until
994 the next @option{-x} option. Possible values for @var{language} are:
996 c c-header c-cpp-output
997 c++ c++-header c++-cpp-output
998 objective-c objective-c-header objective-c-cpp-output
999 objective-c++ objective-c++-header objective-c++-cpp-output
1000 assembler assembler-with-cpp
1008 Turn off any specification of a language, so that subsequent files are
1009 handled according to their file name suffixes (as they are if @option{-x}
1010 has not been used at all).
1012 @item -pass-exit-codes
1013 @opindex pass-exit-codes
1014 Normally the @command{gcc} program will exit with the code of 1 if any
1015 phase of the compiler returns a non-success return code. If you specify
1016 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1017 numerically highest error produced by any phase that returned an error
1018 indication. The C, C++, and Fortran frontends return 4, if an internal
1019 compiler error is encountered.
1022 If you only want some of the stages of compilation, you can use
1023 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1024 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1025 @command{gcc} is to stop. Note that some combinations (for example,
1026 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1031 Compile or assemble the source files, but do not link. The linking
1032 stage simply is not done. The ultimate output is in the form of an
1033 object file for each source file.
1035 By default, the object file name for a source file is made by replacing
1036 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1038 Unrecognized input files, not requiring compilation or assembly, are
1043 Stop after the stage of compilation proper; do not assemble. The output
1044 is in the form of an assembler code file for each non-assembler input
1047 By default, the assembler file name for a source file is made by
1048 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1050 Input files that don't require compilation are ignored.
1054 Stop after the preprocessing stage; do not run the compiler proper. The
1055 output is in the form of preprocessed source code, which is sent to the
1058 Input files which don't require preprocessing are ignored.
1060 @cindex output file option
1063 Place output in file @var{file}. This applies regardless to whatever
1064 sort of output is being produced, whether it be an executable file,
1065 an object file, an assembler file or preprocessed C code.
1067 If @option{-o} is not specified, the default is to put an executable
1068 file in @file{a.out}, the object file for
1069 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1070 assembler file in @file{@var{source}.s}, a precompiled header file in
1071 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1076 Print (on standard error output) the commands executed to run the stages
1077 of compilation. Also print the version number of the compiler driver
1078 program and of the preprocessor and the compiler proper.
1082 Like @option{-v} except the commands are not executed and all command
1083 arguments are quoted. This is useful for shell scripts to capture the
1084 driver-generated command lines.
1088 Use pipes rather than temporary files for communication between the
1089 various stages of compilation. This fails to work on some systems where
1090 the assembler is unable to read from a pipe; but the GNU assembler has
1095 If you are compiling multiple source files, this option tells the driver
1096 to pass all the source files to the compiler at once (for those
1097 languages for which the compiler can handle this). This will allow
1098 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1099 language for which this is supported is C@. If you pass source files for
1100 multiple languages to the driver, using this option, the driver will invoke
1101 the compiler(s) that support IMA once each, passing each compiler all the
1102 source files appropriate for it. For those languages that do not support
1103 IMA this option will be ignored, and the compiler will be invoked once for
1104 each source file in that language. If you use this option in conjunction
1105 with @option{-save-temps}, the compiler will generate multiple
1107 (one for each source file), but only one (combined) @file{.o} or
1112 Print (on the standard output) a description of the command line options
1113 understood by @command{gcc}. If the @option{-v} option is also specified
1114 then @option{--help} will also be passed on to the various processes
1115 invoked by @command{gcc}, so that they can display the command line options
1116 they accept. If the @option{-Wextra} option has also been specified
1117 (prior to the @option{--help} option), then command line options which
1118 have no documentation associated with them will also be displayed.
1121 @opindex target-help
1122 Print (on the standard output) a description of target-specific command
1123 line options for each tool. For some targets extra target-specific
1124 information may also be printed.
1126 @item --help=@var{class}@r{[},@var{qualifier}@r{]}
1127 Print (on the standard output) a description of the command line
1128 options understood by the compiler that fit into a specific class.
1129 The class can be one of @samp{optimizers}, @samp{warnings}, @samp{target},
1130 @samp{params}, or @var{language}:
1133 @item @samp{optimizers}
1134 This will display all of the optimization options supported by the
1137 @item @samp{warnings}
1138 This will display all of the options controlling warning messages
1139 produced by the compiler.
1142 This will display target-specific options. Unlike the
1143 @option{--target-help} option however, target-specific options of the
1144 linker and assembler will not be displayed. This is because those
1145 tools do not currently support the extended @option{--help=} syntax.
1148 This will display the values recognized by the @option{--param}
1151 @item @var{language}
1152 This will display the options supported for @var{language}, where
1153 @var{language} is the name of one of the languages supported in this
1157 This will display the options that are common to all languages.
1160 It is possible to further refine the output of the @option{--help=}
1161 option by adding a comma separated list of qualifiers after the
1162 class. These can be any from the following list:
1165 @item @samp{undocumented}
1166 Display only those options which are undocumented.
1169 Display options which take an argument that appears after an equal
1170 sign in the same continuous piece of text, such as:
1171 @samp{--help=target}.
1173 @item @samp{separate}
1174 Display options which take an argument that appears as a separate word
1175 following the original option, such as: @samp{-o output-file}.
1178 Thus for example to display all the undocumented target-specific
1179 switches supported by the compiler the following can be used:
1182 --help=target,undocumented
1185 The sense of a qualifier can be inverted by prefixing it with the
1186 @var{^} character, so for example to display all binary warning
1187 options (i.e., ones that are either on or off and that do not take an
1188 argument), which have a description the following can be used:
1191 --help=warnings,^joined,^undocumented
1194 A class can also be used as a qualifier, although this usually
1195 restricts the output by so much that there is nothing to display. One
1196 case where it does work however is when one of the classes is
1197 @var{target}. So for example to display all the target-specific
1198 optimization options the following can be used:
1201 --help=target,optimizers
1204 The @option{--help=} option can be repeated on the command line. Each
1205 successive use will display its requested class of options, skipping
1206 those that have already been displayed.
1208 If the @option{-Q} option appears on the command line before the
1209 @option{--help=} option, then the descriptive text displayed by
1210 @option{--help=} is changed. Instead of describing the displayed
1211 options, an indication is given as to whether the option is enabled,
1212 disabled or set to a specific value (assuming that the compiler
1213 knows this at the point where the @option{--help=} option is used).
1215 Here is a truncated example from the ARM port of @command{gcc}:
1218 % gcc -Q -mabi=2 --help=target -c
1219 The following options are target specific:
1221 -mabort-on-noreturn [disabled]
1225 The output is sensitive to the effects of previous command line
1226 options, so for example it is possible to find out which optimizations
1227 are enabled at @option{-O2} by using:
1230 -O2 --help=optimizers
1233 Alternatively you can discover which binary optimizations are enabled
1234 by @option{-O3} by using:
1237 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1238 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1239 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1244 Display the version number and copyrights of the invoked GCC@.
1246 @include @value{srcdir}/../libiberty/at-file.texi
1250 @section Compiling C++ Programs
1252 @cindex suffixes for C++ source
1253 @cindex C++ source file suffixes
1254 C++ source files conventionally use one of the suffixes @samp{.C},
1255 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1256 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1257 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1258 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1259 files with these names and compiles them as C++ programs even if you
1260 call the compiler the same way as for compiling C programs (usually
1261 with the name @command{gcc}).
1265 However, the use of @command{gcc} does not add the C++ library.
1266 @command{g++} is a program that calls GCC and treats @samp{.c},
1267 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1268 files unless @option{-x} is used, and automatically specifies linking
1269 against the C++ library. This program is also useful when
1270 precompiling a C header file with a @samp{.h} extension for use in C++
1271 compilations. On many systems, @command{g++} is also installed with
1272 the name @command{c++}.
1274 @cindex invoking @command{g++}
1275 When you compile C++ programs, you may specify many of the same
1276 command-line options that you use for compiling programs in any
1277 language; or command-line options meaningful for C and related
1278 languages; or options that are meaningful only for C++ programs.
1279 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1280 explanations of options for languages related to C@.
1281 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1282 explanations of options that are meaningful only for C++ programs.
1284 @node C Dialect Options
1285 @section Options Controlling C Dialect
1286 @cindex dialect options
1287 @cindex language dialect options
1288 @cindex options, dialect
1290 The following options control the dialect of C (or languages derived
1291 from C, such as C++, Objective-C and Objective-C++) that the compiler
1295 @cindex ANSI support
1299 In C mode, this is equivalent to @samp{-std=c89}. In C++ mode, it is
1300 equivalent to @samp{-std=c++98}.
1302 This turns off certain features of GCC that are incompatible with ISO
1303 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1304 such as the @code{asm} and @code{typeof} keywords, and
1305 predefined macros such as @code{unix} and @code{vax} that identify the
1306 type of system you are using. It also enables the undesirable and
1307 rarely used ISO trigraph feature. For the C compiler,
1308 it disables recognition of C++ style @samp{//} comments as well as
1309 the @code{inline} keyword.
1311 The alternate keywords @code{__asm__}, @code{__extension__},
1312 @code{__inline__} and @code{__typeof__} continue to work despite
1313 @option{-ansi}. You would not want to use them in an ISO C program, of
1314 course, but it is useful to put them in header files that might be included
1315 in compilations done with @option{-ansi}. Alternate predefined macros
1316 such as @code{__unix__} and @code{__vax__} are also available, with or
1317 without @option{-ansi}.
1319 The @option{-ansi} option does not cause non-ISO programs to be
1320 rejected gratuitously. For that, @option{-pedantic} is required in
1321 addition to @option{-ansi}. @xref{Warning Options}.
1323 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1324 option is used. Some header files may notice this macro and refrain
1325 from declaring certain functions or defining certain macros that the
1326 ISO standard doesn't call for; this is to avoid interfering with any
1327 programs that might use these names for other things.
1329 Functions that would normally be built in but do not have semantics
1330 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1331 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1332 built-in functions provided by GCC}, for details of the functions
1337 Determine the language standard. @xref{Standards,,Language Standards
1338 Supported by GCC}, for details of these standard versions. This option
1339 is currently only supported when compiling C or C++.
1341 The compiler can accept several base standards, such as @samp{c89} or
1342 @samp{c++98}, and GNU dialects of those standards, such as
1343 @samp{gnu89} or @samp{gnu++98}. By specifing a base standard, the
1344 compiler will accept all programs following that standard and those
1345 using GNU extensions that do not contradict it. For example,
1346 @samp{-std=c89} turns off certain features of GCC that are
1347 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1348 keywords, but not other GNU extensions that do not have a meaning in
1349 ISO C90, such as omitting the middle term of a @code{?:}
1350 expression. On the other hand, by specifing a GNU dialect of a
1351 standard, all features the compiler support are enabled, even when
1352 those features change the meaning of the base standard and some
1353 strict-conforming programs may be rejected. The particular standard
1354 is used by @option{-pedantic} to identify which features are GNU
1355 extensions given that version of the standard. For example
1356 @samp{-std=gnu89 -pedantic} would warn about C++ style @samp{//}
1357 comments, while @samp{-std=gnu99 -pedantic} would not.
1359 A value for this option must be provided; possible values are
1364 Support all ISO C90 programs (certain GNU extensions that conflict
1365 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1367 @item iso9899:199409
1368 ISO C90 as modified in amendment 1.
1374 ISO C99. Note that this standard is not yet fully supported; see
1375 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1376 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1379 GNU dialect of ISO C90 (including some C99 features). This
1380 is the default for C code.
1384 GNU dialect of ISO C99. When ISO C99 is fully implemented in GCC,
1385 this will become the default. The name @samp{gnu9x} is deprecated.
1388 The 1998 ISO C++ standard plus amendments. Same as @option{-ansi} for
1392 GNU dialect of @option{-std=c++98}. This is the default for
1396 The working draft of the upcoming ISO C++0x standard. This option
1397 enables experimental features that are likely to be included in
1398 C++0x. The working draft is constantly changing, and any feature that is
1399 enabled by this flag may be removed from future versions of GCC if it is
1400 not part of the C++0x standard.
1403 GNU dialect of @option{-std=c++0x}. This option enables
1404 experimental features that may be removed in future versions of GCC.
1407 @item -fgnu89-inline
1408 @opindex fgnu89-inline
1409 The option @option{-fgnu89-inline} tells GCC to use the traditional
1410 GNU semantics for @code{inline} functions when in C99 mode.
1411 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1412 is accepted and ignored by GCC versions 4.1.3 up to but not including
1413 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1414 C99 mode. Using this option is roughly equivalent to adding the
1415 @code{gnu_inline} function attribute to all inline functions
1416 (@pxref{Function Attributes}).
1418 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1419 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1420 specifies the default behavior). This option was first supported in
1421 GCC 4.3. This option is not supported in C89 or gnu89 mode.
1423 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1424 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1425 in effect for @code{inline} functions. @xref{Common Predefined
1426 Macros,,,cpp,The C Preprocessor}.
1428 @item -aux-info @var{filename}
1430 Output to the given filename prototyped declarations for all functions
1431 declared and/or defined in a translation unit, including those in header
1432 files. This option is silently ignored in any language other than C@.
1434 Besides declarations, the file indicates, in comments, the origin of
1435 each declaration (source file and line), whether the declaration was
1436 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1437 @samp{O} for old, respectively, in the first character after the line
1438 number and the colon), and whether it came from a declaration or a
1439 definition (@samp{C} or @samp{F}, respectively, in the following
1440 character). In the case of function definitions, a K&R-style list of
1441 arguments followed by their declarations is also provided, inside
1442 comments, after the declaration.
1446 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1447 keyword, so that code can use these words as identifiers. You can use
1448 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1449 instead. @option{-ansi} implies @option{-fno-asm}.
1451 In C++, this switch only affects the @code{typeof} keyword, since
1452 @code{asm} and @code{inline} are standard keywords. You may want to
1453 use the @option{-fno-gnu-keywords} flag instead, which has the same
1454 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1455 switch only affects the @code{asm} and @code{typeof} keywords, since
1456 @code{inline} is a standard keyword in ISO C99.
1459 @itemx -fno-builtin-@var{function}
1460 @opindex fno-builtin
1461 @cindex built-in functions
1462 Don't recognize built-in functions that do not begin with
1463 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1464 functions provided by GCC}, for details of the functions affected,
1465 including those which are not built-in functions when @option{-ansi} or
1466 @option{-std} options for strict ISO C conformance are used because they
1467 do not have an ISO standard meaning.
1469 GCC normally generates special code to handle certain built-in functions
1470 more efficiently; for instance, calls to @code{alloca} may become single
1471 instructions that adjust the stack directly, and calls to @code{memcpy}
1472 may become inline copy loops. The resulting code is often both smaller
1473 and faster, but since the function calls no longer appear as such, you
1474 cannot set a breakpoint on those calls, nor can you change the behavior
1475 of the functions by linking with a different library. In addition,
1476 when a function is recognized as a built-in function, GCC may use
1477 information about that function to warn about problems with calls to
1478 that function, or to generate more efficient code, even if the
1479 resulting code still contains calls to that function. For example,
1480 warnings are given with @option{-Wformat} for bad calls to
1481 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1482 known not to modify global memory.
1484 With the @option{-fno-builtin-@var{function}} option
1485 only the built-in function @var{function} is
1486 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1487 function is named this is not built-in in this version of GCC, this
1488 option is ignored. There is no corresponding
1489 @option{-fbuiltin-@var{function}} option; if you wish to enable
1490 built-in functions selectively when using @option{-fno-builtin} or
1491 @option{-ffreestanding}, you may define macros such as:
1494 #define abs(n) __builtin_abs ((n))
1495 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1500 @cindex hosted environment
1502 Assert that compilation takes place in a hosted environment. This implies
1503 @option{-fbuiltin}. A hosted environment is one in which the
1504 entire standard library is available, and in which @code{main} has a return
1505 type of @code{int}. Examples are nearly everything except a kernel.
1506 This is equivalent to @option{-fno-freestanding}.
1508 @item -ffreestanding
1509 @opindex ffreestanding
1510 @cindex hosted environment
1512 Assert that compilation takes place in a freestanding environment. This
1513 implies @option{-fno-builtin}. A freestanding environment
1514 is one in which the standard library may not exist, and program startup may
1515 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1516 This is equivalent to @option{-fno-hosted}.
1518 @xref{Standards,,Language Standards Supported by GCC}, for details of
1519 freestanding and hosted environments.
1523 @cindex openmp parallel
1524 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1525 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1526 compiler generates parallel code according to the OpenMP Application
1527 Program Interface v2.5 @w{@uref{http://www.openmp.org/}}. This option
1528 implies @option{-pthread}, and thus is only supported on targets that
1529 have support for @option{-pthread}.
1531 @item -fms-extensions
1532 @opindex fms-extensions
1533 Accept some non-standard constructs used in Microsoft header files.
1535 Some cases of unnamed fields in structures and unions are only
1536 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1537 fields within structs/unions}, for details.
1541 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1542 options for strict ISO C conformance) implies @option{-trigraphs}.
1544 @item -no-integrated-cpp
1545 @opindex no-integrated-cpp
1546 Performs a compilation in two passes: preprocessing and compiling. This
1547 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1548 @option{-B} option. The user supplied compilation step can then add in
1549 an additional preprocessing step after normal preprocessing but before
1550 compiling. The default is to use the integrated cpp (internal cpp)
1552 The semantics of this option will change if "cc1", "cc1plus", and
1553 "cc1obj" are merged.
1555 @cindex traditional C language
1556 @cindex C language, traditional
1558 @itemx -traditional-cpp
1559 @opindex traditional-cpp
1560 @opindex traditional
1561 Formerly, these options caused GCC to attempt to emulate a pre-standard
1562 C compiler. They are now only supported with the @option{-E} switch.
1563 The preprocessor continues to support a pre-standard mode. See the GNU
1564 CPP manual for details.
1566 @item -fcond-mismatch
1567 @opindex fcond-mismatch
1568 Allow conditional expressions with mismatched types in the second and
1569 third arguments. The value of such an expression is void. This option
1570 is not supported for C++.
1572 @item -flax-vector-conversions
1573 @opindex flax-vector-conversions
1574 Allow implicit conversions between vectors with differing numbers of
1575 elements and/or incompatible element types. This option should not be
1578 @item -funsigned-char
1579 @opindex funsigned-char
1580 Let the type @code{char} be unsigned, like @code{unsigned char}.
1582 Each kind of machine has a default for what @code{char} should
1583 be. It is either like @code{unsigned char} by default or like
1584 @code{signed char} by default.
1586 Ideally, a portable program should always use @code{signed char} or
1587 @code{unsigned char} when it depends on the signedness of an object.
1588 But many programs have been written to use plain @code{char} and
1589 expect it to be signed, or expect it to be unsigned, depending on the
1590 machines they were written for. This option, and its inverse, let you
1591 make such a program work with the opposite default.
1593 The type @code{char} is always a distinct type from each of
1594 @code{signed char} or @code{unsigned char}, even though its behavior
1595 is always just like one of those two.
1598 @opindex fsigned-char
1599 Let the type @code{char} be signed, like @code{signed char}.
1601 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1602 the negative form of @option{-funsigned-char}. Likewise, the option
1603 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1605 @item -fsigned-bitfields
1606 @itemx -funsigned-bitfields
1607 @itemx -fno-signed-bitfields
1608 @itemx -fno-unsigned-bitfields
1609 @opindex fsigned-bitfields
1610 @opindex funsigned-bitfields
1611 @opindex fno-signed-bitfields
1612 @opindex fno-unsigned-bitfields
1613 These options control whether a bit-field is signed or unsigned, when the
1614 declaration does not use either @code{signed} or @code{unsigned}. By
1615 default, such a bit-field is signed, because this is consistent: the
1616 basic integer types such as @code{int} are signed types.
1619 @node C++ Dialect Options
1620 @section Options Controlling C++ Dialect
1622 @cindex compiler options, C++
1623 @cindex C++ options, command line
1624 @cindex options, C++
1625 This section describes the command-line options that are only meaningful
1626 for C++ programs; but you can also use most of the GNU compiler options
1627 regardless of what language your program is in. For example, you
1628 might compile a file @code{firstClass.C} like this:
1631 g++ -g -frepo -O -c firstClass.C
1635 In this example, only @option{-frepo} is an option meant
1636 only for C++ programs; you can use the other options with any
1637 language supported by GCC@.
1639 Here is a list of options that are @emph{only} for compiling C++ programs:
1643 @item -fabi-version=@var{n}
1644 @opindex fabi-version
1645 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1646 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1647 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1648 the version that conforms most closely to the C++ ABI specification.
1649 Therefore, the ABI obtained using version 0 will change as ABI bugs
1652 The default is version 2.
1654 @item -fno-access-control
1655 @opindex fno-access-control
1656 Turn off all access checking. This switch is mainly useful for working
1657 around bugs in the access control code.
1661 Check that the pointer returned by @code{operator new} is non-null
1662 before attempting to modify the storage allocated. This check is
1663 normally unnecessary because the C++ standard specifies that
1664 @code{operator new} will only return @code{0} if it is declared
1665 @samp{throw()}, in which case the compiler will always check the
1666 return value even without this option. In all other cases, when
1667 @code{operator new} has a non-empty exception specification, memory
1668 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1669 @samp{new (nothrow)}.
1671 @item -fconserve-space
1672 @opindex fconserve-space
1673 Put uninitialized or runtime-initialized global variables into the
1674 common segment, as C does. This saves space in the executable at the
1675 cost of not diagnosing duplicate definitions. If you compile with this
1676 flag and your program mysteriously crashes after @code{main()} has
1677 completed, you may have an object that is being destroyed twice because
1678 two definitions were merged.
1680 This option is no longer useful on most targets, now that support has
1681 been added for putting variables into BSS without making them common.
1683 @item -ffriend-injection
1684 @opindex ffriend-injection
1685 Inject friend functions into the enclosing namespace, so that they are
1686 visible outside the scope of the class in which they are declared.
1687 Friend functions were documented to work this way in the old Annotated
1688 C++ Reference Manual, and versions of G++ before 4.1 always worked
1689 that way. However, in ISO C++ a friend function which is not declared
1690 in an enclosing scope can only be found using argument dependent
1691 lookup. This option causes friends to be injected as they were in
1694 This option is for compatibility, and may be removed in a future
1697 @item -fno-elide-constructors
1698 @opindex fno-elide-constructors
1699 The C++ standard allows an implementation to omit creating a temporary
1700 which is only used to initialize another object of the same type.
1701 Specifying this option disables that optimization, and forces G++ to
1702 call the copy constructor in all cases.
1704 @item -fno-enforce-eh-specs
1705 @opindex fno-enforce-eh-specs
1706 Don't generate code to check for violation of exception specifications
1707 at runtime. This option violates the C++ standard, but may be useful
1708 for reducing code size in production builds, much like defining
1709 @samp{NDEBUG}. This does not give user code permission to throw
1710 exceptions in violation of the exception specifications; the compiler
1711 will still optimize based on the specifications, so throwing an
1712 unexpected exception will result in undefined behavior.
1715 @itemx -fno-for-scope
1717 @opindex fno-for-scope
1718 If @option{-ffor-scope} is specified, the scope of variables declared in
1719 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1720 as specified by the C++ standard.
1721 If @option{-fno-for-scope} is specified, the scope of variables declared in
1722 a @i{for-init-statement} extends to the end of the enclosing scope,
1723 as was the case in old versions of G++, and other (traditional)
1724 implementations of C++.
1726 The default if neither flag is given to follow the standard,
1727 but to allow and give a warning for old-style code that would
1728 otherwise be invalid, or have different behavior.
1730 @item -fno-gnu-keywords
1731 @opindex fno-gnu-keywords
1732 Do not recognize @code{typeof} as a keyword, so that code can use this
1733 word as an identifier. You can use the keyword @code{__typeof__} instead.
1734 @option{-ansi} implies @option{-fno-gnu-keywords}.
1736 @item -fno-implicit-templates
1737 @opindex fno-implicit-templates
1738 Never emit code for non-inline templates which are instantiated
1739 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1740 @xref{Template Instantiation}, for more information.
1742 @item -fno-implicit-inline-templates
1743 @opindex fno-implicit-inline-templates
1744 Don't emit code for implicit instantiations of inline templates, either.
1745 The default is to handle inlines differently so that compiles with and
1746 without optimization will need the same set of explicit instantiations.
1748 @item -fno-implement-inlines
1749 @opindex fno-implement-inlines
1750 To save space, do not emit out-of-line copies of inline functions
1751 controlled by @samp{#pragma implementation}. This will cause linker
1752 errors if these functions are not inlined everywhere they are called.
1754 @item -fms-extensions
1755 @opindex fms-extensions
1756 Disable pedantic warnings about constructs used in MFC, such as implicit
1757 int and getting a pointer to member function via non-standard syntax.
1759 @item -fno-nonansi-builtins
1760 @opindex fno-nonansi-builtins
1761 Disable built-in declarations of functions that are not mandated by
1762 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1763 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1765 @item -fno-operator-names
1766 @opindex fno-operator-names
1767 Do not treat the operator name keywords @code{and}, @code{bitand},
1768 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1769 synonyms as keywords.
1771 @item -fno-optional-diags
1772 @opindex fno-optional-diags
1773 Disable diagnostics that the standard says a compiler does not need to
1774 issue. Currently, the only such diagnostic issued by G++ is the one for
1775 a name having multiple meanings within a class.
1778 @opindex fpermissive
1779 Downgrade some diagnostics about nonconformant code from errors to
1780 warnings. Thus, using @option{-fpermissive} will allow some
1781 nonconforming code to compile.
1785 Enable automatic template instantiation at link time. This option also
1786 implies @option{-fno-implicit-templates}. @xref{Template
1787 Instantiation}, for more information.
1791 Disable generation of information about every class with virtual
1792 functions for use by the C++ runtime type identification features
1793 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1794 of the language, you can save some space by using this flag. Note that
1795 exception handling uses the same information, but it will generate it as
1796 needed. The @samp{dynamic_cast} operator can still be used for casts that
1797 do not require runtime type information, i.e.@: casts to @code{void *} or to
1798 unambiguous base classes.
1802 Emit statistics about front-end processing at the end of the compilation.
1803 This information is generally only useful to the G++ development team.
1805 @item -ftemplate-depth-@var{n}
1806 @opindex ftemplate-depth
1807 Set the maximum instantiation depth for template classes to @var{n}.
1808 A limit on the template instantiation depth is needed to detect
1809 endless recursions during template class instantiation. ANSI/ISO C++
1810 conforming programs must not rely on a maximum depth greater than 17.
1812 @item -fno-threadsafe-statics
1813 @opindex fno-threadsafe-statics
1814 Do not emit the extra code to use the routines specified in the C++
1815 ABI for thread-safe initialization of local statics. You can use this
1816 option to reduce code size slightly in code that doesn't need to be
1819 @item -fuse-cxa-atexit
1820 @opindex fuse-cxa-atexit
1821 Register destructors for objects with static storage duration with the
1822 @code{__cxa_atexit} function rather than the @code{atexit} function.
1823 This option is required for fully standards-compliant handling of static
1824 destructors, but will only work if your C library supports
1825 @code{__cxa_atexit}.
1827 @item -fno-use-cxa-get-exception-ptr
1828 @opindex fno-use-cxa-get-exception-ptr
1829 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1830 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1831 if the runtime routine is not available.
1833 @item -fvisibility-inlines-hidden
1834 @opindex fvisibility-inlines-hidden
1835 This switch declares that the user does not attempt to compare
1836 pointers to inline methods where the addresses of the two functions
1837 were taken in different shared objects.
1839 The effect of this is that GCC may, effectively, mark inline methods with
1840 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1841 appear in the export table of a DSO and do not require a PLT indirection
1842 when used within the DSO@. Enabling this option can have a dramatic effect
1843 on load and link times of a DSO as it massively reduces the size of the
1844 dynamic export table when the library makes heavy use of templates.
1846 The behavior of this switch is not quite the same as marking the
1847 methods as hidden directly, because it does not affect static variables
1848 local to the function or cause the compiler to deduce that
1849 the function is defined in only one shared object.
1851 You may mark a method as having a visibility explicitly to negate the
1852 effect of the switch for that method. For example, if you do want to
1853 compare pointers to a particular inline method, you might mark it as
1854 having default visibility. Marking the enclosing class with explicit
1855 visibility will have no effect.
1857 Explicitly instantiated inline methods are unaffected by this option
1858 as their linkage might otherwise cross a shared library boundary.
1859 @xref{Template Instantiation}.
1861 @item -fvisibility-ms-compat
1862 @opindex fvisibility-ms-compat
1863 This flag attempts to use visibility settings to make GCC's C++
1864 linkage model compatible with that of Microsoft Visual Studio.
1866 The flag makes these changes to GCC's linkage model:
1870 It sets the default visibility to @code{hidden}, like
1871 @option{-fvisibility=hidden}.
1874 Types, but not their members, are not hidden by default.
1877 The One Definition Rule is relaxed for types without explicit
1878 visibility specifications which are defined in more than one different
1879 shared object: those declarations are permitted if they would have
1880 been permitted when this option was not used.
1883 In new code it is better to use @option{-fvisibility=hidden} and
1884 export those classes which are intended to be externally visible.
1885 Unfortunately it is possible for code to rely, perhaps accidentally,
1886 on the Visual Studio behavior.
1888 Among the consequences of these changes are that static data members
1889 of the same type with the same name but defined in different shared
1890 objects will be different, so changing one will not change the other;
1891 and that pointers to function members defined in different shared
1892 objects may not compare equal. When this flag is given, it is a
1893 violation of the ODR to define types with the same name differently.
1897 Do not use weak symbol support, even if it is provided by the linker.
1898 By default, G++ will use weak symbols if they are available. This
1899 option exists only for testing, and should not be used by end-users;
1900 it will result in inferior code and has no benefits. This option may
1901 be removed in a future release of G++.
1905 Do not search for header files in the standard directories specific to
1906 C++, but do still search the other standard directories. (This option
1907 is used when building the C++ library.)
1910 In addition, these optimization, warning, and code generation options
1911 have meanings only for C++ programs:
1914 @item -fno-default-inline
1915 @opindex fno-default-inline
1916 Do not assume @samp{inline} for functions defined inside a class scope.
1917 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1918 functions will have linkage like inline functions; they just won't be
1921 @item -Wabi @r{(C++ and Objective-C++ only)}
1924 Warn when G++ generates code that is probably not compatible with the
1925 vendor-neutral C++ ABI@. Although an effort has been made to warn about
1926 all such cases, there are probably some cases that are not warned about,
1927 even though G++ is generating incompatible code. There may also be
1928 cases where warnings are emitted even though the code that is generated
1931 You should rewrite your code to avoid these warnings if you are
1932 concerned about the fact that code generated by G++ may not be binary
1933 compatible with code generated by other compilers.
1935 The known incompatibilities at this point include:
1940 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1941 pack data into the same byte as a base class. For example:
1944 struct A @{ virtual void f(); int f1 : 1; @};
1945 struct B : public A @{ int f2 : 1; @};
1949 In this case, G++ will place @code{B::f2} into the same byte
1950 as@code{A::f1}; other compilers will not. You can avoid this problem
1951 by explicitly padding @code{A} so that its size is a multiple of the
1952 byte size on your platform; that will cause G++ and other compilers to
1953 layout @code{B} identically.
1956 Incorrect handling of tail-padding for virtual bases. G++ does not use
1957 tail padding when laying out virtual bases. For example:
1960 struct A @{ virtual void f(); char c1; @};
1961 struct B @{ B(); char c2; @};
1962 struct C : public A, public virtual B @{@};
1966 In this case, G++ will not place @code{B} into the tail-padding for
1967 @code{A}; other compilers will. You can avoid this problem by
1968 explicitly padding @code{A} so that its size is a multiple of its
1969 alignment (ignoring virtual base classes); that will cause G++ and other
1970 compilers to layout @code{C} identically.
1973 Incorrect handling of bit-fields with declared widths greater than that
1974 of their underlying types, when the bit-fields appear in a union. For
1978 union U @{ int i : 4096; @};
1982 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1983 union too small by the number of bits in an @code{int}.
1986 Empty classes can be placed at incorrect offsets. For example:
1996 struct C : public B, public A @{@};
2000 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
2001 it should be placed at offset zero. G++ mistakenly believes that the
2002 @code{A} data member of @code{B} is already at offset zero.
2005 Names of template functions whose types involve @code{typename} or
2006 template template parameters can be mangled incorrectly.
2009 template <typename Q>
2010 void f(typename Q::X) @{@}
2012 template <template <typename> class Q>
2013 void f(typename Q<int>::X) @{@}
2017 Instantiations of these templates may be mangled incorrectly.
2021 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2022 @opindex Wctor-dtor-privacy
2023 @opindex Wno-ctor-dtor-privacy
2024 Warn when a class seems unusable because all the constructors or
2025 destructors in that class are private, and it has neither friends nor
2026 public static member functions.
2028 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2029 @opindex Wnon-virtual-dtor
2030 @opindex Wno-non-virtual-dtor
2031 Warn when a class has virtual functions and accessible non-virtual
2032 destructor, in which case it would be possible but unsafe to delete
2033 an instance of a derived class through a pointer to the base class.
2034 This warning is also enabled if -Weffc++ is specified.
2036 @item -Wreorder @r{(C++ and Objective-C++ only)}
2038 @opindex Wno-reorder
2039 @cindex reordering, warning
2040 @cindex warning for reordering of member initializers
2041 Warn when the order of member initializers given in the code does not
2042 match the order in which they must be executed. For instance:
2048 A(): j (0), i (1) @{ @}
2052 The compiler will rearrange the member initializers for @samp{i}
2053 and @samp{j} to match the declaration order of the members, emitting
2054 a warning to that effect. This warning is enabled by @option{-Wall}.
2057 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2060 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2063 Warn about violations of the following style guidelines from Scott Meyers'
2064 @cite{Effective C++} book:
2068 Item 11: Define a copy constructor and an assignment operator for classes
2069 with dynamically allocated memory.
2072 Item 12: Prefer initialization to assignment in constructors.
2075 Item 14: Make destructors virtual in base classes.
2078 Item 15: Have @code{operator=} return a reference to @code{*this}.
2081 Item 23: Don't try to return a reference when you must return an object.
2085 Also warn about violations of the following style guidelines from
2086 Scott Meyers' @cite{More Effective C++} book:
2090 Item 6: Distinguish between prefix and postfix forms of increment and
2091 decrement operators.
2094 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2098 When selecting this option, be aware that the standard library
2099 headers do not obey all of these guidelines; use @samp{grep -v}
2100 to filter out those warnings.
2102 @item -Wno-deprecated @r{(C++ and Objective-C++ only)}
2103 @opindex Wno-deprecated
2104 @opindex Wdeprecated
2105 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
2107 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2108 @opindex Wstrict-null-sentinel
2109 @opindex Wno-strict-null-sentinel
2110 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2111 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2112 to @code{__null}. Although it is a null pointer constant not a null pointer,
2113 it is guaranteed to of the same size as a pointer. But this use is
2114 not portable across different compilers.
2116 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2117 @opindex Wno-non-template-friend
2118 @opindex Wnon-template-friend
2119 Disable warnings when non-templatized friend functions are declared
2120 within a template. Since the advent of explicit template specification
2121 support in G++, if the name of the friend is an unqualified-id (i.e.,
2122 @samp{friend foo(int)}), the C++ language specification demands that the
2123 friend declare or define an ordinary, nontemplate function. (Section
2124 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2125 could be interpreted as a particular specialization of a templatized
2126 function. Because this non-conforming behavior is no longer the default
2127 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2128 check existing code for potential trouble spots and is on by default.
2129 This new compiler behavior can be turned off with
2130 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2131 but disables the helpful warning.
2133 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2134 @opindex Wold-style-cast
2135 @opindex Wno-old-style-cast
2136 Warn if an old-style (C-style) cast to a non-void type is used within
2137 a C++ program. The new-style casts (@samp{dynamic_cast},
2138 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2139 less vulnerable to unintended effects and much easier to search for.
2141 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2142 @opindex Woverloaded-virtual
2143 @opindex Wno-overloaded-virtual
2144 @cindex overloaded virtual fn, warning
2145 @cindex warning for overloaded virtual fn
2146 Warn when a function declaration hides virtual functions from a
2147 base class. For example, in:
2154 struct B: public A @{
2159 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2167 will fail to compile.
2169 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
2170 @opindex Wno-pmf-conversions
2171 @opindex Wpmf-conversions
2172 Disable the diagnostic for converting a bound pointer to member function
2175 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
2176 @opindex Wsign-promo
2177 @opindex Wno-sign-promo
2178 Warn when overload resolution chooses a promotion from unsigned or
2179 enumerated type to a signed type, over a conversion to an unsigned type of
2180 the same size. Previous versions of G++ would try to preserve
2181 unsignedness, but the standard mandates the current behavior.
2186 A& operator = (int);
2196 In this example, G++ will synthesize a default @samp{A& operator =
2197 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2200 @node Objective-C and Objective-C++ Dialect Options
2201 @section Options Controlling Objective-C and Objective-C++ Dialects
2203 @cindex compiler options, Objective-C and Objective-C++
2204 @cindex Objective-C and Objective-C++ options, command line
2205 @cindex options, Objective-C and Objective-C++
2206 (NOTE: This manual does not describe the Objective-C and Objective-C++
2207 languages themselves. See @xref{Standards,,Language Standards
2208 Supported by GCC}, for references.)
2210 This section describes the command-line options that are only meaningful
2211 for Objective-C and Objective-C++ programs, but you can also use most of
2212 the language-independent GNU compiler options.
2213 For example, you might compile a file @code{some_class.m} like this:
2216 gcc -g -fgnu-runtime -O -c some_class.m
2220 In this example, @option{-fgnu-runtime} is an option meant only for
2221 Objective-C and Objective-C++ programs; you can use the other options with
2222 any language supported by GCC@.
2224 Note that since Objective-C is an extension of the C language, Objective-C
2225 compilations may also use options specific to the C front-end (e.g.,
2226 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2227 C++-specific options (e.g., @option{-Wabi}).
2229 Here is a list of options that are @emph{only} for compiling Objective-C
2230 and Objective-C++ programs:
2233 @item -fconstant-string-class=@var{class-name}
2234 @opindex fconstant-string-class
2235 Use @var{class-name} as the name of the class to instantiate for each
2236 literal string specified with the syntax @code{@@"@dots{}"}. The default
2237 class name is @code{NXConstantString} if the GNU runtime is being used, and
2238 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2239 @option{-fconstant-cfstrings} option, if also present, will override the
2240 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2241 to be laid out as constant CoreFoundation strings.
2244 @opindex fgnu-runtime
2245 Generate object code compatible with the standard GNU Objective-C
2246 runtime. This is the default for most types of systems.
2248 @item -fnext-runtime
2249 @opindex fnext-runtime
2250 Generate output compatible with the NeXT runtime. This is the default
2251 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2252 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2255 @item -fno-nil-receivers
2256 @opindex fno-nil-receivers
2257 Assume that all Objective-C message dispatches (e.g.,
2258 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2259 is not @code{nil}. This allows for more efficient entry points in the runtime
2260 to be used. Currently, this option is only available in conjunction with
2261 the NeXT runtime on Mac OS X 10.3 and later.
2263 @item -fobjc-call-cxx-cdtors
2264 @opindex fobjc-call-cxx-cdtors
2265 For each Objective-C class, check if any of its instance variables is a
2266 C++ object with a non-trivial default constructor. If so, synthesize a
2267 special @code{- (id) .cxx_construct} instance method that will run
2268 non-trivial default constructors on any such instance variables, in order,
2269 and then return @code{self}. Similarly, check if any instance variable
2270 is a C++ object with a non-trivial destructor, and if so, synthesize a
2271 special @code{- (void) .cxx_destruct} method that will run
2272 all such default destructors, in reverse order.
2274 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2275 thusly generated will only operate on instance variables declared in the
2276 current Objective-C class, and not those inherited from superclasses. It
2277 is the responsibility of the Objective-C runtime to invoke all such methods
2278 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2279 will be invoked by the runtime immediately after a new object
2280 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2281 be invoked immediately before the runtime deallocates an object instance.
2283 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2284 support for invoking the @code{- (id) .cxx_construct} and
2285 @code{- (void) .cxx_destruct} methods.
2287 @item -fobjc-direct-dispatch
2288 @opindex fobjc-direct-dispatch
2289 Allow fast jumps to the message dispatcher. On Darwin this is
2290 accomplished via the comm page.
2292 @item -fobjc-exceptions
2293 @opindex fobjc-exceptions
2294 Enable syntactic support for structured exception handling in Objective-C,
2295 similar to what is offered by C++ and Java. This option is
2296 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2305 @@catch (AnObjCClass *exc) @{
2312 @@catch (AnotherClass *exc) @{
2315 @@catch (id allOthers) @{
2325 The @code{@@throw} statement may appear anywhere in an Objective-C or
2326 Objective-C++ program; when used inside of a @code{@@catch} block, the
2327 @code{@@throw} may appear without an argument (as shown above), in which case
2328 the object caught by the @code{@@catch} will be rethrown.
2330 Note that only (pointers to) Objective-C objects may be thrown and
2331 caught using this scheme. When an object is thrown, it will be caught
2332 by the nearest @code{@@catch} clause capable of handling objects of that type,
2333 analogously to how @code{catch} blocks work in C++ and Java. A
2334 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2335 any and all Objective-C exceptions not caught by previous @code{@@catch}
2338 The @code{@@finally} clause, if present, will be executed upon exit from the
2339 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2340 regardless of whether any exceptions are thrown, caught or rethrown
2341 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2342 of the @code{finally} clause in Java.
2344 There are several caveats to using the new exception mechanism:
2348 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2349 idioms provided by the @code{NSException} class, the new
2350 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2351 systems, due to additional functionality needed in the (NeXT) Objective-C
2355 As mentioned above, the new exceptions do not support handling
2356 types other than Objective-C objects. Furthermore, when used from
2357 Objective-C++, the Objective-C exception model does not interoperate with C++
2358 exceptions at this time. This means you cannot @code{@@throw} an exception
2359 from Objective-C and @code{catch} it in C++, or vice versa
2360 (i.e., @code{throw @dots{} @@catch}).
2363 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2364 blocks for thread-safe execution:
2367 @@synchronized (ObjCClass *guard) @{
2372 Upon entering the @code{@@synchronized} block, a thread of execution shall
2373 first check whether a lock has been placed on the corresponding @code{guard}
2374 object by another thread. If it has, the current thread shall wait until
2375 the other thread relinquishes its lock. Once @code{guard} becomes available,
2376 the current thread will place its own lock on it, execute the code contained in
2377 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2378 making @code{guard} available to other threads).
2380 Unlike Java, Objective-C does not allow for entire methods to be marked
2381 @code{@@synchronized}. Note that throwing exceptions out of
2382 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2383 to be unlocked properly.
2387 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2389 @item -freplace-objc-classes
2390 @opindex freplace-objc-classes
2391 Emit a special marker instructing @command{ld(1)} not to statically link in
2392 the resulting object file, and allow @command{dyld(1)} to load it in at
2393 run time instead. This is used in conjunction with the Fix-and-Continue
2394 debugging mode, where the object file in question may be recompiled and
2395 dynamically reloaded in the course of program execution, without the need
2396 to restart the program itself. Currently, Fix-and-Continue functionality
2397 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2402 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2403 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2404 compile time) with static class references that get initialized at load time,
2405 which improves run-time performance. Specifying the @option{-fzero-link} flag
2406 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2407 to be retained. This is useful in Zero-Link debugging mode, since it allows
2408 for individual class implementations to be modified during program execution.
2412 Dump interface declarations for all classes seen in the source file to a
2413 file named @file{@var{sourcename}.decl}.
2415 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
2416 @opindex Wassign-intercept
2417 @opindex Wno-assign-intercept
2418 Warn whenever an Objective-C assignment is being intercepted by the
2421 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
2422 @opindex Wno-protocol
2424 If a class is declared to implement a protocol, a warning is issued for
2425 every method in the protocol that is not implemented by the class. The
2426 default behavior is to issue a warning for every method not explicitly
2427 implemented in the class, even if a method implementation is inherited
2428 from the superclass. If you use the @option{-Wno-protocol} option, then
2429 methods inherited from the superclass are considered to be implemented,
2430 and no warning is issued for them.
2432 @item -Wselector @r{(Objective-C and Objective-C++ only)}
2434 @opindex Wno-selector
2435 Warn if multiple methods of different types for the same selector are
2436 found during compilation. The check is performed on the list of methods
2437 in the final stage of compilation. Additionally, a check is performed
2438 for each selector appearing in a @code{@@selector(@dots{})}
2439 expression, and a corresponding method for that selector has been found
2440 during compilation. Because these checks scan the method table only at
2441 the end of compilation, these warnings are not produced if the final
2442 stage of compilation is not reached, for example because an error is
2443 found during compilation, or because the @option{-fsyntax-only} option is
2446 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
2447 @opindex Wstrict-selector-match
2448 @opindex Wno-strict-selector-match
2449 Warn if multiple methods with differing argument and/or return types are
2450 found for a given selector when attempting to send a message using this
2451 selector to a receiver of type @code{id} or @code{Class}. When this flag
2452 is off (which is the default behavior), the compiler will omit such warnings
2453 if any differences found are confined to types which share the same size
2456 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
2457 @opindex Wundeclared-selector
2458 @opindex Wno-undeclared-selector
2459 Warn if a @code{@@selector(@dots{})} expression referring to an
2460 undeclared selector is found. A selector is considered undeclared if no
2461 method with that name has been declared before the
2462 @code{@@selector(@dots{})} expression, either explicitly in an
2463 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2464 an @code{@@implementation} section. This option always performs its
2465 checks as soon as a @code{@@selector(@dots{})} expression is found,
2466 while @option{-Wselector} only performs its checks in the final stage of
2467 compilation. This also enforces the coding style convention
2468 that methods and selectors must be declared before being used.
2470 @item -print-objc-runtime-info
2471 @opindex print-objc-runtime-info
2472 Generate C header describing the largest structure that is passed by
2477 @node Language Independent Options
2478 @section Options to Control Diagnostic Messages Formatting
2479 @cindex options to control diagnostics formatting
2480 @cindex diagnostic messages
2481 @cindex message formatting
2483 Traditionally, diagnostic messages have been formatted irrespective of
2484 the output device's aspect (e.g.@: its width, @dots{}). The options described
2485 below can be used to control the diagnostic messages formatting
2486 algorithm, e.g.@: how many characters per line, how often source location
2487 information should be reported. Right now, only the C++ front end can
2488 honor these options. However it is expected, in the near future, that
2489 the remaining front ends would be able to digest them correctly.
2492 @item -fmessage-length=@var{n}
2493 @opindex fmessage-length
2494 Try to format error messages so that they fit on lines of about @var{n}
2495 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2496 the front ends supported by GCC@. If @var{n} is zero, then no
2497 line-wrapping will be done; each error message will appear on a single
2500 @opindex fdiagnostics-show-location
2501 @item -fdiagnostics-show-location=once
2502 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2503 reporter to emit @emph{once} source location information; that is, in
2504 case the message is too long to fit on a single physical line and has to
2505 be wrapped, the source location won't be emitted (as prefix) again,
2506 over and over, in subsequent continuation lines. This is the default
2509 @item -fdiagnostics-show-location=every-line
2510 Only meaningful in line-wrapping mode. Instructs the diagnostic
2511 messages reporter to emit the same source location information (as
2512 prefix) for physical lines that result from the process of breaking
2513 a message which is too long to fit on a single line.
2515 @item -fdiagnostics-show-option
2516 @opindex fdiagnostics-show-option
2517 This option instructs the diagnostic machinery to add text to each
2518 diagnostic emitted, which indicates which command line option directly
2519 controls that diagnostic, when such an option is known to the
2520 diagnostic machinery.
2522 @item -Wcoverage-mismatch
2523 @opindex Wcoverage-mismatch
2524 Warn if feedback profiles do not match when using the
2525 @option{-fprofile-use} option.
2526 If a source file was changed between @option{-fprofile-gen} and
2527 @option{-fprofile-use}, the files with the profile feedback can fail
2528 to match the source file and GCC can not use the profile feedback
2529 information. By default, GCC emits an error message in this case.
2530 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2531 error. GCC does not use appropriate feedback profiles, so using this
2532 option can result in poorly optimized code. This option is useful
2533 only in the case of very minor changes such as bug fixes to an
2538 @node Warning Options
2539 @section Options to Request or Suppress Warnings
2540 @cindex options to control warnings
2541 @cindex warning messages
2542 @cindex messages, warning
2543 @cindex suppressing warnings
2545 Warnings are diagnostic messages that report constructions which
2546 are not inherently erroneous but which are risky or suggest there
2547 may have been an error.
2549 The following language-independent options do not enable specific
2550 warnings but control the kinds of diagnostics produced by GCC.
2553 @cindex syntax checking
2555 @opindex fsyntax-only
2556 Check the code for syntax errors, but don't do anything beyond that.
2560 Inhibit all warning messages.
2565 Make all warnings into errors.
2570 Make the specified warning into an error. The specifier for a warning
2571 is appended, for example @option{-Werror=switch} turns the warnings
2572 controlled by @option{-Wswitch} into errors. This switch takes a
2573 negative form, to be used to negate @option{-Werror} for specific
2574 warnings, for example @option{-Wno-error=switch} makes
2575 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
2576 is in effect. You can use the @option{-fdiagnostics-show-option}
2577 option to have each controllable warning amended with the option which
2578 controls it, to determine what to use with this option.
2580 Note that specifying @option{-Werror=}@var{foo} automatically implies
2581 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
2584 @item -Wfatal-errors
2585 @opindex Wfatal-errors
2586 @opindex Wno-fatal-errors
2587 This option causes the compiler to abort compilation on the first error
2588 occurred rather than trying to keep going and printing further error
2593 You can request many specific warnings with options beginning
2594 @samp{-W}, for example @option{-Wimplicit} to request warnings on
2595 implicit declarations. Each of these specific warning options also
2596 has a negative form beginning @samp{-Wno-} to turn off warnings; for
2597 example, @option{-Wno-implicit}. This manual lists only one of the
2598 two forms, whichever is not the default. For further,
2599 language-specific options also refer to @ref{C++ Dialect Options} and
2600 @ref{Objective-C and Objective-C++ Dialect Options}.
2605 Issue all the warnings demanded by strict ISO C and ISO C++;
2606 reject all programs that use forbidden extensions, and some other
2607 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2608 version of the ISO C standard specified by any @option{-std} option used.
2610 Valid ISO C and ISO C++ programs should compile properly with or without
2611 this option (though a rare few will require @option{-ansi} or a
2612 @option{-std} option specifying the required version of ISO C)@. However,
2613 without this option, certain GNU extensions and traditional C and C++
2614 features are supported as well. With this option, they are rejected.
2616 @option{-pedantic} does not cause warning messages for use of the
2617 alternate keywords whose names begin and end with @samp{__}. Pedantic
2618 warnings are also disabled in the expression that follows
2619 @code{__extension__}. However, only system header files should use
2620 these escape routes; application programs should avoid them.
2621 @xref{Alternate Keywords}.
2623 Some users try to use @option{-pedantic} to check programs for strict ISO
2624 C conformance. They soon find that it does not do quite what they want:
2625 it finds some non-ISO practices, but not all---only those for which
2626 ISO C @emph{requires} a diagnostic, and some others for which
2627 diagnostics have been added.
2629 A feature to report any failure to conform to ISO C might be useful in
2630 some instances, but would require considerable additional work and would
2631 be quite different from @option{-pedantic}. We don't have plans to
2632 support such a feature in the near future.
2634 Where the standard specified with @option{-std} represents a GNU
2635 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2636 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2637 extended dialect is based. Warnings from @option{-pedantic} are given
2638 where they are required by the base standard. (It would not make sense
2639 for such warnings to be given only for features not in the specified GNU
2640 C dialect, since by definition the GNU dialects of C include all
2641 features the compiler supports with the given option, and there would be
2642 nothing to warn about.)
2644 @item -pedantic-errors
2645 @opindex pedantic-errors
2646 Like @option{-pedantic}, except that errors are produced rather than
2652 This enables all the warnings about constructions that some users
2653 consider questionable, and that are easy to avoid (or modify to
2654 prevent the warning), even in conjunction with macros. This also
2655 enables some language-specific warnings described in @ref{C++ Dialect
2656 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
2658 @option{-Wall} turns on the following warning flags:
2660 @gccoptlist{-Waddress @gol
2661 -Warray-bounds @r{(only with} @option{-O2}@r{)} @gol
2663 -Wchar-subscripts @gol
2665 -Wimplicit-function-declaration @gol
2668 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
2669 -Wmissing-braces @gol
2675 -Wsequence-point @gol
2676 -Wsign-compare @r{(only in C++)} @gol
2677 -Wstrict-aliasing @gol
2678 -Wstrict-overflow=1 @gol
2681 -Wuninitialized @r{(only with} @option{-O1} @r{and above)} @gol
2682 -Wunknown-pragmas @gol
2683 -Wunused-function @gol
2686 -Wunused-variable @gol
2687 -Wvolatile-register-var @gol
2690 Note that some warning flags are not implied by @option{-Wall}. Some of
2691 them warn about constructions that users generally do not consider
2692 questionable, but which occasionally you might wish to check for;
2693 others warn about constructions that are necessary or hard to avoid in
2694 some cases, and there is no simple way to modify the code to suppress
2695 the warning. Some of them are enabled by @option{-Wextra} but many of
2696 them must be enabled individually.
2702 This enables some extra warning flags that are not enabled by
2703 @option{-Wall}. (This option used to be called @option{-W}. The older
2704 name is still supported, but the newer name is more descriptive.)
2706 @gccoptlist{-Wclobbered @gol
2708 -Wignored-qualifiers @gol
2709 -Wmissing-field-initializers @gol
2710 -Wmissing-parameter-type @r{(C only)} @gol
2711 -Wold-style-declaration @r{(C only)} @gol
2712 -Woverride-init @gol
2715 -Wuninitialized @r{(only with} @option{-O1} @r{and above)} @gol
2716 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2719 The option @option{-Wextra} also prints warning messages for the
2725 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2726 @samp{>}, or @samp{>=}.
2729 (C++ only) An enumerator and a non-enumerator both appear in a
2730 conditional expression.
2733 (C++ only) A non-static reference or non-static @samp{const} member
2734 appears in a class without constructors.
2737 (C++ only) Ambiguous virtual bases.
2740 (C++ only) Subscripting an array which has been declared @samp{register}.
2743 (C++ only) Taking the address of a variable which has been declared
2747 (C++ only) A base class is not initialized in a derived class' copy
2755 Inhibit warning messages about the use of @samp{#import}.
2757 @item -Wchar-subscripts
2758 @opindex Wchar-subscripts
2759 @opindex Wno-char-subscripts
2760 Warn if an array subscript has type @code{char}. This is a common cause
2761 of error, as programmers often forget that this type is signed on some
2763 This warning is enabled by @option{-Wall}.
2767 @opindex Wno-comment
2768 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2769 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2770 This warning is enabled by @option{-Wall}.
2775 @opindex ffreestanding
2776 @opindex fno-builtin
2777 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2778 the arguments supplied have types appropriate to the format string
2779 specified, and that the conversions specified in the format string make
2780 sense. This includes standard functions, and others specified by format
2781 attributes (@pxref{Function Attributes}), in the @code{printf},
2782 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2783 not in the C standard) families (or other target-specific families).
2784 Which functions are checked without format attributes having been
2785 specified depends on the standard version selected, and such checks of
2786 functions without the attribute specified are disabled by
2787 @option{-ffreestanding} or @option{-fno-builtin}.
2789 The formats are checked against the format features supported by GNU
2790 libc version 2.2. These include all ISO C90 and C99 features, as well
2791 as features from the Single Unix Specification and some BSD and GNU
2792 extensions. Other library implementations may not support all these
2793 features; GCC does not support warning about features that go beyond a
2794 particular library's limitations. However, if @option{-pedantic} is used
2795 with @option{-Wformat}, warnings will be given about format features not
2796 in the selected standard version (but not for @code{strfmon} formats,
2797 since those are not in any version of the C standard). @xref{C Dialect
2798 Options,,Options Controlling C Dialect}.
2800 Since @option{-Wformat} also checks for null format arguments for
2801 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2803 @option{-Wformat} is included in @option{-Wall}. For more control over some
2804 aspects of format checking, the options @option{-Wformat-y2k},
2805 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2806 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2807 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2810 @opindex Wformat-y2k
2811 @opindex Wno-format-y2k
2812 If @option{-Wformat} is specified, also warn about @code{strftime}
2813 formats which may yield only a two-digit year.
2815 @item -Wno-format-extra-args
2816 @opindex Wno-format-extra-args
2817 @opindex Wformat-extra-args
2818 If @option{-Wformat} is specified, do not warn about excess arguments to a
2819 @code{printf} or @code{scanf} format function. The C standard specifies
2820 that such arguments are ignored.
2822 Where the unused arguments lie between used arguments that are
2823 specified with @samp{$} operand number specifications, normally
2824 warnings are still given, since the implementation could not know what
2825 type to pass to @code{va_arg} to skip the unused arguments. However,
2826 in the case of @code{scanf} formats, this option will suppress the
2827 warning if the unused arguments are all pointers, since the Single
2828 Unix Specification says that such unused arguments are allowed.
2830 @item -Wno-format-zero-length @r{(C and Objective-C only)}
2831 @opindex Wno-format-zero-length
2832 @opindex Wformat-zero-length
2833 If @option{-Wformat} is specified, do not warn about zero-length formats.
2834 The C standard specifies that zero-length formats are allowed.
2836 @item -Wformat-nonliteral
2837 @opindex Wformat-nonliteral
2838 @opindex Wno-format-nonliteral
2839 If @option{-Wformat} is specified, also warn if the format string is not a
2840 string literal and so cannot be checked, unless the format function
2841 takes its format arguments as a @code{va_list}.
2843 @item -Wformat-security
2844 @opindex Wformat-security
2845 @opindex Wno-format-security
2846 If @option{-Wformat} is specified, also warn about uses of format
2847 functions that represent possible security problems. At present, this
2848 warns about calls to @code{printf} and @code{scanf} functions where the
2849 format string is not a string literal and there are no format arguments,
2850 as in @code{printf (foo);}. This may be a security hole if the format
2851 string came from untrusted input and contains @samp{%n}. (This is
2852 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2853 in future warnings may be added to @option{-Wformat-security} that are not
2854 included in @option{-Wformat-nonliteral}.)
2858 @opindex Wno-format=2
2859 Enable @option{-Wformat} plus format checks not included in
2860 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2861 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2863 @item -Wnonnull @r{(C and Objective-C only)}
2865 @opindex Wno-nonnull
2866 Warn about passing a null pointer for arguments marked as
2867 requiring a non-null value by the @code{nonnull} function attribute.
2869 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2870 can be disabled with the @option{-Wno-nonnull} option.
2872 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
2874 @opindex Wno-init-self
2875 Warn about uninitialized variables which are initialized with themselves.
2876 Note this option can only be used with the @option{-Wuninitialized} option,
2877 which in turn only works with @option{-O1} and above.
2879 For example, GCC will warn about @code{i} being uninitialized in the
2880 following snippet only when @option{-Winit-self} has been specified:
2891 @item -Wimplicit-int @r{(C and Objective-C only)}
2892 @opindex Wimplicit-int
2893 @opindex Wno-implicit-int
2894 Warn when a declaration does not specify a type.
2895 This warning is enabled by @option{-Wall}.
2897 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
2898 @opindex Wimplicit-function-declaration
2899 @opindex Wno-implicit-function-declaration
2900 Give a warning whenever a function is used before being declared. In
2901 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
2902 enabled by default and it is made into an error by
2903 @option{-pedantic-errors}. This warning is also enabled by
2908 @opindex Wno-implicit
2909 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2910 This warning is enabled by @option{-Wall}.
2912 @item -Wignored-qualifiers @r{(C and C++ only)}
2913 @opindex Wignored-qualifiers
2914 @opindex Wno-ignored-qualifiers
2915 Warn if the return type of a function has a type qualifier
2916 such as @code{const}. For ISO C such a type qualifier has no effect,
2917 since the value returned by a function is not an lvalue.
2918 For C++, the warning is only emitted for scalar types or @code{void}.
2919 ISO C prohibits qualified @code{void} return types on function
2920 definitions, so such return types always receive a warning
2921 even without this option.
2923 This warning is also enabled by @option{-Wextra}.
2928 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2929 function with external linkage, returning int, taking either zero
2930 arguments, two, or three arguments of appropriate types.
2931 This warning is enabled by @option{-Wall}.
2933 @item -Wmissing-braces
2934 @opindex Wmissing-braces
2935 @opindex Wno-missing-braces
2936 Warn if an aggregate or union initializer is not fully bracketed. In
2937 the following example, the initializer for @samp{a} is not fully
2938 bracketed, but that for @samp{b} is fully bracketed.
2941 int a[2][2] = @{ 0, 1, 2, 3 @};
2942 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2945 This warning is enabled by @option{-Wall}.
2947 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
2948 @opindex Wmissing-include-dirs
2949 @opindex Wno-missing-include-dirs
2950 Warn if a user-supplied include directory does not exist.
2953 @opindex Wparentheses
2954 @opindex Wno-parentheses
2955 Warn if parentheses are omitted in certain contexts, such
2956 as when there is an assignment in a context where a truth value
2957 is expected, or when operators are nested whose precedence people
2958 often get confused about.
2960 Also warn if a comparison like @samp{x<=y<=z} appears; this is
2961 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
2962 interpretation from that of ordinary mathematical notation.
2964 Also warn about constructions where there may be confusion to which
2965 @code{if} statement an @code{else} branch belongs. Here is an example of
2980 In C/C++, every @code{else} branch belongs to the innermost possible
2981 @code{if} statement, which in this example is @code{if (b)}. This is
2982 often not what the programmer expected, as illustrated in the above
2983 example by indentation the programmer chose. When there is the
2984 potential for this confusion, GCC will issue a warning when this flag
2985 is specified. To eliminate the warning, add explicit braces around
2986 the innermost @code{if} statement so there is no way the @code{else}
2987 could belong to the enclosing @code{if}. The resulting code would
3004 This warning is enabled by @option{-Wall}.
3006 @item -Wsequence-point
3007 @opindex Wsequence-point
3008 @opindex Wno-sequence-point
3009 Warn about code that may have undefined semantics because of violations
3010 of sequence point rules in the C and C++ standards.
3012 The C and C++ standards defines the order in which expressions in a C/C++
3013 program are evaluated in terms of @dfn{sequence points}, which represent
3014 a partial ordering between the execution of parts of the program: those
3015 executed before the sequence point, and those executed after it. These
3016 occur after the evaluation of a full expression (one which is not part
3017 of a larger expression), after the evaluation of the first operand of a
3018 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
3019 function is called (but after the evaluation of its arguments and the
3020 expression denoting the called function), and in certain other places.
3021 Other than as expressed by the sequence point rules, the order of
3022 evaluation of subexpressions of an expression is not specified. All
3023 these rules describe only a partial order rather than a total order,
3024 since, for example, if two functions are called within one expression
3025 with no sequence point between them, the order in which the functions
3026 are called is not specified. However, the standards committee have
3027 ruled that function calls do not overlap.
3029 It is not specified when between sequence points modifications to the
3030 values of objects take effect. Programs whose behavior depends on this
3031 have undefined behavior; the C and C++ standards specify that ``Between
3032 the previous and next sequence point an object shall have its stored
3033 value modified at most once by the evaluation of an expression.
3034 Furthermore, the prior value shall be read only to determine the value
3035 to be stored.''. If a program breaks these rules, the results on any
3036 particular implementation are entirely unpredictable.
3038 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
3039 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
3040 diagnosed by this option, and it may give an occasional false positive
3041 result, but in general it has been found fairly effective at detecting
3042 this sort of problem in programs.
3044 The standard is worded confusingly, therefore there is some debate
3045 over the precise meaning of the sequence point rules in subtle cases.
3046 Links to discussions of the problem, including proposed formal
3047 definitions, may be found on the GCC readings page, at
3048 @w{@uref{http://gcc.gnu.org/readings.html}}.
3050 This warning is enabled by @option{-Wall} for C and C++.
3053 @opindex Wreturn-type
3054 @opindex Wno-return-type
3055 Warn whenever a function is defined with a return-type that defaults
3056 to @code{int}. Also warn about any @code{return} statement with no
3057 return-value in a function whose return-type is not @code{void}
3058 (falling off the end of the function body is considered returning
3059 without a value), and about a @code{return} statement with a
3060 expression in a function whose return-type is @code{void}.
3062 For C++, a function without return type always produces a diagnostic
3063 message, even when @option{-Wno-return-type} is specified. The only
3064 exceptions are @samp{main} and functions defined in system headers.
3066 This warning is enabled by @option{-Wall}.
3071 Warn whenever a @code{switch} statement has an index of enumerated type
3072 and lacks a @code{case} for one or more of the named codes of that
3073 enumeration. (The presence of a @code{default} label prevents this
3074 warning.) @code{case} labels outside the enumeration range also
3075 provoke warnings when this option is used.
3076 This warning is enabled by @option{-Wall}.
3078 @item -Wswitch-default
3079 @opindex Wswitch-default
3080 @opindex Wno-switch-default
3081 Warn whenever a @code{switch} statement does not have a @code{default}
3085 @opindex Wswitch-enum
3086 @opindex Wno-switch-enum
3087 Warn whenever a @code{switch} statement has an index of enumerated type
3088 and lacks a @code{case} for one or more of the named codes of that
3089 enumeration. @code{case} labels outside the enumeration range also
3090 provoke warnings when this option is used.
3094 @opindex Wno-trigraphs
3095 Warn if any trigraphs are encountered that might change the meaning of
3096 the program (trigraphs within comments are not warned about).
3097 This warning is enabled by @option{-Wall}.
3099 @item -Wunused-function
3100 @opindex Wunused-function
3101 @opindex Wno-unused-function
3102 Warn whenever a static function is declared but not defined or a
3103 non-inline static function is unused.
3104 This warning is enabled by @option{-Wall}.
3106 @item -Wunused-label
3107 @opindex Wunused-label
3108 @opindex Wno-unused-label
3109 Warn whenever a label is declared but not used.
3110 This warning is enabled by @option{-Wall}.
3112 To suppress this warning use the @samp{unused} attribute
3113 (@pxref{Variable Attributes}).
3115 @item -Wunused-parameter
3116 @opindex Wunused-parameter
3117 @opindex Wno-unused-parameter
3118 Warn whenever a function parameter is unused aside from its declaration.
3120 To suppress this warning use the @samp{unused} attribute
3121 (@pxref{Variable Attributes}).
3123 @item -Wunused-variable
3124 @opindex Wunused-variable
3125 @opindex Wno-unused-variable
3126 Warn whenever a local variable or non-constant static variable is unused
3127 aside from its declaration.
3128 This warning is enabled by @option{-Wall}.
3130 To suppress this warning use the @samp{unused} attribute
3131 (@pxref{Variable Attributes}).
3133 @item -Wunused-value
3134 @opindex Wunused-value
3135 @opindex Wno-unused-value
3136 Warn whenever a statement computes a result that is explicitly not
3137 used. To suppress this warning cast the unused expression to
3138 @samp{void}. This includes an expression-statement or the left-hand
3139 side of a comma expression that contains no side effects. For example,
3140 an expression such as @samp{x[i,j]} will cause a warning, while
3141 @samp{x[(void)i,j]} will not.
3143 This warning is enabled by @option{-Wall}.
3148 All the above @option{-Wunused} options combined.
3150 In order to get a warning about an unused function parameter, you must
3151 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
3152 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
3154 @item -Wuninitialized
3155 @opindex Wuninitialized
3156 @opindex Wno-uninitialized
3157 Warn if an automatic variable is used without first being initialized or
3158 if a variable may be clobbered by a @code{setjmp} call.
3160 These warnings are possible only in optimizing compilation,
3161 because they require data flow information that is computed only
3162 when optimizing. If you do not specify @option{-O}, you will not get
3163 these warnings. Instead, GCC will issue a warning about @option{-Wuninitialized}
3164 requiring @option{-O}.
3166 If you want to warn about code which uses the uninitialized value of the
3167 variable in its own initializer, use the @option{-Winit-self} option.
3169 These warnings occur for individual uninitialized or clobbered
3170 elements of structure, union or array variables as well as for
3171 variables which are uninitialized or clobbered as a whole. They do
3172 not occur for variables or elements declared @code{volatile}. Because
3173 these warnings depend on optimization, the exact variables or elements
3174 for which there are warnings will depend on the precise optimization
3175 options and version of GCC used.
3177 Note that there may be no warning about a variable that is used only
3178 to compute a value that itself is never used, because such
3179 computations may be deleted by data flow analysis before the warnings
3182 These warnings are made optional because GCC is not smart
3183 enough to see all the reasons why the code might be correct
3184 despite appearing to have an error. Here is one example of how
3205 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3206 always initialized, but GCC doesn't know this. Here is
3207 another common case:
3212 if (change_y) save_y = y, y = new_y;
3214 if (change_y) y = save_y;
3219 This has no bug because @code{save_y} is used only if it is set.
3221 @cindex @code{longjmp} warnings
3222 This option also warns when a non-volatile automatic variable might be
3223 changed by a call to @code{longjmp}. These warnings as well are possible
3224 only in optimizing compilation.
3226 The compiler sees only the calls to @code{setjmp}. It cannot know
3227 where @code{longjmp} will be called; in fact, a signal handler could
3228 call it at any point in the code. As a result, you may get a warning
3229 even when there is in fact no problem because @code{longjmp} cannot
3230 in fact be called at the place which would cause a problem.
3232 Some spurious warnings can be avoided if you declare all the functions
3233 you use that never return as @code{noreturn}. @xref{Function
3236 This warning is enabled by @option{-Wall} or @option{-Wextra} in
3237 optimizing compilations (@option{-O1} and above).
3239 @item -Wunknown-pragmas
3240 @opindex Wunknown-pragmas
3241 @opindex Wno-unknown-pragmas
3242 @cindex warning for unknown pragmas
3243 @cindex unknown pragmas, warning
3244 @cindex pragmas, warning of unknown
3245 Warn when a #pragma directive is encountered which is not understood by
3246 GCC@. If this command line option is used, warnings will even be issued
3247 for unknown pragmas in system header files. This is not the case if
3248 the warnings were only enabled by the @option{-Wall} command line option.
3251 @opindex Wno-pragmas
3253 Do not warn about misuses of pragmas, such as incorrect parameters,
3254 invalid syntax, or conflicts between pragmas. See also
3255 @samp{-Wunknown-pragmas}.
3257 @item -Wstrict-aliasing
3258 @opindex Wstrict-aliasing
3259 @opindex Wno-strict-aliasing
3260 This option is only active when @option{-fstrict-aliasing} is active.
3261 It warns about code which might break the strict aliasing rules that the
3262 compiler is using for optimization. The warning does not catch all
3263 cases, but does attempt to catch the more common pitfalls. It is
3264 included in @option{-Wall}.
3265 It is equivalent to @option{-Wstrict-aliasing=3}
3267 @item -Wstrict-aliasing=n
3268 @opindex Wstrict-aliasing=n
3269 @opindex Wno-strict-aliasing=n
3270 This option is only active when @option{-fstrict-aliasing} is active.
3271 It warns about code which might break the strict aliasing rules that the
3272 compiler is using for optimization.
3273 Higher levels correspond to higher accuracy (fewer false positives).
3274 Higher levels also correspond to more effort, similar to the way -O works.
3275 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3278 Level 1: Most aggressive, quick, least accurate.
3279 Possibly useful when higher levels
3280 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3281 false negatives. However, it has many false positives.
3282 Warns for all pointer conversions between possibly incompatible types,
3283 even if never dereferenced. Runs in the frontend only.
3285 Level 2: Aggressive, quick, not too precise.
3286 May still have many false positives (not as many as level 1 though),
3287 and few false negatives (but possibly more than level 1).
3288 Unlike level 1, it only warns when an address is taken. Warns about
3289 incomplete types. Runs in the frontend only.
3291 Level 3 (default for @option{-Wstrict-aliasing}):
3292 Should have very few false positives and few false
3293 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3294 Takes care of the common punn+dereference pattern in the frontend:
3295 @code{*(int*)&some_float}.
3296 If optimization is enabled, it also runs in the backend, where it deals
3297 with multiple statement cases using flow-sensitive points-to information.
3298 Only warns when the converted pointer is dereferenced.
3299 Does not warn about incomplete types.
3301 @item -Wstrict-overflow
3302 @itemx -Wstrict-overflow=@var{n}
3303 @opindex Wstrict-overflow
3304 @opindex Wno-strict-overflow
3305 This option is only active when @option{-fstrict-overflow} is active.
3306 It warns about cases where the compiler optimizes based on the
3307 assumption that signed overflow does not occur. Note that it does not
3308 warn about all cases where the code might overflow: it only warns
3309 about cases where the compiler implements some optimization. Thus
3310 this warning depends on the optimization level.
3312 An optimization which assumes that signed overflow does not occur is
3313 perfectly safe if the values of the variables involved are such that
3314 overflow never does, in fact, occur. Therefore this warning can
3315 easily give a false positive: a warning about code which is not
3316 actually a problem. To help focus on important issues, several
3317 warning levels are defined. No warnings are issued for the use of
3318 undefined signed overflow when estimating how many iterations a loop
3319 will require, in particular when determining whether a loop will be
3323 @item -Wstrict-overflow=1
3324 Warn about cases which are both questionable and easy to avoid. For
3325 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3326 compiler will simplify this to @code{1}. This level of
3327 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3328 are not, and must be explicitly requested.
3330 @item -Wstrict-overflow=2
3331 Also warn about other cases where a comparison is simplified to a
3332 constant. For example: @code{abs (x) >= 0}. This can only be
3333 simplified when @option{-fstrict-overflow} is in effect, because
3334 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3335 zero. @option{-Wstrict-overflow} (with no level) is the same as
3336 @option{-Wstrict-overflow=2}.
3338 @item -Wstrict-overflow=3
3339 Also warn about other cases where a comparison is simplified. For
3340 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3342 @item -Wstrict-overflow=4
3343 Also warn about other simplifications not covered by the above cases.
3344 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3346 @item -Wstrict-overflow=5
3347 Also warn about cases where the compiler reduces the magnitude of a
3348 constant involved in a comparison. For example: @code{x + 2 > y} will
3349 be simplified to @code{x + 1 >= y}. This is reported only at the
3350 highest warning level because this simplification applies to many
3351 comparisons, so this warning level will give a very large number of
3355 @item -Warray-bounds
3356 @opindex Wno-array-bounds
3357 @opindex Warray-bounds
3358 This option is only active when @option{-ftree-vrp} is active
3359 (default for -O2 and above). It warns about subscripts to arrays
3360 that are always out of bounds. This warning is enabled by @option{-Wall}.
3362 @item -Wno-div-by-zero
3363 @opindex Wno-div-by-zero
3364 @opindex Wdiv-by-zero
3365 Do not warn about compile-time integer division by zero. Floating point
3366 division by zero is not warned about, as it can be a legitimate way of
3367 obtaining infinities and NaNs.
3369 @item -Wsystem-headers
3370 @opindex Wsystem-headers
3371 @opindex Wno-system-headers
3372 @cindex warnings from system headers
3373 @cindex system headers, warnings from
3374 Print warning messages for constructs found in system header files.
3375 Warnings from system headers are normally suppressed, on the assumption
3376 that they usually do not indicate real problems and would only make the
3377 compiler output harder to read. Using this command line option tells
3378 GCC to emit warnings from system headers as if they occurred in user
3379 code. However, note that using @option{-Wall} in conjunction with this
3380 option will @emph{not} warn about unknown pragmas in system
3381 headers---for that, @option{-Wunknown-pragmas} must also be used.
3384 @opindex Wfloat-equal
3385 @opindex Wno-float-equal
3386 Warn if floating point values are used in equality comparisons.
3388 The idea behind this is that sometimes it is convenient (for the
3389 programmer) to consider floating-point values as approximations to
3390 infinitely precise real numbers. If you are doing this, then you need
3391 to compute (by analyzing the code, or in some other way) the maximum or
3392 likely maximum error that the computation introduces, and allow for it
3393 when performing comparisons (and when producing output, but that's a
3394 different problem). In particular, instead of testing for equality, you
3395 would check to see whether the two values have ranges that overlap; and
3396 this is done with the relational operators, so equality comparisons are
3399 @item -Wtraditional @r{(C and Objective-C only)}
3400 @opindex Wtraditional
3401 @opindex Wno-traditional
3402 Warn about certain constructs that behave differently in traditional and
3403 ISO C@. Also warn about ISO C constructs that have no traditional C
3404 equivalent, and/or problematic constructs which should be avoided.
3408 Macro parameters that appear within string literals in the macro body.
3409 In traditional C macro replacement takes place within string literals,
3410 but does not in ISO C@.
3413 In traditional C, some preprocessor directives did not exist.
3414 Traditional preprocessors would only consider a line to be a directive
3415 if the @samp{#} appeared in column 1 on the line. Therefore
3416 @option{-Wtraditional} warns about directives that traditional C
3417 understands but would ignore because the @samp{#} does not appear as the
3418 first character on the line. It also suggests you hide directives like
3419 @samp{#pragma} not understood by traditional C by indenting them. Some
3420 traditional implementations would not recognize @samp{#elif}, so it
3421 suggests avoiding it altogether.
3424 A function-like macro that appears without arguments.
3427 The unary plus operator.
3430 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3431 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3432 constants.) Note, these suffixes appear in macros defined in the system
3433 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3434 Use of these macros in user code might normally lead to spurious
3435 warnings, however GCC's integrated preprocessor has enough context to
3436 avoid warning in these cases.
3439 A function declared external in one block and then used after the end of
3443 A @code{switch} statement has an operand of type @code{long}.
3446 A non-@code{static} function declaration follows a @code{static} one.
3447 This construct is not accepted by some traditional C compilers.
3450 The ISO type of an integer constant has a different width or
3451 signedness from its traditional type. This warning is only issued if
3452 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3453 typically represent bit patterns, are not warned about.
3456 Usage of ISO string concatenation is detected.
3459 Initialization of automatic aggregates.
3462 Identifier conflicts with labels. Traditional C lacks a separate
3463 namespace for labels.
3466 Initialization of unions. If the initializer is zero, the warning is
3467 omitted. This is done under the assumption that the zero initializer in
3468 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3469 initializer warnings and relies on default initialization to zero in the
3473 Conversions by prototypes between fixed/floating point values and vice
3474 versa. The absence of these prototypes when compiling with traditional
3475 C would cause serious problems. This is a subset of the possible
3476 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3479 Use of ISO C style function definitions. This warning intentionally is
3480 @emph{not} issued for prototype declarations or variadic functions
3481 because these ISO C features will appear in your code when using
3482 libiberty's traditional C compatibility macros, @code{PARAMS} and
3483 @code{VPARAMS}. This warning is also bypassed for nested functions
3484 because that feature is already a GCC extension and thus not relevant to
3485 traditional C compatibility.
3488 @item -Wtraditional-conversion @r{(C and Objective-C only)}
3489 @opindex Wtraditional-conversion
3490 @opindex Wno-traditional-conversion
3491 Warn if a prototype causes a type conversion that is different from what
3492 would happen to the same argument in the absence of a prototype. This
3493 includes conversions of fixed point to floating and vice versa, and
3494 conversions changing the width or signedness of a fixed point argument
3495 except when the same as the default promotion.
3497 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
3498 @opindex Wdeclaration-after-statement
3499 @opindex Wno-declaration-after-statement
3500 Warn when a declaration is found after a statement in a block. This
3501 construct, known from C++, was introduced with ISO C99 and is by default
3502 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3503 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3508 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3510 @item -Wno-endif-labels
3511 @opindex Wno-endif-labels
3512 @opindex Wendif-labels
3513 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3518 Warn whenever a local variable shadows another local variable, parameter or
3519 global variable or whenever a built-in function is shadowed.
3521 @item -Wlarger-than=@var{len}
3522 @opindex Wlarger-than=@var{len}
3523 @opindex Wlarger-than-@var{len}
3524 Warn whenever an object of larger than @var{len} bytes is defined.
3526 @item -Wframe-larger-than=@var{len}
3527 @opindex Wframe-larger-than
3528 Warn whenever the size of a function frame is larger than @var{len} bytes.
3530 @item -Wunsafe-loop-optimizations
3531 @opindex Wunsafe-loop-optimizations
3532 @opindex Wno-unsafe-loop-optimizations
3533 Warn if the loop cannot be optimized because the compiler could not
3534 assume anything on the bounds of the loop indices. With
3535 @option{-funsafe-loop-optimizations} warn if the compiler made
3538 @item -Wpointer-arith
3539 @opindex Wpointer-arith
3540 @opindex Wno-pointer-arith
3541 Warn about anything that depends on the ``size of'' a function type or
3542 of @code{void}. GNU C assigns these types a size of 1, for
3543 convenience in calculations with @code{void *} pointers and pointers
3544 to functions. In C++, warn also when an arithmetic operation involves
3545 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3548 @opindex Wtype-limits
3549 @opindex Wno-type-limits
3550 Warn if a comparison is always true or always false due to the limited
3551 range of the data type, but do not warn for constant expressions. For
3552 example, warn if an unsigned variable is compared against zero with
3553 @samp{<} or @samp{>=}. This warning is also enabled by
3556 @item -Wbad-function-cast @r{(C and Objective-C only)}
3557 @opindex Wbad-function-cast
3558 @opindex Wno-bad-function-cast
3559 Warn whenever a function call is cast to a non-matching type.
3560 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3562 @item -Wc++-compat @r{(C and Objective-C only)}
3563 Warn about ISO C constructs that are outside of the common subset of
3564 ISO C and ISO C++, e.g.@: request for implicit conversion from
3565 @code{void *} to a pointer to non-@code{void} type.
3567 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3568 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3569 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3570 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3574 @opindex Wno-cast-qual
3575 Warn whenever a pointer is cast so as to remove a type qualifier from
3576 the target type. For example, warn if a @code{const char *} is cast
3577 to an ordinary @code{char *}.
3580 @opindex Wcast-align
3581 @opindex Wno-cast-align
3582 Warn whenever a pointer is cast such that the required alignment of the
3583 target is increased. For example, warn if a @code{char *} is cast to
3584 an @code{int *} on machines where integers can only be accessed at
3585 two- or four-byte boundaries.
3587 @item -Wwrite-strings
3588 @opindex Wwrite-strings
3589 @opindex Wno-write-strings
3590 When compiling C, give string constants the type @code{const
3591 char[@var{length}]} so that
3592 copying the address of one into a non-@code{const} @code{char *}
3593 pointer will get a warning; when compiling C++, warn about the
3594 deprecated conversion from string literals to @code{char *}. This
3595 warning, by default, is enabled for C++ programs.
3596 These warnings will help you find at
3597 compile time code that can try to write into a string constant, but
3598 only if you have been very careful about using @code{const} in
3599 declarations and prototypes. Otherwise, it will just be a nuisance;
3600 this is why we did not make @option{-Wall} request these warnings.
3604 @opindex Wno-clobbered
3605 Warn for variables that might be changed by @samp{longjmp} or
3606 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3609 @opindex Wconversion
3610 @opindex Wno-conversion
3611 Warn for implicit conversions that may alter a value. This includes
3612 conversions between real and integer, like @code{abs (x)} when
3613 @code{x} is @code{double}; conversions between signed and unsigned,
3614 like @code{unsigned ui = -1}; and conversions to smaller types, like
3615 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3616 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3617 changed by the conversion like in @code{abs (2.0)}. Warnings about
3618 conversions between signed and unsigned integers can be disabled by
3619 using @option{-Wno-sign-conversion}.
3621 For C++, also warn for conversions between @code{NULL} and non-pointer
3622 types; confusing overload resolution for user-defined conversions; and
3623 conversions that will never use a type conversion operator:
3624 conversions to @code{void}, the same type, a base class or a reference
3625 to them. Warnings about conversions between signed and unsigned
3626 integers are disabled by default in C++ unless
3627 @option{-Wsign-conversion} is explicitly enabled.
3630 @opindex Wempty-body
3631 @opindex Wno-empty-body
3632 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3633 while} statement. Additionally, in C++, warn when an empty body occurs
3634 in a @samp{while} or @samp{for} statement with no whitespacing before
3635 the semicolon. This warning is also enabled by @option{-Wextra}.
3637 @item -Wsign-compare
3638 @opindex Wsign-compare
3639 @opindex Wno-sign-compare
3640 @cindex warning for comparison of signed and unsigned values
3641 @cindex comparison of signed and unsigned values, warning
3642 @cindex signed and unsigned values, comparison warning
3643 Warn when a comparison between signed and unsigned values could produce
3644 an incorrect result when the signed value is converted to unsigned.
3645 This warning is also enabled by @option{-Wextra}; to get the other warnings
3646 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3648 @item -Wsign-conversion
3649 @opindex Wsign-conversion
3650 @opindex Wno-sign-conversion
3651 Warn for implicit conversions that may change the sign of an integer
3652 value, like assigning a signed integer expression to an unsigned
3653 integer variable. An explicit cast silences the warning. In C, this
3654 option is enabled also by @option{-Wconversion}.
3658 @opindex Wno-address
3659 Warn about suspicious uses of memory addresses. These include using
3660 the address of a function in a conditional expression, such as
3661 @code{void func(void); if (func)}, and comparisons against the memory
3662 address of a string literal, such as @code{if (x == "abc")}. Such
3663 uses typically indicate a programmer error: the address of a function
3664 always evaluates to true, so their use in a conditional usually
3665 indicate that the programmer forgot the parentheses in a function
3666 call; and comparisons against string literals result in unspecified
3667 behavior and are not portable in C, so they usually indicate that the
3668 programmer intended to use @code{strcmp}. This warning is enabled by
3672 @opindex Wlogical-op
3673 @opindex Wno-logical-op
3674 Warn about suspicious uses of logical operators in expressions.
3675 This includes using logical operators in contexts where a
3676 bit-wise operator is likely to be expected.
3678 @item -Waggregate-return
3679 @opindex Waggregate-return
3680 @opindex Wno-aggregate-return
3681 Warn if any functions that return structures or unions are defined or
3682 called. (In languages where you can return an array, this also elicits
3685 @item -Wno-attributes
3686 @opindex Wno-attributes
3687 @opindex Wattributes
3688 Do not warn if an unexpected @code{__attribute__} is used, such as
3689 unrecognized attributes, function attributes applied to variables,
3690 etc. This will not stop errors for incorrect use of supported
3693 @item -Wstrict-prototypes @r{(C and Objective-C only)}
3694 @opindex Wstrict-prototypes
3695 @opindex Wno-strict-prototypes
3696 Warn if a function is declared or defined without specifying the
3697 argument types. (An old-style function definition is permitted without
3698 a warning if preceded by a declaration which specifies the argument
3701 @item -Wold-style-declaration @r{(C and Objective-C only)}
3702 @opindex Wold-style-declaration
3703 @opindex Wno-old-style-declaration
3704 Warn for obsolescent usages, according to the C Standard, in a
3705 declaration. For example, warn if storage-class specifiers like
3706 @code{static} are not the first things in a declaration. This warning
3707 is also enabled by @option{-Wextra}.
3709 @item -Wold-style-definition @r{(C and Objective-C only)}
3710 @opindex Wold-style-definition
3711 @opindex Wno-old-style-definition
3712 Warn if an old-style function definition is used. A warning is given
3713 even if there is a previous prototype.
3715 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
3716 @opindex Wmissing-parameter-type
3717 @opindex Wno-missing-parameter-type
3718 A function parameter is declared without a type specifier in K&R-style
3725 This warning is also enabled by @option{-Wextra}.
3727 @item -Wmissing-prototypes @r{(C and Objective-C only)}
3728 @opindex Wmissing-prototypes
3729 @opindex Wno-missing-prototypes
3730 Warn if a global function is defined without a previous prototype
3731 declaration. This warning is issued even if the definition itself
3732 provides a prototype. The aim is to detect global functions that fail
3733 to be declared in header files.
3735 @item -Wmissing-declarations
3736 @opindex Wmissing-declarations
3737 @opindex Wno-missing-declarations
3738 Warn if a global function is defined without a previous declaration.
3739 Do so even if the definition itself provides a prototype.
3740 Use this option to detect global functions that are not declared in
3741 header files. In C++, no warnings are issued for function templates,
3742 or for inline functions, or for functions in anonymous namespaces.
3744 @item -Wmissing-field-initializers
3745 @opindex Wmissing-field-initializers
3746 @opindex Wno-missing-field-initializers
3750 Warn if a structure's initializer has some fields missing. For
3751 example, the following code would cause such a warning, because
3752 @code{x.h} is implicitly zero:
3755 struct s @{ int f, g, h; @};
3756 struct s x = @{ 3, 4 @};
3759 This option does not warn about designated initializers, so the following
3760 modification would not trigger a warning:
3763 struct s @{ int f, g, h; @};
3764 struct s x = @{ .f = 3, .g = 4 @};
3767 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3768 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3770 @item -Wmissing-noreturn
3771 @opindex Wmissing-noreturn
3772 @opindex Wno-missing-noreturn
3773 Warn about functions which might be candidates for attribute @code{noreturn}.
3774 Note these are only possible candidates, not absolute ones. Care should
3775 be taken to manually verify functions actually do not ever return before
3776 adding the @code{noreturn} attribute, otherwise subtle code generation
3777 bugs could be introduced. You will not get a warning for @code{main} in
3778 hosted C environments.
3780 @item -Wmissing-format-attribute
3781 @opindex Wmissing-format-attribute
3782 @opindex Wno-missing-format-attribute
3785 Warn about function pointers which might be candidates for @code{format}
3786 attributes. Note these are only possible candidates, not absolute ones.
3787 GCC will guess that function pointers with @code{format} attributes that
3788 are used in assignment, initialization, parameter passing or return
3789 statements should have a corresponding @code{format} attribute in the
3790 resulting type. I.e.@: the left-hand side of the assignment or
3791 initialization, the type of the parameter variable, or the return type
3792 of the containing function respectively should also have a @code{format}
3793 attribute to avoid the warning.
3795 GCC will also warn about function definitions which might be
3796 candidates for @code{format} attributes. Again, these are only
3797 possible candidates. GCC will guess that @code{format} attributes
3798 might be appropriate for any function that calls a function like
3799 @code{vprintf} or @code{vscanf}, but this might not always be the
3800 case, and some functions for which @code{format} attributes are
3801 appropriate may not be detected.
3803 @item -Wno-multichar
3804 @opindex Wno-multichar
3806 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
3807 Usually they indicate a typo in the user's code, as they have
3808 implementation-defined values, and should not be used in portable code.
3810 @item -Wnormalized=<none|id|nfc|nfkc>
3811 @opindex Wnormalized=
3814 @cindex character set, input normalization
3815 In ISO C and ISO C++, two identifiers are different if they are
3816 different sequences of characters. However, sometimes when characters
3817 outside the basic ASCII character set are used, you can have two
3818 different character sequences that look the same. To avoid confusion,
3819 the ISO 10646 standard sets out some @dfn{normalization rules} which
3820 when applied ensure that two sequences that look the same are turned into
3821 the same sequence. GCC can warn you if you are using identifiers which
3822 have not been normalized; this option controls that warning.
3824 There are four levels of warning that GCC supports. The default is
3825 @option{-Wnormalized=nfc}, which warns about any identifier which is
3826 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
3827 recommended form for most uses.
3829 Unfortunately, there are some characters which ISO C and ISO C++ allow
3830 in identifiers that when turned into NFC aren't allowable as
3831 identifiers. That is, there's no way to use these symbols in portable
3832 ISO C or C++ and have all your identifiers in NFC@.
3833 @option{-Wnormalized=id} suppresses the warning for these characters.
3834 It is hoped that future versions of the standards involved will correct
3835 this, which is why this option is not the default.
3837 You can switch the warning off for all characters by writing
3838 @option{-Wnormalized=none}. You would only want to do this if you
3839 were using some other normalization scheme (like ``D''), because
3840 otherwise you can easily create bugs that are literally impossible to see.
3842 Some characters in ISO 10646 have distinct meanings but look identical
3843 in some fonts or display methodologies, especially once formatting has
3844 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
3845 LETTER N'', will display just like a regular @code{n} which has been
3846 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
3847 normalization scheme to convert all these into a standard form as
3848 well, and GCC will warn if your code is not in NFKC if you use
3849 @option{-Wnormalized=nfkc}. This warning is comparable to warning
3850 about every identifier that contains the letter O because it might be
3851 confused with the digit 0, and so is not the default, but may be
3852 useful as a local coding convention if the programming environment is
3853 unable to be fixed to display these characters distinctly.
3855 @item -Wno-deprecated-declarations
3856 @opindex Wno-deprecated-declarations
3857 @opindex Wdeprecated-declarations
3858 Do not warn about uses of functions (@pxref{Function Attributes}),
3859 variables (@pxref{Variable Attributes}), and types (@pxref{Type
3860 Attributes}) marked as deprecated by using the @code{deprecated}
3864 @opindex Wno-overflow
3866 Do not warn about compile-time overflow in constant expressions.
3868 @item -Woverride-init @r{(C and Objective-C only)}
3869 @opindex Woverride-init
3870 @opindex Wno-override-init
3874 Warn if an initialized field without side effects is overridden when
3875 using designated initializers (@pxref{Designated Inits, , Designated
3878 This warning is included in @option{-Wextra}. To get other
3879 @option{-Wextra} warnings without this one, use @samp{-Wextra
3880 -Wno-override-init}.
3885 Warn if a structure is given the packed attribute, but the packed
3886 attribute has no effect on the layout or size of the structure.
3887 Such structures may be mis-aligned for little benefit. For
3888 instance, in this code, the variable @code{f.x} in @code{struct bar}
3889 will be misaligned even though @code{struct bar} does not itself
3890 have the packed attribute:
3897 @} __attribute__((packed));
3908 Warn if padding is included in a structure, either to align an element
3909 of the structure or to align the whole structure. Sometimes when this
3910 happens it is possible to rearrange the fields of the structure to
3911 reduce the padding and so make the structure smaller.
3913 @item -Wredundant-decls
3914 @opindex Wredundant-decls
3915 @opindex Wno-redundant-decls
3916 Warn if anything is declared more than once in the same scope, even in
3917 cases where multiple declaration is valid and changes nothing.
3919 @item -Wnested-externs @r{(C and Objective-C only)}
3920 @opindex Wnested-externs
3921 @opindex Wno-nested-externs
3922 Warn if an @code{extern} declaration is encountered within a function.
3924 @item -Wunreachable-code
3925 @opindex Wunreachable-code
3926 @opindex Wno-unreachable-code
3927 Warn if the compiler detects that code will never be executed.
3929 This option is intended to warn when the compiler detects that at
3930 least a whole line of source code will never be executed, because
3931 some condition is never satisfied or because it is after a
3932 procedure that never returns.
3934 It is possible for this option to produce a warning even though there
3935 are circumstances under which part of the affected line can be executed,
3936 so care should be taken when removing apparently-unreachable code.
3938 For instance, when a function is inlined, a warning may mean that the
3939 line is unreachable in only one inlined copy of the function.
3941 This option is not made part of @option{-Wall} because in a debugging
3942 version of a program there is often substantial code which checks
3943 correct functioning of the program and is, hopefully, unreachable
3944 because the program does work. Another common use of unreachable
3945 code is to provide behavior which is selectable at compile-time.
3950 Warn if a function can not be inlined and it was declared as inline.
3951 Even with this option, the compiler will not warn about failures to
3952 inline functions declared in system headers.
3954 The compiler uses a variety of heuristics to determine whether or not
3955 to inline a function. For example, the compiler takes into account
3956 the size of the function being inlined and the amount of inlining
3957 that has already been done in the current function. Therefore,
3958 seemingly insignificant changes in the source program can cause the
3959 warnings produced by @option{-Winline} to appear or disappear.
3961 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
3962 @opindex Wno-invalid-offsetof
3963 @opindex Winvalid-offsetof
3964 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
3965 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3966 to a non-POD type is undefined. In existing C++ implementations,
3967 however, @samp{offsetof} typically gives meaningful results even when
3968 applied to certain kinds of non-POD types. (Such as a simple
3969 @samp{struct} that fails to be a POD type only by virtue of having a
3970 constructor.) This flag is for users who are aware that they are
3971 writing nonportable code and who have deliberately chosen to ignore the
3974 The restrictions on @samp{offsetof} may be relaxed in a future version
3975 of the C++ standard.
3977 @item -Wno-int-to-pointer-cast @r{(C and Objective-C only)}
3978 @opindex Wno-int-to-pointer-cast
3979 @opindex Wint-to-pointer-cast
3980 Suppress warnings from casts to pointer type of an integer of a
3983 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
3984 @opindex Wno-pointer-to-int-cast
3985 @opindex Wpointer-to-int-cast
3986 Suppress warnings from casts from a pointer to an integer type of a
3990 @opindex Winvalid-pch
3991 @opindex Wno-invalid-pch
3992 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3993 the search path but can't be used.
3997 @opindex Wno-long-long
3998 Warn if @samp{long long} type is used. This is default. To inhibit
3999 the warning messages, use @option{-Wno-long-long}. Flags
4000 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
4001 only when @option{-pedantic} flag is used.
4003 @item -Wvariadic-macros
4004 @opindex Wvariadic-macros
4005 @opindex Wno-variadic-macros
4006 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
4007 alternate syntax when in pedantic ISO C99 mode. This is default.
4008 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
4013 Warn if variable length array is used in the code.
4014 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
4015 the variable length array.
4017 @item -Wvolatile-register-var
4018 @opindex Wvolatile-register-var
4019 @opindex Wno-volatile-register-var
4020 Warn if a register variable is declared volatile. The volatile
4021 modifier does not inhibit all optimizations that may eliminate reads
4022 and/or writes to register variables. This warning is enabled by
4025 @item -Wdisabled-optimization
4026 @opindex Wdisabled-optimization
4027 @opindex Wno-disabled-optimization
4028 Warn if a requested optimization pass is disabled. This warning does
4029 not generally indicate that there is anything wrong with your code; it
4030 merely indicates that GCC's optimizers were unable to handle the code
4031 effectively. Often, the problem is that your code is too big or too
4032 complex; GCC will refuse to optimize programs when the optimization
4033 itself is likely to take inordinate amounts of time.
4035 @item -Wpointer-sign @r{(C and Objective-C only)}
4036 @opindex Wpointer-sign
4037 @opindex Wno-pointer-sign
4038 Warn for pointer argument passing or assignment with different signedness.
4039 This option is only supported for C and Objective-C@. It is implied by
4040 @option{-Wall} and by @option{-pedantic}, which can be disabled with
4041 @option{-Wno-pointer-sign}.
4043 @item -Wstack-protector
4044 @opindex Wstack-protector
4045 @opindex Wno-stack-protector
4046 This option is only active when @option{-fstack-protector} is active. It
4047 warns about functions that will not be protected against stack smashing.
4049 @item -Woverlength-strings
4050 @opindex Woverlength-strings
4051 @opindex Wno-overlength-strings
4052 Warn about string constants which are longer than the ``minimum
4053 maximum'' length specified in the C standard. Modern compilers
4054 generally allow string constants which are much longer than the
4055 standard's minimum limit, but very portable programs should avoid
4056 using longer strings.
4058 The limit applies @emph{after} string constant concatenation, and does
4059 not count the trailing NUL@. In C89, the limit was 509 characters; in
4060 C99, it was raised to 4095. C++98 does not specify a normative
4061 minimum maximum, so we do not diagnose overlength strings in C++@.
4063 This option is implied by @option{-pedantic}, and can be disabled with
4064 @option{-Wno-overlength-strings}.
4067 @node Debugging Options
4068 @section Options for Debugging Your Program or GCC
4069 @cindex options, debugging
4070 @cindex debugging information options
4072 GCC has various special options that are used for debugging
4073 either your program or GCC:
4078 Produce debugging information in the operating system's native format
4079 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
4082 On most systems that use stabs format, @option{-g} enables use of extra
4083 debugging information that only GDB can use; this extra information
4084 makes debugging work better in GDB but will probably make other debuggers
4086 refuse to read the program. If you want to control for certain whether
4087 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
4088 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
4090 GCC allows you to use @option{-g} with
4091 @option{-O}. The shortcuts taken by optimized code may occasionally
4092 produce surprising results: some variables you declared may not exist
4093 at all; flow of control may briefly move where you did not expect it;
4094 some statements may not be executed because they compute constant
4095 results or their values were already at hand; some statements may
4096 execute in different places because they were moved out of loops.
4098 Nevertheless it proves possible to debug optimized output. This makes
4099 it reasonable to use the optimizer for programs that might have bugs.
4101 The following options are useful when GCC is generated with the
4102 capability for more than one debugging format.
4106 Produce debugging information for use by GDB@. This means to use the
4107 most expressive format available (DWARF 2, stabs, or the native format
4108 if neither of those are supported), including GDB extensions if at all
4113 Produce debugging information in stabs format (if that is supported),
4114 without GDB extensions. This is the format used by DBX on most BSD
4115 systems. On MIPS, Alpha and System V Release 4 systems this option
4116 produces stabs debugging output which is not understood by DBX or SDB@.
4117 On System V Release 4 systems this option requires the GNU assembler.
4119 @item -feliminate-unused-debug-symbols
4120 @opindex feliminate-unused-debug-symbols
4121 Produce debugging information in stabs format (if that is supported),
4122 for only symbols that are actually used.
4124 @item -femit-class-debug-always
4125 Instead of emitting debugging information for a C++ class in only one
4126 object file, emit it in all object files using the class. This option
4127 should be used only with debuggers that are unable to handle the way GCC
4128 normally emits debugging information for classes because using this
4129 option will increase the size of debugging information by as much as a
4134 Produce debugging information in stabs format (if that is supported),
4135 using GNU extensions understood only by the GNU debugger (GDB)@. The
4136 use of these extensions is likely to make other debuggers crash or
4137 refuse to read the program.
4141 Produce debugging information in COFF format (if that is supported).
4142 This is the format used by SDB on most System V systems prior to
4147 Produce debugging information in XCOFF format (if that is supported).
4148 This is the format used by the DBX debugger on IBM RS/6000 systems.
4152 Produce debugging information in XCOFF format (if that is supported),
4153 using GNU extensions understood only by the GNU debugger (GDB)@. The
4154 use of these extensions is likely to make other debuggers crash or
4155 refuse to read the program, and may cause assemblers other than the GNU
4156 assembler (GAS) to fail with an error.
4160 Produce debugging information in DWARF version 2 format (if that is
4161 supported). This is the format used by DBX on IRIX 6. With this
4162 option, GCC uses features of DWARF version 3 when they are useful;
4163 version 3 is upward compatible with version 2, but may still cause
4164 problems for older debuggers.
4168 Produce debugging information in VMS debug format (if that is
4169 supported). This is the format used by DEBUG on VMS systems.
4172 @itemx -ggdb@var{level}
4173 @itemx -gstabs@var{level}
4174 @itemx -gcoff@var{level}
4175 @itemx -gxcoff@var{level}
4176 @itemx -gvms@var{level}
4177 Request debugging information and also use @var{level} to specify how
4178 much information. The default level is 2.
4180 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4183 Level 1 produces minimal information, enough for making backtraces in
4184 parts of the program that you don't plan to debug. This includes
4185 descriptions of functions and external variables, but no information
4186 about local variables and no line numbers.
4188 Level 3 includes extra information, such as all the macro definitions
4189 present in the program. Some debuggers support macro expansion when
4190 you use @option{-g3}.
4192 @option{-gdwarf-2} does not accept a concatenated debug level, because
4193 GCC used to support an option @option{-gdwarf} that meant to generate
4194 debug information in version 1 of the DWARF format (which is very
4195 different from version 2), and it would have been too confusing. That
4196 debug format is long obsolete, but the option cannot be changed now.
4197 Instead use an additional @option{-g@var{level}} option to change the
4198 debug level for DWARF2.
4200 @item -feliminate-dwarf2-dups
4201 @opindex feliminate-dwarf2-dups
4202 Compress DWARF2 debugging information by eliminating duplicated
4203 information about each symbol. This option only makes sense when
4204 generating DWARF2 debugging information with @option{-gdwarf-2}.
4206 @item -femit-struct-debug-baseonly
4207 Emit debug information for struct-like types
4208 only when the base name of the compilation source file
4209 matches the base name of file in which the struct was defined.
4211 This option substantially reduces the size of debugging information,
4212 but at significant potential loss in type information to the debugger.
4213 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4214 See @option{-femit-struct-debug-detailed} for more detailed control.
4216 This option works only with DWARF 2.
4218 @item -femit-struct-debug-reduced
4219 Emit debug information for struct-like types
4220 only when the base name of the compilation source file
4221 matches the base name of file in which the type was defined,
4222 unless the struct is a template or defined in a system header.
4224 This option significantly reduces the size of debugging information,
4225 with some potential loss in type information to the debugger.
4226 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4227 See @option{-femit-struct-debug-detailed} for more detailed control.
4229 This option works only with DWARF 2.
4231 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4232 Specify the struct-like types
4233 for which the compiler will generate debug information.
4234 The intent is to reduce duplicate struct debug information
4235 between different object files within the same program.
4237 This option is a detailed version of
4238 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4239 which will serve for most needs.
4241 A specification has the syntax
4242 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4244 The optional first word limits the specification to
4245 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4246 A struct type is used directly when it is the type of a variable, member.
4247 Indirect uses arise through pointers to structs.
4248 That is, when use of an incomplete struct would be legal, the use is indirect.
4250 @samp{struct one direct; struct two * indirect;}.
4252 The optional second word limits the specification to
4253 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4254 Generic structs are a bit complicated to explain.
4255 For C++, these are non-explicit specializations of template classes,
4256 or non-template classes within the above.
4257 Other programming languages have generics,
4258 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4260 The third word specifies the source files for those
4261 structs for which the compiler will emit debug information.
4262 The values @samp{none} and @samp{any} have the normal meaning.
4263 The value @samp{base} means that
4264 the base of name of the file in which the type declaration appears
4265 must match the base of the name of the main compilation file.
4266 In practice, this means that
4267 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4268 but types declared in other header will not.
4269 The value @samp{sys} means those types satisfying @samp{base}
4270 or declared in system or compiler headers.
4272 You may need to experiment to determine the best settings for your application.
4274 The default is @samp{-femit-struct-debug-detailed=all}.
4276 This option works only with DWARF 2.
4278 @item -fno-merge-debug-strings
4279 @opindex fmerge-debug-strings
4280 @opindex fno-merge-debug-strings
4281 Direct the linker to merge together strings which are identical in
4282 different object files. This is not supported by all assemblers or
4283 linker. This decreases the size of the debug information in the
4284 output file at the cost of increasing link processing time. This is
4287 @item -fdebug-prefix-map=@var{old}=@var{new}
4288 @opindex fdebug-prefix-map
4289 When compiling files in directory @file{@var{old}}, record debugging
4290 information describing them as in @file{@var{new}} instead.
4292 @cindex @command{prof}
4295 Generate extra code to write profile information suitable for the
4296 analysis program @command{prof}. You must use this option when compiling
4297 the source files you want data about, and you must also use it when
4300 @cindex @command{gprof}
4303 Generate extra code to write profile information suitable for the
4304 analysis program @command{gprof}. You must use this option when compiling
4305 the source files you want data about, and you must also use it when
4310 Makes the compiler print out each function name as it is compiled, and
4311 print some statistics about each pass when it finishes.
4314 @opindex ftime-report
4315 Makes the compiler print some statistics about the time consumed by each
4316 pass when it finishes.
4319 @opindex fmem-report
4320 Makes the compiler print some statistics about permanent memory
4321 allocation when it finishes.
4323 @item -fpre-ipa-mem-report
4324 @opindex fpre-ipa-mem-report
4325 @item -fpost-ipa-mem-report
4326 @opindex fpost-ipa-mem-report
4327 Makes the compiler print some statistics about permanent memory
4328 allocation before or after interprocedural optimization.
4330 @item -fprofile-arcs
4331 @opindex fprofile-arcs
4332 Add code so that program flow @dfn{arcs} are instrumented. During
4333 execution the program records how many times each branch and call is
4334 executed and how many times it is taken or returns. When the compiled
4335 program exits it saves this data to a file called
4336 @file{@var{auxname}.gcda} for each source file. The data may be used for
4337 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4338 test coverage analysis (@option{-ftest-coverage}). Each object file's
4339 @var{auxname} is generated from the name of the output file, if
4340 explicitly specified and it is not the final executable, otherwise it is
4341 the basename of the source file. In both cases any suffix is removed
4342 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4343 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4344 @xref{Cross-profiling}.
4346 @cindex @command{gcov}
4350 This option is used to compile and link code instrumented for coverage
4351 analysis. The option is a synonym for @option{-fprofile-arcs}
4352 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4353 linking). See the documentation for those options for more details.
4358 Compile the source files with @option{-fprofile-arcs} plus optimization
4359 and code generation options. For test coverage analysis, use the
4360 additional @option{-ftest-coverage} option. You do not need to profile
4361 every source file in a program.
4364 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4365 (the latter implies the former).
4368 Run the program on a representative workload to generate the arc profile
4369 information. This may be repeated any number of times. You can run
4370 concurrent instances of your program, and provided that the file system
4371 supports locking, the data files will be correctly updated. Also
4372 @code{fork} calls are detected and correctly handled (double counting
4376 For profile-directed optimizations, compile the source files again with
4377 the same optimization and code generation options plus
4378 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4379 Control Optimization}).
4382 For test coverage analysis, use @command{gcov} to produce human readable
4383 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4384 @command{gcov} documentation for further information.
4388 With @option{-fprofile-arcs}, for each function of your program GCC
4389 creates a program flow graph, then finds a spanning tree for the graph.
4390 Only arcs that are not on the spanning tree have to be instrumented: the
4391 compiler adds code to count the number of times that these arcs are
4392 executed. When an arc is the only exit or only entrance to a block, the
4393 instrumentation code can be added to the block; otherwise, a new basic
4394 block must be created to hold the instrumentation code.
4397 @item -ftest-coverage
4398 @opindex ftest-coverage
4399 Produce a notes file that the @command{gcov} code-coverage utility
4400 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4401 show program coverage. Each source file's note file is called
4402 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4403 above for a description of @var{auxname} and instructions on how to
4404 generate test coverage data. Coverage data will match the source files
4405 more closely, if you do not optimize.
4407 @item -fdbg-cnt-list
4408 @opindex fdbg-cnt-list
4409 Print the name and the counter upperbound for all debug counters.
4411 @item -fdbg-cnt=@var{counter-value-list}
4413 Set the internal debug counter upperbound. @var{counter-value-list}
4414 is a comma-separated list of @var{name}:@var{value} pairs
4415 which sets the upperbound of each debug counter @var{name} to @var{value}.
4416 All debug counters have the initial upperbound of @var{UINT_MAX},
4417 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4418 e.g. With -fdbg-cnt=dce:10,tail_call:0
4419 dbg_cnt(dce) will return true only for first 10 invocations
4420 and dbg_cnt(tail_call) will return false always.
4422 @item -d@var{letters}
4423 @item -fdump-rtl-@var{pass}
4425 Says to make debugging dumps during compilation at times specified by
4426 @var{letters}. This is used for debugging the RTL-based passes of the
4427 compiler. The file names for most of the dumps are made by appending a
4428 pass number and a word to the @var{dumpname}. @var{dumpname} is generated
4429 from the name of the output file, if explicitly specified and it is not
4430 an executable, otherwise it is the basename of the source file. These
4431 switches may have different effects when @option{-E} is used for
4434 Most debug dumps can be enabled either passing a letter to the @option{-d}
4435 option, or with a long @option{-fdump-rtl} switch; here are the possible
4436 letters for use in @var{letters} and @var{pass}, and their meanings:
4441 Annotate the assembler output with miscellaneous debugging information.
4444 @itemx -fdump-rtl-bbro
4446 @opindex fdump-rtl-bbro
4447 Dump after block reordering, to @file{@var{file}.148r.bbro}.
4450 @itemx -fdump-rtl-combine
4452 @opindex fdump-rtl-combine
4453 Dump after the RTL instruction combination pass, to the file
4454 @file{@var{file}.129r.combine}.
4457 @itemx -fdump-rtl-ce1
4458 @itemx -fdump-rtl-ce2
4460 @opindex fdump-rtl-ce1
4461 @opindex fdump-rtl-ce2
4462 @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the
4463 first if conversion, to the file @file{@var{file}.117r.ce1}. @option{-dC}
4464 and @option{-fdump-rtl-ce2} enable dumping after the second if
4465 conversion, to the file @file{@var{file}.130r.ce2}.
4468 @itemx -fdump-rtl-btl
4469 @itemx -fdump-rtl-dbr
4471 @opindex fdump-rtl-btl
4472 @opindex fdump-rtl-dbr
4473 @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch
4474 target load optimization, to @file{@var{file}.31.btl}. @option{-dd}
4475 and @option{-fdump-rtl-dbr} enable dumping after delayed branch
4476 scheduling, to @file{@var{file}.36.dbr}.
4480 Dump all macro definitions, at the end of preprocessing, in addition to
4484 @itemx -fdump-rtl-ce3
4486 @opindex fdump-rtl-ce3
4487 Dump after the third if conversion, to @file{@var{file}.146r.ce3}.
4490 @itemx -fdump-rtl-cfg
4491 @itemx -fdump-rtl-life
4493 @opindex fdump-rtl-cfg
4494 @opindex fdump-rtl-life
4495 @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control
4496 and data flow analysis, to @file{@var{file}.116r.cfg}. @option{-df}
4497 and @option{-fdump-rtl-cfg} enable dumping dump after life analysis,
4498 to @file{@var{file}.128r.life1} and @file{@var{file}.135r.life2}.
4501 @itemx -fdump-rtl-greg
4503 @opindex fdump-rtl-greg
4504 Dump after global register allocation, to @file{@var{file}.139r.greg}.
4507 @itemx -fdump-rtl-gcse
4508 @itemx -fdump-rtl-bypass
4510 @opindex fdump-rtl-gcse
4511 @opindex fdump-rtl-bypass
4512 @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to
4513 @file{@var{file}.114r.gcse}. @option{-dG} and @option{-fdump-rtl-bypass}
4514 enable dumping after jump bypassing and control flow optimizations, to
4515 @file{@var{file}.115r.bypass}.
4518 @itemx -fdump-rtl-eh
4520 @opindex fdump-rtl-eh
4521 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
4524 @itemx -fdump-rtl-sibling
4526 @opindex fdump-rtl-sibling
4527 Dump after sibling call optimizations, to @file{@var{file}.106r.sibling}.
4530 @itemx -fdump-rtl-jump
4532 @opindex fdump-rtl-jump
4533 Dump after the first jump optimization, to @file{@var{file}.112r.jump}.
4536 @itemx -fdump-rtl-stack
4538 @opindex fdump-rtl-stack
4539 Dump after conversion from GCC's "flat register file" registers to the
4540 x87's stack-like registers, to @file{@var{file}.152r.stack}.
4543 @itemx -fdump-rtl-lreg
4545 @opindex fdump-rtl-lreg
4546 Dump after local register allocation, to @file{@var{file}.138r.lreg}.
4549 @itemx -fdump-rtl-loop2
4551 @opindex fdump-rtl-loop2
4552 @option{-dL} and @option{-fdump-rtl-loop2} enable dumping after the
4553 loop optimization pass, to @file{@var{file}.119r.loop2},
4554 @file{@var{file}.120r.loop2_init},
4555 @file{@var{file}.121r.loop2_invariant}, and
4556 @file{@var{file}.125r.loop2_done}.
4559 @itemx -fdump-rtl-sms
4561 @opindex fdump-rtl-sms
4562 Dump after modulo scheduling, to @file{@var{file}.136r.sms}.
4565 @itemx -fdump-rtl-mach
4567 @opindex fdump-rtl-mach
4568 Dump after performing the machine dependent reorganization pass, to
4569 @file{@var{file}.155r.mach} if that pass exists.
4572 @itemx -fdump-rtl-rnreg
4574 @opindex fdump-rtl-rnreg
4575 Dump after register renumbering, to @file{@var{file}.147r.rnreg}.
4578 @itemx -fdump-rtl-regmove
4580 @opindex fdump-rtl-regmove
4581 Dump after the register move pass, to @file{@var{file}.132r.regmove}.
4584 @itemx -fdump-rtl-postreload
4586 @opindex fdump-rtl-postreload
4587 Dump after post-reload optimizations, to @file{@var{file}.24.postreload}.
4590 @itemx -fdump-rtl-expand
4592 @opindex fdump-rtl-expand
4593 Dump after RTL generation, to @file{@var{file}.104r.expand}.
4596 @itemx -fdump-rtl-sched2
4598 @opindex fdump-rtl-sched2
4599 Dump after the second scheduling pass, to @file{@var{file}.149r.sched2}.
4602 @itemx -fdump-rtl-cse
4604 @opindex fdump-rtl-cse
4605 Dump after CSE (including the jump optimization that sometimes follows
4606 CSE), to @file{@var{file}.113r.cse}.
4609 @itemx -fdump-rtl-sched1
4611 @opindex fdump-rtl-sched1
4612 Dump after the first scheduling pass, to @file{@var{file}.136r.sched1}.
4615 @itemx -fdump-rtl-cse2
4617 @opindex fdump-rtl-cse2
4618 Dump after the second CSE pass (including the jump optimization that
4619 sometimes follows CSE), to @file{@var{file}.127r.cse2}.
4622 @itemx -fdump-rtl-tracer
4624 @opindex fdump-rtl-tracer
4625 Dump after running tracer, to @file{@var{file}.118r.tracer}.
4628 @itemx -fdump-rtl-vpt
4629 @itemx -fdump-rtl-vartrack
4631 @opindex fdump-rtl-vpt
4632 @opindex fdump-rtl-vartrack
4633 @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value
4634 profile transformations, to @file{@var{file}.10.vpt}. @option{-dV}
4635 and @option{-fdump-rtl-vartrack} enable dumping after variable tracking,
4636 to @file{@var{file}.154r.vartrack}.
4639 @itemx -fdump-rtl-flow2
4641 @opindex fdump-rtl-flow2
4642 Dump after the second flow pass, to @file{@var{file}.142r.flow2}.
4645 @itemx -fdump-rtl-peephole2
4647 @opindex fdump-rtl-peephole2
4648 Dump after the peephole pass, to @file{@var{file}.145r.peephole2}.
4651 @itemx -fdump-rtl-web
4653 @opindex fdump-rtl-web
4654 Dump after live range splitting, to @file{@var{file}.126r.web}.
4657 @itemx -fdump-rtl-all
4659 @opindex fdump-rtl-all
4660 Produce all the dumps listed above.
4664 Produce a core dump whenever an error occurs.
4668 Print statistics on memory usage, at the end of the run, to
4673 Annotate the assembler output with a comment indicating which
4674 pattern and alternative was used. The length of each instruction is
4679 Dump the RTL in the assembler output as a comment before each instruction.
4680 Also turns on @option{-dp} annotation.
4684 For each of the other indicated dump files (either with @option{-d} or
4685 @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow
4686 graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
4690 Just generate RTL for a function instead of compiling it. Usually used
4691 with @samp{r} (@option{-fdump-rtl-expand}).
4695 Dump debugging information during parsing, to standard error.
4699 @opindex fdump-noaddr
4700 When doing debugging dumps (see @option{-d} option above), suppress
4701 address output. This makes it more feasible to use diff on debugging
4702 dumps for compiler invocations with different compiler binaries and/or
4703 different text / bss / data / heap / stack / dso start locations.
4705 @item -fdump-unnumbered
4706 @opindex fdump-unnumbered
4707 When doing debugging dumps (see @option{-d} option above), suppress instruction
4708 numbers and address output. This makes it more feasible to
4709 use diff on debugging dumps for compiler invocations with different
4710 options, in particular with and without @option{-g}.
4712 @item -fdump-translation-unit @r{(C++ only)}
4713 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
4714 @opindex fdump-translation-unit
4715 Dump a representation of the tree structure for the entire translation
4716 unit to a file. The file name is made by appending @file{.tu} to the
4717 source file name. If the @samp{-@var{options}} form is used, @var{options}
4718 controls the details of the dump as described for the
4719 @option{-fdump-tree} options.
4721 @item -fdump-class-hierarchy @r{(C++ only)}
4722 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
4723 @opindex fdump-class-hierarchy
4724 Dump a representation of each class's hierarchy and virtual function
4725 table layout to a file. The file name is made by appending @file{.class}
4726 to the source file name. If the @samp{-@var{options}} form is used,
4727 @var{options} controls the details of the dump as described for the
4728 @option{-fdump-tree} options.
4730 @item -fdump-ipa-@var{switch}
4732 Control the dumping at various stages of inter-procedural analysis
4733 language tree to a file. The file name is generated by appending a switch
4734 specific suffix to the source file name. The following dumps are possible:
4738 Enables all inter-procedural analysis dumps.
4741 Dumps information about call-graph optimization, unused function removal,
4742 and inlining decisions.
4745 Dump after function inlining.
4749 @item -fdump-tree-@var{switch}
4750 @itemx -fdump-tree-@var{switch}-@var{options}
4752 Control the dumping at various stages of processing the intermediate
4753 language tree to a file. The file name is generated by appending a switch
4754 specific suffix to the source file name. If the @samp{-@var{options}}
4755 form is used, @var{options} is a list of @samp{-} separated options that
4756 control the details of the dump. Not all options are applicable to all
4757 dumps, those which are not meaningful will be ignored. The following
4758 options are available
4762 Print the address of each node. Usually this is not meaningful as it
4763 changes according to the environment and source file. Its primary use
4764 is for tying up a dump file with a debug environment.
4766 Inhibit dumping of members of a scope or body of a function merely
4767 because that scope has been reached. Only dump such items when they
4768 are directly reachable by some other path. When dumping pretty-printed
4769 trees, this option inhibits dumping the bodies of control structures.
4771 Print a raw representation of the tree. By default, trees are
4772 pretty-printed into a C-like representation.
4774 Enable more detailed dumps (not honored by every dump option).
4776 Enable dumping various statistics about the pass (not honored by every dump
4779 Enable showing basic block boundaries (disabled in raw dumps).
4781 Enable showing virtual operands for every statement.
4783 Enable showing line numbers for statements.
4785 Enable showing the unique ID (@code{DECL_UID}) for each variable.
4787 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
4790 The following tree dumps are possible:
4794 Dump before any tree based optimization, to @file{@var{file}.original}.
4797 Dump after all tree based optimization, to @file{@var{file}.optimized}.
4800 @opindex fdump-tree-gimple
4801 Dump each function before and after the gimplification pass to a file. The
4802 file name is made by appending @file{.gimple} to the source file name.
4805 @opindex fdump-tree-cfg
4806 Dump the control flow graph of each function to a file. The file name is
4807 made by appending @file{.cfg} to the source file name.
4810 @opindex fdump-tree-vcg
4811 Dump the control flow graph of each function to a file in VCG format. The
4812 file name is made by appending @file{.vcg} to the source file name. Note
4813 that if the file contains more than one function, the generated file cannot
4814 be used directly by VCG@. You will need to cut and paste each function's
4815 graph into its own separate file first.
4818 @opindex fdump-tree-ch
4819 Dump each function after copying loop headers. The file name is made by
4820 appending @file{.ch} to the source file name.
4823 @opindex fdump-tree-ssa
4824 Dump SSA related information to a file. The file name is made by appending
4825 @file{.ssa} to the source file name.
4828 @opindex fdump-tree-salias
4829 Dump structure aliasing variable information to a file. This file name
4830 is made by appending @file{.salias} to the source file name.
4833 @opindex fdump-tree-alias
4834 Dump aliasing information for each function. The file name is made by
4835 appending @file{.alias} to the source file name.
4838 @opindex fdump-tree-ccp
4839 Dump each function after CCP@. The file name is made by appending
4840 @file{.ccp} to the source file name.
4843 @opindex fdump-tree-storeccp
4844 Dump each function after STORE-CCP@. The file name is made by appending
4845 @file{.storeccp} to the source file name.
4848 @opindex fdump-tree-pre
4849 Dump trees after partial redundancy elimination. The file name is made
4850 by appending @file{.pre} to the source file name.
4853 @opindex fdump-tree-fre
4854 Dump trees after full redundancy elimination. The file name is made
4855 by appending @file{.fre} to the source file name.
4858 @opindex fdump-tree-copyprop
4859 Dump trees after copy propagation. The file name is made
4860 by appending @file{.copyprop} to the source file name.
4862 @item store_copyprop
4863 @opindex fdump-tree-store_copyprop
4864 Dump trees after store copy-propagation. The file name is made
4865 by appending @file{.store_copyprop} to the source file name.
4868 @opindex fdump-tree-dce
4869 Dump each function after dead code elimination. The file name is made by
4870 appending @file{.dce} to the source file name.
4873 @opindex fdump-tree-mudflap
4874 Dump each function after adding mudflap instrumentation. The file name is
4875 made by appending @file{.mudflap} to the source file name.
4878 @opindex fdump-tree-sra
4879 Dump each function after performing scalar replacement of aggregates. The
4880 file name is made by appending @file{.sra} to the source file name.
4883 @opindex fdump-tree-sink
4884 Dump each function after performing code sinking. The file name is made
4885 by appending @file{.sink} to the source file name.
4888 @opindex fdump-tree-dom
4889 Dump each function after applying dominator tree optimizations. The file
4890 name is made by appending @file{.dom} to the source file name.
4893 @opindex fdump-tree-dse
4894 Dump each function after applying dead store elimination. The file
4895 name is made by appending @file{.dse} to the source file name.
4898 @opindex fdump-tree-phiopt
4899 Dump each function after optimizing PHI nodes into straightline code. The file
4900 name is made by appending @file{.phiopt} to the source file name.
4903 @opindex fdump-tree-forwprop
4904 Dump each function after forward propagating single use variables. The file
4905 name is made by appending @file{.forwprop} to the source file name.
4908 @opindex fdump-tree-copyrename
4909 Dump each function after applying the copy rename optimization. The file
4910 name is made by appending @file{.copyrename} to the source file name.
4913 @opindex fdump-tree-nrv
4914 Dump each function after applying the named return value optimization on
4915 generic trees. The file name is made by appending @file{.nrv} to the source
4919 @opindex fdump-tree-vect
4920 Dump each function after applying vectorization of loops. The file name is
4921 made by appending @file{.vect} to the source file name.
4924 @opindex fdump-tree-vrp
4925 Dump each function after Value Range Propagation (VRP). The file name
4926 is made by appending @file{.vrp} to the source file name.
4929 @opindex fdump-tree-all
4930 Enable all the available tree dumps with the flags provided in this option.
4933 @item -ftree-vectorizer-verbose=@var{n}
4934 @opindex ftree-vectorizer-verbose
4935 This option controls the amount of debugging output the vectorizer prints.
4936 This information is written to standard error, unless
4937 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
4938 in which case it is output to the usual dump listing file, @file{.vect}.
4939 For @var{n}=0 no diagnostic information is reported.
4940 If @var{n}=1 the vectorizer reports each loop that got vectorized,
4941 and the total number of loops that got vectorized.
4942 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
4943 the first analysis phase (vect_analyze_loop_form) - i.e.@: countable,
4944 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
4945 level that @option{-fdump-tree-vect-stats} uses.
4946 Higher verbosity levels mean either more information dumped for each
4947 reported loop, or same amount of information reported for more loops:
4948 If @var{n}=3, alignment related information is added to the reports.
4949 If @var{n}=4, data-references related information (e.g.@: memory dependences,
4950 memory access-patterns) is added to the reports.
4951 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
4952 that did not pass the first analysis phase (i.e., may not be countable, or
4953 may have complicated control-flow).
4954 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
4955 For @var{n}=7, all the information the vectorizer generates during its
4956 analysis and transformation is reported. This is the same verbosity level
4957 that @option{-fdump-tree-vect-details} uses.
4959 @item -frandom-seed=@var{string}
4960 @opindex frandom-string
4961 This option provides a seed that GCC uses when it would otherwise use
4962 random numbers. It is used to generate certain symbol names
4963 that have to be different in every compiled file. It is also used to
4964 place unique stamps in coverage data files and the object files that
4965 produce them. You can use the @option{-frandom-seed} option to produce
4966 reproducibly identical object files.
4968 The @var{string} should be different for every file you compile.
4970 @item -fsched-verbose=@var{n}
4971 @opindex fsched-verbose
4972 On targets that use instruction scheduling, this option controls the
4973 amount of debugging output the scheduler prints. This information is
4974 written to standard error, unless @option{-dS} or @option{-dR} is
4975 specified, in which case it is output to the usual dump
4976 listing file, @file{.sched} or @file{.sched2} respectively. However
4977 for @var{n} greater than nine, the output is always printed to standard
4980 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
4981 same information as @option{-dRS}. For @var{n} greater than one, it
4982 also output basic block probabilities, detailed ready list information
4983 and unit/insn info. For @var{n} greater than two, it includes RTL
4984 at abort point, control-flow and regions info. And for @var{n} over
4985 four, @option{-fsched-verbose} also includes dependence info.
4989 Store the usual ``temporary'' intermediate files permanently; place them
4990 in the current directory and name them based on the source file. Thus,
4991 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
4992 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
4993 preprocessed @file{foo.i} output file even though the compiler now
4994 normally uses an integrated preprocessor.
4996 When used in combination with the @option{-x} command line option,
4997 @option{-save-temps} is sensible enough to avoid over writing an
4998 input source file with the same extension as an intermediate file.
4999 The corresponding intermediate file may be obtained by renaming the
5000 source file before using @option{-save-temps}.
5004 Report the CPU time taken by each subprocess in the compilation
5005 sequence. For C source files, this is the compiler proper and assembler
5006 (plus the linker if linking is done). The output looks like this:
5013 The first number on each line is the ``user time'', that is time spent
5014 executing the program itself. The second number is ``system time'',
5015 time spent executing operating system routines on behalf of the program.
5016 Both numbers are in seconds.
5018 @item -fvar-tracking
5019 @opindex fvar-tracking
5020 Run variable tracking pass. It computes where variables are stored at each
5021 position in code. Better debugging information is then generated
5022 (if the debugging information format supports this information).
5024 It is enabled by default when compiling with optimization (@option{-Os},
5025 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
5026 the debug info format supports it.
5028 @item -print-file-name=@var{library}
5029 @opindex print-file-name
5030 Print the full absolute name of the library file @var{library} that
5031 would be used when linking---and don't do anything else. With this
5032 option, GCC does not compile or link anything; it just prints the
5035 @item -print-multi-directory
5036 @opindex print-multi-directory
5037 Print the directory name corresponding to the multilib selected by any
5038 other switches present in the command line. This directory is supposed
5039 to exist in @env{GCC_EXEC_PREFIX}.
5041 @item -print-multi-lib
5042 @opindex print-multi-lib
5043 Print the mapping from multilib directory names to compiler switches
5044 that enable them. The directory name is separated from the switches by
5045 @samp{;}, and each switch starts with an @samp{@@} instead of the
5046 @samp{-}, without spaces between multiple switches. This is supposed to
5047 ease shell-processing.
5049 @item -print-prog-name=@var{program}
5050 @opindex print-prog-name
5051 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
5053 @item -print-libgcc-file-name
5054 @opindex print-libgcc-file-name
5055 Same as @option{-print-file-name=libgcc.a}.
5057 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
5058 but you do want to link with @file{libgcc.a}. You can do
5061 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
5064 @item -print-search-dirs
5065 @opindex print-search-dirs
5066 Print the name of the configured installation directory and a list of
5067 program and library directories @command{gcc} will search---and don't do anything else.
5069 This is useful when @command{gcc} prints the error message
5070 @samp{installation problem, cannot exec cpp0: No such file or directory}.
5071 To resolve this you either need to put @file{cpp0} and the other compiler
5072 components where @command{gcc} expects to find them, or you can set the environment
5073 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
5074 Don't forget the trailing @samp{/}.
5075 @xref{Environment Variables}.
5077 @item -print-sysroot-headers-suffix
5078 @opindex print-sysroot-headers-suffix
5079 Print the suffix added to the target sysroot when searching for
5080 headers, or give an error if the compiler is not configured with such
5081 a suffix---and don't do anything else.
5084 @opindex dumpmachine
5085 Print the compiler's target machine (for example,
5086 @samp{i686-pc-linux-gnu})---and don't do anything else.
5089 @opindex dumpversion
5090 Print the compiler version (for example, @samp{3.0})---and don't do
5095 Print the compiler's built-in specs---and don't do anything else. (This
5096 is used when GCC itself is being built.) @xref{Spec Files}.
5098 @item -feliminate-unused-debug-types
5099 @opindex feliminate-unused-debug-types
5100 Normally, when producing DWARF2 output, GCC will emit debugging
5101 information for all types declared in a compilation
5102 unit, regardless of whether or not they are actually used
5103 in that compilation unit. Sometimes this is useful, such as
5104 if, in the debugger, you want to cast a value to a type that is
5105 not actually used in your program (but is declared). More often,
5106 however, this results in a significant amount of wasted space.
5107 With this option, GCC will avoid producing debug symbol output
5108 for types that are nowhere used in the source file being compiled.
5111 @node Optimize Options
5112 @section Options That Control Optimization
5113 @cindex optimize options
5114 @cindex options, optimization
5116 These options control various sorts of optimizations.
5118 Without any optimization option, the compiler's goal is to reduce the
5119 cost of compilation and to make debugging produce the expected
5120 results. Statements are independent: if you stop the program with a
5121 breakpoint between statements, you can then assign a new value to any
5122 variable or change the program counter to any other statement in the
5123 function and get exactly the results you would expect from the source
5126 Turning on optimization flags makes the compiler attempt to improve
5127 the performance and/or code size at the expense of compilation time
5128 and possibly the ability to debug the program.
5130 The compiler performs optimization based on the knowledge it has of
5131 the program. Optimization levels @option{-O} and above, in
5132 particular, enable @emph{unit-at-a-time} mode, which allows the
5133 compiler to consider information gained from later functions in
5134 the file when compiling a function. Compiling multiple files at
5135 once to a single output file in @emph{unit-at-a-time} mode allows
5136 the compiler to use information gained from all of the files when
5137 compiling each of them.
5139 Not all optimizations are controlled directly by a flag. Only
5140 optimizations that have a flag are listed.
5147 Optimize. Optimizing compilation takes somewhat more time, and a lot
5148 more memory for a large function.
5150 With @option{-O}, the compiler tries to reduce code size and execution
5151 time, without performing any optimizations that take a great deal of
5154 @option{-O} turns on the following optimization flags:
5157 -fcprop-registers @gol
5160 -fdelayed-branch @gol
5162 -fguess-branch-probability @gol
5163 -fif-conversion2 @gol
5164 -fif-conversion @gol
5165 -finline-small-functions @gol
5166 -fipa-pure-const @gol
5167 -fipa-reference @gol
5169 -fsplit-wide-types @gol
5172 -ftree-copyrename @gol
5174 -ftree-dominator-opts @gol
5181 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5182 where doing so does not interfere with debugging.
5186 Optimize even more. GCC performs nearly all supported optimizations
5187 that do not involve a space-speed tradeoff. The compiler does not
5188 perform loop unrolling or function inlining when you specify @option{-O2}.
5189 As compared to @option{-O}, this option increases both compilation time
5190 and the performance of the generated code.
5192 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5193 also turns on the following optimization flags:
5194 @gccoptlist{-fthread-jumps @gol
5195 -falign-functions -falign-jumps @gol
5196 -falign-loops -falign-labels @gol
5199 -fcse-follow-jumps -fcse-skip-blocks @gol
5200 -fdelete-null-pointer-checks @gol
5201 -fexpensive-optimizations @gol
5202 -fgcse -fgcse-lm @gol
5203 -foptimize-sibling-calls @gol
5206 -freorder-blocks -freorder-functions @gol
5207 -frerun-cse-after-loop @gol
5208 -fsched-interblock -fsched-spec @gol
5209 -fschedule-insns -fschedule-insns2 @gol
5210 -fstrict-aliasing -fstrict-overflow @gol
5214 Please note the warning under @option{-fgcse} about
5215 invoking @option{-O2} on programs that use computed gotos.
5219 Optimize yet more. @option{-O3} turns on all optimizations specified by
5220 @option{-O2} and also turns on the @option{-finline-functions},
5221 @option{-funswitch-loops}, @option{-fpredictive-commoning} and
5222 @option{-fgcse-after-reload} options.
5226 Reduce compilation time and make debugging produce the expected
5227 results. This is the default.
5231 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5232 do not typically increase code size. It also performs further
5233 optimizations designed to reduce code size.
5235 @option{-Os} disables the following optimization flags:
5236 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5237 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5238 -fprefetch-loop-arrays -ftree-vect-loop-version}
5240 If you use multiple @option{-O} options, with or without level numbers,
5241 the last such option is the one that is effective.
5244 Options of the form @option{-f@var{flag}} specify machine-independent
5245 flags. Most flags have both positive and negative forms; the negative
5246 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5247 below, only one of the forms is listed---the one you typically will
5248 use. You can figure out the other form by either removing @samp{no-}
5251 The following options control specific optimizations. They are either
5252 activated by @option{-O} options or are related to ones that are. You
5253 can use the following flags in the rare cases when ``fine-tuning'' of
5254 optimizations to be performed is desired.
5257 @item -fno-default-inline
5258 @opindex fno-default-inline
5259 Do not make member functions inline by default merely because they are
5260 defined inside the class scope (C++ only). Otherwise, when you specify
5261 @w{@option{-O}}, member functions defined inside class scope are compiled
5262 inline by default; i.e., you don't need to add @samp{inline} in front of
5263 the member function name.
5265 @item -fno-defer-pop
5266 @opindex fno-defer-pop
5267 Always pop the arguments to each function call as soon as that function
5268 returns. For machines which must pop arguments after a function call,
5269 the compiler normally lets arguments accumulate on the stack for several
5270 function calls and pops them all at once.
5272 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5274 @item -fforward-propagate
5275 @opindex fforward-propagate
5276 Perform a forward propagation pass on RTL@. The pass tries to combine two
5277 instructions and checks if the result can be simplified. If loop unrolling
5278 is active, two passes are performed and the second is scheduled after
5281 This option is enabled by default at optimization levels @option{-O2},
5282 @option{-O3}, @option{-Os}.
5284 @item -fomit-frame-pointer
5285 @opindex fomit-frame-pointer
5286 Don't keep the frame pointer in a register for functions that
5287 don't need one. This avoids the instructions to save, set up and
5288 restore frame pointers; it also makes an extra register available
5289 in many functions. @strong{It also makes debugging impossible on
5292 On some machines, such as the VAX, this flag has no effect, because
5293 the standard calling sequence automatically handles the frame pointer
5294 and nothing is saved by pretending it doesn't exist. The
5295 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5296 whether a target machine supports this flag. @xref{Registers,,Register
5297 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5299 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5301 @item -foptimize-sibling-calls
5302 @opindex foptimize-sibling-calls
5303 Optimize sibling and tail recursive calls.
5305 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5309 Don't pay attention to the @code{inline} keyword. Normally this option
5310 is used to keep the compiler from expanding any functions inline.
5311 Note that if you are not optimizing, no functions can be expanded inline.
5313 @item -finline-small-functions
5314 @opindex finline-small-functions
5315 Integrate functions into their callers when their body is smaller than expected
5316 function call code (so overall size of program gets smaller). The compiler
5317 heuristically decides which functions are simple enough to be worth integrating
5320 Enabled at level @option{-O2}.
5322 @item -finline-functions
5323 @opindex finline-functions
5324 Integrate all simple functions into their callers. The compiler
5325 heuristically decides which functions are simple enough to be worth
5326 integrating in this way.
5328 If all calls to a given function are integrated, and the function is
5329 declared @code{static}, then the function is normally not output as
5330 assembler code in its own right.
5332 Enabled at level @option{-O3}.
5334 @item -finline-functions-called-once
5335 @opindex finline-functions-called-once
5336 Consider all @code{static} functions called once for inlining into their
5337 caller even if they are not marked @code{inline}. If a call to a given
5338 function is integrated, then the function is not output as assembler code
5341 Enabled if @option{-funit-at-a-time} is enabled.
5343 @item -fearly-inlining
5344 @opindex fearly-inlining
5345 Inline functions marked by @code{always_inline} and functions whose body seems
5346 smaller than the function call overhead early before doing
5347 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5348 makes profiling significantly cheaper and usually inlining faster on programs
5349 having large chains of nested wrapper functions.
5353 @item -finline-limit=@var{n}
5354 @opindex finline-limit
5355 By default, GCC limits the size of functions that can be inlined. This flag
5356 allows coarse control of this limit. @var{n} is the size of functions that
5357 can be inlined in number of pseudo instructions.
5359 Inlining is actually controlled by a number of parameters, which may be
5360 specified individually by using @option{--param @var{name}=@var{value}}.
5361 The @option{-finline-limit=@var{n}} option sets some of these parameters
5365 @item max-inline-insns-single
5366 is set to @var{n}/2.
5367 @item max-inline-insns-auto
5368 is set to @var{n}/2.
5371 See below for a documentation of the individual
5372 parameters controlling inlining and for the defaults of these parameters.
5374 @emph{Note:} there may be no value to @option{-finline-limit} that results
5375 in default behavior.
5377 @emph{Note:} pseudo instruction represents, in this particular context, an
5378 abstract measurement of function's size. In no way does it represent a count
5379 of assembly instructions and as such its exact meaning might change from one
5380 release to an another.
5382 @item -fkeep-inline-functions
5383 @opindex fkeep-inline-functions
5384 In C, emit @code{static} functions that are declared @code{inline}
5385 into the object file, even if the function has been inlined into all
5386 of its callers. This switch does not affect functions using the
5387 @code{extern inline} extension in GNU C89@. In C++, emit any and all
5388 inline functions into the object file.
5390 @item -fkeep-static-consts
5391 @opindex fkeep-static-consts
5392 Emit variables declared @code{static const} when optimization isn't turned
5393 on, even if the variables aren't referenced.
5395 GCC enables this option by default. If you want to force the compiler to
5396 check if the variable was referenced, regardless of whether or not
5397 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5399 @item -fmerge-constants
5400 @opindex fmerge-constants
5401 Attempt to merge identical constants (string constants and floating point
5402 constants) across compilation units.
5404 This option is the default for optimized compilation if the assembler and
5405 linker support it. Use @option{-fno-merge-constants} to inhibit this
5408 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5410 @item -fmerge-all-constants
5411 @opindex fmerge-all-constants
5412 Attempt to merge identical constants and identical variables.
5414 This option implies @option{-fmerge-constants}. In addition to
5415 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5416 arrays or initialized constant variables with integral or floating point
5417 types. Languages like C or C++ require each non-automatic variable to
5418 have distinct location, so using this option will result in non-conforming
5421 @item -fmodulo-sched
5422 @opindex fmodulo-sched
5423 Perform swing modulo scheduling immediately before the first scheduling
5424 pass. This pass looks at innermost loops and reorders their
5425 instructions by overlapping different iterations.
5427 @item -fmodulo-sched-allow-regmoves
5428 @opindex fmodulo-sched-allow-regmoves
5429 Perform more aggressive SMS based modulo scheduling with register moves
5430 allowed. By setting this flag certain anti-dependences edges will be
5431 deleted which will trigger the generation of reg-moves based on the
5432 life-range analysis. This option is effective only with
5433 @option{-fmodulo-sched} enabled.
5435 @item -fno-branch-count-reg
5436 @opindex fno-branch-count-reg
5437 Do not use ``decrement and branch'' instructions on a count register,
5438 but instead generate a sequence of instructions that decrement a
5439 register, compare it against zero, then branch based upon the result.
5440 This option is only meaningful on architectures that support such
5441 instructions, which include x86, PowerPC, IA-64 and S/390.
5443 The default is @option{-fbranch-count-reg}.
5445 @item -fno-function-cse
5446 @opindex fno-function-cse
5447 Do not put function addresses in registers; make each instruction that
5448 calls a constant function contain the function's address explicitly.
5450 This option results in less efficient code, but some strange hacks
5451 that alter the assembler output may be confused by the optimizations
5452 performed when this option is not used.
5454 The default is @option{-ffunction-cse}
5456 @item -fno-zero-initialized-in-bss
5457 @opindex fno-zero-initialized-in-bss
5458 If the target supports a BSS section, GCC by default puts variables that
5459 are initialized to zero into BSS@. This can save space in the resulting
5462 This option turns off this behavior because some programs explicitly
5463 rely on variables going to the data section. E.g., so that the
5464 resulting executable can find the beginning of that section and/or make
5465 assumptions based on that.
5467 The default is @option{-fzero-initialized-in-bss}.
5469 @item -fmudflap -fmudflapth -fmudflapir
5473 @cindex bounds checking
5475 For front-ends that support it (C and C++), instrument all risky
5476 pointer/array dereferencing operations, some standard library
5477 string/heap functions, and some other associated constructs with
5478 range/validity tests. Modules so instrumented should be immune to
5479 buffer overflows, invalid heap use, and some other classes of C/C++
5480 programming errors. The instrumentation relies on a separate runtime
5481 library (@file{libmudflap}), which will be linked into a program if
5482 @option{-fmudflap} is given at link time. Run-time behavior of the
5483 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
5484 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
5487 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
5488 link if your program is multi-threaded. Use @option{-fmudflapir}, in
5489 addition to @option{-fmudflap} or @option{-fmudflapth}, if
5490 instrumentation should ignore pointer reads. This produces less
5491 instrumentation (and therefore faster execution) and still provides
5492 some protection against outright memory corrupting writes, but allows
5493 erroneously read data to propagate within a program.
5495 @item -fthread-jumps
5496 @opindex fthread-jumps
5497 Perform optimizations where we check to see if a jump branches to a
5498 location where another comparison subsumed by the first is found. If
5499 so, the first branch is redirected to either the destination of the
5500 second branch or a point immediately following it, depending on whether
5501 the condition is known to be true or false.
5503 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5505 @item -fsplit-wide-types
5506 @opindex fsplit-wide-types
5507 When using a type that occupies multiple registers, such as @code{long
5508 long} on a 32-bit system, split the registers apart and allocate them
5509 independently. This normally generates better code for those types,
5510 but may make debugging more difficult.
5512 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
5515 @item -fcse-follow-jumps
5516 @opindex fcse-follow-jumps
5517 In common subexpression elimination (CSE), scan through jump instructions
5518 when the target of the jump is not reached by any other path. For
5519 example, when CSE encounters an @code{if} statement with an
5520 @code{else} clause, CSE will follow the jump when the condition
5523 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5525 @item -fcse-skip-blocks
5526 @opindex fcse-skip-blocks
5527 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
5528 follow jumps which conditionally skip over blocks. When CSE
5529 encounters a simple @code{if} statement with no else clause,
5530 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
5531 body of the @code{if}.
5533 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5535 @item -frerun-cse-after-loop
5536 @opindex frerun-cse-after-loop
5537 Re-run common subexpression elimination after loop optimizations has been
5540 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5544 Perform a global common subexpression elimination pass.
5545 This pass also performs global constant and copy propagation.
5547 @emph{Note:} When compiling a program using computed gotos, a GCC
5548 extension, you may get better runtime performance if you disable
5549 the global common subexpression elimination pass by adding
5550 @option{-fno-gcse} to the command line.
5552 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5556 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
5557 attempt to move loads which are only killed by stores into themselves. This
5558 allows a loop containing a load/store sequence to be changed to a load outside
5559 the loop, and a copy/store within the loop.
5561 Enabled by default when gcse is enabled.
5565 When @option{-fgcse-sm} is enabled, a store motion pass is run after
5566 global common subexpression elimination. This pass will attempt to move
5567 stores out of loops. When used in conjunction with @option{-fgcse-lm},
5568 loops containing a load/store sequence can be changed to a load before
5569 the loop and a store after the loop.
5571 Not enabled at any optimization level.
5575 When @option{-fgcse-las} is enabled, the global common subexpression
5576 elimination pass eliminates redundant loads that come after stores to the
5577 same memory location (both partial and full redundancies).
5579 Not enabled at any optimization level.
5581 @item -fgcse-after-reload
5582 @opindex fgcse-after-reload
5583 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
5584 pass is performed after reload. The purpose of this pass is to cleanup
5587 @item -funsafe-loop-optimizations
5588 @opindex funsafe-loop-optimizations
5589 If given, the loop optimizer will assume that loop indices do not
5590 overflow, and that the loops with nontrivial exit condition are not
5591 infinite. This enables a wider range of loop optimizations even if
5592 the loop optimizer itself cannot prove that these assumptions are valid.
5593 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
5594 if it finds this kind of loop.
5596 @item -fcrossjumping
5597 @opindex fcrossjumping
5598 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
5599 resulting code may or may not perform better than without cross-jumping.
5601 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5603 @item -fauto-inc-dec
5604 @opindex fauto-inc-dec
5605 Combine increments or decrements of addresses with memory accesses.
5606 This pass is always skipped on architectures that do not have
5607 instructions to support this. Enabled by default at @option{-O} and
5608 higher on architectures that support this.
5612 Perform dead code elimination (DCE) on RTL@.
5613 Enabled by default at @option{-O} and higher.
5617 Perform dead store elimination (DSE) on RTL@.
5618 Enabled by default at @option{-O} and higher.
5620 @item -fif-conversion
5621 @opindex fif-conversion
5622 Attempt to transform conditional jumps into branch-less equivalents. This
5623 include use of conditional moves, min, max, set flags and abs instructions, and
5624 some tricks doable by standard arithmetics. The use of conditional execution
5625 on chips where it is available is controlled by @code{if-conversion2}.
5627 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5629 @item -fif-conversion2
5630 @opindex fif-conversion2
5631 Use conditional execution (where available) to transform conditional jumps into
5632 branch-less equivalents.
5634 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5636 @item -fdelete-null-pointer-checks
5637 @opindex fdelete-null-pointer-checks
5638 Use global dataflow analysis to identify and eliminate useless checks
5639 for null pointers. The compiler assumes that dereferencing a null
5640 pointer would have halted the program. If a pointer is checked after
5641 it has already been dereferenced, it cannot be null.
5643 In some environments, this assumption is not true, and programs can
5644 safely dereference null pointers. Use
5645 @option{-fno-delete-null-pointer-checks} to disable this optimization
5646 for programs which depend on that behavior.
5648 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5650 @item -fexpensive-optimizations
5651 @opindex fexpensive-optimizations
5652 Perform a number of minor optimizations that are relatively expensive.
5654 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5656 @item -foptimize-register-move
5658 @opindex foptimize-register-move
5660 Attempt to reassign register numbers in move instructions and as
5661 operands of other simple instructions in order to maximize the amount of
5662 register tying. This is especially helpful on machines with two-operand
5665 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
5668 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5670 @item -fdelayed-branch
5671 @opindex fdelayed-branch
5672 If supported for the target machine, attempt to reorder instructions
5673 to exploit instruction slots available after delayed branch
5676 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5678 @item -fschedule-insns
5679 @opindex fschedule-insns
5680 If supported for the target machine, attempt to reorder instructions to
5681 eliminate execution stalls due to required data being unavailable. This
5682 helps machines that have slow floating point or memory load instructions
5683 by allowing other instructions to be issued until the result of the load
5684 or floating point instruction is required.
5686 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5688 @item -fschedule-insns2
5689 @opindex fschedule-insns2
5690 Similar to @option{-fschedule-insns}, but requests an additional pass of
5691 instruction scheduling after register allocation has been done. This is
5692 especially useful on machines with a relatively small number of
5693 registers and where memory load instructions take more than one cycle.
5695 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5697 @item -fno-sched-interblock
5698 @opindex fno-sched-interblock
5699 Don't schedule instructions across basic blocks. This is normally
5700 enabled by default when scheduling before register allocation, i.e.@:
5701 with @option{-fschedule-insns} or at @option{-O2} or higher.
5703 @item -fno-sched-spec
5704 @opindex fno-sched-spec
5705 Don't allow speculative motion of non-load instructions. This is normally
5706 enabled by default when scheduling before register allocation, i.e.@:
5707 with @option{-fschedule-insns} or at @option{-O2} or higher.
5709 @item -fsched-spec-load
5710 @opindex fsched-spec-load
5711 Allow speculative motion of some load instructions. This only makes
5712 sense when scheduling before register allocation, i.e.@: with
5713 @option{-fschedule-insns} or at @option{-O2} or higher.
5715 @item -fsched-spec-load-dangerous
5716 @opindex fsched-spec-load-dangerous
5717 Allow speculative motion of more load instructions. This only makes
5718 sense when scheduling before register allocation, i.e.@: with
5719 @option{-fschedule-insns} or at @option{-O2} or higher.
5721 @item -fsched-stalled-insns
5722 @itemx -fsched-stalled-insns=@var{n}
5723 @opindex fsched-stalled-insns
5724 Define how many insns (if any) can be moved prematurely from the queue
5725 of stalled insns into the ready list, during the second scheduling pass.
5726 @option{-fno-sched-stalled-insns} means that no insns will be moved
5727 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
5728 on how many queued insns can be moved prematurely.
5729 @option{-fsched-stalled-insns} without a value is equivalent to
5730 @option{-fsched-stalled-insns=1}.
5732 @item -fsched-stalled-insns-dep
5733 @itemx -fsched-stalled-insns-dep=@var{n}
5734 @opindex fsched-stalled-insns-dep
5735 Define how many insn groups (cycles) will be examined for a dependency
5736 on a stalled insn that is candidate for premature removal from the queue
5737 of stalled insns. This has an effect only during the second scheduling pass,
5738 and only if @option{-fsched-stalled-insns} is used.
5739 @option{-fno-sched-stalled-insns-dep} is equivalent to
5740 @option{-fsched-stalled-insns-dep=0}.
5741 @option{-fsched-stalled-insns-dep} without a value is equivalent to
5742 @option{-fsched-stalled-insns-dep=1}.
5744 @item -fsched2-use-superblocks
5745 @opindex fsched2-use-superblocks
5746 When scheduling after register allocation, do use superblock scheduling
5747 algorithm. Superblock scheduling allows motion across basic block boundaries
5748 resulting on faster schedules. This option is experimental, as not all machine
5749 descriptions used by GCC model the CPU closely enough to avoid unreliable
5750 results from the algorithm.
5752 This only makes sense when scheduling after register allocation, i.e.@: with
5753 @option{-fschedule-insns2} or at @option{-O2} or higher.
5755 @item -fsched2-use-traces
5756 @opindex fsched2-use-traces
5757 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
5758 allocation and additionally perform code duplication in order to increase the
5759 size of superblocks using tracer pass. See @option{-ftracer} for details on
5762 This mode should produce faster but significantly longer programs. Also
5763 without @option{-fbranch-probabilities} the traces constructed may not
5764 match the reality and hurt the performance. This only makes
5765 sense when scheduling after register allocation, i.e.@: with
5766 @option{-fschedule-insns2} or at @option{-O2} or higher.
5770 Eliminate redundant sign extension instructions and move the non-redundant
5771 ones to optimal placement using lazy code motion (LCM).
5773 @item -freschedule-modulo-scheduled-loops
5774 @opindex freschedule-modulo-scheduled-loops
5775 The modulo scheduling comes before the traditional scheduling, if a loop
5776 was modulo scheduled we may want to prevent the later scheduling passes
5777 from changing its schedule, we use this option to control that.
5779 @item -fcaller-saves
5780 @opindex fcaller-saves
5781 Enable values to be allocated in registers that will be clobbered by
5782 function calls, by emitting extra instructions to save and restore the
5783 registers around such calls. Such allocation is done only when it
5784 seems to result in better code than would otherwise be produced.
5786 This option is always enabled by default on certain machines, usually
5787 those which have no call-preserved registers to use instead.
5789 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5791 @item -ftree-reassoc
5792 @opindex ftree-reassoc
5793 Perform reassociation on trees. This flag is enabled by default
5794 at @option{-O} and higher.
5798 Perform partial redundancy elimination (PRE) on trees. This flag is
5799 enabled by default at @option{-O2} and @option{-O3}.
5803 Perform full redundancy elimination (FRE) on trees. The difference
5804 between FRE and PRE is that FRE only considers expressions
5805 that are computed on all paths leading to the redundant computation.
5806 This analysis is faster than PRE, though it exposes fewer redundancies.
5807 This flag is enabled by default at @option{-O} and higher.
5809 @item -ftree-copy-prop
5810 @opindex ftree-copy-prop
5811 Perform copy propagation on trees. This pass eliminates unnecessary
5812 copy operations. This flag is enabled by default at @option{-O} and
5816 @opindex ftree-salias
5817 Perform structural alias analysis on trees. This flag
5818 is enabled by default at @option{-O} and higher.
5820 @item -fipa-pure-const
5821 @opindex fipa-pure-const
5822 Discover which functions are pure or constant.
5823 Enabled by default at @option{-O} and higher.
5825 @item -fipa-reference
5826 @opindex fipa-reference
5827 Discover which static variables do not escape cannot escape the
5829 Enabled by default at @option{-O} and higher.
5831 @item -fipa-struct-reorg
5832 @opindex fipa-struct-reorg
5833 Perform structure reorganization optimization, that change C-like structures
5834 layout in order to better utilize spatial locality. This transformation is
5835 affective for programs containing arrays of structures. Available in two
5836 compilation modes: profile-based (enabled with @option{-fprofile-generate})
5837 or static (which uses built-in heuristics). Require @option{-fipa-type-escape}
5838 to provide the safety of this transformation. It works only in whole program
5839 mode, so it requires @option{-fwhole-program} and @option{-combine} to be
5840 enabled. Structures considered @samp{cold} by this transformation are not
5841 affected (see @option{--param struct-reorg-cold-struct-ratio=@var{value}}).
5843 With this flag, the program debug info reflects a new structure layout.
5847 Perform interprocedural pointer analysis.
5851 Perform interprocedural constant propagation.
5852 This optimization analyzes the program to determine when values passed
5853 to functions are constants and then optimizes accordingly.
5854 This optimization can substantially increase performance
5855 if the application has constants passed to functions, but
5856 because this optimization can create multiple copies of functions,
5857 it may significantly increase code size.
5859 @item -fipa-matrix-reorg
5860 @opindex fipa-matrix-reorg
5861 Perform matrix flattening and transposing.
5862 Matrix flattening tries to replace a m-dimensional matrix
5863 with its equivalent n-dimensional matrix, where n < m.
5864 This reduces the level of indirection needed for accessing the elements
5865 of the matrix. The second optimization is matrix transposing that
5866 attemps to change the order of the matrix's dimensions in order to
5867 improve cache locality.
5868 Both optimizations need fwhole-program flag.
5869 Transposing is enabled only if profiling information is avaliable.
5874 Perform forward store motion on trees. This flag is
5875 enabled by default at @option{-O} and higher.
5879 Perform sparse conditional constant propagation (CCP) on trees. This
5880 pass only operates on local scalar variables and is enabled by default
5881 at @option{-O} and higher.
5883 @item -ftree-store-ccp
5884 @opindex ftree-store-ccp
5885 Perform sparse conditional constant propagation (CCP) on trees. This
5886 pass operates on both local scalar variables and memory stores and
5887 loads (global variables, structures, arrays, etc). This flag is
5888 enabled by default at @option{-O2} and higher.
5892 Perform dead code elimination (DCE) on trees. This flag is enabled by
5893 default at @option{-O} and higher.
5895 @item -ftree-dominator-opts
5896 @opindex ftree-dominator-opts
5897 Perform a variety of simple scalar cleanups (constant/copy
5898 propagation, redundancy elimination, range propagation and expression
5899 simplification) based on a dominator tree traversal. This also
5900 performs jump threading (to reduce jumps to jumps). This flag is
5901 enabled by default at @option{-O} and higher.
5905 Perform dead store elimination (DSE) on trees. A dead store is a store into
5906 a memory location which will later be overwritten by another store without
5907 any intervening loads. In this case the earlier store can be deleted. This
5908 flag is enabled by default at @option{-O} and higher.
5912 Perform loop header copying on trees. This is beneficial since it increases
5913 effectiveness of code motion optimizations. It also saves one jump. This flag
5914 is enabled by default at @option{-O} and higher. It is not enabled
5915 for @option{-Os}, since it usually increases code size.
5917 @item -ftree-loop-optimize
5918 @opindex ftree-loop-optimize
5919 Perform loop optimizations on trees. This flag is enabled by default
5920 at @option{-O} and higher.
5922 @item -ftree-loop-linear
5923 @opindex ftree-loop-linear
5924 Perform linear loop transformations on tree. This flag can improve cache
5925 performance and allow further loop optimizations to take place.
5927 @item -fcheck-data-deps
5928 @opindex fcheck-data-deps
5929 Compare the results of several data dependence analyzers. This option
5930 is used for debugging the data dependence analyzers.
5932 @item -ftree-loop-distribution
5933 Perform loop distribution. This flag can improve cache performance on
5934 big loop bodies and allow further loop optimizations, like
5935 parallelization or vectorization, to take place.
5937 @item -ftree-loop-im
5938 @opindex ftree-loop-im
5939 Perform loop invariant motion on trees. This pass moves only invariants that
5940 would be hard to handle at RTL level (function calls, operations that expand to
5941 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
5942 operands of conditions that are invariant out of the loop, so that we can use
5943 just trivial invariantness analysis in loop unswitching. The pass also includes
5946 @item -ftree-loop-ivcanon
5947 @opindex ftree-loop-ivcanon
5948 Create a canonical counter for number of iterations in the loop for that
5949 determining number of iterations requires complicated analysis. Later
5950 optimizations then may determine the number easily. Useful especially
5951 in connection with unrolling.
5955 Perform induction variable optimizations (strength reduction, induction
5956 variable merging and induction variable elimination) on trees.
5958 @item -ftree-parallelize-loops=n
5959 @opindex ftree-parallelize-loops
5960 Parallelize loops, i.e., split their iteration space to run in n threads.
5961 This is only possible for loops whose iterations are independent
5962 and can be arbitrarily reordered. The optimization is only
5963 profitable on multiprocessor machines, for loops that are CPU-intensive,
5964 rather than constrained e.g.@: by memory bandwidth. This option
5965 implies @option{-pthread}, and thus is only supported on targets
5966 that have support for @option{-pthread}.
5970 Perform scalar replacement of aggregates. This pass replaces structure
5971 references with scalars to prevent committing structures to memory too
5972 early. This flag is enabled by default at @option{-O} and higher.
5974 @item -ftree-copyrename
5975 @opindex ftree-copyrename
5976 Perform copy renaming on trees. This pass attempts to rename compiler
5977 temporaries to other variables at copy locations, usually resulting in
5978 variable names which more closely resemble the original variables. This flag
5979 is enabled by default at @option{-O} and higher.
5983 Perform temporary expression replacement during the SSA->normal phase. Single
5984 use/single def temporaries are replaced at their use location with their
5985 defining expression. This results in non-GIMPLE code, but gives the expanders
5986 much more complex trees to work on resulting in better RTL generation. This is
5987 enabled by default at @option{-O} and higher.
5989 @item -ftree-vectorize
5990 @opindex ftree-vectorize
5991 Perform loop vectorization on trees.
5993 @item -ftree-vect-loop-version
5994 @opindex ftree-vect-loop-version
5995 Perform loop versioning when doing loop vectorization on trees. When a loop
5996 appears to be vectorizable except that data alignment or data dependence cannot
5997 be determined at compile time then vectorized and non-vectorized versions of
5998 the loop are generated along with runtime checks for alignment or dependence
5999 to control which version is executed. This option is enabled by default
6000 except at level @option{-Os} where it is disabled.
6002 @item -fvect-cost-model
6003 @opindex fvect-cost-model
6004 Enable cost model for vectorization.
6008 Perform Value Range Propagation on trees. This is similar to the
6009 constant propagation pass, but instead of values, ranges of values are
6010 propagated. This allows the optimizers to remove unnecessary range
6011 checks like array bound checks and null pointer checks. This is
6012 enabled by default at @option{-O2} and higher. Null pointer check
6013 elimination is only done if @option{-fdelete-null-pointer-checks} is
6018 Perform tail duplication to enlarge superblock size. This transformation
6019 simplifies the control flow of the function allowing other optimizations to do
6022 @item -funroll-loops
6023 @opindex funroll-loops
6024 Unroll loops whose number of iterations can be determined at compile
6025 time or upon entry to the loop. @option{-funroll-loops} implies
6026 @option{-frerun-cse-after-loop}. This option makes code larger,
6027 and may or may not make it run faster.
6029 @item -funroll-all-loops
6030 @opindex funroll-all-loops
6031 Unroll all loops, even if their number of iterations is uncertain when
6032 the loop is entered. This usually makes programs run more slowly.
6033 @option{-funroll-all-loops} implies the same options as
6034 @option{-funroll-loops},
6036 @item -fsplit-ivs-in-unroller
6037 @opindex fsplit-ivs-in-unroller
6038 Enables expressing of values of induction variables in later iterations
6039 of the unrolled loop using the value in the first iteration. This breaks
6040 long dependency chains, thus improving efficiency of the scheduling passes.
6042 Combination of @option{-fweb} and CSE is often sufficient to obtain the
6043 same effect. However in cases the loop body is more complicated than
6044 a single basic block, this is not reliable. It also does not work at all
6045 on some of the architectures due to restrictions in the CSE pass.
6047 This optimization is enabled by default.
6049 @item -fvariable-expansion-in-unroller
6050 @opindex fvariable-expansion-in-unroller
6051 With this option, the compiler will create multiple copies of some
6052 local variables when unrolling a loop which can result in superior code.
6054 @item -fpredictive-commoning
6055 @opindex fpredictive-commoning
6056 Perform predictive commoning optimization, i.e., reusing computations
6057 (especially memory loads and stores) performed in previous
6058 iterations of loops.
6060 This option is enabled at level @option{-O3}.
6062 @item -fprefetch-loop-arrays
6063 @opindex fprefetch-loop-arrays
6064 If supported by the target machine, generate instructions to prefetch
6065 memory to improve the performance of loops that access large arrays.
6067 This option may generate better or worse code; results are highly
6068 dependent on the structure of loops within the source code.
6070 Disabled at level @option{-Os}.
6073 @itemx -fno-peephole2
6074 @opindex fno-peephole
6075 @opindex fno-peephole2
6076 Disable any machine-specific peephole optimizations. The difference
6077 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
6078 are implemented in the compiler; some targets use one, some use the
6079 other, a few use both.
6081 @option{-fpeephole} is enabled by default.
6082 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6084 @item -fno-guess-branch-probability
6085 @opindex fno-guess-branch-probability
6086 Do not guess branch probabilities using heuristics.
6088 GCC will use heuristics to guess branch probabilities if they are
6089 not provided by profiling feedback (@option{-fprofile-arcs}). These
6090 heuristics are based on the control flow graph. If some branch probabilities
6091 are specified by @samp{__builtin_expect}, then the heuristics will be
6092 used to guess branch probabilities for the rest of the control flow graph,
6093 taking the @samp{__builtin_expect} info into account. The interactions
6094 between the heuristics and @samp{__builtin_expect} can be complex, and in
6095 some cases, it may be useful to disable the heuristics so that the effects
6096 of @samp{__builtin_expect} are easier to understand.
6098 The default is @option{-fguess-branch-probability} at levels
6099 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6101 @item -freorder-blocks
6102 @opindex freorder-blocks
6103 Reorder basic blocks in the compiled function in order to reduce number of
6104 taken branches and improve code locality.
6106 Enabled at levels @option{-O2}, @option{-O3}.
6108 @item -freorder-blocks-and-partition
6109 @opindex freorder-blocks-and-partition
6110 In addition to reordering basic blocks in the compiled function, in order
6111 to reduce number of taken branches, partitions hot and cold basic blocks
6112 into separate sections of the assembly and .o files, to improve
6113 paging and cache locality performance.
6115 This optimization is automatically turned off in the presence of
6116 exception handling, for linkonce sections, for functions with a user-defined
6117 section attribute and on any architecture that does not support named
6120 @item -freorder-functions
6121 @opindex freorder-functions
6122 Reorder functions in the object file in order to
6123 improve code locality. This is implemented by using special
6124 subsections @code{.text.hot} for most frequently executed functions and
6125 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
6126 the linker so object file format must support named sections and linker must
6127 place them in a reasonable way.
6129 Also profile feedback must be available in to make this option effective. See
6130 @option{-fprofile-arcs} for details.
6132 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6134 @item -fstrict-aliasing
6135 @opindex fstrict-aliasing
6136 Allows the compiler to assume the strictest aliasing rules applicable to
6137 the language being compiled. For C (and C++), this activates
6138 optimizations based on the type of expressions. In particular, an
6139 object of one type is assumed never to reside at the same address as an
6140 object of a different type, unless the types are almost the same. For
6141 example, an @code{unsigned int} can alias an @code{int}, but not a
6142 @code{void*} or a @code{double}. A character type may alias any other
6145 Pay special attention to code like this:
6158 The practice of reading from a different union member than the one most
6159 recently written to (called ``type-punning'') is common. Even with
6160 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
6161 is accessed through the union type. So, the code above will work as
6162 expected. However, this code might not:
6173 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6175 @item -fstrict-overflow
6176 @opindex fstrict-overflow
6177 Allow the compiler to assume strict signed overflow rules, depending
6178 on the language being compiled. For C (and C++) this means that
6179 overflow when doing arithmetic with signed numbers is undefined, which
6180 means that the compiler may assume that it will not happen. This
6181 permits various optimizations. For example, the compiler will assume
6182 that an expression like @code{i + 10 > i} will always be true for
6183 signed @code{i}. This assumption is only valid if signed overflow is
6184 undefined, as the expression is false if @code{i + 10} overflows when
6185 using twos complement arithmetic. When this option is in effect any
6186 attempt to determine whether an operation on signed numbers will
6187 overflow must be written carefully to not actually involve overflow.
6189 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
6190 that signed overflow is fully defined: it wraps. When
6191 @option{-fwrapv} is used, there is no difference between
6192 @option{-fstrict-overflow} and @option{-fno-strict-overflow}. With
6193 @option{-fwrapv} certain types of overflow are permitted. For
6194 example, if the compiler gets an overflow when doing arithmetic on
6195 constants, the overflowed value can still be used with
6196 @option{-fwrapv}, but not otherwise.
6198 The @option{-fstrict-overflow} option is enabled at levels
6199 @option{-O2}, @option{-O3}, @option{-Os}.
6201 @item -falign-functions
6202 @itemx -falign-functions=@var{n}
6203 @opindex falign-functions
6204 Align the start of functions to the next power-of-two greater than
6205 @var{n}, skipping up to @var{n} bytes. For instance,
6206 @option{-falign-functions=32} aligns functions to the next 32-byte
6207 boundary, but @option{-falign-functions=24} would align to the next
6208 32-byte boundary only if this can be done by skipping 23 bytes or less.
6210 @option{-fno-align-functions} and @option{-falign-functions=1} are
6211 equivalent and mean that functions will not be aligned.
6213 Some assemblers only support this flag when @var{n} is a power of two;
6214 in that case, it is rounded up.
6216 If @var{n} is not specified or is zero, use a machine-dependent default.
6218 Enabled at levels @option{-O2}, @option{-O3}.
6220 @item -falign-labels
6221 @itemx -falign-labels=@var{n}
6222 @opindex falign-labels
6223 Align all branch targets to a power-of-two boundary, skipping up to
6224 @var{n} bytes like @option{-falign-functions}. This option can easily
6225 make code slower, because it must insert dummy operations for when the
6226 branch target is reached in the usual flow of the code.
6228 @option{-fno-align-labels} and @option{-falign-labels=1} are
6229 equivalent and mean that labels will not be aligned.
6231 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
6232 are greater than this value, then their values are used instead.
6234 If @var{n} is not specified or is zero, use a machine-dependent default
6235 which is very likely to be @samp{1}, meaning no alignment.
6237 Enabled at levels @option{-O2}, @option{-O3}.
6240 @itemx -falign-loops=@var{n}
6241 @opindex falign-loops
6242 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
6243 like @option{-falign-functions}. The hope is that the loop will be
6244 executed many times, which will make up for any execution of the dummy
6247 @option{-fno-align-loops} and @option{-falign-loops=1} are
6248 equivalent and mean that loops will not be aligned.
6250 If @var{n} is not specified or is zero, use a machine-dependent default.
6252 Enabled at levels @option{-O2}, @option{-O3}.
6255 @itemx -falign-jumps=@var{n}
6256 @opindex falign-jumps
6257 Align branch targets to a power-of-two boundary, for branch targets
6258 where the targets can only be reached by jumping, skipping up to @var{n}
6259 bytes like @option{-falign-functions}. In this case, no dummy operations
6262 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
6263 equivalent and mean that loops will not be aligned.
6265 If @var{n} is not specified or is zero, use a machine-dependent default.
6267 Enabled at levels @option{-O2}, @option{-O3}.
6269 @item -funit-at-a-time
6270 @opindex funit-at-a-time
6271 Parse the whole compilation unit before starting to produce code.
6272 This allows some extra optimizations to take place but consumes
6273 more memory (in general). There are some compatibility issues
6274 with @emph{unit-at-a-time} mode:
6277 enabling @emph{unit-at-a-time} mode may change the order
6278 in which functions, variables, and top-level @code{asm} statements
6279 are emitted, and will likely break code relying on some particular
6280 ordering. The majority of such top-level @code{asm} statements,
6281 though, can be replaced by @code{section} attributes. The
6282 @option{fno-toplevel-reorder} option may be used to keep the ordering
6283 used in the input file, at the cost of some optimizations.
6286 @emph{unit-at-a-time} mode removes unreferenced static variables
6287 and functions. This may result in undefined references
6288 when an @code{asm} statement refers directly to variables or functions
6289 that are otherwise unused. In that case either the variable/function
6290 shall be listed as an operand of the @code{asm} statement operand or,
6291 in the case of top-level @code{asm} statements the attribute @code{used}
6292 shall be used on the declaration.
6295 Static functions now can use non-standard passing conventions that
6296 may break @code{asm} statements calling functions directly. Again,
6297 attribute @code{used} will prevent this behavior.
6300 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
6301 but this scheme may not be supported by future releases of GCC@.
6303 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6305 @item -fno-toplevel-reorder
6306 @opindex fno-toplevel-reorder
6307 Do not reorder top-level functions, variables, and @code{asm}
6308 statements. Output them in the same order that they appear in the
6309 input file. When this option is used, unreferenced static variables
6310 will not be removed. This option is intended to support existing code
6311 which relies on a particular ordering. For new code, it is better to
6316 Constructs webs as commonly used for register allocation purposes and assign
6317 each web individual pseudo register. This allows the register allocation pass
6318 to operate on pseudos directly, but also strengthens several other optimization
6319 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
6320 however, make debugging impossible, since variables will no longer stay in a
6323 Enabled by default with @option{-funroll-loops}.
6325 @item -fwhole-program
6326 @opindex fwhole-program
6327 Assume that the current compilation unit represents whole program being
6328 compiled. All public functions and variables with the exception of @code{main}
6329 and those merged by attribute @code{externally_visible} become static functions
6330 and in a affect gets more aggressively optimized by interprocedural optimizers.
6331 While this option is equivalent to proper use of @code{static} keyword for
6332 programs consisting of single file, in combination with option
6333 @option{--combine} this flag can be used to compile most of smaller scale C
6334 programs since the functions and variables become local for the whole combined
6335 compilation unit, not for the single source file itself.
6337 This option is not supported for Fortran programs.
6339 @item -fcprop-registers
6340 @opindex fcprop-registers
6341 After register allocation and post-register allocation instruction splitting,
6342 we perform a copy-propagation pass to try to reduce scheduling dependencies
6343 and occasionally eliminate the copy.
6345 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6347 @item -fprofile-generate
6348 @opindex fprofile-generate
6350 Enable options usually used for instrumenting application to produce
6351 profile useful for later recompilation with profile feedback based
6352 optimization. You must use @option{-fprofile-generate} both when
6353 compiling and when linking your program.
6355 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
6358 @opindex fprofile-use
6359 Enable profile feedback directed optimizations, and optimizations
6360 generally profitable only with profile feedback available.
6362 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
6363 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
6365 By default, GCC emits an error message if the feedback profiles do not
6366 match the source code. This error can be turned into a warning by using
6367 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
6371 The following options control compiler behavior regarding floating
6372 point arithmetic. These options trade off between speed and
6373 correctness. All must be specifically enabled.
6377 @opindex ffloat-store
6378 Do not store floating point variables in registers, and inhibit other
6379 options that might change whether a floating point value is taken from a
6382 @cindex floating point precision
6383 This option prevents undesirable excess precision on machines such as
6384 the 68000 where the floating registers (of the 68881) keep more
6385 precision than a @code{double} is supposed to have. Similarly for the
6386 x86 architecture. For most programs, the excess precision does only
6387 good, but a few programs rely on the precise definition of IEEE floating
6388 point. Use @option{-ffloat-store} for such programs, after modifying
6389 them to store all pertinent intermediate computations into variables.
6393 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
6394 @option{-ffinite-math-only}, @option{-fno-rounding-math},
6395 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
6397 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
6399 This option is not turned on by any @option{-O} option since
6400 it can result in incorrect output for programs which depend on
6401 an exact implementation of IEEE or ISO rules/specifications for
6402 math functions. It may, however, yield faster code for programs
6403 that do not require the guarantees of these specifications.
6405 @item -fno-math-errno
6406 @opindex fno-math-errno
6407 Do not set ERRNO after calling math functions that are executed
6408 with a single instruction, e.g., sqrt. A program that relies on
6409 IEEE exceptions for math error handling may want to use this flag
6410 for speed while maintaining IEEE arithmetic compatibility.
6412 This option is not turned on by any @option{-O} option since
6413 it can result in incorrect output for programs which depend on
6414 an exact implementation of IEEE or ISO rules/specifications for
6415 math functions. It may, however, yield faster code for programs
6416 that do not require the guarantees of these specifications.
6418 The default is @option{-fmath-errno}.
6420 On Darwin systems, the math library never sets @code{errno}. There is
6421 therefore no reason for the compiler to consider the possibility that
6422 it might, and @option{-fno-math-errno} is the default.
6424 @item -funsafe-math-optimizations
6425 @opindex funsafe-math-optimizations
6427 Allow optimizations for floating-point arithmetic that (a) assume
6428 that arguments and results are valid and (b) may violate IEEE or
6429 ANSI standards. When used at link-time, it may include libraries
6430 or startup files that change the default FPU control word or other
6431 similar optimizations.
6433 This option is not turned on by any @option{-O} option since
6434 it can result in incorrect output for programs which depend on
6435 an exact implementation of IEEE or ISO rules/specifications for
6436 math functions. It may, however, yield faster code for programs
6437 that do not require the guarantees of these specifications.
6438 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
6439 @option{-fassociative-math} and @option{-freciprocal-math}.
6441 The default is @option{-fno-unsafe-math-optimizations}.
6443 @item -fassociative-math
6444 @opindex fassociative-math
6446 Allow re-association of operands in series of floating-point operations.
6447 This violates the ISO C and C++ language standard by possibly changing
6448 computation result. NOTE: re-ordering may change the sign of zero as
6449 well as ignore NaNs and inhibit or create underflow or overflow (and
6450 thus cannot be used on a code which relies on rounding behavior like
6451 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
6452 and thus may not be used when ordered comparisons are required.
6453 This option requires that both @option{-fno-signed-zeros} and
6454 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
6455 much sense with @option{-frounding-math}.
6457 The default is @option{-fno-associative-math}.
6459 @item -freciprocal-math
6460 @opindex freciprocal-math
6462 Allow the reciprocal of a value to be used instead of dividing by
6463 the value if this enables optimizations. For example @code{x / y}
6464 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
6465 is subject to common subexpression elimination. Note that this loses
6466 precision and increases the number of flops operating on the value.
6468 The default is @option{-fno-reciprocal-math}.
6470 @item -ffinite-math-only
6471 @opindex ffinite-math-only
6472 Allow optimizations for floating-point arithmetic that assume
6473 that arguments and results are not NaNs or +-Infs.
6475 This option is not turned on by any @option{-O} option since
6476 it can result in incorrect output for programs which depend on
6477 an exact implementation of IEEE or ISO rules/specifications for
6478 math functions. It may, however, yield faster code for programs
6479 that do not require the guarantees of these specifications.
6481 The default is @option{-fno-finite-math-only}.
6483 @item -fno-signed-zeros
6484 @opindex fno-signed-zeros
6485 Allow optimizations for floating point arithmetic that ignore the
6486 signedness of zero. IEEE arithmetic specifies the behavior of
6487 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
6488 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
6489 This option implies that the sign of a zero result isn't significant.
6491 The default is @option{-fsigned-zeros}.
6493 @item -fno-trapping-math
6494 @opindex fno-trapping-math
6495 Compile code assuming that floating-point operations cannot generate
6496 user-visible traps. These traps include division by zero, overflow,
6497 underflow, inexact result and invalid operation. This option requires
6498 that @option{-fno-signaling-nans} be in effect. Setting this option may
6499 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
6501 This option should never be turned on by any @option{-O} option since
6502 it can result in incorrect output for programs which depend on
6503 an exact implementation of IEEE or ISO rules/specifications for
6506 The default is @option{-ftrapping-math}.
6508 @item -frounding-math
6509 @opindex frounding-math
6510 Disable transformations and optimizations that assume default floating
6511 point rounding behavior. This is round-to-zero for all floating point
6512 to integer conversions, and round-to-nearest for all other arithmetic
6513 truncations. This option should be specified for programs that change
6514 the FP rounding mode dynamically, or that may be executed with a
6515 non-default rounding mode. This option disables constant folding of
6516 floating point expressions at compile-time (which may be affected by
6517 rounding mode) and arithmetic transformations that are unsafe in the
6518 presence of sign-dependent rounding modes.
6520 The default is @option{-fno-rounding-math}.
6522 This option is experimental and does not currently guarantee to
6523 disable all GCC optimizations that are affected by rounding mode.
6524 Future versions of GCC may provide finer control of this setting
6525 using C99's @code{FENV_ACCESS} pragma. This command line option
6526 will be used to specify the default state for @code{FENV_ACCESS}.
6528 @item -frtl-abstract-sequences
6529 @opindex frtl-abstract-sequences
6530 It is a size optimization method. This option is to find identical
6531 sequences of code, which can be turned into pseudo-procedures and
6532 then replace all occurrences with calls to the newly created
6533 subroutine. It is kind of an opposite of @option{-finline-functions}.
6534 This optimization runs at RTL level.
6536 @item -fsignaling-nans
6537 @opindex fsignaling-nans
6538 Compile code assuming that IEEE signaling NaNs may generate user-visible
6539 traps during floating-point operations. Setting this option disables
6540 optimizations that may change the number of exceptions visible with
6541 signaling NaNs. This option implies @option{-ftrapping-math}.
6543 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
6546 The default is @option{-fno-signaling-nans}.
6548 This option is experimental and does not currently guarantee to
6549 disable all GCC optimizations that affect signaling NaN behavior.
6551 @item -fsingle-precision-constant
6552 @opindex fsingle-precision-constant
6553 Treat floating point constant as single precision constant instead of
6554 implicitly converting it to double precision constant.
6556 @item -fcx-limited-range
6557 @opindex fcx-limited-range
6558 When enabled, this option states that a range reduction step is not
6559 needed when performing complex division. Also, there is no checking
6560 whether the result of a complex multiplication or division is @code{NaN
6561 + I*NaN}, with an attempt to rescue the situation in that case. The
6562 default is @option{-fno-cx-limited-range}, but is enabled by
6563 @option{-ffast-math}.
6565 This option controls the default setting of the ISO C99
6566 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
6569 @item -fcx-fortran-rules
6570 @opindex fcx-fortran-rules
6571 Complex multiplication and division follow Fortran rules. Range
6572 reduction is done as part of complex division, but there is no checking
6573 whether the result of a complex multiplication or division is @code{NaN
6574 + I*NaN}, with an attempt to rescue the situation in that case.
6576 The default is @option{-fno-cx-fortran-rules}.
6580 The following options control optimizations that may improve
6581 performance, but are not enabled by any @option{-O} options. This
6582 section includes experimental options that may produce broken code.
6585 @item -fbranch-probabilities
6586 @opindex fbranch-probabilities
6587 After running a program compiled with @option{-fprofile-arcs}
6588 (@pxref{Debugging Options,, Options for Debugging Your Program or
6589 @command{gcc}}), you can compile it a second time using
6590 @option{-fbranch-probabilities}, to improve optimizations based on
6591 the number of times each branch was taken. When the program
6592 compiled with @option{-fprofile-arcs} exits it saves arc execution
6593 counts to a file called @file{@var{sourcename}.gcda} for each source
6594 file. The information in this data file is very dependent on the
6595 structure of the generated code, so you must use the same source code
6596 and the same optimization options for both compilations.
6598 With @option{-fbranch-probabilities}, GCC puts a
6599 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
6600 These can be used to improve optimization. Currently, they are only
6601 used in one place: in @file{reorg.c}, instead of guessing which path a
6602 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
6603 exactly determine which path is taken more often.
6605 @item -fprofile-values
6606 @opindex fprofile-values
6607 If combined with @option{-fprofile-arcs}, it adds code so that some
6608 data about values of expressions in the program is gathered.
6610 With @option{-fbranch-probabilities}, it reads back the data gathered
6611 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
6612 notes to instructions for their later usage in optimizations.
6614 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
6618 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
6619 a code to gather information about values of expressions.
6621 With @option{-fbranch-probabilities}, it reads back the data gathered
6622 and actually performs the optimizations based on them.
6623 Currently the optimizations include specialization of division operation
6624 using the knowledge about the value of the denominator.
6626 @item -frename-registers
6627 @opindex frename-registers
6628 Attempt to avoid false dependencies in scheduled code by making use
6629 of registers left over after register allocation. This optimization
6630 will most benefit processors with lots of registers. Depending on the
6631 debug information format adopted by the target, however, it can
6632 make debugging impossible, since variables will no longer stay in
6633 a ``home register''.
6635 Enabled by default with @option{-funroll-loops}.
6639 Perform tail duplication to enlarge superblock size. This transformation
6640 simplifies the control flow of the function allowing other optimizations to do
6643 Enabled with @option{-fprofile-use}.
6645 @item -funroll-loops
6646 @opindex funroll-loops
6647 Unroll loops whose number of iterations can be determined at compile time or
6648 upon entry to the loop. @option{-funroll-loops} implies
6649 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
6650 It also turns on complete loop peeling (i.e.@: complete removal of loops with
6651 small constant number of iterations). This option makes code larger, and may
6652 or may not make it run faster.
6654 Enabled with @option{-fprofile-use}.
6656 @item -funroll-all-loops
6657 @opindex funroll-all-loops
6658 Unroll all loops, even if their number of iterations is uncertain when
6659 the loop is entered. This usually makes programs run more slowly.
6660 @option{-funroll-all-loops} implies the same options as
6661 @option{-funroll-loops}.
6664 @opindex fpeel-loops
6665 Peels the loops for that there is enough information that they do not
6666 roll much (from profile feedback). It also turns on complete loop peeling
6667 (i.e.@: complete removal of loops with small constant number of iterations).
6669 Enabled with @option{-fprofile-use}.
6671 @item -fmove-loop-invariants
6672 @opindex fmove-loop-invariants
6673 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
6674 at level @option{-O1}
6676 @item -funswitch-loops
6677 @opindex funswitch-loops
6678 Move branches with loop invariant conditions out of the loop, with duplicates
6679 of the loop on both branches (modified according to result of the condition).
6681 @item -ffunction-sections
6682 @itemx -fdata-sections
6683 @opindex ffunction-sections
6684 @opindex fdata-sections
6685 Place each function or data item into its own section in the output
6686 file if the target supports arbitrary sections. The name of the
6687 function or the name of the data item determines the section's name
6690 Use these options on systems where the linker can perform optimizations
6691 to improve locality of reference in the instruction space. Most systems
6692 using the ELF object format and SPARC processors running Solaris 2 have
6693 linkers with such optimizations. AIX may have these optimizations in
6696 Only use these options when there are significant benefits from doing
6697 so. When you specify these options, the assembler and linker will
6698 create larger object and executable files and will also be slower.
6699 You will not be able to use @code{gprof} on all systems if you
6700 specify this option and you may have problems with debugging if
6701 you specify both this option and @option{-g}.
6703 @item -fbranch-target-load-optimize
6704 @opindex fbranch-target-load-optimize
6705 Perform branch target register load optimization before prologue / epilogue
6707 The use of target registers can typically be exposed only during reload,
6708 thus hoisting loads out of loops and doing inter-block scheduling needs
6709 a separate optimization pass.
6711 @item -fbranch-target-load-optimize2
6712 @opindex fbranch-target-load-optimize2
6713 Perform branch target register load optimization after prologue / epilogue
6716 @item -fbtr-bb-exclusive
6717 @opindex fbtr-bb-exclusive
6718 When performing branch target register load optimization, don't reuse
6719 branch target registers in within any basic block.
6721 @item -fstack-protector
6722 @opindex fstack-protector
6723 Emit extra code to check for buffer overflows, such as stack smashing
6724 attacks. This is done by adding a guard variable to functions with
6725 vulnerable objects. This includes functions that call alloca, and
6726 functions with buffers larger than 8 bytes. The guards are initialized
6727 when a function is entered and then checked when the function exits.
6728 If a guard check fails, an error message is printed and the program exits.
6730 @item -fstack-protector-all
6731 @opindex fstack-protector-all
6732 Like @option{-fstack-protector} except that all functions are protected.
6734 @item -fsection-anchors
6735 @opindex fsection-anchors
6736 Try to reduce the number of symbolic address calculations by using
6737 shared ``anchor'' symbols to address nearby objects. This transformation
6738 can help to reduce the number of GOT entries and GOT accesses on some
6741 For example, the implementation of the following function @code{foo}:
6745 int foo (void) @{ return a + b + c; @}
6748 would usually calculate the addresses of all three variables, but if you
6749 compile it with @option{-fsection-anchors}, it will access the variables
6750 from a common anchor point instead. The effect is similar to the
6751 following pseudocode (which isn't valid C):
6756 register int *xr = &x;
6757 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
6761 Not all targets support this option.
6763 @item --param @var{name}=@var{value}
6765 In some places, GCC uses various constants to control the amount of
6766 optimization that is done. For example, GCC will not inline functions
6767 that contain more that a certain number of instructions. You can
6768 control some of these constants on the command-line using the
6769 @option{--param} option.
6771 The names of specific parameters, and the meaning of the values, are
6772 tied to the internals of the compiler, and are subject to change
6773 without notice in future releases.
6775 In each case, the @var{value} is an integer. The allowable choices for
6776 @var{name} are given in the following table:
6779 @item salias-max-implicit-fields
6780 The maximum number of fields in a variable without direct
6781 structure accesses for which structure aliasing will consider trying
6782 to track each field. The default is 5
6784 @item salias-max-array-elements
6785 The maximum number of elements an array can have and its elements
6786 still be tracked individually by structure aliasing. The default is 4
6788 @item sra-max-structure-size
6789 The maximum structure size, in bytes, at which the scalar replacement
6790 of aggregates (SRA) optimization will perform block copies. The
6791 default value, 0, implies that GCC will select the most appropriate
6794 @item sra-field-structure-ratio
6795 The threshold ratio (as a percentage) between instantiated fields and
6796 the complete structure size. We say that if the ratio of the number
6797 of bytes in instantiated fields to the number of bytes in the complete
6798 structure exceeds this parameter, then block copies are not used. The
6801 @item struct-reorg-cold-struct-ratio
6802 The threshold ratio (as a percentage) between a structure frequency
6803 and the frequency of the hottest structure in the program. This parameter
6804 is used by struct-reorg optimization enabled by @option{-fipa-struct-reorg}.
6805 We say that if the ratio of a structure frequency, calculated by profiling,
6806 to the hottest structure frequency in the program is less than this
6807 parameter, then structure reorganization is not applied to this structure.
6810 @item max-crossjump-edges
6811 The maximum number of incoming edges to consider for crossjumping.
6812 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
6813 the number of edges incoming to each block. Increasing values mean
6814 more aggressive optimization, making the compile time increase with
6815 probably small improvement in executable size.
6817 @item min-crossjump-insns
6818 The minimum number of instructions which must be matched at the end
6819 of two blocks before crossjumping will be performed on them. This
6820 value is ignored in the case where all instructions in the block being
6821 crossjumped from are matched. The default value is 5.
6823 @item max-grow-copy-bb-insns
6824 The maximum code size expansion factor when copying basic blocks
6825 instead of jumping. The expansion is relative to a jump instruction.
6826 The default value is 8.
6828 @item max-goto-duplication-insns
6829 The maximum number of instructions to duplicate to a block that jumps
6830 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
6831 passes, GCC factors computed gotos early in the compilation process,
6832 and unfactors them as late as possible. Only computed jumps at the
6833 end of a basic blocks with no more than max-goto-duplication-insns are
6834 unfactored. The default value is 8.
6836 @item max-delay-slot-insn-search
6837 The maximum number of instructions to consider when looking for an
6838 instruction to fill a delay slot. If more than this arbitrary number of
6839 instructions is searched, the time savings from filling the delay slot
6840 will be minimal so stop searching. Increasing values mean more
6841 aggressive optimization, making the compile time increase with probably
6842 small improvement in executable run time.
6844 @item max-delay-slot-live-search
6845 When trying to fill delay slots, the maximum number of instructions to
6846 consider when searching for a block with valid live register
6847 information. Increasing this arbitrarily chosen value means more
6848 aggressive optimization, increasing the compile time. This parameter
6849 should be removed when the delay slot code is rewritten to maintain the
6852 @item max-gcse-memory
6853 The approximate maximum amount of memory that will be allocated in
6854 order to perform the global common subexpression elimination
6855 optimization. If more memory than specified is required, the
6856 optimization will not be done.
6858 @item max-gcse-passes
6859 The maximum number of passes of GCSE to run. The default is 1.
6861 @item max-pending-list-length
6862 The maximum number of pending dependencies scheduling will allow
6863 before flushing the current state and starting over. Large functions
6864 with few branches or calls can create excessively large lists which
6865 needlessly consume memory and resources.
6867 @item max-inline-insns-single
6868 Several parameters control the tree inliner used in gcc.
6869 This number sets the maximum number of instructions (counted in GCC's
6870 internal representation) in a single function that the tree inliner
6871 will consider for inlining. This only affects functions declared
6872 inline and methods implemented in a class declaration (C++).
6873 The default value is 450.
6875 @item max-inline-insns-auto
6876 When you use @option{-finline-functions} (included in @option{-O3}),
6877 a lot of functions that would otherwise not be considered for inlining
6878 by the compiler will be investigated. To those functions, a different
6879 (more restrictive) limit compared to functions declared inline can
6881 The default value is 90.
6883 @item large-function-insns
6884 The limit specifying really large functions. For functions larger than this
6885 limit after inlining inlining is constrained by
6886 @option{--param large-function-growth}. This parameter is useful primarily
6887 to avoid extreme compilation time caused by non-linear algorithms used by the
6889 This parameter is ignored when @option{-funit-at-a-time} is not used.
6890 The default value is 2700.
6892 @item large-function-growth
6893 Specifies maximal growth of large function caused by inlining in percents.
6894 This parameter is ignored when @option{-funit-at-a-time} is not used.
6895 The default value is 100 which limits large function growth to 2.0 times
6898 @item large-unit-insns
6899 The limit specifying large translation unit. Growth caused by inlining of
6900 units larger than this limit is limited by @option{--param inline-unit-growth}.
6901 For small units this might be too tight (consider unit consisting of function A
6902 that is inline and B that just calls A three time. If B is small relative to
6903 A, the growth of unit is 300\% and yet such inlining is very sane. For very
6904 large units consisting of small inlineable functions however the overall unit
6905 growth limit is needed to avoid exponential explosion of code size. Thus for
6906 smaller units, the size is increased to @option{--param large-unit-insns}
6907 before applying @option{--param inline-unit-growth}. The default is 10000
6909 @item inline-unit-growth
6910 Specifies maximal overall growth of the compilation unit caused by inlining.
6911 This parameter is ignored when @option{-funit-at-a-time} is not used.
6912 The default value is 30 which limits unit growth to 1.3 times the original
6915 @item large-stack-frame
6916 The limit specifying large stack frames. While inlining the algorithm is trying
6917 to not grow past this limit too much. Default value is 256 bytes.
6919 @item large-stack-frame-growth
6920 Specifies maximal growth of large stack frames caused by inlining in percents.
6921 The default value is 1000 which limits large stack frame growth to 11 times
6924 @item max-inline-insns-recursive
6925 @itemx max-inline-insns-recursive-auto
6926 Specifies maximum number of instructions out-of-line copy of self recursive inline
6927 function can grow into by performing recursive inlining.
6929 For functions declared inline @option{--param max-inline-insns-recursive} is
6930 taken into account. For function not declared inline, recursive inlining
6931 happens only when @option{-finline-functions} (included in @option{-O3}) is
6932 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
6933 default value is 450.
6935 @item max-inline-recursive-depth
6936 @itemx max-inline-recursive-depth-auto
6937 Specifies maximum recursion depth used by the recursive inlining.
6939 For functions declared inline @option{--param max-inline-recursive-depth} is
6940 taken into account. For function not declared inline, recursive inlining
6941 happens only when @option{-finline-functions} (included in @option{-O3}) is
6942 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
6945 @item min-inline-recursive-probability
6946 Recursive inlining is profitable only for function having deep recursion
6947 in average and can hurt for function having little recursion depth by
6948 increasing the prologue size or complexity of function body to other
6951 When profile feedback is available (see @option{-fprofile-generate}) the actual
6952 recursion depth can be guessed from probability that function will recurse via
6953 given call expression. This parameter limits inlining only to call expression
6954 whose probability exceeds given threshold (in percents). The default value is
6957 @item inline-call-cost
6958 Specify cost of call instruction relative to simple arithmetics operations
6959 (having cost of 1). Increasing this cost disqualifies inlining of non-leaf
6960 functions and at the same time increases size of leaf function that is believed to
6961 reduce function size by being inlined. In effect it increases amount of
6962 inlining for code having large abstraction penalty (many functions that just
6963 pass the arguments to other functions) and decrease inlining for code with low
6964 abstraction penalty. The default value is 12.
6966 @item min-vect-loop-bound
6967 The minimum number of iterations under which a loop will not get vectorized
6968 when @option{-ftree-vectorize} is used. The number of iterations after
6969 vectorization needs to be greater than the value specified by this option
6970 to allow vectorization. The default value is 0.
6972 @item max-unrolled-insns
6973 The maximum number of instructions that a loop should have if that loop
6974 is unrolled, and if the loop is unrolled, it determines how many times
6975 the loop code is unrolled.
6977 @item max-average-unrolled-insns
6978 The maximum number of instructions biased by probabilities of their execution
6979 that a loop should have if that loop is unrolled, and if the loop is unrolled,
6980 it determines how many times the loop code is unrolled.
6982 @item max-unroll-times
6983 The maximum number of unrollings of a single loop.
6985 @item max-peeled-insns
6986 The maximum number of instructions that a loop should have if that loop
6987 is peeled, and if the loop is peeled, it determines how many times
6988 the loop code is peeled.
6990 @item max-peel-times
6991 The maximum number of peelings of a single loop.
6993 @item max-completely-peeled-insns
6994 The maximum number of insns of a completely peeled loop.
6996 @item max-completely-peel-times
6997 The maximum number of iterations of a loop to be suitable for complete peeling.
6999 @item max-unswitch-insns
7000 The maximum number of insns of an unswitched loop.
7002 @item max-unswitch-level
7003 The maximum number of branches unswitched in a single loop.
7006 The minimum cost of an expensive expression in the loop invariant motion.
7008 @item iv-consider-all-candidates-bound
7009 Bound on number of candidates for induction variables below that
7010 all candidates are considered for each use in induction variable
7011 optimizations. Only the most relevant candidates are considered
7012 if there are more candidates, to avoid quadratic time complexity.
7014 @item iv-max-considered-uses
7015 The induction variable optimizations give up on loops that contain more
7016 induction variable uses.
7018 @item iv-always-prune-cand-set-bound
7019 If number of candidates in the set is smaller than this value,
7020 we always try to remove unnecessary ivs from the set during its
7021 optimization when a new iv is added to the set.
7023 @item scev-max-expr-size
7024 Bound on size of expressions used in the scalar evolutions analyzer.
7025 Large expressions slow the analyzer.
7027 @item omega-max-vars
7028 The maximum number of variables in an Omega constraint system.
7029 The default value is 128.
7031 @item omega-max-geqs
7032 The maximum number of inequalities in an Omega constraint system.
7033 The default value is 256.
7036 The maximum number of equalities in an Omega constraint system.
7037 The default value is 128.
7039 @item omega-max-wild-cards
7040 The maximum number of wildcard variables that the Omega solver will
7041 be able to insert. The default value is 18.
7043 @item omega-hash-table-size
7044 The size of the hash table in the Omega solver. The default value is
7047 @item omega-max-keys
7048 The maximal number of keys used by the Omega solver. The default
7051 @item omega-eliminate-redundant-constraints
7052 When set to 1, use expensive methods to eliminate all redundant
7053 constraints. The default value is 0.
7055 @item vect-max-version-for-alignment-checks
7056 The maximum number of runtime checks that can be performed when
7057 doing loop versioning for alignment in the vectorizer. See option
7058 ftree-vect-loop-version for more information.
7060 @item vect-max-version-for-alias-checks
7061 The maximum number of runtime checks that can be performed when
7062 doing loop versioning for alias in the vectorizer. See option
7063 ftree-vect-loop-version for more information.
7065 @item max-iterations-to-track
7067 The maximum number of iterations of a loop the brute force algorithm
7068 for analysis of # of iterations of the loop tries to evaluate.
7070 @item hot-bb-count-fraction
7071 Select fraction of the maximal count of repetitions of basic block in program
7072 given basic block needs to have to be considered hot.
7074 @item hot-bb-frequency-fraction
7075 Select fraction of the maximal frequency of executions of basic block in
7076 function given basic block needs to have to be considered hot
7078 @item max-predicted-iterations
7079 The maximum number of loop iterations we predict statically. This is useful
7080 in cases where function contain single loop with known bound and other loop
7081 with unknown. We predict the known number of iterations correctly, while
7082 the unknown number of iterations average to roughly 10. This means that the
7083 loop without bounds would appear artificially cold relative to the other one.
7085 @item align-threshold
7087 Select fraction of the maximal frequency of executions of basic block in
7088 function given basic block will get aligned.
7090 @item align-loop-iterations
7092 A loop expected to iterate at lest the selected number of iterations will get
7095 @item tracer-dynamic-coverage
7096 @itemx tracer-dynamic-coverage-feedback
7098 This value is used to limit superblock formation once the given percentage of
7099 executed instructions is covered. This limits unnecessary code size
7102 The @option{tracer-dynamic-coverage-feedback} is used only when profile
7103 feedback is available. The real profiles (as opposed to statically estimated
7104 ones) are much less balanced allowing the threshold to be larger value.
7106 @item tracer-max-code-growth
7107 Stop tail duplication once code growth has reached given percentage. This is
7108 rather hokey argument, as most of the duplicates will be eliminated later in
7109 cross jumping, so it may be set to much higher values than is the desired code
7112 @item tracer-min-branch-ratio
7114 Stop reverse growth when the reverse probability of best edge is less than this
7115 threshold (in percent).
7117 @item tracer-min-branch-ratio
7118 @itemx tracer-min-branch-ratio-feedback
7120 Stop forward growth if the best edge do have probability lower than this
7123 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
7124 compilation for profile feedback and one for compilation without. The value
7125 for compilation with profile feedback needs to be more conservative (higher) in
7126 order to make tracer effective.
7128 @item max-cse-path-length
7130 Maximum number of basic blocks on path that cse considers. The default is 10.
7133 The maximum instructions CSE process before flushing. The default is 1000.
7135 @item max-aliased-vops
7137 Maximum number of virtual operands per function allowed to represent
7138 aliases before triggering the alias partitioning heuristic. Alias
7139 partitioning reduces compile times and memory consumption needed for
7140 aliasing at the expense of precision loss in alias information. The
7141 default value for this parameter is 100 for -O1, 500 for -O2 and 1000
7144 Notice that if a function contains more memory statements than the
7145 value of this parameter, it is not really possible to achieve this
7146 reduction. In this case, the compiler will use the number of memory
7147 statements as the value for @option{max-aliased-vops}.
7149 @item avg-aliased-vops
7151 Average number of virtual operands per statement allowed to represent
7152 aliases before triggering the alias partitioning heuristic. This
7153 works in conjunction with @option{max-aliased-vops}. If a function
7154 contains more than @option{max-aliased-vops} virtual operators, then
7155 memory symbols will be grouped into memory partitions until either the
7156 total number of virtual operators is below @option{max-aliased-vops}
7157 or the average number of virtual operators per memory statement is
7158 below @option{avg-aliased-vops}. The default value for this parameter
7159 is 1 for -O1 and -O2, and 3 for -O3.
7161 @item ggc-min-expand
7163 GCC uses a garbage collector to manage its own memory allocation. This
7164 parameter specifies the minimum percentage by which the garbage
7165 collector's heap should be allowed to expand between collections.
7166 Tuning this may improve compilation speed; it has no effect on code
7169 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
7170 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
7171 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
7172 GCC is not able to calculate RAM on a particular platform, the lower
7173 bound of 30% is used. Setting this parameter and
7174 @option{ggc-min-heapsize} to zero causes a full collection to occur at
7175 every opportunity. This is extremely slow, but can be useful for
7178 @item ggc-min-heapsize
7180 Minimum size of the garbage collector's heap before it begins bothering
7181 to collect garbage. The first collection occurs after the heap expands
7182 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
7183 tuning this may improve compilation speed, and has no effect on code
7186 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
7187 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
7188 with a lower bound of 4096 (four megabytes) and an upper bound of
7189 131072 (128 megabytes). If GCC is not able to calculate RAM on a
7190 particular platform, the lower bound is used. Setting this parameter
7191 very large effectively disables garbage collection. Setting this
7192 parameter and @option{ggc-min-expand} to zero causes a full collection
7193 to occur at every opportunity.
7195 @item max-reload-search-insns
7196 The maximum number of instruction reload should look backward for equivalent
7197 register. Increasing values mean more aggressive optimization, making the
7198 compile time increase with probably slightly better performance. The default
7201 @item max-cselib-memory-locations
7202 The maximum number of memory locations cselib should take into account.
7203 Increasing values mean more aggressive optimization, making the compile time
7204 increase with probably slightly better performance. The default value is 500.
7206 @item max-flow-memory-locations
7207 Similar as @option{max-cselib-memory-locations} but for dataflow liveness.
7208 The default value is 100.
7210 @item reorder-blocks-duplicate
7211 @itemx reorder-blocks-duplicate-feedback
7213 Used by basic block reordering pass to decide whether to use unconditional
7214 branch or duplicate the code on its destination. Code is duplicated when its
7215 estimated size is smaller than this value multiplied by the estimated size of
7216 unconditional jump in the hot spots of the program.
7218 The @option{reorder-block-duplicate-feedback} is used only when profile
7219 feedback is available and may be set to higher values than
7220 @option{reorder-block-duplicate} since information about the hot spots is more
7223 @item max-sched-ready-insns
7224 The maximum number of instructions ready to be issued the scheduler should
7225 consider at any given time during the first scheduling pass. Increasing
7226 values mean more thorough searches, making the compilation time increase
7227 with probably little benefit. The default value is 100.
7229 @item max-sched-region-blocks
7230 The maximum number of blocks in a region to be considered for
7231 interblock scheduling. The default value is 10.
7233 @item max-sched-region-insns
7234 The maximum number of insns in a region to be considered for
7235 interblock scheduling. The default value is 100.
7238 The minimum probability (in percents) of reaching a source block
7239 for interblock speculative scheduling. The default value is 40.
7241 @item max-sched-extend-regions-iters
7242 The maximum number of iterations through CFG to extend regions.
7243 0 - disable region extension,
7244 N - do at most N iterations.
7245 The default value is 0.
7247 @item max-sched-insn-conflict-delay
7248 The maximum conflict delay for an insn to be considered for speculative motion.
7249 The default value is 3.
7251 @item sched-spec-prob-cutoff
7252 The minimal probability of speculation success (in percents), so that
7253 speculative insn will be scheduled.
7254 The default value is 40.
7256 @item max-last-value-rtl
7258 The maximum size measured as number of RTLs that can be recorded in an expression
7259 in combiner for a pseudo register as last known value of that register. The default
7262 @item integer-share-limit
7263 Small integer constants can use a shared data structure, reducing the
7264 compiler's memory usage and increasing its speed. This sets the maximum
7265 value of a shared integer constant. The default value is 256.
7267 @item min-virtual-mappings
7268 Specifies the minimum number of virtual mappings in the incremental
7269 SSA updater that should be registered to trigger the virtual mappings
7270 heuristic defined by virtual-mappings-ratio. The default value is
7273 @item virtual-mappings-ratio
7274 If the number of virtual mappings is virtual-mappings-ratio bigger
7275 than the number of virtual symbols to be updated, then the incremental
7276 SSA updater switches to a full update for those symbols. The default
7279 @item ssp-buffer-size
7280 The minimum size of buffers (i.e.@: arrays) that will receive stack smashing
7281 protection when @option{-fstack-protection} is used.
7283 @item max-jump-thread-duplication-stmts
7284 Maximum number of statements allowed in a block that needs to be
7285 duplicated when threading jumps.
7287 @item max-fields-for-field-sensitive
7288 Maximum number of fields in a structure we will treat in
7289 a field sensitive manner during pointer analysis.
7291 @item prefetch-latency
7292 Estimate on average number of instructions that are executed before
7293 prefetch finishes. The distance we prefetch ahead is proportional
7294 to this constant. Increasing this number may also lead to less
7295 streams being prefetched (see @option{simultaneous-prefetches}).
7297 @item simultaneous-prefetches
7298 Maximum number of prefetches that can run at the same time.
7300 @item l1-cache-line-size
7301 The size of cache line in L1 cache, in bytes.
7304 The size of L1 cache, in kilobytes.
7307 The size of L2 cache, in kilobytes.
7309 @item use-canonical-types
7310 Whether the compiler should use the ``canonical'' type system. By
7311 default, this should always be 1, which uses a more efficient internal
7312 mechanism for comparing types in C++ and Objective-C++. However, if
7313 bugs in the canonical type system are causing compilation failures,
7314 set this value to 0 to disable canonical types.
7316 @item max-partial-antic-length
7317 Maximum length of the partial antic set computed during the tree
7318 partial redundancy elimination optimization (@option{-ftree-pre}) when
7319 optimizing at @option{-O3} and above. For some sorts of source code
7320 the enhanced partial redundancy elimination optimization can run away,
7321 consuming all of the memory available on the host machine. This
7322 parameter sets a limit on the length of the sets that are computed,
7323 which prevents the runaway behaviour. Setting a value of 0 for
7324 this paramter will allow an unlimited set length.
7326 @item sccvn-max-scc-size
7327 Maximum size of a strongly connected component (SCC) during SCCVN
7328 processing. If this limit is hit, SCCVN processing for the whole
7329 function will not be done and optimizations depending on it will
7330 be disabled. The default maximum SCC size is 10000.
7335 @node Preprocessor Options
7336 @section Options Controlling the Preprocessor
7337 @cindex preprocessor options
7338 @cindex options, preprocessor
7340 These options control the C preprocessor, which is run on each C source
7341 file before actual compilation.
7343 If you use the @option{-E} option, nothing is done except preprocessing.
7344 Some of these options make sense only together with @option{-E} because
7345 they cause the preprocessor output to be unsuitable for actual
7350 You can use @option{-Wp,@var{option}} to bypass the compiler driver
7351 and pass @var{option} directly through to the preprocessor. If
7352 @var{option} contains commas, it is split into multiple options at the
7353 commas. However, many options are modified, translated or interpreted
7354 by the compiler driver before being passed to the preprocessor, and
7355 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
7356 interface is undocumented and subject to change, so whenever possible
7357 you should avoid using @option{-Wp} and let the driver handle the
7360 @item -Xpreprocessor @var{option}
7361 @opindex preprocessor
7362 Pass @var{option} as an option to the preprocessor. You can use this to
7363 supply system-specific preprocessor options which GCC does not know how to
7366 If you want to pass an option that takes an argument, you must use
7367 @option{-Xpreprocessor} twice, once for the option and once for the argument.
7370 @include cppopts.texi
7372 @node Assembler Options
7373 @section Passing Options to the Assembler
7375 @c prevent bad page break with this line
7376 You can pass options to the assembler.
7379 @item -Wa,@var{option}
7381 Pass @var{option} as an option to the assembler. If @var{option}
7382 contains commas, it is split into multiple options at the commas.
7384 @item -Xassembler @var{option}
7386 Pass @var{option} as an option to the assembler. You can use this to
7387 supply system-specific assembler options which GCC does not know how to
7390 If you want to pass an option that takes an argument, you must use
7391 @option{-Xassembler} twice, once for the option and once for the argument.
7396 @section Options for Linking
7397 @cindex link options
7398 @cindex options, linking
7400 These options come into play when the compiler links object files into
7401 an executable output file. They are meaningless if the compiler is
7402 not doing a link step.
7406 @item @var{object-file-name}
7407 A file name that does not end in a special recognized suffix is
7408 considered to name an object file or library. (Object files are
7409 distinguished from libraries by the linker according to the file
7410 contents.) If linking is done, these object files are used as input
7419 If any of these options is used, then the linker is not run, and
7420 object file names should not be used as arguments. @xref{Overall
7424 @item -l@var{library}
7425 @itemx -l @var{library}
7427 Search the library named @var{library} when linking. (The second
7428 alternative with the library as a separate argument is only for
7429 POSIX compliance and is not recommended.)
7431 It makes a difference where in the command you write this option; the
7432 linker searches and processes libraries and object files in the order they
7433 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
7434 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
7435 to functions in @samp{z}, those functions may not be loaded.
7437 The linker searches a standard list of directories for the library,
7438 which is actually a file named @file{lib@var{library}.a}. The linker
7439 then uses this file as if it had been specified precisely by name.
7441 The directories searched include several standard system directories
7442 plus any that you specify with @option{-L}.
7444 Normally the files found this way are library files---archive files
7445 whose members are object files. The linker handles an archive file by
7446 scanning through it for members which define symbols that have so far
7447 been referenced but not defined. But if the file that is found is an
7448 ordinary object file, it is linked in the usual fashion. The only
7449 difference between using an @option{-l} option and specifying a file name
7450 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
7451 and searches several directories.
7455 You need this special case of the @option{-l} option in order to
7456 link an Objective-C or Objective-C++ program.
7459 @opindex nostartfiles
7460 Do not use the standard system startup files when linking.
7461 The standard system libraries are used normally, unless @option{-nostdlib}
7462 or @option{-nodefaultlibs} is used.
7464 @item -nodefaultlibs
7465 @opindex nodefaultlibs
7466 Do not use the standard system libraries when linking.
7467 Only the libraries you specify will be passed to the linker.
7468 The standard startup files are used normally, unless @option{-nostartfiles}
7469 is used. The compiler may generate calls to @code{memcmp},
7470 @code{memset}, @code{memcpy} and @code{memmove}.
7471 These entries are usually resolved by entries in
7472 libc. These entry points should be supplied through some other
7473 mechanism when this option is specified.
7477 Do not use the standard system startup files or libraries when linking.
7478 No startup files and only the libraries you specify will be passed to
7479 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
7480 @code{memcpy} and @code{memmove}.
7481 These entries are usually resolved by entries in
7482 libc. These entry points should be supplied through some other
7483 mechanism when this option is specified.
7485 @cindex @option{-lgcc}, use with @option{-nostdlib}
7486 @cindex @option{-nostdlib} and unresolved references
7487 @cindex unresolved references and @option{-nostdlib}
7488 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
7489 @cindex @option{-nodefaultlibs} and unresolved references
7490 @cindex unresolved references and @option{-nodefaultlibs}
7491 One of the standard libraries bypassed by @option{-nostdlib} and
7492 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
7493 that GCC uses to overcome shortcomings of particular machines, or special
7494 needs for some languages.
7495 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
7496 Collection (GCC) Internals},
7497 for more discussion of @file{libgcc.a}.)
7498 In most cases, you need @file{libgcc.a} even when you want to avoid
7499 other standard libraries. In other words, when you specify @option{-nostdlib}
7500 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
7501 This ensures that you have no unresolved references to internal GCC
7502 library subroutines. (For example, @samp{__main}, used to ensure C++
7503 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
7504 GNU Compiler Collection (GCC) Internals}.)
7508 Produce a position independent executable on targets which support it.
7509 For predictable results, you must also specify the same set of options
7510 that were used to generate code (@option{-fpie}, @option{-fPIE},
7511 or model suboptions) when you specify this option.
7515 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
7516 that support it. This instructs the linker to add all symbols, not
7517 only used ones, to the dynamic symbol table. This option is needed
7518 for some uses of @code{dlopen} or to allow obtaining backtraces
7519 from within a program.
7523 Remove all symbol table and relocation information from the executable.
7527 On systems that support dynamic linking, this prevents linking with the shared
7528 libraries. On other systems, this option has no effect.
7532 Produce a shared object which can then be linked with other objects to
7533 form an executable. Not all systems support this option. For predictable
7534 results, you must also specify the same set of options that were used to
7535 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
7536 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
7537 needs to build supplementary stub code for constructors to work. On
7538 multi-libbed systems, @samp{gcc -shared} must select the correct support
7539 libraries to link against. Failing to supply the correct flags may lead
7540 to subtle defects. Supplying them in cases where they are not necessary
7543 @item -shared-libgcc
7544 @itemx -static-libgcc
7545 @opindex shared-libgcc
7546 @opindex static-libgcc
7547 On systems that provide @file{libgcc} as a shared library, these options
7548 force the use of either the shared or static version respectively.
7549 If no shared version of @file{libgcc} was built when the compiler was
7550 configured, these options have no effect.
7552 There are several situations in which an application should use the
7553 shared @file{libgcc} instead of the static version. The most common
7554 of these is when the application wishes to throw and catch exceptions
7555 across different shared libraries. In that case, each of the libraries
7556 as well as the application itself should use the shared @file{libgcc}.
7558 Therefore, the G++ and GCJ drivers automatically add
7559 @option{-shared-libgcc} whenever you build a shared library or a main
7560 executable, because C++ and Java programs typically use exceptions, so
7561 this is the right thing to do.
7563 If, instead, you use the GCC driver to create shared libraries, you may
7564 find that they will not always be linked with the shared @file{libgcc}.
7565 If GCC finds, at its configuration time, that you have a non-GNU linker
7566 or a GNU linker that does not support option @option{--eh-frame-hdr},
7567 it will link the shared version of @file{libgcc} into shared libraries
7568 by default. Otherwise, it will take advantage of the linker and optimize
7569 away the linking with the shared version of @file{libgcc}, linking with
7570 the static version of libgcc by default. This allows exceptions to
7571 propagate through such shared libraries, without incurring relocation
7572 costs at library load time.
7574 However, if a library or main executable is supposed to throw or catch
7575 exceptions, you must link it using the G++ or GCJ driver, as appropriate
7576 for the languages used in the program, or using the option
7577 @option{-shared-libgcc}, such that it is linked with the shared
7582 Bind references to global symbols when building a shared object. Warn
7583 about any unresolved references (unless overridden by the link editor
7584 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
7587 @item -Xlinker @var{option}
7589 Pass @var{option} as an option to the linker. You can use this to
7590 supply system-specific linker options which GCC does not know how to
7593 If you want to pass an option that takes an argument, you must use
7594 @option{-Xlinker} twice, once for the option and once for the argument.
7595 For example, to pass @option{-assert definitions}, you must write
7596 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
7597 @option{-Xlinker "-assert definitions"}, because this passes the entire
7598 string as a single argument, which is not what the linker expects.
7600 @item -Wl,@var{option}
7602 Pass @var{option} as an option to the linker. If @var{option} contains
7603 commas, it is split into multiple options at the commas.
7605 @item -u @var{symbol}
7607 Pretend the symbol @var{symbol} is undefined, to force linking of
7608 library modules to define it. You can use @option{-u} multiple times with
7609 different symbols to force loading of additional library modules.
7612 @node Directory Options
7613 @section Options for Directory Search
7614 @cindex directory options
7615 @cindex options, directory search
7618 These options specify directories to search for header files, for
7619 libraries and for parts of the compiler:
7624 Add the directory @var{dir} to the head of the list of directories to be
7625 searched for header files. This can be used to override a system header
7626 file, substituting your own version, since these directories are
7627 searched before the system header file directories. However, you should
7628 not use this option to add directories that contain vendor-supplied
7629 system header files (use @option{-isystem} for that). If you use more than
7630 one @option{-I} option, the directories are scanned in left-to-right
7631 order; the standard system directories come after.
7633 If a standard system include directory, or a directory specified with
7634 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
7635 option will be ignored. The directory will still be searched but as a
7636 system directory at its normal position in the system include chain.
7637 This is to ensure that GCC's procedure to fix buggy system headers and
7638 the ordering for the include_next directive are not inadvertently changed.
7639 If you really need to change the search order for system directories,
7640 use the @option{-nostdinc} and/or @option{-isystem} options.
7642 @item -iquote@var{dir}
7644 Add the directory @var{dir} to the head of the list of directories to
7645 be searched for header files only for the case of @samp{#include
7646 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
7647 otherwise just like @option{-I}.
7651 Add directory @var{dir} to the list of directories to be searched
7654 @item -B@var{prefix}
7656 This option specifies where to find the executables, libraries,
7657 include files, and data files of the compiler itself.
7659 The compiler driver program runs one or more of the subprograms
7660 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
7661 @var{prefix} as a prefix for each program it tries to run, both with and
7662 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
7664 For each subprogram to be run, the compiler driver first tries the
7665 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
7666 was not specified, the driver tries two standard prefixes, which are
7667 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
7668 those results in a file name that is found, the unmodified program
7669 name is searched for using the directories specified in your
7670 @env{PATH} environment variable.
7672 The compiler will check to see if the path provided by the @option{-B}
7673 refers to a directory, and if necessary it will add a directory
7674 separator character at the end of the path.
7676 @option{-B} prefixes that effectively specify directory names also apply
7677 to libraries in the linker, because the compiler translates these
7678 options into @option{-L} options for the linker. They also apply to
7679 includes files in the preprocessor, because the compiler translates these
7680 options into @option{-isystem} options for the preprocessor. In this case,
7681 the compiler appends @samp{include} to the prefix.
7683 The run-time support file @file{libgcc.a} can also be searched for using
7684 the @option{-B} prefix, if needed. If it is not found there, the two
7685 standard prefixes above are tried, and that is all. The file is left
7686 out of the link if it is not found by those means.
7688 Another way to specify a prefix much like the @option{-B} prefix is to use
7689 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
7692 As a special kludge, if the path provided by @option{-B} is
7693 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
7694 9, then it will be replaced by @file{[dir/]include}. This is to help
7695 with boot-strapping the compiler.
7697 @item -specs=@var{file}
7699 Process @var{file} after the compiler reads in the standard @file{specs}
7700 file, in order to override the defaults that the @file{gcc} driver
7701 program uses when determining what switches to pass to @file{cc1},
7702 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
7703 @option{-specs=@var{file}} can be specified on the command line, and they
7704 are processed in order, from left to right.
7706 @item --sysroot=@var{dir}
7708 Use @var{dir} as the logical root directory for headers and libraries.
7709 For example, if the compiler would normally search for headers in
7710 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
7711 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
7713 If you use both this option and the @option{-isysroot} option, then
7714 the @option{--sysroot} option will apply to libraries, but the
7715 @option{-isysroot} option will apply to header files.
7717 The GNU linker (beginning with version 2.16) has the necessary support
7718 for this option. If your linker does not support this option, the
7719 header file aspect of @option{--sysroot} will still work, but the
7720 library aspect will not.
7724 This option has been deprecated. Please use @option{-iquote} instead for
7725 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
7726 Any directories you specify with @option{-I} options before the @option{-I-}
7727 option are searched only for the case of @samp{#include "@var{file}"};
7728 they are not searched for @samp{#include <@var{file}>}.
7730 If additional directories are specified with @option{-I} options after
7731 the @option{-I-}, these directories are searched for all @samp{#include}
7732 directives. (Ordinarily @emph{all} @option{-I} directories are used
7735 In addition, the @option{-I-} option inhibits the use of the current
7736 directory (where the current input file came from) as the first search
7737 directory for @samp{#include "@var{file}"}. There is no way to
7738 override this effect of @option{-I-}. With @option{-I.} you can specify
7739 searching the directory which was current when the compiler was
7740 invoked. That is not exactly the same as what the preprocessor does
7741 by default, but it is often satisfactory.
7743 @option{-I-} does not inhibit the use of the standard system directories
7744 for header files. Thus, @option{-I-} and @option{-nostdinc} are
7751 @section Specifying subprocesses and the switches to pass to them
7754 @command{gcc} is a driver program. It performs its job by invoking a
7755 sequence of other programs to do the work of compiling, assembling and
7756 linking. GCC interprets its command-line parameters and uses these to
7757 deduce which programs it should invoke, and which command-line options
7758 it ought to place on their command lines. This behavior is controlled
7759 by @dfn{spec strings}. In most cases there is one spec string for each
7760 program that GCC can invoke, but a few programs have multiple spec
7761 strings to control their behavior. The spec strings built into GCC can
7762 be overridden by using the @option{-specs=} command-line switch to specify
7765 @dfn{Spec files} are plaintext files that are used to construct spec
7766 strings. They consist of a sequence of directives separated by blank
7767 lines. The type of directive is determined by the first non-whitespace
7768 character on the line and it can be one of the following:
7771 @item %@var{command}
7772 Issues a @var{command} to the spec file processor. The commands that can
7776 @item %include <@var{file}>
7778 Search for @var{file} and insert its text at the current point in the
7781 @item %include_noerr <@var{file}>
7782 @cindex %include_noerr
7783 Just like @samp{%include}, but do not generate an error message if the include
7784 file cannot be found.
7786 @item %rename @var{old_name} @var{new_name}
7788 Rename the spec string @var{old_name} to @var{new_name}.
7792 @item *[@var{spec_name}]:
7793 This tells the compiler to create, override or delete the named spec
7794 string. All lines after this directive up to the next directive or
7795 blank line are considered to be the text for the spec string. If this
7796 results in an empty string then the spec will be deleted. (Or, if the
7797 spec did not exist, then nothing will happened.) Otherwise, if the spec
7798 does not currently exist a new spec will be created. If the spec does
7799 exist then its contents will be overridden by the text of this
7800 directive, unless the first character of that text is the @samp{+}
7801 character, in which case the text will be appended to the spec.
7803 @item [@var{suffix}]:
7804 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
7805 and up to the next directive or blank line are considered to make up the
7806 spec string for the indicated suffix. When the compiler encounters an
7807 input file with the named suffix, it will processes the spec string in
7808 order to work out how to compile that file. For example:
7815 This says that any input file whose name ends in @samp{.ZZ} should be
7816 passed to the program @samp{z-compile}, which should be invoked with the
7817 command-line switch @option{-input} and with the result of performing the
7818 @samp{%i} substitution. (See below.)
7820 As an alternative to providing a spec string, the text that follows a
7821 suffix directive can be one of the following:
7824 @item @@@var{language}
7825 This says that the suffix is an alias for a known @var{language}. This is
7826 similar to using the @option{-x} command-line switch to GCC to specify a
7827 language explicitly. For example:
7834 Says that .ZZ files are, in fact, C++ source files.
7837 This causes an error messages saying:
7840 @var{name} compiler not installed on this system.
7844 GCC already has an extensive list of suffixes built into it.
7845 This directive will add an entry to the end of the list of suffixes, but
7846 since the list is searched from the end backwards, it is effectively
7847 possible to override earlier entries using this technique.
7851 GCC has the following spec strings built into it. Spec files can
7852 override these strings or create their own. Note that individual
7853 targets can also add their own spec strings to this list.
7856 asm Options to pass to the assembler
7857 asm_final Options to pass to the assembler post-processor
7858 cpp Options to pass to the C preprocessor
7859 cc1 Options to pass to the C compiler
7860 cc1plus Options to pass to the C++ compiler
7861 endfile Object files to include at the end of the link
7862 link Options to pass to the linker
7863 lib Libraries to include on the command line to the linker
7864 libgcc Decides which GCC support library to pass to the linker
7865 linker Sets the name of the linker
7866 predefines Defines to be passed to the C preprocessor
7867 signed_char Defines to pass to CPP to say whether @code{char} is signed
7869 startfile Object files to include at the start of the link
7872 Here is a small example of a spec file:
7878 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
7881 This example renames the spec called @samp{lib} to @samp{old_lib} and
7882 then overrides the previous definition of @samp{lib} with a new one.
7883 The new definition adds in some extra command-line options before
7884 including the text of the old definition.
7886 @dfn{Spec strings} are a list of command-line options to be passed to their
7887 corresponding program. In addition, the spec strings can contain
7888 @samp{%}-prefixed sequences to substitute variable text or to
7889 conditionally insert text into the command line. Using these constructs
7890 it is possible to generate quite complex command lines.
7892 Here is a table of all defined @samp{%}-sequences for spec
7893 strings. Note that spaces are not generated automatically around the
7894 results of expanding these sequences. Therefore you can concatenate them
7895 together or combine them with constant text in a single argument.
7899 Substitute one @samp{%} into the program name or argument.
7902 Substitute the name of the input file being processed.
7905 Substitute the basename of the input file being processed.
7906 This is the substring up to (and not including) the last period
7907 and not including the directory.
7910 This is the same as @samp{%b}, but include the file suffix (text after
7914 Marks the argument containing or following the @samp{%d} as a
7915 temporary file name, so that that file will be deleted if GCC exits
7916 successfully. Unlike @samp{%g}, this contributes no text to the
7919 @item %g@var{suffix}
7920 Substitute a file name that has suffix @var{suffix} and is chosen
7921 once per compilation, and mark the argument in the same way as
7922 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
7923 name is now chosen in a way that is hard to predict even when previously
7924 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
7925 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
7926 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
7927 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
7928 was simply substituted with a file name chosen once per compilation,
7929 without regard to any appended suffix (which was therefore treated
7930 just like ordinary text), making such attacks more likely to succeed.
7932 @item %u@var{suffix}
7933 Like @samp{%g}, but generates a new temporary file name even if
7934 @samp{%u@var{suffix}} was already seen.
7936 @item %U@var{suffix}
7937 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
7938 new one if there is no such last file name. In the absence of any
7939 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
7940 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
7941 would involve the generation of two distinct file names, one
7942 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
7943 simply substituted with a file name chosen for the previous @samp{%u},
7944 without regard to any appended suffix.
7946 @item %j@var{suffix}
7947 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
7948 writable, and if save-temps is off; otherwise, substitute the name
7949 of a temporary file, just like @samp{%u}. This temporary file is not
7950 meant for communication between processes, but rather as a junk
7953 @item %|@var{suffix}
7954 @itemx %m@var{suffix}
7955 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
7956 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
7957 all. These are the two most common ways to instruct a program that it
7958 should read from standard input or write to standard output. If you
7959 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
7960 construct: see for example @file{f/lang-specs.h}.
7962 @item %.@var{SUFFIX}
7963 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
7964 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
7965 terminated by the next space or %.
7968 Marks the argument containing or following the @samp{%w} as the
7969 designated output file of this compilation. This puts the argument
7970 into the sequence of arguments that @samp{%o} will substitute later.
7973 Substitutes the names of all the output files, with spaces
7974 automatically placed around them. You should write spaces
7975 around the @samp{%o} as well or the results are undefined.
7976 @samp{%o} is for use in the specs for running the linker.
7977 Input files whose names have no recognized suffix are not compiled
7978 at all, but they are included among the output files, so they will
7982 Substitutes the suffix for object files. Note that this is
7983 handled specially when it immediately follows @samp{%g, %u, or %U},
7984 because of the need for those to form complete file names. The
7985 handling is such that @samp{%O} is treated exactly as if it had already
7986 been substituted, except that @samp{%g, %u, and %U} do not currently
7987 support additional @var{suffix} characters following @samp{%O} as they would
7988 following, for example, @samp{.o}.
7991 Substitutes the standard macro predefinitions for the
7992 current target machine. Use this when running @code{cpp}.
7995 Like @samp{%p}, but puts @samp{__} before and after the name of each
7996 predefined macro, except for macros that start with @samp{__} or with
7997 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
8001 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
8002 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
8003 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
8004 and @option{-imultilib} as necessary.
8007 Current argument is the name of a library or startup file of some sort.
8008 Search for that file in a standard list of directories and substitute
8009 the full name found.
8012 Print @var{str} as an error message. @var{str} is terminated by a newline.
8013 Use this when inconsistent options are detected.
8016 Substitute the contents of spec string @var{name} at this point.
8019 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
8021 @item %x@{@var{option}@}
8022 Accumulate an option for @samp{%X}.
8025 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
8029 Output the accumulated assembler options specified by @option{-Wa}.
8032 Output the accumulated preprocessor options specified by @option{-Wp}.
8035 Process the @code{asm} spec. This is used to compute the
8036 switches to be passed to the assembler.
8039 Process the @code{asm_final} spec. This is a spec string for
8040 passing switches to an assembler post-processor, if such a program is
8044 Process the @code{link} spec. This is the spec for computing the
8045 command line passed to the linker. Typically it will make use of the
8046 @samp{%L %G %S %D and %E} sequences.
8049 Dump out a @option{-L} option for each directory that GCC believes might
8050 contain startup files. If the target supports multilibs then the
8051 current multilib directory will be prepended to each of these paths.
8054 Process the @code{lib} spec. This is a spec string for deciding which
8055 libraries should be included on the command line to the linker.
8058 Process the @code{libgcc} spec. This is a spec string for deciding
8059 which GCC support library should be included on the command line to the linker.
8062 Process the @code{startfile} spec. This is a spec for deciding which
8063 object files should be the first ones passed to the linker. Typically
8064 this might be a file named @file{crt0.o}.
8067 Process the @code{endfile} spec. This is a spec string that specifies
8068 the last object files that will be passed to the linker.
8071 Process the @code{cpp} spec. This is used to construct the arguments
8072 to be passed to the C preprocessor.
8075 Process the @code{cc1} spec. This is used to construct the options to be
8076 passed to the actual C compiler (@samp{cc1}).
8079 Process the @code{cc1plus} spec. This is used to construct the options to be
8080 passed to the actual C++ compiler (@samp{cc1plus}).
8083 Substitute the variable part of a matched option. See below.
8084 Note that each comma in the substituted string is replaced by
8088 Remove all occurrences of @code{-S} from the command line. Note---this
8089 command is position dependent. @samp{%} commands in the spec string
8090 before this one will see @code{-S}, @samp{%} commands in the spec string
8091 after this one will not.
8093 @item %:@var{function}(@var{args})
8094 Call the named function @var{function}, passing it @var{args}.
8095 @var{args} is first processed as a nested spec string, then split
8096 into an argument vector in the usual fashion. The function returns
8097 a string which is processed as if it had appeared literally as part
8098 of the current spec.
8100 The following built-in spec functions are provided:
8104 The @code{getenv} spec function takes two arguments: an environment
8105 variable name and a string. If the environment variable is not
8106 defined, a fatal error is issued. Otherwise, the return value is the
8107 value of the environment variable concatenated with the string. For
8108 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
8111 %:getenv(TOPDIR /include)
8114 expands to @file{/path/to/top/include}.
8116 @item @code{if-exists}
8117 The @code{if-exists} spec function takes one argument, an absolute
8118 pathname to a file. If the file exists, @code{if-exists} returns the
8119 pathname. Here is a small example of its usage:
8123 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
8126 @item @code{if-exists-else}
8127 The @code{if-exists-else} spec function is similar to the @code{if-exists}
8128 spec function, except that it takes two arguments. The first argument is
8129 an absolute pathname to a file. If the file exists, @code{if-exists-else}
8130 returns the pathname. If it does not exist, it returns the second argument.
8131 This way, @code{if-exists-else} can be used to select one file or another,
8132 based on the existence of the first. Here is a small example of its usage:
8136 crt0%O%s %:if-exists(crti%O%s) \
8137 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
8140 @item @code{replace-outfile}
8141 The @code{replace-outfile} spec function takes two arguments. It looks for the
8142 first argument in the outfiles array and replaces it with the second argument. Here
8143 is a small example of its usage:
8146 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
8149 @item @code{print-asm-header}
8150 The @code{print-asm-header} function takes no arguments and simply
8151 prints a banner like:
8157 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
8160 It is used to separate compiler options from assembler options
8161 in the @option{--target-help} output.
8165 Substitutes the @code{-S} switch, if that switch was given to GCC@.
8166 If that switch was not specified, this substitutes nothing. Note that
8167 the leading dash is omitted when specifying this option, and it is
8168 automatically inserted if the substitution is performed. Thus the spec
8169 string @samp{%@{foo@}} would match the command-line option @option{-foo}
8170 and would output the command line option @option{-foo}.
8172 @item %W@{@code{S}@}
8173 Like %@{@code{S}@} but mark last argument supplied within as a file to be
8176 @item %@{@code{S}*@}
8177 Substitutes all the switches specified to GCC whose names start
8178 with @code{-S}, but which also take an argument. This is used for
8179 switches like @option{-o}, @option{-D}, @option{-I}, etc.
8180 GCC considers @option{-o foo} as being
8181 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
8182 text, including the space. Thus two arguments would be generated.
8184 @item %@{@code{S}*&@code{T}*@}
8185 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
8186 (the order of @code{S} and @code{T} in the spec is not significant).
8187 There can be any number of ampersand-separated variables; for each the
8188 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
8190 @item %@{@code{S}:@code{X}@}
8191 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
8193 @item %@{!@code{S}:@code{X}@}
8194 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
8196 @item %@{@code{S}*:@code{X}@}
8197 Substitutes @code{X} if one or more switches whose names start with
8198 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
8199 once, no matter how many such switches appeared. However, if @code{%*}
8200 appears somewhere in @code{X}, then @code{X} will be substituted once
8201 for each matching switch, with the @code{%*} replaced by the part of
8202 that switch that matched the @code{*}.
8204 @item %@{.@code{S}:@code{X}@}
8205 Substitutes @code{X}, if processing a file with suffix @code{S}.
8207 @item %@{!.@code{S}:@code{X}@}
8208 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
8210 @item %@{,@code{S}:@code{X}@}
8211 Substitutes @code{X}, if processing a file for language @code{S}.
8213 @item %@{!,@code{S}:@code{X}@}
8214 Substitutes @code{X}, if not processing a file for language @code{S}.
8216 @item %@{@code{S}|@code{P}:@code{X}@}
8217 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
8218 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
8219 @code{*} sequences as well, although they have a stronger binding than
8220 the @samp{|}. If @code{%*} appears in @code{X}, all of the
8221 alternatives must be starred, and only the first matching alternative
8224 For example, a spec string like this:
8227 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
8230 will output the following command-line options from the following input
8231 command-line options:
8236 -d fred.c -foo -baz -boggle
8237 -d jim.d -bar -baz -boggle
8240 @item %@{S:X; T:Y; :D@}
8242 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
8243 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
8244 be as many clauses as you need. This may be combined with @code{.},
8245 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
8250 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
8251 construct may contain other nested @samp{%} constructs or spaces, or
8252 even newlines. They are processed as usual, as described above.
8253 Trailing white space in @code{X} is ignored. White space may also
8254 appear anywhere on the left side of the colon in these constructs,
8255 except between @code{.} or @code{*} and the corresponding word.
8257 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
8258 handled specifically in these constructs. If another value of
8259 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
8260 @option{-W} switch is found later in the command line, the earlier
8261 switch value is ignored, except with @{@code{S}*@} where @code{S} is
8262 just one letter, which passes all matching options.
8264 The character @samp{|} at the beginning of the predicate text is used to
8265 indicate that a command should be piped to the following command, but
8266 only if @option{-pipe} is specified.
8268 It is built into GCC which switches take arguments and which do not.
8269 (You might think it would be useful to generalize this to allow each
8270 compiler's spec to say which switches take arguments. But this cannot
8271 be done in a consistent fashion. GCC cannot even decide which input
8272 files have been specified without knowing which switches take arguments,
8273 and it must know which input files to compile in order to tell which
8276 GCC also knows implicitly that arguments starting in @option{-l} are to be
8277 treated as compiler output files, and passed to the linker in their
8278 proper position among the other output files.
8280 @c man begin OPTIONS
8282 @node Target Options
8283 @section Specifying Target Machine and Compiler Version
8284 @cindex target options
8285 @cindex cross compiling
8286 @cindex specifying machine version
8287 @cindex specifying compiler version and target machine
8288 @cindex compiler version, specifying
8289 @cindex target machine, specifying
8291 The usual way to run GCC is to run the executable called @file{gcc}, or
8292 @file{<machine>-gcc} when cross-compiling, or
8293 @file{<machine>-gcc-<version>} to run a version other than the one that
8294 was installed last. Sometimes this is inconvenient, so GCC provides
8295 options that will switch to another cross-compiler or version.
8298 @item -b @var{machine}
8300 The argument @var{machine} specifies the target machine for compilation.
8302 The value to use for @var{machine} is the same as was specified as the
8303 machine type when configuring GCC as a cross-compiler. For
8304 example, if a cross-compiler was configured with @samp{configure
8305 arm-elf}, meaning to compile for an arm processor with elf binaries,
8306 then you would specify @option{-b arm-elf} to run that cross compiler.
8307 Because there are other options beginning with @option{-b}, the
8308 configuration must contain a hyphen.
8310 @item -V @var{version}
8312 The argument @var{version} specifies which version of GCC to run.
8313 This is useful when multiple versions are installed. For example,
8314 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
8317 The @option{-V} and @option{-b} options work by running the
8318 @file{<machine>-gcc-<version>} executable, so there's no real reason to
8319 use them if you can just run that directly.
8321 @node Submodel Options
8322 @section Hardware Models and Configurations
8323 @cindex submodel options
8324 @cindex specifying hardware config
8325 @cindex hardware models and configurations, specifying
8326 @cindex machine dependent options
8328 Earlier we discussed the standard option @option{-b} which chooses among
8329 different installed compilers for completely different target
8330 machines, such as VAX vs.@: 68000 vs.@: 80386.
8332 In addition, each of these target machine types can have its own
8333 special options, starting with @samp{-m}, to choose among various
8334 hardware models or configurations---for example, 68010 vs 68020,
8335 floating coprocessor or none. A single installed version of the
8336 compiler can compile for any model or configuration, according to the
8339 Some configurations of the compiler also support additional special
8340 options, usually for compatibility with other compilers on the same
8343 @c This list is ordered alphanumerically by subsection name.
8344 @c It should be the same order and spelling as these options are listed
8345 @c in Machine Dependent Options
8351 * Blackfin Options::
8355 * DEC Alpha Options::
8356 * DEC Alpha/VMS Options::
8358 * GNU/Linux Options::
8361 * i386 and x86-64 Options::
8374 * RS/6000 and PowerPC Options::
8375 * S/390 and zSeries Options::
8380 * System V Options::
8385 * Xstormy16 Options::
8391 @subsection ARC Options
8394 These options are defined for ARC implementations:
8399 Compile code for little endian mode. This is the default.
8403 Compile code for big endian mode.
8406 @opindex mmangle-cpu
8407 Prepend the name of the cpu to all public symbol names.
8408 In multiple-processor systems, there are many ARC variants with different
8409 instruction and register set characteristics. This flag prevents code
8410 compiled for one cpu to be linked with code compiled for another.
8411 No facility exists for handling variants that are ``almost identical''.
8412 This is an all or nothing option.
8414 @item -mcpu=@var{cpu}
8416 Compile code for ARC variant @var{cpu}.
8417 Which variants are supported depend on the configuration.
8418 All variants support @option{-mcpu=base}, this is the default.
8420 @item -mtext=@var{text-section}
8421 @itemx -mdata=@var{data-section}
8422 @itemx -mrodata=@var{readonly-data-section}
8426 Put functions, data, and readonly data in @var{text-section},
8427 @var{data-section}, and @var{readonly-data-section} respectively
8428 by default. This can be overridden with the @code{section} attribute.
8429 @xref{Variable Attributes}.
8434 @subsection ARM Options
8437 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
8441 @item -mabi=@var{name}
8443 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
8444 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
8447 @opindex mapcs-frame
8448 Generate a stack frame that is compliant with the ARM Procedure Call
8449 Standard for all functions, even if this is not strictly necessary for
8450 correct execution of the code. Specifying @option{-fomit-frame-pointer}
8451 with this option will cause the stack frames not to be generated for
8452 leaf functions. The default is @option{-mno-apcs-frame}.
8456 This is a synonym for @option{-mapcs-frame}.
8459 @c not currently implemented
8460 @item -mapcs-stack-check
8461 @opindex mapcs-stack-check
8462 Generate code to check the amount of stack space available upon entry to
8463 every function (that actually uses some stack space). If there is
8464 insufficient space available then either the function
8465 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
8466 called, depending upon the amount of stack space required. The run time
8467 system is required to provide these functions. The default is
8468 @option{-mno-apcs-stack-check}, since this produces smaller code.
8470 @c not currently implemented
8472 @opindex mapcs-float
8473 Pass floating point arguments using the float point registers. This is
8474 one of the variants of the APCS@. This option is recommended if the
8475 target hardware has a floating point unit or if a lot of floating point
8476 arithmetic is going to be performed by the code. The default is
8477 @option{-mno-apcs-float}, since integer only code is slightly increased in
8478 size if @option{-mapcs-float} is used.
8480 @c not currently implemented
8481 @item -mapcs-reentrant
8482 @opindex mapcs-reentrant
8483 Generate reentrant, position independent code. The default is
8484 @option{-mno-apcs-reentrant}.
8487 @item -mthumb-interwork
8488 @opindex mthumb-interwork
8489 Generate code which supports calling between the ARM and Thumb
8490 instruction sets. Without this option the two instruction sets cannot
8491 be reliably used inside one program. The default is
8492 @option{-mno-thumb-interwork}, since slightly larger code is generated
8493 when @option{-mthumb-interwork} is specified.
8495 @item -mno-sched-prolog
8496 @opindex mno-sched-prolog
8497 Prevent the reordering of instructions in the function prolog, or the
8498 merging of those instruction with the instructions in the function's
8499 body. This means that all functions will start with a recognizable set
8500 of instructions (or in fact one of a choice from a small set of
8501 different function prologues), and this information can be used to
8502 locate the start if functions inside an executable piece of code. The
8503 default is @option{-msched-prolog}.
8506 @opindex mhard-float
8507 Generate output containing floating point instructions. This is the
8511 @opindex msoft-float
8512 Generate output containing library calls for floating point.
8513 @strong{Warning:} the requisite libraries are not available for all ARM
8514 targets. Normally the facilities of the machine's usual C compiler are
8515 used, but this cannot be done directly in cross-compilation. You must make
8516 your own arrangements to provide suitable library functions for
8519 @option{-msoft-float} changes the calling convention in the output file;
8520 therefore, it is only useful if you compile @emph{all} of a program with
8521 this option. In particular, you need to compile @file{libgcc.a}, the
8522 library that comes with GCC, with @option{-msoft-float} in order for
8525 @item -mfloat-abi=@var{name}
8527 Specifies which ABI to use for floating point values. Permissible values
8528 are: @samp{soft}, @samp{softfp} and @samp{hard}.
8530 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
8531 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
8532 of floating point instructions, but still uses the soft-float calling
8535 @item -mlittle-endian
8536 @opindex mlittle-endian
8537 Generate code for a processor running in little-endian mode. This is
8538 the default for all standard configurations.
8541 @opindex mbig-endian
8542 Generate code for a processor running in big-endian mode; the default is
8543 to compile code for a little-endian processor.
8545 @item -mwords-little-endian
8546 @opindex mwords-little-endian
8547 This option only applies when generating code for big-endian processors.
8548 Generate code for a little-endian word order but a big-endian byte
8549 order. That is, a byte order of the form @samp{32107654}. Note: this
8550 option should only be used if you require compatibility with code for
8551 big-endian ARM processors generated by versions of the compiler prior to
8554 @item -mcpu=@var{name}
8556 This specifies the name of the target ARM processor. GCC uses this name
8557 to determine what kind of instructions it can emit when generating
8558 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
8559 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
8560 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
8561 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
8562 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
8563 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
8564 @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
8565 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
8566 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
8567 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
8568 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
8569 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
8570 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
8571 @samp{arm1156t2-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
8572 @samp{cortex-a8}, @samp{cortex-r4}, @samp{cortex-m3},
8573 @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
8575 @itemx -mtune=@var{name}
8577 This option is very similar to the @option{-mcpu=} option, except that
8578 instead of specifying the actual target processor type, and hence
8579 restricting which instructions can be used, it specifies that GCC should
8580 tune the performance of the code as if the target were of the type
8581 specified in this option, but still choosing the instructions that it
8582 will generate based on the cpu specified by a @option{-mcpu=} option.
8583 For some ARM implementations better performance can be obtained by using
8586 @item -march=@var{name}
8588 This specifies the name of the target ARM architecture. GCC uses this
8589 name to determine what kind of instructions it can emit when generating
8590 assembly code. This option can be used in conjunction with or instead
8591 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
8592 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
8593 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
8594 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv7}, @samp{armv7-a},
8595 @samp{armv7-r}, @samp{armv7-m}, @samp{iwmmxt}, @samp{ep9312}.
8597 @item -mfpu=@var{name}
8598 @itemx -mfpe=@var{number}
8599 @itemx -mfp=@var{number}
8603 This specifies what floating point hardware (or hardware emulation) is
8604 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
8605 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
8606 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
8607 with older versions of GCC@.
8609 If @option{-msoft-float} is specified this specifies the format of
8610 floating point values.
8612 @item -mstructure-size-boundary=@var{n}
8613 @opindex mstructure-size-boundary
8614 The size of all structures and unions will be rounded up to a multiple
8615 of the number of bits set by this option. Permissible values are 8, 32
8616 and 64. The default value varies for different toolchains. For the COFF
8617 targeted toolchain the default value is 8. A value of 64 is only allowed
8618 if the underlying ABI supports it.
8620 Specifying the larger number can produce faster, more efficient code, but
8621 can also increase the size of the program. Different values are potentially
8622 incompatible. Code compiled with one value cannot necessarily expect to
8623 work with code or libraries compiled with another value, if they exchange
8624 information using structures or unions.
8626 @item -mabort-on-noreturn
8627 @opindex mabort-on-noreturn
8628 Generate a call to the function @code{abort} at the end of a
8629 @code{noreturn} function. It will be executed if the function tries to
8633 @itemx -mno-long-calls
8634 @opindex mlong-calls
8635 @opindex mno-long-calls
8636 Tells the compiler to perform function calls by first loading the
8637 address of the function into a register and then performing a subroutine
8638 call on this register. This switch is needed if the target function
8639 will lie outside of the 64 megabyte addressing range of the offset based
8640 version of subroutine call instruction.
8642 Even if this switch is enabled, not all function calls will be turned
8643 into long calls. The heuristic is that static functions, functions
8644 which have the @samp{short-call} attribute, functions that are inside
8645 the scope of a @samp{#pragma no_long_calls} directive and functions whose
8646 definitions have already been compiled within the current compilation
8647 unit, will not be turned into long calls. The exception to this rule is
8648 that weak function definitions, functions with the @samp{long-call}
8649 attribute or the @samp{section} attribute, and functions that are within
8650 the scope of a @samp{#pragma long_calls} directive, will always be
8651 turned into long calls.
8653 This feature is not enabled by default. Specifying
8654 @option{-mno-long-calls} will restore the default behavior, as will
8655 placing the function calls within the scope of a @samp{#pragma
8656 long_calls_off} directive. Note these switches have no effect on how
8657 the compiler generates code to handle function calls via function
8660 @item -mnop-fun-dllimport
8661 @opindex mnop-fun-dllimport
8662 Disable support for the @code{dllimport} attribute.
8664 @item -msingle-pic-base
8665 @opindex msingle-pic-base
8666 Treat the register used for PIC addressing as read-only, rather than
8667 loading it in the prologue for each function. The run-time system is
8668 responsible for initializing this register with an appropriate value
8669 before execution begins.
8671 @item -mpic-register=@var{reg}
8672 @opindex mpic-register
8673 Specify the register to be used for PIC addressing. The default is R10
8674 unless stack-checking is enabled, when R9 is used.
8676 @item -mcirrus-fix-invalid-insns
8677 @opindex mcirrus-fix-invalid-insns
8678 @opindex mno-cirrus-fix-invalid-insns
8679 Insert NOPs into the instruction stream to in order to work around
8680 problems with invalid Maverick instruction combinations. This option
8681 is only valid if the @option{-mcpu=ep9312} option has been used to
8682 enable generation of instructions for the Cirrus Maverick floating
8683 point co-processor. This option is not enabled by default, since the
8684 problem is only present in older Maverick implementations. The default
8685 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
8688 @item -mpoke-function-name
8689 @opindex mpoke-function-name
8690 Write the name of each function into the text section, directly
8691 preceding the function prologue. The generated code is similar to this:
8695 .ascii "arm_poke_function_name", 0
8698 .word 0xff000000 + (t1 - t0)
8699 arm_poke_function_name
8701 stmfd sp!, @{fp, ip, lr, pc@}
8705 When performing a stack backtrace, code can inspect the value of
8706 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
8707 location @code{pc - 12} and the top 8 bits are set, then we know that
8708 there is a function name embedded immediately preceding this location
8709 and has length @code{((pc[-3]) & 0xff000000)}.
8713 Generate code for the Thumb instruction set. The default is to
8714 use the 32-bit ARM instruction set.
8715 This option automatically enables either 16-bit Thumb-1 or
8716 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
8717 and @option{-march=@var{name}} options.
8720 @opindex mtpcs-frame
8721 Generate a stack frame that is compliant with the Thumb Procedure Call
8722 Standard for all non-leaf functions. (A leaf function is one that does
8723 not call any other functions.) The default is @option{-mno-tpcs-frame}.
8725 @item -mtpcs-leaf-frame
8726 @opindex mtpcs-leaf-frame
8727 Generate a stack frame that is compliant with the Thumb Procedure Call
8728 Standard for all leaf functions. (A leaf function is one that does
8729 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
8731 @item -mcallee-super-interworking
8732 @opindex mcallee-super-interworking
8733 Gives all externally visible functions in the file being compiled an ARM
8734 instruction set header which switches to Thumb mode before executing the
8735 rest of the function. This allows these functions to be called from
8736 non-interworking code.
8738 @item -mcaller-super-interworking
8739 @opindex mcaller-super-interworking
8740 Allows calls via function pointers (including virtual functions) to
8741 execute correctly regardless of whether the target code has been
8742 compiled for interworking or not. There is a small overhead in the cost
8743 of executing a function pointer if this option is enabled.
8745 @item -mtp=@var{name}
8747 Specify the access model for the thread local storage pointer. The valid
8748 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
8749 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
8750 (supported in the arm6k architecture), and @option{auto}, which uses the
8751 best available method for the selected processor. The default setting is
8757 @subsection AVR Options
8760 These options are defined for AVR implementations:
8763 @item -mmcu=@var{mcu}
8765 Specify ATMEL AVR instruction set or MCU type.
8767 Instruction set avr1 is for the minimal AVR core, not supported by the C
8768 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
8769 attiny11, attiny12, attiny15, attiny28).
8771 Instruction set avr2 (default) is for the classic AVR core with up to
8772 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
8773 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
8774 at90c8534, at90s8535).
8776 Instruction set avr3 is for the classic AVR core with up to 128K program
8777 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
8779 Instruction set avr4 is for the enhanced AVR core with up to 8K program
8780 memory space (MCU types: atmega8, atmega83, atmega85).
8782 Instruction set avr5 is for the enhanced AVR core with up to 128K program
8783 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
8784 atmega64, atmega128, at43usb355, at94k).
8788 Output instruction sizes to the asm file.
8790 @item -minit-stack=@var{N}
8791 @opindex minit-stack
8792 Specify the initial stack address, which may be a symbol or numeric value,
8793 @samp{__stack} is the default.
8795 @item -mno-interrupts
8796 @opindex mno-interrupts
8797 Generated code is not compatible with hardware interrupts.
8798 Code size will be smaller.
8800 @item -mcall-prologues
8801 @opindex mcall-prologues
8802 Functions prologues/epilogues expanded as call to appropriate
8803 subroutines. Code size will be smaller.
8805 @item -mno-tablejump
8806 @opindex mno-tablejump
8807 Do not generate tablejump insns which sometimes increase code size.
8810 @opindex mtiny-stack
8811 Change only the low 8 bits of the stack pointer.
8815 Assume int to be 8 bit integer. This affects the sizes of all types: A
8816 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
8817 and long long will be 4 bytes. Please note that this option does not
8818 comply to the C standards, but it will provide you with smaller code
8822 @node Blackfin Options
8823 @subsection Blackfin Options
8824 @cindex Blackfin Options
8827 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
8829 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
8830 can be one of @samp{bf522}, @samp{bf523}, @samp{bf524},
8831 @samp{bf525}, @samp{bf526}, @samp{bf527},
8832 @samp{bf531}, @samp{bf532}, @samp{bf533}, @samp{bf534},
8833 @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
8834 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
8836 The optional @var{sirevision} specifies the silicon revision of the target
8837 Blackfin processor. Any workarounds available for the targeted silicon revision
8838 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
8839 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
8840 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
8841 hexadecimal digits representing the major and minor numbers in the silicon
8842 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
8843 is not defined. If @var{sirevision} is @samp{any}, the
8844 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
8845 If this optional @var{sirevision} is not used, GCC assumes the latest known
8846 silicon revision of the targeted Blackfin processor.
8848 Support for @samp{bf561} is incomplete. For @samp{bf561},
8849 Only the processor macro is defined.
8850 Without this option, @samp{bf532} is used as the processor by default.
8851 The corresponding predefined processor macros for @var{cpu} is to
8852 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
8853 provided by libgloss to be linked in if @option{-msim} is not given.
8857 Specifies that the program will be run on the simulator. This causes
8858 the simulator BSP provided by libgloss to be linked in. This option
8859 has effect only for @samp{bfin-elf} toolchain.
8860 Certain other options, such as @option{-mid-shared-library} and
8861 @option{-mfdpic}, imply @option{-msim}.
8863 @item -momit-leaf-frame-pointer
8864 @opindex momit-leaf-frame-pointer
8865 Don't keep the frame pointer in a register for leaf functions. This
8866 avoids the instructions to save, set up and restore frame pointers and
8867 makes an extra register available in leaf functions. The option
8868 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8869 which might make debugging harder.
8871 @item -mspecld-anomaly
8872 @opindex mspecld-anomaly
8873 When enabled, the compiler will ensure that the generated code does not
8874 contain speculative loads after jump instructions. If this option is used,
8875 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
8877 @item -mno-specld-anomaly
8878 @opindex mno-specld-anomaly
8879 Don't generate extra code to prevent speculative loads from occurring.
8881 @item -mcsync-anomaly
8882 @opindex mcsync-anomaly
8883 When enabled, the compiler will ensure that the generated code does not
8884 contain CSYNC or SSYNC instructions too soon after conditional branches.
8885 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
8887 @item -mno-csync-anomaly
8888 @opindex mno-csync-anomaly
8889 Don't generate extra code to prevent CSYNC or SSYNC instructions from
8890 occurring too soon after a conditional branch.
8894 When enabled, the compiler is free to take advantage of the knowledge that
8895 the entire program fits into the low 64k of memory.
8898 @opindex mno-low-64k
8899 Assume that the program is arbitrarily large. This is the default.
8901 @item -mstack-check-l1
8902 @opindex mstack-check-l1
8903 Do stack checking using information placed into L1 scratchpad memory by the
8906 @item -mid-shared-library
8907 @opindex mid-shared-library
8908 Generate code that supports shared libraries via the library ID method.
8909 This allows for execute in place and shared libraries in an environment
8910 without virtual memory management. This option implies @option{-fPIC}.
8911 With a @samp{bfin-elf} target, this option implies @option{-msim}.
8913 @item -mno-id-shared-library
8914 @opindex mno-id-shared-library
8915 Generate code that doesn't assume ID based shared libraries are being used.
8916 This is the default.
8918 @item -mleaf-id-shared-library
8919 @opindex mleaf-id-shared-library
8920 Generate code that supports shared libraries via the library ID method,
8921 but assumes that this library or executable won't link against any other
8922 ID shared libraries. That allows the compiler to use faster code for jumps
8925 @item -mno-leaf-id-shared-library
8926 @opindex mno-leaf-id-shared-library
8927 Do not assume that the code being compiled won't link against any ID shared
8928 libraries. Slower code will be generated for jump and call insns.
8930 @item -mshared-library-id=n
8931 @opindex mshared-library-id
8932 Specified the identification number of the ID based shared library being
8933 compiled. Specifying a value of 0 will generate more compact code, specifying
8934 other values will force the allocation of that number to the current
8935 library but is no more space or time efficient than omitting this option.
8939 Generate code that allows the data segment to be located in a different
8940 area of memory from the text segment. This allows for execute in place in
8941 an environment without virtual memory management by eliminating relocations
8942 against the text section.
8945 @opindex mno-sep-data
8946 Generate code that assumes that the data segment follows the text segment.
8947 This is the default.
8950 @itemx -mno-long-calls
8951 @opindex mlong-calls
8952 @opindex mno-long-calls
8953 Tells the compiler to perform function calls by first loading the
8954 address of the function into a register and then performing a subroutine
8955 call on this register. This switch is needed if the target function
8956 will lie outside of the 24 bit addressing range of the offset based
8957 version of subroutine call instruction.
8959 This feature is not enabled by default. Specifying
8960 @option{-mno-long-calls} will restore the default behavior. Note these
8961 switches have no effect on how the compiler generates code to handle
8962 function calls via function pointers.
8966 Link with the fast floating-point library. This library relaxes some of
8967 the IEEE floating-point standard's rules for checking inputs against
8968 Not-a-Number (NAN), in the interest of performance.
8971 @opindex minline-plt
8972 Enable inlining of PLT entries in function calls to functions that are
8973 not known to bind locally. It has no effect without @option{-mfdpic}.
8977 @subsection CRIS Options
8978 @cindex CRIS Options
8980 These options are defined specifically for the CRIS ports.
8983 @item -march=@var{architecture-type}
8984 @itemx -mcpu=@var{architecture-type}
8987 Generate code for the specified architecture. The choices for
8988 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
8989 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
8990 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
8993 @item -mtune=@var{architecture-type}
8995 Tune to @var{architecture-type} everything applicable about the generated
8996 code, except for the ABI and the set of available instructions. The
8997 choices for @var{architecture-type} are the same as for
8998 @option{-march=@var{architecture-type}}.
9000 @item -mmax-stack-frame=@var{n}
9001 @opindex mmax-stack-frame
9002 Warn when the stack frame of a function exceeds @var{n} bytes.
9004 @item -melinux-stacksize=@var{n}
9005 @opindex melinux-stacksize
9006 Only available with the @samp{cris-axis-aout} target. Arranges for
9007 indications in the program to the kernel loader that the stack of the
9008 program should be set to @var{n} bytes.
9014 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9015 @option{-march=v3} and @option{-march=v8} respectively.
9017 @item -mmul-bug-workaround
9018 @itemx -mno-mul-bug-workaround
9019 @opindex mmul-bug-workaround
9020 @opindex mno-mul-bug-workaround
9021 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
9022 models where it applies. This option is active by default.
9026 Enable CRIS-specific verbose debug-related information in the assembly
9027 code. This option also has the effect to turn off the @samp{#NO_APP}
9028 formatted-code indicator to the assembler at the beginning of the
9033 Do not use condition-code results from previous instruction; always emit
9034 compare and test instructions before use of condition codes.
9036 @item -mno-side-effects
9037 @opindex mno-side-effects
9038 Do not emit instructions with side-effects in addressing modes other than
9042 @itemx -mno-stack-align
9044 @itemx -mno-data-align
9045 @itemx -mconst-align
9046 @itemx -mno-const-align
9047 @opindex mstack-align
9048 @opindex mno-stack-align
9049 @opindex mdata-align
9050 @opindex mno-data-align
9051 @opindex mconst-align
9052 @opindex mno-const-align
9053 These options (no-options) arranges (eliminate arrangements) for the
9054 stack-frame, individual data and constants to be aligned for the maximum
9055 single data access size for the chosen CPU model. The default is to
9056 arrange for 32-bit alignment. ABI details such as structure layout are
9057 not affected by these options.
9065 Similar to the stack- data- and const-align options above, these options
9066 arrange for stack-frame, writable data and constants to all be 32-bit,
9067 16-bit or 8-bit aligned. The default is 32-bit alignment.
9069 @item -mno-prologue-epilogue
9070 @itemx -mprologue-epilogue
9071 @opindex mno-prologue-epilogue
9072 @opindex mprologue-epilogue
9073 With @option{-mno-prologue-epilogue}, the normal function prologue and
9074 epilogue that sets up the stack-frame are omitted and no return
9075 instructions or return sequences are generated in the code. Use this
9076 option only together with visual inspection of the compiled code: no
9077 warnings or errors are generated when call-saved registers must be saved,
9078 or storage for local variable needs to be allocated.
9084 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9085 instruction sequences that load addresses for functions from the PLT part
9086 of the GOT rather than (traditional on other architectures) calls to the
9087 PLT@. The default is @option{-mgotplt}.
9091 Legacy no-op option only recognized with the cris-axis-aout target.
9095 Legacy no-op option only recognized with the cris-axis-elf and
9096 cris-axis-linux-gnu targets.
9100 Only recognized with the cris-axis-aout target, where it selects a
9101 GNU/linux-like multilib, include files and instruction set for
9106 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9110 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9111 to link with input-output functions from a simulator library. Code,
9112 initialized data and zero-initialized data are allocated consecutively.
9116 Like @option{-sim}, but pass linker options to locate initialized data at
9117 0x40000000 and zero-initialized data at 0x80000000.
9121 @subsection CRX Options
9124 These options are defined specifically for the CRX ports.
9130 Enable the use of multiply-accumulate instructions. Disabled by default.
9134 Push instructions will be used to pass outgoing arguments when functions
9135 are called. Enabled by default.
9138 @node Darwin Options
9139 @subsection Darwin Options
9140 @cindex Darwin options
9142 These options are defined for all architectures running the Darwin operating
9145 FSF GCC on Darwin does not create ``fat'' object files; it will create
9146 an object file for the single architecture that it was built to
9147 target. Apple's GCC on Darwin does create ``fat'' files if multiple
9148 @option{-arch} options are used; it does so by running the compiler or
9149 linker multiple times and joining the results together with
9152 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
9153 @samp{i686}) is determined by the flags that specify the ISA
9154 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
9155 @option{-force_cpusubtype_ALL} option can be used to override this.
9157 The Darwin tools vary in their behavior when presented with an ISA
9158 mismatch. The assembler, @file{as}, will only permit instructions to
9159 be used that are valid for the subtype of the file it is generating,
9160 so you cannot put 64-bit instructions in an @samp{ppc750} object file.
9161 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
9162 and print an error if asked to create a shared library with a less
9163 restrictive subtype than its input files (for instance, trying to put
9164 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
9165 for executables, @file{ld}, will quietly give the executable the most
9166 restrictive subtype of any of its input files.
9171 Add the framework directory @var{dir} to the head of the list of
9172 directories to be searched for header files. These directories are
9173 interleaved with those specified by @option{-I} options and are
9174 scanned in a left-to-right order.
9176 A framework directory is a directory with frameworks in it. A
9177 framework is a directory with a @samp{"Headers"} and/or
9178 @samp{"PrivateHeaders"} directory contained directly in it that ends
9179 in @samp{".framework"}. The name of a framework is the name of this
9180 directory excluding the @samp{".framework"}. Headers associated with
9181 the framework are found in one of those two directories, with
9182 @samp{"Headers"} being searched first. A subframework is a framework
9183 directory that is in a framework's @samp{"Frameworks"} directory.
9184 Includes of subframework headers can only appear in a header of a
9185 framework that contains the subframework, or in a sibling subframework
9186 header. Two subframeworks are siblings if they occur in the same
9187 framework. A subframework should not have the same name as a
9188 framework, a warning will be issued if this is violated. Currently a
9189 subframework cannot have subframeworks, in the future, the mechanism
9190 may be extended to support this. The standard frameworks can be found
9191 in @samp{"/System/Library/Frameworks"} and
9192 @samp{"/Library/Frameworks"}. An example include looks like
9193 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
9194 the name of the framework and header.h is found in the
9195 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
9197 @item -iframework@var{dir}
9199 Like @option{-F} except the directory is a treated as a system
9200 directory. The main difference between this @option{-iframework} and
9201 @option{-F} is that with @option{-iframework} the compiler does not
9202 warn about constructs contained within header files found via
9203 @var{dir}. This option is valid only for the C family of languages.
9207 Emit debugging information for symbols that are used. For STABS
9208 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
9209 This is by default ON@.
9213 Emit debugging information for all symbols and types.
9215 @item -mmacosx-version-min=@var{version}
9216 The earliest version of MacOS X that this executable will run on
9217 is @var{version}. Typical values of @var{version} include @code{10.1},
9218 @code{10.2}, and @code{10.3.9}.
9220 If the compiler was built to use the system's headers by default,
9221 then the default for this option is the system version on which the
9222 compiler is running, otherwise the default is to make choices which
9223 are compatible with as many systems and code bases as possible.
9227 Enable kernel development mode. The @option{-mkernel} option sets
9228 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
9229 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
9230 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
9231 applicable. This mode also sets @option{-mno-altivec},
9232 @option{-msoft-float}, @option{-fno-builtin} and
9233 @option{-mlong-branch} for PowerPC targets.
9235 @item -mone-byte-bool
9236 @opindex mone-byte-bool
9237 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
9238 By default @samp{sizeof(bool)} is @samp{4} when compiling for
9239 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
9240 option has no effect on x86.
9242 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
9243 to generate code that is not binary compatible with code generated
9244 without that switch. Using this switch may require recompiling all
9245 other modules in a program, including system libraries. Use this
9246 switch to conform to a non-default data model.
9248 @item -mfix-and-continue
9249 @itemx -ffix-and-continue
9250 @itemx -findirect-data
9251 @opindex mfix-and-continue
9252 @opindex ffix-and-continue
9253 @opindex findirect-data
9254 Generate code suitable for fast turn around development. Needed to
9255 enable gdb to dynamically load @code{.o} files into already running
9256 programs. @option{-findirect-data} and @option{-ffix-and-continue}
9257 are provided for backwards compatibility.
9261 Loads all members of static archive libraries.
9262 See man ld(1) for more information.
9264 @item -arch_errors_fatal
9265 @opindex arch_errors_fatal
9266 Cause the errors having to do with files that have the wrong architecture
9270 @opindex bind_at_load
9271 Causes the output file to be marked such that the dynamic linker will
9272 bind all undefined references when the file is loaded or launched.
9276 Produce a Mach-o bundle format file.
9277 See man ld(1) for more information.
9279 @item -bundle_loader @var{executable}
9280 @opindex bundle_loader
9281 This option specifies the @var{executable} that will be loading the build
9282 output file being linked. See man ld(1) for more information.
9286 When passed this option, GCC will produce a dynamic library instead of
9287 an executable when linking, using the Darwin @file{libtool} command.
9289 @item -force_cpusubtype_ALL
9290 @opindex force_cpusubtype_ALL
9291 This causes GCC's output file to have the @var{ALL} subtype, instead of
9292 one controlled by the @option{-mcpu} or @option{-march} option.
9294 @item -allowable_client @var{client_name}
9296 @itemx -compatibility_version
9297 @itemx -current_version
9299 @itemx -dependency-file
9301 @itemx -dylinker_install_name
9303 @itemx -exported_symbols_list
9305 @itemx -flat_namespace
9306 @itemx -force_flat_namespace
9307 @itemx -headerpad_max_install_names
9310 @itemx -install_name
9311 @itemx -keep_private_externs
9312 @itemx -multi_module
9313 @itemx -multiply_defined
9314 @itemx -multiply_defined_unused
9316 @itemx -no_dead_strip_inits_and_terms
9317 @itemx -nofixprebinding
9320 @itemx -noseglinkedit
9321 @itemx -pagezero_size
9323 @itemx -prebind_all_twolevel_modules
9324 @itemx -private_bundle
9325 @itemx -read_only_relocs
9327 @itemx -sectobjectsymbols
9331 @itemx -sectobjectsymbols
9334 @itemx -segs_read_only_addr
9335 @itemx -segs_read_write_addr
9336 @itemx -seg_addr_table
9337 @itemx -seg_addr_table_filename
9340 @itemx -segs_read_only_addr
9341 @itemx -segs_read_write_addr
9342 @itemx -single_module
9345 @itemx -sub_umbrella
9346 @itemx -twolevel_namespace
9349 @itemx -unexported_symbols_list
9350 @itemx -weak_reference_mismatches
9353 @opindex allowable_client
9354 @opindex client_name
9355 @opindex compatibility_version
9356 @opindex current_version
9358 @opindex dependency-file
9360 @opindex dylinker_install_name
9362 @opindex exported_symbols_list
9364 @opindex flat_namespace
9365 @opindex force_flat_namespace
9366 @opindex headerpad_max_install_names
9369 @opindex install_name
9370 @opindex keep_private_externs
9371 @opindex multi_module
9372 @opindex multiply_defined
9373 @opindex multiply_defined_unused
9375 @opindex no_dead_strip_inits_and_terms
9376 @opindex nofixprebinding
9377 @opindex nomultidefs
9379 @opindex noseglinkedit
9380 @opindex pagezero_size
9382 @opindex prebind_all_twolevel_modules
9383 @opindex private_bundle
9384 @opindex read_only_relocs
9386 @opindex sectobjectsymbols
9390 @opindex sectobjectsymbols
9393 @opindex segs_read_only_addr
9394 @opindex segs_read_write_addr
9395 @opindex seg_addr_table
9396 @opindex seg_addr_table_filename
9397 @opindex seglinkedit
9399 @opindex segs_read_only_addr
9400 @opindex segs_read_write_addr
9401 @opindex single_module
9403 @opindex sub_library
9404 @opindex sub_umbrella
9405 @opindex twolevel_namespace
9408 @opindex unexported_symbols_list
9409 @opindex weak_reference_mismatches
9410 @opindex whatsloaded
9412 These options are passed to the Darwin linker. The Darwin linker man page
9413 describes them in detail.
9416 @node DEC Alpha Options
9417 @subsection DEC Alpha Options
9419 These @samp{-m} options are defined for the DEC Alpha implementations:
9422 @item -mno-soft-float
9424 @opindex mno-soft-float
9425 @opindex msoft-float
9426 Use (do not use) the hardware floating-point instructions for
9427 floating-point operations. When @option{-msoft-float} is specified,
9428 functions in @file{libgcc.a} will be used to perform floating-point
9429 operations. Unless they are replaced by routines that emulate the
9430 floating-point operations, or compiled in such a way as to call such
9431 emulations routines, these routines will issue floating-point
9432 operations. If you are compiling for an Alpha without floating-point
9433 operations, you must ensure that the library is built so as not to call
9436 Note that Alpha implementations without floating-point operations are
9437 required to have floating-point registers.
9442 @opindex mno-fp-regs
9443 Generate code that uses (does not use) the floating-point register set.
9444 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
9445 register set is not used, floating point operands are passed in integer
9446 registers as if they were integers and floating-point results are passed
9447 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
9448 so any function with a floating-point argument or return value called by code
9449 compiled with @option{-mno-fp-regs} must also be compiled with that
9452 A typical use of this option is building a kernel that does not use,
9453 and hence need not save and restore, any floating-point registers.
9457 The Alpha architecture implements floating-point hardware optimized for
9458 maximum performance. It is mostly compliant with the IEEE floating
9459 point standard. However, for full compliance, software assistance is
9460 required. This option generates code fully IEEE compliant code
9461 @emph{except} that the @var{inexact-flag} is not maintained (see below).
9462 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
9463 defined during compilation. The resulting code is less efficient but is
9464 able to correctly support denormalized numbers and exceptional IEEE
9465 values such as not-a-number and plus/minus infinity. Other Alpha
9466 compilers call this option @option{-ieee_with_no_inexact}.
9468 @item -mieee-with-inexact
9469 @opindex mieee-with-inexact
9470 This is like @option{-mieee} except the generated code also maintains
9471 the IEEE @var{inexact-flag}. Turning on this option causes the
9472 generated code to implement fully-compliant IEEE math. In addition to
9473 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
9474 macro. On some Alpha implementations the resulting code may execute
9475 significantly slower than the code generated by default. Since there is
9476 very little code that depends on the @var{inexact-flag}, you should
9477 normally not specify this option. Other Alpha compilers call this
9478 option @option{-ieee_with_inexact}.
9480 @item -mfp-trap-mode=@var{trap-mode}
9481 @opindex mfp-trap-mode
9482 This option controls what floating-point related traps are enabled.
9483 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9484 The trap mode can be set to one of four values:
9488 This is the default (normal) setting. The only traps that are enabled
9489 are the ones that cannot be disabled in software (e.g., division by zero
9493 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9497 Like @samp{u}, but the instructions are marked to be safe for software
9498 completion (see Alpha architecture manual for details).
9501 Like @samp{su}, but inexact traps are enabled as well.
9504 @item -mfp-rounding-mode=@var{rounding-mode}
9505 @opindex mfp-rounding-mode
9506 Selects the IEEE rounding mode. Other Alpha compilers call this option
9507 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9512 Normal IEEE rounding mode. Floating point numbers are rounded towards
9513 the nearest machine number or towards the even machine number in case
9517 Round towards minus infinity.
9520 Chopped rounding mode. Floating point numbers are rounded towards zero.
9523 Dynamic rounding mode. A field in the floating point control register
9524 (@var{fpcr}, see Alpha architecture reference manual) controls the
9525 rounding mode in effect. The C library initializes this register for
9526 rounding towards plus infinity. Thus, unless your program modifies the
9527 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9530 @item -mtrap-precision=@var{trap-precision}
9531 @opindex mtrap-precision
9532 In the Alpha architecture, floating point traps are imprecise. This
9533 means without software assistance it is impossible to recover from a
9534 floating trap and program execution normally needs to be terminated.
9535 GCC can generate code that can assist operating system trap handlers
9536 in determining the exact location that caused a floating point trap.
9537 Depending on the requirements of an application, different levels of
9538 precisions can be selected:
9542 Program precision. This option is the default and means a trap handler
9543 can only identify which program caused a floating point exception.
9546 Function precision. The trap handler can determine the function that
9547 caused a floating point exception.
9550 Instruction precision. The trap handler can determine the exact
9551 instruction that caused a floating point exception.
9554 Other Alpha compilers provide the equivalent options called
9555 @option{-scope_safe} and @option{-resumption_safe}.
9557 @item -mieee-conformant
9558 @opindex mieee-conformant
9559 This option marks the generated code as IEEE conformant. You must not
9560 use this option unless you also specify @option{-mtrap-precision=i} and either
9561 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9562 is to emit the line @samp{.eflag 48} in the function prologue of the
9563 generated assembly file. Under DEC Unix, this has the effect that
9564 IEEE-conformant math library routines will be linked in.
9566 @item -mbuild-constants
9567 @opindex mbuild-constants
9568 Normally GCC examines a 32- or 64-bit integer constant to
9569 see if it can construct it from smaller constants in two or three
9570 instructions. If it cannot, it will output the constant as a literal and
9571 generate code to load it from the data segment at runtime.
9573 Use this option to require GCC to construct @emph{all} integer constants
9574 using code, even if it takes more instructions (the maximum is six).
9576 You would typically use this option to build a shared library dynamic
9577 loader. Itself a shared library, it must relocate itself in memory
9578 before it can find the variables and constants in its own data segment.
9584 Select whether to generate code to be assembled by the vendor-supplied
9585 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9603 Indicate whether GCC should generate code to use the optional BWX,
9604 CIX, FIX and MAX instruction sets. The default is to use the instruction
9605 sets supported by the CPU type specified via @option{-mcpu=} option or that
9606 of the CPU on which GCC was built if none was specified.
9611 @opindex mfloat-ieee
9612 Generate code that uses (does not use) VAX F and G floating point
9613 arithmetic instead of IEEE single and double precision.
9615 @item -mexplicit-relocs
9616 @itemx -mno-explicit-relocs
9617 @opindex mexplicit-relocs
9618 @opindex mno-explicit-relocs
9619 Older Alpha assemblers provided no way to generate symbol relocations
9620 except via assembler macros. Use of these macros does not allow
9621 optimal instruction scheduling. GNU binutils as of version 2.12
9622 supports a new syntax that allows the compiler to explicitly mark
9623 which relocations should apply to which instructions. This option
9624 is mostly useful for debugging, as GCC detects the capabilities of
9625 the assembler when it is built and sets the default accordingly.
9629 @opindex msmall-data
9630 @opindex mlarge-data
9631 When @option{-mexplicit-relocs} is in effect, static data is
9632 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9633 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9634 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9635 16-bit relocations off of the @code{$gp} register. This limits the
9636 size of the small data area to 64KB, but allows the variables to be
9637 directly accessed via a single instruction.
9639 The default is @option{-mlarge-data}. With this option the data area
9640 is limited to just below 2GB@. Programs that require more than 2GB of
9641 data must use @code{malloc} or @code{mmap} to allocate the data in the
9642 heap instead of in the program's data segment.
9644 When generating code for shared libraries, @option{-fpic} implies
9645 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9649 @opindex msmall-text
9650 @opindex mlarge-text
9651 When @option{-msmall-text} is used, the compiler assumes that the
9652 code of the entire program (or shared library) fits in 4MB, and is
9653 thus reachable with a branch instruction. When @option{-msmall-data}
9654 is used, the compiler can assume that all local symbols share the
9655 same @code{$gp} value, and thus reduce the number of instructions
9656 required for a function call from 4 to 1.
9658 The default is @option{-mlarge-text}.
9660 @item -mcpu=@var{cpu_type}
9662 Set the instruction set and instruction scheduling parameters for
9663 machine type @var{cpu_type}. You can specify either the @samp{EV}
9664 style name or the corresponding chip number. GCC supports scheduling
9665 parameters for the EV4, EV5 and EV6 family of processors and will
9666 choose the default values for the instruction set from the processor
9667 you specify. If you do not specify a processor type, GCC will default
9668 to the processor on which the compiler was built.
9670 Supported values for @var{cpu_type} are
9676 Schedules as an EV4 and has no instruction set extensions.
9680 Schedules as an EV5 and has no instruction set extensions.
9684 Schedules as an EV5 and supports the BWX extension.
9689 Schedules as an EV5 and supports the BWX and MAX extensions.
9693 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9697 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9700 @item -mtune=@var{cpu_type}
9702 Set only the instruction scheduling parameters for machine type
9703 @var{cpu_type}. The instruction set is not changed.
9705 @item -mmemory-latency=@var{time}
9706 @opindex mmemory-latency
9707 Sets the latency the scheduler should assume for typical memory
9708 references as seen by the application. This number is highly
9709 dependent on the memory access patterns used by the application
9710 and the size of the external cache on the machine.
9712 Valid options for @var{time} are
9716 A decimal number representing clock cycles.
9722 The compiler contains estimates of the number of clock cycles for
9723 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9724 (also called Dcache, Scache, and Bcache), as well as to main memory.
9725 Note that L3 is only valid for EV5.
9730 @node DEC Alpha/VMS Options
9731 @subsection DEC Alpha/VMS Options
9733 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9736 @item -mvms-return-codes
9737 @opindex mvms-return-codes
9738 Return VMS condition codes from main. The default is to return POSIX
9739 style condition (e.g.@: error) codes.
9743 @subsection FRV Options
9750 Only use the first 32 general purpose registers.
9755 Use all 64 general purpose registers.
9760 Use only the first 32 floating point registers.
9765 Use all 64 floating point registers
9768 @opindex mhard-float
9770 Use hardware instructions for floating point operations.
9773 @opindex msoft-float
9775 Use library routines for floating point operations.
9780 Dynamically allocate condition code registers.
9785 Do not try to dynamically allocate condition code registers, only
9786 use @code{icc0} and @code{fcc0}.
9791 Change ABI to use double word insns.
9796 Do not use double word instructions.
9801 Use floating point double instructions.
9806 Do not use floating point double instructions.
9811 Use media instructions.
9816 Do not use media instructions.
9821 Use multiply and add/subtract instructions.
9826 Do not use multiply and add/subtract instructions.
9831 Select the FDPIC ABI, that uses function descriptors to represent
9832 pointers to functions. Without any PIC/PIE-related options, it
9833 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
9834 assumes GOT entries and small data are within a 12-bit range from the
9835 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
9836 are computed with 32 bits.
9837 With a @samp{bfin-elf} target, this option implies @option{-msim}.
9840 @opindex minline-plt
9842 Enable inlining of PLT entries in function calls to functions that are
9843 not known to bind locally. It has no effect without @option{-mfdpic}.
9844 It's enabled by default if optimizing for speed and compiling for
9845 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
9846 optimization option such as @option{-O3} or above is present in the
9852 Assume a large TLS segment when generating thread-local code.
9857 Do not assume a large TLS segment when generating thread-local code.
9862 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
9863 that is known to be in read-only sections. It's enabled by default,
9864 except for @option{-fpic} or @option{-fpie}: even though it may help
9865 make the global offset table smaller, it trades 1 instruction for 4.
9866 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
9867 one of which may be shared by multiple symbols, and it avoids the need
9868 for a GOT entry for the referenced symbol, so it's more likely to be a
9869 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
9871 @item -multilib-library-pic
9872 @opindex multilib-library-pic
9874 Link with the (library, not FD) pic libraries. It's implied by
9875 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
9876 @option{-fpic} without @option{-mfdpic}. You should never have to use
9882 Follow the EABI requirement of always creating a frame pointer whenever
9883 a stack frame is allocated. This option is enabled by default and can
9884 be disabled with @option{-mno-linked-fp}.
9887 @opindex mlong-calls
9889 Use indirect addressing to call functions outside the current
9890 compilation unit. This allows the functions to be placed anywhere
9891 within the 32-bit address space.
9893 @item -malign-labels
9894 @opindex malign-labels
9896 Try to align labels to an 8-byte boundary by inserting nops into the
9897 previous packet. This option only has an effect when VLIW packing
9898 is enabled. It doesn't create new packets; it merely adds nops to
9902 @opindex mlibrary-pic
9904 Generate position-independent EABI code.
9909 Use only the first four media accumulator registers.
9914 Use all eight media accumulator registers.
9919 Pack VLIW instructions.
9924 Do not pack VLIW instructions.
9929 Do not mark ABI switches in e_flags.
9934 Enable the use of conditional-move instructions (default).
9936 This switch is mainly for debugging the compiler and will likely be removed
9937 in a future version.
9939 @item -mno-cond-move
9940 @opindex mno-cond-move
9942 Disable the use of conditional-move instructions.
9944 This switch is mainly for debugging the compiler and will likely be removed
9945 in a future version.
9950 Enable the use of conditional set instructions (default).
9952 This switch is mainly for debugging the compiler and will likely be removed
9953 in a future version.
9958 Disable the use of conditional set instructions.
9960 This switch is mainly for debugging the compiler and will likely be removed
9961 in a future version.
9966 Enable the use of conditional execution (default).
9968 This switch is mainly for debugging the compiler and will likely be removed
9969 in a future version.
9971 @item -mno-cond-exec
9972 @opindex mno-cond-exec
9974 Disable the use of conditional execution.
9976 This switch is mainly for debugging the compiler and will likely be removed
9977 in a future version.
9980 @opindex mvliw-branch
9982 Run a pass to pack branches into VLIW instructions (default).
9984 This switch is mainly for debugging the compiler and will likely be removed
9985 in a future version.
9987 @item -mno-vliw-branch
9988 @opindex mno-vliw-branch
9990 Do not run a pass to pack branches into VLIW instructions.
9992 This switch is mainly for debugging the compiler and will likely be removed
9993 in a future version.
9995 @item -mmulti-cond-exec
9996 @opindex mmulti-cond-exec
9998 Enable optimization of @code{&&} and @code{||} in conditional execution
10001 This switch is mainly for debugging the compiler and will likely be removed
10002 in a future version.
10004 @item -mno-multi-cond-exec
10005 @opindex mno-multi-cond-exec
10007 Disable optimization of @code{&&} and @code{||} in conditional execution.
10009 This switch is mainly for debugging the compiler and will likely be removed
10010 in a future version.
10012 @item -mnested-cond-exec
10013 @opindex mnested-cond-exec
10015 Enable nested conditional execution optimizations (default).
10017 This switch is mainly for debugging the compiler and will likely be removed
10018 in a future version.
10020 @item -mno-nested-cond-exec
10021 @opindex mno-nested-cond-exec
10023 Disable nested conditional execution optimizations.
10025 This switch is mainly for debugging the compiler and will likely be removed
10026 in a future version.
10028 @item -moptimize-membar
10029 @opindex moptimize-membar
10031 This switch removes redundant @code{membar} instructions from the
10032 compiler generated code. It is enabled by default.
10034 @item -mno-optimize-membar
10035 @opindex mno-optimize-membar
10037 This switch disables the automatic removal of redundant @code{membar}
10038 instructions from the generated code.
10040 @item -mtomcat-stats
10041 @opindex mtomcat-stats
10043 Cause gas to print out tomcat statistics.
10045 @item -mcpu=@var{cpu}
10048 Select the processor type for which to generate code. Possible values are
10049 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
10050 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
10054 @node GNU/Linux Options
10055 @subsection GNU/Linux Options
10057 These @samp{-m} options are defined for GNU/Linux targets:
10062 Use the GNU C library instead of uClibc. This is the default except
10063 on @samp{*-*-linux-*uclibc*} targets.
10067 Use uClibc instead of the GNU C library. This is the default on
10068 @samp{*-*-linux-*uclibc*} targets.
10071 @node H8/300 Options
10072 @subsection H8/300 Options
10074 These @samp{-m} options are defined for the H8/300 implementations:
10079 Shorten some address references at link time, when possible; uses the
10080 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
10081 ld, Using ld}, for a fuller description.
10085 Generate code for the H8/300H@.
10089 Generate code for the H8S@.
10093 Generate code for the H8S and H8/300H in the normal mode. This switch
10094 must be used either with @option{-mh} or @option{-ms}.
10098 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
10102 Make @code{int} data 32 bits by default.
10105 @opindex malign-300
10106 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
10107 The default for the H8/300H and H8S is to align longs and floats on 4
10109 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
10110 This option has no effect on the H8/300.
10114 @subsection HPPA Options
10115 @cindex HPPA Options
10117 These @samp{-m} options are defined for the HPPA family of computers:
10120 @item -march=@var{architecture-type}
10122 Generate code for the specified architecture. The choices for
10123 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
10124 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
10125 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
10126 architecture option for your machine. Code compiled for lower numbered
10127 architectures will run on higher numbered architectures, but not the
10130 @item -mpa-risc-1-0
10131 @itemx -mpa-risc-1-1
10132 @itemx -mpa-risc-2-0
10133 @opindex mpa-risc-1-0
10134 @opindex mpa-risc-1-1
10135 @opindex mpa-risc-2-0
10136 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
10139 @opindex mbig-switch
10140 Generate code suitable for big switch tables. Use this option only if
10141 the assembler/linker complain about out of range branches within a switch
10144 @item -mjump-in-delay
10145 @opindex mjump-in-delay
10146 Fill delay slots of function calls with unconditional jump instructions
10147 by modifying the return pointer for the function call to be the target
10148 of the conditional jump.
10150 @item -mdisable-fpregs
10151 @opindex mdisable-fpregs
10152 Prevent floating point registers from being used in any manner. This is
10153 necessary for compiling kernels which perform lazy context switching of
10154 floating point registers. If you use this option and attempt to perform
10155 floating point operations, the compiler will abort.
10157 @item -mdisable-indexing
10158 @opindex mdisable-indexing
10159 Prevent the compiler from using indexing address modes. This avoids some
10160 rather obscure problems when compiling MIG generated code under MACH@.
10162 @item -mno-space-regs
10163 @opindex mno-space-regs
10164 Generate code that assumes the target has no space registers. This allows
10165 GCC to generate faster indirect calls and use unscaled index address modes.
10167 Such code is suitable for level 0 PA systems and kernels.
10169 @item -mfast-indirect-calls
10170 @opindex mfast-indirect-calls
10171 Generate code that assumes calls never cross space boundaries. This
10172 allows GCC to emit code which performs faster indirect calls.
10174 This option will not work in the presence of shared libraries or nested
10177 @item -mfixed-range=@var{register-range}
10178 @opindex mfixed-range
10179 Generate code treating the given register range as fixed registers.
10180 A fixed register is one that the register allocator can not use. This is
10181 useful when compiling kernel code. A register range is specified as
10182 two registers separated by a dash. Multiple register ranges can be
10183 specified separated by a comma.
10185 @item -mlong-load-store
10186 @opindex mlong-load-store
10187 Generate 3-instruction load and store sequences as sometimes required by
10188 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
10191 @item -mportable-runtime
10192 @opindex mportable-runtime
10193 Use the portable calling conventions proposed by HP for ELF systems.
10197 Enable the use of assembler directives only GAS understands.
10199 @item -mschedule=@var{cpu-type}
10201 Schedule code according to the constraints for the machine type
10202 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
10203 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
10204 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
10205 proper scheduling option for your machine. The default scheduling is
10209 @opindex mlinker-opt
10210 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
10211 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
10212 linkers in which they give bogus error messages when linking some programs.
10215 @opindex msoft-float
10216 Generate output containing library calls for floating point.
10217 @strong{Warning:} the requisite libraries are not available for all HPPA
10218 targets. Normally the facilities of the machine's usual C compiler are
10219 used, but this cannot be done directly in cross-compilation. You must make
10220 your own arrangements to provide suitable library functions for
10221 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
10222 does provide software floating point support.
10224 @option{-msoft-float} changes the calling convention in the output file;
10225 therefore, it is only useful if you compile @emph{all} of a program with
10226 this option. In particular, you need to compile @file{libgcc.a}, the
10227 library that comes with GCC, with @option{-msoft-float} in order for
10232 Generate the predefine, @code{_SIO}, for server IO@. The default is
10233 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
10234 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
10235 options are available under HP-UX and HI-UX@.
10239 Use GNU ld specific options. This passes @option{-shared} to ld when
10240 building a shared library. It is the default when GCC is configured,
10241 explicitly or implicitly, with the GNU linker. This option does not
10242 have any affect on which ld is called, it only changes what parameters
10243 are passed to that ld. The ld that is called is determined by the
10244 @option{--with-ld} configure option, GCC's program search path, and
10245 finally by the user's @env{PATH}. The linker used by GCC can be printed
10246 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
10247 on the 64 bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
10251 Use HP ld specific options. This passes @option{-b} to ld when building
10252 a shared library and passes @option{+Accept TypeMismatch} to ld on all
10253 links. It is the default when GCC is configured, explicitly or
10254 implicitly, with the HP linker. This option does not have any affect on
10255 which ld is called, it only changes what parameters are passed to that
10256 ld. The ld that is called is determined by the @option{--with-ld}
10257 configure option, GCC's program search path, and finally by the user's
10258 @env{PATH}. The linker used by GCC can be printed using @samp{which
10259 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
10260 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
10263 @opindex mno-long-calls
10264 Generate code that uses long call sequences. This ensures that a call
10265 is always able to reach linker generated stubs. The default is to generate
10266 long calls only when the distance from the call site to the beginning
10267 of the function or translation unit, as the case may be, exceeds a
10268 predefined limit set by the branch type being used. The limits for
10269 normal calls are 7,600,000 and 240,000 bytes, respectively for the
10270 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
10273 Distances are measured from the beginning of functions when using the
10274 @option{-ffunction-sections} option, or when using the @option{-mgas}
10275 and @option{-mno-portable-runtime} options together under HP-UX with
10278 It is normally not desirable to use this option as it will degrade
10279 performance. However, it may be useful in large applications,
10280 particularly when partial linking is used to build the application.
10282 The types of long calls used depends on the capabilities of the
10283 assembler and linker, and the type of code being generated. The
10284 impact on systems that support long absolute calls, and long pic
10285 symbol-difference or pc-relative calls should be relatively small.
10286 However, an indirect call is used on 32-bit ELF systems in pic code
10287 and it is quite long.
10289 @item -munix=@var{unix-std}
10291 Generate compiler predefines and select a startfile for the specified
10292 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
10293 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
10294 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
10295 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
10296 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
10299 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
10300 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
10301 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
10302 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
10303 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
10304 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
10306 It is @emph{important} to note that this option changes the interfaces
10307 for various library routines. It also affects the operational behavior
10308 of the C library. Thus, @emph{extreme} care is needed in using this
10311 Library code that is intended to operate with more than one UNIX
10312 standard must test, set and restore the variable @var{__xpg4_extended_mask}
10313 as appropriate. Most GNU software doesn't provide this capability.
10317 Suppress the generation of link options to search libdld.sl when the
10318 @option{-static} option is specified on HP-UX 10 and later.
10322 The HP-UX implementation of setlocale in libc has a dependency on
10323 libdld.sl. There isn't an archive version of libdld.sl. Thus,
10324 when the @option{-static} option is specified, special link options
10325 are needed to resolve this dependency.
10327 On HP-UX 10 and later, the GCC driver adds the necessary options to
10328 link with libdld.sl when the @option{-static} option is specified.
10329 This causes the resulting binary to be dynamic. On the 64-bit port,
10330 the linkers generate dynamic binaries by default in any case. The
10331 @option{-nolibdld} option can be used to prevent the GCC driver from
10332 adding these link options.
10336 Add support for multithreading with the @dfn{dce thread} library
10337 under HP-UX@. This option sets flags for both the preprocessor and
10341 @node i386 and x86-64 Options
10342 @subsection Intel 386 and AMD x86-64 Options
10343 @cindex i386 Options
10344 @cindex x86-64 Options
10345 @cindex Intel 386 Options
10346 @cindex AMD x86-64 Options
10348 These @samp{-m} options are defined for the i386 and x86-64 family of
10352 @item -mtune=@var{cpu-type}
10354 Tune to @var{cpu-type} everything applicable about the generated code, except
10355 for the ABI and the set of available instructions. The choices for
10356 @var{cpu-type} are:
10359 Produce code optimized for the most common IA32/AMD64/EM64T processors.
10360 If you know the CPU on which your code will run, then you should use
10361 the corresponding @option{-mtune} option instead of
10362 @option{-mtune=generic}. But, if you do not know exactly what CPU users
10363 of your application will have, then you should use this option.
10365 As new processors are deployed in the marketplace, the behavior of this
10366 option will change. Therefore, if you upgrade to a newer version of
10367 GCC, the code generated option will change to reflect the processors
10368 that were most common when that version of GCC was released.
10370 There is no @option{-march=generic} option because @option{-march}
10371 indicates the instruction set the compiler can use, and there is no
10372 generic instruction set applicable to all processors. In contrast,
10373 @option{-mtune} indicates the processor (or, in this case, collection of
10374 processors) for which the code is optimized.
10376 This selects the CPU to tune for at compilation time by determining
10377 the processor type of the compiling machine. Using @option{-mtune=native}
10378 will produce code optimized for the local machine under the constraints
10379 of the selected instruction set. Using @option{-march=native} will
10380 enable all instruction subsets supported by the local machine (hence
10381 the result might not run on different machines).
10383 Original Intel's i386 CPU@.
10385 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
10386 @item i586, pentium
10387 Intel Pentium CPU with no MMX support.
10389 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
10391 Intel PentiumPro CPU@.
10393 Same as @code{generic}, but when used as @code{march} option, PentiumPro
10394 instruction set will be used, so the code will run on all i686 family chips.
10396 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
10397 @item pentium3, pentium3m
10398 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
10401 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
10402 support. Used by Centrino notebooks.
10403 @item pentium4, pentium4m
10404 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
10406 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
10409 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
10410 SSE2 and SSE3 instruction set support.
10412 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
10413 instruction set support.
10415 AMD K6 CPU with MMX instruction set support.
10417 Improved versions of AMD K6 CPU with MMX and 3dNOW!@: instruction set support.
10418 @item athlon, athlon-tbird
10419 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and SSE prefetch instructions
10421 @item athlon-4, athlon-xp, athlon-mp
10422 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and full SSE
10423 instruction set support.
10424 @item k8, opteron, athlon64, athlon-fx
10425 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
10426 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW!@: and 64-bit instruction set extensions.)
10427 @item k8-sse3, opteron-sse3, athlon64-sse3
10428 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
10429 @item amdfam10, barcelona
10430 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
10431 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3dNOW!, enhanced 3dNOW!, ABM and 64-bit
10432 instruction set extensions.)
10434 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
10437 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!@:
10438 instruction set support.
10440 Via C3 CPU with MMX and 3dNOW!@: instruction set support. (No scheduling is
10441 implemented for this chip.)
10443 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
10444 implemented for this chip.)
10446 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
10449 While picking a specific @var{cpu-type} will schedule things appropriately
10450 for that particular chip, the compiler will not generate any code that
10451 does not run on the i386 without the @option{-march=@var{cpu-type}} option
10454 @item -march=@var{cpu-type}
10456 Generate instructions for the machine type @var{cpu-type}. The choices
10457 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
10458 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
10460 @item -mcpu=@var{cpu-type}
10462 A deprecated synonym for @option{-mtune}.
10464 @item -mfpmath=@var{unit}
10466 Generate floating point arithmetics for selected unit @var{unit}. The choices
10467 for @var{unit} are:
10471 Use the standard 387 floating point coprocessor present majority of chips and
10472 emulated otherwise. Code compiled with this option will run almost everywhere.
10473 The temporary results are computed in 80bit precision instead of precision
10474 specified by the type resulting in slightly different results compared to most
10475 of other chips. See @option{-ffloat-store} for more detailed description.
10477 This is the default choice for i386 compiler.
10480 Use scalar floating point instructions present in the SSE instruction set.
10481 This instruction set is supported by Pentium3 and newer chips, in the AMD line
10482 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
10483 instruction set supports only single precision arithmetics, thus the double and
10484 extended precision arithmetics is still done using 387. Later version, present
10485 only in Pentium4 and the future AMD x86-64 chips supports double precision
10488 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
10489 or @option{-msse2} switches to enable SSE extensions and make this option
10490 effective. For the x86-64 compiler, these extensions are enabled by default.
10492 The resulting code should be considerably faster in the majority of cases and avoid
10493 the numerical instability problems of 387 code, but may break some existing
10494 code that expects temporaries to be 80bit.
10496 This is the default choice for the x86-64 compiler.
10499 Attempt to utilize both instruction sets at once. This effectively double the
10500 amount of available registers and on chips with separate execution units for
10501 387 and SSE the execution resources too. Use this option with care, as it is
10502 still experimental, because the GCC register allocator does not model separate
10503 functional units well resulting in instable performance.
10506 @item -masm=@var{dialect}
10507 @opindex masm=@var{dialect}
10508 Output asm instructions using selected @var{dialect}. Supported
10509 choices are @samp{intel} or @samp{att} (the default one). Darwin does
10510 not support @samp{intel}.
10513 @itemx -mno-ieee-fp
10515 @opindex mno-ieee-fp
10516 Control whether or not the compiler uses IEEE floating point
10517 comparisons. These handle correctly the case where the result of a
10518 comparison is unordered.
10521 @opindex msoft-float
10522 Generate output containing library calls for floating point.
10523 @strong{Warning:} the requisite libraries are not part of GCC@.
10524 Normally the facilities of the machine's usual C compiler are used, but
10525 this can't be done directly in cross-compilation. You must make your
10526 own arrangements to provide suitable library functions for
10529 On machines where a function returns floating point results in the 80387
10530 register stack, some floating point opcodes may be emitted even if
10531 @option{-msoft-float} is used.
10533 @item -mno-fp-ret-in-387
10534 @opindex mno-fp-ret-in-387
10535 Do not use the FPU registers for return values of functions.
10537 The usual calling convention has functions return values of types
10538 @code{float} and @code{double} in an FPU register, even if there
10539 is no FPU@. The idea is that the operating system should emulate
10542 The option @option{-mno-fp-ret-in-387} causes such values to be returned
10543 in ordinary CPU registers instead.
10545 @item -mno-fancy-math-387
10546 @opindex mno-fancy-math-387
10547 Some 387 emulators do not support the @code{sin}, @code{cos} and
10548 @code{sqrt} instructions for the 387. Specify this option to avoid
10549 generating those instructions. This option is the default on FreeBSD,
10550 OpenBSD and NetBSD@. This option is overridden when @option{-march}
10551 indicates that the target cpu will always have an FPU and so the
10552 instruction will not need emulation. As of revision 2.6.1, these
10553 instructions are not generated unless you also use the
10554 @option{-funsafe-math-optimizations} switch.
10556 @item -malign-double
10557 @itemx -mno-align-double
10558 @opindex malign-double
10559 @opindex mno-align-double
10560 Control whether GCC aligns @code{double}, @code{long double}, and
10561 @code{long long} variables on a two word boundary or a one word
10562 boundary. Aligning @code{double} variables on a two word boundary will
10563 produce code that runs somewhat faster on a @samp{Pentium} at the
10564 expense of more memory.
10566 On x86-64, @option{-malign-double} is enabled by default.
10568 @strong{Warning:} if you use the @option{-malign-double} switch,
10569 structures containing the above types will be aligned differently than
10570 the published application binary interface specifications for the 386
10571 and will not be binary compatible with structures in code compiled
10572 without that switch.
10574 @item -m96bit-long-double
10575 @itemx -m128bit-long-double
10576 @opindex m96bit-long-double
10577 @opindex m128bit-long-double
10578 These switches control the size of @code{long double} type. The i386
10579 application binary interface specifies the size to be 96 bits,
10580 so @option{-m96bit-long-double} is the default in 32 bit mode.
10582 Modern architectures (Pentium and newer) would prefer @code{long double}
10583 to be aligned to an 8 or 16 byte boundary. In arrays or structures
10584 conforming to the ABI, this would not be possible. So specifying a
10585 @option{-m128bit-long-double} will align @code{long double}
10586 to a 16 byte boundary by padding the @code{long double} with an additional
10589 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
10590 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
10592 Notice that neither of these options enable any extra precision over the x87
10593 standard of 80 bits for a @code{long double}.
10595 @strong{Warning:} if you override the default value for your target ABI, the
10596 structures and arrays containing @code{long double} variables will change
10597 their size as well as function calling convention for function taking
10598 @code{long double} will be modified. Hence they will not be binary
10599 compatible with arrays or structures in code compiled without that switch.
10601 @item -mmlarge-data-threshold=@var{number}
10602 @opindex mlarge-data-threshold=@var{number}
10603 When @option{-mcmodel=medium} is specified, the data greater than
10604 @var{threshold} are placed in large data section. This value must be the
10605 same across all object linked into the binary and defaults to 65535.
10609 Use a different function-calling convention, in which functions that
10610 take a fixed number of arguments return with the @code{ret} @var{num}
10611 instruction, which pops their arguments while returning. This saves one
10612 instruction in the caller since there is no need to pop the arguments
10615 You can specify that an individual function is called with this calling
10616 sequence with the function attribute @samp{stdcall}. You can also
10617 override the @option{-mrtd} option by using the function attribute
10618 @samp{cdecl}. @xref{Function Attributes}.
10620 @strong{Warning:} this calling convention is incompatible with the one
10621 normally used on Unix, so you cannot use it if you need to call
10622 libraries compiled with the Unix compiler.
10624 Also, you must provide function prototypes for all functions that
10625 take variable numbers of arguments (including @code{printf});
10626 otherwise incorrect code will be generated for calls to those
10629 In addition, seriously incorrect code will result if you call a
10630 function with too many arguments. (Normally, extra arguments are
10631 harmlessly ignored.)
10633 @item -mregparm=@var{num}
10635 Control how many registers are used to pass integer arguments. By
10636 default, no registers are used to pass arguments, and at most 3
10637 registers can be used. You can control this behavior for a specific
10638 function by using the function attribute @samp{regparm}.
10639 @xref{Function Attributes}.
10641 @strong{Warning:} if you use this switch, and
10642 @var{num} is nonzero, then you must build all modules with the same
10643 value, including any libraries. This includes the system libraries and
10647 @opindex msseregparm
10648 Use SSE register passing conventions for float and double arguments
10649 and return values. You can control this behavior for a specific
10650 function by using the function attribute @samp{sseregparm}.
10651 @xref{Function Attributes}.
10653 @strong{Warning:} if you use this switch then you must build all
10654 modules with the same value, including any libraries. This includes
10655 the system libraries and startup modules.
10664 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
10665 is specified, the significands of results of floating-point operations are
10666 rounded to 24 bits (single precision); @option{-mpc64} rounds the the
10667 significands of results of floating-point operations to 53 bits (double
10668 precision) and @option{-mpc80} rounds the significands of results of
10669 floating-point operations to 64 bits (extended double precision), which is
10670 the default. When this option is used, floating-point operations in higher
10671 precisions are not available to the programmer without setting the FPU
10672 control word explicitly.
10674 Setting the rounding of floating-point operations to less than the default
10675 80 bits can speed some programs by 2% or more. Note that some mathematical
10676 libraries assume that extended precision (80 bit) floating-point operations
10677 are enabled by default; routines in such libraries could suffer significant
10678 loss of accuracy, typically through so-called "catastrophic cancellation",
10679 when this option is used to set the precision to less than extended precision.
10681 @item -mstackrealign
10682 @opindex mstackrealign
10683 Realign the stack at entry. On the Intel x86, the
10684 @option{-mstackrealign} option will generate an alternate prologue and
10685 epilogue that realigns the runtime stack. This supports mixing legacy
10686 codes that keep a 4-byte aligned stack with modern codes that keep a
10687 16-byte stack for SSE compatibility. The alternate prologue and
10688 epilogue are slower and bigger than the regular ones, and the
10689 alternate prologue requires an extra scratch register; this lowers the
10690 number of registers available if used in conjunction with the
10691 @code{regparm} attribute. The @option{-mstackrealign} option is
10692 incompatible with the nested function prologue; this is considered a
10693 hard error. See also the attribute @code{force_align_arg_pointer},
10694 applicable to individual functions.
10696 @item -mpreferred-stack-boundary=@var{num}
10697 @opindex mpreferred-stack-boundary
10698 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
10699 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
10700 the default is 4 (16 bytes or 128 bits).
10702 On Pentium and PentiumPro, @code{double} and @code{long double} values
10703 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
10704 suffer significant run time performance penalties. On Pentium III, the
10705 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
10706 properly if it is not 16 byte aligned.
10708 To ensure proper alignment of this values on the stack, the stack boundary
10709 must be as aligned as that required by any value stored on the stack.
10710 Further, every function must be generated such that it keeps the stack
10711 aligned. Thus calling a function compiled with a higher preferred
10712 stack boundary from a function compiled with a lower preferred stack
10713 boundary will most likely misalign the stack. It is recommended that
10714 libraries that use callbacks always use the default setting.
10716 This extra alignment does consume extra stack space, and generally
10717 increases code size. Code that is sensitive to stack space usage, such
10718 as embedded systems and operating system kernels, may want to reduce the
10719 preferred alignment to @option{-mpreferred-stack-boundary=2}.
10753 These switches enable or disable the use of instructions in the MMX,
10754 SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4A, SSE5, ABM or 3DNow!@: extended
10756 These extensions are also available as built-in functions: see
10757 @ref{X86 Built-in Functions}, for details of the functions enabled and
10758 disabled by these switches.
10760 To have SSE/SSE2 instructions generated automatically from floating-point
10761 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
10763 These options will enable GCC to use these extended instructions in
10764 generated code, even without @option{-mfpmath=sse}. Applications which
10765 perform runtime CPU detection must compile separate files for each
10766 supported architecture, using the appropriate flags. In particular,
10767 the file containing the CPU detection code should be compiled without
10772 This option will enable GCC to use CMPXCHG16B instruction in generated code.
10773 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
10774 data types. This is useful for high resolution counters that could be updated
10775 by multiple processors (or cores). This instruction is generated as part of
10776 atomic built-in functions: see @ref{Atomic Builtins} for details.
10780 This option will enable GCC to use SAHF instruction in generated 64-bit code.
10781 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
10782 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
10783 SAHF are load and store instructions, respectively, for certain status flags.
10784 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
10785 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
10789 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
10790 vectorized variants RCPPS and RSQRTPS) with an additional Newton-Rhapson step
10791 to increase precision instead of DIVSS and SQRTSS (and their vectorized
10792 variants) for single precision floating point arguments. These instructions
10793 are generated only when @option{-funsafe-math-optimizations} is enabled
10794 together with @option{-finite-math-only} and @option{-fno-trapping-math}.
10795 Note that while the throughput of the sequence is higher than the throughput
10796 of the non-reciprocal instruction, the precision of the sequence can be
10797 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
10799 @item -mveclibabi=@var{type}
10800 @opindex mveclibabi
10801 Specifies the ABI type to use for vectorizing intrinsics using an
10802 external library. Supported types are @code{acml} for the AMD
10803 math core library style of interfacing. GCC will currently emit
10804 calls to @code{__vrd2_sin}, @code{__vrd2_cos}, @code{__vrd2_exp},
10805 @code{__vrd2_log}, @code{__vrd2_log2}, @code{__vrd2_log10},
10806 @code{__vrs4_sinf}, @code{__vrs4_cosf}, @code{__vrs4_expf},
10807 @code{__vrs4_logf}, @code{__vrs4_log2f}, @code{__vrs4_log10f}
10808 and @code{__vrs4_powf} when using this type and @option{-ftree-vectorize}
10809 is enabled. A ACML ABI compatible library will have to be specified
10813 @itemx -mno-push-args
10814 @opindex mpush-args
10815 @opindex mno-push-args
10816 Use PUSH operations to store outgoing parameters. This method is shorter
10817 and usually equally fast as method using SUB/MOV operations and is enabled
10818 by default. In some cases disabling it may improve performance because of
10819 improved scheduling and reduced dependencies.
10821 @item -maccumulate-outgoing-args
10822 @opindex maccumulate-outgoing-args
10823 If enabled, the maximum amount of space required for outgoing arguments will be
10824 computed in the function prologue. This is faster on most modern CPUs
10825 because of reduced dependencies, improved scheduling and reduced stack usage
10826 when preferred stack boundary is not equal to 2. The drawback is a notable
10827 increase in code size. This switch implies @option{-mno-push-args}.
10831 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
10832 on thread-safe exception handling must compile and link all code with the
10833 @option{-mthreads} option. When compiling, @option{-mthreads} defines
10834 @option{-D_MT}; when linking, it links in a special thread helper library
10835 @option{-lmingwthrd} which cleans up per thread exception handling data.
10837 @item -mno-align-stringops
10838 @opindex mno-align-stringops
10839 Do not align destination of inlined string operations. This switch reduces
10840 code size and improves performance in case the destination is already aligned,
10841 but GCC doesn't know about it.
10843 @item -minline-all-stringops
10844 @opindex minline-all-stringops
10845 By default GCC inlines string operations only when destination is known to be
10846 aligned at least to 4 byte boundary. This enables more inlining, increase code
10847 size, but may improve performance of code that depends on fast memcpy, strlen
10848 and memset for short lengths.
10850 @item -minline-stringops-dynamically
10851 @opindex minline-stringops-dynamically
10852 For string operation of unknown size, inline runtime checks so for small
10853 blocks inline code is used, while for large blocks library call is used.
10855 @item -mstringop-strategy=@var{alg}
10856 @opindex mstringop-strategy=@var{alg}
10857 Overwrite internal decision heuristic about particular algorithm to inline
10858 string operation with. The allowed values are @code{rep_byte},
10859 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
10860 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
10861 expanding inline loop, @code{libcall} for always expanding library call.
10863 @item -momit-leaf-frame-pointer
10864 @opindex momit-leaf-frame-pointer
10865 Don't keep the frame pointer in a register for leaf functions. This
10866 avoids the instructions to save, set up and restore frame pointers and
10867 makes an extra register available in leaf functions. The option
10868 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10869 which might make debugging harder.
10871 @item -mtls-direct-seg-refs
10872 @itemx -mno-tls-direct-seg-refs
10873 @opindex mtls-direct-seg-refs
10874 Controls whether TLS variables may be accessed with offsets from the
10875 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
10876 or whether the thread base pointer must be added. Whether or not this
10877 is legal depends on the operating system, and whether it maps the
10878 segment to cover the entire TLS area.
10880 For systems that use GNU libc, the default is on.
10883 @itemx -mno-fused-madd
10884 @opindex mfused-madd
10885 Enable automatic generation of fused floating point multiply-add instructions
10886 if the ISA supports such instructions. The -mfused-madd option is on by
10887 default. The fused multiply-add instructions have a different
10888 rounding behavior compared to executing a multiply followed by an add.
10891 These @samp{-m} switches are supported in addition to the above
10892 on AMD x86-64 processors in 64-bit environments.
10899 Generate code for a 32-bit or 64-bit environment.
10900 The 32-bit environment sets int, long and pointer to 32 bits and
10901 generates code that runs on any i386 system.
10902 The 64-bit environment sets int to 32 bits and long and pointer
10903 to 64 bits and generates code for AMD's x86-64 architecture. For
10904 darwin only the -m64 option turns off the @option{-fno-pic} and
10905 @option{-mdynamic-no-pic} options.
10907 @item -mno-red-zone
10908 @opindex no-red-zone
10909 Do not use a so called red zone for x86-64 code. The red zone is mandated
10910 by the x86-64 ABI, it is a 128-byte area beyond the location of the
10911 stack pointer that will not be modified by signal or interrupt handlers
10912 and therefore can be used for temporary data without adjusting the stack
10913 pointer. The flag @option{-mno-red-zone} disables this red zone.
10915 @item -mcmodel=small
10916 @opindex mcmodel=small
10917 Generate code for the small code model: the program and its symbols must
10918 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
10919 Programs can be statically or dynamically linked. This is the default
10922 @item -mcmodel=kernel
10923 @opindex mcmodel=kernel
10924 Generate code for the kernel code model. The kernel runs in the
10925 negative 2 GB of the address space.
10926 This model has to be used for Linux kernel code.
10928 @item -mcmodel=medium
10929 @opindex mcmodel=medium
10930 Generate code for the medium model: The program is linked in the lower 2
10931 GB of the address space but symbols can be located anywhere in the
10932 address space. Programs can be statically or dynamically linked, but
10933 building of shared libraries are not supported with the medium model.
10935 @item -mcmodel=large
10936 @opindex mcmodel=large
10937 Generate code for the large model: This model makes no assumptions
10938 about addresses and sizes of sections.
10941 @node IA-64 Options
10942 @subsection IA-64 Options
10943 @cindex IA-64 Options
10945 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10949 @opindex mbig-endian
10950 Generate code for a big endian target. This is the default for HP-UX@.
10952 @item -mlittle-endian
10953 @opindex mlittle-endian
10954 Generate code for a little endian target. This is the default for AIX5
10960 @opindex mno-gnu-as
10961 Generate (or don't) code for the GNU assembler. This is the default.
10962 @c Also, this is the default if the configure option @option{--with-gnu-as}
10968 @opindex mno-gnu-ld
10969 Generate (or don't) code for the GNU linker. This is the default.
10970 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10975 Generate code that does not use a global pointer register. The result
10976 is not position independent code, and violates the IA-64 ABI@.
10978 @item -mvolatile-asm-stop
10979 @itemx -mno-volatile-asm-stop
10980 @opindex mvolatile-asm-stop
10981 @opindex mno-volatile-asm-stop
10982 Generate (or don't) a stop bit immediately before and after volatile asm
10985 @item -mregister-names
10986 @itemx -mno-register-names
10987 @opindex mregister-names
10988 @opindex mno-register-names
10989 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
10990 the stacked registers. This may make assembler output more readable.
10996 Disable (or enable) optimizations that use the small data section. This may
10997 be useful for working around optimizer bugs.
10999 @item -mconstant-gp
11000 @opindex mconstant-gp
11001 Generate code that uses a single constant global pointer value. This is
11002 useful when compiling kernel code.
11006 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
11007 This is useful when compiling firmware code.
11009 @item -minline-float-divide-min-latency
11010 @opindex minline-float-divide-min-latency
11011 Generate code for inline divides of floating point values
11012 using the minimum latency algorithm.
11014 @item -minline-float-divide-max-throughput
11015 @opindex minline-float-divide-max-throughput
11016 Generate code for inline divides of floating point values
11017 using the maximum throughput algorithm.
11019 @item -minline-int-divide-min-latency
11020 @opindex minline-int-divide-min-latency
11021 Generate code for inline divides of integer values
11022 using the minimum latency algorithm.
11024 @item -minline-int-divide-max-throughput
11025 @opindex minline-int-divide-max-throughput
11026 Generate code for inline divides of integer values
11027 using the maximum throughput algorithm.
11029 @item -minline-sqrt-min-latency
11030 @opindex minline-sqrt-min-latency
11031 Generate code for inline square roots
11032 using the minimum latency algorithm.
11034 @item -minline-sqrt-max-throughput
11035 @opindex minline-sqrt-max-throughput
11036 Generate code for inline square roots
11037 using the maximum throughput algorithm.
11039 @item -mno-dwarf2-asm
11040 @itemx -mdwarf2-asm
11041 @opindex mno-dwarf2-asm
11042 @opindex mdwarf2-asm
11043 Don't (or do) generate assembler code for the DWARF2 line number debugging
11044 info. This may be useful when not using the GNU assembler.
11046 @item -mearly-stop-bits
11047 @itemx -mno-early-stop-bits
11048 @opindex mearly-stop-bits
11049 @opindex mno-early-stop-bits
11050 Allow stop bits to be placed earlier than immediately preceding the
11051 instruction that triggered the stop bit. This can improve instruction
11052 scheduling, but does not always do so.
11054 @item -mfixed-range=@var{register-range}
11055 @opindex mfixed-range
11056 Generate code treating the given register range as fixed registers.
11057 A fixed register is one that the register allocator can not use. This is
11058 useful when compiling kernel code. A register range is specified as
11059 two registers separated by a dash. Multiple register ranges can be
11060 specified separated by a comma.
11062 @item -mtls-size=@var{tls-size}
11064 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
11067 @item -mtune=@var{cpu-type}
11069 Tune the instruction scheduling for a particular CPU, Valid values are
11070 itanium, itanium1, merced, itanium2, and mckinley.
11076 Add support for multithreading using the POSIX threads library. This
11077 option sets flags for both the preprocessor and linker. It does
11078 not affect the thread safety of object code produced by the compiler or
11079 that of libraries supplied with it. These are HP-UX specific flags.
11085 Generate code for a 32-bit or 64-bit environment.
11086 The 32-bit environment sets int, long and pointer to 32 bits.
11087 The 64-bit environment sets int to 32 bits and long and pointer
11088 to 64 bits. These are HP-UX specific flags.
11090 @item -mno-sched-br-data-spec
11091 @itemx -msched-br-data-spec
11092 @opindex mno-sched-br-data-spec
11093 @opindex msched-br-data-spec
11094 (Dis/En)able data speculative scheduling before reload.
11095 This will result in generation of the ld.a instructions and
11096 the corresponding check instructions (ld.c / chk.a).
11097 The default is 'disable'.
11099 @item -msched-ar-data-spec
11100 @itemx -mno-sched-ar-data-spec
11101 @opindex msched-ar-data-spec
11102 @opindex mno-sched-ar-data-spec
11103 (En/Dis)able data speculative scheduling after reload.
11104 This will result in generation of the ld.a instructions and
11105 the corresponding check instructions (ld.c / chk.a).
11106 The default is 'enable'.
11108 @item -mno-sched-control-spec
11109 @itemx -msched-control-spec
11110 @opindex mno-sched-control-spec
11111 @opindex msched-control-spec
11112 (Dis/En)able control speculative scheduling. This feature is
11113 available only during region scheduling (i.e.@: before reload).
11114 This will result in generation of the ld.s instructions and
11115 the corresponding check instructions chk.s .
11116 The default is 'disable'.
11118 @item -msched-br-in-data-spec
11119 @itemx -mno-sched-br-in-data-spec
11120 @opindex msched-br-in-data-spec
11121 @opindex mno-sched-br-in-data-spec
11122 (En/Dis)able speculative scheduling of the instructions that
11123 are dependent on the data speculative loads before reload.
11124 This is effective only with @option{-msched-br-data-spec} enabled.
11125 The default is 'enable'.
11127 @item -msched-ar-in-data-spec
11128 @itemx -mno-sched-ar-in-data-spec
11129 @opindex msched-ar-in-data-spec
11130 @opindex mno-sched-ar-in-data-spec
11131 (En/Dis)able speculative scheduling of the instructions that
11132 are dependent on the data speculative loads after reload.
11133 This is effective only with @option{-msched-ar-data-spec} enabled.
11134 The default is 'enable'.
11136 @item -msched-in-control-spec
11137 @itemx -mno-sched-in-control-spec
11138 @opindex msched-in-control-spec
11139 @opindex mno-sched-in-control-spec
11140 (En/Dis)able speculative scheduling of the instructions that
11141 are dependent on the control speculative loads.
11142 This is effective only with @option{-msched-control-spec} enabled.
11143 The default is 'enable'.
11146 @itemx -mno-sched-ldc
11147 @opindex msched-ldc
11148 @opindex mno-sched-ldc
11149 (En/Dis)able use of simple data speculation checks ld.c .
11150 If disabled, only chk.a instructions will be emitted to check
11151 data speculative loads.
11152 The default is 'enable'.
11154 @item -mno-sched-control-ldc
11155 @itemx -msched-control-ldc
11156 @opindex mno-sched-control-ldc
11157 @opindex msched-control-ldc
11158 (Dis/En)able use of ld.c instructions to check control speculative loads.
11159 If enabled, in case of control speculative load with no speculatively
11160 scheduled dependent instructions this load will be emitted as ld.sa and
11161 ld.c will be used to check it.
11162 The default is 'disable'.
11164 @item -mno-sched-spec-verbose
11165 @itemx -msched-spec-verbose
11166 @opindex mno-sched-spec-verbose
11167 @opindex msched-spec-verbose
11168 (Dis/En)able printing of the information about speculative motions.
11170 @item -mno-sched-prefer-non-data-spec-insns
11171 @itemx -msched-prefer-non-data-spec-insns
11172 @opindex mno-sched-prefer-non-data-spec-insns
11173 @opindex msched-prefer-non-data-spec-insns
11174 If enabled, data speculative instructions will be chosen for schedule
11175 only if there are no other choices at the moment. This will make
11176 the use of the data speculation much more conservative.
11177 The default is 'disable'.
11179 @item -mno-sched-prefer-non-control-spec-insns
11180 @itemx -msched-prefer-non-control-spec-insns
11181 @opindex mno-sched-prefer-non-control-spec-insns
11182 @opindex msched-prefer-non-control-spec-insns
11183 If enabled, control speculative instructions will be chosen for schedule
11184 only if there are no other choices at the moment. This will make
11185 the use of the control speculation much more conservative.
11186 The default is 'disable'.
11188 @item -mno-sched-count-spec-in-critical-path
11189 @itemx -msched-count-spec-in-critical-path
11190 @opindex mno-sched-count-spec-in-critical-path
11191 @opindex msched-count-spec-in-critical-path
11192 If enabled, speculative dependencies will be considered during
11193 computation of the instructions priorities. This will make the use of the
11194 speculation a bit more conservative.
11195 The default is 'disable'.
11200 @subsection M32C Options
11201 @cindex M32C options
11204 @item -mcpu=@var{name}
11206 Select the CPU for which code is generated. @var{name} may be one of
11207 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
11208 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
11209 the M32C/80 series.
11213 Specifies that the program will be run on the simulator. This causes
11214 an alternate runtime library to be linked in which supports, for
11215 example, file I/O@. You must not use this option when generating
11216 programs that will run on real hardware; you must provide your own
11217 runtime library for whatever I/O functions are needed.
11219 @item -memregs=@var{number}
11221 Specifies the number of memory-based pseudo-registers GCC will use
11222 during code generation. These pseudo-registers will be used like real
11223 registers, so there is a tradeoff between GCC's ability to fit the
11224 code into available registers, and the performance penalty of using
11225 memory instead of registers. Note that all modules in a program must
11226 be compiled with the same value for this option. Because of that, you
11227 must not use this option with the default runtime libraries gcc
11232 @node M32R/D Options
11233 @subsection M32R/D Options
11234 @cindex M32R/D options
11236 These @option{-m} options are defined for Renesas M32R/D architectures:
11241 Generate code for the M32R/2@.
11245 Generate code for the M32R/X@.
11249 Generate code for the M32R@. This is the default.
11251 @item -mmodel=small
11252 @opindex mmodel=small
11253 Assume all objects live in the lower 16MB of memory (so that their addresses
11254 can be loaded with the @code{ld24} instruction), and assume all subroutines
11255 are reachable with the @code{bl} instruction.
11256 This is the default.
11258 The addressability of a particular object can be set with the
11259 @code{model} attribute.
11261 @item -mmodel=medium
11262 @opindex mmodel=medium
11263 Assume objects may be anywhere in the 32-bit address space (the compiler
11264 will generate @code{seth/add3} instructions to load their addresses), and
11265 assume all subroutines are reachable with the @code{bl} instruction.
11267 @item -mmodel=large
11268 @opindex mmodel=large
11269 Assume objects may be anywhere in the 32-bit address space (the compiler
11270 will generate @code{seth/add3} instructions to load their addresses), and
11271 assume subroutines may not be reachable with the @code{bl} instruction
11272 (the compiler will generate the much slower @code{seth/add3/jl}
11273 instruction sequence).
11276 @opindex msdata=none
11277 Disable use of the small data area. Variables will be put into
11278 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
11279 @code{section} attribute has been specified).
11280 This is the default.
11282 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
11283 Objects may be explicitly put in the small data area with the
11284 @code{section} attribute using one of these sections.
11286 @item -msdata=sdata
11287 @opindex msdata=sdata
11288 Put small global and static data in the small data area, but do not
11289 generate special code to reference them.
11292 @opindex msdata=use
11293 Put small global and static data in the small data area, and generate
11294 special instructions to reference them.
11298 @cindex smaller data references
11299 Put global and static objects less than or equal to @var{num} bytes
11300 into the small data or bss sections instead of the normal data or bss
11301 sections. The default value of @var{num} is 8.
11302 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
11303 for this option to have any effect.
11305 All modules should be compiled with the same @option{-G @var{num}} value.
11306 Compiling with different values of @var{num} may or may not work; if it
11307 doesn't the linker will give an error message---incorrect code will not be
11312 Makes the M32R specific code in the compiler display some statistics
11313 that might help in debugging programs.
11315 @item -malign-loops
11316 @opindex malign-loops
11317 Align all loops to a 32-byte boundary.
11319 @item -mno-align-loops
11320 @opindex mno-align-loops
11321 Do not enforce a 32-byte alignment for loops. This is the default.
11323 @item -missue-rate=@var{number}
11324 @opindex missue-rate=@var{number}
11325 Issue @var{number} instructions per cycle. @var{number} can only be 1
11328 @item -mbranch-cost=@var{number}
11329 @opindex mbranch-cost=@var{number}
11330 @var{number} can only be 1 or 2. If it is 1 then branches will be
11331 preferred over conditional code, if it is 2, then the opposite will
11334 @item -mflush-trap=@var{number}
11335 @opindex mflush-trap=@var{number}
11336 Specifies the trap number to use to flush the cache. The default is
11337 12. Valid numbers are between 0 and 15 inclusive.
11339 @item -mno-flush-trap
11340 @opindex mno-flush-trap
11341 Specifies that the cache cannot be flushed by using a trap.
11343 @item -mflush-func=@var{name}
11344 @opindex mflush-func=@var{name}
11345 Specifies the name of the operating system function to call to flush
11346 the cache. The default is @emph{_flush_cache}, but a function call
11347 will only be used if a trap is not available.
11349 @item -mno-flush-func
11350 @opindex mno-flush-func
11351 Indicates that there is no OS function for flushing the cache.
11355 @node M680x0 Options
11356 @subsection M680x0 Options
11357 @cindex M680x0 options
11359 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
11360 The default settings depend on which architecture was selected when
11361 the compiler was configured; the defaults for the most common choices
11365 @item -march=@var{arch}
11367 Generate code for a specific M680x0 or ColdFire instruction set
11368 architecture. Permissible values of @var{arch} for M680x0
11369 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
11370 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
11371 architectures are selected according to Freescale's ISA classification
11372 and the permissible values are: @samp{isaa}, @samp{isaaplus},
11373 @samp{isab} and @samp{isac}.
11375 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
11376 code for a ColdFire target. The @var{arch} in this macro is one of the
11377 @option{-march} arguments given above.
11379 When used together, @option{-march} and @option{-mtune} select code
11380 that runs on a family of similar processors but that is optimized
11381 for a particular microarchitecture.
11383 @item -mcpu=@var{cpu}
11385 Generate code for a specific M680x0 or ColdFire processor.
11386 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
11387 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
11388 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
11389 below, which also classifies the CPUs into families:
11391 @multitable @columnfractions 0.20 0.80
11392 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
11393 @item @samp{51qe} @tab @samp{51qe}
11394 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
11395 @item @samp{5206e} @tab @samp{5206e}
11396 @item @samp{5208} @tab @samp{5207} @samp{5208}
11397 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
11398 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
11399 @item @samp{5216} @tab @samp{5214} @samp{5216}
11400 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
11401 @item @samp{5225} @tab @samp{5224} @samp{5225}
11402 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
11403 @item @samp{5249} @tab @samp{5249}
11404 @item @samp{5250} @tab @samp{5250}
11405 @item @samp{5271} @tab @samp{5270} @samp{5271}
11406 @item @samp{5272} @tab @samp{5272}
11407 @item @samp{5275} @tab @samp{5274} @samp{5275}
11408 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
11409 @item @samp{5307} @tab @samp{5307}
11410 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
11411 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
11412 @item @samp{5407} @tab @samp{5407}
11413 @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}
11416 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
11417 @var{arch} is compatible with @var{cpu}. Other combinations of
11418 @option{-mcpu} and @option{-march} are rejected.
11420 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
11421 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
11422 where the value of @var{family} is given by the table above.
11424 @item -mtune=@var{tune}
11426 Tune the code for a particular microarchitecture, within the
11427 constraints set by @option{-march} and @option{-mcpu}.
11428 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
11429 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
11430 and @samp{cpu32}. The ColdFire microarchitectures
11431 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
11433 You can also use @option{-mtune=68020-40} for code that needs
11434 to run relatively well on 68020, 68030 and 68040 targets.
11435 @option{-mtune=68020-60} is similar but includes 68060 targets
11436 as well. These two options select the same tuning decisions as
11437 @option{-m68020-40} and @option{-m68020-60} respectively.
11439 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
11440 when tuning for 680x0 architecture @var{arch}. It also defines
11441 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
11442 option is used. If gcc is tuning for a range of architectures,
11443 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
11444 it defines the macros for every architecture in the range.
11446 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
11447 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
11448 of the arguments given above.
11454 Generate output for a 68000. This is the default
11455 when the compiler is configured for 68000-based systems.
11456 It is equivalent to @option{-march=68000}.
11458 Use this option for microcontrollers with a 68000 or EC000 core,
11459 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
11463 Generate output for a 68010. This is the default
11464 when the compiler is configured for 68010-based systems.
11465 It is equivalent to @option{-march=68010}.
11471 Generate output for a 68020. This is the default
11472 when the compiler is configured for 68020-based systems.
11473 It is equivalent to @option{-march=68020}.
11477 Generate output for a 68030. This is the default when the compiler is
11478 configured for 68030-based systems. It is equivalent to
11479 @option{-march=68030}.
11483 Generate output for a 68040. This is the default when the compiler is
11484 configured for 68040-based systems. It is equivalent to
11485 @option{-march=68040}.
11487 This option inhibits the use of 68881/68882 instructions that have to be
11488 emulated by software on the 68040. Use this option if your 68040 does not
11489 have code to emulate those instructions.
11493 Generate output for a 68060. This is the default when the compiler is
11494 configured for 68060-based systems. It is equivalent to
11495 @option{-march=68060}.
11497 This option inhibits the use of 68020 and 68881/68882 instructions that
11498 have to be emulated by software on the 68060. Use this option if your 68060
11499 does not have code to emulate those instructions.
11503 Generate output for a CPU32. This is the default
11504 when the compiler is configured for CPU32-based systems.
11505 It is equivalent to @option{-march=cpu32}.
11507 Use this option for microcontrollers with a
11508 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
11509 68336, 68340, 68341, 68349 and 68360.
11513 Generate output for a 520X ColdFire CPU@. This is the default
11514 when the compiler is configured for 520X-based systems.
11515 It is equivalent to @option{-mcpu=5206}, and is now deprecated
11516 in favor of that option.
11518 Use this option for microcontroller with a 5200 core, including
11519 the MCF5202, MCF5203, MCF5204 and MCF5206.
11523 Generate output for a 5206e ColdFire CPU@. The option is now
11524 deprecated in favor of the equivalent @option{-mcpu=5206e}.
11528 Generate output for a member of the ColdFire 528X family.
11529 The option is now deprecated in favor of the equivalent
11530 @option{-mcpu=528x}.
11534 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
11535 in favor of the equivalent @option{-mcpu=5307}.
11539 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
11540 in favor of the equivalent @option{-mcpu=5407}.
11544 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
11545 This includes use of hardware floating point instructions.
11546 The option is equivalent to @option{-mcpu=547x}, and is now
11547 deprecated in favor of that option.
11551 Generate output for a 68040, without using any of the new instructions.
11552 This results in code which can run relatively efficiently on either a
11553 68020/68881 or a 68030 or a 68040. The generated code does use the
11554 68881 instructions that are emulated on the 68040.
11556 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
11560 Generate output for a 68060, without using any of the new instructions.
11561 This results in code which can run relatively efficiently on either a
11562 68020/68881 or a 68030 or a 68040. The generated code does use the
11563 68881 instructions that are emulated on the 68060.
11565 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
11569 @opindex mhard-float
11571 Generate floating-point instructions. This is the default for 68020
11572 and above, and for ColdFire devices that have an FPU@. It defines the
11573 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
11574 on ColdFire targets.
11577 @opindex msoft-float
11578 Do not generate floating-point instructions; use library calls instead.
11579 This is the default for 68000, 68010, and 68832 targets. It is also
11580 the default for ColdFire devices that have no FPU.
11586 Generate (do not generate) ColdFire hardware divide and remainder
11587 instructions. If @option{-march} is used without @option{-mcpu},
11588 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
11589 architectures. Otherwise, the default is taken from the target CPU
11590 (either the default CPU, or the one specified by @option{-mcpu}). For
11591 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
11592 @option{-mcpu=5206e}.
11594 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
11598 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11599 Additionally, parameters passed on the stack are also aligned to a
11600 16-bit boundary even on targets whose API mandates promotion to 32-bit.
11604 Do not consider type @code{int} to be 16 bits wide. This is the default.
11607 @itemx -mno-bitfield
11608 @opindex mnobitfield
11609 @opindex mno-bitfield
11610 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
11611 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
11615 Do use the bit-field instructions. The @option{-m68020} option implies
11616 @option{-mbitfield}. This is the default if you use a configuration
11617 designed for a 68020.
11621 Use a different function-calling convention, in which functions
11622 that take a fixed number of arguments return with the @code{rtd}
11623 instruction, which pops their arguments while returning. This
11624 saves one instruction in the caller since there is no need to pop
11625 the arguments there.
11627 This calling convention is incompatible with the one normally
11628 used on Unix, so you cannot use it if you need to call libraries
11629 compiled with the Unix compiler.
11631 Also, you must provide function prototypes for all functions that
11632 take variable numbers of arguments (including @code{printf});
11633 otherwise incorrect code will be generated for calls to those
11636 In addition, seriously incorrect code will result if you call a
11637 function with too many arguments. (Normally, extra arguments are
11638 harmlessly ignored.)
11640 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
11641 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
11645 Do not use the calling conventions selected by @option{-mrtd}.
11646 This is the default.
11649 @itemx -mno-align-int
11650 @opindex malign-int
11651 @opindex mno-align-int
11652 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
11653 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
11654 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
11655 Aligning variables on 32-bit boundaries produces code that runs somewhat
11656 faster on processors with 32-bit busses at the expense of more memory.
11658 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
11659 align structures containing the above types differently than
11660 most published application binary interface specifications for the m68k.
11664 Use the pc-relative addressing mode of the 68000 directly, instead of
11665 using a global offset table. At present, this option implies @option{-fpic},
11666 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
11667 not presently supported with @option{-mpcrel}, though this could be supported for
11668 68020 and higher processors.
11670 @item -mno-strict-align
11671 @itemx -mstrict-align
11672 @opindex mno-strict-align
11673 @opindex mstrict-align
11674 Do not (do) assume that unaligned memory references will be handled by
11678 Generate code that allows the data segment to be located in a different
11679 area of memory from the text segment. This allows for execute in place in
11680 an environment without virtual memory management. This option implies
11683 @item -mno-sep-data
11684 Generate code that assumes that the data segment follows the text segment.
11685 This is the default.
11687 @item -mid-shared-library
11688 Generate code that supports shared libraries via the library ID method.
11689 This allows for execute in place and shared libraries in an environment
11690 without virtual memory management. This option implies @option{-fPIC}.
11692 @item -mno-id-shared-library
11693 Generate code that doesn't assume ID based shared libraries are being used.
11694 This is the default.
11696 @item -mshared-library-id=n
11697 Specified the identification number of the ID based shared library being
11698 compiled. Specifying a value of 0 will generate more compact code, specifying
11699 other values will force the allocation of that number to the current
11700 library but is no more space or time efficient than omitting this option.
11704 @node M68hc1x Options
11705 @subsection M68hc1x Options
11706 @cindex M68hc1x options
11708 These are the @samp{-m} options defined for the 68hc11 and 68hc12
11709 microcontrollers. The default values for these options depends on
11710 which style of microcontroller was selected when the compiler was configured;
11711 the defaults for the most common choices are given below.
11718 Generate output for a 68HC11. This is the default
11719 when the compiler is configured for 68HC11-based systems.
11725 Generate output for a 68HC12. This is the default
11726 when the compiler is configured for 68HC12-based systems.
11732 Generate output for a 68HCS12.
11734 @item -mauto-incdec
11735 @opindex mauto-incdec
11736 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
11743 Enable the use of 68HC12 min and max instructions.
11746 @itemx -mno-long-calls
11747 @opindex mlong-calls
11748 @opindex mno-long-calls
11749 Treat all calls as being far away (near). If calls are assumed to be
11750 far away, the compiler will use the @code{call} instruction to
11751 call a function and the @code{rtc} instruction for returning.
11755 Consider type @code{int} to be 16 bits wide, like @code{short int}.
11757 @item -msoft-reg-count=@var{count}
11758 @opindex msoft-reg-count
11759 Specify the number of pseudo-soft registers which are used for the
11760 code generation. The maximum number is 32. Using more pseudo-soft
11761 register may or may not result in better code depending on the program.
11762 The default is 4 for 68HC11 and 2 for 68HC12.
11766 @node MCore Options
11767 @subsection MCore Options
11768 @cindex MCore options
11770 These are the @samp{-m} options defined for the Motorola M*Core
11776 @itemx -mno-hardlit
11778 @opindex mno-hardlit
11779 Inline constants into the code stream if it can be done in two
11780 instructions or less.
11786 Use the divide instruction. (Enabled by default).
11788 @item -mrelax-immediate
11789 @itemx -mno-relax-immediate
11790 @opindex mrelax-immediate
11791 @opindex mno-relax-immediate
11792 Allow arbitrary sized immediates in bit operations.
11794 @item -mwide-bitfields
11795 @itemx -mno-wide-bitfields
11796 @opindex mwide-bitfields
11797 @opindex mno-wide-bitfields
11798 Always treat bit-fields as int-sized.
11800 @item -m4byte-functions
11801 @itemx -mno-4byte-functions
11802 @opindex m4byte-functions
11803 @opindex mno-4byte-functions
11804 Force all functions to be aligned to a four byte boundary.
11806 @item -mcallgraph-data
11807 @itemx -mno-callgraph-data
11808 @opindex mcallgraph-data
11809 @opindex mno-callgraph-data
11810 Emit callgraph information.
11813 @itemx -mno-slow-bytes
11814 @opindex mslow-bytes
11815 @opindex mno-slow-bytes
11816 Prefer word access when reading byte quantities.
11818 @item -mlittle-endian
11819 @itemx -mbig-endian
11820 @opindex mlittle-endian
11821 @opindex mbig-endian
11822 Generate code for a little endian target.
11828 Generate code for the 210 processor.
11832 @subsection MIPS Options
11833 @cindex MIPS options
11839 Generate big-endian code.
11843 Generate little-endian code. This is the default for @samp{mips*el-*-*}
11846 @item -march=@var{arch}
11848 Generate code that will run on @var{arch}, which can be the name of a
11849 generic MIPS ISA, or the name of a particular processor.
11851 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
11852 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
11853 The processor names are:
11854 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
11855 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
11856 @samp{5kc}, @samp{5kf},
11858 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
11859 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
11860 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
11861 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
11864 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
11865 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
11866 @samp{rm7000}, @samp{rm9000},
11869 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
11870 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
11871 The special value @samp{from-abi} selects the
11872 most compatible architecture for the selected ABI (that is,
11873 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
11875 In processor names, a final @samp{000} can be abbreviated as @samp{k}
11876 (for example, @samp{-march=r2k}). Prefixes are optional, and
11877 @samp{vr} may be written @samp{r}.
11879 Names of the form @samp{@var{n}f2_1} refer to processors with
11880 FPUs clocked at half the rate of the core, names of the form
11881 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
11882 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
11883 processors with FPUs clocked a ratio of 3:2 with respect to the core.
11884 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
11885 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
11886 accepted as synonyms for @samp{@var{n}f1_1}.
11888 GCC defines two macros based on the value of this option. The first
11889 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
11890 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
11891 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
11892 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
11893 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
11895 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
11896 above. In other words, it will have the full prefix and will not
11897 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
11898 the macro names the resolved architecture (either @samp{"mips1"} or
11899 @samp{"mips3"}). It names the default architecture when no
11900 @option{-march} option is given.
11902 @item -mtune=@var{arch}
11904 Optimize for @var{arch}. Among other things, this option controls
11905 the way instructions are scheduled, and the perceived cost of arithmetic
11906 operations. The list of @var{arch} values is the same as for
11909 When this option is not used, GCC will optimize for the processor
11910 specified by @option{-march}. By using @option{-march} and
11911 @option{-mtune} together, it is possible to generate code that will
11912 run on a family of processors, but optimize the code for one
11913 particular member of that family.
11915 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
11916 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
11917 @samp{-march} ones described above.
11921 Equivalent to @samp{-march=mips1}.
11925 Equivalent to @samp{-march=mips2}.
11929 Equivalent to @samp{-march=mips3}.
11933 Equivalent to @samp{-march=mips4}.
11937 Equivalent to @samp{-march=mips32}.
11941 Equivalent to @samp{-march=mips32r2}.
11945 Equivalent to @samp{-march=mips64}.
11950 @opindex mno-mips16
11951 Generate (do not generate) MIPS16 code. If GCC is targetting a
11952 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
11954 MIPS16 code generation can also be controlled on a per-function basis
11955 by means of @code{mips16} and @code{nomips16} attributes.
11956 @xref{Function Attributes}, for more information.
11958 @item -mflip-mips16
11959 @opindex mflip-mips16
11960 Generate MIPS16 code on alternating functions. This option is provided
11961 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
11962 not intended for ordinary use in compiling user code.
11964 @item -minterlink-mips16
11965 @itemx -mno-interlink-mips16
11966 @opindex minterlink-mips16
11967 @opindex mno-interlink-mips16
11968 Require (do not require) that non-MIPS16 code be link-compatible with
11971 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
11972 it must either use a call or an indirect jump. @option{-minterlink-mips16}
11973 therefore disables direct jumps unless GCC knows that the target of the
11974 jump is not MIPS16.
11986 Generate code for the given ABI@.
11988 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
11989 generates 64-bit code when you select a 64-bit architecture, but you
11990 can use @option{-mgp32} to get 32-bit code instead.
11992 For information about the O64 ABI, see
11993 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
11995 GCC supports a variant of the o32 ABI in which floating-point registers
11996 are 64 rather than 32 bits wide. You can select this combination with
11997 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
11998 and @samp{mfhc1} instructions and is therefore only supported for
11999 MIPS32R2 processors.
12001 The register assignments for arguments and return values remain the
12002 same, but each scalar value is passed in a single 64-bit register
12003 rather than a pair of 32-bit registers. For example, scalar
12004 floating-point values are returned in @samp{$f0} only, not a
12005 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
12006 remains the same, but all 64 bits are saved.
12009 @itemx -mno-abicalls
12011 @opindex mno-abicalls
12012 Generate (do not generate) code that is suitable for SVR4-style
12013 dynamic objects. @option{-mabicalls} is the default for SVR4-based
12018 Generate (do not generate) code that is fully position-independent,
12019 and that can therefore be linked into shared libraries. This option
12020 only affects @option{-mabicalls}.
12022 All @option{-mabicalls} code has traditionally been position-independent,
12023 regardless of options like @option{-fPIC} and @option{-fpic}. However,
12024 as an extension, the GNU toolchain allows executables to use absolute
12025 accesses for locally-binding symbols. It can also use shorter GP
12026 initialization sequences and generate direct calls to locally-defined
12027 functions. This mode is selected by @option{-mno-shared}.
12029 @option{-mno-shared} depends on binutils 2.16 or higher and generates
12030 objects that can only be linked by the GNU linker. However, the option
12031 does not affect the ABI of the final executable; it only affects the ABI
12032 of relocatable objects. Using @option{-mno-shared} will generally make
12033 executables both smaller and quicker.
12035 @option{-mshared} is the default.
12041 Lift (do not lift) the usual restrictions on the size of the global
12044 GCC normally uses a single instruction to load values from the GOT@.
12045 While this is relatively efficient, it will only work if the GOT
12046 is smaller than about 64k. Anything larger will cause the linker
12047 to report an error such as:
12049 @cindex relocation truncated to fit (MIPS)
12051 relocation truncated to fit: R_MIPS_GOT16 foobar
12054 If this happens, you should recompile your code with @option{-mxgot}.
12055 It should then work with very large GOTs, although it will also be
12056 less efficient, since it will take three instructions to fetch the
12057 value of a global symbol.
12059 Note that some linkers can create multiple GOTs. If you have such a
12060 linker, you should only need to use @option{-mxgot} when a single object
12061 file accesses more than 64k's worth of GOT entries. Very few do.
12063 These options have no effect unless GCC is generating position
12068 Assume that general-purpose registers are 32 bits wide.
12072 Assume that general-purpose registers are 64 bits wide.
12076 Assume that floating-point registers are 32 bits wide.
12080 Assume that floating-point registers are 64 bits wide.
12083 @opindex mhard-float
12084 Use floating-point coprocessor instructions.
12087 @opindex msoft-float
12088 Do not use floating-point coprocessor instructions. Implement
12089 floating-point calculations using library calls instead.
12091 @item -msingle-float
12092 @opindex msingle-float
12093 Assume that the floating-point coprocessor only supports single-precision
12096 @item -mdouble-float
12097 @opindex mdouble-float
12098 Assume that the floating-point coprocessor supports double-precision
12099 operations. This is the default.
12105 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
12106 implement atomic memory built-in functions. When neither option is
12107 specified, GCC will use the instructions if the target architecture
12110 @option{-mllsc} is useful if the runtime environment can emulate the
12111 instructions and @option{-mno-llsc} can be useful when compiling for
12112 nonstandard ISAs. You can make either option the default by
12113 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
12114 respectively. @option{--with-llsc} is the default for some
12115 configurations; see the installation documentation for details.
12121 Use (do not use) revision 1 of the MIPS DSP ASE@.
12122 @xref{MIPS DSP Built-in Functions}. This option defines the
12123 preprocessor macro @samp{__mips_dsp}. It also defines
12124 @samp{__mips_dsp_rev} to 1.
12130 Use (do not use) revision 2 of the MIPS DSP ASE@.
12131 @xref{MIPS DSP Built-in Functions}. This option defines the
12132 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
12133 It also defines @samp{__mips_dsp_rev} to 2.
12136 @itemx -mno-smartmips
12137 @opindex msmartmips
12138 @opindex mno-smartmips
12139 Use (do not use) the MIPS SmartMIPS ASE.
12141 @item -mpaired-single
12142 @itemx -mno-paired-single
12143 @opindex mpaired-single
12144 @opindex mno-paired-single
12145 Use (do not use) paired-single floating-point instructions.
12146 @xref{MIPS Paired-Single Support}. This option requires
12147 hardware floating-point support to be enabled.
12153 Use (do not use) MIPS Digital Media Extension instructions.
12154 This option can only be used when generating 64-bit code and requires
12155 hardware floating-point support to be enabled.
12160 @opindex mno-mips3d
12161 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
12162 The option @option{-mips3d} implies @option{-mpaired-single}.
12168 Use (do not use) MT Multithreading instructions.
12172 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
12173 an explanation of the default and the way that the pointer size is
12178 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
12180 The default size of @code{int}s, @code{long}s and pointers depends on
12181 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
12182 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
12183 32-bit @code{long}s. Pointers are the same size as @code{long}s,
12184 or the same size as integer registers, whichever is smaller.
12190 Assume (do not assume) that all symbols have 32-bit values, regardless
12191 of the selected ABI@. This option is useful in combination with
12192 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
12193 to generate shorter and faster references to symbolic addresses.
12197 Put definitions of externally-visible data in a small data section
12198 if that data is no bigger than @var{num} bytes. GCC can then access
12199 the data more efficiently; see @option{-mgpopt} for details.
12201 The default @option{-G} option depends on the configuration.
12203 @item -mlocal-sdata
12204 @itemx -mno-local-sdata
12205 @opindex mlocal-sdata
12206 @opindex mno-local-sdata
12207 Extend (do not extend) the @option{-G} behavior to local data too,
12208 such as to static variables in C@. @option{-mlocal-sdata} is the
12209 default for all configurations.
12211 If the linker complains that an application is using too much small data,
12212 you might want to try rebuilding the less performance-critical parts with
12213 @option{-mno-local-sdata}. You might also want to build large
12214 libraries with @option{-mno-local-sdata}, so that the libraries leave
12215 more room for the main program.
12217 @item -mextern-sdata
12218 @itemx -mno-extern-sdata
12219 @opindex mextern-sdata
12220 @opindex mno-extern-sdata
12221 Assume (do not assume) that externally-defined data will be in
12222 a small data section if that data is within the @option{-G} limit.
12223 @option{-mextern-sdata} is the default for all configurations.
12225 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
12226 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
12227 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
12228 is placed in a small data section. If @var{Var} is defined by another
12229 module, you must either compile that module with a high-enough
12230 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
12231 definition. If @var{Var} is common, you must link the application
12232 with a high-enough @option{-G} setting.
12234 The easiest way of satisfying these restrictions is to compile
12235 and link every module with the same @option{-G} option. However,
12236 you may wish to build a library that supports several different
12237 small data limits. You can do this by compiling the library with
12238 the highest supported @option{-G} setting and additionally using
12239 @option{-mno-extern-sdata} to stop the library from making assumptions
12240 about externally-defined data.
12246 Use (do not use) GP-relative accesses for symbols that are known to be
12247 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
12248 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
12251 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
12252 might not hold the value of @code{_gp}. For example, if the code is
12253 part of a library that might be used in a boot monitor, programs that
12254 call boot monitor routines will pass an unknown value in @code{$gp}.
12255 (In such situations, the boot monitor itself would usually be compiled
12256 with @option{-G0}.)
12258 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
12259 @option{-mno-extern-sdata}.
12261 @item -membedded-data
12262 @itemx -mno-embedded-data
12263 @opindex membedded-data
12264 @opindex mno-embedded-data
12265 Allocate variables to the read-only data section first if possible, then
12266 next in the small data section if possible, otherwise in data. This gives
12267 slightly slower code than the default, but reduces the amount of RAM required
12268 when executing, and thus may be preferred for some embedded systems.
12270 @item -muninit-const-in-rodata
12271 @itemx -mno-uninit-const-in-rodata
12272 @opindex muninit-const-in-rodata
12273 @opindex mno-uninit-const-in-rodata
12274 Put uninitialized @code{const} variables in the read-only data section.
12275 This option is only meaningful in conjunction with @option{-membedded-data}.
12277 @item -mcode-readable=@var{setting}
12278 @opindex mcode-readable
12279 Specify whether GCC may generate code that reads from executable sections.
12280 There are three possible settings:
12283 @item -mcode-readable=yes
12284 Instructions may freely access executable sections. This is the
12287 @item -mcode-readable=pcrel
12288 MIPS16 PC-relative load instructions can access executable sections,
12289 but other instructions must not do so. This option is useful on 4KSc
12290 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
12291 It is also useful on processors that can be configured to have a dual
12292 instruction/data SRAM interface and that, like the M4K, automatically
12293 redirect PC-relative loads to the instruction RAM.
12295 @item -mcode-readable=no
12296 Instructions must not access executable sections. This option can be
12297 useful on targets that are configured to have a dual instruction/data
12298 SRAM interface but that (unlike the M4K) do not automatically redirect
12299 PC-relative loads to the instruction RAM.
12302 @item -msplit-addresses
12303 @itemx -mno-split-addresses
12304 @opindex msplit-addresses
12305 @opindex mno-split-addresses
12306 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
12307 relocation operators. This option has been superseded by
12308 @option{-mexplicit-relocs} but is retained for backwards compatibility.
12310 @item -mexplicit-relocs
12311 @itemx -mno-explicit-relocs
12312 @opindex mexplicit-relocs
12313 @opindex mno-explicit-relocs
12314 Use (do not use) assembler relocation operators when dealing with symbolic
12315 addresses. The alternative, selected by @option{-mno-explicit-relocs},
12316 is to use assembler macros instead.
12318 @option{-mexplicit-relocs} is the default if GCC was configured
12319 to use an assembler that supports relocation operators.
12321 @item -mcheck-zero-division
12322 @itemx -mno-check-zero-division
12323 @opindex mcheck-zero-division
12324 @opindex mno-check-zero-division
12325 Trap (do not trap) on integer division by zero.
12327 The default is @option{-mcheck-zero-division}.
12329 @item -mdivide-traps
12330 @itemx -mdivide-breaks
12331 @opindex mdivide-traps
12332 @opindex mdivide-breaks
12333 MIPS systems check for division by zero by generating either a
12334 conditional trap or a break instruction. Using traps results in
12335 smaller code, but is only supported on MIPS II and later. Also, some
12336 versions of the Linux kernel have a bug that prevents trap from
12337 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
12338 allow conditional traps on architectures that support them and
12339 @option{-mdivide-breaks} to force the use of breaks.
12341 The default is usually @option{-mdivide-traps}, but this can be
12342 overridden at configure time using @option{--with-divide=breaks}.
12343 Divide-by-zero checks can be completely disabled using
12344 @option{-mno-check-zero-division}.
12349 @opindex mno-memcpy
12350 Force (do not force) the use of @code{memcpy()} for non-trivial block
12351 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
12352 most constant-sized copies.
12355 @itemx -mno-long-calls
12356 @opindex mlong-calls
12357 @opindex mno-long-calls
12358 Disable (do not disable) use of the @code{jal} instruction. Calling
12359 functions using @code{jal} is more efficient but requires the caller
12360 and callee to be in the same 256 megabyte segment.
12362 This option has no effect on abicalls code. The default is
12363 @option{-mno-long-calls}.
12369 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
12370 instructions, as provided by the R4650 ISA@.
12373 @itemx -mno-fused-madd
12374 @opindex mfused-madd
12375 @opindex mno-fused-madd
12376 Enable (disable) use of the floating point multiply-accumulate
12377 instructions, when they are available. The default is
12378 @option{-mfused-madd}.
12380 When multiply-accumulate instructions are used, the intermediate
12381 product is calculated to infinite precision and is not subject to
12382 the FCSR Flush to Zero bit. This may be undesirable in some
12387 Tell the MIPS assembler to not run its preprocessor over user
12388 assembler files (with a @samp{.s} suffix) when assembling them.
12391 @itemx -mno-fix-r4000
12392 @opindex mfix-r4000
12393 @opindex mno-fix-r4000
12394 Work around certain R4000 CPU errata:
12397 A double-word or a variable shift may give an incorrect result if executed
12398 immediately after starting an integer division.
12400 A double-word or a variable shift may give an incorrect result if executed
12401 while an integer multiplication is in progress.
12403 An integer division may give an incorrect result if started in a delay slot
12404 of a taken branch or a jump.
12408 @itemx -mno-fix-r4400
12409 @opindex mfix-r4400
12410 @opindex mno-fix-r4400
12411 Work around certain R4400 CPU errata:
12414 A double-word or a variable shift may give an incorrect result if executed
12415 immediately after starting an integer division.
12419 @itemx -mno-fix-vr4120
12420 @opindex mfix-vr4120
12421 Work around certain VR4120 errata:
12424 @code{dmultu} does not always produce the correct result.
12426 @code{div} and @code{ddiv} do not always produce the correct result if one
12427 of the operands is negative.
12429 The workarounds for the division errata rely on special functions in
12430 @file{libgcc.a}. At present, these functions are only provided by
12431 the @code{mips64vr*-elf} configurations.
12433 Other VR4120 errata require a nop to be inserted between certain pairs of
12434 instructions. These errata are handled by the assembler, not by GCC itself.
12437 @opindex mfix-vr4130
12438 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
12439 workarounds are implemented by the assembler rather than by GCC,
12440 although GCC will avoid using @code{mflo} and @code{mfhi} if the
12441 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
12442 instructions are available instead.
12445 @itemx -mno-fix-sb1
12447 Work around certain SB-1 CPU core errata.
12448 (This flag currently works around the SB-1 revision 2
12449 ``F1'' and ``F2'' floating point errata.)
12451 @item -mflush-func=@var{func}
12452 @itemx -mno-flush-func
12453 @opindex mflush-func
12454 Specifies the function to call to flush the I and D caches, or to not
12455 call any such function. If called, the function must take the same
12456 arguments as the common @code{_flush_func()}, that is, the address of the
12457 memory range for which the cache is being flushed, the size of the
12458 memory range, and the number 3 (to flush both caches). The default
12459 depends on the target GCC was configured for, but commonly is either
12460 @samp{_flush_func} or @samp{__cpu_flush}.
12462 @item mbranch-cost=@var{num}
12463 @opindex mbranch-cost
12464 Set the cost of branches to roughly @var{num} ``simple'' instructions.
12465 This cost is only a heuristic and is not guaranteed to produce
12466 consistent results across releases. A zero cost redundantly selects
12467 the default, which is based on the @option{-mtune} setting.
12469 @item -mbranch-likely
12470 @itemx -mno-branch-likely
12471 @opindex mbranch-likely
12472 @opindex mno-branch-likely
12473 Enable or disable use of Branch Likely instructions, regardless of the
12474 default for the selected architecture. By default, Branch Likely
12475 instructions may be generated if they are supported by the selected
12476 architecture. An exception is for the MIPS32 and MIPS64 architectures
12477 and processors which implement those architectures; for those, Branch
12478 Likely instructions will not be generated by default because the MIPS32
12479 and MIPS64 architectures specifically deprecate their use.
12481 @item -mfp-exceptions
12482 @itemx -mno-fp-exceptions
12483 @opindex mfp-exceptions
12484 Specifies whether FP exceptions are enabled. This affects how we schedule
12485 FP instructions for some processors. The default is that FP exceptions are
12488 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
12489 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
12492 @item -mvr4130-align
12493 @itemx -mno-vr4130-align
12494 @opindex mvr4130-align
12495 The VR4130 pipeline is two-way superscalar, but can only issue two
12496 instructions together if the first one is 8-byte aligned. When this
12497 option is enabled, GCC will align pairs of instructions that it
12498 thinks should execute in parallel.
12500 This option only has an effect when optimizing for the VR4130.
12501 It normally makes code faster, but at the expense of making it bigger.
12502 It is enabled by default at optimization level @option{-O3}.
12506 @subsection MMIX Options
12507 @cindex MMIX Options
12509 These options are defined for the MMIX:
12513 @itemx -mno-libfuncs
12515 @opindex mno-libfuncs
12516 Specify that intrinsic library functions are being compiled, passing all
12517 values in registers, no matter the size.
12520 @itemx -mno-epsilon
12522 @opindex mno-epsilon
12523 Generate floating-point comparison instructions that compare with respect
12524 to the @code{rE} epsilon register.
12526 @item -mabi=mmixware
12528 @opindex mabi-mmixware
12530 Generate code that passes function parameters and return values that (in
12531 the called function) are seen as registers @code{$0} and up, as opposed to
12532 the GNU ABI which uses global registers @code{$231} and up.
12534 @item -mzero-extend
12535 @itemx -mno-zero-extend
12536 @opindex mzero-extend
12537 @opindex mno-zero-extend
12538 When reading data from memory in sizes shorter than 64 bits, use (do not
12539 use) zero-extending load instructions by default, rather than
12540 sign-extending ones.
12543 @itemx -mno-knuthdiv
12545 @opindex mno-knuthdiv
12546 Make the result of a division yielding a remainder have the same sign as
12547 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
12548 remainder follows the sign of the dividend. Both methods are
12549 arithmetically valid, the latter being almost exclusively used.
12551 @item -mtoplevel-symbols
12552 @itemx -mno-toplevel-symbols
12553 @opindex mtoplevel-symbols
12554 @opindex mno-toplevel-symbols
12555 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
12556 code can be used with the @code{PREFIX} assembly directive.
12560 Generate an executable in the ELF format, rather than the default
12561 @samp{mmo} format used by the @command{mmix} simulator.
12563 @item -mbranch-predict
12564 @itemx -mno-branch-predict
12565 @opindex mbranch-predict
12566 @opindex mno-branch-predict
12567 Use (do not use) the probable-branch instructions, when static branch
12568 prediction indicates a probable branch.
12570 @item -mbase-addresses
12571 @itemx -mno-base-addresses
12572 @opindex mbase-addresses
12573 @opindex mno-base-addresses
12574 Generate (do not generate) code that uses @emph{base addresses}. Using a
12575 base address automatically generates a request (handled by the assembler
12576 and the linker) for a constant to be set up in a global register. The
12577 register is used for one or more base address requests within the range 0
12578 to 255 from the value held in the register. The generally leads to short
12579 and fast code, but the number of different data items that can be
12580 addressed is limited. This means that a program that uses lots of static
12581 data may require @option{-mno-base-addresses}.
12583 @item -msingle-exit
12584 @itemx -mno-single-exit
12585 @opindex msingle-exit
12586 @opindex mno-single-exit
12587 Force (do not force) generated code to have a single exit point in each
12591 @node MN10300 Options
12592 @subsection MN10300 Options
12593 @cindex MN10300 options
12595 These @option{-m} options are defined for Matsushita MN10300 architectures:
12600 Generate code to avoid bugs in the multiply instructions for the MN10300
12601 processors. This is the default.
12603 @item -mno-mult-bug
12604 @opindex mno-mult-bug
12605 Do not generate code to avoid bugs in the multiply instructions for the
12606 MN10300 processors.
12610 Generate code which uses features specific to the AM33 processor.
12614 Do not generate code which uses features specific to the AM33 processor. This
12617 @item -mreturn-pointer-on-d0
12618 @opindex mreturn-pointer-on-d0
12619 When generating a function which returns a pointer, return the pointer
12620 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
12621 only in a0, and attempts to call such functions without a prototype
12622 would result in errors. Note that this option is on by default; use
12623 @option{-mno-return-pointer-on-d0} to disable it.
12627 Do not link in the C run-time initialization object file.
12631 Indicate to the linker that it should perform a relaxation optimization pass
12632 to shorten branches, calls and absolute memory addresses. This option only
12633 has an effect when used on the command line for the final link step.
12635 This option makes symbolic debugging impossible.
12639 @subsection MT Options
12642 These @option{-m} options are defined for Morpho MT architectures:
12646 @item -march=@var{cpu-type}
12648 Generate code that will run on @var{cpu-type}, which is the name of a system
12649 representing a certain processor type. Possible values for
12650 @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002},
12651 @samp{ms1-16-003} and @samp{ms2}.
12653 When this option is not used, the default is @option{-march=ms1-16-002}.
12657 Use byte loads and stores when generating code.
12661 Do not use byte loads and stores when generating code.
12665 Use simulator runtime
12669 Do not link in the C run-time initialization object file
12670 @file{crti.o}. Other run-time initialization and termination files
12671 such as @file{startup.o} and @file{exit.o} are still included on the
12672 linker command line.
12676 @node PDP-11 Options
12677 @subsection PDP-11 Options
12678 @cindex PDP-11 Options
12680 These options are defined for the PDP-11:
12685 Use hardware FPP floating point. This is the default. (FIS floating
12686 point on the PDP-11/40 is not supported.)
12689 @opindex msoft-float
12690 Do not use hardware floating point.
12694 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
12698 Return floating-point results in memory. This is the default.
12702 Generate code for a PDP-11/40.
12706 Generate code for a PDP-11/45. This is the default.
12710 Generate code for a PDP-11/10.
12712 @item -mbcopy-builtin
12713 @opindex bcopy-builtin
12714 Use inline @code{movmemhi} patterns for copying memory. This is the
12719 Do not use inline @code{movmemhi} patterns for copying memory.
12725 Use 16-bit @code{int}. This is the default.
12731 Use 32-bit @code{int}.
12734 @itemx -mno-float32
12736 @opindex mno-float32
12737 Use 64-bit @code{float}. This is the default.
12740 @itemx -mno-float64
12742 @opindex mno-float64
12743 Use 32-bit @code{float}.
12747 Use @code{abshi2} pattern. This is the default.
12751 Do not use @code{abshi2} pattern.
12753 @item -mbranch-expensive
12754 @opindex mbranch-expensive
12755 Pretend that branches are expensive. This is for experimenting with
12756 code generation only.
12758 @item -mbranch-cheap
12759 @opindex mbranch-cheap
12760 Do not pretend that branches are expensive. This is the default.
12764 Generate code for a system with split I&D@.
12768 Generate code for a system without split I&D@. This is the default.
12772 Use Unix assembler syntax. This is the default when configured for
12773 @samp{pdp11-*-bsd}.
12777 Use DEC assembler syntax. This is the default when configured for any
12778 PDP-11 target other than @samp{pdp11-*-bsd}.
12781 @node PowerPC Options
12782 @subsection PowerPC Options
12783 @cindex PowerPC options
12785 These are listed under @xref{RS/6000 and PowerPC Options}.
12787 @node RS/6000 and PowerPC Options
12788 @subsection IBM RS/6000 and PowerPC Options
12789 @cindex RS/6000 and PowerPC Options
12790 @cindex IBM RS/6000 and PowerPC Options
12792 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
12799 @itemx -mno-powerpc
12800 @itemx -mpowerpc-gpopt
12801 @itemx -mno-powerpc-gpopt
12802 @itemx -mpowerpc-gfxopt
12803 @itemx -mno-powerpc-gfxopt
12805 @itemx -mno-powerpc64
12809 @itemx -mno-popcntb
12817 @itemx -mno-hard-dfp
12821 @opindex mno-power2
12823 @opindex mno-powerpc
12824 @opindex mpowerpc-gpopt
12825 @opindex mno-powerpc-gpopt
12826 @opindex mpowerpc-gfxopt
12827 @opindex mno-powerpc-gfxopt
12828 @opindex mpowerpc64
12829 @opindex mno-powerpc64
12833 @opindex mno-popcntb
12839 @opindex mno-mfpgpr
12841 @opindex mno-hard-dfp
12842 GCC supports two related instruction set architectures for the
12843 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
12844 instructions supported by the @samp{rios} chip set used in the original
12845 RS/6000 systems and the @dfn{PowerPC} instruction set is the
12846 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
12847 the IBM 4xx, 6xx, and follow-on microprocessors.
12849 Neither architecture is a subset of the other. However there is a
12850 large common subset of instructions supported by both. An MQ
12851 register is included in processors supporting the POWER architecture.
12853 You use these options to specify which instructions are available on the
12854 processor you are using. The default value of these options is
12855 determined when configuring GCC@. Specifying the
12856 @option{-mcpu=@var{cpu_type}} overrides the specification of these
12857 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
12858 rather than the options listed above.
12860 The @option{-mpower} option allows GCC to generate instructions that
12861 are found only in the POWER architecture and to use the MQ register.
12862 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
12863 to generate instructions that are present in the POWER2 architecture but
12864 not the original POWER architecture.
12866 The @option{-mpowerpc} option allows GCC to generate instructions that
12867 are found only in the 32-bit subset of the PowerPC architecture.
12868 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
12869 GCC to use the optional PowerPC architecture instructions in the
12870 General Purpose group, including floating-point square root. Specifying
12871 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
12872 use the optional PowerPC architecture instructions in the Graphics
12873 group, including floating-point select.
12875 The @option{-mmfcrf} option allows GCC to generate the move from
12876 condition register field instruction implemented on the POWER4
12877 processor and other processors that support the PowerPC V2.01
12879 The @option{-mpopcntb} option allows GCC to generate the popcount and
12880 double precision FP reciprocal estimate instruction implemented on the
12881 POWER5 processor and other processors that support the PowerPC V2.02
12883 The @option{-mfprnd} option allows GCC to generate the FP round to
12884 integer instructions implemented on the POWER5+ processor and other
12885 processors that support the PowerPC V2.03 architecture.
12886 The @option{-mcmpb} option allows GCC to generate the compare bytes
12887 instruction implemented on the POWER6 processor and other processors
12888 that support the PowerPC V2.05 architecture.
12889 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
12890 general purpose register instructions implemented on the POWER6X
12891 processor and other processors that support the extended PowerPC V2.05
12893 The @option{-mhard-dfp} option allows GCC to generate the decimal floating
12894 point instructions implemented on some POWER processors.
12896 The @option{-mpowerpc64} option allows GCC to generate the additional
12897 64-bit instructions that are found in the full PowerPC64 architecture
12898 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
12899 @option{-mno-powerpc64}.
12901 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
12902 will use only the instructions in the common subset of both
12903 architectures plus some special AIX common-mode calls, and will not use
12904 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
12905 permits GCC to use any instruction from either architecture and to
12906 allow use of the MQ register; specify this for the Motorola MPC601.
12908 @item -mnew-mnemonics
12909 @itemx -mold-mnemonics
12910 @opindex mnew-mnemonics
12911 @opindex mold-mnemonics
12912 Select which mnemonics to use in the generated assembler code. With
12913 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
12914 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
12915 assembler mnemonics defined for the POWER architecture. Instructions
12916 defined in only one architecture have only one mnemonic; GCC uses that
12917 mnemonic irrespective of which of these options is specified.
12919 GCC defaults to the mnemonics appropriate for the architecture in
12920 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
12921 value of these option. Unless you are building a cross-compiler, you
12922 should normally not specify either @option{-mnew-mnemonics} or
12923 @option{-mold-mnemonics}, but should instead accept the default.
12925 @item -mcpu=@var{cpu_type}
12927 Set architecture type, register usage, choice of mnemonics, and
12928 instruction scheduling parameters for machine type @var{cpu_type}.
12929 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
12930 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
12931 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
12932 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
12933 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
12934 @samp{860}, @samp{970}, @samp{8540}, @samp{e300c2}, @samp{e300c3},
12935 @samp{ec603e}, @samp{G3}, @samp{G4}, @samp{G5}, @samp{power},
12936 @samp{power2}, @samp{power3}, @samp{power4}, @samp{power5},
12937 @samp{power5+}, @samp{power6}, @samp{power6x}, @samp{common},
12938 @samp{powerpc}, @samp{powerpc64}, @samp{rios}, @samp{rios1},
12939 @samp{rios2}, @samp{rsc}, and @samp{rs64}.
12941 @option{-mcpu=common} selects a completely generic processor. Code
12942 generated under this option will run on any POWER or PowerPC processor.
12943 GCC will use only the instructions in the common subset of both
12944 architectures, and will not use the MQ register. GCC assumes a generic
12945 processor model for scheduling purposes.
12947 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
12948 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
12949 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
12950 types, with an appropriate, generic processor model assumed for
12951 scheduling purposes.
12953 The other options specify a specific processor. Code generated under
12954 those options will run best on that processor, and may not run at all on
12957 The @option{-mcpu} options automatically enable or disable the
12960 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
12961 -mnew-mnemonics -mpopcntb -mpower -mpower2 -mpowerpc64 @gol
12962 -mpowerpc-gpopt -mpowerpc-gfxopt -mstring -mmulhw -mdlmzb -mmfpgpr}
12964 The particular options set for any particular CPU will vary between
12965 compiler versions, depending on what setting seems to produce optimal
12966 code for that CPU; it doesn't necessarily reflect the actual hardware's
12967 capabilities. If you wish to set an individual option to a particular
12968 value, you may specify it after the @option{-mcpu} option, like
12969 @samp{-mcpu=970 -mno-altivec}.
12971 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
12972 not enabled or disabled by the @option{-mcpu} option at present because
12973 AIX does not have full support for these options. You may still
12974 enable or disable them individually if you're sure it'll work in your
12977 @item -mtune=@var{cpu_type}
12979 Set the instruction scheduling parameters for machine type
12980 @var{cpu_type}, but do not set the architecture type, register usage, or
12981 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
12982 values for @var{cpu_type} are used for @option{-mtune} as for
12983 @option{-mcpu}. If both are specified, the code generated will use the
12984 architecture, registers, and mnemonics set by @option{-mcpu}, but the
12985 scheduling parameters set by @option{-mtune}.
12991 Generate code to compute division as reciprocal estimate and iterative
12992 refinement, creating opportunities for increased throughput. This
12993 feature requires: optional PowerPC Graphics instruction set for single
12994 precision and FRE instruction for double precision, assuming divides
12995 cannot generate user-visible traps, and the domain values not include
12996 Infinities, denormals or zero denominator.
12999 @itemx -mno-altivec
13001 @opindex mno-altivec
13002 Generate code that uses (does not use) AltiVec instructions, and also
13003 enable the use of built-in functions that allow more direct access to
13004 the AltiVec instruction set. You may also need to set
13005 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
13011 @opindex mno-vrsave
13012 Generate VRSAVE instructions when generating AltiVec code.
13015 @opindex msecure-plt
13016 Generate code that allows ld and ld.so to build executables and shared
13017 libraries with non-exec .plt and .got sections. This is a PowerPC
13018 32-bit SYSV ABI option.
13022 Generate code that uses a BSS .plt section that ld.so fills in, and
13023 requires .plt and .got sections that are both writable and executable.
13024 This is a PowerPC 32-bit SYSV ABI option.
13030 This switch enables or disables the generation of ISEL instructions.
13032 @item -misel=@var{yes/no}
13033 This switch has been deprecated. Use @option{-misel} and
13034 @option{-mno-isel} instead.
13040 This switch enables or disables the generation of SPE simd
13046 @opindex mno-paired
13047 This switch enables or disables the generation of PAIRED simd
13050 @item -mspe=@var{yes/no}
13051 This option has been deprecated. Use @option{-mspe} and
13052 @option{-mno-spe} instead.
13054 @item -mfloat-gprs=@var{yes/single/double/no}
13055 @itemx -mfloat-gprs
13056 @opindex mfloat-gprs
13057 This switch enables or disables the generation of floating point
13058 operations on the general purpose registers for architectures that
13061 The argument @var{yes} or @var{single} enables the use of
13062 single-precision floating point operations.
13064 The argument @var{double} enables the use of single and
13065 double-precision floating point operations.
13067 The argument @var{no} disables floating point operations on the
13068 general purpose registers.
13070 This option is currently only available on the MPC854x.
13076 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
13077 targets (including GNU/Linux). The 32-bit environment sets int, long
13078 and pointer to 32 bits and generates code that runs on any PowerPC
13079 variant. The 64-bit environment sets int to 32 bits and long and
13080 pointer to 64 bits, and generates code for PowerPC64, as for
13081 @option{-mpowerpc64}.
13084 @itemx -mno-fp-in-toc
13085 @itemx -mno-sum-in-toc
13086 @itemx -mminimal-toc
13088 @opindex mno-fp-in-toc
13089 @opindex mno-sum-in-toc
13090 @opindex mminimal-toc
13091 Modify generation of the TOC (Table Of Contents), which is created for
13092 every executable file. The @option{-mfull-toc} option is selected by
13093 default. In that case, GCC will allocate at least one TOC entry for
13094 each unique non-automatic variable reference in your program. GCC
13095 will also place floating-point constants in the TOC@. However, only
13096 16,384 entries are available in the TOC@.
13098 If you receive a linker error message that saying you have overflowed
13099 the available TOC space, you can reduce the amount of TOC space used
13100 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
13101 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
13102 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
13103 generate code to calculate the sum of an address and a constant at
13104 run-time instead of putting that sum into the TOC@. You may specify one
13105 or both of these options. Each causes GCC to produce very slightly
13106 slower and larger code at the expense of conserving TOC space.
13108 If you still run out of space in the TOC even when you specify both of
13109 these options, specify @option{-mminimal-toc} instead. This option causes
13110 GCC to make only one TOC entry for every file. When you specify this
13111 option, GCC will produce code that is slower and larger but which
13112 uses extremely little TOC space. You may wish to use this option
13113 only on files that contain less frequently executed code.
13119 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
13120 @code{long} type, and the infrastructure needed to support them.
13121 Specifying @option{-maix64} implies @option{-mpowerpc64} and
13122 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
13123 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
13126 @itemx -mno-xl-compat
13127 @opindex mxl-compat
13128 @opindex mno-xl-compat
13129 Produce code that conforms more closely to IBM XL compiler semantics
13130 when using AIX-compatible ABI@. Pass floating-point arguments to
13131 prototyped functions beyond the register save area (RSA) on the stack
13132 in addition to argument FPRs. Do not assume that most significant
13133 double in 128-bit long double value is properly rounded when comparing
13134 values and converting to double. Use XL symbol names for long double
13137 The AIX calling convention was extended but not initially documented to
13138 handle an obscure K&R C case of calling a function that takes the
13139 address of its arguments with fewer arguments than declared. IBM XL
13140 compilers access floating point arguments which do not fit in the
13141 RSA from the stack when a subroutine is compiled without
13142 optimization. Because always storing floating-point arguments on the
13143 stack is inefficient and rarely needed, this option is not enabled by
13144 default and only is necessary when calling subroutines compiled by IBM
13145 XL compilers without optimization.
13149 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
13150 application written to use message passing with special startup code to
13151 enable the application to run. The system must have PE installed in the
13152 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
13153 must be overridden with the @option{-specs=} option to specify the
13154 appropriate directory location. The Parallel Environment does not
13155 support threads, so the @option{-mpe} option and the @option{-pthread}
13156 option are incompatible.
13158 @item -malign-natural
13159 @itemx -malign-power
13160 @opindex malign-natural
13161 @opindex malign-power
13162 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
13163 @option{-malign-natural} overrides the ABI-defined alignment of larger
13164 types, such as floating-point doubles, on their natural size-based boundary.
13165 The option @option{-malign-power} instructs GCC to follow the ABI-specified
13166 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
13168 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
13172 @itemx -mhard-float
13173 @opindex msoft-float
13174 @opindex mhard-float
13175 Generate code that does not use (uses) the floating-point register set.
13176 Software floating point emulation is provided if you use the
13177 @option{-msoft-float} option, and pass the option to GCC when linking.
13180 @itemx -mno-multiple
13182 @opindex mno-multiple
13183 Generate code that uses (does not use) the load multiple word
13184 instructions and the store multiple word instructions. These
13185 instructions are generated by default on POWER systems, and not
13186 generated on PowerPC systems. Do not use @option{-mmultiple} on little
13187 endian PowerPC systems, since those instructions do not work when the
13188 processor is in little endian mode. The exceptions are PPC740 and
13189 PPC750 which permit the instructions usage in little endian mode.
13194 @opindex mno-string
13195 Generate code that uses (does not use) the load string instructions
13196 and the store string word instructions to save multiple registers and
13197 do small block moves. These instructions are generated by default on
13198 POWER systems, and not generated on PowerPC systems. Do not use
13199 @option{-mstring} on little endian PowerPC systems, since those
13200 instructions do not work when the processor is in little endian mode.
13201 The exceptions are PPC740 and PPC750 which permit the instructions
13202 usage in little endian mode.
13207 @opindex mno-update
13208 Generate code that uses (does not use) the load or store instructions
13209 that update the base register to the address of the calculated memory
13210 location. These instructions are generated by default. If you use
13211 @option{-mno-update}, there is a small window between the time that the
13212 stack pointer is updated and the address of the previous frame is
13213 stored, which means code that walks the stack frame across interrupts or
13214 signals may get corrupted data.
13217 @itemx -mno-fused-madd
13218 @opindex mfused-madd
13219 @opindex mno-fused-madd
13220 Generate code that uses (does not use) the floating point multiply and
13221 accumulate instructions. These instructions are generated by default if
13222 hardware floating is used.
13228 Generate code that uses (does not use) the half-word multiply and
13229 multiply-accumulate instructions on the IBM 405 and 440 processors.
13230 These instructions are generated by default when targetting those
13237 Generate code that uses (does not use) the string-search @samp{dlmzb}
13238 instruction on the IBM 405 and 440 processors. This instruction is
13239 generated by default when targetting those processors.
13241 @item -mno-bit-align
13243 @opindex mno-bit-align
13244 @opindex mbit-align
13245 On System V.4 and embedded PowerPC systems do not (do) force structures
13246 and unions that contain bit-fields to be aligned to the base type of the
13249 For example, by default a structure containing nothing but 8
13250 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
13251 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
13252 the structure would be aligned to a 1 byte boundary and be one byte in
13255 @item -mno-strict-align
13256 @itemx -mstrict-align
13257 @opindex mno-strict-align
13258 @opindex mstrict-align
13259 On System V.4 and embedded PowerPC systems do not (do) assume that
13260 unaligned memory references will be handled by the system.
13262 @item -mrelocatable
13263 @itemx -mno-relocatable
13264 @opindex mrelocatable
13265 @opindex mno-relocatable
13266 On embedded PowerPC systems generate code that allows (does not allow)
13267 the program to be relocated to a different address at runtime. If you
13268 use @option{-mrelocatable} on any module, all objects linked together must
13269 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
13271 @item -mrelocatable-lib
13272 @itemx -mno-relocatable-lib
13273 @opindex mrelocatable-lib
13274 @opindex mno-relocatable-lib
13275 On embedded PowerPC systems generate code that allows (does not allow)
13276 the program to be relocated to a different address at runtime. Modules
13277 compiled with @option{-mrelocatable-lib} can be linked with either modules
13278 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
13279 with modules compiled with the @option{-mrelocatable} options.
13285 On System V.4 and embedded PowerPC systems do not (do) assume that
13286 register 2 contains a pointer to a global area pointing to the addresses
13287 used in the program.
13290 @itemx -mlittle-endian
13292 @opindex mlittle-endian
13293 On System V.4 and embedded PowerPC systems compile code for the
13294 processor in little endian mode. The @option{-mlittle-endian} option is
13295 the same as @option{-mlittle}.
13298 @itemx -mbig-endian
13300 @opindex mbig-endian
13301 On System V.4 and embedded PowerPC systems compile code for the
13302 processor in big endian mode. The @option{-mbig-endian} option is
13303 the same as @option{-mbig}.
13305 @item -mdynamic-no-pic
13306 @opindex mdynamic-no-pic
13307 On Darwin and Mac OS X systems, compile code so that it is not
13308 relocatable, but that its external references are relocatable. The
13309 resulting code is suitable for applications, but not shared
13312 @item -mprioritize-restricted-insns=@var{priority}
13313 @opindex mprioritize-restricted-insns
13314 This option controls the priority that is assigned to
13315 dispatch-slot restricted instructions during the second scheduling
13316 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
13317 @var{no/highest/second-highest} priority to dispatch slot restricted
13320 @item -msched-costly-dep=@var{dependence_type}
13321 @opindex msched-costly-dep
13322 This option controls which dependences are considered costly
13323 by the target during instruction scheduling. The argument
13324 @var{dependence_type} takes one of the following values:
13325 @var{no}: no dependence is costly,
13326 @var{all}: all dependences are costly,
13327 @var{true_store_to_load}: a true dependence from store to load is costly,
13328 @var{store_to_load}: any dependence from store to load is costly,
13329 @var{number}: any dependence which latency >= @var{number} is costly.
13331 @item -minsert-sched-nops=@var{scheme}
13332 @opindex minsert-sched-nops
13333 This option controls which nop insertion scheme will be used during
13334 the second scheduling pass. The argument @var{scheme} takes one of the
13336 @var{no}: Don't insert nops.
13337 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
13338 according to the scheduler's grouping.
13339 @var{regroup_exact}: Insert nops to force costly dependent insns into
13340 separate groups. Insert exactly as many nops as needed to force an insn
13341 to a new group, according to the estimated processor grouping.
13342 @var{number}: Insert nops to force costly dependent insns into
13343 separate groups. Insert @var{number} nops to force an insn to a new group.
13346 @opindex mcall-sysv
13347 On System V.4 and embedded PowerPC systems compile code using calling
13348 conventions that adheres to the March 1995 draft of the System V
13349 Application Binary Interface, PowerPC processor supplement. This is the
13350 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
13352 @item -mcall-sysv-eabi
13353 @opindex mcall-sysv-eabi
13354 Specify both @option{-mcall-sysv} and @option{-meabi} options.
13356 @item -mcall-sysv-noeabi
13357 @opindex mcall-sysv-noeabi
13358 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
13360 @item -mcall-solaris
13361 @opindex mcall-solaris
13362 On System V.4 and embedded PowerPC systems compile code for the Solaris
13366 @opindex mcall-linux
13367 On System V.4 and embedded PowerPC systems compile code for the
13368 Linux-based GNU system.
13372 On System V.4 and embedded PowerPC systems compile code for the
13373 Hurd-based GNU system.
13375 @item -mcall-netbsd
13376 @opindex mcall-netbsd
13377 On System V.4 and embedded PowerPC systems compile code for the
13378 NetBSD operating system.
13380 @item -maix-struct-return
13381 @opindex maix-struct-return
13382 Return all structures in memory (as specified by the AIX ABI)@.
13384 @item -msvr4-struct-return
13385 @opindex msvr4-struct-return
13386 Return structures smaller than 8 bytes in registers (as specified by the
13389 @item -mabi=@var{abi-type}
13391 Extend the current ABI with a particular extension, or remove such extension.
13392 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
13393 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
13397 Extend the current ABI with SPE ABI extensions. This does not change
13398 the default ABI, instead it adds the SPE ABI extensions to the current
13402 @opindex mabi=no-spe
13403 Disable Booke SPE ABI extensions for the current ABI@.
13405 @item -mabi=ibmlongdouble
13406 @opindex mabi=ibmlongdouble
13407 Change the current ABI to use IBM extended precision long double.
13408 This is a PowerPC 32-bit SYSV ABI option.
13410 @item -mabi=ieeelongdouble
13411 @opindex mabi=ieeelongdouble
13412 Change the current ABI to use IEEE extended precision long double.
13413 This is a PowerPC 32-bit Linux ABI option.
13416 @itemx -mno-prototype
13417 @opindex mprototype
13418 @opindex mno-prototype
13419 On System V.4 and embedded PowerPC systems assume that all calls to
13420 variable argument functions are properly prototyped. Otherwise, the
13421 compiler must insert an instruction before every non prototyped call to
13422 set or clear bit 6 of the condition code register (@var{CR}) to
13423 indicate whether floating point values were passed in the floating point
13424 registers in case the function takes a variable arguments. With
13425 @option{-mprototype}, only calls to prototyped variable argument functions
13426 will set or clear the bit.
13430 On embedded PowerPC systems, assume that the startup module is called
13431 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
13432 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
13437 On embedded PowerPC systems, assume that the startup module is called
13438 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
13443 On embedded PowerPC systems, assume that the startup module is called
13444 @file{crt0.o} and the standard C libraries are @file{libads.a} and
13447 @item -myellowknife
13448 @opindex myellowknife
13449 On embedded PowerPC systems, assume that the startup module is called
13450 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
13455 On System V.4 and embedded PowerPC systems, specify that you are
13456 compiling for a VxWorks system.
13460 Specify that you are compiling for the WindISS simulation environment.
13464 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
13465 header to indicate that @samp{eabi} extended relocations are used.
13471 On System V.4 and embedded PowerPC systems do (do not) adhere to the
13472 Embedded Applications Binary Interface (eabi) which is a set of
13473 modifications to the System V.4 specifications. Selecting @option{-meabi}
13474 means that the stack is aligned to an 8 byte boundary, a function
13475 @code{__eabi} is called to from @code{main} to set up the eabi
13476 environment, and the @option{-msdata} option can use both @code{r2} and
13477 @code{r13} to point to two separate small data areas. Selecting
13478 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
13479 do not call an initialization function from @code{main}, and the
13480 @option{-msdata} option will only use @code{r13} to point to a single
13481 small data area. The @option{-meabi} option is on by default if you
13482 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
13485 @opindex msdata=eabi
13486 On System V.4 and embedded PowerPC systems, put small initialized
13487 @code{const} global and static data in the @samp{.sdata2} section, which
13488 is pointed to by register @code{r2}. Put small initialized
13489 non-@code{const} global and static data in the @samp{.sdata} section,
13490 which is pointed to by register @code{r13}. Put small uninitialized
13491 global and static data in the @samp{.sbss} section, which is adjacent to
13492 the @samp{.sdata} section. The @option{-msdata=eabi} option is
13493 incompatible with the @option{-mrelocatable} option. The
13494 @option{-msdata=eabi} option also sets the @option{-memb} option.
13497 @opindex msdata=sysv
13498 On System V.4 and embedded PowerPC systems, put small global and static
13499 data in the @samp{.sdata} section, which is pointed to by register
13500 @code{r13}. Put small uninitialized global and static data in the
13501 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
13502 The @option{-msdata=sysv} option is incompatible with the
13503 @option{-mrelocatable} option.
13505 @item -msdata=default
13507 @opindex msdata=default
13509 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
13510 compile code the same as @option{-msdata=eabi}, otherwise compile code the
13511 same as @option{-msdata=sysv}.
13514 @opindex msdata-data
13515 On System V.4 and embedded PowerPC systems, put small global
13516 data in the @samp{.sdata} section. Put small uninitialized global
13517 data in the @samp{.sbss} section. Do not use register @code{r13}
13518 to address small data however. This is the default behavior unless
13519 other @option{-msdata} options are used.
13523 @opindex msdata=none
13525 On embedded PowerPC systems, put all initialized global and static data
13526 in the @samp{.data} section, and all uninitialized data in the
13527 @samp{.bss} section.
13531 @cindex smaller data references (PowerPC)
13532 @cindex .sdata/.sdata2 references (PowerPC)
13533 On embedded PowerPC systems, put global and static items less than or
13534 equal to @var{num} bytes into the small data or bss sections instead of
13535 the normal data or bss section. By default, @var{num} is 8. The
13536 @option{-G @var{num}} switch is also passed to the linker.
13537 All modules should be compiled with the same @option{-G @var{num}} value.
13540 @itemx -mno-regnames
13542 @opindex mno-regnames
13543 On System V.4 and embedded PowerPC systems do (do not) emit register
13544 names in the assembly language output using symbolic forms.
13547 @itemx -mno-longcall
13549 @opindex mno-longcall
13550 By default assume that all calls are far away so that a longer more
13551 expensive calling sequence is required. This is required for calls
13552 further than 32 megabytes (33,554,432 bytes) from the current location.
13553 A short call will be generated if the compiler knows
13554 the call cannot be that far away. This setting can be overridden by
13555 the @code{shortcall} function attribute, or by @code{#pragma
13558 Some linkers are capable of detecting out-of-range calls and generating
13559 glue code on the fly. On these systems, long calls are unnecessary and
13560 generate slower code. As of this writing, the AIX linker can do this,
13561 as can the GNU linker for PowerPC/64. It is planned to add this feature
13562 to the GNU linker for 32-bit PowerPC systems as well.
13564 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
13565 callee, L42'', plus a ``branch island'' (glue code). The two target
13566 addresses represent the callee and the ``branch island''. The
13567 Darwin/PPC linker will prefer the first address and generate a ``bl
13568 callee'' if the PPC ``bl'' instruction will reach the callee directly;
13569 otherwise, the linker will generate ``bl L42'' to call the ``branch
13570 island''. The ``branch island'' is appended to the body of the
13571 calling function; it computes the full 32-bit address of the callee
13574 On Mach-O (Darwin) systems, this option directs the compiler emit to
13575 the glue for every direct call, and the Darwin linker decides whether
13576 to use or discard it.
13578 In the future, we may cause GCC to ignore all longcall specifications
13579 when the linker is known to generate glue.
13583 Adds support for multithreading with the @dfn{pthreads} library.
13584 This option sets flags for both the preprocessor and linker.
13588 @node S/390 and zSeries Options
13589 @subsection S/390 and zSeries Options
13590 @cindex S/390 and zSeries Options
13592 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
13596 @itemx -msoft-float
13597 @opindex mhard-float
13598 @opindex msoft-float
13599 Use (do not use) the hardware floating-point instructions and registers
13600 for floating-point operations. When @option{-msoft-float} is specified,
13601 functions in @file{libgcc.a} will be used to perform floating-point
13602 operations. When @option{-mhard-float} is specified, the compiler
13603 generates IEEE floating-point instructions. This is the default.
13605 @item -mlong-double-64
13606 @itemx -mlong-double-128
13607 @opindex mlong-double-64
13608 @opindex mlong-double-128
13609 These switches control the size of @code{long double} type. A size
13610 of 64bit makes the @code{long double} type equivalent to the @code{double}
13611 type. This is the default.
13614 @itemx -mno-backchain
13615 @opindex mbackchain
13616 @opindex mno-backchain
13617 Store (do not store) the address of the caller's frame as backchain pointer
13618 into the callee's stack frame.
13619 A backchain may be needed to allow debugging using tools that do not understand
13620 DWARF-2 call frame information.
13621 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
13622 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
13623 the backchain is placed into the topmost word of the 96/160 byte register
13626 In general, code compiled with @option{-mbackchain} is call-compatible with
13627 code compiled with @option{-mmo-backchain}; however, use of the backchain
13628 for debugging purposes usually requires that the whole binary is built with
13629 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
13630 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13631 to build a linux kernel use @option{-msoft-float}.
13633 The default is to not maintain the backchain.
13635 @item -mpacked-stack
13636 @item -mno-packed-stack
13637 @opindex mpacked-stack
13638 @opindex mno-packed-stack
13639 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
13640 specified, the compiler uses the all fields of the 96/160 byte register save
13641 area only for their default purpose; unused fields still take up stack space.
13642 When @option{-mpacked-stack} is specified, register save slots are densely
13643 packed at the top of the register save area; unused space is reused for other
13644 purposes, allowing for more efficient use of the available stack space.
13645 However, when @option{-mbackchain} is also in effect, the topmost word of
13646 the save area is always used to store the backchain, and the return address
13647 register is always saved two words below the backchain.
13649 As long as the stack frame backchain is not used, code generated with
13650 @option{-mpacked-stack} is call-compatible with code generated with
13651 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
13652 S/390 or zSeries generated code that uses the stack frame backchain at run
13653 time, not just for debugging purposes. Such code is not call-compatible
13654 with code compiled with @option{-mpacked-stack}. Also, note that the
13655 combination of @option{-mbackchain},
13656 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
13657 to build a linux kernel use @option{-msoft-float}.
13659 The default is to not use the packed stack layout.
13662 @itemx -mno-small-exec
13663 @opindex msmall-exec
13664 @opindex mno-small-exec
13665 Generate (or do not generate) code using the @code{bras} instruction
13666 to do subroutine calls.
13667 This only works reliably if the total executable size does not
13668 exceed 64k. The default is to use the @code{basr} instruction instead,
13669 which does not have this limitation.
13675 When @option{-m31} is specified, generate code compliant to the
13676 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
13677 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
13678 particular to generate 64-bit instructions. For the @samp{s390}
13679 targets, the default is @option{-m31}, while the @samp{s390x}
13680 targets default to @option{-m64}.
13686 When @option{-mzarch} is specified, generate code using the
13687 instructions available on z/Architecture.
13688 When @option{-mesa} is specified, generate code using the
13689 instructions available on ESA/390. Note that @option{-mesa} is
13690 not possible with @option{-m64}.
13691 When generating code compliant to the GNU/Linux for S/390 ABI,
13692 the default is @option{-mesa}. When generating code compliant
13693 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
13699 Generate (or do not generate) code using the @code{mvcle} instruction
13700 to perform block moves. When @option{-mno-mvcle} is specified,
13701 use a @code{mvc} loop instead. This is the default unless optimizing for
13708 Print (or do not print) additional debug information when compiling.
13709 The default is to not print debug information.
13711 @item -march=@var{cpu-type}
13713 Generate code that will run on @var{cpu-type}, which is the name of a system
13714 representing a certain processor type. Possible values for
13715 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
13716 When generating code using the instructions available on z/Architecture,
13717 the default is @option{-march=z900}. Otherwise, the default is
13718 @option{-march=g5}.
13720 @item -mtune=@var{cpu-type}
13722 Tune to @var{cpu-type} everything applicable about the generated code,
13723 except for the ABI and the set of available instructions.
13724 The list of @var{cpu-type} values is the same as for @option{-march}.
13725 The default is the value used for @option{-march}.
13728 @itemx -mno-tpf-trace
13729 @opindex mtpf-trace
13730 @opindex mno-tpf-trace
13731 Generate code that adds (does not add) in TPF OS specific branches to trace
13732 routines in the operating system. This option is off by default, even
13733 when compiling for the TPF OS@.
13736 @itemx -mno-fused-madd
13737 @opindex mfused-madd
13738 @opindex mno-fused-madd
13739 Generate code that uses (does not use) the floating point multiply and
13740 accumulate instructions. These instructions are generated by default if
13741 hardware floating point is used.
13743 @item -mwarn-framesize=@var{framesize}
13744 @opindex mwarn-framesize
13745 Emit a warning if the current function exceeds the given frame size. Because
13746 this is a compile time check it doesn't need to be a real problem when the program
13747 runs. It is intended to identify functions which most probably cause
13748 a stack overflow. It is useful to be used in an environment with limited stack
13749 size e.g.@: the linux kernel.
13751 @item -mwarn-dynamicstack
13752 @opindex mwarn-dynamicstack
13753 Emit a warning if the function calls alloca or uses dynamically
13754 sized arrays. This is generally a bad idea with a limited stack size.
13756 @item -mstack-guard=@var{stack-guard}
13757 @item -mstack-size=@var{stack-size}
13758 @opindex mstack-guard
13759 @opindex mstack-size
13760 If these options are provided the s390 back end emits additional instructions in
13761 the function prologue which trigger a trap if the stack size is @var{stack-guard}
13762 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
13763 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
13764 the frame size of the compiled function is chosen.
13765 These options are intended to be used to help debugging stack overflow problems.
13766 The additionally emitted code causes only little overhead and hence can also be
13767 used in production like systems without greater performance degradation. The given
13768 values have to be exact powers of 2 and @var{stack-size} has to be greater than
13769 @var{stack-guard} without exceeding 64k.
13770 In order to be efficient the extra code makes the assumption that the stack starts
13771 at an address aligned to the value given by @var{stack-size}.
13772 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
13775 @node Score Options
13776 @subsection Score Options
13777 @cindex Score Options
13779 These options are defined for Score implementations:
13784 Compile code for big endian mode. This is the default.
13788 Compile code for little endian mode.
13792 Disable generate bcnz instruction.
13796 Enable generate unaligned load and store instruction.
13800 Enable the use of multiply-accumulate instructions. Disabled by default.
13804 Specify the SCORE5 as the target architecture.
13808 Specify the SCORE5U of the target architecture.
13812 Specify the SCORE7 as the target architecture. This is the default.
13816 Specify the SCORE7D as the target architecture.
13820 @subsection SH Options
13822 These @samp{-m} options are defined for the SH implementations:
13827 Generate code for the SH1.
13831 Generate code for the SH2.
13834 Generate code for the SH2e.
13838 Generate code for the SH3.
13842 Generate code for the SH3e.
13846 Generate code for the SH4 without a floating-point unit.
13848 @item -m4-single-only
13849 @opindex m4-single-only
13850 Generate code for the SH4 with a floating-point unit that only
13851 supports single-precision arithmetic.
13855 Generate code for the SH4 assuming the floating-point unit is in
13856 single-precision mode by default.
13860 Generate code for the SH4.
13864 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
13865 floating-point unit is not used.
13867 @item -m4a-single-only
13868 @opindex m4a-single-only
13869 Generate code for the SH4a, in such a way that no double-precision
13870 floating point operations are used.
13873 @opindex m4a-single
13874 Generate code for the SH4a assuming the floating-point unit is in
13875 single-precision mode by default.
13879 Generate code for the SH4a.
13883 Same as @option{-m4a-nofpu}, except that it implicitly passes
13884 @option{-dsp} to the assembler. GCC doesn't generate any DSP
13885 instructions at the moment.
13889 Compile code for the processor in big endian mode.
13893 Compile code for the processor in little endian mode.
13897 Align doubles at 64-bit boundaries. Note that this changes the calling
13898 conventions, and thus some functions from the standard C library will
13899 not work unless you recompile it first with @option{-mdalign}.
13903 Shorten some address references at link time, when possible; uses the
13904 linker option @option{-relax}.
13908 Use 32-bit offsets in @code{switch} tables. The default is to use
13913 Enable the use of the instruction @code{fmovd}.
13917 Comply with the calling conventions defined by Renesas.
13921 Comply with the calling conventions defined by Renesas.
13925 Comply with the calling conventions defined for GCC before the Renesas
13926 conventions were available. This option is the default for all
13927 targets of the SH toolchain except for @samp{sh-symbianelf}.
13930 @opindex mnomacsave
13931 Mark the @code{MAC} register as call-clobbered, even if
13932 @option{-mhitachi} is given.
13936 Increase IEEE-compliance of floating-point code.
13937 At the moment, this is equivalent to @option{-fno-finite-math-only}.
13938 When generating 16 bit SH opcodes, getting IEEE-conforming results for
13939 comparisons of NANs / infinities incurs extra overhead in every
13940 floating point comparison, therefore the default is set to
13941 @option{-ffinite-math-only}.
13943 @item -minline-ic_invalidate
13944 @opindex minline-ic_invalidate
13945 Inline code to invalidate instruction cache entries after setting up
13946 nested function trampolines.
13947 This option has no effect if -musermode is in effect and the selected
13948 code generation option (e.g. -m4) does not allow the use of the icbi
13950 If the selected code generation option does not allow the use of the icbi
13951 instruction, and -musermode is not in effect, the inlined code will
13952 manipulate the instruction cache address array directly with an associative
13953 write. This not only requires privileged mode, but it will also
13954 fail if the cache line had been mapped via the TLB and has become unmapped.
13958 Dump instruction size and location in the assembly code.
13961 @opindex mpadstruct
13962 This option is deprecated. It pads structures to multiple of 4 bytes,
13963 which is incompatible with the SH ABI@.
13967 Optimize for space instead of speed. Implied by @option{-Os}.
13970 @opindex mprefergot
13971 When generating position-independent code, emit function calls using
13972 the Global Offset Table instead of the Procedure Linkage Table.
13976 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
13977 if the inlined code would not work in user mode.
13978 This is the default when the target is @code{sh-*-linux*}.
13980 @item -multcost=@var{number}
13981 @opindex multcost=@var{number}
13982 Set the cost to assume for a multiply insn.
13984 @item -mdiv=@var{strategy}
13985 @opindex mdiv=@var{strategy}
13986 Set the division strategy to use for SHmedia code. @var{strategy} must be
13987 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
13988 inv:call2, inv:fp .
13989 "fp" performs the operation in floating point. This has a very high latency,
13990 but needs only a few instructions, so it might be a good choice if
13991 your code has enough easily exploitable ILP to allow the compiler to
13992 schedule the floating point instructions together with other instructions.
13993 Division by zero causes a floating point exception.
13994 "inv" uses integer operations to calculate the inverse of the divisor,
13995 and then multiplies the dividend with the inverse. This strategy allows
13996 cse and hoisting of the inverse calculation. Division by zero calculates
13997 an unspecified result, but does not trap.
13998 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
13999 have been found, or if the entire operation has been hoisted to the same
14000 place, the last stages of the inverse calculation are intertwined with the
14001 final multiply to reduce the overall latency, at the expense of using a few
14002 more instructions, and thus offering fewer scheduling opportunities with
14004 "call" calls a library function that usually implements the inv:minlat
14006 This gives high code density for m5-*media-nofpu compilations.
14007 "call2" uses a different entry point of the same library function, where it
14008 assumes that a pointer to a lookup table has already been set up, which
14009 exposes the pointer load to cse / code hoisting optimizations.
14010 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
14011 code generation, but if the code stays unoptimized, revert to the "call",
14012 "call2", or "fp" strategies, respectively. Note that the
14013 potentially-trapping side effect of division by zero is carried by a
14014 separate instruction, so it is possible that all the integer instructions
14015 are hoisted out, but the marker for the side effect stays where it is.
14016 A recombination to fp operations or a call is not possible in that case.
14017 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
14018 that the inverse calculation was nor separated from the multiply, they speed
14019 up division where the dividend fits into 20 bits (plus sign where applicable),
14020 by inserting a test to skip a number of operations in this case; this test
14021 slows down the case of larger dividends. inv20u assumes the case of a such
14022 a small dividend to be unlikely, and inv20l assumes it to be likely.
14024 @item -mdivsi3_libfunc=@var{name}
14025 @opindex mdivsi3_libfunc=@var{name}
14026 Set the name of the library function used for 32 bit signed division to
14027 @var{name}. This only affect the name used in the call and inv:call
14028 division strategies, and the compiler will still expect the same
14029 sets of input/output/clobbered registers as if this option was not present.
14031 @item -madjust-unroll
14032 @opindex madjust-unroll
14033 Throttle unrolling to avoid thrashing target registers.
14034 This option only has an effect if the gcc code base supports the
14035 TARGET_ADJUST_UNROLL_MAX target hook.
14037 @item -mindexed-addressing
14038 @opindex mindexed-addressing
14039 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
14040 This is only safe if the hardware and/or OS implement 32 bit wrap-around
14041 semantics for the indexed addressing mode. The architecture allows the
14042 implementation of processors with 64 bit MMU, which the OS could use to
14043 get 32 bit addressing, but since no current hardware implementation supports
14044 this or any other way to make the indexed addressing mode safe to use in
14045 the 32 bit ABI, the default is -mno-indexed-addressing.
14047 @item -mgettrcost=@var{number}
14048 @opindex mgettrcost=@var{number}
14049 Set the cost assumed for the gettr instruction to @var{number}.
14050 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
14054 Assume pt* instructions won't trap. This will generally generate better
14055 scheduled code, but is unsafe on current hardware. The current architecture
14056 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
14057 This has the unintentional effect of making it unsafe to schedule ptabs /
14058 ptrel before a branch, or hoist it out of a loop. For example,
14059 __do_global_ctors, a part of libgcc that runs constructors at program
14060 startup, calls functions in a list which is delimited by @minus{}1. With the
14061 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
14062 That means that all the constructors will be run a bit quicker, but when
14063 the loop comes to the end of the list, the program crashes because ptabs
14064 loads @minus{}1 into a target register. Since this option is unsafe for any
14065 hardware implementing the current architecture specification, the default
14066 is -mno-pt-fixed. Unless the user specifies a specific cost with
14067 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
14068 this deters register allocation using target registers for storing
14071 @item -minvalid-symbols
14072 @opindex minvalid-symbols
14073 Assume symbols might be invalid. Ordinary function symbols generated by
14074 the compiler will always be valid to load with movi/shori/ptabs or
14075 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
14076 to generate symbols that will cause ptabs / ptrel to trap.
14077 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
14078 It will then prevent cross-basic-block cse, hoisting and most scheduling
14079 of symbol loads. The default is @option{-mno-invalid-symbols}.
14082 @node SPARC Options
14083 @subsection SPARC Options
14084 @cindex SPARC options
14086 These @samp{-m} options are supported on the SPARC:
14089 @item -mno-app-regs
14091 @opindex mno-app-regs
14093 Specify @option{-mapp-regs} to generate output using the global registers
14094 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
14097 To be fully SVR4 ABI compliant at the cost of some performance loss,
14098 specify @option{-mno-app-regs}. You should compile libraries and system
14099 software with this option.
14102 @itemx -mhard-float
14104 @opindex mhard-float
14105 Generate output containing floating point instructions. This is the
14109 @itemx -msoft-float
14111 @opindex msoft-float
14112 Generate output containing library calls for floating point.
14113 @strong{Warning:} the requisite libraries are not available for all SPARC
14114 targets. Normally the facilities of the machine's usual C compiler are
14115 used, but this cannot be done directly in cross-compilation. You must make
14116 your own arrangements to provide suitable library functions for
14117 cross-compilation. The embedded targets @samp{sparc-*-aout} and
14118 @samp{sparclite-*-*} do provide software floating point support.
14120 @option{-msoft-float} changes the calling convention in the output file;
14121 therefore, it is only useful if you compile @emph{all} of a program with
14122 this option. In particular, you need to compile @file{libgcc.a}, the
14123 library that comes with GCC, with @option{-msoft-float} in order for
14126 @item -mhard-quad-float
14127 @opindex mhard-quad-float
14128 Generate output containing quad-word (long double) floating point
14131 @item -msoft-quad-float
14132 @opindex msoft-quad-float
14133 Generate output containing library calls for quad-word (long double)
14134 floating point instructions. The functions called are those specified
14135 in the SPARC ABI@. This is the default.
14137 As of this writing, there are no SPARC implementations that have hardware
14138 support for the quad-word floating point instructions. They all invoke
14139 a trap handler for one of these instructions, and then the trap handler
14140 emulates the effect of the instruction. Because of the trap handler overhead,
14141 this is much slower than calling the ABI library routines. Thus the
14142 @option{-msoft-quad-float} option is the default.
14144 @item -mno-unaligned-doubles
14145 @itemx -munaligned-doubles
14146 @opindex mno-unaligned-doubles
14147 @opindex munaligned-doubles
14148 Assume that doubles have 8 byte alignment. This is the default.
14150 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
14151 alignment only if they are contained in another type, or if they have an
14152 absolute address. Otherwise, it assumes they have 4 byte alignment.
14153 Specifying this option avoids some rare compatibility problems with code
14154 generated by other compilers. It is not the default because it results
14155 in a performance loss, especially for floating point code.
14157 @item -mno-faster-structs
14158 @itemx -mfaster-structs
14159 @opindex mno-faster-structs
14160 @opindex mfaster-structs
14161 With @option{-mfaster-structs}, the compiler assumes that structures
14162 should have 8 byte alignment. This enables the use of pairs of
14163 @code{ldd} and @code{std} instructions for copies in structure
14164 assignment, in place of twice as many @code{ld} and @code{st} pairs.
14165 However, the use of this changed alignment directly violates the SPARC
14166 ABI@. Thus, it's intended only for use on targets where the developer
14167 acknowledges that their resulting code will not be directly in line with
14168 the rules of the ABI@.
14170 @item -mimpure-text
14171 @opindex mimpure-text
14172 @option{-mimpure-text}, used in addition to @option{-shared}, tells
14173 the compiler to not pass @option{-z text} to the linker when linking a
14174 shared object. Using this option, you can link position-dependent
14175 code into a shared object.
14177 @option{-mimpure-text} suppresses the ``relocations remain against
14178 allocatable but non-writable sections'' linker error message.
14179 However, the necessary relocations will trigger copy-on-write, and the
14180 shared object is not actually shared across processes. Instead of
14181 using @option{-mimpure-text}, you should compile all source code with
14182 @option{-fpic} or @option{-fPIC}.
14184 This option is only available on SunOS and Solaris.
14186 @item -mcpu=@var{cpu_type}
14188 Set the instruction set, register set, and instruction scheduling parameters
14189 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
14190 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
14191 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
14192 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
14193 @samp{ultrasparc3}, @samp{niagara} and @samp{niagara2}.
14195 Default instruction scheduling parameters are used for values that select
14196 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
14197 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
14199 Here is a list of each supported architecture and their supported
14204 v8: supersparc, hypersparc
14205 sparclite: f930, f934, sparclite86x
14207 v9: ultrasparc, ultrasparc3, niagara, niagara2
14210 By default (unless configured otherwise), GCC generates code for the V7
14211 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
14212 additionally optimizes it for the Cypress CY7C602 chip, as used in the
14213 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
14214 SPARCStation 1, 2, IPX etc.
14216 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
14217 architecture. The only difference from V7 code is that the compiler emits
14218 the integer multiply and integer divide instructions which exist in SPARC-V8
14219 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
14220 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
14223 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
14224 the SPARC architecture. This adds the integer multiply, integer divide step
14225 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
14226 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
14227 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
14228 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
14229 MB86934 chip, which is the more recent SPARClite with FPU@.
14231 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
14232 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
14233 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
14234 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
14235 optimizes it for the TEMIC SPARClet chip.
14237 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
14238 architecture. This adds 64-bit integer and floating-point move instructions,
14239 3 additional floating-point condition code registers and conditional move
14240 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
14241 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
14242 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
14243 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
14244 @option{-mcpu=niagara}, the compiler additionally optimizes it for
14245 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
14246 additionally optimizes it for Sun UltraSPARC T2 chips.
14248 @item -mtune=@var{cpu_type}
14250 Set the instruction scheduling parameters for machine type
14251 @var{cpu_type}, but do not set the instruction set or register set that the
14252 option @option{-mcpu=@var{cpu_type}} would.
14254 The same values for @option{-mcpu=@var{cpu_type}} can be used for
14255 @option{-mtune=@var{cpu_type}}, but the only useful values are those
14256 that select a particular cpu implementation. Those are @samp{cypress},
14257 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
14258 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
14259 @samp{ultrasparc3}, @samp{niagara}, and @samp{niagara2}.
14264 @opindex mno-v8plus
14265 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
14266 difference from the V8 ABI is that the global and out registers are
14267 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
14268 mode for all SPARC-V9 processors.
14274 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
14275 Visual Instruction Set extensions. The default is @option{-mno-vis}.
14278 These @samp{-m} options are supported in addition to the above
14279 on SPARC-V9 processors in 64-bit environments:
14282 @item -mlittle-endian
14283 @opindex mlittle-endian
14284 Generate code for a processor running in little-endian mode. It is only
14285 available for a few configurations and most notably not on Solaris and Linux.
14291 Generate code for a 32-bit or 64-bit environment.
14292 The 32-bit environment sets int, long and pointer to 32 bits.
14293 The 64-bit environment sets int to 32 bits and long and pointer
14296 @item -mcmodel=medlow
14297 @opindex mcmodel=medlow
14298 Generate code for the Medium/Low code model: 64-bit addresses, programs
14299 must be linked in the low 32 bits of memory. Programs can be statically
14300 or dynamically linked.
14302 @item -mcmodel=medmid
14303 @opindex mcmodel=medmid
14304 Generate code for the Medium/Middle code model: 64-bit addresses, programs
14305 must be linked in the low 44 bits of memory, the text and data segments must
14306 be less than 2GB in size and the data segment must be located within 2GB of
14309 @item -mcmodel=medany
14310 @opindex mcmodel=medany
14311 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
14312 may be linked anywhere in memory, the text and data segments must be less
14313 than 2GB in size and the data segment must be located within 2GB of the
14316 @item -mcmodel=embmedany
14317 @opindex mcmodel=embmedany
14318 Generate code for the Medium/Anywhere code model for embedded systems:
14319 64-bit addresses, the text and data segments must be less than 2GB in
14320 size, both starting anywhere in memory (determined at link time). The
14321 global register %g4 points to the base of the data segment. Programs
14322 are statically linked and PIC is not supported.
14325 @itemx -mno-stack-bias
14326 @opindex mstack-bias
14327 @opindex mno-stack-bias
14328 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
14329 frame pointer if present, are offset by @minus{}2047 which must be added back
14330 when making stack frame references. This is the default in 64-bit mode.
14331 Otherwise, assume no such offset is present.
14334 These switches are supported in addition to the above on Solaris:
14339 Add support for multithreading using the Solaris threads library. This
14340 option sets flags for both the preprocessor and linker. This option does
14341 not affect the thread safety of object code produced by the compiler or
14342 that of libraries supplied with it.
14346 Add support for multithreading using the POSIX threads library. This
14347 option sets flags for both the preprocessor and linker. This option does
14348 not affect the thread safety of object code produced by the compiler or
14349 that of libraries supplied with it.
14353 This is a synonym for @option{-pthreads}.
14357 @subsection SPU Options
14358 @cindex SPU options
14360 These @samp{-m} options are supported on the SPU:
14364 @itemx -merror-reloc
14365 @opindex mwarn-reloc
14366 @opindex merror-reloc
14368 The loader for SPU does not handle dynamic relocations. By default, GCC
14369 will give an error when it generates code that requires a dynamic
14370 relocation. @option{-mno-error-reloc} disables the error,
14371 @option{-mwarn-reloc} will generate a warning instead.
14374 @itemx -munsafe-dma
14376 @opindex munsafe-dma
14378 Instructions which initiate or test completion of DMA must not be
14379 reordered with respect to loads and stores of the memory which is being
14380 accessed. Users typically address this problem using the volatile
14381 keyword, but that can lead to inefficient code in places where the
14382 memory is known to not change. Rather than mark the memory as volatile
14383 we treat the DMA instructions as potentially effecting all memory. With
14384 @option{-munsafe-dma} users must use the volatile keyword to protect
14387 @item -mbranch-hints
14388 @opindex mbranch-hints
14390 By default, GCC will generate a branch hint instruction to avoid
14391 pipeline stalls for always taken or probably taken branches. A hint
14392 will not be generated closer than 8 instructions away from its branch.
14393 There is little reason to disable them, except for debugging purposes,
14394 or to make an object a little bit smaller.
14398 @opindex msmall-mem
14399 @opindex mlarge-mem
14401 By default, GCC generates code assuming that addresses are never larger
14402 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
14403 a full 32 bit address.
14408 By default, GCC links against startup code that assumes the SPU-style
14409 main function interface (which has an unconventional parameter list).
14410 With @option{-mstdmain}, GCC will link your program against startup
14411 code that assumes a C99-style interface to @code{main}, including a
14412 local copy of @code{argv} strings.
14414 @item -mfixed-range=@var{register-range}
14415 @opindex mfixed-range
14416 Generate code treating the given register range as fixed registers.
14417 A fixed register is one that the register allocator can not use. This is
14418 useful when compiling kernel code. A register range is specified as
14419 two registers separated by a dash. Multiple register ranges can be
14420 specified separated by a comma.
14424 @node System V Options
14425 @subsection Options for System V
14427 These additional options are available on System V Release 4 for
14428 compatibility with other compilers on those systems:
14433 Create a shared object.
14434 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
14438 Identify the versions of each tool used by the compiler, in a
14439 @code{.ident} assembler directive in the output.
14443 Refrain from adding @code{.ident} directives to the output file (this is
14446 @item -YP,@var{dirs}
14448 Search the directories @var{dirs}, and no others, for libraries
14449 specified with @option{-l}.
14451 @item -Ym,@var{dir}
14453 Look in the directory @var{dir} to find the M4 preprocessor.
14454 The assembler uses this option.
14455 @c This is supposed to go with a -Yd for predefined M4 macro files, but
14456 @c the generic assembler that comes with Solaris takes just -Ym.
14460 @subsection V850 Options
14461 @cindex V850 Options
14463 These @samp{-m} options are defined for V850 implementations:
14467 @itemx -mno-long-calls
14468 @opindex mlong-calls
14469 @opindex mno-long-calls
14470 Treat all calls as being far away (near). If calls are assumed to be
14471 far away, the compiler will always load the functions address up into a
14472 register, and call indirect through the pointer.
14478 Do not optimize (do optimize) basic blocks that use the same index
14479 pointer 4 or more times to copy pointer into the @code{ep} register, and
14480 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
14481 option is on by default if you optimize.
14483 @item -mno-prolog-function
14484 @itemx -mprolog-function
14485 @opindex mno-prolog-function
14486 @opindex mprolog-function
14487 Do not use (do use) external functions to save and restore registers
14488 at the prologue and epilogue of a function. The external functions
14489 are slower, but use less code space if more than one function saves
14490 the same number of registers. The @option{-mprolog-function} option
14491 is on by default if you optimize.
14495 Try to make the code as small as possible. At present, this just turns
14496 on the @option{-mep} and @option{-mprolog-function} options.
14498 @item -mtda=@var{n}
14500 Put static or global variables whose size is @var{n} bytes or less into
14501 the tiny data area that register @code{ep} points to. The tiny data
14502 area can hold up to 256 bytes in total (128 bytes for byte references).
14504 @item -msda=@var{n}
14506 Put static or global variables whose size is @var{n} bytes or less into
14507 the small data area that register @code{gp} points to. The small data
14508 area can hold up to 64 kilobytes.
14510 @item -mzda=@var{n}
14512 Put static or global variables whose size is @var{n} bytes or less into
14513 the first 32 kilobytes of memory.
14517 Specify that the target processor is the V850.
14520 @opindex mbig-switch
14521 Generate code suitable for big switch tables. Use this option only if
14522 the assembler/linker complain about out of range branches within a switch
14527 This option will cause r2 and r5 to be used in the code generated by
14528 the compiler. This setting is the default.
14530 @item -mno-app-regs
14531 @opindex mno-app-regs
14532 This option will cause r2 and r5 to be treated as fixed registers.
14536 Specify that the target processor is the V850E1. The preprocessor
14537 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
14538 this option is used.
14542 Specify that the target processor is the V850E@. The preprocessor
14543 constant @samp{__v850e__} will be defined if this option is used.
14545 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
14546 are defined then a default target processor will be chosen and the
14547 relevant @samp{__v850*__} preprocessor constant will be defined.
14549 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
14550 defined, regardless of which processor variant is the target.
14552 @item -mdisable-callt
14553 @opindex mdisable-callt
14554 This option will suppress generation of the CALLT instruction for the
14555 v850e and v850e1 flavors of the v850 architecture. The default is
14556 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
14561 @subsection VAX Options
14562 @cindex VAX options
14564 These @samp{-m} options are defined for the VAX:
14569 Do not output certain jump instructions (@code{aobleq} and so on)
14570 that the Unix assembler for the VAX cannot handle across long
14575 Do output those jump instructions, on the assumption that you
14576 will assemble with the GNU assembler.
14580 Output code for g-format floating point numbers instead of d-format.
14583 @node VxWorks Options
14584 @subsection VxWorks Options
14585 @cindex VxWorks Options
14587 The options in this section are defined for all VxWorks targets.
14588 Options specific to the target hardware are listed with the other
14589 options for that target.
14594 GCC can generate code for both VxWorks kernels and real time processes
14595 (RTPs). This option switches from the former to the latter. It also
14596 defines the preprocessor macro @code{__RTP__}.
14599 @opindex non-static
14600 Link an RTP executable against shared libraries rather than static
14601 libraries. The options @option{-static} and @option{-shared} can
14602 also be used for RTPs (@pxref{Link Options}); @option{-static}
14609 These options are passed down to the linker. They are defined for
14610 compatibility with Diab.
14613 @opindex Xbind-lazy
14614 Enable lazy binding of function calls. This option is equivalent to
14615 @option{-Wl,-z,now} and is defined for compatibility with Diab.
14619 Disable lazy binding of function calls. This option is the default and
14620 is defined for compatibility with Diab.
14623 @node x86-64 Options
14624 @subsection x86-64 Options
14625 @cindex x86-64 options
14627 These are listed under @xref{i386 and x86-64 Options}.
14629 @node Xstormy16 Options
14630 @subsection Xstormy16 Options
14631 @cindex Xstormy16 Options
14633 These options are defined for Xstormy16:
14638 Choose startup files and linker script suitable for the simulator.
14641 @node Xtensa Options
14642 @subsection Xtensa Options
14643 @cindex Xtensa Options
14645 These options are supported for Xtensa targets:
14649 @itemx -mno-const16
14651 @opindex mno-const16
14652 Enable or disable use of @code{CONST16} instructions for loading
14653 constant values. The @code{CONST16} instruction is currently not a
14654 standard option from Tensilica. When enabled, @code{CONST16}
14655 instructions are always used in place of the standard @code{L32R}
14656 instructions. The use of @code{CONST16} is enabled by default only if
14657 the @code{L32R} instruction is not available.
14660 @itemx -mno-fused-madd
14661 @opindex mfused-madd
14662 @opindex mno-fused-madd
14663 Enable or disable use of fused multiply/add and multiply/subtract
14664 instructions in the floating-point option. This has no effect if the
14665 floating-point option is not also enabled. Disabling fused multiply/add
14666 and multiply/subtract instructions forces the compiler to use separate
14667 instructions for the multiply and add/subtract operations. This may be
14668 desirable in some cases where strict IEEE 754-compliant results are
14669 required: the fused multiply add/subtract instructions do not round the
14670 intermediate result, thereby producing results with @emph{more} bits of
14671 precision than specified by the IEEE standard. Disabling fused multiply
14672 add/subtract instructions also ensures that the program output is not
14673 sensitive to the compiler's ability to combine multiply and add/subtract
14676 @item -mtext-section-literals
14677 @itemx -mno-text-section-literals
14678 @opindex mtext-section-literals
14679 @opindex mno-text-section-literals
14680 Control the treatment of literal pools. The default is
14681 @option{-mno-text-section-literals}, which places literals in a separate
14682 section in the output file. This allows the literal pool to be placed
14683 in a data RAM/ROM, and it also allows the linker to combine literal
14684 pools from separate object files to remove redundant literals and
14685 improve code size. With @option{-mtext-section-literals}, the literals
14686 are interspersed in the text section in order to keep them as close as
14687 possible to their references. This may be necessary for large assembly
14690 @item -mtarget-align
14691 @itemx -mno-target-align
14692 @opindex mtarget-align
14693 @opindex mno-target-align
14694 When this option is enabled, GCC instructs the assembler to
14695 automatically align instructions to reduce branch penalties at the
14696 expense of some code density. The assembler attempts to widen density
14697 instructions to align branch targets and the instructions following call
14698 instructions. If there are not enough preceding safe density
14699 instructions to align a target, no widening will be performed. The
14700 default is @option{-mtarget-align}. These options do not affect the
14701 treatment of auto-aligned instructions like @code{LOOP}, which the
14702 assembler will always align, either by widening density instructions or
14703 by inserting no-op instructions.
14706 @itemx -mno-longcalls
14707 @opindex mlongcalls
14708 @opindex mno-longcalls
14709 When this option is enabled, GCC instructs the assembler to translate
14710 direct calls to indirect calls unless it can determine that the target
14711 of a direct call is in the range allowed by the call instruction. This
14712 translation typically occurs for calls to functions in other source
14713 files. Specifically, the assembler translates a direct @code{CALL}
14714 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
14715 The default is @option{-mno-longcalls}. This option should be used in
14716 programs where the call target can potentially be out of range. This
14717 option is implemented in the assembler, not the compiler, so the
14718 assembly code generated by GCC will still show direct call
14719 instructions---look at the disassembled object code to see the actual
14720 instructions. Note that the assembler will use an indirect call for
14721 every cross-file call, not just those that really will be out of range.
14724 @node zSeries Options
14725 @subsection zSeries Options
14726 @cindex zSeries options
14728 These are listed under @xref{S/390 and zSeries Options}.
14730 @node Code Gen Options
14731 @section Options for Code Generation Conventions
14732 @cindex code generation conventions
14733 @cindex options, code generation
14734 @cindex run-time options
14736 These machine-independent options control the interface conventions
14737 used in code generation.
14739 Most of them have both positive and negative forms; the negative form
14740 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
14741 one of the forms is listed---the one which is not the default. You
14742 can figure out the other form by either removing @samp{no-} or adding
14746 @item -fbounds-check
14747 @opindex fbounds-check
14748 For front-ends that support it, generate additional code to check that
14749 indices used to access arrays are within the declared range. This is
14750 currently only supported by the Java and Fortran front-ends, where
14751 this option defaults to true and false respectively.
14755 This option generates traps for signed overflow on addition, subtraction,
14756 multiplication operations.
14760 This option instructs the compiler to assume that signed arithmetic
14761 overflow of addition, subtraction and multiplication wraps around
14762 using twos-complement representation. This flag enables some optimizations
14763 and disables others. This option is enabled by default for the Java
14764 front-end, as required by the Java language specification.
14767 @opindex fexceptions
14768 Enable exception handling. Generates extra code needed to propagate
14769 exceptions. For some targets, this implies GCC will generate frame
14770 unwind information for all functions, which can produce significant data
14771 size overhead, although it does not affect execution. If you do not
14772 specify this option, GCC will enable it by default for languages like
14773 C++ which normally require exception handling, and disable it for
14774 languages like C that do not normally require it. However, you may need
14775 to enable this option when compiling C code that needs to interoperate
14776 properly with exception handlers written in C++. You may also wish to
14777 disable this option if you are compiling older C++ programs that don't
14778 use exception handling.
14780 @item -fnon-call-exceptions
14781 @opindex fnon-call-exceptions
14782 Generate code that allows trapping instructions to throw exceptions.
14783 Note that this requires platform-specific runtime support that does
14784 not exist everywhere. Moreover, it only allows @emph{trapping}
14785 instructions to throw exceptions, i.e.@: memory references or floating
14786 point instructions. It does not allow exceptions to be thrown from
14787 arbitrary signal handlers such as @code{SIGALRM}.
14789 @item -funwind-tables
14790 @opindex funwind-tables
14791 Similar to @option{-fexceptions}, except that it will just generate any needed
14792 static data, but will not affect the generated code in any other way.
14793 You will normally not enable this option; instead, a language processor
14794 that needs this handling would enable it on your behalf.
14796 @item -fasynchronous-unwind-tables
14797 @opindex fasynchronous-unwind-tables
14798 Generate unwind table in dwarf2 format, if supported by target machine. The
14799 table is exact at each instruction boundary, so it can be used for stack
14800 unwinding from asynchronous events (such as debugger or garbage collector).
14802 @item -fpcc-struct-return
14803 @opindex fpcc-struct-return
14804 Return ``short'' @code{struct} and @code{union} values in memory like
14805 longer ones, rather than in registers. This convention is less
14806 efficient, but it has the advantage of allowing intercallability between
14807 GCC-compiled files and files compiled with other compilers, particularly
14808 the Portable C Compiler (pcc).
14810 The precise convention for returning structures in memory depends
14811 on the target configuration macros.
14813 Short structures and unions are those whose size and alignment match
14814 that of some integer type.
14816 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
14817 switch is not binary compatible with code compiled with the
14818 @option{-freg-struct-return} switch.
14819 Use it to conform to a non-default application binary interface.
14821 @item -freg-struct-return
14822 @opindex freg-struct-return
14823 Return @code{struct} and @code{union} values in registers when possible.
14824 This is more efficient for small structures than
14825 @option{-fpcc-struct-return}.
14827 If you specify neither @option{-fpcc-struct-return} nor
14828 @option{-freg-struct-return}, GCC defaults to whichever convention is
14829 standard for the target. If there is no standard convention, GCC
14830 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
14831 the principal compiler. In those cases, we can choose the standard, and
14832 we chose the more efficient register return alternative.
14834 @strong{Warning:} code compiled with the @option{-freg-struct-return}
14835 switch is not binary compatible with code compiled with the
14836 @option{-fpcc-struct-return} switch.
14837 Use it to conform to a non-default application binary interface.
14839 @item -fshort-enums
14840 @opindex fshort-enums
14841 Allocate to an @code{enum} type only as many bytes as it needs for the
14842 declared range of possible values. Specifically, the @code{enum} type
14843 will be equivalent to the smallest integer type which has enough room.
14845 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
14846 code that is not binary compatible with code generated without that switch.
14847 Use it to conform to a non-default application binary interface.
14849 @item -fshort-double
14850 @opindex fshort-double
14851 Use the same size for @code{double} as for @code{float}.
14853 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
14854 code that is not binary compatible with code generated without that switch.
14855 Use it to conform to a non-default application binary interface.
14857 @item -fshort-wchar
14858 @opindex fshort-wchar
14859 Override the underlying type for @samp{wchar_t} to be @samp{short
14860 unsigned int} instead of the default for the target. This option is
14861 useful for building programs to run under WINE@.
14863 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
14864 code that is not binary compatible with code generated without that switch.
14865 Use it to conform to a non-default application binary interface.
14868 @opindex fno-common
14869 In C, allocate even uninitialized global variables in the data section of the
14870 object file, rather than generating them as common blocks. This has the
14871 effect that if the same variable is declared (without @code{extern}) in
14872 two different compilations, you will get an error when you link them.
14873 The only reason this might be useful is if you wish to verify that the
14874 program will work on other systems which always work this way.
14878 Ignore the @samp{#ident} directive.
14880 @item -finhibit-size-directive
14881 @opindex finhibit-size-directive
14882 Don't output a @code{.size} assembler directive, or anything else that
14883 would cause trouble if the function is split in the middle, and the
14884 two halves are placed at locations far apart in memory. This option is
14885 used when compiling @file{crtstuff.c}; you should not need to use it
14888 @item -fverbose-asm
14889 @opindex fverbose-asm
14890 Put extra commentary information in the generated assembly code to
14891 make it more readable. This option is generally only of use to those
14892 who actually need to read the generated assembly code (perhaps while
14893 debugging the compiler itself).
14895 @option{-fno-verbose-asm}, the default, causes the
14896 extra information to be omitted and is useful when comparing two assembler
14899 @item -frecord-gcc-switches
14900 @opindex frecord-gcc-switches
14901 This switch causes the command line that was used to invoke the
14902 compiler to be recorded into the object file that is being created.
14903 This switch is only implemented on some targets and the exact format
14904 of the recording is target and binary file format dependent, but it
14905 usually takes the form of a section containing ASCII text. This
14906 switch is related to the @option{-fverbose-asm} switch, but that
14907 switch only records information in the assembler output file as
14908 comments, so it never reaches the object file.
14912 @cindex global offset table
14914 Generate position-independent code (PIC) suitable for use in a shared
14915 library, if supported for the target machine. Such code accesses all
14916 constant addresses through a global offset table (GOT)@. The dynamic
14917 loader resolves the GOT entries when the program starts (the dynamic
14918 loader is not part of GCC; it is part of the operating system). If
14919 the GOT size for the linked executable exceeds a machine-specific
14920 maximum size, you get an error message from the linker indicating that
14921 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
14922 instead. (These maximums are 8k on the SPARC and 32k
14923 on the m68k and RS/6000. The 386 has no such limit.)
14925 Position-independent code requires special support, and therefore works
14926 only on certain machines. For the 386, GCC supports PIC for System V
14927 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
14928 position-independent.
14930 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14935 If supported for the target machine, emit position-independent code,
14936 suitable for dynamic linking and avoiding any limit on the size of the
14937 global offset table. This option makes a difference on the m68k,
14938 PowerPC and SPARC@.
14940 Position-independent code requires special support, and therefore works
14941 only on certain machines.
14943 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14950 These options are similar to @option{-fpic} and @option{-fPIC}, but
14951 generated position independent code can be only linked into executables.
14952 Usually these options are used when @option{-pie} GCC option will be
14953 used during linking.
14955 @option{-fpie} and @option{-fPIE} both define the macros
14956 @code{__pie__} and @code{__PIE__}. The macros have the value 1
14957 for @option{-fpie} and 2 for @option{-fPIE}.
14959 @item -fno-jump-tables
14960 @opindex fno-jump-tables
14961 Do not use jump tables for switch statements even where it would be
14962 more efficient than other code generation strategies. This option is
14963 of use in conjunction with @option{-fpic} or @option{-fPIC} for
14964 building code which forms part of a dynamic linker and cannot
14965 reference the address of a jump table. On some targets, jump tables
14966 do not require a GOT and this option is not needed.
14968 @item -ffixed-@var{reg}
14970 Treat the register named @var{reg} as a fixed register; generated code
14971 should never refer to it (except perhaps as a stack pointer, frame
14972 pointer or in some other fixed role).
14974 @var{reg} must be the name of a register. The register names accepted
14975 are machine-specific and are defined in the @code{REGISTER_NAMES}
14976 macro in the machine description macro file.
14978 This flag does not have a negative form, because it specifies a
14981 @item -fcall-used-@var{reg}
14982 @opindex fcall-used
14983 Treat the register named @var{reg} as an allocable register that is
14984 clobbered by function calls. It may be allocated for temporaries or
14985 variables that do not live across a call. Functions compiled this way
14986 will not save and restore the register @var{reg}.
14988 It is an error to used this flag with the frame pointer or stack pointer.
14989 Use of this flag for other registers that have fixed pervasive roles in
14990 the machine's execution model will produce disastrous results.
14992 This flag does not have a negative form, because it specifies a
14995 @item -fcall-saved-@var{reg}
14996 @opindex fcall-saved
14997 Treat the register named @var{reg} as an allocable register saved by
14998 functions. It may be allocated even for temporaries or variables that
14999 live across a call. Functions compiled this way will save and restore
15000 the register @var{reg} if they use it.
15002 It is an error to used this flag with the frame pointer or stack pointer.
15003 Use of this flag for other registers that have fixed pervasive roles in
15004 the machine's execution model will produce disastrous results.
15006 A different sort of disaster will result from the use of this flag for
15007 a register in which function values may be returned.
15009 This flag does not have a negative form, because it specifies a
15012 @item -fpack-struct[=@var{n}]
15013 @opindex fpack-struct
15014 Without a value specified, pack all structure members together without
15015 holes. When a value is specified (which must be a small power of two), pack
15016 structure members according to this value, representing the maximum
15017 alignment (that is, objects with default alignment requirements larger than
15018 this will be output potentially unaligned at the next fitting location.
15020 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
15021 code that is not binary compatible with code generated without that switch.
15022 Additionally, it makes the code suboptimal.
15023 Use it to conform to a non-default application binary interface.
15025 @item -finstrument-functions
15026 @opindex finstrument-functions
15027 Generate instrumentation calls for entry and exit to functions. Just
15028 after function entry and just before function exit, the following
15029 profiling functions will be called with the address of the current
15030 function and its call site. (On some platforms,
15031 @code{__builtin_return_address} does not work beyond the current
15032 function, so the call site information may not be available to the
15033 profiling functions otherwise.)
15036 void __cyg_profile_func_enter (void *this_fn,
15038 void __cyg_profile_func_exit (void *this_fn,
15042 The first argument is the address of the start of the current function,
15043 which may be looked up exactly in the symbol table.
15045 This instrumentation is also done for functions expanded inline in other
15046 functions. The profiling calls will indicate where, conceptually, the
15047 inline function is entered and exited. This means that addressable
15048 versions of such functions must be available. If all your uses of a
15049 function are expanded inline, this may mean an additional expansion of
15050 code size. If you use @samp{extern inline} in your C code, an
15051 addressable version of such functions must be provided. (This is
15052 normally the case anyways, but if you get lucky and the optimizer always
15053 expands the functions inline, you might have gotten away without
15054 providing static copies.)
15056 A function may be given the attribute @code{no_instrument_function}, in
15057 which case this instrumentation will not be done. This can be used, for
15058 example, for the profiling functions listed above, high-priority
15059 interrupt routines, and any functions from which the profiling functions
15060 cannot safely be called (perhaps signal handlers, if the profiling
15061 routines generate output or allocate memory).
15063 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
15064 @opindex finstrument-functions-exclude-file-list
15066 Set the list of functions that are excluded from instrumentation (see
15067 the description of @code{-finstrument-functions}). If the file that
15068 contains a function definition matches with one of @var{file}, then
15069 that function is not instrumented. The match is done on substrings:
15070 if the @var{file} parameter is a substring of the file name, it is
15071 considered to be a match.
15074 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
15075 will exclude any inline function defined in files whose pathnames
15076 contain @code{/bits/stl} or @code{include/sys}.
15078 If, for some reason, you want to include letter @code{','} in one of
15079 @var{sym}, write @code{'\,'}. For example,
15080 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
15081 (note the single quote surrounding the option).
15083 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
15084 @opindex finstrument-functions-exclude-function-list
15086 This is similar to @code{-finstrument-functions-exclude-file-list},
15087 but this option sets the list of function names to be excluded from
15088 instrumentation. The function name to be matched is its user-visible
15089 name, such as @code{vector<int> blah(const vector<int> &)}, not the
15090 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
15091 match is done on substrings: if the @var{sym} parameter is a substring
15092 of the function name, it is considered to be a match.
15094 @item -fstack-check
15095 @opindex fstack-check
15096 Generate code to verify that you do not go beyond the boundary of the
15097 stack. You should specify this flag if you are running in an
15098 environment with multiple threads, but only rarely need to specify it in
15099 a single-threaded environment since stack overflow is automatically
15100 detected on nearly all systems if there is only one stack.
15102 Note that this switch does not actually cause checking to be done; the
15103 operating system must do that. The switch causes generation of code
15104 to ensure that the operating system sees the stack being extended.
15106 @item -fstack-limit-register=@var{reg}
15107 @itemx -fstack-limit-symbol=@var{sym}
15108 @itemx -fno-stack-limit
15109 @opindex fstack-limit-register
15110 @opindex fstack-limit-symbol
15111 @opindex fno-stack-limit
15112 Generate code to ensure that the stack does not grow beyond a certain value,
15113 either the value of a register or the address of a symbol. If the stack
15114 would grow beyond the value, a signal is raised. For most targets,
15115 the signal is raised before the stack overruns the boundary, so
15116 it is possible to catch the signal without taking special precautions.
15118 For instance, if the stack starts at absolute address @samp{0x80000000}
15119 and grows downwards, you can use the flags
15120 @option{-fstack-limit-symbol=__stack_limit} and
15121 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
15122 of 128KB@. Note that this may only work with the GNU linker.
15124 @cindex aliasing of parameters
15125 @cindex parameters, aliased
15126 @item -fargument-alias
15127 @itemx -fargument-noalias
15128 @itemx -fargument-noalias-global
15129 @itemx -fargument-noalias-anything
15130 @opindex fargument-alias
15131 @opindex fargument-noalias
15132 @opindex fargument-noalias-global
15133 @opindex fargument-noalias-anything
15134 Specify the possible relationships among parameters and between
15135 parameters and global data.
15137 @option{-fargument-alias} specifies that arguments (parameters) may
15138 alias each other and may alias global storage.@*
15139 @option{-fargument-noalias} specifies that arguments do not alias
15140 each other, but may alias global storage.@*
15141 @option{-fargument-noalias-global} specifies that arguments do not
15142 alias each other and do not alias global storage.
15143 @option{-fargument-noalias-anything} specifies that arguments do not
15144 alias any other storage.
15146 Each language will automatically use whatever option is required by
15147 the language standard. You should not need to use these options yourself.
15149 @item -fleading-underscore
15150 @opindex fleading-underscore
15151 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
15152 change the way C symbols are represented in the object file. One use
15153 is to help link with legacy assembly code.
15155 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
15156 generate code that is not binary compatible with code generated without that
15157 switch. Use it to conform to a non-default application binary interface.
15158 Not all targets provide complete support for this switch.
15160 @item -ftls-model=@var{model}
15161 @opindex ftls-model
15162 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
15163 The @var{model} argument should be one of @code{global-dynamic},
15164 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
15166 The default without @option{-fpic} is @code{initial-exec}; with
15167 @option{-fpic} the default is @code{global-dynamic}.
15169 @item -fvisibility=@var{default|internal|hidden|protected}
15170 @opindex fvisibility
15171 Set the default ELF image symbol visibility to the specified option---all
15172 symbols will be marked with this unless overridden within the code.
15173 Using this feature can very substantially improve linking and
15174 load times of shared object libraries, produce more optimized
15175 code, provide near-perfect API export and prevent symbol clashes.
15176 It is @strong{strongly} recommended that you use this in any shared objects
15179 Despite the nomenclature, @code{default} always means public ie;
15180 available to be linked against from outside the shared object.
15181 @code{protected} and @code{internal} are pretty useless in real-world
15182 usage so the only other commonly used option will be @code{hidden}.
15183 The default if @option{-fvisibility} isn't specified is
15184 @code{default}, i.e., make every
15185 symbol public---this causes the same behavior as previous versions of
15188 A good explanation of the benefits offered by ensuring ELF
15189 symbols have the correct visibility is given by ``How To Write
15190 Shared Libraries'' by Ulrich Drepper (which can be found at
15191 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
15192 solution made possible by this option to marking things hidden when
15193 the default is public is to make the default hidden and mark things
15194 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
15195 and @code{__attribute__ ((visibility("default")))} instead of
15196 @code{__declspec(dllexport)} you get almost identical semantics with
15197 identical syntax. This is a great boon to those working with
15198 cross-platform projects.
15200 For those adding visibility support to existing code, you may find
15201 @samp{#pragma GCC visibility} of use. This works by you enclosing
15202 the declarations you wish to set visibility for with (for example)
15203 @samp{#pragma GCC visibility push(hidden)} and
15204 @samp{#pragma GCC visibility pop}.
15205 Bear in mind that symbol visibility should be viewed @strong{as
15206 part of the API interface contract} and thus all new code should
15207 always specify visibility when it is not the default ie; declarations
15208 only for use within the local DSO should @strong{always} be marked explicitly
15209 as hidden as so to avoid PLT indirection overheads---making this
15210 abundantly clear also aids readability and self-documentation of the code.
15211 Note that due to ISO C++ specification requirements, operator new and
15212 operator delete must always be of default visibility.
15214 Be aware that headers from outside your project, in particular system
15215 headers and headers from any other library you use, may not be
15216 expecting to be compiled with visibility other than the default. You
15217 may need to explicitly say @samp{#pragma GCC visibility push(default)}
15218 before including any such headers.
15220 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
15221 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
15222 no modifications. However, this means that calls to @samp{extern}
15223 functions with no explicit visibility will use the PLT, so it is more
15224 effective to use @samp{__attribute ((visibility))} and/or
15225 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
15226 declarations should be treated as hidden.
15228 Note that @samp{-fvisibility} does affect C++ vague linkage
15229 entities. This means that, for instance, an exception class that will
15230 be thrown between DSOs must be explicitly marked with default
15231 visibility so that the @samp{type_info} nodes will be unified between
15234 An overview of these techniques, their benefits and how to use them
15235 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
15241 @node Environment Variables
15242 @section Environment Variables Affecting GCC
15243 @cindex environment variables
15245 @c man begin ENVIRONMENT
15246 This section describes several environment variables that affect how GCC
15247 operates. Some of them work by specifying directories or prefixes to use
15248 when searching for various kinds of files. Some are used to specify other
15249 aspects of the compilation environment.
15251 Note that you can also specify places to search using options such as
15252 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
15253 take precedence over places specified using environment variables, which
15254 in turn take precedence over those specified by the configuration of GCC@.
15255 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
15256 GNU Compiler Collection (GCC) Internals}.
15261 @c @itemx LC_COLLATE
15263 @c @itemx LC_MONETARY
15264 @c @itemx LC_NUMERIC
15269 @c @findex LC_COLLATE
15270 @findex LC_MESSAGES
15271 @c @findex LC_MONETARY
15272 @c @findex LC_NUMERIC
15276 These environment variables control the way that GCC uses
15277 localization information that allow GCC to work with different
15278 national conventions. GCC inspects the locale categories
15279 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
15280 so. These locale categories can be set to any value supported by your
15281 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
15282 Kingdom encoded in UTF-8.
15284 The @env{LC_CTYPE} environment variable specifies character
15285 classification. GCC uses it to determine the character boundaries in
15286 a string; this is needed for some multibyte encodings that contain quote
15287 and escape characters that would otherwise be interpreted as a string
15290 The @env{LC_MESSAGES} environment variable specifies the language to
15291 use in diagnostic messages.
15293 If the @env{LC_ALL} environment variable is set, it overrides the value
15294 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
15295 and @env{LC_MESSAGES} default to the value of the @env{LANG}
15296 environment variable. If none of these variables are set, GCC
15297 defaults to traditional C English behavior.
15301 If @env{TMPDIR} is set, it specifies the directory to use for temporary
15302 files. GCC uses temporary files to hold the output of one stage of
15303 compilation which is to be used as input to the next stage: for example,
15304 the output of the preprocessor, which is the input to the compiler
15307 @item GCC_EXEC_PREFIX
15308 @findex GCC_EXEC_PREFIX
15309 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
15310 names of the subprograms executed by the compiler. No slash is added
15311 when this prefix is combined with the name of a subprogram, but you can
15312 specify a prefix that ends with a slash if you wish.
15314 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
15315 an appropriate prefix to use based on the pathname it was invoked with.
15317 If GCC cannot find the subprogram using the specified prefix, it
15318 tries looking in the usual places for the subprogram.
15320 The default value of @env{GCC_EXEC_PREFIX} is
15321 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
15322 the installed compiler. In many cases @var{prefix} is the value
15323 of @code{prefix} when you ran the @file{configure} script.
15325 Other prefixes specified with @option{-B} take precedence over this prefix.
15327 This prefix is also used for finding files such as @file{crt0.o} that are
15330 In addition, the prefix is used in an unusual way in finding the
15331 directories to search for header files. For each of the standard
15332 directories whose name normally begins with @samp{/usr/local/lib/gcc}
15333 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
15334 replacing that beginning with the specified prefix to produce an
15335 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
15336 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
15337 These alternate directories are searched first; the standard directories
15338 come next. If a standard directory begins with the configured
15339 @var{prefix} then the value of @var{prefix} is replaced by
15340 @env{GCC_EXEC_PREFIX} when looking for header files.
15342 @item COMPILER_PATH
15343 @findex COMPILER_PATH
15344 The value of @env{COMPILER_PATH} is a colon-separated list of
15345 directories, much like @env{PATH}. GCC tries the directories thus
15346 specified when searching for subprograms, if it can't find the
15347 subprograms using @env{GCC_EXEC_PREFIX}.
15350 @findex LIBRARY_PATH
15351 The value of @env{LIBRARY_PATH} is a colon-separated list of
15352 directories, much like @env{PATH}. When configured as a native compiler,
15353 GCC tries the directories thus specified when searching for special
15354 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
15355 using GCC also uses these directories when searching for ordinary
15356 libraries for the @option{-l} option (but directories specified with
15357 @option{-L} come first).
15361 @cindex locale definition
15362 This variable is used to pass locale information to the compiler. One way in
15363 which this information is used is to determine the character set to be used
15364 when character literals, string literals and comments are parsed in C and C++.
15365 When the compiler is configured to allow multibyte characters,
15366 the following values for @env{LANG} are recognized:
15370 Recognize JIS characters.
15372 Recognize SJIS characters.
15374 Recognize EUCJP characters.
15377 If @env{LANG} is not defined, or if it has some other value, then the
15378 compiler will use mblen and mbtowc as defined by the default locale to
15379 recognize and translate multibyte characters.
15383 Some additional environments variables affect the behavior of the
15386 @include cppenv.texi
15390 @node Precompiled Headers
15391 @section Using Precompiled Headers
15392 @cindex precompiled headers
15393 @cindex speed of compilation
15395 Often large projects have many header files that are included in every
15396 source file. The time the compiler takes to process these header files
15397 over and over again can account for nearly all of the time required to
15398 build the project. To make builds faster, GCC allows users to
15399 `precompile' a header file; then, if builds can use the precompiled
15400 header file they will be much faster.
15402 To create a precompiled header file, simply compile it as you would any
15403 other file, if necessary using the @option{-x} option to make the driver
15404 treat it as a C or C++ header file. You will probably want to use a
15405 tool like @command{make} to keep the precompiled header up-to-date when
15406 the headers it contains change.
15408 A precompiled header file will be searched for when @code{#include} is
15409 seen in the compilation. As it searches for the included file
15410 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
15411 compiler looks for a precompiled header in each directory just before it
15412 looks for the include file in that directory. The name searched for is
15413 the name specified in the @code{#include} with @samp{.gch} appended. If
15414 the precompiled header file can't be used, it is ignored.
15416 For instance, if you have @code{#include "all.h"}, and you have
15417 @file{all.h.gch} in the same directory as @file{all.h}, then the
15418 precompiled header file will be used if possible, and the original
15419 header will be used otherwise.
15421 Alternatively, you might decide to put the precompiled header file in a
15422 directory and use @option{-I} to ensure that directory is searched
15423 before (or instead of) the directory containing the original header.
15424 Then, if you want to check that the precompiled header file is always
15425 used, you can put a file of the same name as the original header in this
15426 directory containing an @code{#error} command.
15428 This also works with @option{-include}. So yet another way to use
15429 precompiled headers, good for projects not designed with precompiled
15430 header files in mind, is to simply take most of the header files used by
15431 a project, include them from another header file, precompile that header
15432 file, and @option{-include} the precompiled header. If the header files
15433 have guards against multiple inclusion, they will be skipped because
15434 they've already been included (in the precompiled header).
15436 If you need to precompile the same header file for different
15437 languages, targets, or compiler options, you can instead make a
15438 @emph{directory} named like @file{all.h.gch}, and put each precompiled
15439 header in the directory, perhaps using @option{-o}. It doesn't matter
15440 what you call the files in the directory, every precompiled header in
15441 the directory will be considered. The first precompiled header
15442 encountered in the directory that is valid for this compilation will
15443 be used; they're searched in no particular order.
15445 There are many other possibilities, limited only by your imagination,
15446 good sense, and the constraints of your build system.
15448 A precompiled header file can be used only when these conditions apply:
15452 Only one precompiled header can be used in a particular compilation.
15455 A precompiled header can't be used once the first C token is seen. You
15456 can have preprocessor directives before a precompiled header; you can
15457 even include a precompiled header from inside another header, so long as
15458 there are no C tokens before the @code{#include}.
15461 The precompiled header file must be produced for the same language as
15462 the current compilation. You can't use a C precompiled header for a C++
15466 The precompiled header file must have been produced by the same compiler
15467 binary as the current compilation is using.
15470 Any macros defined before the precompiled header is included must
15471 either be defined in the same way as when the precompiled header was
15472 generated, or must not affect the precompiled header, which usually
15473 means that they don't appear in the precompiled header at all.
15475 The @option{-D} option is one way to define a macro before a
15476 precompiled header is included; using a @code{#define} can also do it.
15477 There are also some options that define macros implicitly, like
15478 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
15481 @item If debugging information is output when using the precompiled
15482 header, using @option{-g} or similar, the same kind of debugging information
15483 must have been output when building the precompiled header. However,
15484 a precompiled header built using @option{-g} can be used in a compilation
15485 when no debugging information is being output.
15487 @item The same @option{-m} options must generally be used when building
15488 and using the precompiled header. @xref{Submodel Options},
15489 for any cases where this rule is relaxed.
15491 @item Each of the following options must be the same when building and using
15492 the precompiled header:
15494 @gccoptlist{-fexceptions -funit-at-a-time}
15497 Some other command-line options starting with @option{-f},
15498 @option{-p}, or @option{-O} must be defined in the same way as when
15499 the precompiled header was generated. At present, it's not clear
15500 which options are safe to change and which are not; the safest choice
15501 is to use exactly the same options when generating and using the
15502 precompiled header. The following are known to be safe:
15504 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
15505 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
15506 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
15511 For all of these except the last, the compiler will automatically
15512 ignore the precompiled header if the conditions aren't met. If you
15513 find an option combination that doesn't work and doesn't cause the
15514 precompiled header to be ignored, please consider filing a bug report,
15517 If you do use differing options when generating and using the
15518 precompiled header, the actual behavior will be a mixture of the
15519 behavior for the options. For instance, if you use @option{-g} to
15520 generate the precompiled header but not when using it, you may or may
15521 not get debugging information for routines in the precompiled header.
15523 @node Running Protoize
15524 @section Running Protoize
15526 The program @code{protoize} is an optional part of GCC@. You can use
15527 it to add prototypes to a program, thus converting the program to ISO
15528 C in one respect. The companion program @code{unprotoize} does the
15529 reverse: it removes argument types from any prototypes that are found.
15531 When you run these programs, you must specify a set of source files as
15532 command line arguments. The conversion programs start out by compiling
15533 these files to see what functions they define. The information gathered
15534 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
15536 After scanning comes actual conversion. The specified files are all
15537 eligible to be converted; any files they include (whether sources or
15538 just headers) are eligible as well.
15540 But not all the eligible files are converted. By default,
15541 @code{protoize} and @code{unprotoize} convert only source and header
15542 files in the current directory. You can specify additional directories
15543 whose files should be converted with the @option{-d @var{directory}}
15544 option. You can also specify particular files to exclude with the
15545 @option{-x @var{file}} option. A file is converted if it is eligible, its
15546 directory name matches one of the specified directory names, and its
15547 name within the directory has not been excluded.
15549 Basic conversion with @code{protoize} consists of rewriting most
15550 function definitions and function declarations to specify the types of
15551 the arguments. The only ones not rewritten are those for varargs
15554 @code{protoize} optionally inserts prototype declarations at the
15555 beginning of the source file, to make them available for any calls that
15556 precede the function's definition. Or it can insert prototype
15557 declarations with block scope in the blocks where undeclared functions
15560 Basic conversion with @code{unprotoize} consists of rewriting most
15561 function declarations to remove any argument types, and rewriting
15562 function definitions to the old-style pre-ISO form.
15564 Both conversion programs print a warning for any function declaration or
15565 definition that they can't convert. You can suppress these warnings
15568 The output from @code{protoize} or @code{unprotoize} replaces the
15569 original source file. The original file is renamed to a name ending
15570 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
15571 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
15572 for DOS) file already exists, then the source file is simply discarded.
15574 @code{protoize} and @code{unprotoize} both depend on GCC itself to
15575 scan the program and collect information about the functions it uses.
15576 So neither of these programs will work until GCC is installed.
15578 Here is a table of the options you can use with @code{protoize} and
15579 @code{unprotoize}. Each option works with both programs unless
15583 @item -B @var{directory}
15584 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
15585 usual directory (normally @file{/usr/local/lib}). This file contains
15586 prototype information about standard system functions. This option
15587 applies only to @code{protoize}.
15589 @item -c @var{compilation-options}
15590 Use @var{compilation-options} as the options when running @command{gcc} to
15591 produce the @samp{.X} files. The special option @option{-aux-info} is
15592 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
15594 Note that the compilation options must be given as a single argument to
15595 @code{protoize} or @code{unprotoize}. If you want to specify several
15596 @command{gcc} options, you must quote the entire set of compilation options
15597 to make them a single word in the shell.
15599 There are certain @command{gcc} arguments that you cannot use, because they
15600 would produce the wrong kind of output. These include @option{-g},
15601 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
15602 the @var{compilation-options}, they are ignored.
15605 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
15606 systems) instead of @samp{.c}. This is convenient if you are converting
15607 a C program to C++. This option applies only to @code{protoize}.
15610 Add explicit global declarations. This means inserting explicit
15611 declarations at the beginning of each source file for each function
15612 that is called in the file and was not declared. These declarations
15613 precede the first function definition that contains a call to an
15614 undeclared function. This option applies only to @code{protoize}.
15616 @item -i @var{string}
15617 Indent old-style parameter declarations with the string @var{string}.
15618 This option applies only to @code{protoize}.
15620 @code{unprotoize} converts prototyped function definitions to old-style
15621 function definitions, where the arguments are declared between the
15622 argument list and the initial @samp{@{}. By default, @code{unprotoize}
15623 uses five spaces as the indentation. If you want to indent with just
15624 one space instead, use @option{-i " "}.
15627 Keep the @samp{.X} files. Normally, they are deleted after conversion
15631 Add explicit local declarations. @code{protoize} with @option{-l} inserts
15632 a prototype declaration for each function in each block which calls the
15633 function without any declaration. This option applies only to
15637 Make no real changes. This mode just prints information about the conversions
15638 that would have been done without @option{-n}.
15641 Make no @samp{.save} files. The original files are simply deleted.
15642 Use this option with caution.
15644 @item -p @var{program}
15645 Use the program @var{program} as the compiler. Normally, the name
15646 @file{gcc} is used.
15649 Work quietly. Most warnings are suppressed.
15652 Print the version number, just like @option{-v} for @command{gcc}.
15655 If you need special compiler options to compile one of your program's
15656 source files, then you should generate that file's @samp{.X} file
15657 specially, by running @command{gcc} on that source file with the
15658 appropriate options and the option @option{-aux-info}. Then run
15659 @code{protoize} on the entire set of files. @code{protoize} will use
15660 the existing @samp{.X} file because it is newer than the source file.
15664 gcc -Dfoo=bar file1.c -aux-info file1.X
15669 You need to include the special files along with the rest in the
15670 @code{protoize} command, even though their @samp{.X} files already
15671 exist, because otherwise they won't get converted.
15673 @xref{Protoize Caveats}, for more information on how to use
15674 @code{protoize} successfully.