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, 2009
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, 2009
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{-dv} 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.
156 @section Option Summary
158 Here is a summary of all the options, grouped by type. Explanations are
159 in the following sections.
162 @item Overall Options
163 @xref{Overall Options,,Options Controlling the Kind of Output}.
164 @gccoptlist{-c -S -E -o @var{file} -combine -no-canonical-prefixes @gol
165 -pipe -pass-exit-codes @gol
166 -x @var{language} -v -### --help@r{[}=@var{class}@r{[},@dots{}@r{]]} --target-help @gol
167 --version -wrapper@@@var{file} -fplugin=@var{file} -fplugin-arg-@var{name}=@var{arg}}
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 -fno-pretty-templates @gol
193 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
194 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
195 -fno-default-inline -fvisibility-inlines-hidden @gol
196 -fvisibility-ms-compat @gol
197 -Wabi -Wctor-dtor-privacy @gol
198 -Wnon-virtual-dtor -Wreorder @gol
199 -Weffc++ -Wstrict-null-sentinel @gol
200 -Wno-non-template-friend -Wold-style-cast @gol
201 -Woverloaded-virtual -Wno-pmf-conversions @gol
204 @item Objective-C and Objective-C++ Language Options
205 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
206 Objective-C and Objective-C++ Dialects}.
207 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
208 -fgnu-runtime -fnext-runtime @gol
209 -fno-nil-receivers @gol
210 -fobjc-call-cxx-cdtors @gol
211 -fobjc-direct-dispatch @gol
212 -fobjc-exceptions @gol
214 -freplace-objc-classes @gol
217 -Wassign-intercept @gol
218 -Wno-protocol -Wselector @gol
219 -Wstrict-selector-match @gol
220 -Wundeclared-selector}
222 @item Language Independent Options
223 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
224 @gccoptlist{-fmessage-length=@var{n} @gol
225 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
226 -fdiagnostics-show-option}
228 @item Warning Options
229 @xref{Warning Options,,Options to Request or Suppress Warnings}.
230 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
231 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
232 -Wno-attributes -Wno-builtin-macro-redefined @gol
233 -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
234 -Wchar-subscripts -Wclobbered -Wcomment @gol
235 -Wconversion -Wcoverage-mismatch -Wno-deprecated @gol
236 -Wno-deprecated-declarations -Wdisabled-optimization @gol
237 -Wno-div-by-zero -Wempty-body -Wenum-compare -Wno-endif-labels @gol
238 -Werror -Werror=* @gol
239 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
240 -Wno-format-contains-nul -Wno-format-extra-args -Wformat-nonliteral @gol
241 -Wformat-security -Wformat-y2k @gol
242 -Wframe-larger-than=@var{len} -Wjump-misses-init -Wignored-qualifiers @gol
243 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
244 -Winit-self -Winline @gol
245 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
246 -Winvalid-pch -Wlarger-than=@var{len} -Wunsafe-loop-optimizations @gol
247 -Wlogical-op -Wlong-long @gol
248 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
249 -Wmissing-format-attribute -Wmissing-include-dirs @gol
250 -Wmissing-noreturn -Wno-mudflap @gol
251 -Wno-multichar -Wnonnull -Wno-overflow @gol
252 -Woverlength-strings -Wpacked -Wpacked-bitfield-compat -Wpadded @gol
253 -Wparentheses -Wpedantic-ms-format -Wno-pedantic-ms-format @gol
254 -Wpointer-arith -Wno-pointer-to-int-cast @gol
255 -Wredundant-decls @gol
256 -Wreturn-type -Wsequence-point -Wshadow @gol
257 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
258 -Wstrict-aliasing -Wstrict-aliasing=n @gol
259 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
260 -Wswitch -Wswitch-default -Wswitch-enum -Wsync-nand @gol
261 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
262 -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol
263 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
264 -Wunused-label -Wunused-parameter -Wno-unused-result -Wunused-value -Wunused-variable @gol
265 -Wvariadic-macros -Wvla @gol
266 -Wvolatile-register-var -Wwrite-strings}
268 @item C and Objective-C-only Warning Options
269 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
270 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
271 -Wold-style-declaration -Wold-style-definition @gol
272 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
273 -Wdeclaration-after-statement -Wpointer-sign}
275 @item Debugging Options
276 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
277 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
278 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
279 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
280 -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
281 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
282 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
283 -fdump-statistics @gol
285 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
287 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
289 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
290 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
291 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
293 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
294 -fdump-tree-phiprop@r{[}-@var{n}@r{]} @gol
295 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
296 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
297 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
298 -fdump-tree-nrv -fdump-tree-vect @gol
299 -fdump-tree-sink @gol
300 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
301 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
302 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
303 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
304 -ftree-vectorizer-verbose=@var{n} @gol
305 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
306 -fdump-final-insns=@var{file} @gol
307 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
308 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
309 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
310 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
311 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
312 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
313 -ftest-coverage -ftime-report -fvar-tracking @gol
314 -fvar-tracking-assigments -fvar-tracking-assignments-toggle @gol
315 -g -g@var{level} -gtoggle -gcoff -gdwarf-@var{version} @gol
316 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
317 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
318 -fdebug-prefix-map=@var{old}=@var{new} @gol
319 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
320 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
321 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
322 -print-multi-directory -print-multi-lib @gol
323 -print-prog-name=@var{program} -print-search-dirs -Q @gol
324 -print-sysroot -print-sysroot-headers-suffix @gol
325 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
327 @item Optimization Options
328 @xref{Optimize Options,,Options that Control Optimization}.
330 -falign-functions[=@var{n}] -falign-jumps[=@var{n}] @gol
331 -falign-labels[=@var{n}] -falign-loops[=@var{n}] -fassociative-math @gol
332 -fauto-inc-dec -fbranch-probabilities -fbranch-target-load-optimize @gol
333 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fcaller-saves @gol
334 -fcheck-data-deps -fconserve-stack -fcprop-registers -fcrossjumping @gol
335 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules -fcx-limited-range @gol
336 -fdata-sections -fdce -fdce @gol
337 -fdelayed-branch -fdelete-null-pointer-checks -fdse -fdse @gol
338 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffast-math @gol
339 -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
340 -fforward-propagate -ffunction-sections @gol
341 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol
342 -fgcse-sm -fif-conversion -fif-conversion2 -findirect-inlining @gol
343 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
344 -finline-small-functions -fipa-cp -fipa-cp-clone -fipa-matrix-reorg -fipa-pta @gol
345 -fipa-pure-const -fipa-reference -fipa-struct-reorg @gol
346 -fipa-type-escape -fira-algorithm=@var{algorithm} @gol
347 -fira-region=@var{region} -fira-coalesce -fno-ira-share-save-slots @gol
348 -fno-ira-share-spill-slots -fira-verbose=@var{n} @gol
349 -fivopts -fkeep-inline-functions -fkeep-static-consts @gol
350 -floop-block -floop-interchange -floop-strip-mine -fgraphite-identity @gol
351 -floop-parallelize-all @gol
352 -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol
353 -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol
354 -fmudflapir -fmudflapth -fno-branch-count-reg -fno-default-inline @gol
355 -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
356 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
357 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
358 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
359 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
360 -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
361 -fprofile-correction -fprofile-dir=@var{path} -fprofile-generate @gol
362 -fprofile-generate=@var{path} @gol
363 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
364 -freciprocal-math -fregmove -frename-registers -freorder-blocks @gol
365 -freorder-blocks-and-partition -freorder-functions @gol
366 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
367 -frounding-math -fsched2-use-superblocks @gol
368 -fsched2-use-traces -fsched-pressure @gol
369 -fsched-spec-load -fsched-spec-load-dangerous @gol
370 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
371 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
372 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
373 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
374 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
375 -fselective-scheduling -fselective-scheduling2 @gol
376 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
377 -fsignaling-nans -fsingle-precision-constant -fsplit-ivs-in-unroller @gol
378 -fsplit-wide-types -fstack-protector -fstack-protector-all @gol
379 -fstrict-aliasing -fstrict-overflow -fthread-jumps -ftracer @gol
380 -ftree-builtin-call-dce -ftree-ccp -ftree-ch -ftree-copy-prop @gol
381 -ftree-copyrename -ftree-dce @gol
382 -ftree-dominator-opts -ftree-dse -ftree-forwprop -ftree-fre -ftree-loop-im @gol
383 -ftree-phiprop -ftree-loop-distribution @gol
384 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
385 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-pta -ftree-reassoc @gol
386 -ftree-sink -ftree-sra -ftree-switch-conversion @gol
387 -ftree-ter -ftree-vect-loop-version -ftree-vectorize -ftree-vrp @gol
388 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
389 -funsafe-loop-optimizations -funsafe-math-optimizations -funswitch-loops @gol
390 -fvariable-expansion-in-unroller -fvect-cost-model -fvpt -fweb @gol
392 --param @var{name}=@var{value}
393 -O -O0 -O1 -O2 -O3 -Os}
395 @item Preprocessor Options
396 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
397 @gccoptlist{-A@var{question}=@var{answer} @gol
398 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
399 -C -dD -dI -dM -dN @gol
400 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
401 -idirafter @var{dir} @gol
402 -include @var{file} -imacros @var{file} @gol
403 -iprefix @var{file} -iwithprefix @var{dir} @gol
404 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
405 -imultilib @var{dir} -isysroot @var{dir} @gol
406 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
407 -P -fworking-directory -remap @gol
408 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
409 -Xpreprocessor @var{option}}
411 @item Assembler Option
412 @xref{Assembler Options,,Passing Options to the Assembler}.
413 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
416 @xref{Link Options,,Options for Linking}.
417 @gccoptlist{@var{object-file-name} -l@var{library} @gol
418 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
419 -s -static -static-libgcc -static-libstdc++ -shared @gol
420 -shared-libgcc -symbolic @gol
421 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
424 @item Directory Options
425 @xref{Directory Options,,Options for Directory Search}.
426 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
427 -specs=@var{file} -I- --sysroot=@var{dir}}
430 @c I wrote this xref this way to avoid overfull hbox. -- rms
431 @xref{Target Options}.
432 @gccoptlist{-V @var{version} -b @var{machine}}
434 @item Machine Dependent Options
435 @xref{Submodel Options,,Hardware Models and Configurations}.
436 @c This list is ordered alphanumerically by subsection name.
437 @c Try and put the significant identifier (CPU or system) first,
438 @c so users have a clue at guessing where the ones they want will be.
441 @gccoptlist{-EB -EL @gol
442 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
443 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
446 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
447 -mabi=@var{name} @gol
448 -mapcs-stack-check -mno-apcs-stack-check @gol
449 -mapcs-float -mno-apcs-float @gol
450 -mapcs-reentrant -mno-apcs-reentrant @gol
451 -msched-prolog -mno-sched-prolog @gol
452 -mlittle-endian -mbig-endian -mwords-little-endian @gol
453 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
454 -mfp16-format=@var{name}
455 -mthumb-interwork -mno-thumb-interwork @gol
456 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
457 -mstructure-size-boundary=@var{n} @gol
458 -mabort-on-noreturn @gol
459 -mlong-calls -mno-long-calls @gol
460 -msingle-pic-base -mno-single-pic-base @gol
461 -mpic-register=@var{reg} @gol
462 -mnop-fun-dllimport @gol
463 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
464 -mpoke-function-name @gol
466 -mtpcs-frame -mtpcs-leaf-frame @gol
467 -mcaller-super-interworking -mcallee-super-interworking @gol
469 -mword-relocations @gol
470 -mfix-cortex-m3-ldrd}
473 @gccoptlist{-mmcu=@var{mcu} -msize -mno-interrupts @gol
474 -mcall-prologues -mtiny-stack -mint8}
476 @emph{Blackfin Options}
477 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
478 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
479 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
480 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
481 -mno-id-shared-library -mshared-library-id=@var{n} @gol
482 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
483 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
484 -mfast-fp -minline-plt -mmulticore -mcorea -mcoreb -msdram @gol
488 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
489 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
490 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
491 -mstack-align -mdata-align -mconst-align @gol
492 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
493 -melf -maout -melinux -mlinux -sim -sim2 @gol
494 -mmul-bug-workaround -mno-mul-bug-workaround}
497 @gccoptlist{-mmac -mpush-args}
499 @emph{Darwin Options}
500 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
501 -arch_only -bind_at_load -bundle -bundle_loader @gol
502 -client_name -compatibility_version -current_version @gol
504 -dependency-file -dylib_file -dylinker_install_name @gol
505 -dynamic -dynamiclib -exported_symbols_list @gol
506 -filelist -flat_namespace -force_cpusubtype_ALL @gol
507 -force_flat_namespace -headerpad_max_install_names @gol
509 -image_base -init -install_name -keep_private_externs @gol
510 -multi_module -multiply_defined -multiply_defined_unused @gol
511 -noall_load -no_dead_strip_inits_and_terms @gol
512 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
513 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
514 -private_bundle -read_only_relocs -sectalign @gol
515 -sectobjectsymbols -whyload -seg1addr @gol
516 -sectcreate -sectobjectsymbols -sectorder @gol
517 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
518 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
519 -segprot -segs_read_only_addr -segs_read_write_addr @gol
520 -single_module -static -sub_library -sub_umbrella @gol
521 -twolevel_namespace -umbrella -undefined @gol
522 -unexported_symbols_list -weak_reference_mismatches @gol
523 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
524 -mkernel -mone-byte-bool}
526 @emph{DEC Alpha Options}
527 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
528 -mieee -mieee-with-inexact -mieee-conformant @gol
529 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
530 -mtrap-precision=@var{mode} -mbuild-constants @gol
531 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
532 -mbwx -mmax -mfix -mcix @gol
533 -mfloat-vax -mfloat-ieee @gol
534 -mexplicit-relocs -msmall-data -mlarge-data @gol
535 -msmall-text -mlarge-text @gol
536 -mmemory-latency=@var{time}}
538 @emph{DEC Alpha/VMS Options}
539 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
542 @gccoptlist{-msmall-model -mno-lsim}
545 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
546 -mhard-float -msoft-float @gol
547 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
548 -mdouble -mno-double @gol
549 -mmedia -mno-media -mmuladd -mno-muladd @gol
550 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
551 -mlinked-fp -mlong-calls -malign-labels @gol
552 -mlibrary-pic -macc-4 -macc-8 @gol
553 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
554 -moptimize-membar -mno-optimize-membar @gol
555 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
556 -mvliw-branch -mno-vliw-branch @gol
557 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
558 -mno-nested-cond-exec -mtomcat-stats @gol
562 @emph{GNU/Linux Options}
563 @gccoptlist{-muclibc}
565 @emph{H8/300 Options}
566 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
569 @gccoptlist{-march=@var{architecture-type} @gol
570 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
571 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
572 -mfixed-range=@var{register-range} @gol
573 -mjump-in-delay -mlinker-opt -mlong-calls @gol
574 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
575 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
576 -mno-jump-in-delay -mno-long-load-store @gol
577 -mno-portable-runtime -mno-soft-float @gol
578 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
579 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
580 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
581 -munix=@var{unix-std} -nolibdld -static -threads}
583 @emph{i386 and x86-64 Options}
584 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
585 -mfpmath=@var{unit} @gol
586 -masm=@var{dialect} -mno-fancy-math-387 @gol
587 -mno-fp-ret-in-387 -msoft-float @gol
588 -mno-wide-multiply -mrtd -malign-double @gol
589 -mpreferred-stack-boundary=@var{num}
590 -mincoming-stack-boundary=@var{num}
591 -mcld -mcx16 -msahf -mmovbe -mcrc32 -mrecip @gol
592 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
594 -msse4a -m3dnow -mpopcnt -mabm @gol
595 -mthreads -mno-align-stringops -minline-all-stringops @gol
596 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
597 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
598 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
599 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
600 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
601 -mcmodel=@var{code-model} -mabi=@var{name} @gol
602 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
606 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
607 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
608 -mconstant-gp -mauto-pic -mfused-madd @gol
609 -minline-float-divide-min-latency @gol
610 -minline-float-divide-max-throughput @gol
611 -mno-inline-float-divide @gol
612 -minline-int-divide-min-latency @gol
613 -minline-int-divide-max-throughput @gol
614 -mno-inline-int-divide @gol
615 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
616 -mno-inline-sqrt @gol
617 -mdwarf2-asm -mearly-stop-bits @gol
618 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
619 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
620 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
621 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
622 -msched-spec-ldc -msched-spec-control-ldc @gol
623 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
624 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
625 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
626 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
628 @emph{IA-64/VMS Options}
629 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
631 @emph{M32R/D Options}
632 @gccoptlist{-m32r2 -m32rx -m32r @gol
634 -malign-loops -mno-align-loops @gol
635 -missue-rate=@var{number} @gol
636 -mbranch-cost=@var{number} @gol
637 -mmodel=@var{code-size-model-type} @gol
638 -msdata=@var{sdata-type} @gol
639 -mno-flush-func -mflush-func=@var{name} @gol
640 -mno-flush-trap -mflush-trap=@var{number} @gol
644 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
646 @emph{M680x0 Options}
647 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
648 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
649 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
650 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
651 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
652 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
653 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
654 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
657 @emph{M68hc1x Options}
658 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
659 -mauto-incdec -minmax -mlong-calls -mshort @gol
660 -msoft-reg-count=@var{count}}
663 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
664 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
665 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
666 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
667 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
670 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
671 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
672 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
673 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
677 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
678 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 @gol
679 -mips64 -mips64r2 @gol
680 -mips16 -mno-mips16 -mflip-mips16 @gol
681 -minterlink-mips16 -mno-interlink-mips16 @gol
682 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
683 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
684 -mgp32 -mgp64 -mfp32 -mfp64 -mhard-float -msoft-float @gol
685 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
686 -mfpu=@var{fpu-type} @gol
687 -msmartmips -mno-smartmips @gol
688 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
689 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
690 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
691 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
692 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
693 -membedded-data -mno-embedded-data @gol
694 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
695 -mcode-readable=@var{setting} @gol
696 -msplit-addresses -mno-split-addresses @gol
697 -mexplicit-relocs -mno-explicit-relocs @gol
698 -mcheck-zero-division -mno-check-zero-division @gol
699 -mdivide-traps -mdivide-breaks @gol
700 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
701 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
702 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
703 -mfix-r10000 -mno-fix-r10000 -mfix-vr4120 -mno-fix-vr4120 @gol
704 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
705 -mflush-func=@var{func} -mno-flush-func @gol
706 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
707 -mfp-exceptions -mno-fp-exceptions @gol
708 -mvr4130-align -mno-vr4130-align -msynci -mno-synci}
711 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
712 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
713 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
714 -mno-base-addresses -msingle-exit -mno-single-exit}
716 @emph{MN10300 Options}
717 @gccoptlist{-mmult-bug -mno-mult-bug @gol
718 -mam33 -mno-am33 @gol
719 -mam33-2 -mno-am33-2 @gol
720 -mreturn-pointer-on-d0 @gol
723 @emph{PDP-11 Options}
724 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
725 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
726 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
727 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
728 -mbranch-expensive -mbranch-cheap @gol
729 -msplit -mno-split -munix-asm -mdec-asm}
731 @emph{picoChip Options}
732 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N}
733 -msymbol-as-address -mno-inefficient-warnings}
735 @emph{PowerPC Options}
736 See RS/6000 and PowerPC Options.
738 @emph{RS/6000 and PowerPC Options}
739 @gccoptlist{-mcpu=@var{cpu-type} @gol
740 -mtune=@var{cpu-type} @gol
741 -mpower -mno-power -mpower2 -mno-power2 @gol
742 -mpowerpc -mpowerpc64 -mno-powerpc @gol
743 -maltivec -mno-altivec @gol
744 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
745 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
746 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mpopcntd -mno-popcntd @gol
747 -mfprnd -mno-fprnd @gol
748 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
749 -mnew-mnemonics -mold-mnemonics @gol
750 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
751 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
752 -malign-power -malign-natural @gol
753 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
754 -msingle-float -mdouble-float -msimple-fpu @gol
755 -mstring -mno-string -mupdate -mno-update @gol
756 -mavoid-indexed-addresses -mno-avoid-indexed-addresses @gol
757 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
758 -mstrict-align -mno-strict-align -mrelocatable @gol
759 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
760 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
761 -mdynamic-no-pic -maltivec -mswdiv @gol
762 -mprioritize-restricted-insns=@var{priority} @gol
763 -msched-costly-dep=@var{dependence_type} @gol
764 -minsert-sched-nops=@var{scheme} @gol
765 -mcall-sysv -mcall-netbsd @gol
766 -maix-struct-return -msvr4-struct-return @gol
767 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
768 -misel -mno-isel @gol
769 -misel=yes -misel=no @gol
771 -mspe=yes -mspe=no @gol
773 -mgen-cell-microcode -mwarn-cell-microcode @gol
774 -mvrsave -mno-vrsave @gol
775 -mmulhw -mno-mulhw @gol
776 -mdlmzb -mno-dlmzb @gol
777 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
778 -mprototype -mno-prototype @gol
779 -msim -mmvme -mads -myellowknife -memb -msdata @gol
780 -msdata=@var{opt} -mvxworks -G @var{num} -pthread}
782 @emph{S/390 and zSeries Options}
783 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
784 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
785 -mlong-double-64 -mlong-double-128 @gol
786 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
787 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
788 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
789 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
790 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
793 @gccoptlist{-meb -mel @gol
797 -mscore5 -mscore5u -mscore7 -mscore7d}
800 @gccoptlist{-m1 -m2 -m2e @gol
801 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
803 -m4-nofpu -m4-single-only -m4-single -m4 @gol
804 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
805 -m5-64media -m5-64media-nofpu @gol
806 -m5-32media -m5-32media-nofpu @gol
807 -m5-compact -m5-compact-nofpu @gol
808 -mb -ml -mdalign -mrelax @gol
809 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
810 -mieee -mbitops -misize -minline-ic_invalidate -mpadstruct -mspace @gol
811 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
812 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
813 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
817 @gccoptlist{-mcpu=@var{cpu-type} @gol
818 -mtune=@var{cpu-type} @gol
819 -mcmodel=@var{code-model} @gol
820 -m32 -m64 -mapp-regs -mno-app-regs @gol
821 -mfaster-structs -mno-faster-structs @gol
822 -mfpu -mno-fpu -mhard-float -msoft-float @gol
823 -mhard-quad-float -msoft-quad-float @gol
824 -mimpure-text -mno-impure-text -mlittle-endian @gol
825 -mstack-bias -mno-stack-bias @gol
826 -munaligned-doubles -mno-unaligned-doubles @gol
827 -mv8plus -mno-v8plus -mvis -mno-vis
828 -threads -pthreads -pthread}
831 @gccoptlist{-mwarn-reloc -merror-reloc @gol
832 -msafe-dma -munsafe-dma @gol
834 -msmall-mem -mlarge-mem -mstdmain @gol
835 -mfixed-range=@var{register-range}}
837 @emph{System V Options}
838 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
841 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
842 -mprolog-function -mno-prolog-function -mspace @gol
843 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
844 -mapp-regs -mno-app-regs @gol
845 -mdisable-callt -mno-disable-callt @gol
851 @gccoptlist{-mg -mgnu -munix}
853 @emph{VxWorks Options}
854 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
855 -Xbind-lazy -Xbind-now}
857 @emph{x86-64 Options}
858 See i386 and x86-64 Options.
860 @emph{i386 and x86-64 Windows Options}
861 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll
862 -mnop-fun-dllimport -mthread -municode -mwin32 -mwindows}
864 @emph{Xstormy16 Options}
867 @emph{Xtensa Options}
868 @gccoptlist{-mconst16 -mno-const16 @gol
869 -mfused-madd -mno-fused-madd @gol
870 -mserialize-volatile -mno-serialize-volatile @gol
871 -mtext-section-literals -mno-text-section-literals @gol
872 -mtarget-align -mno-target-align @gol
873 -mlongcalls -mno-longcalls}
875 @emph{zSeries Options}
876 See S/390 and zSeries Options.
878 @item Code Generation Options
879 @xref{Code Gen Options,,Options for Code Generation Conventions}.
880 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
881 -ffixed-@var{reg} -fexceptions @gol
882 -fnon-call-exceptions -funwind-tables @gol
883 -fasynchronous-unwind-tables @gol
884 -finhibit-size-directive -finstrument-functions @gol
885 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
886 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
887 -fno-common -fno-ident @gol
888 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
889 -fno-jump-tables @gol
890 -frecord-gcc-switches @gol
891 -freg-struct-return -fshort-enums @gol
892 -fshort-double -fshort-wchar @gol
893 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
894 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
895 -fno-stack-limit -fargument-alias -fargument-noalias @gol
896 -fargument-noalias-global -fargument-noalias-anything @gol
897 -fleading-underscore -ftls-model=@var{model} @gol
898 -ftrapv -fwrapv -fbounds-check @gol
903 * Overall Options:: Controlling the kind of output:
904 an executable, object files, assembler files,
905 or preprocessed source.
906 * C Dialect Options:: Controlling the variant of C language compiled.
907 * C++ Dialect Options:: Variations on C++.
908 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
910 * Language Independent Options:: Controlling how diagnostics should be
912 * Warning Options:: How picky should the compiler be?
913 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
914 * Optimize Options:: How much optimization?
915 * Preprocessor Options:: Controlling header files and macro definitions.
916 Also, getting dependency information for Make.
917 * Assembler Options:: Passing options to the assembler.
918 * Link Options:: Specifying libraries and so on.
919 * Directory Options:: Where to find header files and libraries.
920 Where to find the compiler executable files.
921 * Spec Files:: How to pass switches to sub-processes.
922 * Target Options:: Running a cross-compiler, or an old version of GCC.
925 @node Overall Options
926 @section Options Controlling the Kind of Output
928 Compilation can involve up to four stages: preprocessing, compilation
929 proper, assembly and linking, always in that order. GCC is capable of
930 preprocessing and compiling several files either into several
931 assembler input files, or into one assembler input file; then each
932 assembler input file produces an object file, and linking combines all
933 the object files (those newly compiled, and those specified as input)
934 into an executable file.
936 @cindex file name suffix
937 For any given input file, the file name suffix determines what kind of
942 C source code which must be preprocessed.
945 C source code which should not be preprocessed.
948 C++ source code which should not be preprocessed.
951 Objective-C source code. Note that you must link with the @file{libobjc}
952 library to make an Objective-C program work.
955 Objective-C source code which should not be preprocessed.
959 Objective-C++ source code. Note that you must link with the @file{libobjc}
960 library to make an Objective-C++ program work. Note that @samp{.M} refers
961 to a literal capital M@.
964 Objective-C++ source code which should not be preprocessed.
967 C, C++, Objective-C or Objective-C++ header file to be turned into a
972 @itemx @var{file}.cxx
973 @itemx @var{file}.cpp
974 @itemx @var{file}.CPP
975 @itemx @var{file}.c++
977 C++ source code which must be preprocessed. Note that in @samp{.cxx},
978 the last two letters must both be literally @samp{x}. Likewise,
979 @samp{.C} refers to a literal capital C@.
983 Objective-C++ source code which must be preprocessed.
986 Objective-C++ source code which should not be preprocessed.
991 @itemx @var{file}.hxx
992 @itemx @var{file}.hpp
993 @itemx @var{file}.HPP
994 @itemx @var{file}.h++
995 @itemx @var{file}.tcc
996 C++ header file to be turned into a precompiled header.
999 @itemx @var{file}.for
1000 @itemx @var{file}.ftn
1001 Fixed form Fortran source code which should not be preprocessed.
1004 @itemx @var{file}.FOR
1005 @itemx @var{file}.fpp
1006 @itemx @var{file}.FPP
1007 @itemx @var{file}.FTN
1008 Fixed form Fortran source code which must be preprocessed (with the traditional
1011 @item @var{file}.f90
1012 @itemx @var{file}.f95
1013 @itemx @var{file}.f03
1014 @itemx @var{file}.f08
1015 Free form Fortran source code which should not be preprocessed.
1017 @item @var{file}.F90
1018 @itemx @var{file}.F95
1019 @itemx @var{file}.F03
1020 @itemx @var{file}.F08
1021 Free form Fortran source code which must be preprocessed (with the
1022 traditional preprocessor).
1024 @c FIXME: Descriptions of Java file types.
1030 @item @var{file}.ads
1031 Ada source code file which contains a library unit declaration (a
1032 declaration of a package, subprogram, or generic, or a generic
1033 instantiation), or a library unit renaming declaration (a package,
1034 generic, or subprogram renaming declaration). Such files are also
1037 @item @var{file}.adb
1038 Ada source code file containing a library unit body (a subprogram or
1039 package body). Such files are also called @dfn{bodies}.
1041 @c GCC also knows about some suffixes for languages not yet included:
1052 @itemx @var{file}.sx
1053 Assembler code which must be preprocessed.
1056 An object file to be fed straight into linking.
1057 Any file name with no recognized suffix is treated this way.
1061 You can specify the input language explicitly with the @option{-x} option:
1064 @item -x @var{language}
1065 Specify explicitly the @var{language} for the following input files
1066 (rather than letting the compiler choose a default based on the file
1067 name suffix). This option applies to all following input files until
1068 the next @option{-x} option. Possible values for @var{language} are:
1070 c c-header c-cpp-output
1071 c++ c++-header c++-cpp-output
1072 objective-c objective-c-header objective-c-cpp-output
1073 objective-c++ objective-c++-header objective-c++-cpp-output
1074 assembler assembler-with-cpp
1076 f77 f77-cpp-input f95 f95-cpp-input
1081 Turn off any specification of a language, so that subsequent files are
1082 handled according to their file name suffixes (as they are if @option{-x}
1083 has not been used at all).
1085 @item -pass-exit-codes
1086 @opindex pass-exit-codes
1087 Normally the @command{gcc} program will exit with the code of 1 if any
1088 phase of the compiler returns a non-success return code. If you specify
1089 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1090 numerically highest error produced by any phase that returned an error
1091 indication. The C, C++, and Fortran frontends return 4, if an internal
1092 compiler error is encountered.
1095 If you only want some of the stages of compilation, you can use
1096 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1097 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1098 @command{gcc} is to stop. Note that some combinations (for example,
1099 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1104 Compile or assemble the source files, but do not link. The linking
1105 stage simply is not done. The ultimate output is in the form of an
1106 object file for each source file.
1108 By default, the object file name for a source file is made by replacing
1109 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1111 Unrecognized input files, not requiring compilation or assembly, are
1116 Stop after the stage of compilation proper; do not assemble. The output
1117 is in the form of an assembler code file for each non-assembler input
1120 By default, the assembler file name for a source file is made by
1121 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1123 Input files that don't require compilation are ignored.
1127 Stop after the preprocessing stage; do not run the compiler proper. The
1128 output is in the form of preprocessed source code, which is sent to the
1131 Input files which don't require preprocessing are ignored.
1133 @cindex output file option
1136 Place output in file @var{file}. This applies regardless to whatever
1137 sort of output is being produced, whether it be an executable file,
1138 an object file, an assembler file or preprocessed C code.
1140 If @option{-o} is not specified, the default is to put an executable
1141 file in @file{a.out}, the object file for
1142 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1143 assembler file in @file{@var{source}.s}, a precompiled header file in
1144 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1149 Print (on standard error output) the commands executed to run the stages
1150 of compilation. Also print the version number of the compiler driver
1151 program and of the preprocessor and the compiler proper.
1155 Like @option{-v} except the commands are not executed and all command
1156 arguments are quoted. This is useful for shell scripts to capture the
1157 driver-generated command lines.
1161 Use pipes rather than temporary files for communication between the
1162 various stages of compilation. This fails to work on some systems where
1163 the assembler is unable to read from a pipe; but the GNU assembler has
1168 If you are compiling multiple source files, this option tells the driver
1169 to pass all the source files to the compiler at once (for those
1170 languages for which the compiler can handle this). This will allow
1171 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1172 language for which this is supported is C@. If you pass source files for
1173 multiple languages to the driver, using this option, the driver will invoke
1174 the compiler(s) that support IMA once each, passing each compiler all the
1175 source files appropriate for it. For those languages that do not support
1176 IMA this option will be ignored, and the compiler will be invoked once for
1177 each source file in that language. If you use this option in conjunction
1178 with @option{-save-temps}, the compiler will generate multiple
1180 (one for each source file), but only one (combined) @file{.o} or
1185 Print (on the standard output) a description of the command line options
1186 understood by @command{gcc}. If the @option{-v} option is also specified
1187 then @option{--help} will also be passed on to the various processes
1188 invoked by @command{gcc}, so that they can display the command line options
1189 they accept. If the @option{-Wextra} option has also been specified
1190 (prior to the @option{--help} option), then command line options which
1191 have no documentation associated with them will also be displayed.
1194 @opindex target-help
1195 Print (on the standard output) a description of target-specific command
1196 line options for each tool. For some targets extra target-specific
1197 information may also be printed.
1199 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1200 Print (on the standard output) a description of the command line
1201 options understood by the compiler that fit into all specified classes
1202 and qualifiers. These are the supported classes:
1205 @item @samp{optimizers}
1206 This will display all of the optimization options supported by the
1209 @item @samp{warnings}
1210 This will display all of the options controlling warning messages
1211 produced by the compiler.
1214 This will display target-specific options. Unlike the
1215 @option{--target-help} option however, target-specific options of the
1216 linker and assembler will not be displayed. This is because those
1217 tools do not currently support the extended @option{--help=} syntax.
1220 This will display the values recognized by the @option{--param}
1223 @item @var{language}
1224 This will display the options supported for @var{language}, where
1225 @var{language} is the name of one of the languages supported in this
1229 This will display the options that are common to all languages.
1232 These are the supported qualifiers:
1235 @item @samp{undocumented}
1236 Display only those options which are undocumented.
1239 Display options which take an argument that appears after an equal
1240 sign in the same continuous piece of text, such as:
1241 @samp{--help=target}.
1243 @item @samp{separate}
1244 Display options which take an argument that appears as a separate word
1245 following the original option, such as: @samp{-o output-file}.
1248 Thus for example to display all the undocumented target-specific
1249 switches supported by the compiler the following can be used:
1252 --help=target,undocumented
1255 The sense of a qualifier can be inverted by prefixing it with the
1256 @samp{^} character, so for example to display all binary warning
1257 options (i.e., ones that are either on or off and that do not take an
1258 argument), which have a description the following can be used:
1261 --help=warnings,^joined,^undocumented
1264 The argument to @option{--help=} should not consist solely of inverted
1267 Combining several classes is possible, although this usually
1268 restricts the output by so much that there is nothing to display. One
1269 case where it does work however is when one of the classes is
1270 @var{target}. So for example to display all the target-specific
1271 optimization options the following can be used:
1274 --help=target,optimizers
1277 The @option{--help=} option can be repeated on the command line. Each
1278 successive use will display its requested class of options, skipping
1279 those that have already been displayed.
1281 If the @option{-Q} option appears on the command line before the
1282 @option{--help=} option, then the descriptive text displayed by
1283 @option{--help=} is changed. Instead of describing the displayed
1284 options, an indication is given as to whether the option is enabled,
1285 disabled or set to a specific value (assuming that the compiler
1286 knows this at the point where the @option{--help=} option is used).
1288 Here is a truncated example from the ARM port of @command{gcc}:
1291 % gcc -Q -mabi=2 --help=target -c
1292 The following options are target specific:
1294 -mabort-on-noreturn [disabled]
1298 The output is sensitive to the effects of previous command line
1299 options, so for example it is possible to find out which optimizations
1300 are enabled at @option{-O2} by using:
1303 -Q -O2 --help=optimizers
1306 Alternatively you can discover which binary optimizations are enabled
1307 by @option{-O3} by using:
1310 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1311 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1312 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1315 @item -no-canonical-prefixes
1316 @opindex no-canonical-prefixes
1317 Do not expand any symbolic links, resolve references to @samp{/../}
1318 or @samp{/./}, or make the path absolute when generating a relative
1323 Display the version number and copyrights of the invoked GCC@.
1327 Invoke all subcommands under a wrapper program. It takes a single
1328 comma separated list as an argument, which will be used to invoke
1332 gcc -c t.c -wrapper gdb,--args
1335 This will invoke all subprograms of gcc under "gdb --args",
1336 thus cc1 invocation will be "gdb --args cc1 ...".
1338 @item -fplugin=@var{name}.so
1339 Load the plugin code in file @var{name}.so, assumed to be a
1340 shared object to be dlopen'd by the compiler. The base name of
1341 the shared object file is used to identify the plugin for the
1342 purposes of argument parsing (See
1343 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1344 Each plugin should define the callback functions specified in the
1347 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1348 Define an argument called @var{key} with a value of @var{value}
1349 for the plugin called @var{name}.
1351 @include @value{srcdir}/../libiberty/at-file.texi
1355 @section Compiling C++ Programs
1357 @cindex suffixes for C++ source
1358 @cindex C++ source file suffixes
1359 C++ source files conventionally use one of the suffixes @samp{.C},
1360 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1361 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1362 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1363 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1364 files with these names and compiles them as C++ programs even if you
1365 call the compiler the same way as for compiling C programs (usually
1366 with the name @command{gcc}).
1370 However, the use of @command{gcc} does not add the C++ library.
1371 @command{g++} is a program that calls GCC and treats @samp{.c},
1372 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1373 files unless @option{-x} is used, and automatically specifies linking
1374 against the C++ library. This program is also useful when
1375 precompiling a C header file with a @samp{.h} extension for use in C++
1376 compilations. On many systems, @command{g++} is also installed with
1377 the name @command{c++}.
1379 @cindex invoking @command{g++}
1380 When you compile C++ programs, you may specify many of the same
1381 command-line options that you use for compiling programs in any
1382 language; or command-line options meaningful for C and related
1383 languages; or options that are meaningful only for C++ programs.
1384 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1385 explanations of options for languages related to C@.
1386 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1387 explanations of options that are meaningful only for C++ programs.
1389 @node C Dialect Options
1390 @section Options Controlling C Dialect
1391 @cindex dialect options
1392 @cindex language dialect options
1393 @cindex options, dialect
1395 The following options control the dialect of C (or languages derived
1396 from C, such as C++, Objective-C and Objective-C++) that the compiler
1400 @cindex ANSI support
1404 In C mode, this is equivalent to @samp{-std=c89}. In C++ mode, it is
1405 equivalent to @samp{-std=c++98}.
1407 This turns off certain features of GCC that are incompatible with ISO
1408 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1409 such as the @code{asm} and @code{typeof} keywords, and
1410 predefined macros such as @code{unix} and @code{vax} that identify the
1411 type of system you are using. It also enables the undesirable and
1412 rarely used ISO trigraph feature. For the C compiler,
1413 it disables recognition of C++ style @samp{//} comments as well as
1414 the @code{inline} keyword.
1416 The alternate keywords @code{__asm__}, @code{__extension__},
1417 @code{__inline__} and @code{__typeof__} continue to work despite
1418 @option{-ansi}. You would not want to use them in an ISO C program, of
1419 course, but it is useful to put them in header files that might be included
1420 in compilations done with @option{-ansi}. Alternate predefined macros
1421 such as @code{__unix__} and @code{__vax__} are also available, with or
1422 without @option{-ansi}.
1424 The @option{-ansi} option does not cause non-ISO programs to be
1425 rejected gratuitously. For that, @option{-pedantic} is required in
1426 addition to @option{-ansi}. @xref{Warning Options}.
1428 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1429 option is used. Some header files may notice this macro and refrain
1430 from declaring certain functions or defining certain macros that the
1431 ISO standard doesn't call for; this is to avoid interfering with any
1432 programs that might use these names for other things.
1434 Functions that would normally be built in but do not have semantics
1435 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1436 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1437 built-in functions provided by GCC}, for details of the functions
1442 Determine the language standard. @xref{Standards,,Language Standards
1443 Supported by GCC}, for details of these standard versions. This option
1444 is currently only supported when compiling C or C++.
1446 The compiler can accept several base standards, such as @samp{c89} or
1447 @samp{c++98}, and GNU dialects of those standards, such as
1448 @samp{gnu89} or @samp{gnu++98}. By specifying a base standard, the
1449 compiler will accept all programs following that standard and those
1450 using GNU extensions that do not contradict it. For example,
1451 @samp{-std=c89} turns off certain features of GCC that are
1452 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1453 keywords, but not other GNU extensions that do not have a meaning in
1454 ISO C90, such as omitting the middle term of a @code{?:}
1455 expression. On the other hand, by specifying a GNU dialect of a
1456 standard, all features the compiler support are enabled, even when
1457 those features change the meaning of the base standard and some
1458 strict-conforming programs may be rejected. The particular standard
1459 is used by @option{-pedantic} to identify which features are GNU
1460 extensions given that version of the standard. For example
1461 @samp{-std=gnu89 -pedantic} would warn about C++ style @samp{//}
1462 comments, while @samp{-std=gnu99 -pedantic} would not.
1464 A value for this option must be provided; possible values are
1469 Support all ISO C90 programs (certain GNU extensions that conflict
1470 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1472 @item iso9899:199409
1473 ISO C90 as modified in amendment 1.
1479 ISO C99. Note that this standard is not yet fully supported; see
1480 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1481 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1484 GNU dialect of ISO C90 (including some C99 features). This
1485 is the default for C code.
1489 GNU dialect of ISO C99. When ISO C99 is fully implemented in GCC,
1490 this will become the default. The name @samp{gnu9x} is deprecated.
1493 The 1998 ISO C++ standard plus amendments. Same as @option{-ansi} for
1497 GNU dialect of @option{-std=c++98}. This is the default for
1501 The working draft of the upcoming ISO C++0x standard. This option
1502 enables experimental features that are likely to be included in
1503 C++0x. The working draft is constantly changing, and any feature that is
1504 enabled by this flag may be removed from future versions of GCC if it is
1505 not part of the C++0x standard.
1508 GNU dialect of @option{-std=c++0x}. This option enables
1509 experimental features that may be removed in future versions of GCC.
1512 @item -fgnu89-inline
1513 @opindex fgnu89-inline
1514 The option @option{-fgnu89-inline} tells GCC to use the traditional
1515 GNU semantics for @code{inline} functions when in C99 mode.
1516 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1517 is accepted and ignored by GCC versions 4.1.3 up to but not including
1518 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1519 C99 mode. Using this option is roughly equivalent to adding the
1520 @code{gnu_inline} function attribute to all inline functions
1521 (@pxref{Function Attributes}).
1523 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1524 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1525 specifies the default behavior). This option was first supported in
1526 GCC 4.3. This option is not supported in C89 or gnu89 mode.
1528 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1529 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1530 in effect for @code{inline} functions. @xref{Common Predefined
1531 Macros,,,cpp,The C Preprocessor}.
1533 @item -aux-info @var{filename}
1535 Output to the given filename prototyped declarations for all functions
1536 declared and/or defined in a translation unit, including those in header
1537 files. This option is silently ignored in any language other than C@.
1539 Besides declarations, the file indicates, in comments, the origin of
1540 each declaration (source file and line), whether the declaration was
1541 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1542 @samp{O} for old, respectively, in the first character after the line
1543 number and the colon), and whether it came from a declaration or a
1544 definition (@samp{C} or @samp{F}, respectively, in the following
1545 character). In the case of function definitions, a K&R-style list of
1546 arguments followed by their declarations is also provided, inside
1547 comments, after the declaration.
1551 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1552 keyword, so that code can use these words as identifiers. You can use
1553 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1554 instead. @option{-ansi} implies @option{-fno-asm}.
1556 In C++, this switch only affects the @code{typeof} keyword, since
1557 @code{asm} and @code{inline} are standard keywords. You may want to
1558 use the @option{-fno-gnu-keywords} flag instead, which has the same
1559 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1560 switch only affects the @code{asm} and @code{typeof} keywords, since
1561 @code{inline} is a standard keyword in ISO C99.
1564 @itemx -fno-builtin-@var{function}
1565 @opindex fno-builtin
1566 @cindex built-in functions
1567 Don't recognize built-in functions that do not begin with
1568 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1569 functions provided by GCC}, for details of the functions affected,
1570 including those which are not built-in functions when @option{-ansi} or
1571 @option{-std} options for strict ISO C conformance are used because they
1572 do not have an ISO standard meaning.
1574 GCC normally generates special code to handle certain built-in functions
1575 more efficiently; for instance, calls to @code{alloca} may become single
1576 instructions that adjust the stack directly, and calls to @code{memcpy}
1577 may become inline copy loops. The resulting code is often both smaller
1578 and faster, but since the function calls no longer appear as such, you
1579 cannot set a breakpoint on those calls, nor can you change the behavior
1580 of the functions by linking with a different library. In addition,
1581 when a function is recognized as a built-in function, GCC may use
1582 information about that function to warn about problems with calls to
1583 that function, or to generate more efficient code, even if the
1584 resulting code still contains calls to that function. For example,
1585 warnings are given with @option{-Wformat} for bad calls to
1586 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1587 known not to modify global memory.
1589 With the @option{-fno-builtin-@var{function}} option
1590 only the built-in function @var{function} is
1591 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1592 function is named that is not built-in in this version of GCC, this
1593 option is ignored. There is no corresponding
1594 @option{-fbuiltin-@var{function}} option; if you wish to enable
1595 built-in functions selectively when using @option{-fno-builtin} or
1596 @option{-ffreestanding}, you may define macros such as:
1599 #define abs(n) __builtin_abs ((n))
1600 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1605 @cindex hosted environment
1607 Assert that compilation takes place in a hosted environment. This implies
1608 @option{-fbuiltin}. A hosted environment is one in which the
1609 entire standard library is available, and in which @code{main} has a return
1610 type of @code{int}. Examples are nearly everything except a kernel.
1611 This is equivalent to @option{-fno-freestanding}.
1613 @item -ffreestanding
1614 @opindex ffreestanding
1615 @cindex hosted environment
1617 Assert that compilation takes place in a freestanding environment. This
1618 implies @option{-fno-builtin}. A freestanding environment
1619 is one in which the standard library may not exist, and program startup may
1620 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1621 This is equivalent to @option{-fno-hosted}.
1623 @xref{Standards,,Language Standards Supported by GCC}, for details of
1624 freestanding and hosted environments.
1628 @cindex openmp parallel
1629 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1630 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1631 compiler generates parallel code according to the OpenMP Application
1632 Program Interface v3.0 @w{@uref{http://www.openmp.org/}}. This option
1633 implies @option{-pthread}, and thus is only supported on targets that
1634 have support for @option{-pthread}.
1636 @item -fms-extensions
1637 @opindex fms-extensions
1638 Accept some non-standard constructs used in Microsoft header files.
1640 Some cases of unnamed fields in structures and unions are only
1641 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1642 fields within structs/unions}, for details.
1646 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1647 options for strict ISO C conformance) implies @option{-trigraphs}.
1649 @item -no-integrated-cpp
1650 @opindex no-integrated-cpp
1651 Performs a compilation in two passes: preprocessing and compiling. This
1652 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1653 @option{-B} option. The user supplied compilation step can then add in
1654 an additional preprocessing step after normal preprocessing but before
1655 compiling. The default is to use the integrated cpp (internal cpp)
1657 The semantics of this option will change if "cc1", "cc1plus", and
1658 "cc1obj" are merged.
1660 @cindex traditional C language
1661 @cindex C language, traditional
1663 @itemx -traditional-cpp
1664 @opindex traditional-cpp
1665 @opindex traditional
1666 Formerly, these options caused GCC to attempt to emulate a pre-standard
1667 C compiler. They are now only supported with the @option{-E} switch.
1668 The preprocessor continues to support a pre-standard mode. See the GNU
1669 CPP manual for details.
1671 @item -fcond-mismatch
1672 @opindex fcond-mismatch
1673 Allow conditional expressions with mismatched types in the second and
1674 third arguments. The value of such an expression is void. This option
1675 is not supported for C++.
1677 @item -flax-vector-conversions
1678 @opindex flax-vector-conversions
1679 Allow implicit conversions between vectors with differing numbers of
1680 elements and/or incompatible element types. This option should not be
1683 @item -funsigned-char
1684 @opindex funsigned-char
1685 Let the type @code{char} be unsigned, like @code{unsigned char}.
1687 Each kind of machine has a default for what @code{char} should
1688 be. It is either like @code{unsigned char} by default or like
1689 @code{signed char} by default.
1691 Ideally, a portable program should always use @code{signed char} or
1692 @code{unsigned char} when it depends on the signedness of an object.
1693 But many programs have been written to use plain @code{char} and
1694 expect it to be signed, or expect it to be unsigned, depending on the
1695 machines they were written for. This option, and its inverse, let you
1696 make such a program work with the opposite default.
1698 The type @code{char} is always a distinct type from each of
1699 @code{signed char} or @code{unsigned char}, even though its behavior
1700 is always just like one of those two.
1703 @opindex fsigned-char
1704 Let the type @code{char} be signed, like @code{signed char}.
1706 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1707 the negative form of @option{-funsigned-char}. Likewise, the option
1708 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1710 @item -fsigned-bitfields
1711 @itemx -funsigned-bitfields
1712 @itemx -fno-signed-bitfields
1713 @itemx -fno-unsigned-bitfields
1714 @opindex fsigned-bitfields
1715 @opindex funsigned-bitfields
1716 @opindex fno-signed-bitfields
1717 @opindex fno-unsigned-bitfields
1718 These options control whether a bit-field is signed or unsigned, when the
1719 declaration does not use either @code{signed} or @code{unsigned}. By
1720 default, such a bit-field is signed, because this is consistent: the
1721 basic integer types such as @code{int} are signed types.
1724 @node C++ Dialect Options
1725 @section Options Controlling C++ Dialect
1727 @cindex compiler options, C++
1728 @cindex C++ options, command line
1729 @cindex options, C++
1730 This section describes the command-line options that are only meaningful
1731 for C++ programs; but you can also use most of the GNU compiler options
1732 regardless of what language your program is in. For example, you
1733 might compile a file @code{firstClass.C} like this:
1736 g++ -g -frepo -O -c firstClass.C
1740 In this example, only @option{-frepo} is an option meant
1741 only for C++ programs; you can use the other options with any
1742 language supported by GCC@.
1744 Here is a list of options that are @emph{only} for compiling C++ programs:
1748 @item -fabi-version=@var{n}
1749 @opindex fabi-version
1750 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1751 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1752 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1753 the version that conforms most closely to the C++ ABI specification.
1754 Therefore, the ABI obtained using version 0 will change as ABI bugs
1757 The default is version 2.
1759 @item -fno-access-control
1760 @opindex fno-access-control
1761 Turn off all access checking. This switch is mainly useful for working
1762 around bugs in the access control code.
1766 Check that the pointer returned by @code{operator new} is non-null
1767 before attempting to modify the storage allocated. This check is
1768 normally unnecessary because the C++ standard specifies that
1769 @code{operator new} will only return @code{0} if it is declared
1770 @samp{throw()}, in which case the compiler will always check the
1771 return value even without this option. In all other cases, when
1772 @code{operator new} has a non-empty exception specification, memory
1773 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1774 @samp{new (nothrow)}.
1776 @item -fconserve-space
1777 @opindex fconserve-space
1778 Put uninitialized or runtime-initialized global variables into the
1779 common segment, as C does. This saves space in the executable at the
1780 cost of not diagnosing duplicate definitions. If you compile with this
1781 flag and your program mysteriously crashes after @code{main()} has
1782 completed, you may have an object that is being destroyed twice because
1783 two definitions were merged.
1785 This option is no longer useful on most targets, now that support has
1786 been added for putting variables into BSS without making them common.
1788 @item -fno-deduce-init-list
1789 @opindex fno-deduce-init-list
1790 Disable deduction of a template type parameter as
1791 std::initializer_list from a brace-enclosed initializer list, i.e.
1794 template <class T> auto forward(T t) -> decltype (realfn (t))
1801 forward(@{1,2@}); // call forward<std::initializer_list<int>>
1805 This option is present because this deduction is an extension to the
1806 current specification in the C++0x working draft, and there was
1807 some concern about potential overload resolution problems.
1809 @item -ffriend-injection
1810 @opindex ffriend-injection
1811 Inject friend functions into the enclosing namespace, so that they are
1812 visible outside the scope of the class in which they are declared.
1813 Friend functions were documented to work this way in the old Annotated
1814 C++ Reference Manual, and versions of G++ before 4.1 always worked
1815 that way. However, in ISO C++ a friend function which is not declared
1816 in an enclosing scope can only be found using argument dependent
1817 lookup. This option causes friends to be injected as they were in
1820 This option is for compatibility, and may be removed in a future
1823 @item -fno-elide-constructors
1824 @opindex fno-elide-constructors
1825 The C++ standard allows an implementation to omit creating a temporary
1826 which is only used to initialize another object of the same type.
1827 Specifying this option disables that optimization, and forces G++ to
1828 call the copy constructor in all cases.
1830 @item -fno-enforce-eh-specs
1831 @opindex fno-enforce-eh-specs
1832 Don't generate code to check for violation of exception specifications
1833 at runtime. This option violates the C++ standard, but may be useful
1834 for reducing code size in production builds, much like defining
1835 @samp{NDEBUG}. This does not give user code permission to throw
1836 exceptions in violation of the exception specifications; the compiler
1837 will still optimize based on the specifications, so throwing an
1838 unexpected exception will result in undefined behavior.
1841 @itemx -fno-for-scope
1843 @opindex fno-for-scope
1844 If @option{-ffor-scope} is specified, the scope of variables declared in
1845 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1846 as specified by the C++ standard.
1847 If @option{-fno-for-scope} is specified, the scope of variables declared in
1848 a @i{for-init-statement} extends to the end of the enclosing scope,
1849 as was the case in old versions of G++, and other (traditional)
1850 implementations of C++.
1852 The default if neither flag is given to follow the standard,
1853 but to allow and give a warning for old-style code that would
1854 otherwise be invalid, or have different behavior.
1856 @item -fno-gnu-keywords
1857 @opindex fno-gnu-keywords
1858 Do not recognize @code{typeof} as a keyword, so that code can use this
1859 word as an identifier. You can use the keyword @code{__typeof__} instead.
1860 @option{-ansi} implies @option{-fno-gnu-keywords}.
1862 @item -fno-implicit-templates
1863 @opindex fno-implicit-templates
1864 Never emit code for non-inline templates which are instantiated
1865 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1866 @xref{Template Instantiation}, for more information.
1868 @item -fno-implicit-inline-templates
1869 @opindex fno-implicit-inline-templates
1870 Don't emit code for implicit instantiations of inline templates, either.
1871 The default is to handle inlines differently so that compiles with and
1872 without optimization will need the same set of explicit instantiations.
1874 @item -fno-implement-inlines
1875 @opindex fno-implement-inlines
1876 To save space, do not emit out-of-line copies of inline functions
1877 controlled by @samp{#pragma implementation}. This will cause linker
1878 errors if these functions are not inlined everywhere they are called.
1880 @item -fms-extensions
1881 @opindex fms-extensions
1882 Disable pedantic warnings about constructs used in MFC, such as implicit
1883 int and getting a pointer to member function via non-standard syntax.
1885 @item -fno-nonansi-builtins
1886 @opindex fno-nonansi-builtins
1887 Disable built-in declarations of functions that are not mandated by
1888 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1889 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1891 @item -fno-operator-names
1892 @opindex fno-operator-names
1893 Do not treat the operator name keywords @code{and}, @code{bitand},
1894 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1895 synonyms as keywords.
1897 @item -fno-optional-diags
1898 @opindex fno-optional-diags
1899 Disable diagnostics that the standard says a compiler does not need to
1900 issue. Currently, the only such diagnostic issued by G++ is the one for
1901 a name having multiple meanings within a class.
1904 @opindex fpermissive
1905 Downgrade some diagnostics about nonconformant code from errors to
1906 warnings. Thus, using @option{-fpermissive} will allow some
1907 nonconforming code to compile.
1909 @item -fno-pretty-templates
1910 @opindex fno-pretty-templates
1911 When an error message refers to a specialization of a function
1912 template, the compiler will normally print the signature of the
1913 template followed by the template arguments and any typedefs or
1914 typenames in the signature (e.g. @code{void f(T) [with T = int]}
1915 rather than @code{void f(int)}) so that it's clear which template is
1916 involved. When an error message refers to a specialization of a class
1917 template, the compiler will omit any template arguments which match
1918 the default template arguments for that template. If either of these
1919 behaviors make it harder to understand the error message rather than
1920 easier, using @option{-fno-pretty-templates} will disable them.
1924 Enable automatic template instantiation at link time. This option also
1925 implies @option{-fno-implicit-templates}. @xref{Template
1926 Instantiation}, for more information.
1930 Disable generation of information about every class with virtual
1931 functions for use by the C++ runtime type identification features
1932 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1933 of the language, you can save some space by using this flag. Note that
1934 exception handling uses the same information, but it will generate it as
1935 needed. The @samp{dynamic_cast} operator can still be used for casts that
1936 do not require runtime type information, i.e.@: casts to @code{void *} or to
1937 unambiguous base classes.
1941 Emit statistics about front-end processing at the end of the compilation.
1942 This information is generally only useful to the G++ development team.
1944 @item -ftemplate-depth-@var{n}
1945 @opindex ftemplate-depth
1946 Set the maximum instantiation depth for template classes to @var{n}.
1947 A limit on the template instantiation depth is needed to detect
1948 endless recursions during template class instantiation. ANSI/ISO C++
1949 conforming programs must not rely on a maximum depth greater than 17
1950 (changed to 1024 in C++0x).
1952 @item -fno-threadsafe-statics
1953 @opindex fno-threadsafe-statics
1954 Do not emit the extra code to use the routines specified in the C++
1955 ABI for thread-safe initialization of local statics. You can use this
1956 option to reduce code size slightly in code that doesn't need to be
1959 @item -fuse-cxa-atexit
1960 @opindex fuse-cxa-atexit
1961 Register destructors for objects with static storage duration with the
1962 @code{__cxa_atexit} function rather than the @code{atexit} function.
1963 This option is required for fully standards-compliant handling of static
1964 destructors, but will only work if your C library supports
1965 @code{__cxa_atexit}.
1967 @item -fno-use-cxa-get-exception-ptr
1968 @opindex fno-use-cxa-get-exception-ptr
1969 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
1970 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
1971 if the runtime routine is not available.
1973 @item -fvisibility-inlines-hidden
1974 @opindex fvisibility-inlines-hidden
1975 This switch declares that the user does not attempt to compare
1976 pointers to inline methods where the addresses of the two functions
1977 were taken in different shared objects.
1979 The effect of this is that GCC may, effectively, mark inline methods with
1980 @code{__attribute__ ((visibility ("hidden")))} so that they do not
1981 appear in the export table of a DSO and do not require a PLT indirection
1982 when used within the DSO@. Enabling this option can have a dramatic effect
1983 on load and link times of a DSO as it massively reduces the size of the
1984 dynamic export table when the library makes heavy use of templates.
1986 The behavior of this switch is not quite the same as marking the
1987 methods as hidden directly, because it does not affect static variables
1988 local to the function or cause the compiler to deduce that
1989 the function is defined in only one shared object.
1991 You may mark a method as having a visibility explicitly to negate the
1992 effect of the switch for that method. For example, if you do want to
1993 compare pointers to a particular inline method, you might mark it as
1994 having default visibility. Marking the enclosing class with explicit
1995 visibility will have no effect.
1997 Explicitly instantiated inline methods are unaffected by this option
1998 as their linkage might otherwise cross a shared library boundary.
1999 @xref{Template Instantiation}.
2001 @item -fvisibility-ms-compat
2002 @opindex fvisibility-ms-compat
2003 This flag attempts to use visibility settings to make GCC's C++
2004 linkage model compatible with that of Microsoft Visual Studio.
2006 The flag makes these changes to GCC's linkage model:
2010 It sets the default visibility to @code{hidden}, like
2011 @option{-fvisibility=hidden}.
2014 Types, but not their members, are not hidden by default.
2017 The One Definition Rule is relaxed for types without explicit
2018 visibility specifications which are defined in more than one different
2019 shared object: those declarations are permitted if they would have
2020 been permitted when this option was not used.
2023 In new code it is better to use @option{-fvisibility=hidden} and
2024 export those classes which are intended to be externally visible.
2025 Unfortunately it is possible for code to rely, perhaps accidentally,
2026 on the Visual Studio behavior.
2028 Among the consequences of these changes are that static data members
2029 of the same type with the same name but defined in different shared
2030 objects will be different, so changing one will not change the other;
2031 and that pointers to function members defined in different shared
2032 objects may not compare equal. When this flag is given, it is a
2033 violation of the ODR to define types with the same name differently.
2037 Do not use weak symbol support, even if it is provided by the linker.
2038 By default, G++ will use weak symbols if they are available. This
2039 option exists only for testing, and should not be used by end-users;
2040 it will result in inferior code and has no benefits. This option may
2041 be removed in a future release of G++.
2045 Do not search for header files in the standard directories specific to
2046 C++, but do still search the other standard directories. (This option
2047 is used when building the C++ library.)
2050 In addition, these optimization, warning, and code generation options
2051 have meanings only for C++ programs:
2054 @item -fno-default-inline
2055 @opindex fno-default-inline
2056 Do not assume @samp{inline} for functions defined inside a class scope.
2057 @xref{Optimize Options,,Options That Control Optimization}. Note that these
2058 functions will have linkage like inline functions; they just won't be
2061 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2064 Warn when G++ generates code that is probably not compatible with the
2065 vendor-neutral C++ ABI@. Although an effort has been made to warn about
2066 all such cases, there are probably some cases that are not warned about,
2067 even though G++ is generating incompatible code. There may also be
2068 cases where warnings are emitted even though the code that is generated
2071 You should rewrite your code to avoid these warnings if you are
2072 concerned about the fact that code generated by G++ may not be binary
2073 compatible with code generated by other compilers.
2075 The known incompatibilities at this point include:
2080 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
2081 pack data into the same byte as a base class. For example:
2084 struct A @{ virtual void f(); int f1 : 1; @};
2085 struct B : public A @{ int f2 : 1; @};
2089 In this case, G++ will place @code{B::f2} into the same byte
2090 as@code{A::f1}; other compilers will not. You can avoid this problem
2091 by explicitly padding @code{A} so that its size is a multiple of the
2092 byte size on your platform; that will cause G++ and other compilers to
2093 layout @code{B} identically.
2096 Incorrect handling of tail-padding for virtual bases. G++ does not use
2097 tail padding when laying out virtual bases. For example:
2100 struct A @{ virtual void f(); char c1; @};
2101 struct B @{ B(); char c2; @};
2102 struct C : public A, public virtual B @{@};
2106 In this case, G++ will not place @code{B} into the tail-padding for
2107 @code{A}; other compilers will. You can avoid this problem by
2108 explicitly padding @code{A} so that its size is a multiple of its
2109 alignment (ignoring virtual base classes); that will cause G++ and other
2110 compilers to layout @code{C} identically.
2113 Incorrect handling of bit-fields with declared widths greater than that
2114 of their underlying types, when the bit-fields appear in a union. For
2118 union U @{ int i : 4096; @};
2122 Assuming that an @code{int} does not have 4096 bits, G++ will make the
2123 union too small by the number of bits in an @code{int}.
2126 Empty classes can be placed at incorrect offsets. For example:
2136 struct C : public B, public A @{@};
2140 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
2141 it should be placed at offset zero. G++ mistakenly believes that the
2142 @code{A} data member of @code{B} is already at offset zero.
2145 Names of template functions whose types involve @code{typename} or
2146 template template parameters can be mangled incorrectly.
2149 template <typename Q>
2150 void f(typename Q::X) @{@}
2152 template <template <typename> class Q>
2153 void f(typename Q<int>::X) @{@}
2157 Instantiations of these templates may be mangled incorrectly.
2161 It also warns psABI related changes. The known psABI changes at this
2167 For SYSV/x86-64, when passing union with long double, it is changed to
2168 pass in memory as specified in psABI. For example:
2178 @code{union U} will always be passed in memory.
2182 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2183 @opindex Wctor-dtor-privacy
2184 @opindex Wno-ctor-dtor-privacy
2185 Warn when a class seems unusable because all the constructors or
2186 destructors in that class are private, and it has neither friends nor
2187 public static member functions.
2189 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2190 @opindex Wnon-virtual-dtor
2191 @opindex Wno-non-virtual-dtor
2192 Warn when a class has virtual functions and accessible non-virtual
2193 destructor, in which case it would be possible but unsafe to delete
2194 an instance of a derived class through a pointer to the base class.
2195 This warning is also enabled if -Weffc++ is specified.
2197 @item -Wreorder @r{(C++ and Objective-C++ only)}
2199 @opindex Wno-reorder
2200 @cindex reordering, warning
2201 @cindex warning for reordering of member initializers
2202 Warn when the order of member initializers given in the code does not
2203 match the order in which they must be executed. For instance:
2209 A(): j (0), i (1) @{ @}
2213 The compiler will rearrange the member initializers for @samp{i}
2214 and @samp{j} to match the declaration order of the members, emitting
2215 a warning to that effect. This warning is enabled by @option{-Wall}.
2218 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2221 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2224 Warn about violations of the following style guidelines from Scott Meyers'
2225 @cite{Effective C++} book:
2229 Item 11: Define a copy constructor and an assignment operator for classes
2230 with dynamically allocated memory.
2233 Item 12: Prefer initialization to assignment in constructors.
2236 Item 14: Make destructors virtual in base classes.
2239 Item 15: Have @code{operator=} return a reference to @code{*this}.
2242 Item 23: Don't try to return a reference when you must return an object.
2246 Also warn about violations of the following style guidelines from
2247 Scott Meyers' @cite{More Effective C++} book:
2251 Item 6: Distinguish between prefix and postfix forms of increment and
2252 decrement operators.
2255 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2259 When selecting this option, be aware that the standard library
2260 headers do not obey all of these guidelines; use @samp{grep -v}
2261 to filter out those warnings.
2263 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2264 @opindex Wstrict-null-sentinel
2265 @opindex Wno-strict-null-sentinel
2266 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2267 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2268 to @code{__null}. Although it is a null pointer constant not a null pointer,
2269 it is guaranteed to be of the same size as a pointer. But this use is
2270 not portable across different compilers.
2272 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2273 @opindex Wno-non-template-friend
2274 @opindex Wnon-template-friend
2275 Disable warnings when non-templatized friend functions are declared
2276 within a template. Since the advent of explicit template specification
2277 support in G++, if the name of the friend is an unqualified-id (i.e.,
2278 @samp{friend foo(int)}), the C++ language specification demands that the
2279 friend declare or define an ordinary, nontemplate function. (Section
2280 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2281 could be interpreted as a particular specialization of a templatized
2282 function. Because this non-conforming behavior is no longer the default
2283 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2284 check existing code for potential trouble spots and is on by default.
2285 This new compiler behavior can be turned off with
2286 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2287 but disables the helpful warning.
2289 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2290 @opindex Wold-style-cast
2291 @opindex Wno-old-style-cast
2292 Warn if an old-style (C-style) cast to a non-void type is used within
2293 a C++ program. The new-style casts (@samp{dynamic_cast},
2294 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2295 less vulnerable to unintended effects and much easier to search for.
2297 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2298 @opindex Woverloaded-virtual
2299 @opindex Wno-overloaded-virtual
2300 @cindex overloaded virtual fn, warning
2301 @cindex warning for overloaded virtual fn
2302 Warn when a function declaration hides virtual functions from a
2303 base class. For example, in:
2310 struct B: public A @{
2315 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2323 will fail to compile.
2325 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
2326 @opindex Wno-pmf-conversions
2327 @opindex Wpmf-conversions
2328 Disable the diagnostic for converting a bound pointer to member function
2331 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
2332 @opindex Wsign-promo
2333 @opindex Wno-sign-promo
2334 Warn when overload resolution chooses a promotion from unsigned or
2335 enumerated type to a signed type, over a conversion to an unsigned type of
2336 the same size. Previous versions of G++ would try to preserve
2337 unsignedness, but the standard mandates the current behavior.
2342 A& operator = (int);
2352 In this example, G++ will synthesize a default @samp{A& operator =
2353 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2356 @node Objective-C and Objective-C++ Dialect Options
2357 @section Options Controlling Objective-C and Objective-C++ Dialects
2359 @cindex compiler options, Objective-C and Objective-C++
2360 @cindex Objective-C and Objective-C++ options, command line
2361 @cindex options, Objective-C and Objective-C++
2362 (NOTE: This manual does not describe the Objective-C and Objective-C++
2363 languages themselves. See @xref{Standards,,Language Standards
2364 Supported by GCC}, for references.)
2366 This section describes the command-line options that are only meaningful
2367 for Objective-C and Objective-C++ programs, but you can also use most of
2368 the language-independent GNU compiler options.
2369 For example, you might compile a file @code{some_class.m} like this:
2372 gcc -g -fgnu-runtime -O -c some_class.m
2376 In this example, @option{-fgnu-runtime} is an option meant only for
2377 Objective-C and Objective-C++ programs; you can use the other options with
2378 any language supported by GCC@.
2380 Note that since Objective-C is an extension of the C language, Objective-C
2381 compilations may also use options specific to the C front-end (e.g.,
2382 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2383 C++-specific options (e.g., @option{-Wabi}).
2385 Here is a list of options that are @emph{only} for compiling Objective-C
2386 and Objective-C++ programs:
2389 @item -fconstant-string-class=@var{class-name}
2390 @opindex fconstant-string-class
2391 Use @var{class-name} as the name of the class to instantiate for each
2392 literal string specified with the syntax @code{@@"@dots{}"}. The default
2393 class name is @code{NXConstantString} if the GNU runtime is being used, and
2394 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2395 @option{-fconstant-cfstrings} option, if also present, will override the
2396 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2397 to be laid out as constant CoreFoundation strings.
2400 @opindex fgnu-runtime
2401 Generate object code compatible with the standard GNU Objective-C
2402 runtime. This is the default for most types of systems.
2404 @item -fnext-runtime
2405 @opindex fnext-runtime
2406 Generate output compatible with the NeXT runtime. This is the default
2407 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2408 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2411 @item -fno-nil-receivers
2412 @opindex fno-nil-receivers
2413 Assume that all Objective-C message dispatches (e.g.,
2414 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2415 is not @code{nil}. This allows for more efficient entry points in the runtime
2416 to be used. Currently, this option is only available in conjunction with
2417 the NeXT runtime on Mac OS X 10.3 and later.
2419 @item -fobjc-call-cxx-cdtors
2420 @opindex fobjc-call-cxx-cdtors
2421 For each Objective-C class, check if any of its instance variables is a
2422 C++ object with a non-trivial default constructor. If so, synthesize a
2423 special @code{- (id) .cxx_construct} instance method that will run
2424 non-trivial default constructors on any such instance variables, in order,
2425 and then return @code{self}. Similarly, check if any instance variable
2426 is a C++ object with a non-trivial destructor, and if so, synthesize a
2427 special @code{- (void) .cxx_destruct} method that will run
2428 all such default destructors, in reverse order.
2430 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2431 thusly generated will only operate on instance variables declared in the
2432 current Objective-C class, and not those inherited from superclasses. It
2433 is the responsibility of the Objective-C runtime to invoke all such methods
2434 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2435 will be invoked by the runtime immediately after a new object
2436 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2437 be invoked immediately before the runtime deallocates an object instance.
2439 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2440 support for invoking the @code{- (id) .cxx_construct} and
2441 @code{- (void) .cxx_destruct} methods.
2443 @item -fobjc-direct-dispatch
2444 @opindex fobjc-direct-dispatch
2445 Allow fast jumps to the message dispatcher. On Darwin this is
2446 accomplished via the comm page.
2448 @item -fobjc-exceptions
2449 @opindex fobjc-exceptions
2450 Enable syntactic support for structured exception handling in Objective-C,
2451 similar to what is offered by C++ and Java. This option is
2452 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2461 @@catch (AnObjCClass *exc) @{
2468 @@catch (AnotherClass *exc) @{
2471 @@catch (id allOthers) @{
2481 The @code{@@throw} statement may appear anywhere in an Objective-C or
2482 Objective-C++ program; when used inside of a @code{@@catch} block, the
2483 @code{@@throw} may appear without an argument (as shown above), in which case
2484 the object caught by the @code{@@catch} will be rethrown.
2486 Note that only (pointers to) Objective-C objects may be thrown and
2487 caught using this scheme. When an object is thrown, it will be caught
2488 by the nearest @code{@@catch} clause capable of handling objects of that type,
2489 analogously to how @code{catch} blocks work in C++ and Java. A
2490 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2491 any and all Objective-C exceptions not caught by previous @code{@@catch}
2494 The @code{@@finally} clause, if present, will be executed upon exit from the
2495 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2496 regardless of whether any exceptions are thrown, caught or rethrown
2497 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2498 of the @code{finally} clause in Java.
2500 There are several caveats to using the new exception mechanism:
2504 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2505 idioms provided by the @code{NSException} class, the new
2506 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2507 systems, due to additional functionality needed in the (NeXT) Objective-C
2511 As mentioned above, the new exceptions do not support handling
2512 types other than Objective-C objects. Furthermore, when used from
2513 Objective-C++, the Objective-C exception model does not interoperate with C++
2514 exceptions at this time. This means you cannot @code{@@throw} an exception
2515 from Objective-C and @code{catch} it in C++, or vice versa
2516 (i.e., @code{throw @dots{} @@catch}).
2519 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2520 blocks for thread-safe execution:
2523 @@synchronized (ObjCClass *guard) @{
2528 Upon entering the @code{@@synchronized} block, a thread of execution shall
2529 first check whether a lock has been placed on the corresponding @code{guard}
2530 object by another thread. If it has, the current thread shall wait until
2531 the other thread relinquishes its lock. Once @code{guard} becomes available,
2532 the current thread will place its own lock on it, execute the code contained in
2533 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2534 making @code{guard} available to other threads).
2536 Unlike Java, Objective-C does not allow for entire methods to be marked
2537 @code{@@synchronized}. Note that throwing exceptions out of
2538 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2539 to be unlocked properly.
2543 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2545 @item -freplace-objc-classes
2546 @opindex freplace-objc-classes
2547 Emit a special marker instructing @command{ld(1)} not to statically link in
2548 the resulting object file, and allow @command{dyld(1)} to load it in at
2549 run time instead. This is used in conjunction with the Fix-and-Continue
2550 debugging mode, where the object file in question may be recompiled and
2551 dynamically reloaded in the course of program execution, without the need
2552 to restart the program itself. Currently, Fix-and-Continue functionality
2553 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2558 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2559 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2560 compile time) with static class references that get initialized at load time,
2561 which improves run-time performance. Specifying the @option{-fzero-link} flag
2562 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2563 to be retained. This is useful in Zero-Link debugging mode, since it allows
2564 for individual class implementations to be modified during program execution.
2568 Dump interface declarations for all classes seen in the source file to a
2569 file named @file{@var{sourcename}.decl}.
2571 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
2572 @opindex Wassign-intercept
2573 @opindex Wno-assign-intercept
2574 Warn whenever an Objective-C assignment is being intercepted by the
2577 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
2578 @opindex Wno-protocol
2580 If a class is declared to implement a protocol, a warning is issued for
2581 every method in the protocol that is not implemented by the class. The
2582 default behavior is to issue a warning for every method not explicitly
2583 implemented in the class, even if a method implementation is inherited
2584 from the superclass. If you use the @option{-Wno-protocol} option, then
2585 methods inherited from the superclass are considered to be implemented,
2586 and no warning is issued for them.
2588 @item -Wselector @r{(Objective-C and Objective-C++ only)}
2590 @opindex Wno-selector
2591 Warn if multiple methods of different types for the same selector are
2592 found during compilation. The check is performed on the list of methods
2593 in the final stage of compilation. Additionally, a check is performed
2594 for each selector appearing in a @code{@@selector(@dots{})}
2595 expression, and a corresponding method for that selector has been found
2596 during compilation. Because these checks scan the method table only at
2597 the end of compilation, these warnings are not produced if the final
2598 stage of compilation is not reached, for example because an error is
2599 found during compilation, or because the @option{-fsyntax-only} option is
2602 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
2603 @opindex Wstrict-selector-match
2604 @opindex Wno-strict-selector-match
2605 Warn if multiple methods with differing argument and/or return types are
2606 found for a given selector when attempting to send a message using this
2607 selector to a receiver of type @code{id} or @code{Class}. When this flag
2608 is off (which is the default behavior), the compiler will omit such warnings
2609 if any differences found are confined to types which share the same size
2612 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
2613 @opindex Wundeclared-selector
2614 @opindex Wno-undeclared-selector
2615 Warn if a @code{@@selector(@dots{})} expression referring to an
2616 undeclared selector is found. A selector is considered undeclared if no
2617 method with that name has been declared before the
2618 @code{@@selector(@dots{})} expression, either explicitly in an
2619 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2620 an @code{@@implementation} section. This option always performs its
2621 checks as soon as a @code{@@selector(@dots{})} expression is found,
2622 while @option{-Wselector} only performs its checks in the final stage of
2623 compilation. This also enforces the coding style convention
2624 that methods and selectors must be declared before being used.
2626 @item -print-objc-runtime-info
2627 @opindex print-objc-runtime-info
2628 Generate C header describing the largest structure that is passed by
2633 @node Language Independent Options
2634 @section Options to Control Diagnostic Messages Formatting
2635 @cindex options to control diagnostics formatting
2636 @cindex diagnostic messages
2637 @cindex message formatting
2639 Traditionally, diagnostic messages have been formatted irrespective of
2640 the output device's aspect (e.g.@: its width, @dots{}). The options described
2641 below can be used to control the diagnostic messages formatting
2642 algorithm, e.g.@: how many characters per line, how often source location
2643 information should be reported. Right now, only the C++ front end can
2644 honor these options. However it is expected, in the near future, that
2645 the remaining front ends would be able to digest them correctly.
2648 @item -fmessage-length=@var{n}
2649 @opindex fmessage-length
2650 Try to format error messages so that they fit on lines of about @var{n}
2651 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2652 the front ends supported by GCC@. If @var{n} is zero, then no
2653 line-wrapping will be done; each error message will appear on a single
2656 @opindex fdiagnostics-show-location
2657 @item -fdiagnostics-show-location=once
2658 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2659 reporter to emit @emph{once} source location information; that is, in
2660 case the message is too long to fit on a single physical line and has to
2661 be wrapped, the source location won't be emitted (as prefix) again,
2662 over and over, in subsequent continuation lines. This is the default
2665 @item -fdiagnostics-show-location=every-line
2666 Only meaningful in line-wrapping mode. Instructs the diagnostic
2667 messages reporter to emit the same source location information (as
2668 prefix) for physical lines that result from the process of breaking
2669 a message which is too long to fit on a single line.
2671 @item -fdiagnostics-show-option
2672 @opindex fdiagnostics-show-option
2673 This option instructs the diagnostic machinery to add text to each
2674 diagnostic emitted, which indicates which command line option directly
2675 controls that diagnostic, when such an option is known to the
2676 diagnostic machinery.
2678 @item -Wcoverage-mismatch
2679 @opindex Wcoverage-mismatch
2680 Warn if feedback profiles do not match when using the
2681 @option{-fprofile-use} option.
2682 If a source file was changed between @option{-fprofile-gen} and
2683 @option{-fprofile-use}, the files with the profile feedback can fail
2684 to match the source file and GCC can not use the profile feedback
2685 information. By default, GCC emits an error message in this case.
2686 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2687 error. GCC does not use appropriate feedback profiles, so using this
2688 option can result in poorly optimized code. This option is useful
2689 only in the case of very minor changes such as bug fixes to an
2694 @node Warning Options
2695 @section Options to Request or Suppress Warnings
2696 @cindex options to control warnings
2697 @cindex warning messages
2698 @cindex messages, warning
2699 @cindex suppressing warnings
2701 Warnings are diagnostic messages that report constructions which
2702 are not inherently erroneous but which are risky or suggest there
2703 may have been an error.
2705 The following language-independent options do not enable specific
2706 warnings but control the kinds of diagnostics produced by GCC.
2709 @cindex syntax checking
2711 @opindex fsyntax-only
2712 Check the code for syntax errors, but don't do anything beyond that.
2716 Inhibit all warning messages.
2721 Make all warnings into errors.
2726 Make the specified warning into an error. The specifier for a warning
2727 is appended, for example @option{-Werror=switch} turns the warnings
2728 controlled by @option{-Wswitch} into errors. This switch takes a
2729 negative form, to be used to negate @option{-Werror} for specific
2730 warnings, for example @option{-Wno-error=switch} makes
2731 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
2732 is in effect. You can use the @option{-fdiagnostics-show-option}
2733 option to have each controllable warning amended with the option which
2734 controls it, to determine what to use with this option.
2736 Note that specifying @option{-Werror=}@var{foo} automatically implies
2737 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
2740 @item -Wfatal-errors
2741 @opindex Wfatal-errors
2742 @opindex Wno-fatal-errors
2743 This option causes the compiler to abort compilation on the first error
2744 occurred rather than trying to keep going and printing further error
2749 You can request many specific warnings with options beginning
2750 @samp{-W}, for example @option{-Wimplicit} to request warnings on
2751 implicit declarations. Each of these specific warning options also
2752 has a negative form beginning @samp{-Wno-} to turn off warnings; for
2753 example, @option{-Wno-implicit}. This manual lists only one of the
2754 two forms, whichever is not the default. For further,
2755 language-specific options also refer to @ref{C++ Dialect Options} and
2756 @ref{Objective-C and Objective-C++ Dialect Options}.
2761 Issue all the warnings demanded by strict ISO C and ISO C++;
2762 reject all programs that use forbidden extensions, and some other
2763 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2764 version of the ISO C standard specified by any @option{-std} option used.
2766 Valid ISO C and ISO C++ programs should compile properly with or without
2767 this option (though a rare few will require @option{-ansi} or a
2768 @option{-std} option specifying the required version of ISO C)@. However,
2769 without this option, certain GNU extensions and traditional C and C++
2770 features are supported as well. With this option, they are rejected.
2772 @option{-pedantic} does not cause warning messages for use of the
2773 alternate keywords whose names begin and end with @samp{__}. Pedantic
2774 warnings are also disabled in the expression that follows
2775 @code{__extension__}. However, only system header files should use
2776 these escape routes; application programs should avoid them.
2777 @xref{Alternate Keywords}.
2779 Some users try to use @option{-pedantic} to check programs for strict ISO
2780 C conformance. They soon find that it does not do quite what they want:
2781 it finds some non-ISO practices, but not all---only those for which
2782 ISO C @emph{requires} a diagnostic, and some others for which
2783 diagnostics have been added.
2785 A feature to report any failure to conform to ISO C might be useful in
2786 some instances, but would require considerable additional work and would
2787 be quite different from @option{-pedantic}. We don't have plans to
2788 support such a feature in the near future.
2790 Where the standard specified with @option{-std} represents a GNU
2791 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2792 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2793 extended dialect is based. Warnings from @option{-pedantic} are given
2794 where they are required by the base standard. (It would not make sense
2795 for such warnings to be given only for features not in the specified GNU
2796 C dialect, since by definition the GNU dialects of C include all
2797 features the compiler supports with the given option, and there would be
2798 nothing to warn about.)
2800 @item -pedantic-errors
2801 @opindex pedantic-errors
2802 Like @option{-pedantic}, except that errors are produced rather than
2808 This enables all the warnings about constructions that some users
2809 consider questionable, and that are easy to avoid (or modify to
2810 prevent the warning), even in conjunction with macros. This also
2811 enables some language-specific warnings described in @ref{C++ Dialect
2812 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
2814 @option{-Wall} turns on the following warning flags:
2816 @gccoptlist{-Waddress @gol
2817 -Warray-bounds @r{(only with} @option{-O2}@r{)} @gol
2819 -Wchar-subscripts @gol
2820 -Wenum-compare @r{(in C/Objc; this is on by default in C++)} @gol
2822 -Wimplicit-function-declaration @gol
2825 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
2826 -Wmissing-braces @gol
2832 -Wsequence-point @gol
2833 -Wsign-compare @r{(only in C++)} @gol
2834 -Wstrict-aliasing @gol
2835 -Wstrict-overflow=1 @gol
2838 -Wuninitialized @gol
2839 -Wunknown-pragmas @gol
2840 -Wunused-function @gol
2843 -Wunused-variable @gol
2844 -Wvolatile-register-var @gol
2847 Note that some warning flags are not implied by @option{-Wall}. Some of
2848 them warn about constructions that users generally do not consider
2849 questionable, but which occasionally you might wish to check for;
2850 others warn about constructions that are necessary or hard to avoid in
2851 some cases, and there is no simple way to modify the code to suppress
2852 the warning. Some of them are enabled by @option{-Wextra} but many of
2853 them must be enabled individually.
2859 This enables some extra warning flags that are not enabled by
2860 @option{-Wall}. (This option used to be called @option{-W}. The older
2861 name is still supported, but the newer name is more descriptive.)
2863 @gccoptlist{-Wclobbered @gol
2865 -Wignored-qualifiers @gol
2866 -Wmissing-field-initializers @gol
2867 -Wmissing-parameter-type @r{(C only)} @gol
2868 -Wold-style-declaration @r{(C only)} @gol
2869 -Woverride-init @gol
2872 -Wuninitialized @gol
2873 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2876 The option @option{-Wextra} also prints warning messages for the
2882 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2883 @samp{>}, or @samp{>=}.
2886 (C++ only) An enumerator and a non-enumerator both appear in a
2887 conditional expression.
2890 (C++ only) Ambiguous virtual bases.
2893 (C++ only) Subscripting an array which has been declared @samp{register}.
2896 (C++ only) Taking the address of a variable which has been declared
2900 (C++ only) A base class is not initialized in a derived class' copy
2905 @item -Wchar-subscripts
2906 @opindex Wchar-subscripts
2907 @opindex Wno-char-subscripts
2908 Warn if an array subscript has type @code{char}. This is a common cause
2909 of error, as programmers often forget that this type is signed on some
2911 This warning is enabled by @option{-Wall}.
2915 @opindex Wno-comment
2916 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2917 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2918 This warning is enabled by @option{-Wall}.
2923 @opindex ffreestanding
2924 @opindex fno-builtin
2925 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2926 the arguments supplied have types appropriate to the format string
2927 specified, and that the conversions specified in the format string make
2928 sense. This includes standard functions, and others specified by format
2929 attributes (@pxref{Function Attributes}), in the @code{printf},
2930 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2931 not in the C standard) families (or other target-specific families).
2932 Which functions are checked without format attributes having been
2933 specified depends on the standard version selected, and such checks of
2934 functions without the attribute specified are disabled by
2935 @option{-ffreestanding} or @option{-fno-builtin}.
2937 The formats are checked against the format features supported by GNU
2938 libc version 2.2. These include all ISO C90 and C99 features, as well
2939 as features from the Single Unix Specification and some BSD and GNU
2940 extensions. Other library implementations may not support all these
2941 features; GCC does not support warning about features that go beyond a
2942 particular library's limitations. However, if @option{-pedantic} is used
2943 with @option{-Wformat}, warnings will be given about format features not
2944 in the selected standard version (but not for @code{strfmon} formats,
2945 since those are not in any version of the C standard). @xref{C Dialect
2946 Options,,Options Controlling C Dialect}.
2948 Since @option{-Wformat} also checks for null format arguments for
2949 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2951 @option{-Wformat} is included in @option{-Wall}. For more control over some
2952 aspects of format checking, the options @option{-Wformat-y2k},
2953 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2954 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2955 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2958 @opindex Wformat-y2k
2959 @opindex Wno-format-y2k
2960 If @option{-Wformat} is specified, also warn about @code{strftime}
2961 formats which may yield only a two-digit year.
2963 @item -Wno-format-contains-nul
2964 @opindex Wno-format-contains-nul
2965 @opindex Wformat-contains-nul
2966 If @option{-Wformat} is specified, do not warn about format strings that
2969 @item -Wno-format-extra-args
2970 @opindex Wno-format-extra-args
2971 @opindex Wformat-extra-args
2972 If @option{-Wformat} is specified, do not warn about excess arguments to a
2973 @code{printf} or @code{scanf} format function. The C standard specifies
2974 that such arguments are ignored.
2976 Where the unused arguments lie between used arguments that are
2977 specified with @samp{$} operand number specifications, normally
2978 warnings are still given, since the implementation could not know what
2979 type to pass to @code{va_arg} to skip the unused arguments. However,
2980 in the case of @code{scanf} formats, this option will suppress the
2981 warning if the unused arguments are all pointers, since the Single
2982 Unix Specification says that such unused arguments are allowed.
2984 @item -Wno-format-zero-length @r{(C and Objective-C only)}
2985 @opindex Wno-format-zero-length
2986 @opindex Wformat-zero-length
2987 If @option{-Wformat} is specified, do not warn about zero-length formats.
2988 The C standard specifies that zero-length formats are allowed.
2990 @item -Wformat-nonliteral
2991 @opindex Wformat-nonliteral
2992 @opindex Wno-format-nonliteral
2993 If @option{-Wformat} is specified, also warn if the format string is not a
2994 string literal and so cannot be checked, unless the format function
2995 takes its format arguments as a @code{va_list}.
2997 @item -Wformat-security
2998 @opindex Wformat-security
2999 @opindex Wno-format-security
3000 If @option{-Wformat} is specified, also warn about uses of format
3001 functions that represent possible security problems. At present, this
3002 warns about calls to @code{printf} and @code{scanf} functions where the
3003 format string is not a string literal and there are no format arguments,
3004 as in @code{printf (foo);}. This may be a security hole if the format
3005 string came from untrusted input and contains @samp{%n}. (This is
3006 currently a subset of what @option{-Wformat-nonliteral} warns about, but
3007 in future warnings may be added to @option{-Wformat-security} that are not
3008 included in @option{-Wformat-nonliteral}.)
3012 @opindex Wno-format=2
3013 Enable @option{-Wformat} plus format checks not included in
3014 @option{-Wformat}. Currently equivalent to @samp{-Wformat
3015 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
3017 @item -Wnonnull @r{(C and Objective-C only)}
3019 @opindex Wno-nonnull
3020 Warn about passing a null pointer for arguments marked as
3021 requiring a non-null value by the @code{nonnull} function attribute.
3023 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
3024 can be disabled with the @option{-Wno-nonnull} option.
3026 @item -Wjump-misses-init @r{(C, Objective-C only)}
3027 @opindex Wjump-misses-init
3028 @opindex Wno-jump-misses-init
3029 Warn if a @code{goto} statement or a @code{switch} statement jumps
3030 forward across the initialization of a variable, or jumps backward to a
3031 label after the variable has been initialized. This only warns about
3032 variables which are initialized when they are declared. This warning is
3033 only supported for C and Objective C; in C++ this sort of branch is an
3036 @option{-Wjump-misses-init} is included in @option{-Wall} and
3037 @option{-Wc++-compat}. It can be disabled with the
3038 @option{-Wno-jump-misses-init} option.
3040 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
3042 @opindex Wno-init-self
3043 Warn about uninitialized variables which are initialized with themselves.
3044 Note this option can only be used with the @option{-Wuninitialized} option.
3046 For example, GCC will warn about @code{i} being uninitialized in the
3047 following snippet only when @option{-Winit-self} has been specified:
3058 @item -Wimplicit-int @r{(C and Objective-C only)}
3059 @opindex Wimplicit-int
3060 @opindex Wno-implicit-int
3061 Warn when a declaration does not specify a type.
3062 This warning is enabled by @option{-Wall}.
3064 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
3065 @opindex Wimplicit-function-declaration
3066 @opindex Wno-implicit-function-declaration
3067 Give a warning whenever a function is used before being declared. In
3068 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
3069 enabled by default and it is made into an error by
3070 @option{-pedantic-errors}. This warning is also enabled by
3075 @opindex Wno-implicit
3076 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
3077 This warning is enabled by @option{-Wall}.
3079 @item -Wignored-qualifiers @r{(C and C++ only)}
3080 @opindex Wignored-qualifiers
3081 @opindex Wno-ignored-qualifiers
3082 Warn if the return type of a function has a type qualifier
3083 such as @code{const}. For ISO C such a type qualifier has no effect,
3084 since the value returned by a function is not an lvalue.
3085 For C++, the warning is only emitted for scalar types or @code{void}.
3086 ISO C prohibits qualified @code{void} return types on function
3087 definitions, so such return types always receive a warning
3088 even without this option.
3090 This warning is also enabled by @option{-Wextra}.
3095 Warn if the type of @samp{main} is suspicious. @samp{main} should be
3096 a function with external linkage, returning int, taking either zero
3097 arguments, two, or three arguments of appropriate types. This warning
3098 is enabled by default in C++ and is enabled by either @option{-Wall}
3099 or @option{-pedantic}.
3101 @item -Wmissing-braces
3102 @opindex Wmissing-braces
3103 @opindex Wno-missing-braces
3104 Warn if an aggregate or union initializer is not fully bracketed. In
3105 the following example, the initializer for @samp{a} is not fully
3106 bracketed, but that for @samp{b} is fully bracketed.
3109 int a[2][2] = @{ 0, 1, 2, 3 @};
3110 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
3113 This warning is enabled by @option{-Wall}.
3115 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
3116 @opindex Wmissing-include-dirs
3117 @opindex Wno-missing-include-dirs
3118 Warn if a user-supplied include directory does not exist.
3121 @opindex Wparentheses
3122 @opindex Wno-parentheses
3123 Warn if parentheses are omitted in certain contexts, such
3124 as when there is an assignment in a context where a truth value
3125 is expected, or when operators are nested whose precedence people
3126 often get confused about.
3128 Also warn if a comparison like @samp{x<=y<=z} appears; this is
3129 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
3130 interpretation from that of ordinary mathematical notation.
3132 Also warn about constructions where there may be confusion to which
3133 @code{if} statement an @code{else} branch belongs. Here is an example of
3148 In C/C++, every @code{else} branch belongs to the innermost possible
3149 @code{if} statement, which in this example is @code{if (b)}. This is
3150 often not what the programmer expected, as illustrated in the above
3151 example by indentation the programmer chose. When there is the
3152 potential for this confusion, GCC will issue a warning when this flag
3153 is specified. To eliminate the warning, add explicit braces around
3154 the innermost @code{if} statement so there is no way the @code{else}
3155 could belong to the enclosing @code{if}. The resulting code would
3172 This warning is enabled by @option{-Wall}.
3174 @item -Wsequence-point
3175 @opindex Wsequence-point
3176 @opindex Wno-sequence-point
3177 Warn about code that may have undefined semantics because of violations
3178 of sequence point rules in the C and C++ standards.
3180 The C and C++ standards defines the order in which expressions in a C/C++
3181 program are evaluated in terms of @dfn{sequence points}, which represent
3182 a partial ordering between the execution of parts of the program: those
3183 executed before the sequence point, and those executed after it. These
3184 occur after the evaluation of a full expression (one which is not part
3185 of a larger expression), after the evaluation of the first operand of a
3186 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
3187 function is called (but after the evaluation of its arguments and the
3188 expression denoting the called function), and in certain other places.
3189 Other than as expressed by the sequence point rules, the order of
3190 evaluation of subexpressions of an expression is not specified. All
3191 these rules describe only a partial order rather than a total order,
3192 since, for example, if two functions are called within one expression
3193 with no sequence point between them, the order in which the functions
3194 are called is not specified. However, the standards committee have
3195 ruled that function calls do not overlap.
3197 It is not specified when between sequence points modifications to the
3198 values of objects take effect. Programs whose behavior depends on this
3199 have undefined behavior; the C and C++ standards specify that ``Between
3200 the previous and next sequence point an object shall have its stored
3201 value modified at most once by the evaluation of an expression.
3202 Furthermore, the prior value shall be read only to determine the value
3203 to be stored.''. If a program breaks these rules, the results on any
3204 particular implementation are entirely unpredictable.
3206 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
3207 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
3208 diagnosed by this option, and it may give an occasional false positive
3209 result, but in general it has been found fairly effective at detecting
3210 this sort of problem in programs.
3212 The standard is worded confusingly, therefore there is some debate
3213 over the precise meaning of the sequence point rules in subtle cases.
3214 Links to discussions of the problem, including proposed formal
3215 definitions, may be found on the GCC readings page, at
3216 @w{@uref{http://gcc.gnu.org/readings.html}}.
3218 This warning is enabled by @option{-Wall} for C and C++.
3221 @opindex Wreturn-type
3222 @opindex Wno-return-type
3223 Warn whenever a function is defined with a return-type that defaults
3224 to @code{int}. Also warn about any @code{return} statement with no
3225 return-value in a function whose return-type is not @code{void}
3226 (falling off the end of the function body is considered returning
3227 without a value), and about a @code{return} statement with an
3228 expression in a function whose return-type is @code{void}.
3230 For C++, a function without return type always produces a diagnostic
3231 message, even when @option{-Wno-return-type} is specified. The only
3232 exceptions are @samp{main} and functions defined in system headers.
3234 This warning is enabled by @option{-Wall}.
3239 Warn whenever a @code{switch} statement has an index of enumerated type
3240 and lacks a @code{case} for one or more of the named codes of that
3241 enumeration. (The presence of a @code{default} label prevents this
3242 warning.) @code{case} labels outside the enumeration range also
3243 provoke warnings when this option is used (even if there is a
3244 @code{default} label).
3245 This warning is enabled by @option{-Wall}.
3247 @item -Wswitch-default
3248 @opindex Wswitch-default
3249 @opindex Wno-switch-default
3250 Warn whenever a @code{switch} statement does not have a @code{default}
3254 @opindex Wswitch-enum
3255 @opindex Wno-switch-enum
3256 Warn whenever a @code{switch} statement has an index of enumerated type
3257 and lacks a @code{case} for one or more of the named codes of that
3258 enumeration. @code{case} labels outside the enumeration range also
3259 provoke warnings when this option is used. The only difference
3260 between @option{-Wswitch} and this option is that this option gives a
3261 warning about an omitted enumeration code even if there is a
3262 @code{default} label.
3264 @item -Wsync-nand @r{(C and C++ only)}
3266 @opindex Wno-sync-nand
3267 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
3268 built-in functions are used. These functions changed semantics in GCC 4.4.
3272 @opindex Wno-trigraphs
3273 Warn if any trigraphs are encountered that might change the meaning of
3274 the program (trigraphs within comments are not warned about).
3275 This warning is enabled by @option{-Wall}.
3277 @item -Wunused-function
3278 @opindex Wunused-function
3279 @opindex Wno-unused-function
3280 Warn whenever a static function is declared but not defined or a
3281 non-inline static function is unused.
3282 This warning is enabled by @option{-Wall}.
3284 @item -Wunused-label
3285 @opindex Wunused-label
3286 @opindex Wno-unused-label
3287 Warn whenever a label is declared but not used.
3288 This warning is enabled by @option{-Wall}.
3290 To suppress this warning use the @samp{unused} attribute
3291 (@pxref{Variable Attributes}).
3293 @item -Wunused-parameter
3294 @opindex Wunused-parameter
3295 @opindex Wno-unused-parameter
3296 Warn whenever a function parameter is unused aside from its declaration.
3298 To suppress this warning use the @samp{unused} attribute
3299 (@pxref{Variable Attributes}).
3301 @item -Wno-unused-result
3302 @opindex Wunused-result
3303 @opindex Wno-unused-result
3304 Do not warn if a caller of a function marked with attribute
3305 @code{warn_unused_result} (@pxref{Variable Attributes}) does not use
3306 its return value. The default is @option{-Wunused-result}.
3308 @item -Wunused-variable
3309 @opindex Wunused-variable
3310 @opindex Wno-unused-variable
3311 Warn whenever a local variable or non-constant static variable is unused
3312 aside from its declaration.
3313 This warning is enabled by @option{-Wall}.
3315 To suppress this warning use the @samp{unused} attribute
3316 (@pxref{Variable Attributes}).
3318 @item -Wunused-value
3319 @opindex Wunused-value
3320 @opindex Wno-unused-value
3321 Warn whenever a statement computes a result that is explicitly not
3322 used. To suppress this warning cast the unused expression to
3323 @samp{void}. This includes an expression-statement or the left-hand
3324 side of a comma expression that contains no side effects. For example,
3325 an expression such as @samp{x[i,j]} will cause a warning, while
3326 @samp{x[(void)i,j]} will not.
3328 This warning is enabled by @option{-Wall}.
3333 All the above @option{-Wunused} options combined.
3335 In order to get a warning about an unused function parameter, you must
3336 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
3337 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
3339 @item -Wuninitialized
3340 @opindex Wuninitialized
3341 @opindex Wno-uninitialized
3342 Warn if an automatic variable is used without first being initialized
3343 or if a variable may be clobbered by a @code{setjmp} call. In C++,
3344 warn if a non-static reference or non-static @samp{const} member
3345 appears in a class without constructors.
3347 If you want to warn about code which uses the uninitialized value of the
3348 variable in its own initializer, use the @option{-Winit-self} option.
3350 These warnings occur for individual uninitialized or clobbered
3351 elements of structure, union or array variables as well as for
3352 variables which are uninitialized or clobbered as a whole. They do
3353 not occur for variables or elements declared @code{volatile}. Because
3354 these warnings depend on optimization, the exact variables or elements
3355 for which there are warnings will depend on the precise optimization
3356 options and version of GCC used.
3358 Note that there may be no warning about a variable that is used only
3359 to compute a value that itself is never used, because such
3360 computations may be deleted by data flow analysis before the warnings
3363 These warnings are made optional because GCC is not smart
3364 enough to see all the reasons why the code might be correct
3365 despite appearing to have an error. Here is one example of how
3386 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3387 always initialized, but GCC doesn't know this. Here is
3388 another common case:
3393 if (change_y) save_y = y, y = new_y;
3395 if (change_y) y = save_y;
3400 This has no bug because @code{save_y} is used only if it is set.
3402 @cindex @code{longjmp} warnings
3403 This option also warns when a non-volatile automatic variable might be
3404 changed by a call to @code{longjmp}. These warnings as well are possible
3405 only in optimizing compilation.
3407 The compiler sees only the calls to @code{setjmp}. It cannot know
3408 where @code{longjmp} will be called; in fact, a signal handler could
3409 call it at any point in the code. As a result, you may get a warning
3410 even when there is in fact no problem because @code{longjmp} cannot
3411 in fact be called at the place which would cause a problem.
3413 Some spurious warnings can be avoided if you declare all the functions
3414 you use that never return as @code{noreturn}. @xref{Function
3417 This warning is enabled by @option{-Wall} or @option{-Wextra}.
3419 @item -Wunknown-pragmas
3420 @opindex Wunknown-pragmas
3421 @opindex Wno-unknown-pragmas
3422 @cindex warning for unknown pragmas
3423 @cindex unknown pragmas, warning
3424 @cindex pragmas, warning of unknown
3425 Warn when a #pragma directive is encountered which is not understood by
3426 GCC@. If this command line option is used, warnings will even be issued
3427 for unknown pragmas in system header files. This is not the case if
3428 the warnings were only enabled by the @option{-Wall} command line option.
3431 @opindex Wno-pragmas
3433 Do not warn about misuses of pragmas, such as incorrect parameters,
3434 invalid syntax, or conflicts between pragmas. See also
3435 @samp{-Wunknown-pragmas}.
3437 @item -Wstrict-aliasing
3438 @opindex Wstrict-aliasing
3439 @opindex Wno-strict-aliasing
3440 This option is only active when @option{-fstrict-aliasing} is active.
3441 It warns about code which might break the strict aliasing rules that the
3442 compiler is using for optimization. The warning does not catch all
3443 cases, but does attempt to catch the more common pitfalls. It is
3444 included in @option{-Wall}.
3445 It is equivalent to @option{-Wstrict-aliasing=3}
3447 @item -Wstrict-aliasing=n
3448 @opindex Wstrict-aliasing=n
3449 @opindex Wno-strict-aliasing=n
3450 This option is only active when @option{-fstrict-aliasing} is active.
3451 It warns about code which might break the strict aliasing rules that the
3452 compiler is using for optimization.
3453 Higher levels correspond to higher accuracy (fewer false positives).
3454 Higher levels also correspond to more effort, similar to the way -O works.
3455 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3458 Level 1: Most aggressive, quick, least accurate.
3459 Possibly useful when higher levels
3460 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3461 false negatives. However, it has many false positives.
3462 Warns for all pointer conversions between possibly incompatible types,
3463 even if never dereferenced. Runs in the frontend only.
3465 Level 2: Aggressive, quick, not too precise.
3466 May still have many false positives (not as many as level 1 though),
3467 and few false negatives (but possibly more than level 1).
3468 Unlike level 1, it only warns when an address is taken. Warns about
3469 incomplete types. Runs in the frontend only.
3471 Level 3 (default for @option{-Wstrict-aliasing}):
3472 Should have very few false positives and few false
3473 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3474 Takes care of the common punn+dereference pattern in the frontend:
3475 @code{*(int*)&some_float}.
3476 If optimization is enabled, it also runs in the backend, where it deals
3477 with multiple statement cases using flow-sensitive points-to information.
3478 Only warns when the converted pointer is dereferenced.
3479 Does not warn about incomplete types.
3481 @item -Wstrict-overflow
3482 @itemx -Wstrict-overflow=@var{n}
3483 @opindex Wstrict-overflow
3484 @opindex Wno-strict-overflow
3485 This option is only active when @option{-fstrict-overflow} is active.
3486 It warns about cases where the compiler optimizes based on the
3487 assumption that signed overflow does not occur. Note that it does not
3488 warn about all cases where the code might overflow: it only warns
3489 about cases where the compiler implements some optimization. Thus
3490 this warning depends on the optimization level.
3492 An optimization which assumes that signed overflow does not occur is
3493 perfectly safe if the values of the variables involved are such that
3494 overflow never does, in fact, occur. Therefore this warning can
3495 easily give a false positive: a warning about code which is not
3496 actually a problem. To help focus on important issues, several
3497 warning levels are defined. No warnings are issued for the use of
3498 undefined signed overflow when estimating how many iterations a loop
3499 will require, in particular when determining whether a loop will be
3503 @item -Wstrict-overflow=1
3504 Warn about cases which are both questionable and easy to avoid. For
3505 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3506 compiler will simplify this to @code{1}. This level of
3507 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3508 are not, and must be explicitly requested.
3510 @item -Wstrict-overflow=2
3511 Also warn about other cases where a comparison is simplified to a
3512 constant. For example: @code{abs (x) >= 0}. This can only be
3513 simplified when @option{-fstrict-overflow} is in effect, because
3514 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3515 zero. @option{-Wstrict-overflow} (with no level) is the same as
3516 @option{-Wstrict-overflow=2}.
3518 @item -Wstrict-overflow=3
3519 Also warn about other cases where a comparison is simplified. For
3520 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3522 @item -Wstrict-overflow=4
3523 Also warn about other simplifications not covered by the above cases.
3524 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3526 @item -Wstrict-overflow=5
3527 Also warn about cases where the compiler reduces the magnitude of a
3528 constant involved in a comparison. For example: @code{x + 2 > y} will
3529 be simplified to @code{x + 1 >= y}. This is reported only at the
3530 highest warning level because this simplification applies to many
3531 comparisons, so this warning level will give a very large number of
3535 @item -Warray-bounds
3536 @opindex Wno-array-bounds
3537 @opindex Warray-bounds
3538 This option is only active when @option{-ftree-vrp} is active
3539 (default for -O2 and above). It warns about subscripts to arrays
3540 that are always out of bounds. This warning is enabled by @option{-Wall}.
3542 @item -Wno-div-by-zero
3543 @opindex Wno-div-by-zero
3544 @opindex Wdiv-by-zero
3545 Do not warn about compile-time integer division by zero. Floating point
3546 division by zero is not warned about, as it can be a legitimate way of
3547 obtaining infinities and NaNs.
3549 @item -Wsystem-headers
3550 @opindex Wsystem-headers
3551 @opindex Wno-system-headers
3552 @cindex warnings from system headers
3553 @cindex system headers, warnings from
3554 Print warning messages for constructs found in system header files.
3555 Warnings from system headers are normally suppressed, on the assumption
3556 that they usually do not indicate real problems and would only make the
3557 compiler output harder to read. Using this command line option tells
3558 GCC to emit warnings from system headers as if they occurred in user
3559 code. However, note that using @option{-Wall} in conjunction with this
3560 option will @emph{not} warn about unknown pragmas in system
3561 headers---for that, @option{-Wunknown-pragmas} must also be used.
3564 @opindex Wfloat-equal
3565 @opindex Wno-float-equal
3566 Warn if floating point values are used in equality comparisons.
3568 The idea behind this is that sometimes it is convenient (for the
3569 programmer) to consider floating-point values as approximations to
3570 infinitely precise real numbers. If you are doing this, then you need
3571 to compute (by analyzing the code, or in some other way) the maximum or
3572 likely maximum error that the computation introduces, and allow for it
3573 when performing comparisons (and when producing output, but that's a
3574 different problem). In particular, instead of testing for equality, you
3575 would check to see whether the two values have ranges that overlap; and
3576 this is done with the relational operators, so equality comparisons are
3579 @item -Wtraditional @r{(C and Objective-C only)}
3580 @opindex Wtraditional
3581 @opindex Wno-traditional
3582 Warn about certain constructs that behave differently in traditional and
3583 ISO C@. Also warn about ISO C constructs that have no traditional C
3584 equivalent, and/or problematic constructs which should be avoided.
3588 Macro parameters that appear within string literals in the macro body.
3589 In traditional C macro replacement takes place within string literals,
3590 but does not in ISO C@.
3593 In traditional C, some preprocessor directives did not exist.
3594 Traditional preprocessors would only consider a line to be a directive
3595 if the @samp{#} appeared in column 1 on the line. Therefore
3596 @option{-Wtraditional} warns about directives that traditional C
3597 understands but would ignore because the @samp{#} does not appear as the
3598 first character on the line. It also suggests you hide directives like
3599 @samp{#pragma} not understood by traditional C by indenting them. Some
3600 traditional implementations would not recognize @samp{#elif}, so it
3601 suggests avoiding it altogether.
3604 A function-like macro that appears without arguments.
3607 The unary plus operator.
3610 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3611 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3612 constants.) Note, these suffixes appear in macros defined in the system
3613 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3614 Use of these macros in user code might normally lead to spurious
3615 warnings, however GCC's integrated preprocessor has enough context to
3616 avoid warning in these cases.
3619 A function declared external in one block and then used after the end of
3623 A @code{switch} statement has an operand of type @code{long}.
3626 A non-@code{static} function declaration follows a @code{static} one.
3627 This construct is not accepted by some traditional C compilers.
3630 The ISO type of an integer constant has a different width or
3631 signedness from its traditional type. This warning is only issued if
3632 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3633 typically represent bit patterns, are not warned about.
3636 Usage of ISO string concatenation is detected.
3639 Initialization of automatic aggregates.
3642 Identifier conflicts with labels. Traditional C lacks a separate
3643 namespace for labels.
3646 Initialization of unions. If the initializer is zero, the warning is
3647 omitted. This is done under the assumption that the zero initializer in
3648 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3649 initializer warnings and relies on default initialization to zero in the
3653 Conversions by prototypes between fixed/floating point values and vice
3654 versa. The absence of these prototypes when compiling with traditional
3655 C would cause serious problems. This is a subset of the possible
3656 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3659 Use of ISO C style function definitions. This warning intentionally is
3660 @emph{not} issued for prototype declarations or variadic functions
3661 because these ISO C features will appear in your code when using
3662 libiberty's traditional C compatibility macros, @code{PARAMS} and
3663 @code{VPARAMS}. This warning is also bypassed for nested functions
3664 because that feature is already a GCC extension and thus not relevant to
3665 traditional C compatibility.
3668 @item -Wtraditional-conversion @r{(C and Objective-C only)}
3669 @opindex Wtraditional-conversion
3670 @opindex Wno-traditional-conversion
3671 Warn if a prototype causes a type conversion that is different from what
3672 would happen to the same argument in the absence of a prototype. This
3673 includes conversions of fixed point to floating and vice versa, and
3674 conversions changing the width or signedness of a fixed point argument
3675 except when the same as the default promotion.
3677 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
3678 @opindex Wdeclaration-after-statement
3679 @opindex Wno-declaration-after-statement
3680 Warn when a declaration is found after a statement in a block. This
3681 construct, known from C++, was introduced with ISO C99 and is by default
3682 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3683 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3688 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3690 @item -Wno-endif-labels
3691 @opindex Wno-endif-labels
3692 @opindex Wendif-labels
3693 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3698 Warn whenever a local variable shadows another local variable, parameter or
3699 global variable or whenever a built-in function is shadowed.
3701 @item -Wlarger-than=@var{len}
3702 @opindex Wlarger-than=@var{len}
3703 @opindex Wlarger-than-@var{len}
3704 Warn whenever an object of larger than @var{len} bytes is defined.
3706 @item -Wframe-larger-than=@var{len}
3707 @opindex Wframe-larger-than
3708 Warn if the size of a function frame is larger than @var{len} bytes.
3709 The computation done to determine the stack frame size is approximate
3710 and not conservative.
3711 The actual requirements may be somewhat greater than @var{len}
3712 even if you do not get a warning. In addition, any space allocated
3713 via @code{alloca}, variable-length arrays, or related constructs
3714 is not included by the compiler when determining
3715 whether or not to issue a warning.
3717 @item -Wunsafe-loop-optimizations
3718 @opindex Wunsafe-loop-optimizations
3719 @opindex Wno-unsafe-loop-optimizations
3720 Warn if the loop cannot be optimized because the compiler could not
3721 assume anything on the bounds of the loop indices. With
3722 @option{-funsafe-loop-optimizations} warn if the compiler made
3725 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
3726 @opindex Wno-pedantic-ms-format
3727 @opindex Wpedantic-ms-format
3728 Disables the warnings about non-ISO @code{printf} / @code{scanf} format
3729 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets
3730 depending on the MS runtime, when you are using the options @option{-Wformat}
3731 and @option{-pedantic} without gnu-extensions.
3733 @item -Wpointer-arith
3734 @opindex Wpointer-arith
3735 @opindex Wno-pointer-arith
3736 Warn about anything that depends on the ``size of'' a function type or
3737 of @code{void}. GNU C assigns these types a size of 1, for
3738 convenience in calculations with @code{void *} pointers and pointers
3739 to functions. In C++, warn also when an arithmetic operation involves
3740 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3743 @opindex Wtype-limits
3744 @opindex Wno-type-limits
3745 Warn if a comparison is always true or always false due to the limited
3746 range of the data type, but do not warn for constant expressions. For
3747 example, warn if an unsigned variable is compared against zero with
3748 @samp{<} or @samp{>=}. This warning is also enabled by
3751 @item -Wbad-function-cast @r{(C and Objective-C only)}
3752 @opindex Wbad-function-cast
3753 @opindex Wno-bad-function-cast
3754 Warn whenever a function call is cast to a non-matching type.
3755 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3757 @item -Wc++-compat @r{(C and Objective-C only)}
3758 Warn about ISO C constructs that are outside of the common subset of
3759 ISO C and ISO C++, e.g.@: request for implicit conversion from
3760 @code{void *} to a pointer to non-@code{void} type.
3762 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3763 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3764 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3765 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3769 @opindex Wno-cast-qual
3770 Warn whenever a pointer is cast so as to remove a type qualifier from
3771 the target type. For example, warn if a @code{const char *} is cast
3772 to an ordinary @code{char *}.
3774 Also warn when making a cast which introduces a type qualifier in an
3775 unsafe way. For example, casting @code{char **} to @code{const char **}
3776 is unsafe, as in this example:
3779 /* p is char ** value. */
3780 const char **q = (const char **) p;
3781 /* Assignment of readonly string to const char * is OK. */
3783 /* Now char** pointer points to read-only memory. */
3788 @opindex Wcast-align
3789 @opindex Wno-cast-align
3790 Warn whenever a pointer is cast such that the required alignment of the
3791 target is increased. For example, warn if a @code{char *} is cast to
3792 an @code{int *} on machines where integers can only be accessed at
3793 two- or four-byte boundaries.
3795 @item -Wwrite-strings
3796 @opindex Wwrite-strings
3797 @opindex Wno-write-strings
3798 When compiling C, give string constants the type @code{const
3799 char[@var{length}]} so that copying the address of one into a
3800 non-@code{const} @code{char *} pointer will get a warning. These
3801 warnings will help you find at compile time code that can try to write
3802 into a string constant, but only if you have been very careful about
3803 using @code{const} in declarations and prototypes. Otherwise, it will
3804 just be a nuisance. This is why we did not make @option{-Wall} request
3807 When compiling C++, warn about the deprecated conversion from string
3808 literals to @code{char *}. This warning is enabled by default for C++
3813 @opindex Wno-clobbered
3814 Warn for variables that might be changed by @samp{longjmp} or
3815 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3818 @opindex Wconversion
3819 @opindex Wno-conversion
3820 Warn for implicit conversions that may alter a value. This includes
3821 conversions between real and integer, like @code{abs (x)} when
3822 @code{x} is @code{double}; conversions between signed and unsigned,
3823 like @code{unsigned ui = -1}; and conversions to smaller types, like
3824 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3825 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3826 changed by the conversion like in @code{abs (2.0)}. Warnings about
3827 conversions between signed and unsigned integers can be disabled by
3828 using @option{-Wno-sign-conversion}.
3830 For C++, also warn for conversions between @code{NULL} and non-pointer
3831 types; confusing overload resolution for user-defined conversions; and
3832 conversions that will never use a type conversion operator:
3833 conversions to @code{void}, the same type, a base class or a reference
3834 to them. Warnings about conversions between signed and unsigned
3835 integers are disabled by default in C++ unless
3836 @option{-Wsign-conversion} is explicitly enabled.
3839 @opindex Wempty-body
3840 @opindex Wno-empty-body
3841 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3842 while} statement. This warning is also enabled by @option{-Wextra}.
3844 @item -Wenum-compare
3845 @opindex Wenum-compare
3846 @opindex Wno-enum-compare
3847 Warn about a comparison between values of different enum types. In C++
3848 this warning is enabled by default. In C this warning is enabled by
3851 @item -Wsign-compare
3852 @opindex Wsign-compare
3853 @opindex Wno-sign-compare
3854 @cindex warning for comparison of signed and unsigned values
3855 @cindex comparison of signed and unsigned values, warning
3856 @cindex signed and unsigned values, comparison warning
3857 Warn when a comparison between signed and unsigned values could produce
3858 an incorrect result when the signed value is converted to unsigned.
3859 This warning is also enabled by @option{-Wextra}; to get the other warnings
3860 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3862 @item -Wsign-conversion
3863 @opindex Wsign-conversion
3864 @opindex Wno-sign-conversion
3865 Warn for implicit conversions that may change the sign of an integer
3866 value, like assigning a signed integer expression to an unsigned
3867 integer variable. An explicit cast silences the warning. In C, this
3868 option is enabled also by @option{-Wconversion}.
3872 @opindex Wno-address
3873 Warn about suspicious uses of memory addresses. These include using
3874 the address of a function in a conditional expression, such as
3875 @code{void func(void); if (func)}, and comparisons against the memory
3876 address of a string literal, such as @code{if (x == "abc")}. Such
3877 uses typically indicate a programmer error: the address of a function
3878 always evaluates to true, so their use in a conditional usually
3879 indicate that the programmer forgot the parentheses in a function
3880 call; and comparisons against string literals result in unspecified
3881 behavior and are not portable in C, so they usually indicate that the
3882 programmer intended to use @code{strcmp}. This warning is enabled by
3886 @opindex Wlogical-op
3887 @opindex Wno-logical-op
3888 Warn about suspicious uses of logical operators in expressions.
3889 This includes using logical operators in contexts where a
3890 bit-wise operator is likely to be expected.
3892 @item -Waggregate-return
3893 @opindex Waggregate-return
3894 @opindex Wno-aggregate-return
3895 Warn if any functions that return structures or unions are defined or
3896 called. (In languages where you can return an array, this also elicits
3899 @item -Wno-attributes
3900 @opindex Wno-attributes
3901 @opindex Wattributes
3902 Do not warn if an unexpected @code{__attribute__} is used, such as
3903 unrecognized attributes, function attributes applied to variables,
3904 etc. This will not stop errors for incorrect use of supported
3907 @item -Wno-builtin-macro-redefined
3908 @opindex Wno-builtin-macro-redefined
3909 @opindex Wbuiltin-macro-redefined
3910 Do not warn if certain built-in macros are redefined. This suppresses
3911 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
3912 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
3914 @item -Wstrict-prototypes @r{(C and Objective-C only)}
3915 @opindex Wstrict-prototypes
3916 @opindex Wno-strict-prototypes
3917 Warn if a function is declared or defined without specifying the
3918 argument types. (An old-style function definition is permitted without
3919 a warning if preceded by a declaration which specifies the argument
3922 @item -Wold-style-declaration @r{(C and Objective-C only)}
3923 @opindex Wold-style-declaration
3924 @opindex Wno-old-style-declaration
3925 Warn for obsolescent usages, according to the C Standard, in a
3926 declaration. For example, warn if storage-class specifiers like
3927 @code{static} are not the first things in a declaration. This warning
3928 is also enabled by @option{-Wextra}.
3930 @item -Wold-style-definition @r{(C and Objective-C only)}
3931 @opindex Wold-style-definition
3932 @opindex Wno-old-style-definition
3933 Warn if an old-style function definition is used. A warning is given
3934 even if there is a previous prototype.
3936 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
3937 @opindex Wmissing-parameter-type
3938 @opindex Wno-missing-parameter-type
3939 A function parameter is declared without a type specifier in K&R-style
3946 This warning is also enabled by @option{-Wextra}.
3948 @item -Wmissing-prototypes @r{(C and Objective-C only)}
3949 @opindex Wmissing-prototypes
3950 @opindex Wno-missing-prototypes
3951 Warn if a global function is defined without a previous prototype
3952 declaration. This warning is issued even if the definition itself
3953 provides a prototype. The aim is to detect global functions that fail
3954 to be declared in header files.
3956 @item -Wmissing-declarations
3957 @opindex Wmissing-declarations
3958 @opindex Wno-missing-declarations
3959 Warn if a global function is defined without a previous declaration.
3960 Do so even if the definition itself provides a prototype.
3961 Use this option to detect global functions that are not declared in
3962 header files. In C++, no warnings are issued for function templates,
3963 or for inline functions, or for functions in anonymous namespaces.
3965 @item -Wmissing-field-initializers
3966 @opindex Wmissing-field-initializers
3967 @opindex Wno-missing-field-initializers
3971 Warn if a structure's initializer has some fields missing. For
3972 example, the following code would cause such a warning, because
3973 @code{x.h} is implicitly zero:
3976 struct s @{ int f, g, h; @};
3977 struct s x = @{ 3, 4 @};
3980 This option does not warn about designated initializers, so the following
3981 modification would not trigger a warning:
3984 struct s @{ int f, g, h; @};
3985 struct s x = @{ .f = 3, .g = 4 @};
3988 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
3989 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
3991 @item -Wmissing-noreturn
3992 @opindex Wmissing-noreturn
3993 @opindex Wno-missing-noreturn
3994 Warn about functions which might be candidates for attribute @code{noreturn}.
3995 Note these are only possible candidates, not absolute ones. Care should
3996 be taken to manually verify functions actually do not ever return before
3997 adding the @code{noreturn} attribute, otherwise subtle code generation
3998 bugs could be introduced. You will not get a warning for @code{main} in
3999 hosted C environments.
4001 @item -Wmissing-format-attribute
4002 @opindex Wmissing-format-attribute
4003 @opindex Wno-missing-format-attribute
4006 Warn about function pointers which might be candidates for @code{format}
4007 attributes. Note these are only possible candidates, not absolute ones.
4008 GCC will guess that function pointers with @code{format} attributes that
4009 are used in assignment, initialization, parameter passing or return
4010 statements should have a corresponding @code{format} attribute in the
4011 resulting type. I.e.@: the left-hand side of the assignment or
4012 initialization, the type of the parameter variable, or the return type
4013 of the containing function respectively should also have a @code{format}
4014 attribute to avoid the warning.
4016 GCC will also warn about function definitions which might be
4017 candidates for @code{format} attributes. Again, these are only
4018 possible candidates. GCC will guess that @code{format} attributes
4019 might be appropriate for any function that calls a function like
4020 @code{vprintf} or @code{vscanf}, but this might not always be the
4021 case, and some functions for which @code{format} attributes are
4022 appropriate may not be detected.
4024 @item -Wno-multichar
4025 @opindex Wno-multichar
4027 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
4028 Usually they indicate a typo in the user's code, as they have
4029 implementation-defined values, and should not be used in portable code.
4031 @item -Wnormalized=<none|id|nfc|nfkc>
4032 @opindex Wnormalized=
4035 @cindex character set, input normalization
4036 In ISO C and ISO C++, two identifiers are different if they are
4037 different sequences of characters. However, sometimes when characters
4038 outside the basic ASCII character set are used, you can have two
4039 different character sequences that look the same. To avoid confusion,
4040 the ISO 10646 standard sets out some @dfn{normalization rules} which
4041 when applied ensure that two sequences that look the same are turned into
4042 the same sequence. GCC can warn you if you are using identifiers which
4043 have not been normalized; this option controls that warning.
4045 There are four levels of warning that GCC supports. The default is
4046 @option{-Wnormalized=nfc}, which warns about any identifier which is
4047 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
4048 recommended form for most uses.
4050 Unfortunately, there are some characters which ISO C and ISO C++ allow
4051 in identifiers that when turned into NFC aren't allowable as
4052 identifiers. That is, there's no way to use these symbols in portable
4053 ISO C or C++ and have all your identifiers in NFC@.
4054 @option{-Wnormalized=id} suppresses the warning for these characters.
4055 It is hoped that future versions of the standards involved will correct
4056 this, which is why this option is not the default.
4058 You can switch the warning off for all characters by writing
4059 @option{-Wnormalized=none}. You would only want to do this if you
4060 were using some other normalization scheme (like ``D''), because
4061 otherwise you can easily create bugs that are literally impossible to see.
4063 Some characters in ISO 10646 have distinct meanings but look identical
4064 in some fonts or display methodologies, especially once formatting has
4065 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
4066 LETTER N'', will display just like a regular @code{n} which has been
4067 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
4068 normalization scheme to convert all these into a standard form as
4069 well, and GCC will warn if your code is not in NFKC if you use
4070 @option{-Wnormalized=nfkc}. This warning is comparable to warning
4071 about every identifier that contains the letter O because it might be
4072 confused with the digit 0, and so is not the default, but may be
4073 useful as a local coding convention if the programming environment is
4074 unable to be fixed to display these characters distinctly.
4076 @item -Wno-deprecated
4077 @opindex Wno-deprecated
4078 @opindex Wdeprecated
4079 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
4081 @item -Wno-deprecated-declarations
4082 @opindex Wno-deprecated-declarations
4083 @opindex Wdeprecated-declarations
4084 Do not warn about uses of functions (@pxref{Function Attributes}),
4085 variables (@pxref{Variable Attributes}), and types (@pxref{Type
4086 Attributes}) marked as deprecated by using the @code{deprecated}
4090 @opindex Wno-overflow
4092 Do not warn about compile-time overflow in constant expressions.
4094 @item -Woverride-init @r{(C and Objective-C only)}
4095 @opindex Woverride-init
4096 @opindex Wno-override-init
4100 Warn if an initialized field without side effects is overridden when
4101 using designated initializers (@pxref{Designated Inits, , Designated
4104 This warning is included in @option{-Wextra}. To get other
4105 @option{-Wextra} warnings without this one, use @samp{-Wextra
4106 -Wno-override-init}.
4111 Warn if a structure is given the packed attribute, but the packed
4112 attribute has no effect on the layout or size of the structure.
4113 Such structures may be mis-aligned for little benefit. For
4114 instance, in this code, the variable @code{f.x} in @code{struct bar}
4115 will be misaligned even though @code{struct bar} does not itself
4116 have the packed attribute:
4123 @} __attribute__((packed));
4131 @item -Wpacked-bitfield-compat
4132 @opindex Wpacked-bitfield-compat
4133 @opindex Wno-packed-bitfield-compat
4134 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
4135 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
4136 the change can lead to differences in the structure layout. GCC
4137 informs you when the offset of such a field has changed in GCC 4.4.
4138 For example there is no longer a 4-bit padding between field @code{a}
4139 and @code{b} in this structure:
4146 @} __attribute__ ((packed));
4149 This warning is enabled by default. Use
4150 @option{-Wno-packed-bitfield-compat} to disable this warning.
4155 Warn if padding is included in a structure, either to align an element
4156 of the structure or to align the whole structure. Sometimes when this
4157 happens it is possible to rearrange the fields of the structure to
4158 reduce the padding and so make the structure smaller.
4160 @item -Wredundant-decls
4161 @opindex Wredundant-decls
4162 @opindex Wno-redundant-decls
4163 Warn if anything is declared more than once in the same scope, even in
4164 cases where multiple declaration is valid and changes nothing.
4166 @item -Wnested-externs @r{(C and Objective-C only)}
4167 @opindex Wnested-externs
4168 @opindex Wno-nested-externs
4169 Warn if an @code{extern} declaration is encountered within a function.
4171 @item -Wunreachable-code
4172 @opindex Wunreachable-code
4173 @opindex Wno-unreachable-code
4174 Warn if the compiler detects that code will never be executed.
4176 This option is intended to warn when the compiler detects that at
4177 least a whole line of source code will never be executed, because
4178 some condition is never satisfied or because it is after a
4179 procedure that never returns.
4181 It is possible for this option to produce a warning even though there
4182 are circumstances under which part of the affected line can be executed,
4183 so care should be taken when removing apparently-unreachable code.
4185 For instance, when a function is inlined, a warning may mean that the
4186 line is unreachable in only one inlined copy of the function.
4188 This option is not made part of @option{-Wall} because in a debugging
4189 version of a program there is often substantial code which checks
4190 correct functioning of the program and is, hopefully, unreachable
4191 because the program does work. Another common use of unreachable
4192 code is to provide behavior which is selectable at compile-time.
4197 Warn if a function can not be inlined and it was declared as inline.
4198 Even with this option, the compiler will not warn about failures to
4199 inline functions declared in system headers.
4201 The compiler uses a variety of heuristics to determine whether or not
4202 to inline a function. For example, the compiler takes into account
4203 the size of the function being inlined and the amount of inlining
4204 that has already been done in the current function. Therefore,
4205 seemingly insignificant changes in the source program can cause the
4206 warnings produced by @option{-Winline} to appear or disappear.
4208 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
4209 @opindex Wno-invalid-offsetof
4210 @opindex Winvalid-offsetof
4211 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
4212 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
4213 to a non-POD type is undefined. In existing C++ implementations,
4214 however, @samp{offsetof} typically gives meaningful results even when
4215 applied to certain kinds of non-POD types. (Such as a simple
4216 @samp{struct} that fails to be a POD type only by virtue of having a
4217 constructor.) This flag is for users who are aware that they are
4218 writing nonportable code and who have deliberately chosen to ignore the
4221 The restrictions on @samp{offsetof} may be relaxed in a future version
4222 of the C++ standard.
4224 @item -Wno-int-to-pointer-cast @r{(C and Objective-C only)}
4225 @opindex Wno-int-to-pointer-cast
4226 @opindex Wint-to-pointer-cast
4227 Suppress warnings from casts to pointer type of an integer of a
4230 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
4231 @opindex Wno-pointer-to-int-cast
4232 @opindex Wpointer-to-int-cast
4233 Suppress warnings from casts from a pointer to an integer type of a
4237 @opindex Winvalid-pch
4238 @opindex Wno-invalid-pch
4239 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
4240 the search path but can't be used.
4244 @opindex Wno-long-long
4245 Warn if @samp{long long} type is used. This is enabled by either
4246 @option{-pedantic} or @option{-Wtraditional} in ISO C90 and C++98
4247 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
4249 @item -Wvariadic-macros
4250 @opindex Wvariadic-macros
4251 @opindex Wno-variadic-macros
4252 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
4253 alternate syntax when in pedantic ISO C99 mode. This is default.
4254 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
4259 Warn if variable length array is used in the code.
4260 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
4261 the variable length array.
4263 @item -Wvolatile-register-var
4264 @opindex Wvolatile-register-var
4265 @opindex Wno-volatile-register-var
4266 Warn if a register variable is declared volatile. The volatile
4267 modifier does not inhibit all optimizations that may eliminate reads
4268 and/or writes to register variables. This warning is enabled by
4271 @item -Wdisabled-optimization
4272 @opindex Wdisabled-optimization
4273 @opindex Wno-disabled-optimization
4274 Warn if a requested optimization pass is disabled. This warning does
4275 not generally indicate that there is anything wrong with your code; it
4276 merely indicates that GCC's optimizers were unable to handle the code
4277 effectively. Often, the problem is that your code is too big or too
4278 complex; GCC will refuse to optimize programs when the optimization
4279 itself is likely to take inordinate amounts of time.
4281 @item -Wpointer-sign @r{(C and Objective-C only)}
4282 @opindex Wpointer-sign
4283 @opindex Wno-pointer-sign
4284 Warn for pointer argument passing or assignment with different signedness.
4285 This option is only supported for C and Objective-C@. It is implied by
4286 @option{-Wall} and by @option{-pedantic}, which can be disabled with
4287 @option{-Wno-pointer-sign}.
4289 @item -Wstack-protector
4290 @opindex Wstack-protector
4291 @opindex Wno-stack-protector
4292 This option is only active when @option{-fstack-protector} is active. It
4293 warns about functions that will not be protected against stack smashing.
4296 @opindex Wno-mudflap
4297 Suppress warnings about constructs that cannot be instrumented by
4300 @item -Woverlength-strings
4301 @opindex Woverlength-strings
4302 @opindex Wno-overlength-strings
4303 Warn about string constants which are longer than the ``minimum
4304 maximum'' length specified in the C standard. Modern compilers
4305 generally allow string constants which are much longer than the
4306 standard's minimum limit, but very portable programs should avoid
4307 using longer strings.
4309 The limit applies @emph{after} string constant concatenation, and does
4310 not count the trailing NUL@. In C89, the limit was 509 characters; in
4311 C99, it was raised to 4095. C++98 does not specify a normative
4312 minimum maximum, so we do not diagnose overlength strings in C++@.
4314 This option is implied by @option{-pedantic}, and can be disabled with
4315 @option{-Wno-overlength-strings}.
4317 @item -Wunsuffixed-float-constants
4318 @opindex Wunsuffixed-float-constants
4320 GCC will issue a warning for any floating constant that does not have
4321 a suffix. When used together with @option{-Wsystem-headers} it will
4322 warn about such constants in system header files. This can be useful
4323 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
4324 from the decimal floating-point extension to C99.
4327 @node Debugging Options
4328 @section Options for Debugging Your Program or GCC
4329 @cindex options, debugging
4330 @cindex debugging information options
4332 GCC has various special options that are used for debugging
4333 either your program or GCC:
4338 Produce debugging information in the operating system's native format
4339 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
4342 On most systems that use stabs format, @option{-g} enables use of extra
4343 debugging information that only GDB can use; this extra information
4344 makes debugging work better in GDB but will probably make other debuggers
4346 refuse to read the program. If you want to control for certain whether
4347 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
4348 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
4350 GCC allows you to use @option{-g} with
4351 @option{-O}. The shortcuts taken by optimized code may occasionally
4352 produce surprising results: some variables you declared may not exist
4353 at all; flow of control may briefly move where you did not expect it;
4354 some statements may not be executed because they compute constant
4355 results or their values were already at hand; some statements may
4356 execute in different places because they were moved out of loops.
4358 Nevertheless it proves possible to debug optimized output. This makes
4359 it reasonable to use the optimizer for programs that might have bugs.
4361 The following options are useful when GCC is generated with the
4362 capability for more than one debugging format.
4366 Produce debugging information for use by GDB@. This means to use the
4367 most expressive format available (DWARF 2, stabs, or the native format
4368 if neither of those are supported), including GDB extensions if at all
4373 Produce debugging information in stabs format (if that is supported),
4374 without GDB extensions. This is the format used by DBX on most BSD
4375 systems. On MIPS, Alpha and System V Release 4 systems this option
4376 produces stabs debugging output which is not understood by DBX or SDB@.
4377 On System V Release 4 systems this option requires the GNU assembler.
4379 @item -feliminate-unused-debug-symbols
4380 @opindex feliminate-unused-debug-symbols
4381 Produce debugging information in stabs format (if that is supported),
4382 for only symbols that are actually used.
4384 @item -femit-class-debug-always
4385 Instead of emitting debugging information for a C++ class in only one
4386 object file, emit it in all object files using the class. This option
4387 should be used only with debuggers that are unable to handle the way GCC
4388 normally emits debugging information for classes because using this
4389 option will increase the size of debugging information by as much as a
4394 Produce debugging information in stabs format (if that is supported),
4395 using GNU extensions understood only by the GNU debugger (GDB)@. The
4396 use of these extensions is likely to make other debuggers crash or
4397 refuse to read the program.
4401 Produce debugging information in COFF format (if that is supported).
4402 This is the format used by SDB on most System V systems prior to
4407 Produce debugging information in XCOFF format (if that is supported).
4408 This is the format used by the DBX debugger on IBM RS/6000 systems.
4412 Produce debugging information in XCOFF format (if that is supported),
4413 using GNU extensions understood only by the GNU debugger (GDB)@. The
4414 use of these extensions is likely to make other debuggers crash or
4415 refuse to read the program, and may cause assemblers other than the GNU
4416 assembler (GAS) to fail with an error.
4418 @item -gdwarf-@var{version}
4419 @opindex gdwarf-@var{version}
4420 Produce debugging information in DWARF format (if that is
4421 supported). This is the format used by DBX on IRIX 6. The value
4422 of @var{version} may be either 2, 3 or 4; the default version is 2.
4424 Note that with DWARF version 2 some ports require, and will always
4425 use, some non-conflicting DWARF 3 extensions in the unwind tables.
4427 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
4428 for maximum benefit.
4432 Produce debugging information in VMS debug format (if that is
4433 supported). This is the format used by DEBUG on VMS systems.
4436 @itemx -ggdb@var{level}
4437 @itemx -gstabs@var{level}
4438 @itemx -gcoff@var{level}
4439 @itemx -gxcoff@var{level}
4440 @itemx -gvms@var{level}
4441 Request debugging information and also use @var{level} to specify how
4442 much information. The default level is 2.
4444 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4447 Level 1 produces minimal information, enough for making backtraces in
4448 parts of the program that you don't plan to debug. This includes
4449 descriptions of functions and external variables, but no information
4450 about local variables and no line numbers.
4452 Level 3 includes extra information, such as all the macro definitions
4453 present in the program. Some debuggers support macro expansion when
4454 you use @option{-g3}.
4456 @option{-gdwarf-2} does not accept a concatenated debug level, because
4457 GCC used to support an option @option{-gdwarf} that meant to generate
4458 debug information in version 1 of the DWARF format (which is very
4459 different from version 2), and it would have been too confusing. That
4460 debug format is long obsolete, but the option cannot be changed now.
4461 Instead use an additional @option{-g@var{level}} option to change the
4462 debug level for DWARF.
4466 Turn off generation of debug info, if leaving out this option would have
4467 generated it, or turn it on at level 2 otherwise. The position of this
4468 argument in the command line does not matter, it takes effect after all
4469 other options are processed, and it does so only once, no matter how
4470 many times it is given. This is mainly intended to be used with
4471 @option{-fcompare-debug}.
4473 @item -fdump-final-insns@r{[}=@var{file}@r{]}
4474 @opindex fdump-final-insns
4475 Dump the final internal representation (RTL) to @var{file}. If the
4476 optional argument is omitted (or if @var{file} is @code{.}), the name
4477 of the dump file will be determined by appending @code{.gkd} to the
4478 compilation output file name.
4480 @item -fcompare-debug@r{[}=@var{opts}@r{]}
4481 @opindex fcompare-debug
4482 @opindex fno-compare-debug
4483 If no error occurs during compilation, run the compiler a second time,
4484 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
4485 passed to the second compilation. Dump the final internal
4486 representation in both compilations, and print an error if they differ.
4488 If the equal sign is omitted, the default @option{-gtoggle} is used.
4490 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
4491 and nonzero, implicitly enables @option{-fcompare-debug}. If
4492 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
4493 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
4496 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
4497 is equivalent to @option{-fno-compare-debug}, which disables the dumping
4498 of the final representation and the second compilation, preventing even
4499 @env{GCC_COMPARE_DEBUG} from taking effect.
4501 To verify full coverage during @option{-fcompare-debug} testing, set
4502 @env{GCC_COMPARE_DEBUG} to say @samp{-fcompare-debug-not-overridden},
4503 which GCC will reject as an invalid option in any actual compilation
4504 (rather than preprocessing, assembly or linking). To get just a
4505 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
4506 not overridden} will do.
4508 @item -fcompare-debug-second
4509 @opindex fcompare-debug-second
4510 This option is implicitly passed to the compiler for the second
4511 compilation requested by @option{-fcompare-debug}, along with options to
4512 silence warnings, and omitting other options that would cause
4513 side-effect compiler outputs to files or to the standard output. Dump
4514 files and preserved temporary files are renamed so as to contain the
4515 @code{.gk} additional extension during the second compilation, to avoid
4516 overwriting those generated by the first.
4518 When this option is passed to the compiler driver, it causes the
4519 @emph{first} compilation to be skipped, which makes it useful for little
4520 other than debugging the compiler proper.
4522 @item -feliminate-dwarf2-dups
4523 @opindex feliminate-dwarf2-dups
4524 Compress DWARF2 debugging information by eliminating duplicated
4525 information about each symbol. This option only makes sense when
4526 generating DWARF2 debugging information with @option{-gdwarf-2}.
4528 @item -femit-struct-debug-baseonly
4529 Emit debug information for struct-like types
4530 only when the base name of the compilation source file
4531 matches the base name of file in which the struct was defined.
4533 This option substantially reduces the size of debugging information,
4534 but at significant potential loss in type information to the debugger.
4535 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4536 See @option{-femit-struct-debug-detailed} for more detailed control.
4538 This option works only with DWARF 2.
4540 @item -femit-struct-debug-reduced
4541 Emit debug information for struct-like types
4542 only when the base name of the compilation source file
4543 matches the base name of file in which the type was defined,
4544 unless the struct is a template or defined in a system header.
4546 This option significantly reduces the size of debugging information,
4547 with some potential loss in type information to the debugger.
4548 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4549 See @option{-femit-struct-debug-detailed} for more detailed control.
4551 This option works only with DWARF 2.
4553 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4554 Specify the struct-like types
4555 for which the compiler will generate debug information.
4556 The intent is to reduce duplicate struct debug information
4557 between different object files within the same program.
4559 This option is a detailed version of
4560 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4561 which will serve for most needs.
4563 A specification has the syntax
4564 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4566 The optional first word limits the specification to
4567 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4568 A struct type is used directly when it is the type of a variable, member.
4569 Indirect uses arise through pointers to structs.
4570 That is, when use of an incomplete struct would be legal, the use is indirect.
4572 @samp{struct one direct; struct two * indirect;}.
4574 The optional second word limits the specification to
4575 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4576 Generic structs are a bit complicated to explain.
4577 For C++, these are non-explicit specializations of template classes,
4578 or non-template classes within the above.
4579 Other programming languages have generics,
4580 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4582 The third word specifies the source files for those
4583 structs for which the compiler will emit debug information.
4584 The values @samp{none} and @samp{any} have the normal meaning.
4585 The value @samp{base} means that
4586 the base of name of the file in which the type declaration appears
4587 must match the base of the name of the main compilation file.
4588 In practice, this means that
4589 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4590 but types declared in other header will not.
4591 The value @samp{sys} means those types satisfying @samp{base}
4592 or declared in system or compiler headers.
4594 You may need to experiment to determine the best settings for your application.
4596 The default is @samp{-femit-struct-debug-detailed=all}.
4598 This option works only with DWARF 2.
4600 @item -fno-merge-debug-strings
4601 @opindex fmerge-debug-strings
4602 @opindex fno-merge-debug-strings
4603 Direct the linker to not merge together strings in the debugging
4604 information which are identical in different object files. Merging is
4605 not supported by all assemblers or linkers. Merging decreases the size
4606 of the debug information in the output file at the cost of increasing
4607 link processing time. Merging is enabled by default.
4609 @item -fdebug-prefix-map=@var{old}=@var{new}
4610 @opindex fdebug-prefix-map
4611 When compiling files in directory @file{@var{old}}, record debugging
4612 information describing them as in @file{@var{new}} instead.
4614 @item -fno-dwarf2-cfi-asm
4615 @opindex fdwarf2-cfi-asm
4616 @opindex fno-dwarf2-cfi-asm
4617 Emit DWARF 2 unwind info as compiler generated @code{.eh_frame} section
4618 instead of using GAS @code{.cfi_*} directives.
4620 @cindex @command{prof}
4623 Generate extra code to write profile information suitable for the
4624 analysis program @command{prof}. You must use this option when compiling
4625 the source files you want data about, and you must also use it when
4628 @cindex @command{gprof}
4631 Generate extra code to write profile information suitable for the
4632 analysis program @command{gprof}. You must use this option when compiling
4633 the source files you want data about, and you must also use it when
4638 Makes the compiler print out each function name as it is compiled, and
4639 print some statistics about each pass when it finishes.
4642 @opindex ftime-report
4643 Makes the compiler print some statistics about the time consumed by each
4644 pass when it finishes.
4647 @opindex fmem-report
4648 Makes the compiler print some statistics about permanent memory
4649 allocation when it finishes.
4651 @item -fpre-ipa-mem-report
4652 @opindex fpre-ipa-mem-report
4653 @item -fpost-ipa-mem-report
4654 @opindex fpost-ipa-mem-report
4655 Makes the compiler print some statistics about permanent memory
4656 allocation before or after interprocedural optimization.
4658 @item -fprofile-arcs
4659 @opindex fprofile-arcs
4660 Add code so that program flow @dfn{arcs} are instrumented. During
4661 execution the program records how many times each branch and call is
4662 executed and how many times it is taken or returns. When the compiled
4663 program exits it saves this data to a file called
4664 @file{@var{auxname}.gcda} for each source file. The data may be used for
4665 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4666 test coverage analysis (@option{-ftest-coverage}). Each object file's
4667 @var{auxname} is generated from the name of the output file, if
4668 explicitly specified and it is not the final executable, otherwise it is
4669 the basename of the source file. In both cases any suffix is removed
4670 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4671 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4672 @xref{Cross-profiling}.
4674 @cindex @command{gcov}
4678 This option is used to compile and link code instrumented for coverage
4679 analysis. The option is a synonym for @option{-fprofile-arcs}
4680 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4681 linking). See the documentation for those options for more details.
4686 Compile the source files with @option{-fprofile-arcs} plus optimization
4687 and code generation options. For test coverage analysis, use the
4688 additional @option{-ftest-coverage} option. You do not need to profile
4689 every source file in a program.
4692 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4693 (the latter implies the former).
4696 Run the program on a representative workload to generate the arc profile
4697 information. This may be repeated any number of times. You can run
4698 concurrent instances of your program, and provided that the file system
4699 supports locking, the data files will be correctly updated. Also
4700 @code{fork} calls are detected and correctly handled (double counting
4704 For profile-directed optimizations, compile the source files again with
4705 the same optimization and code generation options plus
4706 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4707 Control Optimization}).
4710 For test coverage analysis, use @command{gcov} to produce human readable
4711 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4712 @command{gcov} documentation for further information.
4716 With @option{-fprofile-arcs}, for each function of your program GCC
4717 creates a program flow graph, then finds a spanning tree for the graph.
4718 Only arcs that are not on the spanning tree have to be instrumented: the
4719 compiler adds code to count the number of times that these arcs are
4720 executed. When an arc is the only exit or only entrance to a block, the
4721 instrumentation code can be added to the block; otherwise, a new basic
4722 block must be created to hold the instrumentation code.
4725 @item -ftest-coverage
4726 @opindex ftest-coverage
4727 Produce a notes file that the @command{gcov} code-coverage utility
4728 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4729 show program coverage. Each source file's note file is called
4730 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4731 above for a description of @var{auxname} and instructions on how to
4732 generate test coverage data. Coverage data will match the source files
4733 more closely, if you do not optimize.
4735 @item -fdbg-cnt-list
4736 @opindex fdbg-cnt-list
4737 Print the name and the counter upperbound for all debug counters.
4739 @item -fdbg-cnt=@var{counter-value-list}
4741 Set the internal debug counter upperbound. @var{counter-value-list}
4742 is a comma-separated list of @var{name}:@var{value} pairs
4743 which sets the upperbound of each debug counter @var{name} to @var{value}.
4744 All debug counters have the initial upperbound of @var{UINT_MAX},
4745 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4746 e.g. With -fdbg-cnt=dce:10,tail_call:0
4747 dbg_cnt(dce) will return true only for first 10 invocations
4748 and dbg_cnt(tail_call) will return false always.
4750 @item -d@var{letters}
4751 @itemx -fdump-rtl-@var{pass}
4753 Says to make debugging dumps during compilation at times specified by
4754 @var{letters}. This is used for debugging the RTL-based passes of the
4755 compiler. The file names for most of the dumps are made by appending
4756 a pass number and a word to the @var{dumpname}, and the files are
4757 created in the directory of the output file. @var{dumpname} is
4758 generated from the name of the output file, if explicitly specified
4759 and it is not an executable, otherwise it is the basename of the
4760 source file. These switches may have different effects when
4761 @option{-E} is used for preprocessing.
4763 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
4764 @option{-d} option @var{letters}. Here are the possible
4765 letters for use in @var{pass} and @var{letters}, and their meanings:
4769 @item -fdump-rtl-alignments
4770 @opindex fdump-rtl-alignments
4771 Dump after branch alignments have been computed.
4773 @item -fdump-rtl-asmcons
4774 @opindex fdump-rtl-asmcons
4775 Dump after fixing rtl statements that have unsatisfied in/out constraints.
4777 @item -fdump-rtl-auto_inc_dec
4778 @opindex fdump-rtl-auto_inc_dec
4779 Dump after auto-inc-dec discovery. This pass is only run on
4780 architectures that have auto inc or auto dec instructions.
4782 @item -fdump-rtl-barriers
4783 @opindex fdump-rtl-barriers
4784 Dump after cleaning up the barrier instructions.
4786 @item -fdump-rtl-bbpart
4787 @opindex fdump-rtl-bbpart
4788 Dump after partitioning hot and cold basic blocks.
4790 @item -fdump-rtl-bbro
4791 @opindex fdump-rtl-bbro
4792 Dump after block reordering.
4794 @item -fdump-rtl-btl1
4795 @itemx -fdump-rtl-btl2
4796 @opindex fdump-rtl-btl2
4797 @opindex fdump-rtl-btl2
4798 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
4799 after the two branch
4800 target load optimization passes.
4802 @item -fdump-rtl-bypass
4803 @opindex fdump-rtl-bypass
4804 Dump after jump bypassing and control flow optimizations.
4806 @item -fdump-rtl-combine
4807 @opindex fdump-rtl-combine
4808 Dump after the RTL instruction combination pass.
4810 @item -fdump-rtl-compgotos
4811 @opindex fdump-rtl-compgotos
4812 Dump after duplicating the computed gotos.
4814 @item -fdump-rtl-ce1
4815 @itemx -fdump-rtl-ce2
4816 @itemx -fdump-rtl-ce3
4817 @opindex fdump-rtl-ce1
4818 @opindex fdump-rtl-ce2
4819 @opindex fdump-rtl-ce3
4820 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
4821 @option{-fdump-rtl-ce3} enable dumping after the three
4822 if conversion passes.
4824 @itemx -fdump-rtl-cprop_hardreg
4825 @opindex fdump-rtl-cprop_hardreg
4826 Dump after hard register copy propagation.
4828 @itemx -fdump-rtl-csa
4829 @opindex fdump-rtl-csa
4830 Dump after combining stack adjustments.
4832 @item -fdump-rtl-cse1
4833 @itemx -fdump-rtl-cse2
4834 @opindex fdump-rtl-cse1
4835 @opindex fdump-rtl-cse2
4836 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
4837 the two common sub-expression elimination passes.
4839 @itemx -fdump-rtl-dce
4840 @opindex fdump-rtl-dce
4841 Dump after the standalone dead code elimination passes.
4843 @itemx -fdump-rtl-dbr
4844 @opindex fdump-rtl-dbr
4845 Dump after delayed branch scheduling.
4847 @item -fdump-rtl-dce1
4848 @itemx -fdump-rtl-dce2
4849 @opindex fdump-rtl-dce1
4850 @opindex fdump-rtl-dce2
4851 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
4852 the two dead store elimination passes.
4855 @opindex fdump-rtl-eh
4856 Dump after finalization of EH handling code.
4858 @item -fdump-rtl-eh_ranges
4859 @opindex fdump-rtl-eh_ranges
4860 Dump after conversion of EH handling range regions.
4862 @item -fdump-rtl-expand
4863 @opindex fdump-rtl-expand
4864 Dump after RTL generation.
4866 @item -fdump-rtl-fwprop1
4867 @itemx -fdump-rtl-fwprop2
4868 @opindex fdump-rtl-fwprop1
4869 @opindex fdump-rtl-fwprop2
4870 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
4871 dumping after the two forward propagation passes.
4873 @item -fdump-rtl-gcse1
4874 @itemx -fdump-rtl-gcse2
4875 @opindex fdump-rtl-gcse1
4876 @opindex fdump-rtl-gcse2
4877 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
4878 after global common subexpression elimination.
4880 @item -fdump-rtl-init-regs
4881 @opindex fdump-rtl-init-regs
4882 Dump after the initialization of the registers.
4884 @item -fdump-rtl-initvals
4885 @opindex fdump-rtl-initvals
4886 Dump after the computation of the initial value sets.
4888 @itemx -fdump-rtl-into_cfglayout
4889 @opindex fdump-rtl-into_cfglayout
4890 Dump after converting to cfglayout mode.
4892 @item -fdump-rtl-ira
4893 @opindex fdump-rtl-ira
4894 Dump after iterated register allocation.
4896 @item -fdump-rtl-jump
4897 @opindex fdump-rtl-jump
4898 Dump after the second jump optimization.
4900 @item -fdump-rtl-loop2
4901 @opindex fdump-rtl-loop2
4902 @option{-fdump-rtl-loop2} enables dumping after the rtl
4903 loop optimization passes.
4905 @item -fdump-rtl-mach
4906 @opindex fdump-rtl-mach
4907 Dump after performing the machine dependent reorganization pass, if that
4910 @item -fdump-rtl-mode_sw
4911 @opindex fdump-rtl-mode_sw
4912 Dump after removing redundant mode switches.
4914 @item -fdump-rtl-rnreg
4915 @opindex fdump-rtl-rnreg
4916 Dump after register renumbering.
4918 @itemx -fdump-rtl-outof_cfglayout
4919 @opindex fdump-rtl-outof_cfglayout
4920 Dump after converting from cfglayout mode.
4922 @item -fdump-rtl-peephole2
4923 @opindex fdump-rtl-peephole2
4924 Dump after the peephole pass.
4926 @item -fdump-rtl-postreload
4927 @opindex fdump-rtl-postreload
4928 Dump after post-reload optimizations.
4930 @itemx -fdump-rtl-pro_and_epilogue
4931 @opindex fdump-rtl-pro_and_epilogue
4932 Dump after generating the function pro and epilogues.
4934 @item -fdump-rtl-regmove
4935 @opindex fdump-rtl-regmove
4936 Dump after the register move pass.
4938 @item -fdump-rtl-sched1
4939 @itemx -fdump-rtl-sched2
4940 @opindex fdump-rtl-sched1
4941 @opindex fdump-rtl-sched2
4942 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
4943 after the basic block scheduling passes.
4945 @item -fdump-rtl-see
4946 @opindex fdump-rtl-see
4947 Dump after sign extension elimination.
4949 @item -fdump-rtl-seqabstr
4950 @opindex fdump-rtl-seqabstr
4951 Dump after common sequence discovery.
4953 @item -fdump-rtl-shorten
4954 @opindex fdump-rtl-shorten
4955 Dump after shortening branches.
4957 @item -fdump-rtl-sibling
4958 @opindex fdump-rtl-sibling
4959 Dump after sibling call optimizations.
4961 @item -fdump-rtl-split1
4962 @itemx -fdump-rtl-split2
4963 @itemx -fdump-rtl-split3
4964 @itemx -fdump-rtl-split4
4965 @itemx -fdump-rtl-split5
4966 @opindex fdump-rtl-split1
4967 @opindex fdump-rtl-split2
4968 @opindex fdump-rtl-split3
4969 @opindex fdump-rtl-split4
4970 @opindex fdump-rtl-split5
4971 @option{-fdump-rtl-split1}, @option{-fdump-rtl-split2},
4972 @option{-fdump-rtl-split3}, @option{-fdump-rtl-split4} and
4973 @option{-fdump-rtl-split5} enable dumping after five rounds of
4974 instruction splitting.
4976 @item -fdump-rtl-sms
4977 @opindex fdump-rtl-sms
4978 Dump after modulo scheduling. This pass is only run on some
4981 @item -fdump-rtl-stack
4982 @opindex fdump-rtl-stack
4983 Dump after conversion from GCC's "flat register file" registers to the
4984 x87's stack-like registers. This pass is only run on x86 variants.
4986 @item -fdump-rtl-subreg1
4987 @itemx -fdump-rtl-subreg2
4988 @opindex fdump-rtl-subreg1
4989 @opindex fdump-rtl-subreg2
4990 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
4991 the two subreg expansion passes.
4993 @item -fdump-rtl-unshare
4994 @opindex fdump-rtl-unshare
4995 Dump after all rtl has been unshared.
4997 @item -fdump-rtl-vartrack
4998 @opindex fdump-rtl-vartrack
4999 Dump after variable tracking.
5001 @item -fdump-rtl-vregs
5002 @opindex fdump-rtl-vregs
5003 Dump after converting virtual registers to hard registers.
5005 @item -fdump-rtl-web
5006 @opindex fdump-rtl-web
5007 Dump after live range splitting.
5009 @item -fdump-rtl-regclass
5010 @itemx -fdump-rtl-subregs_of_mode_init
5011 @itemx -fdump-rtl-subregs_of_mode_finish
5012 @itemx -fdump-rtl-dfinit
5013 @itemx -fdump-rtl-dfinish
5014 @opindex fdump-rtl-regclass
5015 @opindex fdump-rtl-subregs_of_mode_init
5016 @opindex fdump-rtl-subregs_of_mode_finish
5017 @opindex fdump-rtl-dfinit
5018 @opindex fdump-rtl-dfinish
5019 These dumps are defined but always produce empty files.
5021 @item -fdump-rtl-all
5022 @opindex fdump-rtl-all
5023 Produce all the dumps listed above.
5027 Annotate the assembler output with miscellaneous debugging information.
5031 Dump all macro definitions, at the end of preprocessing, in addition to
5036 Produce a core dump whenever an error occurs.
5040 Print statistics on memory usage, at the end of the run, to
5045 Annotate the assembler output with a comment indicating which
5046 pattern and alternative was used. The length of each instruction is
5051 Dump the RTL in the assembler output as a comment before each instruction.
5052 Also turns on @option{-dp} annotation.
5056 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
5057 dump a representation of the control flow graph suitable for viewing with VCG
5058 to @file{@var{file}.@var{pass}.vcg}.
5062 Just generate RTL for a function instead of compiling it. Usually used
5063 with @option{-fdump-rtl-expand}.
5067 Dump debugging information during parsing, to standard error.
5071 @opindex fdump-noaddr
5072 When doing debugging dumps, suppress address output. This makes it more
5073 feasible to use diff on debugging dumps for compiler invocations with
5074 different compiler binaries and/or different
5075 text / bss / data / heap / stack / dso start locations.
5077 @item -fdump-unnumbered
5078 @opindex fdump-unnumbered
5079 When doing debugging dumps, suppress instruction numbers and address output.
5080 This makes it more feasible to use diff on debugging dumps for compiler
5081 invocations with different options, in particular with and without
5084 @item -fdump-unnumbered-links
5085 @opindex fdump-unnumbered-links
5086 When doing debugging dumps (see @option{-d} option above), suppress
5087 instruction numbers for the links to the previous and next instructions
5090 @item -fdump-translation-unit @r{(C++ only)}
5091 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
5092 @opindex fdump-translation-unit
5093 Dump a representation of the tree structure for the entire translation
5094 unit to a file. The file name is made by appending @file{.tu} to the
5095 source file name, and the file is created in the same directory as the
5096 output file. If the @samp{-@var{options}} form is used, @var{options}
5097 controls the details of the dump as described for the
5098 @option{-fdump-tree} options.
5100 @item -fdump-class-hierarchy @r{(C++ only)}
5101 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
5102 @opindex fdump-class-hierarchy
5103 Dump a representation of each class's hierarchy and virtual function
5104 table layout to a file. The file name is made by appending
5105 @file{.class} to the source file name, and the file is created in the
5106 same directory as the output file. If the @samp{-@var{options}} form
5107 is used, @var{options} controls the details of the dump as described
5108 for the @option{-fdump-tree} options.
5110 @item -fdump-ipa-@var{switch}
5112 Control the dumping at various stages of inter-procedural analysis
5113 language tree to a file. The file name is generated by appending a
5114 switch specific suffix to the source file name, and the file is created
5115 in the same directory as the output file. The following dumps are
5120 Enables all inter-procedural analysis dumps.
5123 Dumps information about call-graph optimization, unused function removal,
5124 and inlining decisions.
5127 Dump after function inlining.
5131 @item -fdump-statistics-@var{option}
5132 @opindex fdump-statistics
5133 Enable and control dumping of pass statistics in a separate file. The
5134 file name is generated by appending a suffix ending in
5135 @samp{.statistics} to the source file name, and the file is created in
5136 the same directory as the output file. If the @samp{-@var{option}}
5137 form is used, @samp{-stats} will cause counters to be summed over the
5138 whole compilation unit while @samp{-details} will dump every event as
5139 the passes generate them. The default with no option is to sum
5140 counters for each function compiled.
5142 @item -fdump-tree-@var{switch}
5143 @itemx -fdump-tree-@var{switch}-@var{options}
5145 Control the dumping at various stages of processing the intermediate
5146 language tree to a file. The file name is generated by appending a
5147 switch specific suffix to the source file name, and the file is
5148 created in the same directory as the output file. If the
5149 @samp{-@var{options}} form is used, @var{options} is a list of
5150 @samp{-} separated options that control the details of the dump. Not
5151 all options are applicable to all dumps, those which are not
5152 meaningful will be ignored. The following options are available
5156 Print the address of each node. Usually this is not meaningful as it
5157 changes according to the environment and source file. Its primary use
5158 is for tying up a dump file with a debug environment.
5160 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
5161 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
5162 use working backward from mangled names in the assembly file.
5164 Inhibit dumping of members of a scope or body of a function merely
5165 because that scope has been reached. Only dump such items when they
5166 are directly reachable by some other path. When dumping pretty-printed
5167 trees, this option inhibits dumping the bodies of control structures.
5169 Print a raw representation of the tree. By default, trees are
5170 pretty-printed into a C-like representation.
5172 Enable more detailed dumps (not honored by every dump option).
5174 Enable dumping various statistics about the pass (not honored by every dump
5177 Enable showing basic block boundaries (disabled in raw dumps).
5179 Enable showing virtual operands for every statement.
5181 Enable showing line numbers for statements.
5183 Enable showing the unique ID (@code{DECL_UID}) for each variable.
5185 Enable showing the tree dump for each statement.
5187 Enable showing the EH region number holding each statement.
5189 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
5190 and @option{lineno}.
5193 The following tree dumps are possible:
5197 @opindex fdump-tree-original
5198 Dump before any tree based optimization, to @file{@var{file}.original}.
5201 @opindex fdump-tree-optimized
5202 Dump after all tree based optimization, to @file{@var{file}.optimized}.
5205 @opindex fdump-tree-gimple
5206 Dump each function before and after the gimplification pass to a file. The
5207 file name is made by appending @file{.gimple} to the source file name.
5210 @opindex fdump-tree-cfg
5211 Dump the control flow graph of each function to a file. The file name is
5212 made by appending @file{.cfg} to the source file name.
5215 @opindex fdump-tree-vcg
5216 Dump the control flow graph of each function to a file in VCG format. The
5217 file name is made by appending @file{.vcg} to the source file name. Note
5218 that if the file contains more than one function, the generated file cannot
5219 be used directly by VCG@. You will need to cut and paste each function's
5220 graph into its own separate file first.
5223 @opindex fdump-tree-ch
5224 Dump each function after copying loop headers. The file name is made by
5225 appending @file{.ch} to the source file name.
5228 @opindex fdump-tree-ssa
5229 Dump SSA related information to a file. The file name is made by appending
5230 @file{.ssa} to the source file name.
5233 @opindex fdump-tree-alias
5234 Dump aliasing information for each function. The file name is made by
5235 appending @file{.alias} to the source file name.
5238 @opindex fdump-tree-ccp
5239 Dump each function after CCP@. The file name is made by appending
5240 @file{.ccp} to the source file name.
5243 @opindex fdump-tree-storeccp
5244 Dump each function after STORE-CCP@. The file name is made by appending
5245 @file{.storeccp} to the source file name.
5248 @opindex fdump-tree-pre
5249 Dump trees after partial redundancy elimination. The file name is made
5250 by appending @file{.pre} to the source file name.
5253 @opindex fdump-tree-fre
5254 Dump trees after full redundancy elimination. The file name is made
5255 by appending @file{.fre} to the source file name.
5258 @opindex fdump-tree-copyprop
5259 Dump trees after copy propagation. The file name is made
5260 by appending @file{.copyprop} to the source file name.
5262 @item store_copyprop
5263 @opindex fdump-tree-store_copyprop
5264 Dump trees after store copy-propagation. The file name is made
5265 by appending @file{.store_copyprop} to the source file name.
5268 @opindex fdump-tree-dce
5269 Dump each function after dead code elimination. The file name is made by
5270 appending @file{.dce} to the source file name.
5273 @opindex fdump-tree-mudflap
5274 Dump each function after adding mudflap instrumentation. The file name is
5275 made by appending @file{.mudflap} to the source file name.
5278 @opindex fdump-tree-sra
5279 Dump each function after performing scalar replacement of aggregates. The
5280 file name is made by appending @file{.sra} to the source file name.
5283 @opindex fdump-tree-sink
5284 Dump each function after performing code sinking. The file name is made
5285 by appending @file{.sink} to the source file name.
5288 @opindex fdump-tree-dom
5289 Dump each function after applying dominator tree optimizations. The file
5290 name is made by appending @file{.dom} to the source file name.
5293 @opindex fdump-tree-dse
5294 Dump each function after applying dead store elimination. The file
5295 name is made by appending @file{.dse} to the source file name.
5298 @opindex fdump-tree-phiopt
5299 Dump each function after optimizing PHI nodes into straightline code. The file
5300 name is made by appending @file{.phiopt} to the source file name.
5303 @opindex fdump-tree-forwprop
5304 Dump each function after forward propagating single use variables. The file
5305 name is made by appending @file{.forwprop} to the source file name.
5308 @opindex fdump-tree-copyrename
5309 Dump each function after applying the copy rename optimization. The file
5310 name is made by appending @file{.copyrename} to the source file name.
5313 @opindex fdump-tree-nrv
5314 Dump each function after applying the named return value optimization on
5315 generic trees. The file name is made by appending @file{.nrv} to the source
5319 @opindex fdump-tree-vect
5320 Dump each function after applying vectorization of loops. The file name is
5321 made by appending @file{.vect} to the source file name.
5324 @opindex fdump-tree-vrp
5325 Dump each function after Value Range Propagation (VRP). The file name
5326 is made by appending @file{.vrp} to the source file name.
5329 @opindex fdump-tree-all
5330 Enable all the available tree dumps with the flags provided in this option.
5333 @item -ftree-vectorizer-verbose=@var{n}
5334 @opindex ftree-vectorizer-verbose
5335 This option controls the amount of debugging output the vectorizer prints.
5336 This information is written to standard error, unless
5337 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
5338 in which case it is output to the usual dump listing file, @file{.vect}.
5339 For @var{n}=0 no diagnostic information is reported.
5340 If @var{n}=1 the vectorizer reports each loop that got vectorized,
5341 and the total number of loops that got vectorized.
5342 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
5343 the first analysis phase (vect_analyze_loop_form) - i.e.@: countable,
5344 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
5345 level that @option{-fdump-tree-vect-stats} uses.
5346 Higher verbosity levels mean either more information dumped for each
5347 reported loop, or same amount of information reported for more loops:
5348 If @var{n}=3, alignment related information is added to the reports.
5349 If @var{n}=4, data-references related information (e.g.@: memory dependences,
5350 memory access-patterns) is added to the reports.
5351 If @var{n}=5, the vectorizer reports also non-vectorized inner-most loops
5352 that did not pass the first analysis phase (i.e., may not be countable, or
5353 may have complicated control-flow).
5354 If @var{n}=6, the vectorizer reports also non-vectorized nested loops.
5355 For @var{n}=7, all the information the vectorizer generates during its
5356 analysis and transformation is reported. This is the same verbosity level
5357 that @option{-fdump-tree-vect-details} uses.
5359 @item -frandom-seed=@var{string}
5360 @opindex frandom-seed
5361 This option provides a seed that GCC uses when it would otherwise use
5362 random numbers. It is used to generate certain symbol names
5363 that have to be different in every compiled file. It is also used to
5364 place unique stamps in coverage data files and the object files that
5365 produce them. You can use the @option{-frandom-seed} option to produce
5366 reproducibly identical object files.
5368 The @var{string} should be different for every file you compile.
5370 @item -fsched-verbose=@var{n}
5371 @opindex fsched-verbose
5372 On targets that use instruction scheduling, this option controls the
5373 amount of debugging output the scheduler prints. This information is
5374 written to standard error, unless @option{-fdump-rtl-sched1} or
5375 @option{-fdump-rtl-sched2} is specified, in which case it is output
5376 to the usual dump listing file, @file{.sched} or @file{.sched2}
5377 respectively. However for @var{n} greater than nine, the output is
5378 always printed to standard error.
5380 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
5381 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
5382 For @var{n} greater than one, it also output basic block probabilities,
5383 detailed ready list information and unit/insn info. For @var{n} greater
5384 than two, it includes RTL at abort point, control-flow and regions info.
5385 And for @var{n} over four, @option{-fsched-verbose} also includes
5389 @itemx -save-temps=cwd
5391 Store the usual ``temporary'' intermediate files permanently; place them
5392 in the current directory and name them based on the source file. Thus,
5393 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
5394 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
5395 preprocessed @file{foo.i} output file even though the compiler now
5396 normally uses an integrated preprocessor.
5398 When used in combination with the @option{-x} command line option,
5399 @option{-save-temps} is sensible enough to avoid over writing an
5400 input source file with the same extension as an intermediate file.
5401 The corresponding intermediate file may be obtained by renaming the
5402 source file before using @option{-save-temps}.
5404 If you invoke GCC in parallel, compiling several different source
5405 files that share a common base name in different subdirectories or the
5406 same source file compiled for multiple output destinations, it is
5407 likely that the different parallel compilers will interfere with each
5408 other, and overwrite the temporary files. For instance:
5411 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
5412 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
5415 may result in @file{foo.i} and @file{foo.o} being written to
5416 simultaneously by both compilers.
5418 @item -save-temps=obj
5419 @opindex save-temps=obj
5420 Store the usual ``temporary'' intermediate files permanently. If the
5421 @option{-o} option is used, the temporary files are based on the
5422 object file. If the @option{-o} option is not used, the
5423 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
5428 gcc -save-temps=obj -c foo.c
5429 gcc -save-temps=obj -c bar.c -o dir/xbar.o
5430 gcc -save-temps=obj foobar.c -o dir2/yfoobar
5433 would create @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
5434 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
5435 @file{dir2/yfoobar.o}.
5437 @item -time@r{[}=@var{file}@r{]}
5439 Report the CPU time taken by each subprocess in the compilation
5440 sequence. For C source files, this is the compiler proper and assembler
5441 (plus the linker if linking is done).
5443 Without the specification of an output file, the output looks like this:
5450 The first number on each line is the ``user time'', that is time spent
5451 executing the program itself. The second number is ``system time'',
5452 time spent executing operating system routines on behalf of the program.
5453 Both numbers are in seconds.
5455 With the specification of an output file, the output is appended to the
5456 named file, and it looks like this:
5459 0.12 0.01 cc1 @var{options}
5460 0.00 0.01 as @var{options}
5463 The ``user time'' and the ``system time'' are moved before the program
5464 name, and the options passed to the program are displayed, so that one
5465 can later tell what file was being compiled, and with which options.
5467 @item -fvar-tracking
5468 @opindex fvar-tracking
5469 Run variable tracking pass. It computes where variables are stored at each
5470 position in code. Better debugging information is then generated
5471 (if the debugging information format supports this information).
5473 It is enabled by default when compiling with optimization (@option{-Os},
5474 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
5475 the debug info format supports it.
5477 @item -fvar-tracking-assignments
5478 @opindex fvar-tracking-assignments
5479 @opindex fno-var-tracking-assignments
5480 Annotate assignments to user variables early in the compilation and
5481 attempt to carry the annotations over throughout the compilation all the
5482 way to the end, in an attempt to improve debug information while
5483 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
5485 It can be enabled even if var-tracking is disabled, in which case
5486 annotations will be created and maintained, but discarded at the end.
5488 @item -fvar-tracking-assignments-toggle
5489 @opindex fvar-tracking-assignments-toggle
5490 @opindex fno-var-tracking-assignments-toggle
5491 Toggle @option{-fvar-tracking-assignments}, in the same way that
5492 @option{-gtoggle} toggles @option{-g}.
5494 @item -print-file-name=@var{library}
5495 @opindex print-file-name
5496 Print the full absolute name of the library file @var{library} that
5497 would be used when linking---and don't do anything else. With this
5498 option, GCC does not compile or link anything; it just prints the
5501 @item -print-multi-directory
5502 @opindex print-multi-directory
5503 Print the directory name corresponding to the multilib selected by any
5504 other switches present in the command line. This directory is supposed
5505 to exist in @env{GCC_EXEC_PREFIX}.
5507 @item -print-multi-lib
5508 @opindex print-multi-lib
5509 Print the mapping from multilib directory names to compiler switches
5510 that enable them. The directory name is separated from the switches by
5511 @samp{;}, and each switch starts with an @samp{@@} instead of the
5512 @samp{-}, without spaces between multiple switches. This is supposed to
5513 ease shell-processing.
5515 @item -print-prog-name=@var{program}
5516 @opindex print-prog-name
5517 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
5519 @item -print-libgcc-file-name
5520 @opindex print-libgcc-file-name
5521 Same as @option{-print-file-name=libgcc.a}.
5523 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
5524 but you do want to link with @file{libgcc.a}. You can do
5527 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
5530 @item -print-search-dirs
5531 @opindex print-search-dirs
5532 Print the name of the configured installation directory and a list of
5533 program and library directories @command{gcc} will search---and don't do anything else.
5535 This is useful when @command{gcc} prints the error message
5536 @samp{installation problem, cannot exec cpp0: No such file or directory}.
5537 To resolve this you either need to put @file{cpp0} and the other compiler
5538 components where @command{gcc} expects to find them, or you can set the environment
5539 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
5540 Don't forget the trailing @samp{/}.
5541 @xref{Environment Variables}.
5543 @item -print-sysroot
5544 @opindex print-sysroot
5545 Print the target sysroot directory that will be used during
5546 compilation. This is the target sysroot specified either at configure
5547 time or using the @option{--sysroot} option, possibly with an extra
5548 suffix that depends on compilation options. If no target sysroot is
5549 specified, the option prints nothing.
5551 @item -print-sysroot-headers-suffix
5552 @opindex print-sysroot-headers-suffix
5553 Print the suffix added to the target sysroot when searching for
5554 headers, or give an error if the compiler is not configured with such
5555 a suffix---and don't do anything else.
5558 @opindex dumpmachine
5559 Print the compiler's target machine (for example,
5560 @samp{i686-pc-linux-gnu})---and don't do anything else.
5563 @opindex dumpversion
5564 Print the compiler version (for example, @samp{3.0})---and don't do
5569 Print the compiler's built-in specs---and don't do anything else. (This
5570 is used when GCC itself is being built.) @xref{Spec Files}.
5572 @item -feliminate-unused-debug-types
5573 @opindex feliminate-unused-debug-types
5574 Normally, when producing DWARF2 output, GCC will emit debugging
5575 information for all types declared in a compilation
5576 unit, regardless of whether or not they are actually used
5577 in that compilation unit. Sometimes this is useful, such as
5578 if, in the debugger, you want to cast a value to a type that is
5579 not actually used in your program (but is declared). More often,
5580 however, this results in a significant amount of wasted space.
5581 With this option, GCC will avoid producing debug symbol output
5582 for types that are nowhere used in the source file being compiled.
5585 @node Optimize Options
5586 @section Options That Control Optimization
5587 @cindex optimize options
5588 @cindex options, optimization
5590 These options control various sorts of optimizations.
5592 Without any optimization option, the compiler's goal is to reduce the
5593 cost of compilation and to make debugging produce the expected
5594 results. Statements are independent: if you stop the program with a
5595 breakpoint between statements, you can then assign a new value to any
5596 variable or change the program counter to any other statement in the
5597 function and get exactly the results you would expect from the source
5600 Turning on optimization flags makes the compiler attempt to improve
5601 the performance and/or code size at the expense of compilation time
5602 and possibly the ability to debug the program.
5604 The compiler performs optimization based on the knowledge it has of the
5605 program. Compiling multiple files at once to a single output file mode allows
5606 the compiler to use information gained from all of the files when compiling
5609 Not all optimizations are controlled directly by a flag. Only
5610 optimizations that have a flag are listed in this section.
5612 Depending on the target and how GCC was configured, a slightly different
5613 set of optimizations may be enabled at each @option{-O} level than
5614 those listed here. You can invoke GCC with @samp{-Q --help=optimizers}
5615 to find out the exact set of optimizations that are enabled at each level.
5616 @xref{Overall Options}, for examples.
5623 Optimize. Optimizing compilation takes somewhat more time, and a lot
5624 more memory for a large function.
5626 With @option{-O}, the compiler tries to reduce code size and execution
5627 time, without performing any optimizations that take a great deal of
5630 @option{-O} turns on the following optimization flags:
5633 -fcprop-registers @gol
5636 -fdelayed-branch @gol
5638 -fguess-branch-probability @gol
5639 -fif-conversion2 @gol
5640 -fif-conversion @gol
5641 -fipa-pure-const @gol
5642 -fipa-reference @gol
5644 -fsplit-wide-types @gol
5645 -ftree-builtin-call-dce @gol
5648 -ftree-copyrename @gol
5650 -ftree-dominator-opts @gol
5652 -ftree-forwprop @gol
5660 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5661 where doing so does not interfere with debugging.
5665 Optimize even more. GCC performs nearly all supported optimizations
5666 that do not involve a space-speed tradeoff.
5667 As compared to @option{-O}, this option increases both compilation time
5668 and the performance of the generated code.
5670 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5671 also turns on the following optimization flags:
5672 @gccoptlist{-fthread-jumps @gol
5673 -falign-functions -falign-jumps @gol
5674 -falign-loops -falign-labels @gol
5677 -fcse-follow-jumps -fcse-skip-blocks @gol
5678 -fdelete-null-pointer-checks @gol
5679 -fexpensive-optimizations @gol
5680 -fgcse -fgcse-lm @gol
5681 -finline-small-functions @gol
5682 -findirect-inlining @gol
5684 -foptimize-sibling-calls @gol
5687 -freorder-blocks -freorder-functions @gol
5688 -frerun-cse-after-loop @gol
5689 -fsched-interblock -fsched-spec @gol
5690 -fschedule-insns -fschedule-insns2 @gol
5691 -fstrict-aliasing -fstrict-overflow @gol
5692 -ftree-switch-conversion @gol
5696 Please note the warning under @option{-fgcse} about
5697 invoking @option{-O2} on programs that use computed gotos.
5701 Optimize yet more. @option{-O3} turns on all optimizations specified
5702 by @option{-O2} and also turns on the @option{-finline-functions},
5703 @option{-funswitch-loops}, @option{-fpredictive-commoning},
5704 @option{-fgcse-after-reload} and @option{-ftree-vectorize} options.
5708 Reduce compilation time and make debugging produce the expected
5709 results. This is the default.
5713 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5714 do not typically increase code size. It also performs further
5715 optimizations designed to reduce code size.
5717 @option{-Os} disables the following optimization flags:
5718 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5719 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5720 -fprefetch-loop-arrays -ftree-vect-loop-version}
5722 If you use multiple @option{-O} options, with or without level numbers,
5723 the last such option is the one that is effective.
5726 Options of the form @option{-f@var{flag}} specify machine-independent
5727 flags. Most flags have both positive and negative forms; the negative
5728 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5729 below, only one of the forms is listed---the one you typically will
5730 use. You can figure out the other form by either removing @samp{no-}
5733 The following options control specific optimizations. They are either
5734 activated by @option{-O} options or are related to ones that are. You
5735 can use the following flags in the rare cases when ``fine-tuning'' of
5736 optimizations to be performed is desired.
5739 @item -fno-default-inline
5740 @opindex fno-default-inline
5741 Do not make member functions inline by default merely because they are
5742 defined inside the class scope (C++ only). Otherwise, when you specify
5743 @w{@option{-O}}, member functions defined inside class scope are compiled
5744 inline by default; i.e., you don't need to add @samp{inline} in front of
5745 the member function name.
5747 @item -fno-defer-pop
5748 @opindex fno-defer-pop
5749 Always pop the arguments to each function call as soon as that function
5750 returns. For machines which must pop arguments after a function call,
5751 the compiler normally lets arguments accumulate on the stack for several
5752 function calls and pops them all at once.
5754 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5756 @item -fforward-propagate
5757 @opindex fforward-propagate
5758 Perform a forward propagation pass on RTL@. The pass tries to combine two
5759 instructions and checks if the result can be simplified. If loop unrolling
5760 is active, two passes are performed and the second is scheduled after
5763 This option is enabled by default at optimization levels @option{-O},
5764 @option{-O2}, @option{-O3}, @option{-Os}.
5766 @item -fomit-frame-pointer
5767 @opindex fomit-frame-pointer
5768 Don't keep the frame pointer in a register for functions that
5769 don't need one. This avoids the instructions to save, set up and
5770 restore frame pointers; it also makes an extra register available
5771 in many functions. @strong{It also makes debugging impossible on
5774 On some machines, such as the VAX, this flag has no effect, because
5775 the standard calling sequence automatically handles the frame pointer
5776 and nothing is saved by pretending it doesn't exist. The
5777 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5778 whether a target machine supports this flag. @xref{Registers,,Register
5779 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5781 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5783 @item -foptimize-sibling-calls
5784 @opindex foptimize-sibling-calls
5785 Optimize sibling and tail recursive calls.
5787 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5791 Don't pay attention to the @code{inline} keyword. Normally this option
5792 is used to keep the compiler from expanding any functions inline.
5793 Note that if you are not optimizing, no functions can be expanded inline.
5795 @item -finline-small-functions
5796 @opindex finline-small-functions
5797 Integrate functions into their callers when their body is smaller than expected
5798 function call code (so overall size of program gets smaller). The compiler
5799 heuristically decides which functions are simple enough to be worth integrating
5802 Enabled at level @option{-O2}.
5804 @item -findirect-inlining
5805 @opindex findirect-inlining
5806 Inline also indirect calls that are discovered to be known at compile
5807 time thanks to previous inlining. This option has any effect only
5808 when inlining itself is turned on by the @option{-finline-functions}
5809 or @option{-finline-small-functions} options.
5811 Enabled at level @option{-O2}.
5813 @item -finline-functions
5814 @opindex finline-functions
5815 Integrate all simple functions into their callers. The compiler
5816 heuristically decides which functions are simple enough to be worth
5817 integrating in this way.
5819 If all calls to a given function are integrated, and the function is
5820 declared @code{static}, then the function is normally not output as
5821 assembler code in its own right.
5823 Enabled at level @option{-O3}.
5825 @item -finline-functions-called-once
5826 @opindex finline-functions-called-once
5827 Consider all @code{static} functions called once for inlining into their
5828 caller even if they are not marked @code{inline}. If a call to a given
5829 function is integrated, then the function is not output as assembler code
5832 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os}.
5834 @item -fearly-inlining
5835 @opindex fearly-inlining
5836 Inline functions marked by @code{always_inline} and functions whose body seems
5837 smaller than the function call overhead early before doing
5838 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5839 makes profiling significantly cheaper and usually inlining faster on programs
5840 having large chains of nested wrapper functions.
5846 Perform interprocedural scalar replacement of aggregates, removal of
5847 unused parameters and replacement of parameters passed by reference
5848 by parameters passed by value.
5850 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
5852 @item -finline-limit=@var{n}
5853 @opindex finline-limit
5854 By default, GCC limits the size of functions that can be inlined. This flag
5855 allows coarse control of this limit. @var{n} is the size of functions that
5856 can be inlined in number of pseudo instructions.
5858 Inlining is actually controlled by a number of parameters, which may be
5859 specified individually by using @option{--param @var{name}=@var{value}}.
5860 The @option{-finline-limit=@var{n}} option sets some of these parameters
5864 @item max-inline-insns-single
5865 is set to @var{n}/2.
5866 @item max-inline-insns-auto
5867 is set to @var{n}/2.
5870 See below for a documentation of the individual
5871 parameters controlling inlining and for the defaults of these parameters.
5873 @emph{Note:} there may be no value to @option{-finline-limit} that results
5874 in default behavior.
5876 @emph{Note:} pseudo instruction represents, in this particular context, an
5877 abstract measurement of function's size. In no way does it represent a count
5878 of assembly instructions and as such its exact meaning might change from one
5879 release to an another.
5881 @item -fkeep-inline-functions
5882 @opindex fkeep-inline-functions
5883 In C, emit @code{static} functions that are declared @code{inline}
5884 into the object file, even if the function has been inlined into all
5885 of its callers. This switch does not affect functions using the
5886 @code{extern inline} extension in GNU C89@. In C++, emit any and all
5887 inline functions into the object file.
5889 @item -fkeep-static-consts
5890 @opindex fkeep-static-consts
5891 Emit variables declared @code{static const} when optimization isn't turned
5892 on, even if the variables aren't referenced.
5894 GCC enables this option by default. If you want to force the compiler to
5895 check if the variable was referenced, regardless of whether or not
5896 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5898 @item -fmerge-constants
5899 @opindex fmerge-constants
5900 Attempt to merge identical constants (string constants and floating point
5901 constants) across compilation units.
5903 This option is the default for optimized compilation if the assembler and
5904 linker support it. Use @option{-fno-merge-constants} to inhibit this
5907 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5909 @item -fmerge-all-constants
5910 @opindex fmerge-all-constants
5911 Attempt to merge identical constants and identical variables.
5913 This option implies @option{-fmerge-constants}. In addition to
5914 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5915 arrays or initialized constant variables with integral or floating point
5916 types. Languages like C or C++ require each variable, including multiple
5917 instances of the same variable in recursive calls, to have distinct locations,
5918 so using this option will result in non-conforming
5921 @item -fmodulo-sched
5922 @opindex fmodulo-sched
5923 Perform swing modulo scheduling immediately before the first scheduling
5924 pass. This pass looks at innermost loops and reorders their
5925 instructions by overlapping different iterations.
5927 @item -fmodulo-sched-allow-regmoves
5928 @opindex fmodulo-sched-allow-regmoves
5929 Perform more aggressive SMS based modulo scheduling with register moves
5930 allowed. By setting this flag certain anti-dependences edges will be
5931 deleted which will trigger the generation of reg-moves based on the
5932 life-range analysis. This option is effective only with
5933 @option{-fmodulo-sched} enabled.
5935 @item -fno-branch-count-reg
5936 @opindex fno-branch-count-reg
5937 Do not use ``decrement and branch'' instructions on a count register,
5938 but instead generate a sequence of instructions that decrement a
5939 register, compare it against zero, then branch based upon the result.
5940 This option is only meaningful on architectures that support such
5941 instructions, which include x86, PowerPC, IA-64 and S/390.
5943 The default is @option{-fbranch-count-reg}.
5945 @item -fno-function-cse
5946 @opindex fno-function-cse
5947 Do not put function addresses in registers; make each instruction that
5948 calls a constant function contain the function's address explicitly.
5950 This option results in less efficient code, but some strange hacks
5951 that alter the assembler output may be confused by the optimizations
5952 performed when this option is not used.
5954 The default is @option{-ffunction-cse}
5956 @item -fno-zero-initialized-in-bss
5957 @opindex fno-zero-initialized-in-bss
5958 If the target supports a BSS section, GCC by default puts variables that
5959 are initialized to zero into BSS@. This can save space in the resulting
5962 This option turns off this behavior because some programs explicitly
5963 rely on variables going to the data section. E.g., so that the
5964 resulting executable can find the beginning of that section and/or make
5965 assumptions based on that.
5967 The default is @option{-fzero-initialized-in-bss}.
5969 @item -fmudflap -fmudflapth -fmudflapir
5973 @cindex bounds checking
5975 For front-ends that support it (C and C++), instrument all risky
5976 pointer/array dereferencing operations, some standard library
5977 string/heap functions, and some other associated constructs with
5978 range/validity tests. Modules so instrumented should be immune to
5979 buffer overflows, invalid heap use, and some other classes of C/C++
5980 programming errors. The instrumentation relies on a separate runtime
5981 library (@file{libmudflap}), which will be linked into a program if
5982 @option{-fmudflap} is given at link time. Run-time behavior of the
5983 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
5984 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
5987 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
5988 link if your program is multi-threaded. Use @option{-fmudflapir}, in
5989 addition to @option{-fmudflap} or @option{-fmudflapth}, if
5990 instrumentation should ignore pointer reads. This produces less
5991 instrumentation (and therefore faster execution) and still provides
5992 some protection against outright memory corrupting writes, but allows
5993 erroneously read data to propagate within a program.
5995 @item -fthread-jumps
5996 @opindex fthread-jumps
5997 Perform optimizations where we check to see if a jump branches to a
5998 location where another comparison subsumed by the first is found. If
5999 so, the first branch is redirected to either the destination of the
6000 second branch or a point immediately following it, depending on whether
6001 the condition is known to be true or false.
6003 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6005 @item -fsplit-wide-types
6006 @opindex fsplit-wide-types
6007 When using a type that occupies multiple registers, such as @code{long
6008 long} on a 32-bit system, split the registers apart and allocate them
6009 independently. This normally generates better code for those types,
6010 but may make debugging more difficult.
6012 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
6015 @item -fcse-follow-jumps
6016 @opindex fcse-follow-jumps
6017 In common subexpression elimination (CSE), scan through jump instructions
6018 when the target of the jump is not reached by any other path. For
6019 example, when CSE encounters an @code{if} statement with an
6020 @code{else} clause, CSE will follow the jump when the condition
6023 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6025 @item -fcse-skip-blocks
6026 @opindex fcse-skip-blocks
6027 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
6028 follow jumps which conditionally skip over blocks. When CSE
6029 encounters a simple @code{if} statement with no else clause,
6030 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
6031 body of the @code{if}.
6033 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6035 @item -frerun-cse-after-loop
6036 @opindex frerun-cse-after-loop
6037 Re-run common subexpression elimination after loop optimizations has been
6040 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6044 Perform a global common subexpression elimination pass.
6045 This pass also performs global constant and copy propagation.
6047 @emph{Note:} When compiling a program using computed gotos, a GCC
6048 extension, you may get better runtime performance if you disable
6049 the global common subexpression elimination pass by adding
6050 @option{-fno-gcse} to the command line.
6052 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6056 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
6057 attempt to move loads which are only killed by stores into themselves. This
6058 allows a loop containing a load/store sequence to be changed to a load outside
6059 the loop, and a copy/store within the loop.
6061 Enabled by default when gcse is enabled.
6065 When @option{-fgcse-sm} is enabled, a store motion pass is run after
6066 global common subexpression elimination. This pass will attempt to move
6067 stores out of loops. When used in conjunction with @option{-fgcse-lm},
6068 loops containing a load/store sequence can be changed to a load before
6069 the loop and a store after the loop.
6071 Not enabled at any optimization level.
6075 When @option{-fgcse-las} is enabled, the global common subexpression
6076 elimination pass eliminates redundant loads that come after stores to the
6077 same memory location (both partial and full redundancies).
6079 Not enabled at any optimization level.
6081 @item -fgcse-after-reload
6082 @opindex fgcse-after-reload
6083 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
6084 pass is performed after reload. The purpose of this pass is to cleanup
6087 @item -funsafe-loop-optimizations
6088 @opindex funsafe-loop-optimizations
6089 If given, the loop optimizer will assume that loop indices do not
6090 overflow, and that the loops with nontrivial exit condition are not
6091 infinite. This enables a wider range of loop optimizations even if
6092 the loop optimizer itself cannot prove that these assumptions are valid.
6093 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
6094 if it finds this kind of loop.
6096 @item -fcrossjumping
6097 @opindex fcrossjumping
6098 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
6099 resulting code may or may not perform better than without cross-jumping.
6101 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6103 @item -fauto-inc-dec
6104 @opindex fauto-inc-dec
6105 Combine increments or decrements of addresses with memory accesses.
6106 This pass is always skipped on architectures that do not have
6107 instructions to support this. Enabled by default at @option{-O} and
6108 higher on architectures that support this.
6112 Perform dead code elimination (DCE) on RTL@.
6113 Enabled by default at @option{-O} and higher.
6117 Perform dead store elimination (DSE) on RTL@.
6118 Enabled by default at @option{-O} and higher.
6120 @item -fif-conversion
6121 @opindex fif-conversion
6122 Attempt to transform conditional jumps into branch-less equivalents. This
6123 include use of conditional moves, min, max, set flags and abs instructions, and
6124 some tricks doable by standard arithmetics. The use of conditional execution
6125 on chips where it is available is controlled by @code{if-conversion2}.
6127 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6129 @item -fif-conversion2
6130 @opindex fif-conversion2
6131 Use conditional execution (where available) to transform conditional jumps into
6132 branch-less equivalents.
6134 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6136 @item -fdelete-null-pointer-checks
6137 @opindex fdelete-null-pointer-checks
6138 Assume that programs cannot safely dereference null pointers, and that
6139 no code or data element resides there. This enables simple constant
6140 folding optimizations at all optimization levels. In addition, other
6141 optimization passes in GCC use this flag to control global dataflow
6142 analyses that eliminate useless checks for null pointers; these assume
6143 that if a pointer is checked after it has already been dereferenced,
6146 Note however that in some environments this assumption is not true.
6147 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
6148 for programs which depend on that behavior.
6150 Some targets, especially embedded ones, disable this option at all levels.
6151 Otherwise it is enabled at all levels: @option{-O0}, @option{-O1},
6152 @option{-O2}, @option{-O3}, @option{-Os}. Passes that use the information
6153 are enabled independently at different optimization levels.
6155 @item -fexpensive-optimizations
6156 @opindex fexpensive-optimizations
6157 Perform a number of minor optimizations that are relatively expensive.
6159 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6161 @item -foptimize-register-move
6163 @opindex foptimize-register-move
6165 Attempt to reassign register numbers in move instructions and as
6166 operands of other simple instructions in order to maximize the amount of
6167 register tying. This is especially helpful on machines with two-operand
6170 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
6173 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6175 @item -fira-algorithm=@var{algorithm}
6176 Use specified coloring algorithm for the integrated register
6177 allocator. The @var{algorithm} argument should be @code{priority} or
6178 @code{CB}. The first algorithm specifies Chow's priority coloring,
6179 the second one specifies Chaitin-Briggs coloring. The second
6180 algorithm can be unimplemented for some architectures. If it is
6181 implemented, it is the default because Chaitin-Briggs coloring as a
6182 rule generates a better code.
6184 @item -fira-region=@var{region}
6185 Use specified regions for the integrated register allocator. The
6186 @var{region} argument should be one of @code{all}, @code{mixed}, or
6187 @code{one}. The first value means using all loops as register
6188 allocation regions, the second value which is the default means using
6189 all loops except for loops with small register pressure as the
6190 regions, and third one means using all function as a single region.
6191 The first value can give best result for machines with small size and
6192 irregular register set, the third one results in faster and generates
6193 decent code and the smallest size code, and the default value usually
6194 give the best results in most cases and for most architectures.
6196 @item -fira-coalesce
6197 @opindex fira-coalesce
6198 Do optimistic register coalescing. This option might be profitable for
6199 architectures with big regular register files.
6201 @item -fno-ira-share-save-slots
6202 @opindex fno-ira-share-save-slots
6203 Switch off sharing stack slots used for saving call used hard
6204 registers living through a call. Each hard register will get a
6205 separate stack slot and as a result function stack frame will be
6208 @item -fno-ira-share-spill-slots
6209 @opindex fno-ira-share-spill-slots
6210 Switch off sharing stack slots allocated for pseudo-registers. Each
6211 pseudo-register which did not get a hard register will get a separate
6212 stack slot and as a result function stack frame will be bigger.
6214 @item -fira-verbose=@var{n}
6215 @opindex fira-verbose
6216 Set up how verbose dump file for the integrated register allocator
6217 will be. Default value is 5. If the value is greater or equal to 10,
6218 the dump file will be stderr as if the value were @var{n} minus 10.
6220 @item -fdelayed-branch
6221 @opindex fdelayed-branch
6222 If supported for the target machine, attempt to reorder instructions
6223 to exploit instruction slots available after delayed branch
6226 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6228 @item -fschedule-insns
6229 @opindex fschedule-insns
6230 If supported for the target machine, attempt to reorder instructions to
6231 eliminate execution stalls due to required data being unavailable. This
6232 helps machines that have slow floating point or memory load instructions
6233 by allowing other instructions to be issued until the result of the load
6234 or floating point instruction is required.
6236 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6238 @item -fschedule-insns2
6239 @opindex fschedule-insns2
6240 Similar to @option{-fschedule-insns}, but requests an additional pass of
6241 instruction scheduling after register allocation has been done. This is
6242 especially useful on machines with a relatively small number of
6243 registers and where memory load instructions take more than one cycle.
6245 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6247 @item -fno-sched-interblock
6248 @opindex fno-sched-interblock
6249 Don't schedule instructions across basic blocks. This is normally
6250 enabled by default when scheduling before register allocation, i.e.@:
6251 with @option{-fschedule-insns} or at @option{-O2} or higher.
6253 @item -fno-sched-spec
6254 @opindex fno-sched-spec
6255 Don't allow speculative motion of non-load instructions. This is normally
6256 enabled by default when scheduling before register allocation, i.e.@:
6257 with @option{-fschedule-insns} or at @option{-O2} or higher.
6259 @item -fsched-pressure
6260 @opindex fsched-pressure
6261 Enable register pressure sensitive insn scheduling before the register
6262 allocation. This only makes sense when scheduling before register
6263 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
6264 @option{-O2} or higher. Usage of this option can improve the
6265 generated code and decrease its size by preventing register pressure
6266 increase above the number of available hard registers and as a
6267 consequence register spills in the register allocation.
6269 @item -fsched-spec-load
6270 @opindex fsched-spec-load
6271 Allow speculative motion of some load instructions. This only makes
6272 sense when scheduling before register allocation, i.e.@: with
6273 @option{-fschedule-insns} or at @option{-O2} or higher.
6275 @item -fsched-spec-load-dangerous
6276 @opindex fsched-spec-load-dangerous
6277 Allow speculative motion of more load instructions. This only makes
6278 sense when scheduling before register allocation, i.e.@: with
6279 @option{-fschedule-insns} or at @option{-O2} or higher.
6281 @item -fsched-stalled-insns
6282 @itemx -fsched-stalled-insns=@var{n}
6283 @opindex fsched-stalled-insns
6284 Define how many insns (if any) can be moved prematurely from the queue
6285 of stalled insns into the ready list, during the second scheduling pass.
6286 @option{-fno-sched-stalled-insns} means that no insns will be moved
6287 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
6288 on how many queued insns can be moved prematurely.
6289 @option{-fsched-stalled-insns} without a value is equivalent to
6290 @option{-fsched-stalled-insns=1}.
6292 @item -fsched-stalled-insns-dep
6293 @itemx -fsched-stalled-insns-dep=@var{n}
6294 @opindex fsched-stalled-insns-dep
6295 Define how many insn groups (cycles) will be examined for a dependency
6296 on a stalled insn that is candidate for premature removal from the queue
6297 of stalled insns. This has an effect only during the second scheduling pass,
6298 and only if @option{-fsched-stalled-insns} is used.
6299 @option{-fno-sched-stalled-insns-dep} is equivalent to
6300 @option{-fsched-stalled-insns-dep=0}.
6301 @option{-fsched-stalled-insns-dep} without a value is equivalent to
6302 @option{-fsched-stalled-insns-dep=1}.
6304 @item -fsched2-use-superblocks
6305 @opindex fsched2-use-superblocks
6306 When scheduling after register allocation, do use superblock scheduling
6307 algorithm. Superblock scheduling allows motion across basic block boundaries
6308 resulting on faster schedules. This option is experimental, as not all machine
6309 descriptions used by GCC model the CPU closely enough to avoid unreliable
6310 results from the algorithm.
6312 This only makes sense when scheduling after register allocation, i.e.@: with
6313 @option{-fschedule-insns2} or at @option{-O2} or higher.
6315 @item -fsched-group-heuristic
6316 @opindex fsched-group-heuristic
6317 Enable the group heuristic in the scheduler. This heuristic favors
6318 the instruction that belongs to a schedule group. This is enabled
6319 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6320 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6322 @item -fsched-critical-path-heuristic
6323 @opindex fsched-critical-path-heuristic
6324 Enable the critical-path heuristic in the scheduler. This heuristic favors
6325 instructions on the critical path. This is enabled by default when
6326 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6327 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6329 @item -fsched-spec-insn-heuristic
6330 @opindex fsched-spec-insn-heuristic
6331 Enable the speculative instruction heuristic in the scheduler. This
6332 heuristic favors speculative instructions with greater dependency weakness.
6333 This is enabled by default when scheduling is enabled, i.e.@:
6334 with @option{-fschedule-insns} or @option{-fschedule-insns2}
6335 or at @option{-O2} or higher.
6337 @item -fsched-rank-heuristic
6338 @opindex fsched-rank-heuristic
6339 Enable the rank heuristic in the scheduler. This heuristic favors
6340 the instruction belonging to a basic block with greater size or frequency.
6341 This is enabled by default when scheduling is enabled, i.e.@:
6342 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6343 at @option{-O2} or higher.
6345 @item -fsched-last-insn-heuristic
6346 @opindex fsched-last-insn-heuristic
6347 Enable the last-instruction heuristic in the scheduler. This heuristic
6348 favors the instruction that is less dependent on the last instruction
6349 scheduled. This is enabled by default when scheduling is enabled,
6350 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6351 at @option{-O2} or higher.
6353 @item -fsched-dep-count-heuristic
6354 @opindex fsched-dep-count-heuristic
6355 Enable the dependent-count heuristic in the scheduler. This heuristic
6356 favors the instruction that has more instructions depending on it.
6357 This is enabled by default when scheduling is enabled, i.e.@:
6358 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6359 at @option{-O2} or higher.
6361 @item -fsched2-use-traces
6362 @opindex fsched2-use-traces
6363 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
6364 allocation and additionally perform code duplication in order to increase the
6365 size of superblocks using tracer pass. See @option{-ftracer} for details on
6368 This mode should produce faster but significantly longer programs. Also
6369 without @option{-fbranch-probabilities} the traces constructed may not
6370 match the reality and hurt the performance. This only makes
6371 sense when scheduling after register allocation, i.e.@: with
6372 @option{-fschedule-insns2} or at @option{-O2} or higher.
6374 @item -freschedule-modulo-scheduled-loops
6375 @opindex freschedule-modulo-scheduled-loops
6376 The modulo scheduling comes before the traditional scheduling, if a loop
6377 was modulo scheduled we may want to prevent the later scheduling passes
6378 from changing its schedule, we use this option to control that.
6380 @item -fselective-scheduling
6381 @opindex fselective-scheduling
6382 Schedule instructions using selective scheduling algorithm. Selective
6383 scheduling runs instead of the first scheduler pass.
6385 @item -fselective-scheduling2
6386 @opindex fselective-scheduling2
6387 Schedule instructions using selective scheduling algorithm. Selective
6388 scheduling runs instead of the second scheduler pass.
6390 @item -fsel-sched-pipelining
6391 @opindex fsel-sched-pipelining
6392 Enable software pipelining of innermost loops during selective scheduling.
6393 This option has no effect until one of @option{-fselective-scheduling} or
6394 @option{-fselective-scheduling2} is turned on.
6396 @item -fsel-sched-pipelining-outer-loops
6397 @opindex fsel-sched-pipelining-outer-loops
6398 When pipelining loops during selective scheduling, also pipeline outer loops.
6399 This option has no effect until @option{-fsel-sched-pipelining} is turned on.
6401 @item -fcaller-saves
6402 @opindex fcaller-saves
6403 Enable values to be allocated in registers that will be clobbered by
6404 function calls, by emitting extra instructions to save and restore the
6405 registers around such calls. Such allocation is done only when it
6406 seems to result in better code than would otherwise be produced.
6408 This option is always enabled by default on certain machines, usually
6409 those which have no call-preserved registers to use instead.
6411 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6413 @item -fconserve-stack
6414 @opindex fconserve-stack
6415 Attempt to minimize stack usage. The compiler will attempt to use less
6416 stack space, even if that makes the program slower. This option
6417 implies setting the @option{large-stack-frame} parameter to 100
6418 and the @option{large-stack-frame-growth} parameter to 400.
6420 @item -ftree-reassoc
6421 @opindex ftree-reassoc
6422 Perform reassociation on trees. This flag is enabled by default
6423 at @option{-O} and higher.
6427 Perform partial redundancy elimination (PRE) on trees. This flag is
6428 enabled by default at @option{-O2} and @option{-O3}.
6430 @item -ftree-forwprop
6431 @opindex ftree-forwprop
6432 Perform forward propagation on trees. This flag is enabled by default
6433 at @option{-O} and higher.
6437 Perform full redundancy elimination (FRE) on trees. The difference
6438 between FRE and PRE is that FRE only considers expressions
6439 that are computed on all paths leading to the redundant computation.
6440 This analysis is faster than PRE, though it exposes fewer redundancies.
6441 This flag is enabled by default at @option{-O} and higher.
6443 @item -ftree-phiprop
6444 @opindex ftree-phiprop
6445 Perform hoisting of loads from conditional pointers on trees. This
6446 pass is enabled by default at @option{-O} and higher.
6448 @item -ftree-copy-prop
6449 @opindex ftree-copy-prop
6450 Perform copy propagation on trees. This pass eliminates unnecessary
6451 copy operations. This flag is enabled by default at @option{-O} and
6454 @item -fipa-pure-const
6455 @opindex fipa-pure-const
6456 Discover which functions are pure or constant.
6457 Enabled by default at @option{-O} and higher.
6459 @item -fipa-reference
6460 @opindex fipa-reference
6461 Discover which static variables do not escape cannot escape the
6463 Enabled by default at @option{-O} and higher.
6465 @item -fipa-struct-reorg
6466 @opindex fipa-struct-reorg
6467 Perform structure reorganization optimization, that change C-like structures
6468 layout in order to better utilize spatial locality. This transformation is
6469 affective for programs containing arrays of structures. Available in two
6470 compilation modes: profile-based (enabled with @option{-fprofile-generate})
6471 or static (which uses built-in heuristics). Require @option{-fipa-type-escape}
6472 to provide the safety of this transformation. It works only in whole program
6473 mode, so it requires @option{-fwhole-program} and @option{-combine} to be
6474 enabled. Structures considered @samp{cold} by this transformation are not
6475 affected (see @option{--param struct-reorg-cold-struct-ratio=@var{value}}).
6477 With this flag, the program debug info reflects a new structure layout.
6481 Perform interprocedural pointer analysis. This option is experimental
6482 and does not affect generated code.
6486 Perform interprocedural constant propagation.
6487 This optimization analyzes the program to determine when values passed
6488 to functions are constants and then optimizes accordingly.
6489 This optimization can substantially increase performance
6490 if the application has constants passed to functions.
6491 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
6493 @item -fipa-cp-clone
6494 @opindex fipa-cp-clone
6495 Perform function cloning to make interprocedural constant propagation stronger.
6496 When enabled, interprocedural constant propagation will perform function cloning
6497 when externally visible function can be called with constant arguments.
6498 Because this optimization can create multiple copies of functions,
6499 it may significantly increase code size
6500 (see @option{--param ipcp-unit-growth=@var{value}}).
6501 This flag is enabled by default at @option{-O3}.
6503 @item -fipa-matrix-reorg
6504 @opindex fipa-matrix-reorg
6505 Perform matrix flattening and transposing.
6506 Matrix flattening tries to replace an @math{m}-dimensional matrix
6507 with its equivalent @math{n}-dimensional matrix, where @math{n < m}.
6508 This reduces the level of indirection needed for accessing the elements
6509 of the matrix. The second optimization is matrix transposing that
6510 attempts to change the order of the matrix's dimensions in order to
6511 improve cache locality.
6512 Both optimizations need the @option{-fwhole-program} flag.
6513 Transposing is enabled only if profiling information is available.
6517 Perform forward store motion on trees. This flag is
6518 enabled by default at @option{-O} and higher.
6522 Perform sparse conditional constant propagation (CCP) on trees. This
6523 pass only operates on local scalar variables and is enabled by default
6524 at @option{-O} and higher.
6526 @item -ftree-switch-conversion
6527 Perform conversion of simple initializations in a switch to
6528 initializations from a scalar array. This flag is enabled by default
6529 at @option{-O2} and higher.
6533 Perform dead code elimination (DCE) on trees. This flag is enabled by
6534 default at @option{-O} and higher.
6536 @item -ftree-builtin-call-dce
6537 @opindex ftree-builtin-call-dce
6538 Perform conditional dead code elimination (DCE) for calls to builtin functions
6539 that may set @code{errno} but are otherwise side-effect free. This flag is
6540 enabled by default at @option{-O2} and higher if @option{-Os} is not also
6543 @item -ftree-dominator-opts
6544 @opindex ftree-dominator-opts
6545 Perform a variety of simple scalar cleanups (constant/copy
6546 propagation, redundancy elimination, range propagation and expression
6547 simplification) based on a dominator tree traversal. This also
6548 performs jump threading (to reduce jumps to jumps). This flag is
6549 enabled by default at @option{-O} and higher.
6553 Perform dead store elimination (DSE) on trees. A dead store is a store into
6554 a memory location which will later be overwritten by another store without
6555 any intervening loads. In this case the earlier store can be deleted. This
6556 flag is enabled by default at @option{-O} and higher.
6560 Perform loop header copying on trees. This is beneficial since it increases
6561 effectiveness of code motion optimizations. It also saves one jump. This flag
6562 is enabled by default at @option{-O} and higher. It is not enabled
6563 for @option{-Os}, since it usually increases code size.
6565 @item -ftree-loop-optimize
6566 @opindex ftree-loop-optimize
6567 Perform loop optimizations on trees. This flag is enabled by default
6568 at @option{-O} and higher.
6570 @item -ftree-loop-linear
6571 @opindex ftree-loop-linear
6572 Perform linear loop transformations on tree. This flag can improve cache
6573 performance and allow further loop optimizations to take place.
6575 @item -floop-interchange
6576 Perform loop interchange transformations on loops. Interchanging two
6577 nested loops switches the inner and outer loops. For example, given a
6582 A(J, I) = A(J, I) * C
6586 loop interchange will transform the loop as if the user had written:
6590 A(J, I) = A(J, I) * C
6594 which can be beneficial when @code{N} is larger than the caches,
6595 because in Fortran, the elements of an array are stored in memory
6596 contiguously by column, and the original loop iterates over rows,
6597 potentially creating at each access a cache miss. This optimization
6598 applies to all the languages supported by GCC and is not limited to
6599 Fortran. To use this code transformation, GCC has to be configured
6600 with @option{--with-ppl} and @option{--with-cloog} to enable the
6601 Graphite loop transformation infrastructure.
6603 @item -floop-strip-mine
6604 Perform loop strip mining transformations on loops. Strip mining
6605 splits a loop into two nested loops. The outer loop has strides
6606 equal to the strip size and the inner loop has strides of the
6607 original loop within a strip. For example, given a loop like:
6613 loop strip mining will transform the loop as if the user had written:
6616 DO I = II, min (II + 3, N)
6621 This optimization applies to all the languages supported by GCC and is
6622 not limited to Fortran. To use this code transformation, GCC has to
6623 be configured with @option{--with-ppl} and @option{--with-cloog} to
6624 enable the Graphite loop transformation infrastructure.
6627 Perform loop blocking transformations on loops. Blocking strip mines
6628 each loop in the loop nest such that the memory accesses of the
6629 element loops fit inside caches. For example, given a loop like:
6633 A(J, I) = B(I) + C(J)
6637 loop blocking will transform the loop as if the user had written:
6641 DO I = II, min (II + 63, N)
6642 DO J = JJ, min (JJ + 63, M)
6643 A(J, I) = B(I) + C(J)
6649 which can be beneficial when @code{M} is larger than the caches,
6650 because the innermost loop will iterate over a smaller amount of data
6651 that can be kept in the caches. This optimization applies to all the
6652 languages supported by GCC and is not limited to Fortran. To use this
6653 code transformation, GCC has to be configured with @option{--with-ppl}
6654 and @option{--with-cloog} to enable the Graphite loop transformation
6657 @item -fgraphite-identity
6658 @opindex fgraphite-identity
6659 Enable the identity transformation for graphite. For every SCoP we generate
6660 the polyhedral representation and transform it back to gimple. Using
6661 @option{-fgraphite-identity} we can check the costs or benefits of the
6662 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
6663 are also performed by the code generator CLooG, like index splitting and
6664 dead code elimination in loops.
6666 @item -floop-parallelize-all
6667 Use the Graphite data dependence analysis to identify loops that can
6668 be parallelized. Parallelize all the loops that can be analyzed to
6669 not contain loop carried dependences without checking that it is
6670 profitable to parallelize the loops.
6672 @item -fcheck-data-deps
6673 @opindex fcheck-data-deps
6674 Compare the results of several data dependence analyzers. This option
6675 is used for debugging the data dependence analyzers.
6677 @item -ftree-loop-distribution
6678 Perform loop distribution. This flag can improve cache performance on
6679 big loop bodies and allow further loop optimizations, like
6680 parallelization or vectorization, to take place. For example, the loop
6697 @item -ftree-loop-im
6698 @opindex ftree-loop-im
6699 Perform loop invariant motion on trees. This pass moves only invariants that
6700 would be hard to handle at RTL level (function calls, operations that expand to
6701 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
6702 operands of conditions that are invariant out of the loop, so that we can use
6703 just trivial invariantness analysis in loop unswitching. The pass also includes
6706 @item -ftree-loop-ivcanon
6707 @opindex ftree-loop-ivcanon
6708 Create a canonical counter for number of iterations in the loop for that
6709 determining number of iterations requires complicated analysis. Later
6710 optimizations then may determine the number easily. Useful especially
6711 in connection with unrolling.
6715 Perform induction variable optimizations (strength reduction, induction
6716 variable merging and induction variable elimination) on trees.
6718 @item -ftree-parallelize-loops=n
6719 @opindex ftree-parallelize-loops
6720 Parallelize loops, i.e., split their iteration space to run in n threads.
6721 This is only possible for loops whose iterations are independent
6722 and can be arbitrarily reordered. The optimization is only
6723 profitable on multiprocessor machines, for loops that are CPU-intensive,
6724 rather than constrained e.g.@: by memory bandwidth. This option
6725 implies @option{-pthread}, and thus is only supported on targets
6726 that have support for @option{-pthread}.
6730 Perform function-local points-to analysis on trees. This flag is
6731 enabled by default at @option{-O} and higher.
6735 Perform scalar replacement of aggregates. This pass replaces structure
6736 references with scalars to prevent committing structures to memory too
6737 early. This flag is enabled by default at @option{-O} and higher.
6739 @item -ftree-copyrename
6740 @opindex ftree-copyrename
6741 Perform copy renaming on trees. This pass attempts to rename compiler
6742 temporaries to other variables at copy locations, usually resulting in
6743 variable names which more closely resemble the original variables. This flag
6744 is enabled by default at @option{-O} and higher.
6748 Perform temporary expression replacement during the SSA->normal phase. Single
6749 use/single def temporaries are replaced at their use location with their
6750 defining expression. This results in non-GIMPLE code, but gives the expanders
6751 much more complex trees to work on resulting in better RTL generation. This is
6752 enabled by default at @option{-O} and higher.
6754 @item -ftree-vectorize
6755 @opindex ftree-vectorize
6756 Perform loop vectorization on trees. This flag is enabled by default at
6759 @item -ftree-vect-loop-version
6760 @opindex ftree-vect-loop-version
6761 Perform loop versioning when doing loop vectorization on trees. When a loop
6762 appears to be vectorizable except that data alignment or data dependence cannot
6763 be determined at compile time then vectorized and non-vectorized versions of
6764 the loop are generated along with runtime checks for alignment or dependence
6765 to control which version is executed. This option is enabled by default
6766 except at level @option{-Os} where it is disabled.
6768 @item -fvect-cost-model
6769 @opindex fvect-cost-model
6770 Enable cost model for vectorization.
6774 Perform Value Range Propagation on trees. This is similar to the
6775 constant propagation pass, but instead of values, ranges of values are
6776 propagated. This allows the optimizers to remove unnecessary range
6777 checks like array bound checks and null pointer checks. This is
6778 enabled by default at @option{-O2} and higher. Null pointer check
6779 elimination is only done if @option{-fdelete-null-pointer-checks} is
6784 Perform tail duplication to enlarge superblock size. This transformation
6785 simplifies the control flow of the function allowing other optimizations to do
6788 @item -funroll-loops
6789 @opindex funroll-loops
6790 Unroll loops whose number of iterations can be determined at compile
6791 time or upon entry to the loop. @option{-funroll-loops} implies
6792 @option{-frerun-cse-after-loop}. This option makes code larger,
6793 and may or may not make it run faster.
6795 @item -funroll-all-loops
6796 @opindex funroll-all-loops
6797 Unroll all loops, even if their number of iterations is uncertain when
6798 the loop is entered. This usually makes programs run more slowly.
6799 @option{-funroll-all-loops} implies the same options as
6800 @option{-funroll-loops},
6802 @item -fsplit-ivs-in-unroller
6803 @opindex fsplit-ivs-in-unroller
6804 Enables expressing of values of induction variables in later iterations
6805 of the unrolled loop using the value in the first iteration. This breaks
6806 long dependency chains, thus improving efficiency of the scheduling passes.
6808 Combination of @option{-fweb} and CSE is often sufficient to obtain the
6809 same effect. However in cases the loop body is more complicated than
6810 a single basic block, this is not reliable. It also does not work at all
6811 on some of the architectures due to restrictions in the CSE pass.
6813 This optimization is enabled by default.
6815 @item -fvariable-expansion-in-unroller
6816 @opindex fvariable-expansion-in-unroller
6817 With this option, the compiler will create multiple copies of some
6818 local variables when unrolling a loop which can result in superior code.
6820 @item -fpredictive-commoning
6821 @opindex fpredictive-commoning
6822 Perform predictive commoning optimization, i.e., reusing computations
6823 (especially memory loads and stores) performed in previous
6824 iterations of loops.
6826 This option is enabled at level @option{-O3}.
6828 @item -fprefetch-loop-arrays
6829 @opindex fprefetch-loop-arrays
6830 If supported by the target machine, generate instructions to prefetch
6831 memory to improve the performance of loops that access large arrays.
6833 This option may generate better or worse code; results are highly
6834 dependent on the structure of loops within the source code.
6836 Disabled at level @option{-Os}.
6839 @itemx -fno-peephole2
6840 @opindex fno-peephole
6841 @opindex fno-peephole2
6842 Disable any machine-specific peephole optimizations. The difference
6843 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
6844 are implemented in the compiler; some targets use one, some use the
6845 other, a few use both.
6847 @option{-fpeephole} is enabled by default.
6848 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6850 @item -fno-guess-branch-probability
6851 @opindex fno-guess-branch-probability
6852 Do not guess branch probabilities using heuristics.
6854 GCC will use heuristics to guess branch probabilities if they are
6855 not provided by profiling feedback (@option{-fprofile-arcs}). These
6856 heuristics are based on the control flow graph. If some branch probabilities
6857 are specified by @samp{__builtin_expect}, then the heuristics will be
6858 used to guess branch probabilities for the rest of the control flow graph,
6859 taking the @samp{__builtin_expect} info into account. The interactions
6860 between the heuristics and @samp{__builtin_expect} can be complex, and in
6861 some cases, it may be useful to disable the heuristics so that the effects
6862 of @samp{__builtin_expect} are easier to understand.
6864 The default is @option{-fguess-branch-probability} at levels
6865 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6867 @item -freorder-blocks
6868 @opindex freorder-blocks
6869 Reorder basic blocks in the compiled function in order to reduce number of
6870 taken branches and improve code locality.
6872 Enabled at levels @option{-O2}, @option{-O3}.
6874 @item -freorder-blocks-and-partition
6875 @opindex freorder-blocks-and-partition
6876 In addition to reordering basic blocks in the compiled function, in order
6877 to reduce number of taken branches, partitions hot and cold basic blocks
6878 into separate sections of the assembly and .o files, to improve
6879 paging and cache locality performance.
6881 This optimization is automatically turned off in the presence of
6882 exception handling, for linkonce sections, for functions with a user-defined
6883 section attribute and on any architecture that does not support named
6886 @item -freorder-functions
6887 @opindex freorder-functions
6888 Reorder functions in the object file in order to
6889 improve code locality. This is implemented by using special
6890 subsections @code{.text.hot} for most frequently executed functions and
6891 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
6892 the linker so object file format must support named sections and linker must
6893 place them in a reasonable way.
6895 Also profile feedback must be available in to make this option effective. See
6896 @option{-fprofile-arcs} for details.
6898 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6900 @item -fstrict-aliasing
6901 @opindex fstrict-aliasing
6902 Allow the compiler to assume the strictest aliasing rules applicable to
6903 the language being compiled. For C (and C++), this activates
6904 optimizations based on the type of expressions. In particular, an
6905 object of one type is assumed never to reside at the same address as an
6906 object of a different type, unless the types are almost the same. For
6907 example, an @code{unsigned int} can alias an @code{int}, but not a
6908 @code{void*} or a @code{double}. A character type may alias any other
6911 @anchor{Type-punning}Pay special attention to code like this:
6924 The practice of reading from a different union member than the one most
6925 recently written to (called ``type-punning'') is common. Even with
6926 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
6927 is accessed through the union type. So, the code above will work as
6928 expected. @xref{Structures unions enumerations and bit-fields
6929 implementation}. However, this code might not:
6940 Similarly, access by taking the address, casting the resulting pointer
6941 and dereferencing the result has undefined behavior, even if the cast
6942 uses a union type, e.g.:
6946 return ((union a_union *) &d)->i;
6950 The @option{-fstrict-aliasing} option is enabled at levels
6951 @option{-O2}, @option{-O3}, @option{-Os}.
6953 @item -fstrict-overflow
6954 @opindex fstrict-overflow
6955 Allow the compiler to assume strict signed overflow rules, depending
6956 on the language being compiled. For C (and C++) this means that
6957 overflow when doing arithmetic with signed numbers is undefined, which
6958 means that the compiler may assume that it will not happen. This
6959 permits various optimizations. For example, the compiler will assume
6960 that an expression like @code{i + 10 > i} will always be true for
6961 signed @code{i}. This assumption is only valid if signed overflow is
6962 undefined, as the expression is false if @code{i + 10} overflows when
6963 using twos complement arithmetic. When this option is in effect any
6964 attempt to determine whether an operation on signed numbers will
6965 overflow must be written carefully to not actually involve overflow.
6967 This option also allows the compiler to assume strict pointer
6968 semantics: given a pointer to an object, if adding an offset to that
6969 pointer does not produce a pointer to the same object, the addition is
6970 undefined. This permits the compiler to conclude that @code{p + u >
6971 p} is always true for a pointer @code{p} and unsigned integer
6972 @code{u}. This assumption is only valid because pointer wraparound is
6973 undefined, as the expression is false if @code{p + u} overflows using
6974 twos complement arithmetic.
6976 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
6977 that integer signed overflow is fully defined: it wraps. When
6978 @option{-fwrapv} is used, there is no difference between
6979 @option{-fstrict-overflow} and @option{-fno-strict-overflow} for
6980 integers. With @option{-fwrapv} certain types of overflow are
6981 permitted. For example, if the compiler gets an overflow when doing
6982 arithmetic on constants, the overflowed value can still be used with
6983 @option{-fwrapv}, but not otherwise.
6985 The @option{-fstrict-overflow} option is enabled at levels
6986 @option{-O2}, @option{-O3}, @option{-Os}.
6988 @item -falign-functions
6989 @itemx -falign-functions=@var{n}
6990 @opindex falign-functions
6991 Align the start of functions to the next power-of-two greater than
6992 @var{n}, skipping up to @var{n} bytes. For instance,
6993 @option{-falign-functions=32} aligns functions to the next 32-byte
6994 boundary, but @option{-falign-functions=24} would align to the next
6995 32-byte boundary only if this can be done by skipping 23 bytes or less.
6997 @option{-fno-align-functions} and @option{-falign-functions=1} are
6998 equivalent and mean that functions will not be aligned.
7000 Some assemblers only support this flag when @var{n} is a power of two;
7001 in that case, it is rounded up.
7003 If @var{n} is not specified or is zero, use a machine-dependent default.
7005 Enabled at levels @option{-O2}, @option{-O3}.
7007 @item -falign-labels
7008 @itemx -falign-labels=@var{n}
7009 @opindex falign-labels
7010 Align all branch targets to a power-of-two boundary, skipping up to
7011 @var{n} bytes like @option{-falign-functions}. This option can easily
7012 make code slower, because it must insert dummy operations for when the
7013 branch target is reached in the usual flow of the code.
7015 @option{-fno-align-labels} and @option{-falign-labels=1} are
7016 equivalent and mean that labels will not be aligned.
7018 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
7019 are greater than this value, then their values are used instead.
7021 If @var{n} is not specified or is zero, use a machine-dependent default
7022 which is very likely to be @samp{1}, meaning no alignment.
7024 Enabled at levels @option{-O2}, @option{-O3}.
7027 @itemx -falign-loops=@var{n}
7028 @opindex falign-loops
7029 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
7030 like @option{-falign-functions}. The hope is that the loop will be
7031 executed many times, which will make up for any execution of the dummy
7034 @option{-fno-align-loops} and @option{-falign-loops=1} are
7035 equivalent and mean that loops will not be aligned.
7037 If @var{n} is not specified or is zero, use a machine-dependent default.
7039 Enabled at levels @option{-O2}, @option{-O3}.
7042 @itemx -falign-jumps=@var{n}
7043 @opindex falign-jumps
7044 Align branch targets to a power-of-two boundary, for branch targets
7045 where the targets can only be reached by jumping, skipping up to @var{n}
7046 bytes like @option{-falign-functions}. In this case, no dummy operations
7049 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
7050 equivalent and mean that loops will not be aligned.
7052 If @var{n} is not specified or is zero, use a machine-dependent default.
7054 Enabled at levels @option{-O2}, @option{-O3}.
7056 @item -funit-at-a-time
7057 @opindex funit-at-a-time
7058 This option is left for compatibility reasons. @option{-funit-at-a-time}
7059 has no effect, while @option{-fno-unit-at-a-time} implies
7060 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
7064 @item -fno-toplevel-reorder
7065 @opindex fno-toplevel-reorder
7066 Do not reorder top-level functions, variables, and @code{asm}
7067 statements. Output them in the same order that they appear in the
7068 input file. When this option is used, unreferenced static variables
7069 will not be removed. This option is intended to support existing code
7070 which relies on a particular ordering. For new code, it is better to
7073 Enabled at level @option{-O0}. When disabled explicitly, it also imply
7074 @option{-fno-section-anchors} that is otherwise enabled at @option{-O0} on some
7079 Constructs webs as commonly used for register allocation purposes and assign
7080 each web individual pseudo register. This allows the register allocation pass
7081 to operate on pseudos directly, but also strengthens several other optimization
7082 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
7083 however, make debugging impossible, since variables will no longer stay in a
7086 Enabled by default with @option{-funroll-loops}.
7088 @item -fwhole-program
7089 @opindex fwhole-program
7090 Assume that the current compilation unit represents the whole program being
7091 compiled. All public functions and variables with the exception of @code{main}
7092 and those merged by attribute @code{externally_visible} become static functions
7093 and in effect are optimized more aggressively by interprocedural optimizers.
7094 While this option is equivalent to proper use of the @code{static} keyword for
7095 programs consisting of a single file, in combination with option
7096 @option{--combine} this flag can be used to compile many smaller scale C
7097 programs since the functions and variables become local for the whole combined
7098 compilation unit, not for the single source file itself.
7100 This option implies @option{-fwhole-file} for Fortran programs.
7102 @item -fcprop-registers
7103 @opindex fcprop-registers
7104 After register allocation and post-register allocation instruction splitting,
7105 we perform a copy-propagation pass to try to reduce scheduling dependencies
7106 and occasionally eliminate the copy.
7108 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7110 @item -fprofile-correction
7111 @opindex fprofile-correction
7112 Profiles collected using an instrumented binary for multi-threaded programs may
7113 be inconsistent due to missed counter updates. When this option is specified,
7114 GCC will use heuristics to correct or smooth out such inconsistencies. By
7115 default, GCC will emit an error message when an inconsistent profile is detected.
7117 @item -fprofile-dir=@var{path}
7118 @opindex fprofile-dir
7120 Set the directory to search the profile data files in to @var{path}.
7121 This option affects only the profile data generated by
7122 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
7123 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
7124 and its related options.
7125 By default, GCC will use the current directory as @var{path}
7126 thus the profile data file will appear in the same directory as the object file.
7128 @item -fprofile-generate
7129 @itemx -fprofile-generate=@var{path}
7130 @opindex fprofile-generate
7132 Enable options usually used for instrumenting application to produce
7133 profile useful for later recompilation with profile feedback based
7134 optimization. You must use @option{-fprofile-generate} both when
7135 compiling and when linking your program.
7137 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
7139 If @var{path} is specified, GCC will look at the @var{path} to find
7140 the profile feedback data files. See @option{-fprofile-dir}.
7143 @itemx -fprofile-use=@var{path}
7144 @opindex fprofile-use
7145 Enable profile feedback directed optimizations, and optimizations
7146 generally profitable only with profile feedback available.
7148 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
7149 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
7151 By default, GCC emits an error message if the feedback profiles do not
7152 match the source code. This error can be turned into a warning by using
7153 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
7156 If @var{path} is specified, GCC will look at the @var{path} to find
7157 the profile feedback data files. See @option{-fprofile-dir}.
7160 The following options control compiler behavior regarding floating
7161 point arithmetic. These options trade off between speed and
7162 correctness. All must be specifically enabled.
7166 @opindex ffloat-store
7167 Do not store floating point variables in registers, and inhibit other
7168 options that might change whether a floating point value is taken from a
7171 @cindex floating point precision
7172 This option prevents undesirable excess precision on machines such as
7173 the 68000 where the floating registers (of the 68881) keep more
7174 precision than a @code{double} is supposed to have. Similarly for the
7175 x86 architecture. For most programs, the excess precision does only
7176 good, but a few programs rely on the precise definition of IEEE floating
7177 point. Use @option{-ffloat-store} for such programs, after modifying
7178 them to store all pertinent intermediate computations into variables.
7180 @item -fexcess-precision=@var{style}
7181 @opindex fexcess-precision
7182 This option allows further control over excess precision on machines
7183 where floating-point registers have more precision than the IEEE
7184 @code{float} and @code{double} types and the processor does not
7185 support operations rounding to those types. By default,
7186 @option{-fexcess-precision=fast} is in effect; this means that
7187 operations are carried out in the precision of the registers and that
7188 it is unpredictable when rounding to the types specified in the source
7189 code takes place. When compiling C, if
7190 @option{-fexcess-precision=standard} is specified then excess
7191 precision will follow the rules specified in ISO C99; in particular,
7192 both casts and assignments cause values to be rounded to their
7193 semantic types (whereas @option{-ffloat-store} only affects
7194 assignments). This option is enabled by default for C if a strict
7195 conformance option such as @option{-std=c99} is used.
7198 @option{-fexcess-precision=standard} is not implemented for languages
7199 other than C, and has no effect if
7200 @option{-funsafe-math-optimizations} or @option{-ffast-math} is
7201 specified. On the x86, it also has no effect if @option{-mfpmath=sse}
7202 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
7203 semantics apply without excess precision, and in the latter, rounding
7208 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
7209 @option{-ffinite-math-only}, @option{-fno-rounding-math},
7210 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
7212 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
7214 This option is not turned on by any @option{-O} option since
7215 it can result in incorrect output for programs which depend on
7216 an exact implementation of IEEE or ISO rules/specifications for
7217 math functions. It may, however, yield faster code for programs
7218 that do not require the guarantees of these specifications.
7220 @item -fno-math-errno
7221 @opindex fno-math-errno
7222 Do not set ERRNO after calling math functions that are executed
7223 with a single instruction, e.g., sqrt. A program that relies on
7224 IEEE exceptions for math error handling may want to use this flag
7225 for speed while maintaining IEEE arithmetic compatibility.
7227 This option is not turned on by any @option{-O} option since
7228 it can result in incorrect output for programs which depend on
7229 an exact implementation of IEEE or ISO rules/specifications for
7230 math functions. It may, however, yield faster code for programs
7231 that do not require the guarantees of these specifications.
7233 The default is @option{-fmath-errno}.
7235 On Darwin systems, the math library never sets @code{errno}. There is
7236 therefore no reason for the compiler to consider the possibility that
7237 it might, and @option{-fno-math-errno} is the default.
7239 @item -funsafe-math-optimizations
7240 @opindex funsafe-math-optimizations
7242 Allow optimizations for floating-point arithmetic that (a) assume
7243 that arguments and results are valid and (b) may violate IEEE or
7244 ANSI standards. When used at link-time, it may include libraries
7245 or startup files that change the default FPU control word or other
7246 similar optimizations.
7248 This option is not turned on by any @option{-O} option since
7249 it can result in incorrect output for programs which depend on
7250 an exact implementation of IEEE or ISO rules/specifications for
7251 math functions. It may, however, yield faster code for programs
7252 that do not require the guarantees of these specifications.
7253 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
7254 @option{-fassociative-math} and @option{-freciprocal-math}.
7256 The default is @option{-fno-unsafe-math-optimizations}.
7258 @item -fassociative-math
7259 @opindex fassociative-math
7261 Allow re-association of operands in series of floating-point operations.
7262 This violates the ISO C and C++ language standard by possibly changing
7263 computation result. NOTE: re-ordering may change the sign of zero as
7264 well as ignore NaNs and inhibit or create underflow or overflow (and
7265 thus cannot be used on a code which relies on rounding behavior like
7266 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
7267 and thus may not be used when ordered comparisons are required.
7268 This option requires that both @option{-fno-signed-zeros} and
7269 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
7270 much sense with @option{-frounding-math}.
7272 The default is @option{-fno-associative-math}.
7274 @item -freciprocal-math
7275 @opindex freciprocal-math
7277 Allow the reciprocal of a value to be used instead of dividing by
7278 the value if this enables optimizations. For example @code{x / y}
7279 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
7280 is subject to common subexpression elimination. Note that this loses
7281 precision and increases the number of flops operating on the value.
7283 The default is @option{-fno-reciprocal-math}.
7285 @item -ffinite-math-only
7286 @opindex ffinite-math-only
7287 Allow optimizations for floating-point arithmetic that assume
7288 that arguments and results are not NaNs or +-Infs.
7290 This option is not turned on by any @option{-O} option since
7291 it can result in incorrect output for programs which depend on
7292 an exact implementation of IEEE or ISO rules/specifications for
7293 math functions. It may, however, yield faster code for programs
7294 that do not require the guarantees of these specifications.
7296 The default is @option{-fno-finite-math-only}.
7298 @item -fno-signed-zeros
7299 @opindex fno-signed-zeros
7300 Allow optimizations for floating point arithmetic that ignore the
7301 signedness of zero. IEEE arithmetic specifies the behavior of
7302 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
7303 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
7304 This option implies that the sign of a zero result isn't significant.
7306 The default is @option{-fsigned-zeros}.
7308 @item -fno-trapping-math
7309 @opindex fno-trapping-math
7310 Compile code assuming that floating-point operations cannot generate
7311 user-visible traps. These traps include division by zero, overflow,
7312 underflow, inexact result and invalid operation. This option requires
7313 that @option{-fno-signaling-nans} be in effect. Setting this option may
7314 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
7316 This option should never be turned on by any @option{-O} option since
7317 it can result in incorrect output for programs which depend on
7318 an exact implementation of IEEE or ISO rules/specifications for
7321 The default is @option{-ftrapping-math}.
7323 @item -frounding-math
7324 @opindex frounding-math
7325 Disable transformations and optimizations that assume default floating
7326 point rounding behavior. This is round-to-zero for all floating point
7327 to integer conversions, and round-to-nearest for all other arithmetic
7328 truncations. This option should be specified for programs that change
7329 the FP rounding mode dynamically, or that may be executed with a
7330 non-default rounding mode. This option disables constant folding of
7331 floating point expressions at compile-time (which may be affected by
7332 rounding mode) and arithmetic transformations that are unsafe in the
7333 presence of sign-dependent rounding modes.
7335 The default is @option{-fno-rounding-math}.
7337 This option is experimental and does not currently guarantee to
7338 disable all GCC optimizations that are affected by rounding mode.
7339 Future versions of GCC may provide finer control of this setting
7340 using C99's @code{FENV_ACCESS} pragma. This command line option
7341 will be used to specify the default state for @code{FENV_ACCESS}.
7343 @item -fsignaling-nans
7344 @opindex fsignaling-nans
7345 Compile code assuming that IEEE signaling NaNs may generate user-visible
7346 traps during floating-point operations. Setting this option disables
7347 optimizations that may change the number of exceptions visible with
7348 signaling NaNs. This option implies @option{-ftrapping-math}.
7350 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
7353 The default is @option{-fno-signaling-nans}.
7355 This option is experimental and does not currently guarantee to
7356 disable all GCC optimizations that affect signaling NaN behavior.
7358 @item -fsingle-precision-constant
7359 @opindex fsingle-precision-constant
7360 Treat floating point constant as single precision constant instead of
7361 implicitly converting it to double precision constant.
7363 @item -fcx-limited-range
7364 @opindex fcx-limited-range
7365 When enabled, this option states that a range reduction step is not
7366 needed when performing complex division. Also, there is no checking
7367 whether the result of a complex multiplication or division is @code{NaN
7368 + I*NaN}, with an attempt to rescue the situation in that case. The
7369 default is @option{-fno-cx-limited-range}, but is enabled by
7370 @option{-ffast-math}.
7372 This option controls the default setting of the ISO C99
7373 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
7376 @item -fcx-fortran-rules
7377 @opindex fcx-fortran-rules
7378 Complex multiplication and division follow Fortran rules. Range
7379 reduction is done as part of complex division, but there is no checking
7380 whether the result of a complex multiplication or division is @code{NaN
7381 + I*NaN}, with an attempt to rescue the situation in that case.
7383 The default is @option{-fno-cx-fortran-rules}.
7387 The following options control optimizations that may improve
7388 performance, but are not enabled by any @option{-O} options. This
7389 section includes experimental options that may produce broken code.
7392 @item -fbranch-probabilities
7393 @opindex fbranch-probabilities
7394 After running a program compiled with @option{-fprofile-arcs}
7395 (@pxref{Debugging Options,, Options for Debugging Your Program or
7396 @command{gcc}}), you can compile it a second time using
7397 @option{-fbranch-probabilities}, to improve optimizations based on
7398 the number of times each branch was taken. When the program
7399 compiled with @option{-fprofile-arcs} exits it saves arc execution
7400 counts to a file called @file{@var{sourcename}.gcda} for each source
7401 file. The information in this data file is very dependent on the
7402 structure of the generated code, so you must use the same source code
7403 and the same optimization options for both compilations.
7405 With @option{-fbranch-probabilities}, GCC puts a
7406 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
7407 These can be used to improve optimization. Currently, they are only
7408 used in one place: in @file{reorg.c}, instead of guessing which path a
7409 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
7410 exactly determine which path is taken more often.
7412 @item -fprofile-values
7413 @opindex fprofile-values
7414 If combined with @option{-fprofile-arcs}, it adds code so that some
7415 data about values of expressions in the program is gathered.
7417 With @option{-fbranch-probabilities}, it reads back the data gathered
7418 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
7419 notes to instructions for their later usage in optimizations.
7421 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
7425 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
7426 a code to gather information about values of expressions.
7428 With @option{-fbranch-probabilities}, it reads back the data gathered
7429 and actually performs the optimizations based on them.
7430 Currently the optimizations include specialization of division operation
7431 using the knowledge about the value of the denominator.
7433 @item -frename-registers
7434 @opindex frename-registers
7435 Attempt to avoid false dependencies in scheduled code by making use
7436 of registers left over after register allocation. This optimization
7437 will most benefit processors with lots of registers. Depending on the
7438 debug information format adopted by the target, however, it can
7439 make debugging impossible, since variables will no longer stay in
7440 a ``home register''.
7442 Enabled by default with @option{-funroll-loops}.
7446 Perform tail duplication to enlarge superblock size. This transformation
7447 simplifies the control flow of the function allowing other optimizations to do
7450 Enabled with @option{-fprofile-use}.
7452 @item -funroll-loops
7453 @opindex funroll-loops
7454 Unroll loops whose number of iterations can be determined at compile time or
7455 upon entry to the loop. @option{-funroll-loops} implies
7456 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
7457 It also turns on complete loop peeling (i.e.@: complete removal of loops with
7458 small constant number of iterations). This option makes code larger, and may
7459 or may not make it run faster.
7461 Enabled with @option{-fprofile-use}.
7463 @item -funroll-all-loops
7464 @opindex funroll-all-loops
7465 Unroll all loops, even if their number of iterations is uncertain when
7466 the loop is entered. This usually makes programs run more slowly.
7467 @option{-funroll-all-loops} implies the same options as
7468 @option{-funroll-loops}.
7471 @opindex fpeel-loops
7472 Peels the loops for that there is enough information that they do not
7473 roll much (from profile feedback). It also turns on complete loop peeling
7474 (i.e.@: complete removal of loops with small constant number of iterations).
7476 Enabled with @option{-fprofile-use}.
7478 @item -fmove-loop-invariants
7479 @opindex fmove-loop-invariants
7480 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
7481 at level @option{-O1}
7483 @item -funswitch-loops
7484 @opindex funswitch-loops
7485 Move branches with loop invariant conditions out of the loop, with duplicates
7486 of the loop on both branches (modified according to result of the condition).
7488 @item -ffunction-sections
7489 @itemx -fdata-sections
7490 @opindex ffunction-sections
7491 @opindex fdata-sections
7492 Place each function or data item into its own section in the output
7493 file if the target supports arbitrary sections. The name of the
7494 function or the name of the data item determines the section's name
7497 Use these options on systems where the linker can perform optimizations
7498 to improve locality of reference in the instruction space. Most systems
7499 using the ELF object format and SPARC processors running Solaris 2 have
7500 linkers with such optimizations. AIX may have these optimizations in
7503 Only use these options when there are significant benefits from doing
7504 so. When you specify these options, the assembler and linker will
7505 create larger object and executable files and will also be slower.
7506 You will not be able to use @code{gprof} on all systems if you
7507 specify this option and you may have problems with debugging if
7508 you specify both this option and @option{-g}.
7510 @item -fbranch-target-load-optimize
7511 @opindex fbranch-target-load-optimize
7512 Perform branch target register load optimization before prologue / epilogue
7514 The use of target registers can typically be exposed only during reload,
7515 thus hoisting loads out of loops and doing inter-block scheduling needs
7516 a separate optimization pass.
7518 @item -fbranch-target-load-optimize2
7519 @opindex fbranch-target-load-optimize2
7520 Perform branch target register load optimization after prologue / epilogue
7523 @item -fbtr-bb-exclusive
7524 @opindex fbtr-bb-exclusive
7525 When performing branch target register load optimization, don't reuse
7526 branch target registers in within any basic block.
7528 @item -fstack-protector
7529 @opindex fstack-protector
7530 Emit extra code to check for buffer overflows, such as stack smashing
7531 attacks. This is done by adding a guard variable to functions with
7532 vulnerable objects. This includes functions that call alloca, and
7533 functions with buffers larger than 8 bytes. The guards are initialized
7534 when a function is entered and then checked when the function exits.
7535 If a guard check fails, an error message is printed and the program exits.
7537 @item -fstack-protector-all
7538 @opindex fstack-protector-all
7539 Like @option{-fstack-protector} except that all functions are protected.
7541 @item -fsection-anchors
7542 @opindex fsection-anchors
7543 Try to reduce the number of symbolic address calculations by using
7544 shared ``anchor'' symbols to address nearby objects. This transformation
7545 can help to reduce the number of GOT entries and GOT accesses on some
7548 For example, the implementation of the following function @code{foo}:
7552 int foo (void) @{ return a + b + c; @}
7555 would usually calculate the addresses of all three variables, but if you
7556 compile it with @option{-fsection-anchors}, it will access the variables
7557 from a common anchor point instead. The effect is similar to the
7558 following pseudocode (which isn't valid C):
7563 register int *xr = &x;
7564 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
7568 Not all targets support this option.
7570 @item --param @var{name}=@var{value}
7572 In some places, GCC uses various constants to control the amount of
7573 optimization that is done. For example, GCC will not inline functions
7574 that contain more that a certain number of instructions. You can
7575 control some of these constants on the command-line using the
7576 @option{--param} option.
7578 The names of specific parameters, and the meaning of the values, are
7579 tied to the internals of the compiler, and are subject to change
7580 without notice in future releases.
7582 In each case, the @var{value} is an integer. The allowable choices for
7583 @var{name} are given in the following table:
7586 @item struct-reorg-cold-struct-ratio
7587 The threshold ratio (as a percentage) between a structure frequency
7588 and the frequency of the hottest structure in the program. This parameter
7589 is used by struct-reorg optimization enabled by @option{-fipa-struct-reorg}.
7590 We say that if the ratio of a structure frequency, calculated by profiling,
7591 to the hottest structure frequency in the program is less than this
7592 parameter, then structure reorganization is not applied to this structure.
7595 @item predictable-branch-cost-outcome
7596 When branch is predicted to be taken with probability lower than this threshold
7597 (in percent), then it is considered well predictable. The default is 10.
7599 @item max-crossjump-edges
7600 The maximum number of incoming edges to consider for crossjumping.
7601 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
7602 the number of edges incoming to each block. Increasing values mean
7603 more aggressive optimization, making the compile time increase with
7604 probably small improvement in executable size.
7606 @item min-crossjump-insns
7607 The minimum number of instructions which must be matched at the end
7608 of two blocks before crossjumping will be performed on them. This
7609 value is ignored in the case where all instructions in the block being
7610 crossjumped from are matched. The default value is 5.
7612 @item max-grow-copy-bb-insns
7613 The maximum code size expansion factor when copying basic blocks
7614 instead of jumping. The expansion is relative to a jump instruction.
7615 The default value is 8.
7617 @item max-goto-duplication-insns
7618 The maximum number of instructions to duplicate to a block that jumps
7619 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
7620 passes, GCC factors computed gotos early in the compilation process,
7621 and unfactors them as late as possible. Only computed jumps at the
7622 end of a basic blocks with no more than max-goto-duplication-insns are
7623 unfactored. The default value is 8.
7625 @item max-delay-slot-insn-search
7626 The maximum number of instructions to consider when looking for an
7627 instruction to fill a delay slot. If more than this arbitrary number of
7628 instructions is searched, the time savings from filling the delay slot
7629 will be minimal so stop searching. Increasing values mean more
7630 aggressive optimization, making the compile time increase with probably
7631 small improvement in executable run time.
7633 @item max-delay-slot-live-search
7634 When trying to fill delay slots, the maximum number of instructions to
7635 consider when searching for a block with valid live register
7636 information. Increasing this arbitrarily chosen value means more
7637 aggressive optimization, increasing the compile time. This parameter
7638 should be removed when the delay slot code is rewritten to maintain the
7641 @item max-gcse-memory
7642 The approximate maximum amount of memory that will be allocated in
7643 order to perform the global common subexpression elimination
7644 optimization. If more memory than specified is required, the
7645 optimization will not be done.
7647 @item max-pending-list-length
7648 The maximum number of pending dependencies scheduling will allow
7649 before flushing the current state and starting over. Large functions
7650 with few branches or calls can create excessively large lists which
7651 needlessly consume memory and resources.
7653 @item max-inline-insns-single
7654 Several parameters control the tree inliner used in gcc.
7655 This number sets the maximum number of instructions (counted in GCC's
7656 internal representation) in a single function that the tree inliner
7657 will consider for inlining. This only affects functions declared
7658 inline and methods implemented in a class declaration (C++).
7659 The default value is 300.
7661 @item max-inline-insns-auto
7662 When you use @option{-finline-functions} (included in @option{-O3}),
7663 a lot of functions that would otherwise not be considered for inlining
7664 by the compiler will be investigated. To those functions, a different
7665 (more restrictive) limit compared to functions declared inline can
7667 The default value is 50.
7669 @item large-function-insns
7670 The limit specifying really large functions. For functions larger than this
7671 limit after inlining, inlining is constrained by
7672 @option{--param large-function-growth}. This parameter is useful primarily
7673 to avoid extreme compilation time caused by non-linear algorithms used by the
7675 The default value is 2700.
7677 @item large-function-growth
7678 Specifies maximal growth of large function caused by inlining in percents.
7679 The default value is 100 which limits large function growth to 2.0 times
7682 @item large-unit-insns
7683 The limit specifying large translation unit. Growth caused by inlining of
7684 units larger than this limit is limited by @option{--param inline-unit-growth}.
7685 For small units this might be too tight (consider unit consisting of function A
7686 that is inline and B that just calls A three time. If B is small relative to
7687 A, the growth of unit is 300\% and yet such inlining is very sane. For very
7688 large units consisting of small inlineable functions however the overall unit
7689 growth limit is needed to avoid exponential explosion of code size. Thus for
7690 smaller units, the size is increased to @option{--param large-unit-insns}
7691 before applying @option{--param inline-unit-growth}. The default is 10000
7693 @item inline-unit-growth
7694 Specifies maximal overall growth of the compilation unit caused by inlining.
7695 The default value is 30 which limits unit growth to 1.3 times the original
7698 @item ipcp-unit-growth
7699 Specifies maximal overall growth of the compilation unit caused by
7700 interprocedural constant propagation. The default value is 10 which limits
7701 unit growth to 1.1 times the original size.
7703 @item large-stack-frame
7704 The limit specifying large stack frames. While inlining the algorithm is trying
7705 to not grow past this limit too much. Default value is 256 bytes.
7707 @item large-stack-frame-growth
7708 Specifies maximal growth of large stack frames caused by inlining in percents.
7709 The default value is 1000 which limits large stack frame growth to 11 times
7712 @item max-inline-insns-recursive
7713 @itemx max-inline-insns-recursive-auto
7714 Specifies maximum number of instructions out-of-line copy of self recursive inline
7715 function can grow into by performing recursive inlining.
7717 For functions declared inline @option{--param max-inline-insns-recursive} is
7718 taken into account. For function not declared inline, recursive inlining
7719 happens only when @option{-finline-functions} (included in @option{-O3}) is
7720 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
7721 default value is 450.
7723 @item max-inline-recursive-depth
7724 @itemx max-inline-recursive-depth-auto
7725 Specifies maximum recursion depth used by the recursive inlining.
7727 For functions declared inline @option{--param max-inline-recursive-depth} is
7728 taken into account. For function not declared inline, recursive inlining
7729 happens only when @option{-finline-functions} (included in @option{-O3}) is
7730 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
7733 @item min-inline-recursive-probability
7734 Recursive inlining is profitable only for function having deep recursion
7735 in average and can hurt for function having little recursion depth by
7736 increasing the prologue size or complexity of function body to other
7739 When profile feedback is available (see @option{-fprofile-generate}) the actual
7740 recursion depth can be guessed from probability that function will recurse via
7741 given call expression. This parameter limits inlining only to call expression
7742 whose probability exceeds given threshold (in percents). The default value is
7745 @item early-inlining-insns
7746 Specify growth that early inliner can make. In effect it increases amount of
7747 inlining for code having large abstraction penalty. The default value is 8.
7749 @item max-early-inliner-iterations
7750 @itemx max-early-inliner-iterations
7751 Limit of iterations of early inliner. This basically bounds number of nested
7752 indirect calls early inliner can resolve. Deeper chains are still handled by
7755 @item min-vect-loop-bound
7756 The minimum number of iterations under which a loop will not get vectorized
7757 when @option{-ftree-vectorize} is used. The number of iterations after
7758 vectorization needs to be greater than the value specified by this option
7759 to allow vectorization. The default value is 0.
7761 @item max-unrolled-insns
7762 The maximum number of instructions that a loop should have if that loop
7763 is unrolled, and if the loop is unrolled, it determines how many times
7764 the loop code is unrolled.
7766 @item max-average-unrolled-insns
7767 The maximum number of instructions biased by probabilities of their execution
7768 that a loop should have if that loop is unrolled, and if the loop is unrolled,
7769 it determines how many times the loop code is unrolled.
7771 @item max-unroll-times
7772 The maximum number of unrollings of a single loop.
7774 @item max-peeled-insns
7775 The maximum number of instructions that a loop should have if that loop
7776 is peeled, and if the loop is peeled, it determines how many times
7777 the loop code is peeled.
7779 @item max-peel-times
7780 The maximum number of peelings of a single loop.
7782 @item max-completely-peeled-insns
7783 The maximum number of insns of a completely peeled loop.
7785 @item max-completely-peel-times
7786 The maximum number of iterations of a loop to be suitable for complete peeling.
7788 @item max-unswitch-insns
7789 The maximum number of insns of an unswitched loop.
7791 @item max-unswitch-level
7792 The maximum number of branches unswitched in a single loop.
7795 The minimum cost of an expensive expression in the loop invariant motion.
7797 @item iv-consider-all-candidates-bound
7798 Bound on number of candidates for induction variables below that
7799 all candidates are considered for each use in induction variable
7800 optimizations. Only the most relevant candidates are considered
7801 if there are more candidates, to avoid quadratic time complexity.
7803 @item iv-max-considered-uses
7804 The induction variable optimizations give up on loops that contain more
7805 induction variable uses.
7807 @item iv-always-prune-cand-set-bound
7808 If number of candidates in the set is smaller than this value,
7809 we always try to remove unnecessary ivs from the set during its
7810 optimization when a new iv is added to the set.
7812 @item scev-max-expr-size
7813 Bound on size of expressions used in the scalar evolutions analyzer.
7814 Large expressions slow the analyzer.
7816 @item omega-max-vars
7817 The maximum number of variables in an Omega constraint system.
7818 The default value is 128.
7820 @item omega-max-geqs
7821 The maximum number of inequalities in an Omega constraint system.
7822 The default value is 256.
7825 The maximum number of equalities in an Omega constraint system.
7826 The default value is 128.
7828 @item omega-max-wild-cards
7829 The maximum number of wildcard variables that the Omega solver will
7830 be able to insert. The default value is 18.
7832 @item omega-hash-table-size
7833 The size of the hash table in the Omega solver. The default value is
7836 @item omega-max-keys
7837 The maximal number of keys used by the Omega solver. The default
7840 @item omega-eliminate-redundant-constraints
7841 When set to 1, use expensive methods to eliminate all redundant
7842 constraints. The default value is 0.
7844 @item vect-max-version-for-alignment-checks
7845 The maximum number of runtime checks that can be performed when
7846 doing loop versioning for alignment in the vectorizer. See option
7847 ftree-vect-loop-version for more information.
7849 @item vect-max-version-for-alias-checks
7850 The maximum number of runtime checks that can be performed when
7851 doing loop versioning for alias in the vectorizer. See option
7852 ftree-vect-loop-version for more information.
7854 @item max-iterations-to-track
7856 The maximum number of iterations of a loop the brute force algorithm
7857 for analysis of # of iterations of the loop tries to evaluate.
7859 @item hot-bb-count-fraction
7860 Select fraction of the maximal count of repetitions of basic block in program
7861 given basic block needs to have to be considered hot.
7863 @item hot-bb-frequency-fraction
7864 Select fraction of the maximal frequency of executions of basic block in
7865 function given basic block needs to have to be considered hot
7867 @item max-predicted-iterations
7868 The maximum number of loop iterations we predict statically. This is useful
7869 in cases where function contain single loop with known bound and other loop
7870 with unknown. We predict the known number of iterations correctly, while
7871 the unknown number of iterations average to roughly 10. This means that the
7872 loop without bounds would appear artificially cold relative to the other one.
7874 @item align-threshold
7876 Select fraction of the maximal frequency of executions of basic block in
7877 function given basic block will get aligned.
7879 @item align-loop-iterations
7881 A loop expected to iterate at lest the selected number of iterations will get
7884 @item tracer-dynamic-coverage
7885 @itemx tracer-dynamic-coverage-feedback
7887 This value is used to limit superblock formation once the given percentage of
7888 executed instructions is covered. This limits unnecessary code size
7891 The @option{tracer-dynamic-coverage-feedback} is used only when profile
7892 feedback is available. The real profiles (as opposed to statically estimated
7893 ones) are much less balanced allowing the threshold to be larger value.
7895 @item tracer-max-code-growth
7896 Stop tail duplication once code growth has reached given percentage. This is
7897 rather hokey argument, as most of the duplicates will be eliminated later in
7898 cross jumping, so it may be set to much higher values than is the desired code
7901 @item tracer-min-branch-ratio
7903 Stop reverse growth when the reverse probability of best edge is less than this
7904 threshold (in percent).
7906 @item tracer-min-branch-ratio
7907 @itemx tracer-min-branch-ratio-feedback
7909 Stop forward growth if the best edge do have probability lower than this
7912 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
7913 compilation for profile feedback and one for compilation without. The value
7914 for compilation with profile feedback needs to be more conservative (higher) in
7915 order to make tracer effective.
7917 @item max-cse-path-length
7919 Maximum number of basic blocks on path that cse considers. The default is 10.
7922 The maximum instructions CSE process before flushing. The default is 1000.
7924 @item ggc-min-expand
7926 GCC uses a garbage collector to manage its own memory allocation. This
7927 parameter specifies the minimum percentage by which the garbage
7928 collector's heap should be allowed to expand between collections.
7929 Tuning this may improve compilation speed; it has no effect on code
7932 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
7933 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
7934 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
7935 GCC is not able to calculate RAM on a particular platform, the lower
7936 bound of 30% is used. Setting this parameter and
7937 @option{ggc-min-heapsize} to zero causes a full collection to occur at
7938 every opportunity. This is extremely slow, but can be useful for
7941 @item ggc-min-heapsize
7943 Minimum size of the garbage collector's heap before it begins bothering
7944 to collect garbage. The first collection occurs after the heap expands
7945 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
7946 tuning this may improve compilation speed, and has no effect on code
7949 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
7950 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
7951 with a lower bound of 4096 (four megabytes) and an upper bound of
7952 131072 (128 megabytes). If GCC is not able to calculate RAM on a
7953 particular platform, the lower bound is used. Setting this parameter
7954 very large effectively disables garbage collection. Setting this
7955 parameter and @option{ggc-min-expand} to zero causes a full collection
7956 to occur at every opportunity.
7958 @item max-reload-search-insns
7959 The maximum number of instruction reload should look backward for equivalent
7960 register. Increasing values mean more aggressive optimization, making the
7961 compile time increase with probably slightly better performance. The default
7964 @item max-cselib-memory-locations
7965 The maximum number of memory locations cselib should take into account.
7966 Increasing values mean more aggressive optimization, making the compile time
7967 increase with probably slightly better performance. The default value is 500.
7969 @item reorder-blocks-duplicate
7970 @itemx reorder-blocks-duplicate-feedback
7972 Used by basic block reordering pass to decide whether to use unconditional
7973 branch or duplicate the code on its destination. Code is duplicated when its
7974 estimated size is smaller than this value multiplied by the estimated size of
7975 unconditional jump in the hot spots of the program.
7977 The @option{reorder-block-duplicate-feedback} is used only when profile
7978 feedback is available and may be set to higher values than
7979 @option{reorder-block-duplicate} since information about the hot spots is more
7982 @item max-sched-ready-insns
7983 The maximum number of instructions ready to be issued the scheduler should
7984 consider at any given time during the first scheduling pass. Increasing
7985 values mean more thorough searches, making the compilation time increase
7986 with probably little benefit. The default value is 100.
7988 @item max-sched-region-blocks
7989 The maximum number of blocks in a region to be considered for
7990 interblock scheduling. The default value is 10.
7992 @item max-pipeline-region-blocks
7993 The maximum number of blocks in a region to be considered for
7994 pipelining in the selective scheduler. The default value is 15.
7996 @item max-sched-region-insns
7997 The maximum number of insns in a region to be considered for
7998 interblock scheduling. The default value is 100.
8000 @item max-pipeline-region-insns
8001 The maximum number of insns in a region to be considered for
8002 pipelining in the selective scheduler. The default value is 200.
8005 The minimum probability (in percents) of reaching a source block
8006 for interblock speculative scheduling. The default value is 40.
8008 @item max-sched-extend-regions-iters
8009 The maximum number of iterations through CFG to extend regions.
8010 0 - disable region extension,
8011 N - do at most N iterations.
8012 The default value is 0.
8014 @item max-sched-insn-conflict-delay
8015 The maximum conflict delay for an insn to be considered for speculative motion.
8016 The default value is 3.
8018 @item sched-spec-prob-cutoff
8019 The minimal probability of speculation success (in percents), so that
8020 speculative insn will be scheduled.
8021 The default value is 40.
8023 @item sched-mem-true-dep-cost
8024 Minimal distance (in CPU cycles) between store and load targeting same
8025 memory locations. The default value is 1.
8027 @item selsched-max-lookahead
8028 The maximum size of the lookahead window of selective scheduling. It is a
8029 depth of search for available instructions.
8030 The default value is 50.
8032 @item selsched-max-sched-times
8033 The maximum number of times that an instruction will be scheduled during
8034 selective scheduling. This is the limit on the number of iterations
8035 through which the instruction may be pipelined. The default value is 2.
8037 @item selsched-max-insns-to-rename
8038 The maximum number of best instructions in the ready list that are considered
8039 for renaming in the selective scheduler. The default value is 2.
8041 @item max-last-value-rtl
8042 The maximum size measured as number of RTLs that can be recorded in an expression
8043 in combiner for a pseudo register as last known value of that register. The default
8046 @item integer-share-limit
8047 Small integer constants can use a shared data structure, reducing the
8048 compiler's memory usage and increasing its speed. This sets the maximum
8049 value of a shared integer constant. The default value is 256.
8051 @item min-virtual-mappings
8052 Specifies the minimum number of virtual mappings in the incremental
8053 SSA updater that should be registered to trigger the virtual mappings
8054 heuristic defined by virtual-mappings-ratio. The default value is
8057 @item virtual-mappings-ratio
8058 If the number of virtual mappings is virtual-mappings-ratio bigger
8059 than the number of virtual symbols to be updated, then the incremental
8060 SSA updater switches to a full update for those symbols. The default
8063 @item ssp-buffer-size
8064 The minimum size of buffers (i.e.@: arrays) that will receive stack smashing
8065 protection when @option{-fstack-protection} is used.
8067 @item max-jump-thread-duplication-stmts
8068 Maximum number of statements allowed in a block that needs to be
8069 duplicated when threading jumps.
8071 @item max-fields-for-field-sensitive
8072 Maximum number of fields in a structure we will treat in
8073 a field sensitive manner during pointer analysis. The default is zero
8074 for -O0, and -O1 and 100 for -Os, -O2, and -O3.
8076 @item prefetch-latency
8077 Estimate on average number of instructions that are executed before
8078 prefetch finishes. The distance we prefetch ahead is proportional
8079 to this constant. Increasing this number may also lead to less
8080 streams being prefetched (see @option{simultaneous-prefetches}).
8082 @item simultaneous-prefetches
8083 Maximum number of prefetches that can run at the same time.
8085 @item l1-cache-line-size
8086 The size of cache line in L1 cache, in bytes.
8089 The size of L1 cache, in kilobytes.
8092 The size of L2 cache, in kilobytes.
8094 @item min-insn-to-prefetch-ratio
8095 The minimum ratio between the number of instructions and the
8096 number of prefetches to enable prefetching in a loop with an
8099 @item prefetch-min-insn-to-mem-ratio
8100 The minimum ratio between the number of instructions and the
8101 number of memory references to enable prefetching in a loop.
8103 @item use-canonical-types
8104 Whether the compiler should use the ``canonical'' type system. By
8105 default, this should always be 1, which uses a more efficient internal
8106 mechanism for comparing types in C++ and Objective-C++. However, if
8107 bugs in the canonical type system are causing compilation failures,
8108 set this value to 0 to disable canonical types.
8110 @item switch-conversion-max-branch-ratio
8111 Switch initialization conversion will refuse to create arrays that are
8112 bigger than @option{switch-conversion-max-branch-ratio} times the number of
8113 branches in the switch.
8115 @item max-partial-antic-length
8116 Maximum length of the partial antic set computed during the tree
8117 partial redundancy elimination optimization (@option{-ftree-pre}) when
8118 optimizing at @option{-O3} and above. For some sorts of source code
8119 the enhanced partial redundancy elimination optimization can run away,
8120 consuming all of the memory available on the host machine. This
8121 parameter sets a limit on the length of the sets that are computed,
8122 which prevents the runaway behavior. Setting a value of 0 for
8123 this parameter will allow an unlimited set length.
8125 @item sccvn-max-scc-size
8126 Maximum size of a strongly connected component (SCC) during SCCVN
8127 processing. If this limit is hit, SCCVN processing for the whole
8128 function will not be done and optimizations depending on it will
8129 be disabled. The default maximum SCC size is 10000.
8131 @item ira-max-loops-num
8132 IRA uses a regional register allocation by default. If a function
8133 contains loops more than number given by the parameter, only at most
8134 given number of the most frequently executed loops will form regions
8135 for the regional register allocation. The default value of the
8138 @item ira-max-conflict-table-size
8139 Although IRA uses a sophisticated algorithm of compression conflict
8140 table, the table can be still big for huge functions. If the conflict
8141 table for a function could be more than size in MB given by the
8142 parameter, the conflict table is not built and faster, simpler, and
8143 lower quality register allocation algorithm will be used. The
8144 algorithm do not use pseudo-register conflicts. The default value of
8145 the parameter is 2000.
8147 @item loop-invariant-max-bbs-in-loop
8148 Loop invariant motion can be very expensive, both in compile time and
8149 in amount of needed compile time memory, with very large loops. Loops
8150 with more basic blocks than this parameter won't have loop invariant
8151 motion optimization performed on them. The default value of the
8152 parameter is 1000 for -O1 and 10000 for -O2 and above.
8154 @item min-nondebug-insn-uid
8155 Use uids starting at this parameter for nondebug insns. The range below
8156 the parameter is reserved exclusively for debug insns created by
8157 @option{-fvar-tracking-assignments}, but debug insns may get
8158 (non-overlapping) uids above it if the reserved range is exhausted.
8160 @item ipa-sra-ptr-growth-factor
8161 IPA-SRA will replace a pointer to an aggregate with one or more new
8162 parameters only when their cumulative size is less or equal to
8163 @option{ipa-sra-ptr-growth-factor} times the size of the original
8169 @node Preprocessor Options
8170 @section Options Controlling the Preprocessor
8171 @cindex preprocessor options
8172 @cindex options, preprocessor
8174 These options control the C preprocessor, which is run on each C source
8175 file before actual compilation.
8177 If you use the @option{-E} option, nothing is done except preprocessing.
8178 Some of these options make sense only together with @option{-E} because
8179 they cause the preprocessor output to be unsuitable for actual
8183 @item -Wp,@var{option}
8185 You can use @option{-Wp,@var{option}} to bypass the compiler driver
8186 and pass @var{option} directly through to the preprocessor. If
8187 @var{option} contains commas, it is split into multiple options at the
8188 commas. However, many options are modified, translated or interpreted
8189 by the compiler driver before being passed to the preprocessor, and
8190 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
8191 interface is undocumented and subject to change, so whenever possible
8192 you should avoid using @option{-Wp} and let the driver handle the
8195 @item -Xpreprocessor @var{option}
8196 @opindex Xpreprocessor
8197 Pass @var{option} as an option to the preprocessor. You can use this to
8198 supply system-specific preprocessor options which GCC does not know how to
8201 If you want to pass an option that takes an argument, you must use
8202 @option{-Xpreprocessor} twice, once for the option and once for the argument.
8205 @include cppopts.texi
8207 @node Assembler Options
8208 @section Passing Options to the Assembler
8210 @c prevent bad page break with this line
8211 You can pass options to the assembler.
8214 @item -Wa,@var{option}
8216 Pass @var{option} as an option to the assembler. If @var{option}
8217 contains commas, it is split into multiple options at the commas.
8219 @item -Xassembler @var{option}
8221 Pass @var{option} as an option to the assembler. You can use this to
8222 supply system-specific assembler options which GCC does not know how to
8225 If you want to pass an option that takes an argument, you must use
8226 @option{-Xassembler} twice, once for the option and once for the argument.
8231 @section Options for Linking
8232 @cindex link options
8233 @cindex options, linking
8235 These options come into play when the compiler links object files into
8236 an executable output file. They are meaningless if the compiler is
8237 not doing a link step.
8241 @item @var{object-file-name}
8242 A file name that does not end in a special recognized suffix is
8243 considered to name an object file or library. (Object files are
8244 distinguished from libraries by the linker according to the file
8245 contents.) If linking is done, these object files are used as input
8254 If any of these options is used, then the linker is not run, and
8255 object file names should not be used as arguments. @xref{Overall
8259 @item -l@var{library}
8260 @itemx -l @var{library}
8262 Search the library named @var{library} when linking. (The second
8263 alternative with the library as a separate argument is only for
8264 POSIX compliance and is not recommended.)
8266 It makes a difference where in the command you write this option; the
8267 linker searches and processes libraries and object files in the order they
8268 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
8269 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
8270 to functions in @samp{z}, those functions may not be loaded.
8272 The linker searches a standard list of directories for the library,
8273 which is actually a file named @file{lib@var{library}.a}. The linker
8274 then uses this file as if it had been specified precisely by name.
8276 The directories searched include several standard system directories
8277 plus any that you specify with @option{-L}.
8279 Normally the files found this way are library files---archive files
8280 whose members are object files. The linker handles an archive file by
8281 scanning through it for members which define symbols that have so far
8282 been referenced but not defined. But if the file that is found is an
8283 ordinary object file, it is linked in the usual fashion. The only
8284 difference between using an @option{-l} option and specifying a file name
8285 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
8286 and searches several directories.
8290 You need this special case of the @option{-l} option in order to
8291 link an Objective-C or Objective-C++ program.
8294 @opindex nostartfiles
8295 Do not use the standard system startup files when linking.
8296 The standard system libraries are used normally, unless @option{-nostdlib}
8297 or @option{-nodefaultlibs} is used.
8299 @item -nodefaultlibs
8300 @opindex nodefaultlibs
8301 Do not use the standard system libraries when linking.
8302 Only the libraries you specify will be passed to the linker, options
8303 specifying linkage of the system libraries, such as @code{-static-libgcc}
8304 or @code{-shared-libgcc}, will be ignored.
8305 The standard startup files are used normally, unless @option{-nostartfiles}
8306 is used. The compiler may generate calls to @code{memcmp},
8307 @code{memset}, @code{memcpy} and @code{memmove}.
8308 These entries are usually resolved by entries in
8309 libc. These entry points should be supplied through some other
8310 mechanism when this option is specified.
8314 Do not use the standard system startup files or libraries when linking.
8315 No startup files and only the libraries you specify will be passed to
8316 the linker, options specifying linkage of the system libraries, such as
8317 @code{-static-libgcc} or @code{-shared-libgcc}, will be ignored.
8318 The compiler may generate calls to @code{memcmp}, @code{memset},
8319 @code{memcpy} and @code{memmove}.
8320 These entries are usually resolved by entries in
8321 libc. These entry points should be supplied through some other
8322 mechanism when this option is specified.
8324 @cindex @option{-lgcc}, use with @option{-nostdlib}
8325 @cindex @option{-nostdlib} and unresolved references
8326 @cindex unresolved references and @option{-nostdlib}
8327 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
8328 @cindex @option{-nodefaultlibs} and unresolved references
8329 @cindex unresolved references and @option{-nodefaultlibs}
8330 One of the standard libraries bypassed by @option{-nostdlib} and
8331 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
8332 that GCC uses to overcome shortcomings of particular machines, or special
8333 needs for some languages.
8334 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
8335 Collection (GCC) Internals},
8336 for more discussion of @file{libgcc.a}.)
8337 In most cases, you need @file{libgcc.a} even when you want to avoid
8338 other standard libraries. In other words, when you specify @option{-nostdlib}
8339 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
8340 This ensures that you have no unresolved references to internal GCC
8341 library subroutines. (For example, @samp{__main}, used to ensure C++
8342 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
8343 GNU Compiler Collection (GCC) Internals}.)
8347 Produce a position independent executable on targets which support it.
8348 For predictable results, you must also specify the same set of options
8349 that were used to generate code (@option{-fpie}, @option{-fPIE},
8350 or model suboptions) when you specify this option.
8354 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
8355 that support it. This instructs the linker to add all symbols, not
8356 only used ones, to the dynamic symbol table. This option is needed
8357 for some uses of @code{dlopen} or to allow obtaining backtraces
8358 from within a program.
8362 Remove all symbol table and relocation information from the executable.
8366 On systems that support dynamic linking, this prevents linking with the shared
8367 libraries. On other systems, this option has no effect.
8371 Produce a shared object which can then be linked with other objects to
8372 form an executable. Not all systems support this option. For predictable
8373 results, you must also specify the same set of options that were used to
8374 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
8375 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
8376 needs to build supplementary stub code for constructors to work. On
8377 multi-libbed systems, @samp{gcc -shared} must select the correct support
8378 libraries to link against. Failing to supply the correct flags may lead
8379 to subtle defects. Supplying them in cases where they are not necessary
8382 @item -shared-libgcc
8383 @itemx -static-libgcc
8384 @opindex shared-libgcc
8385 @opindex static-libgcc
8386 On systems that provide @file{libgcc} as a shared library, these options
8387 force the use of either the shared or static version respectively.
8388 If no shared version of @file{libgcc} was built when the compiler was
8389 configured, these options have no effect.
8391 There are several situations in which an application should use the
8392 shared @file{libgcc} instead of the static version. The most common
8393 of these is when the application wishes to throw and catch exceptions
8394 across different shared libraries. In that case, each of the libraries
8395 as well as the application itself should use the shared @file{libgcc}.
8397 Therefore, the G++ and GCJ drivers automatically add
8398 @option{-shared-libgcc} whenever you build a shared library or a main
8399 executable, because C++ and Java programs typically use exceptions, so
8400 this is the right thing to do.
8402 If, instead, you use the GCC driver to create shared libraries, you may
8403 find that they will not always be linked with the shared @file{libgcc}.
8404 If GCC finds, at its configuration time, that you have a non-GNU linker
8405 or a GNU linker that does not support option @option{--eh-frame-hdr},
8406 it will link the shared version of @file{libgcc} into shared libraries
8407 by default. Otherwise, it will take advantage of the linker and optimize
8408 away the linking with the shared version of @file{libgcc}, linking with
8409 the static version of libgcc by default. This allows exceptions to
8410 propagate through such shared libraries, without incurring relocation
8411 costs at library load time.
8413 However, if a library or main executable is supposed to throw or catch
8414 exceptions, you must link it using the G++ or GCJ driver, as appropriate
8415 for the languages used in the program, or using the option
8416 @option{-shared-libgcc}, such that it is linked with the shared
8419 @item -static-libstdc++
8420 When the @command{g++} program is used to link a C++ program, it will
8421 normally automatically link against @option{libstdc++}. If
8422 @file{libstdc++} is available as a shared library, and the
8423 @option{-static} option is not used, then this will link against the
8424 shared version of @file{libstdc++}. That is normally fine. However, it
8425 is sometimes useful to freeze the version of @file{libstdc++} used by
8426 the program without going all the way to a fully static link. The
8427 @option{-static-libstdc++} option directs the @command{g++} driver to
8428 link @file{libstdc++} statically, without necessarily linking other
8429 libraries statically.
8433 Bind references to global symbols when building a shared object. Warn
8434 about any unresolved references (unless overridden by the link editor
8435 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
8438 @item -T @var{script}
8440 @cindex linker script
8441 Use @var{script} as the linker script. This option is supported by most
8442 systems using the GNU linker. On some targets, such as bare-board
8443 targets without an operating system, the @option{-T} option may be required
8444 when linking to avoid references to undefined symbols.
8446 @item -Xlinker @var{option}
8448 Pass @var{option} as an option to the linker. You can use this to
8449 supply system-specific linker options which GCC does not know how to
8452 If you want to pass an option that takes a separate argument, you must use
8453 @option{-Xlinker} twice, once for the option and once for the argument.
8454 For example, to pass @option{-assert definitions}, you must write
8455 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
8456 @option{-Xlinker "-assert definitions"}, because this passes the entire
8457 string as a single argument, which is not what the linker expects.
8459 When using the GNU linker, it is usually more convenient to pass
8460 arguments to linker options using the @option{@var{option}=@var{value}}
8461 syntax than as separate arguments. For example, you can specify
8462 @samp{-Xlinker -Map=output.map} rather than
8463 @samp{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
8464 this syntax for command-line options.
8466 @item -Wl,@var{option}
8468 Pass @var{option} as an option to the linker. If @var{option} contains
8469 commas, it is split into multiple options at the commas. You can use this
8470 syntax to pass an argument to the option.
8471 For example, @samp{-Wl,-Map,output.map} passes @samp{-Map output.map} to the
8472 linker. When using the GNU linker, you can also get the same effect with
8473 @samp{-Wl,-Map=output.map}.
8475 @item -u @var{symbol}
8477 Pretend the symbol @var{symbol} is undefined, to force linking of
8478 library modules to define it. You can use @option{-u} multiple times with
8479 different symbols to force loading of additional library modules.
8482 @node Directory Options
8483 @section Options for Directory Search
8484 @cindex directory options
8485 @cindex options, directory search
8488 These options specify directories to search for header files, for
8489 libraries and for parts of the compiler:
8494 Add the directory @var{dir} to the head of the list of directories to be
8495 searched for header files. This can be used to override a system header
8496 file, substituting your own version, since these directories are
8497 searched before the system header file directories. However, you should
8498 not use this option to add directories that contain vendor-supplied
8499 system header files (use @option{-isystem} for that). If you use more than
8500 one @option{-I} option, the directories are scanned in left-to-right
8501 order; the standard system directories come after.
8503 If a standard system include directory, or a directory specified with
8504 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
8505 option will be ignored. The directory will still be searched but as a
8506 system directory at its normal position in the system include chain.
8507 This is to ensure that GCC's procedure to fix buggy system headers and
8508 the ordering for the include_next directive are not inadvertently changed.
8509 If you really need to change the search order for system directories,
8510 use the @option{-nostdinc} and/or @option{-isystem} options.
8512 @item -iquote@var{dir}
8514 Add the directory @var{dir} to the head of the list of directories to
8515 be searched for header files only for the case of @samp{#include
8516 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
8517 otherwise just like @option{-I}.
8521 Add directory @var{dir} to the list of directories to be searched
8524 @item -B@var{prefix}
8526 This option specifies where to find the executables, libraries,
8527 include files, and data files of the compiler itself.
8529 The compiler driver program runs one or more of the subprograms
8530 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
8531 @var{prefix} as a prefix for each program it tries to run, both with and
8532 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
8534 For each subprogram to be run, the compiler driver first tries the
8535 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
8536 was not specified, the driver tries two standard prefixes, which are
8537 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
8538 those results in a file name that is found, the unmodified program
8539 name is searched for using the directories specified in your
8540 @env{PATH} environment variable.
8542 The compiler will check to see if the path provided by the @option{-B}
8543 refers to a directory, and if necessary it will add a directory
8544 separator character at the end of the path.
8546 @option{-B} prefixes that effectively specify directory names also apply
8547 to libraries in the linker, because the compiler translates these
8548 options into @option{-L} options for the linker. They also apply to
8549 includes files in the preprocessor, because the compiler translates these
8550 options into @option{-isystem} options for the preprocessor. In this case,
8551 the compiler appends @samp{include} to the prefix.
8553 The run-time support file @file{libgcc.a} can also be searched for using
8554 the @option{-B} prefix, if needed. If it is not found there, the two
8555 standard prefixes above are tried, and that is all. The file is left
8556 out of the link if it is not found by those means.
8558 Another way to specify a prefix much like the @option{-B} prefix is to use
8559 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
8562 As a special kludge, if the path provided by @option{-B} is
8563 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
8564 9, then it will be replaced by @file{[dir/]include}. This is to help
8565 with boot-strapping the compiler.
8567 @item -specs=@var{file}
8569 Process @var{file} after the compiler reads in the standard @file{specs}
8570 file, in order to override the defaults that the @file{gcc} driver
8571 program uses when determining what switches to pass to @file{cc1},
8572 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
8573 @option{-specs=@var{file}} can be specified on the command line, and they
8574 are processed in order, from left to right.
8576 @item --sysroot=@var{dir}
8578 Use @var{dir} as the logical root directory for headers and libraries.
8579 For example, if the compiler would normally search for headers in
8580 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
8581 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
8583 If you use both this option and the @option{-isysroot} option, then
8584 the @option{--sysroot} option will apply to libraries, but the
8585 @option{-isysroot} option will apply to header files.
8587 The GNU linker (beginning with version 2.16) has the necessary support
8588 for this option. If your linker does not support this option, the
8589 header file aspect of @option{--sysroot} will still work, but the
8590 library aspect will not.
8594 This option has been deprecated. Please use @option{-iquote} instead for
8595 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
8596 Any directories you specify with @option{-I} options before the @option{-I-}
8597 option are searched only for the case of @samp{#include "@var{file}"};
8598 they are not searched for @samp{#include <@var{file}>}.
8600 If additional directories are specified with @option{-I} options after
8601 the @option{-I-}, these directories are searched for all @samp{#include}
8602 directives. (Ordinarily @emph{all} @option{-I} directories are used
8605 In addition, the @option{-I-} option inhibits the use of the current
8606 directory (where the current input file came from) as the first search
8607 directory for @samp{#include "@var{file}"}. There is no way to
8608 override this effect of @option{-I-}. With @option{-I.} you can specify
8609 searching the directory which was current when the compiler was
8610 invoked. That is not exactly the same as what the preprocessor does
8611 by default, but it is often satisfactory.
8613 @option{-I-} does not inhibit the use of the standard system directories
8614 for header files. Thus, @option{-I-} and @option{-nostdinc} are
8621 @section Specifying subprocesses and the switches to pass to them
8624 @command{gcc} is a driver program. It performs its job by invoking a
8625 sequence of other programs to do the work of compiling, assembling and
8626 linking. GCC interprets its command-line parameters and uses these to
8627 deduce which programs it should invoke, and which command-line options
8628 it ought to place on their command lines. This behavior is controlled
8629 by @dfn{spec strings}. In most cases there is one spec string for each
8630 program that GCC can invoke, but a few programs have multiple spec
8631 strings to control their behavior. The spec strings built into GCC can
8632 be overridden by using the @option{-specs=} command-line switch to specify
8635 @dfn{Spec files} are plaintext files that are used to construct spec
8636 strings. They consist of a sequence of directives separated by blank
8637 lines. The type of directive is determined by the first non-whitespace
8638 character on the line and it can be one of the following:
8641 @item %@var{command}
8642 Issues a @var{command} to the spec file processor. The commands that can
8646 @item %include <@var{file}>
8648 Search for @var{file} and insert its text at the current point in the
8651 @item %include_noerr <@var{file}>
8652 @cindex %include_noerr
8653 Just like @samp{%include}, but do not generate an error message if the include
8654 file cannot be found.
8656 @item %rename @var{old_name} @var{new_name}
8658 Rename the spec string @var{old_name} to @var{new_name}.
8662 @item *[@var{spec_name}]:
8663 This tells the compiler to create, override or delete the named spec
8664 string. All lines after this directive up to the next directive or
8665 blank line are considered to be the text for the spec string. If this
8666 results in an empty string then the spec will be deleted. (Or, if the
8667 spec did not exist, then nothing will happened.) Otherwise, if the spec
8668 does not currently exist a new spec will be created. If the spec does
8669 exist then its contents will be overridden by the text of this
8670 directive, unless the first character of that text is the @samp{+}
8671 character, in which case the text will be appended to the spec.
8673 @item [@var{suffix}]:
8674 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
8675 and up to the next directive or blank line are considered to make up the
8676 spec string for the indicated suffix. When the compiler encounters an
8677 input file with the named suffix, it will processes the spec string in
8678 order to work out how to compile that file. For example:
8685 This says that any input file whose name ends in @samp{.ZZ} should be
8686 passed to the program @samp{z-compile}, which should be invoked with the
8687 command-line switch @option{-input} and with the result of performing the
8688 @samp{%i} substitution. (See below.)
8690 As an alternative to providing a spec string, the text that follows a
8691 suffix directive can be one of the following:
8694 @item @@@var{language}
8695 This says that the suffix is an alias for a known @var{language}. This is
8696 similar to using the @option{-x} command-line switch to GCC to specify a
8697 language explicitly. For example:
8704 Says that .ZZ files are, in fact, C++ source files.
8707 This causes an error messages saying:
8710 @var{name} compiler not installed on this system.
8714 GCC already has an extensive list of suffixes built into it.
8715 This directive will add an entry to the end of the list of suffixes, but
8716 since the list is searched from the end backwards, it is effectively
8717 possible to override earlier entries using this technique.
8721 GCC has the following spec strings built into it. Spec files can
8722 override these strings or create their own. Note that individual
8723 targets can also add their own spec strings to this list.
8726 asm Options to pass to the assembler
8727 asm_final Options to pass to the assembler post-processor
8728 cpp Options to pass to the C preprocessor
8729 cc1 Options to pass to the C compiler
8730 cc1plus Options to pass to the C++ compiler
8731 endfile Object files to include at the end of the link
8732 link Options to pass to the linker
8733 lib Libraries to include on the command line to the linker
8734 libgcc Decides which GCC support library to pass to the linker
8735 linker Sets the name of the linker
8736 predefines Defines to be passed to the C preprocessor
8737 signed_char Defines to pass to CPP to say whether @code{char} is signed
8739 startfile Object files to include at the start of the link
8742 Here is a small example of a spec file:
8748 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
8751 This example renames the spec called @samp{lib} to @samp{old_lib} and
8752 then overrides the previous definition of @samp{lib} with a new one.
8753 The new definition adds in some extra command-line options before
8754 including the text of the old definition.
8756 @dfn{Spec strings} are a list of command-line options to be passed to their
8757 corresponding program. In addition, the spec strings can contain
8758 @samp{%}-prefixed sequences to substitute variable text or to
8759 conditionally insert text into the command line. Using these constructs
8760 it is possible to generate quite complex command lines.
8762 Here is a table of all defined @samp{%}-sequences for spec
8763 strings. Note that spaces are not generated automatically around the
8764 results of expanding these sequences. Therefore you can concatenate them
8765 together or combine them with constant text in a single argument.
8769 Substitute one @samp{%} into the program name or argument.
8772 Substitute the name of the input file being processed.
8775 Substitute the basename of the input file being processed.
8776 This is the substring up to (and not including) the last period
8777 and not including the directory.
8780 This is the same as @samp{%b}, but include the file suffix (text after
8784 Marks the argument containing or following the @samp{%d} as a
8785 temporary file name, so that that file will be deleted if GCC exits
8786 successfully. Unlike @samp{%g}, this contributes no text to the
8789 @item %g@var{suffix}
8790 Substitute a file name that has suffix @var{suffix} and is chosen
8791 once per compilation, and mark the argument in the same way as
8792 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
8793 name is now chosen in a way that is hard to predict even when previously
8794 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
8795 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
8796 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
8797 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
8798 was simply substituted with a file name chosen once per compilation,
8799 without regard to any appended suffix (which was therefore treated
8800 just like ordinary text), making such attacks more likely to succeed.
8802 @item %u@var{suffix}
8803 Like @samp{%g}, but generates a new temporary file name even if
8804 @samp{%u@var{suffix}} was already seen.
8806 @item %U@var{suffix}
8807 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
8808 new one if there is no such last file name. In the absence of any
8809 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
8810 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
8811 would involve the generation of two distinct file names, one
8812 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
8813 simply substituted with a file name chosen for the previous @samp{%u},
8814 without regard to any appended suffix.
8816 @item %j@var{suffix}
8817 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
8818 writable, and if save-temps is off; otherwise, substitute the name
8819 of a temporary file, just like @samp{%u}. This temporary file is not
8820 meant for communication between processes, but rather as a junk
8823 @item %|@var{suffix}
8824 @itemx %m@var{suffix}
8825 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
8826 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
8827 all. These are the two most common ways to instruct a program that it
8828 should read from standard input or write to standard output. If you
8829 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
8830 construct: see for example @file{f/lang-specs.h}.
8832 @item %.@var{SUFFIX}
8833 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
8834 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
8835 terminated by the next space or %.
8838 Marks the argument containing or following the @samp{%w} as the
8839 designated output file of this compilation. This puts the argument
8840 into the sequence of arguments that @samp{%o} will substitute later.
8843 Substitutes the names of all the output files, with spaces
8844 automatically placed around them. You should write spaces
8845 around the @samp{%o} as well or the results are undefined.
8846 @samp{%o} is for use in the specs for running the linker.
8847 Input files whose names have no recognized suffix are not compiled
8848 at all, but they are included among the output files, so they will
8852 Substitutes the suffix for object files. Note that this is
8853 handled specially when it immediately follows @samp{%g, %u, or %U},
8854 because of the need for those to form complete file names. The
8855 handling is such that @samp{%O} is treated exactly as if it had already
8856 been substituted, except that @samp{%g, %u, and %U} do not currently
8857 support additional @var{suffix} characters following @samp{%O} as they would
8858 following, for example, @samp{.o}.
8861 Substitutes the standard macro predefinitions for the
8862 current target machine. Use this when running @code{cpp}.
8865 Like @samp{%p}, but puts @samp{__} before and after the name of each
8866 predefined macro, except for macros that start with @samp{__} or with
8867 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
8871 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
8872 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
8873 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
8874 and @option{-imultilib} as necessary.
8877 Current argument is the name of a library or startup file of some sort.
8878 Search for that file in a standard list of directories and substitute
8879 the full name found. The current working directory is included in the
8880 list of directories scanned.
8883 Current argument is the name of a linker script. Search for that file
8884 in the current list of directories to scan for libraries. If the file
8885 is located insert a @option{--script} option into the command line
8886 followed by the full path name found. If the file is not found then
8887 generate an error message. Note: the current working directory is not
8891 Print @var{str} as an error message. @var{str} is terminated by a newline.
8892 Use this when inconsistent options are detected.
8895 Substitute the contents of spec string @var{name} at this point.
8898 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
8900 @item %x@{@var{option}@}
8901 Accumulate an option for @samp{%X}.
8904 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
8908 Output the accumulated assembler options specified by @option{-Wa}.
8911 Output the accumulated preprocessor options specified by @option{-Wp}.
8914 Process the @code{asm} spec. This is used to compute the
8915 switches to be passed to the assembler.
8918 Process the @code{asm_final} spec. This is a spec string for
8919 passing switches to an assembler post-processor, if such a program is
8923 Process the @code{link} spec. This is the spec for computing the
8924 command line passed to the linker. Typically it will make use of the
8925 @samp{%L %G %S %D and %E} sequences.
8928 Dump out a @option{-L} option for each directory that GCC believes might
8929 contain startup files. If the target supports multilibs then the
8930 current multilib directory will be prepended to each of these paths.
8933 Process the @code{lib} spec. This is a spec string for deciding which
8934 libraries should be included on the command line to the linker.
8937 Process the @code{libgcc} spec. This is a spec string for deciding
8938 which GCC support library should be included on the command line to the linker.
8941 Process the @code{startfile} spec. This is a spec for deciding which
8942 object files should be the first ones passed to the linker. Typically
8943 this might be a file named @file{crt0.o}.
8946 Process the @code{endfile} spec. This is a spec string that specifies
8947 the last object files that will be passed to the linker.
8950 Process the @code{cpp} spec. This is used to construct the arguments
8951 to be passed to the C preprocessor.
8954 Process the @code{cc1} spec. This is used to construct the options to be
8955 passed to the actual C compiler (@samp{cc1}).
8958 Process the @code{cc1plus} spec. This is used to construct the options to be
8959 passed to the actual C++ compiler (@samp{cc1plus}).
8962 Substitute the variable part of a matched option. See below.
8963 Note that each comma in the substituted string is replaced by
8967 Remove all occurrences of @code{-S} from the command line. Note---this
8968 command is position dependent. @samp{%} commands in the spec string
8969 before this one will see @code{-S}, @samp{%} commands in the spec string
8970 after this one will not.
8972 @item %:@var{function}(@var{args})
8973 Call the named function @var{function}, passing it @var{args}.
8974 @var{args} is first processed as a nested spec string, then split
8975 into an argument vector in the usual fashion. The function returns
8976 a string which is processed as if it had appeared literally as part
8977 of the current spec.
8979 The following built-in spec functions are provided:
8983 The @code{getenv} spec function takes two arguments: an environment
8984 variable name and a string. If the environment variable is not
8985 defined, a fatal error is issued. Otherwise, the return value is the
8986 value of the environment variable concatenated with the string. For
8987 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
8990 %:getenv(TOPDIR /include)
8993 expands to @file{/path/to/top/include}.
8995 @item @code{if-exists}
8996 The @code{if-exists} spec function takes one argument, an absolute
8997 pathname to a file. If the file exists, @code{if-exists} returns the
8998 pathname. Here is a small example of its usage:
9002 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
9005 @item @code{if-exists-else}
9006 The @code{if-exists-else} spec function is similar to the @code{if-exists}
9007 spec function, except that it takes two arguments. The first argument is
9008 an absolute pathname to a file. If the file exists, @code{if-exists-else}
9009 returns the pathname. If it does not exist, it returns the second argument.
9010 This way, @code{if-exists-else} can be used to select one file or another,
9011 based on the existence of the first. Here is a small example of its usage:
9015 crt0%O%s %:if-exists(crti%O%s) \
9016 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
9019 @item @code{replace-outfile}
9020 The @code{replace-outfile} spec function takes two arguments. It looks for the
9021 first argument in the outfiles array and replaces it with the second argument. Here
9022 is a small example of its usage:
9025 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
9028 @item @code{print-asm-header}
9029 The @code{print-asm-header} function takes no arguments and simply
9030 prints a banner like:
9036 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
9039 It is used to separate compiler options from assembler options
9040 in the @option{--target-help} output.
9044 Substitutes the @code{-S} switch, if that switch was given to GCC@.
9045 If that switch was not specified, this substitutes nothing. Note that
9046 the leading dash is omitted when specifying this option, and it is
9047 automatically inserted if the substitution is performed. Thus the spec
9048 string @samp{%@{foo@}} would match the command-line option @option{-foo}
9049 and would output the command line option @option{-foo}.
9051 @item %W@{@code{S}@}
9052 Like %@{@code{S}@} but mark last argument supplied within as a file to be
9055 @item %@{@code{S}*@}
9056 Substitutes all the switches specified to GCC whose names start
9057 with @code{-S}, but which also take an argument. This is used for
9058 switches like @option{-o}, @option{-D}, @option{-I}, etc.
9059 GCC considers @option{-o foo} as being
9060 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
9061 text, including the space. Thus two arguments would be generated.
9063 @item %@{@code{S}*&@code{T}*@}
9064 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
9065 (the order of @code{S} and @code{T} in the spec is not significant).
9066 There can be any number of ampersand-separated variables; for each the
9067 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
9069 @item %@{@code{S}:@code{X}@}
9070 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
9072 @item %@{!@code{S}:@code{X}@}
9073 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
9075 @item %@{@code{S}*:@code{X}@}
9076 Substitutes @code{X} if one or more switches whose names start with
9077 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
9078 once, no matter how many such switches appeared. However, if @code{%*}
9079 appears somewhere in @code{X}, then @code{X} will be substituted once
9080 for each matching switch, with the @code{%*} replaced by the part of
9081 that switch that matched the @code{*}.
9083 @item %@{.@code{S}:@code{X}@}
9084 Substitutes @code{X}, if processing a file with suffix @code{S}.
9086 @item %@{!.@code{S}:@code{X}@}
9087 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
9089 @item %@{,@code{S}:@code{X}@}
9090 Substitutes @code{X}, if processing a file for language @code{S}.
9092 @item %@{!,@code{S}:@code{X}@}
9093 Substitutes @code{X}, if not processing a file for language @code{S}.
9095 @item %@{@code{S}|@code{P}:@code{X}@}
9096 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
9097 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
9098 @code{*} sequences as well, although they have a stronger binding than
9099 the @samp{|}. If @code{%*} appears in @code{X}, all of the
9100 alternatives must be starred, and only the first matching alternative
9103 For example, a spec string like this:
9106 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
9109 will output the following command-line options from the following input
9110 command-line options:
9115 -d fred.c -foo -baz -boggle
9116 -d jim.d -bar -baz -boggle
9119 @item %@{S:X; T:Y; :D@}
9121 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
9122 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
9123 be as many clauses as you need. This may be combined with @code{.},
9124 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
9129 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
9130 construct may contain other nested @samp{%} constructs or spaces, or
9131 even newlines. They are processed as usual, as described above.
9132 Trailing white space in @code{X} is ignored. White space may also
9133 appear anywhere on the left side of the colon in these constructs,
9134 except between @code{.} or @code{*} and the corresponding word.
9136 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
9137 handled specifically in these constructs. If another value of
9138 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
9139 @option{-W} switch is found later in the command line, the earlier
9140 switch value is ignored, except with @{@code{S}*@} where @code{S} is
9141 just one letter, which passes all matching options.
9143 The character @samp{|} at the beginning of the predicate text is used to
9144 indicate that a command should be piped to the following command, but
9145 only if @option{-pipe} is specified.
9147 It is built into GCC which switches take arguments and which do not.
9148 (You might think it would be useful to generalize this to allow each
9149 compiler's spec to say which switches take arguments. But this cannot
9150 be done in a consistent fashion. GCC cannot even decide which input
9151 files have been specified without knowing which switches take arguments,
9152 and it must know which input files to compile in order to tell which
9155 GCC also knows implicitly that arguments starting in @option{-l} are to be
9156 treated as compiler output files, and passed to the linker in their
9157 proper position among the other output files.
9159 @c man begin OPTIONS
9161 @node Target Options
9162 @section Specifying Target Machine and Compiler Version
9163 @cindex target options
9164 @cindex cross compiling
9165 @cindex specifying machine version
9166 @cindex specifying compiler version and target machine
9167 @cindex compiler version, specifying
9168 @cindex target machine, specifying
9170 The usual way to run GCC is to run the executable called @file{gcc}, or
9171 @file{<machine>-gcc} when cross-compiling, or
9172 @file{<machine>-gcc-<version>} to run a version other than the one that
9173 was installed last. Sometimes this is inconvenient, so GCC provides
9174 options that will switch to another cross-compiler or version.
9177 @item -b @var{machine}
9179 The argument @var{machine} specifies the target machine for compilation.
9181 The value to use for @var{machine} is the same as was specified as the
9182 machine type when configuring GCC as a cross-compiler. For
9183 example, if a cross-compiler was configured with @samp{configure
9184 arm-elf}, meaning to compile for an arm processor with elf binaries,
9185 then you would specify @option{-b arm-elf} to run that cross compiler.
9186 Because there are other options beginning with @option{-b}, the
9187 configuration must contain a hyphen, or @option{-b} alone should be one
9188 argument followed by the configuration in the next argument.
9190 @item -V @var{version}
9192 The argument @var{version} specifies which version of GCC to run.
9193 This is useful when multiple versions are installed. For example,
9194 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
9197 The @option{-V} and @option{-b} options work by running the
9198 @file{<machine>-gcc-<version>} executable, so there's no real reason to
9199 use them if you can just run that directly.
9201 @node Submodel Options
9202 @section Hardware Models and Configurations
9203 @cindex submodel options
9204 @cindex specifying hardware config
9205 @cindex hardware models and configurations, specifying
9206 @cindex machine dependent options
9208 Earlier we discussed the standard option @option{-b} which chooses among
9209 different installed compilers for completely different target
9210 machines, such as VAX vs.@: 68000 vs.@: 80386.
9212 In addition, each of these target machine types can have its own
9213 special options, starting with @samp{-m}, to choose among various
9214 hardware models or configurations---for example, 68010 vs 68020,
9215 floating coprocessor or none. A single installed version of the
9216 compiler can compile for any model or configuration, according to the
9219 Some configurations of the compiler also support additional special
9220 options, usually for compatibility with other compilers on the same
9223 @c This list is ordered alphanumerically by subsection name.
9224 @c It should be the same order and spelling as these options are listed
9225 @c in Machine Dependent Options
9231 * Blackfin Options::
9235 * DEC Alpha Options::
9236 * DEC Alpha/VMS Options::
9239 * GNU/Linux Options::
9242 * i386 and x86-64 Options::
9243 * i386 and x86-64 Windows Options::
9245 * IA-64/VMS Options::
9256 * picoChip Options::
9258 * RS/6000 and PowerPC Options::
9259 * S/390 and zSeries Options::
9264 * System V Options::
9269 * Xstormy16 Options::
9275 @subsection ARC Options
9278 These options are defined for ARC implementations:
9283 Compile code for little endian mode. This is the default.
9287 Compile code for big endian mode.
9290 @opindex mmangle-cpu
9291 Prepend the name of the cpu to all public symbol names.
9292 In multiple-processor systems, there are many ARC variants with different
9293 instruction and register set characteristics. This flag prevents code
9294 compiled for one cpu to be linked with code compiled for another.
9295 No facility exists for handling variants that are ``almost identical''.
9296 This is an all or nothing option.
9298 @item -mcpu=@var{cpu}
9300 Compile code for ARC variant @var{cpu}.
9301 Which variants are supported depend on the configuration.
9302 All variants support @option{-mcpu=base}, this is the default.
9304 @item -mtext=@var{text-section}
9305 @itemx -mdata=@var{data-section}
9306 @itemx -mrodata=@var{readonly-data-section}
9310 Put functions, data, and readonly data in @var{text-section},
9311 @var{data-section}, and @var{readonly-data-section} respectively
9312 by default. This can be overridden with the @code{section} attribute.
9313 @xref{Variable Attributes}.
9315 @item -mfix-cortex-m3-ldrd
9316 @opindex mfix-cortex-m3-ldrd
9317 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
9318 with overlapping destination and base registers are used. This option avoids
9319 generating these instructions. This option is enabled by default when
9320 @option{-mcpu=cortex-m3} is specified.
9325 @subsection ARM Options
9328 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
9332 @item -mabi=@var{name}
9334 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
9335 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
9338 @opindex mapcs-frame
9339 Generate a stack frame that is compliant with the ARM Procedure Call
9340 Standard for all functions, even if this is not strictly necessary for
9341 correct execution of the code. Specifying @option{-fomit-frame-pointer}
9342 with this option will cause the stack frames not to be generated for
9343 leaf functions. The default is @option{-mno-apcs-frame}.
9347 This is a synonym for @option{-mapcs-frame}.
9350 @c not currently implemented
9351 @item -mapcs-stack-check
9352 @opindex mapcs-stack-check
9353 Generate code to check the amount of stack space available upon entry to
9354 every function (that actually uses some stack space). If there is
9355 insufficient space available then either the function
9356 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
9357 called, depending upon the amount of stack space required. The run time
9358 system is required to provide these functions. The default is
9359 @option{-mno-apcs-stack-check}, since this produces smaller code.
9361 @c not currently implemented
9363 @opindex mapcs-float
9364 Pass floating point arguments using the float point registers. This is
9365 one of the variants of the APCS@. This option is recommended if the
9366 target hardware has a floating point unit or if a lot of floating point
9367 arithmetic is going to be performed by the code. The default is
9368 @option{-mno-apcs-float}, since integer only code is slightly increased in
9369 size if @option{-mapcs-float} is used.
9371 @c not currently implemented
9372 @item -mapcs-reentrant
9373 @opindex mapcs-reentrant
9374 Generate reentrant, position independent code. The default is
9375 @option{-mno-apcs-reentrant}.
9378 @item -mthumb-interwork
9379 @opindex mthumb-interwork
9380 Generate code which supports calling between the ARM and Thumb
9381 instruction sets. Without this option the two instruction sets cannot
9382 be reliably used inside one program. The default is
9383 @option{-mno-thumb-interwork}, since slightly larger code is generated
9384 when @option{-mthumb-interwork} is specified.
9386 @item -mno-sched-prolog
9387 @opindex mno-sched-prolog
9388 Prevent the reordering of instructions in the function prolog, or the
9389 merging of those instruction with the instructions in the function's
9390 body. This means that all functions will start with a recognizable set
9391 of instructions (or in fact one of a choice from a small set of
9392 different function prologues), and this information can be used to
9393 locate the start if functions inside an executable piece of code. The
9394 default is @option{-msched-prolog}.
9396 @item -mfloat-abi=@var{name}
9398 Specifies which floating-point ABI to use. Permissible values
9399 are: @samp{soft}, @samp{softfp} and @samp{hard}.
9401 Specifying @samp{soft} causes GCC to generate output containing
9402 library calls for floating-point operations.
9403 @samp{softfp} allows the generation of code using hardware floating-point
9404 instructions, but still uses the soft-float calling conventions.
9405 @samp{hard} allows generation of floating-point instructions
9406 and uses FPU-specific calling conventions.
9408 The default depends on the specific target configuration. Note that
9409 the hard-float and soft-float ABIs are not link-compatible; you must
9410 compile your entire program with the same ABI, and link with a
9411 compatible set of libraries.
9414 @opindex mhard-float
9415 Equivalent to @option{-mfloat-abi=hard}.
9418 @opindex msoft-float
9419 Equivalent to @option{-mfloat-abi=soft}.
9421 @item -mlittle-endian
9422 @opindex mlittle-endian
9423 Generate code for a processor running in little-endian mode. This is
9424 the default for all standard configurations.
9427 @opindex mbig-endian
9428 Generate code for a processor running in big-endian mode; the default is
9429 to compile code for a little-endian processor.
9431 @item -mwords-little-endian
9432 @opindex mwords-little-endian
9433 This option only applies when generating code for big-endian processors.
9434 Generate code for a little-endian word order but a big-endian byte
9435 order. That is, a byte order of the form @samp{32107654}. Note: this
9436 option should only be used if you require compatibility with code for
9437 big-endian ARM processors generated by versions of the compiler prior to
9440 @item -mcpu=@var{name}
9442 This specifies the name of the target ARM processor. GCC uses this name
9443 to determine what kind of instructions it can emit when generating
9444 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
9445 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
9446 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
9447 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
9448 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
9450 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
9451 @samp{arm710t}, @samp{arm720t}, @samp{arm740t},
9452 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
9453 @samp{strongarm1110},
9454 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
9455 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
9456 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
9457 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
9458 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
9459 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
9460 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
9461 @samp{cortex-a8}, @samp{cortex-a9},
9462 @samp{cortex-r4}, @samp{cortex-r4f}, @samp{cortex-m3},
9465 @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9467 @item -mtune=@var{name}
9469 This option is very similar to the @option{-mcpu=} option, except that
9470 instead of specifying the actual target processor type, and hence
9471 restricting which instructions can be used, it specifies that GCC should
9472 tune the performance of the code as if the target were of the type
9473 specified in this option, but still choosing the instructions that it
9474 will generate based on the cpu specified by a @option{-mcpu=} option.
9475 For some ARM implementations better performance can be obtained by using
9478 @item -march=@var{name}
9480 This specifies the name of the target ARM architecture. GCC uses this
9481 name to determine what kind of instructions it can emit when generating
9482 assembly code. This option can be used in conjunction with or instead
9483 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
9484 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
9485 @samp{armv5}, @samp{armv5t}, @samp{armv5e}, @samp{armv5te},
9486 @samp{armv6}, @samp{armv6j},
9487 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv6-m},
9488 @samp{armv7}, @samp{armv7-a}, @samp{armv7-r}, @samp{armv7-m},
9489 @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9491 @item -mfpu=@var{name}
9492 @itemx -mfpe=@var{number}
9493 @itemx -mfp=@var{number}
9497 This specifies what floating point hardware (or hardware emulation) is
9498 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
9499 @samp{fpe3}, @samp{maverick}, @samp{vfp}, @samp{vfpv3}, @samp{vfpv3-d16},
9500 @samp{neon}, and @samp{neon-fp16}. @option{-mfp} and @option{-mfpe}
9501 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
9502 with older versions of GCC@.
9504 If @option{-msoft-float} is specified this specifies the format of
9505 floating point values.
9507 @item -mfp16-format=@var{name}
9508 @opindex mfp16-format
9509 Specify the format of the @code{__fp16} half-precision floating-point type.
9510 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
9511 the default is @samp{none}, in which case the @code{__fp16} type is not
9512 defined. @xref{Half-Precision}, for more information.
9514 @item -mstructure-size-boundary=@var{n}
9515 @opindex mstructure-size-boundary
9516 The size of all structures and unions will be rounded up to a multiple
9517 of the number of bits set by this option. Permissible values are 8, 32
9518 and 64. The default value varies for different toolchains. For the COFF
9519 targeted toolchain the default value is 8. A value of 64 is only allowed
9520 if the underlying ABI supports it.
9522 Specifying the larger number can produce faster, more efficient code, but
9523 can also increase the size of the program. Different values are potentially
9524 incompatible. Code compiled with one value cannot necessarily expect to
9525 work with code or libraries compiled with another value, if they exchange
9526 information using structures or unions.
9528 @item -mabort-on-noreturn
9529 @opindex mabort-on-noreturn
9530 Generate a call to the function @code{abort} at the end of a
9531 @code{noreturn} function. It will be executed if the function tries to
9535 @itemx -mno-long-calls
9536 @opindex mlong-calls
9537 @opindex mno-long-calls
9538 Tells the compiler to perform function calls by first loading the
9539 address of the function into a register and then performing a subroutine
9540 call on this register. This switch is needed if the target function
9541 will lie outside of the 64 megabyte addressing range of the offset based
9542 version of subroutine call instruction.
9544 Even if this switch is enabled, not all function calls will be turned
9545 into long calls. The heuristic is that static functions, functions
9546 which have the @samp{short-call} attribute, functions that are inside
9547 the scope of a @samp{#pragma no_long_calls} directive and functions whose
9548 definitions have already been compiled within the current compilation
9549 unit, will not be turned into long calls. The exception to this rule is
9550 that weak function definitions, functions with the @samp{long-call}
9551 attribute or the @samp{section} attribute, and functions that are within
9552 the scope of a @samp{#pragma long_calls} directive, will always be
9553 turned into long calls.
9555 This feature is not enabled by default. Specifying
9556 @option{-mno-long-calls} will restore the default behavior, as will
9557 placing the function calls within the scope of a @samp{#pragma
9558 long_calls_off} directive. Note these switches have no effect on how
9559 the compiler generates code to handle function calls via function
9562 @item -msingle-pic-base
9563 @opindex msingle-pic-base
9564 Treat the register used for PIC addressing as read-only, rather than
9565 loading it in the prologue for each function. The run-time system is
9566 responsible for initializing this register with an appropriate value
9567 before execution begins.
9569 @item -mpic-register=@var{reg}
9570 @opindex mpic-register
9571 Specify the register to be used for PIC addressing. The default is R10
9572 unless stack-checking is enabled, when R9 is used.
9574 @item -mcirrus-fix-invalid-insns
9575 @opindex mcirrus-fix-invalid-insns
9576 @opindex mno-cirrus-fix-invalid-insns
9577 Insert NOPs into the instruction stream to in order to work around
9578 problems with invalid Maverick instruction combinations. This option
9579 is only valid if the @option{-mcpu=ep9312} option has been used to
9580 enable generation of instructions for the Cirrus Maverick floating
9581 point co-processor. This option is not enabled by default, since the
9582 problem is only present in older Maverick implementations. The default
9583 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
9586 @item -mpoke-function-name
9587 @opindex mpoke-function-name
9588 Write the name of each function into the text section, directly
9589 preceding the function prologue. The generated code is similar to this:
9593 .ascii "arm_poke_function_name", 0
9596 .word 0xff000000 + (t1 - t0)
9597 arm_poke_function_name
9599 stmfd sp!, @{fp, ip, lr, pc@}
9603 When performing a stack backtrace, code can inspect the value of
9604 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
9605 location @code{pc - 12} and the top 8 bits are set, then we know that
9606 there is a function name embedded immediately preceding this location
9607 and has length @code{((pc[-3]) & 0xff000000)}.
9611 Generate code for the Thumb instruction set. The default is to
9612 use the 32-bit ARM instruction set.
9613 This option automatically enables either 16-bit Thumb-1 or
9614 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
9615 and @option{-march=@var{name}} options. This option is not passed to the
9616 assembler. If you want to force assembler files to be interpreted as Thumb code,
9617 either add a @samp{.thumb} directive to the source or pass the @option{-mthumb}
9618 option directly to the assembler by prefixing it with @option{-Wa}.
9621 @opindex mtpcs-frame
9622 Generate a stack frame that is compliant with the Thumb Procedure Call
9623 Standard for all non-leaf functions. (A leaf function is one that does
9624 not call any other functions.) The default is @option{-mno-tpcs-frame}.
9626 @item -mtpcs-leaf-frame
9627 @opindex mtpcs-leaf-frame
9628 Generate a stack frame that is compliant with the Thumb Procedure Call
9629 Standard for all leaf functions. (A leaf function is one that does
9630 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
9632 @item -mcallee-super-interworking
9633 @opindex mcallee-super-interworking
9634 Gives all externally visible functions in the file being compiled an ARM
9635 instruction set header which switches to Thumb mode before executing the
9636 rest of the function. This allows these functions to be called from
9637 non-interworking code. This option is not valid in AAPCS configurations
9638 because interworking is enabled by default.
9640 @item -mcaller-super-interworking
9641 @opindex mcaller-super-interworking
9642 Allows calls via function pointers (including virtual functions) to
9643 execute correctly regardless of whether the target code has been
9644 compiled for interworking or not. There is a small overhead in the cost
9645 of executing a function pointer if this option is enabled. This option
9646 is not valid in AAPCS configurations because interworking is enabled
9649 @item -mtp=@var{name}
9651 Specify the access model for the thread local storage pointer. The valid
9652 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
9653 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
9654 (supported in the arm6k architecture), and @option{auto}, which uses the
9655 best available method for the selected processor. The default setting is
9658 @item -mword-relocations
9659 @opindex mword-relocations
9660 Only generate absolute relocations on word sized values (i.e. R_ARM_ABS32).
9661 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
9662 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
9668 @subsection AVR Options
9671 These options are defined for AVR implementations:
9674 @item -mmcu=@var{mcu}
9676 Specify ATMEL AVR instruction set or MCU type.
9678 Instruction set avr1 is for the minimal AVR core, not supported by the C
9679 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
9680 attiny11, attiny12, attiny15, attiny28).
9682 Instruction set avr2 (default) is for the classic AVR core with up to
9683 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
9684 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
9685 at90c8534, at90s8535).
9687 Instruction set avr3 is for the classic AVR core with up to 128K program
9688 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
9690 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9691 memory space (MCU types: atmega8, atmega83, atmega85).
9693 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9694 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9695 atmega64, atmega128, at43usb355, at94k).
9699 Output instruction sizes to the asm file.
9701 @item -mno-interrupts
9702 @opindex mno-interrupts
9703 Generated code is not compatible with hardware interrupts.
9704 Code size will be smaller.
9706 @item -mcall-prologues
9707 @opindex mcall-prologues
9708 Functions prologues/epilogues expanded as call to appropriate
9709 subroutines. Code size will be smaller.
9712 @opindex mtiny-stack
9713 Change only the low 8 bits of the stack pointer.
9717 Assume int to be 8 bit integer. This affects the sizes of all types: A
9718 char will be 1 byte, an int will be 1 byte, a long will be 2 bytes
9719 and long long will be 4 bytes. Please note that this option does not
9720 comply to the C standards, but it will provide you with smaller code
9724 @node Blackfin Options
9725 @subsection Blackfin Options
9726 @cindex Blackfin Options
9729 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
9731 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
9732 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
9733 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
9734 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
9735 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
9736 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
9737 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
9739 The optional @var{sirevision} specifies the silicon revision of the target
9740 Blackfin processor. Any workarounds available for the targeted silicon revision
9741 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
9742 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
9743 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
9744 hexadecimal digits representing the major and minor numbers in the silicon
9745 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
9746 is not defined. If @var{sirevision} is @samp{any}, the
9747 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
9748 If this optional @var{sirevision} is not used, GCC assumes the latest known
9749 silicon revision of the targeted Blackfin processor.
9751 Support for @samp{bf561} is incomplete. For @samp{bf561},
9752 Only the processor macro is defined.
9753 Without this option, @samp{bf532} is used as the processor by default.
9754 The corresponding predefined processor macros for @var{cpu} is to
9755 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
9756 provided by libgloss to be linked in if @option{-msim} is not given.
9760 Specifies that the program will be run on the simulator. This causes
9761 the simulator BSP provided by libgloss to be linked in. This option
9762 has effect only for @samp{bfin-elf} toolchain.
9763 Certain other options, such as @option{-mid-shared-library} and
9764 @option{-mfdpic}, imply @option{-msim}.
9766 @item -momit-leaf-frame-pointer
9767 @opindex momit-leaf-frame-pointer
9768 Don't keep the frame pointer in a register for leaf functions. This
9769 avoids the instructions to save, set up and restore frame pointers and
9770 makes an extra register available in leaf functions. The option
9771 @option{-fomit-frame-pointer} removes the frame pointer for all functions
9772 which might make debugging harder.
9774 @item -mspecld-anomaly
9775 @opindex mspecld-anomaly
9776 When enabled, the compiler will ensure that the generated code does not
9777 contain speculative loads after jump instructions. If this option is used,
9778 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
9780 @item -mno-specld-anomaly
9781 @opindex mno-specld-anomaly
9782 Don't generate extra code to prevent speculative loads from occurring.
9784 @item -mcsync-anomaly
9785 @opindex mcsync-anomaly
9786 When enabled, the compiler will ensure that the generated code does not
9787 contain CSYNC or SSYNC instructions too soon after conditional branches.
9788 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
9790 @item -mno-csync-anomaly
9791 @opindex mno-csync-anomaly
9792 Don't generate extra code to prevent CSYNC or SSYNC instructions from
9793 occurring too soon after a conditional branch.
9797 When enabled, the compiler is free to take advantage of the knowledge that
9798 the entire program fits into the low 64k of memory.
9801 @opindex mno-low-64k
9802 Assume that the program is arbitrarily large. This is the default.
9804 @item -mstack-check-l1
9805 @opindex mstack-check-l1
9806 Do stack checking using information placed into L1 scratchpad memory by the
9809 @item -mid-shared-library
9810 @opindex mid-shared-library
9811 Generate code that supports shared libraries via the library ID method.
9812 This allows for execute in place and shared libraries in an environment
9813 without virtual memory management. This option implies @option{-fPIC}.
9814 With a @samp{bfin-elf} target, this option implies @option{-msim}.
9816 @item -mno-id-shared-library
9817 @opindex mno-id-shared-library
9818 Generate code that doesn't assume ID based shared libraries are being used.
9819 This is the default.
9821 @item -mleaf-id-shared-library
9822 @opindex mleaf-id-shared-library
9823 Generate code that supports shared libraries via the library ID method,
9824 but assumes that this library or executable won't link against any other
9825 ID shared libraries. That allows the compiler to use faster code for jumps
9828 @item -mno-leaf-id-shared-library
9829 @opindex mno-leaf-id-shared-library
9830 Do not assume that the code being compiled won't link against any ID shared
9831 libraries. Slower code will be generated for jump and call insns.
9833 @item -mshared-library-id=n
9834 @opindex mshared-library-id
9835 Specified the identification number of the ID based shared library being
9836 compiled. Specifying a value of 0 will generate more compact code, specifying
9837 other values will force the allocation of that number to the current
9838 library but is no more space or time efficient than omitting this option.
9842 Generate code that allows the data segment to be located in a different
9843 area of memory from the text segment. This allows for execute in place in
9844 an environment without virtual memory management by eliminating relocations
9845 against the text section.
9848 @opindex mno-sep-data
9849 Generate code that assumes that the data segment follows the text segment.
9850 This is the default.
9853 @itemx -mno-long-calls
9854 @opindex mlong-calls
9855 @opindex mno-long-calls
9856 Tells the compiler to perform function calls by first loading the
9857 address of the function into a register and then performing a subroutine
9858 call on this register. This switch is needed if the target function
9859 will lie outside of the 24 bit addressing range of the offset based
9860 version of subroutine call instruction.
9862 This feature is not enabled by default. Specifying
9863 @option{-mno-long-calls} will restore the default behavior. Note these
9864 switches have no effect on how the compiler generates code to handle
9865 function calls via function pointers.
9869 Link with the fast floating-point library. This library relaxes some of
9870 the IEEE floating-point standard's rules for checking inputs against
9871 Not-a-Number (NAN), in the interest of performance.
9874 @opindex minline-plt
9875 Enable inlining of PLT entries in function calls to functions that are
9876 not known to bind locally. It has no effect without @option{-mfdpic}.
9880 Build standalone application for multicore Blackfin processor. Proper
9881 start files and link scripts will be used to support multicore.
9882 This option defines @code{__BFIN_MULTICORE}. It can only be used with
9883 @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}. It can be used with
9884 @option{-mcorea} or @option{-mcoreb}. If it's used without
9885 @option{-mcorea} or @option{-mcoreb}, single application/dual core
9886 programming model is used. In this model, the main function of Core B
9887 should be named as coreb_main. If it's used with @option{-mcorea} or
9888 @option{-mcoreb}, one application per core programming model is used.
9889 If this option is not used, single core application programming
9894 Build standalone application for Core A of BF561 when using
9895 one application per core programming model. Proper start files
9896 and link scripts will be used to support Core A. This option
9897 defines @code{__BFIN_COREA}. It must be used with @option{-mmulticore}.
9901 Build standalone application for Core B of BF561 when using
9902 one application per core programming model. Proper start files
9903 and link scripts will be used to support Core B. This option
9904 defines @code{__BFIN_COREB}. When this option is used, coreb_main
9905 should be used instead of main. It must be used with
9906 @option{-mmulticore}.
9910 Build standalone application for SDRAM. Proper start files and
9911 link scripts will be used to put the application into SDRAM.
9912 Loader should initialize SDRAM before loading the application
9913 into SDRAM. This option defines @code{__BFIN_SDRAM}.
9917 Assume that ICPLBs are enabled at runtime. This has an effect on certain
9918 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
9919 are enabled; for standalone applications the default is off.
9923 @subsection CRIS Options
9924 @cindex CRIS Options
9926 These options are defined specifically for the CRIS ports.
9929 @item -march=@var{architecture-type}
9930 @itemx -mcpu=@var{architecture-type}
9933 Generate code for the specified architecture. The choices for
9934 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9935 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
9936 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9939 @item -mtune=@var{architecture-type}
9941 Tune to @var{architecture-type} everything applicable about the generated
9942 code, except for the ABI and the set of available instructions. The
9943 choices for @var{architecture-type} are the same as for
9944 @option{-march=@var{architecture-type}}.
9946 @item -mmax-stack-frame=@var{n}
9947 @opindex mmax-stack-frame
9948 Warn when the stack frame of a function exceeds @var{n} bytes.
9954 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9955 @option{-march=v3} and @option{-march=v8} respectively.
9957 @item -mmul-bug-workaround
9958 @itemx -mno-mul-bug-workaround
9959 @opindex mmul-bug-workaround
9960 @opindex mno-mul-bug-workaround
9961 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
9962 models where it applies. This option is active by default.
9966 Enable CRIS-specific verbose debug-related information in the assembly
9967 code. This option also has the effect to turn off the @samp{#NO_APP}
9968 formatted-code indicator to the assembler at the beginning of the
9973 Do not use condition-code results from previous instruction; always emit
9974 compare and test instructions before use of condition codes.
9976 @item -mno-side-effects
9977 @opindex mno-side-effects
9978 Do not emit instructions with side-effects in addressing modes other than
9982 @itemx -mno-stack-align
9984 @itemx -mno-data-align
9985 @itemx -mconst-align
9986 @itemx -mno-const-align
9987 @opindex mstack-align
9988 @opindex mno-stack-align
9989 @opindex mdata-align
9990 @opindex mno-data-align
9991 @opindex mconst-align
9992 @opindex mno-const-align
9993 These options (no-options) arranges (eliminate arrangements) for the
9994 stack-frame, individual data and constants to be aligned for the maximum
9995 single data access size for the chosen CPU model. The default is to
9996 arrange for 32-bit alignment. ABI details such as structure layout are
9997 not affected by these options.
10005 Similar to the stack- data- and const-align options above, these options
10006 arrange for stack-frame, writable data and constants to all be 32-bit,
10007 16-bit or 8-bit aligned. The default is 32-bit alignment.
10009 @item -mno-prologue-epilogue
10010 @itemx -mprologue-epilogue
10011 @opindex mno-prologue-epilogue
10012 @opindex mprologue-epilogue
10013 With @option{-mno-prologue-epilogue}, the normal function prologue and
10014 epilogue that sets up the stack-frame are omitted and no return
10015 instructions or return sequences are generated in the code. Use this
10016 option only together with visual inspection of the compiled code: no
10017 warnings or errors are generated when call-saved registers must be saved,
10018 or storage for local variable needs to be allocated.
10022 @opindex mno-gotplt
10024 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10025 instruction sequences that load addresses for functions from the PLT part
10026 of the GOT rather than (traditional on other architectures) calls to the
10027 PLT@. The default is @option{-mgotplt}.
10031 Legacy no-op option only recognized with the cris-axis-elf and
10032 cris-axis-linux-gnu targets.
10036 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10040 This option, recognized for the cris-axis-elf arranges
10041 to link with input-output functions from a simulator library. Code,
10042 initialized data and zero-initialized data are allocated consecutively.
10046 Like @option{-sim}, but pass linker options to locate initialized data at
10047 0x40000000 and zero-initialized data at 0x80000000.
10051 @subsection CRX Options
10052 @cindex CRX Options
10054 These options are defined specifically for the CRX ports.
10060 Enable the use of multiply-accumulate instructions. Disabled by default.
10063 @opindex mpush-args
10064 Push instructions will be used to pass outgoing arguments when functions
10065 are called. Enabled by default.
10068 @node Darwin Options
10069 @subsection Darwin Options
10070 @cindex Darwin options
10072 These options are defined for all architectures running the Darwin operating
10075 FSF GCC on Darwin does not create ``fat'' object files; it will create
10076 an object file for the single architecture that it was built to
10077 target. Apple's GCC on Darwin does create ``fat'' files if multiple
10078 @option{-arch} options are used; it does so by running the compiler or
10079 linker multiple times and joining the results together with
10082 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
10083 @samp{i686}) is determined by the flags that specify the ISA
10084 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
10085 @option{-force_cpusubtype_ALL} option can be used to override this.
10087 The Darwin tools vary in their behavior when presented with an ISA
10088 mismatch. The assembler, @file{as}, will only permit instructions to
10089 be used that are valid for the subtype of the file it is generating,
10090 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
10091 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
10092 and print an error if asked to create a shared library with a less
10093 restrictive subtype than its input files (for instance, trying to put
10094 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
10095 for executables, @file{ld}, will quietly give the executable the most
10096 restrictive subtype of any of its input files.
10101 Add the framework directory @var{dir} to the head of the list of
10102 directories to be searched for header files. These directories are
10103 interleaved with those specified by @option{-I} options and are
10104 scanned in a left-to-right order.
10106 A framework directory is a directory with frameworks in it. A
10107 framework is a directory with a @samp{"Headers"} and/or
10108 @samp{"PrivateHeaders"} directory contained directly in it that ends
10109 in @samp{".framework"}. The name of a framework is the name of this
10110 directory excluding the @samp{".framework"}. Headers associated with
10111 the framework are found in one of those two directories, with
10112 @samp{"Headers"} being searched first. A subframework is a framework
10113 directory that is in a framework's @samp{"Frameworks"} directory.
10114 Includes of subframework headers can only appear in a header of a
10115 framework that contains the subframework, or in a sibling subframework
10116 header. Two subframeworks are siblings if they occur in the same
10117 framework. A subframework should not have the same name as a
10118 framework, a warning will be issued if this is violated. Currently a
10119 subframework cannot have subframeworks, in the future, the mechanism
10120 may be extended to support this. The standard frameworks can be found
10121 in @samp{"/System/Library/Frameworks"} and
10122 @samp{"/Library/Frameworks"}. An example include looks like
10123 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
10124 the name of the framework and header.h is found in the
10125 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
10127 @item -iframework@var{dir}
10128 @opindex iframework
10129 Like @option{-F} except the directory is a treated as a system
10130 directory. The main difference between this @option{-iframework} and
10131 @option{-F} is that with @option{-iframework} the compiler does not
10132 warn about constructs contained within header files found via
10133 @var{dir}. This option is valid only for the C family of languages.
10137 Emit debugging information for symbols that are used. For STABS
10138 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
10139 This is by default ON@.
10143 Emit debugging information for all symbols and types.
10145 @item -mmacosx-version-min=@var{version}
10146 The earliest version of MacOS X that this executable will run on
10147 is @var{version}. Typical values of @var{version} include @code{10.1},
10148 @code{10.2}, and @code{10.3.9}.
10150 If the compiler was built to use the system's headers by default,
10151 then the default for this option is the system version on which the
10152 compiler is running, otherwise the default is to make choices which
10153 are compatible with as many systems and code bases as possible.
10157 Enable kernel development mode. The @option{-mkernel} option sets
10158 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
10159 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
10160 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
10161 applicable. This mode also sets @option{-mno-altivec},
10162 @option{-msoft-float}, @option{-fno-builtin} and
10163 @option{-mlong-branch} for PowerPC targets.
10165 @item -mone-byte-bool
10166 @opindex mone-byte-bool
10167 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
10168 By default @samp{sizeof(bool)} is @samp{4} when compiling for
10169 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
10170 option has no effect on x86.
10172 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
10173 to generate code that is not binary compatible with code generated
10174 without that switch. Using this switch may require recompiling all
10175 other modules in a program, including system libraries. Use this
10176 switch to conform to a non-default data model.
10178 @item -mfix-and-continue
10179 @itemx -ffix-and-continue
10180 @itemx -findirect-data
10181 @opindex mfix-and-continue
10182 @opindex ffix-and-continue
10183 @opindex findirect-data
10184 Generate code suitable for fast turn around development. Needed to
10185 enable gdb to dynamically load @code{.o} files into already running
10186 programs. @option{-findirect-data} and @option{-ffix-and-continue}
10187 are provided for backwards compatibility.
10191 Loads all members of static archive libraries.
10192 See man ld(1) for more information.
10194 @item -arch_errors_fatal
10195 @opindex arch_errors_fatal
10196 Cause the errors having to do with files that have the wrong architecture
10199 @item -bind_at_load
10200 @opindex bind_at_load
10201 Causes the output file to be marked such that the dynamic linker will
10202 bind all undefined references when the file is loaded or launched.
10206 Produce a Mach-o bundle format file.
10207 See man ld(1) for more information.
10209 @item -bundle_loader @var{executable}
10210 @opindex bundle_loader
10211 This option specifies the @var{executable} that will be loading the build
10212 output file being linked. See man ld(1) for more information.
10215 @opindex dynamiclib
10216 When passed this option, GCC will produce a dynamic library instead of
10217 an executable when linking, using the Darwin @file{libtool} command.
10219 @item -force_cpusubtype_ALL
10220 @opindex force_cpusubtype_ALL
10221 This causes GCC's output file to have the @var{ALL} subtype, instead of
10222 one controlled by the @option{-mcpu} or @option{-march} option.
10224 @item -allowable_client @var{client_name}
10225 @itemx -client_name
10226 @itemx -compatibility_version
10227 @itemx -current_version
10229 @itemx -dependency-file
10231 @itemx -dylinker_install_name
10233 @itemx -exported_symbols_list
10235 @itemx -flat_namespace
10236 @itemx -force_flat_namespace
10237 @itemx -headerpad_max_install_names
10240 @itemx -install_name
10241 @itemx -keep_private_externs
10242 @itemx -multi_module
10243 @itemx -multiply_defined
10244 @itemx -multiply_defined_unused
10246 @itemx -no_dead_strip_inits_and_terms
10247 @itemx -nofixprebinding
10248 @itemx -nomultidefs
10250 @itemx -noseglinkedit
10251 @itemx -pagezero_size
10253 @itemx -prebind_all_twolevel_modules
10254 @itemx -private_bundle
10255 @itemx -read_only_relocs
10257 @itemx -sectobjectsymbols
10261 @itemx -sectobjectsymbols
10264 @itemx -segs_read_only_addr
10265 @itemx -segs_read_write_addr
10266 @itemx -seg_addr_table
10267 @itemx -seg_addr_table_filename
10268 @itemx -seglinkedit
10270 @itemx -segs_read_only_addr
10271 @itemx -segs_read_write_addr
10272 @itemx -single_module
10274 @itemx -sub_library
10275 @itemx -sub_umbrella
10276 @itemx -twolevel_namespace
10279 @itemx -unexported_symbols_list
10280 @itemx -weak_reference_mismatches
10281 @itemx -whatsloaded
10282 @opindex allowable_client
10283 @opindex client_name
10284 @opindex compatibility_version
10285 @opindex current_version
10286 @opindex dead_strip
10287 @opindex dependency-file
10288 @opindex dylib_file
10289 @opindex dylinker_install_name
10291 @opindex exported_symbols_list
10293 @opindex flat_namespace
10294 @opindex force_flat_namespace
10295 @opindex headerpad_max_install_names
10296 @opindex image_base
10298 @opindex install_name
10299 @opindex keep_private_externs
10300 @opindex multi_module
10301 @opindex multiply_defined
10302 @opindex multiply_defined_unused
10303 @opindex noall_load
10304 @opindex no_dead_strip_inits_and_terms
10305 @opindex nofixprebinding
10306 @opindex nomultidefs
10308 @opindex noseglinkedit
10309 @opindex pagezero_size
10311 @opindex prebind_all_twolevel_modules
10312 @opindex private_bundle
10313 @opindex read_only_relocs
10315 @opindex sectobjectsymbols
10318 @opindex sectcreate
10319 @opindex sectobjectsymbols
10322 @opindex segs_read_only_addr
10323 @opindex segs_read_write_addr
10324 @opindex seg_addr_table
10325 @opindex seg_addr_table_filename
10326 @opindex seglinkedit
10328 @opindex segs_read_only_addr
10329 @opindex segs_read_write_addr
10330 @opindex single_module
10332 @opindex sub_library
10333 @opindex sub_umbrella
10334 @opindex twolevel_namespace
10337 @opindex unexported_symbols_list
10338 @opindex weak_reference_mismatches
10339 @opindex whatsloaded
10340 These options are passed to the Darwin linker. The Darwin linker man page
10341 describes them in detail.
10344 @node DEC Alpha Options
10345 @subsection DEC Alpha Options
10347 These @samp{-m} options are defined for the DEC Alpha implementations:
10350 @item -mno-soft-float
10351 @itemx -msoft-float
10352 @opindex mno-soft-float
10353 @opindex msoft-float
10354 Use (do not use) the hardware floating-point instructions for
10355 floating-point operations. When @option{-msoft-float} is specified,
10356 functions in @file{libgcc.a} will be used to perform floating-point
10357 operations. Unless they are replaced by routines that emulate the
10358 floating-point operations, or compiled in such a way as to call such
10359 emulations routines, these routines will issue floating-point
10360 operations. If you are compiling for an Alpha without floating-point
10361 operations, you must ensure that the library is built so as not to call
10364 Note that Alpha implementations without floating-point operations are
10365 required to have floating-point registers.
10368 @itemx -mno-fp-regs
10370 @opindex mno-fp-regs
10371 Generate code that uses (does not use) the floating-point register set.
10372 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
10373 register set is not used, floating point operands are passed in integer
10374 registers as if they were integers and floating-point results are passed
10375 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
10376 so any function with a floating-point argument or return value called by code
10377 compiled with @option{-mno-fp-regs} must also be compiled with that
10380 A typical use of this option is building a kernel that does not use,
10381 and hence need not save and restore, any floating-point registers.
10385 The Alpha architecture implements floating-point hardware optimized for
10386 maximum performance. It is mostly compliant with the IEEE floating
10387 point standard. However, for full compliance, software assistance is
10388 required. This option generates code fully IEEE compliant code
10389 @emph{except} that the @var{inexact-flag} is not maintained (see below).
10390 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
10391 defined during compilation. The resulting code is less efficient but is
10392 able to correctly support denormalized numbers and exceptional IEEE
10393 values such as not-a-number and plus/minus infinity. Other Alpha
10394 compilers call this option @option{-ieee_with_no_inexact}.
10396 @item -mieee-with-inexact
10397 @opindex mieee-with-inexact
10398 This is like @option{-mieee} except the generated code also maintains
10399 the IEEE @var{inexact-flag}. Turning on this option causes the
10400 generated code to implement fully-compliant IEEE math. In addition to
10401 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
10402 macro. On some Alpha implementations the resulting code may execute
10403 significantly slower than the code generated by default. Since there is
10404 very little code that depends on the @var{inexact-flag}, you should
10405 normally not specify this option. Other Alpha compilers call this
10406 option @option{-ieee_with_inexact}.
10408 @item -mfp-trap-mode=@var{trap-mode}
10409 @opindex mfp-trap-mode
10410 This option controls what floating-point related traps are enabled.
10411 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
10412 The trap mode can be set to one of four values:
10416 This is the default (normal) setting. The only traps that are enabled
10417 are the ones that cannot be disabled in software (e.g., division by zero
10421 In addition to the traps enabled by @samp{n}, underflow traps are enabled
10425 Like @samp{u}, but the instructions are marked to be safe for software
10426 completion (see Alpha architecture manual for details).
10429 Like @samp{su}, but inexact traps are enabled as well.
10432 @item -mfp-rounding-mode=@var{rounding-mode}
10433 @opindex mfp-rounding-mode
10434 Selects the IEEE rounding mode. Other Alpha compilers call this option
10435 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
10440 Normal IEEE rounding mode. Floating point numbers are rounded towards
10441 the nearest machine number or towards the even machine number in case
10445 Round towards minus infinity.
10448 Chopped rounding mode. Floating point numbers are rounded towards zero.
10451 Dynamic rounding mode. A field in the floating point control register
10452 (@var{fpcr}, see Alpha architecture reference manual) controls the
10453 rounding mode in effect. The C library initializes this register for
10454 rounding towards plus infinity. Thus, unless your program modifies the
10455 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
10458 @item -mtrap-precision=@var{trap-precision}
10459 @opindex mtrap-precision
10460 In the Alpha architecture, floating point traps are imprecise. This
10461 means without software assistance it is impossible to recover from a
10462 floating trap and program execution normally needs to be terminated.
10463 GCC can generate code that can assist operating system trap handlers
10464 in determining the exact location that caused a floating point trap.
10465 Depending on the requirements of an application, different levels of
10466 precisions can be selected:
10470 Program precision. This option is the default and means a trap handler
10471 can only identify which program caused a floating point exception.
10474 Function precision. The trap handler can determine the function that
10475 caused a floating point exception.
10478 Instruction precision. The trap handler can determine the exact
10479 instruction that caused a floating point exception.
10482 Other Alpha compilers provide the equivalent options called
10483 @option{-scope_safe} and @option{-resumption_safe}.
10485 @item -mieee-conformant
10486 @opindex mieee-conformant
10487 This option marks the generated code as IEEE conformant. You must not
10488 use this option unless you also specify @option{-mtrap-precision=i} and either
10489 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
10490 is to emit the line @samp{.eflag 48} in the function prologue of the
10491 generated assembly file. Under DEC Unix, this has the effect that
10492 IEEE-conformant math library routines will be linked in.
10494 @item -mbuild-constants
10495 @opindex mbuild-constants
10496 Normally GCC examines a 32- or 64-bit integer constant to
10497 see if it can construct it from smaller constants in two or three
10498 instructions. If it cannot, it will output the constant as a literal and
10499 generate code to load it from the data segment at runtime.
10501 Use this option to require GCC to construct @emph{all} integer constants
10502 using code, even if it takes more instructions (the maximum is six).
10504 You would typically use this option to build a shared library dynamic
10505 loader. Itself a shared library, it must relocate itself in memory
10506 before it can find the variables and constants in its own data segment.
10512 Select whether to generate code to be assembled by the vendor-supplied
10513 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
10531 Indicate whether GCC should generate code to use the optional BWX,
10532 CIX, FIX and MAX instruction sets. The default is to use the instruction
10533 sets supported by the CPU type specified via @option{-mcpu=} option or that
10534 of the CPU on which GCC was built if none was specified.
10537 @itemx -mfloat-ieee
10538 @opindex mfloat-vax
10539 @opindex mfloat-ieee
10540 Generate code that uses (does not use) VAX F and G floating point
10541 arithmetic instead of IEEE single and double precision.
10543 @item -mexplicit-relocs
10544 @itemx -mno-explicit-relocs
10545 @opindex mexplicit-relocs
10546 @opindex mno-explicit-relocs
10547 Older Alpha assemblers provided no way to generate symbol relocations
10548 except via assembler macros. Use of these macros does not allow
10549 optimal instruction scheduling. GNU binutils as of version 2.12
10550 supports a new syntax that allows the compiler to explicitly mark
10551 which relocations should apply to which instructions. This option
10552 is mostly useful for debugging, as GCC detects the capabilities of
10553 the assembler when it is built and sets the default accordingly.
10556 @itemx -mlarge-data
10557 @opindex msmall-data
10558 @opindex mlarge-data
10559 When @option{-mexplicit-relocs} is in effect, static data is
10560 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
10561 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
10562 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
10563 16-bit relocations off of the @code{$gp} register. This limits the
10564 size of the small data area to 64KB, but allows the variables to be
10565 directly accessed via a single instruction.
10567 The default is @option{-mlarge-data}. With this option the data area
10568 is limited to just below 2GB@. Programs that require more than 2GB of
10569 data must use @code{malloc} or @code{mmap} to allocate the data in the
10570 heap instead of in the program's data segment.
10572 When generating code for shared libraries, @option{-fpic} implies
10573 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
10576 @itemx -mlarge-text
10577 @opindex msmall-text
10578 @opindex mlarge-text
10579 When @option{-msmall-text} is used, the compiler assumes that the
10580 code of the entire program (or shared library) fits in 4MB, and is
10581 thus reachable with a branch instruction. When @option{-msmall-data}
10582 is used, the compiler can assume that all local symbols share the
10583 same @code{$gp} value, and thus reduce the number of instructions
10584 required for a function call from 4 to 1.
10586 The default is @option{-mlarge-text}.
10588 @item -mcpu=@var{cpu_type}
10590 Set the instruction set and instruction scheduling parameters for
10591 machine type @var{cpu_type}. You can specify either the @samp{EV}
10592 style name or the corresponding chip number. GCC supports scheduling
10593 parameters for the EV4, EV5 and EV6 family of processors and will
10594 choose the default values for the instruction set from the processor
10595 you specify. If you do not specify a processor type, GCC will default
10596 to the processor on which the compiler was built.
10598 Supported values for @var{cpu_type} are
10604 Schedules as an EV4 and has no instruction set extensions.
10608 Schedules as an EV5 and has no instruction set extensions.
10612 Schedules as an EV5 and supports the BWX extension.
10617 Schedules as an EV5 and supports the BWX and MAX extensions.
10621 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
10625 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
10628 Native Linux/GNU toolchains also support the value @samp{native},
10629 which selects the best architecture option for the host processor.
10630 @option{-mcpu=native} has no effect if GCC does not recognize
10633 @item -mtune=@var{cpu_type}
10635 Set only the instruction scheduling parameters for machine type
10636 @var{cpu_type}. The instruction set is not changed.
10638 Native Linux/GNU toolchains also support the value @samp{native},
10639 which selects the best architecture option for the host processor.
10640 @option{-mtune=native} has no effect if GCC does not recognize
10643 @item -mmemory-latency=@var{time}
10644 @opindex mmemory-latency
10645 Sets the latency the scheduler should assume for typical memory
10646 references as seen by the application. This number is highly
10647 dependent on the memory access patterns used by the application
10648 and the size of the external cache on the machine.
10650 Valid options for @var{time} are
10654 A decimal number representing clock cycles.
10660 The compiler contains estimates of the number of clock cycles for
10661 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
10662 (also called Dcache, Scache, and Bcache), as well as to main memory.
10663 Note that L3 is only valid for EV5.
10668 @node DEC Alpha/VMS Options
10669 @subsection DEC Alpha/VMS Options
10671 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
10674 @item -mvms-return-codes
10675 @opindex mvms-return-codes
10676 Return VMS condition codes from main. The default is to return POSIX
10677 style condition (e.g.@: error) codes.
10679 @item -mdebug-main=@var{prefix}
10680 @opindex mdebug-main=@var{prefix}
10681 Flag the first routine whose name starts with @var{prefix} as the main
10682 routine for the debugger.
10686 Default to 64bit memory allocation routines.
10690 @subsection FR30 Options
10691 @cindex FR30 Options
10693 These options are defined specifically for the FR30 port.
10697 @item -msmall-model
10698 @opindex msmall-model
10699 Use the small address space model. This can produce smaller code, but
10700 it does assume that all symbolic values and addresses will fit into a
10705 Assume that run-time support has been provided and so there is no need
10706 to include the simulator library (@file{libsim.a}) on the linker
10712 @subsection FRV Options
10713 @cindex FRV Options
10719 Only use the first 32 general purpose registers.
10724 Use all 64 general purpose registers.
10729 Use only the first 32 floating point registers.
10734 Use all 64 floating point registers
10737 @opindex mhard-float
10739 Use hardware instructions for floating point operations.
10742 @opindex msoft-float
10744 Use library routines for floating point operations.
10749 Dynamically allocate condition code registers.
10754 Do not try to dynamically allocate condition code registers, only
10755 use @code{icc0} and @code{fcc0}.
10760 Change ABI to use double word insns.
10765 Do not use double word instructions.
10770 Use floating point double instructions.
10773 @opindex mno-double
10775 Do not use floating point double instructions.
10780 Use media instructions.
10785 Do not use media instructions.
10790 Use multiply and add/subtract instructions.
10793 @opindex mno-muladd
10795 Do not use multiply and add/subtract instructions.
10800 Select the FDPIC ABI, that uses function descriptors to represent
10801 pointers to functions. Without any PIC/PIE-related options, it
10802 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
10803 assumes GOT entries and small data are within a 12-bit range from the
10804 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
10805 are computed with 32 bits.
10806 With a @samp{bfin-elf} target, this option implies @option{-msim}.
10809 @opindex minline-plt
10811 Enable inlining of PLT entries in function calls to functions that are
10812 not known to bind locally. It has no effect without @option{-mfdpic}.
10813 It's enabled by default if optimizing for speed and compiling for
10814 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
10815 optimization option such as @option{-O3} or above is present in the
10821 Assume a large TLS segment when generating thread-local code.
10826 Do not assume a large TLS segment when generating thread-local code.
10831 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
10832 that is known to be in read-only sections. It's enabled by default,
10833 except for @option{-fpic} or @option{-fpie}: even though it may help
10834 make the global offset table smaller, it trades 1 instruction for 4.
10835 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
10836 one of which may be shared by multiple symbols, and it avoids the need
10837 for a GOT entry for the referenced symbol, so it's more likely to be a
10838 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
10840 @item -multilib-library-pic
10841 @opindex multilib-library-pic
10843 Link with the (library, not FD) pic libraries. It's implied by
10844 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
10845 @option{-fpic} without @option{-mfdpic}. You should never have to use
10849 @opindex mlinked-fp
10851 Follow the EABI requirement of always creating a frame pointer whenever
10852 a stack frame is allocated. This option is enabled by default and can
10853 be disabled with @option{-mno-linked-fp}.
10856 @opindex mlong-calls
10858 Use indirect addressing to call functions outside the current
10859 compilation unit. This allows the functions to be placed anywhere
10860 within the 32-bit address space.
10862 @item -malign-labels
10863 @opindex malign-labels
10865 Try to align labels to an 8-byte boundary by inserting nops into the
10866 previous packet. This option only has an effect when VLIW packing
10867 is enabled. It doesn't create new packets; it merely adds nops to
10870 @item -mlibrary-pic
10871 @opindex mlibrary-pic
10873 Generate position-independent EABI code.
10878 Use only the first four media accumulator registers.
10883 Use all eight media accumulator registers.
10888 Pack VLIW instructions.
10893 Do not pack VLIW instructions.
10896 @opindex mno-eflags
10898 Do not mark ABI switches in e_flags.
10901 @opindex mcond-move
10903 Enable the use of conditional-move instructions (default).
10905 This switch is mainly for debugging the compiler and will likely be removed
10906 in a future version.
10908 @item -mno-cond-move
10909 @opindex mno-cond-move
10911 Disable the use of conditional-move instructions.
10913 This switch is mainly for debugging the compiler and will likely be removed
10914 in a future version.
10919 Enable the use of conditional set instructions (default).
10921 This switch is mainly for debugging the compiler and will likely be removed
10922 in a future version.
10927 Disable the use of conditional set instructions.
10929 This switch is mainly for debugging the compiler and will likely be removed
10930 in a future version.
10933 @opindex mcond-exec
10935 Enable the use of conditional execution (default).
10937 This switch is mainly for debugging the compiler and will likely be removed
10938 in a future version.
10940 @item -mno-cond-exec
10941 @opindex mno-cond-exec
10943 Disable the use of conditional execution.
10945 This switch is mainly for debugging the compiler and will likely be removed
10946 in a future version.
10948 @item -mvliw-branch
10949 @opindex mvliw-branch
10951 Run a pass to pack branches into VLIW instructions (default).
10953 This switch is mainly for debugging the compiler and will likely be removed
10954 in a future version.
10956 @item -mno-vliw-branch
10957 @opindex mno-vliw-branch
10959 Do not run a pass to pack branches into VLIW instructions.
10961 This switch is mainly for debugging the compiler and will likely be removed
10962 in a future version.
10964 @item -mmulti-cond-exec
10965 @opindex mmulti-cond-exec
10967 Enable optimization of @code{&&} and @code{||} in conditional execution
10970 This switch is mainly for debugging the compiler and will likely be removed
10971 in a future version.
10973 @item -mno-multi-cond-exec
10974 @opindex mno-multi-cond-exec
10976 Disable optimization of @code{&&} and @code{||} in conditional execution.
10978 This switch is mainly for debugging the compiler and will likely be removed
10979 in a future version.
10981 @item -mnested-cond-exec
10982 @opindex mnested-cond-exec
10984 Enable nested conditional execution optimizations (default).
10986 This switch is mainly for debugging the compiler and will likely be removed
10987 in a future version.
10989 @item -mno-nested-cond-exec
10990 @opindex mno-nested-cond-exec
10992 Disable nested conditional execution optimizations.
10994 This switch is mainly for debugging the compiler and will likely be removed
10995 in a future version.
10997 @item -moptimize-membar
10998 @opindex moptimize-membar
11000 This switch removes redundant @code{membar} instructions from the
11001 compiler generated code. It is enabled by default.
11003 @item -mno-optimize-membar
11004 @opindex mno-optimize-membar
11006 This switch disables the automatic removal of redundant @code{membar}
11007 instructions from the generated code.
11009 @item -mtomcat-stats
11010 @opindex mtomcat-stats
11012 Cause gas to print out tomcat statistics.
11014 @item -mcpu=@var{cpu}
11017 Select the processor type for which to generate code. Possible values are
11018 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
11019 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
11023 @node GNU/Linux Options
11024 @subsection GNU/Linux Options
11026 These @samp{-m} options are defined for GNU/Linux targets:
11031 Use the GNU C library instead of uClibc. This is the default except
11032 on @samp{*-*-linux-*uclibc*} targets.
11036 Use uClibc instead of the GNU C library. This is the default on
11037 @samp{*-*-linux-*uclibc*} targets.
11040 @node H8/300 Options
11041 @subsection H8/300 Options
11043 These @samp{-m} options are defined for the H8/300 implementations:
11048 Shorten some address references at link time, when possible; uses the
11049 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
11050 ld, Using ld}, for a fuller description.
11054 Generate code for the H8/300H@.
11058 Generate code for the H8S@.
11062 Generate code for the H8S and H8/300H in the normal mode. This switch
11063 must be used either with @option{-mh} or @option{-ms}.
11067 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
11071 Make @code{int} data 32 bits by default.
11074 @opindex malign-300
11075 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
11076 The default for the H8/300H and H8S is to align longs and floats on 4
11078 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
11079 This option has no effect on the H8/300.
11083 @subsection HPPA Options
11084 @cindex HPPA Options
11086 These @samp{-m} options are defined for the HPPA family of computers:
11089 @item -march=@var{architecture-type}
11091 Generate code for the specified architecture. The choices for
11092 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
11093 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
11094 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
11095 architecture option for your machine. Code compiled for lower numbered
11096 architectures will run on higher numbered architectures, but not the
11099 @item -mpa-risc-1-0
11100 @itemx -mpa-risc-1-1
11101 @itemx -mpa-risc-2-0
11102 @opindex mpa-risc-1-0
11103 @opindex mpa-risc-1-1
11104 @opindex mpa-risc-2-0
11105 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
11108 @opindex mbig-switch
11109 Generate code suitable for big switch tables. Use this option only if
11110 the assembler/linker complain about out of range branches within a switch
11113 @item -mjump-in-delay
11114 @opindex mjump-in-delay
11115 Fill delay slots of function calls with unconditional jump instructions
11116 by modifying the return pointer for the function call to be the target
11117 of the conditional jump.
11119 @item -mdisable-fpregs
11120 @opindex mdisable-fpregs
11121 Prevent floating point registers from being used in any manner. This is
11122 necessary for compiling kernels which perform lazy context switching of
11123 floating point registers. If you use this option and attempt to perform
11124 floating point operations, the compiler will abort.
11126 @item -mdisable-indexing
11127 @opindex mdisable-indexing
11128 Prevent the compiler from using indexing address modes. This avoids some
11129 rather obscure problems when compiling MIG generated code under MACH@.
11131 @item -mno-space-regs
11132 @opindex mno-space-regs
11133 Generate code that assumes the target has no space registers. This allows
11134 GCC to generate faster indirect calls and use unscaled index address modes.
11136 Such code is suitable for level 0 PA systems and kernels.
11138 @item -mfast-indirect-calls
11139 @opindex mfast-indirect-calls
11140 Generate code that assumes calls never cross space boundaries. This
11141 allows GCC to emit code which performs faster indirect calls.
11143 This option will not work in the presence of shared libraries or nested
11146 @item -mfixed-range=@var{register-range}
11147 @opindex mfixed-range
11148 Generate code treating the given register range as fixed registers.
11149 A fixed register is one that the register allocator can not use. This is
11150 useful when compiling kernel code. A register range is specified as
11151 two registers separated by a dash. Multiple register ranges can be
11152 specified separated by a comma.
11154 @item -mlong-load-store
11155 @opindex mlong-load-store
11156 Generate 3-instruction load and store sequences as sometimes required by
11157 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
11160 @item -mportable-runtime
11161 @opindex mportable-runtime
11162 Use the portable calling conventions proposed by HP for ELF systems.
11166 Enable the use of assembler directives only GAS understands.
11168 @item -mschedule=@var{cpu-type}
11170 Schedule code according to the constraints for the machine type
11171 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
11172 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
11173 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
11174 proper scheduling option for your machine. The default scheduling is
11178 @opindex mlinker-opt
11179 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
11180 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
11181 linkers in which they give bogus error messages when linking some programs.
11184 @opindex msoft-float
11185 Generate output containing library calls for floating point.
11186 @strong{Warning:} the requisite libraries are not available for all HPPA
11187 targets. Normally the facilities of the machine's usual C compiler are
11188 used, but this cannot be done directly in cross-compilation. You must make
11189 your own arrangements to provide suitable library functions for
11192 @option{-msoft-float} changes the calling convention in the output file;
11193 therefore, it is only useful if you compile @emph{all} of a program with
11194 this option. In particular, you need to compile @file{libgcc.a}, the
11195 library that comes with GCC, with @option{-msoft-float} in order for
11200 Generate the predefine, @code{_SIO}, for server IO@. The default is
11201 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
11202 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
11203 options are available under HP-UX and HI-UX@.
11207 Use GNU ld specific options. This passes @option{-shared} to ld when
11208 building a shared library. It is the default when GCC is configured,
11209 explicitly or implicitly, with the GNU linker. This option does not
11210 have any affect on which ld is called, it only changes what parameters
11211 are passed to that ld. The ld that is called is determined by the
11212 @option{--with-ld} configure option, GCC's program search path, and
11213 finally by the user's @env{PATH}. The linker used by GCC can be printed
11214 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
11215 on the 64 bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11219 Use HP ld specific options. This passes @option{-b} to ld when building
11220 a shared library and passes @option{+Accept TypeMismatch} to ld on all
11221 links. It is the default when GCC is configured, explicitly or
11222 implicitly, with the HP linker. This option does not have any affect on
11223 which ld is called, it only changes what parameters are passed to that
11224 ld. The ld that is called is determined by the @option{--with-ld}
11225 configure option, GCC's program search path, and finally by the user's
11226 @env{PATH}. The linker used by GCC can be printed using @samp{which
11227 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
11228 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11231 @opindex mno-long-calls
11232 Generate code that uses long call sequences. This ensures that a call
11233 is always able to reach linker generated stubs. The default is to generate
11234 long calls only when the distance from the call site to the beginning
11235 of the function or translation unit, as the case may be, exceeds a
11236 predefined limit set by the branch type being used. The limits for
11237 normal calls are 7,600,000 and 240,000 bytes, respectively for the
11238 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
11241 Distances are measured from the beginning of functions when using the
11242 @option{-ffunction-sections} option, or when using the @option{-mgas}
11243 and @option{-mno-portable-runtime} options together under HP-UX with
11246 It is normally not desirable to use this option as it will degrade
11247 performance. However, it may be useful in large applications,
11248 particularly when partial linking is used to build the application.
11250 The types of long calls used depends on the capabilities of the
11251 assembler and linker, and the type of code being generated. The
11252 impact on systems that support long absolute calls, and long pic
11253 symbol-difference or pc-relative calls should be relatively small.
11254 However, an indirect call is used on 32-bit ELF systems in pic code
11255 and it is quite long.
11257 @item -munix=@var{unix-std}
11259 Generate compiler predefines and select a startfile for the specified
11260 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
11261 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
11262 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
11263 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
11264 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
11267 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
11268 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
11269 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
11270 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
11271 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
11272 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
11274 It is @emph{important} to note that this option changes the interfaces
11275 for various library routines. It also affects the operational behavior
11276 of the C library. Thus, @emph{extreme} care is needed in using this
11279 Library code that is intended to operate with more than one UNIX
11280 standard must test, set and restore the variable @var{__xpg4_extended_mask}
11281 as appropriate. Most GNU software doesn't provide this capability.
11285 Suppress the generation of link options to search libdld.sl when the
11286 @option{-static} option is specified on HP-UX 10 and later.
11290 The HP-UX implementation of setlocale in libc has a dependency on
11291 libdld.sl. There isn't an archive version of libdld.sl. Thus,
11292 when the @option{-static} option is specified, special link options
11293 are needed to resolve this dependency.
11295 On HP-UX 10 and later, the GCC driver adds the necessary options to
11296 link with libdld.sl when the @option{-static} option is specified.
11297 This causes the resulting binary to be dynamic. On the 64-bit port,
11298 the linkers generate dynamic binaries by default in any case. The
11299 @option{-nolibdld} option can be used to prevent the GCC driver from
11300 adding these link options.
11304 Add support for multithreading with the @dfn{dce thread} library
11305 under HP-UX@. This option sets flags for both the preprocessor and
11309 @node i386 and x86-64 Options
11310 @subsection Intel 386 and AMD x86-64 Options
11311 @cindex i386 Options
11312 @cindex x86-64 Options
11313 @cindex Intel 386 Options
11314 @cindex AMD x86-64 Options
11316 These @samp{-m} options are defined for the i386 and x86-64 family of
11320 @item -mtune=@var{cpu-type}
11322 Tune to @var{cpu-type} everything applicable about the generated code, except
11323 for the ABI and the set of available instructions. The choices for
11324 @var{cpu-type} are:
11327 Produce code optimized for the most common IA32/AMD64/EM64T processors.
11328 If you know the CPU on which your code will run, then you should use
11329 the corresponding @option{-mtune} option instead of
11330 @option{-mtune=generic}. But, if you do not know exactly what CPU users
11331 of your application will have, then you should use this option.
11333 As new processors are deployed in the marketplace, the behavior of this
11334 option will change. Therefore, if you upgrade to a newer version of
11335 GCC, the code generated option will change to reflect the processors
11336 that were most common when that version of GCC was released.
11338 There is no @option{-march=generic} option because @option{-march}
11339 indicates the instruction set the compiler can use, and there is no
11340 generic instruction set applicable to all processors. In contrast,
11341 @option{-mtune} indicates the processor (or, in this case, collection of
11342 processors) for which the code is optimized.
11344 This selects the CPU to tune for at compilation time by determining
11345 the processor type of the compiling machine. Using @option{-mtune=native}
11346 will produce code optimized for the local machine under the constraints
11347 of the selected instruction set. Using @option{-march=native} will
11348 enable all instruction subsets supported by the local machine (hence
11349 the result might not run on different machines).
11351 Original Intel's i386 CPU@.
11353 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
11354 @item i586, pentium
11355 Intel Pentium CPU with no MMX support.
11357 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
11359 Intel PentiumPro CPU@.
11361 Same as @code{generic}, but when used as @code{march} option, PentiumPro
11362 instruction set will be used, so the code will run on all i686 family chips.
11364 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
11365 @item pentium3, pentium3m
11366 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
11369 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
11370 support. Used by Centrino notebooks.
11371 @item pentium4, pentium4m
11372 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
11374 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
11377 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
11378 SSE2 and SSE3 instruction set support.
11380 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11381 instruction set support.
11383 Intel Atom CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11384 instruction set support.
11386 AMD K6 CPU with MMX instruction set support.
11388 Improved versions of AMD K6 CPU with MMX and 3dNOW!@: instruction set support.
11389 @item athlon, athlon-tbird
11390 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and SSE prefetch instructions
11392 @item athlon-4, athlon-xp, athlon-mp
11393 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW!@: and full SSE
11394 instruction set support.
11395 @item k8, opteron, athlon64, athlon-fx
11396 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
11397 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW!@: and 64-bit instruction set extensions.)
11398 @item k8-sse3, opteron-sse3, athlon64-sse3
11399 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
11400 @item amdfam10, barcelona
11401 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
11402 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3dNOW!, enhanced 3dNOW!, ABM and 64-bit
11403 instruction set extensions.)
11405 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
11408 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!@:
11409 instruction set support.
11411 Via C3 CPU with MMX and 3dNOW!@: instruction set support. (No scheduling is
11412 implemented for this chip.)
11414 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
11415 implemented for this chip.)
11417 Embedded AMD CPU with MMX and 3dNOW! instruction set support.
11420 While picking a specific @var{cpu-type} will schedule things appropriately
11421 for that particular chip, the compiler will not generate any code that
11422 does not run on the i386 without the @option{-march=@var{cpu-type}} option
11425 @item -march=@var{cpu-type}
11427 Generate instructions for the machine type @var{cpu-type}. The choices
11428 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
11429 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
11431 @item -mcpu=@var{cpu-type}
11433 A deprecated synonym for @option{-mtune}.
11435 @item -mfpmath=@var{unit}
11437 Generate floating point arithmetics for selected unit @var{unit}. The choices
11438 for @var{unit} are:
11442 Use the standard 387 floating point coprocessor present majority of chips and
11443 emulated otherwise. Code compiled with this option will run almost everywhere.
11444 The temporary results are computed in 80bit precision instead of precision
11445 specified by the type resulting in slightly different results compared to most
11446 of other chips. See @option{-ffloat-store} for more detailed description.
11448 This is the default choice for i386 compiler.
11451 Use scalar floating point instructions present in the SSE instruction set.
11452 This instruction set is supported by Pentium3 and newer chips, in the AMD line
11453 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
11454 instruction set supports only single precision arithmetics, thus the double and
11455 extended precision arithmetics is still done using 387. Later version, present
11456 only in Pentium4 and the future AMD x86-64 chips supports double precision
11459 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
11460 or @option{-msse2} switches to enable SSE extensions and make this option
11461 effective. For the x86-64 compiler, these extensions are enabled by default.
11463 The resulting code should be considerably faster in the majority of cases and avoid
11464 the numerical instability problems of 387 code, but may break some existing
11465 code that expects temporaries to be 80bit.
11467 This is the default choice for the x86-64 compiler.
11472 Attempt to utilize both instruction sets at once. This effectively double the
11473 amount of available registers and on chips with separate execution units for
11474 387 and SSE the execution resources too. Use this option with care, as it is
11475 still experimental, because the GCC register allocator does not model separate
11476 functional units well resulting in instable performance.
11479 @item -masm=@var{dialect}
11480 @opindex masm=@var{dialect}
11481 Output asm instructions using selected @var{dialect}. Supported
11482 choices are @samp{intel} or @samp{att} (the default one). Darwin does
11483 not support @samp{intel}.
11486 @itemx -mno-ieee-fp
11488 @opindex mno-ieee-fp
11489 Control whether or not the compiler uses IEEE floating point
11490 comparisons. These handle correctly the case where the result of a
11491 comparison is unordered.
11494 @opindex msoft-float
11495 Generate output containing library calls for floating point.
11496 @strong{Warning:} the requisite libraries are not part of GCC@.
11497 Normally the facilities of the machine's usual C compiler are used, but
11498 this can't be done directly in cross-compilation. You must make your
11499 own arrangements to provide suitable library functions for
11502 On machines where a function returns floating point results in the 80387
11503 register stack, some floating point opcodes may be emitted even if
11504 @option{-msoft-float} is used.
11506 @item -mno-fp-ret-in-387
11507 @opindex mno-fp-ret-in-387
11508 Do not use the FPU registers for return values of functions.
11510 The usual calling convention has functions return values of types
11511 @code{float} and @code{double} in an FPU register, even if there
11512 is no FPU@. The idea is that the operating system should emulate
11515 The option @option{-mno-fp-ret-in-387} causes such values to be returned
11516 in ordinary CPU registers instead.
11518 @item -mno-fancy-math-387
11519 @opindex mno-fancy-math-387
11520 Some 387 emulators do not support the @code{sin}, @code{cos} and
11521 @code{sqrt} instructions for the 387. Specify this option to avoid
11522 generating those instructions. This option is the default on FreeBSD,
11523 OpenBSD and NetBSD@. This option is overridden when @option{-march}
11524 indicates that the target cpu will always have an FPU and so the
11525 instruction will not need emulation. As of revision 2.6.1, these
11526 instructions are not generated unless you also use the
11527 @option{-funsafe-math-optimizations} switch.
11529 @item -malign-double
11530 @itemx -mno-align-double
11531 @opindex malign-double
11532 @opindex mno-align-double
11533 Control whether GCC aligns @code{double}, @code{long double}, and
11534 @code{long long} variables on a two word boundary or a one word
11535 boundary. Aligning @code{double} variables on a two word boundary will
11536 produce code that runs somewhat faster on a @samp{Pentium} at the
11537 expense of more memory.
11539 On x86-64, @option{-malign-double} is enabled by default.
11541 @strong{Warning:} if you use the @option{-malign-double} switch,
11542 structures containing the above types will be aligned differently than
11543 the published application binary interface specifications for the 386
11544 and will not be binary compatible with structures in code compiled
11545 without that switch.
11547 @item -m96bit-long-double
11548 @itemx -m128bit-long-double
11549 @opindex m96bit-long-double
11550 @opindex m128bit-long-double
11551 These switches control the size of @code{long double} type. The i386
11552 application binary interface specifies the size to be 96 bits,
11553 so @option{-m96bit-long-double} is the default in 32 bit mode.
11555 Modern architectures (Pentium and newer) would prefer @code{long double}
11556 to be aligned to an 8 or 16 byte boundary. In arrays or structures
11557 conforming to the ABI, this would not be possible. So specifying a
11558 @option{-m128bit-long-double} will align @code{long double}
11559 to a 16 byte boundary by padding the @code{long double} with an additional
11562 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
11563 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
11565 Notice that neither of these options enable any extra precision over the x87
11566 standard of 80 bits for a @code{long double}.
11568 @strong{Warning:} if you override the default value for your target ABI, the
11569 structures and arrays containing @code{long double} variables will change
11570 their size as well as function calling convention for function taking
11571 @code{long double} will be modified. Hence they will not be binary
11572 compatible with arrays or structures in code compiled without that switch.
11574 @item -mlarge-data-threshold=@var{number}
11575 @opindex mlarge-data-threshold=@var{number}
11576 When @option{-mcmodel=medium} is specified, the data greater than
11577 @var{threshold} are placed in large data section. This value must be the
11578 same across all object linked into the binary and defaults to 65535.
11582 Use a different function-calling convention, in which functions that
11583 take a fixed number of arguments return with the @code{ret} @var{num}
11584 instruction, which pops their arguments while returning. This saves one
11585 instruction in the caller since there is no need to pop the arguments
11588 You can specify that an individual function is called with this calling
11589 sequence with the function attribute @samp{stdcall}. You can also
11590 override the @option{-mrtd} option by using the function attribute
11591 @samp{cdecl}. @xref{Function Attributes}.
11593 @strong{Warning:} this calling convention is incompatible with the one
11594 normally used on Unix, so you cannot use it if you need to call
11595 libraries compiled with the Unix compiler.
11597 Also, you must provide function prototypes for all functions that
11598 take variable numbers of arguments (including @code{printf});
11599 otherwise incorrect code will be generated for calls to those
11602 In addition, seriously incorrect code will result if you call a
11603 function with too many arguments. (Normally, extra arguments are
11604 harmlessly ignored.)
11606 @item -mregparm=@var{num}
11608 Control how many registers are used to pass integer arguments. By
11609 default, no registers are used to pass arguments, and at most 3
11610 registers can be used. You can control this behavior for a specific
11611 function by using the function attribute @samp{regparm}.
11612 @xref{Function Attributes}.
11614 @strong{Warning:} if you use this switch, and
11615 @var{num} is nonzero, then you must build all modules with the same
11616 value, including any libraries. This includes the system libraries and
11620 @opindex msseregparm
11621 Use SSE register passing conventions for float and double arguments
11622 and return values. You can control this behavior for a specific
11623 function by using the function attribute @samp{sseregparm}.
11624 @xref{Function Attributes}.
11626 @strong{Warning:} if you use this switch then you must build all
11627 modules with the same value, including any libraries. This includes
11628 the system libraries and startup modules.
11637 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
11638 is specified, the significands of results of floating-point operations are
11639 rounded to 24 bits (single precision); @option{-mpc64} rounds the
11640 significands of results of floating-point operations to 53 bits (double
11641 precision) and @option{-mpc80} rounds the significands of results of
11642 floating-point operations to 64 bits (extended double precision), which is
11643 the default. When this option is used, floating-point operations in higher
11644 precisions are not available to the programmer without setting the FPU
11645 control word explicitly.
11647 Setting the rounding of floating-point operations to less than the default
11648 80 bits can speed some programs by 2% or more. Note that some mathematical
11649 libraries assume that extended precision (80 bit) floating-point operations
11650 are enabled by default; routines in such libraries could suffer significant
11651 loss of accuracy, typically through so-called "catastrophic cancellation",
11652 when this option is used to set the precision to less than extended precision.
11654 @item -mstackrealign
11655 @opindex mstackrealign
11656 Realign the stack at entry. On the Intel x86, the @option{-mstackrealign}
11657 option will generate an alternate prologue and epilogue that realigns the
11658 runtime stack if necessary. This supports mixing legacy codes that keep
11659 a 4-byte aligned stack with modern codes that keep a 16-byte stack for
11660 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
11661 applicable to individual functions.
11663 @item -mpreferred-stack-boundary=@var{num}
11664 @opindex mpreferred-stack-boundary
11665 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
11666 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
11667 the default is 4 (16 bytes or 128 bits).
11669 @item -mincoming-stack-boundary=@var{num}
11670 @opindex mincoming-stack-boundary
11671 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
11672 boundary. If @option{-mincoming-stack-boundary} is not specified,
11673 the one specified by @option{-mpreferred-stack-boundary} will be used.
11675 On Pentium and PentiumPro, @code{double} and @code{long double} values
11676 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
11677 suffer significant run time performance penalties. On Pentium III, the
11678 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
11679 properly if it is not 16 byte aligned.
11681 To ensure proper alignment of this values on the stack, the stack boundary
11682 must be as aligned as that required by any value stored on the stack.
11683 Further, every function must be generated such that it keeps the stack
11684 aligned. Thus calling a function compiled with a higher preferred
11685 stack boundary from a function compiled with a lower preferred stack
11686 boundary will most likely misalign the stack. It is recommended that
11687 libraries that use callbacks always use the default setting.
11689 This extra alignment does consume extra stack space, and generally
11690 increases code size. Code that is sensitive to stack space usage, such
11691 as embedded systems and operating system kernels, may want to reduce the
11692 preferred alignment to @option{-mpreferred-stack-boundary=2}.
11730 These switches enable or disable the use of instructions in the MMX,
11731 SSE, SSE2, SSE3, SSSE3, SSE4.1, AVX, AES, PCLMUL, SSE4A, ABM or
11732 3DNow!@: extended instruction sets.
11733 These extensions are also available as built-in functions: see
11734 @ref{X86 Built-in Functions}, for details of the functions enabled and
11735 disabled by these switches.
11737 To have SSE/SSE2 instructions generated automatically from floating-point
11738 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
11740 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
11741 generates new AVX instructions or AVX equivalence for all SSEx instructions
11744 These options will enable GCC to use these extended instructions in
11745 generated code, even without @option{-mfpmath=sse}. Applications which
11746 perform runtime CPU detection must compile separate files for each
11747 supported architecture, using the appropriate flags. In particular,
11748 the file containing the CPU detection code should be compiled without
11753 This option instructs GCC to emit a @code{cld} instruction in the prologue
11754 of functions that use string instructions. String instructions depend on
11755 the DF flag to select between autoincrement or autodecrement mode. While the
11756 ABI specifies the DF flag to be cleared on function entry, some operating
11757 systems violate this specification by not clearing the DF flag in their
11758 exception dispatchers. The exception handler can be invoked with the DF flag
11759 set which leads to wrong direction mode, when string instructions are used.
11760 This option can be enabled by default on 32-bit x86 targets by configuring
11761 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
11762 instructions can be suppressed with the @option{-mno-cld} compiler option
11767 This option will enable GCC to use CMPXCHG16B instruction in generated code.
11768 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
11769 data types. This is useful for high resolution counters that could be updated
11770 by multiple processors (or cores). This instruction is generated as part of
11771 atomic built-in functions: see @ref{Atomic Builtins} for details.
11775 This option will enable GCC to use SAHF instruction in generated 64-bit code.
11776 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
11777 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
11778 SAHF are load and store instructions, respectively, for certain status flags.
11779 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
11780 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
11784 This option will enable GCC to use movbe instruction to implement
11785 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
11789 This option will enable built-in functions, @code{__builtin_ia32_crc32qi},
11790 @code{__builtin_ia32_crc32hi}. @code{__builtin_ia32_crc32si} and
11791 @code{__builtin_ia32_crc32di} to generate the crc32 machine instruction.
11795 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
11796 vectorized variants RCPPS and RSQRTPS) with an additional Newton-Raphson step
11797 to increase precision instead of DIVSS and SQRTSS (and their vectorized
11798 variants) for single precision floating point arguments. These instructions
11799 are generated only when @option{-funsafe-math-optimizations} is enabled
11800 together with @option{-finite-math-only} and @option{-fno-trapping-math}.
11801 Note that while the throughput of the sequence is higher than the throughput
11802 of the non-reciprocal instruction, the precision of the sequence can be
11803 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
11805 @item -mveclibabi=@var{type}
11806 @opindex mveclibabi
11807 Specifies the ABI type to use for vectorizing intrinsics using an
11808 external library. Supported types are @code{svml} for the Intel short
11809 vector math library and @code{acml} for the AMD math core library style
11810 of interfacing. GCC will currently emit calls to @code{vmldExp2},
11811 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
11812 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
11813 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
11814 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
11815 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
11816 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
11817 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
11818 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
11819 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
11820 function type when @option{-mveclibabi=svml} is used and @code{__vrd2_sin},
11821 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
11822 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
11823 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
11824 @code{__vrs4_log10f} and @code{__vrs4_powf} for corresponding function type
11825 when @option{-mveclibabi=acml} is used. Both @option{-ftree-vectorize} and
11826 @option{-funsafe-math-optimizations} have to be enabled. A SVML or ACML ABI
11827 compatible library will have to be specified at link time.
11829 @item -mabi=@var{name}
11831 Generate code for the specified calling convention. Permissible values
11832 are: @samp{sysv} for the ABI used on GNU/Linux and other systems and
11833 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
11834 ABI when targeting Windows. On all other systems, the default is the
11835 SYSV ABI. You can control this behavior for a specific function by
11836 using the function attribute @samp{ms_abi}/@samp{sysv_abi}.
11837 @xref{Function Attributes}.
11840 @itemx -mno-push-args
11841 @opindex mpush-args
11842 @opindex mno-push-args
11843 Use PUSH operations to store outgoing parameters. This method is shorter
11844 and usually equally fast as method using SUB/MOV operations and is enabled
11845 by default. In some cases disabling it may improve performance because of
11846 improved scheduling and reduced dependencies.
11848 @item -maccumulate-outgoing-args
11849 @opindex maccumulate-outgoing-args
11850 If enabled, the maximum amount of space required for outgoing arguments will be
11851 computed in the function prologue. This is faster on most modern CPUs
11852 because of reduced dependencies, improved scheduling and reduced stack usage
11853 when preferred stack boundary is not equal to 2. The drawback is a notable
11854 increase in code size. This switch implies @option{-mno-push-args}.
11858 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
11859 on thread-safe exception handling must compile and link all code with the
11860 @option{-mthreads} option. When compiling, @option{-mthreads} defines
11861 @option{-D_MT}; when linking, it links in a special thread helper library
11862 @option{-lmingwthrd} which cleans up per thread exception handling data.
11864 @item -mno-align-stringops
11865 @opindex mno-align-stringops
11866 Do not align destination of inlined string operations. This switch reduces
11867 code size and improves performance in case the destination is already aligned,
11868 but GCC doesn't know about it.
11870 @item -minline-all-stringops
11871 @opindex minline-all-stringops
11872 By default GCC inlines string operations only when destination is known to be
11873 aligned at least to 4 byte boundary. This enables more inlining, increase code
11874 size, but may improve performance of code that depends on fast memcpy, strlen
11875 and memset for short lengths.
11877 @item -minline-stringops-dynamically
11878 @opindex minline-stringops-dynamically
11879 For string operation of unknown size, inline runtime checks so for small
11880 blocks inline code is used, while for large blocks library call is used.
11882 @item -mstringop-strategy=@var{alg}
11883 @opindex mstringop-strategy=@var{alg}
11884 Overwrite internal decision heuristic about particular algorithm to inline
11885 string operation with. The allowed values are @code{rep_byte},
11886 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
11887 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
11888 expanding inline loop, @code{libcall} for always expanding library call.
11890 @item -momit-leaf-frame-pointer
11891 @opindex momit-leaf-frame-pointer
11892 Don't keep the frame pointer in a register for leaf functions. This
11893 avoids the instructions to save, set up and restore frame pointers and
11894 makes an extra register available in leaf functions. The option
11895 @option{-fomit-frame-pointer} removes the frame pointer for all functions
11896 which might make debugging harder.
11898 @item -mtls-direct-seg-refs
11899 @itemx -mno-tls-direct-seg-refs
11900 @opindex mtls-direct-seg-refs
11901 Controls whether TLS variables may be accessed with offsets from the
11902 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
11903 or whether the thread base pointer must be added. Whether or not this
11904 is legal depends on the operating system, and whether it maps the
11905 segment to cover the entire TLS area.
11907 For systems that use GNU libc, the default is on.
11910 @itemx -mno-sse2avx
11912 Specify that the assembler should encode SSE instructions with VEX
11913 prefix. The option @option{-mavx} turns this on by default.
11916 These @samp{-m} switches are supported in addition to the above
11917 on AMD x86-64 processors in 64-bit environments.
11924 Generate code for a 32-bit or 64-bit environment.
11925 The 32-bit environment sets int, long and pointer to 32 bits and
11926 generates code that runs on any i386 system.
11927 The 64-bit environment sets int to 32 bits and long and pointer
11928 to 64 bits and generates code for AMD's x86-64 architecture. For
11929 darwin only the -m64 option turns off the @option{-fno-pic} and
11930 @option{-mdynamic-no-pic} options.
11932 @item -mno-red-zone
11933 @opindex mno-red-zone
11934 Do not use a so called red zone for x86-64 code. The red zone is mandated
11935 by the x86-64 ABI, it is a 128-byte area beyond the location of the
11936 stack pointer that will not be modified by signal or interrupt handlers
11937 and therefore can be used for temporary data without adjusting the stack
11938 pointer. The flag @option{-mno-red-zone} disables this red zone.
11940 @item -mcmodel=small
11941 @opindex mcmodel=small
11942 Generate code for the small code model: the program and its symbols must
11943 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
11944 Programs can be statically or dynamically linked. This is the default
11947 @item -mcmodel=kernel
11948 @opindex mcmodel=kernel
11949 Generate code for the kernel code model. The kernel runs in the
11950 negative 2 GB of the address space.
11951 This model has to be used for Linux kernel code.
11953 @item -mcmodel=medium
11954 @opindex mcmodel=medium
11955 Generate code for the medium model: The program is linked in the lower 2
11956 GB of the address space. Small symbols are also placed there. Symbols
11957 with sizes larger than @option{-mlarge-data-threshold} are put into
11958 large data or bss sections and can be located above 2GB. Programs can
11959 be statically or dynamically linked.
11961 @item -mcmodel=large
11962 @opindex mcmodel=large
11963 Generate code for the large model: This model makes no assumptions
11964 about addresses and sizes of sections.
11967 @node IA-64 Options
11968 @subsection IA-64 Options
11969 @cindex IA-64 Options
11971 These are the @samp{-m} options defined for the Intel IA-64 architecture.
11975 @opindex mbig-endian
11976 Generate code for a big endian target. This is the default for HP-UX@.
11978 @item -mlittle-endian
11979 @opindex mlittle-endian
11980 Generate code for a little endian target. This is the default for AIX5
11986 @opindex mno-gnu-as
11987 Generate (or don't) code for the GNU assembler. This is the default.
11988 @c Also, this is the default if the configure option @option{--with-gnu-as}
11994 @opindex mno-gnu-ld
11995 Generate (or don't) code for the GNU linker. This is the default.
11996 @c Also, this is the default if the configure option @option{--with-gnu-ld}
12001 Generate code that does not use a global pointer register. The result
12002 is not position independent code, and violates the IA-64 ABI@.
12004 @item -mvolatile-asm-stop
12005 @itemx -mno-volatile-asm-stop
12006 @opindex mvolatile-asm-stop
12007 @opindex mno-volatile-asm-stop
12008 Generate (or don't) a stop bit immediately before and after volatile asm
12011 @item -mregister-names
12012 @itemx -mno-register-names
12013 @opindex mregister-names
12014 @opindex mno-register-names
12015 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
12016 the stacked registers. This may make assembler output more readable.
12022 Disable (or enable) optimizations that use the small data section. This may
12023 be useful for working around optimizer bugs.
12025 @item -mconstant-gp
12026 @opindex mconstant-gp
12027 Generate code that uses a single constant global pointer value. This is
12028 useful when compiling kernel code.
12032 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
12033 This is useful when compiling firmware code.
12035 @item -minline-float-divide-min-latency
12036 @opindex minline-float-divide-min-latency
12037 Generate code for inline divides of floating point values
12038 using the minimum latency algorithm.
12040 @item -minline-float-divide-max-throughput
12041 @opindex minline-float-divide-max-throughput
12042 Generate code for inline divides of floating point values
12043 using the maximum throughput algorithm.
12045 @item -mno-inline-float-divide
12046 @opindex mno-inline-float-divide
12047 Do not generate inline code for divides of floating point values.
12049 @item -minline-int-divide-min-latency
12050 @opindex minline-int-divide-min-latency
12051 Generate code for inline divides of integer values
12052 using the minimum latency algorithm.
12054 @item -minline-int-divide-max-throughput
12055 @opindex minline-int-divide-max-throughput
12056 Generate code for inline divides of integer values
12057 using the maximum throughput algorithm.
12059 @item -mno-inline-int-divide
12060 @opindex mno-inline-int-divide
12061 Do not generate inline code for divides of integer values.
12063 @item -minline-sqrt-min-latency
12064 @opindex minline-sqrt-min-latency
12065 Generate code for inline square roots
12066 using the minimum latency algorithm.
12068 @item -minline-sqrt-max-throughput
12069 @opindex minline-sqrt-max-throughput
12070 Generate code for inline square roots
12071 using the maximum throughput algorithm.
12073 @item -mno-inline-sqrt
12074 @opindex mno-inline-sqrt
12075 Do not generate inline code for sqrt.
12078 @itemx -mno-fused-madd
12079 @opindex mfused-madd
12080 @opindex mno-fused-madd
12081 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12082 instructions. The default is to use these instructions.
12084 @item -mno-dwarf2-asm
12085 @itemx -mdwarf2-asm
12086 @opindex mno-dwarf2-asm
12087 @opindex mdwarf2-asm
12088 Don't (or do) generate assembler code for the DWARF2 line number debugging
12089 info. This may be useful when not using the GNU assembler.
12091 @item -mearly-stop-bits
12092 @itemx -mno-early-stop-bits
12093 @opindex mearly-stop-bits
12094 @opindex mno-early-stop-bits
12095 Allow stop bits to be placed earlier than immediately preceding the
12096 instruction that triggered the stop bit. This can improve instruction
12097 scheduling, but does not always do so.
12099 @item -mfixed-range=@var{register-range}
12100 @opindex mfixed-range
12101 Generate code treating the given register range as fixed registers.
12102 A fixed register is one that the register allocator can not use. This is
12103 useful when compiling kernel code. A register range is specified as
12104 two registers separated by a dash. Multiple register ranges can be
12105 specified separated by a comma.
12107 @item -mtls-size=@var{tls-size}
12109 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
12112 @item -mtune=@var{cpu-type}
12114 Tune the instruction scheduling for a particular CPU, Valid values are
12115 itanium, itanium1, merced, itanium2, and mckinley.
12121 Generate code for a 32-bit or 64-bit environment.
12122 The 32-bit environment sets int, long and pointer to 32 bits.
12123 The 64-bit environment sets int to 32 bits and long and pointer
12124 to 64 bits. These are HP-UX specific flags.
12126 @item -mno-sched-br-data-spec
12127 @itemx -msched-br-data-spec
12128 @opindex mno-sched-br-data-spec
12129 @opindex msched-br-data-spec
12130 (Dis/En)able data speculative scheduling before reload.
12131 This will result in generation of the ld.a instructions and
12132 the corresponding check instructions (ld.c / chk.a).
12133 The default is 'disable'.
12135 @item -msched-ar-data-spec
12136 @itemx -mno-sched-ar-data-spec
12137 @opindex msched-ar-data-spec
12138 @opindex mno-sched-ar-data-spec
12139 (En/Dis)able data speculative scheduling after reload.
12140 This will result in generation of the ld.a instructions and
12141 the corresponding check instructions (ld.c / chk.a).
12142 The default is 'enable'.
12144 @item -mno-sched-control-spec
12145 @itemx -msched-control-spec
12146 @opindex mno-sched-control-spec
12147 @opindex msched-control-spec
12148 (Dis/En)able control speculative scheduling. This feature is
12149 available only during region scheduling (i.e.@: before reload).
12150 This will result in generation of the ld.s instructions and
12151 the corresponding check instructions chk.s .
12152 The default is 'disable'.
12154 @item -msched-br-in-data-spec
12155 @itemx -mno-sched-br-in-data-spec
12156 @opindex msched-br-in-data-spec
12157 @opindex mno-sched-br-in-data-spec
12158 (En/Dis)able speculative scheduling of the instructions that
12159 are dependent on the data speculative loads before reload.
12160 This is effective only with @option{-msched-br-data-spec} enabled.
12161 The default is 'enable'.
12163 @item -msched-ar-in-data-spec
12164 @itemx -mno-sched-ar-in-data-spec
12165 @opindex msched-ar-in-data-spec
12166 @opindex mno-sched-ar-in-data-spec
12167 (En/Dis)able speculative scheduling of the instructions that
12168 are dependent on the data speculative loads after reload.
12169 This is effective only with @option{-msched-ar-data-spec} enabled.
12170 The default is 'enable'.
12172 @item -msched-in-control-spec
12173 @itemx -mno-sched-in-control-spec
12174 @opindex msched-in-control-spec
12175 @opindex mno-sched-in-control-spec
12176 (En/Dis)able speculative scheduling of the instructions that
12177 are dependent on the control speculative loads.
12178 This is effective only with @option{-msched-control-spec} enabled.
12179 The default is 'enable'.
12181 @item -mno-sched-prefer-non-data-spec-insns
12182 @itemx -msched-prefer-non-data-spec-insns
12183 @opindex mno-sched-prefer-non-data-spec-insns
12184 @opindex msched-prefer-non-data-spec-insns
12185 If enabled, data speculative instructions will be chosen for schedule
12186 only if there are no other choices at the moment. This will make
12187 the use of the data speculation much more conservative.
12188 The default is 'disable'.
12190 @item -mno-sched-prefer-non-control-spec-insns
12191 @itemx -msched-prefer-non-control-spec-insns
12192 @opindex mno-sched-prefer-non-control-spec-insns
12193 @opindex msched-prefer-non-control-spec-insns
12194 If enabled, control speculative instructions will be chosen for schedule
12195 only if there are no other choices at the moment. This will make
12196 the use of the control speculation much more conservative.
12197 The default is 'disable'.
12199 @item -mno-sched-count-spec-in-critical-path
12200 @itemx -msched-count-spec-in-critical-path
12201 @opindex mno-sched-count-spec-in-critical-path
12202 @opindex msched-count-spec-in-critical-path
12203 If enabled, speculative dependencies will be considered during
12204 computation of the instructions priorities. This will make the use of the
12205 speculation a bit more conservative.
12206 The default is 'disable'.
12208 @item -msched-spec-ldc
12209 @opindex msched-spec-ldc
12210 Use a simple data speculation check. This option is on by default.
12212 @item -msched-control-spec-ldc
12213 @opindex msched-spec-ldc
12214 Use a simple check for control speculation. This option is on by default.
12216 @item -msched-stop-bits-after-every-cycle
12217 @opindex msched-stop-bits-after-every-cycle
12218 Place a stop bit after every cycle when scheduling. This option is on
12221 @item -msched-fp-mem-deps-zero-cost
12222 @opindex msched-fp-mem-deps-zero-cost
12223 Assume that floating-point stores and loads are not likely to cause a conflict
12224 when placed into the same instruction group. This option is disabled by
12227 @item -msel-sched-dont-check-control-spec
12228 @opindex msel-sched-dont-check-control-spec
12229 Generate checks for control speculation in selective scheduling.
12230 This flag is disabled by default.
12232 @item -msched-max-memory-insns=@var{max-insns}
12233 @opindex msched-max-memory-insns
12234 Limit on the number of memory insns per instruction group, giving lower
12235 priority to subsequent memory insns attempting to schedule in the same
12236 instruction group. Frequently useful to prevent cache bank conflicts.
12237 The default value is 1.
12239 @item -msched-max-memory-insns-hard-limit
12240 @opindex msched-max-memory-insns-hard-limit
12241 Disallow more than `msched-max-memory-insns' in instruction group.
12242 Otherwise, limit is `soft' meaning that we would prefer non-memory operations
12243 when limit is reached but may still schedule memory operations.
12247 @node IA-64/VMS Options
12248 @subsection IA-64/VMS Options
12250 These @samp{-m} options are defined for the IA-64/VMS implementations:
12253 @item -mvms-return-codes
12254 @opindex mvms-return-codes
12255 Return VMS condition codes from main. The default is to return POSIX
12256 style condition (e.g.@ error) codes.
12258 @item -mdebug-main=@var{prefix}
12259 @opindex mdebug-main=@var{prefix}
12260 Flag the first routine whose name starts with @var{prefix} as the main
12261 routine for the debugger.
12265 Default to 64bit memory allocation routines.
12269 @subsection M32C Options
12270 @cindex M32C options
12273 @item -mcpu=@var{name}
12275 Select the CPU for which code is generated. @var{name} may be one of
12276 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
12277 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
12278 the M32C/80 series.
12282 Specifies that the program will be run on the simulator. This causes
12283 an alternate runtime library to be linked in which supports, for
12284 example, file I/O@. You must not use this option when generating
12285 programs that will run on real hardware; you must provide your own
12286 runtime library for whatever I/O functions are needed.
12288 @item -memregs=@var{number}
12290 Specifies the number of memory-based pseudo-registers GCC will use
12291 during code generation. These pseudo-registers will be used like real
12292 registers, so there is a tradeoff between GCC's ability to fit the
12293 code into available registers, and the performance penalty of using
12294 memory instead of registers. Note that all modules in a program must
12295 be compiled with the same value for this option. Because of that, you
12296 must not use this option with the default runtime libraries gcc
12301 @node M32R/D Options
12302 @subsection M32R/D Options
12303 @cindex M32R/D options
12305 These @option{-m} options are defined for Renesas M32R/D architectures:
12310 Generate code for the M32R/2@.
12314 Generate code for the M32R/X@.
12318 Generate code for the M32R@. This is the default.
12320 @item -mmodel=small
12321 @opindex mmodel=small
12322 Assume all objects live in the lower 16MB of memory (so that their addresses
12323 can be loaded with the @code{ld24} instruction), and assume all subroutines
12324 are reachable with the @code{bl} instruction.
12325 This is the default.
12327 The addressability of a particular object can be set with the
12328 @code{model} attribute.
12330 @item -mmodel=medium
12331 @opindex mmodel=medium
12332 Assume objects may be anywhere in the 32-bit address space (the compiler
12333 will generate @code{seth/add3} instructions to load their addresses), and
12334 assume all subroutines are reachable with the @code{bl} instruction.
12336 @item -mmodel=large
12337 @opindex mmodel=large
12338 Assume objects may be anywhere in the 32-bit address space (the compiler
12339 will generate @code{seth/add3} instructions to load their addresses), and
12340 assume subroutines may not be reachable with the @code{bl} instruction
12341 (the compiler will generate the much slower @code{seth/add3/jl}
12342 instruction sequence).
12345 @opindex msdata=none
12346 Disable use of the small data area. Variables will be put into
12347 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
12348 @code{section} attribute has been specified).
12349 This is the default.
12351 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
12352 Objects may be explicitly put in the small data area with the
12353 @code{section} attribute using one of these sections.
12355 @item -msdata=sdata
12356 @opindex msdata=sdata
12357 Put small global and static data in the small data area, but do not
12358 generate special code to reference them.
12361 @opindex msdata=use
12362 Put small global and static data in the small data area, and generate
12363 special instructions to reference them.
12367 @cindex smaller data references
12368 Put global and static objects less than or equal to @var{num} bytes
12369 into the small data or bss sections instead of the normal data or bss
12370 sections. The default value of @var{num} is 8.
12371 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
12372 for this option to have any effect.
12374 All modules should be compiled with the same @option{-G @var{num}} value.
12375 Compiling with different values of @var{num} may or may not work; if it
12376 doesn't the linker will give an error message---incorrect code will not be
12381 Makes the M32R specific code in the compiler display some statistics
12382 that might help in debugging programs.
12384 @item -malign-loops
12385 @opindex malign-loops
12386 Align all loops to a 32-byte boundary.
12388 @item -mno-align-loops
12389 @opindex mno-align-loops
12390 Do not enforce a 32-byte alignment for loops. This is the default.
12392 @item -missue-rate=@var{number}
12393 @opindex missue-rate=@var{number}
12394 Issue @var{number} instructions per cycle. @var{number} can only be 1
12397 @item -mbranch-cost=@var{number}
12398 @opindex mbranch-cost=@var{number}
12399 @var{number} can only be 1 or 2. If it is 1 then branches will be
12400 preferred over conditional code, if it is 2, then the opposite will
12403 @item -mflush-trap=@var{number}
12404 @opindex mflush-trap=@var{number}
12405 Specifies the trap number to use to flush the cache. The default is
12406 12. Valid numbers are between 0 and 15 inclusive.
12408 @item -mno-flush-trap
12409 @opindex mno-flush-trap
12410 Specifies that the cache cannot be flushed by using a trap.
12412 @item -mflush-func=@var{name}
12413 @opindex mflush-func=@var{name}
12414 Specifies the name of the operating system function to call to flush
12415 the cache. The default is @emph{_flush_cache}, but a function call
12416 will only be used if a trap is not available.
12418 @item -mno-flush-func
12419 @opindex mno-flush-func
12420 Indicates that there is no OS function for flushing the cache.
12424 @node M680x0 Options
12425 @subsection M680x0 Options
12426 @cindex M680x0 options
12428 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
12429 The default settings depend on which architecture was selected when
12430 the compiler was configured; the defaults for the most common choices
12434 @item -march=@var{arch}
12436 Generate code for a specific M680x0 or ColdFire instruction set
12437 architecture. Permissible values of @var{arch} for M680x0
12438 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
12439 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
12440 architectures are selected according to Freescale's ISA classification
12441 and the permissible values are: @samp{isaa}, @samp{isaaplus},
12442 @samp{isab} and @samp{isac}.
12444 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
12445 code for a ColdFire target. The @var{arch} in this macro is one of the
12446 @option{-march} arguments given above.
12448 When used together, @option{-march} and @option{-mtune} select code
12449 that runs on a family of similar processors but that is optimized
12450 for a particular microarchitecture.
12452 @item -mcpu=@var{cpu}
12454 Generate code for a specific M680x0 or ColdFire processor.
12455 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
12456 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
12457 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
12458 below, which also classifies the CPUs into families:
12460 @multitable @columnfractions 0.20 0.80
12461 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
12462 @item @samp{51} @tab @samp{51} @samp{51ac} @samp{51cn} @samp{51em} @samp{51qe}
12463 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
12464 @item @samp{5206e} @tab @samp{5206e}
12465 @item @samp{5208} @tab @samp{5207} @samp{5208}
12466 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
12467 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
12468 @item @samp{5216} @tab @samp{5214} @samp{5216}
12469 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
12470 @item @samp{5225} @tab @samp{5224} @samp{5225}
12471 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
12472 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
12473 @item @samp{5249} @tab @samp{5249}
12474 @item @samp{5250} @tab @samp{5250}
12475 @item @samp{5271} @tab @samp{5270} @samp{5271}
12476 @item @samp{5272} @tab @samp{5272}
12477 @item @samp{5275} @tab @samp{5274} @samp{5275}
12478 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
12479 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
12480 @item @samp{5307} @tab @samp{5307}
12481 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
12482 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
12483 @item @samp{5407} @tab @samp{5407}
12484 @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}
12487 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
12488 @var{arch} is compatible with @var{cpu}. Other combinations of
12489 @option{-mcpu} and @option{-march} are rejected.
12491 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
12492 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
12493 where the value of @var{family} is given by the table above.
12495 @item -mtune=@var{tune}
12497 Tune the code for a particular microarchitecture, within the
12498 constraints set by @option{-march} and @option{-mcpu}.
12499 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
12500 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
12501 and @samp{cpu32}. The ColdFire microarchitectures
12502 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
12504 You can also use @option{-mtune=68020-40} for code that needs
12505 to run relatively well on 68020, 68030 and 68040 targets.
12506 @option{-mtune=68020-60} is similar but includes 68060 targets
12507 as well. These two options select the same tuning decisions as
12508 @option{-m68020-40} and @option{-m68020-60} respectively.
12510 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
12511 when tuning for 680x0 architecture @var{arch}. It also defines
12512 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
12513 option is used. If gcc is tuning for a range of architectures,
12514 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
12515 it defines the macros for every architecture in the range.
12517 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
12518 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
12519 of the arguments given above.
12525 Generate output for a 68000. This is the default
12526 when the compiler is configured for 68000-based systems.
12527 It is equivalent to @option{-march=68000}.
12529 Use this option for microcontrollers with a 68000 or EC000 core,
12530 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
12534 Generate output for a 68010. This is the default
12535 when the compiler is configured for 68010-based systems.
12536 It is equivalent to @option{-march=68010}.
12542 Generate output for a 68020. This is the default
12543 when the compiler is configured for 68020-based systems.
12544 It is equivalent to @option{-march=68020}.
12548 Generate output for a 68030. This is the default when the compiler is
12549 configured for 68030-based systems. It is equivalent to
12550 @option{-march=68030}.
12554 Generate output for a 68040. This is the default when the compiler is
12555 configured for 68040-based systems. It is equivalent to
12556 @option{-march=68040}.
12558 This option inhibits the use of 68881/68882 instructions that have to be
12559 emulated by software on the 68040. Use this option if your 68040 does not
12560 have code to emulate those instructions.
12564 Generate output for a 68060. This is the default when the compiler is
12565 configured for 68060-based systems. It is equivalent to
12566 @option{-march=68060}.
12568 This option inhibits the use of 68020 and 68881/68882 instructions that
12569 have to be emulated by software on the 68060. Use this option if your 68060
12570 does not have code to emulate those instructions.
12574 Generate output for a CPU32. This is the default
12575 when the compiler is configured for CPU32-based systems.
12576 It is equivalent to @option{-march=cpu32}.
12578 Use this option for microcontrollers with a
12579 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
12580 68336, 68340, 68341, 68349 and 68360.
12584 Generate output for a 520X ColdFire CPU@. This is the default
12585 when the compiler is configured for 520X-based systems.
12586 It is equivalent to @option{-mcpu=5206}, and is now deprecated
12587 in favor of that option.
12589 Use this option for microcontroller with a 5200 core, including
12590 the MCF5202, MCF5203, MCF5204 and MCF5206.
12594 Generate output for a 5206e ColdFire CPU@. The option is now
12595 deprecated in favor of the equivalent @option{-mcpu=5206e}.
12599 Generate output for a member of the ColdFire 528X family.
12600 The option is now deprecated in favor of the equivalent
12601 @option{-mcpu=528x}.
12605 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
12606 in favor of the equivalent @option{-mcpu=5307}.
12610 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
12611 in favor of the equivalent @option{-mcpu=5407}.
12615 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
12616 This includes use of hardware floating point instructions.
12617 The option is equivalent to @option{-mcpu=547x}, and is now
12618 deprecated in favor of that option.
12622 Generate output for a 68040, without using any of the new instructions.
12623 This results in code which can run relatively efficiently on either a
12624 68020/68881 or a 68030 or a 68040. The generated code does use the
12625 68881 instructions that are emulated on the 68040.
12627 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
12631 Generate output for a 68060, without using any of the new instructions.
12632 This results in code which can run relatively efficiently on either a
12633 68020/68881 or a 68030 or a 68040. The generated code does use the
12634 68881 instructions that are emulated on the 68060.
12636 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
12640 @opindex mhard-float
12642 Generate floating-point instructions. This is the default for 68020
12643 and above, and for ColdFire devices that have an FPU@. It defines the
12644 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
12645 on ColdFire targets.
12648 @opindex msoft-float
12649 Do not generate floating-point instructions; use library calls instead.
12650 This is the default for 68000, 68010, and 68832 targets. It is also
12651 the default for ColdFire devices that have no FPU.
12657 Generate (do not generate) ColdFire hardware divide and remainder
12658 instructions. If @option{-march} is used without @option{-mcpu},
12659 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
12660 architectures. Otherwise, the default is taken from the target CPU
12661 (either the default CPU, or the one specified by @option{-mcpu}). For
12662 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
12663 @option{-mcpu=5206e}.
12665 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
12669 Consider type @code{int} to be 16 bits wide, like @code{short int}.
12670 Additionally, parameters passed on the stack are also aligned to a
12671 16-bit boundary even on targets whose API mandates promotion to 32-bit.
12675 Do not consider type @code{int} to be 16 bits wide. This is the default.
12678 @itemx -mno-bitfield
12679 @opindex mnobitfield
12680 @opindex mno-bitfield
12681 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
12682 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
12686 Do use the bit-field instructions. The @option{-m68020} option implies
12687 @option{-mbitfield}. This is the default if you use a configuration
12688 designed for a 68020.
12692 Use a different function-calling convention, in which functions
12693 that take a fixed number of arguments return with the @code{rtd}
12694 instruction, which pops their arguments while returning. This
12695 saves one instruction in the caller since there is no need to pop
12696 the arguments there.
12698 This calling convention is incompatible with the one normally
12699 used on Unix, so you cannot use it if you need to call libraries
12700 compiled with the Unix compiler.
12702 Also, you must provide function prototypes for all functions that
12703 take variable numbers of arguments (including @code{printf});
12704 otherwise incorrect code will be generated for calls to those
12707 In addition, seriously incorrect code will result if you call a
12708 function with too many arguments. (Normally, extra arguments are
12709 harmlessly ignored.)
12711 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
12712 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
12716 Do not use the calling conventions selected by @option{-mrtd}.
12717 This is the default.
12720 @itemx -mno-align-int
12721 @opindex malign-int
12722 @opindex mno-align-int
12723 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
12724 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
12725 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
12726 Aligning variables on 32-bit boundaries produces code that runs somewhat
12727 faster on processors with 32-bit busses at the expense of more memory.
12729 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
12730 align structures containing the above types differently than
12731 most published application binary interface specifications for the m68k.
12735 Use the pc-relative addressing mode of the 68000 directly, instead of
12736 using a global offset table. At present, this option implies @option{-fpic},
12737 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
12738 not presently supported with @option{-mpcrel}, though this could be supported for
12739 68020 and higher processors.
12741 @item -mno-strict-align
12742 @itemx -mstrict-align
12743 @opindex mno-strict-align
12744 @opindex mstrict-align
12745 Do not (do) assume that unaligned memory references will be handled by
12749 Generate code that allows the data segment to be located in a different
12750 area of memory from the text segment. This allows for execute in place in
12751 an environment without virtual memory management. This option implies
12754 @item -mno-sep-data
12755 Generate code that assumes that the data segment follows the text segment.
12756 This is the default.
12758 @item -mid-shared-library
12759 Generate code that supports shared libraries via the library ID method.
12760 This allows for execute in place and shared libraries in an environment
12761 without virtual memory management. This option implies @option{-fPIC}.
12763 @item -mno-id-shared-library
12764 Generate code that doesn't assume ID based shared libraries are being used.
12765 This is the default.
12767 @item -mshared-library-id=n
12768 Specified the identification number of the ID based shared library being
12769 compiled. Specifying a value of 0 will generate more compact code, specifying
12770 other values will force the allocation of that number to the current
12771 library but is no more space or time efficient than omitting this option.
12777 When generating position-independent code for ColdFire, generate code
12778 that works if the GOT has more than 8192 entries. This code is
12779 larger and slower than code generated without this option. On M680x0
12780 processors, this option is not needed; @option{-fPIC} suffices.
12782 GCC normally uses a single instruction to load values from the GOT@.
12783 While this is relatively efficient, it only works if the GOT
12784 is smaller than about 64k. Anything larger causes the linker
12785 to report an error such as:
12787 @cindex relocation truncated to fit (ColdFire)
12789 relocation truncated to fit: R_68K_GOT16O foobar
12792 If this happens, you should recompile your code with @option{-mxgot}.
12793 It should then work with very large GOTs. However, code generated with
12794 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
12795 the value of a global symbol.
12797 Note that some linkers, including newer versions of the GNU linker,
12798 can create multiple GOTs and sort GOT entries. If you have such a linker,
12799 you should only need to use @option{-mxgot} when compiling a single
12800 object file that accesses more than 8192 GOT entries. Very few do.
12802 These options have no effect unless GCC is generating
12803 position-independent code.
12807 @node M68hc1x Options
12808 @subsection M68hc1x Options
12809 @cindex M68hc1x options
12811 These are the @samp{-m} options defined for the 68hc11 and 68hc12
12812 microcontrollers. The default values for these options depends on
12813 which style of microcontroller was selected when the compiler was configured;
12814 the defaults for the most common choices are given below.
12821 Generate output for a 68HC11. This is the default
12822 when the compiler is configured for 68HC11-based systems.
12828 Generate output for a 68HC12. This is the default
12829 when the compiler is configured for 68HC12-based systems.
12835 Generate output for a 68HCS12.
12837 @item -mauto-incdec
12838 @opindex mauto-incdec
12839 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
12846 Enable the use of 68HC12 min and max instructions.
12849 @itemx -mno-long-calls
12850 @opindex mlong-calls
12851 @opindex mno-long-calls
12852 Treat all calls as being far away (near). If calls are assumed to be
12853 far away, the compiler will use the @code{call} instruction to
12854 call a function and the @code{rtc} instruction for returning.
12858 Consider type @code{int} to be 16 bits wide, like @code{short int}.
12860 @item -msoft-reg-count=@var{count}
12861 @opindex msoft-reg-count
12862 Specify the number of pseudo-soft registers which are used for the
12863 code generation. The maximum number is 32. Using more pseudo-soft
12864 register may or may not result in better code depending on the program.
12865 The default is 4 for 68HC11 and 2 for 68HC12.
12869 @node MCore Options
12870 @subsection MCore Options
12871 @cindex MCore options
12873 These are the @samp{-m} options defined for the Motorola M*Core
12879 @itemx -mno-hardlit
12881 @opindex mno-hardlit
12882 Inline constants into the code stream if it can be done in two
12883 instructions or less.
12889 Use the divide instruction. (Enabled by default).
12891 @item -mrelax-immediate
12892 @itemx -mno-relax-immediate
12893 @opindex mrelax-immediate
12894 @opindex mno-relax-immediate
12895 Allow arbitrary sized immediates in bit operations.
12897 @item -mwide-bitfields
12898 @itemx -mno-wide-bitfields
12899 @opindex mwide-bitfields
12900 @opindex mno-wide-bitfields
12901 Always treat bit-fields as int-sized.
12903 @item -m4byte-functions
12904 @itemx -mno-4byte-functions
12905 @opindex m4byte-functions
12906 @opindex mno-4byte-functions
12907 Force all functions to be aligned to a four byte boundary.
12909 @item -mcallgraph-data
12910 @itemx -mno-callgraph-data
12911 @opindex mcallgraph-data
12912 @opindex mno-callgraph-data
12913 Emit callgraph information.
12916 @itemx -mno-slow-bytes
12917 @opindex mslow-bytes
12918 @opindex mno-slow-bytes
12919 Prefer word access when reading byte quantities.
12921 @item -mlittle-endian
12922 @itemx -mbig-endian
12923 @opindex mlittle-endian
12924 @opindex mbig-endian
12925 Generate code for a little endian target.
12931 Generate code for the 210 processor.
12935 Assume that run-time support has been provided and so omit the
12936 simulator library (@file{libsim.a)} from the linker command line.
12938 @item -mstack-increment=@var{size}
12939 @opindex mstack-increment
12940 Set the maximum amount for a single stack increment operation. Large
12941 values can increase the speed of programs which contain functions
12942 that need a large amount of stack space, but they can also trigger a
12943 segmentation fault if the stack is extended too much. The default
12949 @subsection MeP Options
12950 @cindex MeP options
12956 Enables the @code{abs} instruction, which is the absolute difference
12957 between two registers.
12961 Enables all the optional instructions - average, multiply, divide, bit
12962 operations, leading zero, absolute difference, min/max, clip, and
12968 Enables the @code{ave} instruction, which computes the average of two
12971 @item -mbased=@var{n}
12973 Variables of size @var{n} bytes or smaller will be placed in the
12974 @code{.based} section by default. Based variables use the @code{$tp}
12975 register as a base register, and there is a 128 byte limit to the
12976 @code{.based} section.
12980 Enables the bit operation instructions - bit test (@code{btstm}), set
12981 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
12982 test-and-set (@code{tas}).
12984 @item -mc=@var{name}
12986 Selects which section constant data will be placed in. @var{name} may
12987 be @code{tiny}, @code{near}, or @code{far}.
12991 Enables the @code{clip} instruction. Note that @code{-mclip} is not
12992 useful unless you also provide @code{-mminmax}.
12994 @item -mconfig=@var{name}
12996 Selects one of the build-in core configurations. Each MeP chip has
12997 one or more modules in it; each module has a core CPU and a variety of
12998 coprocessors, optional instructions, and peripherals. The
12999 @code{MeP-Integrator} tool, not part of GCC, provides these
13000 configurations through this option; using this option is the same as
13001 using all the corresponding command line options. The default
13002 configuration is @code{default}.
13006 Enables the coprocessor instructions. By default, this is a 32-bit
13007 coprocessor. Note that the coprocessor is normally enabled via the
13008 @code{-mconfig=} option.
13012 Enables the 32-bit coprocessor's instructions.
13016 Enables the 64-bit coprocessor's instructions.
13020 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
13024 Causes constant variables to be placed in the @code{.near} section.
13028 Enables the @code{div} and @code{divu} instructions.
13032 Generate big-endian code.
13036 Generate little-endian code.
13038 @item -mio-volatile
13039 @opindex mio-volatile
13040 Tells the compiler that any variable marked with the @code{io}
13041 attribute is to be considered volatile.
13045 Causes variables to be assigned to the @code{.far} section by default.
13049 Enables the @code{leadz} (leading zero) instruction.
13053 Causes variables to be assigned to the @code{.near} section by default.
13057 Enables the @code{min} and @code{max} instructions.
13061 Enables the multiplication and multiply-accumulate instructions.
13065 Disables all the optional instructions enabled by @code{-mall-opts}.
13069 Enables the @code{repeat} and @code{erepeat} instructions, used for
13070 low-overhead looping.
13074 Causes all variables to default to the @code{.tiny} section. Note
13075 that there is a 65536 byte limit to this section. Accesses to these
13076 variables use the @code{%gp} base register.
13080 Enables the saturation instructions. Note that the compiler does not
13081 currently generate these itself, but this option is included for
13082 compatibility with other tools, like @code{as}.
13086 Link the SDRAM-based runtime instead of the default ROM-based runtime.
13090 Link the simulator runtime libraries.
13094 Link the simulator runtime libraries, excluding built-in support
13095 for reset and exception vectors and tables.
13099 Causes all functions to default to the @code{.far} section. Without
13100 this option, functions default to the @code{.near} section.
13102 @item -mtiny=@var{n}
13104 Variables that are @var{n} bytes or smaller will be allocated to the
13105 @code{.tiny} section. These variables use the @code{$gp} base
13106 register. The default for this option is 4, but note that there's a
13107 65536 byte limit to the @code{.tiny} section.
13112 @subsection MIPS Options
13113 @cindex MIPS options
13119 Generate big-endian code.
13123 Generate little-endian code. This is the default for @samp{mips*el-*-*}
13126 @item -march=@var{arch}
13128 Generate code that will run on @var{arch}, which can be the name of a
13129 generic MIPS ISA, or the name of a particular processor.
13131 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
13132 @samp{mips32}, @samp{mips32r2}, @samp{mips64} and @samp{mips64r2}.
13133 The processor names are:
13134 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
13135 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
13136 @samp{5kc}, @samp{5kf},
13138 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
13139 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
13140 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
13141 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
13142 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
13143 @samp{loongson2e}, @samp{loongson2f},
13147 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
13148 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
13149 @samp{rm7000}, @samp{rm9000},
13150 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
13153 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
13154 @samp{vr5000}, @samp{vr5400}, @samp{vr5500}
13156 The special value @samp{from-abi} selects the
13157 most compatible architecture for the selected ABI (that is,
13158 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
13160 Native Linux/GNU toolchains also support the value @samp{native},
13161 which selects the best architecture option for the host processor.
13162 @option{-march=native} has no effect if GCC does not recognize
13165 In processor names, a final @samp{000} can be abbreviated as @samp{k}
13166 (for example, @samp{-march=r2k}). Prefixes are optional, and
13167 @samp{vr} may be written @samp{r}.
13169 Names of the form @samp{@var{n}f2_1} refer to processors with
13170 FPUs clocked at half the rate of the core, names of the form
13171 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
13172 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
13173 processors with FPUs clocked a ratio of 3:2 with respect to the core.
13174 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
13175 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
13176 accepted as synonyms for @samp{@var{n}f1_1}.
13178 GCC defines two macros based on the value of this option. The first
13179 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
13180 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
13181 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
13182 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
13183 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
13185 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
13186 above. In other words, it will have the full prefix and will not
13187 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
13188 the macro names the resolved architecture (either @samp{"mips1"} or
13189 @samp{"mips3"}). It names the default architecture when no
13190 @option{-march} option is given.
13192 @item -mtune=@var{arch}
13194 Optimize for @var{arch}. Among other things, this option controls
13195 the way instructions are scheduled, and the perceived cost of arithmetic
13196 operations. The list of @var{arch} values is the same as for
13199 When this option is not used, GCC will optimize for the processor
13200 specified by @option{-march}. By using @option{-march} and
13201 @option{-mtune} together, it is possible to generate code that will
13202 run on a family of processors, but optimize the code for one
13203 particular member of that family.
13205 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
13206 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
13207 @samp{-march} ones described above.
13211 Equivalent to @samp{-march=mips1}.
13215 Equivalent to @samp{-march=mips2}.
13219 Equivalent to @samp{-march=mips3}.
13223 Equivalent to @samp{-march=mips4}.
13227 Equivalent to @samp{-march=mips32}.
13231 Equivalent to @samp{-march=mips32r2}.
13235 Equivalent to @samp{-march=mips64}.
13239 Equivalent to @samp{-march=mips64r2}.
13244 @opindex mno-mips16
13245 Generate (do not generate) MIPS16 code. If GCC is targetting a
13246 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
13248 MIPS16 code generation can also be controlled on a per-function basis
13249 by means of @code{mips16} and @code{nomips16} attributes.
13250 @xref{Function Attributes}, for more information.
13252 @item -mflip-mips16
13253 @opindex mflip-mips16
13254 Generate MIPS16 code on alternating functions. This option is provided
13255 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
13256 not intended for ordinary use in compiling user code.
13258 @item -minterlink-mips16
13259 @itemx -mno-interlink-mips16
13260 @opindex minterlink-mips16
13261 @opindex mno-interlink-mips16
13262 Require (do not require) that non-MIPS16 code be link-compatible with
13265 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
13266 it must either use a call or an indirect jump. @option{-minterlink-mips16}
13267 therefore disables direct jumps unless GCC knows that the target of the
13268 jump is not MIPS16.
13280 Generate code for the given ABI@.
13282 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
13283 generates 64-bit code when you select a 64-bit architecture, but you
13284 can use @option{-mgp32} to get 32-bit code instead.
13286 For information about the O64 ABI, see
13287 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
13289 GCC supports a variant of the o32 ABI in which floating-point registers
13290 are 64 rather than 32 bits wide. You can select this combination with
13291 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
13292 and @samp{mfhc1} instructions and is therefore only supported for
13293 MIPS32R2 processors.
13295 The register assignments for arguments and return values remain the
13296 same, but each scalar value is passed in a single 64-bit register
13297 rather than a pair of 32-bit registers. For example, scalar
13298 floating-point values are returned in @samp{$f0} only, not a
13299 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
13300 remains the same, but all 64 bits are saved.
13303 @itemx -mno-abicalls
13305 @opindex mno-abicalls
13306 Generate (do not generate) code that is suitable for SVR4-style
13307 dynamic objects. @option{-mabicalls} is the default for SVR4-based
13312 Generate (do not generate) code that is fully position-independent,
13313 and that can therefore be linked into shared libraries. This option
13314 only affects @option{-mabicalls}.
13316 All @option{-mabicalls} code has traditionally been position-independent,
13317 regardless of options like @option{-fPIC} and @option{-fpic}. However,
13318 as an extension, the GNU toolchain allows executables to use absolute
13319 accesses for locally-binding symbols. It can also use shorter GP
13320 initialization sequences and generate direct calls to locally-defined
13321 functions. This mode is selected by @option{-mno-shared}.
13323 @option{-mno-shared} depends on binutils 2.16 or higher and generates
13324 objects that can only be linked by the GNU linker. However, the option
13325 does not affect the ABI of the final executable; it only affects the ABI
13326 of relocatable objects. Using @option{-mno-shared} will generally make
13327 executables both smaller and quicker.
13329 @option{-mshared} is the default.
13335 Assume (do not assume) that the static and dynamic linkers
13336 support PLTs and copy relocations. This option only affects
13337 @samp{-mno-shared -mabicalls}. For the n64 ABI, this option
13338 has no effect without @samp{-msym32}.
13340 You can make @option{-mplt} the default by configuring
13341 GCC with @option{--with-mips-plt}. The default is
13342 @option{-mno-plt} otherwise.
13348 Lift (do not lift) the usual restrictions on the size of the global
13351 GCC normally uses a single instruction to load values from the GOT@.
13352 While this is relatively efficient, it will only work if the GOT
13353 is smaller than about 64k. Anything larger will cause the linker
13354 to report an error such as:
13356 @cindex relocation truncated to fit (MIPS)
13358 relocation truncated to fit: R_MIPS_GOT16 foobar
13361 If this happens, you should recompile your code with @option{-mxgot}.
13362 It should then work with very large GOTs, although it will also be
13363 less efficient, since it will take three instructions to fetch the
13364 value of a global symbol.
13366 Note that some linkers can create multiple GOTs. If you have such a
13367 linker, you should only need to use @option{-mxgot} when a single object
13368 file accesses more than 64k's worth of GOT entries. Very few do.
13370 These options have no effect unless GCC is generating position
13375 Assume that general-purpose registers are 32 bits wide.
13379 Assume that general-purpose registers are 64 bits wide.
13383 Assume that floating-point registers are 32 bits wide.
13387 Assume that floating-point registers are 64 bits wide.
13390 @opindex mhard-float
13391 Use floating-point coprocessor instructions.
13394 @opindex msoft-float
13395 Do not use floating-point coprocessor instructions. Implement
13396 floating-point calculations using library calls instead.
13398 @item -msingle-float
13399 @opindex msingle-float
13400 Assume that the floating-point coprocessor only supports single-precision
13403 @item -mdouble-float
13404 @opindex mdouble-float
13405 Assume that the floating-point coprocessor supports double-precision
13406 operations. This is the default.
13412 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
13413 implement atomic memory built-in functions. When neither option is
13414 specified, GCC will use the instructions if the target architecture
13417 @option{-mllsc} is useful if the runtime environment can emulate the
13418 instructions and @option{-mno-llsc} can be useful when compiling for
13419 nonstandard ISAs. You can make either option the default by
13420 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
13421 respectively. @option{--with-llsc} is the default for some
13422 configurations; see the installation documentation for details.
13428 Use (do not use) revision 1 of the MIPS DSP ASE@.
13429 @xref{MIPS DSP Built-in Functions}. This option defines the
13430 preprocessor macro @samp{__mips_dsp}. It also defines
13431 @samp{__mips_dsp_rev} to 1.
13437 Use (do not use) revision 2 of the MIPS DSP ASE@.
13438 @xref{MIPS DSP Built-in Functions}. This option defines the
13439 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
13440 It also defines @samp{__mips_dsp_rev} to 2.
13443 @itemx -mno-smartmips
13444 @opindex msmartmips
13445 @opindex mno-smartmips
13446 Use (do not use) the MIPS SmartMIPS ASE.
13448 @item -mpaired-single
13449 @itemx -mno-paired-single
13450 @opindex mpaired-single
13451 @opindex mno-paired-single
13452 Use (do not use) paired-single floating-point instructions.
13453 @xref{MIPS Paired-Single Support}. This option requires
13454 hardware floating-point support to be enabled.
13460 Use (do not use) MIPS Digital Media Extension instructions.
13461 This option can only be used when generating 64-bit code and requires
13462 hardware floating-point support to be enabled.
13467 @opindex mno-mips3d
13468 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
13469 The option @option{-mips3d} implies @option{-mpaired-single}.
13475 Use (do not use) MT Multithreading instructions.
13479 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
13480 an explanation of the default and the way that the pointer size is
13485 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
13487 The default size of @code{int}s, @code{long}s and pointers depends on
13488 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
13489 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
13490 32-bit @code{long}s. Pointers are the same size as @code{long}s,
13491 or the same size as integer registers, whichever is smaller.
13497 Assume (do not assume) that all symbols have 32-bit values, regardless
13498 of the selected ABI@. This option is useful in combination with
13499 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
13500 to generate shorter and faster references to symbolic addresses.
13504 Put definitions of externally-visible data in a small data section
13505 if that data is no bigger than @var{num} bytes. GCC can then access
13506 the data more efficiently; see @option{-mgpopt} for details.
13508 The default @option{-G} option depends on the configuration.
13510 @item -mlocal-sdata
13511 @itemx -mno-local-sdata
13512 @opindex mlocal-sdata
13513 @opindex mno-local-sdata
13514 Extend (do not extend) the @option{-G} behavior to local data too,
13515 such as to static variables in C@. @option{-mlocal-sdata} is the
13516 default for all configurations.
13518 If the linker complains that an application is using too much small data,
13519 you might want to try rebuilding the less performance-critical parts with
13520 @option{-mno-local-sdata}. You might also want to build large
13521 libraries with @option{-mno-local-sdata}, so that the libraries leave
13522 more room for the main program.
13524 @item -mextern-sdata
13525 @itemx -mno-extern-sdata
13526 @opindex mextern-sdata
13527 @opindex mno-extern-sdata
13528 Assume (do not assume) that externally-defined data will be in
13529 a small data section if that data is within the @option{-G} limit.
13530 @option{-mextern-sdata} is the default for all configurations.
13532 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
13533 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
13534 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
13535 is placed in a small data section. If @var{Var} is defined by another
13536 module, you must either compile that module with a high-enough
13537 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
13538 definition. If @var{Var} is common, you must link the application
13539 with a high-enough @option{-G} setting.
13541 The easiest way of satisfying these restrictions is to compile
13542 and link every module with the same @option{-G} option. However,
13543 you may wish to build a library that supports several different
13544 small data limits. You can do this by compiling the library with
13545 the highest supported @option{-G} setting and additionally using
13546 @option{-mno-extern-sdata} to stop the library from making assumptions
13547 about externally-defined data.
13553 Use (do not use) GP-relative accesses for symbols that are known to be
13554 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
13555 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
13558 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
13559 might not hold the value of @code{_gp}. For example, if the code is
13560 part of a library that might be used in a boot monitor, programs that
13561 call boot monitor routines will pass an unknown value in @code{$gp}.
13562 (In such situations, the boot monitor itself would usually be compiled
13563 with @option{-G0}.)
13565 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
13566 @option{-mno-extern-sdata}.
13568 @item -membedded-data
13569 @itemx -mno-embedded-data
13570 @opindex membedded-data
13571 @opindex mno-embedded-data
13572 Allocate variables to the read-only data section first if possible, then
13573 next in the small data section if possible, otherwise in data. This gives
13574 slightly slower code than the default, but reduces the amount of RAM required
13575 when executing, and thus may be preferred for some embedded systems.
13577 @item -muninit-const-in-rodata
13578 @itemx -mno-uninit-const-in-rodata
13579 @opindex muninit-const-in-rodata
13580 @opindex mno-uninit-const-in-rodata
13581 Put uninitialized @code{const} variables in the read-only data section.
13582 This option is only meaningful in conjunction with @option{-membedded-data}.
13584 @item -mcode-readable=@var{setting}
13585 @opindex mcode-readable
13586 Specify whether GCC may generate code that reads from executable sections.
13587 There are three possible settings:
13590 @item -mcode-readable=yes
13591 Instructions may freely access executable sections. This is the
13594 @item -mcode-readable=pcrel
13595 MIPS16 PC-relative load instructions can access executable sections,
13596 but other instructions must not do so. This option is useful on 4KSc
13597 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
13598 It is also useful on processors that can be configured to have a dual
13599 instruction/data SRAM interface and that, like the M4K, automatically
13600 redirect PC-relative loads to the instruction RAM.
13602 @item -mcode-readable=no
13603 Instructions must not access executable sections. This option can be
13604 useful on targets that are configured to have a dual instruction/data
13605 SRAM interface but that (unlike the M4K) do not automatically redirect
13606 PC-relative loads to the instruction RAM.
13609 @item -msplit-addresses
13610 @itemx -mno-split-addresses
13611 @opindex msplit-addresses
13612 @opindex mno-split-addresses
13613 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
13614 relocation operators. This option has been superseded by
13615 @option{-mexplicit-relocs} but is retained for backwards compatibility.
13617 @item -mexplicit-relocs
13618 @itemx -mno-explicit-relocs
13619 @opindex mexplicit-relocs
13620 @opindex mno-explicit-relocs
13621 Use (do not use) assembler relocation operators when dealing with symbolic
13622 addresses. The alternative, selected by @option{-mno-explicit-relocs},
13623 is to use assembler macros instead.
13625 @option{-mexplicit-relocs} is the default if GCC was configured
13626 to use an assembler that supports relocation operators.
13628 @item -mcheck-zero-division
13629 @itemx -mno-check-zero-division
13630 @opindex mcheck-zero-division
13631 @opindex mno-check-zero-division
13632 Trap (do not trap) on integer division by zero.
13634 The default is @option{-mcheck-zero-division}.
13636 @item -mdivide-traps
13637 @itemx -mdivide-breaks
13638 @opindex mdivide-traps
13639 @opindex mdivide-breaks
13640 MIPS systems check for division by zero by generating either a
13641 conditional trap or a break instruction. Using traps results in
13642 smaller code, but is only supported on MIPS II and later. Also, some
13643 versions of the Linux kernel have a bug that prevents trap from
13644 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
13645 allow conditional traps on architectures that support them and
13646 @option{-mdivide-breaks} to force the use of breaks.
13648 The default is usually @option{-mdivide-traps}, but this can be
13649 overridden at configure time using @option{--with-divide=breaks}.
13650 Divide-by-zero checks can be completely disabled using
13651 @option{-mno-check-zero-division}.
13656 @opindex mno-memcpy
13657 Force (do not force) the use of @code{memcpy()} for non-trivial block
13658 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
13659 most constant-sized copies.
13662 @itemx -mno-long-calls
13663 @opindex mlong-calls
13664 @opindex mno-long-calls
13665 Disable (do not disable) use of the @code{jal} instruction. Calling
13666 functions using @code{jal} is more efficient but requires the caller
13667 and callee to be in the same 256 megabyte segment.
13669 This option has no effect on abicalls code. The default is
13670 @option{-mno-long-calls}.
13676 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
13677 instructions, as provided by the R4650 ISA@.
13680 @itemx -mno-fused-madd
13681 @opindex mfused-madd
13682 @opindex mno-fused-madd
13683 Enable (disable) use of the floating point multiply-accumulate
13684 instructions, when they are available. The default is
13685 @option{-mfused-madd}.
13687 When multiply-accumulate instructions are used, the intermediate
13688 product is calculated to infinite precision and is not subject to
13689 the FCSR Flush to Zero bit. This may be undesirable in some
13694 Tell the MIPS assembler to not run its preprocessor over user
13695 assembler files (with a @samp{.s} suffix) when assembling them.
13698 @itemx -mno-fix-r4000
13699 @opindex mfix-r4000
13700 @opindex mno-fix-r4000
13701 Work around certain R4000 CPU errata:
13704 A double-word or a variable shift may give an incorrect result if executed
13705 immediately after starting an integer division.
13707 A double-word or a variable shift may give an incorrect result if executed
13708 while an integer multiplication is in progress.
13710 An integer division may give an incorrect result if started in a delay slot
13711 of a taken branch or a jump.
13715 @itemx -mno-fix-r4400
13716 @opindex mfix-r4400
13717 @opindex mno-fix-r4400
13718 Work around certain R4400 CPU errata:
13721 A double-word or a variable shift may give an incorrect result if executed
13722 immediately after starting an integer division.
13726 @itemx -mno-fix-r10000
13727 @opindex mfix-r10000
13728 @opindex mno-fix-r10000
13729 Work around certain R10000 errata:
13732 @code{ll}/@code{sc} sequences may not behave atomically on revisions
13733 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
13736 This option can only be used if the target architecture supports
13737 branch-likely instructions. @option{-mfix-r10000} is the default when
13738 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
13742 @itemx -mno-fix-vr4120
13743 @opindex mfix-vr4120
13744 Work around certain VR4120 errata:
13747 @code{dmultu} does not always produce the correct result.
13749 @code{div} and @code{ddiv} do not always produce the correct result if one
13750 of the operands is negative.
13752 The workarounds for the division errata rely on special functions in
13753 @file{libgcc.a}. At present, these functions are only provided by
13754 the @code{mips64vr*-elf} configurations.
13756 Other VR4120 errata require a nop to be inserted between certain pairs of
13757 instructions. These errata are handled by the assembler, not by GCC itself.
13760 @opindex mfix-vr4130
13761 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
13762 workarounds are implemented by the assembler rather than by GCC,
13763 although GCC will avoid using @code{mflo} and @code{mfhi} if the
13764 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
13765 instructions are available instead.
13768 @itemx -mno-fix-sb1
13770 Work around certain SB-1 CPU core errata.
13771 (This flag currently works around the SB-1 revision 2
13772 ``F1'' and ``F2'' floating point errata.)
13774 @item -mr10k-cache-barrier=@var{setting}
13775 @opindex mr10k-cache-barrier
13776 Specify whether GCC should insert cache barriers to avoid the
13777 side-effects of speculation on R10K processors.
13779 In common with many processors, the R10K tries to predict the outcome
13780 of a conditional branch and speculatively executes instructions from
13781 the ``taken'' branch. It later aborts these instructions if the
13782 predicted outcome was wrong. However, on the R10K, even aborted
13783 instructions can have side effects.
13785 This problem only affects kernel stores and, depending on the system,
13786 kernel loads. As an example, a speculatively-executed store may load
13787 the target memory into cache and mark the cache line as dirty, even if
13788 the store itself is later aborted. If a DMA operation writes to the
13789 same area of memory before the ``dirty'' line is flushed, the cached
13790 data will overwrite the DMA-ed data. See the R10K processor manual
13791 for a full description, including other potential problems.
13793 One workaround is to insert cache barrier instructions before every memory
13794 access that might be speculatively executed and that might have side
13795 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
13796 controls GCC's implementation of this workaround. It assumes that
13797 aborted accesses to any byte in the following regions will not have
13802 the memory occupied by the current function's stack frame;
13805 the memory occupied by an incoming stack argument;
13808 the memory occupied by an object with a link-time-constant address.
13811 It is the kernel's responsibility to ensure that speculative
13812 accesses to these regions are indeed safe.
13814 If the input program contains a function declaration such as:
13820 then the implementation of @code{foo} must allow @code{j foo} and
13821 @code{jal foo} to be executed speculatively. GCC honors this
13822 restriction for functions it compiles itself. It expects non-GCC
13823 functions (such as hand-written assembly code) to do the same.
13825 The option has three forms:
13828 @item -mr10k-cache-barrier=load-store
13829 Insert a cache barrier before a load or store that might be
13830 speculatively executed and that might have side effects even
13833 @item -mr10k-cache-barrier=store
13834 Insert a cache barrier before a store that might be speculatively
13835 executed and that might have side effects even if aborted.
13837 @item -mr10k-cache-barrier=none
13838 Disable the insertion of cache barriers. This is the default setting.
13841 @item -mflush-func=@var{func}
13842 @itemx -mno-flush-func
13843 @opindex mflush-func
13844 Specifies the function to call to flush the I and D caches, or to not
13845 call any such function. If called, the function must take the same
13846 arguments as the common @code{_flush_func()}, that is, the address of the
13847 memory range for which the cache is being flushed, the size of the
13848 memory range, and the number 3 (to flush both caches). The default
13849 depends on the target GCC was configured for, but commonly is either
13850 @samp{_flush_func} or @samp{__cpu_flush}.
13852 @item mbranch-cost=@var{num}
13853 @opindex mbranch-cost
13854 Set the cost of branches to roughly @var{num} ``simple'' instructions.
13855 This cost is only a heuristic and is not guaranteed to produce
13856 consistent results across releases. A zero cost redundantly selects
13857 the default, which is based on the @option{-mtune} setting.
13859 @item -mbranch-likely
13860 @itemx -mno-branch-likely
13861 @opindex mbranch-likely
13862 @opindex mno-branch-likely
13863 Enable or disable use of Branch Likely instructions, regardless of the
13864 default for the selected architecture. By default, Branch Likely
13865 instructions may be generated if they are supported by the selected
13866 architecture. An exception is for the MIPS32 and MIPS64 architectures
13867 and processors which implement those architectures; for those, Branch
13868 Likely instructions will not be generated by default because the MIPS32
13869 and MIPS64 architectures specifically deprecate their use.
13871 @item -mfp-exceptions
13872 @itemx -mno-fp-exceptions
13873 @opindex mfp-exceptions
13874 Specifies whether FP exceptions are enabled. This affects how we schedule
13875 FP instructions for some processors. The default is that FP exceptions are
13878 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
13879 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
13882 @item -mvr4130-align
13883 @itemx -mno-vr4130-align
13884 @opindex mvr4130-align
13885 The VR4130 pipeline is two-way superscalar, but can only issue two
13886 instructions together if the first one is 8-byte aligned. When this
13887 option is enabled, GCC will align pairs of instructions that it
13888 thinks should execute in parallel.
13890 This option only has an effect when optimizing for the VR4130.
13891 It normally makes code faster, but at the expense of making it bigger.
13892 It is enabled by default at optimization level @option{-O3}.
13897 Enable (disable) generation of @code{synci} instructions on
13898 architectures that support it. The @code{synci} instructions (if
13899 enabled) will be generated when @code{__builtin___clear_cache()} is
13902 This option defaults to @code{-mno-synci}, but the default can be
13903 overridden by configuring with @code{--with-synci}.
13905 When compiling code for single processor systems, it is generally safe
13906 to use @code{synci}. However, on many multi-core (SMP) systems, it
13907 will not invalidate the instruction caches on all cores and may lead
13908 to undefined behavior.
13912 @subsection MMIX Options
13913 @cindex MMIX Options
13915 These options are defined for the MMIX:
13919 @itemx -mno-libfuncs
13921 @opindex mno-libfuncs
13922 Specify that intrinsic library functions are being compiled, passing all
13923 values in registers, no matter the size.
13926 @itemx -mno-epsilon
13928 @opindex mno-epsilon
13929 Generate floating-point comparison instructions that compare with respect
13930 to the @code{rE} epsilon register.
13932 @item -mabi=mmixware
13934 @opindex mabi=mmixware
13936 Generate code that passes function parameters and return values that (in
13937 the called function) are seen as registers @code{$0} and up, as opposed to
13938 the GNU ABI which uses global registers @code{$231} and up.
13940 @item -mzero-extend
13941 @itemx -mno-zero-extend
13942 @opindex mzero-extend
13943 @opindex mno-zero-extend
13944 When reading data from memory in sizes shorter than 64 bits, use (do not
13945 use) zero-extending load instructions by default, rather than
13946 sign-extending ones.
13949 @itemx -mno-knuthdiv
13951 @opindex mno-knuthdiv
13952 Make the result of a division yielding a remainder have the same sign as
13953 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
13954 remainder follows the sign of the dividend. Both methods are
13955 arithmetically valid, the latter being almost exclusively used.
13957 @item -mtoplevel-symbols
13958 @itemx -mno-toplevel-symbols
13959 @opindex mtoplevel-symbols
13960 @opindex mno-toplevel-symbols
13961 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
13962 code can be used with the @code{PREFIX} assembly directive.
13966 Generate an executable in the ELF format, rather than the default
13967 @samp{mmo} format used by the @command{mmix} simulator.
13969 @item -mbranch-predict
13970 @itemx -mno-branch-predict
13971 @opindex mbranch-predict
13972 @opindex mno-branch-predict
13973 Use (do not use) the probable-branch instructions, when static branch
13974 prediction indicates a probable branch.
13976 @item -mbase-addresses
13977 @itemx -mno-base-addresses
13978 @opindex mbase-addresses
13979 @opindex mno-base-addresses
13980 Generate (do not generate) code that uses @emph{base addresses}. Using a
13981 base address automatically generates a request (handled by the assembler
13982 and the linker) for a constant to be set up in a global register. The
13983 register is used for one or more base address requests within the range 0
13984 to 255 from the value held in the register. The generally leads to short
13985 and fast code, but the number of different data items that can be
13986 addressed is limited. This means that a program that uses lots of static
13987 data may require @option{-mno-base-addresses}.
13989 @item -msingle-exit
13990 @itemx -mno-single-exit
13991 @opindex msingle-exit
13992 @opindex mno-single-exit
13993 Force (do not force) generated code to have a single exit point in each
13997 @node MN10300 Options
13998 @subsection MN10300 Options
13999 @cindex MN10300 options
14001 These @option{-m} options are defined for Matsushita MN10300 architectures:
14006 Generate code to avoid bugs in the multiply instructions for the MN10300
14007 processors. This is the default.
14009 @item -mno-mult-bug
14010 @opindex mno-mult-bug
14011 Do not generate code to avoid bugs in the multiply instructions for the
14012 MN10300 processors.
14016 Generate code which uses features specific to the AM33 processor.
14020 Do not generate code which uses features specific to the AM33 processor. This
14023 @item -mreturn-pointer-on-d0
14024 @opindex mreturn-pointer-on-d0
14025 When generating a function which returns a pointer, return the pointer
14026 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
14027 only in a0, and attempts to call such functions without a prototype
14028 would result in errors. Note that this option is on by default; use
14029 @option{-mno-return-pointer-on-d0} to disable it.
14033 Do not link in the C run-time initialization object file.
14037 Indicate to the linker that it should perform a relaxation optimization pass
14038 to shorten branches, calls and absolute memory addresses. This option only
14039 has an effect when used on the command line for the final link step.
14041 This option makes symbolic debugging impossible.
14044 @node PDP-11 Options
14045 @subsection PDP-11 Options
14046 @cindex PDP-11 Options
14048 These options are defined for the PDP-11:
14053 Use hardware FPP floating point. This is the default. (FIS floating
14054 point on the PDP-11/40 is not supported.)
14057 @opindex msoft-float
14058 Do not use hardware floating point.
14062 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
14066 Return floating-point results in memory. This is the default.
14070 Generate code for a PDP-11/40.
14074 Generate code for a PDP-11/45. This is the default.
14078 Generate code for a PDP-11/10.
14080 @item -mbcopy-builtin
14081 @opindex mbcopy-builtin
14082 Use inline @code{movmemhi} patterns for copying memory. This is the
14087 Do not use inline @code{movmemhi} patterns for copying memory.
14093 Use 16-bit @code{int}. This is the default.
14099 Use 32-bit @code{int}.
14102 @itemx -mno-float32
14104 @opindex mno-float32
14105 Use 64-bit @code{float}. This is the default.
14108 @itemx -mno-float64
14110 @opindex mno-float64
14111 Use 32-bit @code{float}.
14115 Use @code{abshi2} pattern. This is the default.
14119 Do not use @code{abshi2} pattern.
14121 @item -mbranch-expensive
14122 @opindex mbranch-expensive
14123 Pretend that branches are expensive. This is for experimenting with
14124 code generation only.
14126 @item -mbranch-cheap
14127 @opindex mbranch-cheap
14128 Do not pretend that branches are expensive. This is the default.
14132 Generate code for a system with split I&D@.
14136 Generate code for a system without split I&D@. This is the default.
14140 Use Unix assembler syntax. This is the default when configured for
14141 @samp{pdp11-*-bsd}.
14145 Use DEC assembler syntax. This is the default when configured for any
14146 PDP-11 target other than @samp{pdp11-*-bsd}.
14149 @node picoChip Options
14150 @subsection picoChip Options
14151 @cindex picoChip options
14153 These @samp{-m} options are defined for picoChip implementations:
14157 @item -mae=@var{ae_type}
14159 Set the instruction set, register set, and instruction scheduling
14160 parameters for array element type @var{ae_type}. Supported values
14161 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
14163 @option{-mae=ANY} selects a completely generic AE type. Code
14164 generated with this option will run on any of the other AE types. The
14165 code will not be as efficient as it would be if compiled for a specific
14166 AE type, and some types of operation (e.g., multiplication) will not
14167 work properly on all types of AE.
14169 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
14170 for compiled code, and is the default.
14172 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
14173 option may suffer from poor performance of byte (char) manipulation,
14174 since the DSP AE does not provide hardware support for byte load/stores.
14176 @item -msymbol-as-address
14177 Enable the compiler to directly use a symbol name as an address in a
14178 load/store instruction, without first loading it into a
14179 register. Typically, the use of this option will generate larger
14180 programs, which run faster than when the option isn't used. However, the
14181 results vary from program to program, so it is left as a user option,
14182 rather than being permanently enabled.
14184 @item -mno-inefficient-warnings
14185 Disables warnings about the generation of inefficient code. These
14186 warnings can be generated, for example, when compiling code which
14187 performs byte-level memory operations on the MAC AE type. The MAC AE has
14188 no hardware support for byte-level memory operations, so all byte
14189 load/stores must be synthesized from word load/store operations. This is
14190 inefficient and a warning will be generated indicating to the programmer
14191 that they should rewrite the code to avoid byte operations, or to target
14192 an AE type which has the necessary hardware support. This option enables
14193 the warning to be turned off.
14197 @node PowerPC Options
14198 @subsection PowerPC Options
14199 @cindex PowerPC options
14201 These are listed under @xref{RS/6000 and PowerPC Options}.
14203 @node RS/6000 and PowerPC Options
14204 @subsection IBM RS/6000 and PowerPC Options
14205 @cindex RS/6000 and PowerPC Options
14206 @cindex IBM RS/6000 and PowerPC Options
14208 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
14215 @itemx -mno-powerpc
14216 @itemx -mpowerpc-gpopt
14217 @itemx -mno-powerpc-gpopt
14218 @itemx -mpowerpc-gfxopt
14219 @itemx -mno-powerpc-gfxopt
14221 @itemx -mno-powerpc64
14225 @itemx -mno-popcntb
14227 @itemx -mno-popcntd
14235 @itemx -mno-hard-dfp
14239 @opindex mno-power2
14241 @opindex mno-powerpc
14242 @opindex mpowerpc-gpopt
14243 @opindex mno-powerpc-gpopt
14244 @opindex mpowerpc-gfxopt
14245 @opindex mno-powerpc-gfxopt
14246 @opindex mpowerpc64
14247 @opindex mno-powerpc64
14251 @opindex mno-popcntb
14253 @opindex mno-popcntd
14259 @opindex mno-mfpgpr
14261 @opindex mno-hard-dfp
14262 GCC supports two related instruction set architectures for the
14263 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
14264 instructions supported by the @samp{rios} chip set used in the original
14265 RS/6000 systems and the @dfn{PowerPC} instruction set is the
14266 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
14267 the IBM 4xx, 6xx, and follow-on microprocessors.
14269 Neither architecture is a subset of the other. However there is a
14270 large common subset of instructions supported by both. An MQ
14271 register is included in processors supporting the POWER architecture.
14273 You use these options to specify which instructions are available on the
14274 processor you are using. The default value of these options is
14275 determined when configuring GCC@. Specifying the
14276 @option{-mcpu=@var{cpu_type}} overrides the specification of these
14277 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
14278 rather than the options listed above.
14280 The @option{-mpower} option allows GCC to generate instructions that
14281 are found only in the POWER architecture and to use the MQ register.
14282 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
14283 to generate instructions that are present in the POWER2 architecture but
14284 not the original POWER architecture.
14286 The @option{-mpowerpc} option allows GCC to generate instructions that
14287 are found only in the 32-bit subset of the PowerPC architecture.
14288 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
14289 GCC to use the optional PowerPC architecture instructions in the
14290 General Purpose group, including floating-point square root. Specifying
14291 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
14292 use the optional PowerPC architecture instructions in the Graphics
14293 group, including floating-point select.
14295 The @option{-mmfcrf} option allows GCC to generate the move from
14296 condition register field instruction implemented on the POWER4
14297 processor and other processors that support the PowerPC V2.01
14299 The @option{-mpopcntb} option allows GCC to generate the popcount and
14300 double precision FP reciprocal estimate instruction implemented on the
14301 POWER5 processor and other processors that support the PowerPC V2.02
14303 The @option{-mpopcntd} option allows GCC to generate the popcount
14304 instruction implemented on the POWER7 processor and other processors
14305 that support the PowerPC V2.06 architecture.
14306 The @option{-mfprnd} option allows GCC to generate the FP round to
14307 integer instructions implemented on the POWER5+ processor and other
14308 processors that support the PowerPC V2.03 architecture.
14309 The @option{-mcmpb} option allows GCC to generate the compare bytes
14310 instruction implemented on the POWER6 processor and other processors
14311 that support the PowerPC V2.05 architecture.
14312 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
14313 general purpose register instructions implemented on the POWER6X
14314 processor and other processors that support the extended PowerPC V2.05
14316 The @option{-mhard-dfp} option allows GCC to generate the decimal floating
14317 point instructions implemented on some POWER processors.
14319 The @option{-mpowerpc64} option allows GCC to generate the additional
14320 64-bit instructions that are found in the full PowerPC64 architecture
14321 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
14322 @option{-mno-powerpc64}.
14324 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
14325 will use only the instructions in the common subset of both
14326 architectures plus some special AIX common-mode calls, and will not use
14327 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
14328 permits GCC to use any instruction from either architecture and to
14329 allow use of the MQ register; specify this for the Motorola MPC601.
14331 @item -mnew-mnemonics
14332 @itemx -mold-mnemonics
14333 @opindex mnew-mnemonics
14334 @opindex mold-mnemonics
14335 Select which mnemonics to use in the generated assembler code. With
14336 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
14337 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
14338 assembler mnemonics defined for the POWER architecture. Instructions
14339 defined in only one architecture have only one mnemonic; GCC uses that
14340 mnemonic irrespective of which of these options is specified.
14342 GCC defaults to the mnemonics appropriate for the architecture in
14343 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
14344 value of these option. Unless you are building a cross-compiler, you
14345 should normally not specify either @option{-mnew-mnemonics} or
14346 @option{-mold-mnemonics}, but should instead accept the default.
14348 @item -mcpu=@var{cpu_type}
14350 Set architecture type, register usage, choice of mnemonics, and
14351 instruction scheduling parameters for machine type @var{cpu_type}.
14352 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
14353 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
14354 @samp{505}, @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
14355 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
14356 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
14357 @samp{860}, @samp{970}, @samp{8540}, @samp{e300c2}, @samp{e300c3},
14358 @samp{e500mc}, @samp{ec603e}, @samp{G3}, @samp{G4}, @samp{G5},
14359 @samp{power}, @samp{power2}, @samp{power3}, @samp{power4},
14360 @samp{power5}, @samp{power5+}, @samp{power6}, @samp{power6x}, @samp{power7}
14361 @samp{common}, @samp{powerpc}, @samp{powerpc64}, @samp{rios},
14362 @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
14364 @option{-mcpu=common} selects a completely generic processor. Code
14365 generated under this option will run on any POWER or PowerPC processor.
14366 GCC will use only the instructions in the common subset of both
14367 architectures, and will not use the MQ register. GCC assumes a generic
14368 processor model for scheduling purposes.
14370 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
14371 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
14372 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
14373 types, with an appropriate, generic processor model assumed for
14374 scheduling purposes.
14376 The other options specify a specific processor. Code generated under
14377 those options will run best on that processor, and may not run at all on
14380 The @option{-mcpu} options automatically enable or disable the
14383 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
14384 -mnew-mnemonics -mpopcntb -mpopcntd -mpower -mpower2 -mpowerpc64 @gol
14385 -mpowerpc-gpopt -mpowerpc-gfxopt -msingle-float -mdouble-float @gol
14386 -msimple-fpu -mstring -mmulhw -mdlmzb -mmfpgpr -mvsx}
14388 The particular options set for any particular CPU will vary between
14389 compiler versions, depending on what setting seems to produce optimal
14390 code for that CPU; it doesn't necessarily reflect the actual hardware's
14391 capabilities. If you wish to set an individual option to a particular
14392 value, you may specify it after the @option{-mcpu} option, like
14393 @samp{-mcpu=970 -mno-altivec}.
14395 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
14396 not enabled or disabled by the @option{-mcpu} option at present because
14397 AIX does not have full support for these options. You may still
14398 enable or disable them individually if you're sure it'll work in your
14401 @item -mtune=@var{cpu_type}
14403 Set the instruction scheduling parameters for machine type
14404 @var{cpu_type}, but do not set the architecture type, register usage, or
14405 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
14406 values for @var{cpu_type} are used for @option{-mtune} as for
14407 @option{-mcpu}. If both are specified, the code generated will use the
14408 architecture, registers, and mnemonics set by @option{-mcpu}, but the
14409 scheduling parameters set by @option{-mtune}.
14415 Generate code to compute division as reciprocal estimate and iterative
14416 refinement, creating opportunities for increased throughput. This
14417 feature requires: optional PowerPC Graphics instruction set for single
14418 precision and FRE instruction for double precision, assuming divides
14419 cannot generate user-visible traps, and the domain values not include
14420 Infinities, denormals or zero denominator.
14423 @itemx -mno-altivec
14425 @opindex mno-altivec
14426 Generate code that uses (does not use) AltiVec instructions, and also
14427 enable the use of built-in functions that allow more direct access to
14428 the AltiVec instruction set. You may also need to set
14429 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
14435 @opindex mno-vrsave
14436 Generate VRSAVE instructions when generating AltiVec code.
14438 @item -mgen-cell-microcode
14439 @opindex mgen-cell-microcode
14440 Generate Cell microcode instructions
14442 @item -mwarn-cell-microcode
14443 @opindex mwarn-cell-microcode
14444 Warning when a Cell microcode instruction is going to emitted. An example
14445 of a Cell microcode instruction is a variable shift.
14448 @opindex msecure-plt
14449 Generate code that allows ld and ld.so to build executables and shared
14450 libraries with non-exec .plt and .got sections. This is a PowerPC
14451 32-bit SYSV ABI option.
14455 Generate code that uses a BSS .plt section that ld.so fills in, and
14456 requires .plt and .got sections that are both writable and executable.
14457 This is a PowerPC 32-bit SYSV ABI option.
14463 This switch enables or disables the generation of ISEL instructions.
14465 @item -misel=@var{yes/no}
14466 This switch has been deprecated. Use @option{-misel} and
14467 @option{-mno-isel} instead.
14473 This switch enables or disables the generation of SPE simd
14479 @opindex mno-paired
14480 This switch enables or disables the generation of PAIRED simd
14483 @item -mspe=@var{yes/no}
14484 This option has been deprecated. Use @option{-mspe} and
14485 @option{-mno-spe} instead.
14491 Generate code that uses (does not use) vector/scalar (VSX)
14492 instructions, and also enable the use of built-in functions that allow
14493 more direct access to the VSX instruction set.
14495 @item -mfloat-gprs=@var{yes/single/double/no}
14496 @itemx -mfloat-gprs
14497 @opindex mfloat-gprs
14498 This switch enables or disables the generation of floating point
14499 operations on the general purpose registers for architectures that
14502 The argument @var{yes} or @var{single} enables the use of
14503 single-precision floating point operations.
14505 The argument @var{double} enables the use of single and
14506 double-precision floating point operations.
14508 The argument @var{no} disables floating point operations on the
14509 general purpose registers.
14511 This option is currently only available on the MPC854x.
14517 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
14518 targets (including GNU/Linux). The 32-bit environment sets int, long
14519 and pointer to 32 bits and generates code that runs on any PowerPC
14520 variant. The 64-bit environment sets int to 32 bits and long and
14521 pointer to 64 bits, and generates code for PowerPC64, as for
14522 @option{-mpowerpc64}.
14525 @itemx -mno-fp-in-toc
14526 @itemx -mno-sum-in-toc
14527 @itemx -mminimal-toc
14529 @opindex mno-fp-in-toc
14530 @opindex mno-sum-in-toc
14531 @opindex mminimal-toc
14532 Modify generation of the TOC (Table Of Contents), which is created for
14533 every executable file. The @option{-mfull-toc} option is selected by
14534 default. In that case, GCC will allocate at least one TOC entry for
14535 each unique non-automatic variable reference in your program. GCC
14536 will also place floating-point constants in the TOC@. However, only
14537 16,384 entries are available in the TOC@.
14539 If you receive a linker error message that saying you have overflowed
14540 the available TOC space, you can reduce the amount of TOC space used
14541 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
14542 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
14543 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
14544 generate code to calculate the sum of an address and a constant at
14545 run-time instead of putting that sum into the TOC@. You may specify one
14546 or both of these options. Each causes GCC to produce very slightly
14547 slower and larger code at the expense of conserving TOC space.
14549 If you still run out of space in the TOC even when you specify both of
14550 these options, specify @option{-mminimal-toc} instead. This option causes
14551 GCC to make only one TOC entry for every file. When you specify this
14552 option, GCC will produce code that is slower and larger but which
14553 uses extremely little TOC space. You may wish to use this option
14554 only on files that contain less frequently executed code.
14560 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
14561 @code{long} type, and the infrastructure needed to support them.
14562 Specifying @option{-maix64} implies @option{-mpowerpc64} and
14563 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
14564 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
14567 @itemx -mno-xl-compat
14568 @opindex mxl-compat
14569 @opindex mno-xl-compat
14570 Produce code that conforms more closely to IBM XL compiler semantics
14571 when using AIX-compatible ABI@. Pass floating-point arguments to
14572 prototyped functions beyond the register save area (RSA) on the stack
14573 in addition to argument FPRs. Do not assume that most significant
14574 double in 128-bit long double value is properly rounded when comparing
14575 values and converting to double. Use XL symbol names for long double
14578 The AIX calling convention was extended but not initially documented to
14579 handle an obscure K&R C case of calling a function that takes the
14580 address of its arguments with fewer arguments than declared. IBM XL
14581 compilers access floating point arguments which do not fit in the
14582 RSA from the stack when a subroutine is compiled without
14583 optimization. Because always storing floating-point arguments on the
14584 stack is inefficient and rarely needed, this option is not enabled by
14585 default and only is necessary when calling subroutines compiled by IBM
14586 XL compilers without optimization.
14590 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
14591 application written to use message passing with special startup code to
14592 enable the application to run. The system must have PE installed in the
14593 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
14594 must be overridden with the @option{-specs=} option to specify the
14595 appropriate directory location. The Parallel Environment does not
14596 support threads, so the @option{-mpe} option and the @option{-pthread}
14597 option are incompatible.
14599 @item -malign-natural
14600 @itemx -malign-power
14601 @opindex malign-natural
14602 @opindex malign-power
14603 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
14604 @option{-malign-natural} overrides the ABI-defined alignment of larger
14605 types, such as floating-point doubles, on their natural size-based boundary.
14606 The option @option{-malign-power} instructs GCC to follow the ABI-specified
14607 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
14609 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
14613 @itemx -mhard-float
14614 @opindex msoft-float
14615 @opindex mhard-float
14616 Generate code that does not use (uses) the floating-point register set.
14617 Software floating point emulation is provided if you use the
14618 @option{-msoft-float} option, and pass the option to GCC when linking.
14620 @item -msingle-float
14621 @itemx -mdouble-float
14622 @opindex msingle-float
14623 @opindex mdouble-float
14624 Generate code for single or double-precision floating point operations.
14625 @option{-mdouble-float} implies @option{-msingle-float}.
14628 @opindex msimple-fpu
14629 Do not generate sqrt and div instructions for hardware floating point unit.
14633 Specify type of floating point unit. Valid values are @var{sp_lite}
14634 (equivalent to -msingle-float -msimple-fpu), @var{dp_lite} (equivalent
14635 to -mdouble-float -msimple-fpu), @var{sp_full} (equivalent to -msingle-float),
14636 and @var{dp_full} (equivalent to -mdouble-float).
14639 @opindex mxilinx-fpu
14640 Perform optimizations for floating point unit on Xilinx PPC 405/440.
14643 @itemx -mno-multiple
14645 @opindex mno-multiple
14646 Generate code that uses (does not use) the load multiple word
14647 instructions and the store multiple word instructions. These
14648 instructions are generated by default on POWER systems, and not
14649 generated on PowerPC systems. Do not use @option{-mmultiple} on little
14650 endian PowerPC systems, since those instructions do not work when the
14651 processor is in little endian mode. The exceptions are PPC740 and
14652 PPC750 which permit the instructions usage in little endian mode.
14657 @opindex mno-string
14658 Generate code that uses (does not use) the load string instructions
14659 and the store string word instructions to save multiple registers and
14660 do small block moves. These instructions are generated by default on
14661 POWER systems, and not generated on PowerPC systems. Do not use
14662 @option{-mstring} on little endian PowerPC systems, since those
14663 instructions do not work when the processor is in little endian mode.
14664 The exceptions are PPC740 and PPC750 which permit the instructions
14665 usage in little endian mode.
14670 @opindex mno-update
14671 Generate code that uses (does not use) the load or store instructions
14672 that update the base register to the address of the calculated memory
14673 location. These instructions are generated by default. If you use
14674 @option{-mno-update}, there is a small window between the time that the
14675 stack pointer is updated and the address of the previous frame is
14676 stored, which means code that walks the stack frame across interrupts or
14677 signals may get corrupted data.
14679 @item -mavoid-indexed-addresses
14680 @item -mno-avoid-indexed-addresses
14681 @opindex mavoid-indexed-addresses
14682 @opindex mno-avoid-indexed-addresses
14683 Generate code that tries to avoid (not avoid) the use of indexed load
14684 or store instructions. These instructions can incur a performance
14685 penalty on Power6 processors in certain situations, such as when
14686 stepping through large arrays that cross a 16M boundary. This option
14687 is enabled by default when targetting Power6 and disabled otherwise.
14690 @itemx -mno-fused-madd
14691 @opindex mfused-madd
14692 @opindex mno-fused-madd
14693 Generate code that uses (does not use) the floating point multiply and
14694 accumulate instructions. These instructions are generated by default if
14695 hardware floating is used.
14701 Generate code that uses (does not use) the half-word multiply and
14702 multiply-accumulate instructions on the IBM 405, 440 and 464 processors.
14703 These instructions are generated by default when targetting those
14710 Generate code that uses (does not use) the string-search @samp{dlmzb}
14711 instruction on the IBM 405, 440 and 464 processors. This instruction is
14712 generated by default when targetting those processors.
14714 @item -mno-bit-align
14716 @opindex mno-bit-align
14717 @opindex mbit-align
14718 On System V.4 and embedded PowerPC systems do not (do) force structures
14719 and unions that contain bit-fields to be aligned to the base type of the
14722 For example, by default a structure containing nothing but 8
14723 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
14724 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
14725 the structure would be aligned to a 1 byte boundary and be one byte in
14728 @item -mno-strict-align
14729 @itemx -mstrict-align
14730 @opindex mno-strict-align
14731 @opindex mstrict-align
14732 On System V.4 and embedded PowerPC systems do not (do) assume that
14733 unaligned memory references will be handled by the system.
14735 @item -mrelocatable
14736 @itemx -mno-relocatable
14737 @opindex mrelocatable
14738 @opindex mno-relocatable
14739 On embedded PowerPC systems generate code that allows (does not allow)
14740 the program to be relocated to a different address at runtime. If you
14741 use @option{-mrelocatable} on any module, all objects linked together must
14742 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
14744 @item -mrelocatable-lib
14745 @itemx -mno-relocatable-lib
14746 @opindex mrelocatable-lib
14747 @opindex mno-relocatable-lib
14748 On embedded PowerPC systems generate code that allows (does not allow)
14749 the program to be relocated to a different address at runtime. Modules
14750 compiled with @option{-mrelocatable-lib} can be linked with either modules
14751 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
14752 with modules compiled with the @option{-mrelocatable} options.
14758 On System V.4 and embedded PowerPC systems do not (do) assume that
14759 register 2 contains a pointer to a global area pointing to the addresses
14760 used in the program.
14763 @itemx -mlittle-endian
14765 @opindex mlittle-endian
14766 On System V.4 and embedded PowerPC systems compile code for the
14767 processor in little endian mode. The @option{-mlittle-endian} option is
14768 the same as @option{-mlittle}.
14771 @itemx -mbig-endian
14773 @opindex mbig-endian
14774 On System V.4 and embedded PowerPC systems compile code for the
14775 processor in big endian mode. The @option{-mbig-endian} option is
14776 the same as @option{-mbig}.
14778 @item -mdynamic-no-pic
14779 @opindex mdynamic-no-pic
14780 On Darwin and Mac OS X systems, compile code so that it is not
14781 relocatable, but that its external references are relocatable. The
14782 resulting code is suitable for applications, but not shared
14785 @item -mprioritize-restricted-insns=@var{priority}
14786 @opindex mprioritize-restricted-insns
14787 This option controls the priority that is assigned to
14788 dispatch-slot restricted instructions during the second scheduling
14789 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
14790 @var{no/highest/second-highest} priority to dispatch slot restricted
14793 @item -msched-costly-dep=@var{dependence_type}
14794 @opindex msched-costly-dep
14795 This option controls which dependences are considered costly
14796 by the target during instruction scheduling. The argument
14797 @var{dependence_type} takes one of the following values:
14798 @var{no}: no dependence is costly,
14799 @var{all}: all dependences are costly,
14800 @var{true_store_to_load}: a true dependence from store to load is costly,
14801 @var{store_to_load}: any dependence from store to load is costly,
14802 @var{number}: any dependence which latency >= @var{number} is costly.
14804 @item -minsert-sched-nops=@var{scheme}
14805 @opindex minsert-sched-nops
14806 This option controls which nop insertion scheme will be used during
14807 the second scheduling pass. The argument @var{scheme} takes one of the
14809 @var{no}: Don't insert nops.
14810 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
14811 according to the scheduler's grouping.
14812 @var{regroup_exact}: Insert nops to force costly dependent insns into
14813 separate groups. Insert exactly as many nops as needed to force an insn
14814 to a new group, according to the estimated processor grouping.
14815 @var{number}: Insert nops to force costly dependent insns into
14816 separate groups. Insert @var{number} nops to force an insn to a new group.
14819 @opindex mcall-sysv
14820 On System V.4 and embedded PowerPC systems compile code using calling
14821 conventions that adheres to the March 1995 draft of the System V
14822 Application Binary Interface, PowerPC processor supplement. This is the
14823 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
14825 @item -mcall-sysv-eabi
14827 @opindex mcall-sysv-eabi
14828 @opindex mcall-eabi
14829 Specify both @option{-mcall-sysv} and @option{-meabi} options.
14831 @item -mcall-sysv-noeabi
14832 @opindex mcall-sysv-noeabi
14833 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
14835 @item -mcall-aixdesc
14837 On System V.4 and embedded PowerPC systems compile code for the AIX
14841 @opindex mcall-linux
14842 On System V.4 and embedded PowerPC systems compile code for the
14843 Linux-based GNU system.
14847 On System V.4 and embedded PowerPC systems compile code for the
14848 Hurd-based GNU system.
14850 @item -mcall-freebsd
14851 @opindex mcall-freebsd
14852 On System V.4 and embedded PowerPC systems compile code for the
14853 FreeBSD operating system.
14855 @item -mcall-netbsd
14856 @opindex mcall-netbsd
14857 On System V.4 and embedded PowerPC systems compile code for the
14858 NetBSD operating system.
14860 @item -mcall-openbsd
14861 @opindex mcall-netbsd
14862 On System V.4 and embedded PowerPC systems compile code for the
14863 OpenBSD operating system.
14865 @item -maix-struct-return
14866 @opindex maix-struct-return
14867 Return all structures in memory (as specified by the AIX ABI)@.
14869 @item -msvr4-struct-return
14870 @opindex msvr4-struct-return
14871 Return structures smaller than 8 bytes in registers (as specified by the
14874 @item -mabi=@var{abi-type}
14876 Extend the current ABI with a particular extension, or remove such extension.
14877 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
14878 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
14882 Extend the current ABI with SPE ABI extensions. This does not change
14883 the default ABI, instead it adds the SPE ABI extensions to the current
14887 @opindex mabi=no-spe
14888 Disable Booke SPE ABI extensions for the current ABI@.
14890 @item -mabi=ibmlongdouble
14891 @opindex mabi=ibmlongdouble
14892 Change the current ABI to use IBM extended precision long double.
14893 This is a PowerPC 32-bit SYSV ABI option.
14895 @item -mabi=ieeelongdouble
14896 @opindex mabi=ieeelongdouble
14897 Change the current ABI to use IEEE extended precision long double.
14898 This is a PowerPC 32-bit Linux ABI option.
14901 @itemx -mno-prototype
14902 @opindex mprototype
14903 @opindex mno-prototype
14904 On System V.4 and embedded PowerPC systems assume that all calls to
14905 variable argument functions are properly prototyped. Otherwise, the
14906 compiler must insert an instruction before every non prototyped call to
14907 set or clear bit 6 of the condition code register (@var{CR}) to
14908 indicate whether floating point values were passed in the floating point
14909 registers in case the function takes a variable arguments. With
14910 @option{-mprototype}, only calls to prototyped variable argument functions
14911 will set or clear the bit.
14915 On embedded PowerPC systems, assume that the startup module is called
14916 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
14917 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
14922 On embedded PowerPC systems, assume that the startup module is called
14923 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
14928 On embedded PowerPC systems, assume that the startup module is called
14929 @file{crt0.o} and the standard C libraries are @file{libads.a} and
14932 @item -myellowknife
14933 @opindex myellowknife
14934 On embedded PowerPC systems, assume that the startup module is called
14935 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
14940 On System V.4 and embedded PowerPC systems, specify that you are
14941 compiling for a VxWorks system.
14945 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
14946 header to indicate that @samp{eabi} extended relocations are used.
14952 On System V.4 and embedded PowerPC systems do (do not) adhere to the
14953 Embedded Applications Binary Interface (eabi) which is a set of
14954 modifications to the System V.4 specifications. Selecting @option{-meabi}
14955 means that the stack is aligned to an 8 byte boundary, a function
14956 @code{__eabi} is called to from @code{main} to set up the eabi
14957 environment, and the @option{-msdata} option can use both @code{r2} and
14958 @code{r13} to point to two separate small data areas. Selecting
14959 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
14960 do not call an initialization function from @code{main}, and the
14961 @option{-msdata} option will only use @code{r13} to point to a single
14962 small data area. The @option{-meabi} option is on by default if you
14963 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
14966 @opindex msdata=eabi
14967 On System V.4 and embedded PowerPC systems, put small initialized
14968 @code{const} global and static data in the @samp{.sdata2} section, which
14969 is pointed to by register @code{r2}. Put small initialized
14970 non-@code{const} global and static data in the @samp{.sdata} section,
14971 which is pointed to by register @code{r13}. Put small uninitialized
14972 global and static data in the @samp{.sbss} section, which is adjacent to
14973 the @samp{.sdata} section. The @option{-msdata=eabi} option is
14974 incompatible with the @option{-mrelocatable} option. The
14975 @option{-msdata=eabi} option also sets the @option{-memb} option.
14978 @opindex msdata=sysv
14979 On System V.4 and embedded PowerPC systems, put small global and static
14980 data in the @samp{.sdata} section, which is pointed to by register
14981 @code{r13}. Put small uninitialized global and static data in the
14982 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
14983 The @option{-msdata=sysv} option is incompatible with the
14984 @option{-mrelocatable} option.
14986 @item -msdata=default
14988 @opindex msdata=default
14990 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
14991 compile code the same as @option{-msdata=eabi}, otherwise compile code the
14992 same as @option{-msdata=sysv}.
14995 @opindex msdata=data
14996 On System V.4 and embedded PowerPC systems, put small global
14997 data in the @samp{.sdata} section. Put small uninitialized global
14998 data in the @samp{.sbss} section. Do not use register @code{r13}
14999 to address small data however. This is the default behavior unless
15000 other @option{-msdata} options are used.
15004 @opindex msdata=none
15006 On embedded PowerPC systems, put all initialized global and static data
15007 in the @samp{.data} section, and all uninitialized data in the
15008 @samp{.bss} section.
15012 @cindex smaller data references (PowerPC)
15013 @cindex .sdata/.sdata2 references (PowerPC)
15014 On embedded PowerPC systems, put global and static items less than or
15015 equal to @var{num} bytes into the small data or bss sections instead of
15016 the normal data or bss section. By default, @var{num} is 8. The
15017 @option{-G @var{num}} switch is also passed to the linker.
15018 All modules should be compiled with the same @option{-G @var{num}} value.
15021 @itemx -mno-regnames
15023 @opindex mno-regnames
15024 On System V.4 and embedded PowerPC systems do (do not) emit register
15025 names in the assembly language output using symbolic forms.
15028 @itemx -mno-longcall
15030 @opindex mno-longcall
15031 By default assume that all calls are far away so that a longer more
15032 expensive calling sequence is required. This is required for calls
15033 further than 32 megabytes (33,554,432 bytes) from the current location.
15034 A short call will be generated if the compiler knows
15035 the call cannot be that far away. This setting can be overridden by
15036 the @code{shortcall} function attribute, or by @code{#pragma
15039 Some linkers are capable of detecting out-of-range calls and generating
15040 glue code on the fly. On these systems, long calls are unnecessary and
15041 generate slower code. As of this writing, the AIX linker can do this,
15042 as can the GNU linker for PowerPC/64. It is planned to add this feature
15043 to the GNU linker for 32-bit PowerPC systems as well.
15045 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
15046 callee, L42'', plus a ``branch island'' (glue code). The two target
15047 addresses represent the callee and the ``branch island''. The
15048 Darwin/PPC linker will prefer the first address and generate a ``bl
15049 callee'' if the PPC ``bl'' instruction will reach the callee directly;
15050 otherwise, the linker will generate ``bl L42'' to call the ``branch
15051 island''. The ``branch island'' is appended to the body of the
15052 calling function; it computes the full 32-bit address of the callee
15055 On Mach-O (Darwin) systems, this option directs the compiler emit to
15056 the glue for every direct call, and the Darwin linker decides whether
15057 to use or discard it.
15059 In the future, we may cause GCC to ignore all longcall specifications
15060 when the linker is known to generate glue.
15062 @item -mtls-markers
15063 @itemx -mno-tls-markers
15064 @opindex mtls-markers
15065 @opindex mno-tls-markers
15066 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
15067 specifying the function argument. The relocation allows ld to
15068 reliably associate function call with argument setup instructions for
15069 TLS optimization, which in turn allows gcc to better schedule the
15074 Adds support for multithreading with the @dfn{pthreads} library.
15075 This option sets flags for both the preprocessor and linker.
15079 @node S/390 and zSeries Options
15080 @subsection S/390 and zSeries Options
15081 @cindex S/390 and zSeries Options
15083 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
15087 @itemx -msoft-float
15088 @opindex mhard-float
15089 @opindex msoft-float
15090 Use (do not use) the hardware floating-point instructions and registers
15091 for floating-point operations. When @option{-msoft-float} is specified,
15092 functions in @file{libgcc.a} will be used to perform floating-point
15093 operations. When @option{-mhard-float} is specified, the compiler
15094 generates IEEE floating-point instructions. This is the default.
15097 @itemx -mno-hard-dfp
15099 @opindex mno-hard-dfp
15100 Use (do not use) the hardware decimal-floating-point instructions for
15101 decimal-floating-point operations. When @option{-mno-hard-dfp} is
15102 specified, functions in @file{libgcc.a} will be used to perform
15103 decimal-floating-point operations. When @option{-mhard-dfp} is
15104 specified, the compiler generates decimal-floating-point hardware
15105 instructions. This is the default for @option{-march=z9-ec} or higher.
15107 @item -mlong-double-64
15108 @itemx -mlong-double-128
15109 @opindex mlong-double-64
15110 @opindex mlong-double-128
15111 These switches control the size of @code{long double} type. A size
15112 of 64bit makes the @code{long double} type equivalent to the @code{double}
15113 type. This is the default.
15116 @itemx -mno-backchain
15117 @opindex mbackchain
15118 @opindex mno-backchain
15119 Store (do not store) the address of the caller's frame as backchain pointer
15120 into the callee's stack frame.
15121 A backchain may be needed to allow debugging using tools that do not understand
15122 DWARF-2 call frame information.
15123 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
15124 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
15125 the backchain is placed into the topmost word of the 96/160 byte register
15128 In general, code compiled with @option{-mbackchain} is call-compatible with
15129 code compiled with @option{-mmo-backchain}; however, use of the backchain
15130 for debugging purposes usually requires that the whole binary is built with
15131 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
15132 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15133 to build a linux kernel use @option{-msoft-float}.
15135 The default is to not maintain the backchain.
15137 @item -mpacked-stack
15138 @itemx -mno-packed-stack
15139 @opindex mpacked-stack
15140 @opindex mno-packed-stack
15141 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
15142 specified, the compiler uses the all fields of the 96/160 byte register save
15143 area only for their default purpose; unused fields still take up stack space.
15144 When @option{-mpacked-stack} is specified, register save slots are densely
15145 packed at the top of the register save area; unused space is reused for other
15146 purposes, allowing for more efficient use of the available stack space.
15147 However, when @option{-mbackchain} is also in effect, the topmost word of
15148 the save area is always used to store the backchain, and the return address
15149 register is always saved two words below the backchain.
15151 As long as the stack frame backchain is not used, code generated with
15152 @option{-mpacked-stack} is call-compatible with code generated with
15153 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
15154 S/390 or zSeries generated code that uses the stack frame backchain at run
15155 time, not just for debugging purposes. Such code is not call-compatible
15156 with code compiled with @option{-mpacked-stack}. Also, note that the
15157 combination of @option{-mbackchain},
15158 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15159 to build a linux kernel use @option{-msoft-float}.
15161 The default is to not use the packed stack layout.
15164 @itemx -mno-small-exec
15165 @opindex msmall-exec
15166 @opindex mno-small-exec
15167 Generate (or do not generate) code using the @code{bras} instruction
15168 to do subroutine calls.
15169 This only works reliably if the total executable size does not
15170 exceed 64k. The default is to use the @code{basr} instruction instead,
15171 which does not have this limitation.
15177 When @option{-m31} is specified, generate code compliant to the
15178 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
15179 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
15180 particular to generate 64-bit instructions. For the @samp{s390}
15181 targets, the default is @option{-m31}, while the @samp{s390x}
15182 targets default to @option{-m64}.
15188 When @option{-mzarch} is specified, generate code using the
15189 instructions available on z/Architecture.
15190 When @option{-mesa} is specified, generate code using the
15191 instructions available on ESA/390. Note that @option{-mesa} is
15192 not possible with @option{-m64}.
15193 When generating code compliant to the GNU/Linux for S/390 ABI,
15194 the default is @option{-mesa}. When generating code compliant
15195 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
15201 Generate (or do not generate) code using the @code{mvcle} instruction
15202 to perform block moves. When @option{-mno-mvcle} is specified,
15203 use a @code{mvc} loop instead. This is the default unless optimizing for
15210 Print (or do not print) additional debug information when compiling.
15211 The default is to not print debug information.
15213 @item -march=@var{cpu-type}
15215 Generate code that will run on @var{cpu-type}, which is the name of a system
15216 representing a certain processor type. Possible values for
15217 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, @samp{z990},
15218 @samp{z9-109}, @samp{z9-ec} and @samp{z10}.
15219 When generating code using the instructions available on z/Architecture,
15220 the default is @option{-march=z900}. Otherwise, the default is
15221 @option{-march=g5}.
15223 @item -mtune=@var{cpu-type}
15225 Tune to @var{cpu-type} everything applicable about the generated code,
15226 except for the ABI and the set of available instructions.
15227 The list of @var{cpu-type} values is the same as for @option{-march}.
15228 The default is the value used for @option{-march}.
15231 @itemx -mno-tpf-trace
15232 @opindex mtpf-trace
15233 @opindex mno-tpf-trace
15234 Generate code that adds (does not add) in TPF OS specific branches to trace
15235 routines in the operating system. This option is off by default, even
15236 when compiling for the TPF OS@.
15239 @itemx -mno-fused-madd
15240 @opindex mfused-madd
15241 @opindex mno-fused-madd
15242 Generate code that uses (does not use) the floating point multiply and
15243 accumulate instructions. These instructions are generated by default if
15244 hardware floating point is used.
15246 @item -mwarn-framesize=@var{framesize}
15247 @opindex mwarn-framesize
15248 Emit a warning if the current function exceeds the given frame size. Because
15249 this is a compile time check it doesn't need to be a real problem when the program
15250 runs. It is intended to identify functions which most probably cause
15251 a stack overflow. It is useful to be used in an environment with limited stack
15252 size e.g.@: the linux kernel.
15254 @item -mwarn-dynamicstack
15255 @opindex mwarn-dynamicstack
15256 Emit a warning if the function calls alloca or uses dynamically
15257 sized arrays. This is generally a bad idea with a limited stack size.
15259 @item -mstack-guard=@var{stack-guard}
15260 @itemx -mstack-size=@var{stack-size}
15261 @opindex mstack-guard
15262 @opindex mstack-size
15263 If these options are provided the s390 back end emits additional instructions in
15264 the function prologue which trigger a trap if the stack size is @var{stack-guard}
15265 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
15266 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
15267 the frame size of the compiled function is chosen.
15268 These options are intended to be used to help debugging stack overflow problems.
15269 The additionally emitted code causes only little overhead and hence can also be
15270 used in production like systems without greater performance degradation. The given
15271 values have to be exact powers of 2 and @var{stack-size} has to be greater than
15272 @var{stack-guard} without exceeding 64k.
15273 In order to be efficient the extra code makes the assumption that the stack starts
15274 at an address aligned to the value given by @var{stack-size}.
15275 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
15278 @node Score Options
15279 @subsection Score Options
15280 @cindex Score Options
15282 These options are defined for Score implementations:
15287 Compile code for big endian mode. This is the default.
15291 Compile code for little endian mode.
15295 Disable generate bcnz instruction.
15299 Enable generate unaligned load and store instruction.
15303 Enable the use of multiply-accumulate instructions. Disabled by default.
15307 Specify the SCORE5 as the target architecture.
15311 Specify the SCORE5U of the target architecture.
15315 Specify the SCORE7 as the target architecture. This is the default.
15319 Specify the SCORE7D as the target architecture.
15323 @subsection SH Options
15325 These @samp{-m} options are defined for the SH implementations:
15330 Generate code for the SH1.
15334 Generate code for the SH2.
15337 Generate code for the SH2e.
15341 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
15342 that the floating-point unit is not used.
15344 @item -m2a-single-only
15345 @opindex m2a-single-only
15346 Generate code for the SH2a-FPU, in such a way that no double-precision
15347 floating point operations are used.
15350 @opindex m2a-single
15351 Generate code for the SH2a-FPU assuming the floating-point unit is in
15352 single-precision mode by default.
15356 Generate code for the SH2a-FPU assuming the floating-point unit is in
15357 double-precision mode by default.
15361 Generate code for the SH3.
15365 Generate code for the SH3e.
15369 Generate code for the SH4 without a floating-point unit.
15371 @item -m4-single-only
15372 @opindex m4-single-only
15373 Generate code for the SH4 with a floating-point unit that only
15374 supports single-precision arithmetic.
15378 Generate code for the SH4 assuming the floating-point unit is in
15379 single-precision mode by default.
15383 Generate code for the SH4.
15387 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
15388 floating-point unit is not used.
15390 @item -m4a-single-only
15391 @opindex m4a-single-only
15392 Generate code for the SH4a, in such a way that no double-precision
15393 floating point operations are used.
15396 @opindex m4a-single
15397 Generate code for the SH4a assuming the floating-point unit is in
15398 single-precision mode by default.
15402 Generate code for the SH4a.
15406 Same as @option{-m4a-nofpu}, except that it implicitly passes
15407 @option{-dsp} to the assembler. GCC doesn't generate any DSP
15408 instructions at the moment.
15412 Compile code for the processor in big endian mode.
15416 Compile code for the processor in little endian mode.
15420 Align doubles at 64-bit boundaries. Note that this changes the calling
15421 conventions, and thus some functions from the standard C library will
15422 not work unless you recompile it first with @option{-mdalign}.
15426 Shorten some address references at link time, when possible; uses the
15427 linker option @option{-relax}.
15431 Use 32-bit offsets in @code{switch} tables. The default is to use
15436 Enable the use of bit manipulation instructions on SH2A.
15440 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
15441 alignment constraints.
15445 Comply with the calling conventions defined by Renesas.
15449 Comply with the calling conventions defined by Renesas.
15453 Comply with the calling conventions defined for GCC before the Renesas
15454 conventions were available. This option is the default for all
15455 targets of the SH toolchain except for @samp{sh-symbianelf}.
15458 @opindex mnomacsave
15459 Mark the @code{MAC} register as call-clobbered, even if
15460 @option{-mhitachi} is given.
15464 Increase IEEE-compliance of floating-point code.
15465 At the moment, this is equivalent to @option{-fno-finite-math-only}.
15466 When generating 16 bit SH opcodes, getting IEEE-conforming results for
15467 comparisons of NANs / infinities incurs extra overhead in every
15468 floating point comparison, therefore the default is set to
15469 @option{-ffinite-math-only}.
15471 @item -minline-ic_invalidate
15472 @opindex minline-ic_invalidate
15473 Inline code to invalidate instruction cache entries after setting up
15474 nested function trampolines.
15475 This option has no effect if -musermode is in effect and the selected
15476 code generation option (e.g. -m4) does not allow the use of the icbi
15478 If the selected code generation option does not allow the use of the icbi
15479 instruction, and -musermode is not in effect, the inlined code will
15480 manipulate the instruction cache address array directly with an associative
15481 write. This not only requires privileged mode, but it will also
15482 fail if the cache line had been mapped via the TLB and has become unmapped.
15486 Dump instruction size and location in the assembly code.
15489 @opindex mpadstruct
15490 This option is deprecated. It pads structures to multiple of 4 bytes,
15491 which is incompatible with the SH ABI@.
15495 Optimize for space instead of speed. Implied by @option{-Os}.
15498 @opindex mprefergot
15499 When generating position-independent code, emit function calls using
15500 the Global Offset Table instead of the Procedure Linkage Table.
15504 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
15505 if the inlined code would not work in user mode.
15506 This is the default when the target is @code{sh-*-linux*}.
15508 @item -multcost=@var{number}
15509 @opindex multcost=@var{number}
15510 Set the cost to assume for a multiply insn.
15512 @item -mdiv=@var{strategy}
15513 @opindex mdiv=@var{strategy}
15514 Set the division strategy to use for SHmedia code. @var{strategy} must be
15515 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
15516 inv:call2, inv:fp .
15517 "fp" performs the operation in floating point. This has a very high latency,
15518 but needs only a few instructions, so it might be a good choice if
15519 your code has enough easily exploitable ILP to allow the compiler to
15520 schedule the floating point instructions together with other instructions.
15521 Division by zero causes a floating point exception.
15522 "inv" uses integer operations to calculate the inverse of the divisor,
15523 and then multiplies the dividend with the inverse. This strategy allows
15524 cse and hoisting of the inverse calculation. Division by zero calculates
15525 an unspecified result, but does not trap.
15526 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
15527 have been found, or if the entire operation has been hoisted to the same
15528 place, the last stages of the inverse calculation are intertwined with the
15529 final multiply to reduce the overall latency, at the expense of using a few
15530 more instructions, and thus offering fewer scheduling opportunities with
15532 "call" calls a library function that usually implements the inv:minlat
15534 This gives high code density for m5-*media-nofpu compilations.
15535 "call2" uses a different entry point of the same library function, where it
15536 assumes that a pointer to a lookup table has already been set up, which
15537 exposes the pointer load to cse / code hoisting optimizations.
15538 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
15539 code generation, but if the code stays unoptimized, revert to the "call",
15540 "call2", or "fp" strategies, respectively. Note that the
15541 potentially-trapping side effect of division by zero is carried by a
15542 separate instruction, so it is possible that all the integer instructions
15543 are hoisted out, but the marker for the side effect stays where it is.
15544 A recombination to fp operations or a call is not possible in that case.
15545 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
15546 that the inverse calculation was nor separated from the multiply, they speed
15547 up division where the dividend fits into 20 bits (plus sign where applicable),
15548 by inserting a test to skip a number of operations in this case; this test
15549 slows down the case of larger dividends. inv20u assumes the case of a such
15550 a small dividend to be unlikely, and inv20l assumes it to be likely.
15552 @item -mdivsi3_libfunc=@var{name}
15553 @opindex mdivsi3_libfunc=@var{name}
15554 Set the name of the library function used for 32 bit signed division to
15555 @var{name}. This only affect the name used in the call and inv:call
15556 division strategies, and the compiler will still expect the same
15557 sets of input/output/clobbered registers as if this option was not present.
15559 @item -mfixed-range=@var{register-range}
15560 @opindex mfixed-range
15561 Generate code treating the given register range as fixed registers.
15562 A fixed register is one that the register allocator can not use. This is
15563 useful when compiling kernel code. A register range is specified as
15564 two registers separated by a dash. Multiple register ranges can be
15565 specified separated by a comma.
15567 @item -madjust-unroll
15568 @opindex madjust-unroll
15569 Throttle unrolling to avoid thrashing target registers.
15570 This option only has an effect if the gcc code base supports the
15571 TARGET_ADJUST_UNROLL_MAX target hook.
15573 @item -mindexed-addressing
15574 @opindex mindexed-addressing
15575 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
15576 This is only safe if the hardware and/or OS implement 32 bit wrap-around
15577 semantics for the indexed addressing mode. The architecture allows the
15578 implementation of processors with 64 bit MMU, which the OS could use to
15579 get 32 bit addressing, but since no current hardware implementation supports
15580 this or any other way to make the indexed addressing mode safe to use in
15581 the 32 bit ABI, the default is -mno-indexed-addressing.
15583 @item -mgettrcost=@var{number}
15584 @opindex mgettrcost=@var{number}
15585 Set the cost assumed for the gettr instruction to @var{number}.
15586 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
15590 Assume pt* instructions won't trap. This will generally generate better
15591 scheduled code, but is unsafe on current hardware. The current architecture
15592 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
15593 This has the unintentional effect of making it unsafe to schedule ptabs /
15594 ptrel before a branch, or hoist it out of a loop. For example,
15595 __do_global_ctors, a part of libgcc that runs constructors at program
15596 startup, calls functions in a list which is delimited by @minus{}1. With the
15597 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
15598 That means that all the constructors will be run a bit quicker, but when
15599 the loop comes to the end of the list, the program crashes because ptabs
15600 loads @minus{}1 into a target register. Since this option is unsafe for any
15601 hardware implementing the current architecture specification, the default
15602 is -mno-pt-fixed. Unless the user specifies a specific cost with
15603 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
15604 this deters register allocation using target registers for storing
15607 @item -minvalid-symbols
15608 @opindex minvalid-symbols
15609 Assume symbols might be invalid. Ordinary function symbols generated by
15610 the compiler will always be valid to load with movi/shori/ptabs or
15611 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
15612 to generate symbols that will cause ptabs / ptrel to trap.
15613 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
15614 It will then prevent cross-basic-block cse, hoisting and most scheduling
15615 of symbol loads. The default is @option{-mno-invalid-symbols}.
15618 @node SPARC Options
15619 @subsection SPARC Options
15620 @cindex SPARC options
15622 These @samp{-m} options are supported on the SPARC:
15625 @item -mno-app-regs
15627 @opindex mno-app-regs
15629 Specify @option{-mapp-regs} to generate output using the global registers
15630 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
15633 To be fully SVR4 ABI compliant at the cost of some performance loss,
15634 specify @option{-mno-app-regs}. You should compile libraries and system
15635 software with this option.
15638 @itemx -mhard-float
15640 @opindex mhard-float
15641 Generate output containing floating point instructions. This is the
15645 @itemx -msoft-float
15647 @opindex msoft-float
15648 Generate output containing library calls for floating point.
15649 @strong{Warning:} the requisite libraries are not available for all SPARC
15650 targets. Normally the facilities of the machine's usual C compiler are
15651 used, but this cannot be done directly in cross-compilation. You must make
15652 your own arrangements to provide suitable library functions for
15653 cross-compilation. The embedded targets @samp{sparc-*-aout} and
15654 @samp{sparclite-*-*} do provide software floating point support.
15656 @option{-msoft-float} changes the calling convention in the output file;
15657 therefore, it is only useful if you compile @emph{all} of a program with
15658 this option. In particular, you need to compile @file{libgcc.a}, the
15659 library that comes with GCC, with @option{-msoft-float} in order for
15662 @item -mhard-quad-float
15663 @opindex mhard-quad-float
15664 Generate output containing quad-word (long double) floating point
15667 @item -msoft-quad-float
15668 @opindex msoft-quad-float
15669 Generate output containing library calls for quad-word (long double)
15670 floating point instructions. The functions called are those specified
15671 in the SPARC ABI@. This is the default.
15673 As of this writing, there are no SPARC implementations that have hardware
15674 support for the quad-word floating point instructions. They all invoke
15675 a trap handler for one of these instructions, and then the trap handler
15676 emulates the effect of the instruction. Because of the trap handler overhead,
15677 this is much slower than calling the ABI library routines. Thus the
15678 @option{-msoft-quad-float} option is the default.
15680 @item -mno-unaligned-doubles
15681 @itemx -munaligned-doubles
15682 @opindex mno-unaligned-doubles
15683 @opindex munaligned-doubles
15684 Assume that doubles have 8 byte alignment. This is the default.
15686 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
15687 alignment only if they are contained in another type, or if they have an
15688 absolute address. Otherwise, it assumes they have 4 byte alignment.
15689 Specifying this option avoids some rare compatibility problems with code
15690 generated by other compilers. It is not the default because it results
15691 in a performance loss, especially for floating point code.
15693 @item -mno-faster-structs
15694 @itemx -mfaster-structs
15695 @opindex mno-faster-structs
15696 @opindex mfaster-structs
15697 With @option{-mfaster-structs}, the compiler assumes that structures
15698 should have 8 byte alignment. This enables the use of pairs of
15699 @code{ldd} and @code{std} instructions for copies in structure
15700 assignment, in place of twice as many @code{ld} and @code{st} pairs.
15701 However, the use of this changed alignment directly violates the SPARC
15702 ABI@. Thus, it's intended only for use on targets where the developer
15703 acknowledges that their resulting code will not be directly in line with
15704 the rules of the ABI@.
15706 @item -mimpure-text
15707 @opindex mimpure-text
15708 @option{-mimpure-text}, used in addition to @option{-shared}, tells
15709 the compiler to not pass @option{-z text} to the linker when linking a
15710 shared object. Using this option, you can link position-dependent
15711 code into a shared object.
15713 @option{-mimpure-text} suppresses the ``relocations remain against
15714 allocatable but non-writable sections'' linker error message.
15715 However, the necessary relocations will trigger copy-on-write, and the
15716 shared object is not actually shared across processes. Instead of
15717 using @option{-mimpure-text}, you should compile all source code with
15718 @option{-fpic} or @option{-fPIC}.
15720 This option is only available on SunOS and Solaris.
15722 @item -mcpu=@var{cpu_type}
15724 Set the instruction set, register set, and instruction scheduling parameters
15725 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
15726 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
15727 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
15728 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
15729 @samp{ultrasparc3}, @samp{niagara} and @samp{niagara2}.
15731 Default instruction scheduling parameters are used for values that select
15732 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
15733 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
15735 Here is a list of each supported architecture and their supported
15740 v8: supersparc, hypersparc
15741 sparclite: f930, f934, sparclite86x
15743 v9: ultrasparc, ultrasparc3, niagara, niagara2
15746 By default (unless configured otherwise), GCC generates code for the V7
15747 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
15748 additionally optimizes it for the Cypress CY7C602 chip, as used in the
15749 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
15750 SPARCStation 1, 2, IPX etc.
15752 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
15753 architecture. The only difference from V7 code is that the compiler emits
15754 the integer multiply and integer divide instructions which exist in SPARC-V8
15755 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
15756 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
15759 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
15760 the SPARC architecture. This adds the integer multiply, integer divide step
15761 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
15762 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
15763 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
15764 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
15765 MB86934 chip, which is the more recent SPARClite with FPU@.
15767 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
15768 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
15769 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
15770 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
15771 optimizes it for the TEMIC SPARClet chip.
15773 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
15774 architecture. This adds 64-bit integer and floating-point move instructions,
15775 3 additional floating-point condition code registers and conditional move
15776 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
15777 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
15778 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
15779 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
15780 @option{-mcpu=niagara}, the compiler additionally optimizes it for
15781 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
15782 additionally optimizes it for Sun UltraSPARC T2 chips.
15784 @item -mtune=@var{cpu_type}
15786 Set the instruction scheduling parameters for machine type
15787 @var{cpu_type}, but do not set the instruction set or register set that the
15788 option @option{-mcpu=@var{cpu_type}} would.
15790 The same values for @option{-mcpu=@var{cpu_type}} can be used for
15791 @option{-mtune=@var{cpu_type}}, but the only useful values are those
15792 that select a particular cpu implementation. Those are @samp{cypress},
15793 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
15794 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
15795 @samp{ultrasparc3}, @samp{niagara}, and @samp{niagara2}.
15800 @opindex mno-v8plus
15801 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
15802 difference from the V8 ABI is that the global and out registers are
15803 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
15804 mode for all SPARC-V9 processors.
15810 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
15811 Visual Instruction Set extensions. The default is @option{-mno-vis}.
15814 These @samp{-m} options are supported in addition to the above
15815 on SPARC-V9 processors in 64-bit environments:
15818 @item -mlittle-endian
15819 @opindex mlittle-endian
15820 Generate code for a processor running in little-endian mode. It is only
15821 available for a few configurations and most notably not on Solaris and Linux.
15827 Generate code for a 32-bit or 64-bit environment.
15828 The 32-bit environment sets int, long and pointer to 32 bits.
15829 The 64-bit environment sets int to 32 bits and long and pointer
15832 @item -mcmodel=medlow
15833 @opindex mcmodel=medlow
15834 Generate code for the Medium/Low code model: 64-bit addresses, programs
15835 must be linked in the low 32 bits of memory. Programs can be statically
15836 or dynamically linked.
15838 @item -mcmodel=medmid
15839 @opindex mcmodel=medmid
15840 Generate code for the Medium/Middle code model: 64-bit addresses, programs
15841 must be linked in the low 44 bits of memory, the text and data segments must
15842 be less than 2GB in size and the data segment must be located within 2GB of
15845 @item -mcmodel=medany
15846 @opindex mcmodel=medany
15847 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
15848 may be linked anywhere in memory, the text and data segments must be less
15849 than 2GB in size and the data segment must be located within 2GB of the
15852 @item -mcmodel=embmedany
15853 @opindex mcmodel=embmedany
15854 Generate code for the Medium/Anywhere code model for embedded systems:
15855 64-bit addresses, the text and data segments must be less than 2GB in
15856 size, both starting anywhere in memory (determined at link time). The
15857 global register %g4 points to the base of the data segment. Programs
15858 are statically linked and PIC is not supported.
15861 @itemx -mno-stack-bias
15862 @opindex mstack-bias
15863 @opindex mno-stack-bias
15864 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
15865 frame pointer if present, are offset by @minus{}2047 which must be added back
15866 when making stack frame references. This is the default in 64-bit mode.
15867 Otherwise, assume no such offset is present.
15870 These switches are supported in addition to the above on Solaris:
15875 Add support for multithreading using the Solaris threads library. This
15876 option sets flags for both the preprocessor and linker. This option does
15877 not affect the thread safety of object code produced by the compiler or
15878 that of libraries supplied with it.
15882 Add support for multithreading using the POSIX threads library. This
15883 option sets flags for both the preprocessor and linker. This option does
15884 not affect the thread safety of object code produced by the compiler or
15885 that of libraries supplied with it.
15889 This is a synonym for @option{-pthreads}.
15893 @subsection SPU Options
15894 @cindex SPU options
15896 These @samp{-m} options are supported on the SPU:
15900 @itemx -merror-reloc
15901 @opindex mwarn-reloc
15902 @opindex merror-reloc
15904 The loader for SPU does not handle dynamic relocations. By default, GCC
15905 will give an error when it generates code that requires a dynamic
15906 relocation. @option{-mno-error-reloc} disables the error,
15907 @option{-mwarn-reloc} will generate a warning instead.
15910 @itemx -munsafe-dma
15912 @opindex munsafe-dma
15914 Instructions which initiate or test completion of DMA must not be
15915 reordered with respect to loads and stores of the memory which is being
15916 accessed. Users typically address this problem using the volatile
15917 keyword, but that can lead to inefficient code in places where the
15918 memory is known to not change. Rather than mark the memory as volatile
15919 we treat the DMA instructions as potentially effecting all memory. With
15920 @option{-munsafe-dma} users must use the volatile keyword to protect
15923 @item -mbranch-hints
15924 @opindex mbranch-hints
15926 By default, GCC will generate a branch hint instruction to avoid
15927 pipeline stalls for always taken or probably taken branches. A hint
15928 will not be generated closer than 8 instructions away from its branch.
15929 There is little reason to disable them, except for debugging purposes,
15930 or to make an object a little bit smaller.
15934 @opindex msmall-mem
15935 @opindex mlarge-mem
15937 By default, GCC generates code assuming that addresses are never larger
15938 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
15939 a full 32 bit address.
15944 By default, GCC links against startup code that assumes the SPU-style
15945 main function interface (which has an unconventional parameter list).
15946 With @option{-mstdmain}, GCC will link your program against startup
15947 code that assumes a C99-style interface to @code{main}, including a
15948 local copy of @code{argv} strings.
15950 @item -mfixed-range=@var{register-range}
15951 @opindex mfixed-range
15952 Generate code treating the given register range as fixed registers.
15953 A fixed register is one that the register allocator can not use. This is
15954 useful when compiling kernel code. A register range is specified as
15955 two registers separated by a dash. Multiple register ranges can be
15956 specified separated by a comma.
15959 @itemx -mdual-nops=@var{n}
15960 @opindex mdual-nops
15961 By default, GCC will insert nops to increase dual issue when it expects
15962 it to increase performance. @var{n} can be a value from 0 to 10. A
15963 smaller @var{n} will insert fewer nops. 10 is the default, 0 is the
15964 same as @option{-mno-dual-nops}. Disabled with @option{-Os}.
15966 @item -mhint-max-nops=@var{n}
15967 @opindex mhint-max-nops
15968 Maximum number of nops to insert for a branch hint. A branch hint must
15969 be at least 8 instructions away from the branch it is effecting. GCC
15970 will insert up to @var{n} nops to enforce this, otherwise it will not
15971 generate the branch hint.
15973 @item -mhint-max-distance=@var{n}
15974 @opindex mhint-max-distance
15975 The encoding of the branch hint instruction limits the hint to be within
15976 256 instructions of the branch it is effecting. By default, GCC makes
15977 sure it is within 125.
15980 @opindex msafe-hints
15981 Work around a hardware bug which causes the SPU to stall indefinitely.
15982 By default, GCC will insert the @code{hbrp} instruction to make sure
15983 this stall won't happen.
15987 @node System V Options
15988 @subsection Options for System V
15990 These additional options are available on System V Release 4 for
15991 compatibility with other compilers on those systems:
15996 Create a shared object.
15997 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
16001 Identify the versions of each tool used by the compiler, in a
16002 @code{.ident} assembler directive in the output.
16006 Refrain from adding @code{.ident} directives to the output file (this is
16009 @item -YP,@var{dirs}
16011 Search the directories @var{dirs}, and no others, for libraries
16012 specified with @option{-l}.
16014 @item -Ym,@var{dir}
16016 Look in the directory @var{dir} to find the M4 preprocessor.
16017 The assembler uses this option.
16018 @c This is supposed to go with a -Yd for predefined M4 macro files, but
16019 @c the generic assembler that comes with Solaris takes just -Ym.
16023 @subsection V850 Options
16024 @cindex V850 Options
16026 These @samp{-m} options are defined for V850 implementations:
16030 @itemx -mno-long-calls
16031 @opindex mlong-calls
16032 @opindex mno-long-calls
16033 Treat all calls as being far away (near). If calls are assumed to be
16034 far away, the compiler will always load the functions address up into a
16035 register, and call indirect through the pointer.
16041 Do not optimize (do optimize) basic blocks that use the same index
16042 pointer 4 or more times to copy pointer into the @code{ep} register, and
16043 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
16044 option is on by default if you optimize.
16046 @item -mno-prolog-function
16047 @itemx -mprolog-function
16048 @opindex mno-prolog-function
16049 @opindex mprolog-function
16050 Do not use (do use) external functions to save and restore registers
16051 at the prologue and epilogue of a function. The external functions
16052 are slower, but use less code space if more than one function saves
16053 the same number of registers. The @option{-mprolog-function} option
16054 is on by default if you optimize.
16058 Try to make the code as small as possible. At present, this just turns
16059 on the @option{-mep} and @option{-mprolog-function} options.
16061 @item -mtda=@var{n}
16063 Put static or global variables whose size is @var{n} bytes or less into
16064 the tiny data area that register @code{ep} points to. The tiny data
16065 area can hold up to 256 bytes in total (128 bytes for byte references).
16067 @item -msda=@var{n}
16069 Put static or global variables whose size is @var{n} bytes or less into
16070 the small data area that register @code{gp} points to. The small data
16071 area can hold up to 64 kilobytes.
16073 @item -mzda=@var{n}
16075 Put static or global variables whose size is @var{n} bytes or less into
16076 the first 32 kilobytes of memory.
16080 Specify that the target processor is the V850.
16083 @opindex mbig-switch
16084 Generate code suitable for big switch tables. Use this option only if
16085 the assembler/linker complain about out of range branches within a switch
16090 This option will cause r2 and r5 to be used in the code generated by
16091 the compiler. This setting is the default.
16093 @item -mno-app-regs
16094 @opindex mno-app-regs
16095 This option will cause r2 and r5 to be treated as fixed registers.
16099 Specify that the target processor is the V850E1. The preprocessor
16100 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
16101 this option is used.
16105 Specify that the target processor is the V850E@. The preprocessor
16106 constant @samp{__v850e__} will be defined if this option is used.
16108 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
16109 are defined then a default target processor will be chosen and the
16110 relevant @samp{__v850*__} preprocessor constant will be defined.
16112 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
16113 defined, regardless of which processor variant is the target.
16115 @item -mdisable-callt
16116 @opindex mdisable-callt
16117 This option will suppress generation of the CALLT instruction for the
16118 v850e and v850e1 flavors of the v850 architecture. The default is
16119 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
16124 @subsection VAX Options
16125 @cindex VAX options
16127 These @samp{-m} options are defined for the VAX:
16132 Do not output certain jump instructions (@code{aobleq} and so on)
16133 that the Unix assembler for the VAX cannot handle across long
16138 Do output those jump instructions, on the assumption that you
16139 will assemble with the GNU assembler.
16143 Output code for g-format floating point numbers instead of d-format.
16146 @node VxWorks Options
16147 @subsection VxWorks Options
16148 @cindex VxWorks Options
16150 The options in this section are defined for all VxWorks targets.
16151 Options specific to the target hardware are listed with the other
16152 options for that target.
16157 GCC can generate code for both VxWorks kernels and real time processes
16158 (RTPs). This option switches from the former to the latter. It also
16159 defines the preprocessor macro @code{__RTP__}.
16162 @opindex non-static
16163 Link an RTP executable against shared libraries rather than static
16164 libraries. The options @option{-static} and @option{-shared} can
16165 also be used for RTPs (@pxref{Link Options}); @option{-static}
16172 These options are passed down to the linker. They are defined for
16173 compatibility with Diab.
16176 @opindex Xbind-lazy
16177 Enable lazy binding of function calls. This option is equivalent to
16178 @option{-Wl,-z,now} and is defined for compatibility with Diab.
16182 Disable lazy binding of function calls. This option is the default and
16183 is defined for compatibility with Diab.
16186 @node x86-64 Options
16187 @subsection x86-64 Options
16188 @cindex x86-64 options
16190 These are listed under @xref{i386 and x86-64 Options}.
16192 @node i386 and x86-64 Windows Options
16193 @subsection i386 and x86-64 Windows Options
16194 @cindex i386 and x86-64 Windows Options
16196 These additional options are available for Windows targets:
16201 This option is available for Cygwin and MinGW targets. It
16202 specifies that a console application is to be generated, by
16203 instructing the linker to set the PE header subsystem type
16204 required for console applications.
16205 This is the default behavior for Cygwin and MinGW targets.
16209 This option is available for Cygwin targets. It specifies that
16210 the Cygwin internal interface is to be used for predefined
16211 preprocessor macros, C runtime libraries and related linker
16212 paths and options. For Cygwin targets this is the default behavior.
16213 This option is deprecated and will be removed in a future release.
16216 @opindex mno-cygwin
16217 This option is available for Cygwin targets. It specifies that
16218 the MinGW internal interface is to be used instead of Cygwin's, by
16219 setting MinGW-related predefined macros and linker paths and default
16221 This option is deprecated and will be removed in a future release.
16225 This option is available for Cygwin and MinGW targets. It
16226 specifies that a DLL - a dynamic link library - is to be
16227 generated, enabling the selection of the required runtime
16228 startup object and entry point.
16230 @item -mnop-fun-dllimport
16231 @opindex mnop-fun-dllimport
16232 This option is available for Cygwin and MinGW targets. It
16233 specifies that the dllimport attribute should be ignored.
16237 This option is available for MinGW targets. It specifies
16238 that MinGW-specific thread support is to be used.
16242 This option is available for mingw-w64 targets. It specifies
16243 that the UNICODE macro is getting pre-defined and that the
16244 unicode capable runtime startup code is choosen.
16248 This option is available for Cygwin and MinGW targets. It
16249 specifies that the typical Windows pre-defined macros are to
16250 be set in the pre-processor, but does not influence the choice
16251 of runtime library/startup code.
16255 This option is available for Cygwin and MinGW targets. It
16256 specifies that a GUI application is to be generated by
16257 instructing the linker to set the PE header subsystem type
16260 @item -mpe-aligned-commons
16261 @opindex mpe-aligned-commons
16262 This option is available for Cygwin and MinGW targets. It
16263 specifies that the GNU extension to the PE file format that
16264 permits the correct alignment of COMMON variables should be
16265 used when generating code. It will be enabled by default if
16266 GCC detects that the target assembler found during configuration
16267 supports the feature.
16270 See also under @ref{i386 and x86-64 Options} for standard options.
16272 @node Xstormy16 Options
16273 @subsection Xstormy16 Options
16274 @cindex Xstormy16 Options
16276 These options are defined for Xstormy16:
16281 Choose startup files and linker script suitable for the simulator.
16284 @node Xtensa Options
16285 @subsection Xtensa Options
16286 @cindex Xtensa Options
16288 These options are supported for Xtensa targets:
16292 @itemx -mno-const16
16294 @opindex mno-const16
16295 Enable or disable use of @code{CONST16} instructions for loading
16296 constant values. The @code{CONST16} instruction is currently not a
16297 standard option from Tensilica. When enabled, @code{CONST16}
16298 instructions are always used in place of the standard @code{L32R}
16299 instructions. The use of @code{CONST16} is enabled by default only if
16300 the @code{L32R} instruction is not available.
16303 @itemx -mno-fused-madd
16304 @opindex mfused-madd
16305 @opindex mno-fused-madd
16306 Enable or disable use of fused multiply/add and multiply/subtract
16307 instructions in the floating-point option. This has no effect if the
16308 floating-point option is not also enabled. Disabling fused multiply/add
16309 and multiply/subtract instructions forces the compiler to use separate
16310 instructions for the multiply and add/subtract operations. This may be
16311 desirable in some cases where strict IEEE 754-compliant results are
16312 required: the fused multiply add/subtract instructions do not round the
16313 intermediate result, thereby producing results with @emph{more} bits of
16314 precision than specified by the IEEE standard. Disabling fused multiply
16315 add/subtract instructions also ensures that the program output is not
16316 sensitive to the compiler's ability to combine multiply and add/subtract
16319 @item -mserialize-volatile
16320 @itemx -mno-serialize-volatile
16321 @opindex mserialize-volatile
16322 @opindex mno-serialize-volatile
16323 When this option is enabled, GCC inserts @code{MEMW} instructions before
16324 @code{volatile} memory references to guarantee sequential consistency.
16325 The default is @option{-mserialize-volatile}. Use
16326 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
16328 @item -mtext-section-literals
16329 @itemx -mno-text-section-literals
16330 @opindex mtext-section-literals
16331 @opindex mno-text-section-literals
16332 Control the treatment of literal pools. The default is
16333 @option{-mno-text-section-literals}, which places literals in a separate
16334 section in the output file. This allows the literal pool to be placed
16335 in a data RAM/ROM, and it also allows the linker to combine literal
16336 pools from separate object files to remove redundant literals and
16337 improve code size. With @option{-mtext-section-literals}, the literals
16338 are interspersed in the text section in order to keep them as close as
16339 possible to their references. This may be necessary for large assembly
16342 @item -mtarget-align
16343 @itemx -mno-target-align
16344 @opindex mtarget-align
16345 @opindex mno-target-align
16346 When this option is enabled, GCC instructs the assembler to
16347 automatically align instructions to reduce branch penalties at the
16348 expense of some code density. The assembler attempts to widen density
16349 instructions to align branch targets and the instructions following call
16350 instructions. If there are not enough preceding safe density
16351 instructions to align a target, no widening will be performed. The
16352 default is @option{-mtarget-align}. These options do not affect the
16353 treatment of auto-aligned instructions like @code{LOOP}, which the
16354 assembler will always align, either by widening density instructions or
16355 by inserting no-op instructions.
16358 @itemx -mno-longcalls
16359 @opindex mlongcalls
16360 @opindex mno-longcalls
16361 When this option is enabled, GCC instructs the assembler to translate
16362 direct calls to indirect calls unless it can determine that the target
16363 of a direct call is in the range allowed by the call instruction. This
16364 translation typically occurs for calls to functions in other source
16365 files. Specifically, the assembler translates a direct @code{CALL}
16366 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
16367 The default is @option{-mno-longcalls}. This option should be used in
16368 programs where the call target can potentially be out of range. This
16369 option is implemented in the assembler, not the compiler, so the
16370 assembly code generated by GCC will still show direct call
16371 instructions---look at the disassembled object code to see the actual
16372 instructions. Note that the assembler will use an indirect call for
16373 every cross-file call, not just those that really will be out of range.
16376 @node zSeries Options
16377 @subsection zSeries Options
16378 @cindex zSeries options
16380 These are listed under @xref{S/390 and zSeries Options}.
16382 @node Code Gen Options
16383 @section Options for Code Generation Conventions
16384 @cindex code generation conventions
16385 @cindex options, code generation
16386 @cindex run-time options
16388 These machine-independent options control the interface conventions
16389 used in code generation.
16391 Most of them have both positive and negative forms; the negative form
16392 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
16393 one of the forms is listed---the one which is not the default. You
16394 can figure out the other form by either removing @samp{no-} or adding
16398 @item -fbounds-check
16399 @opindex fbounds-check
16400 For front-ends that support it, generate additional code to check that
16401 indices used to access arrays are within the declared range. This is
16402 currently only supported by the Java and Fortran front-ends, where
16403 this option defaults to true and false respectively.
16407 This option generates traps for signed overflow on addition, subtraction,
16408 multiplication operations.
16412 This option instructs the compiler to assume that signed arithmetic
16413 overflow of addition, subtraction and multiplication wraps around
16414 using twos-complement representation. This flag enables some optimizations
16415 and disables others. This option is enabled by default for the Java
16416 front-end, as required by the Java language specification.
16419 @opindex fexceptions
16420 Enable exception handling. Generates extra code needed to propagate
16421 exceptions. For some targets, this implies GCC will generate frame
16422 unwind information for all functions, which can produce significant data
16423 size overhead, although it does not affect execution. If you do not
16424 specify this option, GCC will enable it by default for languages like
16425 C++ which normally require exception handling, and disable it for
16426 languages like C that do not normally require it. However, you may need
16427 to enable this option when compiling C code that needs to interoperate
16428 properly with exception handlers written in C++. You may also wish to
16429 disable this option if you are compiling older C++ programs that don't
16430 use exception handling.
16432 @item -fnon-call-exceptions
16433 @opindex fnon-call-exceptions
16434 Generate code that allows trapping instructions to throw exceptions.
16435 Note that this requires platform-specific runtime support that does
16436 not exist everywhere. Moreover, it only allows @emph{trapping}
16437 instructions to throw exceptions, i.e.@: memory references or floating
16438 point instructions. It does not allow exceptions to be thrown from
16439 arbitrary signal handlers such as @code{SIGALRM}.
16441 @item -funwind-tables
16442 @opindex funwind-tables
16443 Similar to @option{-fexceptions}, except that it will just generate any needed
16444 static data, but will not affect the generated code in any other way.
16445 You will normally not enable this option; instead, a language processor
16446 that needs this handling would enable it on your behalf.
16448 @item -fasynchronous-unwind-tables
16449 @opindex fasynchronous-unwind-tables
16450 Generate unwind table in dwarf2 format, if supported by target machine. The
16451 table is exact at each instruction boundary, so it can be used for stack
16452 unwinding from asynchronous events (such as debugger or garbage collector).
16454 @item -fpcc-struct-return
16455 @opindex fpcc-struct-return
16456 Return ``short'' @code{struct} and @code{union} values in memory like
16457 longer ones, rather than in registers. This convention is less
16458 efficient, but it has the advantage of allowing intercallability between
16459 GCC-compiled files and files compiled with other compilers, particularly
16460 the Portable C Compiler (pcc).
16462 The precise convention for returning structures in memory depends
16463 on the target configuration macros.
16465 Short structures and unions are those whose size and alignment match
16466 that of some integer type.
16468 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
16469 switch is not binary compatible with code compiled with the
16470 @option{-freg-struct-return} switch.
16471 Use it to conform to a non-default application binary interface.
16473 @item -freg-struct-return
16474 @opindex freg-struct-return
16475 Return @code{struct} and @code{union} values in registers when possible.
16476 This is more efficient for small structures than
16477 @option{-fpcc-struct-return}.
16479 If you specify neither @option{-fpcc-struct-return} nor
16480 @option{-freg-struct-return}, GCC defaults to whichever convention is
16481 standard for the target. If there is no standard convention, GCC
16482 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
16483 the principal compiler. In those cases, we can choose the standard, and
16484 we chose the more efficient register return alternative.
16486 @strong{Warning:} code compiled with the @option{-freg-struct-return}
16487 switch is not binary compatible with code compiled with the
16488 @option{-fpcc-struct-return} switch.
16489 Use it to conform to a non-default application binary interface.
16491 @item -fshort-enums
16492 @opindex fshort-enums
16493 Allocate to an @code{enum} type only as many bytes as it needs for the
16494 declared range of possible values. Specifically, the @code{enum} type
16495 will be equivalent to the smallest integer type which has enough room.
16497 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
16498 code that is not binary compatible with code generated without that switch.
16499 Use it to conform to a non-default application binary interface.
16501 @item -fshort-double
16502 @opindex fshort-double
16503 Use the same size for @code{double} as for @code{float}.
16505 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
16506 code that is not binary compatible with code generated without that switch.
16507 Use it to conform to a non-default application binary interface.
16509 @item -fshort-wchar
16510 @opindex fshort-wchar
16511 Override the underlying type for @samp{wchar_t} to be @samp{short
16512 unsigned int} instead of the default for the target. This option is
16513 useful for building programs to run under WINE@.
16515 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
16516 code that is not binary compatible with code generated without that switch.
16517 Use it to conform to a non-default application binary interface.
16520 @opindex fno-common
16521 In C code, controls the placement of uninitialized global variables.
16522 Unix C compilers have traditionally permitted multiple definitions of
16523 such variables in different compilation units by placing the variables
16525 This is the behavior specified by @option{-fcommon}, and is the default
16526 for GCC on most targets.
16527 On the other hand, this behavior is not required by ISO C, and on some
16528 targets may carry a speed or code size penalty on variable references.
16529 The @option{-fno-common} option specifies that the compiler should place
16530 uninitialized global variables in the data section of the object file,
16531 rather than generating them as common blocks.
16532 This has the effect that if the same variable is declared
16533 (without @code{extern}) in two different compilations,
16534 you will get a multiple-definition error when you link them.
16535 In this case, you must compile with @option{-fcommon} instead.
16536 Compiling with @option{-fno-common} is useful on targets for which
16537 it provides better performance, or if you wish to verify that the
16538 program will work on other systems which always treat uninitialized
16539 variable declarations this way.
16543 Ignore the @samp{#ident} directive.
16545 @item -finhibit-size-directive
16546 @opindex finhibit-size-directive
16547 Don't output a @code{.size} assembler directive, or anything else that
16548 would cause trouble if the function is split in the middle, and the
16549 two halves are placed at locations far apart in memory. This option is
16550 used when compiling @file{crtstuff.c}; you should not need to use it
16553 @item -fverbose-asm
16554 @opindex fverbose-asm
16555 Put extra commentary information in the generated assembly code to
16556 make it more readable. This option is generally only of use to those
16557 who actually need to read the generated assembly code (perhaps while
16558 debugging the compiler itself).
16560 @option{-fno-verbose-asm}, the default, causes the
16561 extra information to be omitted and is useful when comparing two assembler
16564 @item -frecord-gcc-switches
16565 @opindex frecord-gcc-switches
16566 This switch causes the command line that was used to invoke the
16567 compiler to be recorded into the object file that is being created.
16568 This switch is only implemented on some targets and the exact format
16569 of the recording is target and binary file format dependent, but it
16570 usually takes the form of a section containing ASCII text. This
16571 switch is related to the @option{-fverbose-asm} switch, but that
16572 switch only records information in the assembler output file as
16573 comments, so it never reaches the object file.
16577 @cindex global offset table
16579 Generate position-independent code (PIC) suitable for use in a shared
16580 library, if supported for the target machine. Such code accesses all
16581 constant addresses through a global offset table (GOT)@. The dynamic
16582 loader resolves the GOT entries when the program starts (the dynamic
16583 loader is not part of GCC; it is part of the operating system). If
16584 the GOT size for the linked executable exceeds a machine-specific
16585 maximum size, you get an error message from the linker indicating that
16586 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
16587 instead. (These maximums are 8k on the SPARC and 32k
16588 on the m68k and RS/6000. The 386 has no such limit.)
16590 Position-independent code requires special support, and therefore works
16591 only on certain machines. For the 386, GCC supports PIC for System V
16592 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
16593 position-independent.
16595 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
16600 If supported for the target machine, emit position-independent code,
16601 suitable for dynamic linking and avoiding any limit on the size of the
16602 global offset table. This option makes a difference on the m68k,
16603 PowerPC and SPARC@.
16605 Position-independent code requires special support, and therefore works
16606 only on certain machines.
16608 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
16615 These options are similar to @option{-fpic} and @option{-fPIC}, but
16616 generated position independent code can be only linked into executables.
16617 Usually these options are used when @option{-pie} GCC option will be
16618 used during linking.
16620 @option{-fpie} and @option{-fPIE} both define the macros
16621 @code{__pie__} and @code{__PIE__}. The macros have the value 1
16622 for @option{-fpie} and 2 for @option{-fPIE}.
16624 @item -fno-jump-tables
16625 @opindex fno-jump-tables
16626 Do not use jump tables for switch statements even where it would be
16627 more efficient than other code generation strategies. This option is
16628 of use in conjunction with @option{-fpic} or @option{-fPIC} for
16629 building code which forms part of a dynamic linker and cannot
16630 reference the address of a jump table. On some targets, jump tables
16631 do not require a GOT and this option is not needed.
16633 @item -ffixed-@var{reg}
16635 Treat the register named @var{reg} as a fixed register; generated code
16636 should never refer to it (except perhaps as a stack pointer, frame
16637 pointer or in some other fixed role).
16639 @var{reg} must be the name of a register. The register names accepted
16640 are machine-specific and are defined in the @code{REGISTER_NAMES}
16641 macro in the machine description macro file.
16643 This flag does not have a negative form, because it specifies a
16646 @item -fcall-used-@var{reg}
16647 @opindex fcall-used
16648 Treat the register named @var{reg} as an allocable register that is
16649 clobbered by function calls. It may be allocated for temporaries or
16650 variables that do not live across a call. Functions compiled this way
16651 will not save and restore the register @var{reg}.
16653 It is an error to used this flag with the frame pointer or stack pointer.
16654 Use of this flag for other registers that have fixed pervasive roles in
16655 the machine's execution model will produce disastrous results.
16657 This flag does not have a negative form, because it specifies a
16660 @item -fcall-saved-@var{reg}
16661 @opindex fcall-saved
16662 Treat the register named @var{reg} as an allocable register saved by
16663 functions. It may be allocated even for temporaries or variables that
16664 live across a call. Functions compiled this way will save and restore
16665 the register @var{reg} if they use it.
16667 It is an error to used this flag with the frame pointer or stack pointer.
16668 Use of this flag for other registers that have fixed pervasive roles in
16669 the machine's execution model will produce disastrous results.
16671 A different sort of disaster will result from the use of this flag for
16672 a register in which function values may be returned.
16674 This flag does not have a negative form, because it specifies a
16677 @item -fpack-struct[=@var{n}]
16678 @opindex fpack-struct
16679 Without a value specified, pack all structure members together without
16680 holes. When a value is specified (which must be a small power of two), pack
16681 structure members according to this value, representing the maximum
16682 alignment (that is, objects with default alignment requirements larger than
16683 this will be output potentially unaligned at the next fitting location.
16685 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
16686 code that is not binary compatible with code generated without that switch.
16687 Additionally, it makes the code suboptimal.
16688 Use it to conform to a non-default application binary interface.
16690 @item -finstrument-functions
16691 @opindex finstrument-functions
16692 Generate instrumentation calls for entry and exit to functions. Just
16693 after function entry and just before function exit, the following
16694 profiling functions will be called with the address of the current
16695 function and its call site. (On some platforms,
16696 @code{__builtin_return_address} does not work beyond the current
16697 function, so the call site information may not be available to the
16698 profiling functions otherwise.)
16701 void __cyg_profile_func_enter (void *this_fn,
16703 void __cyg_profile_func_exit (void *this_fn,
16707 The first argument is the address of the start of the current function,
16708 which may be looked up exactly in the symbol table.
16710 This instrumentation is also done for functions expanded inline in other
16711 functions. The profiling calls will indicate where, conceptually, the
16712 inline function is entered and exited. This means that addressable
16713 versions of such functions must be available. If all your uses of a
16714 function are expanded inline, this may mean an additional expansion of
16715 code size. If you use @samp{extern inline} in your C code, an
16716 addressable version of such functions must be provided. (This is
16717 normally the case anyways, but if you get lucky and the optimizer always
16718 expands the functions inline, you might have gotten away without
16719 providing static copies.)
16721 A function may be given the attribute @code{no_instrument_function}, in
16722 which case this instrumentation will not be done. This can be used, for
16723 example, for the profiling functions listed above, high-priority
16724 interrupt routines, and any functions from which the profiling functions
16725 cannot safely be called (perhaps signal handlers, if the profiling
16726 routines generate output or allocate memory).
16728 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
16729 @opindex finstrument-functions-exclude-file-list
16731 Set the list of functions that are excluded from instrumentation (see
16732 the description of @code{-finstrument-functions}). If the file that
16733 contains a function definition matches with one of @var{file}, then
16734 that function is not instrumented. The match is done on substrings:
16735 if the @var{file} parameter is a substring of the file name, it is
16736 considered to be a match.
16739 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
16740 will exclude any inline function defined in files whose pathnames
16741 contain @code{/bits/stl} or @code{include/sys}.
16743 If, for some reason, you want to include letter @code{','} in one of
16744 @var{sym}, write @code{'\,'}. For example,
16745 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
16746 (note the single quote surrounding the option).
16748 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
16749 @opindex finstrument-functions-exclude-function-list
16751 This is similar to @code{-finstrument-functions-exclude-file-list},
16752 but this option sets the list of function names to be excluded from
16753 instrumentation. The function name to be matched is its user-visible
16754 name, such as @code{vector<int> blah(const vector<int> &)}, not the
16755 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
16756 match is done on substrings: if the @var{sym} parameter is a substring
16757 of the function name, it is considered to be a match. For C99 and C++
16758 extended identifiers, the function name must be given in UTF-8, not
16759 using universal character names.
16761 @item -fstack-check
16762 @opindex fstack-check
16763 Generate code to verify that you do not go beyond the boundary of the
16764 stack. You should specify this flag if you are running in an
16765 environment with multiple threads, but only rarely need to specify it in
16766 a single-threaded environment since stack overflow is automatically
16767 detected on nearly all systems if there is only one stack.
16769 Note that this switch does not actually cause checking to be done; the
16770 operating system or the language runtime must do that. The switch causes
16771 generation of code to ensure that they see the stack being extended.
16773 You can additionally specify a string parameter: @code{no} means no
16774 checking, @code{generic} means force the use of old-style checking,
16775 @code{specific} means use the best checking method and is equivalent
16776 to bare @option{-fstack-check}.
16778 Old-style checking is a generic mechanism that requires no specific
16779 target support in the compiler but comes with the following drawbacks:
16783 Modified allocation strategy for large objects: they will always be
16784 allocated dynamically if their size exceeds a fixed threshold.
16787 Fixed limit on the size of the static frame of functions: when it is
16788 topped by a particular function, stack checking is not reliable and
16789 a warning is issued by the compiler.
16792 Inefficiency: because of both the modified allocation strategy and the
16793 generic implementation, the performances of the code are hampered.
16796 Note that old-style stack checking is also the fallback method for
16797 @code{specific} if no target support has been added in the compiler.
16799 @item -fstack-limit-register=@var{reg}
16800 @itemx -fstack-limit-symbol=@var{sym}
16801 @itemx -fno-stack-limit
16802 @opindex fstack-limit-register
16803 @opindex fstack-limit-symbol
16804 @opindex fno-stack-limit
16805 Generate code to ensure that the stack does not grow beyond a certain value,
16806 either the value of a register or the address of a symbol. If the stack
16807 would grow beyond the value, a signal is raised. For most targets,
16808 the signal is raised before the stack overruns the boundary, so
16809 it is possible to catch the signal without taking special precautions.
16811 For instance, if the stack starts at absolute address @samp{0x80000000}
16812 and grows downwards, you can use the flags
16813 @option{-fstack-limit-symbol=__stack_limit} and
16814 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
16815 of 128KB@. Note that this may only work with the GNU linker.
16817 @cindex aliasing of parameters
16818 @cindex parameters, aliased
16819 @item -fargument-alias
16820 @itemx -fargument-noalias
16821 @itemx -fargument-noalias-global
16822 @itemx -fargument-noalias-anything
16823 @opindex fargument-alias
16824 @opindex fargument-noalias
16825 @opindex fargument-noalias-global
16826 @opindex fargument-noalias-anything
16827 Specify the possible relationships among parameters and between
16828 parameters and global data.
16830 @option{-fargument-alias} specifies that arguments (parameters) may
16831 alias each other and may alias global storage.@*
16832 @option{-fargument-noalias} specifies that arguments do not alias
16833 each other, but may alias global storage.@*
16834 @option{-fargument-noalias-global} specifies that arguments do not
16835 alias each other and do not alias global storage.
16836 @option{-fargument-noalias-anything} specifies that arguments do not
16837 alias any other storage.
16839 Each language will automatically use whatever option is required by
16840 the language standard. You should not need to use these options yourself.
16842 @item -fleading-underscore
16843 @opindex fleading-underscore
16844 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
16845 change the way C symbols are represented in the object file. One use
16846 is to help link with legacy assembly code.
16848 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
16849 generate code that is not binary compatible with code generated without that
16850 switch. Use it to conform to a non-default application binary interface.
16851 Not all targets provide complete support for this switch.
16853 @item -ftls-model=@var{model}
16854 @opindex ftls-model
16855 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
16856 The @var{model} argument should be one of @code{global-dynamic},
16857 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
16859 The default without @option{-fpic} is @code{initial-exec}; with
16860 @option{-fpic} the default is @code{global-dynamic}.
16862 @item -fvisibility=@var{default|internal|hidden|protected}
16863 @opindex fvisibility
16864 Set the default ELF image symbol visibility to the specified option---all
16865 symbols will be marked with this unless overridden within the code.
16866 Using this feature can very substantially improve linking and
16867 load times of shared object libraries, produce more optimized
16868 code, provide near-perfect API export and prevent symbol clashes.
16869 It is @strong{strongly} recommended that you use this in any shared objects
16872 Despite the nomenclature, @code{default} always means public ie;
16873 available to be linked against from outside the shared object.
16874 @code{protected} and @code{internal} are pretty useless in real-world
16875 usage so the only other commonly used option will be @code{hidden}.
16876 The default if @option{-fvisibility} isn't specified is
16877 @code{default}, i.e., make every
16878 symbol public---this causes the same behavior as previous versions of
16881 A good explanation of the benefits offered by ensuring ELF
16882 symbols have the correct visibility is given by ``How To Write
16883 Shared Libraries'' by Ulrich Drepper (which can be found at
16884 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
16885 solution made possible by this option to marking things hidden when
16886 the default is public is to make the default hidden and mark things
16887 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
16888 and @code{__attribute__ ((visibility("default")))} instead of
16889 @code{__declspec(dllexport)} you get almost identical semantics with
16890 identical syntax. This is a great boon to those working with
16891 cross-platform projects.
16893 For those adding visibility support to existing code, you may find
16894 @samp{#pragma GCC visibility} of use. This works by you enclosing
16895 the declarations you wish to set visibility for with (for example)
16896 @samp{#pragma GCC visibility push(hidden)} and
16897 @samp{#pragma GCC visibility pop}.
16898 Bear in mind that symbol visibility should be viewed @strong{as
16899 part of the API interface contract} and thus all new code should
16900 always specify visibility when it is not the default ie; declarations
16901 only for use within the local DSO should @strong{always} be marked explicitly
16902 as hidden as so to avoid PLT indirection overheads---making this
16903 abundantly clear also aids readability and self-documentation of the code.
16904 Note that due to ISO C++ specification requirements, operator new and
16905 operator delete must always be of default visibility.
16907 Be aware that headers from outside your project, in particular system
16908 headers and headers from any other library you use, may not be
16909 expecting to be compiled with visibility other than the default. You
16910 may need to explicitly say @samp{#pragma GCC visibility push(default)}
16911 before including any such headers.
16913 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
16914 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
16915 no modifications. However, this means that calls to @samp{extern}
16916 functions with no explicit visibility will use the PLT, so it is more
16917 effective to use @samp{__attribute ((visibility))} and/or
16918 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
16919 declarations should be treated as hidden.
16921 Note that @samp{-fvisibility} does affect C++ vague linkage
16922 entities. This means that, for instance, an exception class that will
16923 be thrown between DSOs must be explicitly marked with default
16924 visibility so that the @samp{type_info} nodes will be unified between
16927 An overview of these techniques, their benefits and how to use them
16928 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
16934 @node Environment Variables
16935 @section Environment Variables Affecting GCC
16936 @cindex environment variables
16938 @c man begin ENVIRONMENT
16939 This section describes several environment variables that affect how GCC
16940 operates. Some of them work by specifying directories or prefixes to use
16941 when searching for various kinds of files. Some are used to specify other
16942 aspects of the compilation environment.
16944 Note that you can also specify places to search using options such as
16945 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
16946 take precedence over places specified using environment variables, which
16947 in turn take precedence over those specified by the configuration of GCC@.
16948 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
16949 GNU Compiler Collection (GCC) Internals}.
16954 @c @itemx LC_COLLATE
16956 @c @itemx LC_MONETARY
16957 @c @itemx LC_NUMERIC
16962 @c @findex LC_COLLATE
16963 @findex LC_MESSAGES
16964 @c @findex LC_MONETARY
16965 @c @findex LC_NUMERIC
16969 These environment variables control the way that GCC uses
16970 localization information that allow GCC to work with different
16971 national conventions. GCC inspects the locale categories
16972 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
16973 so. These locale categories can be set to any value supported by your
16974 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
16975 Kingdom encoded in UTF-8.
16977 The @env{LC_CTYPE} environment variable specifies character
16978 classification. GCC uses it to determine the character boundaries in
16979 a string; this is needed for some multibyte encodings that contain quote
16980 and escape characters that would otherwise be interpreted as a string
16983 The @env{LC_MESSAGES} environment variable specifies the language to
16984 use in diagnostic messages.
16986 If the @env{LC_ALL} environment variable is set, it overrides the value
16987 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
16988 and @env{LC_MESSAGES} default to the value of the @env{LANG}
16989 environment variable. If none of these variables are set, GCC
16990 defaults to traditional C English behavior.
16994 If @env{TMPDIR} is set, it specifies the directory to use for temporary
16995 files. GCC uses temporary files to hold the output of one stage of
16996 compilation which is to be used as input to the next stage: for example,
16997 the output of the preprocessor, which is the input to the compiler
17000 @item GCC_EXEC_PREFIX
17001 @findex GCC_EXEC_PREFIX
17002 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
17003 names of the subprograms executed by the compiler. No slash is added
17004 when this prefix is combined with the name of a subprogram, but you can
17005 specify a prefix that ends with a slash if you wish.
17007 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
17008 an appropriate prefix to use based on the pathname it was invoked with.
17010 If GCC cannot find the subprogram using the specified prefix, it
17011 tries looking in the usual places for the subprogram.
17013 The default value of @env{GCC_EXEC_PREFIX} is
17014 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
17015 the installed compiler. In many cases @var{prefix} is the value
17016 of @code{prefix} when you ran the @file{configure} script.
17018 Other prefixes specified with @option{-B} take precedence over this prefix.
17020 This prefix is also used for finding files such as @file{crt0.o} that are
17023 In addition, the prefix is used in an unusual way in finding the
17024 directories to search for header files. For each of the standard
17025 directories whose name normally begins with @samp{/usr/local/lib/gcc}
17026 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
17027 replacing that beginning with the specified prefix to produce an
17028 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
17029 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
17030 These alternate directories are searched first; the standard directories
17031 come next. If a standard directory begins with the configured
17032 @var{prefix} then the value of @var{prefix} is replaced by
17033 @env{GCC_EXEC_PREFIX} when looking for header files.
17035 @item COMPILER_PATH
17036 @findex COMPILER_PATH
17037 The value of @env{COMPILER_PATH} is a colon-separated list of
17038 directories, much like @env{PATH}. GCC tries the directories thus
17039 specified when searching for subprograms, if it can't find the
17040 subprograms using @env{GCC_EXEC_PREFIX}.
17043 @findex LIBRARY_PATH
17044 The value of @env{LIBRARY_PATH} is a colon-separated list of
17045 directories, much like @env{PATH}. When configured as a native compiler,
17046 GCC tries the directories thus specified when searching for special
17047 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
17048 using GCC also uses these directories when searching for ordinary
17049 libraries for the @option{-l} option (but directories specified with
17050 @option{-L} come first).
17054 @cindex locale definition
17055 This variable is used to pass locale information to the compiler. One way in
17056 which this information is used is to determine the character set to be used
17057 when character literals, string literals and comments are parsed in C and C++.
17058 When the compiler is configured to allow multibyte characters,
17059 the following values for @env{LANG} are recognized:
17063 Recognize JIS characters.
17065 Recognize SJIS characters.
17067 Recognize EUCJP characters.
17070 If @env{LANG} is not defined, or if it has some other value, then the
17071 compiler will use mblen and mbtowc as defined by the default locale to
17072 recognize and translate multibyte characters.
17076 Some additional environments variables affect the behavior of the
17079 @include cppenv.texi
17083 @node Precompiled Headers
17084 @section Using Precompiled Headers
17085 @cindex precompiled headers
17086 @cindex speed of compilation
17088 Often large projects have many header files that are included in every
17089 source file. The time the compiler takes to process these header files
17090 over and over again can account for nearly all of the time required to
17091 build the project. To make builds faster, GCC allows users to
17092 `precompile' a header file; then, if builds can use the precompiled
17093 header file they will be much faster.
17095 To create a precompiled header file, simply compile it as you would any
17096 other file, if necessary using the @option{-x} option to make the driver
17097 treat it as a C or C++ header file. You will probably want to use a
17098 tool like @command{make} to keep the precompiled header up-to-date when
17099 the headers it contains change.
17101 A precompiled header file will be searched for when @code{#include} is
17102 seen in the compilation. As it searches for the included file
17103 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
17104 compiler looks for a precompiled header in each directory just before it
17105 looks for the include file in that directory. The name searched for is
17106 the name specified in the @code{#include} with @samp{.gch} appended. If
17107 the precompiled header file can't be used, it is ignored.
17109 For instance, if you have @code{#include "all.h"}, and you have
17110 @file{all.h.gch} in the same directory as @file{all.h}, then the
17111 precompiled header file will be used if possible, and the original
17112 header will be used otherwise.
17114 Alternatively, you might decide to put the precompiled header file in a
17115 directory and use @option{-I} to ensure that directory is searched
17116 before (or instead of) the directory containing the original header.
17117 Then, if you want to check that the precompiled header file is always
17118 used, you can put a file of the same name as the original header in this
17119 directory containing an @code{#error} command.
17121 This also works with @option{-include}. So yet another way to use
17122 precompiled headers, good for projects not designed with precompiled
17123 header files in mind, is to simply take most of the header files used by
17124 a project, include them from another header file, precompile that header
17125 file, and @option{-include} the precompiled header. If the header files
17126 have guards against multiple inclusion, they will be skipped because
17127 they've already been included (in the precompiled header).
17129 If you need to precompile the same header file for different
17130 languages, targets, or compiler options, you can instead make a
17131 @emph{directory} named like @file{all.h.gch}, and put each precompiled
17132 header in the directory, perhaps using @option{-o}. It doesn't matter
17133 what you call the files in the directory, every precompiled header in
17134 the directory will be considered. The first precompiled header
17135 encountered in the directory that is valid for this compilation will
17136 be used; they're searched in no particular order.
17138 There are many other possibilities, limited only by your imagination,
17139 good sense, and the constraints of your build system.
17141 A precompiled header file can be used only when these conditions apply:
17145 Only one precompiled header can be used in a particular compilation.
17148 A precompiled header can't be used once the first C token is seen. You
17149 can have preprocessor directives before a precompiled header; you can
17150 even include a precompiled header from inside another header, so long as
17151 there are no C tokens before the @code{#include}.
17154 The precompiled header file must be produced for the same language as
17155 the current compilation. You can't use a C precompiled header for a C++
17159 The precompiled header file must have been produced by the same compiler
17160 binary as the current compilation is using.
17163 Any macros defined before the precompiled header is included must
17164 either be defined in the same way as when the precompiled header was
17165 generated, or must not affect the precompiled header, which usually
17166 means that they don't appear in the precompiled header at all.
17168 The @option{-D} option is one way to define a macro before a
17169 precompiled header is included; using a @code{#define} can also do it.
17170 There are also some options that define macros implicitly, like
17171 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
17174 @item If debugging information is output when using the precompiled
17175 header, using @option{-g} or similar, the same kind of debugging information
17176 must have been output when building the precompiled header. However,
17177 a precompiled header built using @option{-g} can be used in a compilation
17178 when no debugging information is being output.
17180 @item The same @option{-m} options must generally be used when building
17181 and using the precompiled header. @xref{Submodel Options},
17182 for any cases where this rule is relaxed.
17184 @item Each of the following options must be the same when building and using
17185 the precompiled header:
17187 @gccoptlist{-fexceptions}
17190 Some other command-line options starting with @option{-f},
17191 @option{-p}, or @option{-O} must be defined in the same way as when
17192 the precompiled header was generated. At present, it's not clear
17193 which options are safe to change and which are not; the safest choice
17194 is to use exactly the same options when generating and using the
17195 precompiled header. The following are known to be safe:
17197 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
17198 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
17199 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
17204 For all of these except the last, the compiler will automatically
17205 ignore the precompiled header if the conditions aren't met. If you
17206 find an option combination that doesn't work and doesn't cause the
17207 precompiled header to be ignored, please consider filing a bug report,
17210 If you do use differing options when generating and using the
17211 precompiled header, the actual behavior will be a mixture of the
17212 behavior for the options. For instance, if you use @option{-g} to
17213 generate the precompiled header but not when using it, you may or may
17214 not get debugging information for routines in the precompiled header.